Merge commit 'v2.6.29-rc1' into perfcounters/core

Conflicts:
	include/linux/kernel_stat.h

526 files changed, 77358 insertions, 22493 deletions

diff --git a/fs/Kconfig b/fs/Kconfig
index 522469a7eca..51307b0fdf0 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -189,6 +189,8 @@ config OCFS2_FS
 	select CONFIGFS_FS
 	select JBD2
 	select CRC32
+	select QUOTA
+	select QUOTA_TREE
 	help
 	  OCFS2 is a general purpose extent based shared disk cluster file
 	  system with many similarities to ext3. It supports 64 bit inode
@@ -258,56 +260,37 @@ config OCFS2_DEBUG_FS
 	  this option for debugging only as it is likely to decrease
 	  performance of the filesystem.
 
-config OCFS2_COMPAT_JBD
-	bool "Use JBD for compatibility"
+config OCFS2_FS_POSIX_ACL
+	bool "OCFS2 POSIX Access Control Lists"
 	depends on OCFS2_FS
+	select FS_POSIX_ACL
 	default n
-	select JBD
 	help
-	  The ocfs2 filesystem now uses JBD2 for its journalling.  JBD2
-	  is backwards compatible with JBD.  It is safe to say N here.
-	  However, if you really want to use the original JBD, say Y here.
-
-endif # BLOCK
+	  Posix Access Control Lists (ACLs) support permissions for users and
+	  groups beyond the owner/group/world scheme.
 
-config DNOTIFY
-	bool "Dnotify support"
-	default y
+config BTRFS_FS
+	tristate "Btrfs filesystem (EXPERIMENTAL) Unstable disk format"
+	depends on EXPERIMENTAL
+	select LIBCRC32C
+	select ZLIB_INFLATE
+	select ZLIB_DEFLATE
 	help
-	  Dnotify is a directory-based per-fd file change notification system
-	  that uses signals to communicate events to user-space.  There exist
-	  superior alternatives, but some applications may still rely on
-	  dnotify.
-
-	  If unsure, say Y.
-
-config INOTIFY
-	bool "Inotify file change notification support"
-	default y
-	---help---
-	  Say Y here to enable inotify support.  Inotify is a file change
-	  notification system and a replacement for dnotify.  Inotify fixes
-	  numerous shortcomings in dnotify and introduces several new features
-	  including multiple file events, one-shot support, and unmount
-	  notification.
+	  Btrfs is a new filesystem with extents, writable snapshotting,
+	  support for multiple devices and many more features.
 
-	  For more information, see <file:Documentation/filesystems/inotify.txt>
+	  Btrfs is highly experimental, and THE DISK FORMAT IS NOT YET
+	  FINALIZED.  You should say N here unless you are interested in
+	  testing Btrfs with non-critical data.
 
-	  If unsure, say Y.
+	  To compile this file system support as a module, choose M here. The
+	  module will be called btrfs.
 
-config INOTIFY_USER
-	bool "Inotify support for userspace"
-	depends on INOTIFY
-	default y
-	---help---
-	  Say Y here to enable inotify support for userspace, including the
-	  associated system calls.  Inotify allows monitoring of both files and
-	  directories via a single open fd.  Events are read from the file
-	  descriptor, which is also select()- and poll()-able.
+	  If unsure, say N.
 
-	  For more information, see <file:Documentation/filesystems/inotify.txt>
+endif # BLOCK
 
-	  If unsure, say Y.
+source "fs/notify/Kconfig"
 
 config QUOTA
 	bool "Quota support"
@@ -340,6 +323,10 @@ config PRINT_QUOTA_WARNING
 	  Note that this behavior is currently deprecated and may go away in
 	  future. Please use notification via netlink socket instead.
 
+# Generic support for tree structured quota files. Seleted when needed.
+config QUOTA_TREE
+	 tristate
+
 config QFMT_V1
 	tristate "Old quota format support"
 	depends on QUOTA
@@ -351,6 +338,7 @@ config QFMT_V1
 config QFMT_V2
 	tristate "Quota format v2 support"
 	depends on QUOTA
+	select QUOTA_TREE
 	help
 	  This quota format allows using quotas with 32-bit UIDs/GIDs. If you
 	  need this functionality say Y here.
@@ -752,7 +740,20 @@ config CONFIGFS_FS
 
 endmenu
 
-menu "Miscellaneous filesystems"
+menuconfig MISC_FILESYSTEMS
+	bool "Miscellaneous filesystems"
+	default y
+	---help---
+	  Say Y here to get to see options for various miscellaneous
+	  filesystems, such as filesystems that came from other
+	  operating systems.
+
+	  This option alone does not add any kernel code.
+
+	  If you say N, all options in this submenu will be skipped and
+	  disabled; if unsure, say Y here.
+
+if MISC_FILESYSTEMS
 
 config ADFS_FS
 	tristate "ADFS file system support (EXPERIMENTAL)"
@@ -931,6 +932,58 @@ config CRAMFS
 
 	  If unsure, say N.
 
+config SQUASHFS
+	tristate "SquashFS 4.0 - Squashed file system support"
+	depends on BLOCK
+	select ZLIB_INFLATE
+	help
+	  Saying Y here includes support for SquashFS 4.0 (a Compressed
+	  Read-Only File System).  Squashfs is a highly compressed read-only
+	  filesystem for Linux.  It uses zlib compression to compress both
+	  files, inodes and directories.  Inodes in the system are very small
+	  and all blocks are packed to minimise data overhead. Block sizes
+	  greater than 4K are supported up to a maximum of 1 Mbytes (default
+	  block size 128K).  SquashFS 4.0 supports 64 bit filesystems and files
+	  (larger than 4GB), full uid/gid information, hard links and
+	  timestamps.  
+
+	  Squashfs is intended for general read-only filesystem use, for
+	  archival use (i.e. in cases where a .tar.gz file may be used), and in
+	  embedded systems where low overhead is needed.  Further information
+	  and tools are available from http://squashfs.sourceforge.net.
+
+	  If you want to compile this as a module ( = code which can be
+	  inserted in and removed from the running kernel whenever you want),
+	  say M here and read <file:Documentation/modules.txt>.  The module
+	  will be called squashfs.  Note that the root file system (the one
+	  containing the directory /) cannot be compiled as a module.
+
+	  If unsure, say N.
+
+config SQUASHFS_EMBEDDED
+
+	bool "Additional option for memory-constrained systems" 
+	depends on SQUASHFS
+	default n
+	help
+	  Saying Y here allows you to specify cache size.
+
+	  If unsure, say N.
+
+config SQUASHFS_FRAGMENT_CACHE_SIZE
+	int "Number of fragments cached" if SQUASHFS_EMBEDDED
+	depends on SQUASHFS
+	default "3"
+	help
+	  By default SquashFS caches the last 3 fragments read from
+	  the filesystem.  Increasing this amount may mean SquashFS
+	  has to re-read fragments less often from disk, at the expense
+	  of extra system memory.  Decreasing this amount will mean
+	  SquashFS uses less memory at the expense of extra reads from disk.
+
+	  Note there must be at least one cached fragment.  Anything
+	  much more than three will probably not make much difference.
+
 config VXFS_FS
 	tristate "FreeVxFS file system support (VERITAS VxFS(TM) compatible)"
 	depends on BLOCK
@@ -1122,7 +1175,7 @@ config UFS_DEBUG
 	  Y here.  This will result in _many_ additional debugging messages to be
 	  written to the system log.
 
-endmenu
+endif # MISC_FILESYSTEMS
 
 menuconfig NETWORK_FILESYSTEMS
 	bool "Network File Systems"
diff --git a/fs/Kconfig.binfmt b/fs/Kconfig.binfmt
index ce9fb3fbfae..bb4cc5b8abc 100644
--- a/fs/Kconfig.binfmt
+++ b/fs/Kconfig.binfmt
@@ -43,7 +43,7 @@ config BINFMT_ELF_FDPIC
 config CORE_DUMP_DEFAULT_ELF_HEADERS
 	bool "Write ELF core dumps with partial segments"
 	default n
-	depends on BINFMT_ELF
+	depends on BINFMT_ELF && ELF_CORE
 	help
 	  ELF core dump files describe each memory mapping of the crashed
 	  process, and can contain or omit the memory contents of each one.
diff --git a/fs/Makefile b/fs/Makefile
index d9f8afe6f0c..38bc735c67a 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -20,8 +20,7 @@ obj-y +=	no-block.o
 endif
 
 obj-$(CONFIG_BLK_DEV_INTEGRITY) += bio-integrity.o
-obj-$(CONFIG_INOTIFY)		+= inotify.o
-obj-$(CONFIG_INOTIFY_USER)	+= inotify_user.o
+obj-y				+= notify/
 obj-$(CONFIG_EPOLL)		+= eventpoll.o
 obj-$(CONFIG_ANON_INODES)	+= anon_inodes.o
 obj-$(CONFIG_SIGNALFD)		+= signalfd.o
@@ -55,10 +54,9 @@ obj-$(CONFIG_GENERIC_ACL)	+= generic_acl.o
 obj-$(CONFIG_QUOTA)		+= dquot.o
 obj-$(CONFIG_QFMT_V1)		+= quota_v1.o
 obj-$(CONFIG_QFMT_V2)		+= quota_v2.o
+obj-$(CONFIG_QUOTA_TREE)	+= quota_tree.o
 obj-$(CONFIG_QUOTACTL)		+= quota.o
 
-obj-$(CONFIG_DNOTIFY)		+= dnotify.o
-
 obj-$(CONFIG_PROC_FS)		+= proc/
 obj-y				+= partitions/
 obj-$(CONFIG_SYSFS)		+= sysfs/
@@ -76,6 +74,7 @@ obj-$(CONFIG_JBD)		+= jbd/
 obj-$(CONFIG_JBD2)		+= jbd2/
 obj-$(CONFIG_EXT2_FS)		+= ext2/
 obj-$(CONFIG_CRAMFS)		+= cramfs/
+obj-$(CONFIG_SQUASHFS)		+= squashfs/
 obj-y				+= ramfs/
 obj-$(CONFIG_HUGETLBFS)		+= hugetlbfs/
 obj-$(CONFIG_CODA_FS)		+= coda/
@@ -121,4 +120,5 @@ obj-$(CONFIG_HOSTFS)		+= hostfs/
 obj-$(CONFIG_HPPFS)		+= hppfs/
 obj-$(CONFIG_DEBUG_FS)		+= debugfs/
 obj-$(CONFIG_OCFS2_FS)		+= ocfs2/
+obj-$(CONFIG_BTRFS_FS)		+= btrfs/
 obj-$(CONFIG_GFS2_FS)           += gfs2/
diff --git a/fs/affs/file.c b/fs/affs/file.c
index 1377b1240b6..9246cb4aa01 100644
--- a/fs/affs/file.c
+++ b/fs/affs/file.c
@@ -628,7 +628,7 @@ static int affs_write_begin_ofs(struct file *file, struct address_space *mapping
 	}
 
 	index = pos >> PAGE_CACHE_SHIFT;
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
diff --git a/fs/affs/inode.c b/fs/affs/inode.c
index 415d9c67ac1..3c4ec7d864c 100644
--- a/fs/affs/inode.c
+++ b/fs/affs/inode.c
@@ -119,8 +119,7 @@ struct inode *affs_iget(struct super_block *sb, unsigned long ino)
 		goto bad_inode;
 #else
 		inode->i_mode |= S_IFDIR;
-		inode->i_op = NULL;
-		inode->i_fop = NULL;
+		/* ... and leave ->i_op and ->i_fop pointing to empty */
 		break;
 #endif
 	case ST_LINKFILE:
diff --git a/fs/afs/write.c b/fs/afs/write.c
index d6b85dab35f..3fb36d43362 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -144,7 +144,7 @@ int afs_write_begin(struct file *file, struct address_space *mapping,
 	candidate->state = AFS_WBACK_PENDING;
 	init_waitqueue_head(&candidate->waitq);
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
 		kfree(candidate);
 		return -ENOMEM;
diff --git a/fs/aio.c b/fs/aio.c
index f658441d566..d6f89d3c15e 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -191,6 +191,20 @@ static int aio_setup_ring(struct kioctx *ctx)
 	kunmap_atomic((void *)((unsigned long)__event & PAGE_MASK), km); \
 } while(0)
 
+static void ctx_rcu_free(struct rcu_head *head)
+{
+	struct kioctx *ctx = container_of(head, struct kioctx, rcu_head);
+	unsigned nr_events = ctx->max_reqs;
+
+	kmem_cache_free(kioctx_cachep, ctx);
+
+	if (nr_events) {
+		spin_lock(&aio_nr_lock);
+		BUG_ON(aio_nr - nr_events > aio_nr);
+		aio_nr -= nr_events;
+		spin_unlock(&aio_nr_lock);
+	}
+}
 
 /* __put_ioctx
  *	Called when the last user of an aio context has gone away,
@@ -198,8 +212,6 @@ static int aio_setup_ring(struct kioctx *ctx)
  */
 static void __put_ioctx(struct kioctx *ctx)
 {
-	unsigned nr_events = ctx->max_reqs;
-
 	BUG_ON(ctx->reqs_active);
 
 	cancel_delayed_work(&ctx->wq);
@@ -208,14 +220,7 @@ static void __put_ioctx(struct kioctx *ctx)
 	mmdrop(ctx->mm);
 	ctx->mm = NULL;
 	pr_debug("__put_ioctx: freeing %p\n", ctx);
-	kmem_cache_free(kioctx_cachep, ctx);
-
-	if (nr_events) {
-		spin_lock(&aio_nr_lock);
-		BUG_ON(aio_nr - nr_events > aio_nr);
-		aio_nr -= nr_events;
-		spin_unlock(&aio_nr_lock);
-	}
+	call_rcu(&ctx->rcu_head, ctx_rcu_free);
 }
 
 #define get_ioctx(kioctx) do {						\
@@ -235,6 +240,7 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 {
 	struct mm_struct *mm;
 	struct kioctx *ctx;
+	int did_sync = 0;
 
 	/* Prevent overflows */
 	if ((nr_events > (0x10000000U / sizeof(struct io_event))) ||
@@ -267,21 +273,30 @@ static struct kioctx *ioctx_alloc(unsigned nr_events)
 		goto out_freectx;
 
 	/* limit the number of system wide aios */
-	spin_lock(&aio_nr_lock);
-	if (aio_nr + ctx->max_reqs > aio_max_nr ||
-	    aio_nr + ctx->max_reqs < aio_nr)
-		ctx->max_reqs = 0;
-	else
-		aio_nr += ctx->max_reqs;
-	spin_unlock(&aio_nr_lock);
+	do {
+		spin_lock_bh(&aio_nr_lock);
+		if (aio_nr + nr_events > aio_max_nr ||
+		    aio_nr + nr_events < aio_nr)
+			ctx->max_reqs = 0;
+		else
+			aio_nr += ctx->max_reqs;
+		spin_unlock_bh(&aio_nr_lock);
+		if (ctx->max_reqs || did_sync)
+			break;
+
+		/* wait for rcu callbacks to have completed before giving up */
+		synchronize_rcu();
+		did_sync = 1;
+		ctx->max_reqs = nr_events;
+	} while (1);
+
 	if (ctx->max_reqs == 0)
 		goto out_cleanup;
 
 	/* now link into global list. */
-	write_lock(&mm->ioctx_list_lock);
-	ctx->next = mm->ioctx_list;
-	mm->ioctx_list = ctx;
-	write_unlock(&mm->ioctx_list_lock);
+	spin_lock(&mm->ioctx_lock);
+	hlist_add_head_rcu(&ctx->list, &mm->ioctx_list);
+	spin_unlock(&mm->ioctx_lock);
 
 	dprintk("aio: allocated ioctx %p[%ld]: mm=%p mask=0x%x\n",
 		ctx, ctx->user_id, current->mm, ctx->ring_info.nr);
@@ -375,11 +390,12 @@ ssize_t wait_on_sync_kiocb(struct kiocb *iocb)
  */
 void exit_aio(struct mm_struct *mm)
 {
-	struct kioctx *ctx = mm->ioctx_list;
-	mm->ioctx_list = NULL;
-	while (ctx) {
-		struct kioctx *next = ctx->next;
-		ctx->next = NULL;
+	struct kioctx *ctx;
+
+	while (!hlist_empty(&mm->ioctx_list)) {
+		ctx = hlist_entry(mm->ioctx_list.first, struct kioctx, list);
+		hlist_del_rcu(&ctx->list);
+
 		aio_cancel_all(ctx);
 
 		wait_for_all_aios(ctx);
@@ -394,7 +410,6 @@ void exit_aio(struct mm_struct *mm)
 				atomic_read(&ctx->users), ctx->dead,
 				ctx->reqs_active);
 		put_ioctx(ctx);
-		ctx = next;
 	}
 }
 
@@ -555,19 +570,21 @@ int aio_put_req(struct kiocb *req)
 
 static struct kioctx *lookup_ioctx(unsigned long ctx_id)
 {
-	struct kioctx *ioctx;
-	struct mm_struct *mm;
+	struct mm_struct *mm = current->mm;
+	struct kioctx *ctx = NULL;
+	struct hlist_node *n;
 
-	mm = current->mm;
-	read_lock(&mm->ioctx_list_lock);
-	for (ioctx = mm->ioctx_list; ioctx; ioctx = ioctx->next)
-		if (likely(ioctx->user_id == ctx_id && !ioctx->dead)) {
-			get_ioctx(ioctx);
+	rcu_read_lock();
+
+	hlist_for_each_entry_rcu(ctx, n, &mm->ioctx_list, list) {
+		if (ctx->user_id == ctx_id && !ctx->dead) {
+			get_ioctx(ctx);
 			break;
 		}
-	read_unlock(&mm->ioctx_list_lock);
+	}
 
-	return ioctx;
+	rcu_read_unlock();
+	return ctx;
 }
 
 /*
@@ -1215,19 +1232,14 @@ out:
 static void io_destroy(struct kioctx *ioctx)
 {
 	struct mm_struct *mm = current->mm;
-	struct kioctx **tmp;
 	int was_dead;
 
 	/* delete the entry from the list is someone else hasn't already */
-	write_lock(&mm->ioctx_list_lock);
+	spin_lock(&mm->ioctx_lock);
 	was_dead = ioctx->dead;
 	ioctx->dead = 1;
-	for (tmp = &mm->ioctx_list; *tmp && *tmp != ioctx;
-	     tmp = &(*tmp)->next)
-		;
-	if (*tmp)
-		*tmp = ioctx->next;
-	write_unlock(&mm->ioctx_list_lock);
+	hlist_del_rcu(&ioctx->list);
+	spin_unlock(&mm->ioctx_lock);
 
 	dprintk("aio_release(%p)\n", ioctx);
 	if (likely(!was_dead))
diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c
index c16d9be1b01..3bbdb9d0237 100644
--- a/fs/anon_inodes.c
+++ b/fs/anon_inodes.c
@@ -79,9 +79,12 @@ int anon_inode_getfd(const char *name, const struct file_operations *fops,
 	if (IS_ERR(anon_inode_inode))
 		return -ENODEV;
 
+	if (fops->owner && !try_module_get(fops->owner))
+		return -ENOENT;
+
 	error = get_unused_fd_flags(flags);
 	if (error < 0)
-		return error;
+		goto err_module;
 	fd = error;
 
 	/*
@@ -128,6 +131,8 @@ err_dput:
 	dput(dentry);
 err_put_unused_fd:
 	put_unused_fd(fd);
+err_module:
+	module_put(fops->owner);
 	return error;
 }
 EXPORT_SYMBOL_GPL(anon_inode_getfd);
diff --git a/fs/autofs/inode.c b/fs/autofs/inode.c
index c773680d5c6..e1734f2d6e2 100644
--- a/fs/autofs/inode.c
+++ b/fs/autofs/inode.c
@@ -251,13 +251,11 @@ struct inode *autofs_iget(struct super_block *sb, unsigned long ino)
 	inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO;
 	inode->i_nlink = 2;
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
-	inode->i_blocks = 0;
 
 	if (ino == AUTOFS_ROOT_INO) {
 		inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
 		inode->i_op = &autofs_root_inode_operations;
 		inode->i_fop = &autofs_root_operations;
-		inode->i_uid = inode->i_gid = 0; /* Changed in read_super */
 		goto done;
 	} 
 	
diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h
index e0f16da00e5..a76803108d0 100644
--- a/fs/autofs4/autofs_i.h
+++ b/fs/autofs4/autofs_i.h
@@ -25,8 +25,6 @@
 #define AUTOFS_DEV_IOCTL_IOC_FIRST	(AUTOFS_DEV_IOCTL_VERSION)
 #define AUTOFS_DEV_IOCTL_IOC_COUNT	(AUTOFS_IOC_COUNT - 11)
 
-#define AUTOFS_TYPE_TRIGGER	(AUTOFS_TYPE_DIRECT|AUTOFS_TYPE_OFFSET)
-
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/time.h>
diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c
index 63b7c7afe8d..025e105bffe 100644
--- a/fs/autofs4/dev-ioctl.c
+++ b/fs/autofs4/dev-ioctl.c
@@ -124,7 +124,7 @@ static inline void free_dev_ioctl(struct autofs_dev_ioctl *param)
 
 /*
  * Check sanity of parameter control fields and if a path is present
- * check that it has a "/" and is terminated.
+ * check that it is terminated and contains at least one "/".
  */
 static int validate_dev_ioctl(int cmd, struct autofs_dev_ioctl *param)
 {
@@ -138,15 +138,16 @@ static int validate_dev_ioctl(int cmd, struct autofs_dev_ioctl *param)
 	}
 
 	if (param->size > sizeof(*param)) {
-		err = check_name(param->path);
+		err = invalid_str(param->path,
+				 (void *) ((size_t) param + param->size));
 		if (err) {
-			AUTOFS_WARN("invalid path supplied for cmd(0x%08x)",
-				    cmd);
+			AUTOFS_WARN(
+			  "path string terminator missing for cmd(0x%08x)",
+			  cmd);
 			goto out;
 		}
 
-		err = invalid_str(param->path,
-				 (void *) ((size_t) param + param->size));
+		err = check_name(param->path);
 		if (err) {
 			AUTOFS_WARN("invalid path supplied for cmd(0x%08x)",
 				    cmd);
@@ -180,7 +181,7 @@ static int autofs_dev_ioctl_protover(struct file *fp,
 				     struct autofs_sb_info *sbi,
 				     struct autofs_dev_ioctl *param)
 {
-	param->arg1 = sbi->version;
+	param->protover.version = sbi->version;
 	return 0;
 }
 
@@ -189,7 +190,7 @@ static int autofs_dev_ioctl_protosubver(struct file *fp,
 					struct autofs_sb_info *sbi,
 					struct autofs_dev_ioctl *param)
 {
-	param->arg1 = sbi->sub_version;
+	param->protosubver.sub_version = sbi->sub_version;
 	return 0;
 }
 
@@ -335,13 +336,13 @@ static int autofs_dev_ioctl_openmount(struct file *fp,
 	int err, fd;
 
 	/* param->path has already been checked */
-	if (!param->arg1)
+	if (!param->openmount.devid)
 		return -EINVAL;
 
 	param->ioctlfd = -1;
 
 	path = param->path;
-	devid = param->arg1;
+	devid = param->openmount.devid;
 
 	err = 0;
 	fd = autofs_dev_ioctl_open_mountpoint(path, devid);
@@ -373,7 +374,7 @@ static int autofs_dev_ioctl_ready(struct file *fp,
 {
 	autofs_wqt_t token;
 
-	token = (autofs_wqt_t) param->arg1;
+	token = (autofs_wqt_t) param->ready.token;
 	return autofs4_wait_release(sbi, token, 0);
 }
 
@@ -388,8 +389,8 @@ static int autofs_dev_ioctl_fail(struct file *fp,
 	autofs_wqt_t token;
 	int status;
 
-	token = (autofs_wqt_t) param->arg1;
-	status = param->arg2 ? param->arg2 : -ENOENT;
+	token = (autofs_wqt_t) param->fail.token;
+	status = param->fail.status ? param->fail.status : -ENOENT;
 	return autofs4_wait_release(sbi, token, status);
 }
 
@@ -412,10 +413,10 @@ static int autofs_dev_ioctl_setpipefd(struct file *fp,
 	int pipefd;
 	int err = 0;
 
-	if (param->arg1 == -1)
+	if (param->setpipefd.pipefd == -1)
 		return -EINVAL;
 
-	pipefd = param->arg1;
+	pipefd = param->setpipefd.pipefd;
 
 	mutex_lock(&sbi->wq_mutex);
 	if (!sbi->catatonic) {
@@ -457,8 +458,8 @@ static int autofs_dev_ioctl_timeout(struct file *fp,
 {
 	unsigned long timeout;
 
-	timeout = param->arg1;
-	param->arg1 = sbi->exp_timeout / HZ;
+	timeout = param->timeout.timeout;
+	param->timeout.timeout = sbi->exp_timeout / HZ;
 	sbi->exp_timeout = timeout * HZ;
 	return 0;
 }
@@ -489,7 +490,7 @@ static int autofs_dev_ioctl_requester(struct file *fp,
 	path = param->path;
 	devid = sbi->sb->s_dev;
 
-	param->arg1 = param->arg2 = -1;
+	param->requester.uid = param->requester.gid = -1;
 
 	/* Get nameidata of the parent directory */
 	err = path_lookup(path, LOOKUP_PARENT, &nd);
@@ -505,8 +506,8 @@ static int autofs_dev_ioctl_requester(struct file *fp,
 		err = 0;
 		autofs4_expire_wait(nd.path.dentry);
 		spin_lock(&sbi->fs_lock);
-		param->arg1 = ino->uid;
-		param->arg2 = ino->gid;
+		param->requester.uid = ino->uid;
+		param->requester.gid = ino->gid;
 		spin_unlock(&sbi->fs_lock);
 	}
 
@@ -529,10 +530,10 @@ static int autofs_dev_ioctl_expire(struct file *fp,
 	int err = -EAGAIN;
 	int how;
 
-	how = param->arg1;
+	how = param->expire.how;
 	mnt = fp->f_path.mnt;
 
-	if (sbi->type & AUTOFS_TYPE_TRIGGER)
+	if (autofs_type_trigger(sbi->type))
 		dentry = autofs4_expire_direct(sbi->sb, mnt, sbi, how);
 	else
 		dentry = autofs4_expire_indirect(sbi->sb, mnt, sbi, how);
@@ -565,9 +566,9 @@ static int autofs_dev_ioctl_askumount(struct file *fp,
 				      struct autofs_sb_info *sbi,
 				      struct autofs_dev_ioctl *param)
 {
-	param->arg1 = 0;
+	param->askumount.may_umount = 0;
 	if (may_umount(fp->f_path.mnt))
-		param->arg1 = 1;
+		param->askumount.may_umount = 1;
 	return 0;
 }
 
@@ -600,6 +601,7 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp,
 	struct nameidata nd;
 	const char *path;
 	unsigned int type;
+	unsigned int devid, magic;
 	int err = -ENOENT;
 
 	if (param->size <= sizeof(*param)) {
@@ -608,13 +610,13 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp,
 	}
 
 	path = param->path;
-	type = param->arg1;
+	type = param->ismountpoint.in.type;
 
-	param->arg1 = 0;
-	param->arg2 = 0;
+	param->ismountpoint.out.devid = devid = 0;
+	param->ismountpoint.out.magic = magic = 0;
 
 	if (!fp || param->ioctlfd == -1) {
-		if (type == AUTOFS_TYPE_ANY) {
+		if (autofs_type_any(type)) {
 			struct super_block *sb;
 
 			err = path_lookup(path, LOOKUP_FOLLOW, &nd);
@@ -622,7 +624,7 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp,
 				goto out;
 
 			sb = nd.path.dentry->d_sb;
-			param->arg1 = new_encode_dev(sb->s_dev);
+			devid = new_encode_dev(sb->s_dev);
 		} else {
 			struct autofs_info *ino;
 
@@ -635,38 +637,41 @@ static int autofs_dev_ioctl_ismountpoint(struct file *fp,
 				goto out_release;
 
 			ino = autofs4_dentry_ino(nd.path.dentry);
-			param->arg1 = autofs4_get_dev(ino->sbi);
+			devid = autofs4_get_dev(ino->sbi);
 		}
 
 		err = 0;
 		if (nd.path.dentry->d_inode &&
 		    nd.path.mnt->mnt_root == nd.path.dentry) {
 			err = 1;
-			param->arg2 = nd.path.dentry->d_inode->i_sb->s_magic;
+			magic = nd.path.dentry->d_inode->i_sb->s_magic;
 		}
 	} else {
-		dev_t devid = new_encode_dev(sbi->sb->s_dev);
+		dev_t dev = autofs4_get_dev(sbi);
 
 		err = path_lookup(path, LOOKUP_PARENT, &nd);
 		if (err)
 			goto out;
 
-		err = autofs_dev_ioctl_find_super(&nd, devid);
+		err = autofs_dev_ioctl_find_super(&nd, dev);
 		if (err)
 			goto out_release;
 
-		param->arg1 = autofs4_get_dev(sbi);
+		devid = dev;
 
 		err = have_submounts(nd.path.dentry);
 
 		if (nd.path.mnt->mnt_mountpoint != nd.path.mnt->mnt_root) {
 			if (follow_down(&nd.path.mnt, &nd.path.dentry)) {
 				struct inode *inode = nd.path.dentry->d_inode;
-				param->arg2 = inode->i_sb->s_magic;
+				magic = inode->i_sb->s_magic;
 			}
 		}
 	}
 
+	param->ismountpoint.out.devid = devid;
+	param->ismountpoint.out.magic = magic;
+
 out_release:
 	path_put(&nd.path);
 out:
diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c
index 4b6fb3f628c..e3bd50776f9 100644
--- a/fs/autofs4/expire.c
+++ b/fs/autofs4/expire.c
@@ -63,7 +63,7 @@ static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry)
 		struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
 
 		/* This is an autofs submount, we can't expire it */
-		if (sbi->type == AUTOFS_TYPE_INDIRECT)
+		if (autofs_type_indirect(sbi->type))
 			goto done;
 
 		/*
@@ -490,7 +490,7 @@ int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt,
 	if (arg && get_user(do_now, arg))
 		return -EFAULT;
 
-	if (sbi->type & AUTOFS_TYPE_TRIGGER)
+	if (autofs_type_trigger(sbi->type))
 		dentry = autofs4_expire_direct(sb, mnt, sbi, do_now);
 	else
 		dentry = autofs4_expire_indirect(sb, mnt, sbi, do_now);
diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c
index 7b19802cfef..716e12b627b 100644
--- a/fs/autofs4/inode.c
+++ b/fs/autofs4/inode.c
@@ -197,9 +197,9 @@ static int autofs4_show_options(struct seq_file *m, struct vfsmount *mnt)
 	seq_printf(m, ",minproto=%d", sbi->min_proto);
 	seq_printf(m, ",maxproto=%d", sbi->max_proto);
 
-	if (sbi->type & AUTOFS_TYPE_OFFSET)
+	if (autofs_type_offset(sbi->type))
 		seq_printf(m, ",offset");
-	else if (sbi->type & AUTOFS_TYPE_DIRECT)
+	else if (autofs_type_direct(sbi->type))
 		seq_printf(m, ",direct");
 	else
 		seq_printf(m, ",indirect");
@@ -284,13 +284,13 @@ static int parse_options(char *options, int *pipefd, uid_t *uid, gid_t *gid,
 			*maxproto = option;
 			break;
 		case Opt_indirect:
-			*type = AUTOFS_TYPE_INDIRECT;
+			set_autofs_type_indirect(type);
 			break;
 		case Opt_direct:
-			*type = AUTOFS_TYPE_DIRECT;
+			set_autofs_type_direct(type);
 			break;
 		case Opt_offset:
-			*type = AUTOFS_TYPE_OFFSET;
+			set_autofs_type_offset(type);
 			break;
 		default:
 			return 1;
@@ -338,7 +338,7 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	sbi->sb = s;
 	sbi->version = 0;
 	sbi->sub_version = 0;
-	sbi->type = AUTOFS_TYPE_INDIRECT;
+	set_autofs_type_indirect(&sbi->type);
 	sbi->min_proto = 0;
 	sbi->max_proto = 0;
 	mutex_init(&sbi->wq_mutex);
@@ -380,7 +380,7 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	}
 
 	root_inode->i_fop = &autofs4_root_operations;
-	root_inode->i_op = sbi->type & AUTOFS_TYPE_TRIGGER ?
+	root_inode->i_op = autofs_type_trigger(sbi->type) ?
 			&autofs4_direct_root_inode_operations :
 			&autofs4_indirect_root_inode_operations;
 
@@ -455,11 +455,7 @@ struct inode *autofs4_get_inode(struct super_block *sb,
 	if (sb->s_root) {
 		inode->i_uid = sb->s_root->d_inode->i_uid;
 		inode->i_gid = sb->s_root->d_inode->i_gid;
-	} else {
-		inode->i_uid = 0;
-		inode->i_gid = 0;
 	}
-	inode->i_blocks = 0;
 	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
 	if (S_ISDIR(inf->mode)) {
diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c
index e02cc8ae5eb..eeb24684590 100644
--- a/fs/autofs4/waitq.c
+++ b/fs/autofs4/waitq.c
@@ -337,7 +337,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		 * is very similar for indirect mounts except only dentrys
 		 * in the root of the autofs file system may be negative.
 		 */
-		if (sbi->type & AUTOFS_TYPE_TRIGGER)
+		if (autofs_type_trigger(sbi->type))
 			return -ENOENT;
 		else if (!IS_ROOT(dentry->d_parent))
 			return -ENOENT;
@@ -348,7 +348,7 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		return -ENOMEM;
 
 	/* If this is a direct mount request create a dummy name */
-	if (IS_ROOT(dentry) && sbi->type & AUTOFS_TYPE_TRIGGER)
+	if (IS_ROOT(dentry) && autofs_type_trigger(sbi->type))
 		qstr.len = sprintf(name, "%p", dentry);
 	else {
 		qstr.len = autofs4_getpath(sbi, dentry, &name);
@@ -406,11 +406,11 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 				type = autofs_ptype_expire_multi;
 		} else {
 			if (notify == NFY_MOUNT)
-				type = (sbi->type & AUTOFS_TYPE_TRIGGER) ?
+				type = autofs_type_trigger(sbi->type) ?
 					autofs_ptype_missing_direct :
 					 autofs_ptype_missing_indirect;
 			else
-				type = (sbi->type & AUTOFS_TYPE_TRIGGER) ?
+				type = autofs_type_trigger(sbi->type) ?
 					autofs_ptype_expire_direct :
 					autofs_ptype_expire_indirect;
 		}
diff --git a/fs/bad_inode.c b/fs/bad_inode.c
index 5f1538c03b1..a05287a23f6 100644
--- a/fs/bad_inode.c
+++ b/fs/bad_inode.c
@@ -132,11 +132,6 @@ static int bad_file_check_flags(int flags)
 	return -EIO;
 }
 
-static int bad_file_dir_notify(struct file *file, unsigned long arg)
-{
-	return -EIO;
-}
-
 static int bad_file_flock(struct file *filp, int cmd, struct file_lock *fl)
 {
 	return -EIO;
@@ -179,7 +174,6 @@ static const struct file_operations bad_file_ops =
 	.sendpage	= bad_file_sendpage,
 	.get_unmapped_area = bad_file_get_unmapped_area,
 	.check_flags	= bad_file_check_flags,
-	.dir_notify	= bad_file_dir_notify,
 	.flock		= bad_file_flock,
 	.splice_write	= bad_file_splice_write,
 	.splice_read	= bad_file_splice_read,
diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c
index b6dfee37c7b..d06cb023ad0 100644
--- a/fs/befs/linuxvfs.c
+++ b/fs/befs/linuxvfs.c
@@ -378,7 +378,8 @@ static struct inode *befs_iget(struct super_block *sb, unsigned long ino)
 		inode->i_size = 0;
 		inode->i_blocks = befs_sb->block_size / VFS_BLOCK_SIZE;
 		strncpy(befs_ino->i_data.symlink, raw_inode->data.symlink,
-			BEFS_SYMLINK_LEN);
+			BEFS_SYMLINK_LEN - 1);
+		befs_ino->i_data.symlink[BEFS_SYMLINK_LEN - 1] = '\0';
 	} else {
 		int num_blks;
 
@@ -477,6 +478,8 @@ befs_follow_link(struct dentry *dentry, struct nameidata *nd)
 			kfree(link);
 			befs_error(sb, "Failed to read entire long symlink");
 			link = ERR_PTR(-EIO);
+		} else {
+			link[len - 1] = '\0';
 		}
 	} else {
 		link = befs_ino->i_data.symlink;
diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c
index 0ed57b5ee01..cc4062d12ca 100644
--- a/fs/bfs/inode.c
+++ b/fs/bfs/inode.c
@@ -213,6 +213,9 @@ static void bfs_put_super(struct super_block *s)
 {
 	struct bfs_sb_info *info = BFS_SB(s);
 
+	if (!info)
+		return;
+
 	brelse(info->si_sbh);
 	mutex_destroy(&info->bfs_lock);
 	kfree(info->si_imap);
@@ -327,6 +330,7 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent)
 	unsigned i, imap_len;
 	struct bfs_sb_info *info;
 	long ret = -EINVAL;
+	unsigned long i_sblock, i_eblock, i_eoff, s_size;
 
 	info = kzalloc(sizeof(*info), GFP_KERNEL);
 	if (!info)
@@ -350,6 +354,12 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent)
 
 	s->s_magic = BFS_MAGIC;
 	info->si_sbh = bh;
+
+	if (le32_to_cpu(bfs_sb->s_start) > le32_to_cpu(bfs_sb->s_end)) {
+		printf("Superblock is corrupted\n");
+		goto out;
+	}
+
 	info->si_lasti = (le32_to_cpu(bfs_sb->s_start) - BFS_BSIZE) /
 					sizeof(struct bfs_inode)
 					+ BFS_ROOT_INO - 1;
@@ -380,6 +390,18 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent)
 			- le32_to_cpu(bfs_sb->s_start)) >> BFS_BSIZE_BITS;
 	info->si_freei = 0;
 	info->si_lf_eblk = 0;
+
+	/* can we read the last block? */
+	bh = sb_bread(s, info->si_blocks - 1);
+	if (!bh) {
+		printf("Last block not available: %lu\n", info->si_blocks - 1);
+		iput(inode);
+		ret = -EIO;
+		kfree(info->si_imap);
+		goto out;
+	}
+	brelse(bh);
+
 	bh = NULL;
 	for (i = BFS_ROOT_INO; i <= info->si_lasti; i++) {
 		struct bfs_inode *di;
@@ -397,6 +419,29 @@ static int bfs_fill_super(struct super_block *s, void *data, int silent)
 
 		di = (struct bfs_inode *)bh->b_data + off;
 
+		/* test if filesystem is not corrupted */
+
+		i_eoff = le32_to_cpu(di->i_eoffset);
+		i_sblock = le32_to_cpu(di->i_sblock);
+		i_eblock = le32_to_cpu(di->i_eblock);
+		s_size = le32_to_cpu(bfs_sb->s_end);
+
+		if (i_sblock > info->si_blocks ||
+			i_eblock > info->si_blocks ||
+			i_sblock > i_eblock ||
+			i_eoff > s_size ||
+			i_sblock * BFS_BSIZE > i_eoff) {
+
+			printf("Inode 0x%08x corrupted\n", i);
+
+			brelse(bh);
+			s->s_root = NULL;
+			kfree(info->si_imap);
+			kfree(info);
+			s->s_fs_info = NULL;
+			return -EIO;
+		}
+
 		if (!di->i_ino) {
 			info->si_freei++;
 			continue;
diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c
index f1f3f4192a6..b639dcf7c77 100644
--- a/fs/binfmt_aout.c
+++ b/fs/binfmt_aout.c
@@ -95,92 +95,55 @@ static int aout_core_dump(long signr, struct pt_regs *regs, struct file *file, u
 	int has_dumped = 0;
 	unsigned long dump_start, dump_size;
 	struct user dump;
-#if defined(__alpha__)
+#ifdef __alpha__
 #       define START_DATA(u)	(u.start_data)
-#elif defined(__arm__)
+#else
 #	define START_DATA(u)	((u.u_tsize << PAGE_SHIFT) + u.start_code)
-#elif defined(__sparc__)
-#       define START_DATA(u)    (u.u_tsize)
-#elif defined(__i386__) || defined(__mc68000__) || defined(__arch_um__)
-#       define START_DATA(u)	(u.u_tsize << PAGE_SHIFT)
 #endif
-#ifdef __sparc__
-#       define START_STACK(u)   ((regs->u_regs[UREG_FP]) & ~(PAGE_SIZE - 1))
-#else
 #       define START_STACK(u)   (u.start_stack)
-#endif
 
 	fs = get_fs();
 	set_fs(KERNEL_DS);
 	has_dumped = 1;
 	current->flags |= PF_DUMPCORE;
        	strncpy(dump.u_comm, current->comm, sizeof(dump.u_comm));
-#ifndef __sparc__
 	dump.u_ar0 = offsetof(struct user, regs);
-#endif
 	dump.signal = signr;
 	aout_dump_thread(regs, &dump);
 
 /* If the size of the dump file exceeds the rlimit, then see what would happen
    if we wrote the stack, but not the data area.  */
-#ifdef __sparc__
-	if ((dump.u_dsize + dump.u_ssize) > limit)
-		dump.u_dsize = 0;
-#else
 	if ((dump.u_dsize + dump.u_ssize+1) * PAGE_SIZE > limit)
 		dump.u_dsize = 0;
-#endif
 
 /* Make sure we have enough room to write the stack and data areas. */
-#ifdef __sparc__
-	if (dump.u_ssize > limit)
-		dump.u_ssize = 0;
-#else
 	if ((dump.u_ssize + 1) * PAGE_SIZE > limit)
 		dump.u_ssize = 0;
-#endif
 
 /* make sure we actually have a data and stack area to dump */
 	set_fs(USER_DS);
-#ifdef __sparc__
-	if (!access_ok(VERIFY_READ, (void __user *)START_DATA(dump), dump.u_dsize))
-		dump.u_dsize = 0;
-	if (!access_ok(VERIFY_READ, (void __user *)START_STACK(dump), dump.u_ssize))
-		dump.u_ssize = 0;
-#else
 	if (!access_ok(VERIFY_READ, (void __user *)START_DATA(dump), dump.u_dsize << PAGE_SHIFT))
 		dump.u_dsize = 0;
 	if (!access_ok(VERIFY_READ, (void __user *)START_STACK(dump), dump.u_ssize << PAGE_SHIFT))
 		dump.u_ssize = 0;
-#endif
 
 	set_fs(KERNEL_DS);
 /* struct user */
 	DUMP_WRITE(&dump,sizeof(dump));
 /* Now dump all of the user data.  Include malloced stuff as well */
-#ifndef __sparc__
 	DUMP_SEEK(PAGE_SIZE);
-#endif
 /* now we start writing out the user space info */
 	set_fs(USER_DS);
 /* Dump the data area */
 	if (dump.u_dsize != 0) {
 		dump_start = START_DATA(dump);
-#ifdef __sparc__
-		dump_size = dump.u_dsize;
-#else
 		dump_size = dump.u_dsize << PAGE_SHIFT;
-#endif
 		DUMP_WRITE(dump_start,dump_size);
 	}
 /* Now prepare to dump the stack area */
 	if (dump.u_ssize != 0) {
 		dump_start = START_STACK(dump);
-#ifdef __sparc__
-		dump_size = dump.u_ssize;
-#else
 		dump_size = dump.u_ssize << PAGE_SHIFT;
-#endif
 		DUMP_WRITE(dump_start,dump_size);
 	}
 /* Finally dump the task struct.  Not be used by gdb, but could be useful */
@@ -205,29 +168,24 @@ static unsigned long __user *create_aout_tables(char __user *p, struct linux_bin
 	int envc = bprm->envc;
 
 	sp = (void __user *)((-(unsigned long)sizeof(char *)) & (unsigned long) p);
-#ifdef __sparc__
-	/* This imposes the proper stack alignment for a new process. */
-	sp = (void __user *) (((unsigned long) sp) & ~7);
-	if ((envc+argc+3)&1) --sp;
-#endif
 #ifdef __alpha__
 /* whee.. test-programs are so much fun. */
 	put_user(0, --sp);
 	put_user(0, --sp);
 	if (bprm->loader) {
 		put_user(0, --sp);
-		put_user(0x3eb, --sp);
+		put_user(1003, --sp);
 		put_user(bprm->loader, --sp);
-		put_user(0x3ea, --sp);
+		put_user(1002, --sp);
 	}
 	put_user(bprm->exec, --sp);
-	put_user(0x3e9, --sp);
+	put_user(1001, --sp);
 #endif
 	sp -= envc+1;
 	envp = (char __user * __user *) sp;
 	sp -= argc+1;
 	argv = (char __user * __user *) sp;
-#if defined(__i386__) || defined(__mc68000__) || defined(__arm__) || defined(__arch_um__)
+#ifndef __alpha__
 	put_user((unsigned long) envp,--sp);
 	put_user((unsigned long) argv,--sp);
 #endif
@@ -300,13 +258,8 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 		return retval;
 
 	/* OK, This is the point of no return */
-#if defined(__alpha__)
+#ifdef __alpha__
 	SET_AOUT_PERSONALITY(bprm, ex);
-#elif defined(__sparc__)
-	set_personality(PER_SUNOS);
-#if !defined(__sparc_v9__)
-	memcpy(&current->thread.core_exec, &ex, sizeof(struct exec));
-#endif
 #else
 	set_personality(PER_LINUX);
 #endif
@@ -322,24 +275,6 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 
 	install_exec_creds(bprm);
  	current->flags &= ~PF_FORKNOEXEC;
-#ifdef __sparc__
-	if (N_MAGIC(ex) == NMAGIC) {
-		loff_t pos = fd_offset;
-		/* Fuck me plenty... */
-		/* <AOL></AOL> */
-		down_write(&current->mm->mmap_sem);	
-		error = do_brk(N_TXTADDR(ex), ex.a_text);
-		up_write(&current->mm->mmap_sem);
-		bprm->file->f_op->read(bprm->file, (char *) N_TXTADDR(ex),
-			  ex.a_text, &pos);
-		down_write(&current->mm->mmap_sem);
-		error = do_brk(N_DATADDR(ex), ex.a_data);
-		up_write(&current->mm->mmap_sem);
-		bprm->file->f_op->read(bprm->file, (char *) N_DATADDR(ex),
-			  ex.a_data, &pos);
-		goto beyond_if;
-	}
-#endif
 
 	if (N_MAGIC(ex) == OMAGIC) {
 		unsigned long text_addr, map_size;
@@ -347,7 +282,7 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs)
 
 		text_addr = N_TXTADDR(ex);
 
-#if defined(__alpha__) || defined(__sparc__)
+#ifdef __alpha__
 		pos = fd_offset;
 		map_size = ex.a_text+ex.a_data + PAGE_SIZE - 1;
 #else
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index c41fa2af767..e3ff2b9e602 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -152,8 +152,10 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec,
 	elf_addr_t __user *sp;
 	elf_addr_t __user *u_platform;
 	elf_addr_t __user *u_base_platform;
+	elf_addr_t __user *u_rand_bytes;
 	const char *k_platform = ELF_PLATFORM;
 	const char *k_base_platform = ELF_BASE_PLATFORM;
+	unsigned char k_rand_bytes[16];
 	int items;
 	elf_addr_t *elf_info;
 	int ei_index = 0;
@@ -196,6 +198,15 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec,
 			return -EFAULT;
 	}
 
+	/*
+	 * Generate 16 random bytes for userspace PRNG seeding.
+	 */
+	get_random_bytes(k_rand_bytes, sizeof(k_rand_bytes));
+	u_rand_bytes = (elf_addr_t __user *)
+		       STACK_ALLOC(p, sizeof(k_rand_bytes));
+	if (__copy_to_user(u_rand_bytes, k_rand_bytes, sizeof(k_rand_bytes)))
+		return -EFAULT;
+
 	/* Create the ELF interpreter info */
 	elf_info = (elf_addr_t *)current->mm->saved_auxv;
 	/* update AT_VECTOR_SIZE_BASE if the number of NEW_AUX_ENT() changes */
@@ -228,6 +239,7 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec,
 	NEW_AUX_ENT(AT_GID, cred->gid);
 	NEW_AUX_ENT(AT_EGID, cred->egid);
  	NEW_AUX_ENT(AT_SECURE, security_bprm_secureexec(bprm));
+	NEW_AUX_ENT(AT_RANDOM, (elf_addr_t)(unsigned long)u_rand_bytes);
 	NEW_AUX_ENT(AT_EXECFN, bprm->exec);
 	if (k_platform) {
 		NEW_AUX_ENT(AT_PLATFORM,
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index aa5b43205e3..f3e72c5c19f 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -168,9 +168,6 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 	struct elf_fdpic_params exec_params, interp_params;
 	struct elf_phdr *phdr;
 	unsigned long stack_size, entryaddr;
-#ifndef CONFIG_MMU
-	unsigned long fullsize;
-#endif
 #ifdef ELF_FDPIC_PLAT_INIT
 	unsigned long dynaddr;
 #endif
@@ -390,11 +387,6 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 		goto error_kill;
 	}
 
-	/* expand the stack mapping to use up the entire allocation granule */
-	fullsize = kobjsize((char *) current->mm->start_brk);
-	if (!IS_ERR_VALUE(do_mremap(current->mm->start_brk, stack_size,
-				    fullsize, 0, 0)))
-		stack_size = fullsize;
 	up_write(&current->mm->mmap_sem);
 
 	current->mm->brk = current->mm->start_brk;
@@ -1567,11 +1559,9 @@ end_coredump:
 static int elf_fdpic_dump_segments(struct file *file, size_t *size,
 			   unsigned long *limit, unsigned long mm_flags)
 {
-	struct vm_list_struct *vml;
-
-	for (vml = current->mm->context.vmlist; vml; vml = vml->next) {
-	struct vm_area_struct *vma = vml->vma;
+	struct vm_area_struct *vma;
 
+	for (vma = current->mm->mmap; vma; vma = vma->vm_next) {
 		if (!maydump(vma, mm_flags))
 			continue;
 
@@ -1617,9 +1607,6 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs,
 	elf_fpxregset_t *xfpu = NULL;
 #endif
 	int thread_status_size = 0;
-#ifndef CONFIG_MMU
-	struct vm_list_struct *vml;
-#endif
 	elf_addr_t *auxv;
 	unsigned long mm_flags;
 
@@ -1685,13 +1672,7 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs,
 	fill_prstatus(prstatus, current, signr);
 	elf_core_copy_regs(&prstatus->pr_reg, regs);
 
-#ifdef CONFIG_MMU
 	segs = current->mm->map_count;
-#else
-	segs = 0;
-	for (vml = current->mm->context.vmlist; vml; vml = vml->next)
-	    segs++;
-#endif
 #ifdef ELF_CORE_EXTRA_PHDRS
 	segs += ELF_CORE_EXTRA_PHDRS;
 #endif
@@ -1766,20 +1747,10 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs,
 	mm_flags = current->mm->flags;
 
 	/* write program headers for segments dump */
-	for (
-#ifdef CONFIG_MMU
-		vma = current->mm->mmap; vma; vma = vma->vm_next
-#else
-			vml = current->mm->context.vmlist; vml; vml = vml->next
-#endif
-	     ) {
+	for (vma = current->mm->mmap; vma; vma = vma->vm_next) {
 		struct elf_phdr phdr;
 		size_t sz;
 
-#ifndef CONFIG_MMU
-		vma = vml->vma;
-#endif
-
 		sz = vma->vm_end - vma->vm_start;
 
 		phdr.p_type = PT_LOAD;
diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c
index 7bbd5c6b372..5cebf0b3779 100644
--- a/fs/binfmt_flat.c
+++ b/fs/binfmt_flat.c
@@ -417,8 +417,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 	unsigned long textpos = 0, datapos = 0, result;
 	unsigned long realdatastart = 0;
 	unsigned long text_len, data_len, bss_len, stack_len, flags;
-	unsigned long len, reallen, memp = 0;
-	unsigned long extra, rlim;
+	unsigned long len, memp = 0;
+	unsigned long memp_size, extra, rlim;
 	unsigned long *reloc = 0, *rp;
 	struct inode *inode;
 	int i, rev, relocs = 0;
@@ -543,17 +543,10 @@ static int load_flat_file(struct linux_binprm * bprm,
 		}
 
 		len = data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long);
+		len = PAGE_ALIGN(len);
 		down_write(&current->mm->mmap_sem);
 		realdatastart = do_mmap(0, 0, len,
 			PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE, 0);
-		/* Remap to use all availabe slack region space */
-		if (realdatastart && (realdatastart < (unsigned long)-4096)) {
-			reallen = kobjsize((void *)realdatastart);
-			if (reallen > len) {
-				realdatastart = do_mremap(realdatastart, len,
-					reallen, MREMAP_FIXED, realdatastart);
-			}
-		}
 		up_write(&current->mm->mmap_sem);
 
 		if (realdatastart == 0 || realdatastart >= (unsigned long)-4096) {
@@ -591,21 +584,14 @@ static int load_flat_file(struct linux_binprm * bprm,
 
 		reloc = (unsigned long *) (datapos+(ntohl(hdr->reloc_start)-text_len));
 		memp = realdatastart;
-
+		memp_size = len;
 	} else {
 
 		len = text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long);
+		len = PAGE_ALIGN(len);
 		down_write(&current->mm->mmap_sem);
 		textpos = do_mmap(0, 0, len,
 			PROT_READ | PROT_EXEC | PROT_WRITE, MAP_PRIVATE, 0);
-		/* Remap to use all availabe slack region space */
-		if (textpos && (textpos < (unsigned long) -4096)) {
-			reallen = kobjsize((void *)textpos);
-			if (reallen > len) {
-				textpos = do_mremap(textpos, len, reallen,
-					MREMAP_FIXED, textpos);
-			}
-		}
 		up_write(&current->mm->mmap_sem);
 
 		if (!textpos  || textpos >= (unsigned long) -4096) {
@@ -622,7 +608,7 @@ static int load_flat_file(struct linux_binprm * bprm,
 		reloc = (unsigned long *) (textpos + ntohl(hdr->reloc_start) +
 				MAX_SHARED_LIBS * sizeof(unsigned long));
 		memp = textpos;
-
+		memp_size = len;
 #ifdef CONFIG_BINFMT_ZFLAT
 		/*
 		 * load it all in and treat it like a RAM load from now on
@@ -680,10 +666,12 @@ static int load_flat_file(struct linux_binprm * bprm,
 		 * set up the brk stuff, uses any slack left in data/bss/stack
 		 * allocation.  We put the brk after the bss (between the bss
 		 * and stack) like other platforms.
+		 * Userspace code relies on the stack pointer starting out at
+		 * an address right at the end of a page.
 		 */
 		current->mm->start_brk = datapos + data_len + bss_len;
 		current->mm->brk = (current->mm->start_brk + 3) & ~3;
-		current->mm->context.end_brk = memp + kobjsize((void *) memp) - stack_len;
+		current->mm->context.end_brk = memp + memp_size - stack_len;
 	}
 
 	if (flags & FLAT_FLAG_KTRACE)
@@ -790,8 +778,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 
 	/* zero the BSS,  BRK and stack areas */
 	memset((void*)(datapos + data_len), 0, bss_len + 
-			(memp + kobjsize((void *) memp) - stack_len -	/* end brk */
-			libinfo->lib_list[id].start_brk) +		/* start brk */
+			(memp + memp_size - stack_len -		/* end brk */
+			libinfo->lib_list[id].start_brk) +	/* start brk */
 			stack_len);
 
 	return 0;
diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c
index f2744ab4e5b..c4e83537ead 100644
--- a/fs/binfmt_misc.c
+++ b/fs/binfmt_misc.c
@@ -496,9 +496,6 @@ static struct inode *bm_get_inode(struct super_block *sb, int mode)
 
 	if (inode) {
 		inode->i_mode = mode;
-		inode->i_uid = 0;
-		inode->i_gid = 0;
-		inode->i_blocks = 0;
 		inode->i_atime = inode->i_mtime = inode->i_ctime =
 			current_fs_time(inode->i_sb);
 	}
@@ -652,7 +649,7 @@ static const struct file_operations bm_register_operations = {
 static ssize_t
 bm_status_read(struct file *file, char __user *buf, size_t nbytes, loff_t *ppos)
 {
-	char *s = enabled ? "enabled" : "disabled";
+	char *s = enabled ? "enabled\n" : "disabled\n";
 
 	return simple_read_from_buffer(buf, nbytes, ppos, s, strlen(s));
 }
diff --git a/fs/bio-integrity.c b/fs/bio-integrity.c
index 19caf7c962a..77ebc3c263d 100644
--- a/fs/bio-integrity.c
+++ b/fs/bio-integrity.c
@@ -111,7 +111,7 @@ void bio_integrity_free(struct bio *bio, struct bio_set *bs)
 	    && bip->bip_buf != NULL)
 		kfree(bip->bip_buf);
 
-	mempool_free(bip->bip_vec, bs->bvec_pools[bip->bip_pool]);
+	bvec_free_bs(bs, bip->bip_vec, bip->bip_pool);
 	mempool_free(bip, bs->bio_integrity_pool);
 
 	bio->bi_integrity = NULL;
diff --git a/fs/bio.c b/fs/bio.c
index df99c882b80..062299acbcc 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -31,7 +31,11 @@
 
 DEFINE_TRACE(block_split);
 
-static struct kmem_cache *bio_slab __read_mostly;
+/*
+ * Test patch to inline a certain number of bi_io_vec's inside the bio
+ * itself, to shrink a bio data allocation from two mempool calls to one
+ */
+#define BIO_INLINE_VECS		4
 
 static mempool_t *bio_split_pool __read_mostly;
 
@@ -40,9 +44,8 @@ static mempool_t *bio_split_pool __read_mostly;
  * break badly! cannot be bigger than what you can fit into an
  * unsigned short
  */
-
 #define BV(x) { .nr_vecs = x, .name = "biovec-"__stringify(x) }
-static struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
+struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
 	BV(1), BV(4), BV(16), BV(64), BV(128), BV(BIO_MAX_PAGES),
 };
 #undef BV
@@ -53,12 +56,121 @@ static struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
  */
 struct bio_set *fs_bio_set;
 
+/*
+ * Our slab pool management
+ */
+struct bio_slab {
+	struct kmem_cache *slab;
+	unsigned int slab_ref;
+	unsigned int slab_size;
+	char name[8];
+};
+static DEFINE_MUTEX(bio_slab_lock);
+static struct bio_slab *bio_slabs;
+static unsigned int bio_slab_nr, bio_slab_max;
+
+static struct kmem_cache *bio_find_or_create_slab(unsigned int extra_size)
+{
+	unsigned int sz = sizeof(struct bio) + extra_size;
+	struct kmem_cache *slab = NULL;
+	struct bio_slab *bslab;
+	unsigned int i, entry = -1;
+
+	mutex_lock(&bio_slab_lock);
+
+	i = 0;
+	while (i < bio_slab_nr) {
+		struct bio_slab *bslab = &bio_slabs[i];
+
+		if (!bslab->slab && entry == -1)
+			entry = i;
+		else if (bslab->slab_size == sz) {
+			slab = bslab->slab;
+			bslab->slab_ref++;
+			break;
+		}
+		i++;
+	}
+
+	if (slab)
+		goto out_unlock;
+
+	if (bio_slab_nr == bio_slab_max && entry == -1) {
+		bio_slab_max <<= 1;
+		bio_slabs = krealloc(bio_slabs,
+				     bio_slab_max * sizeof(struct bio_slab),
+				     GFP_KERNEL);
+		if (!bio_slabs)
+			goto out_unlock;
+	}
+	if (entry == -1)
+		entry = bio_slab_nr++;
+
+	bslab = &bio_slabs[entry];
+
+	snprintf(bslab->name, sizeof(bslab->name), "bio-%d", entry);
+	slab = kmem_cache_create(bslab->name, sz, 0, SLAB_HWCACHE_ALIGN, NULL);
+	if (!slab)
+		goto out_unlock;
+
+	printk("bio: create slab <%s> at %d\n", bslab->name, entry);
+	bslab->slab = slab;
+	bslab->slab_ref = 1;
+	bslab->slab_size = sz;
+out_unlock:
+	mutex_unlock(&bio_slab_lock);
+	return slab;
+}
+
+static void bio_put_slab(struct bio_set *bs)
+{
+	struct bio_slab *bslab = NULL;
+	unsigned int i;
+
+	mutex_lock(&bio_slab_lock);
+
+	for (i = 0; i < bio_slab_nr; i++) {
+		if (bs->bio_slab == bio_slabs[i].slab) {
+			bslab = &bio_slabs[i];
+			break;
+		}
+	}
+
+	if (WARN(!bslab, KERN_ERR "bio: unable to find slab!\n"))
+		goto out;
+
+	WARN_ON(!bslab->slab_ref);
+
+	if (--bslab->slab_ref)
+		goto out;
+
+	kmem_cache_destroy(bslab->slab);
+	bslab->slab = NULL;
+
+out:
+	mutex_unlock(&bio_slab_lock);
+}
+
 unsigned int bvec_nr_vecs(unsigned short idx)
 {
 	return bvec_slabs[idx].nr_vecs;
 }
 
-struct bio_vec *bvec_alloc_bs(gfp_t gfp_mask, int nr, unsigned long *idx, struct bio_set *bs)
+void bvec_free_bs(struct bio_set *bs, struct bio_vec *bv, unsigned int idx)
+{
+	BIO_BUG_ON(idx >= BIOVEC_NR_POOLS);
+
+	if (idx == BIOVEC_MAX_IDX)
+		mempool_free(bv, bs->bvec_pool);
+	else {
+		struct biovec_slab *bvs = bvec_slabs + idx;
+
+		kmem_cache_free(bvs->slab, bv);
+	}
+}
+
+struct bio_vec *bvec_alloc_bs(gfp_t gfp_mask, int nr, unsigned long *idx,
+			      struct bio_set *bs)
 {
 	struct bio_vec *bvl;
 
@@ -67,60 +179,85 @@ struct bio_vec *bvec_alloc_bs(gfp_t gfp_mask, int nr, unsigned long *idx, struct
 	 * If not, this is a bio_kmalloc() allocation and just do a
 	 * kzalloc() for the exact number of vecs right away.
 	 */
-	if (bs) {
+	if (!bs)
+		bvl = kmalloc(nr * sizeof(struct bio_vec), gfp_mask);
+
+	/*
+	 * see comment near bvec_array define!
+	 */
+	switch (nr) {
+	case 1:
+		*idx = 0;
+		break;
+	case 2 ... 4:
+		*idx = 1;
+		break;
+	case 5 ... 16:
+		*idx = 2;
+		break;
+	case 17 ... 64:
+		*idx = 3;
+		break;
+	case 65 ... 128:
+		*idx = 4;
+		break;
+	case 129 ... BIO_MAX_PAGES:
+		*idx = 5;
+		break;
+	default:
+		return NULL;
+	}
+
+	/*
+	 * idx now points to the pool we want to allocate from. only the
+	 * 1-vec entry pool is mempool backed.
+	 */
+	if (*idx == BIOVEC_MAX_IDX) {
+fallback:
+		bvl = mempool_alloc(bs->bvec_pool, gfp_mask);
+	} else {
+		struct biovec_slab *bvs = bvec_slabs + *idx;
+		gfp_t __gfp_mask = gfp_mask & ~(__GFP_WAIT | __GFP_IO);
+
 		/*
-		 * see comment near bvec_array define!
+		 * Make this allocation restricted and don't dump info on
+		 * allocation failures, since we'll fallback to the mempool
+		 * in case of failure.
 		 */
-		switch (nr) {
-		case 1:
-			*idx = 0;
-			break;
-		case 2 ... 4:
-			*idx = 1;
-			break;
-		case 5 ... 16:
-			*idx = 2;
-			break;
-		case 17 ... 64:
-			*idx = 3;
-			break;
-		case 65 ... 128:
-			*idx = 4;
-			break;
-		case 129 ... BIO_MAX_PAGES:
-			*idx = 5;
-			break;
-		default:
-			return NULL;
-		}
+		__gfp_mask |= __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN;
 
 		/*
-		 * idx now points to the pool we want to allocate from
+		 * Try a slab allocation. If this fails and __GFP_WAIT
+		 * is set, retry with the 1-entry mempool
 		 */
-		bvl = mempool_alloc(bs->bvec_pools[*idx], gfp_mask);
-		if (bvl)
-			memset(bvl, 0,
-				bvec_nr_vecs(*idx) * sizeof(struct bio_vec));
-	} else
-		bvl = kzalloc(nr * sizeof(struct bio_vec), gfp_mask);
+		bvl = kmem_cache_alloc(bvs->slab, __gfp_mask);
+		if (unlikely(!bvl && (gfp_mask & __GFP_WAIT))) {
+			*idx = BIOVEC_MAX_IDX;
+			goto fallback;
+		}
+	}
 
 	return bvl;
 }
 
-void bio_free(struct bio *bio, struct bio_set *bio_set)
+void bio_free(struct bio *bio, struct bio_set *bs)
 {
-	if (bio->bi_io_vec) {
-		const int pool_idx = BIO_POOL_IDX(bio);
-
-		BIO_BUG_ON(pool_idx >= BIOVEC_NR_POOLS);
+	void *p;
 
-		mempool_free(bio->bi_io_vec, bio_set->bvec_pools[pool_idx]);
-	}
+	if (bio_has_allocated_vec(bio))
+		bvec_free_bs(bs, bio->bi_io_vec, BIO_POOL_IDX(bio));
 
 	if (bio_integrity(bio))
-		bio_integrity_free(bio, bio_set);
+		bio_integrity_free(bio, bs);
+
+	/*
+	 * If we have front padding, adjust the bio pointer before freeing
+	 */
+	p = bio;
+	if (bs->front_pad)
+		p -= bs->front_pad;
 
-	mempool_free(bio, bio_set->bio_pool);
+	mempool_free(p, bs->bio_pool);
 }
 
 /*
@@ -133,7 +270,8 @@ static void bio_fs_destructor(struct bio *bio)
 
 static void bio_kmalloc_destructor(struct bio *bio)
 {
-	kfree(bio->bi_io_vec);
+	if (bio_has_allocated_vec(bio))
+		kfree(bio->bi_io_vec);
 	kfree(bio);
 }
 
@@ -157,16 +295,20 @@ void bio_init(struct bio *bio)
  *   for a &struct bio to become free. If a %NULL @bs is passed in, we will
  *   fall back to just using @kmalloc to allocate the required memory.
  *
- *   allocate bio and iovecs from the memory pools specified by the
- *   bio_set structure, or @kmalloc if none given.
+ *   Note that the caller must set ->bi_destructor on succesful return
+ *   of a bio, to do the appropriate freeing of the bio once the reference
+ *   count drops to zero.
  **/
 struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
 {
-	struct bio *bio;
+	struct bio *bio = NULL;
+
+	if (bs) {
+		void *p = mempool_alloc(bs->bio_pool, gfp_mask);
 
-	if (bs)
-		bio = mempool_alloc(bs->bio_pool, gfp_mask);
-	else
+		if (p)
+			bio = p + bs->front_pad;
+	} else
 		bio = kmalloc(sizeof(*bio), gfp_mask);
 
 	if (likely(bio)) {
@@ -176,7 +318,15 @@ struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
 		if (likely(nr_iovecs)) {
 			unsigned long uninitialized_var(idx);
 
-			bvl = bvec_alloc_bs(gfp_mask, nr_iovecs, &idx, bs);
+			if (nr_iovecs <= BIO_INLINE_VECS) {
+				idx = 0;
+				bvl = bio->bi_inline_vecs;
+				nr_iovecs = BIO_INLINE_VECS;
+			} else {
+				bvl = bvec_alloc_bs(gfp_mask, nr_iovecs, &idx,
+							bs);
+				nr_iovecs = bvec_nr_vecs(idx);
+			}
 			if (unlikely(!bvl)) {
 				if (bs)
 					mempool_free(bio, bs->bio_pool);
@@ -186,7 +336,7 @@ struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
 				goto out;
 			}
 			bio->bi_flags |= idx << BIO_POOL_OFFSET;
-			bio->bi_max_vecs = bvec_nr_vecs(idx);
+			bio->bi_max_vecs = nr_iovecs;
 		}
 		bio->bi_io_vec = bvl;
 	}
@@ -638,6 +788,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	int i, ret;
 	int nr_pages = 0;
 	unsigned int len = 0;
+	unsigned int offset = map_data ? map_data->offset & ~PAGE_MASK : 0;
 
 	for (i = 0; i < iov_count; i++) {
 		unsigned long uaddr;
@@ -664,35 +815,42 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	bio->bi_rw |= (!write_to_vm << BIO_RW);
 
 	ret = 0;
-	i = 0;
+
+	if (map_data) {
+		nr_pages = 1 << map_data->page_order;
+		i = map_data->offset / PAGE_SIZE;
+	}
 	while (len) {
-		unsigned int bytes;
+		unsigned int bytes = PAGE_SIZE;
 
-		if (map_data)
-			bytes = 1U << (PAGE_SHIFT + map_data->page_order);
-		else
-			bytes = PAGE_SIZE;
+		bytes -= offset;
 
 		if (bytes > len)
 			bytes = len;
 
 		if (map_data) {
-			if (i == map_data->nr_entries) {
+			if (i == map_data->nr_entries * nr_pages) {
 				ret = -ENOMEM;
 				break;
 			}
-			page = map_data->pages[i++];
-		} else
+
+			page = map_data->pages[i / nr_pages];
+			page += (i % nr_pages);
+
+			i++;
+		} else {
 			page = alloc_page(q->bounce_gfp | gfp_mask);
-		if (!page) {
-			ret = -ENOMEM;
-			break;
+			if (!page) {
+				ret = -ENOMEM;
+				break;
+			}
 		}
 
-		if (bio_add_pc_page(q, bio, page, bytes, 0) < bytes)
+		if (bio_add_pc_page(q, bio, page, bytes, offset) < bytes)
 			break;
 
 		len -= bytes;
+		offset = 0;
 	}
 
 	if (ret)
@@ -701,7 +859,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	/*
 	 * success
 	 */
-	if (!write_to_vm) {
+	if (!write_to_vm && (!map_data || !map_data->null_mapped)) {
 		ret = __bio_copy_iov(bio, bio->bi_io_vec, iov, iov_count, 0, 0);
 		if (ret)
 			goto cleanup;
@@ -1346,30 +1504,18 @@ EXPORT_SYMBOL(bio_sector_offset);
  */
 static int biovec_create_pools(struct bio_set *bs, int pool_entries)
 {
-	int i;
+	struct biovec_slab *bp = bvec_slabs + BIOVEC_MAX_IDX;
 
-	for (i = 0; i < BIOVEC_NR_POOLS; i++) {
-		struct biovec_slab *bp = bvec_slabs + i;
-		mempool_t **bvp = bs->bvec_pools + i;
+	bs->bvec_pool = mempool_create_slab_pool(pool_entries, bp->slab);
+	if (!bs->bvec_pool)
+		return -ENOMEM;
 
-		*bvp = mempool_create_slab_pool(pool_entries, bp->slab);
-		if (!*bvp)
-			return -ENOMEM;
-	}
 	return 0;
 }
 
 static void biovec_free_pools(struct bio_set *bs)
 {
-	int i;
-
-	for (i = 0; i < BIOVEC_NR_POOLS; i++) {
-		mempool_t *bvp = bs->bvec_pools[i];
-
-		if (bvp)
-			mempool_destroy(bvp);
-	}
-
+	mempool_destroy(bs->bvec_pool);
 }
 
 void bioset_free(struct bio_set *bs)
@@ -1379,25 +1525,49 @@ void bioset_free(struct bio_set *bs)
 
 	bioset_integrity_free(bs);
 	biovec_free_pools(bs);
+	bio_put_slab(bs);
 
 	kfree(bs);
 }
 
-struct bio_set *bioset_create(int bio_pool_size, int bvec_pool_size)
+/**
+ * bioset_create  - Create a bio_set
+ * @pool_size:	Number of bio and bio_vecs to cache in the mempool
+ * @front_pad:	Number of bytes to allocate in front of the returned bio
+ *
+ * Description:
+ *    Set up a bio_set to be used with @bio_alloc_bioset. Allows the caller
+ *    to ask for a number of bytes to be allocated in front of the bio.
+ *    Front pad allocation is useful for embedding the bio inside
+ *    another structure, to avoid allocating extra data to go with the bio.
+ *    Note that the bio must be embedded at the END of that structure always,
+ *    or things will break badly.
+ */
+struct bio_set *bioset_create(unsigned int pool_size, unsigned int front_pad)
 {
-	struct bio_set *bs = kzalloc(sizeof(*bs), GFP_KERNEL);
+	unsigned int back_pad = BIO_INLINE_VECS * sizeof(struct bio_vec);
+	struct bio_set *bs;
 
+	bs = kzalloc(sizeof(*bs), GFP_KERNEL);
 	if (!bs)
 		return NULL;
 
-	bs->bio_pool = mempool_create_slab_pool(bio_pool_size, bio_slab);
+	bs->front_pad = front_pad;
+
+	bs->bio_slab = bio_find_or_create_slab(front_pad + back_pad);
+	if (!bs->bio_slab) {
+		kfree(bs);
+		return NULL;
+	}
+
+	bs->bio_pool = mempool_create_slab_pool(pool_size, bs->bio_slab);
 	if (!bs->bio_pool)
 		goto bad;
 
-	if (bioset_integrity_create(bs, bio_pool_size))
+	if (bioset_integrity_create(bs, pool_size))
 		goto bad;
 
-	if (!biovec_create_pools(bs, bvec_pool_size))
+	if (!biovec_create_pools(bs, pool_size))
 		return bs;
 
 bad:
@@ -1421,12 +1591,16 @@ static void __init biovec_init_slabs(void)
 
 static int __init init_bio(void)
 {
-	bio_slab = KMEM_CACHE(bio, SLAB_HWCACHE_ALIGN|SLAB_PANIC);
+	bio_slab_max = 2;
+	bio_slab_nr = 0;
+	bio_slabs = kzalloc(bio_slab_max * sizeof(struct bio_slab), GFP_KERNEL);
+	if (!bio_slabs)
+		panic("bio: can't allocate bios\n");
 
 	bio_integrity_init_slab();
 	biovec_init_slabs();
 
-	fs_bio_set = bioset_create(BIO_POOL_SIZE, 2);
+	fs_bio_set = bioset_create(BIO_POOL_SIZE, 0);
 	if (!fs_bio_set)
 		panic("bio: can't allocate bios\n");
 
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 99e0ae1a4c7..b3c1efff5e1 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -285,6 +285,8 @@ static void init_once(void *foo)
 	INIT_LIST_HEAD(&bdev->bd_holder_list);
 #endif
 	inode_init_once(&ei->vfs_inode);
+	/* Initialize mutex for freeze. */
+	mutex_init(&bdev->bd_fsfreeze_mutex);
 }
 
 static inline void __bd_forget(struct inode *inode)
@@ -326,12 +328,13 @@ static struct file_system_type bd_type = {
 	.kill_sb	= kill_anon_super,
 };
 
-static struct vfsmount *bd_mnt __read_mostly;
-struct super_block *blockdev_superblock;
+struct super_block *blockdev_superblock __read_mostly;
 
 void __init bdev_cache_init(void)
 {
 	int err;
+	struct vfsmount *bd_mnt;
+
 	bdev_cachep = kmem_cache_create("bdev_cache", sizeof(struct bdev_inode),
 			0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
 				SLAB_MEM_SPREAD|SLAB_PANIC),
@@ -373,7 +376,7 @@ struct block_device *bdget(dev_t dev)
 	struct block_device *bdev;
 	struct inode *inode;
 
-	inode = iget5_locked(bd_mnt->mnt_sb, hash(dev),
+	inode = iget5_locked(blockdev_superblock, hash(dev),
 			bdev_test, bdev_set, &dev);
 
 	if (!inode)
@@ -463,7 +466,7 @@ void bd_forget(struct inode *inode)
 
 	spin_lock(&bdev_lock);
 	if (inode->i_bdev) {
-		if (inode->i_sb != blockdev_superblock)
+		if (!sb_is_blkdev_sb(inode->i_sb))
 			bdev = inode->i_bdev;
 		__bd_forget(inode);
 	}
@@ -1004,6 +1007,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
 	}
 
 	lock_kernel();
+ restart:
 
 	ret = -ENXIO;
 	disk = get_gendisk(bdev->bd_dev, &partno);
@@ -1024,6 +1028,19 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
 
 			if (disk->fops->open) {
 				ret = disk->fops->open(bdev, mode);
+				if (ret == -ERESTARTSYS) {
+					/* Lost a race with 'disk' being
+					 * deleted, try again.
+					 * See md.c
+					 */
+					disk_put_part(bdev->bd_part);
+					bdev->bd_part = NULL;
+					module_put(disk->fops->owner);
+					put_disk(disk);
+					bdev->bd_disk = NULL;
+					mutex_unlock(&bdev->bd_mutex);
+					goto restart;
+				}
 				if (ret)
 					goto out_clear;
 			}
@@ -1219,6 +1236,20 @@ static long block_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 	return blkdev_ioctl(bdev, mode, cmd, arg);
 }
 
+/*
+ * Try to release a page associated with block device when the system
+ * is under memory pressure.
+ */
+static int blkdev_releasepage(struct page *page, gfp_t wait)
+{
+	struct super_block *super = BDEV_I(page->mapping->host)->bdev.bd_super;
+
+	if (super && super->s_op->bdev_try_to_free_page)
+		return super->s_op->bdev_try_to_free_page(super, page, wait);
+
+	return try_to_free_buffers(page);
+}
+
 static const struct address_space_operations def_blk_aops = {
 	.readpage	= blkdev_readpage,
 	.writepage	= blkdev_writepage,
@@ -1226,6 +1257,7 @@ static const struct address_space_operations def_blk_aops = {
 	.write_begin	= blkdev_write_begin,
 	.write_end	= blkdev_write_end,
 	.writepages	= generic_writepages,
+	.releasepage	= blkdev_releasepage,
 	.direct_IO	= blkdev_direct_IO,
 };
 
@@ -1261,7 +1293,7 @@ EXPORT_SYMBOL(ioctl_by_bdev);
 
 /**
  * lookup_bdev  - lookup a struct block_device by name
- * @path:	special file representing the block device
+ * @pathname:	special file representing the block device
  *
  * Get a reference to the blockdevice at @pathname in the current
  * namespace if possible and return it.  Return ERR_PTR(error)
diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
new file mode 100644
index 00000000000..d2cf5a54a4b
--- /dev/null
+++ b/fs/btrfs/Makefile
@@ -0,0 +1,25 @@
+ifneq ($(KERNELRELEASE),)
+# kbuild part of makefile
+
+obj-$(CONFIG_BTRFS_FS) := btrfs.o
+btrfs-y := super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
+	   file-item.o inode-item.o inode-map.o disk-io.o \
+	   transaction.o inode.o file.o tree-defrag.o \
+	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
+	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
+	   ref-cache.o export.o tree-log.o acl.o free-space-cache.o zlib.o \
+	   compression.o
+else
+
+# Normal Makefile
+
+KERNELDIR := /lib/modules/`uname -r`/build
+all:
+	$(MAKE) -C $(KERNELDIR) M=`pwd` CONFIG_BTRFS_FS=m modules
+
+modules_install:
+	$(MAKE) -C $(KERNELDIR) M=`pwd` modules_install
+clean:
+	$(MAKE) -C $(KERNELDIR) M=`pwd` clean
+
+endif
diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c
new file mode 100644
index 00000000000..1d53b62dbba
--- /dev/null
+++ b/fs/btrfs/acl.c
@@ -0,0 +1,351 @@
+/*
+ * Copyright (C) 2007 Red Hat.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/fs.h>
+#include <linux/string.h>
+#include <linux/xattr.h>
+#include <linux/posix_acl_xattr.h>
+#include <linux/posix_acl.h>
+#include <linux/sched.h>
+
+#include "ctree.h"
+#include "btrfs_inode.h"
+#include "xattr.h"
+
+#ifdef CONFIG_FS_POSIX_ACL
+
+static void btrfs_update_cached_acl(struct inode *inode,
+				    struct posix_acl **p_acl,
+				    struct posix_acl *acl)
+{
+	spin_lock(&inode->i_lock);
+	if (*p_acl && *p_acl != BTRFS_ACL_NOT_CACHED)
+		posix_acl_release(*p_acl);
+	*p_acl = posix_acl_dup(acl);
+	spin_unlock(&inode->i_lock);
+}
+
+static struct posix_acl *btrfs_get_acl(struct inode *inode, int type)
+{
+	int size;
+	const char *name;
+	char *value = NULL;
+	struct posix_acl *acl = NULL, **p_acl;
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		name = POSIX_ACL_XATTR_ACCESS;
+		p_acl = &BTRFS_I(inode)->i_acl;
+		break;
+	case ACL_TYPE_DEFAULT:
+		name = POSIX_ACL_XATTR_DEFAULT;
+		p_acl = &BTRFS_I(inode)->i_default_acl;
+		break;
+	default:
+		return ERR_PTR(-EINVAL);
+	}
+
+	spin_lock(&inode->i_lock);
+	if (*p_acl != BTRFS_ACL_NOT_CACHED)
+		acl = posix_acl_dup(*p_acl);
+	spin_unlock(&inode->i_lock);
+
+	if (acl)
+		return acl;
+
+
+	size = __btrfs_getxattr(inode, name, "", 0);
+	if (size > 0) {
+		value = kzalloc(size, GFP_NOFS);
+		if (!value)
+			return ERR_PTR(-ENOMEM);
+		size = __btrfs_getxattr(inode, name, value, size);
+		if (size > 0) {
+			acl = posix_acl_from_xattr(value, size);
+			btrfs_update_cached_acl(inode, p_acl, acl);
+		}
+		kfree(value);
+	} else if (size == -ENOENT) {
+		acl = NULL;
+		btrfs_update_cached_acl(inode, p_acl, acl);
+	}
+
+	return acl;
+}
+
+static int btrfs_xattr_get_acl(struct inode *inode, int type,
+			       void *value, size_t size)
+{
+	struct posix_acl *acl;
+	int ret = 0;
+
+	acl = btrfs_get_acl(inode, type);
+
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl == NULL)
+		return -ENODATA;
+	ret = posix_acl_to_xattr(acl, value, size);
+	posix_acl_release(acl);
+
+	return ret;
+}
+
+/*
+ * Needs to be called with fs_mutex held
+ */
+static int btrfs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
+{
+	int ret, size = 0;
+	const char *name;
+	struct posix_acl **p_acl;
+	char *value = NULL;
+	mode_t mode;
+
+	if (acl) {
+		ret = posix_acl_valid(acl);
+		if (ret < 0)
+			return ret;
+		ret = 0;
+	}
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		mode = inode->i_mode;
+		ret = posix_acl_equiv_mode(acl, &mode);
+		if (ret < 0)
+			return ret;
+		ret = 0;
+		inode->i_mode = mode;
+		name = POSIX_ACL_XATTR_ACCESS;
+		p_acl = &BTRFS_I(inode)->i_acl;
+		break;
+	case ACL_TYPE_DEFAULT:
+		if (!S_ISDIR(inode->i_mode))
+			return acl ? -EINVAL : 0;
+		name = POSIX_ACL_XATTR_DEFAULT;
+		p_acl = &BTRFS_I(inode)->i_default_acl;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (acl) {
+		size = posix_acl_xattr_size(acl->a_count);
+		value = kmalloc(size, GFP_NOFS);
+		if (!value) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		ret = posix_acl_to_xattr(acl, value, size);
+		if (ret < 0)
+			goto out;
+	}
+
+	ret = __btrfs_setxattr(inode, name, value, size, 0);
+
+out:
+	kfree(value);
+
+	if (!ret)
+		btrfs_update_cached_acl(inode, p_acl, acl);
+
+	return ret;
+}
+
+static int btrfs_xattr_set_acl(struct inode *inode, int type,
+			       const void *value, size_t size)
+{
+	int ret = 0;
+	struct posix_acl *acl = NULL;
+
+	if (value) {
+		acl = posix_acl_from_xattr(value, size);
+		if (acl == NULL) {
+			value = NULL;
+			size = 0;
+		} else if (IS_ERR(acl)) {
+			return PTR_ERR(acl);
+		}
+	}
+
+	ret = btrfs_set_acl(inode, acl, type);
+
+	posix_acl_release(acl);
+
+	return ret;
+}
+
+
+static int btrfs_xattr_acl_access_get(struct inode *inode, const char *name,
+				      void *value, size_t size)
+{
+	return btrfs_xattr_get_acl(inode, ACL_TYPE_ACCESS, value, size);
+}
+
+static int btrfs_xattr_acl_access_set(struct inode *inode, const char *name,
+				      const void *value, size_t size, int flags)
+{
+	return btrfs_xattr_set_acl(inode, ACL_TYPE_ACCESS, value, size);
+}
+
+static int btrfs_xattr_acl_default_get(struct inode *inode, const char *name,
+				       void *value, size_t size)
+{
+	return btrfs_xattr_get_acl(inode, ACL_TYPE_DEFAULT, value, size);
+}
+
+static int btrfs_xattr_acl_default_set(struct inode *inode, const char *name,
+			       const void *value, size_t size, int flags)
+{
+	return btrfs_xattr_set_acl(inode, ACL_TYPE_DEFAULT, value, size);
+}
+
+int btrfs_check_acl(struct inode *inode, int mask)
+{
+	struct posix_acl *acl;
+	int error = -EAGAIN;
+
+	acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS);
+
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl) {
+		error = posix_acl_permission(inode, acl, mask);
+		posix_acl_release(acl);
+	}
+
+	return error;
+}
+
+/*
+ * btrfs_init_acl is already generally called under fs_mutex, so the locking
+ * stuff has been fixed to work with that.  If the locking stuff changes, we
+ * need to re-evaluate the acl locking stuff.
+ */
+int btrfs_init_acl(struct inode *inode, struct inode *dir)
+{
+	struct posix_acl *acl = NULL;
+	int ret = 0;
+
+	/* this happens with subvols */
+	if (!dir)
+		return 0;
+
+	if (!S_ISLNK(inode->i_mode)) {
+		if (IS_POSIXACL(dir)) {
+			acl = btrfs_get_acl(dir, ACL_TYPE_DEFAULT);
+			if (IS_ERR(acl))
+				return PTR_ERR(acl);
+		}
+
+		if (!acl)
+			inode->i_mode &= ~current->fs->umask;
+	}
+
+	if (IS_POSIXACL(dir) && acl) {
+		struct posix_acl *clone;
+		mode_t mode;
+
+		if (S_ISDIR(inode->i_mode)) {
+			ret = btrfs_set_acl(inode, acl, ACL_TYPE_DEFAULT);
+			if (ret)
+				goto failed;
+		}
+		clone = posix_acl_clone(acl, GFP_NOFS);
+		ret = -ENOMEM;
+		if (!clone)
+			goto failed;
+
+		mode = inode->i_mode;
+		ret = posix_acl_create_masq(clone, &mode);
+		if (ret >= 0) {
+			inode->i_mode = mode;
+			if (ret > 0) {
+				/* we need an acl */
+				ret = btrfs_set_acl(inode, clone,
+						    ACL_TYPE_ACCESS);
+			}
+		}
+	}
+failed:
+	posix_acl_release(acl);
+
+	return ret;
+}
+
+int btrfs_acl_chmod(struct inode *inode)
+{
+	struct posix_acl *acl, *clone;
+	int ret = 0;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+
+	if (!IS_POSIXACL(inode))
+		return 0;
+
+	acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS);
+	if (IS_ERR(acl) || !acl)
+		return PTR_ERR(acl);
+
+	clone = posix_acl_clone(acl, GFP_KERNEL);
+	posix_acl_release(acl);
+	if (!clone)
+		return -ENOMEM;
+
+	ret = posix_acl_chmod_masq(clone, inode->i_mode);
+	if (!ret)
+		ret = btrfs_set_acl(inode, clone, ACL_TYPE_ACCESS);
+
+	posix_acl_release(clone);
+
+	return ret;
+}
+
+struct xattr_handler btrfs_xattr_acl_default_handler = {
+	.prefix = POSIX_ACL_XATTR_DEFAULT,
+	.get	= btrfs_xattr_acl_default_get,
+	.set	= btrfs_xattr_acl_default_set,
+};
+
+struct xattr_handler btrfs_xattr_acl_access_handler = {
+	.prefix = POSIX_ACL_XATTR_ACCESS,
+	.get	= btrfs_xattr_acl_access_get,
+	.set	= btrfs_xattr_acl_access_set,
+};
+
+#else /* CONFIG_FS_POSIX_ACL */
+
+int btrfs_acl_chmod(struct inode *inode)
+{
+	return 0;
+}
+
+int btrfs_init_acl(struct inode *inode, struct inode *dir)
+{
+	return 0;
+}
+
+int btrfs_check_acl(struct inode *inode, int mask)
+{
+	return 0;
+}
+
+#endif /* CONFIG_FS_POSIX_ACL */
diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c
new file mode 100644
index 00000000000..8e2fec05dbe
--- /dev/null
+++ b/fs/btrfs/async-thread.c
@@ -0,0 +1,419 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/version.h>
+#include <linux/kthread.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+# include <linux/freezer.h>
+#include "async-thread.h"
+
+#define WORK_QUEUED_BIT 0
+#define WORK_DONE_BIT 1
+#define WORK_ORDER_DONE_BIT 2
+
+/*
+ * container for the kthread task pointer and the list of pending work
+ * One of these is allocated per thread.
+ */
+struct btrfs_worker_thread {
+	/* pool we belong to */
+	struct btrfs_workers *workers;
+
+	/* list of struct btrfs_work that are waiting for service */
+	struct list_head pending;
+
+	/* list of worker threads from struct btrfs_workers */
+	struct list_head worker_list;
+
+	/* kthread */
+	struct task_struct *task;
+
+	/* number of things on the pending list */
+	atomic_t num_pending;
+
+	unsigned long sequence;
+
+	/* protects the pending list. */
+	spinlock_t lock;
+
+	/* set to non-zero when this thread is already awake and kicking */
+	int working;
+
+	/* are we currently idle */
+	int idle;
+};
+
+/*
+ * helper function to move a thread onto the idle list after it
+ * has finished some requests.
+ */
+static void check_idle_worker(struct btrfs_worker_thread *worker)
+{
+	if (!worker->idle && atomic_read(&worker->num_pending) <
+	    worker->workers->idle_thresh / 2) {
+		unsigned long flags;
+		spin_lock_irqsave(&worker->workers->lock, flags);
+		worker->idle = 1;
+		list_move(&worker->worker_list, &worker->workers->idle_list);
+		spin_unlock_irqrestore(&worker->workers->lock, flags);
+	}
+}
+
+/*
+ * helper function to move a thread off the idle list after new
+ * pending work is added.
+ */
+static void check_busy_worker(struct btrfs_worker_thread *worker)
+{
+	if (worker->idle && atomic_read(&worker->num_pending) >=
+	    worker->workers->idle_thresh) {
+		unsigned long flags;
+		spin_lock_irqsave(&worker->workers->lock, flags);
+		worker->idle = 0;
+		list_move_tail(&worker->worker_list,
+			       &worker->workers->worker_list);
+		spin_unlock_irqrestore(&worker->workers->lock, flags);
+	}
+}
+
+static noinline int run_ordered_completions(struct btrfs_workers *workers,
+					    struct btrfs_work *work)
+{
+	unsigned long flags;
+
+	if (!workers->ordered)
+		return 0;
+
+	set_bit(WORK_DONE_BIT, &work->flags);
+
+	spin_lock_irqsave(&workers->lock, flags);
+
+	while (!list_empty(&workers->order_list)) {
+		work = list_entry(workers->order_list.next,
+				  struct btrfs_work, order_list);
+
+		if (!test_bit(WORK_DONE_BIT, &work->flags))
+			break;
+
+		/* we are going to call the ordered done function, but
+		 * we leave the work item on the list as a barrier so
+		 * that later work items that are done don't have their
+		 * functions called before this one returns
+		 */
+		if (test_and_set_bit(WORK_ORDER_DONE_BIT, &work->flags))
+			break;
+
+		spin_unlock_irqrestore(&workers->lock, flags);
+
+		work->ordered_func(work);
+
+		/* now take the lock again and call the freeing code */
+		spin_lock_irqsave(&workers->lock, flags);
+		list_del(&work->order_list);
+		work->ordered_free(work);
+	}
+
+	spin_unlock_irqrestore(&workers->lock, flags);
+	return 0;
+}
+
+/*
+ * main loop for servicing work items
+ */
+static int worker_loop(void *arg)
+{
+	struct btrfs_worker_thread *worker = arg;
+	struct list_head *cur;
+	struct btrfs_work *work;
+	do {
+		spin_lock_irq(&worker->lock);
+		while (!list_empty(&worker->pending)) {
+			cur = worker->pending.next;
+			work = list_entry(cur, struct btrfs_work, list);
+			list_del(&work->list);
+			clear_bit(WORK_QUEUED_BIT, &work->flags);
+
+			work->worker = worker;
+			spin_unlock_irq(&worker->lock);
+
+			work->func(work);
+
+			atomic_dec(&worker->num_pending);
+			/*
+			 * unless this is an ordered work queue,
+			 * 'work' was probably freed by func above.
+			 */
+			run_ordered_completions(worker->workers, work);
+
+			spin_lock_irq(&worker->lock);
+			check_idle_worker(worker);
+
+		}
+		worker->working = 0;
+		if (freezing(current)) {
+			refrigerator();
+		} else {
+			set_current_state(TASK_INTERRUPTIBLE);
+			spin_unlock_irq(&worker->lock);
+			if (!kthread_should_stop())
+				schedule();
+			__set_current_state(TASK_RUNNING);
+		}
+	} while (!kthread_should_stop());
+	return 0;
+}
+
+/*
+ * this will wait for all the worker threads to shutdown
+ */
+int btrfs_stop_workers(struct btrfs_workers *workers)
+{
+	struct list_head *cur;
+	struct btrfs_worker_thread *worker;
+
+	list_splice_init(&workers->idle_list, &workers->worker_list);
+	while (!list_empty(&workers->worker_list)) {
+		cur = workers->worker_list.next;
+		worker = list_entry(cur, struct btrfs_worker_thread,
+				    worker_list);
+		kthread_stop(worker->task);
+		list_del(&worker->worker_list);
+		kfree(worker);
+	}
+	return 0;
+}
+
+/*
+ * simple init on struct btrfs_workers
+ */
+void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max)
+{
+	workers->num_workers = 0;
+	INIT_LIST_HEAD(&workers->worker_list);
+	INIT_LIST_HEAD(&workers->idle_list);
+	INIT_LIST_HEAD(&workers->order_list);
+	spin_lock_init(&workers->lock);
+	workers->max_workers = max;
+	workers->idle_thresh = 32;
+	workers->name = name;
+	workers->ordered = 0;
+}
+
+/*
+ * starts new worker threads.  This does not enforce the max worker
+ * count in case you need to temporarily go past it.
+ */
+int btrfs_start_workers(struct btrfs_workers *workers, int num_workers)
+{
+	struct btrfs_worker_thread *worker;
+	int ret = 0;
+	int i;
+
+	for (i = 0; i < num_workers; i++) {
+		worker = kzalloc(sizeof(*worker), GFP_NOFS);
+		if (!worker) {
+			ret = -ENOMEM;
+			goto fail;
+		}
+
+		INIT_LIST_HEAD(&worker->pending);
+		INIT_LIST_HEAD(&worker->worker_list);
+		spin_lock_init(&worker->lock);
+		atomic_set(&worker->num_pending, 0);
+		worker->task = kthread_run(worker_loop, worker,
+					   "btrfs-%s-%d", workers->name,
+					   workers->num_workers + i);
+		worker->workers = workers;
+		if (IS_ERR(worker->task)) {
+			kfree(worker);
+			ret = PTR_ERR(worker->task);
+			goto fail;
+		}
+
+		spin_lock_irq(&workers->lock);
+		list_add_tail(&worker->worker_list, &workers->idle_list);
+		worker->idle = 1;
+		workers->num_workers++;
+		spin_unlock_irq(&workers->lock);
+	}
+	return 0;
+fail:
+	btrfs_stop_workers(workers);
+	return ret;
+}
+
+/*
+ * run through the list and find a worker thread that doesn't have a lot
+ * to do right now.  This can return null if we aren't yet at the thread
+ * count limit and all of the threads are busy.
+ */
+static struct btrfs_worker_thread *next_worker(struct btrfs_workers *workers)
+{
+	struct btrfs_worker_thread *worker;
+	struct list_head *next;
+	int enforce_min = workers->num_workers < workers->max_workers;
+
+	/*
+	 * if we find an idle thread, don't move it to the end of the
+	 * idle list.  This improves the chance that the next submission
+	 * will reuse the same thread, and maybe catch it while it is still
+	 * working
+	 */
+	if (!list_empty(&workers->idle_list)) {
+		next = workers->idle_list.next;
+		worker = list_entry(next, struct btrfs_worker_thread,
+				    worker_list);
+		return worker;
+	}
+	if (enforce_min || list_empty(&workers->worker_list))
+		return NULL;
+
+	/*
+	 * if we pick a busy task, move the task to the end of the list.
+	 * hopefully this will keep things somewhat evenly balanced.
+	 * Do the move in batches based on the sequence number.  This groups
+	 * requests submitted at roughly the same time onto the same worker.
+	 */
+	next = workers->worker_list.next;
+	worker = list_entry(next, struct btrfs_worker_thread, worker_list);
+	atomic_inc(&worker->num_pending);
+	worker->sequence++;
+
+	if (worker->sequence % workers->idle_thresh == 0)
+		list_move_tail(next, &workers->worker_list);
+	return worker;
+}
+
+/*
+ * selects a worker thread to take the next job.  This will either find
+ * an idle worker, start a new worker up to the max count, or just return
+ * one of the existing busy workers.
+ */
+static struct btrfs_worker_thread *find_worker(struct btrfs_workers *workers)
+{
+	struct btrfs_worker_thread *worker;
+	unsigned long flags;
+
+again:
+	spin_lock_irqsave(&workers->lock, flags);
+	worker = next_worker(workers);
+	spin_unlock_irqrestore(&workers->lock, flags);
+
+	if (!worker) {
+		spin_lock_irqsave(&workers->lock, flags);
+		if (workers->num_workers >= workers->max_workers) {
+			struct list_head *fallback = NULL;
+			/*
+			 * we have failed to find any workers, just
+			 * return the force one
+			 */
+			if (!list_empty(&workers->worker_list))
+				fallback = workers->worker_list.next;
+			if (!list_empty(&workers->idle_list))
+				fallback = workers->idle_list.next;
+			BUG_ON(!fallback);
+			worker = list_entry(fallback,
+				  struct btrfs_worker_thread, worker_list);
+			spin_unlock_irqrestore(&workers->lock, flags);
+		} else {
+			spin_unlock_irqrestore(&workers->lock, flags);
+			/* we're below the limit, start another worker */
+			btrfs_start_workers(workers, 1);
+			goto again;
+		}
+	}
+	return worker;
+}
+
+/*
+ * btrfs_requeue_work just puts the work item back on the tail of the list
+ * it was taken from.  It is intended for use with long running work functions
+ * that make some progress and want to give the cpu up for others.
+ */
+int btrfs_requeue_work(struct btrfs_work *work)
+{
+	struct btrfs_worker_thread *worker = work->worker;
+	unsigned long flags;
+
+	if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags))
+		goto out;
+
+	spin_lock_irqsave(&worker->lock, flags);
+	atomic_inc(&worker->num_pending);
+	list_add_tail(&work->list, &worker->pending);
+
+	/* by definition we're busy, take ourselves off the idle
+	 * list
+	 */
+	if (worker->idle) {
+		spin_lock_irqsave(&worker->workers->lock, flags);
+		worker->idle = 0;
+		list_move_tail(&worker->worker_list,
+			       &worker->workers->worker_list);
+		spin_unlock_irqrestore(&worker->workers->lock, flags);
+	}
+
+	spin_unlock_irqrestore(&worker->lock, flags);
+
+out:
+	return 0;
+}
+
+/*
+ * places a struct btrfs_work into the pending queue of one of the kthreads
+ */
+int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work)
+{
+	struct btrfs_worker_thread *worker;
+	unsigned long flags;
+	int wake = 0;
+
+	/* don't requeue something already on a list */
+	if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags))
+		goto out;
+
+	worker = find_worker(workers);
+	if (workers->ordered) {
+		spin_lock_irqsave(&workers->lock, flags);
+		list_add_tail(&work->order_list, &workers->order_list);
+		spin_unlock_irqrestore(&workers->lock, flags);
+	} else {
+		INIT_LIST_HEAD(&work->order_list);
+	}
+
+	spin_lock_irqsave(&worker->lock, flags);
+	atomic_inc(&worker->num_pending);
+	check_busy_worker(worker);
+	list_add_tail(&work->list, &worker->pending);
+
+	/*
+	 * avoid calling into wake_up_process if this thread has already
+	 * been kicked
+	 */
+	if (!worker->working)
+		wake = 1;
+	worker->working = 1;
+
+	spin_unlock_irqrestore(&worker->lock, flags);
+
+	if (wake)
+		wake_up_process(worker->task);
+out:
+	return 0;
+}
diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h
new file mode 100644
index 00000000000..31be4ed8b63
--- /dev/null
+++ b/fs/btrfs/async-thread.h
@@ -0,0 +1,101 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_ASYNC_THREAD_
+#define __BTRFS_ASYNC_THREAD_
+
+struct btrfs_worker_thread;
+
+/*
+ * This is similar to a workqueue, but it is meant to spread the operations
+ * across all available cpus instead of just the CPU that was used to
+ * queue the work.  There is also some batching introduced to try and
+ * cut down on context switches.
+ *
+ * By default threads are added on demand up to 2 * the number of cpus.
+ * Changing struct btrfs_workers->max_workers is one way to prevent
+ * demand creation of kthreads.
+ *
+ * the basic model of these worker threads is to embed a btrfs_work
+ * structure in your own data struct, and use container_of in a
+ * work function to get back to your data struct.
+ */
+struct btrfs_work {
+	/*
+	 * func should be set to the function you want called
+	 * your work struct is passed as the only arg
+	 *
+	 * ordered_func must be set for work sent to an ordered work queue,
+	 * and it is called to complete a given work item in the same
+	 * order they were sent to the queue.
+	 */
+	void (*func)(struct btrfs_work *work);
+	void (*ordered_func)(struct btrfs_work *work);
+	void (*ordered_free)(struct btrfs_work *work);
+
+	/*
+	 * flags should be set to zero.  It is used to make sure the
+	 * struct is only inserted once into the list.
+	 */
+	unsigned long flags;
+
+	/* don't touch these */
+	struct btrfs_worker_thread *worker;
+	struct list_head list;
+	struct list_head order_list;
+};
+
+struct btrfs_workers {
+	/* current number of running workers */
+	int num_workers;
+
+	/* max number of workers allowed.  changed by btrfs_start_workers */
+	int max_workers;
+
+	/* once a worker has this many requests or fewer, it is idle */
+	int idle_thresh;
+
+	/* force completions in the order they were queued */
+	int ordered;
+
+	/* list with all the work threads.  The workers on the idle thread
+	 * may be actively servicing jobs, but they haven't yet hit the
+	 * idle thresh limit above.
+	 */
+	struct list_head worker_list;
+	struct list_head idle_list;
+
+	/*
+	 * when operating in ordered mode, this maintains the list
+	 * of work items waiting for completion
+	 */
+	struct list_head order_list;
+
+	/* lock for finding the next worker thread to queue on */
+	spinlock_t lock;
+
+	/* extra name for this worker, used for current->name */
+	char *name;
+};
+
+int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work);
+int btrfs_start_workers(struct btrfs_workers *workers, int num_workers);
+int btrfs_stop_workers(struct btrfs_workers *workers);
+void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max);
+int btrfs_requeue_work(struct btrfs_work *work);
+#endif
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
new file mode 100644
index 00000000000..a8c9693b75a
--- /dev/null
+++ b/fs/btrfs/btrfs_inode.h
@@ -0,0 +1,131 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_I__
+#define __BTRFS_I__
+
+#include "extent_map.h"
+#include "extent_io.h"
+#include "ordered-data.h"
+
+/* in memory btrfs inode */
+struct btrfs_inode {
+	/* which subvolume this inode belongs to */
+	struct btrfs_root *root;
+
+	/* key used to find this inode on disk.  This is used by the code
+	 * to read in roots of subvolumes
+	 */
+	struct btrfs_key location;
+
+	/* the extent_tree has caches of all the extent mappings to disk */
+	struct extent_map_tree extent_tree;
+
+	/* the io_tree does range state (DIRTY, LOCKED etc) */
+	struct extent_io_tree io_tree;
+
+	/* special utility tree used to record which mirrors have already been
+	 * tried when checksums fail for a given block
+	 */
+	struct extent_io_tree io_failure_tree;
+
+	/* held while inesrting or deleting extents from files */
+	struct mutex extent_mutex;
+
+	/* held while logging the inode in tree-log.c */
+	struct mutex log_mutex;
+
+	/* used to order data wrt metadata */
+	struct btrfs_ordered_inode_tree ordered_tree;
+
+	/* standard acl pointers */
+	struct posix_acl *i_acl;
+	struct posix_acl *i_default_acl;
+
+	/* for keeping track of orphaned inodes */
+	struct list_head i_orphan;
+
+	/* list of all the delalloc inodes in the FS.  There are times we need
+	 * to write all the delalloc pages to disk, and this list is used
+	 * to walk them all.
+	 */
+	struct list_head delalloc_inodes;
+
+	/* full 64 bit generation number, struct vfs_inode doesn't have a big
+	 * enough field for this.
+	 */
+	u64 generation;
+
+	/* sequence number for NFS changes */
+	u64 sequence;
+
+	/*
+	 * transid of the trans_handle that last modified this inode
+	 */
+	u64 last_trans;
+	/*
+	 * transid that last logged this inode
+	 */
+	u64 logged_trans;
+
+	/*
+	 * trans that last made a change that should be fully fsync'd.  This
+	 * gets reset to zero each time the inode is logged
+	 */
+	u64 log_dirty_trans;
+
+	/* total number of bytes pending delalloc, used by stat to calc the
+	 * real block usage of the file
+	 */
+	u64 delalloc_bytes;
+
+	/*
+	 * the size of the file stored in the metadata on disk.  data=ordered
+	 * means the in-memory i_size might be larger than the size on disk
+	 * because not all the blocks are written yet.
+	 */
+	u64 disk_i_size;
+
+	/* flags field from the on disk inode */
+	u32 flags;
+
+	/*
+	 * if this is a directory then index_cnt is the counter for the index
+	 * number for new files that are created
+	 */
+	u64 index_cnt;
+
+	/* the start of block group preferred for allocations. */
+	u64 block_group;
+
+	struct inode vfs_inode;
+};
+
+static inline struct btrfs_inode *BTRFS_I(struct inode *inode)
+{
+	return container_of(inode, struct btrfs_inode, vfs_inode);
+}
+
+static inline void btrfs_i_size_write(struct inode *inode, u64 size)
+{
+	inode->i_size = size;
+	BTRFS_I(inode)->disk_i_size = size;
+}
+
+
+#endif
diff --git a/fs/btrfs/compat.h b/fs/btrfs/compat.h
new file mode 100644
index 00000000000..7c4503ef6ef
--- /dev/null
+++ b/fs/btrfs/compat.h
@@ -0,0 +1,7 @@
+#ifndef _COMPAT_H_
+#define _COMPAT_H_
+
+#define btrfs_drop_nlink(inode) drop_nlink(inode)
+#define btrfs_inc_nlink(inode)	inc_nlink(inode)
+
+#endif /* _COMPAT_H_ */
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
new file mode 100644
index 00000000000..ee848d8585d
--- /dev/null
+++ b/fs/btrfs/compression.c
@@ -0,0 +1,709 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/kernel.h>
+#include <linux/bio.h>
+#include <linux/buffer_head.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/pagemap.h>
+#include <linux/highmem.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/smp_lock.h>
+#include <linux/backing-dev.h>
+#include <linux/mpage.h>
+#include <linux/swap.h>
+#include <linux/writeback.h>
+#include <linux/bit_spinlock.h>
+#include <linux/version.h>
+#include <linux/pagevec.h>
+#include "compat.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "volumes.h"
+#include "ordered-data.h"
+#include "compression.h"
+#include "extent_io.h"
+#include "extent_map.h"
+
+struct compressed_bio {
+	/* number of bios pending for this compressed extent */
+	atomic_t pending_bios;
+
+	/* the pages with the compressed data on them */
+	struct page **compressed_pages;
+
+	/* inode that owns this data */
+	struct inode *inode;
+
+	/* starting offset in the inode for our pages */
+	u64 start;
+
+	/* number of bytes in the inode we're working on */
+	unsigned long len;
+
+	/* number of bytes on disk */
+	unsigned long compressed_len;
+
+	/* number of compressed pages in the array */
+	unsigned long nr_pages;
+
+	/* IO errors */
+	int errors;
+	int mirror_num;
+
+	/* for reads, this is the bio we are copying the data into */
+	struct bio *orig_bio;
+
+	/*
+	 * the start of a variable length array of checksums only
+	 * used by reads
+	 */
+	u32 sums;
+};
+
+static inline int compressed_bio_size(struct btrfs_root *root,
+				      unsigned long disk_size)
+{
+	u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy);
+	return sizeof(struct compressed_bio) +
+		((disk_size + root->sectorsize - 1) / root->sectorsize) *
+		csum_size;
+}
+
+static struct bio *compressed_bio_alloc(struct block_device *bdev,
+					u64 first_byte, gfp_t gfp_flags)
+{
+	struct bio *bio;
+	int nr_vecs;
+
+	nr_vecs = bio_get_nr_vecs(bdev);
+	bio = bio_alloc(gfp_flags, nr_vecs);
+
+	if (bio == NULL && (current->flags & PF_MEMALLOC)) {
+		while (!bio && (nr_vecs /= 2))
+			bio = bio_alloc(gfp_flags, nr_vecs);
+	}
+
+	if (bio) {
+		bio->bi_size = 0;
+		bio->bi_bdev = bdev;
+		bio->bi_sector = first_byte >> 9;
+	}
+	return bio;
+}
+
+static int check_compressed_csum(struct inode *inode,
+				 struct compressed_bio *cb,
+				 u64 disk_start)
+{
+	int ret;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct page *page;
+	unsigned long i;
+	char *kaddr;
+	u32 csum;
+	u32 *cb_sum = &cb->sums;
+
+	if (btrfs_test_flag(inode, NODATASUM))
+		return 0;
+
+	for (i = 0; i < cb->nr_pages; i++) {
+		page = cb->compressed_pages[i];
+		csum = ~(u32)0;
+
+		kaddr = kmap_atomic(page, KM_USER0);
+		csum = btrfs_csum_data(root, kaddr, csum, PAGE_CACHE_SIZE);
+		btrfs_csum_final(csum, (char *)&csum);
+		kunmap_atomic(kaddr, KM_USER0);
+
+		if (csum != *cb_sum) {
+			printk(KERN_INFO "btrfs csum failed ino %lu "
+			       "extent %llu csum %u "
+			       "wanted %u mirror %d\n", inode->i_ino,
+			       (unsigned long long)disk_start,
+			       csum, *cb_sum, cb->mirror_num);
+			ret = -EIO;
+			goto fail;
+		}
+		cb_sum++;
+
+	}
+	ret = 0;
+fail:
+	return ret;
+}
+
+/* when we finish reading compressed pages from the disk, we
+ * decompress them and then run the bio end_io routines on the
+ * decompressed pages (in the inode address space).
+ *
+ * This allows the checksumming and other IO error handling routines
+ * to work normally
+ *
+ * The compressed pages are freed here, and it must be run
+ * in process context
+ */
+static void end_compressed_bio_read(struct bio *bio, int err)
+{
+	struct extent_io_tree *tree;
+	struct compressed_bio *cb = bio->bi_private;
+	struct inode *inode;
+	struct page *page;
+	unsigned long index;
+	int ret;
+
+	if (err)
+		cb->errors = 1;
+
+	/* if there are more bios still pending for this compressed
+	 * extent, just exit
+	 */
+	if (!atomic_dec_and_test(&cb->pending_bios))
+		goto out;
+
+	inode = cb->inode;
+	ret = check_compressed_csum(inode, cb, (u64)bio->bi_sector << 9);
+	if (ret)
+		goto csum_failed;
+
+	/* ok, we're the last bio for this extent, lets start
+	 * the decompression.
+	 */
+	tree = &BTRFS_I(inode)->io_tree;
+	ret = btrfs_zlib_decompress_biovec(cb->compressed_pages,
+					cb->start,
+					cb->orig_bio->bi_io_vec,
+					cb->orig_bio->bi_vcnt,
+					cb->compressed_len);
+csum_failed:
+	if (ret)
+		cb->errors = 1;
+
+	/* release the compressed pages */
+	index = 0;
+	for (index = 0; index < cb->nr_pages; index++) {
+		page = cb->compressed_pages[index];
+		page->mapping = NULL;
+		page_cache_release(page);
+	}
+
+	/* do io completion on the original bio */
+	if (cb->errors) {
+		bio_io_error(cb->orig_bio);
+	} else {
+		int bio_index = 0;
+		struct bio_vec *bvec = cb->orig_bio->bi_io_vec;
+
+		/*
+		 * we have verified the checksum already, set page
+		 * checked so the end_io handlers know about it
+		 */
+		while (bio_index < cb->orig_bio->bi_vcnt) {
+			SetPageChecked(bvec->bv_page);
+			bvec++;
+			bio_index++;
+		}
+		bio_endio(cb->orig_bio, 0);
+	}
+
+	/* finally free the cb struct */
+	kfree(cb->compressed_pages);
+	kfree(cb);
+out:
+	bio_put(bio);
+}
+
+/*
+ * Clear the writeback bits on all of the file
+ * pages for a compressed write
+ */
+static noinline int end_compressed_writeback(struct inode *inode, u64 start,
+					     unsigned long ram_size)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = (start + ram_size - 1) >> PAGE_CACHE_SHIFT;
+	struct page *pages[16];
+	unsigned long nr_pages = end_index - index + 1;
+	int i;
+	int ret;
+
+	while (nr_pages > 0) {
+		ret = find_get_pages_contig(inode->i_mapping, index,
+				     min_t(unsigned long,
+				     nr_pages, ARRAY_SIZE(pages)), pages);
+		if (ret == 0) {
+			nr_pages -= 1;
+			index += 1;
+			continue;
+		}
+		for (i = 0; i < ret; i++) {
+			end_page_writeback(pages[i]);
+			page_cache_release(pages[i]);
+		}
+		nr_pages -= ret;
+		index += ret;
+	}
+	/* the inode may be gone now */
+	return 0;
+}
+
+/*
+ * do the cleanup once all the compressed pages hit the disk.
+ * This will clear writeback on the file pages and free the compressed
+ * pages.
+ *
+ * This also calls the writeback end hooks for the file pages so that
+ * metadata and checksums can be updated in the file.
+ */
+static void end_compressed_bio_write(struct bio *bio, int err)
+{
+	struct extent_io_tree *tree;
+	struct compressed_bio *cb = bio->bi_private;
+	struct inode *inode;
+	struct page *page;
+	unsigned long index;
+
+	if (err)
+		cb->errors = 1;
+
+	/* if there are more bios still pending for this compressed
+	 * extent, just exit
+	 */
+	if (!atomic_dec_and_test(&cb->pending_bios))
+		goto out;
+
+	/* ok, we're the last bio for this extent, step one is to
+	 * call back into the FS and do all the end_io operations
+	 */
+	inode = cb->inode;
+	tree = &BTRFS_I(inode)->io_tree;
+	cb->compressed_pages[0]->mapping = cb->inode->i_mapping;
+	tree->ops->writepage_end_io_hook(cb->compressed_pages[0],
+					 cb->start,
+					 cb->start + cb->len - 1,
+					 NULL, 1);
+	cb->compressed_pages[0]->mapping = NULL;
+
+	end_compressed_writeback(inode, cb->start, cb->len);
+	/* note, our inode could be gone now */
+
+	/*
+	 * release the compressed pages, these came from alloc_page and
+	 * are not attached to the inode at all
+	 */
+	index = 0;
+	for (index = 0; index < cb->nr_pages; index++) {
+		page = cb->compressed_pages[index];
+		page->mapping = NULL;
+		page_cache_release(page);
+	}
+
+	/* finally free the cb struct */
+	kfree(cb->compressed_pages);
+	kfree(cb);
+out:
+	bio_put(bio);
+}
+
+/*
+ * worker function to build and submit bios for previously compressed pages.
+ * The corresponding pages in the inode should be marked for writeback
+ * and the compressed pages should have a reference on them for dropping
+ * when the IO is complete.
+ *
+ * This also checksums the file bytes and gets things ready for
+ * the end io hooks.
+ */
+int btrfs_submit_compressed_write(struct inode *inode, u64 start,
+				 unsigned long len, u64 disk_start,
+				 unsigned long compressed_len,
+				 struct page **compressed_pages,
+				 unsigned long nr_pages)
+{
+	struct bio *bio = NULL;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct compressed_bio *cb;
+	unsigned long bytes_left;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	int page_index = 0;
+	struct page *page;
+	u64 first_byte = disk_start;
+	struct block_device *bdev;
+	int ret;
+
+	WARN_ON(start & ((u64)PAGE_CACHE_SIZE - 1));
+	cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS);
+	atomic_set(&cb->pending_bios, 0);
+	cb->errors = 0;
+	cb->inode = inode;
+	cb->start = start;
+	cb->len = len;
+	cb->mirror_num = 0;
+	cb->compressed_pages = compressed_pages;
+	cb->compressed_len = compressed_len;
+	cb->orig_bio = NULL;
+	cb->nr_pages = nr_pages;
+
+	bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
+
+	bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS);
+	bio->bi_private = cb;
+	bio->bi_end_io = end_compressed_bio_write;
+	atomic_inc(&cb->pending_bios);
+
+	/* create and submit bios for the compressed pages */
+	bytes_left = compressed_len;
+	for (page_index = 0; page_index < cb->nr_pages; page_index++) {
+		page = compressed_pages[page_index];
+		page->mapping = inode->i_mapping;
+		if (bio->bi_size)
+			ret = io_tree->ops->merge_bio_hook(page, 0,
+							   PAGE_CACHE_SIZE,
+							   bio, 0);
+		else
+			ret = 0;
+
+		page->mapping = NULL;
+		if (ret || bio_add_page(bio, page, PAGE_CACHE_SIZE, 0) <
+		    PAGE_CACHE_SIZE) {
+			bio_get(bio);
+
+			/*
+			 * inc the count before we submit the bio so
+			 * we know the end IO handler won't happen before
+			 * we inc the count.  Otherwise, the cb might get
+			 * freed before we're done setting it up
+			 */
+			atomic_inc(&cb->pending_bios);
+			ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
+			BUG_ON(ret);
+
+			ret = btrfs_csum_one_bio(root, inode, bio, start, 1);
+			BUG_ON(ret);
+
+			ret = btrfs_map_bio(root, WRITE, bio, 0, 1);
+			BUG_ON(ret);
+
+			bio_put(bio);
+
+			bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS);
+			bio->bi_private = cb;
+			bio->bi_end_io = end_compressed_bio_write;
+			bio_add_page(bio, page, PAGE_CACHE_SIZE, 0);
+		}
+		if (bytes_left < PAGE_CACHE_SIZE) {
+			printk("bytes left %lu compress len %lu nr %lu\n",
+			       bytes_left, cb->compressed_len, cb->nr_pages);
+		}
+		bytes_left -= PAGE_CACHE_SIZE;
+		first_byte += PAGE_CACHE_SIZE;
+		cond_resched();
+	}
+	bio_get(bio);
+
+	ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
+	BUG_ON(ret);
+
+	ret = btrfs_csum_one_bio(root, inode, bio, start, 1);
+	BUG_ON(ret);
+
+	ret = btrfs_map_bio(root, WRITE, bio, 0, 1);
+	BUG_ON(ret);
+
+	bio_put(bio);
+	return 0;
+}
+
+static noinline int add_ra_bio_pages(struct inode *inode,
+				     u64 compressed_end,
+				     struct compressed_bio *cb)
+{
+	unsigned long end_index;
+	unsigned long page_index;
+	u64 last_offset;
+	u64 isize = i_size_read(inode);
+	int ret;
+	struct page *page;
+	unsigned long nr_pages = 0;
+	struct extent_map *em;
+	struct address_space *mapping = inode->i_mapping;
+	struct pagevec pvec;
+	struct extent_map_tree *em_tree;
+	struct extent_io_tree *tree;
+	u64 end;
+	int misses = 0;
+
+	page = cb->orig_bio->bi_io_vec[cb->orig_bio->bi_vcnt - 1].bv_page;
+	last_offset = (page_offset(page) + PAGE_CACHE_SIZE);
+	em_tree = &BTRFS_I(inode)->extent_tree;
+	tree = &BTRFS_I(inode)->io_tree;
+
+	if (isize == 0)
+		return 0;
+
+	end_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
+
+	pagevec_init(&pvec, 0);
+	while (last_offset < compressed_end) {
+		page_index = last_offset >> PAGE_CACHE_SHIFT;
+
+		if (page_index > end_index)
+			break;
+
+		rcu_read_lock();
+		page = radix_tree_lookup(&mapping->page_tree, page_index);
+		rcu_read_unlock();
+		if (page) {
+			misses++;
+			if (misses > 4)
+				break;
+			goto next;
+		}
+
+		page = alloc_page(mapping_gfp_mask(mapping) | GFP_NOFS);
+		if (!page)
+			break;
+
+		page->index = page_index;
+		/*
+		 * what we want to do here is call add_to_page_cache_lru,
+		 * but that isn't exported, so we reproduce it here
+		 */
+		if (add_to_page_cache(page, mapping,
+				      page->index, GFP_NOFS)) {
+			page_cache_release(page);
+			goto next;
+		}
+
+		/* open coding of lru_cache_add, also not exported */
+		page_cache_get(page);
+		if (!pagevec_add(&pvec, page))
+			__pagevec_lru_add_file(&pvec);
+
+		end = last_offset + PAGE_CACHE_SIZE - 1;
+		/*
+		 * at this point, we have a locked page in the page cache
+		 * for these bytes in the file.  But, we have to make
+		 * sure they map to this compressed extent on disk.
+		 */
+		set_page_extent_mapped(page);
+		lock_extent(tree, last_offset, end, GFP_NOFS);
+		spin_lock(&em_tree->lock);
+		em = lookup_extent_mapping(em_tree, last_offset,
+					   PAGE_CACHE_SIZE);
+		spin_unlock(&em_tree->lock);
+
+		if (!em || last_offset < em->start ||
+		    (last_offset + PAGE_CACHE_SIZE > extent_map_end(em)) ||
+		    (em->block_start >> 9) != cb->orig_bio->bi_sector) {
+			free_extent_map(em);
+			unlock_extent(tree, last_offset, end, GFP_NOFS);
+			unlock_page(page);
+			page_cache_release(page);
+			break;
+		}
+		free_extent_map(em);
+
+		if (page->index == end_index) {
+			char *userpage;
+			size_t zero_offset = isize & (PAGE_CACHE_SIZE - 1);
+
+			if (zero_offset) {
+				int zeros;
+				zeros = PAGE_CACHE_SIZE - zero_offset;
+				userpage = kmap_atomic(page, KM_USER0);
+				memset(userpage + zero_offset, 0, zeros);
+				flush_dcache_page(page);
+				kunmap_atomic(userpage, KM_USER0);
+			}
+		}
+
+		ret = bio_add_page(cb->orig_bio, page,
+				   PAGE_CACHE_SIZE, 0);
+
+		if (ret == PAGE_CACHE_SIZE) {
+			nr_pages++;
+			page_cache_release(page);
+		} else {
+			unlock_extent(tree, last_offset, end, GFP_NOFS);
+			unlock_page(page);
+			page_cache_release(page);
+			break;
+		}
+next:
+		last_offset += PAGE_CACHE_SIZE;
+	}
+	if (pagevec_count(&pvec))
+		__pagevec_lru_add_file(&pvec);
+	return 0;
+}
+
+/*
+ * for a compressed read, the bio we get passed has all the inode pages
+ * in it.  We don't actually do IO on those pages but allocate new ones
+ * to hold the compressed pages on disk.
+ *
+ * bio->bi_sector points to the compressed extent on disk
+ * bio->bi_io_vec points to all of the inode pages
+ * bio->bi_vcnt is a count of pages
+ *
+ * After the compressed pages are read, we copy the bytes into the
+ * bio we were passed and then call the bio end_io calls
+ */
+int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
+				 int mirror_num, unsigned long bio_flags)
+{
+	struct extent_io_tree *tree;
+	struct extent_map_tree *em_tree;
+	struct compressed_bio *cb;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	unsigned long uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE;
+	unsigned long compressed_len;
+	unsigned long nr_pages;
+	unsigned long page_index;
+	struct page *page;
+	struct block_device *bdev;
+	struct bio *comp_bio;
+	u64 cur_disk_byte = (u64)bio->bi_sector << 9;
+	u64 em_len;
+	u64 em_start;
+	struct extent_map *em;
+	int ret;
+	u32 *sums;
+
+	tree = &BTRFS_I(inode)->io_tree;
+	em_tree = &BTRFS_I(inode)->extent_tree;
+
+	/* we need the actual starting offset of this extent in the file */
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree,
+				   page_offset(bio->bi_io_vec->bv_page),
+				   PAGE_CACHE_SIZE);
+	spin_unlock(&em_tree->lock);
+
+	compressed_len = em->block_len;
+	cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS);
+	atomic_set(&cb->pending_bios, 0);
+	cb->errors = 0;
+	cb->inode = inode;
+	cb->mirror_num = mirror_num;
+	sums = &cb->sums;
+
+	cb->start = em->orig_start;
+	em_len = em->len;
+	em_start = em->start;
+
+	free_extent_map(em);
+	em = NULL;
+
+	cb->len = uncompressed_len;
+	cb->compressed_len = compressed_len;
+	cb->orig_bio = bio;
+
+	nr_pages = (compressed_len + PAGE_CACHE_SIZE - 1) /
+				 PAGE_CACHE_SIZE;
+	cb->compressed_pages = kmalloc(sizeof(struct page *) * nr_pages,
+				       GFP_NOFS);
+	bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
+
+	for (page_index = 0; page_index < nr_pages; page_index++) {
+		cb->compressed_pages[page_index] = alloc_page(GFP_NOFS |
+							      __GFP_HIGHMEM);
+	}
+	cb->nr_pages = nr_pages;
+
+	add_ra_bio_pages(inode, em_start + em_len, cb);
+
+	/* include any pages we added in add_ra-bio_pages */
+	uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE;
+	cb->len = uncompressed_len;
+
+	comp_bio = compressed_bio_alloc(bdev, cur_disk_byte, GFP_NOFS);
+	comp_bio->bi_private = cb;
+	comp_bio->bi_end_io = end_compressed_bio_read;
+	atomic_inc(&cb->pending_bios);
+
+	for (page_index = 0; page_index < nr_pages; page_index++) {
+		page = cb->compressed_pages[page_index];
+		page->mapping = inode->i_mapping;
+		page->index = em_start >> PAGE_CACHE_SHIFT;
+
+		if (comp_bio->bi_size)
+			ret = tree->ops->merge_bio_hook(page, 0,
+							PAGE_CACHE_SIZE,
+							comp_bio, 0);
+		else
+			ret = 0;
+
+		page->mapping = NULL;
+		if (ret || bio_add_page(comp_bio, page, PAGE_CACHE_SIZE, 0) <
+		    PAGE_CACHE_SIZE) {
+			bio_get(comp_bio);
+
+			ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0);
+			BUG_ON(ret);
+
+			/*
+			 * inc the count before we submit the bio so
+			 * we know the end IO handler won't happen before
+			 * we inc the count.  Otherwise, the cb might get
+			 * freed before we're done setting it up
+			 */
+			atomic_inc(&cb->pending_bios);
+
+			if (!btrfs_test_flag(inode, NODATASUM)) {
+				btrfs_lookup_bio_sums(root, inode, comp_bio,
+						      sums);
+			}
+			sums += (comp_bio->bi_size + root->sectorsize - 1) /
+				root->sectorsize;
+
+			ret = btrfs_map_bio(root, READ, comp_bio,
+					    mirror_num, 0);
+			BUG_ON(ret);
+
+			bio_put(comp_bio);
+
+			comp_bio = compressed_bio_alloc(bdev, cur_disk_byte,
+							GFP_NOFS);
+			comp_bio->bi_private = cb;
+			comp_bio->bi_end_io = end_compressed_bio_read;
+
+			bio_add_page(comp_bio, page, PAGE_CACHE_SIZE, 0);
+		}
+		cur_disk_byte += PAGE_CACHE_SIZE;
+	}
+	bio_get(comp_bio);
+
+	ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0);
+	BUG_ON(ret);
+
+	if (!btrfs_test_flag(inode, NODATASUM))
+		btrfs_lookup_bio_sums(root, inode, comp_bio, sums);
+
+	ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0);
+	BUG_ON(ret);
+
+	bio_put(comp_bio);
+	return 0;
+}
diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h
new file mode 100644
index 00000000000..421f5b4aa71
--- /dev/null
+++ b/fs/btrfs/compression.h
@@ -0,0 +1,47 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_COMPRESSION_
+#define __BTRFS_COMPRESSION_
+
+int btrfs_zlib_decompress(unsigned char *data_in,
+			  struct page *dest_page,
+			  unsigned long start_byte,
+			  size_t srclen, size_t destlen);
+int btrfs_zlib_compress_pages(struct address_space *mapping,
+			      u64 start, unsigned long len,
+			      struct page **pages,
+			      unsigned long nr_dest_pages,
+			      unsigned long *out_pages,
+			      unsigned long *total_in,
+			      unsigned long *total_out,
+			      unsigned long max_out);
+int btrfs_zlib_decompress_biovec(struct page **pages_in,
+			      u64 disk_start,
+			      struct bio_vec *bvec,
+			      int vcnt,
+			      size_t srclen);
+void btrfs_zlib_exit(void);
+int btrfs_submit_compressed_write(struct inode *inode, u64 start,
+				  unsigned long len, u64 disk_start,
+				  unsigned long compressed_len,
+				  struct page **compressed_pages,
+				  unsigned long nr_pages);
+int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
+				 int mirror_num, unsigned long bio_flags);
+#endif
diff --git a/fs/btrfs/crc32c.h b/fs/btrfs/crc32c.h
new file mode 100644
index 00000000000..6e1b3de3670
--- /dev/null
+++ b/fs/btrfs/crc32c.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_CRC32C__
+#define __BTRFS_CRC32C__
+#include <linux/crc32c.h>
+
+/*
+ * this file used to do more for selecting the HW version of crc32c,
+ * perhaps it will one day again soon.
+ */
+#define btrfs_crc32c(seed, data, length) crc32c(seed, data, length)
+#endif
+
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
new file mode 100644
index 00000000000..9e46c077681
--- /dev/null
+++ b/fs/btrfs/ctree.c
@@ -0,0 +1,3953 @@
+/*
+ * Copyright (C) 2007,2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "print-tree.h"
+#include "locking.h"
+
+static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_path *path, int level);
+static int split_leaf(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *ins_key,
+		      struct btrfs_path *path, int data_size, int extend);
+static int push_node_left(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct extent_buffer *dst,
+			  struct extent_buffer *src, int empty);
+static int balance_node_right(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct extent_buffer *dst_buf,
+			      struct extent_buffer *src_buf);
+static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		   struct btrfs_path *path, int level, int slot);
+
+inline void btrfs_init_path(struct btrfs_path *p)
+{
+	memset(p, 0, sizeof(*p));
+}
+
+struct btrfs_path *btrfs_alloc_path(void)
+{
+	struct btrfs_path *path;
+	path = kmem_cache_alloc(btrfs_path_cachep, GFP_NOFS);
+	if (path) {
+		btrfs_init_path(path);
+		path->reada = 1;
+	}
+	return path;
+}
+
+/* this also releases the path */
+void btrfs_free_path(struct btrfs_path *p)
+{
+	btrfs_release_path(NULL, p);
+	kmem_cache_free(btrfs_path_cachep, p);
+}
+
+/*
+ * path release drops references on the extent buffers in the path
+ * and it drops any locks held by this path
+ *
+ * It is safe to call this on paths that no locks or extent buffers held.
+ */
+noinline void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
+{
+	int i;
+
+	for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
+		p->slots[i] = 0;
+		if (!p->nodes[i])
+			continue;
+		if (p->locks[i]) {
+			btrfs_tree_unlock(p->nodes[i]);
+			p->locks[i] = 0;
+		}
+		free_extent_buffer(p->nodes[i]);
+		p->nodes[i] = NULL;
+	}
+}
+
+/*
+ * safely gets a reference on the root node of a tree.  A lock
+ * is not taken, so a concurrent writer may put a different node
+ * at the root of the tree.  See btrfs_lock_root_node for the
+ * looping required.
+ *
+ * The extent buffer returned by this has a reference taken, so
+ * it won't disappear.  It may stop being the root of the tree
+ * at any time because there are no locks held.
+ */
+struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
+{
+	struct extent_buffer *eb;
+	spin_lock(&root->node_lock);
+	eb = root->node;
+	extent_buffer_get(eb);
+	spin_unlock(&root->node_lock);
+	return eb;
+}
+
+/* loop around taking references on and locking the root node of the
+ * tree until you end up with a lock on the root.  A locked buffer
+ * is returned, with a reference held.
+ */
+struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root)
+{
+	struct extent_buffer *eb;
+
+	while (1) {
+		eb = btrfs_root_node(root);
+		btrfs_tree_lock(eb);
+
+		spin_lock(&root->node_lock);
+		if (eb == root->node) {
+			spin_unlock(&root->node_lock);
+			break;
+		}
+		spin_unlock(&root->node_lock);
+
+		btrfs_tree_unlock(eb);
+		free_extent_buffer(eb);
+	}
+	return eb;
+}
+
+/* cowonly root (everything not a reference counted cow subvolume), just get
+ * put onto a simple dirty list.  transaction.c walks this to make sure they
+ * get properly updated on disk.
+ */
+static void add_root_to_dirty_list(struct btrfs_root *root)
+{
+	if (root->track_dirty && list_empty(&root->dirty_list)) {
+		list_add(&root->dirty_list,
+			 &root->fs_info->dirty_cowonly_roots);
+	}
+}
+
+/*
+ * used by snapshot creation to make a copy of a root for a tree with
+ * a given objectid.  The buffer with the new root node is returned in
+ * cow_ret, and this func returns zero on success or a negative error code.
+ */
+int btrfs_copy_root(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root,
+		      struct extent_buffer *buf,
+		      struct extent_buffer **cow_ret, u64 new_root_objectid)
+{
+	struct extent_buffer *cow;
+	u32 nritems;
+	int ret = 0;
+	int level;
+	struct btrfs_root *new_root;
+
+	new_root = kmalloc(sizeof(*new_root), GFP_NOFS);
+	if (!new_root)
+		return -ENOMEM;
+
+	memcpy(new_root, root, sizeof(*new_root));
+	new_root->root_key.objectid = new_root_objectid;
+
+	WARN_ON(root->ref_cows && trans->transid !=
+		root->fs_info->running_transaction->transid);
+	WARN_ON(root->ref_cows && trans->transid != root->last_trans);
+
+	level = btrfs_header_level(buf);
+	nritems = btrfs_header_nritems(buf);
+
+	cow = btrfs_alloc_free_block(trans, new_root, buf->len, 0,
+				     new_root_objectid, trans->transid,
+				     level, buf->start, 0);
+	if (IS_ERR(cow)) {
+		kfree(new_root);
+		return PTR_ERR(cow);
+	}
+
+	copy_extent_buffer(cow, buf, 0, 0, cow->len);
+	btrfs_set_header_bytenr(cow, cow->start);
+	btrfs_set_header_generation(cow, trans->transid);
+	btrfs_set_header_owner(cow, new_root_objectid);
+	btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN);
+
+	write_extent_buffer(cow, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(cow),
+			    BTRFS_FSID_SIZE);
+
+	WARN_ON(btrfs_header_generation(buf) > trans->transid);
+	ret = btrfs_inc_ref(trans, new_root, buf, cow, NULL);
+	kfree(new_root);
+
+	if (ret)
+		return ret;
+
+	btrfs_mark_buffer_dirty(cow);
+	*cow_ret = cow;
+	return 0;
+}
+
+/*
+ * does the dirty work in cow of a single block.  The parent block (if
+ * supplied) is updated to point to the new cow copy.  The new buffer is marked
+ * dirty and returned locked.  If you modify the block it needs to be marked
+ * dirty again.
+ *
+ * search_start -- an allocation hint for the new block
+ *
+ * empty_size -- a hint that you plan on doing more cow.  This is the size in
+ * bytes the allocator should try to find free next to the block it returns.
+ * This is just a hint and may be ignored by the allocator.
+ *
+ * prealloc_dest -- if you have already reserved a destination for the cow,
+ * this uses that block instead of allocating a new one.
+ * btrfs_alloc_reserved_extent is used to finish the allocation.
+ */
+static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct extent_buffer *buf,
+			     struct extent_buffer *parent, int parent_slot,
+			     struct extent_buffer **cow_ret,
+			     u64 search_start, u64 empty_size,
+			     u64 prealloc_dest)
+{
+	u64 parent_start;
+	struct extent_buffer *cow;
+	u32 nritems;
+	int ret = 0;
+	int level;
+	int unlock_orig = 0;
+
+	if (*cow_ret == buf)
+		unlock_orig = 1;
+
+	WARN_ON(!btrfs_tree_locked(buf));
+
+	if (parent)
+		parent_start = parent->start;
+	else
+		parent_start = 0;
+
+	WARN_ON(root->ref_cows && trans->transid !=
+		root->fs_info->running_transaction->transid);
+	WARN_ON(root->ref_cows && trans->transid != root->last_trans);
+
+	level = btrfs_header_level(buf);
+	nritems = btrfs_header_nritems(buf);
+
+	if (prealloc_dest) {
+		struct btrfs_key ins;
+
+		ins.objectid = prealloc_dest;
+		ins.offset = buf->len;
+		ins.type = BTRFS_EXTENT_ITEM_KEY;
+
+		ret = btrfs_alloc_reserved_extent(trans, root, parent_start,
+						  root->root_key.objectid,
+						  trans->transid, level, &ins);
+		BUG_ON(ret);
+		cow = btrfs_init_new_buffer(trans, root, prealloc_dest,
+					    buf->len);
+	} else {
+		cow = btrfs_alloc_free_block(trans, root, buf->len,
+					     parent_start,
+					     root->root_key.objectid,
+					     trans->transid, level,
+					     search_start, empty_size);
+	}
+	if (IS_ERR(cow))
+		return PTR_ERR(cow);
+
+	copy_extent_buffer(cow, buf, 0, 0, cow->len);
+	btrfs_set_header_bytenr(cow, cow->start);
+	btrfs_set_header_generation(cow, trans->transid);
+	btrfs_set_header_owner(cow, root->root_key.objectid);
+	btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN);
+
+	write_extent_buffer(cow, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(cow),
+			    BTRFS_FSID_SIZE);
+
+	WARN_ON(btrfs_header_generation(buf) > trans->transid);
+	if (btrfs_header_generation(buf) != trans->transid) {
+		u32 nr_extents;
+		ret = btrfs_inc_ref(trans, root, buf, cow, &nr_extents);
+		if (ret)
+			return ret;
+
+		ret = btrfs_cache_ref(trans, root, buf, nr_extents);
+		WARN_ON(ret);
+	} else if (btrfs_header_owner(buf) == BTRFS_TREE_RELOC_OBJECTID) {
+		/*
+		 * There are only two places that can drop reference to
+		 * tree blocks owned by living reloc trees, one is here,
+		 * the other place is btrfs_drop_subtree. In both places,
+		 * we check reference count while tree block is locked.
+		 * Furthermore, if reference count is one, it won't get
+		 * increased by someone else.
+		 */
+		u32 refs;
+		ret = btrfs_lookup_extent_ref(trans, root, buf->start,
+					      buf->len, &refs);
+		BUG_ON(ret);
+		if (refs == 1) {
+			ret = btrfs_update_ref(trans, root, buf, cow,
+					       0, nritems);
+			clean_tree_block(trans, root, buf);
+		} else {
+			ret = btrfs_inc_ref(trans, root, buf, cow, NULL);
+		}
+		BUG_ON(ret);
+	} else {
+		ret = btrfs_update_ref(trans, root, buf, cow, 0, nritems);
+		if (ret)
+			return ret;
+		clean_tree_block(trans, root, buf);
+	}
+
+	if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
+		ret = btrfs_reloc_tree_cache_ref(trans, root, cow, buf->start);
+		WARN_ON(ret);
+	}
+
+	if (buf == root->node) {
+		WARN_ON(parent && parent != buf);
+
+		spin_lock(&root->node_lock);
+		root->node = cow;
+		extent_buffer_get(cow);
+		spin_unlock(&root->node_lock);
+
+		if (buf != root->commit_root) {
+			btrfs_free_extent(trans, root, buf->start,
+					  buf->len, buf->start,
+					  root->root_key.objectid,
+					  btrfs_header_generation(buf),
+					  level, 1);
+		}
+		free_extent_buffer(buf);
+		add_root_to_dirty_list(root);
+	} else {
+		btrfs_set_node_blockptr(parent, parent_slot,
+					cow->start);
+		WARN_ON(trans->transid == 0);
+		btrfs_set_node_ptr_generation(parent, parent_slot,
+					      trans->transid);
+		btrfs_mark_buffer_dirty(parent);
+		WARN_ON(btrfs_header_generation(parent) != trans->transid);
+		btrfs_free_extent(trans, root, buf->start, buf->len,
+				  parent_start, btrfs_header_owner(parent),
+				  btrfs_header_generation(parent), level, 1);
+	}
+	if (unlock_orig)
+		btrfs_tree_unlock(buf);
+	free_extent_buffer(buf);
+	btrfs_mark_buffer_dirty(cow);
+	*cow_ret = cow;
+	return 0;
+}
+
+/*
+ * cows a single block, see __btrfs_cow_block for the real work.
+ * This version of it has extra checks so that a block isn't cow'd more than
+ * once per transaction, as long as it hasn't been written yet
+ */
+noinline int btrfs_cow_block(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, struct extent_buffer *buf,
+		    struct extent_buffer *parent, int parent_slot,
+		    struct extent_buffer **cow_ret, u64 prealloc_dest)
+{
+	u64 search_start;
+	int ret;
+
+	if (trans->transaction != root->fs_info->running_transaction) {
+		printk(KERN_CRIT "trans %llu running %llu\n",
+		       (unsigned long long)trans->transid,
+		       (unsigned long long)
+		       root->fs_info->running_transaction->transid);
+		WARN_ON(1);
+	}
+	if (trans->transid != root->fs_info->generation) {
+		printk(KERN_CRIT "trans %llu running %llu\n",
+		       (unsigned long long)trans->transid,
+		       (unsigned long long)root->fs_info->generation);
+		WARN_ON(1);
+	}
+
+	spin_lock(&root->fs_info->hash_lock);
+	if (btrfs_header_generation(buf) == trans->transid &&
+	    btrfs_header_owner(buf) == root->root_key.objectid &&
+	    !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
+		*cow_ret = buf;
+		spin_unlock(&root->fs_info->hash_lock);
+		WARN_ON(prealloc_dest);
+		return 0;
+	}
+	spin_unlock(&root->fs_info->hash_lock);
+	search_start = buf->start & ~((u64)(1024 * 1024 * 1024) - 1);
+	ret = __btrfs_cow_block(trans, root, buf, parent,
+				 parent_slot, cow_ret, search_start, 0,
+				 prealloc_dest);
+	return ret;
+}
+
+/*
+ * helper function for defrag to decide if two blocks pointed to by a
+ * node are actually close by
+ */
+static int close_blocks(u64 blocknr, u64 other, u32 blocksize)
+{
+	if (blocknr < other && other - (blocknr + blocksize) < 32768)
+		return 1;
+	if (blocknr > other && blocknr - (other + blocksize) < 32768)
+		return 1;
+	return 0;
+}
+
+/*
+ * compare two keys in a memcmp fashion
+ */
+static int comp_keys(struct btrfs_disk_key *disk, struct btrfs_key *k2)
+{
+	struct btrfs_key k1;
+
+	btrfs_disk_key_to_cpu(&k1, disk);
+
+	if (k1.objectid > k2->objectid)
+		return 1;
+	if (k1.objectid < k2->objectid)
+		return -1;
+	if (k1.type > k2->type)
+		return 1;
+	if (k1.type < k2->type)
+		return -1;
+	if (k1.offset > k2->offset)
+		return 1;
+	if (k1.offset < k2->offset)
+		return -1;
+	return 0;
+}
+
+/*
+ * same as comp_keys only with two btrfs_key's
+ */
+static int comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2)
+{
+	if (k1->objectid > k2->objectid)
+		return 1;
+	if (k1->objectid < k2->objectid)
+		return -1;
+	if (k1->type > k2->type)
+		return 1;
+	if (k1->type < k2->type)
+		return -1;
+	if (k1->offset > k2->offset)
+		return 1;
+	if (k1->offset < k2->offset)
+		return -1;
+	return 0;
+}
+
+/*
+ * this is used by the defrag code to go through all the
+ * leaves pointed to by a node and reallocate them so that
+ * disk order is close to key order
+ */
+int btrfs_realloc_node(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct extent_buffer *parent,
+		       int start_slot, int cache_only, u64 *last_ret,
+		       struct btrfs_key *progress)
+{
+	struct extent_buffer *cur;
+	u64 blocknr;
+	u64 gen;
+	u64 search_start = *last_ret;
+	u64 last_block = 0;
+	u64 other;
+	u32 parent_nritems;
+	int end_slot;
+	int i;
+	int err = 0;
+	int parent_level;
+	int uptodate;
+	u32 blocksize;
+	int progress_passed = 0;
+	struct btrfs_disk_key disk_key;
+
+	parent_level = btrfs_header_level(parent);
+	if (cache_only && parent_level != 1)
+		return 0;
+
+	if (trans->transaction != root->fs_info->running_transaction)
+		WARN_ON(1);
+	if (trans->transid != root->fs_info->generation)
+		WARN_ON(1);
+
+	parent_nritems = btrfs_header_nritems(parent);
+	blocksize = btrfs_level_size(root, parent_level - 1);
+	end_slot = parent_nritems;
+
+	if (parent_nritems == 1)
+		return 0;
+
+	for (i = start_slot; i < end_slot; i++) {
+		int close = 1;
+
+		if (!parent->map_token) {
+			map_extent_buffer(parent,
+					btrfs_node_key_ptr_offset(i),
+					sizeof(struct btrfs_key_ptr),
+					&parent->map_token, &parent->kaddr,
+					&parent->map_start, &parent->map_len,
+					KM_USER1);
+		}
+		btrfs_node_key(parent, &disk_key, i);
+		if (!progress_passed && comp_keys(&disk_key, progress) < 0)
+			continue;
+
+		progress_passed = 1;
+		blocknr = btrfs_node_blockptr(parent, i);
+		gen = btrfs_node_ptr_generation(parent, i);
+		if (last_block == 0)
+			last_block = blocknr;
+
+		if (i > 0) {
+			other = btrfs_node_blockptr(parent, i - 1);
+			close = close_blocks(blocknr, other, blocksize);
+		}
+		if (!close && i < end_slot - 2) {
+			other = btrfs_node_blockptr(parent, i + 1);
+			close = close_blocks(blocknr, other, blocksize);
+		}
+		if (close) {
+			last_block = blocknr;
+			continue;
+		}
+		if (parent->map_token) {
+			unmap_extent_buffer(parent, parent->map_token,
+					    KM_USER1);
+			parent->map_token = NULL;
+		}
+
+		cur = btrfs_find_tree_block(root, blocknr, blocksize);
+		if (cur)
+			uptodate = btrfs_buffer_uptodate(cur, gen);
+		else
+			uptodate = 0;
+		if (!cur || !uptodate) {
+			if (cache_only) {
+				free_extent_buffer(cur);
+				continue;
+			}
+			if (!cur) {
+				cur = read_tree_block(root, blocknr,
+							 blocksize, gen);
+			} else if (!uptodate) {
+				btrfs_read_buffer(cur, gen);
+			}
+		}
+		if (search_start == 0)
+			search_start = last_block;
+
+		btrfs_tree_lock(cur);
+		err = __btrfs_cow_block(trans, root, cur, parent, i,
+					&cur, search_start,
+					min(16 * blocksize,
+					    (end_slot - i) * blocksize), 0);
+		if (err) {
+			btrfs_tree_unlock(cur);
+			free_extent_buffer(cur);
+			break;
+		}
+		search_start = cur->start;
+		last_block = cur->start;
+		*last_ret = search_start;
+		btrfs_tree_unlock(cur);
+		free_extent_buffer(cur);
+	}
+	if (parent->map_token) {
+		unmap_extent_buffer(parent, parent->map_token,
+				    KM_USER1);
+		parent->map_token = NULL;
+	}
+	return err;
+}
+
+/*
+ * The leaf data grows from end-to-front in the node.
+ * this returns the address of the start of the last item,
+ * which is the stop of the leaf data stack
+ */
+static inline unsigned int leaf_data_end(struct btrfs_root *root,
+					 struct extent_buffer *leaf)
+{
+	u32 nr = btrfs_header_nritems(leaf);
+	if (nr == 0)
+		return BTRFS_LEAF_DATA_SIZE(root);
+	return btrfs_item_offset_nr(leaf, nr - 1);
+}
+
+/*
+ * extra debugging checks to make sure all the items in a key are
+ * well formed and in the proper order
+ */
+static int check_node(struct btrfs_root *root, struct btrfs_path *path,
+		      int level)
+{
+	struct extent_buffer *parent = NULL;
+	struct extent_buffer *node = path->nodes[level];
+	struct btrfs_disk_key parent_key;
+	struct btrfs_disk_key node_key;
+	int parent_slot;
+	int slot;
+	struct btrfs_key cpukey;
+	u32 nritems = btrfs_header_nritems(node);
+
+	if (path->nodes[level + 1])
+		parent = path->nodes[level + 1];
+
+	slot = path->slots[level];
+	BUG_ON(nritems == 0);
+	if (parent) {
+		parent_slot = path->slots[level + 1];
+		btrfs_node_key(parent, &parent_key, parent_slot);
+		btrfs_node_key(node, &node_key, 0);
+		BUG_ON(memcmp(&parent_key, &node_key,
+			      sizeof(struct btrfs_disk_key)));
+		BUG_ON(btrfs_node_blockptr(parent, parent_slot) !=
+		       btrfs_header_bytenr(node));
+	}
+	BUG_ON(nritems > BTRFS_NODEPTRS_PER_BLOCK(root));
+	if (slot != 0) {
+		btrfs_node_key_to_cpu(node, &cpukey, slot - 1);
+		btrfs_node_key(node, &node_key, slot);
+		BUG_ON(comp_keys(&node_key, &cpukey) <= 0);
+	}
+	if (slot < nritems - 1) {
+		btrfs_node_key_to_cpu(node, &cpukey, slot + 1);
+		btrfs_node_key(node, &node_key, slot);
+		BUG_ON(comp_keys(&node_key, &cpukey) >= 0);
+	}
+	return 0;
+}
+
+/*
+ * extra checking to make sure all the items in a leaf are
+ * well formed and in the proper order
+ */
+static int check_leaf(struct btrfs_root *root, struct btrfs_path *path,
+		      int level)
+{
+	struct extent_buffer *leaf = path->nodes[level];
+	struct extent_buffer *parent = NULL;
+	int parent_slot;
+	struct btrfs_key cpukey;
+	struct btrfs_disk_key parent_key;
+	struct btrfs_disk_key leaf_key;
+	int slot = path->slots[0];
+
+	u32 nritems = btrfs_header_nritems(leaf);
+
+	if (path->nodes[level + 1])
+		parent = path->nodes[level + 1];
+
+	if (nritems == 0)
+		return 0;
+
+	if (parent) {
+		parent_slot = path->slots[level + 1];
+		btrfs_node_key(parent, &parent_key, parent_slot);
+		btrfs_item_key(leaf, &leaf_key, 0);
+
+		BUG_ON(memcmp(&parent_key, &leaf_key,
+		       sizeof(struct btrfs_disk_key)));
+		BUG_ON(btrfs_node_blockptr(parent, parent_slot) !=
+		       btrfs_header_bytenr(leaf));
+	}
+	if (slot != 0 && slot < nritems - 1) {
+		btrfs_item_key(leaf, &leaf_key, slot);
+		btrfs_item_key_to_cpu(leaf, &cpukey, slot - 1);
+		if (comp_keys(&leaf_key, &cpukey) <= 0) {
+			btrfs_print_leaf(root, leaf);
+			printk(KERN_CRIT "slot %d offset bad key\n", slot);
+			BUG_ON(1);
+		}
+		if (btrfs_item_offset_nr(leaf, slot - 1) !=
+		       btrfs_item_end_nr(leaf, slot)) {
+			btrfs_print_leaf(root, leaf);
+			printk(KERN_CRIT "slot %d offset bad\n", slot);
+			BUG_ON(1);
+		}
+	}
+	if (slot < nritems - 1) {
+		btrfs_item_key(leaf, &leaf_key, slot);
+		btrfs_item_key_to_cpu(leaf, &cpukey, slot + 1);
+		BUG_ON(comp_keys(&leaf_key, &cpukey) >= 0);
+		if (btrfs_item_offset_nr(leaf, slot) !=
+			btrfs_item_end_nr(leaf, slot + 1)) {
+			btrfs_print_leaf(root, leaf);
+			printk(KERN_CRIT "slot %d offset bad\n", slot);
+			BUG_ON(1);
+		}
+	}
+	BUG_ON(btrfs_item_offset_nr(leaf, 0) +
+	       btrfs_item_size_nr(leaf, 0) != BTRFS_LEAF_DATA_SIZE(root));
+	return 0;
+}
+
+static noinline int check_block(struct btrfs_root *root,
+				struct btrfs_path *path, int level)
+{
+	return 0;
+	if (level == 0)
+		return check_leaf(root, path, level);
+	return check_node(root, path, level);
+}
+
+/*
+ * search for key in the extent_buffer.  The items start at offset p,
+ * and they are item_size apart.  There are 'max' items in p.
+ *
+ * the slot in the array is returned via slot, and it points to
+ * the place where you would insert key if it is not found in
+ * the array.
+ *
+ * slot may point to max if the key is bigger than all of the keys
+ */
+static noinline int generic_bin_search(struct extent_buffer *eb,
+				       unsigned long p,
+				       int item_size, struct btrfs_key *key,
+				       int max, int *slot)
+{
+	int low = 0;
+	int high = max;
+	int mid;
+	int ret;
+	struct btrfs_disk_key *tmp = NULL;
+	struct btrfs_disk_key unaligned;
+	unsigned long offset;
+	char *map_token = NULL;
+	char *kaddr = NULL;
+	unsigned long map_start = 0;
+	unsigned long map_len = 0;
+	int err;
+
+	while (low < high) {
+		mid = (low + high) / 2;
+		offset = p + mid * item_size;
+
+		if (!map_token || offset < map_start ||
+		    (offset + sizeof(struct btrfs_disk_key)) >
+		    map_start + map_len) {
+			if (map_token) {
+				unmap_extent_buffer(eb, map_token, KM_USER0);
+				map_token = NULL;
+			}
+
+			err = map_private_extent_buffer(eb, offset,
+						sizeof(struct btrfs_disk_key),
+						&map_token, &kaddr,
+						&map_start, &map_len, KM_USER0);
+
+			if (!err) {
+				tmp = (struct btrfs_disk_key *)(kaddr + offset -
+							map_start);
+			} else {
+				read_extent_buffer(eb, &unaligned,
+						   offset, sizeof(unaligned));
+				tmp = &unaligned;
+			}
+
+		} else {
+			tmp = (struct btrfs_disk_key *)(kaddr + offset -
+							map_start);
+		}
+		ret = comp_keys(tmp, key);
+
+		if (ret < 0)
+			low = mid + 1;
+		else if (ret > 0)
+			high = mid;
+		else {
+			*slot = mid;
+			if (map_token)
+				unmap_extent_buffer(eb, map_token, KM_USER0);
+			return 0;
+		}
+	}
+	*slot = low;
+	if (map_token)
+		unmap_extent_buffer(eb, map_token, KM_USER0);
+	return 1;
+}
+
+/*
+ * simple bin_search frontend that does the right thing for
+ * leaves vs nodes
+ */
+static int bin_search(struct extent_buffer *eb, struct btrfs_key *key,
+		      int level, int *slot)
+{
+	if (level == 0) {
+		return generic_bin_search(eb,
+					  offsetof(struct btrfs_leaf, items),
+					  sizeof(struct btrfs_item),
+					  key, btrfs_header_nritems(eb),
+					  slot);
+	} else {
+		return generic_bin_search(eb,
+					  offsetof(struct btrfs_node, ptrs),
+					  sizeof(struct btrfs_key_ptr),
+					  key, btrfs_header_nritems(eb),
+					  slot);
+	}
+	return -1;
+}
+
+/* given a node and slot number, this reads the blocks it points to.  The
+ * extent buffer is returned with a reference taken (but unlocked).
+ * NULL is returned on error.
+ */
+static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
+				   struct extent_buffer *parent, int slot)
+{
+	int level = btrfs_header_level(parent);
+	if (slot < 0)
+		return NULL;
+	if (slot >= btrfs_header_nritems(parent))
+		return NULL;
+
+	BUG_ON(level == 0);
+
+	return read_tree_block(root, btrfs_node_blockptr(parent, slot),
+		       btrfs_level_size(root, level - 1),
+		       btrfs_node_ptr_generation(parent, slot));
+}
+
+/*
+ * node level balancing, used to make sure nodes are in proper order for
+ * item deletion.  We balance from the top down, so we have to make sure
+ * that a deletion won't leave an node completely empty later on.
+ */
+static noinline int balance_level(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root,
+			 struct btrfs_path *path, int level)
+{
+	struct extent_buffer *right = NULL;
+	struct extent_buffer *mid;
+	struct extent_buffer *left = NULL;
+	struct extent_buffer *parent = NULL;
+	int ret = 0;
+	int wret;
+	int pslot;
+	int orig_slot = path->slots[level];
+	int err_on_enospc = 0;
+	u64 orig_ptr;
+
+	if (level == 0)
+		return 0;
+
+	mid = path->nodes[level];
+	WARN_ON(!path->locks[level]);
+	WARN_ON(btrfs_header_generation(mid) != trans->transid);
+
+	orig_ptr = btrfs_node_blockptr(mid, orig_slot);
+
+	if (level < BTRFS_MAX_LEVEL - 1)
+		parent = path->nodes[level + 1];
+	pslot = path->slots[level + 1];
+
+	/*
+	 * deal with the case where there is only one pointer in the root
+	 * by promoting the node below to a root
+	 */
+	if (!parent) {
+		struct extent_buffer *child;
+
+		if (btrfs_header_nritems(mid) != 1)
+			return 0;
+
+		/* promote the child to a root */
+		child = read_node_slot(root, mid, 0);
+		btrfs_tree_lock(child);
+		BUG_ON(!child);
+		ret = btrfs_cow_block(trans, root, child, mid, 0, &child, 0);
+		BUG_ON(ret);
+
+		spin_lock(&root->node_lock);
+		root->node = child;
+		spin_unlock(&root->node_lock);
+
+		ret = btrfs_update_extent_ref(trans, root, child->start,
+					      mid->start, child->start,
+					      root->root_key.objectid,
+					      trans->transid, level - 1);
+		BUG_ON(ret);
+
+		add_root_to_dirty_list(root);
+		btrfs_tree_unlock(child);
+		path->locks[level] = 0;
+		path->nodes[level] = NULL;
+		clean_tree_block(trans, root, mid);
+		btrfs_tree_unlock(mid);
+		/* once for the path */
+		free_extent_buffer(mid);
+		ret = btrfs_free_extent(trans, root, mid->start, mid->len,
+					mid->start, root->root_key.objectid,
+					btrfs_header_generation(mid),
+					level, 1);
+		/* once for the root ptr */
+		free_extent_buffer(mid);
+		return ret;
+	}
+	if (btrfs_header_nritems(mid) >
+	    BTRFS_NODEPTRS_PER_BLOCK(root) / 4)
+		return 0;
+
+	if (btrfs_header_nritems(mid) < 2)
+		err_on_enospc = 1;
+
+	left = read_node_slot(root, parent, pslot - 1);
+	if (left) {
+		btrfs_tree_lock(left);
+		wret = btrfs_cow_block(trans, root, left,
+				       parent, pslot - 1, &left, 0);
+		if (wret) {
+			ret = wret;
+			goto enospc;
+		}
+	}
+	right = read_node_slot(root, parent, pslot + 1);
+	if (right) {
+		btrfs_tree_lock(right);
+		wret = btrfs_cow_block(trans, root, right,
+				       parent, pslot + 1, &right, 0);
+		if (wret) {
+			ret = wret;
+			goto enospc;
+		}
+	}
+
+	/* first, try to make some room in the middle buffer */
+	if (left) {
+		orig_slot += btrfs_header_nritems(left);
+		wret = push_node_left(trans, root, left, mid, 1);
+		if (wret < 0)
+			ret = wret;
+		if (btrfs_header_nritems(mid) < 2)
+			err_on_enospc = 1;
+	}
+
+	/*
+	 * then try to empty the right most buffer into the middle
+	 */
+	if (right) {
+		wret = push_node_left(trans, root, mid, right, 1);
+		if (wret < 0 && wret != -ENOSPC)
+			ret = wret;
+		if (btrfs_header_nritems(right) == 0) {
+			u64 bytenr = right->start;
+			u64 generation = btrfs_header_generation(parent);
+			u32 blocksize = right->len;
+
+			clean_tree_block(trans, root, right);
+			btrfs_tree_unlock(right);
+			free_extent_buffer(right);
+			right = NULL;
+			wret = del_ptr(trans, root, path, level + 1, pslot +
+				       1);
+			if (wret)
+				ret = wret;
+			wret = btrfs_free_extent(trans, root, bytenr,
+						 blocksize, parent->start,
+						 btrfs_header_owner(parent),
+						 generation, level, 1);
+			if (wret)
+				ret = wret;
+		} else {
+			struct btrfs_disk_key right_key;
+			btrfs_node_key(right, &right_key, 0);
+			btrfs_set_node_key(parent, &right_key, pslot + 1);
+			btrfs_mark_buffer_dirty(parent);
+		}
+	}
+	if (btrfs_header_nritems(mid) == 1) {
+		/*
+		 * we're not allowed to leave a node with one item in the
+		 * tree during a delete.  A deletion from lower in the tree
+		 * could try to delete the only pointer in this node.
+		 * So, pull some keys from the left.
+		 * There has to be a left pointer at this point because
+		 * otherwise we would have pulled some pointers from the
+		 * right
+		 */
+		BUG_ON(!left);
+		wret = balance_node_right(trans, root, mid, left);
+		if (wret < 0) {
+			ret = wret;
+			goto enospc;
+		}
+		if (wret == 1) {
+			wret = push_node_left(trans, root, left, mid, 1);
+			if (wret < 0)
+				ret = wret;
+		}
+		BUG_ON(wret == 1);
+	}
+	if (btrfs_header_nritems(mid) == 0) {
+		/* we've managed to empty the middle node, drop it */
+		u64 root_gen = btrfs_header_generation(parent);
+		u64 bytenr = mid->start;
+		u32 blocksize = mid->len;
+
+		clean_tree_block(trans, root, mid);
+		btrfs_tree_unlock(mid);
+		free_extent_buffer(mid);
+		mid = NULL;
+		wret = del_ptr(trans, root, path, level + 1, pslot);
+		if (wret)
+			ret = wret;
+		wret = btrfs_free_extent(trans, root, bytenr, blocksize,
+					 parent->start,
+					 btrfs_header_owner(parent),
+					 root_gen, level, 1);
+		if (wret)
+			ret = wret;
+	} else {
+		/* update the parent key to reflect our changes */
+		struct btrfs_disk_key mid_key;
+		btrfs_node_key(mid, &mid_key, 0);
+		btrfs_set_node_key(parent, &mid_key, pslot);
+		btrfs_mark_buffer_dirty(parent);
+	}
+
+	/* update the path */
+	if (left) {
+		if (btrfs_header_nritems(left) > orig_slot) {
+			extent_buffer_get(left);
+			/* left was locked after cow */
+			path->nodes[level] = left;
+			path->slots[level + 1] -= 1;
+			path->slots[level] = orig_slot;
+			if (mid) {
+				btrfs_tree_unlock(mid);
+				free_extent_buffer(mid);
+			}
+		} else {
+			orig_slot -= btrfs_header_nritems(left);
+			path->slots[level] = orig_slot;
+		}
+	}
+	/* double check we haven't messed things up */
+	check_block(root, path, level);
+	if (orig_ptr !=
+	    btrfs_node_blockptr(path->nodes[level], path->slots[level]))
+		BUG();
+enospc:
+	if (right) {
+		btrfs_tree_unlock(right);
+		free_extent_buffer(right);
+	}
+	if (left) {
+		if (path->nodes[level] != left)
+			btrfs_tree_unlock(left);
+		free_extent_buffer(left);
+	}
+	return ret;
+}
+
+/* Node balancing for insertion.  Here we only split or push nodes around
+ * when they are completely full.  This is also done top down, so we
+ * have to be pessimistic.
+ */
+static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path, int level)
+{
+	struct extent_buffer *right = NULL;
+	struct extent_buffer *mid;
+	struct extent_buffer *left = NULL;
+	struct extent_buffer *parent = NULL;
+	int ret = 0;
+	int wret;
+	int pslot;
+	int orig_slot = path->slots[level];
+	u64 orig_ptr;
+
+	if (level == 0)
+		return 1;
+
+	mid = path->nodes[level];
+	WARN_ON(btrfs_header_generation(mid) != trans->transid);
+	orig_ptr = btrfs_node_blockptr(mid, orig_slot);
+
+	if (level < BTRFS_MAX_LEVEL - 1)
+		parent = path->nodes[level + 1];
+	pslot = path->slots[level + 1];
+
+	if (!parent)
+		return 1;
+
+	left = read_node_slot(root, parent, pslot - 1);
+
+	/* first, try to make some room in the middle buffer */
+	if (left) {
+		u32 left_nr;
+
+		btrfs_tree_lock(left);
+		left_nr = btrfs_header_nritems(left);
+		if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
+			wret = 1;
+		} else {
+			ret = btrfs_cow_block(trans, root, left, parent,
+					      pslot - 1, &left, 0);
+			if (ret)
+				wret = 1;
+			else {
+				wret = push_node_left(trans, root,
+						      left, mid, 0);
+			}
+		}
+		if (wret < 0)
+			ret = wret;
+		if (wret == 0) {
+			struct btrfs_disk_key disk_key;
+			orig_slot += left_nr;
+			btrfs_node_key(mid, &disk_key, 0);
+			btrfs_set_node_key(parent, &disk_key, pslot);
+			btrfs_mark_buffer_dirty(parent);
+			if (btrfs_header_nritems(left) > orig_slot) {
+				path->nodes[level] = left;
+				path->slots[level + 1] -= 1;
+				path->slots[level] = orig_slot;
+				btrfs_tree_unlock(mid);
+				free_extent_buffer(mid);
+			} else {
+				orig_slot -=
+					btrfs_header_nritems(left);
+				path->slots[level] = orig_slot;
+				btrfs_tree_unlock(left);
+				free_extent_buffer(left);
+			}
+			return 0;
+		}
+		btrfs_tree_unlock(left);
+		free_extent_buffer(left);
+	}
+	right = read_node_slot(root, parent, pslot + 1);
+
+	/*
+	 * then try to empty the right most buffer into the middle
+	 */
+	if (right) {
+		u32 right_nr;
+		btrfs_tree_lock(right);
+		right_nr = btrfs_header_nritems(right);
+		if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
+			wret = 1;
+		} else {
+			ret = btrfs_cow_block(trans, root, right,
+					      parent, pslot + 1,
+					      &right, 0);
+			if (ret)
+				wret = 1;
+			else {
+				wret = balance_node_right(trans, root,
+							  right, mid);
+			}
+		}
+		if (wret < 0)
+			ret = wret;
+		if (wret == 0) {
+			struct btrfs_disk_key disk_key;
+
+			btrfs_node_key(right, &disk_key, 0);
+			btrfs_set_node_key(parent, &disk_key, pslot + 1);
+			btrfs_mark_buffer_dirty(parent);
+
+			if (btrfs_header_nritems(mid) <= orig_slot) {
+				path->nodes[level] = right;
+				path->slots[level + 1] += 1;
+				path->slots[level] = orig_slot -
+					btrfs_header_nritems(mid);
+				btrfs_tree_unlock(mid);
+				free_extent_buffer(mid);
+			} else {
+				btrfs_tree_unlock(right);
+				free_extent_buffer(right);
+			}
+			return 0;
+		}
+		btrfs_tree_unlock(right);
+		free_extent_buffer(right);
+	}
+	return 1;
+}
+
+/*
+ * readahead one full node of leaves, finding things that are close
+ * to the block in 'slot', and triggering ra on them.
+ */
+static noinline void reada_for_search(struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      int level, int slot, u64 objectid)
+{
+	struct extent_buffer *node;
+	struct btrfs_disk_key disk_key;
+	u32 nritems;
+	u64 search;
+	u64 lowest_read;
+	u64 highest_read;
+	u64 nread = 0;
+	int direction = path->reada;
+	struct extent_buffer *eb;
+	u32 nr;
+	u32 blocksize;
+	u32 nscan = 0;
+
+	if (level != 1)
+		return;
+
+	if (!path->nodes[level])
+		return;
+
+	node = path->nodes[level];
+
+	search = btrfs_node_blockptr(node, slot);
+	blocksize = btrfs_level_size(root, level - 1);
+	eb = btrfs_find_tree_block(root, search, blocksize);
+	if (eb) {
+		free_extent_buffer(eb);
+		return;
+	}
+
+	highest_read = search;
+	lowest_read = search;
+
+	nritems = btrfs_header_nritems(node);
+	nr = slot;
+	while (1) {
+		if (direction < 0) {
+			if (nr == 0)
+				break;
+			nr--;
+		} else if (direction > 0) {
+			nr++;
+			if (nr >= nritems)
+				break;
+		}
+		if (path->reada < 0 && objectid) {
+			btrfs_node_key(node, &disk_key, nr);
+			if (btrfs_disk_key_objectid(&disk_key) != objectid)
+				break;
+		}
+		search = btrfs_node_blockptr(node, nr);
+		if ((search >= lowest_read && search <= highest_read) ||
+		    (search < lowest_read && lowest_read - search <= 16384) ||
+		    (search > highest_read && search - highest_read <= 16384)) {
+			readahead_tree_block(root, search, blocksize,
+				     btrfs_node_ptr_generation(node, nr));
+			nread += blocksize;
+		}
+		nscan++;
+		if (path->reada < 2 && (nread > (64 * 1024) || nscan > 32))
+			break;
+
+		if (nread > (256 * 1024) || nscan > 128)
+			break;
+
+		if (search < lowest_read)
+			lowest_read = search;
+		if (search > highest_read)
+			highest_read = search;
+	}
+}
+
+/*
+ * when we walk down the tree, it is usually safe to unlock the higher layers
+ * in the tree.  The exceptions are when our path goes through slot 0, because
+ * operations on the tree might require changing key pointers higher up in the
+ * tree.
+ *
+ * callers might also have set path->keep_locks, which tells this code to keep
+ * the lock if the path points to the last slot in the block.  This is part of
+ * walking through the tree, and selecting the next slot in the higher block.
+ *
+ * lowest_unlock sets the lowest level in the tree we're allowed to unlock.  so
+ * if lowest_unlock is 1, level 0 won't be unlocked
+ */
+static noinline void unlock_up(struct btrfs_path *path, int level,
+			       int lowest_unlock)
+{
+	int i;
+	int skip_level = level;
+	int no_skips = 0;
+	struct extent_buffer *t;
+
+	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
+		if (!path->nodes[i])
+			break;
+		if (!path->locks[i])
+			break;
+		if (!no_skips && path->slots[i] == 0) {
+			skip_level = i + 1;
+			continue;
+		}
+		if (!no_skips && path->keep_locks) {
+			u32 nritems;
+			t = path->nodes[i];
+			nritems = btrfs_header_nritems(t);
+			if (nritems < 1 || path->slots[i] >= nritems - 1) {
+				skip_level = i + 1;
+				continue;
+			}
+		}
+		if (skip_level < i && i >= lowest_unlock)
+			no_skips = 1;
+
+		t = path->nodes[i];
+		if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
+			btrfs_tree_unlock(t);
+			path->locks[i] = 0;
+		}
+	}
+}
+
+/*
+ * look for key in the tree.  path is filled in with nodes along the way
+ * if key is found, we return zero and you can find the item in the leaf
+ * level of the path (level 0)
+ *
+ * If the key isn't found, the path points to the slot where it should
+ * be inserted, and 1 is returned.  If there are other errors during the
+ * search a negative error number is returned.
+ *
+ * if ins_len > 0, nodes and leaves will be split as we walk down the
+ * tree.  if ins_len < 0, nodes will be merged as we walk down the tree (if
+ * possible)
+ */
+int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, struct btrfs_path *p, int
+		      ins_len, int cow)
+{
+	struct extent_buffer *b;
+	struct extent_buffer *tmp;
+	int slot;
+	int ret;
+	int level;
+	int should_reada = p->reada;
+	int lowest_unlock = 1;
+	int blocksize;
+	u8 lowest_level = 0;
+	u64 blocknr;
+	u64 gen;
+	struct btrfs_key prealloc_block;
+
+	lowest_level = p->lowest_level;
+	WARN_ON(lowest_level && ins_len > 0);
+	WARN_ON(p->nodes[0] != NULL);
+
+	if (ins_len < 0)
+		lowest_unlock = 2;
+
+	prealloc_block.objectid = 0;
+
+again:
+	if (p->skip_locking)
+		b = btrfs_root_node(root);
+	else
+		b = btrfs_lock_root_node(root);
+
+	while (b) {
+		level = btrfs_header_level(b);
+
+		/*
+		 * setup the path here so we can release it under lock
+		 * contention with the cow code
+		 */
+		p->nodes[level] = b;
+		if (!p->skip_locking)
+			p->locks[level] = 1;
+
+		if (cow) {
+			int wret;
+
+			/* is a cow on this block not required */
+			spin_lock(&root->fs_info->hash_lock);
+			if (btrfs_header_generation(b) == trans->transid &&
+			    btrfs_header_owner(b) == root->root_key.objectid &&
+			    !btrfs_header_flag(b, BTRFS_HEADER_FLAG_WRITTEN)) {
+				spin_unlock(&root->fs_info->hash_lock);
+				goto cow_done;
+			}
+			spin_unlock(&root->fs_info->hash_lock);
+
+			/* ok, we have to cow, is our old prealloc the right
+			 * size?
+			 */
+			if (prealloc_block.objectid &&
+			    prealloc_block.offset != b->len) {
+				btrfs_free_reserved_extent(root,
+					   prealloc_block.objectid,
+					   prealloc_block.offset);
+				prealloc_block.objectid = 0;
+			}
+
+			/*
+			 * for higher level blocks, try not to allocate blocks
+			 * with the block and the parent locks held.
+			 */
+			if (level > 1 && !prealloc_block.objectid &&
+			    btrfs_path_lock_waiting(p, level)) {
+				u32 size = b->len;
+				u64 hint = b->start;
+
+				btrfs_release_path(root, p);
+				ret = btrfs_reserve_extent(trans, root,
+							   size, size, 0,
+							   hint, (u64)-1,
+							   &prealloc_block, 0);
+				BUG_ON(ret);
+				goto again;
+			}
+
+			wret = btrfs_cow_block(trans, root, b,
+					       p->nodes[level + 1],
+					       p->slots[level + 1],
+					       &b, prealloc_block.objectid);
+			prealloc_block.objectid = 0;
+			if (wret) {
+				free_extent_buffer(b);
+				ret = wret;
+				goto done;
+			}
+		}
+cow_done:
+		BUG_ON(!cow && ins_len);
+		if (level != btrfs_header_level(b))
+			WARN_ON(1);
+		level = btrfs_header_level(b);
+
+		p->nodes[level] = b;
+		if (!p->skip_locking)
+			p->locks[level] = 1;
+
+		ret = check_block(root, p, level);
+		if (ret) {
+			ret = -1;
+			goto done;
+		}
+
+		ret = bin_search(b, key, level, &slot);
+		if (level != 0) {
+			if (ret && slot > 0)
+				slot -= 1;
+			p->slots[level] = slot;
+			if ((p->search_for_split || ins_len > 0) &&
+			    btrfs_header_nritems(b) >=
+			    BTRFS_NODEPTRS_PER_BLOCK(root) - 3) {
+				int sret = split_node(trans, root, p, level);
+				BUG_ON(sret > 0);
+				if (sret) {
+					ret = sret;
+					goto done;
+				}
+				b = p->nodes[level];
+				slot = p->slots[level];
+			} else if (ins_len < 0) {
+				int sret = balance_level(trans, root, p,
+							 level);
+				if (sret) {
+					ret = sret;
+					goto done;
+				}
+				b = p->nodes[level];
+				if (!b) {
+					btrfs_release_path(NULL, p);
+					goto again;
+				}
+				slot = p->slots[level];
+				BUG_ON(btrfs_header_nritems(b) == 1);
+			}
+			unlock_up(p, level, lowest_unlock);
+
+			/* this is only true while dropping a snapshot */
+			if (level == lowest_level) {
+				ret = 0;
+				goto done;
+			}
+
+			blocknr = btrfs_node_blockptr(b, slot);
+			gen = btrfs_node_ptr_generation(b, slot);
+			blocksize = btrfs_level_size(root, level - 1);
+
+			tmp = btrfs_find_tree_block(root, blocknr, blocksize);
+			if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+				b = tmp;
+			} else {
+				/*
+				 * reduce lock contention at high levels
+				 * of the btree by dropping locks before
+				 * we read.
+				 */
+				if (level > 1) {
+					btrfs_release_path(NULL, p);
+					if (tmp)
+						free_extent_buffer(tmp);
+					if (should_reada)
+						reada_for_search(root, p,
+								 level, slot,
+								 key->objectid);
+
+					tmp = read_tree_block(root, blocknr,
+							 blocksize, gen);
+					if (tmp)
+						free_extent_buffer(tmp);
+					goto again;
+				} else {
+					if (tmp)
+						free_extent_buffer(tmp);
+					if (should_reada)
+						reada_for_search(root, p,
+								 level, slot,
+								 key->objectid);
+					b = read_node_slot(root, b, slot);
+				}
+			}
+			if (!p->skip_locking)
+				btrfs_tree_lock(b);
+		} else {
+			p->slots[level] = slot;
+			if (ins_len > 0 &&
+			    btrfs_leaf_free_space(root, b) < ins_len) {
+				int sret = split_leaf(trans, root, key,
+						      p, ins_len, ret == 0);
+				BUG_ON(sret > 0);
+				if (sret) {
+					ret = sret;
+					goto done;
+				}
+			}
+			if (!p->search_for_split)
+				unlock_up(p, level, lowest_unlock);
+			goto done;
+		}
+	}
+	ret = 1;
+done:
+	if (prealloc_block.objectid) {
+		btrfs_free_reserved_extent(root,
+			   prealloc_block.objectid,
+			   prealloc_block.offset);
+	}
+
+	return ret;
+}
+
+int btrfs_merge_path(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root,
+		     struct btrfs_key *node_keys,
+		     u64 *nodes, int lowest_level)
+{
+	struct extent_buffer *eb;
+	struct extent_buffer *parent;
+	struct btrfs_key key;
+	u64 bytenr;
+	u64 generation;
+	u32 blocksize;
+	int level;
+	int slot;
+	int key_match;
+	int ret;
+
+	eb = btrfs_lock_root_node(root);
+	ret = btrfs_cow_block(trans, root, eb, NULL, 0, &eb, 0);
+	BUG_ON(ret);
+
+	parent = eb;
+	while (1) {
+		level = btrfs_header_level(parent);
+		if (level == 0 || level <= lowest_level)
+			break;
+
+		ret = bin_search(parent, &node_keys[lowest_level], level,
+				 &slot);
+		if (ret && slot > 0)
+			slot--;
+
+		bytenr = btrfs_node_blockptr(parent, slot);
+		if (nodes[level - 1] == bytenr)
+			break;
+
+		blocksize = btrfs_level_size(root, level - 1);
+		generation = btrfs_node_ptr_generation(parent, slot);
+		btrfs_node_key_to_cpu(eb, &key, slot);
+		key_match = !memcmp(&key, &node_keys[level - 1], sizeof(key));
+
+		if (generation == trans->transid) {
+			eb = read_tree_block(root, bytenr, blocksize,
+					     generation);
+			btrfs_tree_lock(eb);
+		}
+
+		/*
+		 * if node keys match and node pointer hasn't been modified
+		 * in the running transaction, we can merge the path. for
+		 * blocks owened by reloc trees, the node pointer check is
+		 * skipped, this is because these blocks are fully controlled
+		 * by the space balance code, no one else can modify them.
+		 */
+		if (!nodes[level - 1] || !key_match ||
+		    (generation == trans->transid &&
+		     btrfs_header_owner(eb) != BTRFS_TREE_RELOC_OBJECTID)) {
+			if (level == 1 || level == lowest_level + 1) {
+				if (generation == trans->transid) {
+					btrfs_tree_unlock(eb);
+					free_extent_buffer(eb);
+				}
+				break;
+			}
+
+			if (generation != trans->transid) {
+				eb = read_tree_block(root, bytenr, blocksize,
+						generation);
+				btrfs_tree_lock(eb);
+			}
+
+			ret = btrfs_cow_block(trans, root, eb, parent, slot,
+					      &eb, 0);
+			BUG_ON(ret);
+
+			if (root->root_key.objectid ==
+			    BTRFS_TREE_RELOC_OBJECTID) {
+				if (!nodes[level - 1]) {
+					nodes[level - 1] = eb->start;
+					memcpy(&node_keys[level - 1], &key,
+					       sizeof(node_keys[0]));
+				} else {
+					WARN_ON(1);
+				}
+			}
+
+			btrfs_tree_unlock(parent);
+			free_extent_buffer(parent);
+			parent = eb;
+			continue;
+		}
+
+		btrfs_set_node_blockptr(parent, slot, nodes[level - 1]);
+		btrfs_set_node_ptr_generation(parent, slot, trans->transid);
+		btrfs_mark_buffer_dirty(parent);
+
+		ret = btrfs_inc_extent_ref(trans, root,
+					nodes[level - 1],
+					blocksize, parent->start,
+					btrfs_header_owner(parent),
+					btrfs_header_generation(parent),
+					level - 1);
+		BUG_ON(ret);
+
+		/*
+		 * If the block was created in the running transaction,
+		 * it's possible this is the last reference to it, so we
+		 * should drop the subtree.
+		 */
+		if (generation == trans->transid) {
+			ret = btrfs_drop_subtree(trans, root, eb, parent);
+			BUG_ON(ret);
+			btrfs_tree_unlock(eb);
+			free_extent_buffer(eb);
+		} else {
+			ret = btrfs_free_extent(trans, root, bytenr,
+					blocksize, parent->start,
+					btrfs_header_owner(parent),
+					btrfs_header_generation(parent),
+					level - 1, 1);
+			BUG_ON(ret);
+		}
+		break;
+	}
+	btrfs_tree_unlock(parent);
+	free_extent_buffer(parent);
+	return 0;
+}
+
+/*
+ * adjust the pointers going up the tree, starting at level
+ * making sure the right key of each node is points to 'key'.
+ * This is used after shifting pointers to the left, so it stops
+ * fixing up pointers when a given leaf/node is not in slot 0 of the
+ * higher levels
+ *
+ * If this fails to write a tree block, it returns -1, but continues
+ * fixing up the blocks in ram so the tree is consistent.
+ */
+static int fixup_low_keys(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct btrfs_path *path,
+			  struct btrfs_disk_key *key, int level)
+{
+	int i;
+	int ret = 0;
+	struct extent_buffer *t;
+
+	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
+		int tslot = path->slots[i];
+		if (!path->nodes[i])
+			break;
+		t = path->nodes[i];
+		btrfs_set_node_key(t, key, tslot);
+		btrfs_mark_buffer_dirty(path->nodes[i]);
+		if (tslot != 0)
+			break;
+	}
+	return ret;
+}
+
+/*
+ * update item key.
+ *
+ * This function isn't completely safe. It's the caller's responsibility
+ * that the new key won't break the order
+ */
+int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, struct btrfs_path *path,
+			    struct btrfs_key *new_key)
+{
+	struct btrfs_disk_key disk_key;
+	struct extent_buffer *eb;
+	int slot;
+
+	eb = path->nodes[0];
+	slot = path->slots[0];
+	if (slot > 0) {
+		btrfs_item_key(eb, &disk_key, slot - 1);
+		if (comp_keys(&disk_key, new_key) >= 0)
+			return -1;
+	}
+	if (slot < btrfs_header_nritems(eb) - 1) {
+		btrfs_item_key(eb, &disk_key, slot + 1);
+		if (comp_keys(&disk_key, new_key) <= 0)
+			return -1;
+	}
+
+	btrfs_cpu_key_to_disk(&disk_key, new_key);
+	btrfs_set_item_key(eb, &disk_key, slot);
+	btrfs_mark_buffer_dirty(eb);
+	if (slot == 0)
+		fixup_low_keys(trans, root, path, &disk_key, 1);
+	return 0;
+}
+
+/*
+ * try to push data from one node into the next node left in the
+ * tree.
+ *
+ * returns 0 if some ptrs were pushed left, < 0 if there was some horrible
+ * error, and > 0 if there was no room in the left hand block.
+ */
+static int push_node_left(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct extent_buffer *dst,
+			  struct extent_buffer *src, int empty)
+{
+	int push_items = 0;
+	int src_nritems;
+	int dst_nritems;
+	int ret = 0;
+
+	src_nritems = btrfs_header_nritems(src);
+	dst_nritems = btrfs_header_nritems(dst);
+	push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems;
+	WARN_ON(btrfs_header_generation(src) != trans->transid);
+	WARN_ON(btrfs_header_generation(dst) != trans->transid);
+
+	if (!empty && src_nritems <= 8)
+		return 1;
+
+	if (push_items <= 0)
+		return 1;
+
+	if (empty) {
+		push_items = min(src_nritems, push_items);
+		if (push_items < src_nritems) {
+			/* leave at least 8 pointers in the node if
+			 * we aren't going to empty it
+			 */
+			if (src_nritems - push_items < 8) {
+				if (push_items <= 8)
+					return 1;
+				push_items -= 8;
+			}
+		}
+	} else
+		push_items = min(src_nritems - 8, push_items);
+
+	copy_extent_buffer(dst, src,
+			   btrfs_node_key_ptr_offset(dst_nritems),
+			   btrfs_node_key_ptr_offset(0),
+			   push_items * sizeof(struct btrfs_key_ptr));
+
+	if (push_items < src_nritems) {
+		memmove_extent_buffer(src, btrfs_node_key_ptr_offset(0),
+				      btrfs_node_key_ptr_offset(push_items),
+				      (src_nritems - push_items) *
+				      sizeof(struct btrfs_key_ptr));
+	}
+	btrfs_set_header_nritems(src, src_nritems - push_items);
+	btrfs_set_header_nritems(dst, dst_nritems + push_items);
+	btrfs_mark_buffer_dirty(src);
+	btrfs_mark_buffer_dirty(dst);
+
+	ret = btrfs_update_ref(trans, root, src, dst, dst_nritems, push_items);
+	BUG_ON(ret);
+
+	return ret;
+}
+
+/*
+ * try to push data from one node into the next node right in the
+ * tree.
+ *
+ * returns 0 if some ptrs were pushed, < 0 if there was some horrible
+ * error, and > 0 if there was no room in the right hand block.
+ *
+ * this will  only push up to 1/2 the contents of the left node over
+ */
+static int balance_node_right(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct extent_buffer *dst,
+			      struct extent_buffer *src)
+{
+	int push_items = 0;
+	int max_push;
+	int src_nritems;
+	int dst_nritems;
+	int ret = 0;
+
+	WARN_ON(btrfs_header_generation(src) != trans->transid);
+	WARN_ON(btrfs_header_generation(dst) != trans->transid);
+
+	src_nritems = btrfs_header_nritems(src);
+	dst_nritems = btrfs_header_nritems(dst);
+	push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems;
+	if (push_items <= 0)
+		return 1;
+
+	if (src_nritems < 4)
+		return 1;
+
+	max_push = src_nritems / 2 + 1;
+	/* don't try to empty the node */
+	if (max_push >= src_nritems)
+		return 1;
+
+	if (max_push < push_items)
+		push_items = max_push;
+
+	memmove_extent_buffer(dst, btrfs_node_key_ptr_offset(push_items),
+				      btrfs_node_key_ptr_offset(0),
+				      (dst_nritems) *
+				      sizeof(struct btrfs_key_ptr));
+
+	copy_extent_buffer(dst, src,
+			   btrfs_node_key_ptr_offset(0),
+			   btrfs_node_key_ptr_offset(src_nritems - push_items),
+			   push_items * sizeof(struct btrfs_key_ptr));
+
+	btrfs_set_header_nritems(src, src_nritems - push_items);
+	btrfs_set_header_nritems(dst, dst_nritems + push_items);
+
+	btrfs_mark_buffer_dirty(src);
+	btrfs_mark_buffer_dirty(dst);
+
+	ret = btrfs_update_ref(trans, root, src, dst, 0, push_items);
+	BUG_ON(ret);
+
+	return ret;
+}
+
+/*
+ * helper function to insert a new root level in the tree.
+ * A new node is allocated, and a single item is inserted to
+ * point to the existing root
+ *
+ * returns zero on success or < 0 on failure.
+ */
+static noinline int insert_new_root(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct btrfs_path *path, int level)
+{
+	u64 lower_gen;
+	struct extent_buffer *lower;
+	struct extent_buffer *c;
+	struct extent_buffer *old;
+	struct btrfs_disk_key lower_key;
+	int ret;
+
+	BUG_ON(path->nodes[level]);
+	BUG_ON(path->nodes[level-1] != root->node);
+
+	lower = path->nodes[level-1];
+	if (level == 1)
+		btrfs_item_key(lower, &lower_key, 0);
+	else
+		btrfs_node_key(lower, &lower_key, 0);
+
+	c = btrfs_alloc_free_block(trans, root, root->nodesize, 0,
+				   root->root_key.objectid, trans->transid,
+				   level, root->node->start, 0);
+	if (IS_ERR(c))
+		return PTR_ERR(c);
+
+	memset_extent_buffer(c, 0, 0, root->nodesize);
+	btrfs_set_header_nritems(c, 1);
+	btrfs_set_header_level(c, level);
+	btrfs_set_header_bytenr(c, c->start);
+	btrfs_set_header_generation(c, trans->transid);
+	btrfs_set_header_owner(c, root->root_key.objectid);
+
+	write_extent_buffer(c, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(c),
+			    BTRFS_FSID_SIZE);
+
+	write_extent_buffer(c, root->fs_info->chunk_tree_uuid,
+			    (unsigned long)btrfs_header_chunk_tree_uuid(c),
+			    BTRFS_UUID_SIZE);
+
+	btrfs_set_node_key(c, &lower_key, 0);
+	btrfs_set_node_blockptr(c, 0, lower->start);
+	lower_gen = btrfs_header_generation(lower);
+	WARN_ON(lower_gen != trans->transid);
+
+	btrfs_set_node_ptr_generation(c, 0, lower_gen);
+
+	btrfs_mark_buffer_dirty(c);
+
+	spin_lock(&root->node_lock);
+	old = root->node;
+	root->node = c;
+	spin_unlock(&root->node_lock);
+
+	ret = btrfs_update_extent_ref(trans, root, lower->start,
+				      lower->start, c->start,
+				      root->root_key.objectid,
+				      trans->transid, level - 1);
+	BUG_ON(ret);
+
+	/* the super has an extra ref to root->node */
+	free_extent_buffer(old);
+
+	add_root_to_dirty_list(root);
+	extent_buffer_get(c);
+	path->nodes[level] = c;
+	path->locks[level] = 1;
+	path->slots[level] = 0;
+	return 0;
+}
+
+/*
+ * worker function to insert a single pointer in a node.
+ * the node should have enough room for the pointer already
+ *
+ * slot and level indicate where you want the key to go, and
+ * blocknr is the block the key points to.
+ *
+ * returns zero on success and < 0 on any error
+ */
+static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_path *path, struct btrfs_disk_key
+		      *key, u64 bytenr, int slot, int level)
+{
+	struct extent_buffer *lower;
+	int nritems;
+
+	BUG_ON(!path->nodes[level]);
+	lower = path->nodes[level];
+	nritems = btrfs_header_nritems(lower);
+	if (slot > nritems)
+		BUG();
+	if (nritems == BTRFS_NODEPTRS_PER_BLOCK(root))
+		BUG();
+	if (slot != nritems) {
+		memmove_extent_buffer(lower,
+			      btrfs_node_key_ptr_offset(slot + 1),
+			      btrfs_node_key_ptr_offset(slot),
+			      (nritems - slot) * sizeof(struct btrfs_key_ptr));
+	}
+	btrfs_set_node_key(lower, key, slot);
+	btrfs_set_node_blockptr(lower, slot, bytenr);
+	WARN_ON(trans->transid == 0);
+	btrfs_set_node_ptr_generation(lower, slot, trans->transid);
+	btrfs_set_header_nritems(lower, nritems + 1);
+	btrfs_mark_buffer_dirty(lower);
+	return 0;
+}
+
+/*
+ * split the node at the specified level in path in two.
+ * The path is corrected to point to the appropriate node after the split
+ *
+ * Before splitting this tries to make some room in the node by pushing
+ * left and right, if either one works, it returns right away.
+ *
+ * returns 0 on success and < 0 on failure
+ */
+static noinline int split_node(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct btrfs_path *path, int level)
+{
+	struct extent_buffer *c;
+	struct extent_buffer *split;
+	struct btrfs_disk_key disk_key;
+	int mid;
+	int ret;
+	int wret;
+	u32 c_nritems;
+
+	c = path->nodes[level];
+	WARN_ON(btrfs_header_generation(c) != trans->transid);
+	if (c == root->node) {
+		/* trying to split the root, lets make a new one */
+		ret = insert_new_root(trans, root, path, level + 1);
+		if (ret)
+			return ret;
+	} else {
+		ret = push_nodes_for_insert(trans, root, path, level);
+		c = path->nodes[level];
+		if (!ret && btrfs_header_nritems(c) <
+		    BTRFS_NODEPTRS_PER_BLOCK(root) - 3)
+			return 0;
+		if (ret < 0)
+			return ret;
+	}
+
+	c_nritems = btrfs_header_nritems(c);
+
+	split = btrfs_alloc_free_block(trans, root, root->nodesize,
+					path->nodes[level + 1]->start,
+					root->root_key.objectid,
+					trans->transid, level, c->start, 0);
+	if (IS_ERR(split))
+		return PTR_ERR(split);
+
+	btrfs_set_header_flags(split, btrfs_header_flags(c));
+	btrfs_set_header_level(split, btrfs_header_level(c));
+	btrfs_set_header_bytenr(split, split->start);
+	btrfs_set_header_generation(split, trans->transid);
+	btrfs_set_header_owner(split, root->root_key.objectid);
+	btrfs_set_header_flags(split, 0);
+	write_extent_buffer(split, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(split),
+			    BTRFS_FSID_SIZE);
+	write_extent_buffer(split, root->fs_info->chunk_tree_uuid,
+			    (unsigned long)btrfs_header_chunk_tree_uuid(split),
+			    BTRFS_UUID_SIZE);
+
+	mid = (c_nritems + 1) / 2;
+
+	copy_extent_buffer(split, c,
+			   btrfs_node_key_ptr_offset(0),
+			   btrfs_node_key_ptr_offset(mid),
+			   (c_nritems - mid) * sizeof(struct btrfs_key_ptr));
+	btrfs_set_header_nritems(split, c_nritems - mid);
+	btrfs_set_header_nritems(c, mid);
+	ret = 0;
+
+	btrfs_mark_buffer_dirty(c);
+	btrfs_mark_buffer_dirty(split);
+
+	btrfs_node_key(split, &disk_key, 0);
+	wret = insert_ptr(trans, root, path, &disk_key, split->start,
+			  path->slots[level + 1] + 1,
+			  level + 1);
+	if (wret)
+		ret = wret;
+
+	ret = btrfs_update_ref(trans, root, c, split, 0, c_nritems - mid);
+	BUG_ON(ret);
+
+	if (path->slots[level] >= mid) {
+		path->slots[level] -= mid;
+		btrfs_tree_unlock(c);
+		free_extent_buffer(c);
+		path->nodes[level] = split;
+		path->slots[level + 1] += 1;
+	} else {
+		btrfs_tree_unlock(split);
+		free_extent_buffer(split);
+	}
+	return ret;
+}
+
+/*
+ * how many bytes are required to store the items in a leaf.  start
+ * and nr indicate which items in the leaf to check.  This totals up the
+ * space used both by the item structs and the item data
+ */
+static int leaf_space_used(struct extent_buffer *l, int start, int nr)
+{
+	int data_len;
+	int nritems = btrfs_header_nritems(l);
+	int end = min(nritems, start + nr) - 1;
+
+	if (!nr)
+		return 0;
+	data_len = btrfs_item_end_nr(l, start);
+	data_len = data_len - btrfs_item_offset_nr(l, end);
+	data_len += sizeof(struct btrfs_item) * nr;
+	WARN_ON(data_len < 0);
+	return data_len;
+}
+
+/*
+ * The space between the end of the leaf items and
+ * the start of the leaf data.  IOW, how much room
+ * the leaf has left for both items and data
+ */
+noinline int btrfs_leaf_free_space(struct btrfs_root *root,
+				   struct extent_buffer *leaf)
+{
+	int nritems = btrfs_header_nritems(leaf);
+	int ret;
+	ret = BTRFS_LEAF_DATA_SIZE(root) - leaf_space_used(leaf, 0, nritems);
+	if (ret < 0) {
+		printk(KERN_CRIT "leaf free space ret %d, leaf data size %lu, "
+		       "used %d nritems %d\n",
+		       ret, (unsigned long) BTRFS_LEAF_DATA_SIZE(root),
+		       leaf_space_used(leaf, 0, nritems), nritems);
+	}
+	return ret;
+}
+
+/*
+ * push some data in the path leaf to the right, trying to free up at
+ * least data_size bytes.  returns zero if the push worked, nonzero otherwise
+ *
+ * returns 1 if the push failed because the other node didn't have enough
+ * room, 0 if everything worked out and < 0 if there were major errors.
+ */
+static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root
+			   *root, struct btrfs_path *path, int data_size,
+			   int empty)
+{
+	struct extent_buffer *left = path->nodes[0];
+	struct extent_buffer *right;
+	struct extent_buffer *upper;
+	struct btrfs_disk_key disk_key;
+	int slot;
+	u32 i;
+	int free_space;
+	int push_space = 0;
+	int push_items = 0;
+	struct btrfs_item *item;
+	u32 left_nritems;
+	u32 nr;
+	u32 right_nritems;
+	u32 data_end;
+	u32 this_item_size;
+	int ret;
+
+	slot = path->slots[1];
+	if (!path->nodes[1])
+		return 1;
+
+	upper = path->nodes[1];
+	if (slot >= btrfs_header_nritems(upper) - 1)
+		return 1;
+
+	WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+
+	right = read_node_slot(root, upper, slot + 1);
+	btrfs_tree_lock(right);
+	free_space = btrfs_leaf_free_space(root, right);
+	if (free_space < data_size)
+		goto out_unlock;
+
+	/* cow and double check */
+	ret = btrfs_cow_block(trans, root, right, upper,
+			      slot + 1, &right, 0);
+	if (ret)
+		goto out_unlock;
+
+	free_space = btrfs_leaf_free_space(root, right);
+	if (free_space < data_size)
+		goto out_unlock;
+
+	left_nritems = btrfs_header_nritems(left);
+	if (left_nritems == 0)
+		goto out_unlock;
+
+	if (empty)
+		nr = 0;
+	else
+		nr = 1;
+
+	if (path->slots[0] >= left_nritems)
+		push_space += data_size;
+
+	i = left_nritems - 1;
+	while (i >= nr) {
+		item = btrfs_item_nr(left, i);
+
+		if (!empty && push_items > 0) {
+			if (path->slots[0] > i)
+				break;
+			if (path->slots[0] == i) {
+				int space = btrfs_leaf_free_space(root, left);
+				if (space + push_space * 2 > free_space)
+					break;
+			}
+		}
+
+		if (path->slots[0] == i)
+			push_space += data_size;
+
+		if (!left->map_token) {
+			map_extent_buffer(left, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&left->map_token, &left->kaddr,
+					&left->map_start, &left->map_len,
+					KM_USER1);
+		}
+
+		this_item_size = btrfs_item_size(left, item);
+		if (this_item_size + sizeof(*item) + push_space > free_space)
+			break;
+
+		push_items++;
+		push_space += this_item_size + sizeof(*item);
+		if (i == 0)
+			break;
+		i--;
+	}
+	if (left->map_token) {
+		unmap_extent_buffer(left, left->map_token, KM_USER1);
+		left->map_token = NULL;
+	}
+
+	if (push_items == 0)
+		goto out_unlock;
+
+	if (!empty && push_items == left_nritems)
+		WARN_ON(1);
+
+	/* push left to right */
+	right_nritems = btrfs_header_nritems(right);
+
+	push_space = btrfs_item_end_nr(left, left_nritems - push_items);
+	push_space -= leaf_data_end(root, left);
+
+	/* make room in the right data area */
+	data_end = leaf_data_end(root, right);
+	memmove_extent_buffer(right,
+			      btrfs_leaf_data(right) + data_end - push_space,
+			      btrfs_leaf_data(right) + data_end,
+			      BTRFS_LEAF_DATA_SIZE(root) - data_end);
+
+	/* copy from the left data area */
+	copy_extent_buffer(right, left, btrfs_leaf_data(right) +
+		     BTRFS_LEAF_DATA_SIZE(root) - push_space,
+		     btrfs_leaf_data(left) + leaf_data_end(root, left),
+		     push_space);
+
+	memmove_extent_buffer(right, btrfs_item_nr_offset(push_items),
+			      btrfs_item_nr_offset(0),
+			      right_nritems * sizeof(struct btrfs_item));
+
+	/* copy the items from left to right */
+	copy_extent_buffer(right, left, btrfs_item_nr_offset(0),
+		   btrfs_item_nr_offset(left_nritems - push_items),
+		   push_items * sizeof(struct btrfs_item));
+
+	/* update the item pointers */
+	right_nritems += push_items;
+	btrfs_set_header_nritems(right, right_nritems);
+	push_space = BTRFS_LEAF_DATA_SIZE(root);
+	for (i = 0; i < right_nritems; i++) {
+		item = btrfs_item_nr(right, i);
+		if (!right->map_token) {
+			map_extent_buffer(right, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&right->map_token, &right->kaddr,
+					&right->map_start, &right->map_len,
+					KM_USER1);
+		}
+		push_space -= btrfs_item_size(right, item);
+		btrfs_set_item_offset(right, item, push_space);
+	}
+
+	if (right->map_token) {
+		unmap_extent_buffer(right, right->map_token, KM_USER1);
+		right->map_token = NULL;
+	}
+	left_nritems -= push_items;
+	btrfs_set_header_nritems(left, left_nritems);
+
+	if (left_nritems)
+		btrfs_mark_buffer_dirty(left);
+	btrfs_mark_buffer_dirty(right);
+
+	ret = btrfs_update_ref(trans, root, left, right, 0, push_items);
+	BUG_ON(ret);
+
+	btrfs_item_key(right, &disk_key, 0);
+	btrfs_set_node_key(upper, &disk_key, slot + 1);
+	btrfs_mark_buffer_dirty(upper);
+
+	/* then fixup the leaf pointer in the path */
+	if (path->slots[0] >= left_nritems) {
+		path->slots[0] -= left_nritems;
+		if (btrfs_header_nritems(path->nodes[0]) == 0)
+			clean_tree_block(trans, root, path->nodes[0]);
+		btrfs_tree_unlock(path->nodes[0]);
+		free_extent_buffer(path->nodes[0]);
+		path->nodes[0] = right;
+		path->slots[1] += 1;
+	} else {
+		btrfs_tree_unlock(right);
+		free_extent_buffer(right);
+	}
+	return 0;
+
+out_unlock:
+	btrfs_tree_unlock(right);
+	free_extent_buffer(right);
+	return 1;
+}
+
+/*
+ * push some data in the path leaf to the left, trying to free up at
+ * least data_size bytes.  returns zero if the push worked, nonzero otherwise
+ */
+static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
+			  *root, struct btrfs_path *path, int data_size,
+			  int empty)
+{
+	struct btrfs_disk_key disk_key;
+	struct extent_buffer *right = path->nodes[0];
+	struct extent_buffer *left;
+	int slot;
+	int i;
+	int free_space;
+	int push_space = 0;
+	int push_items = 0;
+	struct btrfs_item *item;
+	u32 old_left_nritems;
+	u32 right_nritems;
+	u32 nr;
+	int ret = 0;
+	int wret;
+	u32 this_item_size;
+	u32 old_left_item_size;
+
+	slot = path->slots[1];
+	if (slot == 0)
+		return 1;
+	if (!path->nodes[1])
+		return 1;
+
+	right_nritems = btrfs_header_nritems(right);
+	if (right_nritems == 0)
+		return 1;
+
+	WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+
+	left = read_node_slot(root, path->nodes[1], slot - 1);
+	btrfs_tree_lock(left);
+	free_space = btrfs_leaf_free_space(root, left);
+	if (free_space < data_size) {
+		ret = 1;
+		goto out;
+	}
+
+	/* cow and double check */
+	ret = btrfs_cow_block(trans, root, left,
+			      path->nodes[1], slot - 1, &left, 0);
+	if (ret) {
+		/* we hit -ENOSPC, but it isn't fatal here */
+		ret = 1;
+		goto out;
+	}
+
+	free_space = btrfs_leaf_free_space(root, left);
+	if (free_space < data_size) {
+		ret = 1;
+		goto out;
+	}
+
+	if (empty)
+		nr = right_nritems;
+	else
+		nr = right_nritems - 1;
+
+	for (i = 0; i < nr; i++) {
+		item = btrfs_item_nr(right, i);
+		if (!right->map_token) {
+			map_extent_buffer(right, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&right->map_token, &right->kaddr,
+					&right->map_start, &right->map_len,
+					KM_USER1);
+		}
+
+		if (!empty && push_items > 0) {
+			if (path->slots[0] < i)
+				break;
+			if (path->slots[0] == i) {
+				int space = btrfs_leaf_free_space(root, right);
+				if (space + push_space * 2 > free_space)
+					break;
+			}
+		}
+
+		if (path->slots[0] == i)
+			push_space += data_size;
+
+		this_item_size = btrfs_item_size(right, item);
+		if (this_item_size + sizeof(*item) + push_space > free_space)
+			break;
+
+		push_items++;
+		push_space += this_item_size + sizeof(*item);
+	}
+
+	if (right->map_token) {
+		unmap_extent_buffer(right, right->map_token, KM_USER1);
+		right->map_token = NULL;
+	}
+
+	if (push_items == 0) {
+		ret = 1;
+		goto out;
+	}
+	if (!empty && push_items == btrfs_header_nritems(right))
+		WARN_ON(1);
+
+	/* push data from right to left */
+	copy_extent_buffer(left, right,
+			   btrfs_item_nr_offset(btrfs_header_nritems(left)),
+			   btrfs_item_nr_offset(0),
+			   push_items * sizeof(struct btrfs_item));
+
+	push_space = BTRFS_LEAF_DATA_SIZE(root) -
+		     btrfs_item_offset_nr(right, push_items - 1);
+
+	copy_extent_buffer(left, right, btrfs_leaf_data(left) +
+		     leaf_data_end(root, left) - push_space,
+		     btrfs_leaf_data(right) +
+		     btrfs_item_offset_nr(right, push_items - 1),
+		     push_space);
+	old_left_nritems = btrfs_header_nritems(left);
+	BUG_ON(old_left_nritems <= 0);
+
+	old_left_item_size = btrfs_item_offset_nr(left, old_left_nritems - 1);
+	for (i = old_left_nritems; i < old_left_nritems + push_items; i++) {
+		u32 ioff;
+
+		item = btrfs_item_nr(left, i);
+		if (!left->map_token) {
+			map_extent_buffer(left, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&left->map_token, &left->kaddr,
+					&left->map_start, &left->map_len,
+					KM_USER1);
+		}
+
+		ioff = btrfs_item_offset(left, item);
+		btrfs_set_item_offset(left, item,
+		      ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size));
+	}
+	btrfs_set_header_nritems(left, old_left_nritems + push_items);
+	if (left->map_token) {
+		unmap_extent_buffer(left, left->map_token, KM_USER1);
+		left->map_token = NULL;
+	}
+
+	/* fixup right node */
+	if (push_items > right_nritems) {
+		printk(KERN_CRIT "push items %d nr %u\n", push_items,
+		       right_nritems);
+		WARN_ON(1);
+	}
+
+	if (push_items < right_nritems) {
+		push_space = btrfs_item_offset_nr(right, push_items - 1) -
+						  leaf_data_end(root, right);
+		memmove_extent_buffer(right, btrfs_leaf_data(right) +
+				      BTRFS_LEAF_DATA_SIZE(root) - push_space,
+				      btrfs_leaf_data(right) +
+				      leaf_data_end(root, right), push_space);
+
+		memmove_extent_buffer(right, btrfs_item_nr_offset(0),
+			      btrfs_item_nr_offset(push_items),
+			     (btrfs_header_nritems(right) - push_items) *
+			     sizeof(struct btrfs_item));
+	}
+	right_nritems -= push_items;
+	btrfs_set_header_nritems(right, right_nritems);
+	push_space = BTRFS_LEAF_DATA_SIZE(root);
+	for (i = 0; i < right_nritems; i++) {
+		item = btrfs_item_nr(right, i);
+
+		if (!right->map_token) {
+			map_extent_buffer(right, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&right->map_token, &right->kaddr,
+					&right->map_start, &right->map_len,
+					KM_USER1);
+		}
+
+		push_space = push_space - btrfs_item_size(right, item);
+		btrfs_set_item_offset(right, item, push_space);
+	}
+	if (right->map_token) {
+		unmap_extent_buffer(right, right->map_token, KM_USER1);
+		right->map_token = NULL;
+	}
+
+	btrfs_mark_buffer_dirty(left);
+	if (right_nritems)
+		btrfs_mark_buffer_dirty(right);
+
+	ret = btrfs_update_ref(trans, root, right, left,
+			       old_left_nritems, push_items);
+	BUG_ON(ret);
+
+	btrfs_item_key(right, &disk_key, 0);
+	wret = fixup_low_keys(trans, root, path, &disk_key, 1);
+	if (wret)
+		ret = wret;
+
+	/* then fixup the leaf pointer in the path */
+	if (path->slots[0] < push_items) {
+		path->slots[0] += old_left_nritems;
+		if (btrfs_header_nritems(path->nodes[0]) == 0)
+			clean_tree_block(trans, root, path->nodes[0]);
+		btrfs_tree_unlock(path->nodes[0]);
+		free_extent_buffer(path->nodes[0]);
+		path->nodes[0] = left;
+		path->slots[1] -= 1;
+	} else {
+		btrfs_tree_unlock(left);
+		free_extent_buffer(left);
+		path->slots[0] -= push_items;
+	}
+	BUG_ON(path->slots[0] < 0);
+	return ret;
+out:
+	btrfs_tree_unlock(left);
+	free_extent_buffer(left);
+	return ret;
+}
+
+/*
+ * split the path's leaf in two, making sure there is at least data_size
+ * available for the resulting leaf level of the path.
+ *
+ * returns 0 if all went well and < 0 on failure.
+ */
+static noinline int split_leaf(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct btrfs_key *ins_key,
+			       struct btrfs_path *path, int data_size,
+			       int extend)
+{
+	struct extent_buffer *l;
+	u32 nritems;
+	int mid;
+	int slot;
+	struct extent_buffer *right;
+	int data_copy_size;
+	int rt_data_off;
+	int i;
+	int ret = 0;
+	int wret;
+	int double_split;
+	int num_doubles = 0;
+	struct btrfs_disk_key disk_key;
+
+	/* first try to make some room by pushing left and right */
+	if (data_size && ins_key->type != BTRFS_DIR_ITEM_KEY) {
+		wret = push_leaf_right(trans, root, path, data_size, 0);
+		if (wret < 0)
+			return wret;
+		if (wret) {
+			wret = push_leaf_left(trans, root, path, data_size, 0);
+			if (wret < 0)
+				return wret;
+		}
+		l = path->nodes[0];
+
+		/* did the pushes work? */
+		if (btrfs_leaf_free_space(root, l) >= data_size)
+			return 0;
+	}
+
+	if (!path->nodes[1]) {
+		ret = insert_new_root(trans, root, path, 1);
+		if (ret)
+			return ret;
+	}
+again:
+	double_split = 0;
+	l = path->nodes[0];
+	slot = path->slots[0];
+	nritems = btrfs_header_nritems(l);
+	mid = (nritems + 1) / 2;
+
+	right = btrfs_alloc_free_block(trans, root, root->leafsize,
+					path->nodes[1]->start,
+					root->root_key.objectid,
+					trans->transid, 0, l->start, 0);
+	if (IS_ERR(right)) {
+		BUG_ON(1);
+		return PTR_ERR(right);
+	}
+
+	memset_extent_buffer(right, 0, 0, sizeof(struct btrfs_header));
+	btrfs_set_header_bytenr(right, right->start);
+	btrfs_set_header_generation(right, trans->transid);
+	btrfs_set_header_owner(right, root->root_key.objectid);
+	btrfs_set_header_level(right, 0);
+	write_extent_buffer(right, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(right),
+			    BTRFS_FSID_SIZE);
+
+	write_extent_buffer(right, root->fs_info->chunk_tree_uuid,
+			    (unsigned long)btrfs_header_chunk_tree_uuid(right),
+			    BTRFS_UUID_SIZE);
+	if (mid <= slot) {
+		if (nritems == 1 ||
+		    leaf_space_used(l, mid, nritems - mid) + data_size >
+			BTRFS_LEAF_DATA_SIZE(root)) {
+			if (slot >= nritems) {
+				btrfs_cpu_key_to_disk(&disk_key, ins_key);
+				btrfs_set_header_nritems(right, 0);
+				wret = insert_ptr(trans, root, path,
+						  &disk_key, right->start,
+						  path->slots[1] + 1, 1);
+				if (wret)
+					ret = wret;
+
+				btrfs_tree_unlock(path->nodes[0]);
+				free_extent_buffer(path->nodes[0]);
+				path->nodes[0] = right;
+				path->slots[0] = 0;
+				path->slots[1] += 1;
+				btrfs_mark_buffer_dirty(right);
+				return ret;
+			}
+			mid = slot;
+			if (mid != nritems &&
+			    leaf_space_used(l, mid, nritems - mid) +
+			    data_size > BTRFS_LEAF_DATA_SIZE(root)) {
+				double_split = 1;
+			}
+		}
+	} else {
+		if (leaf_space_used(l, 0, mid) + data_size >
+			BTRFS_LEAF_DATA_SIZE(root)) {
+			if (!extend && data_size && slot == 0) {
+				btrfs_cpu_key_to_disk(&disk_key, ins_key);
+				btrfs_set_header_nritems(right, 0);
+				wret = insert_ptr(trans, root, path,
+						  &disk_key,
+						  right->start,
+						  path->slots[1], 1);
+				if (wret)
+					ret = wret;
+				btrfs_tree_unlock(path->nodes[0]);
+				free_extent_buffer(path->nodes[0]);
+				path->nodes[0] = right;
+				path->slots[0] = 0;
+				if (path->slots[1] == 0) {
+					wret = fixup_low_keys(trans, root,
+						      path, &disk_key, 1);
+					if (wret)
+						ret = wret;
+				}
+				btrfs_mark_buffer_dirty(right);
+				return ret;
+			} else if ((extend || !data_size) && slot == 0) {
+				mid = 1;
+			} else {
+				mid = slot;
+				if (mid != nritems &&
+				    leaf_space_used(l, mid, nritems - mid) +
+				    data_size > BTRFS_LEAF_DATA_SIZE(root)) {
+					double_split = 1;
+				}
+			}
+		}
+	}
+	nritems = nritems - mid;
+	btrfs_set_header_nritems(right, nritems);
+	data_copy_size = btrfs_item_end_nr(l, mid) - leaf_data_end(root, l);
+
+	copy_extent_buffer(right, l, btrfs_item_nr_offset(0),
+			   btrfs_item_nr_offset(mid),
+			   nritems * sizeof(struct btrfs_item));
+
+	copy_extent_buffer(right, l,
+		     btrfs_leaf_data(right) + BTRFS_LEAF_DATA_SIZE(root) -
+		     data_copy_size, btrfs_leaf_data(l) +
+		     leaf_data_end(root, l), data_copy_size);
+
+	rt_data_off = BTRFS_LEAF_DATA_SIZE(root) -
+		      btrfs_item_end_nr(l, mid);
+
+	for (i = 0; i < nritems; i++) {
+		struct btrfs_item *item = btrfs_item_nr(right, i);
+		u32 ioff;
+
+		if (!right->map_token) {
+			map_extent_buffer(right, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&right->map_token, &right->kaddr,
+					&right->map_start, &right->map_len,
+					KM_USER1);
+		}
+
+		ioff = btrfs_item_offset(right, item);
+		btrfs_set_item_offset(right, item, ioff + rt_data_off);
+	}
+
+	if (right->map_token) {
+		unmap_extent_buffer(right, right->map_token, KM_USER1);
+		right->map_token = NULL;
+	}
+
+	btrfs_set_header_nritems(l, mid);
+	ret = 0;
+	btrfs_item_key(right, &disk_key, 0);
+	wret = insert_ptr(trans, root, path, &disk_key, right->start,
+			  path->slots[1] + 1, 1);
+	if (wret)
+		ret = wret;
+
+	btrfs_mark_buffer_dirty(right);
+	btrfs_mark_buffer_dirty(l);
+	BUG_ON(path->slots[0] != slot);
+
+	ret = btrfs_update_ref(trans, root, l, right, 0, nritems);
+	BUG_ON(ret);
+
+	if (mid <= slot) {
+		btrfs_tree_unlock(path->nodes[0]);
+		free_extent_buffer(path->nodes[0]);
+		path->nodes[0] = right;
+		path->slots[0] -= mid;
+		path->slots[1] += 1;
+	} else {
+		btrfs_tree_unlock(right);
+		free_extent_buffer(right);
+	}
+
+	BUG_ON(path->slots[0] < 0);
+
+	if (double_split) {
+		BUG_ON(num_doubles != 0);
+		num_doubles++;
+		goto again;
+	}
+	return ret;
+}
+
+/*
+ * This function splits a single item into two items,
+ * giving 'new_key' to the new item and splitting the
+ * old one at split_offset (from the start of the item).
+ *
+ * The path may be released by this operation.  After
+ * the split, the path is pointing to the old item.  The
+ * new item is going to be in the same node as the old one.
+ *
+ * Note, the item being split must be smaller enough to live alone on
+ * a tree block with room for one extra struct btrfs_item
+ *
+ * This allows us to split the item in place, keeping a lock on the
+ * leaf the entire time.
+ */
+int btrfs_split_item(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root,
+		     struct btrfs_path *path,
+		     struct btrfs_key *new_key,
+		     unsigned long split_offset)
+{
+	u32 item_size;
+	struct extent_buffer *leaf;
+	struct btrfs_key orig_key;
+	struct btrfs_item *item;
+	struct btrfs_item *new_item;
+	int ret = 0;
+	int slot;
+	u32 nritems;
+	u32 orig_offset;
+	struct btrfs_disk_key disk_key;
+	char *buf;
+
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &orig_key, path->slots[0]);
+	if (btrfs_leaf_free_space(root, leaf) >= sizeof(struct btrfs_item))
+		goto split;
+
+	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+	btrfs_release_path(root, path);
+
+	path->search_for_split = 1;
+	path->keep_locks = 1;
+
+	ret = btrfs_search_slot(trans, root, &orig_key, path, 0, 1);
+	path->search_for_split = 0;
+
+	/* if our item isn't there or got smaller, return now */
+	if (ret != 0 || item_size != btrfs_item_size_nr(path->nodes[0],
+							path->slots[0])) {
+		path->keep_locks = 0;
+		return -EAGAIN;
+	}
+
+	ret = split_leaf(trans, root, &orig_key, path,
+			 sizeof(struct btrfs_item), 1);
+	path->keep_locks = 0;
+	BUG_ON(ret);
+
+	leaf = path->nodes[0];
+	BUG_ON(btrfs_leaf_free_space(root, leaf) < sizeof(struct btrfs_item));
+
+split:
+	item = btrfs_item_nr(leaf, path->slots[0]);
+	orig_offset = btrfs_item_offset(leaf, item);
+	item_size = btrfs_item_size(leaf, item);
+
+
+	buf = kmalloc(item_size, GFP_NOFS);
+	read_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf,
+			    path->slots[0]), item_size);
+	slot = path->slots[0] + 1;
+	leaf = path->nodes[0];
+
+	nritems = btrfs_header_nritems(leaf);
+
+	if (slot != nritems) {
+		/* shift the items */
+		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + 1),
+			      btrfs_item_nr_offset(slot),
+			      (nritems - slot) * sizeof(struct btrfs_item));
+
+	}
+
+	btrfs_cpu_key_to_disk(&disk_key, new_key);
+	btrfs_set_item_key(leaf, &disk_key, slot);
+
+	new_item = btrfs_item_nr(leaf, slot);
+
+	btrfs_set_item_offset(leaf, new_item, orig_offset);
+	btrfs_set_item_size(leaf, new_item, item_size - split_offset);
+
+	btrfs_set_item_offset(leaf, item,
+			      orig_offset + item_size - split_offset);
+	btrfs_set_item_size(leaf, item, split_offset);
+
+	btrfs_set_header_nritems(leaf, nritems + 1);
+
+	/* write the data for the start of the original item */
+	write_extent_buffer(leaf, buf,
+			    btrfs_item_ptr_offset(leaf, path->slots[0]),
+			    split_offset);
+
+	/* write the data for the new item */
+	write_extent_buffer(leaf, buf + split_offset,
+			    btrfs_item_ptr_offset(leaf, slot),
+			    item_size - split_offset);
+	btrfs_mark_buffer_dirty(leaf);
+
+	ret = 0;
+	if (btrfs_leaf_free_space(root, leaf) < 0) {
+		btrfs_print_leaf(root, leaf);
+		BUG();
+	}
+	kfree(buf);
+	return ret;
+}
+
+/*
+ * make the item pointed to by the path smaller.  new_size indicates
+ * how small to make it, and from_end tells us if we just chop bytes
+ * off the end of the item or if we shift the item to chop bytes off
+ * the front.
+ */
+int btrfs_truncate_item(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct btrfs_path *path,
+			u32 new_size, int from_end)
+{
+	int ret = 0;
+	int slot;
+	int slot_orig;
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	u32 nritems;
+	unsigned int data_end;
+	unsigned int old_data_start;
+	unsigned int old_size;
+	unsigned int size_diff;
+	int i;
+
+	slot_orig = path->slots[0];
+	leaf = path->nodes[0];
+	slot = path->slots[0];
+
+	old_size = btrfs_item_size_nr(leaf, slot);
+	if (old_size == new_size)
+		return 0;
+
+	nritems = btrfs_header_nritems(leaf);
+	data_end = leaf_data_end(root, leaf);
+
+	old_data_start = btrfs_item_offset_nr(leaf, slot);
+
+	size_diff = old_size - new_size;
+
+	BUG_ON(slot < 0);
+	BUG_ON(slot >= nritems);
+
+	/*
+	 * item0..itemN ... dataN.offset..dataN.size .. data0.size
+	 */
+	/* first correct the data pointers */
+	for (i = slot; i < nritems; i++) {
+		u32 ioff;
+		item = btrfs_item_nr(leaf, i);
+
+		if (!leaf->map_token) {
+			map_extent_buffer(leaf, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&leaf->map_token, &leaf->kaddr,
+					&leaf->map_start, &leaf->map_len,
+					KM_USER1);
+		}
+
+		ioff = btrfs_item_offset(leaf, item);
+		btrfs_set_item_offset(leaf, item, ioff + size_diff);
+	}
+
+	if (leaf->map_token) {
+		unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+		leaf->map_token = NULL;
+	}
+
+	/* shift the data */
+	if (from_end) {
+		memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+			      data_end + size_diff, btrfs_leaf_data(leaf) +
+			      data_end, old_data_start + new_size - data_end);
+	} else {
+		struct btrfs_disk_key disk_key;
+		u64 offset;
+
+		btrfs_item_key(leaf, &disk_key, slot);
+
+		if (btrfs_disk_key_type(&disk_key) == BTRFS_EXTENT_DATA_KEY) {
+			unsigned long ptr;
+			struct btrfs_file_extent_item *fi;
+
+			fi = btrfs_item_ptr(leaf, slot,
+					    struct btrfs_file_extent_item);
+			fi = (struct btrfs_file_extent_item *)(
+			     (unsigned long)fi - size_diff);
+
+			if (btrfs_file_extent_type(leaf, fi) ==
+			    BTRFS_FILE_EXTENT_INLINE) {
+				ptr = btrfs_item_ptr_offset(leaf, slot);
+				memmove_extent_buffer(leaf, ptr,
+				      (unsigned long)fi,
+				      offsetof(struct btrfs_file_extent_item,
+						 disk_bytenr));
+			}
+		}
+
+		memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+			      data_end + size_diff, btrfs_leaf_data(leaf) +
+			      data_end, old_data_start - data_end);
+
+		offset = btrfs_disk_key_offset(&disk_key);
+		btrfs_set_disk_key_offset(&disk_key, offset + size_diff);
+		btrfs_set_item_key(leaf, &disk_key, slot);
+		if (slot == 0)
+			fixup_low_keys(trans, root, path, &disk_key, 1);
+	}
+
+	item = btrfs_item_nr(leaf, slot);
+	btrfs_set_item_size(leaf, item, new_size);
+	btrfs_mark_buffer_dirty(leaf);
+
+	ret = 0;
+	if (btrfs_leaf_free_space(root, leaf) < 0) {
+		btrfs_print_leaf(root, leaf);
+		BUG();
+	}
+	return ret;
+}
+
+/*
+ * make the item pointed to by the path bigger, data_size is the new size.
+ */
+int btrfs_extend_item(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root, struct btrfs_path *path,
+		      u32 data_size)
+{
+	int ret = 0;
+	int slot;
+	int slot_orig;
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	u32 nritems;
+	unsigned int data_end;
+	unsigned int old_data;
+	unsigned int old_size;
+	int i;
+
+	slot_orig = path->slots[0];
+	leaf = path->nodes[0];
+
+	nritems = btrfs_header_nritems(leaf);
+	data_end = leaf_data_end(root, leaf);
+
+	if (btrfs_leaf_free_space(root, leaf) < data_size) {
+		btrfs_print_leaf(root, leaf);
+		BUG();
+	}
+	slot = path->slots[0];
+	old_data = btrfs_item_end_nr(leaf, slot);
+
+	BUG_ON(slot < 0);
+	if (slot >= nritems) {
+		btrfs_print_leaf(root, leaf);
+		printk(KERN_CRIT "slot %d too large, nritems %d\n",
+		       slot, nritems);
+		BUG_ON(1);
+	}
+
+	/*
+	 * item0..itemN ... dataN.offset..dataN.size .. data0.size
+	 */
+	/* first correct the data pointers */
+	for (i = slot; i < nritems; i++) {
+		u32 ioff;
+		item = btrfs_item_nr(leaf, i);
+
+		if (!leaf->map_token) {
+			map_extent_buffer(leaf, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&leaf->map_token, &leaf->kaddr,
+					&leaf->map_start, &leaf->map_len,
+					KM_USER1);
+		}
+		ioff = btrfs_item_offset(leaf, item);
+		btrfs_set_item_offset(leaf, item, ioff - data_size);
+	}
+
+	if (leaf->map_token) {
+		unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+		leaf->map_token = NULL;
+	}
+
+	/* shift the data */
+	memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+		      data_end - data_size, btrfs_leaf_data(leaf) +
+		      data_end, old_data - data_end);
+
+	data_end = old_data;
+	old_size = btrfs_item_size_nr(leaf, slot);
+	item = btrfs_item_nr(leaf, slot);
+	btrfs_set_item_size(leaf, item, old_size + data_size);
+	btrfs_mark_buffer_dirty(leaf);
+
+	ret = 0;
+	if (btrfs_leaf_free_space(root, leaf) < 0) {
+		btrfs_print_leaf(root, leaf);
+		BUG();
+	}
+	return ret;
+}
+
+/*
+ * Given a key and some data, insert items into the tree.
+ * This does all the path init required, making room in the tree if needed.
+ * Returns the number of keys that were inserted.
+ */
+int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path,
+			    struct btrfs_key *cpu_key, u32 *data_size,
+			    int nr)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	int ret = 0;
+	int slot;
+	int i;
+	u32 nritems;
+	u32 total_data = 0;
+	u32 total_size = 0;
+	unsigned int data_end;
+	struct btrfs_disk_key disk_key;
+	struct btrfs_key found_key;
+
+	for (i = 0; i < nr; i++) {
+		if (total_size + data_size[i] + sizeof(struct btrfs_item) >
+		    BTRFS_LEAF_DATA_SIZE(root)) {
+			break;
+			nr = i;
+		}
+		total_data += data_size[i];
+		total_size += data_size[i] + sizeof(struct btrfs_item);
+	}
+	BUG_ON(nr == 0);
+
+	ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
+	if (ret == 0)
+		return -EEXIST;
+	if (ret < 0)
+		goto out;
+
+	leaf = path->nodes[0];
+
+	nritems = btrfs_header_nritems(leaf);
+	data_end = leaf_data_end(root, leaf);
+
+	if (btrfs_leaf_free_space(root, leaf) < total_size) {
+		for (i = nr; i >= 0; i--) {
+			total_data -= data_size[i];
+			total_size -= data_size[i] + sizeof(struct btrfs_item);
+			if (total_size < btrfs_leaf_free_space(root, leaf))
+				break;
+		}
+		nr = i;
+	}
+
+	slot = path->slots[0];
+	BUG_ON(slot < 0);
+
+	if (slot != nritems) {
+		unsigned int old_data = btrfs_item_end_nr(leaf, slot);
+
+		item = btrfs_item_nr(leaf, slot);
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		/* figure out how many keys we can insert in here */
+		total_data = data_size[0];
+		for (i = 1; i < nr; i++) {
+			if (comp_cpu_keys(&found_key, cpu_key + i) <= 0)
+				break;
+			total_data += data_size[i];
+		}
+		nr = i;
+
+		if (old_data < data_end) {
+			btrfs_print_leaf(root, leaf);
+			printk(KERN_CRIT "slot %d old_data %d data_end %d\n",
+			       slot, old_data, data_end);
+			BUG_ON(1);
+		}
+		/*
+		 * item0..itemN ... dataN.offset..dataN.size .. data0.size
+		 */
+		/* first correct the data pointers */
+		WARN_ON(leaf->map_token);
+		for (i = slot; i < nritems; i++) {
+			u32 ioff;
+
+			item = btrfs_item_nr(leaf, i);
+			if (!leaf->map_token) {
+				map_extent_buffer(leaf, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&leaf->map_token, &leaf->kaddr,
+					&leaf->map_start, &leaf->map_len,
+					KM_USER1);
+			}
+
+			ioff = btrfs_item_offset(leaf, item);
+			btrfs_set_item_offset(leaf, item, ioff - total_data);
+		}
+		if (leaf->map_token) {
+			unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+			leaf->map_token = NULL;
+		}
+
+		/* shift the items */
+		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
+			      btrfs_item_nr_offset(slot),
+			      (nritems - slot) * sizeof(struct btrfs_item));
+
+		/* shift the data */
+		memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+			      data_end - total_data, btrfs_leaf_data(leaf) +
+			      data_end, old_data - data_end);
+		data_end = old_data;
+	} else {
+		/*
+		 * this sucks but it has to be done, if we are inserting at
+		 * the end of the leaf only insert 1 of the items, since we
+		 * have no way of knowing whats on the next leaf and we'd have
+		 * to drop our current locks to figure it out
+		 */
+		nr = 1;
+	}
+
+	/* setup the item for the new data */
+	for (i = 0; i < nr; i++) {
+		btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
+		btrfs_set_item_key(leaf, &disk_key, slot + i);
+		item = btrfs_item_nr(leaf, slot + i);
+		btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
+		data_end -= data_size[i];
+		btrfs_set_item_size(leaf, item, data_size[i]);
+	}
+	btrfs_set_header_nritems(leaf, nritems + nr);
+	btrfs_mark_buffer_dirty(leaf);
+
+	ret = 0;
+	if (slot == 0) {
+		btrfs_cpu_key_to_disk(&disk_key, cpu_key);
+		ret = fixup_low_keys(trans, root, path, &disk_key, 1);
+	}
+
+	if (btrfs_leaf_free_space(root, leaf) < 0) {
+		btrfs_print_leaf(root, leaf);
+		BUG();
+	}
+out:
+	if (!ret)
+		ret = nr;
+	return ret;
+}
+
+/*
+ * Given a key and some data, insert items into the tree.
+ * This does all the path init required, making room in the tree if needed.
+ */
+int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path,
+			    struct btrfs_key *cpu_key, u32 *data_size,
+			    int nr)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	int ret = 0;
+	int slot;
+	int slot_orig;
+	int i;
+	u32 nritems;
+	u32 total_size = 0;
+	u32 total_data = 0;
+	unsigned int data_end;
+	struct btrfs_disk_key disk_key;
+
+	for (i = 0; i < nr; i++)
+		total_data += data_size[i];
+
+	total_size = total_data + (nr * sizeof(struct btrfs_item));
+	ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
+	if (ret == 0)
+		return -EEXIST;
+	if (ret < 0)
+		goto out;
+
+	slot_orig = path->slots[0];
+	leaf = path->nodes[0];
+
+	nritems = btrfs_header_nritems(leaf);
+	data_end = leaf_data_end(root, leaf);
+
+	if (btrfs_leaf_free_space(root, leaf) < total_size) {
+		btrfs_print_leaf(root, leaf);
+		printk(KERN_CRIT "not enough freespace need %u have %d\n",
+		       total_size, btrfs_leaf_free_space(root, leaf));
+		BUG();
+	}
+
+	slot = path->slots[0];
+	BUG_ON(slot < 0);
+
+	if (slot != nritems) {
+		unsigned int old_data = btrfs_item_end_nr(leaf, slot);
+
+		if (old_data < data_end) {
+			btrfs_print_leaf(root, leaf);
+			printk(KERN_CRIT "slot %d old_data %d data_end %d\n",
+			       slot, old_data, data_end);
+			BUG_ON(1);
+		}
+		/*
+		 * item0..itemN ... dataN.offset..dataN.size .. data0.size
+		 */
+		/* first correct the data pointers */
+		WARN_ON(leaf->map_token);
+		for (i = slot; i < nritems; i++) {
+			u32 ioff;
+
+			item = btrfs_item_nr(leaf, i);
+			if (!leaf->map_token) {
+				map_extent_buffer(leaf, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&leaf->map_token, &leaf->kaddr,
+					&leaf->map_start, &leaf->map_len,
+					KM_USER1);
+			}
+
+			ioff = btrfs_item_offset(leaf, item);
+			btrfs_set_item_offset(leaf, item, ioff - total_data);
+		}
+		if (leaf->map_token) {
+			unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+			leaf->map_token = NULL;
+		}
+
+		/* shift the items */
+		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
+			      btrfs_item_nr_offset(slot),
+			      (nritems - slot) * sizeof(struct btrfs_item));
+
+		/* shift the data */
+		memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+			      data_end - total_data, btrfs_leaf_data(leaf) +
+			      data_end, old_data - data_end);
+		data_end = old_data;
+	}
+
+	/* setup the item for the new data */
+	for (i = 0; i < nr; i++) {
+		btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
+		btrfs_set_item_key(leaf, &disk_key, slot + i);
+		item = btrfs_item_nr(leaf, slot + i);
+		btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
+		data_end -= data_size[i];
+		btrfs_set_item_size(leaf, item, data_size[i]);
+	}
+	btrfs_set_header_nritems(leaf, nritems + nr);
+	btrfs_mark_buffer_dirty(leaf);
+
+	ret = 0;
+	if (slot == 0) {
+		btrfs_cpu_key_to_disk(&disk_key, cpu_key);
+		ret = fixup_low_keys(trans, root, path, &disk_key, 1);
+	}
+
+	if (btrfs_leaf_free_space(root, leaf) < 0) {
+		btrfs_print_leaf(root, leaf);
+		BUG();
+	}
+out:
+	return ret;
+}
+
+/*
+ * Given a key and some data, insert an item into the tree.
+ * This does all the path init required, making room in the tree if needed.
+ */
+int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *cpu_key, void *data, u32
+		      data_size)
+{
+	int ret = 0;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	unsigned long ptr;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size);
+	if (!ret) {
+		leaf = path->nodes[0];
+		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+		write_extent_buffer(leaf, data, ptr, data_size);
+		btrfs_mark_buffer_dirty(leaf);
+	}
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * delete the pointer from a given node.
+ *
+ * the tree should have been previously balanced so the deletion does not
+ * empty a node.
+ */
+static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		   struct btrfs_path *path, int level, int slot)
+{
+	struct extent_buffer *parent = path->nodes[level];
+	u32 nritems;
+	int ret = 0;
+	int wret;
+
+	nritems = btrfs_header_nritems(parent);
+	if (slot != nritems - 1) {
+		memmove_extent_buffer(parent,
+			      btrfs_node_key_ptr_offset(slot),
+			      btrfs_node_key_ptr_offset(slot + 1),
+			      sizeof(struct btrfs_key_ptr) *
+			      (nritems - slot - 1));
+	}
+	nritems--;
+	btrfs_set_header_nritems(parent, nritems);
+	if (nritems == 0 && parent == root->node) {
+		BUG_ON(btrfs_header_level(root->node) != 1);
+		/* just turn the root into a leaf and break */
+		btrfs_set_header_level(root->node, 0);
+	} else if (slot == 0) {
+		struct btrfs_disk_key disk_key;
+
+		btrfs_node_key(parent, &disk_key, 0);
+		wret = fixup_low_keys(trans, root, path, &disk_key, level + 1);
+		if (wret)
+			ret = wret;
+	}
+	btrfs_mark_buffer_dirty(parent);
+	return ret;
+}
+
+/*
+ * a helper function to delete the leaf pointed to by path->slots[1] and
+ * path->nodes[1].  bytenr is the node block pointer, but since the callers
+ * already know it, it is faster to have them pass it down than to
+ * read it out of the node again.
+ *
+ * This deletes the pointer in path->nodes[1] and frees the leaf
+ * block extent.  zero is returned if it all worked out, < 0 otherwise.
+ *
+ * The path must have already been setup for deleting the leaf, including
+ * all the proper balancing.  path->nodes[1] must be locked.
+ */
+noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path, u64 bytenr)
+{
+	int ret;
+	u64 root_gen = btrfs_header_generation(path->nodes[1]);
+
+	ret = del_ptr(trans, root, path, 1, path->slots[1]);
+	if (ret)
+		return ret;
+
+	ret = btrfs_free_extent(trans, root, bytenr,
+				btrfs_level_size(root, 0),
+				path->nodes[1]->start,
+				btrfs_header_owner(path->nodes[1]),
+				root_gen, 0, 1);
+	return ret;
+}
+/*
+ * delete the item at the leaf level in path.  If that empties
+ * the leaf, remove it from the tree
+ */
+int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		    struct btrfs_path *path, int slot, int nr)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	int last_off;
+	int dsize = 0;
+	int ret = 0;
+	int wret;
+	int i;
+	u32 nritems;
+
+	leaf = path->nodes[0];
+	last_off = btrfs_item_offset_nr(leaf, slot + nr - 1);
+
+	for (i = 0; i < nr; i++)
+		dsize += btrfs_item_size_nr(leaf, slot + i);
+
+	nritems = btrfs_header_nritems(leaf);
+
+	if (slot + nr != nritems) {
+		int data_end = leaf_data_end(root, leaf);
+
+		memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+			      data_end + dsize,
+			      btrfs_leaf_data(leaf) + data_end,
+			      last_off - data_end);
+
+		for (i = slot + nr; i < nritems; i++) {
+			u32 ioff;
+
+			item = btrfs_item_nr(leaf, i);
+			if (!leaf->map_token) {
+				map_extent_buffer(leaf, (unsigned long)item,
+					sizeof(struct btrfs_item),
+					&leaf->map_token, &leaf->kaddr,
+					&leaf->map_start, &leaf->map_len,
+					KM_USER1);
+			}
+			ioff = btrfs_item_offset(leaf, item);
+			btrfs_set_item_offset(leaf, item, ioff + dsize);
+		}
+
+		if (leaf->map_token) {
+			unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+			leaf->map_token = NULL;
+		}
+
+		memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot),
+			      btrfs_item_nr_offset(slot + nr),
+			      sizeof(struct btrfs_item) *
+			      (nritems - slot - nr));
+	}
+	btrfs_set_header_nritems(leaf, nritems - nr);
+	nritems -= nr;
+
+	/* delete the leaf if we've emptied it */
+	if (nritems == 0) {
+		if (leaf == root->node) {
+			btrfs_set_header_level(leaf, 0);
+		} else {
+			ret = btrfs_del_leaf(trans, root, path, leaf->start);
+			BUG_ON(ret);
+		}
+	} else {
+		int used = leaf_space_used(leaf, 0, nritems);
+		if (slot == 0) {
+			struct btrfs_disk_key disk_key;
+
+			btrfs_item_key(leaf, &disk_key, 0);
+			wret = fixup_low_keys(trans, root, path,
+					      &disk_key, 1);
+			if (wret)
+				ret = wret;
+		}
+
+		/* delete the leaf if it is mostly empty */
+		if (used < BTRFS_LEAF_DATA_SIZE(root) / 4) {
+			/* push_leaf_left fixes the path.
+			 * make sure the path still points to our leaf
+			 * for possible call to del_ptr below
+			 */
+			slot = path->slots[1];
+			extent_buffer_get(leaf);
+
+			wret = push_leaf_left(trans, root, path, 1, 1);
+			if (wret < 0 && wret != -ENOSPC)
+				ret = wret;
+
+			if (path->nodes[0] == leaf &&
+			    btrfs_header_nritems(leaf)) {
+				wret = push_leaf_right(trans, root, path, 1, 1);
+				if (wret < 0 && wret != -ENOSPC)
+					ret = wret;
+			}
+
+			if (btrfs_header_nritems(leaf) == 0) {
+				path->slots[1] = slot;
+				ret = btrfs_del_leaf(trans, root, path,
+						     leaf->start);
+				BUG_ON(ret);
+				free_extent_buffer(leaf);
+			} else {
+				/* if we're still in the path, make sure
+				 * we're dirty.  Otherwise, one of the
+				 * push_leaf functions must have already
+				 * dirtied this buffer
+				 */
+				if (path->nodes[0] == leaf)
+					btrfs_mark_buffer_dirty(leaf);
+				free_extent_buffer(leaf);
+			}
+		} else {
+			btrfs_mark_buffer_dirty(leaf);
+		}
+	}
+	return ret;
+}
+
+/*
+ * search the tree again to find a leaf with lesser keys
+ * returns 0 if it found something or 1 if there are no lesser leaves.
+ * returns < 0 on io errors.
+ *
+ * This may release the path, and so you may lose any locks held at the
+ * time you call it.
+ */
+int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
+{
+	struct btrfs_key key;
+	struct btrfs_disk_key found_key;
+	int ret;
+
+	btrfs_item_key_to_cpu(path->nodes[0], &key, 0);
+
+	if (key.offset > 0)
+		key.offset--;
+	else if (key.type > 0)
+		key.type--;
+	else if (key.objectid > 0)
+		key.objectid--;
+	else
+		return 1;
+
+	btrfs_release_path(root, path);
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		return ret;
+	btrfs_item_key(path->nodes[0], &found_key, 0);
+	ret = comp_keys(&found_key, &key);
+	if (ret < 0)
+		return 0;
+	return 1;
+}
+
+/*
+ * A helper function to walk down the tree starting at min_key, and looking
+ * for nodes or leaves that are either in cache or have a minimum
+ * transaction id.  This is used by the btree defrag code, and tree logging
+ *
+ * This does not cow, but it does stuff the starting key it finds back
+ * into min_key, so you can call btrfs_search_slot with cow=1 on the
+ * key and get a writable path.
+ *
+ * This does lock as it descends, and path->keep_locks should be set
+ * to 1 by the caller.
+ *
+ * This honors path->lowest_level to prevent descent past a given level
+ * of the tree.
+ *
+ * min_trans indicates the oldest transaction that you are interested
+ * in walking through.  Any nodes or leaves older than min_trans are
+ * skipped over (without reading them).
+ *
+ * returns zero if something useful was found, < 0 on error and 1 if there
+ * was nothing in the tree that matched the search criteria.
+ */
+int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
+			 struct btrfs_key *max_key,
+			 struct btrfs_path *path, int cache_only,
+			 u64 min_trans)
+{
+	struct extent_buffer *cur;
+	struct btrfs_key found_key;
+	int slot;
+	int sret;
+	u32 nritems;
+	int level;
+	int ret = 1;
+
+	WARN_ON(!path->keep_locks);
+again:
+	cur = btrfs_lock_root_node(root);
+	level = btrfs_header_level(cur);
+	WARN_ON(path->nodes[level]);
+	path->nodes[level] = cur;
+	path->locks[level] = 1;
+
+	if (btrfs_header_generation(cur) < min_trans) {
+		ret = 1;
+		goto out;
+	}
+	while (1) {
+		nritems = btrfs_header_nritems(cur);
+		level = btrfs_header_level(cur);
+		sret = bin_search(cur, min_key, level, &slot);
+
+		/* at the lowest level, we're done, setup the path and exit */
+		if (level == path->lowest_level) {
+			if (slot >= nritems)
+				goto find_next_key;
+			ret = 0;
+			path->slots[level] = slot;
+			btrfs_item_key_to_cpu(cur, &found_key, slot);
+			goto out;
+		}
+		if (sret && slot > 0)
+			slot--;
+		/*
+		 * check this node pointer against the cache_only and
+		 * min_trans parameters.  If it isn't in cache or is too
+		 * old, skip to the next one.
+		 */
+		while (slot < nritems) {
+			u64 blockptr;
+			u64 gen;
+			struct extent_buffer *tmp;
+			struct btrfs_disk_key disk_key;
+
+			blockptr = btrfs_node_blockptr(cur, slot);
+			gen = btrfs_node_ptr_generation(cur, slot);
+			if (gen < min_trans) {
+				slot++;
+				continue;
+			}
+			if (!cache_only)
+				break;
+
+			if (max_key) {
+				btrfs_node_key(cur, &disk_key, slot);
+				if (comp_keys(&disk_key, max_key) >= 0) {
+					ret = 1;
+					goto out;
+				}
+			}
+
+			tmp = btrfs_find_tree_block(root, blockptr,
+					    btrfs_level_size(root, level - 1));
+
+			if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+				free_extent_buffer(tmp);
+				break;
+			}
+			if (tmp)
+				free_extent_buffer(tmp);
+			slot++;
+		}
+find_next_key:
+		/*
+		 * we didn't find a candidate key in this node, walk forward
+		 * and find another one
+		 */
+		if (slot >= nritems) {
+			path->slots[level] = slot;
+			sret = btrfs_find_next_key(root, path, min_key, level,
+						  cache_only, min_trans);
+			if (sret == 0) {
+				btrfs_release_path(root, path);
+				goto again;
+			} else {
+				goto out;
+			}
+		}
+		/* save our key for returning back */
+		btrfs_node_key_to_cpu(cur, &found_key, slot);
+		path->slots[level] = slot;
+		if (level == path->lowest_level) {
+			ret = 0;
+			unlock_up(path, level, 1);
+			goto out;
+		}
+		cur = read_node_slot(root, cur, slot);
+
+		btrfs_tree_lock(cur);
+		path->locks[level - 1] = 1;
+		path->nodes[level - 1] = cur;
+		unlock_up(path, level, 1);
+	}
+out:
+	if (ret == 0)
+		memcpy(min_key, &found_key, sizeof(found_key));
+	return ret;
+}
+
+/*
+ * this is similar to btrfs_next_leaf, but does not try to preserve
+ * and fixup the path.  It looks for and returns the next key in the
+ * tree based on the current path and the cache_only and min_trans
+ * parameters.
+ *
+ * 0 is returned if another key is found, < 0 if there are any errors
+ * and 1 is returned if there are no higher keys in the tree
+ *
+ * path->keep_locks should be set to 1 on the search made before
+ * calling this function.
+ */
+int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
+			struct btrfs_key *key, int lowest_level,
+			int cache_only, u64 min_trans)
+{
+	int level = lowest_level;
+	int slot;
+	struct extent_buffer *c;
+
+	WARN_ON(!path->keep_locks);
+	while (level < BTRFS_MAX_LEVEL) {
+		if (!path->nodes[level])
+			return 1;
+
+		slot = path->slots[level] + 1;
+		c = path->nodes[level];
+next:
+		if (slot >= btrfs_header_nritems(c)) {
+			level++;
+			if (level == BTRFS_MAX_LEVEL)
+				return 1;
+			continue;
+		}
+		if (level == 0)
+			btrfs_item_key_to_cpu(c, key, slot);
+		else {
+			u64 blockptr = btrfs_node_blockptr(c, slot);
+			u64 gen = btrfs_node_ptr_generation(c, slot);
+
+			if (cache_only) {
+				struct extent_buffer *cur;
+				cur = btrfs_find_tree_block(root, blockptr,
+					    btrfs_level_size(root, level - 1));
+				if (!cur || !btrfs_buffer_uptodate(cur, gen)) {
+					slot++;
+					if (cur)
+						free_extent_buffer(cur);
+					goto next;
+				}
+				free_extent_buffer(cur);
+			}
+			if (gen < min_trans) {
+				slot++;
+				goto next;
+			}
+			btrfs_node_key_to_cpu(c, key, slot);
+		}
+		return 0;
+	}
+	return 1;
+}
+
+/*
+ * search the tree again to find a leaf with greater keys
+ * returns 0 if it found something or 1 if there are no greater leaves.
+ * returns < 0 on io errors.
+ */
+int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
+{
+	int slot;
+	int level = 1;
+	struct extent_buffer *c;
+	struct extent_buffer *next = NULL;
+	struct btrfs_key key;
+	u32 nritems;
+	int ret;
+
+	nritems = btrfs_header_nritems(path->nodes[0]);
+	if (nritems == 0)
+		return 1;
+
+	btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1);
+
+	btrfs_release_path(root, path);
+	path->keep_locks = 1;
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	path->keep_locks = 0;
+
+	if (ret < 0)
+		return ret;
+
+	nritems = btrfs_header_nritems(path->nodes[0]);
+	/*
+	 * by releasing the path above we dropped all our locks.  A balance
+	 * could have added more items next to the key that used to be
+	 * at the very end of the block.  So, check again here and
+	 * advance the path if there are now more items available.
+	 */
+	if (nritems > 0 && path->slots[0] < nritems - 1) {
+		path->slots[0]++;
+		goto done;
+	}
+
+	while (level < BTRFS_MAX_LEVEL) {
+		if (!path->nodes[level])
+			return 1;
+
+		slot = path->slots[level] + 1;
+		c = path->nodes[level];
+		if (slot >= btrfs_header_nritems(c)) {
+			level++;
+			if (level == BTRFS_MAX_LEVEL)
+				return 1;
+			continue;
+		}
+
+		if (next) {
+			btrfs_tree_unlock(next);
+			free_extent_buffer(next);
+		}
+
+		if (level == 1 && (path->locks[1] || path->skip_locking) &&
+		    path->reada)
+			reada_for_search(root, path, level, slot, 0);
+
+		next = read_node_slot(root, c, slot);
+		if (!path->skip_locking) {
+			WARN_ON(!btrfs_tree_locked(c));
+			btrfs_tree_lock(next);
+		}
+		break;
+	}
+	path->slots[level] = slot;
+	while (1) {
+		level--;
+		c = path->nodes[level];
+		if (path->locks[level])
+			btrfs_tree_unlock(c);
+		free_extent_buffer(c);
+		path->nodes[level] = next;
+		path->slots[level] = 0;
+		if (!path->skip_locking)
+			path->locks[level] = 1;
+		if (!level)
+			break;
+		if (level == 1 && path->locks[1] && path->reada)
+			reada_for_search(root, path, level, slot, 0);
+		next = read_node_slot(root, next, 0);
+		if (!path->skip_locking) {
+			WARN_ON(!btrfs_tree_locked(path->nodes[level]));
+			btrfs_tree_lock(next);
+		}
+	}
+done:
+	unlock_up(path, 0, 1);
+	return 0;
+}
+
+/*
+ * this uses btrfs_prev_leaf to walk backwards in the tree, and keeps
+ * searching until it gets past min_objectid or finds an item of 'type'
+ *
+ * returns 0 if something is found, 1 if nothing was found and < 0 on error
+ */
+int btrfs_previous_item(struct btrfs_root *root,
+			struct btrfs_path *path, u64 min_objectid,
+			int type)
+{
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf;
+	u32 nritems;
+	int ret;
+
+	while (1) {
+		if (path->slots[0] == 0) {
+			ret = btrfs_prev_leaf(root, path);
+			if (ret != 0)
+				return ret;
+		} else {
+			path->slots[0]--;
+		}
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		if (nritems == 0)
+			return 1;
+		if (path->slots[0] == nritems)
+			path->slots[0]--;
+
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (found_key.type == type)
+			return 0;
+		if (found_key.objectid < min_objectid)
+			break;
+		if (found_key.objectid == min_objectid &&
+		    found_key.type < type)
+			break;
+	}
+	return 1;
+}
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
new file mode 100644
index 00000000000..eee060f8811
--- /dev/null
+++ b/fs/btrfs/ctree.h
@@ -0,0 +1,2129 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_CTREE__
+#define __BTRFS_CTREE__
+
+#include <linux/version.h>
+#include <linux/mm.h>
+#include <linux/highmem.h>
+#include <linux/fs.h>
+#include <linux/completion.h>
+#include <linux/backing-dev.h>
+#include <linux/wait.h>
+#include <asm/kmap_types.h>
+#include "extent_io.h"
+#include "extent_map.h"
+#include "async-thread.h"
+
+struct btrfs_trans_handle;
+struct btrfs_transaction;
+extern struct kmem_cache *btrfs_trans_handle_cachep;
+extern struct kmem_cache *btrfs_transaction_cachep;
+extern struct kmem_cache *btrfs_bit_radix_cachep;
+extern struct kmem_cache *btrfs_path_cachep;
+struct btrfs_ordered_sum;
+
+#define BTRFS_MAGIC "_BHRfS_M"
+
+#define BTRFS_ACL_NOT_CACHED    ((void *)-1)
+
+#ifdef CONFIG_LOCKDEP
+# define BTRFS_MAX_LEVEL 7
+#else
+# define BTRFS_MAX_LEVEL 8
+#endif
+
+/* holds pointers to all of the tree roots */
+#define BTRFS_ROOT_TREE_OBJECTID 1ULL
+
+/* stores information about which extents are in use, and reference counts */
+#define BTRFS_EXTENT_TREE_OBJECTID 2ULL
+
+/*
+ * chunk tree stores translations from logical -> physical block numbering
+ * the super block points to the chunk tree
+ */
+#define BTRFS_CHUNK_TREE_OBJECTID 3ULL
+
+/*
+ * stores information about which areas of a given device are in use.
+ * one per device.  The tree of tree roots points to the device tree
+ */
+#define BTRFS_DEV_TREE_OBJECTID 4ULL
+
+/* one per subvolume, storing files and directories */
+#define BTRFS_FS_TREE_OBJECTID 5ULL
+
+/* directory objectid inside the root tree */
+#define BTRFS_ROOT_TREE_DIR_OBJECTID 6ULL
+
+/* holds checksums of all the data extents */
+#define BTRFS_CSUM_TREE_OBJECTID 7ULL
+
+/* orhpan objectid for tracking unlinked/truncated files */
+#define BTRFS_ORPHAN_OBJECTID -5ULL
+
+/* does write ahead logging to speed up fsyncs */
+#define BTRFS_TREE_LOG_OBJECTID -6ULL
+#define BTRFS_TREE_LOG_FIXUP_OBJECTID -7ULL
+
+/* for space balancing */
+#define BTRFS_TREE_RELOC_OBJECTID -8ULL
+#define BTRFS_DATA_RELOC_TREE_OBJECTID -9ULL
+
+/*
+ * extent checksums all have this objectid
+ * this allows them to share the logging tree
+ * for fsyncs
+ */
+#define BTRFS_EXTENT_CSUM_OBJECTID -10ULL
+
+/* dummy objectid represents multiple objectids */
+#define BTRFS_MULTIPLE_OBJECTIDS -255ULL
+
+/*
+ * All files have objectids in this range.
+ */
+#define BTRFS_FIRST_FREE_OBJECTID 256ULL
+#define BTRFS_LAST_FREE_OBJECTID -256ULL
+#define BTRFS_FIRST_CHUNK_TREE_OBJECTID 256ULL
+
+
+/*
+ * the device items go into the chunk tree.  The key is in the form
+ * [ 1 BTRFS_DEV_ITEM_KEY device_id ]
+ */
+#define BTRFS_DEV_ITEMS_OBJECTID 1ULL
+
+/*
+ * we can actually store much bigger names, but lets not confuse the rest
+ * of linux
+ */
+#define BTRFS_NAME_LEN 255
+
+/* 32 bytes in various csum fields */
+#define BTRFS_CSUM_SIZE 32
+
+/* csum types */
+#define BTRFS_CSUM_TYPE_CRC32	0
+
+static int btrfs_csum_sizes[] = { 4, 0 };
+
+/* four bytes for CRC32 */
+#define BTRFS_EMPTY_DIR_SIZE 0
+
+#define BTRFS_FT_UNKNOWN	0
+#define BTRFS_FT_REG_FILE	1
+#define BTRFS_FT_DIR		2
+#define BTRFS_FT_CHRDEV		3
+#define BTRFS_FT_BLKDEV		4
+#define BTRFS_FT_FIFO		5
+#define BTRFS_FT_SOCK		6
+#define BTRFS_FT_SYMLINK	7
+#define BTRFS_FT_XATTR		8
+#define BTRFS_FT_MAX		9
+
+/*
+ * the key defines the order in the tree, and so it also defines (optimal)
+ * block layout.  objectid corresonds to the inode number.  The flags
+ * tells us things about the object, and is a kind of stream selector.
+ * so for a given inode, keys with flags of 1 might refer to the inode
+ * data, flags of 2 may point to file data in the btree and flags == 3
+ * may point to extents.
+ *
+ * offset is the starting byte offset for this key in the stream.
+ *
+ * btrfs_disk_key is in disk byte order.  struct btrfs_key is always
+ * in cpu native order.  Otherwise they are identical and their sizes
+ * should be the same (ie both packed)
+ */
+struct btrfs_disk_key {
+	__le64 objectid;
+	u8 type;
+	__le64 offset;
+} __attribute__ ((__packed__));
+
+struct btrfs_key {
+	u64 objectid;
+	u8 type;
+	u64 offset;
+} __attribute__ ((__packed__));
+
+struct btrfs_mapping_tree {
+	struct extent_map_tree map_tree;
+};
+
+#define BTRFS_UUID_SIZE 16
+struct btrfs_dev_item {
+	/* the internal btrfs device id */
+	__le64 devid;
+
+	/* size of the device */
+	__le64 total_bytes;
+
+	/* bytes used */
+	__le64 bytes_used;
+
+	/* optimal io alignment for this device */
+	__le32 io_align;
+
+	/* optimal io width for this device */
+	__le32 io_width;
+
+	/* minimal io size for this device */
+	__le32 sector_size;
+
+	/* type and info about this device */
+	__le64 type;
+
+	/* expected generation for this device */
+	__le64 generation;
+
+	/*
+	 * starting byte of this partition on the device,
+	 * to allowr for stripe alignment in the future
+	 */
+	__le64 start_offset;
+
+	/* grouping information for allocation decisions */
+	__le32 dev_group;
+
+	/* seek speed 0-100 where 100 is fastest */
+	u8 seek_speed;
+
+	/* bandwidth 0-100 where 100 is fastest */
+	u8 bandwidth;
+
+	/* btrfs generated uuid for this device */
+	u8 uuid[BTRFS_UUID_SIZE];
+
+	/* uuid of FS who owns this device */
+	u8 fsid[BTRFS_UUID_SIZE];
+} __attribute__ ((__packed__));
+
+struct btrfs_stripe {
+	__le64 devid;
+	__le64 offset;
+	u8 dev_uuid[BTRFS_UUID_SIZE];
+} __attribute__ ((__packed__));
+
+struct btrfs_chunk {
+	/* size of this chunk in bytes */
+	__le64 length;
+
+	/* objectid of the root referencing this chunk */
+	__le64 owner;
+
+	__le64 stripe_len;
+	__le64 type;
+
+	/* optimal io alignment for this chunk */
+	__le32 io_align;
+
+	/* optimal io width for this chunk */
+	__le32 io_width;
+
+	/* minimal io size for this chunk */
+	__le32 sector_size;
+
+	/* 2^16 stripes is quite a lot, a second limit is the size of a single
+	 * item in the btree
+	 */
+	__le16 num_stripes;
+
+	/* sub stripes only matter for raid10 */
+	__le16 sub_stripes;
+	struct btrfs_stripe stripe;
+	/* additional stripes go here */
+} __attribute__ ((__packed__));
+
+static inline unsigned long btrfs_chunk_item_size(int num_stripes)
+{
+	BUG_ON(num_stripes == 0);
+	return sizeof(struct btrfs_chunk) +
+		sizeof(struct btrfs_stripe) * (num_stripes - 1);
+}
+
+#define BTRFS_FSID_SIZE 16
+#define BTRFS_HEADER_FLAG_WRITTEN (1 << 0)
+
+/*
+ * every tree block (leaf or node) starts with this header.
+ */
+struct btrfs_header {
+	/* these first four must match the super block */
+	u8 csum[BTRFS_CSUM_SIZE];
+	u8 fsid[BTRFS_FSID_SIZE]; /* FS specific uuid */
+	__le64 bytenr; /* which block this node is supposed to live in */
+	__le64 flags;
+
+	/* allowed to be different from the super from here on down */
+	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
+	__le64 generation;
+	__le64 owner;
+	__le32 nritems;
+	u8 level;
+} __attribute__ ((__packed__));
+
+#define BTRFS_NODEPTRS_PER_BLOCK(r) (((r)->nodesize - \
+				      sizeof(struct btrfs_header)) / \
+				     sizeof(struct btrfs_key_ptr))
+#define __BTRFS_LEAF_DATA_SIZE(bs) ((bs) - sizeof(struct btrfs_header))
+#define BTRFS_LEAF_DATA_SIZE(r) (__BTRFS_LEAF_DATA_SIZE(r->leafsize))
+#define BTRFS_MAX_INLINE_DATA_SIZE(r) (BTRFS_LEAF_DATA_SIZE(r) - \
+					sizeof(struct btrfs_item) - \
+					sizeof(struct btrfs_file_extent_item))
+
+#define BTRFS_SUPER_FLAG_SEEDING (1ULL << 32)
+
+/*
+ * this is a very generous portion of the super block, giving us
+ * room to translate 14 chunks with 3 stripes each.
+ */
+#define BTRFS_SYSTEM_CHUNK_ARRAY_SIZE 2048
+#define BTRFS_LABEL_SIZE 256
+
+/*
+ * the super block basically lists the main trees of the FS
+ * it currently lacks any block count etc etc
+ */
+struct btrfs_super_block {
+	u8 csum[BTRFS_CSUM_SIZE];
+	/* the first 4 fields must match struct btrfs_header */
+	u8 fsid[BTRFS_FSID_SIZE];    /* FS specific uuid */
+	__le64 bytenr; /* this block number */
+	__le64 flags;
+
+	/* allowed to be different from the btrfs_header from here own down */
+	__le64 magic;
+	__le64 generation;
+	__le64 root;
+	__le64 chunk_root;
+	__le64 log_root;
+
+	/* this will help find the new super based on the log root */
+	__le64 log_root_transid;
+	__le64 total_bytes;
+	__le64 bytes_used;
+	__le64 root_dir_objectid;
+	__le64 num_devices;
+	__le32 sectorsize;
+	__le32 nodesize;
+	__le32 leafsize;
+	__le32 stripesize;
+	__le32 sys_chunk_array_size;
+	__le64 chunk_root_generation;
+	__le64 compat_flags;
+	__le64 compat_ro_flags;
+	__le64 incompat_flags;
+	__le16 csum_type;
+	u8 root_level;
+	u8 chunk_root_level;
+	u8 log_root_level;
+	struct btrfs_dev_item dev_item;
+
+	char label[BTRFS_LABEL_SIZE];
+
+	/* future expansion */
+	__le64 reserved[32];
+	u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE];
+} __attribute__ ((__packed__));
+
+/*
+ * Compat flags that we support.  If any incompat flags are set other than the
+ * ones specified below then we will fail to mount
+ */
+#define BTRFS_FEATURE_COMPAT_SUPP	0x0
+#define BTRFS_FEATURE_COMPAT_RO_SUPP	0x0
+#define BTRFS_FEATURE_INCOMPAT_SUPP	0x0
+
+/*
+ * A leaf is full of items. offset and size tell us where to find
+ * the item in the leaf (relative to the start of the data area)
+ */
+struct btrfs_item {
+	struct btrfs_disk_key key;
+	__le32 offset;
+	__le32 size;
+} __attribute__ ((__packed__));
+
+/*
+ * leaves have an item area and a data area:
+ * [item0, item1....itemN] [free space] [dataN...data1, data0]
+ *
+ * The data is separate from the items to get the keys closer together
+ * during searches.
+ */
+struct btrfs_leaf {
+	struct btrfs_header header;
+	struct btrfs_item items[];
+} __attribute__ ((__packed__));
+
+/*
+ * all non-leaf blocks are nodes, they hold only keys and pointers to
+ * other blocks
+ */
+struct btrfs_key_ptr {
+	struct btrfs_disk_key key;
+	__le64 blockptr;
+	__le64 generation;
+} __attribute__ ((__packed__));
+
+struct btrfs_node {
+	struct btrfs_header header;
+	struct btrfs_key_ptr ptrs[];
+} __attribute__ ((__packed__));
+
+/*
+ * btrfs_paths remember the path taken from the root down to the leaf.
+ * level 0 is always the leaf, and nodes[1...BTRFS_MAX_LEVEL] will point
+ * to any other levels that are present.
+ *
+ * The slots array records the index of the item or block pointer
+ * used while walking the tree.
+ */
+struct btrfs_path {
+	struct extent_buffer *nodes[BTRFS_MAX_LEVEL];
+	int slots[BTRFS_MAX_LEVEL];
+	/* if there is real range locking, this locks field will change */
+	int locks[BTRFS_MAX_LEVEL];
+	int reada;
+	/* keep some upper locks as we walk down */
+	int keep_locks;
+	int skip_locking;
+	int lowest_level;
+
+	/*
+	 * set by btrfs_split_item, tells search_slot to keep all locks
+	 * and to force calls to keep space in the nodes
+	 */
+	int search_for_split;
+};
+
+/*
+ * items in the extent btree are used to record the objectid of the
+ * owner of the block and the number of references
+ */
+struct btrfs_extent_item {
+	__le32 refs;
+} __attribute__ ((__packed__));
+
+struct btrfs_extent_ref {
+	__le64 root;
+	__le64 generation;
+	__le64 objectid;
+	__le32 num_refs;
+} __attribute__ ((__packed__));
+
+/* dev extents record free space on individual devices.  The owner
+ * field points back to the chunk allocation mapping tree that allocated
+ * the extent.  The chunk tree uuid field is a way to double check the owner
+ */
+struct btrfs_dev_extent {
+	__le64 chunk_tree;
+	__le64 chunk_objectid;
+	__le64 chunk_offset;
+	__le64 length;
+	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
+} __attribute__ ((__packed__));
+
+struct btrfs_inode_ref {
+	__le64 index;
+	__le16 name_len;
+	/* name goes here */
+} __attribute__ ((__packed__));
+
+struct btrfs_timespec {
+	__le64 sec;
+	__le32 nsec;
+} __attribute__ ((__packed__));
+
+typedef enum {
+	BTRFS_COMPRESS_NONE = 0,
+	BTRFS_COMPRESS_ZLIB = 1,
+	BTRFS_COMPRESS_LAST = 2,
+} btrfs_compression_type;
+
+/* we don't understand any encryption methods right now */
+typedef enum {
+	BTRFS_ENCRYPTION_NONE = 0,
+	BTRFS_ENCRYPTION_LAST = 1,
+} btrfs_encryption_type;
+
+struct btrfs_inode_item {
+	/* nfs style generation number */
+	__le64 generation;
+	/* transid that last touched this inode */
+	__le64 transid;
+	__le64 size;
+	__le64 nbytes;
+	__le64 block_group;
+	__le32 nlink;
+	__le32 uid;
+	__le32 gid;
+	__le32 mode;
+	__le64 rdev;
+	__le64 flags;
+
+	/* modification sequence number for NFS */
+	__le64 sequence;
+
+	/*
+	 * a little future expansion, for more than this we can
+	 * just grow the inode item and version it
+	 */
+	__le64 reserved[4];
+	struct btrfs_timespec atime;
+	struct btrfs_timespec ctime;
+	struct btrfs_timespec mtime;
+	struct btrfs_timespec otime;
+} __attribute__ ((__packed__));
+
+struct btrfs_dir_log_item {
+	__le64 end;
+} __attribute__ ((__packed__));
+
+struct btrfs_dir_item {
+	struct btrfs_disk_key location;
+	__le64 transid;
+	__le16 data_len;
+	__le16 name_len;
+	u8 type;
+} __attribute__ ((__packed__));
+
+struct btrfs_root_item {
+	struct btrfs_inode_item inode;
+	__le64 generation;
+	__le64 root_dirid;
+	__le64 bytenr;
+	__le64 byte_limit;
+	__le64 bytes_used;
+	__le64 last_snapshot;
+	__le64 flags;
+	__le32 refs;
+	struct btrfs_disk_key drop_progress;
+	u8 drop_level;
+	u8 level;
+} __attribute__ ((__packed__));
+
+/*
+ * this is used for both forward and backward root refs
+ */
+struct btrfs_root_ref {
+	__le64 dirid;
+	__le64 sequence;
+	__le16 name_len;
+} __attribute__ ((__packed__));
+
+#define BTRFS_FILE_EXTENT_INLINE 0
+#define BTRFS_FILE_EXTENT_REG 1
+#define BTRFS_FILE_EXTENT_PREALLOC 2
+
+struct btrfs_file_extent_item {
+	/*
+	 * transaction id that created this extent
+	 */
+	__le64 generation;
+	/*
+	 * max number of bytes to hold this extent in ram
+	 * when we split a compressed extent we can't know how big
+	 * each of the resulting pieces will be.  So, this is
+	 * an upper limit on the size of the extent in ram instead of
+	 * an exact limit.
+	 */
+	__le64 ram_bytes;
+
+	/*
+	 * 32 bits for the various ways we might encode the data,
+	 * including compression and encryption.  If any of these
+	 * are set to something a given disk format doesn't understand
+	 * it is treated like an incompat flag for reading and writing,
+	 * but not for stat.
+	 */
+	u8 compression;
+	u8 encryption;
+	__le16 other_encoding; /* spare for later use */
+
+	/* are we inline data or a real extent? */
+	u8 type;
+
+	/*
+	 * disk space consumed by the extent, checksum blocks are included
+	 * in these numbers
+	 */
+	__le64 disk_bytenr;
+	__le64 disk_num_bytes;
+	/*
+	 * the logical offset in file blocks (no csums)
+	 * this extent record is for.  This allows a file extent to point
+	 * into the middle of an existing extent on disk, sharing it
+	 * between two snapshots (useful if some bytes in the middle of the
+	 * extent have changed
+	 */
+	__le64 offset;
+	/*
+	 * the logical number of file blocks (no csums included).  This
+	 * always reflects the size uncompressed and without encoding.
+	 */
+	__le64 num_bytes;
+
+} __attribute__ ((__packed__));
+
+struct btrfs_csum_item {
+	u8 csum;
+} __attribute__ ((__packed__));
+
+/* different types of block groups (and chunks) */
+#define BTRFS_BLOCK_GROUP_DATA     (1 << 0)
+#define BTRFS_BLOCK_GROUP_SYSTEM   (1 << 1)
+#define BTRFS_BLOCK_GROUP_METADATA (1 << 2)
+#define BTRFS_BLOCK_GROUP_RAID0    (1 << 3)
+#define BTRFS_BLOCK_GROUP_RAID1    (1 << 4)
+#define BTRFS_BLOCK_GROUP_DUP	   (1 << 5)
+#define BTRFS_BLOCK_GROUP_RAID10   (1 << 6)
+
+struct btrfs_block_group_item {
+	__le64 used;
+	__le64 chunk_objectid;
+	__le64 flags;
+} __attribute__ ((__packed__));
+
+struct btrfs_space_info {
+	u64 flags;
+	u64 total_bytes;
+	u64 bytes_used;
+	u64 bytes_pinned;
+	u64 bytes_reserved;
+	u64 bytes_readonly;
+	int full;
+	int force_alloc;
+	struct list_head list;
+
+	/* for block groups in our same type */
+	struct list_head block_groups;
+	spinlock_t lock;
+	struct rw_semaphore groups_sem;
+};
+
+struct btrfs_free_space {
+	struct rb_node bytes_index;
+	struct rb_node offset_index;
+	u64 offset;
+	u64 bytes;
+};
+
+struct btrfs_block_group_cache {
+	struct btrfs_key key;
+	struct btrfs_block_group_item item;
+	spinlock_t lock;
+	struct mutex alloc_mutex;
+	struct mutex cache_mutex;
+	u64 pinned;
+	u64 reserved;
+	u64 flags;
+	int cached;
+	int ro;
+	int dirty;
+
+	struct btrfs_space_info *space_info;
+
+	/* free space cache stuff */
+	struct rb_root free_space_bytes;
+	struct rb_root free_space_offset;
+
+	/* block group cache stuff */
+	struct rb_node cache_node;
+
+	/* for block groups in the same raid type */
+	struct list_head list;
+
+	/* usage count */
+	atomic_t count;
+};
+
+struct btrfs_leaf_ref_tree {
+	struct rb_root root;
+	struct list_head list;
+	spinlock_t lock;
+};
+
+struct btrfs_device;
+struct btrfs_fs_devices;
+struct btrfs_fs_info {
+	u8 fsid[BTRFS_FSID_SIZE];
+	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
+	struct btrfs_root *extent_root;
+	struct btrfs_root *tree_root;
+	struct btrfs_root *chunk_root;
+	struct btrfs_root *dev_root;
+	struct btrfs_root *fs_root;
+	struct btrfs_root *csum_root;
+
+	/* the log root tree is a directory of all the other log roots */
+	struct btrfs_root *log_root_tree;
+	struct radix_tree_root fs_roots_radix;
+
+	/* block group cache stuff */
+	spinlock_t block_group_cache_lock;
+	struct rb_root block_group_cache_tree;
+
+	struct extent_io_tree pinned_extents;
+	struct extent_io_tree pending_del;
+	struct extent_io_tree extent_ins;
+
+	/* logical->physical extent mapping */
+	struct btrfs_mapping_tree mapping_tree;
+
+	u64 generation;
+	u64 last_trans_committed;
+	u64 last_trans_new_blockgroup;
+	u64 open_ioctl_trans;
+	unsigned long mount_opt;
+	u64 max_extent;
+	u64 max_inline;
+	u64 alloc_start;
+	struct btrfs_transaction *running_transaction;
+	wait_queue_head_t transaction_throttle;
+	wait_queue_head_t transaction_wait;
+
+	wait_queue_head_t async_submit_wait;
+	wait_queue_head_t tree_log_wait;
+
+	struct btrfs_super_block super_copy;
+	struct btrfs_super_block super_for_commit;
+	struct block_device *__bdev;
+	struct super_block *sb;
+	struct inode *btree_inode;
+	struct backing_dev_info bdi;
+	spinlock_t hash_lock;
+	struct mutex trans_mutex;
+	struct mutex tree_log_mutex;
+	struct mutex transaction_kthread_mutex;
+	struct mutex cleaner_mutex;
+	struct mutex extent_ins_mutex;
+	struct mutex pinned_mutex;
+	struct mutex chunk_mutex;
+	struct mutex drop_mutex;
+	struct mutex volume_mutex;
+	struct mutex tree_reloc_mutex;
+	struct list_head trans_list;
+	struct list_head hashers;
+	struct list_head dead_roots;
+
+	atomic_t nr_async_submits;
+	atomic_t async_submit_draining;
+	atomic_t nr_async_bios;
+	atomic_t async_delalloc_pages;
+	atomic_t tree_log_writers;
+	atomic_t tree_log_commit;
+	unsigned long tree_log_batch;
+	u64 tree_log_transid;
+
+	/*
+	 * this is used by the balancing code to wait for all the pending
+	 * ordered extents
+	 */
+	spinlock_t ordered_extent_lock;
+	struct list_head ordered_extents;
+	struct list_head delalloc_inodes;
+
+	/*
+	 * there is a pool of worker threads for checksumming during writes
+	 * and a pool for checksumming after reads.  This is because readers
+	 * can run with FS locks held, and the writers may be waiting for
+	 * those locks.  We don't want ordering in the pending list to cause
+	 * deadlocks, and so the two are serviced separately.
+	 *
+	 * A third pool does submit_bio to avoid deadlocking with the other
+	 * two
+	 */
+	struct btrfs_workers workers;
+	struct btrfs_workers delalloc_workers;
+	struct btrfs_workers endio_workers;
+	struct btrfs_workers endio_meta_workers;
+	struct btrfs_workers endio_meta_write_workers;
+	struct btrfs_workers endio_write_workers;
+	struct btrfs_workers submit_workers;
+	/*
+	 * fixup workers take dirty pages that didn't properly go through
+	 * the cow mechanism and make them safe to write.  It happens
+	 * for the sys_munmap function call path
+	 */
+	struct btrfs_workers fixup_workers;
+	struct task_struct *transaction_kthread;
+	struct task_struct *cleaner_kthread;
+	int thread_pool_size;
+
+	/* tree relocation relocated fields */
+	struct list_head dead_reloc_roots;
+	struct btrfs_leaf_ref_tree reloc_ref_tree;
+	struct btrfs_leaf_ref_tree shared_ref_tree;
+
+	struct kobject super_kobj;
+	struct completion kobj_unregister;
+	int do_barriers;
+	int closing;
+	int log_root_recovering;
+	atomic_t throttles;
+	atomic_t throttle_gen;
+
+	u64 total_pinned;
+	struct list_head dirty_cowonly_roots;
+
+	struct btrfs_fs_devices *fs_devices;
+	struct list_head space_info;
+	spinlock_t delalloc_lock;
+	spinlock_t new_trans_lock;
+	u64 delalloc_bytes;
+	u64 last_alloc;
+	u64 last_data_alloc;
+
+	spinlock_t ref_cache_lock;
+	u64 total_ref_cache_size;
+
+	u64 avail_data_alloc_bits;
+	u64 avail_metadata_alloc_bits;
+	u64 avail_system_alloc_bits;
+	u64 data_alloc_profile;
+	u64 metadata_alloc_profile;
+	u64 system_alloc_profile;
+
+	void *bdev_holder;
+};
+
+/*
+ * in ram representation of the tree.  extent_root is used for all allocations
+ * and for the extent tree extent_root root.
+ */
+struct btrfs_dirty_root;
+struct btrfs_root {
+	struct extent_buffer *node;
+
+	/* the node lock is held while changing the node pointer */
+	spinlock_t node_lock;
+
+	struct extent_buffer *commit_root;
+	struct btrfs_leaf_ref_tree *ref_tree;
+	struct btrfs_leaf_ref_tree ref_tree_struct;
+	struct btrfs_dirty_root *dirty_root;
+	struct btrfs_root *log_root;
+	struct btrfs_root *reloc_root;
+
+	struct btrfs_root_item root_item;
+	struct btrfs_key root_key;
+	struct btrfs_fs_info *fs_info;
+	struct extent_io_tree dirty_log_pages;
+
+	struct kobject root_kobj;
+	struct completion kobj_unregister;
+	struct mutex objectid_mutex;
+	struct mutex log_mutex;
+
+	u64 objectid;
+	u64 last_trans;
+
+	/* data allocations are done in sectorsize units */
+	u32 sectorsize;
+
+	/* node allocations are done in nodesize units */
+	u32 nodesize;
+
+	/* leaf allocations are done in leafsize units */
+	u32 leafsize;
+
+	u32 stripesize;
+
+	u32 type;
+	u64 highest_inode;
+	u64 last_inode_alloc;
+	int ref_cows;
+	int track_dirty;
+	u64 defrag_trans_start;
+	struct btrfs_key defrag_progress;
+	struct btrfs_key defrag_max;
+	int defrag_running;
+	int defrag_level;
+	char *name;
+	int in_sysfs;
+
+	/* the dirty list is only used by non-reference counted roots */
+	struct list_head dirty_list;
+
+	spinlock_t list_lock;
+	struct list_head dead_list;
+	struct list_head orphan_list;
+
+	/*
+	 * right now this just gets used so that a root has its own devid
+	 * for stat.  It may be used for more later
+	 */
+	struct super_block anon_super;
+};
+
+/*
+
+ * inode items have the data typically returned from stat and store other
+ * info about object characteristics.  There is one for every file and dir in
+ * the FS
+ */
+#define BTRFS_INODE_ITEM_KEY		1
+#define BTRFS_INODE_REF_KEY		12
+#define BTRFS_XATTR_ITEM_KEY		24
+#define BTRFS_ORPHAN_ITEM_KEY		48
+/* reserve 2-15 close to the inode for later flexibility */
+
+/*
+ * dir items are the name -> inode pointers in a directory.  There is one
+ * for every name in a directory.
+ */
+#define BTRFS_DIR_LOG_ITEM_KEY  60
+#define BTRFS_DIR_LOG_INDEX_KEY 72
+#define BTRFS_DIR_ITEM_KEY	84
+#define BTRFS_DIR_INDEX_KEY	96
+/*
+ * extent data is for file data
+ */
+#define BTRFS_EXTENT_DATA_KEY	108
+
+/*
+ * extent csums are stored in a separate tree and hold csums for
+ * an entire extent on disk.
+ */
+#define BTRFS_EXTENT_CSUM_KEY	128
+
+/*
+ * root items point to tree roots.  There are typically in the root
+ * tree used by the super block to find all the other trees
+ */
+#define BTRFS_ROOT_ITEM_KEY	132
+
+/*
+ * root backrefs tie subvols and snapshots to the directory entries that
+ * reference them
+ */
+#define BTRFS_ROOT_BACKREF_KEY	144
+
+/*
+ * root refs make a fast index for listing all of the snapshots and
+ * subvolumes referenced by a given root.  They point directly to the
+ * directory item in the root that references the subvol
+ */
+#define BTRFS_ROOT_REF_KEY	156
+
+/*
+ * extent items are in the extent map tree.  These record which blocks
+ * are used, and how many references there are to each block
+ */
+#define BTRFS_EXTENT_ITEM_KEY	168
+#define BTRFS_EXTENT_REF_KEY	180
+
+/*
+ * block groups give us hints into the extent allocation trees.  Which
+ * blocks are free etc etc
+ */
+#define BTRFS_BLOCK_GROUP_ITEM_KEY 192
+
+#define BTRFS_DEV_EXTENT_KEY	204
+#define BTRFS_DEV_ITEM_KEY	216
+#define BTRFS_CHUNK_ITEM_KEY	228
+
+/*
+ * string items are for debugging.  They just store a short string of
+ * data in the FS
+ */
+#define BTRFS_STRING_ITEM_KEY	253
+
+#define BTRFS_MOUNT_NODATASUM		(1 << 0)
+#define BTRFS_MOUNT_NODATACOW		(1 << 1)
+#define BTRFS_MOUNT_NOBARRIER		(1 << 2)
+#define BTRFS_MOUNT_SSD			(1 << 3)
+#define BTRFS_MOUNT_DEGRADED		(1 << 4)
+#define BTRFS_MOUNT_COMPRESS		(1 << 5)
+
+#define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
+#define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
+#define btrfs_test_opt(root, opt)	((root)->fs_info->mount_opt & \
+					 BTRFS_MOUNT_##opt)
+/*
+ * Inode flags
+ */
+#define BTRFS_INODE_NODATASUM		(1 << 0)
+#define BTRFS_INODE_NODATACOW		(1 << 1)
+#define BTRFS_INODE_READONLY		(1 << 2)
+#define BTRFS_INODE_NOCOMPRESS		(1 << 3)
+#define BTRFS_INODE_PREALLOC		(1 << 4)
+#define btrfs_clear_flag(inode, flag)	(BTRFS_I(inode)->flags &= \
+					 ~BTRFS_INODE_##flag)
+#define btrfs_set_flag(inode, flag)	(BTRFS_I(inode)->flags |= \
+					 BTRFS_INODE_##flag)
+#define btrfs_test_flag(inode, flag)	(BTRFS_I(inode)->flags & \
+					 BTRFS_INODE_##flag)
+/* some macros to generate set/get funcs for the struct fields.  This
+ * assumes there is a lefoo_to_cpu for every type, so lets make a simple
+ * one for u8:
+ */
+#define le8_to_cpu(v) (v)
+#define cpu_to_le8(v) (v)
+#define __le8 u8
+
+#define read_eb_member(eb, ptr, type, member, result) (			\
+	read_extent_buffer(eb, (char *)(result),			\
+			   ((unsigned long)(ptr)) +			\
+			    offsetof(type, member),			\
+			   sizeof(((type *)0)->member)))
+
+#define write_eb_member(eb, ptr, type, member, result) (		\
+	write_extent_buffer(eb, (char *)(result),			\
+			   ((unsigned long)(ptr)) +			\
+			    offsetof(type, member),			\
+			   sizeof(((type *)0)->member)))
+
+#ifndef BTRFS_SETGET_FUNCS
+#define BTRFS_SETGET_FUNCS(name, type, member, bits)			\
+u##bits btrfs_##name(struct extent_buffer *eb, type *s);		\
+void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val);
+#endif
+
+#define BTRFS_SETGET_HEADER_FUNCS(name, type, member, bits)		\
+static inline u##bits btrfs_##name(struct extent_buffer *eb)		\
+{									\
+	type *p = kmap_atomic(eb->first_page, KM_USER0);		\
+	u##bits res = le##bits##_to_cpu(p->member);			\
+	kunmap_atomic(p, KM_USER0);					\
+	return res;							\
+}									\
+static inline void btrfs_set_##name(struct extent_buffer *eb,		\
+				    u##bits val)			\
+{									\
+	type *p = kmap_atomic(eb->first_page, KM_USER0);		\
+	p->member = cpu_to_le##bits(val);				\
+	kunmap_atomic(p, KM_USER0);					\
+}
+
+#define BTRFS_SETGET_STACK_FUNCS(name, type, member, bits)		\
+static inline u##bits btrfs_##name(type *s)				\
+{									\
+	return le##bits##_to_cpu(s->member);				\
+}									\
+static inline void btrfs_set_##name(type *s, u##bits val)		\
+{									\
+	s->member = cpu_to_le##bits(val);				\
+}
+
+BTRFS_SETGET_FUNCS(device_type, struct btrfs_dev_item, type, 64);
+BTRFS_SETGET_FUNCS(device_total_bytes, struct btrfs_dev_item, total_bytes, 64);
+BTRFS_SETGET_FUNCS(device_bytes_used, struct btrfs_dev_item, bytes_used, 64);
+BTRFS_SETGET_FUNCS(device_io_align, struct btrfs_dev_item, io_align, 32);
+BTRFS_SETGET_FUNCS(device_io_width, struct btrfs_dev_item, io_width, 32);
+BTRFS_SETGET_FUNCS(device_start_offset, struct btrfs_dev_item,
+		   start_offset, 64);
+BTRFS_SETGET_FUNCS(device_sector_size, struct btrfs_dev_item, sector_size, 32);
+BTRFS_SETGET_FUNCS(device_id, struct btrfs_dev_item, devid, 64);
+BTRFS_SETGET_FUNCS(device_group, struct btrfs_dev_item, dev_group, 32);
+BTRFS_SETGET_FUNCS(device_seek_speed, struct btrfs_dev_item, seek_speed, 8);
+BTRFS_SETGET_FUNCS(device_bandwidth, struct btrfs_dev_item, bandwidth, 8);
+BTRFS_SETGET_FUNCS(device_generation, struct btrfs_dev_item, generation, 64);
+
+BTRFS_SETGET_STACK_FUNCS(stack_device_type, struct btrfs_dev_item, type, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_device_total_bytes, struct btrfs_dev_item,
+			 total_bytes, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_device_bytes_used, struct btrfs_dev_item,
+			 bytes_used, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_device_io_align, struct btrfs_dev_item,
+			 io_align, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_device_io_width, struct btrfs_dev_item,
+			 io_width, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_device_sector_size, struct btrfs_dev_item,
+			 sector_size, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_device_id, struct btrfs_dev_item, devid, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_device_group, struct btrfs_dev_item,
+			 dev_group, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_device_seek_speed, struct btrfs_dev_item,
+			 seek_speed, 8);
+BTRFS_SETGET_STACK_FUNCS(stack_device_bandwidth, struct btrfs_dev_item,
+			 bandwidth, 8);
+BTRFS_SETGET_STACK_FUNCS(stack_device_generation, struct btrfs_dev_item,
+			 generation, 64);
+
+static inline char *btrfs_device_uuid(struct btrfs_dev_item *d)
+{
+	return (char *)d + offsetof(struct btrfs_dev_item, uuid);
+}
+
+static inline char *btrfs_device_fsid(struct btrfs_dev_item *d)
+{
+	return (char *)d + offsetof(struct btrfs_dev_item, fsid);
+}
+
+BTRFS_SETGET_FUNCS(chunk_length, struct btrfs_chunk, length, 64);
+BTRFS_SETGET_FUNCS(chunk_owner, struct btrfs_chunk, owner, 64);
+BTRFS_SETGET_FUNCS(chunk_stripe_len, struct btrfs_chunk, stripe_len, 64);
+BTRFS_SETGET_FUNCS(chunk_io_align, struct btrfs_chunk, io_align, 32);
+BTRFS_SETGET_FUNCS(chunk_io_width, struct btrfs_chunk, io_width, 32);
+BTRFS_SETGET_FUNCS(chunk_sector_size, struct btrfs_chunk, sector_size, 32);
+BTRFS_SETGET_FUNCS(chunk_type, struct btrfs_chunk, type, 64);
+BTRFS_SETGET_FUNCS(chunk_num_stripes, struct btrfs_chunk, num_stripes, 16);
+BTRFS_SETGET_FUNCS(chunk_sub_stripes, struct btrfs_chunk, sub_stripes, 16);
+BTRFS_SETGET_FUNCS(stripe_devid, struct btrfs_stripe, devid, 64);
+BTRFS_SETGET_FUNCS(stripe_offset, struct btrfs_stripe, offset, 64);
+
+static inline char *btrfs_stripe_dev_uuid(struct btrfs_stripe *s)
+{
+	return (char *)s + offsetof(struct btrfs_stripe, dev_uuid);
+}
+
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_length, struct btrfs_chunk, length, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_owner, struct btrfs_chunk, owner, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_stripe_len, struct btrfs_chunk,
+			 stripe_len, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_io_align, struct btrfs_chunk,
+			 io_align, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_io_width, struct btrfs_chunk,
+			 io_width, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_sector_size, struct btrfs_chunk,
+			 sector_size, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_type, struct btrfs_chunk, type, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_num_stripes, struct btrfs_chunk,
+			 num_stripes, 16);
+BTRFS_SETGET_STACK_FUNCS(stack_chunk_sub_stripes, struct btrfs_chunk,
+			 sub_stripes, 16);
+BTRFS_SETGET_STACK_FUNCS(stack_stripe_devid, struct btrfs_stripe, devid, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_stripe_offset, struct btrfs_stripe, offset, 64);
+
+static inline struct btrfs_stripe *btrfs_stripe_nr(struct btrfs_chunk *c,
+						   int nr)
+{
+	unsigned long offset = (unsigned long)c;
+	offset += offsetof(struct btrfs_chunk, stripe);
+	offset += nr * sizeof(struct btrfs_stripe);
+	return (struct btrfs_stripe *)offset;
+}
+
+static inline char *btrfs_stripe_dev_uuid_nr(struct btrfs_chunk *c, int nr)
+{
+	return btrfs_stripe_dev_uuid(btrfs_stripe_nr(c, nr));
+}
+
+static inline u64 btrfs_stripe_offset_nr(struct extent_buffer *eb,
+					 struct btrfs_chunk *c, int nr)
+{
+	return btrfs_stripe_offset(eb, btrfs_stripe_nr(c, nr));
+}
+
+static inline void btrfs_set_stripe_offset_nr(struct extent_buffer *eb,
+					     struct btrfs_chunk *c, int nr,
+					     u64 val)
+{
+	btrfs_set_stripe_offset(eb, btrfs_stripe_nr(c, nr), val);
+}
+
+static inline u64 btrfs_stripe_devid_nr(struct extent_buffer *eb,
+					 struct btrfs_chunk *c, int nr)
+{
+	return btrfs_stripe_devid(eb, btrfs_stripe_nr(c, nr));
+}
+
+static inline void btrfs_set_stripe_devid_nr(struct extent_buffer *eb,
+					     struct btrfs_chunk *c, int nr,
+					     u64 val)
+{
+	btrfs_set_stripe_devid(eb, btrfs_stripe_nr(c, nr), val);
+}
+
+/* struct btrfs_block_group_item */
+BTRFS_SETGET_STACK_FUNCS(block_group_used, struct btrfs_block_group_item,
+			 used, 64);
+BTRFS_SETGET_FUNCS(disk_block_group_used, struct btrfs_block_group_item,
+			 used, 64);
+BTRFS_SETGET_STACK_FUNCS(block_group_chunk_objectid,
+			struct btrfs_block_group_item, chunk_objectid, 64);
+
+BTRFS_SETGET_FUNCS(disk_block_group_chunk_objectid,
+		   struct btrfs_block_group_item, chunk_objectid, 64);
+BTRFS_SETGET_FUNCS(disk_block_group_flags,
+		   struct btrfs_block_group_item, flags, 64);
+BTRFS_SETGET_STACK_FUNCS(block_group_flags,
+			struct btrfs_block_group_item, flags, 64);
+
+/* struct btrfs_inode_ref */
+BTRFS_SETGET_FUNCS(inode_ref_name_len, struct btrfs_inode_ref, name_len, 16);
+BTRFS_SETGET_FUNCS(inode_ref_index, struct btrfs_inode_ref, index, 64);
+
+/* struct btrfs_inode_item */
+BTRFS_SETGET_FUNCS(inode_generation, struct btrfs_inode_item, generation, 64);
+BTRFS_SETGET_FUNCS(inode_sequence, struct btrfs_inode_item, sequence, 64);
+BTRFS_SETGET_FUNCS(inode_transid, struct btrfs_inode_item, transid, 64);
+BTRFS_SETGET_FUNCS(inode_size, struct btrfs_inode_item, size, 64);
+BTRFS_SETGET_FUNCS(inode_nbytes, struct btrfs_inode_item, nbytes, 64);
+BTRFS_SETGET_FUNCS(inode_block_group, struct btrfs_inode_item, block_group, 64);
+BTRFS_SETGET_FUNCS(inode_nlink, struct btrfs_inode_item, nlink, 32);
+BTRFS_SETGET_FUNCS(inode_uid, struct btrfs_inode_item, uid, 32);
+BTRFS_SETGET_FUNCS(inode_gid, struct btrfs_inode_item, gid, 32);
+BTRFS_SETGET_FUNCS(inode_mode, struct btrfs_inode_item, mode, 32);
+BTRFS_SETGET_FUNCS(inode_rdev, struct btrfs_inode_item, rdev, 64);
+BTRFS_SETGET_FUNCS(inode_flags, struct btrfs_inode_item, flags, 64);
+
+static inline struct btrfs_timespec *
+btrfs_inode_atime(struct btrfs_inode_item *inode_item)
+{
+	unsigned long ptr = (unsigned long)inode_item;
+	ptr += offsetof(struct btrfs_inode_item, atime);
+	return (struct btrfs_timespec *)ptr;
+}
+
+static inline struct btrfs_timespec *
+btrfs_inode_mtime(struct btrfs_inode_item *inode_item)
+{
+	unsigned long ptr = (unsigned long)inode_item;
+	ptr += offsetof(struct btrfs_inode_item, mtime);
+	return (struct btrfs_timespec *)ptr;
+}
+
+static inline struct btrfs_timespec *
+btrfs_inode_ctime(struct btrfs_inode_item *inode_item)
+{
+	unsigned long ptr = (unsigned long)inode_item;
+	ptr += offsetof(struct btrfs_inode_item, ctime);
+	return (struct btrfs_timespec *)ptr;
+}
+
+static inline struct btrfs_timespec *
+btrfs_inode_otime(struct btrfs_inode_item *inode_item)
+{
+	unsigned long ptr = (unsigned long)inode_item;
+	ptr += offsetof(struct btrfs_inode_item, otime);
+	return (struct btrfs_timespec *)ptr;
+}
+
+BTRFS_SETGET_FUNCS(timespec_sec, struct btrfs_timespec, sec, 64);
+BTRFS_SETGET_FUNCS(timespec_nsec, struct btrfs_timespec, nsec, 32);
+
+/* struct btrfs_dev_extent */
+BTRFS_SETGET_FUNCS(dev_extent_chunk_tree, struct btrfs_dev_extent,
+		   chunk_tree, 64);
+BTRFS_SETGET_FUNCS(dev_extent_chunk_objectid, struct btrfs_dev_extent,
+		   chunk_objectid, 64);
+BTRFS_SETGET_FUNCS(dev_extent_chunk_offset, struct btrfs_dev_extent,
+		   chunk_offset, 64);
+BTRFS_SETGET_FUNCS(dev_extent_length, struct btrfs_dev_extent, length, 64);
+
+static inline u8 *btrfs_dev_extent_chunk_tree_uuid(struct btrfs_dev_extent *dev)
+{
+	unsigned long ptr = offsetof(struct btrfs_dev_extent, chunk_tree_uuid);
+	return (u8 *)((unsigned long)dev + ptr);
+}
+
+/* struct btrfs_extent_ref */
+BTRFS_SETGET_FUNCS(ref_root, struct btrfs_extent_ref, root, 64);
+BTRFS_SETGET_FUNCS(ref_generation, struct btrfs_extent_ref, generation, 64);
+BTRFS_SETGET_FUNCS(ref_objectid, struct btrfs_extent_ref, objectid, 64);
+BTRFS_SETGET_FUNCS(ref_num_refs, struct btrfs_extent_ref, num_refs, 32);
+
+BTRFS_SETGET_STACK_FUNCS(stack_ref_root, struct btrfs_extent_ref, root, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_ref_generation, struct btrfs_extent_ref,
+			 generation, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_ref_objectid, struct btrfs_extent_ref,
+			 objectid, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_ref_num_refs, struct btrfs_extent_ref,
+			 num_refs, 32);
+
+/* struct btrfs_extent_item */
+BTRFS_SETGET_FUNCS(extent_refs, struct btrfs_extent_item, refs, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_extent_refs, struct btrfs_extent_item,
+			 refs, 32);
+
+/* struct btrfs_node */
+BTRFS_SETGET_FUNCS(key_blockptr, struct btrfs_key_ptr, blockptr, 64);
+BTRFS_SETGET_FUNCS(key_generation, struct btrfs_key_ptr, generation, 64);
+
+static inline u64 btrfs_node_blockptr(struct extent_buffer *eb, int nr)
+{
+	unsigned long ptr;
+	ptr = offsetof(struct btrfs_node, ptrs) +
+		sizeof(struct btrfs_key_ptr) * nr;
+	return btrfs_key_blockptr(eb, (struct btrfs_key_ptr *)ptr);
+}
+
+static inline void btrfs_set_node_blockptr(struct extent_buffer *eb,
+					   int nr, u64 val)
+{
+	unsigned long ptr;
+	ptr = offsetof(struct btrfs_node, ptrs) +
+		sizeof(struct btrfs_key_ptr) * nr;
+	btrfs_set_key_blockptr(eb, (struct btrfs_key_ptr *)ptr, val);
+}
+
+static inline u64 btrfs_node_ptr_generation(struct extent_buffer *eb, int nr)
+{
+	unsigned long ptr;
+	ptr = offsetof(struct btrfs_node, ptrs) +
+		sizeof(struct btrfs_key_ptr) * nr;
+	return btrfs_key_generation(eb, (struct btrfs_key_ptr *)ptr);
+}
+
+static inline void btrfs_set_node_ptr_generation(struct extent_buffer *eb,
+						 int nr, u64 val)
+{
+	unsigned long ptr;
+	ptr = offsetof(struct btrfs_node, ptrs) +
+		sizeof(struct btrfs_key_ptr) * nr;
+	btrfs_set_key_generation(eb, (struct btrfs_key_ptr *)ptr, val);
+}
+
+static inline unsigned long btrfs_node_key_ptr_offset(int nr)
+{
+	return offsetof(struct btrfs_node, ptrs) +
+		sizeof(struct btrfs_key_ptr) * nr;
+}
+
+void btrfs_node_key(struct extent_buffer *eb,
+		    struct btrfs_disk_key *disk_key, int nr);
+
+static inline void btrfs_set_node_key(struct extent_buffer *eb,
+				      struct btrfs_disk_key *disk_key, int nr)
+{
+	unsigned long ptr;
+	ptr = btrfs_node_key_ptr_offset(nr);
+	write_eb_member(eb, (struct btrfs_key_ptr *)ptr,
+		       struct btrfs_key_ptr, key, disk_key);
+}
+
+/* struct btrfs_item */
+BTRFS_SETGET_FUNCS(item_offset, struct btrfs_item, offset, 32);
+BTRFS_SETGET_FUNCS(item_size, struct btrfs_item, size, 32);
+
+static inline unsigned long btrfs_item_nr_offset(int nr)
+{
+	return offsetof(struct btrfs_leaf, items) +
+		sizeof(struct btrfs_item) * nr;
+}
+
+static inline struct btrfs_item *btrfs_item_nr(struct extent_buffer *eb,
+					       int nr)
+{
+	return (struct btrfs_item *)btrfs_item_nr_offset(nr);
+}
+
+static inline u32 btrfs_item_end(struct extent_buffer *eb,
+				 struct btrfs_item *item)
+{
+	return btrfs_item_offset(eb, item) + btrfs_item_size(eb, item);
+}
+
+static inline u32 btrfs_item_end_nr(struct extent_buffer *eb, int nr)
+{
+	return btrfs_item_end(eb, btrfs_item_nr(eb, nr));
+}
+
+static inline u32 btrfs_item_offset_nr(struct extent_buffer *eb, int nr)
+{
+	return btrfs_item_offset(eb, btrfs_item_nr(eb, nr));
+}
+
+static inline u32 btrfs_item_size_nr(struct extent_buffer *eb, int nr)
+{
+	return btrfs_item_size(eb, btrfs_item_nr(eb, nr));
+}
+
+static inline void btrfs_item_key(struct extent_buffer *eb,
+			   struct btrfs_disk_key *disk_key, int nr)
+{
+	struct btrfs_item *item = btrfs_item_nr(eb, nr);
+	read_eb_member(eb, item, struct btrfs_item, key, disk_key);
+}
+
+static inline void btrfs_set_item_key(struct extent_buffer *eb,
+			       struct btrfs_disk_key *disk_key, int nr)
+{
+	struct btrfs_item *item = btrfs_item_nr(eb, nr);
+	write_eb_member(eb, item, struct btrfs_item, key, disk_key);
+}
+
+BTRFS_SETGET_FUNCS(dir_log_end, struct btrfs_dir_log_item, end, 64);
+
+/*
+ * struct btrfs_root_ref
+ */
+BTRFS_SETGET_FUNCS(root_ref_dirid, struct btrfs_root_ref, dirid, 64);
+BTRFS_SETGET_FUNCS(root_ref_sequence, struct btrfs_root_ref, sequence, 64);
+BTRFS_SETGET_FUNCS(root_ref_name_len, struct btrfs_root_ref, name_len, 16);
+
+/* struct btrfs_dir_item */
+BTRFS_SETGET_FUNCS(dir_data_len, struct btrfs_dir_item, data_len, 16);
+BTRFS_SETGET_FUNCS(dir_type, struct btrfs_dir_item, type, 8);
+BTRFS_SETGET_FUNCS(dir_name_len, struct btrfs_dir_item, name_len, 16);
+BTRFS_SETGET_FUNCS(dir_transid, struct btrfs_dir_item, transid, 64);
+
+static inline void btrfs_dir_item_key(struct extent_buffer *eb,
+				      struct btrfs_dir_item *item,
+				      struct btrfs_disk_key *key)
+{
+	read_eb_member(eb, item, struct btrfs_dir_item, location, key);
+}
+
+static inline void btrfs_set_dir_item_key(struct extent_buffer *eb,
+					  struct btrfs_dir_item *item,
+					  struct btrfs_disk_key *key)
+{
+	write_eb_member(eb, item, struct btrfs_dir_item, location, key);
+}
+
+/* struct btrfs_disk_key */
+BTRFS_SETGET_STACK_FUNCS(disk_key_objectid, struct btrfs_disk_key,
+			 objectid, 64);
+BTRFS_SETGET_STACK_FUNCS(disk_key_offset, struct btrfs_disk_key, offset, 64);
+BTRFS_SETGET_STACK_FUNCS(disk_key_type, struct btrfs_disk_key, type, 8);
+
+static inline void btrfs_disk_key_to_cpu(struct btrfs_key *cpu,
+					 struct btrfs_disk_key *disk)
+{
+	cpu->offset = le64_to_cpu(disk->offset);
+	cpu->type = disk->type;
+	cpu->objectid = le64_to_cpu(disk->objectid);
+}
+
+static inline void btrfs_cpu_key_to_disk(struct btrfs_disk_key *disk,
+					 struct btrfs_key *cpu)
+{
+	disk->offset = cpu_to_le64(cpu->offset);
+	disk->type = cpu->type;
+	disk->objectid = cpu_to_le64(cpu->objectid);
+}
+
+static inline void btrfs_node_key_to_cpu(struct extent_buffer *eb,
+				  struct btrfs_key *key, int nr)
+{
+	struct btrfs_disk_key disk_key;
+	btrfs_node_key(eb, &disk_key, nr);
+	btrfs_disk_key_to_cpu(key, &disk_key);
+}
+
+static inline void btrfs_item_key_to_cpu(struct extent_buffer *eb,
+				  struct btrfs_key *key, int nr)
+{
+	struct btrfs_disk_key disk_key;
+	btrfs_item_key(eb, &disk_key, nr);
+	btrfs_disk_key_to_cpu(key, &disk_key);
+}
+
+static inline void btrfs_dir_item_key_to_cpu(struct extent_buffer *eb,
+				      struct btrfs_dir_item *item,
+				      struct btrfs_key *key)
+{
+	struct btrfs_disk_key disk_key;
+	btrfs_dir_item_key(eb, item, &disk_key);
+	btrfs_disk_key_to_cpu(key, &disk_key);
+}
+
+
+static inline u8 btrfs_key_type(struct btrfs_key *key)
+{
+	return key->type;
+}
+
+static inline void btrfs_set_key_type(struct btrfs_key *key, u8 val)
+{
+	key->type = val;
+}
+
+/* struct btrfs_header */
+BTRFS_SETGET_HEADER_FUNCS(header_bytenr, struct btrfs_header, bytenr, 64);
+BTRFS_SETGET_HEADER_FUNCS(header_generation, struct btrfs_header,
+			  generation, 64);
+BTRFS_SETGET_HEADER_FUNCS(header_owner, struct btrfs_header, owner, 64);
+BTRFS_SETGET_HEADER_FUNCS(header_nritems, struct btrfs_header, nritems, 32);
+BTRFS_SETGET_HEADER_FUNCS(header_flags, struct btrfs_header, flags, 64);
+BTRFS_SETGET_HEADER_FUNCS(header_level, struct btrfs_header, level, 8);
+
+static inline int btrfs_header_flag(struct extent_buffer *eb, u64 flag)
+{
+	return (btrfs_header_flags(eb) & flag) == flag;
+}
+
+static inline int btrfs_set_header_flag(struct extent_buffer *eb, u64 flag)
+{
+	u64 flags = btrfs_header_flags(eb);
+	btrfs_set_header_flags(eb, flags | flag);
+	return (flags & flag) == flag;
+}
+
+static inline int btrfs_clear_header_flag(struct extent_buffer *eb, u64 flag)
+{
+	u64 flags = btrfs_header_flags(eb);
+	btrfs_set_header_flags(eb, flags & ~flag);
+	return (flags & flag) == flag;
+}
+
+static inline u8 *btrfs_header_fsid(struct extent_buffer *eb)
+{
+	unsigned long ptr = offsetof(struct btrfs_header, fsid);
+	return (u8 *)ptr;
+}
+
+static inline u8 *btrfs_header_chunk_tree_uuid(struct extent_buffer *eb)
+{
+	unsigned long ptr = offsetof(struct btrfs_header, chunk_tree_uuid);
+	return (u8 *)ptr;
+}
+
+static inline u8 *btrfs_super_fsid(struct extent_buffer *eb)
+{
+	unsigned long ptr = offsetof(struct btrfs_super_block, fsid);
+	return (u8 *)ptr;
+}
+
+static inline u8 *btrfs_header_csum(struct extent_buffer *eb)
+{
+	unsigned long ptr = offsetof(struct btrfs_header, csum);
+	return (u8 *)ptr;
+}
+
+static inline struct btrfs_node *btrfs_buffer_node(struct extent_buffer *eb)
+{
+	return NULL;
+}
+
+static inline struct btrfs_leaf *btrfs_buffer_leaf(struct extent_buffer *eb)
+{
+	return NULL;
+}
+
+static inline struct btrfs_header *btrfs_buffer_header(struct extent_buffer *eb)
+{
+	return NULL;
+}
+
+static inline int btrfs_is_leaf(struct extent_buffer *eb)
+{
+	return btrfs_header_level(eb) == 0;
+}
+
+/* struct btrfs_root_item */
+BTRFS_SETGET_FUNCS(disk_root_generation, struct btrfs_root_item,
+		   generation, 64);
+BTRFS_SETGET_FUNCS(disk_root_refs, struct btrfs_root_item, refs, 32);
+BTRFS_SETGET_FUNCS(disk_root_bytenr, struct btrfs_root_item, bytenr, 64);
+BTRFS_SETGET_FUNCS(disk_root_level, struct btrfs_root_item, level, 8);
+
+BTRFS_SETGET_STACK_FUNCS(root_generation, struct btrfs_root_item,
+			 generation, 64);
+BTRFS_SETGET_STACK_FUNCS(root_bytenr, struct btrfs_root_item, bytenr, 64);
+BTRFS_SETGET_STACK_FUNCS(root_level, struct btrfs_root_item, level, 8);
+BTRFS_SETGET_STACK_FUNCS(root_dirid, struct btrfs_root_item, root_dirid, 64);
+BTRFS_SETGET_STACK_FUNCS(root_refs, struct btrfs_root_item, refs, 32);
+BTRFS_SETGET_STACK_FUNCS(root_flags, struct btrfs_root_item, flags, 64);
+BTRFS_SETGET_STACK_FUNCS(root_used, struct btrfs_root_item, bytes_used, 64);
+BTRFS_SETGET_STACK_FUNCS(root_limit, struct btrfs_root_item, byte_limit, 64);
+BTRFS_SETGET_STACK_FUNCS(root_last_snapshot, struct btrfs_root_item,
+			 last_snapshot, 64);
+
+/* struct btrfs_super_block */
+
+BTRFS_SETGET_STACK_FUNCS(super_bytenr, struct btrfs_super_block, bytenr, 64);
+BTRFS_SETGET_STACK_FUNCS(super_flags, struct btrfs_super_block, flags, 64);
+BTRFS_SETGET_STACK_FUNCS(super_generation, struct btrfs_super_block,
+			 generation, 64);
+BTRFS_SETGET_STACK_FUNCS(super_root, struct btrfs_super_block, root, 64);
+BTRFS_SETGET_STACK_FUNCS(super_sys_array_size,
+			 struct btrfs_super_block, sys_chunk_array_size, 32);
+BTRFS_SETGET_STACK_FUNCS(super_chunk_root_generation,
+			 struct btrfs_super_block, chunk_root_generation, 64);
+BTRFS_SETGET_STACK_FUNCS(super_root_level, struct btrfs_super_block,
+			 root_level, 8);
+BTRFS_SETGET_STACK_FUNCS(super_chunk_root, struct btrfs_super_block,
+			 chunk_root, 64);
+BTRFS_SETGET_STACK_FUNCS(super_chunk_root_level, struct btrfs_super_block,
+			 chunk_root_level, 8);
+BTRFS_SETGET_STACK_FUNCS(super_log_root, struct btrfs_super_block,
+			 log_root, 64);
+BTRFS_SETGET_STACK_FUNCS(super_log_root_transid, struct btrfs_super_block,
+			 log_root_transid, 64);
+BTRFS_SETGET_STACK_FUNCS(super_log_root_level, struct btrfs_super_block,
+			 log_root_level, 8);
+BTRFS_SETGET_STACK_FUNCS(super_total_bytes, struct btrfs_super_block,
+			 total_bytes, 64);
+BTRFS_SETGET_STACK_FUNCS(super_bytes_used, struct btrfs_super_block,
+			 bytes_used, 64);
+BTRFS_SETGET_STACK_FUNCS(super_sectorsize, struct btrfs_super_block,
+			 sectorsize, 32);
+BTRFS_SETGET_STACK_FUNCS(super_nodesize, struct btrfs_super_block,
+			 nodesize, 32);
+BTRFS_SETGET_STACK_FUNCS(super_leafsize, struct btrfs_super_block,
+			 leafsize, 32);
+BTRFS_SETGET_STACK_FUNCS(super_stripesize, struct btrfs_super_block,
+			 stripesize, 32);
+BTRFS_SETGET_STACK_FUNCS(super_root_dir, struct btrfs_super_block,
+			 root_dir_objectid, 64);
+BTRFS_SETGET_STACK_FUNCS(super_num_devices, struct btrfs_super_block,
+			 num_devices, 64);
+BTRFS_SETGET_STACK_FUNCS(super_compat_flags, struct btrfs_super_block,
+			 compat_flags, 64);
+BTRFS_SETGET_STACK_FUNCS(super_compat_ro_flags, struct btrfs_super_block,
+			 compat_flags, 64);
+BTRFS_SETGET_STACK_FUNCS(super_incompat_flags, struct btrfs_super_block,
+			 incompat_flags, 64);
+BTRFS_SETGET_STACK_FUNCS(super_csum_type, struct btrfs_super_block,
+			 csum_type, 16);
+
+static inline int btrfs_super_csum_size(struct btrfs_super_block *s)
+{
+	int t = btrfs_super_csum_type(s);
+	BUG_ON(t >= ARRAY_SIZE(btrfs_csum_sizes));
+	return btrfs_csum_sizes[t];
+}
+
+static inline unsigned long btrfs_leaf_data(struct extent_buffer *l)
+{
+	return offsetof(struct btrfs_leaf, items);
+}
+
+/* struct btrfs_file_extent_item */
+BTRFS_SETGET_FUNCS(file_extent_type, struct btrfs_file_extent_item, type, 8);
+
+static inline unsigned long
+btrfs_file_extent_inline_start(struct btrfs_file_extent_item *e)
+{
+	unsigned long offset = (unsigned long)e;
+	offset += offsetof(struct btrfs_file_extent_item, disk_bytenr);
+	return offset;
+}
+
+static inline u32 btrfs_file_extent_calc_inline_size(u32 datasize)
+{
+	return offsetof(struct btrfs_file_extent_item, disk_bytenr) + datasize;
+}
+
+BTRFS_SETGET_FUNCS(file_extent_disk_bytenr, struct btrfs_file_extent_item,
+		   disk_bytenr, 64);
+BTRFS_SETGET_FUNCS(file_extent_generation, struct btrfs_file_extent_item,
+		   generation, 64);
+BTRFS_SETGET_FUNCS(file_extent_disk_num_bytes, struct btrfs_file_extent_item,
+		   disk_num_bytes, 64);
+BTRFS_SETGET_FUNCS(file_extent_offset, struct btrfs_file_extent_item,
+		  offset, 64);
+BTRFS_SETGET_FUNCS(file_extent_num_bytes, struct btrfs_file_extent_item,
+		   num_bytes, 64);
+BTRFS_SETGET_FUNCS(file_extent_ram_bytes, struct btrfs_file_extent_item,
+		   ram_bytes, 64);
+BTRFS_SETGET_FUNCS(file_extent_compression, struct btrfs_file_extent_item,
+		   compression, 8);
+BTRFS_SETGET_FUNCS(file_extent_encryption, struct btrfs_file_extent_item,
+		   encryption, 8);
+BTRFS_SETGET_FUNCS(file_extent_other_encoding, struct btrfs_file_extent_item,
+		   other_encoding, 16);
+
+/* this returns the number of file bytes represented by the inline item.
+ * If an item is compressed, this is the uncompressed size
+ */
+static inline u32 btrfs_file_extent_inline_len(struct extent_buffer *eb,
+					       struct btrfs_file_extent_item *e)
+{
+	return btrfs_file_extent_ram_bytes(eb, e);
+}
+
+/*
+ * this returns the number of bytes used by the item on disk, minus the
+ * size of any extent headers.  If a file is compressed on disk, this is
+ * the compressed size
+ */
+static inline u32 btrfs_file_extent_inline_item_len(struct extent_buffer *eb,
+						    struct btrfs_item *e)
+{
+	unsigned long offset;
+	offset = offsetof(struct btrfs_file_extent_item, disk_bytenr);
+	return btrfs_item_size(eb, e) - offset;
+}
+
+static inline struct btrfs_root *btrfs_sb(struct super_block *sb)
+{
+	return sb->s_fs_info;
+}
+
+static inline int btrfs_set_root_name(struct btrfs_root *root,
+				      const char *name, int len)
+{
+	/* if we already have a name just free it */
+	kfree(root->name);
+
+	root->name = kmalloc(len+1, GFP_KERNEL);
+	if (!root->name)
+		return -ENOMEM;
+
+	memcpy(root->name, name, len);
+	root->name[len] = '\0';
+
+	return 0;
+}
+
+static inline u32 btrfs_level_size(struct btrfs_root *root, int level)
+{
+	if (level == 0)
+		return root->leafsize;
+	return root->nodesize;
+}
+
+/* helper function to cast into the data area of the leaf. */
+#define btrfs_item_ptr(leaf, slot, type) \
+	((type *)(btrfs_leaf_data(leaf) + \
+	btrfs_item_offset_nr(leaf, slot)))
+
+#define btrfs_item_ptr_offset(leaf, slot) \
+	((unsigned long)(btrfs_leaf_data(leaf) + \
+	btrfs_item_offset_nr(leaf, slot)))
+
+static inline struct dentry *fdentry(struct file *file)
+{
+	return file->f_path.dentry;
+}
+
+/* extent-tree.c */
+int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len);
+int btrfs_lookup_extent_ref(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, u64 bytenr,
+			    u64 num_bytes, u32 *refs);
+int btrfs_update_pinned_extents(struct btrfs_root *root,
+				u64 bytenr, u64 num, int pin);
+int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root, struct extent_buffer *leaf);
+int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, u64 objectid, u64 bytenr);
+int btrfs_extent_post_op(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root);
+int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy);
+struct btrfs_block_group_cache *btrfs_lookup_block_group(
+						 struct btrfs_fs_info *info,
+						 u64 bytenr);
+u64 btrfs_find_block_group(struct btrfs_root *root,
+			   u64 search_start, u64 search_hint, int owner);
+struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
+					     struct btrfs_root *root,
+					     u32 blocksize, u64 parent,
+					     u64 root_objectid,
+					     u64 ref_generation,
+					     int level,
+					     u64 hint,
+					     u64 empty_size);
+struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
+					    struct btrfs_root *root,
+					    u64 bytenr, u32 blocksize);
+int btrfs_alloc_extent(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root,
+		       u64 num_bytes, u64 parent, u64 min_bytes,
+		       u64 root_objectid, u64 ref_generation,
+		       u64 owner, u64 empty_size, u64 hint_byte,
+		       u64 search_end, struct btrfs_key *ins, u64 data);
+int btrfs_alloc_reserved_extent(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root, u64 parent,
+				u64 root_objectid, u64 ref_generation,
+				u64 owner, struct btrfs_key *ins);
+int btrfs_alloc_logged_extent(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root, u64 parent,
+				u64 root_objectid, u64 ref_generation,
+				u64 owner, struct btrfs_key *ins);
+int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  u64 num_bytes, u64 min_alloc_size,
+				  u64 empty_size, u64 hint_byte,
+				  u64 search_end, struct btrfs_key *ins,
+				  u64 data);
+int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		  struct extent_buffer *orig_buf, struct extent_buffer *buf,
+		  u32 *nr_extents);
+int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		    struct extent_buffer *buf, u32 nr_extents);
+int btrfs_update_ref(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root, struct extent_buffer *orig_buf,
+		     struct extent_buffer *buf, int start_slot, int nr);
+int btrfs_free_extent(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root,
+		      u64 bytenr, u64 num_bytes, u64 parent,
+		      u64 root_objectid, u64 ref_generation,
+		      u64 owner_objectid, int pin);
+int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len);
+int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct extent_io_tree *unpin);
+int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root,
+			 u64 bytenr, u64 num_bytes, u64 parent,
+			 u64 root_objectid, u64 ref_generation,
+			 u64 owner_objectid);
+int btrfs_update_extent_ref(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, u64 bytenr,
+			    u64 orig_parent, u64 parent,
+			    u64 root_objectid, u64 ref_generation,
+			    u64 owner_objectid);
+int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root);
+int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr);
+int btrfs_free_block_groups(struct btrfs_fs_info *info);
+int btrfs_read_block_groups(struct btrfs_root *root);
+int btrfs_make_block_group(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, u64 bytes_used,
+			   u64 type, u64 chunk_objectid, u64 chunk_offset,
+			   u64 size);
+int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, u64 group_start);
+int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start);
+int btrfs_free_reloc_root(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root);
+int btrfs_drop_dead_reloc_roots(struct btrfs_root *root);
+int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct extent_buffer *buf, u64 orig_start);
+int btrfs_add_dead_reloc_root(struct btrfs_root *root);
+int btrfs_cleanup_reloc_trees(struct btrfs_root *root);
+int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
+u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags);
+/* ctree.c */
+int btrfs_previous_item(struct btrfs_root *root,
+			struct btrfs_path *path, u64 min_objectid,
+			int type);
+int btrfs_merge_path(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root,
+		     struct btrfs_key *node_keys,
+		     u64 *nodes, int lowest_level);
+int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, struct btrfs_path *path,
+			    struct btrfs_key *new_key);
+struct extent_buffer *btrfs_root_node(struct btrfs_root *root);
+struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root);
+int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
+			struct btrfs_key *key, int lowest_level,
+			int cache_only, u64 min_trans);
+int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
+			 struct btrfs_key *max_key,
+			 struct btrfs_path *path, int cache_only,
+			 u64 min_trans);
+int btrfs_cow_block(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, struct extent_buffer *buf,
+		    struct extent_buffer *parent, int parent_slot,
+		    struct extent_buffer **cow_ret, u64 prealloc_dest);
+int btrfs_copy_root(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root,
+		      struct extent_buffer *buf,
+		      struct extent_buffer **cow_ret, u64 new_root_objectid);
+int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_path *path, u32 data_size);
+int btrfs_truncate_item(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct btrfs_path *path,
+			u32 new_size, int from_end);
+int btrfs_split_item(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root,
+		     struct btrfs_path *path,
+		     struct btrfs_key *new_key,
+		     unsigned long split_offset);
+int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, struct btrfs_path *p, int
+		      ins_len, int cow);
+int btrfs_realloc_node(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct extent_buffer *parent,
+		       int start_slot, int cache_only, u64 *last_ret,
+		       struct btrfs_key *progress);
+void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p);
+struct btrfs_path *btrfs_alloc_path(void);
+void btrfs_free_path(struct btrfs_path *p);
+void btrfs_init_path(struct btrfs_path *p);
+int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		   struct btrfs_path *path, int slot, int nr);
+int btrfs_del_leaf(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path, u64 bytenr);
+static inline int btrfs_del_item(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct btrfs_path *path)
+{
+	return btrfs_del_items(trans, root, path, path->slots[0], 1);
+}
+
+int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, void *data, u32 data_size);
+int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path,
+			    struct btrfs_key *cpu_key, u32 *data_size,
+			    int nr);
+int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct btrfs_path *path,
+			     struct btrfs_key *cpu_key, u32 *data_size, int nr);
+
+static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path,
+					  struct btrfs_key *key,
+					  u32 data_size)
+{
+	return btrfs_insert_empty_items(trans, root, path, key, &data_size, 1);
+}
+
+int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path);
+int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path);
+int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf);
+int btrfs_drop_snapshot(struct btrfs_trans_handle *trans, struct btrfs_root
+			*root);
+int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct extent_buffer *node,
+			struct extent_buffer *parent);
+/* root-item.c */
+int btrfs_find_root_ref(struct btrfs_root *tree_root,
+		   struct btrfs_path *path,
+		   u64 root_id, u64 ref_id);
+int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *tree_root,
+		       u64 root_id, u8 type, u64 ref_id,
+		       u64 dirid, u64 sequence,
+		       const char *name, int name_len);
+int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		   struct btrfs_key *key);
+int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, struct btrfs_root_item
+		      *item);
+int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, struct btrfs_root_item
+		      *item);
+int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct
+			 btrfs_root_item *item, struct btrfs_key *key);
+int btrfs_search_root(struct btrfs_root *root, u64 search_start,
+		      u64 *found_objectid);
+int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid,
+			  struct btrfs_root *latest_root);
+/* dir-item.c */
+int btrfs_insert_dir_item(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, const char *name,
+			  int name_len, u64 dir,
+			  struct btrfs_key *location, u8 type, u64 index);
+struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans,
+					     struct btrfs_root *root,
+					     struct btrfs_path *path, u64 dir,
+					     const char *name, int name_len,
+					     int mod);
+struct btrfs_dir_item *
+btrfs_lookup_dir_index_item(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path, u64 dir,
+			    u64 objectid, const char *name, int name_len,
+			    int mod);
+struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_root *root,
+			      struct btrfs_path *path,
+			      const char *name, int name_len);
+int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct btrfs_path *path,
+			      struct btrfs_dir_item *di);
+int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, const char *name,
+			    u16 name_len, const void *data, u16 data_len,
+			    u64 dir);
+struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path, u64 dir,
+					  const char *name, u16 name_len,
+					  int mod);
+
+/* orphan.c */
+int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, u64 offset);
+int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, u64 offset);
+
+/* inode-map.c */
+int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *fs_root,
+			     u64 dirid, u64 *objectid);
+int btrfs_find_highest_inode(struct btrfs_root *fs_root, u64 *objectid);
+
+/* inode-item.c */
+int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   const char *name, int name_len,
+			   u64 inode_objectid, u64 ref_objectid, u64 index);
+int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   const char *name, int name_len,
+			   u64 inode_objectid, u64 ref_objectid, u64 *index);
+int btrfs_insert_empty_inode(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct btrfs_path *path, u64 objectid);
+int btrfs_lookup_inode(struct btrfs_trans_handle *trans, struct btrfs_root
+		       *root, struct btrfs_path *path,
+		       struct btrfs_key *location, int mod);
+
+/* file-item.c */
+int btrfs_del_csums(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, u64 bytenr, u64 len);
+int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
+			  struct bio *bio, u32 *dst);
+int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     u64 objectid, u64 pos,
+			     u64 disk_offset, u64 disk_num_bytes,
+			     u64 num_bytes, u64 offset, u64 ram_bytes,
+			     u8 compression, u8 encryption, u16 other_encoding);
+int btrfs_lookup_file_extent(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct btrfs_path *path, u64 objectid,
+			     u64 bytenr, int mod);
+int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct btrfs_ordered_sum *sums);
+int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
+		       struct bio *bio, u64 file_start, int contig);
+int btrfs_csum_file_bytes(struct btrfs_root *root, struct inode *inode,
+			  u64 start, unsigned long len);
+struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path,
+					  u64 bytenr, int cow);
+int btrfs_csum_truncate(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root, struct btrfs_path *path,
+			u64 isize);
+int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start,
+			     u64 end, struct list_head *list);
+/* inode.c */
+
+/* RHEL and EL kernels have a patch that renames PG_checked to FsMisc */
+#if defined(ClearPageFsMisc) && !defined(ClearPageChecked)
+#define ClearPageChecked ClearPageFsMisc
+#define SetPageChecked SetPageFsMisc
+#define PageChecked PageFsMisc
+#endif
+
+struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry);
+int btrfs_set_inode_index(struct inode *dir, u64 *index);
+int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root,
+		       struct inode *dir, struct inode *inode,
+		       const char *name, int name_len);
+int btrfs_add_link(struct btrfs_trans_handle *trans,
+		   struct inode *parent_inode, struct inode *inode,
+		   const char *name, int name_len, int add_backref, u64 index);
+int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct inode *inode, u64 new_size,
+			       u32 min_type);
+
+int btrfs_start_delalloc_inodes(struct btrfs_root *root);
+int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end);
+int btrfs_writepages(struct address_space *mapping,
+		     struct writeback_control *wbc);
+int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *new_root, struct dentry *dentry,
+			     u64 new_dirid, u64 alloc_hint);
+int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
+			 size_t size, struct bio *bio, unsigned long bio_flags);
+
+unsigned long btrfs_force_ra(struct address_space *mapping,
+			      struct file_ra_state *ra, struct file *file,
+			      pgoff_t offset, pgoff_t last_index);
+int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
+			   int for_del);
+int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page);
+int btrfs_readpage(struct file *file, struct page *page);
+void btrfs_delete_inode(struct inode *inode);
+void btrfs_put_inode(struct inode *inode);
+void btrfs_read_locked_inode(struct inode *inode);
+int btrfs_write_inode(struct inode *inode, int wait);
+void btrfs_dirty_inode(struct inode *inode);
+struct inode *btrfs_alloc_inode(struct super_block *sb);
+void btrfs_destroy_inode(struct inode *inode);
+int btrfs_init_cachep(void);
+void btrfs_destroy_cachep(void);
+long btrfs_ioctl_trans_end(struct file *file);
+struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
+			    struct btrfs_root *root, int wait);
+struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
+				struct btrfs_root *root);
+struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
+			 struct btrfs_root *root, int *is_new);
+int btrfs_commit_write(struct file *file, struct page *page,
+		       unsigned from, unsigned to);
+struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
+				    size_t page_offset, u64 start, u64 end,
+				    int create);
+int btrfs_update_inode(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct inode *inode);
+int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode);
+int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode);
+void btrfs_orphan_cleanup(struct btrfs_root *root);
+int btrfs_cont_expand(struct inode *inode, loff_t size);
+
+/* ioctl.c */
+long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
+
+/* file.c */
+int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync);
+int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
+			    int skip_pinned);
+int btrfs_check_file(struct btrfs_root *root, struct inode *inode);
+extern struct file_operations btrfs_file_operations;
+int btrfs_drop_extents(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct inode *inode,
+		       u64 start, u64 end, u64 inline_limit, u64 *hint_block);
+int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct inode *inode, u64 start, u64 end);
+int btrfs_release_file(struct inode *inode, struct file *file);
+
+/* tree-defrag.c */
+int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root, int cache_only);
+
+/* sysfs.c */
+int btrfs_init_sysfs(void);
+void btrfs_exit_sysfs(void);
+int btrfs_sysfs_add_super(struct btrfs_fs_info *fs);
+int btrfs_sysfs_add_root(struct btrfs_root *root);
+void btrfs_sysfs_del_root(struct btrfs_root *root);
+void btrfs_sysfs_del_super(struct btrfs_fs_info *root);
+
+/* xattr.c */
+ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size);
+
+/* super.c */
+u64 btrfs_parse_size(char *str);
+int btrfs_parse_options(struct btrfs_root *root, char *options);
+int btrfs_sync_fs(struct super_block *sb, int wait);
+
+/* acl.c */
+int btrfs_check_acl(struct inode *inode, int mask);
+int btrfs_init_acl(struct inode *inode, struct inode *dir);
+int btrfs_acl_chmod(struct inode *inode);
+
+/* free-space-cache.c */
+int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
+			 u64 bytenr, u64 size);
+int btrfs_add_free_space_lock(struct btrfs_block_group_cache *block_group,
+			      u64 offset, u64 bytes);
+int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
+			    u64 bytenr, u64 size);
+int btrfs_remove_free_space_lock(struct btrfs_block_group_cache *block_group,
+				 u64 offset, u64 bytes);
+void btrfs_remove_free_space_cache(struct btrfs_block_group_cache
+				   *block_group);
+struct btrfs_free_space *btrfs_find_free_space(struct btrfs_block_group_cache
+					       *block_group, u64 offset,
+					       u64 bytes);
+void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
+			   u64 bytes);
+u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group);
+#endif
diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c
new file mode 100644
index 00000000000..926a0b287a7
--- /dev/null
+++ b/fs/btrfs/dir-item.c
@@ -0,0 +1,386 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "hash.h"
+#include "transaction.h"
+
+/*
+ * insert a name into a directory, doing overflow properly if there is a hash
+ * collision.  data_size indicates how big the item inserted should be.  On
+ * success a struct btrfs_dir_item pointer is returned, otherwise it is
+ * an ERR_PTR.
+ *
+ * The name is not copied into the dir item, you have to do that yourself.
+ */
+static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle
+						   *trans,
+						   struct btrfs_root *root,
+						   struct btrfs_path *path,
+						   struct btrfs_key *cpu_key,
+						   u32 data_size,
+						   const char *name,
+						   int name_len)
+{
+	int ret;
+	char *ptr;
+	struct btrfs_item *item;
+	struct extent_buffer *leaf;
+
+	ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size);
+	if (ret == -EEXIST) {
+		struct btrfs_dir_item *di;
+		di = btrfs_match_dir_item_name(root, path, name, name_len);
+		if (di)
+			return ERR_PTR(-EEXIST);
+		ret = btrfs_extend_item(trans, root, path, data_size);
+		WARN_ON(ret > 0);
+	}
+	if (ret < 0)
+		return ERR_PTR(ret);
+	WARN_ON(ret > 0);
+	leaf = path->nodes[0];
+	item = btrfs_item_nr(leaf, path->slots[0]);
+	ptr = btrfs_item_ptr(leaf, path->slots[0], char);
+	BUG_ON(data_size > btrfs_item_size(leaf, item));
+	ptr += btrfs_item_size(leaf, item) - data_size;
+	return (struct btrfs_dir_item *)ptr;
+}
+
+/*
+ * xattrs work a lot like directories, this inserts an xattr item
+ * into the tree
+ */
+int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, const char *name,
+			    u16 name_len, const void *data, u16 data_len,
+			    u64 dir)
+{
+	int ret = 0;
+	struct btrfs_path *path;
+	struct btrfs_dir_item *dir_item;
+	unsigned long name_ptr, data_ptr;
+	struct btrfs_key key, location;
+	struct btrfs_disk_key disk_key;
+	struct extent_buffer *leaf;
+	u32 data_size;
+
+	key.objectid = dir;
+	btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY);
+	key.offset = btrfs_name_hash(name, name_len);
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	if (name_len + data_len + sizeof(struct btrfs_dir_item) >
+	    BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item))
+		return -ENOSPC;
+
+	data_size = sizeof(*dir_item) + name_len + data_len;
+	dir_item = insert_with_overflow(trans, root, path, &key, data_size,
+					name, name_len);
+	/*
+	 * FIXME: at some point we should handle xattr's that are larger than
+	 * what we can fit in our leaf.  We set location to NULL b/c we arent
+	 * pointing at anything else, that will change if we store the xattr
+	 * data in a separate inode.
+	 */
+	BUG_ON(IS_ERR(dir_item));
+	memset(&location, 0, sizeof(location));
+
+	leaf = path->nodes[0];
+	btrfs_cpu_key_to_disk(&disk_key, &location);
+	btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
+	btrfs_set_dir_type(leaf, dir_item, BTRFS_FT_XATTR);
+	btrfs_set_dir_name_len(leaf, dir_item, name_len);
+	btrfs_set_dir_transid(leaf, dir_item, trans->transid);
+	btrfs_set_dir_data_len(leaf, dir_item, data_len);
+	name_ptr = (unsigned long)(dir_item + 1);
+	data_ptr = (unsigned long)((char *)name_ptr + name_len);
+
+	write_extent_buffer(leaf, name, name_ptr, name_len);
+	write_extent_buffer(leaf, data, data_ptr, data_len);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * insert a directory item in the tree, doing all the magic for
+ * both indexes. 'dir' indicates which objectid to insert it into,
+ * 'location' is the key to stuff into the directory item, 'type' is the
+ * type of the inode we're pointing to, and 'index' is the sequence number
+ * to use for the second index (if one is created).
+ */
+int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root
+			  *root, const char *name, int name_len, u64 dir,
+			  struct btrfs_key *location, u8 type, u64 index)
+{
+	int ret = 0;
+	int ret2 = 0;
+	struct btrfs_path *path;
+	struct btrfs_dir_item *dir_item;
+	struct extent_buffer *leaf;
+	unsigned long name_ptr;
+	struct btrfs_key key;
+	struct btrfs_disk_key disk_key;
+	u32 data_size;
+
+	key.objectid = dir;
+	btrfs_set_key_type(&key, BTRFS_DIR_ITEM_KEY);
+	key.offset = btrfs_name_hash(name, name_len);
+	path = btrfs_alloc_path();
+	data_size = sizeof(*dir_item) + name_len;
+	dir_item = insert_with_overflow(trans, root, path, &key, data_size,
+					name, name_len);
+	if (IS_ERR(dir_item)) {
+		ret = PTR_ERR(dir_item);
+		if (ret == -EEXIST)
+			goto second_insert;
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	btrfs_cpu_key_to_disk(&disk_key, location);
+	btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
+	btrfs_set_dir_type(leaf, dir_item, type);
+	btrfs_set_dir_data_len(leaf, dir_item, 0);
+	btrfs_set_dir_name_len(leaf, dir_item, name_len);
+	btrfs_set_dir_transid(leaf, dir_item, trans->transid);
+	name_ptr = (unsigned long)(dir_item + 1);
+
+	write_extent_buffer(leaf, name, name_ptr, name_len);
+	btrfs_mark_buffer_dirty(leaf);
+
+second_insert:
+	/* FIXME, use some real flag for selecting the extra index */
+	if (root == root->fs_info->tree_root) {
+		ret = 0;
+		goto out;
+	}
+	btrfs_release_path(root, path);
+
+	btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY);
+	key.offset = index;
+	dir_item = insert_with_overflow(trans, root, path, &key, data_size,
+					name, name_len);
+	if (IS_ERR(dir_item)) {
+		ret2 = PTR_ERR(dir_item);
+		goto out;
+	}
+	leaf = path->nodes[0];
+	btrfs_cpu_key_to_disk(&disk_key, location);
+	btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
+	btrfs_set_dir_type(leaf, dir_item, type);
+	btrfs_set_dir_data_len(leaf, dir_item, 0);
+	btrfs_set_dir_name_len(leaf, dir_item, name_len);
+	btrfs_set_dir_transid(leaf, dir_item, trans->transid);
+	name_ptr = (unsigned long)(dir_item + 1);
+	write_extent_buffer(leaf, name, name_ptr, name_len);
+	btrfs_mark_buffer_dirty(leaf);
+out:
+	btrfs_free_path(path);
+	if (ret)
+		return ret;
+	if (ret2)
+		return ret2;
+	return 0;
+}
+
+/*
+ * lookup a directory item based on name.  'dir' is the objectid
+ * we're searching in, and 'mod' tells us if you plan on deleting the
+ * item (use mod < 0) or changing the options (use mod > 0)
+ */
+struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans,
+					     struct btrfs_root *root,
+					     struct btrfs_path *path, u64 dir,
+					     const char *name, int name_len,
+					     int mod)
+{
+	int ret;
+	struct btrfs_key key;
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf;
+
+	key.objectid = dir;
+	btrfs_set_key_type(&key, BTRFS_DIR_ITEM_KEY);
+
+	key.offset = btrfs_name_hash(name, name_len);
+
+	ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
+	if (ret < 0)
+		return ERR_PTR(ret);
+	if (ret > 0) {
+		if (path->slots[0] == 0)
+			return NULL;
+		path->slots[0]--;
+	}
+
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+	if (found_key.objectid != dir ||
+	    btrfs_key_type(&found_key) != BTRFS_DIR_ITEM_KEY ||
+	    found_key.offset != key.offset)
+		return NULL;
+
+	return btrfs_match_dir_item_name(root, path, name, name_len);
+}
+
+/*
+ * lookup a directory item based on index.  'dir' is the objectid
+ * we're searching in, and 'mod' tells us if you plan on deleting the
+ * item (use mod < 0) or changing the options (use mod > 0)
+ *
+ * The name is used to make sure the index really points to the name you were
+ * looking for.
+ */
+struct btrfs_dir_item *
+btrfs_lookup_dir_index_item(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    struct btrfs_path *path, u64 dir,
+			    u64 objectid, const char *name, int name_len,
+			    int mod)
+{
+	int ret;
+	struct btrfs_key key;
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+
+	key.objectid = dir;
+	btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY);
+	key.offset = objectid;
+
+	ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
+	if (ret < 0)
+		return ERR_PTR(ret);
+	if (ret > 0)
+		return ERR_PTR(-ENOENT);
+	return btrfs_match_dir_item_name(root, path, name, name_len);
+}
+
+struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path, u64 dir,
+					  const char *name, u16 name_len,
+					  int mod)
+{
+	int ret;
+	struct btrfs_key key;
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf;
+
+	key.objectid = dir;
+	btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY);
+	key.offset = btrfs_name_hash(name, name_len);
+	ret = btrfs_search_slot(trans, root, &key, path, ins_len, cow);
+	if (ret < 0)
+		return ERR_PTR(ret);
+	if (ret > 0) {
+		if (path->slots[0] == 0)
+			return NULL;
+		path->slots[0]--;
+	}
+
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+	if (found_key.objectid != dir ||
+	    btrfs_key_type(&found_key) != BTRFS_XATTR_ITEM_KEY ||
+	    found_key.offset != key.offset)
+		return NULL;
+
+	return btrfs_match_dir_item_name(root, path, name, name_len);
+}
+
+/*
+ * helper function to look at the directory item pointed to by 'path'
+ * this walks through all the entries in a dir item and finds one
+ * for a specific name.
+ */
+struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_root *root,
+			      struct btrfs_path *path,
+			      const char *name, int name_len)
+{
+	struct btrfs_dir_item *dir_item;
+	unsigned long name_ptr;
+	u32 total_len;
+	u32 cur = 0;
+	u32 this_len;
+	struct extent_buffer *leaf;
+
+	leaf = path->nodes[0];
+	dir_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dir_item);
+	total_len = btrfs_item_size_nr(leaf, path->slots[0]);
+	while (cur < total_len) {
+		this_len = sizeof(*dir_item) +
+			btrfs_dir_name_len(leaf, dir_item) +
+			btrfs_dir_data_len(leaf, dir_item);
+		name_ptr = (unsigned long)(dir_item + 1);
+
+		if (btrfs_dir_name_len(leaf, dir_item) == name_len &&
+		    memcmp_extent_buffer(leaf, name, name_ptr, name_len) == 0)
+			return dir_item;
+
+		cur += this_len;
+		dir_item = (struct btrfs_dir_item *)((char *)dir_item +
+						     this_len);
+	}
+	return NULL;
+}
+
+/*
+ * given a pointer into a directory item, delete it.  This
+ * handles items that have more than one entry in them.
+ */
+int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct btrfs_path *path,
+			      struct btrfs_dir_item *di)
+{
+
+	struct extent_buffer *leaf;
+	u32 sub_item_len;
+	u32 item_len;
+	int ret = 0;
+
+	leaf = path->nodes[0];
+	sub_item_len = sizeof(*di) + btrfs_dir_name_len(leaf, di) +
+		btrfs_dir_data_len(leaf, di);
+	item_len = btrfs_item_size_nr(leaf, path->slots[0]);
+	if (sub_item_len == item_len) {
+		ret = btrfs_del_item(trans, root, path);
+	} else {
+		/* MARKER */
+		unsigned long ptr = (unsigned long)di;
+		unsigned long start;
+
+		start = btrfs_item_ptr_offset(leaf, path->slots[0]);
+		memmove_extent_buffer(leaf, ptr, ptr + sub_item_len,
+			item_len - (ptr + sub_item_len - start));
+		ret = btrfs_truncate_item(trans, root, path,
+					  item_len - sub_item_len, 1);
+	}
+	return 0;
+}
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
new file mode 100644
index 00000000000..81a313874ae
--- /dev/null
+++ b/fs/btrfs/disk-io.c
@@ -0,0 +1,2343 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/version.h>
+#include <linux/fs.h>
+#include <linux/blkdev.h>
+#include <linux/scatterlist.h>
+#include <linux/swap.h>
+#include <linux/radix-tree.h>
+#include <linux/writeback.h>
+#include <linux/buffer_head.h>
+#include <linux/workqueue.h>
+#include <linux/kthread.h>
+#include <linux/freezer.h>
+#include "compat.h"
+#include "crc32c.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "volumes.h"
+#include "print-tree.h"
+#include "async-thread.h"
+#include "locking.h"
+#include "ref-cache.h"
+#include "tree-log.h"
+
+static struct extent_io_ops btree_extent_io_ops;
+static void end_workqueue_fn(struct btrfs_work *work);
+
+/*
+ * end_io_wq structs are used to do processing in task context when an IO is
+ * complete.  This is used during reads to verify checksums, and it is used
+ * by writes to insert metadata for new file extents after IO is complete.
+ */
+struct end_io_wq {
+	struct bio *bio;
+	bio_end_io_t *end_io;
+	void *private;
+	struct btrfs_fs_info *info;
+	int error;
+	int metadata;
+	struct list_head list;
+	struct btrfs_work work;
+};
+
+/*
+ * async submit bios are used to offload expensive checksumming
+ * onto the worker threads.  They checksum file and metadata bios
+ * just before they are sent down the IO stack.
+ */
+struct async_submit_bio {
+	struct inode *inode;
+	struct bio *bio;
+	struct list_head list;
+	extent_submit_bio_hook_t *submit_bio_start;
+	extent_submit_bio_hook_t *submit_bio_done;
+	int rw;
+	int mirror_num;
+	unsigned long bio_flags;
+	struct btrfs_work work;
+};
+
+/*
+ * extents on the btree inode are pretty simple, there's one extent
+ * that covers the entire device
+ */
+static struct extent_map *btree_get_extent(struct inode *inode,
+		struct page *page, size_t page_offset, u64 start, u64 len,
+		int create)
+{
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct extent_map *em;
+	int ret;
+
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, start, len);
+	if (em) {
+		em->bdev =
+			BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
+		spin_unlock(&em_tree->lock);
+		goto out;
+	}
+	spin_unlock(&em_tree->lock);
+
+	em = alloc_extent_map(GFP_NOFS);
+	if (!em) {
+		em = ERR_PTR(-ENOMEM);
+		goto out;
+	}
+	em->start = 0;
+	em->len = (u64)-1;
+	em->block_len = (u64)-1;
+	em->block_start = 0;
+	em->bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
+
+	spin_lock(&em_tree->lock);
+	ret = add_extent_mapping(em_tree, em);
+	if (ret == -EEXIST) {
+		u64 failed_start = em->start;
+		u64 failed_len = em->len;
+
+		free_extent_map(em);
+		em = lookup_extent_mapping(em_tree, start, len);
+		if (em) {
+			ret = 0;
+		} else {
+			em = lookup_extent_mapping(em_tree, failed_start,
+						   failed_len);
+			ret = -EIO;
+		}
+	} else if (ret) {
+		free_extent_map(em);
+		em = NULL;
+	}
+	spin_unlock(&em_tree->lock);
+
+	if (ret)
+		em = ERR_PTR(ret);
+out:
+	return em;
+}
+
+u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len)
+{
+	return btrfs_crc32c(seed, data, len);
+}
+
+void btrfs_csum_final(u32 crc, char *result)
+{
+	*(__le32 *)result = ~cpu_to_le32(crc);
+}
+
+/*
+ * compute the csum for a btree block, and either verify it or write it
+ * into the csum field of the block.
+ */
+static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
+			   int verify)
+{
+	u16 csum_size =
+		btrfs_super_csum_size(&root->fs_info->super_copy);
+	char *result = NULL;
+	unsigned long len;
+	unsigned long cur_len;
+	unsigned long offset = BTRFS_CSUM_SIZE;
+	char *map_token = NULL;
+	char *kaddr;
+	unsigned long map_start;
+	unsigned long map_len;
+	int err;
+	u32 crc = ~(u32)0;
+	unsigned long inline_result;
+
+	len = buf->len - offset;
+	while (len > 0) {
+		err = map_private_extent_buffer(buf, offset, 32,
+					&map_token, &kaddr,
+					&map_start, &map_len, KM_USER0);
+		if (err)
+			return 1;
+		cur_len = min(len, map_len - (offset - map_start));
+		crc = btrfs_csum_data(root, kaddr + offset - map_start,
+				      crc, cur_len);
+		len -= cur_len;
+		offset += cur_len;
+		unmap_extent_buffer(buf, map_token, KM_USER0);
+	}
+	if (csum_size > sizeof(inline_result)) {
+		result = kzalloc(csum_size * sizeof(char), GFP_NOFS);
+		if (!result)
+			return 1;
+	} else {
+		result = (char *)&inline_result;
+	}
+
+	btrfs_csum_final(crc, result);
+
+	if (verify) {
+		if (memcmp_extent_buffer(buf, result, 0, csum_size)) {
+			u32 val;
+			u32 found = 0;
+			memcpy(&found, result, csum_size);
+
+			read_extent_buffer(buf, &val, 0, csum_size);
+			printk(KERN_INFO "btrfs: %s checksum verify failed "
+			       "on %llu wanted %X found %X level %d\n",
+			       root->fs_info->sb->s_id,
+			       buf->start, val, found, btrfs_header_level(buf));
+			if (result != (char *)&inline_result)
+				kfree(result);
+			return 1;
+		}
+	} else {
+		write_extent_buffer(buf, result, 0, csum_size);
+	}
+	if (result != (char *)&inline_result)
+		kfree(result);
+	return 0;
+}
+
+/*
+ * we can't consider a given block up to date unless the transid of the
+ * block matches the transid in the parent node's pointer.  This is how we
+ * detect blocks that either didn't get written at all or got written
+ * in the wrong place.
+ */
+static int verify_parent_transid(struct extent_io_tree *io_tree,
+				 struct extent_buffer *eb, u64 parent_transid)
+{
+	int ret;
+
+	if (!parent_transid || btrfs_header_generation(eb) == parent_transid)
+		return 0;
+
+	lock_extent(io_tree, eb->start, eb->start + eb->len - 1, GFP_NOFS);
+	if (extent_buffer_uptodate(io_tree, eb) &&
+	    btrfs_header_generation(eb) == parent_transid) {
+		ret = 0;
+		goto out;
+	}
+	printk("parent transid verify failed on %llu wanted %llu found %llu\n",
+	       (unsigned long long)eb->start,
+	       (unsigned long long)parent_transid,
+	       (unsigned long long)btrfs_header_generation(eb));
+	ret = 1;
+	clear_extent_buffer_uptodate(io_tree, eb);
+out:
+	unlock_extent(io_tree, eb->start, eb->start + eb->len - 1,
+		      GFP_NOFS);
+	return ret;
+}
+
+/*
+ * helper to read a given tree block, doing retries as required when
+ * the checksums don't match and we have alternate mirrors to try.
+ */
+static int btree_read_extent_buffer_pages(struct btrfs_root *root,
+					  struct extent_buffer *eb,
+					  u64 start, u64 parent_transid)
+{
+	struct extent_io_tree *io_tree;
+	int ret;
+	int num_copies = 0;
+	int mirror_num = 0;
+
+	io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
+	while (1) {
+		ret = read_extent_buffer_pages(io_tree, eb, start, 1,
+					       btree_get_extent, mirror_num);
+		if (!ret &&
+		    !verify_parent_transid(io_tree, eb, parent_transid))
+			return ret;
+
+		num_copies = btrfs_num_copies(&root->fs_info->mapping_tree,
+					      eb->start, eb->len);
+		if (num_copies == 1)
+			return ret;
+
+		mirror_num++;
+		if (mirror_num > num_copies)
+			return ret;
+	}
+	return -EIO;
+}
+
+/*
+ * checksum a dirty tree block before IO.  This has extra checks to make sure
+ * we only fill in the checksum field in the first page of a multi-page block
+ */
+
+static int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
+{
+	struct extent_io_tree *tree;
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 found_start;
+	int found_level;
+	unsigned long len;
+	struct extent_buffer *eb;
+	int ret;
+
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+
+	if (page->private == EXTENT_PAGE_PRIVATE)
+		goto out;
+	if (!page->private)
+		goto out;
+	len = page->private >> 2;
+	WARN_ON(len == 0);
+
+	eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
+	ret = btree_read_extent_buffer_pages(root, eb, start + PAGE_CACHE_SIZE,
+					     btrfs_header_generation(eb));
+	BUG_ON(ret);
+	found_start = btrfs_header_bytenr(eb);
+	if (found_start != start) {
+		WARN_ON(1);
+		goto err;
+	}
+	if (eb->first_page != page) {
+		WARN_ON(1);
+		goto err;
+	}
+	if (!PageUptodate(page)) {
+		WARN_ON(1);
+		goto err;
+	}
+	found_level = btrfs_header_level(eb);
+
+	csum_tree_block(root, eb, 0);
+err:
+	free_extent_buffer(eb);
+out:
+	return 0;
+}
+
+static int check_tree_block_fsid(struct btrfs_root *root,
+				 struct extent_buffer *eb)
+{
+	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
+	u8 fsid[BTRFS_UUID_SIZE];
+	int ret = 1;
+
+	read_extent_buffer(eb, fsid, (unsigned long)btrfs_header_fsid(eb),
+			   BTRFS_FSID_SIZE);
+	while (fs_devices) {
+		if (!memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE)) {
+			ret = 0;
+			break;
+		}
+		fs_devices = fs_devices->seed;
+	}
+	return ret;
+}
+
+static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
+			       struct extent_state *state)
+{
+	struct extent_io_tree *tree;
+	u64 found_start;
+	int found_level;
+	unsigned long len;
+	struct extent_buffer *eb;
+	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
+	int ret = 0;
+
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	if (page->private == EXTENT_PAGE_PRIVATE)
+		goto out;
+	if (!page->private)
+		goto out;
+
+	len = page->private >> 2;
+	WARN_ON(len == 0);
+
+	eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
+
+	found_start = btrfs_header_bytenr(eb);
+	if (found_start != start) {
+		printk(KERN_INFO "btrfs bad tree block start %llu %llu\n",
+		       (unsigned long long)found_start,
+		       (unsigned long long)eb->start);
+		ret = -EIO;
+		goto err;
+	}
+	if (eb->first_page != page) {
+		printk(KERN_INFO "btrfs bad first page %lu %lu\n",
+		       eb->first_page->index, page->index);
+		WARN_ON(1);
+		ret = -EIO;
+		goto err;
+	}
+	if (check_tree_block_fsid(root, eb)) {
+		printk(KERN_INFO "btrfs bad fsid on block %llu\n",
+		       (unsigned long long)eb->start);
+		ret = -EIO;
+		goto err;
+	}
+	found_level = btrfs_header_level(eb);
+
+	ret = csum_tree_block(root, eb, 1);
+	if (ret)
+		ret = -EIO;
+
+	end = min_t(u64, eb->len, PAGE_CACHE_SIZE);
+	end = eb->start + end - 1;
+err:
+	free_extent_buffer(eb);
+out:
+	return ret;
+}
+
+static void end_workqueue_bio(struct bio *bio, int err)
+{
+	struct end_io_wq *end_io_wq = bio->bi_private;
+	struct btrfs_fs_info *fs_info;
+
+	fs_info = end_io_wq->info;
+	end_io_wq->error = err;
+	end_io_wq->work.func = end_workqueue_fn;
+	end_io_wq->work.flags = 0;
+
+	if (bio->bi_rw & (1 << BIO_RW)) {
+		if (end_io_wq->metadata)
+			btrfs_queue_worker(&fs_info->endio_meta_write_workers,
+					   &end_io_wq->work);
+		else
+			btrfs_queue_worker(&fs_info->endio_write_workers,
+					   &end_io_wq->work);
+	} else {
+		if (end_io_wq->metadata)
+			btrfs_queue_worker(&fs_info->endio_meta_workers,
+					   &end_io_wq->work);
+		else
+			btrfs_queue_worker(&fs_info->endio_workers,
+					   &end_io_wq->work);
+	}
+}
+
+int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
+			int metadata)
+{
+	struct end_io_wq *end_io_wq;
+	end_io_wq = kmalloc(sizeof(*end_io_wq), GFP_NOFS);
+	if (!end_io_wq)
+		return -ENOMEM;
+
+	end_io_wq->private = bio->bi_private;
+	end_io_wq->end_io = bio->bi_end_io;
+	end_io_wq->info = info;
+	end_io_wq->error = 0;
+	end_io_wq->bio = bio;
+	end_io_wq->metadata = metadata;
+
+	bio->bi_private = end_io_wq;
+	bio->bi_end_io = end_workqueue_bio;
+	return 0;
+}
+
+unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info)
+{
+	unsigned long limit = min_t(unsigned long,
+				    info->workers.max_workers,
+				    info->fs_devices->open_devices);
+	return 256 * limit;
+}
+
+int btrfs_congested_async(struct btrfs_fs_info *info, int iodone)
+{
+	return atomic_read(&info->nr_async_bios) >
+		btrfs_async_submit_limit(info);
+}
+
+static void run_one_async_start(struct btrfs_work *work)
+{
+	struct btrfs_fs_info *fs_info;
+	struct async_submit_bio *async;
+
+	async = container_of(work, struct  async_submit_bio, work);
+	fs_info = BTRFS_I(async->inode)->root->fs_info;
+	async->submit_bio_start(async->inode, async->rw, async->bio,
+			       async->mirror_num, async->bio_flags);
+}
+
+static void run_one_async_done(struct btrfs_work *work)
+{
+	struct btrfs_fs_info *fs_info;
+	struct async_submit_bio *async;
+	int limit;
+
+	async = container_of(work, struct  async_submit_bio, work);
+	fs_info = BTRFS_I(async->inode)->root->fs_info;
+
+	limit = btrfs_async_submit_limit(fs_info);
+	limit = limit * 2 / 3;
+
+	atomic_dec(&fs_info->nr_async_submits);
+
+	if (atomic_read(&fs_info->nr_async_submits) < limit &&
+	    waitqueue_active(&fs_info->async_submit_wait))
+		wake_up(&fs_info->async_submit_wait);
+
+	async->submit_bio_done(async->inode, async->rw, async->bio,
+			       async->mirror_num, async->bio_flags);
+}
+
+static void run_one_async_free(struct btrfs_work *work)
+{
+	struct async_submit_bio *async;
+
+	async = container_of(work, struct  async_submit_bio, work);
+	kfree(async);
+}
+
+int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
+			int rw, struct bio *bio, int mirror_num,
+			unsigned long bio_flags,
+			extent_submit_bio_hook_t *submit_bio_start,
+			extent_submit_bio_hook_t *submit_bio_done)
+{
+	struct async_submit_bio *async;
+
+	async = kmalloc(sizeof(*async), GFP_NOFS);
+	if (!async)
+		return -ENOMEM;
+
+	async->inode = inode;
+	async->rw = rw;
+	async->bio = bio;
+	async->mirror_num = mirror_num;
+	async->submit_bio_start = submit_bio_start;
+	async->submit_bio_done = submit_bio_done;
+
+	async->work.func = run_one_async_start;
+	async->work.ordered_func = run_one_async_done;
+	async->work.ordered_free = run_one_async_free;
+
+	async->work.flags = 0;
+	async->bio_flags = bio_flags;
+
+	atomic_inc(&fs_info->nr_async_submits);
+	btrfs_queue_worker(&fs_info->workers, &async->work);
+#if 0
+	int limit = btrfs_async_submit_limit(fs_info);
+	if (atomic_read(&fs_info->nr_async_submits) > limit) {
+		wait_event_timeout(fs_info->async_submit_wait,
+			   (atomic_read(&fs_info->nr_async_submits) < limit),
+			   HZ/10);
+
+		wait_event_timeout(fs_info->async_submit_wait,
+			   (atomic_read(&fs_info->nr_async_bios) < limit),
+			   HZ/10);
+	}
+#endif
+	while (atomic_read(&fs_info->async_submit_draining) &&
+	      atomic_read(&fs_info->nr_async_submits)) {
+		wait_event(fs_info->async_submit_wait,
+			   (atomic_read(&fs_info->nr_async_submits) == 0));
+	}
+
+	return 0;
+}
+
+static int btree_csum_one_bio(struct bio *bio)
+{
+	struct bio_vec *bvec = bio->bi_io_vec;
+	int bio_index = 0;
+	struct btrfs_root *root;
+
+	WARN_ON(bio->bi_vcnt <= 0);
+	while (bio_index < bio->bi_vcnt) {
+		root = BTRFS_I(bvec->bv_page->mapping->host)->root;
+		csum_dirty_buffer(root, bvec->bv_page);
+		bio_index++;
+		bvec++;
+	}
+	return 0;
+}
+
+static int __btree_submit_bio_start(struct inode *inode, int rw,
+				    struct bio *bio, int mirror_num,
+				    unsigned long bio_flags)
+{
+	/*
+	 * when we're called for a write, we're already in the async
+	 * submission context.  Just jump into btrfs_map_bio
+	 */
+	btree_csum_one_bio(bio);
+	return 0;
+}
+
+static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
+				 int mirror_num, unsigned long bio_flags)
+{
+	/*
+	 * when we're called for a write, we're already in the async
+	 * submission context.  Just jump into btrfs_map_bio
+	 */
+	return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, mirror_num, 1);
+}
+
+static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
+				 int mirror_num, unsigned long bio_flags)
+{
+	int ret;
+
+	ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info,
+					  bio, 1);
+	BUG_ON(ret);
+
+	if (!(rw & (1 << BIO_RW))) {
+		/*
+		 * called for a read, do the setup so that checksum validation
+		 * can happen in the async kernel threads
+		 */
+		return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio,
+				     mirror_num, 0);
+	}
+	/*
+	 * kthread helpers are used to submit writes so that checksumming
+	 * can happen in parallel across all CPUs
+	 */
+	return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
+				   inode, rw, bio, mirror_num, 0,
+				   __btree_submit_bio_start,
+				   __btree_submit_bio_done);
+}
+
+static int btree_writepage(struct page *page, struct writeback_control *wbc)
+{
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+
+	if (current->flags & PF_MEMALLOC) {
+		redirty_page_for_writepage(wbc, page);
+		unlock_page(page);
+		return 0;
+	}
+	return extent_write_full_page(tree, page, btree_get_extent, wbc);
+}
+
+static int btree_writepages(struct address_space *mapping,
+			    struct writeback_control *wbc)
+{
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(mapping->host)->io_tree;
+	if (wbc->sync_mode == WB_SYNC_NONE) {
+		u64 num_dirty;
+		u64 start = 0;
+		unsigned long thresh = 32 * 1024 * 1024;
+
+		if (wbc->for_kupdate)
+			return 0;
+
+		num_dirty = count_range_bits(tree, &start, (u64)-1,
+					     thresh, EXTENT_DIRTY);
+		if (num_dirty < thresh)
+			return 0;
+	}
+	return extent_writepages(tree, mapping, btree_get_extent, wbc);
+}
+
+static int btree_readpage(struct file *file, struct page *page)
+{
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	return extent_read_full_page(tree, page, btree_get_extent);
+}
+
+static int btree_releasepage(struct page *page, gfp_t gfp_flags)
+{
+	struct extent_io_tree *tree;
+	struct extent_map_tree *map;
+	int ret;
+
+	if (PageWriteback(page) || PageDirty(page))
+		return 0;
+
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	map = &BTRFS_I(page->mapping->host)->extent_tree;
+
+	ret = try_release_extent_state(map, tree, page, gfp_flags);
+	if (!ret)
+		return 0;
+
+	ret = try_release_extent_buffer(tree, page);
+	if (ret == 1) {
+		ClearPagePrivate(page);
+		set_page_private(page, 0);
+		page_cache_release(page);
+	}
+
+	return ret;
+}
+
+static void btree_invalidatepage(struct page *page, unsigned long offset)
+{
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	extent_invalidatepage(tree, page, offset);
+	btree_releasepage(page, GFP_NOFS);
+	if (PagePrivate(page)) {
+		printk(KERN_WARNING "btrfs warning page private not zero "
+		       "on page %llu\n", (unsigned long long)page_offset(page));
+		ClearPagePrivate(page);
+		set_page_private(page, 0);
+		page_cache_release(page);
+	}
+}
+
+#if 0
+static int btree_writepage(struct page *page, struct writeback_control *wbc)
+{
+	struct buffer_head *bh;
+	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
+	struct buffer_head *head;
+	if (!page_has_buffers(page)) {
+		create_empty_buffers(page, root->fs_info->sb->s_blocksize,
+					(1 << BH_Dirty)|(1 << BH_Uptodate));
+	}
+	head = page_buffers(page);
+	bh = head;
+	do {
+		if (buffer_dirty(bh))
+			csum_tree_block(root, bh, 0);
+		bh = bh->b_this_page;
+	} while (bh != head);
+	return block_write_full_page(page, btree_get_block, wbc);
+}
+#endif
+
+static struct address_space_operations btree_aops = {
+	.readpage	= btree_readpage,
+	.writepage	= btree_writepage,
+	.writepages	= btree_writepages,
+	.releasepage	= btree_releasepage,
+	.invalidatepage = btree_invalidatepage,
+	.sync_page	= block_sync_page,
+};
+
+int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize,
+			 u64 parent_transid)
+{
+	struct extent_buffer *buf = NULL;
+	struct inode *btree_inode = root->fs_info->btree_inode;
+	int ret = 0;
+
+	buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
+	if (!buf)
+		return 0;
+	read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
+				 buf, 0, 0, btree_get_extent, 0);
+	free_extent_buffer(buf);
+	return ret;
+}
+
+struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
+					    u64 bytenr, u32 blocksize)
+{
+	struct inode *btree_inode = root->fs_info->btree_inode;
+	struct extent_buffer *eb;
+	eb = find_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
+				bytenr, blocksize, GFP_NOFS);
+	return eb;
+}
+
+struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
+						 u64 bytenr, u32 blocksize)
+{
+	struct inode *btree_inode = root->fs_info->btree_inode;
+	struct extent_buffer *eb;
+
+	eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
+				 bytenr, blocksize, NULL, GFP_NOFS);
+	return eb;
+}
+
+
+int btrfs_write_tree_block(struct extent_buffer *buf)
+{
+	return btrfs_fdatawrite_range(buf->first_page->mapping, buf->start,
+				      buf->start + buf->len - 1, WB_SYNC_ALL);
+}
+
+int btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
+{
+	return btrfs_wait_on_page_writeback_range(buf->first_page->mapping,
+				  buf->start, buf->start + buf->len - 1);
+}
+
+struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
+				      u32 blocksize, u64 parent_transid)
+{
+	struct extent_buffer *buf = NULL;
+	struct inode *btree_inode = root->fs_info->btree_inode;
+	struct extent_io_tree *io_tree;
+	int ret;
+
+	io_tree = &BTRFS_I(btree_inode)->io_tree;
+
+	buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
+	if (!buf)
+		return NULL;
+
+	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
+
+	if (ret == 0)
+		buf->flags |= EXTENT_UPTODATE;
+	else
+		WARN_ON(1);
+	return buf;
+
+}
+
+int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		     struct extent_buffer *buf)
+{
+	struct inode *btree_inode = root->fs_info->btree_inode;
+	if (btrfs_header_generation(buf) ==
+	    root->fs_info->running_transaction->transid) {
+		WARN_ON(!btrfs_tree_locked(buf));
+		clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
+					  buf);
+	}
+	return 0;
+}
+
+static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
+			u32 stripesize, struct btrfs_root *root,
+			struct btrfs_fs_info *fs_info,
+			u64 objectid)
+{
+	root->node = NULL;
+	root->commit_root = NULL;
+	root->ref_tree = NULL;
+	root->sectorsize = sectorsize;
+	root->nodesize = nodesize;
+	root->leafsize = leafsize;
+	root->stripesize = stripesize;
+	root->ref_cows = 0;
+	root->track_dirty = 0;
+
+	root->fs_info = fs_info;
+	root->objectid = objectid;
+	root->last_trans = 0;
+	root->highest_inode = 0;
+	root->last_inode_alloc = 0;
+	root->name = NULL;
+	root->in_sysfs = 0;
+
+	INIT_LIST_HEAD(&root->dirty_list);
+	INIT_LIST_HEAD(&root->orphan_list);
+	INIT_LIST_HEAD(&root->dead_list);
+	spin_lock_init(&root->node_lock);
+	spin_lock_init(&root->list_lock);
+	mutex_init(&root->objectid_mutex);
+	mutex_init(&root->log_mutex);
+	extent_io_tree_init(&root->dirty_log_pages,
+			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+
+	btrfs_leaf_ref_tree_init(&root->ref_tree_struct);
+	root->ref_tree = &root->ref_tree_struct;
+
+	memset(&root->root_key, 0, sizeof(root->root_key));
+	memset(&root->root_item, 0, sizeof(root->root_item));
+	memset(&root->defrag_progress, 0, sizeof(root->defrag_progress));
+	memset(&root->root_kobj, 0, sizeof(root->root_kobj));
+	root->defrag_trans_start = fs_info->generation;
+	init_completion(&root->kobj_unregister);
+	root->defrag_running = 0;
+	root->defrag_level = 0;
+	root->root_key.objectid = objectid;
+	root->anon_super.s_root = NULL;
+	root->anon_super.s_dev = 0;
+	INIT_LIST_HEAD(&root->anon_super.s_list);
+	INIT_LIST_HEAD(&root->anon_super.s_instances);
+	init_rwsem(&root->anon_super.s_umount);
+
+	return 0;
+}
+
+static int find_and_setup_root(struct btrfs_root *tree_root,
+			       struct btrfs_fs_info *fs_info,
+			       u64 objectid,
+			       struct btrfs_root *root)
+{
+	int ret;
+	u32 blocksize;
+	u64 generation;
+
+	__setup_root(tree_root->nodesize, tree_root->leafsize,
+		     tree_root->sectorsize, tree_root->stripesize,
+		     root, fs_info, objectid);
+	ret = btrfs_find_last_root(tree_root, objectid,
+				   &root->root_item, &root->root_key);
+	BUG_ON(ret);
+
+	generation = btrfs_root_generation(&root->root_item);
+	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
+	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
+				     blocksize, generation);
+	BUG_ON(!root->node);
+	return 0;
+}
+
+int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info)
+{
+	struct extent_buffer *eb;
+	struct btrfs_root *log_root_tree = fs_info->log_root_tree;
+	u64 start = 0;
+	u64 end = 0;
+	int ret;
+
+	if (!log_root_tree)
+		return 0;
+
+	while (1) {
+		ret = find_first_extent_bit(&log_root_tree->dirty_log_pages,
+				    0, &start, &end, EXTENT_DIRTY);
+		if (ret)
+			break;
+
+		clear_extent_dirty(&log_root_tree->dirty_log_pages,
+				   start, end, GFP_NOFS);
+	}
+	eb = fs_info->log_root_tree->node;
+
+	WARN_ON(btrfs_header_level(eb) != 0);
+	WARN_ON(btrfs_header_nritems(eb) != 0);
+
+	ret = btrfs_free_reserved_extent(fs_info->tree_root,
+				eb->start, eb->len);
+	BUG_ON(ret);
+
+	free_extent_buffer(eb);
+	kfree(fs_info->log_root_tree);
+	fs_info->log_root_tree = NULL;
+	return 0;
+}
+
+int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_root *root;
+	struct btrfs_root *tree_root = fs_info->tree_root;
+
+	root = kzalloc(sizeof(*root), GFP_NOFS);
+	if (!root)
+		return -ENOMEM;
+
+	__setup_root(tree_root->nodesize, tree_root->leafsize,
+		     tree_root->sectorsize, tree_root->stripesize,
+		     root, fs_info, BTRFS_TREE_LOG_OBJECTID);
+
+	root->root_key.objectid = BTRFS_TREE_LOG_OBJECTID;
+	root->root_key.type = BTRFS_ROOT_ITEM_KEY;
+	root->root_key.offset = BTRFS_TREE_LOG_OBJECTID;
+	root->ref_cows = 0;
+
+	root->node = btrfs_alloc_free_block(trans, root, root->leafsize,
+					    0, BTRFS_TREE_LOG_OBJECTID,
+					    trans->transid, 0, 0, 0);
+
+	btrfs_set_header_nritems(root->node, 0);
+	btrfs_set_header_level(root->node, 0);
+	btrfs_set_header_bytenr(root->node, root->node->start);
+	btrfs_set_header_generation(root->node, trans->transid);
+	btrfs_set_header_owner(root->node, BTRFS_TREE_LOG_OBJECTID);
+
+	write_extent_buffer(root->node, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(root->node),
+			    BTRFS_FSID_SIZE);
+	btrfs_mark_buffer_dirty(root->node);
+	btrfs_tree_unlock(root->node);
+	fs_info->log_root_tree = root;
+	return 0;
+}
+
+struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
+					       struct btrfs_key *location)
+{
+	struct btrfs_root *root;
+	struct btrfs_fs_info *fs_info = tree_root->fs_info;
+	struct btrfs_path *path;
+	struct extent_buffer *l;
+	u64 highest_inode;
+	u64 generation;
+	u32 blocksize;
+	int ret = 0;
+
+	root = kzalloc(sizeof(*root), GFP_NOFS);
+	if (!root)
+		return ERR_PTR(-ENOMEM);
+	if (location->offset == (u64)-1) {
+		ret = find_and_setup_root(tree_root, fs_info,
+					  location->objectid, root);
+		if (ret) {
+			kfree(root);
+			return ERR_PTR(ret);
+		}
+		goto insert;
+	}
+
+	__setup_root(tree_root->nodesize, tree_root->leafsize,
+		     tree_root->sectorsize, tree_root->stripesize,
+		     root, fs_info, location->objectid);
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	ret = btrfs_search_slot(NULL, tree_root, location, path, 0, 0);
+	if (ret != 0) {
+		if (ret > 0)
+			ret = -ENOENT;
+		goto out;
+	}
+	l = path->nodes[0];
+	read_extent_buffer(l, &root->root_item,
+	       btrfs_item_ptr_offset(l, path->slots[0]),
+	       sizeof(root->root_item));
+	memcpy(&root->root_key, location, sizeof(*location));
+	ret = 0;
+out:
+	btrfs_release_path(root, path);
+	btrfs_free_path(path);
+	if (ret) {
+		kfree(root);
+		return ERR_PTR(ret);
+	}
+	generation = btrfs_root_generation(&root->root_item);
+	blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
+	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
+				     blocksize, generation);
+	BUG_ON(!root->node);
+insert:
+	if (location->objectid != BTRFS_TREE_LOG_OBJECTID) {
+		root->ref_cows = 1;
+		ret = btrfs_find_highest_inode(root, &highest_inode);
+		if (ret == 0) {
+			root->highest_inode = highest_inode;
+			root->last_inode_alloc = highest_inode;
+		}
+	}
+	return root;
+}
+
+struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
+					u64 root_objectid)
+{
+	struct btrfs_root *root;
+
+	if (root_objectid == BTRFS_ROOT_TREE_OBJECTID)
+		return fs_info->tree_root;
+	if (root_objectid == BTRFS_EXTENT_TREE_OBJECTID)
+		return fs_info->extent_root;
+
+	root = radix_tree_lookup(&fs_info->fs_roots_radix,
+				 (unsigned long)root_objectid);
+	return root;
+}
+
+struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
+					      struct btrfs_key *location)
+{
+	struct btrfs_root *root;
+	int ret;
+
+	if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
+		return fs_info->tree_root;
+	if (location->objectid == BTRFS_EXTENT_TREE_OBJECTID)
+		return fs_info->extent_root;
+	if (location->objectid == BTRFS_CHUNK_TREE_OBJECTID)
+		return fs_info->chunk_root;
+	if (location->objectid == BTRFS_DEV_TREE_OBJECTID)
+		return fs_info->dev_root;
+	if (location->objectid == BTRFS_CSUM_TREE_OBJECTID)
+		return fs_info->csum_root;
+
+	root = radix_tree_lookup(&fs_info->fs_roots_radix,
+				 (unsigned long)location->objectid);
+	if (root)
+		return root;
+
+	root = btrfs_read_fs_root_no_radix(fs_info->tree_root, location);
+	if (IS_ERR(root))
+		return root;
+
+	set_anon_super(&root->anon_super, NULL);
+
+	ret = radix_tree_insert(&fs_info->fs_roots_radix,
+				(unsigned long)root->root_key.objectid,
+				root);
+	if (ret) {
+		free_extent_buffer(root->node);
+		kfree(root);
+		return ERR_PTR(ret);
+	}
+	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
+		ret = btrfs_find_dead_roots(fs_info->tree_root,
+					    root->root_key.objectid, root);
+		BUG_ON(ret);
+		btrfs_orphan_cleanup(root);
+	}
+	return root;
+}
+
+struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
+				      struct btrfs_key *location,
+				      const char *name, int namelen)
+{
+	struct btrfs_root *root;
+	int ret;
+
+	root = btrfs_read_fs_root_no_name(fs_info, location);
+	if (!root)
+		return NULL;
+
+	if (root->in_sysfs)
+		return root;
+
+	ret = btrfs_set_root_name(root, name, namelen);
+	if (ret) {
+		free_extent_buffer(root->node);
+		kfree(root);
+		return ERR_PTR(ret);
+	}
+#if 0
+	ret = btrfs_sysfs_add_root(root);
+	if (ret) {
+		free_extent_buffer(root->node);
+		kfree(root->name);
+		kfree(root);
+		return ERR_PTR(ret);
+	}
+#endif
+	root->in_sysfs = 1;
+	return root;
+}
+
+static int btrfs_congested_fn(void *congested_data, int bdi_bits)
+{
+	struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data;
+	int ret = 0;
+	struct list_head *cur;
+	struct btrfs_device *device;
+	struct backing_dev_info *bdi;
+#if 0
+	if ((bdi_bits & (1 << BDI_write_congested)) &&
+	    btrfs_congested_async(info, 0))
+		return 1;
+#endif
+	list_for_each(cur, &info->fs_devices->devices) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		if (!device->bdev)
+			continue;
+		bdi = blk_get_backing_dev_info(device->bdev);
+		if (bdi && bdi_congested(bdi, bdi_bits)) {
+			ret = 1;
+			break;
+		}
+	}
+	return ret;
+}
+
+/*
+ * this unplugs every device on the box, and it is only used when page
+ * is null
+ */
+static void __unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
+{
+	struct list_head *cur;
+	struct btrfs_device *device;
+	struct btrfs_fs_info *info;
+
+	info = (struct btrfs_fs_info *)bdi->unplug_io_data;
+	list_for_each(cur, &info->fs_devices->devices) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		if (!device->bdev)
+			continue;
+
+		bdi = blk_get_backing_dev_info(device->bdev);
+		if (bdi->unplug_io_fn)
+			bdi->unplug_io_fn(bdi, page);
+	}
+}
+
+static void btrfs_unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
+{
+	struct inode *inode;
+	struct extent_map_tree *em_tree;
+	struct extent_map *em;
+	struct address_space *mapping;
+	u64 offset;
+
+	/* the generic O_DIRECT read code does this */
+	if (1 || !page) {
+		__unplug_io_fn(bdi, page);
+		return;
+	}
+
+	/*
+	 * page->mapping may change at any time.  Get a consistent copy
+	 * and use that for everything below
+	 */
+	smp_mb();
+	mapping = page->mapping;
+	if (!mapping)
+		return;
+
+	inode = mapping->host;
+
+	/*
+	 * don't do the expensive searching for a small number of
+	 * devices
+	 */
+	if (BTRFS_I(inode)->root->fs_info->fs_devices->open_devices <= 2) {
+		__unplug_io_fn(bdi, page);
+		return;
+	}
+
+	offset = page_offset(page);
+
+	em_tree = &BTRFS_I(inode)->extent_tree;
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, offset, PAGE_CACHE_SIZE);
+	spin_unlock(&em_tree->lock);
+	if (!em) {
+		__unplug_io_fn(bdi, page);
+		return;
+	}
+
+	if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
+		free_extent_map(em);
+		__unplug_io_fn(bdi, page);
+		return;
+	}
+	offset = offset - em->start;
+	btrfs_unplug_page(&BTRFS_I(inode)->root->fs_info->mapping_tree,
+			  em->block_start + offset, page);
+	free_extent_map(em);
+}
+
+static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
+{
+	bdi_init(bdi);
+	bdi->ra_pages	= default_backing_dev_info.ra_pages;
+	bdi->state		= 0;
+	bdi->capabilities	= default_backing_dev_info.capabilities;
+	bdi->unplug_io_fn	= btrfs_unplug_io_fn;
+	bdi->unplug_io_data	= info;
+	bdi->congested_fn	= btrfs_congested_fn;
+	bdi->congested_data	= info;
+	return 0;
+}
+
+static int bio_ready_for_csum(struct bio *bio)
+{
+	u64 length = 0;
+	u64 buf_len = 0;
+	u64 start = 0;
+	struct page *page;
+	struct extent_io_tree *io_tree = NULL;
+	struct btrfs_fs_info *info = NULL;
+	struct bio_vec *bvec;
+	int i;
+	int ret;
+
+	bio_for_each_segment(bvec, bio, i) {
+		page = bvec->bv_page;
+		if (page->private == EXTENT_PAGE_PRIVATE) {
+			length += bvec->bv_len;
+			continue;
+		}
+		if (!page->private) {
+			length += bvec->bv_len;
+			continue;
+		}
+		length = bvec->bv_len;
+		buf_len = page->private >> 2;
+		start = page_offset(page) + bvec->bv_offset;
+		io_tree = &BTRFS_I(page->mapping->host)->io_tree;
+		info = BTRFS_I(page->mapping->host)->root->fs_info;
+	}
+	/* are we fully contained in this bio? */
+	if (buf_len <= length)
+		return 1;
+
+	ret = extent_range_uptodate(io_tree, start + length,
+				    start + buf_len - 1);
+	if (ret == 1)
+		return ret;
+	return ret;
+}
+
+/*
+ * called by the kthread helper functions to finally call the bio end_io
+ * functions.  This is where read checksum verification actually happens
+ */
+static void end_workqueue_fn(struct btrfs_work *work)
+{
+	struct bio *bio;
+	struct end_io_wq *end_io_wq;
+	struct btrfs_fs_info *fs_info;
+	int error;
+
+	end_io_wq = container_of(work, struct end_io_wq, work);
+	bio = end_io_wq->bio;
+	fs_info = end_io_wq->info;
+
+	/* metadata bio reads are special because the whole tree block must
+	 * be checksummed at once.  This makes sure the entire block is in
+	 * ram and up to date before trying to verify things.  For
+	 * blocksize <= pagesize, it is basically a noop
+	 */
+	if (!(bio->bi_rw & (1 << BIO_RW)) && end_io_wq->metadata &&
+	    !bio_ready_for_csum(bio)) {
+		btrfs_queue_worker(&fs_info->endio_meta_workers,
+				   &end_io_wq->work);
+		return;
+	}
+	error = end_io_wq->error;
+	bio->bi_private = end_io_wq->private;
+	bio->bi_end_io = end_io_wq->end_io;
+	kfree(end_io_wq);
+	bio_endio(bio, error);
+}
+
+static int cleaner_kthread(void *arg)
+{
+	struct btrfs_root *root = arg;
+
+	do {
+		smp_mb();
+		if (root->fs_info->closing)
+			break;
+
+		vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
+		mutex_lock(&root->fs_info->cleaner_mutex);
+		btrfs_clean_old_snapshots(root);
+		mutex_unlock(&root->fs_info->cleaner_mutex);
+
+		if (freezing(current)) {
+			refrigerator();
+		} else {
+			smp_mb();
+			if (root->fs_info->closing)
+				break;
+			set_current_state(TASK_INTERRUPTIBLE);
+			schedule();
+			__set_current_state(TASK_RUNNING);
+		}
+	} while (!kthread_should_stop());
+	return 0;
+}
+
+static int transaction_kthread(void *arg)
+{
+	struct btrfs_root *root = arg;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_transaction *cur;
+	unsigned long now;
+	unsigned long delay;
+	int ret;
+
+	do {
+		smp_mb();
+		if (root->fs_info->closing)
+			break;
+
+		delay = HZ * 30;
+		vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE);
+		mutex_lock(&root->fs_info->transaction_kthread_mutex);
+
+		if (root->fs_info->total_ref_cache_size > 20 * 1024 * 1024) {
+			printk(KERN_INFO "btrfs: total reference cache "
+			       "size %llu\n",
+			       root->fs_info->total_ref_cache_size);
+		}
+
+		mutex_lock(&root->fs_info->trans_mutex);
+		cur = root->fs_info->running_transaction;
+		if (!cur) {
+			mutex_unlock(&root->fs_info->trans_mutex);
+			goto sleep;
+		}
+
+		now = get_seconds();
+		if (now < cur->start_time || now - cur->start_time < 30) {
+			mutex_unlock(&root->fs_info->trans_mutex);
+			delay = HZ * 5;
+			goto sleep;
+		}
+		mutex_unlock(&root->fs_info->trans_mutex);
+		trans = btrfs_start_transaction(root, 1);
+		ret = btrfs_commit_transaction(trans, root);
+sleep:
+		wake_up_process(root->fs_info->cleaner_kthread);
+		mutex_unlock(&root->fs_info->transaction_kthread_mutex);
+
+		if (freezing(current)) {
+			refrigerator();
+		} else {
+			if (root->fs_info->closing)
+				break;
+			set_current_state(TASK_INTERRUPTIBLE);
+			schedule_timeout(delay);
+			__set_current_state(TASK_RUNNING);
+		}
+	} while (!kthread_should_stop());
+	return 0;
+}
+
+struct btrfs_root *open_ctree(struct super_block *sb,
+			      struct btrfs_fs_devices *fs_devices,
+			      char *options)
+{
+	u32 sectorsize;
+	u32 nodesize;
+	u32 leafsize;
+	u32 blocksize;
+	u32 stripesize;
+	u64 generation;
+	u64 features;
+	struct btrfs_key location;
+	struct buffer_head *bh;
+	struct btrfs_root *extent_root = kzalloc(sizeof(struct btrfs_root),
+						 GFP_NOFS);
+	struct btrfs_root *csum_root = kzalloc(sizeof(struct btrfs_root),
+						 GFP_NOFS);
+	struct btrfs_root *tree_root = kzalloc(sizeof(struct btrfs_root),
+					       GFP_NOFS);
+	struct btrfs_fs_info *fs_info = kzalloc(sizeof(*fs_info),
+						GFP_NOFS);
+	struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root),
+						GFP_NOFS);
+	struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root),
+					      GFP_NOFS);
+	struct btrfs_root *log_tree_root;
+
+	int ret;
+	int err = -EINVAL;
+
+	struct btrfs_super_block *disk_super;
+
+	if (!extent_root || !tree_root || !fs_info ||
+	    !chunk_root || !dev_root || !csum_root) {
+		err = -ENOMEM;
+		goto fail;
+	}
+	INIT_RADIX_TREE(&fs_info->fs_roots_radix, GFP_NOFS);
+	INIT_LIST_HEAD(&fs_info->trans_list);
+	INIT_LIST_HEAD(&fs_info->dead_roots);
+	INIT_LIST_HEAD(&fs_info->hashers);
+	INIT_LIST_HEAD(&fs_info->delalloc_inodes);
+	spin_lock_init(&fs_info->hash_lock);
+	spin_lock_init(&fs_info->delalloc_lock);
+	spin_lock_init(&fs_info->new_trans_lock);
+	spin_lock_init(&fs_info->ref_cache_lock);
+
+	init_completion(&fs_info->kobj_unregister);
+	fs_info->tree_root = tree_root;
+	fs_info->extent_root = extent_root;
+	fs_info->csum_root = csum_root;
+	fs_info->chunk_root = chunk_root;
+	fs_info->dev_root = dev_root;
+	fs_info->fs_devices = fs_devices;
+	INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
+	INIT_LIST_HEAD(&fs_info->space_info);
+	btrfs_mapping_init(&fs_info->mapping_tree);
+	atomic_set(&fs_info->nr_async_submits, 0);
+	atomic_set(&fs_info->async_delalloc_pages, 0);
+	atomic_set(&fs_info->async_submit_draining, 0);
+	atomic_set(&fs_info->nr_async_bios, 0);
+	atomic_set(&fs_info->throttles, 0);
+	atomic_set(&fs_info->throttle_gen, 0);
+	fs_info->sb = sb;
+	fs_info->max_extent = (u64)-1;
+	fs_info->max_inline = 8192 * 1024;
+	setup_bdi(fs_info, &fs_info->bdi);
+	fs_info->btree_inode = new_inode(sb);
+	fs_info->btree_inode->i_ino = 1;
+	fs_info->btree_inode->i_nlink = 1;
+
+	fs_info->thread_pool_size = min_t(unsigned long,
+					  num_online_cpus() + 2, 8);
+
+	INIT_LIST_HEAD(&fs_info->ordered_extents);
+	spin_lock_init(&fs_info->ordered_extent_lock);
+
+	sb->s_blocksize = 4096;
+	sb->s_blocksize_bits = blksize_bits(4096);
+
+	/*
+	 * we set the i_size on the btree inode to the max possible int.
+	 * the real end of the address space is determined by all of
+	 * the devices in the system
+	 */
+	fs_info->btree_inode->i_size = OFFSET_MAX;
+	fs_info->btree_inode->i_mapping->a_ops = &btree_aops;
+	fs_info->btree_inode->i_mapping->backing_dev_info = &fs_info->bdi;
+
+	extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
+			     fs_info->btree_inode->i_mapping,
+			     GFP_NOFS);
+	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree,
+			     GFP_NOFS);
+
+	BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops;
+
+	spin_lock_init(&fs_info->block_group_cache_lock);
+	fs_info->block_group_cache_tree.rb_node = NULL;
+
+	extent_io_tree_init(&fs_info->pinned_extents,
+			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+	extent_io_tree_init(&fs_info->pending_del,
+			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+	extent_io_tree_init(&fs_info->extent_ins,
+			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+	fs_info->do_barriers = 1;
+
+	INIT_LIST_HEAD(&fs_info->dead_reloc_roots);
+	btrfs_leaf_ref_tree_init(&fs_info->reloc_ref_tree);
+	btrfs_leaf_ref_tree_init(&fs_info->shared_ref_tree);
+
+	BTRFS_I(fs_info->btree_inode)->root = tree_root;
+	memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
+	       sizeof(struct btrfs_key));
+	insert_inode_hash(fs_info->btree_inode);
+
+	mutex_init(&fs_info->trans_mutex);
+	mutex_init(&fs_info->tree_log_mutex);
+	mutex_init(&fs_info->drop_mutex);
+	mutex_init(&fs_info->extent_ins_mutex);
+	mutex_init(&fs_info->pinned_mutex);
+	mutex_init(&fs_info->chunk_mutex);
+	mutex_init(&fs_info->transaction_kthread_mutex);
+	mutex_init(&fs_info->cleaner_mutex);
+	mutex_init(&fs_info->volume_mutex);
+	mutex_init(&fs_info->tree_reloc_mutex);
+	init_waitqueue_head(&fs_info->transaction_throttle);
+	init_waitqueue_head(&fs_info->transaction_wait);
+	init_waitqueue_head(&fs_info->async_submit_wait);
+	init_waitqueue_head(&fs_info->tree_log_wait);
+	atomic_set(&fs_info->tree_log_commit, 0);
+	atomic_set(&fs_info->tree_log_writers, 0);
+	fs_info->tree_log_transid = 0;
+
+	__setup_root(4096, 4096, 4096, 4096, tree_root,
+		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
+
+
+	bh = btrfs_read_dev_super(fs_devices->latest_bdev);
+	if (!bh)
+		goto fail_iput;
+
+	memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy));
+	memcpy(&fs_info->super_for_commit, &fs_info->super_copy,
+	       sizeof(fs_info->super_for_commit));
+	brelse(bh);
+
+	memcpy(fs_info->fsid, fs_info->super_copy.fsid, BTRFS_FSID_SIZE);
+
+	disk_super = &fs_info->super_copy;
+	if (!btrfs_super_root(disk_super))
+		goto fail_iput;
+
+	ret = btrfs_parse_options(tree_root, options);
+	if (ret) {
+		err = ret;
+		goto fail_iput;
+	}
+
+	features = btrfs_super_incompat_flags(disk_super) &
+		~BTRFS_FEATURE_INCOMPAT_SUPP;
+	if (features) {
+		printk(KERN_ERR "BTRFS: couldn't mount because of "
+		       "unsupported optional features (%Lx).\n",
+		       features);
+		err = -EINVAL;
+		goto fail_iput;
+	}
+
+	features = btrfs_super_compat_ro_flags(disk_super) &
+		~BTRFS_FEATURE_COMPAT_RO_SUPP;
+	if (!(sb->s_flags & MS_RDONLY) && features) {
+		printk(KERN_ERR "BTRFS: couldn't mount RDWR because of "
+		       "unsupported option features (%Lx).\n",
+		       features);
+		err = -EINVAL;
+		goto fail_iput;
+	}
+
+	/*
+	 * we need to start all the end_io workers up front because the
+	 * queue work function gets called at interrupt time, and so it
+	 * cannot dynamically grow.
+	 */
+	btrfs_init_workers(&fs_info->workers, "worker",
+			   fs_info->thread_pool_size);
+
+	btrfs_init_workers(&fs_info->delalloc_workers, "delalloc",
+			   fs_info->thread_pool_size);
+
+	btrfs_init_workers(&fs_info->submit_workers, "submit",
+			   min_t(u64, fs_devices->num_devices,
+			   fs_info->thread_pool_size));
+
+	/* a higher idle thresh on the submit workers makes it much more
+	 * likely that bios will be send down in a sane order to the
+	 * devices
+	 */
+	fs_info->submit_workers.idle_thresh = 64;
+
+	fs_info->workers.idle_thresh = 16;
+	fs_info->workers.ordered = 1;
+
+	fs_info->delalloc_workers.idle_thresh = 2;
+	fs_info->delalloc_workers.ordered = 1;
+
+	btrfs_init_workers(&fs_info->fixup_workers, "fixup", 1);
+	btrfs_init_workers(&fs_info->endio_workers, "endio",
+			   fs_info->thread_pool_size);
+	btrfs_init_workers(&fs_info->endio_meta_workers, "endio-meta",
+			   fs_info->thread_pool_size);
+	btrfs_init_workers(&fs_info->endio_meta_write_workers,
+			   "endio-meta-write", fs_info->thread_pool_size);
+	btrfs_init_workers(&fs_info->endio_write_workers, "endio-write",
+			   fs_info->thread_pool_size);
+
+	/*
+	 * endios are largely parallel and should have a very
+	 * low idle thresh
+	 */
+	fs_info->endio_workers.idle_thresh = 4;
+	fs_info->endio_write_workers.idle_thresh = 64;
+	fs_info->endio_meta_write_workers.idle_thresh = 64;
+
+	btrfs_start_workers(&fs_info->workers, 1);
+	btrfs_start_workers(&fs_info->submit_workers, 1);
+	btrfs_start_workers(&fs_info->delalloc_workers, 1);
+	btrfs_start_workers(&fs_info->fixup_workers, 1);
+	btrfs_start_workers(&fs_info->endio_workers, fs_info->thread_pool_size);
+	btrfs_start_workers(&fs_info->endio_meta_workers,
+			    fs_info->thread_pool_size);
+	btrfs_start_workers(&fs_info->endio_meta_write_workers,
+			    fs_info->thread_pool_size);
+	btrfs_start_workers(&fs_info->endio_write_workers,
+			    fs_info->thread_pool_size);
+
+	fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super);
+	fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages,
+				    4 * 1024 * 1024 / PAGE_CACHE_SIZE);
+
+	nodesize = btrfs_super_nodesize(disk_super);
+	leafsize = btrfs_super_leafsize(disk_super);
+	sectorsize = btrfs_super_sectorsize(disk_super);
+	stripesize = btrfs_super_stripesize(disk_super);
+	tree_root->nodesize = nodesize;
+	tree_root->leafsize = leafsize;
+	tree_root->sectorsize = sectorsize;
+	tree_root->stripesize = stripesize;
+
+	sb->s_blocksize = sectorsize;
+	sb->s_blocksize_bits = blksize_bits(sectorsize);
+
+	if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
+		    sizeof(disk_super->magic))) {
+		printk(KERN_INFO "btrfs: valid FS not found on %s\n", sb->s_id);
+		goto fail_sb_buffer;
+	}
+
+	mutex_lock(&fs_info->chunk_mutex);
+	ret = btrfs_read_sys_array(tree_root);
+	mutex_unlock(&fs_info->chunk_mutex);
+	if (ret) {
+		printk(KERN_WARNING "btrfs: failed to read the system "
+		       "array on %s\n", sb->s_id);
+		goto fail_sys_array;
+	}
+
+	blocksize = btrfs_level_size(tree_root,
+				     btrfs_super_chunk_root_level(disk_super));
+	generation = btrfs_super_chunk_root_generation(disk_super);
+
+	__setup_root(nodesize, leafsize, sectorsize, stripesize,
+		     chunk_root, fs_info, BTRFS_CHUNK_TREE_OBJECTID);
+
+	chunk_root->node = read_tree_block(chunk_root,
+					   btrfs_super_chunk_root(disk_super),
+					   blocksize, generation);
+	BUG_ON(!chunk_root->node);
+
+	read_extent_buffer(chunk_root->node, fs_info->chunk_tree_uuid,
+	   (unsigned long)btrfs_header_chunk_tree_uuid(chunk_root->node),
+	   BTRFS_UUID_SIZE);
+
+	mutex_lock(&fs_info->chunk_mutex);
+	ret = btrfs_read_chunk_tree(chunk_root);
+	mutex_unlock(&fs_info->chunk_mutex);
+	if (ret) {
+		printk(KERN_WARNING "btrfs: failed to read chunk tree on %s\n",
+		       sb->s_id);
+		goto fail_chunk_root;
+	}
+
+	btrfs_close_extra_devices(fs_devices);
+
+	blocksize = btrfs_level_size(tree_root,
+				     btrfs_super_root_level(disk_super));
+	generation = btrfs_super_generation(disk_super);
+
+	tree_root->node = read_tree_block(tree_root,
+					  btrfs_super_root(disk_super),
+					  blocksize, generation);
+	if (!tree_root->node)
+		goto fail_chunk_root;
+
+
+	ret = find_and_setup_root(tree_root, fs_info,
+				  BTRFS_EXTENT_TREE_OBJECTID, extent_root);
+	if (ret)
+		goto fail_tree_root;
+	extent_root->track_dirty = 1;
+
+	ret = find_and_setup_root(tree_root, fs_info,
+				  BTRFS_DEV_TREE_OBJECTID, dev_root);
+	dev_root->track_dirty = 1;
+
+	if (ret)
+		goto fail_extent_root;
+
+	ret = find_and_setup_root(tree_root, fs_info,
+				  BTRFS_CSUM_TREE_OBJECTID, csum_root);
+	if (ret)
+		goto fail_extent_root;
+
+	csum_root->track_dirty = 1;
+
+	btrfs_read_block_groups(extent_root);
+
+	fs_info->generation = generation;
+	fs_info->last_trans_committed = generation;
+	fs_info->data_alloc_profile = (u64)-1;
+	fs_info->metadata_alloc_profile = (u64)-1;
+	fs_info->system_alloc_profile = fs_info->metadata_alloc_profile;
+	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root,
+					       "btrfs-cleaner");
+	if (!fs_info->cleaner_kthread)
+		goto fail_csum_root;
+
+	fs_info->transaction_kthread = kthread_run(transaction_kthread,
+						   tree_root,
+						   "btrfs-transaction");
+	if (!fs_info->transaction_kthread)
+		goto fail_cleaner;
+
+	if (btrfs_super_log_root(disk_super) != 0) {
+		u64 bytenr = btrfs_super_log_root(disk_super);
+
+		if (fs_devices->rw_devices == 0) {
+			printk(KERN_WARNING "Btrfs log replay required "
+			       "on RO media\n");
+			err = -EIO;
+			goto fail_trans_kthread;
+		}
+		blocksize =
+		     btrfs_level_size(tree_root,
+				      btrfs_super_log_root_level(disk_super));
+
+		log_tree_root = kzalloc(sizeof(struct btrfs_root),
+						      GFP_NOFS);
+
+		__setup_root(nodesize, leafsize, sectorsize, stripesize,
+			     log_tree_root, fs_info, BTRFS_TREE_LOG_OBJECTID);
+
+		log_tree_root->node = read_tree_block(tree_root, bytenr,
+						      blocksize,
+						      generation + 1);
+		ret = btrfs_recover_log_trees(log_tree_root);
+		BUG_ON(ret);
+
+		if (sb->s_flags & MS_RDONLY) {
+			ret =  btrfs_commit_super(tree_root);
+			BUG_ON(ret);
+		}
+	}
+
+	if (!(sb->s_flags & MS_RDONLY)) {
+		ret = btrfs_cleanup_reloc_trees(tree_root);
+		BUG_ON(ret);
+	}
+
+	location.objectid = BTRFS_FS_TREE_OBJECTID;
+	location.type = BTRFS_ROOT_ITEM_KEY;
+	location.offset = (u64)-1;
+
+	fs_info->fs_root = btrfs_read_fs_root_no_name(fs_info, &location);
+	if (!fs_info->fs_root)
+		goto fail_trans_kthread;
+	return tree_root;
+
+fail_trans_kthread:
+	kthread_stop(fs_info->transaction_kthread);
+fail_cleaner:
+	kthread_stop(fs_info->cleaner_kthread);
+
+	/*
+	 * make sure we're done with the btree inode before we stop our
+	 * kthreads
+	 */
+	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
+	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
+
+fail_csum_root:
+	free_extent_buffer(csum_root->node);
+fail_extent_root:
+	free_extent_buffer(extent_root->node);
+fail_tree_root:
+	free_extent_buffer(tree_root->node);
+fail_chunk_root:
+	free_extent_buffer(chunk_root->node);
+fail_sys_array:
+	free_extent_buffer(dev_root->node);
+fail_sb_buffer:
+	btrfs_stop_workers(&fs_info->fixup_workers);
+	btrfs_stop_workers(&fs_info->delalloc_workers);
+	btrfs_stop_workers(&fs_info->workers);
+	btrfs_stop_workers(&fs_info->endio_workers);
+	btrfs_stop_workers(&fs_info->endio_meta_workers);
+	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
+	btrfs_stop_workers(&fs_info->endio_write_workers);
+	btrfs_stop_workers(&fs_info->submit_workers);
+fail_iput:
+	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
+	iput(fs_info->btree_inode);
+fail:
+	btrfs_close_devices(fs_info->fs_devices);
+	btrfs_mapping_tree_free(&fs_info->mapping_tree);
+
+	kfree(extent_root);
+	kfree(tree_root);
+	bdi_destroy(&fs_info->bdi);
+	kfree(fs_info);
+	kfree(chunk_root);
+	kfree(dev_root);
+	kfree(csum_root);
+	return ERR_PTR(err);
+}
+
+static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate)
+{
+	char b[BDEVNAME_SIZE];
+
+	if (uptodate) {
+		set_buffer_uptodate(bh);
+	} else {
+		if (!buffer_eopnotsupp(bh) && printk_ratelimit()) {
+			printk(KERN_WARNING "lost page write due to "
+					"I/O error on %s\n",
+				       bdevname(bh->b_bdev, b));
+		}
+		/* note, we dont' set_buffer_write_io_error because we have
+		 * our own ways of dealing with the IO errors
+		 */
+		clear_buffer_uptodate(bh);
+	}
+	unlock_buffer(bh);
+	put_bh(bh);
+}
+
+struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
+{
+	struct buffer_head *bh;
+	struct buffer_head *latest = NULL;
+	struct btrfs_super_block *super;
+	int i;
+	u64 transid = 0;
+	u64 bytenr;
+
+	/* we would like to check all the supers, but that would make
+	 * a btrfs mount succeed after a mkfs from a different FS.
+	 * So, we need to add a special mount option to scan for
+	 * later supers, using BTRFS_SUPER_MIRROR_MAX instead
+	 */
+	for (i = 0; i < 1; i++) {
+		bytenr = btrfs_sb_offset(i);
+		if (bytenr + 4096 >= i_size_read(bdev->bd_inode))
+			break;
+		bh = __bread(bdev, bytenr / 4096, 4096);
+		if (!bh)
+			continue;
+
+		super = (struct btrfs_super_block *)bh->b_data;
+		if (btrfs_super_bytenr(super) != bytenr ||
+		    strncmp((char *)(&super->magic), BTRFS_MAGIC,
+			    sizeof(super->magic))) {
+			brelse(bh);
+			continue;
+		}
+
+		if (!latest || btrfs_super_generation(super) > transid) {
+			brelse(latest);
+			latest = bh;
+			transid = btrfs_super_generation(super);
+		} else {
+			brelse(bh);
+		}
+	}
+	return latest;
+}
+
+static int write_dev_supers(struct btrfs_device *device,
+			    struct btrfs_super_block *sb,
+			    int do_barriers, int wait, int max_mirrors)
+{
+	struct buffer_head *bh;
+	int i;
+	int ret;
+	int errors = 0;
+	u32 crc;
+	u64 bytenr;
+	int last_barrier = 0;
+
+	if (max_mirrors == 0)
+		max_mirrors = BTRFS_SUPER_MIRROR_MAX;
+
+	/* make sure only the last submit_bh does a barrier */
+	if (do_barriers) {
+		for (i = 0; i < max_mirrors; i++) {
+			bytenr = btrfs_sb_offset(i);
+			if (bytenr + BTRFS_SUPER_INFO_SIZE >=
+			    device->total_bytes)
+				break;
+			last_barrier = i;
+		}
+	}
+
+	for (i = 0; i < max_mirrors; i++) {
+		bytenr = btrfs_sb_offset(i);
+		if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes)
+			break;
+
+		if (wait) {
+			bh = __find_get_block(device->bdev, bytenr / 4096,
+					      BTRFS_SUPER_INFO_SIZE);
+			BUG_ON(!bh);
+			brelse(bh);
+			wait_on_buffer(bh);
+			if (buffer_uptodate(bh)) {
+				brelse(bh);
+				continue;
+			}
+		} else {
+			btrfs_set_super_bytenr(sb, bytenr);
+
+			crc = ~(u32)0;
+			crc = btrfs_csum_data(NULL, (char *)sb +
+					      BTRFS_CSUM_SIZE, crc,
+					      BTRFS_SUPER_INFO_SIZE -
+					      BTRFS_CSUM_SIZE);
+			btrfs_csum_final(crc, sb->csum);
+
+			bh = __getblk(device->bdev, bytenr / 4096,
+				      BTRFS_SUPER_INFO_SIZE);
+			memcpy(bh->b_data, sb, BTRFS_SUPER_INFO_SIZE);
+
+			set_buffer_uptodate(bh);
+			get_bh(bh);
+			lock_buffer(bh);
+			bh->b_end_io = btrfs_end_buffer_write_sync;
+		}
+
+		if (i == last_barrier && do_barriers && device->barriers) {
+			ret = submit_bh(WRITE_BARRIER, bh);
+			if (ret == -EOPNOTSUPP) {
+				printk("btrfs: disabling barriers on dev %s\n",
+				       device->name);
+				set_buffer_uptodate(bh);
+				device->barriers = 0;
+				get_bh(bh);
+				lock_buffer(bh);
+				ret = submit_bh(WRITE, bh);
+			}
+		} else {
+			ret = submit_bh(WRITE, bh);
+		}
+
+		if (!ret && wait) {
+			wait_on_buffer(bh);
+			if (!buffer_uptodate(bh))
+				errors++;
+		} else if (ret) {
+			errors++;
+		}
+		if (wait)
+			brelse(bh);
+	}
+	return errors < i ? 0 : -1;
+}
+
+int write_all_supers(struct btrfs_root *root, int max_mirrors)
+{
+	struct list_head *cur;
+	struct list_head *head = &root->fs_info->fs_devices->devices;
+	struct btrfs_device *dev;
+	struct btrfs_super_block *sb;
+	struct btrfs_dev_item *dev_item;
+	int ret;
+	int do_barriers;
+	int max_errors;
+	int total_errors = 0;
+	u64 flags;
+
+	max_errors = btrfs_super_num_devices(&root->fs_info->super_copy) - 1;
+	do_barriers = !btrfs_test_opt(root, NOBARRIER);
+
+	sb = &root->fs_info->super_for_commit;
+	dev_item = &sb->dev_item;
+	list_for_each(cur, head) {
+		dev = list_entry(cur, struct btrfs_device, dev_list);
+		if (!dev->bdev) {
+			total_errors++;
+			continue;
+		}
+		if (!dev->in_fs_metadata || !dev->writeable)
+			continue;
+
+		btrfs_set_stack_device_generation(dev_item, 0);
+		btrfs_set_stack_device_type(dev_item, dev->type);
+		btrfs_set_stack_device_id(dev_item, dev->devid);
+		btrfs_set_stack_device_total_bytes(dev_item, dev->total_bytes);
+		btrfs_set_stack_device_bytes_used(dev_item, dev->bytes_used);
+		btrfs_set_stack_device_io_align(dev_item, dev->io_align);
+		btrfs_set_stack_device_io_width(dev_item, dev->io_width);
+		btrfs_set_stack_device_sector_size(dev_item, dev->sector_size);
+		memcpy(dev_item->uuid, dev->uuid, BTRFS_UUID_SIZE);
+		memcpy(dev_item->fsid, dev->fs_devices->fsid, BTRFS_UUID_SIZE);
+
+		flags = btrfs_super_flags(sb);
+		btrfs_set_super_flags(sb, flags | BTRFS_HEADER_FLAG_WRITTEN);
+
+		ret = write_dev_supers(dev, sb, do_barriers, 0, max_mirrors);
+		if (ret)
+			total_errors++;
+	}
+	if (total_errors > max_errors) {
+		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
+		       total_errors);
+		BUG();
+	}
+
+	total_errors = 0;
+	list_for_each(cur, head) {
+		dev = list_entry(cur, struct btrfs_device, dev_list);
+		if (!dev->bdev)
+			continue;
+		if (!dev->in_fs_metadata || !dev->writeable)
+			continue;
+
+		ret = write_dev_supers(dev, sb, do_barriers, 1, max_mirrors);
+		if (ret)
+			total_errors++;
+	}
+	if (total_errors > max_errors) {
+		printk(KERN_ERR "btrfs: %d errors while writing supers\n",
+		       total_errors);
+		BUG();
+	}
+	return 0;
+}
+
+int write_ctree_super(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root, int max_mirrors)
+{
+	int ret;
+
+	ret = write_all_supers(root, max_mirrors);
+	return ret;
+}
+
+int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
+{
+	radix_tree_delete(&fs_info->fs_roots_radix,
+			  (unsigned long)root->root_key.objectid);
+	if (root->anon_super.s_dev) {
+		down_write(&root->anon_super.s_umount);
+		kill_anon_super(&root->anon_super);
+	}
+	if (root->node)
+		free_extent_buffer(root->node);
+	if (root->commit_root)
+		free_extent_buffer(root->commit_root);
+	kfree(root->name);
+	kfree(root);
+	return 0;
+}
+
+static int del_fs_roots(struct btrfs_fs_info *fs_info)
+{
+	int ret;
+	struct btrfs_root *gang[8];
+	int i;
+
+	while (1) {
+		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
+					     (void **)gang, 0,
+					     ARRAY_SIZE(gang));
+		if (!ret)
+			break;
+		for (i = 0; i < ret; i++)
+			btrfs_free_fs_root(fs_info, gang[i]);
+	}
+	return 0;
+}
+
+int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
+{
+	u64 root_objectid = 0;
+	struct btrfs_root *gang[8];
+	int i;
+	int ret;
+
+	while (1) {
+		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
+					     (void **)gang, root_objectid,
+					     ARRAY_SIZE(gang));
+		if (!ret)
+			break;
+		for (i = 0; i < ret; i++) {
+			root_objectid = gang[i]->root_key.objectid;
+			ret = btrfs_find_dead_roots(fs_info->tree_root,
+						    root_objectid, gang[i]);
+			BUG_ON(ret);
+			btrfs_orphan_cleanup(gang[i]);
+		}
+		root_objectid++;
+	}
+	return 0;
+}
+
+int btrfs_commit_super(struct btrfs_root *root)
+{
+	struct btrfs_trans_handle *trans;
+	int ret;
+
+	mutex_lock(&root->fs_info->cleaner_mutex);
+	btrfs_clean_old_snapshots(root);
+	mutex_unlock(&root->fs_info->cleaner_mutex);
+	trans = btrfs_start_transaction(root, 1);
+	ret = btrfs_commit_transaction(trans, root);
+	BUG_ON(ret);
+	/* run commit again to drop the original snapshot */
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_commit_transaction(trans, root);
+	ret = btrfs_write_and_wait_transaction(NULL, root);
+	BUG_ON(ret);
+
+	ret = write_ctree_super(NULL, root, 0);
+	return ret;
+}
+
+int close_ctree(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	int ret;
+
+	fs_info->closing = 1;
+	smp_mb();
+
+	kthread_stop(root->fs_info->transaction_kthread);
+	kthread_stop(root->fs_info->cleaner_kthread);
+
+	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
+		ret =  btrfs_commit_super(root);
+		if (ret)
+			printk(KERN_ERR "btrfs: commit super ret %d\n", ret);
+	}
+
+	if (fs_info->delalloc_bytes) {
+		printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n",
+		       fs_info->delalloc_bytes);
+	}
+	if (fs_info->total_ref_cache_size) {
+		printk(KERN_INFO "btrfs: at umount reference cache size %llu\n",
+		       (unsigned long long)fs_info->total_ref_cache_size);
+	}
+
+	if (fs_info->extent_root->node)
+		free_extent_buffer(fs_info->extent_root->node);
+
+	if (fs_info->tree_root->node)
+		free_extent_buffer(fs_info->tree_root->node);
+
+	if (root->fs_info->chunk_root->node)
+		free_extent_buffer(root->fs_info->chunk_root->node);
+
+	if (root->fs_info->dev_root->node)
+		free_extent_buffer(root->fs_info->dev_root->node);
+
+	if (root->fs_info->csum_root->node)
+		free_extent_buffer(root->fs_info->csum_root->node);
+
+	btrfs_free_block_groups(root->fs_info);
+
+	del_fs_roots(fs_info);
+
+	iput(fs_info->btree_inode);
+
+	btrfs_stop_workers(&fs_info->fixup_workers);
+	btrfs_stop_workers(&fs_info->delalloc_workers);
+	btrfs_stop_workers(&fs_info->workers);
+	btrfs_stop_workers(&fs_info->endio_workers);
+	btrfs_stop_workers(&fs_info->endio_meta_workers);
+	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
+	btrfs_stop_workers(&fs_info->endio_write_workers);
+	btrfs_stop_workers(&fs_info->submit_workers);
+
+#if 0
+	while (!list_empty(&fs_info->hashers)) {
+		struct btrfs_hasher *hasher;
+		hasher = list_entry(fs_info->hashers.next, struct btrfs_hasher,
+				    hashers);
+		list_del(&hasher->hashers);
+		crypto_free_hash(&fs_info->hash_tfm);
+		kfree(hasher);
+	}
+#endif
+	btrfs_close_devices(fs_info->fs_devices);
+	btrfs_mapping_tree_free(&fs_info->mapping_tree);
+
+	bdi_destroy(&fs_info->bdi);
+
+	kfree(fs_info->extent_root);
+	kfree(fs_info->tree_root);
+	kfree(fs_info->chunk_root);
+	kfree(fs_info->dev_root);
+	kfree(fs_info->csum_root);
+	return 0;
+}
+
+int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid)
+{
+	int ret;
+	struct inode *btree_inode = buf->first_page->mapping->host;
+
+	ret = extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree, buf);
+	if (!ret)
+		return ret;
+
+	ret = verify_parent_transid(&BTRFS_I(btree_inode)->io_tree, buf,
+				    parent_transid);
+	return !ret;
+}
+
+int btrfs_set_buffer_uptodate(struct extent_buffer *buf)
+{
+	struct inode *btree_inode = buf->first_page->mapping->host;
+	return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree,
+					  buf);
+}
+
+void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
+{
+	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+	u64 transid = btrfs_header_generation(buf);
+	struct inode *btree_inode = root->fs_info->btree_inode;
+
+	WARN_ON(!btrfs_tree_locked(buf));
+	if (transid != root->fs_info->generation) {
+		printk(KERN_CRIT "btrfs transid mismatch buffer %llu, "
+		       "found %llu running %llu\n",
+			(unsigned long long)buf->start,
+			(unsigned long long)transid,
+			(unsigned long long)root->fs_info->generation);
+		WARN_ON(1);
+	}
+	set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, buf);
+}
+
+void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
+{
+	/*
+	 * looks as though older kernels can get into trouble with
+	 * this code, they end up stuck in balance_dirty_pages forever
+	 */
+	struct extent_io_tree *tree;
+	u64 num_dirty;
+	u64 start = 0;
+	unsigned long thresh = 32 * 1024 * 1024;
+	tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
+
+	if (current_is_pdflush() || current->flags & PF_MEMALLOC)
+		return;
+
+	num_dirty = count_range_bits(tree, &start, (u64)-1,
+				     thresh, EXTENT_DIRTY);
+	if (num_dirty > thresh) {
+		balance_dirty_pages_ratelimited_nr(
+				   root->fs_info->btree_inode->i_mapping, 1);
+	}
+	return;
+}
+
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
+{
+	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+	int ret;
+	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
+	if (ret == 0)
+		buf->flags |= EXTENT_UPTODATE;
+	return ret;
+}
+
+int btree_lock_page_hook(struct page *page)
+{
+	struct inode *inode = page->mapping->host;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct extent_buffer *eb;
+	unsigned long len;
+	u64 bytenr = page_offset(page);
+
+	if (page->private == EXTENT_PAGE_PRIVATE)
+		goto out;
+
+	len = page->private >> 2;
+	eb = find_extent_buffer(io_tree, bytenr, len, GFP_NOFS);
+	if (!eb)
+		goto out;
+
+	btrfs_tree_lock(eb);
+	spin_lock(&root->fs_info->hash_lock);
+	btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
+	spin_unlock(&root->fs_info->hash_lock);
+	btrfs_tree_unlock(eb);
+	free_extent_buffer(eb);
+out:
+	lock_page(page);
+	return 0;
+}
+
+static struct extent_io_ops btree_extent_io_ops = {
+	.write_cache_pages_lock_hook = btree_lock_page_hook,
+	.readpage_end_io_hook = btree_readpage_end_io_hook,
+	.submit_bio_hook = btree_submit_bio_hook,
+	/* note we're sharing with inode.c for the merge bio hook */
+	.merge_bio_hook = btrfs_merge_bio_hook,
+};
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
new file mode 100644
index 00000000000..c0ff404c31b
--- /dev/null
+++ b/fs/btrfs/disk-io.h
@@ -0,0 +1,102 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __DISKIO__
+#define __DISKIO__
+
+#define BTRFS_SUPER_INFO_OFFSET (64 * 1024)
+#define BTRFS_SUPER_INFO_SIZE 4096
+
+#define BTRFS_SUPER_MIRROR_MAX	 3
+#define BTRFS_SUPER_MIRROR_SHIFT 12
+
+static inline u64 btrfs_sb_offset(int mirror)
+{
+	u64 start = 16 * 1024;
+	if (mirror)
+		return start << (BTRFS_SUPER_MIRROR_SHIFT * mirror);
+	return BTRFS_SUPER_INFO_OFFSET;
+}
+
+struct btrfs_device;
+struct btrfs_fs_devices;
+
+struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
+				      u32 blocksize, u64 parent_transid);
+int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize,
+			 u64 parent_transid);
+struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
+						   u64 bytenr, u32 blocksize);
+int clean_tree_block(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root, struct extent_buffer *buf);
+struct btrfs_root *open_ctree(struct super_block *sb,
+			      struct btrfs_fs_devices *fs_devices,
+			      char *options);
+int close_ctree(struct btrfs_root *root);
+int write_ctree_super(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root, int max_mirrors);
+struct buffer_head *btrfs_read_dev_super(struct block_device *bdev);
+int btrfs_commit_super(struct btrfs_root *root);
+struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
+					    u64 bytenr, u32 blocksize);
+struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
+					u64 root_objectid);
+struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
+				      struct btrfs_key *location,
+				      const char *name, int namelen);
+struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
+					       struct btrfs_key *location);
+struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
+					      struct btrfs_key *location);
+int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info);
+int btrfs_insert_dev_radix(struct btrfs_root *root,
+			   struct block_device *bdev,
+			   u64 device_id,
+			   u64 block_start,
+			   u64 num_blocks);
+void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr);
+int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root);
+void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
+int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid);
+int btrfs_set_buffer_uptodate(struct extent_buffer *buf);
+int wait_on_tree_block_writeback(struct btrfs_root *root,
+				 struct extent_buffer *buf);
+int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid);
+u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len);
+void btrfs_csum_final(u32 crc, char *result);
+int btrfs_open_device(struct btrfs_device *dev);
+int btrfs_verify_block_csum(struct btrfs_root *root,
+			    struct extent_buffer *buf);
+int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
+			int metadata);
+int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
+			int rw, struct bio *bio, int mirror_num,
+			unsigned long bio_flags,
+			extent_submit_bio_hook_t *submit_bio_start,
+			extent_submit_bio_hook_t *submit_bio_done);
+
+int btrfs_congested_async(struct btrfs_fs_info *info, int iodone);
+unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info);
+int btrfs_write_tree_block(struct extent_buffer *buf);
+int btrfs_wait_tree_block_writeback(struct extent_buffer *buf);
+int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info);
+int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
+			     struct btrfs_fs_info *fs_info);
+int btree_lock_page_hook(struct page *page);
+#endif
diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c
new file mode 100644
index 00000000000..85315d2c90d
--- /dev/null
+++ b/fs/btrfs/export.c
@@ -0,0 +1,203 @@
+#include <linux/fs.h>
+#include <linux/types.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "btrfs_inode.h"
+#include "print-tree.h"
+#include "export.h"
+#include "compat.h"
+
+#define BTRFS_FID_SIZE_NON_CONNECTABLE (offsetof(struct btrfs_fid, \
+						 parent_objectid) / 4)
+#define BTRFS_FID_SIZE_CONNECTABLE (offsetof(struct btrfs_fid, \
+					     parent_root_objectid) / 4)
+#define BTRFS_FID_SIZE_CONNECTABLE_ROOT (sizeof(struct btrfs_fid) / 4)
+
+static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len,
+			   int connectable)
+{
+	struct btrfs_fid *fid = (struct btrfs_fid *)fh;
+	struct inode *inode = dentry->d_inode;
+	int len = *max_len;
+	int type;
+
+	if ((len < BTRFS_FID_SIZE_NON_CONNECTABLE) ||
+	    (connectable && len < BTRFS_FID_SIZE_CONNECTABLE))
+		return 255;
+
+	len  = BTRFS_FID_SIZE_NON_CONNECTABLE;
+	type = FILEID_BTRFS_WITHOUT_PARENT;
+
+	fid->objectid = BTRFS_I(inode)->location.objectid;
+	fid->root_objectid = BTRFS_I(inode)->root->objectid;
+	fid->gen = inode->i_generation;
+
+	if (connectable && !S_ISDIR(inode->i_mode)) {
+		struct inode *parent;
+		u64 parent_root_id;
+
+		spin_lock(&dentry->d_lock);
+
+		parent = dentry->d_parent->d_inode;
+		fid->parent_objectid = BTRFS_I(parent)->location.objectid;
+		fid->parent_gen = parent->i_generation;
+		parent_root_id = BTRFS_I(parent)->root->objectid;
+
+		spin_unlock(&dentry->d_lock);
+
+		if (parent_root_id != fid->root_objectid) {
+			fid->parent_root_objectid = parent_root_id;
+			len = BTRFS_FID_SIZE_CONNECTABLE_ROOT;
+			type = FILEID_BTRFS_WITH_PARENT_ROOT;
+		} else {
+			len = BTRFS_FID_SIZE_CONNECTABLE;
+			type = FILEID_BTRFS_WITH_PARENT;
+		}
+	}
+
+	*max_len = len;
+	return type;
+}
+
+static struct dentry *btrfs_get_dentry(struct super_block *sb, u64 objectid,
+				       u64 root_objectid, u32 generation)
+{
+	struct btrfs_root *root;
+	struct inode *inode;
+	struct btrfs_key key;
+
+	key.objectid = root_objectid;
+	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+	key.offset = (u64)-1;
+
+	root = btrfs_read_fs_root_no_name(btrfs_sb(sb)->fs_info, &key);
+	if (IS_ERR(root))
+		return ERR_CAST(root);
+
+	key.objectid = objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+	key.offset = 0;
+
+	inode = btrfs_iget(sb, &key, root, NULL);
+	if (IS_ERR(inode))
+		return (void *)inode;
+
+	if (generation != inode->i_generation) {
+		iput(inode);
+		return ERR_PTR(-ESTALE);
+	}
+
+	return d_obtain_alias(inode);
+}
+
+static struct dentry *btrfs_fh_to_parent(struct super_block *sb, struct fid *fh,
+					 int fh_len, int fh_type)
+{
+	struct btrfs_fid *fid = (struct btrfs_fid *) fh;
+	u64 objectid, root_objectid;
+	u32 generation;
+
+	if (fh_type == FILEID_BTRFS_WITH_PARENT) {
+		if (fh_len !=  BTRFS_FID_SIZE_CONNECTABLE)
+			return NULL;
+		root_objectid = fid->root_objectid;
+	} else if (fh_type == FILEID_BTRFS_WITH_PARENT_ROOT) {
+		if (fh_len != BTRFS_FID_SIZE_CONNECTABLE_ROOT)
+			return NULL;
+		root_objectid = fid->parent_root_objectid;
+	} else
+		return NULL;
+
+	objectid = fid->parent_objectid;
+	generation = fid->parent_gen;
+
+	return btrfs_get_dentry(sb, objectid, root_objectid, generation);
+}
+
+static struct dentry *btrfs_fh_to_dentry(struct super_block *sb, struct fid *fh,
+					 int fh_len, int fh_type)
+{
+	struct btrfs_fid *fid = (struct btrfs_fid *) fh;
+	u64 objectid, root_objectid;
+	u32 generation;
+
+	if ((fh_type != FILEID_BTRFS_WITH_PARENT ||
+	     fh_len != BTRFS_FID_SIZE_CONNECTABLE) &&
+	    (fh_type != FILEID_BTRFS_WITH_PARENT_ROOT ||
+	     fh_len != BTRFS_FID_SIZE_CONNECTABLE_ROOT) &&
+	    (fh_type != FILEID_BTRFS_WITHOUT_PARENT ||
+	     fh_len != BTRFS_FID_SIZE_NON_CONNECTABLE))
+		return NULL;
+
+	objectid = fid->objectid;
+	root_objectid = fid->root_objectid;
+	generation = fid->gen;
+
+	return btrfs_get_dentry(sb, objectid, root_objectid, generation);
+}
+
+static struct dentry *btrfs_get_parent(struct dentry *child)
+{
+	struct inode *dir = child->d_inode;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	int slot;
+	u64 objectid;
+	int ret;
+
+	path = btrfs_alloc_path();
+
+	key.objectid = dir->i_ino;
+	btrfs_set_key_type(&key, BTRFS_INODE_REF_KEY);
+	key.offset = (u64)-1;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0) {
+		/* Error */
+		btrfs_free_path(path);
+		return ERR_PTR(ret);
+	}
+	leaf = path->nodes[0];
+	slot = path->slots[0];
+	if (ret) {
+		/* btrfs_search_slot() returns the slot where we'd want to
+		   insert a backref for parent inode #0xFFFFFFFFFFFFFFFF.
+		   The _real_ backref, telling us what the parent inode
+		   _actually_ is, will be in the slot _before_ the one
+		   that btrfs_search_slot() returns. */
+		if (!slot) {
+			/* Unless there is _no_ key in the tree before... */
+			btrfs_free_path(path);
+			return ERR_PTR(-EIO);
+		}
+		slot--;
+	}
+
+	btrfs_item_key_to_cpu(leaf, &key, slot);
+	btrfs_free_path(path);
+
+	if (key.objectid != dir->i_ino || key.type != BTRFS_INODE_REF_KEY)
+		return ERR_PTR(-EINVAL);
+
+	objectid = key.offset;
+
+	/* If we are already at the root of a subvol, return the real root */
+	if (objectid == dir->i_ino)
+		return dget(dir->i_sb->s_root);
+
+	/* Build a new key for the inode item */
+	key.objectid = objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+	key.offset = 0;
+
+	return d_obtain_alias(btrfs_iget(root->fs_info->sb, &key, root, NULL));
+}
+
+const struct export_operations btrfs_export_ops = {
+	.encode_fh	= btrfs_encode_fh,
+	.fh_to_dentry	= btrfs_fh_to_dentry,
+	.fh_to_parent	= btrfs_fh_to_parent,
+	.get_parent	= btrfs_get_parent,
+};
diff --git a/fs/btrfs/export.h b/fs/btrfs/export.h
new file mode 100644
index 00000000000..074348a9584
--- /dev/null
+++ b/fs/btrfs/export.h
@@ -0,0 +1,19 @@
+#ifndef BTRFS_EXPORT_H
+#define BTRFS_EXPORT_H
+
+#include <linux/exportfs.h>
+
+extern const struct export_operations btrfs_export_ops;
+
+struct btrfs_fid {
+	u64 objectid;
+	u64 root_objectid;
+	u32 gen;
+
+	u64 parent_objectid;
+	u32 parent_gen;
+
+	u64 parent_root_objectid;
+} __attribute__ ((packed));
+
+#endif
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
new file mode 100644
index 00000000000..293da650873
--- /dev/null
+++ b/fs/btrfs/extent-tree.c
@@ -0,0 +1,5986 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+#include <linux/sched.h>
+#include <linux/pagemap.h>
+#include <linux/writeback.h>
+#include <linux/blkdev.h>
+#include <linux/version.h>
+#include "compat.h"
+#include "hash.h"
+#include "crc32c.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "print-tree.h"
+#include "transaction.h"
+#include "volumes.h"
+#include "locking.h"
+#include "ref-cache.h"
+#include "compat.h"
+
+#define PENDING_EXTENT_INSERT 0
+#define PENDING_EXTENT_DELETE 1
+#define PENDING_BACKREF_UPDATE 2
+
+struct pending_extent_op {
+	int type;
+	u64 bytenr;
+	u64 num_bytes;
+	u64 parent;
+	u64 orig_parent;
+	u64 generation;
+	u64 orig_generation;
+	int level;
+	struct list_head list;
+	int del;
+};
+
+static int finish_current_insert(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *extent_root, int all);
+static int del_pending_extents(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *extent_root, int all);
+static int pin_down_bytes(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root,
+			  u64 bytenr, u64 num_bytes, int is_data);
+static int update_block_group(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      u64 bytenr, u64 num_bytes, int alloc,
+			      int mark_free);
+
+static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
+{
+	return (cache->flags & bits) == bits;
+}
+
+/*
+ * this adds the block group to the fs_info rb tree for the block group
+ * cache
+ */
+static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
+				struct btrfs_block_group_cache *block_group)
+{
+	struct rb_node **p;
+	struct rb_node *parent = NULL;
+	struct btrfs_block_group_cache *cache;
+
+	spin_lock(&info->block_group_cache_lock);
+	p = &info->block_group_cache_tree.rb_node;
+
+	while (*p) {
+		parent = *p;
+		cache = rb_entry(parent, struct btrfs_block_group_cache,
+				 cache_node);
+		if (block_group->key.objectid < cache->key.objectid) {
+			p = &(*p)->rb_left;
+		} else if (block_group->key.objectid > cache->key.objectid) {
+			p = &(*p)->rb_right;
+		} else {
+			spin_unlock(&info->block_group_cache_lock);
+			return -EEXIST;
+		}
+	}
+
+	rb_link_node(&block_group->cache_node, parent, p);
+	rb_insert_color(&block_group->cache_node,
+			&info->block_group_cache_tree);
+	spin_unlock(&info->block_group_cache_lock);
+
+	return 0;
+}
+
+/*
+ * This will return the block group at or after bytenr if contains is 0, else
+ * it will return the block group that contains the bytenr
+ */
+static struct btrfs_block_group_cache *
+block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr,
+			      int contains)
+{
+	struct btrfs_block_group_cache *cache, *ret = NULL;
+	struct rb_node *n;
+	u64 end, start;
+
+	spin_lock(&info->block_group_cache_lock);
+	n = info->block_group_cache_tree.rb_node;
+
+	while (n) {
+		cache = rb_entry(n, struct btrfs_block_group_cache,
+				 cache_node);
+		end = cache->key.objectid + cache->key.offset - 1;
+		start = cache->key.objectid;
+
+		if (bytenr < start) {
+			if (!contains && (!ret || start < ret->key.objectid))
+				ret = cache;
+			n = n->rb_left;
+		} else if (bytenr > start) {
+			if (contains && bytenr <= end) {
+				ret = cache;
+				break;
+			}
+			n = n->rb_right;
+		} else {
+			ret = cache;
+			break;
+		}
+	}
+	if (ret)
+		atomic_inc(&ret->count);
+	spin_unlock(&info->block_group_cache_lock);
+
+	return ret;
+}
+
+/*
+ * this is only called by cache_block_group, since we could have freed extents
+ * we need to check the pinned_extents for any extents that can't be used yet
+ * since their free space will be released as soon as the transaction commits.
+ */
+static int add_new_free_space(struct btrfs_block_group_cache *block_group,
+			      struct btrfs_fs_info *info, u64 start, u64 end)
+{
+	u64 extent_start, extent_end, size;
+	int ret;
+
+	mutex_lock(&info->pinned_mutex);
+	while (start < end) {
+		ret = find_first_extent_bit(&info->pinned_extents, start,
+					    &extent_start, &extent_end,
+					    EXTENT_DIRTY);
+		if (ret)
+			break;
+
+		if (extent_start == start) {
+			start = extent_end + 1;
+		} else if (extent_start > start && extent_start < end) {
+			size = extent_start - start;
+			ret = btrfs_add_free_space(block_group, start,
+						   size);
+			BUG_ON(ret);
+			start = extent_end + 1;
+		} else {
+			break;
+		}
+	}
+
+	if (start < end) {
+		size = end - start;
+		ret = btrfs_add_free_space(block_group, start, size);
+		BUG_ON(ret);
+	}
+	mutex_unlock(&info->pinned_mutex);
+
+	return 0;
+}
+
+static int remove_sb_from_cache(struct btrfs_root *root,
+				struct btrfs_block_group_cache *cache)
+{
+	u64 bytenr;
+	u64 *logical;
+	int stripe_len;
+	int i, nr, ret;
+
+	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
+		bytenr = btrfs_sb_offset(i);
+		ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
+				       cache->key.objectid, bytenr, 0,
+				       &logical, &nr, &stripe_len);
+		BUG_ON(ret);
+		while (nr--) {
+			btrfs_remove_free_space(cache, logical[nr],
+						stripe_len);
+		}
+		kfree(logical);
+	}
+	return 0;
+}
+
+static int cache_block_group(struct btrfs_root *root,
+			     struct btrfs_block_group_cache *block_group)
+{
+	struct btrfs_path *path;
+	int ret = 0;
+	struct btrfs_key key;
+	struct extent_buffer *leaf;
+	int slot;
+	u64 last;
+
+	if (!block_group)
+		return 0;
+
+	root = root->fs_info->extent_root;
+
+	if (block_group->cached)
+		return 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->reada = 2;
+	/*
+	 * we get into deadlocks with paths held by callers of this function.
+	 * since the alloc_mutex is protecting things right now, just
+	 * skip the locking here
+	 */
+	path->skip_locking = 1;
+	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
+	key.objectid = last;
+	key.offset = 0;
+	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto err;
+
+	while (1) {
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto err;
+			if (ret == 0)
+				continue;
+			else
+				break;
+		}
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (key.objectid < block_group->key.objectid)
+			goto next;
+
+		if (key.objectid >= block_group->key.objectid +
+		    block_group->key.offset)
+			break;
+
+		if (btrfs_key_type(&key) == BTRFS_EXTENT_ITEM_KEY) {
+			add_new_free_space(block_group, root->fs_info, last,
+					   key.objectid);
+
+			last = key.objectid + key.offset;
+		}
+next:
+		path->slots[0]++;
+	}
+
+	add_new_free_space(block_group, root->fs_info, last,
+			   block_group->key.objectid +
+			   block_group->key.offset);
+
+	remove_sb_from_cache(root, block_group);
+	block_group->cached = 1;
+	ret = 0;
+err:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * return the block group that starts at or after bytenr
+ */
+static struct btrfs_block_group_cache *
+btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr)
+{
+	struct btrfs_block_group_cache *cache;
+
+	cache = block_group_cache_tree_search(info, bytenr, 0);
+
+	return cache;
+}
+
+/*
+ * return the block group that contains teh given bytenr
+ */
+struct btrfs_block_group_cache *btrfs_lookup_block_group(
+						 struct btrfs_fs_info *info,
+						 u64 bytenr)
+{
+	struct btrfs_block_group_cache *cache;
+
+	cache = block_group_cache_tree_search(info, bytenr, 1);
+
+	return cache;
+}
+
+static inline void put_block_group(struct btrfs_block_group_cache *cache)
+{
+	if (atomic_dec_and_test(&cache->count))
+		kfree(cache);
+}
+
+static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
+						  u64 flags)
+{
+	struct list_head *head = &info->space_info;
+	struct list_head *cur;
+	struct btrfs_space_info *found;
+	list_for_each(cur, head) {
+		found = list_entry(cur, struct btrfs_space_info, list);
+		if (found->flags == flags)
+			return found;
+	}
+	return NULL;
+}
+
+static u64 div_factor(u64 num, int factor)
+{
+	if (factor == 10)
+		return num;
+	num *= factor;
+	do_div(num, 10);
+	return num;
+}
+
+u64 btrfs_find_block_group(struct btrfs_root *root,
+			   u64 search_start, u64 search_hint, int owner)
+{
+	struct btrfs_block_group_cache *cache;
+	u64 used;
+	u64 last = max(search_hint, search_start);
+	u64 group_start = 0;
+	int full_search = 0;
+	int factor = 9;
+	int wrapped = 0;
+again:
+	while (1) {
+		cache = btrfs_lookup_first_block_group(root->fs_info, last);
+		if (!cache)
+			break;
+
+		spin_lock(&cache->lock);
+		last = cache->key.objectid + cache->key.offset;
+		used = btrfs_block_group_used(&cache->item);
+
+		if ((full_search || !cache->ro) &&
+		    block_group_bits(cache, BTRFS_BLOCK_GROUP_METADATA)) {
+			if (used + cache->pinned + cache->reserved <
+			    div_factor(cache->key.offset, factor)) {
+				group_start = cache->key.objectid;
+				spin_unlock(&cache->lock);
+				put_block_group(cache);
+				goto found;
+			}
+		}
+		spin_unlock(&cache->lock);
+		put_block_group(cache);
+		cond_resched();
+	}
+	if (!wrapped) {
+		last = search_start;
+		wrapped = 1;
+		goto again;
+	}
+	if (!full_search && factor < 10) {
+		last = search_start;
+		full_search = 1;
+		factor = 10;
+		goto again;
+	}
+found:
+	return group_start;
+}
+
+/* simple helper to search for an existing extent at a given offset */
+int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len)
+{
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_path *path;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	key.objectid = start;
+	key.offset = len;
+	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
+	ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path,
+				0, 0);
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * Back reference rules.  Back refs have three main goals:
+ *
+ * 1) differentiate between all holders of references to an extent so that
+ *    when a reference is dropped we can make sure it was a valid reference
+ *    before freeing the extent.
+ *
+ * 2) Provide enough information to quickly find the holders of an extent
+ *    if we notice a given block is corrupted or bad.
+ *
+ * 3) Make it easy to migrate blocks for FS shrinking or storage pool
+ *    maintenance.  This is actually the same as #2, but with a slightly
+ *    different use case.
+ *
+ * File extents can be referenced by:
+ *
+ * - multiple snapshots, subvolumes, or different generations in one subvol
+ * - different files inside a single subvolume
+ * - different offsets inside a file (bookend extents in file.c)
+ *
+ * The extent ref structure has fields for:
+ *
+ * - Objectid of the subvolume root
+ * - Generation number of the tree holding the reference
+ * - objectid of the file holding the reference
+ * - number of references holding by parent node (alway 1 for tree blocks)
+ *
+ * Btree leaf may hold multiple references to a file extent. In most cases,
+ * these references are from same file and the corresponding offsets inside
+ * the file are close together.
+ *
+ * When a file extent is allocated the fields are filled in:
+ *     (root_key.objectid, trans->transid, inode objectid, 1)
+ *
+ * When a leaf is cow'd new references are added for every file extent found
+ * in the leaf.  It looks similar to the create case, but trans->transid will
+ * be different when the block is cow'd.
+ *
+ *     (root_key.objectid, trans->transid, inode objectid,
+ *      number of references in the leaf)
+ *
+ * When a file extent is removed either during snapshot deletion or
+ * file truncation, we find the corresponding back reference and check
+ * the following fields:
+ *
+ *     (btrfs_header_owner(leaf), btrfs_header_generation(leaf),
+ *      inode objectid)
+ *
+ * Btree extents can be referenced by:
+ *
+ * - Different subvolumes
+ * - Different generations of the same subvolume
+ *
+ * When a tree block is created, back references are inserted:
+ *
+ * (root->root_key.objectid, trans->transid, level, 1)
+ *
+ * When a tree block is cow'd, new back references are added for all the
+ * blocks it points to. If the tree block isn't in reference counted root,
+ * the old back references are removed. These new back references are of
+ * the form (trans->transid will have increased since creation):
+ *
+ * (root->root_key.objectid, trans->transid, level, 1)
+ *
+ * When a backref is in deleting, the following fields are checked:
+ *
+ * if backref was for a tree root:
+ *     (btrfs_header_owner(itself), btrfs_header_generation(itself), level)
+ * else
+ *     (btrfs_header_owner(parent), btrfs_header_generation(parent), level)
+ *
+ * Back Reference Key composing:
+ *
+ * The key objectid corresponds to the first byte in the extent, the key
+ * type is set to BTRFS_EXTENT_REF_KEY, and the key offset is the first
+ * byte of parent extent. If a extent is tree root, the key offset is set
+ * to the key objectid.
+ */
+
+static noinline int lookup_extent_backref(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path,
+					  u64 bytenr, u64 parent,
+					  u64 ref_root, u64 ref_generation,
+					  u64 owner_objectid, int del)
+{
+	struct btrfs_key key;
+	struct btrfs_extent_ref *ref;
+	struct extent_buffer *leaf;
+	u64 ref_objectid;
+	int ret;
+
+	key.objectid = bytenr;
+	key.type = BTRFS_EXTENT_REF_KEY;
+	key.offset = parent;
+
+	ret = btrfs_search_slot(trans, root, &key, path, del ? -1 : 0, 1);
+	if (ret < 0)
+		goto out;
+	if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_ref);
+	ref_objectid = btrfs_ref_objectid(leaf, ref);
+	if (btrfs_ref_root(leaf, ref) != ref_root ||
+	    btrfs_ref_generation(leaf, ref) != ref_generation ||
+	    (ref_objectid != owner_objectid &&
+	     ref_objectid != BTRFS_MULTIPLE_OBJECTIDS)) {
+		ret = -EIO;
+		WARN_ON(1);
+		goto out;
+	}
+	ret = 0;
+out:
+	return ret;
+}
+
+/*
+ * updates all the backrefs that are pending on update_list for the
+ * extent_root
+ */
+static noinline int update_backrefs(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *extent_root,
+				    struct btrfs_path *path,
+				    struct list_head *update_list)
+{
+	struct btrfs_key key;
+	struct btrfs_extent_ref *ref;
+	struct btrfs_fs_info *info = extent_root->fs_info;
+	struct pending_extent_op *op;
+	struct extent_buffer *leaf;
+	int ret = 0;
+	struct list_head *cur = update_list->next;
+	u64 ref_objectid;
+	u64 ref_root = extent_root->root_key.objectid;
+
+	op = list_entry(cur, struct pending_extent_op, list);
+
+search:
+	key.objectid = op->bytenr;
+	key.type = BTRFS_EXTENT_REF_KEY;
+	key.offset = op->orig_parent;
+
+	ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 1);
+	BUG_ON(ret);
+
+	leaf = path->nodes[0];
+
+loop:
+	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_ref);
+
+	ref_objectid = btrfs_ref_objectid(leaf, ref);
+
+	if (btrfs_ref_root(leaf, ref) != ref_root ||
+	    btrfs_ref_generation(leaf, ref) != op->orig_generation ||
+	    (ref_objectid != op->level &&
+	     ref_objectid != BTRFS_MULTIPLE_OBJECTIDS)) {
+		printk(KERN_ERR "btrfs couldn't find %llu, parent %llu, "
+		       "root %llu, owner %u\n",
+		       (unsigned long long)op->bytenr,
+		       (unsigned long long)op->orig_parent,
+		       (unsigned long long)ref_root, op->level);
+		btrfs_print_leaf(extent_root, leaf);
+		BUG();
+	}
+
+	key.objectid = op->bytenr;
+	key.offset = op->parent;
+	key.type = BTRFS_EXTENT_REF_KEY;
+	ret = btrfs_set_item_key_safe(trans, extent_root, path, &key);
+	BUG_ON(ret);
+	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_ref);
+	btrfs_set_ref_generation(leaf, ref, op->generation);
+
+	cur = cur->next;
+
+	list_del_init(&op->list);
+	unlock_extent(&info->extent_ins, op->bytenr,
+		      op->bytenr + op->num_bytes - 1, GFP_NOFS);
+	kfree(op);
+
+	if (cur == update_list) {
+		btrfs_mark_buffer_dirty(path->nodes[0]);
+		btrfs_release_path(extent_root, path);
+		goto out;
+	}
+
+	op = list_entry(cur, struct pending_extent_op, list);
+
+	path->slots[0]++;
+	while (path->slots[0] < btrfs_header_nritems(leaf)) {
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+		if (key.objectid == op->bytenr &&
+		    key.type == BTRFS_EXTENT_REF_KEY)
+			goto loop;
+		path->slots[0]++;
+	}
+
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	btrfs_release_path(extent_root, path);
+	goto search;
+
+out:
+	return 0;
+}
+
+static noinline int insert_extents(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *extent_root,
+				   struct btrfs_path *path,
+				   struct list_head *insert_list, int nr)
+{
+	struct btrfs_key *keys;
+	u32 *data_size;
+	struct pending_extent_op *op;
+	struct extent_buffer *leaf;
+	struct list_head *cur = insert_list->next;
+	struct btrfs_fs_info *info = extent_root->fs_info;
+	u64 ref_root = extent_root->root_key.objectid;
+	int i = 0, last = 0, ret;
+	int total = nr * 2;
+
+	if (!nr)
+		return 0;
+
+	keys = kzalloc(total * sizeof(struct btrfs_key), GFP_NOFS);
+	if (!keys)
+		return -ENOMEM;
+
+	data_size = kzalloc(total * sizeof(u32), GFP_NOFS);
+	if (!data_size) {
+		kfree(keys);
+		return -ENOMEM;
+	}
+
+	list_for_each_entry(op, insert_list, list) {
+		keys[i].objectid = op->bytenr;
+		keys[i].offset = op->num_bytes;
+		keys[i].type = BTRFS_EXTENT_ITEM_KEY;
+		data_size[i] = sizeof(struct btrfs_extent_item);
+		i++;
+
+		keys[i].objectid = op->bytenr;
+		keys[i].offset = op->parent;
+		keys[i].type = BTRFS_EXTENT_REF_KEY;
+		data_size[i] = sizeof(struct btrfs_extent_ref);
+		i++;
+	}
+
+	op = list_entry(cur, struct pending_extent_op, list);
+	i = 0;
+	while (i < total) {
+		int c;
+		ret = btrfs_insert_some_items(trans, extent_root, path,
+					      keys+i, data_size+i, total-i);
+		BUG_ON(ret < 0);
+
+		if (last && ret > 1)
+			BUG();
+
+		leaf = path->nodes[0];
+		for (c = 0; c < ret; c++) {
+			int ref_first = keys[i].type == BTRFS_EXTENT_REF_KEY;
+
+			/*
+			 * if the first item we inserted was a backref, then
+			 * the EXTENT_ITEM will be the odd c's, else it will
+			 * be the even c's
+			 */
+			if ((ref_first && (c % 2)) ||
+			    (!ref_first && !(c % 2))) {
+				struct btrfs_extent_item *itm;
+
+				itm = btrfs_item_ptr(leaf, path->slots[0] + c,
+						     struct btrfs_extent_item);
+				btrfs_set_extent_refs(path->nodes[0], itm, 1);
+				op->del++;
+			} else {
+				struct btrfs_extent_ref *ref;
+
+				ref = btrfs_item_ptr(leaf, path->slots[0] + c,
+						     struct btrfs_extent_ref);
+				btrfs_set_ref_root(leaf, ref, ref_root);
+				btrfs_set_ref_generation(leaf, ref,
+							 op->generation);
+				btrfs_set_ref_objectid(leaf, ref, op->level);
+				btrfs_set_ref_num_refs(leaf, ref, 1);
+				op->del++;
+			}
+
+			/*
+			 * using del to see when its ok to free up the
+			 * pending_extent_op.  In the case where we insert the
+			 * last item on the list in order to help do batching
+			 * we need to not free the extent op until we actually
+			 * insert the extent_item
+			 */
+			if (op->del == 2) {
+				unlock_extent(&info->extent_ins, op->bytenr,
+					      op->bytenr + op->num_bytes - 1,
+					      GFP_NOFS);
+				cur = cur->next;
+				list_del_init(&op->list);
+				kfree(op);
+				if (cur != insert_list)
+					op = list_entry(cur,
+						struct pending_extent_op,
+						list);
+			}
+		}
+		btrfs_mark_buffer_dirty(leaf);
+		btrfs_release_path(extent_root, path);
+
+		/*
+		 * Ok backref's and items usually go right next to eachother,
+		 * but if we could only insert 1 item that means that we
+		 * inserted on the end of a leaf, and we have no idea what may
+		 * be on the next leaf so we just play it safe.  In order to
+		 * try and help this case we insert the last thing on our
+		 * insert list so hopefully it will end up being the last
+		 * thing on the leaf and everything else will be before it,
+		 * which will let us insert a whole bunch of items at the same
+		 * time.
+		 */
+		if (ret == 1 && !last && (i + ret < total)) {
+			/*
+			 * last: where we will pick up the next time around
+			 * i: our current key to insert, will be total - 1
+			 * cur: the current op we are screwing with
+			 * op: duh
+			 */
+			last = i + ret;
+			i = total - 1;
+			cur = insert_list->prev;
+			op = list_entry(cur, struct pending_extent_op, list);
+		} else if (last) {
+			/*
+			 * ok we successfully inserted the last item on the
+			 * list, lets reset everything
+			 *
+			 * i: our current key to insert, so where we left off
+			 *    last time
+			 * last: done with this
+			 * cur: the op we are messing with
+			 * op: duh
+			 * total: since we inserted the last key, we need to
+			 *        decrement total so we dont overflow
+			 */
+			i = last;
+			last = 0;
+			total--;
+			if (i < total) {
+				cur = insert_list->next;
+				op = list_entry(cur, struct pending_extent_op,
+						list);
+			}
+		} else {
+			i += ret;
+		}
+
+		cond_resched();
+	}
+	ret = 0;
+	kfree(keys);
+	kfree(data_size);
+	return ret;
+}
+
+static noinline int insert_extent_backref(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path,
+					  u64 bytenr, u64 parent,
+					  u64 ref_root, u64 ref_generation,
+					  u64 owner_objectid)
+{
+	struct btrfs_key key;
+	struct extent_buffer *leaf;
+	struct btrfs_extent_ref *ref;
+	u32 num_refs;
+	int ret;
+
+	key.objectid = bytenr;
+	key.type = BTRFS_EXTENT_REF_KEY;
+	key.offset = parent;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*ref));
+	if (ret == 0) {
+		leaf = path->nodes[0];
+		ref = btrfs_item_ptr(leaf, path->slots[0],
+				     struct btrfs_extent_ref);
+		btrfs_set_ref_root(leaf, ref, ref_root);
+		btrfs_set_ref_generation(leaf, ref, ref_generation);
+		btrfs_set_ref_objectid(leaf, ref, owner_objectid);
+		btrfs_set_ref_num_refs(leaf, ref, 1);
+	} else if (ret == -EEXIST) {
+		u64 existing_owner;
+		BUG_ON(owner_objectid < BTRFS_FIRST_FREE_OBJECTID);
+		leaf = path->nodes[0];
+		ref = btrfs_item_ptr(leaf, path->slots[0],
+				     struct btrfs_extent_ref);
+		if (btrfs_ref_root(leaf, ref) != ref_root ||
+		    btrfs_ref_generation(leaf, ref) != ref_generation) {
+			ret = -EIO;
+			WARN_ON(1);
+			goto out;
+		}
+
+		num_refs = btrfs_ref_num_refs(leaf, ref);
+		BUG_ON(num_refs == 0);
+		btrfs_set_ref_num_refs(leaf, ref, num_refs + 1);
+
+		existing_owner = btrfs_ref_objectid(leaf, ref);
+		if (existing_owner != owner_objectid &&
+		    existing_owner != BTRFS_MULTIPLE_OBJECTIDS) {
+			btrfs_set_ref_objectid(leaf, ref,
+					BTRFS_MULTIPLE_OBJECTIDS);
+		}
+		ret = 0;
+	} else {
+		goto out;
+	}
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+out:
+	btrfs_release_path(root, path);
+	return ret;
+}
+
+static noinline int remove_extent_backref(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_extent_ref *ref;
+	u32 num_refs;
+	int ret = 0;
+
+	leaf = path->nodes[0];
+	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_ref);
+	num_refs = btrfs_ref_num_refs(leaf, ref);
+	BUG_ON(num_refs == 0);
+	num_refs -= 1;
+	if (num_refs == 0) {
+		ret = btrfs_del_item(trans, root, path);
+	} else {
+		btrfs_set_ref_num_refs(leaf, ref, num_refs);
+		btrfs_mark_buffer_dirty(leaf);
+	}
+	btrfs_release_path(root, path);
+	return ret;
+}
+
+#ifdef BIO_RW_DISCARD
+static void btrfs_issue_discard(struct block_device *bdev,
+				u64 start, u64 len)
+{
+	blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL);
+}
+#endif
+
+static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
+				u64 num_bytes)
+{
+#ifdef BIO_RW_DISCARD
+	int ret;
+	u64 map_length = num_bytes;
+	struct btrfs_multi_bio *multi = NULL;
+
+	/* Tell the block device(s) that the sectors can be discarded */
+	ret = btrfs_map_block(&root->fs_info->mapping_tree, READ,
+			      bytenr, &map_length, &multi, 0);
+	if (!ret) {
+		struct btrfs_bio_stripe *stripe = multi->stripes;
+		int i;
+
+		if (map_length > num_bytes)
+			map_length = num_bytes;
+
+		for (i = 0; i < multi->num_stripes; i++, stripe++) {
+			btrfs_issue_discard(stripe->dev->bdev,
+					    stripe->physical,
+					    map_length);
+		}
+		kfree(multi);
+	}
+
+	return ret;
+#else
+	return 0;
+#endif
+}
+
+static noinline int free_extents(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *extent_root,
+				 struct list_head *del_list)
+{
+	struct btrfs_fs_info *info = extent_root->fs_info;
+	struct btrfs_path *path;
+	struct btrfs_key key, found_key;
+	struct extent_buffer *leaf;
+	struct list_head *cur;
+	struct pending_extent_op *op;
+	struct btrfs_extent_item *ei;
+	int ret, num_to_del, extent_slot = 0, found_extent = 0;
+	u32 refs;
+	u64 bytes_freed = 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	path->reada = 1;
+
+search:
+	/* search for the backref for the current ref we want to delete */
+	cur = del_list->next;
+	op = list_entry(cur, struct pending_extent_op, list);
+	ret = lookup_extent_backref(trans, extent_root, path, op->bytenr,
+				    op->orig_parent,
+				    extent_root->root_key.objectid,
+				    op->orig_generation, op->level, 1);
+	if (ret) {
+		printk(KERN_ERR "btrfs unable to find backref byte nr %llu "
+		       "root %llu gen %llu owner %u\n",
+		       (unsigned long long)op->bytenr,
+		       (unsigned long long)extent_root->root_key.objectid,
+		       (unsigned long long)op->orig_generation, op->level);
+		btrfs_print_leaf(extent_root, path->nodes[0]);
+		WARN_ON(1);
+		goto out;
+	}
+
+	extent_slot = path->slots[0];
+	num_to_del = 1;
+	found_extent = 0;
+
+	/*
+	 * if we aren't the first item on the leaf we can move back one and see
+	 * if our ref is right next to our extent item
+	 */
+	if (likely(extent_slot)) {
+		extent_slot--;
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+				      extent_slot);
+		if (found_key.objectid == op->bytenr &&
+		    found_key.type == BTRFS_EXTENT_ITEM_KEY &&
+		    found_key.offset == op->num_bytes) {
+			num_to_del++;
+			found_extent = 1;
+		}
+	}
+
+	/*
+	 * if we didn't find the extent we need to delete the backref and then
+	 * search for the extent item key so we can update its ref count
+	 */
+	if (!found_extent) {
+		key.objectid = op->bytenr;
+		key.type = BTRFS_EXTENT_ITEM_KEY;
+		key.offset = op->num_bytes;
+
+		ret = remove_extent_backref(trans, extent_root, path);
+		BUG_ON(ret);
+		btrfs_release_path(extent_root, path);
+		ret = btrfs_search_slot(trans, extent_root, &key, path, -1, 1);
+		BUG_ON(ret);
+		extent_slot = path->slots[0];
+	}
+
+	/* this is where we update the ref count for the extent */
+	leaf = path->nodes[0];
+	ei = btrfs_item_ptr(leaf, extent_slot, struct btrfs_extent_item);
+	refs = btrfs_extent_refs(leaf, ei);
+	BUG_ON(refs == 0);
+	refs--;
+	btrfs_set_extent_refs(leaf, ei, refs);
+
+	btrfs_mark_buffer_dirty(leaf);
+
+	/*
+	 * This extent needs deleting.  The reason cur_slot is extent_slot +
+	 * num_to_del is because extent_slot points to the slot where the extent
+	 * is, and if the backref was not right next to the extent we will be
+	 * deleting at least 1 item, and will want to start searching at the
+	 * slot directly next to extent_slot.  However if we did find the
+	 * backref next to the extent item them we will be deleting at least 2
+	 * items and will want to start searching directly after the ref slot
+	 */
+	if (!refs) {
+		struct list_head *pos, *n, *end;
+		int cur_slot = extent_slot+num_to_del;
+		u64 super_used;
+		u64 root_used;
+
+		path->slots[0] = extent_slot;
+		bytes_freed = op->num_bytes;
+
+		mutex_lock(&info->pinned_mutex);
+		ret = pin_down_bytes(trans, extent_root, op->bytenr,
+				     op->num_bytes, op->level >=
+				     BTRFS_FIRST_FREE_OBJECTID);
+		mutex_unlock(&info->pinned_mutex);
+		BUG_ON(ret < 0);
+		op->del = ret;
+
+		/*
+		 * we need to see if we can delete multiple things at once, so
+		 * start looping through the list of extents we are wanting to
+		 * delete and see if their extent/backref's are right next to
+		 * eachother and the extents only have 1 ref
+		 */
+		for (pos = cur->next; pos != del_list; pos = pos->next) {
+			struct pending_extent_op *tmp;
+
+			tmp = list_entry(pos, struct pending_extent_op, list);
+
+			/* we only want to delete extent+ref at this stage */
+			if (cur_slot >= btrfs_header_nritems(leaf) - 1)
+				break;
+
+			btrfs_item_key_to_cpu(leaf, &found_key, cur_slot);
+			if (found_key.objectid != tmp->bytenr ||
+			    found_key.type != BTRFS_EXTENT_ITEM_KEY ||
+			    found_key.offset != tmp->num_bytes)
+				break;
+
+			/* check to make sure this extent only has one ref */
+			ei = btrfs_item_ptr(leaf, cur_slot,
+					    struct btrfs_extent_item);
+			if (btrfs_extent_refs(leaf, ei) != 1)
+				break;
+
+			btrfs_item_key_to_cpu(leaf, &found_key, cur_slot+1);
+			if (found_key.objectid != tmp->bytenr ||
+			    found_key.type != BTRFS_EXTENT_REF_KEY ||
+			    found_key.offset != tmp->orig_parent)
+				break;
+
+			/*
+			 * the ref is right next to the extent, we can set the
+			 * ref count to 0 since we will delete them both now
+			 */
+			btrfs_set_extent_refs(leaf, ei, 0);
+
+			/* pin down the bytes for this extent */
+			mutex_lock(&info->pinned_mutex);
+			ret = pin_down_bytes(trans, extent_root, tmp->bytenr,
+					     tmp->num_bytes, tmp->level >=
+					     BTRFS_FIRST_FREE_OBJECTID);
+			mutex_unlock(&info->pinned_mutex);
+			BUG_ON(ret < 0);
+
+			/*
+			 * use the del field to tell if we need to go ahead and
+			 * free up the extent when we delete the item or not.
+			 */
+			tmp->del = ret;
+			bytes_freed += tmp->num_bytes;
+
+			num_to_del += 2;
+			cur_slot += 2;
+		}
+		end = pos;
+
+		/* update the free space counters */
+		spin_lock(&info->delalloc_lock);
+		super_used = btrfs_super_bytes_used(&info->super_copy);
+		btrfs_set_super_bytes_used(&info->super_copy,
+					   super_used - bytes_freed);
+
+		root_used = btrfs_root_used(&extent_root->root_item);
+		btrfs_set_root_used(&extent_root->root_item,
+				    root_used - bytes_freed);
+		spin_unlock(&info->delalloc_lock);
+
+		/* delete the items */
+		ret = btrfs_del_items(trans, extent_root, path,
+				      path->slots[0], num_to_del);
+		BUG_ON(ret);
+
+		/*
+		 * loop through the extents we deleted and do the cleanup work
+		 * on them
+		 */
+		for (pos = cur, n = pos->next; pos != end;
+		     pos = n, n = pos->next) {
+			struct pending_extent_op *tmp;
+			tmp = list_entry(pos, struct pending_extent_op, list);
+
+			/*
+			 * remember tmp->del tells us wether or not we pinned
+			 * down the extent
+			 */
+			ret = update_block_group(trans, extent_root,
+						 tmp->bytenr, tmp->num_bytes, 0,
+						 tmp->del);
+			BUG_ON(ret);
+
+			list_del_init(&tmp->list);
+			unlock_extent(&info->extent_ins, tmp->bytenr,
+				      tmp->bytenr + tmp->num_bytes - 1,
+				      GFP_NOFS);
+			kfree(tmp);
+		}
+	} else if (refs && found_extent) {
+		/*
+		 * the ref and extent were right next to eachother, but the
+		 * extent still has a ref, so just free the backref and keep
+		 * going
+		 */
+		ret = remove_extent_backref(trans, extent_root, path);
+		BUG_ON(ret);
+
+		list_del_init(&op->list);
+		unlock_extent(&info->extent_ins, op->bytenr,
+			      op->bytenr + op->num_bytes - 1, GFP_NOFS);
+		kfree(op);
+	} else {
+		/*
+		 * the extent has multiple refs and the backref we were looking
+		 * for was not right next to it, so just unlock and go next,
+		 * we're good to go
+		 */
+		list_del_init(&op->list);
+		unlock_extent(&info->extent_ins, op->bytenr,
+			      op->bytenr + op->num_bytes - 1, GFP_NOFS);
+		kfree(op);
+	}
+
+	btrfs_release_path(extent_root, path);
+	if (!list_empty(del_list))
+		goto search;
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int __btrfs_update_extent_ref(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root, u64 bytenr,
+				     u64 orig_parent, u64 parent,
+				     u64 orig_root, u64 ref_root,
+				     u64 orig_generation, u64 ref_generation,
+				     u64 owner_objectid)
+{
+	int ret;
+	struct btrfs_root *extent_root = root->fs_info->extent_root;
+	struct btrfs_path *path;
+
+	if (root == root->fs_info->extent_root) {
+		struct pending_extent_op *extent_op;
+		u64 num_bytes;
+
+		BUG_ON(owner_objectid >= BTRFS_MAX_LEVEL);
+		num_bytes = btrfs_level_size(root, (int)owner_objectid);
+		mutex_lock(&root->fs_info->extent_ins_mutex);
+		if (test_range_bit(&root->fs_info->extent_ins, bytenr,
+				bytenr + num_bytes - 1, EXTENT_WRITEBACK, 0)) {
+			u64 priv;
+			ret = get_state_private(&root->fs_info->extent_ins,
+						bytenr, &priv);
+			BUG_ON(ret);
+			extent_op = (struct pending_extent_op *)
+							(unsigned long)priv;
+			BUG_ON(extent_op->parent != orig_parent);
+			BUG_ON(extent_op->generation != orig_generation);
+
+			extent_op->parent = parent;
+			extent_op->generation = ref_generation;
+		} else {
+			extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
+			BUG_ON(!extent_op);
+
+			extent_op->type = PENDING_BACKREF_UPDATE;
+			extent_op->bytenr = bytenr;
+			extent_op->num_bytes = num_bytes;
+			extent_op->parent = parent;
+			extent_op->orig_parent = orig_parent;
+			extent_op->generation = ref_generation;
+			extent_op->orig_generation = orig_generation;
+			extent_op->level = (int)owner_objectid;
+			INIT_LIST_HEAD(&extent_op->list);
+			extent_op->del = 0;
+
+			set_extent_bits(&root->fs_info->extent_ins,
+					bytenr, bytenr + num_bytes - 1,
+					EXTENT_WRITEBACK, GFP_NOFS);
+			set_state_private(&root->fs_info->extent_ins,
+					  bytenr, (unsigned long)extent_op);
+		}
+		mutex_unlock(&root->fs_info->extent_ins_mutex);
+		return 0;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	ret = lookup_extent_backref(trans, extent_root, path,
+				    bytenr, orig_parent, orig_root,
+				    orig_generation, owner_objectid, 1);
+	if (ret)
+		goto out;
+	ret = remove_extent_backref(trans, extent_root, path);
+	if (ret)
+		goto out;
+	ret = insert_extent_backref(trans, extent_root, path, bytenr,
+				    parent, ref_root, ref_generation,
+				    owner_objectid);
+	BUG_ON(ret);
+	finish_current_insert(trans, extent_root, 0);
+	del_pending_extents(trans, extent_root, 0);
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_update_extent_ref(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, u64 bytenr,
+			    u64 orig_parent, u64 parent,
+			    u64 ref_root, u64 ref_generation,
+			    u64 owner_objectid)
+{
+	int ret;
+	if (ref_root == BTRFS_TREE_LOG_OBJECTID &&
+	    owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
+		return 0;
+	ret = __btrfs_update_extent_ref(trans, root, bytenr, orig_parent,
+					parent, ref_root, ref_root,
+					ref_generation, ref_generation,
+					owner_objectid);
+	return ret;
+}
+
+static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root, u64 bytenr,
+				  u64 orig_parent, u64 parent,
+				  u64 orig_root, u64 ref_root,
+				  u64 orig_generation, u64 ref_generation,
+				  u64 owner_objectid)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct btrfs_key key;
+	struct extent_buffer *l;
+	struct btrfs_extent_item *item;
+	u32 refs;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->reada = 1;
+	key.objectid = bytenr;
+	key.type = BTRFS_EXTENT_ITEM_KEY;
+	key.offset = (u64)-1;
+
+	ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, path,
+				0, 1);
+	if (ret < 0)
+		return ret;
+	BUG_ON(ret == 0 || path->slots[0] == 0);
+
+	path->slots[0]--;
+	l = path->nodes[0];
+
+	btrfs_item_key_to_cpu(l, &key, path->slots[0]);
+	if (key.objectid != bytenr) {
+		btrfs_print_leaf(root->fs_info->extent_root, path->nodes[0]);
+		printk(KERN_ERR "btrfs wanted %llu found %llu\n",
+		       (unsigned long long)bytenr,
+		       (unsigned long long)key.objectid);
+		BUG();
+	}
+	BUG_ON(key.type != BTRFS_EXTENT_ITEM_KEY);
+
+	item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
+	refs = btrfs_extent_refs(l, item);
+	btrfs_set_extent_refs(l, item, refs + 1);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+
+	btrfs_release_path(root->fs_info->extent_root, path);
+
+	path->reada = 1;
+	ret = insert_extent_backref(trans, root->fs_info->extent_root,
+				    path, bytenr, parent,
+				    ref_root, ref_generation,
+				    owner_objectid);
+	BUG_ON(ret);
+	finish_current_insert(trans, root->fs_info->extent_root, 0);
+	del_pending_extents(trans, root->fs_info->extent_root, 0);
+
+	btrfs_free_path(path);
+	return 0;
+}
+
+int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root,
+			 u64 bytenr, u64 num_bytes, u64 parent,
+			 u64 ref_root, u64 ref_generation,
+			 u64 owner_objectid)
+{
+	int ret;
+	if (ref_root == BTRFS_TREE_LOG_OBJECTID &&
+	    owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
+		return 0;
+	ret = __btrfs_inc_extent_ref(trans, root, bytenr, 0, parent,
+				     0, ref_root, 0, ref_generation,
+				     owner_objectid);
+	return ret;
+}
+
+int btrfs_extent_post_op(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root)
+{
+	finish_current_insert(trans, root->fs_info->extent_root, 1);
+	del_pending_extents(trans, root->fs_info->extent_root, 1);
+	return 0;
+}
+
+int btrfs_lookup_extent_ref(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root, u64 bytenr,
+			    u64 num_bytes, u32 *refs)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct btrfs_key key;
+	struct extent_buffer *l;
+	struct btrfs_extent_item *item;
+
+	WARN_ON(num_bytes < root->sectorsize);
+	path = btrfs_alloc_path();
+	path->reada = 1;
+	key.objectid = bytenr;
+	key.offset = num_bytes;
+	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
+	ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, path,
+				0, 0);
+	if (ret < 0)
+		goto out;
+	if (ret != 0) {
+		btrfs_print_leaf(root, path->nodes[0]);
+		printk(KERN_INFO "btrfs failed to find block number %llu\n",
+		       (unsigned long long)bytenr);
+		BUG();
+	}
+	l = path->nodes[0];
+	item = btrfs_item_ptr(l, path->slots[0], struct btrfs_extent_item);
+	*refs = btrfs_extent_refs(l, item);
+out:
+	btrfs_free_path(path);
+	return 0;
+}
+
+int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, u64 objectid, u64 bytenr)
+{
+	struct btrfs_root *extent_root = root->fs_info->extent_root;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_extent_ref *ref_item;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	u64 ref_root;
+	u64 last_snapshot;
+	u32 nritems;
+	int ret;
+
+	key.objectid = bytenr;
+	key.offset = (u64)-1;
+	key.type = BTRFS_EXTENT_ITEM_KEY;
+
+	path = btrfs_alloc_path();
+	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	BUG_ON(ret == 0);
+
+	ret = -ENOENT;
+	if (path->slots[0] == 0)
+		goto out;
+
+	path->slots[0]--;
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+	if (found_key.objectid != bytenr ||
+	    found_key.type != BTRFS_EXTENT_ITEM_KEY)
+		goto out;
+
+	last_snapshot = btrfs_root_last_snapshot(&root->root_item);
+	while (1) {
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(extent_root, path);
+			if (ret < 0)
+				goto out;
+			if (ret == 0)
+				continue;
+			break;
+		}
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (found_key.objectid != bytenr)
+			break;
+
+		if (found_key.type != BTRFS_EXTENT_REF_KEY) {
+			path->slots[0]++;
+			continue;
+		}
+
+		ref_item = btrfs_item_ptr(leaf, path->slots[0],
+					  struct btrfs_extent_ref);
+		ref_root = btrfs_ref_root(leaf, ref_item);
+		if ((ref_root != root->root_key.objectid &&
+		     ref_root != BTRFS_TREE_LOG_OBJECTID) ||
+		     objectid != btrfs_ref_objectid(leaf, ref_item)) {
+			ret = 1;
+			goto out;
+		}
+		if (btrfs_ref_generation(leaf, ref_item) <= last_snapshot) {
+			ret = 1;
+			goto out;
+		}
+
+		path->slots[0]++;
+	}
+	ret = 0;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		    struct extent_buffer *buf, u32 nr_extents)
+{
+	struct btrfs_key key;
+	struct btrfs_file_extent_item *fi;
+	u64 root_gen;
+	u32 nritems;
+	int i;
+	int level;
+	int ret = 0;
+	int shared = 0;
+
+	if (!root->ref_cows)
+		return 0;
+
+	if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
+		shared = 0;
+		root_gen = root->root_key.offset;
+	} else {
+		shared = 1;
+		root_gen = trans->transid - 1;
+	}
+
+	level = btrfs_header_level(buf);
+	nritems = btrfs_header_nritems(buf);
+
+	if (level == 0) {
+		struct btrfs_leaf_ref *ref;
+		struct btrfs_extent_info *info;
+
+		ref = btrfs_alloc_leaf_ref(root, nr_extents);
+		if (!ref) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		ref->root_gen = root_gen;
+		ref->bytenr = buf->start;
+		ref->owner = btrfs_header_owner(buf);
+		ref->generation = btrfs_header_generation(buf);
+		ref->nritems = nr_extents;
+		info = ref->extents;
+
+		for (i = 0; nr_extents > 0 && i < nritems; i++) {
+			u64 disk_bytenr;
+			btrfs_item_key_to_cpu(buf, &key, i);
+			if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
+				continue;
+			fi = btrfs_item_ptr(buf, i,
+					    struct btrfs_file_extent_item);
+			if (btrfs_file_extent_type(buf, fi) ==
+			    BTRFS_FILE_EXTENT_INLINE)
+				continue;
+			disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
+			if (disk_bytenr == 0)
+				continue;
+
+			info->bytenr = disk_bytenr;
+			info->num_bytes =
+				btrfs_file_extent_disk_num_bytes(buf, fi);
+			info->objectid = key.objectid;
+			info->offset = key.offset;
+			info++;
+		}
+
+		ret = btrfs_add_leaf_ref(root, ref, shared);
+		if (ret == -EEXIST && shared) {
+			struct btrfs_leaf_ref *old;
+			old = btrfs_lookup_leaf_ref(root, ref->bytenr);
+			BUG_ON(!old);
+			btrfs_remove_leaf_ref(root, old);
+			btrfs_free_leaf_ref(root, old);
+			ret = btrfs_add_leaf_ref(root, ref, shared);
+		}
+		WARN_ON(ret);
+		btrfs_free_leaf_ref(root, ref);
+	}
+out:
+	return ret;
+}
+
+int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		  struct extent_buffer *orig_buf, struct extent_buffer *buf,
+		  u32 *nr_extents)
+{
+	u64 bytenr;
+	u64 ref_root;
+	u64 orig_root;
+	u64 ref_generation;
+	u64 orig_generation;
+	u32 nritems;
+	u32 nr_file_extents = 0;
+	struct btrfs_key key;
+	struct btrfs_file_extent_item *fi;
+	int i;
+	int level;
+	int ret = 0;
+	int faili = 0;
+	int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *,
+			    u64, u64, u64, u64, u64, u64, u64, u64);
+
+	ref_root = btrfs_header_owner(buf);
+	ref_generation = btrfs_header_generation(buf);
+	orig_root = btrfs_header_owner(orig_buf);
+	orig_generation = btrfs_header_generation(orig_buf);
+
+	nritems = btrfs_header_nritems(buf);
+	level = btrfs_header_level(buf);
+
+	if (root->ref_cows) {
+		process_func = __btrfs_inc_extent_ref;
+	} else {
+		if (level == 0 &&
+		    root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+			goto out;
+		if (level != 0 &&
+		    root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID)
+			goto out;
+		process_func = __btrfs_update_extent_ref;
+	}
+
+	for (i = 0; i < nritems; i++) {
+		cond_resched();
+		if (level == 0) {
+			btrfs_item_key_to_cpu(buf, &key, i);
+			if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
+				continue;
+			fi = btrfs_item_ptr(buf, i,
+					    struct btrfs_file_extent_item);
+			if (btrfs_file_extent_type(buf, fi) ==
+			    BTRFS_FILE_EXTENT_INLINE)
+				continue;
+			bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
+			if (bytenr == 0)
+				continue;
+
+			nr_file_extents++;
+
+			ret = process_func(trans, root, bytenr,
+					   orig_buf->start, buf->start,
+					   orig_root, ref_root,
+					   orig_generation, ref_generation,
+					   key.objectid);
+
+			if (ret) {
+				faili = i;
+				WARN_ON(1);
+				goto fail;
+			}
+		} else {
+			bytenr = btrfs_node_blockptr(buf, i);
+			ret = process_func(trans, root, bytenr,
+					   orig_buf->start, buf->start,
+					   orig_root, ref_root,
+					   orig_generation, ref_generation,
+					   level - 1);
+			if (ret) {
+				faili = i;
+				WARN_ON(1);
+				goto fail;
+			}
+		}
+	}
+out:
+	if (nr_extents) {
+		if (level == 0)
+			*nr_extents = nr_file_extents;
+		else
+			*nr_extents = nritems;
+	}
+	return 0;
+fail:
+	WARN_ON(1);
+	return ret;
+}
+
+int btrfs_update_ref(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root, struct extent_buffer *orig_buf,
+		     struct extent_buffer *buf, int start_slot, int nr)
+
+{
+	u64 bytenr;
+	u64 ref_root;
+	u64 orig_root;
+	u64 ref_generation;
+	u64 orig_generation;
+	struct btrfs_key key;
+	struct btrfs_file_extent_item *fi;
+	int i;
+	int ret;
+	int slot;
+	int level;
+
+	BUG_ON(start_slot < 0);
+	BUG_ON(start_slot + nr > btrfs_header_nritems(buf));
+
+	ref_root = btrfs_header_owner(buf);
+	ref_generation = btrfs_header_generation(buf);
+	orig_root = btrfs_header_owner(orig_buf);
+	orig_generation = btrfs_header_generation(orig_buf);
+	level = btrfs_header_level(buf);
+
+	if (!root->ref_cows) {
+		if (level == 0 &&
+		    root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+			return 0;
+		if (level != 0 &&
+		    root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID)
+			return 0;
+	}
+
+	for (i = 0, slot = start_slot; i < nr; i++, slot++) {
+		cond_resched();
+		if (level == 0) {
+			btrfs_item_key_to_cpu(buf, &key, slot);
+			if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
+				continue;
+			fi = btrfs_item_ptr(buf, slot,
+					    struct btrfs_file_extent_item);
+			if (btrfs_file_extent_type(buf, fi) ==
+			    BTRFS_FILE_EXTENT_INLINE)
+				continue;
+			bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
+			if (bytenr == 0)
+				continue;
+			ret = __btrfs_update_extent_ref(trans, root, bytenr,
+					    orig_buf->start, buf->start,
+					    orig_root, ref_root,
+					    orig_generation, ref_generation,
+					    key.objectid);
+			if (ret)
+				goto fail;
+		} else {
+			bytenr = btrfs_node_blockptr(buf, slot);
+			ret = __btrfs_update_extent_ref(trans, root, bytenr,
+					    orig_buf->start, buf->start,
+					    orig_root, ref_root,
+					    orig_generation, ref_generation,
+					    level - 1);
+			if (ret)
+				goto fail;
+		}
+	}
+	return 0;
+fail:
+	WARN_ON(1);
+	return -1;
+}
+
+static int write_one_cache_group(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct btrfs_path *path,
+				 struct btrfs_block_group_cache *cache)
+{
+	int ret;
+	int pending_ret;
+	struct btrfs_root *extent_root = root->fs_info->extent_root;
+	unsigned long bi;
+	struct extent_buffer *leaf;
+
+	ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1);
+	if (ret < 0)
+		goto fail;
+	BUG_ON(ret);
+
+	leaf = path->nodes[0];
+	bi = btrfs_item_ptr_offset(leaf, path->slots[0]);
+	write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item));
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_release_path(extent_root, path);
+fail:
+	finish_current_insert(trans, extent_root, 0);
+	pending_ret = del_pending_extents(trans, extent_root, 0);
+	if (ret)
+		return ret;
+	if (pending_ret)
+		return pending_ret;
+	return 0;
+
+}
+
+int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root)
+{
+	struct btrfs_block_group_cache *cache, *entry;
+	struct rb_node *n;
+	int err = 0;
+	int werr = 0;
+	struct btrfs_path *path;
+	u64 last = 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	while (1) {
+		cache = NULL;
+		spin_lock(&root->fs_info->block_group_cache_lock);
+		for (n = rb_first(&root->fs_info->block_group_cache_tree);
+		     n; n = rb_next(n)) {
+			entry = rb_entry(n, struct btrfs_block_group_cache,
+					 cache_node);
+			if (entry->dirty) {
+				cache = entry;
+				break;
+			}
+		}
+		spin_unlock(&root->fs_info->block_group_cache_lock);
+
+		if (!cache)
+			break;
+
+		cache->dirty = 0;
+		last += cache->key.offset;
+
+		err = write_one_cache_group(trans, root,
+					    path, cache);
+		/*
+		 * if we fail to write the cache group, we want
+		 * to keep it marked dirty in hopes that a later
+		 * write will work
+		 */
+		if (err) {
+			werr = err;
+			continue;
+		}
+	}
+	btrfs_free_path(path);
+	return werr;
+}
+
+int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr)
+{
+	struct btrfs_block_group_cache *block_group;
+	int readonly = 0;
+
+	block_group = btrfs_lookup_block_group(root->fs_info, bytenr);
+	if (!block_group || block_group->ro)
+		readonly = 1;
+	if (block_group)
+		put_block_group(block_group);
+	return readonly;
+}
+
+static int update_space_info(struct btrfs_fs_info *info, u64 flags,
+			     u64 total_bytes, u64 bytes_used,
+			     struct btrfs_space_info **space_info)
+{
+	struct btrfs_space_info *found;
+
+	found = __find_space_info(info, flags);
+	if (found) {
+		spin_lock(&found->lock);
+		found->total_bytes += total_bytes;
+		found->bytes_used += bytes_used;
+		found->full = 0;
+		spin_unlock(&found->lock);
+		*space_info = found;
+		return 0;
+	}
+	found = kzalloc(sizeof(*found), GFP_NOFS);
+	if (!found)
+		return -ENOMEM;
+
+	list_add(&found->list, &info->space_info);
+	INIT_LIST_HEAD(&found->block_groups);
+	init_rwsem(&found->groups_sem);
+	spin_lock_init(&found->lock);
+	found->flags = flags;
+	found->total_bytes = total_bytes;
+	found->bytes_used = bytes_used;
+	found->bytes_pinned = 0;
+	found->bytes_reserved = 0;
+	found->bytes_readonly = 0;
+	found->full = 0;
+	found->force_alloc = 0;
+	*space_info = found;
+	return 0;
+}
+
+static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
+{
+	u64 extra_flags = flags & (BTRFS_BLOCK_GROUP_RAID0 |
+				   BTRFS_BLOCK_GROUP_RAID1 |
+				   BTRFS_BLOCK_GROUP_RAID10 |
+				   BTRFS_BLOCK_GROUP_DUP);
+	if (extra_flags) {
+		if (flags & BTRFS_BLOCK_GROUP_DATA)
+			fs_info->avail_data_alloc_bits |= extra_flags;
+		if (flags & BTRFS_BLOCK_GROUP_METADATA)
+			fs_info->avail_metadata_alloc_bits |= extra_flags;
+		if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
+			fs_info->avail_system_alloc_bits |= extra_flags;
+	}
+}
+
+static void set_block_group_readonly(struct btrfs_block_group_cache *cache)
+{
+	spin_lock(&cache->space_info->lock);
+	spin_lock(&cache->lock);
+	if (!cache->ro) {
+		cache->space_info->bytes_readonly += cache->key.offset -
+					btrfs_block_group_used(&cache->item);
+		cache->ro = 1;
+	}
+	spin_unlock(&cache->lock);
+	spin_unlock(&cache->space_info->lock);
+}
+
+u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
+{
+	u64 num_devices = root->fs_info->fs_devices->rw_devices;
+
+	if (num_devices == 1)
+		flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0);
+	if (num_devices < 4)
+		flags &= ~BTRFS_BLOCK_GROUP_RAID10;
+
+	if ((flags & BTRFS_BLOCK_GROUP_DUP) &&
+	    (flags & (BTRFS_BLOCK_GROUP_RAID1 |
+		      BTRFS_BLOCK_GROUP_RAID10))) {
+		flags &= ~BTRFS_BLOCK_GROUP_DUP;
+	}
+
+	if ((flags & BTRFS_BLOCK_GROUP_RAID1) &&
+	    (flags & BTRFS_BLOCK_GROUP_RAID10)) {
+		flags &= ~BTRFS_BLOCK_GROUP_RAID1;
+	}
+
+	if ((flags & BTRFS_BLOCK_GROUP_RAID0) &&
+	    ((flags & BTRFS_BLOCK_GROUP_RAID1) |
+	     (flags & BTRFS_BLOCK_GROUP_RAID10) |
+	     (flags & BTRFS_BLOCK_GROUP_DUP)))
+		flags &= ~BTRFS_BLOCK_GROUP_RAID0;
+	return flags;
+}
+
+static int do_chunk_alloc(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *extent_root, u64 alloc_bytes,
+			  u64 flags, int force)
+{
+	struct btrfs_space_info *space_info;
+	u64 thresh;
+	int ret = 0;
+
+	mutex_lock(&extent_root->fs_info->chunk_mutex);
+
+	flags = btrfs_reduce_alloc_profile(extent_root, flags);
+
+	space_info = __find_space_info(extent_root->fs_info, flags);
+	if (!space_info) {
+		ret = update_space_info(extent_root->fs_info, flags,
+					0, 0, &space_info);
+		BUG_ON(ret);
+	}
+	BUG_ON(!space_info);
+
+	spin_lock(&space_info->lock);
+	if (space_info->force_alloc) {
+		force = 1;
+		space_info->force_alloc = 0;
+	}
+	if (space_info->full) {
+		spin_unlock(&space_info->lock);
+		goto out;
+	}
+
+	thresh = space_info->total_bytes - space_info->bytes_readonly;
+	thresh = div_factor(thresh, 6);
+	if (!force &&
+	   (space_info->bytes_used + space_info->bytes_pinned +
+	    space_info->bytes_reserved + alloc_bytes) < thresh) {
+		spin_unlock(&space_info->lock);
+		goto out;
+	}
+	spin_unlock(&space_info->lock);
+
+	ret = btrfs_alloc_chunk(trans, extent_root, flags);
+	if (ret)
+		space_info->full = 1;
+out:
+	mutex_unlock(&extent_root->fs_info->chunk_mutex);
+	return ret;
+}
+
+static int update_block_group(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      u64 bytenr, u64 num_bytes, int alloc,
+			      int mark_free)
+{
+	struct btrfs_block_group_cache *cache;
+	struct btrfs_fs_info *info = root->fs_info;
+	u64 total = num_bytes;
+	u64 old_val;
+	u64 byte_in_group;
+
+	while (total) {
+		cache = btrfs_lookup_block_group(info, bytenr);
+		if (!cache)
+			return -1;
+		byte_in_group = bytenr - cache->key.objectid;
+		WARN_ON(byte_in_group > cache->key.offset);
+
+		spin_lock(&cache->space_info->lock);
+		spin_lock(&cache->lock);
+		cache->dirty = 1;
+		old_val = btrfs_block_group_used(&cache->item);
+		num_bytes = min(total, cache->key.offset - byte_in_group);
+		if (alloc) {
+			old_val += num_bytes;
+			cache->space_info->bytes_used += num_bytes;
+			if (cache->ro)
+				cache->space_info->bytes_readonly -= num_bytes;
+			btrfs_set_block_group_used(&cache->item, old_val);
+			spin_unlock(&cache->lock);
+			spin_unlock(&cache->space_info->lock);
+		} else {
+			old_val -= num_bytes;
+			cache->space_info->bytes_used -= num_bytes;
+			if (cache->ro)
+				cache->space_info->bytes_readonly += num_bytes;
+			btrfs_set_block_group_used(&cache->item, old_val);
+			spin_unlock(&cache->lock);
+			spin_unlock(&cache->space_info->lock);
+			if (mark_free) {
+				int ret;
+
+				ret = btrfs_discard_extent(root, bytenr,
+							   num_bytes);
+				WARN_ON(ret);
+
+				ret = btrfs_add_free_space(cache, bytenr,
+							   num_bytes);
+				WARN_ON(ret);
+			}
+		}
+		put_block_group(cache);
+		total -= num_bytes;
+		bytenr += num_bytes;
+	}
+	return 0;
+}
+
+static u64 first_logical_byte(struct btrfs_root *root, u64 search_start)
+{
+	struct btrfs_block_group_cache *cache;
+	u64 bytenr;
+
+	cache = btrfs_lookup_first_block_group(root->fs_info, search_start);
+	if (!cache)
+		return 0;
+
+	bytenr = cache->key.objectid;
+	put_block_group(cache);
+
+	return bytenr;
+}
+
+int btrfs_update_pinned_extents(struct btrfs_root *root,
+				u64 bytenr, u64 num, int pin)
+{
+	u64 len;
+	struct btrfs_block_group_cache *cache;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	WARN_ON(!mutex_is_locked(&root->fs_info->pinned_mutex));
+	if (pin) {
+		set_extent_dirty(&fs_info->pinned_extents,
+				bytenr, bytenr + num - 1, GFP_NOFS);
+	} else {
+		clear_extent_dirty(&fs_info->pinned_extents,
+				bytenr, bytenr + num - 1, GFP_NOFS);
+	}
+	while (num > 0) {
+		cache = btrfs_lookup_block_group(fs_info, bytenr);
+		BUG_ON(!cache);
+		len = min(num, cache->key.offset -
+			  (bytenr - cache->key.objectid));
+		if (pin) {
+			spin_lock(&cache->space_info->lock);
+			spin_lock(&cache->lock);
+			cache->pinned += len;
+			cache->space_info->bytes_pinned += len;
+			spin_unlock(&cache->lock);
+			spin_unlock(&cache->space_info->lock);
+			fs_info->total_pinned += len;
+		} else {
+			spin_lock(&cache->space_info->lock);
+			spin_lock(&cache->lock);
+			cache->pinned -= len;
+			cache->space_info->bytes_pinned -= len;
+			spin_unlock(&cache->lock);
+			spin_unlock(&cache->space_info->lock);
+			fs_info->total_pinned -= len;
+			if (cache->cached)
+				btrfs_add_free_space(cache, bytenr, len);
+		}
+		put_block_group(cache);
+		bytenr += len;
+		num -= len;
+	}
+	return 0;
+}
+
+static int update_reserved_extents(struct btrfs_root *root,
+				   u64 bytenr, u64 num, int reserve)
+{
+	u64 len;
+	struct btrfs_block_group_cache *cache;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	while (num > 0) {
+		cache = btrfs_lookup_block_group(fs_info, bytenr);
+		BUG_ON(!cache);
+		len = min(num, cache->key.offset -
+			  (bytenr - cache->key.objectid));
+
+		spin_lock(&cache->space_info->lock);
+		spin_lock(&cache->lock);
+		if (reserve) {
+			cache->reserved += len;
+			cache->space_info->bytes_reserved += len;
+		} else {
+			cache->reserved -= len;
+			cache->space_info->bytes_reserved -= len;
+		}
+		spin_unlock(&cache->lock);
+		spin_unlock(&cache->space_info->lock);
+		put_block_group(cache);
+		bytenr += len;
+		num -= len;
+	}
+	return 0;
+}
+
+int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy)
+{
+	u64 last = 0;
+	u64 start;
+	u64 end;
+	struct extent_io_tree *pinned_extents = &root->fs_info->pinned_extents;
+	int ret;
+
+	mutex_lock(&root->fs_info->pinned_mutex);
+	while (1) {
+		ret = find_first_extent_bit(pinned_extents, last,
+					    &start, &end, EXTENT_DIRTY);
+		if (ret)
+			break;
+		set_extent_dirty(copy, start, end, GFP_NOFS);
+		last = end + 1;
+	}
+	mutex_unlock(&root->fs_info->pinned_mutex);
+	return 0;
+}
+
+int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct extent_io_tree *unpin)
+{
+	u64 start;
+	u64 end;
+	int ret;
+
+	mutex_lock(&root->fs_info->pinned_mutex);
+	while (1) {
+		ret = find_first_extent_bit(unpin, 0, &start, &end,
+					    EXTENT_DIRTY);
+		if (ret)
+			break;
+
+		ret = btrfs_discard_extent(root, start, end + 1 - start);
+
+		btrfs_update_pinned_extents(root, start, end + 1 - start, 0);
+		clear_extent_dirty(unpin, start, end, GFP_NOFS);
+
+		if (need_resched()) {
+			mutex_unlock(&root->fs_info->pinned_mutex);
+			cond_resched();
+			mutex_lock(&root->fs_info->pinned_mutex);
+		}
+	}
+	mutex_unlock(&root->fs_info->pinned_mutex);
+	return ret;
+}
+
+static int finish_current_insert(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *extent_root, int all)
+{
+	u64 start;
+	u64 end;
+	u64 priv;
+	u64 search = 0;
+	u64 skipped = 0;
+	struct btrfs_fs_info *info = extent_root->fs_info;
+	struct btrfs_path *path;
+	struct pending_extent_op *extent_op, *tmp;
+	struct list_head insert_list, update_list;
+	int ret;
+	int num_inserts = 0, max_inserts;
+
+	path = btrfs_alloc_path();
+	INIT_LIST_HEAD(&insert_list);
+	INIT_LIST_HEAD(&update_list);
+
+	max_inserts = extent_root->leafsize /
+		(2 * sizeof(struct btrfs_key) + 2 * sizeof(struct btrfs_item) +
+		 sizeof(struct btrfs_extent_ref) +
+		 sizeof(struct btrfs_extent_item));
+again:
+	mutex_lock(&info->extent_ins_mutex);
+	while (1) {
+		ret = find_first_extent_bit(&info->extent_ins, search, &start,
+					    &end, EXTENT_WRITEBACK);
+		if (ret) {
+			if (skipped && all && !num_inserts) {
+				skipped = 0;
+				search = 0;
+				continue;
+			}
+			mutex_unlock(&info->extent_ins_mutex);
+			break;
+		}
+
+		ret = try_lock_extent(&info->extent_ins, start, end, GFP_NOFS);
+		if (!ret) {
+			skipped = 1;
+			search = end + 1;
+			if (need_resched()) {
+				mutex_unlock(&info->extent_ins_mutex);
+				cond_resched();
+				mutex_lock(&info->extent_ins_mutex);
+			}
+			continue;
+		}
+
+		ret = get_state_private(&info->extent_ins, start, &priv);
+		BUG_ON(ret);
+		extent_op = (struct pending_extent_op *)(unsigned long) priv;
+
+		if (extent_op->type == PENDING_EXTENT_INSERT) {
+			num_inserts++;
+			list_add_tail(&extent_op->list, &insert_list);
+			search = end + 1;
+			if (num_inserts == max_inserts) {
+				mutex_unlock(&info->extent_ins_mutex);
+				break;
+			}
+		} else if (extent_op->type == PENDING_BACKREF_UPDATE) {
+			list_add_tail(&extent_op->list, &update_list);
+			search = end + 1;
+		} else {
+			BUG();
+		}
+	}
+
+	/*
+	 * process the update list, clear the writeback bit for it, and if
+	 * somebody marked this thing for deletion then just unlock it and be
+	 * done, the free_extents will handle it
+	 */
+	mutex_lock(&info->extent_ins_mutex);
+	list_for_each_entry_safe(extent_op, tmp, &update_list, list) {
+		clear_extent_bits(&info->extent_ins, extent_op->bytenr,
+				  extent_op->bytenr + extent_op->num_bytes - 1,
+				  EXTENT_WRITEBACK, GFP_NOFS);
+		if (extent_op->del) {
+			list_del_init(&extent_op->list);
+			unlock_extent(&info->extent_ins, extent_op->bytenr,
+				      extent_op->bytenr + extent_op->num_bytes
+				      - 1, GFP_NOFS);
+			kfree(extent_op);
+		}
+	}
+	mutex_unlock(&info->extent_ins_mutex);
+
+	/*
+	 * still have things left on the update list, go ahead an update
+	 * everything
+	 */
+	if (!list_empty(&update_list)) {
+		ret = update_backrefs(trans, extent_root, path, &update_list);
+		BUG_ON(ret);
+	}
+
+	/*
+	 * if no inserts need to be done, but we skipped some extents and we
+	 * need to make sure everything is cleaned then reset everything and
+	 * go back to the beginning
+	 */
+	if (!num_inserts && all && skipped) {
+		search = 0;
+		skipped = 0;
+		INIT_LIST_HEAD(&update_list);
+		INIT_LIST_HEAD(&insert_list);
+		goto again;
+	} else if (!num_inserts) {
+		goto out;
+	}
+
+	/*
+	 * process the insert extents list.  Again if we are deleting this
+	 * extent, then just unlock it, pin down the bytes if need be, and be
+	 * done with it.  Saves us from having to actually insert the extent
+	 * into the tree and then subsequently come along and delete it
+	 */
+	mutex_lock(&info->extent_ins_mutex);
+	list_for_each_entry_safe(extent_op, tmp, &insert_list, list) {
+		clear_extent_bits(&info->extent_ins, extent_op->bytenr,
+				  extent_op->bytenr + extent_op->num_bytes - 1,
+				  EXTENT_WRITEBACK, GFP_NOFS);
+		if (extent_op->del) {
+			u64 used;
+			list_del_init(&extent_op->list);
+			unlock_extent(&info->extent_ins, extent_op->bytenr,
+				      extent_op->bytenr + extent_op->num_bytes
+				      - 1, GFP_NOFS);
+
+			mutex_lock(&extent_root->fs_info->pinned_mutex);
+			ret = pin_down_bytes(trans, extent_root,
+					     extent_op->bytenr,
+					     extent_op->num_bytes, 0);
+			mutex_unlock(&extent_root->fs_info->pinned_mutex);
+
+			spin_lock(&info->delalloc_lock);
+			used = btrfs_super_bytes_used(&info->super_copy);
+			btrfs_set_super_bytes_used(&info->super_copy,
+					used - extent_op->num_bytes);
+			used = btrfs_root_used(&extent_root->root_item);
+			btrfs_set_root_used(&extent_root->root_item,
+					used - extent_op->num_bytes);
+			spin_unlock(&info->delalloc_lock);
+
+			ret = update_block_group(trans, extent_root,
+						 extent_op->bytenr,
+						 extent_op->num_bytes,
+						 0, ret > 0);
+			BUG_ON(ret);
+			kfree(extent_op);
+			num_inserts--;
+		}
+	}
+	mutex_unlock(&info->extent_ins_mutex);
+
+	ret = insert_extents(trans, extent_root, path, &insert_list,
+			     num_inserts);
+	BUG_ON(ret);
+
+	/*
+	 * if we broke out of the loop in order to insert stuff because we hit
+	 * the maximum number of inserts at a time we can handle, then loop
+	 * back and pick up where we left off
+	 */
+	if (num_inserts == max_inserts) {
+		INIT_LIST_HEAD(&insert_list);
+		INIT_LIST_HEAD(&update_list);
+		num_inserts = 0;
+		goto again;
+	}
+
+	/*
+	 * again, if we need to make absolutely sure there are no more pending
+	 * extent operations left and we know that we skipped some, go back to
+	 * the beginning and do it all again
+	 */
+	if (all && skipped) {
+		INIT_LIST_HEAD(&insert_list);
+		INIT_LIST_HEAD(&update_list);
+		search = 0;
+		skipped = 0;
+		num_inserts = 0;
+		goto again;
+	}
+out:
+	btrfs_free_path(path);
+	return 0;
+}
+
+static int pin_down_bytes(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root,
+			  u64 bytenr, u64 num_bytes, int is_data)
+{
+	int err = 0;
+	struct extent_buffer *buf;
+
+	if (is_data)
+		goto pinit;
+
+	buf = btrfs_find_tree_block(root, bytenr, num_bytes);
+	if (!buf)
+		goto pinit;
+
+	/* we can reuse a block if it hasn't been written
+	 * and it is from this transaction.  We can't
+	 * reuse anything from the tree log root because
+	 * it has tiny sub-transactions.
+	 */
+	if (btrfs_buffer_uptodate(buf, 0) &&
+	    btrfs_try_tree_lock(buf)) {
+		u64 header_owner = btrfs_header_owner(buf);
+		u64 header_transid = btrfs_header_generation(buf);
+		if (header_owner != BTRFS_TREE_LOG_OBJECTID &&
+		    header_owner != BTRFS_TREE_RELOC_OBJECTID &&
+		    header_transid == trans->transid &&
+		    !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
+			clean_tree_block(NULL, root, buf);
+			btrfs_tree_unlock(buf);
+			free_extent_buffer(buf);
+			return 1;
+		}
+		btrfs_tree_unlock(buf);
+	}
+	free_extent_buffer(buf);
+pinit:
+	btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);
+
+	BUG_ON(err < 0);
+	return 0;
+}
+
+/*
+ * remove an extent from the root, returns 0 on success
+ */
+static int __free_extent(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *root,
+			 u64 bytenr, u64 num_bytes, u64 parent,
+			 u64 root_objectid, u64 ref_generation,
+			 u64 owner_objectid, int pin, int mark_free)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_fs_info *info = root->fs_info;
+	struct btrfs_root *extent_root = info->extent_root;
+	struct extent_buffer *leaf;
+	int ret;
+	int extent_slot = 0;
+	int found_extent = 0;
+	int num_to_del = 1;
+	struct btrfs_extent_item *ei;
+	u32 refs;
+
+	key.objectid = bytenr;
+	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
+	key.offset = num_bytes;
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->reada = 1;
+	ret = lookup_extent_backref(trans, extent_root, path,
+				    bytenr, parent, root_objectid,
+				    ref_generation, owner_objectid, 1);
+	if (ret == 0) {
+		struct btrfs_key found_key;
+		extent_slot = path->slots[0];
+		while (extent_slot > 0) {
+			extent_slot--;
+			btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+					      extent_slot);
+			if (found_key.objectid != bytenr)
+				break;
+			if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
+			    found_key.offset == num_bytes) {
+				found_extent = 1;
+				break;
+			}
+			if (path->slots[0] - extent_slot > 5)
+				break;
+		}
+		if (!found_extent) {
+			ret = remove_extent_backref(trans, extent_root, path);
+			BUG_ON(ret);
+			btrfs_release_path(extent_root, path);
+			ret = btrfs_search_slot(trans, extent_root,
+						&key, path, -1, 1);
+			if (ret) {
+				printk(KERN_ERR "umm, got %d back from search"
+				       ", was looking for %llu\n", ret,
+				       (unsigned long long)bytenr);
+				btrfs_print_leaf(extent_root, path->nodes[0]);
+			}
+			BUG_ON(ret);
+			extent_slot = path->slots[0];
+		}
+	} else {
+		btrfs_print_leaf(extent_root, path->nodes[0]);
+		WARN_ON(1);
+		printk(KERN_ERR "btrfs unable to find ref byte nr %llu "
+		       "root %llu gen %llu owner %llu\n",
+		       (unsigned long long)bytenr,
+		       (unsigned long long)root_objectid,
+		       (unsigned long long)ref_generation,
+		       (unsigned long long)owner_objectid);
+	}
+
+	leaf = path->nodes[0];
+	ei = btrfs_item_ptr(leaf, extent_slot,
+			    struct btrfs_extent_item);
+	refs = btrfs_extent_refs(leaf, ei);
+	BUG_ON(refs == 0);
+	refs -= 1;
+	btrfs_set_extent_refs(leaf, ei, refs);
+
+	btrfs_mark_buffer_dirty(leaf);
+
+	if (refs == 0 && found_extent && path->slots[0] == extent_slot + 1) {
+		struct btrfs_extent_ref *ref;
+		ref = btrfs_item_ptr(leaf, path->slots[0],
+				     struct btrfs_extent_ref);
+		BUG_ON(btrfs_ref_num_refs(leaf, ref) != 1);
+		/* if the back ref and the extent are next to each other
+		 * they get deleted below in one shot
+		 */
+		path->slots[0] = extent_slot;
+		num_to_del = 2;
+	} else if (found_extent) {
+		/* otherwise delete the extent back ref */
+		ret = remove_extent_backref(trans, extent_root, path);
+		BUG_ON(ret);
+		/* if refs are 0, we need to setup the path for deletion */
+		if (refs == 0) {
+			btrfs_release_path(extent_root, path);
+			ret = btrfs_search_slot(trans, extent_root, &key, path,
+						-1, 1);
+			BUG_ON(ret);
+		}
+	}
+
+	if (refs == 0) {
+		u64 super_used;
+		u64 root_used;
+
+		if (pin) {
+			mutex_lock(&root->fs_info->pinned_mutex);
+			ret = pin_down_bytes(trans, root, bytenr, num_bytes,
+				owner_objectid >= BTRFS_FIRST_FREE_OBJECTID);
+			mutex_unlock(&root->fs_info->pinned_mutex);
+			if (ret > 0)
+				mark_free = 1;
+			BUG_ON(ret < 0);
+		}
+		/* block accounting for super block */
+		spin_lock(&info->delalloc_lock);
+		super_used = btrfs_super_bytes_used(&info->super_copy);
+		btrfs_set_super_bytes_used(&info->super_copy,
+					   super_used - num_bytes);
+
+		/* block accounting for root item */
+		root_used = btrfs_root_used(&root->root_item);
+		btrfs_set_root_used(&root->root_item,
+					   root_used - num_bytes);
+		spin_unlock(&info->delalloc_lock);
+		ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
+				      num_to_del);
+		BUG_ON(ret);
+		btrfs_release_path(extent_root, path);
+
+		if (owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
+			ret = btrfs_del_csums(trans, root, bytenr, num_bytes);
+			BUG_ON(ret);
+		}
+
+		ret = update_block_group(trans, root, bytenr, num_bytes, 0,
+					 mark_free);
+		BUG_ON(ret);
+	}
+	btrfs_free_path(path);
+	finish_current_insert(trans, extent_root, 0);
+	return ret;
+}
+
+/*
+ * find all the blocks marked as pending in the radix tree and remove
+ * them from the extent map
+ */
+static int del_pending_extents(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *extent_root, int all)
+{
+	int ret;
+	int err = 0;
+	u64 start;
+	u64 end;
+	u64 priv;
+	u64 search = 0;
+	int nr = 0, skipped = 0;
+	struct extent_io_tree *pending_del;
+	struct extent_io_tree *extent_ins;
+	struct pending_extent_op *extent_op;
+	struct btrfs_fs_info *info = extent_root->fs_info;
+	struct list_head delete_list;
+
+	INIT_LIST_HEAD(&delete_list);
+	extent_ins = &extent_root->fs_info->extent_ins;
+	pending_del = &extent_root->fs_info->pending_del;
+
+again:
+	mutex_lock(&info->extent_ins_mutex);
+	while (1) {
+		ret = find_first_extent_bit(pending_del, search, &start, &end,
+					    EXTENT_WRITEBACK);
+		if (ret) {
+			if (all && skipped && !nr) {
+				search = 0;
+				continue;
+			}
+			mutex_unlock(&info->extent_ins_mutex);
+			break;
+		}
+
+		ret = try_lock_extent(extent_ins, start, end, GFP_NOFS);
+		if (!ret) {
+			search = end+1;
+			skipped = 1;
+
+			if (need_resched()) {
+				mutex_unlock(&info->extent_ins_mutex);
+				cond_resched();
+				mutex_lock(&info->extent_ins_mutex);
+			}
+
+			continue;
+		}
+		BUG_ON(ret < 0);
+
+		ret = get_state_private(pending_del, start, &priv);
+		BUG_ON(ret);
+		extent_op = (struct pending_extent_op *)(unsigned long)priv;
+
+		clear_extent_bits(pending_del, start, end, EXTENT_WRITEBACK,
+				  GFP_NOFS);
+		if (!test_range_bit(extent_ins, start, end,
+				    EXTENT_WRITEBACK, 0)) {
+			list_add_tail(&extent_op->list, &delete_list);
+			nr++;
+		} else {
+			kfree(extent_op);
+
+			ret = get_state_private(&info->extent_ins, start,
+						&priv);
+			BUG_ON(ret);
+			extent_op = (struct pending_extent_op *)
+						(unsigned long)priv;
+
+			clear_extent_bits(&info->extent_ins, start, end,
+					  EXTENT_WRITEBACK, GFP_NOFS);
+
+			if (extent_op->type == PENDING_BACKREF_UPDATE) {
+				list_add_tail(&extent_op->list, &delete_list);
+				search = end + 1;
+				nr++;
+				continue;
+			}
+
+			mutex_lock(&extent_root->fs_info->pinned_mutex);
+			ret = pin_down_bytes(trans, extent_root, start,
+					     end + 1 - start, 0);
+			mutex_unlock(&extent_root->fs_info->pinned_mutex);
+
+			ret = update_block_group(trans, extent_root, start,
+						end + 1 - start, 0, ret > 0);
+
+			unlock_extent(extent_ins, start, end, GFP_NOFS);
+			BUG_ON(ret);
+			kfree(extent_op);
+		}
+		if (ret)
+			err = ret;
+
+		search = end + 1;
+
+		if (need_resched()) {
+			mutex_unlock(&info->extent_ins_mutex);
+			cond_resched();
+			mutex_lock(&info->extent_ins_mutex);
+		}
+	}
+
+	if (nr) {
+		ret = free_extents(trans, extent_root, &delete_list);
+		BUG_ON(ret);
+	}
+
+	if (all && skipped) {
+		INIT_LIST_HEAD(&delete_list);
+		search = 0;
+		nr = 0;
+		goto again;
+	}
+
+	return err;
+}
+
+/*
+ * remove an extent from the root, returns 0 on success
+ */
+static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       u64 bytenr, u64 num_bytes, u64 parent,
+			       u64 root_objectid, u64 ref_generation,
+			       u64 owner_objectid, int pin)
+{
+	struct btrfs_root *extent_root = root->fs_info->extent_root;
+	int pending_ret;
+	int ret;
+
+	WARN_ON(num_bytes < root->sectorsize);
+	if (root == extent_root) {
+		struct pending_extent_op *extent_op = NULL;
+
+		mutex_lock(&root->fs_info->extent_ins_mutex);
+		if (test_range_bit(&root->fs_info->extent_ins, bytenr,
+				bytenr + num_bytes - 1, EXTENT_WRITEBACK, 0)) {
+			u64 priv;
+			ret = get_state_private(&root->fs_info->extent_ins,
+						bytenr, &priv);
+			BUG_ON(ret);
+			extent_op = (struct pending_extent_op *)
+						(unsigned long)priv;
+
+			extent_op->del = 1;
+			if (extent_op->type == PENDING_EXTENT_INSERT) {
+				mutex_unlock(&root->fs_info->extent_ins_mutex);
+				return 0;
+			}
+		}
+
+		if (extent_op) {
+			ref_generation = extent_op->orig_generation;
+			parent = extent_op->orig_parent;
+		}
+
+		extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
+		BUG_ON(!extent_op);
+
+		extent_op->type = PENDING_EXTENT_DELETE;
+		extent_op->bytenr = bytenr;
+		extent_op->num_bytes = num_bytes;
+		extent_op->parent = parent;
+		extent_op->orig_parent = parent;
+		extent_op->generation = ref_generation;
+		extent_op->orig_generation = ref_generation;
+		extent_op->level = (int)owner_objectid;
+		INIT_LIST_HEAD(&extent_op->list);
+		extent_op->del = 0;
+
+		set_extent_bits(&root->fs_info->pending_del,
+				bytenr, bytenr + num_bytes - 1,
+				EXTENT_WRITEBACK, GFP_NOFS);
+		set_state_private(&root->fs_info->pending_del,
+				  bytenr, (unsigned long)extent_op);
+		mutex_unlock(&root->fs_info->extent_ins_mutex);
+		return 0;
+	}
+	/* if metadata always pin */
+	if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID) {
+		if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
+			struct btrfs_block_group_cache *cache;
+
+			/* btrfs_free_reserved_extent */
+			cache = btrfs_lookup_block_group(root->fs_info, bytenr);
+			BUG_ON(!cache);
+			btrfs_add_free_space(cache, bytenr, num_bytes);
+			put_block_group(cache);
+			update_reserved_extents(root, bytenr, num_bytes, 0);
+			return 0;
+		}
+		pin = 1;
+	}
+
+	/* if data pin when any transaction has committed this */
+	if (ref_generation != trans->transid)
+		pin = 1;
+
+	ret = __free_extent(trans, root, bytenr, num_bytes, parent,
+			    root_objectid, ref_generation,
+			    owner_objectid, pin, pin == 0);
+
+	finish_current_insert(trans, root->fs_info->extent_root, 0);
+	pending_ret = del_pending_extents(trans, root->fs_info->extent_root, 0);
+	return ret ? ret : pending_ret;
+}
+
+int btrfs_free_extent(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root,
+		      u64 bytenr, u64 num_bytes, u64 parent,
+		      u64 root_objectid, u64 ref_generation,
+		      u64 owner_objectid, int pin)
+{
+	int ret;
+
+	ret = __btrfs_free_extent(trans, root, bytenr, num_bytes, parent,
+				  root_objectid, ref_generation,
+				  owner_objectid, pin);
+	return ret;
+}
+
+static u64 stripe_align(struct btrfs_root *root, u64 val)
+{
+	u64 mask = ((u64)root->stripesize - 1);
+	u64 ret = (val + mask) & ~mask;
+	return ret;
+}
+
+/*
+ * walks the btree of allocated extents and find a hole of a given size.
+ * The key ins is changed to record the hole:
+ * ins->objectid == block start
+ * ins->flags = BTRFS_EXTENT_ITEM_KEY
+ * ins->offset == number of blocks
+ * Any available blocks before search_start are skipped.
+ */
+static noinline int find_free_extent(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *orig_root,
+				     u64 num_bytes, u64 empty_size,
+				     u64 search_start, u64 search_end,
+				     u64 hint_byte, struct btrfs_key *ins,
+				     u64 exclude_start, u64 exclude_nr,
+				     int data)
+{
+	int ret = 0;
+	struct btrfs_root *root = orig_root->fs_info->extent_root;
+	u64 total_needed = num_bytes;
+	u64 *last_ptr = NULL;
+	u64 last_wanted = 0;
+	struct btrfs_block_group_cache *block_group = NULL;
+	int chunk_alloc_done = 0;
+	int empty_cluster = 2 * 1024 * 1024;
+	int allowed_chunk_alloc = 0;
+	struct list_head *head = NULL, *cur = NULL;
+	int loop = 0;
+	int extra_loop = 0;
+	struct btrfs_space_info *space_info;
+
+	WARN_ON(num_bytes < root->sectorsize);
+	btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
+	ins->objectid = 0;
+	ins->offset = 0;
+
+	if (orig_root->ref_cows || empty_size)
+		allowed_chunk_alloc = 1;
+
+	if (data & BTRFS_BLOCK_GROUP_METADATA) {
+		last_ptr = &root->fs_info->last_alloc;
+		empty_cluster = 64 * 1024;
+	}
+
+	if ((data & BTRFS_BLOCK_GROUP_DATA) && btrfs_test_opt(root, SSD))
+		last_ptr = &root->fs_info->last_data_alloc;
+
+	if (last_ptr) {
+		if (*last_ptr) {
+			hint_byte = *last_ptr;
+			last_wanted = *last_ptr;
+		} else
+			empty_size += empty_cluster;
+	} else {
+		empty_cluster = 0;
+	}
+	search_start = max(search_start, first_logical_byte(root, 0));
+	search_start = max(search_start, hint_byte);
+
+	if (last_wanted && search_start != last_wanted) {
+		last_wanted = 0;
+		empty_size += empty_cluster;
+	}
+
+	total_needed += empty_size;
+	block_group = btrfs_lookup_block_group(root->fs_info, search_start);
+	if (!block_group)
+		block_group = btrfs_lookup_first_block_group(root->fs_info,
+							     search_start);
+	space_info = __find_space_info(root->fs_info, data);
+
+	down_read(&space_info->groups_sem);
+	while (1) {
+		struct btrfs_free_space *free_space;
+		/*
+		 * the only way this happens if our hint points to a block
+		 * group thats not of the proper type, while looping this
+		 * should never happen
+		 */
+		if (empty_size)
+			extra_loop = 1;
+
+		if (!block_group)
+			goto new_group_no_lock;
+
+		if (unlikely(!block_group->cached)) {
+			mutex_lock(&block_group->cache_mutex);
+			ret = cache_block_group(root, block_group);
+			mutex_unlock(&block_group->cache_mutex);
+			if (ret)
+				break;
+		}
+
+		mutex_lock(&block_group->alloc_mutex);
+		if (unlikely(!block_group_bits(block_group, data)))
+			goto new_group;
+
+		if (unlikely(block_group->ro))
+			goto new_group;
+
+		free_space = btrfs_find_free_space(block_group, search_start,
+						   total_needed);
+		if (free_space) {
+			u64 start = block_group->key.objectid;
+			u64 end = block_group->key.objectid +
+				block_group->key.offset;
+
+			search_start = stripe_align(root, free_space->offset);
+
+			/* move on to the next group */
+			if (search_start + num_bytes >= search_end)
+				goto new_group;
+
+			/* move on to the next group */
+			if (search_start + num_bytes > end)
+				goto new_group;
+
+			if (last_wanted && search_start != last_wanted) {
+				total_needed += empty_cluster;
+				empty_size += empty_cluster;
+				last_wanted = 0;
+				/*
+				 * if search_start is still in this block group
+				 * then we just re-search this block group
+				 */
+				if (search_start >= start &&
+				    search_start < end) {
+					mutex_unlock(&block_group->alloc_mutex);
+					continue;
+				}
+
+				/* else we go to the next block group */
+				goto new_group;
+			}
+
+			if (exclude_nr > 0 &&
+			    (search_start + num_bytes > exclude_start &&
+			     search_start < exclude_start + exclude_nr)) {
+				search_start = exclude_start + exclude_nr;
+				/*
+				 * if search_start is still in this block group
+				 * then we just re-search this block group
+				 */
+				if (search_start >= start &&
+				    search_start < end) {
+					mutex_unlock(&block_group->alloc_mutex);
+					last_wanted = 0;
+					continue;
+				}
+
+				/* else we go to the next block group */
+				goto new_group;
+			}
+
+			ins->objectid = search_start;
+			ins->offset = num_bytes;
+
+			btrfs_remove_free_space_lock(block_group, search_start,
+						     num_bytes);
+			/* we are all good, lets return */
+			mutex_unlock(&block_group->alloc_mutex);
+			break;
+		}
+new_group:
+		mutex_unlock(&block_group->alloc_mutex);
+		put_block_group(block_group);
+		block_group = NULL;
+new_group_no_lock:
+		/* don't try to compare new allocations against the
+		 * last allocation any more
+		 */
+		last_wanted = 0;
+
+		/*
+		 * Here's how this works.
+		 * loop == 0: we were searching a block group via a hint
+		 *		and didn't find anything, so we start at
+		 *		the head of the block groups and keep searching
+		 * loop == 1: we're searching through all of the block groups
+		 *		if we hit the head again we have searched
+		 *		all of the block groups for this space and we
+		 *		need to try and allocate, if we cant error out.
+		 * loop == 2: we allocated more space and are looping through
+		 *		all of the block groups again.
+		 */
+		if (loop == 0) {
+			head = &space_info->block_groups;
+			cur = head->next;
+			loop++;
+		} else if (loop == 1 && cur == head) {
+			int keep_going;
+
+			/* at this point we give up on the empty_size
+			 * allocations and just try to allocate the min
+			 * space.
+			 *
+			 * The extra_loop field was set if an empty_size
+			 * allocation was attempted above, and if this
+			 * is try we need to try the loop again without
+			 * the additional empty_size.
+			 */
+			total_needed -= empty_size;
+			empty_size = 0;
+			keep_going = extra_loop;
+			loop++;
+
+			if (allowed_chunk_alloc && !chunk_alloc_done) {
+				up_read(&space_info->groups_sem);
+				ret = do_chunk_alloc(trans, root, num_bytes +
+						     2 * 1024 * 1024, data, 1);
+				down_read(&space_info->groups_sem);
+				if (ret < 0)
+					goto loop_check;
+				head = &space_info->block_groups;
+				/*
+				 * we've allocated a new chunk, keep
+				 * trying
+				 */
+				keep_going = 1;
+				chunk_alloc_done = 1;
+			} else if (!allowed_chunk_alloc) {
+				space_info->force_alloc = 1;
+			}
+loop_check:
+			if (keep_going) {
+				cur = head->next;
+				extra_loop = 0;
+			} else {
+				break;
+			}
+		} else if (cur == head) {
+			break;
+		}
+
+		block_group = list_entry(cur, struct btrfs_block_group_cache,
+					 list);
+		atomic_inc(&block_group->count);
+
+		search_start = block_group->key.objectid;
+		cur = cur->next;
+	}
+
+	/* we found what we needed */
+	if (ins->objectid) {
+		if (!(data & BTRFS_BLOCK_GROUP_DATA))
+			trans->block_group = block_group->key.objectid;
+
+		if (last_ptr)
+			*last_ptr = ins->objectid + ins->offset;
+		ret = 0;
+	} else if (!ret) {
+		printk(KERN_ERR "btrfs searching for %llu bytes, "
+		       "num_bytes %llu, loop %d, allowed_alloc %d\n",
+		       (unsigned long long)total_needed,
+		       (unsigned long long)num_bytes,
+		       loop, allowed_chunk_alloc);
+		ret = -ENOSPC;
+	}
+	if (block_group)
+		put_block_group(block_group);
+
+	up_read(&space_info->groups_sem);
+	return ret;
+}
+
+static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
+{
+	struct btrfs_block_group_cache *cache;
+	struct list_head *l;
+
+	printk(KERN_INFO "space_info has %llu free, is %sfull\n",
+	       (unsigned long long)(info->total_bytes - info->bytes_used -
+				    info->bytes_pinned - info->bytes_reserved),
+	       (info->full) ? "" : "not ");
+
+	down_read(&info->groups_sem);
+	list_for_each(l, &info->block_groups) {
+		cache = list_entry(l, struct btrfs_block_group_cache, list);
+		spin_lock(&cache->lock);
+		printk(KERN_INFO "block group %llu has %llu bytes, %llu used "
+		       "%llu pinned %llu reserved\n",
+		       (unsigned long long)cache->key.objectid,
+		       (unsigned long long)cache->key.offset,
+		       (unsigned long long)btrfs_block_group_used(&cache->item),
+		       (unsigned long long)cache->pinned,
+		       (unsigned long long)cache->reserved);
+		btrfs_dump_free_space(cache, bytes);
+		spin_unlock(&cache->lock);
+	}
+	up_read(&info->groups_sem);
+}
+
+static int __btrfs_reserve_extent(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  u64 num_bytes, u64 min_alloc_size,
+				  u64 empty_size, u64 hint_byte,
+				  u64 search_end, struct btrfs_key *ins,
+				  u64 data)
+{
+	int ret;
+	u64 search_start = 0;
+	u64 alloc_profile;
+	struct btrfs_fs_info *info = root->fs_info;
+
+	if (data) {
+		alloc_profile = info->avail_data_alloc_bits &
+			info->data_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
+	} else if (root == root->fs_info->chunk_root) {
+		alloc_profile = info->avail_system_alloc_bits &
+			info->system_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
+	} else {
+		alloc_profile = info->avail_metadata_alloc_bits &
+			info->metadata_alloc_profile;
+		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
+	}
+again:
+	data = btrfs_reduce_alloc_profile(root, data);
+	/*
+	 * the only place that sets empty_size is btrfs_realloc_node, which
+	 * is not called recursively on allocations
+	 */
+	if (empty_size || root->ref_cows) {
+		if (!(data & BTRFS_BLOCK_GROUP_METADATA)) {
+			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+				     2 * 1024 * 1024,
+				     BTRFS_BLOCK_GROUP_METADATA |
+				     (info->metadata_alloc_profile &
+				      info->avail_metadata_alloc_bits), 0);
+		}
+		ret = do_chunk_alloc(trans, root->fs_info->extent_root,
+				     num_bytes + 2 * 1024 * 1024, data, 0);
+	}
+
+	WARN_ON(num_bytes < root->sectorsize);
+	ret = find_free_extent(trans, root, num_bytes, empty_size,
+			       search_start, search_end, hint_byte, ins,
+			       trans->alloc_exclude_start,
+			       trans->alloc_exclude_nr, data);
+
+	if (ret == -ENOSPC && num_bytes > min_alloc_size) {
+		num_bytes = num_bytes >> 1;
+		num_bytes = num_bytes & ~(root->sectorsize - 1);
+		num_bytes = max(num_bytes, min_alloc_size);
+		do_chunk_alloc(trans, root->fs_info->extent_root,
+			       num_bytes, data, 1);
+		goto again;
+	}
+	if (ret) {
+		struct btrfs_space_info *sinfo;
+
+		sinfo = __find_space_info(root->fs_info, data);
+		printk(KERN_ERR "btrfs allocation failed flags %llu, "
+		       "wanted %llu\n", (unsigned long long)data,
+		       (unsigned long long)num_bytes);
+		dump_space_info(sinfo, num_bytes);
+		BUG();
+	}
+
+	return ret;
+}
+
+int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
+{
+	struct btrfs_block_group_cache *cache;
+	int ret = 0;
+
+	cache = btrfs_lookup_block_group(root->fs_info, start);
+	if (!cache) {
+		printk(KERN_ERR "Unable to find block group for %llu\n",
+		       (unsigned long long)start);
+		return -ENOSPC;
+	}
+
+	ret = btrfs_discard_extent(root, start, len);
+
+	btrfs_add_free_space(cache, start, len);
+	put_block_group(cache);
+	update_reserved_extents(root, start, len, 0);
+
+	return ret;
+}
+
+int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  u64 num_bytes, u64 min_alloc_size,
+				  u64 empty_size, u64 hint_byte,
+				  u64 search_end, struct btrfs_key *ins,
+				  u64 data)
+{
+	int ret;
+	ret = __btrfs_reserve_extent(trans, root, num_bytes, min_alloc_size,
+				     empty_size, hint_byte, search_end, ins,
+				     data);
+	update_reserved_extents(root, ins->objectid, ins->offset, 1);
+	return ret;
+}
+
+static int __btrfs_alloc_reserved_extent(struct btrfs_trans_handle *trans,
+					 struct btrfs_root *root, u64 parent,
+					 u64 root_objectid, u64 ref_generation,
+					 u64 owner, struct btrfs_key *ins)
+{
+	int ret;
+	int pending_ret;
+	u64 super_used;
+	u64 root_used;
+	u64 num_bytes = ins->offset;
+	u32 sizes[2];
+	struct btrfs_fs_info *info = root->fs_info;
+	struct btrfs_root *extent_root = info->extent_root;
+	struct btrfs_extent_item *extent_item;
+	struct btrfs_extent_ref *ref;
+	struct btrfs_path *path;
+	struct btrfs_key keys[2];
+
+	if (parent == 0)
+		parent = ins->objectid;
+
+	/* block accounting for super block */
+	spin_lock(&info->delalloc_lock);
+	super_used = btrfs_super_bytes_used(&info->super_copy);
+	btrfs_set_super_bytes_used(&info->super_copy, super_used + num_bytes);
+
+	/* block accounting for root item */
+	root_used = btrfs_root_used(&root->root_item);
+	btrfs_set_root_used(&root->root_item, root_used + num_bytes);
+	spin_unlock(&info->delalloc_lock);
+
+	if (root == extent_root) {
+		struct pending_extent_op *extent_op;
+
+		extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
+		BUG_ON(!extent_op);
+
+		extent_op->type = PENDING_EXTENT_INSERT;
+		extent_op->bytenr = ins->objectid;
+		extent_op->num_bytes = ins->offset;
+		extent_op->parent = parent;
+		extent_op->orig_parent = 0;
+		extent_op->generation = ref_generation;
+		extent_op->orig_generation = 0;
+		extent_op->level = (int)owner;
+		INIT_LIST_HEAD(&extent_op->list);
+		extent_op->del = 0;
+
+		mutex_lock(&root->fs_info->extent_ins_mutex);
+		set_extent_bits(&root->fs_info->extent_ins, ins->objectid,
+				ins->objectid + ins->offset - 1,
+				EXTENT_WRITEBACK, GFP_NOFS);
+		set_state_private(&root->fs_info->extent_ins,
+				  ins->objectid, (unsigned long)extent_op);
+		mutex_unlock(&root->fs_info->extent_ins_mutex);
+		goto update_block;
+	}
+
+	memcpy(&keys[0], ins, sizeof(*ins));
+	keys[1].objectid = ins->objectid;
+	keys[1].type = BTRFS_EXTENT_REF_KEY;
+	keys[1].offset = parent;
+	sizes[0] = sizeof(*extent_item);
+	sizes[1] = sizeof(*ref);
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	ret = btrfs_insert_empty_items(trans, extent_root, path, keys,
+				       sizes, 2);
+	BUG_ON(ret);
+
+	extent_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				     struct btrfs_extent_item);
+	btrfs_set_extent_refs(path->nodes[0], extent_item, 1);
+	ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1,
+			     struct btrfs_extent_ref);
+
+	btrfs_set_ref_root(path->nodes[0], ref, root_objectid);
+	btrfs_set_ref_generation(path->nodes[0], ref, ref_generation);
+	btrfs_set_ref_objectid(path->nodes[0], ref, owner);
+	btrfs_set_ref_num_refs(path->nodes[0], ref, 1);
+
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+
+	trans->alloc_exclude_start = 0;
+	trans->alloc_exclude_nr = 0;
+	btrfs_free_path(path);
+	finish_current_insert(trans, extent_root, 0);
+	pending_ret = del_pending_extents(trans, extent_root, 0);
+
+	if (ret)
+		goto out;
+	if (pending_ret) {
+		ret = pending_ret;
+		goto out;
+	}
+
+update_block:
+	ret = update_block_group(trans, root, ins->objectid,
+				 ins->offset, 1, 0);
+	if (ret) {
+		printk(KERN_ERR "btrfs update block group failed for %llu "
+		       "%llu\n", (unsigned long long)ins->objectid,
+		       (unsigned long long)ins->offset);
+		BUG();
+	}
+out:
+	return ret;
+}
+
+int btrfs_alloc_reserved_extent(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root, u64 parent,
+				u64 root_objectid, u64 ref_generation,
+				u64 owner, struct btrfs_key *ins)
+{
+	int ret;
+
+	if (root_objectid == BTRFS_TREE_LOG_OBJECTID)
+		return 0;
+	ret = __btrfs_alloc_reserved_extent(trans, root, parent, root_objectid,
+					    ref_generation, owner, ins);
+	update_reserved_extents(root, ins->objectid, ins->offset, 0);
+	return ret;
+}
+
+/*
+ * this is used by the tree logging recovery code.  It records that
+ * an extent has been allocated and makes sure to clear the free
+ * space cache bits as well
+ */
+int btrfs_alloc_logged_extent(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root, u64 parent,
+				u64 root_objectid, u64 ref_generation,
+				u64 owner, struct btrfs_key *ins)
+{
+	int ret;
+	struct btrfs_block_group_cache *block_group;
+
+	block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
+	mutex_lock(&block_group->cache_mutex);
+	cache_block_group(root, block_group);
+	mutex_unlock(&block_group->cache_mutex);
+
+	ret = btrfs_remove_free_space(block_group, ins->objectid,
+				      ins->offset);
+	BUG_ON(ret);
+	put_block_group(block_group);
+	ret = __btrfs_alloc_reserved_extent(trans, root, parent, root_objectid,
+					    ref_generation, owner, ins);
+	return ret;
+}
+
+/*
+ * finds a free extent and does all the dirty work required for allocation
+ * returns the key for the extent through ins, and a tree buffer for
+ * the first block of the extent through buf.
+ *
+ * returns 0 if everything worked, non-zero otherwise.
+ */
+int btrfs_alloc_extent(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root,
+		       u64 num_bytes, u64 parent, u64 min_alloc_size,
+		       u64 root_objectid, u64 ref_generation,
+		       u64 owner_objectid, u64 empty_size, u64 hint_byte,
+		       u64 search_end, struct btrfs_key *ins, u64 data)
+{
+	int ret;
+
+	ret = __btrfs_reserve_extent(trans, root, num_bytes,
+				     min_alloc_size, empty_size, hint_byte,
+				     search_end, ins, data);
+	BUG_ON(ret);
+	if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
+		ret = __btrfs_alloc_reserved_extent(trans, root, parent,
+					root_objectid, ref_generation,
+					owner_objectid, ins);
+		BUG_ON(ret);
+
+	} else {
+		update_reserved_extents(root, ins->objectid, ins->offset, 1);
+	}
+	return ret;
+}
+
+struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
+					    struct btrfs_root *root,
+					    u64 bytenr, u32 blocksize)
+{
+	struct extent_buffer *buf;
+
+	buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
+	if (!buf)
+		return ERR_PTR(-ENOMEM);
+	btrfs_set_header_generation(buf, trans->transid);
+	btrfs_tree_lock(buf);
+	clean_tree_block(trans, root, buf);
+	btrfs_set_buffer_uptodate(buf);
+	if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
+		set_extent_dirty(&root->dirty_log_pages, buf->start,
+			 buf->start + buf->len - 1, GFP_NOFS);
+	} else {
+		set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
+			 buf->start + buf->len - 1, GFP_NOFS);
+	}
+	trans->blocks_used++;
+	return buf;
+}
+
+/*
+ * helper function to allocate a block for a given tree
+ * returns the tree buffer or NULL.
+ */
+struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
+					     struct btrfs_root *root,
+					     u32 blocksize, u64 parent,
+					     u64 root_objectid,
+					     u64 ref_generation,
+					     int level,
+					     u64 hint,
+					     u64 empty_size)
+{
+	struct btrfs_key ins;
+	int ret;
+	struct extent_buffer *buf;
+
+	ret = btrfs_alloc_extent(trans, root, blocksize, parent, blocksize,
+				 root_objectid, ref_generation, level,
+				 empty_size, hint, (u64)-1, &ins, 0);
+	if (ret) {
+		BUG_ON(ret > 0);
+		return ERR_PTR(ret);
+	}
+
+	buf = btrfs_init_new_buffer(trans, root, ins.objectid, blocksize);
+	return buf;
+}
+
+int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root, struct extent_buffer *leaf)
+{
+	u64 leaf_owner;
+	u64 leaf_generation;
+	struct btrfs_key key;
+	struct btrfs_file_extent_item *fi;
+	int i;
+	int nritems;
+	int ret;
+
+	BUG_ON(!btrfs_is_leaf(leaf));
+	nritems = btrfs_header_nritems(leaf);
+	leaf_owner = btrfs_header_owner(leaf);
+	leaf_generation = btrfs_header_generation(leaf);
+
+	for (i = 0; i < nritems; i++) {
+		u64 disk_bytenr;
+		cond_resched();
+
+		btrfs_item_key_to_cpu(leaf, &key, i);
+		if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
+			continue;
+		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
+		if (btrfs_file_extent_type(leaf, fi) ==
+		    BTRFS_FILE_EXTENT_INLINE)
+			continue;
+		/*
+		 * FIXME make sure to insert a trans record that
+		 * repeats the snapshot del on crash
+		 */
+		disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+		if (disk_bytenr == 0)
+			continue;
+
+		ret = __btrfs_free_extent(trans, root, disk_bytenr,
+				btrfs_file_extent_disk_num_bytes(leaf, fi),
+				leaf->start, leaf_owner, leaf_generation,
+				key.objectid, 0);
+		BUG_ON(ret);
+
+		atomic_inc(&root->fs_info->throttle_gen);
+		wake_up(&root->fs_info->transaction_throttle);
+		cond_resched();
+	}
+	return 0;
+}
+
+static noinline int cache_drop_leaf_ref(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct btrfs_leaf_ref *ref)
+{
+	int i;
+	int ret;
+	struct btrfs_extent_info *info = ref->extents;
+
+	for (i = 0; i < ref->nritems; i++) {
+		ret = __btrfs_free_extent(trans, root, info->bytenr,
+					  info->num_bytes, ref->bytenr,
+					  ref->owner, ref->generation,
+					  info->objectid, 0);
+
+		atomic_inc(&root->fs_info->throttle_gen);
+		wake_up(&root->fs_info->transaction_throttle);
+		cond_resched();
+
+		BUG_ON(ret);
+		info++;
+	}
+
+	return 0;
+}
+
+static int drop_snap_lookup_refcount(struct btrfs_root *root, u64 start,
+				     u64 len, u32 *refs)
+{
+	int ret;
+
+	ret = btrfs_lookup_extent_ref(NULL, root, start, len, refs);
+	BUG_ON(ret);
+
+#if 0 /* some debugging code in case we see problems here */
+	/* if the refs count is one, it won't get increased again.  But
+	 * if the ref count is > 1, someone may be decreasing it at
+	 * the same time we are.
+	 */
+	if (*refs != 1) {
+		struct extent_buffer *eb = NULL;
+		eb = btrfs_find_create_tree_block(root, start, len);
+		if (eb)
+			btrfs_tree_lock(eb);
+
+		mutex_lock(&root->fs_info->alloc_mutex);
+		ret = lookup_extent_ref(NULL, root, start, len, refs);
+		BUG_ON(ret);
+		mutex_unlock(&root->fs_info->alloc_mutex);
+
+		if (eb) {
+			btrfs_tree_unlock(eb);
+			free_extent_buffer(eb);
+		}
+		if (*refs == 1) {
+			printk(KERN_ERR "btrfs block %llu went down to one "
+			       "during drop_snap\n", (unsigned long long)start);
+		}
+
+	}
+#endif
+
+	cond_resched();
+	return ret;
+}
+
+/*
+ * helper function for drop_snapshot, this walks down the tree dropping ref
+ * counts as it goes.
+ */
+static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root,
+				   struct btrfs_path *path, int *level)
+{
+	u64 root_owner;
+	u64 root_gen;
+	u64 bytenr;
+	u64 ptr_gen;
+	struct extent_buffer *next;
+	struct extent_buffer *cur;
+	struct extent_buffer *parent;
+	struct btrfs_leaf_ref *ref;
+	u32 blocksize;
+	int ret;
+	u32 refs;
+
+	WARN_ON(*level < 0);
+	WARN_ON(*level >= BTRFS_MAX_LEVEL);
+	ret = drop_snap_lookup_refcount(root, path->nodes[*level]->start,
+				path->nodes[*level]->len, &refs);
+	BUG_ON(ret);
+	if (refs > 1)
+		goto out;
+
+	/*
+	 * walk down to the last node level and free all the leaves
+	 */
+	while (*level >= 0) {
+		WARN_ON(*level < 0);
+		WARN_ON(*level >= BTRFS_MAX_LEVEL);
+		cur = path->nodes[*level];
+
+		if (btrfs_header_level(cur) != *level)
+			WARN_ON(1);
+
+		if (path->slots[*level] >=
+		    btrfs_header_nritems(cur))
+			break;
+		if (*level == 0) {
+			ret = btrfs_drop_leaf_ref(trans, root, cur);
+			BUG_ON(ret);
+			break;
+		}
+		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
+		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+		blocksize = btrfs_level_size(root, *level - 1);
+
+		ret = drop_snap_lookup_refcount(root, bytenr, blocksize, &refs);
+		BUG_ON(ret);
+		if (refs != 1) {
+			parent = path->nodes[*level];
+			root_owner = btrfs_header_owner(parent);
+			root_gen = btrfs_header_generation(parent);
+			path->slots[*level]++;
+
+			ret = __btrfs_free_extent(trans, root, bytenr,
+						blocksize, parent->start,
+						root_owner, root_gen,
+						*level - 1, 1);
+			BUG_ON(ret);
+
+			atomic_inc(&root->fs_info->throttle_gen);
+			wake_up(&root->fs_info->transaction_throttle);
+			cond_resched();
+
+			continue;
+		}
+		/*
+		 * at this point, we have a single ref, and since the
+		 * only place referencing this extent is a dead root
+		 * the reference count should never go higher.
+		 * So, we don't need to check it again
+		 */
+		if (*level == 1) {
+			ref = btrfs_lookup_leaf_ref(root, bytenr);
+			if (ref && ref->generation != ptr_gen) {
+				btrfs_free_leaf_ref(root, ref);
+				ref = NULL;
+			}
+			if (ref) {
+				ret = cache_drop_leaf_ref(trans, root, ref);
+				BUG_ON(ret);
+				btrfs_remove_leaf_ref(root, ref);
+				btrfs_free_leaf_ref(root, ref);
+				*level = 0;
+				break;
+			}
+		}
+		next = btrfs_find_tree_block(root, bytenr, blocksize);
+		if (!next || !btrfs_buffer_uptodate(next, ptr_gen)) {
+			free_extent_buffer(next);
+
+			next = read_tree_block(root, bytenr, blocksize,
+					       ptr_gen);
+			cond_resched();
+#if 0
+			/*
+			 * this is a debugging check and can go away
+			 * the ref should never go all the way down to 1
+			 * at this point
+			 */
+			ret = lookup_extent_ref(NULL, root, bytenr, blocksize,
+						&refs);
+			BUG_ON(ret);
+			WARN_ON(refs != 1);
+#endif
+		}
+		WARN_ON(*level <= 0);
+		if (path->nodes[*level-1])
+			free_extent_buffer(path->nodes[*level-1]);
+		path->nodes[*level-1] = next;
+		*level = btrfs_header_level(next);
+		path->slots[*level] = 0;
+		cond_resched();
+	}
+out:
+	WARN_ON(*level < 0);
+	WARN_ON(*level >= BTRFS_MAX_LEVEL);
+
+	if (path->nodes[*level] == root->node) {
+		parent = path->nodes[*level];
+		bytenr = path->nodes[*level]->start;
+	} else {
+		parent = path->nodes[*level + 1];
+		bytenr = btrfs_node_blockptr(parent, path->slots[*level + 1]);
+	}
+
+	blocksize = btrfs_level_size(root, *level);
+	root_owner = btrfs_header_owner(parent);
+	root_gen = btrfs_header_generation(parent);
+
+	ret = __btrfs_free_extent(trans, root, bytenr, blocksize,
+				  parent->start, root_owner, root_gen,
+				  *level, 1);
+	free_extent_buffer(path->nodes[*level]);
+	path->nodes[*level] = NULL;
+	*level += 1;
+	BUG_ON(ret);
+
+	cond_resched();
+	return 0;
+}
+
+/*
+ * helper function for drop_subtree, this function is similar to
+ * walk_down_tree. The main difference is that it checks reference
+ * counts while tree blocks are locked.
+ */
+static noinline int walk_down_subtree(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path, int *level)
+{
+	struct extent_buffer *next;
+	struct extent_buffer *cur;
+	struct extent_buffer *parent;
+	u64 bytenr;
+	u64 ptr_gen;
+	u32 blocksize;
+	u32 refs;
+	int ret;
+
+	cur = path->nodes[*level];
+	ret = btrfs_lookup_extent_ref(trans, root, cur->start, cur->len,
+				      &refs);
+	BUG_ON(ret);
+	if (refs > 1)
+		goto out;
+
+	while (*level >= 0) {
+		cur = path->nodes[*level];
+		if (*level == 0) {
+			ret = btrfs_drop_leaf_ref(trans, root, cur);
+			BUG_ON(ret);
+			clean_tree_block(trans, root, cur);
+			break;
+		}
+		if (path->slots[*level] >= btrfs_header_nritems(cur)) {
+			clean_tree_block(trans, root, cur);
+			break;
+		}
+
+		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
+		blocksize = btrfs_level_size(root, *level - 1);
+		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+
+		next = read_tree_block(root, bytenr, blocksize, ptr_gen);
+		btrfs_tree_lock(next);
+
+		ret = btrfs_lookup_extent_ref(trans, root, bytenr, blocksize,
+					      &refs);
+		BUG_ON(ret);
+		if (refs > 1) {
+			parent = path->nodes[*level];
+			ret = btrfs_free_extent(trans, root, bytenr,
+					blocksize, parent->start,
+					btrfs_header_owner(parent),
+					btrfs_header_generation(parent),
+					*level - 1, 1);
+			BUG_ON(ret);
+			path->slots[*level]++;
+			btrfs_tree_unlock(next);
+			free_extent_buffer(next);
+			continue;
+		}
+
+		*level = btrfs_header_level(next);
+		path->nodes[*level] = next;
+		path->slots[*level] = 0;
+		path->locks[*level] = 1;
+		cond_resched();
+	}
+out:
+	parent = path->nodes[*level + 1];
+	bytenr = path->nodes[*level]->start;
+	blocksize = path->nodes[*level]->len;
+
+	ret = btrfs_free_extent(trans, root, bytenr, blocksize,
+			parent->start, btrfs_header_owner(parent),
+			btrfs_header_generation(parent), *level, 1);
+	BUG_ON(ret);
+
+	if (path->locks[*level]) {
+		btrfs_tree_unlock(path->nodes[*level]);
+		path->locks[*level] = 0;
+	}
+	free_extent_buffer(path->nodes[*level]);
+	path->nodes[*level] = NULL;
+	*level += 1;
+	cond_resched();
+	return 0;
+}
+
+/*
+ * helper for dropping snapshots.  This walks back up the tree in the path
+ * to find the first node higher up where we haven't yet gone through
+ * all the slots
+ */
+static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct btrfs_path *path,
+				 int *level, int max_level)
+{
+	u64 root_owner;
+	u64 root_gen;
+	struct btrfs_root_item *root_item = &root->root_item;
+	int i;
+	int slot;
+	int ret;
+
+	for (i = *level; i < max_level && path->nodes[i]; i++) {
+		slot = path->slots[i];
+		if (slot < btrfs_header_nritems(path->nodes[i]) - 1) {
+			struct extent_buffer *node;
+			struct btrfs_disk_key disk_key;
+			node = path->nodes[i];
+			path->slots[i]++;
+			*level = i;
+			WARN_ON(*level == 0);
+			btrfs_node_key(node, &disk_key, path->slots[i]);
+			memcpy(&root_item->drop_progress,
+			       &disk_key, sizeof(disk_key));
+			root_item->drop_level = i;
+			return 0;
+		} else {
+			struct extent_buffer *parent;
+			if (path->nodes[*level] == root->node)
+				parent = path->nodes[*level];
+			else
+				parent = path->nodes[*level + 1];
+
+			root_owner = btrfs_header_owner(parent);
+			root_gen = btrfs_header_generation(parent);
+
+			clean_tree_block(trans, root, path->nodes[*level]);
+			ret = btrfs_free_extent(trans, root,
+						path->nodes[*level]->start,
+						path->nodes[*level]->len,
+						parent->start, root_owner,
+						root_gen, *level, 1);
+			BUG_ON(ret);
+			if (path->locks[*level]) {
+				btrfs_tree_unlock(path->nodes[*level]);
+				path->locks[*level] = 0;
+			}
+			free_extent_buffer(path->nodes[*level]);
+			path->nodes[*level] = NULL;
+			*level = i + 1;
+		}
+	}
+	return 1;
+}
+
+/*
+ * drop the reference count on the tree rooted at 'snap'.  This traverses
+ * the tree freeing any blocks that have a ref count of zero after being
+ * decremented.
+ */
+int btrfs_drop_snapshot(struct btrfs_trans_handle *trans, struct btrfs_root
+			*root)
+{
+	int ret = 0;
+	int wret;
+	int level;
+	struct btrfs_path *path;
+	int i;
+	int orig_level;
+	struct btrfs_root_item *root_item = &root->root_item;
+
+	WARN_ON(!mutex_is_locked(&root->fs_info->drop_mutex));
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	level = btrfs_header_level(root->node);
+	orig_level = level;
+	if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
+		path->nodes[level] = root->node;
+		extent_buffer_get(root->node);
+		path->slots[level] = 0;
+	} else {
+		struct btrfs_key key;
+		struct btrfs_disk_key found_key;
+		struct extent_buffer *node;
+
+		btrfs_disk_key_to_cpu(&key, &root_item->drop_progress);
+		level = root_item->drop_level;
+		path->lowest_level = level;
+		wret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (wret < 0) {
+			ret = wret;
+			goto out;
+		}
+		node = path->nodes[level];
+		btrfs_node_key(node, &found_key, path->slots[level]);
+		WARN_ON(memcmp(&found_key, &root_item->drop_progress,
+			       sizeof(found_key)));
+		/*
+		 * unlock our path, this is safe because only this
+		 * function is allowed to delete this snapshot
+		 */
+		for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
+			if (path->nodes[i] && path->locks[i]) {
+				path->locks[i] = 0;
+				btrfs_tree_unlock(path->nodes[i]);
+			}
+		}
+	}
+	while (1) {
+		wret = walk_down_tree(trans, root, path, &level);
+		if (wret > 0)
+			break;
+		if (wret < 0)
+			ret = wret;
+
+		wret = walk_up_tree(trans, root, path, &level,
+				    BTRFS_MAX_LEVEL);
+		if (wret > 0)
+			break;
+		if (wret < 0)
+			ret = wret;
+		if (trans->transaction->in_commit) {
+			ret = -EAGAIN;
+			break;
+		}
+		atomic_inc(&root->fs_info->throttle_gen);
+		wake_up(&root->fs_info->transaction_throttle);
+	}
+	for (i = 0; i <= orig_level; i++) {
+		if (path->nodes[i]) {
+			free_extent_buffer(path->nodes[i]);
+			path->nodes[i] = NULL;
+		}
+	}
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			struct extent_buffer *node,
+			struct extent_buffer *parent)
+{
+	struct btrfs_path *path;
+	int level;
+	int parent_level;
+	int ret = 0;
+	int wret;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	BUG_ON(!btrfs_tree_locked(parent));
+	parent_level = btrfs_header_level(parent);
+	extent_buffer_get(parent);
+	path->nodes[parent_level] = parent;
+	path->slots[parent_level] = btrfs_header_nritems(parent);
+
+	BUG_ON(!btrfs_tree_locked(node));
+	level = btrfs_header_level(node);
+	extent_buffer_get(node);
+	path->nodes[level] = node;
+	path->slots[level] = 0;
+
+	while (1) {
+		wret = walk_down_subtree(trans, root, path, &level);
+		if (wret < 0)
+			ret = wret;
+		if (wret != 0)
+			break;
+
+		wret = walk_up_tree(trans, root, path, &level, parent_level);
+		if (wret < 0)
+			ret = wret;
+		if (wret != 0)
+			break;
+	}
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+static unsigned long calc_ra(unsigned long start, unsigned long last,
+			     unsigned long nr)
+{
+	return min(last, start + nr - 1);
+}
+
+static noinline int relocate_inode_pages(struct inode *inode, u64 start,
+					 u64 len)
+{
+	u64 page_start;
+	u64 page_end;
+	unsigned long first_index;
+	unsigned long last_index;
+	unsigned long i;
+	struct page *page;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct file_ra_state *ra;
+	struct btrfs_ordered_extent *ordered;
+	unsigned int total_read = 0;
+	unsigned int total_dirty = 0;
+	int ret = 0;
+
+	ra = kzalloc(sizeof(*ra), GFP_NOFS);
+
+	mutex_lock(&inode->i_mutex);
+	first_index = start >> PAGE_CACHE_SHIFT;
+	last_index = (start + len - 1) >> PAGE_CACHE_SHIFT;
+
+	/* make sure the dirty trick played by the caller work */
+	ret = invalidate_inode_pages2_range(inode->i_mapping,
+					    first_index, last_index);
+	if (ret)
+		goto out_unlock;
+
+	file_ra_state_init(ra, inode->i_mapping);
+
+	for (i = first_index ; i <= last_index; i++) {
+		if (total_read % ra->ra_pages == 0) {
+			btrfs_force_ra(inode->i_mapping, ra, NULL, i,
+				       calc_ra(i, last_index, ra->ra_pages));
+		}
+		total_read++;
+again:
+		if (((u64)i << PAGE_CACHE_SHIFT) > i_size_read(inode))
+			BUG_ON(1);
+		page = grab_cache_page(inode->i_mapping, i);
+		if (!page) {
+			ret = -ENOMEM;
+			goto out_unlock;
+		}
+		if (!PageUptodate(page)) {
+			btrfs_readpage(NULL, page);
+			lock_page(page);
+			if (!PageUptodate(page)) {
+				unlock_page(page);
+				page_cache_release(page);
+				ret = -EIO;
+				goto out_unlock;
+			}
+		}
+		wait_on_page_writeback(page);
+
+		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
+		page_end = page_start + PAGE_CACHE_SIZE - 1;
+		lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+
+		ordered = btrfs_lookup_ordered_extent(inode, page_start);
+		if (ordered) {
+			unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+			unlock_page(page);
+			page_cache_release(page);
+			btrfs_start_ordered_extent(inode, ordered, 1);
+			btrfs_put_ordered_extent(ordered);
+			goto again;
+		}
+		set_page_extent_mapped(page);
+
+		if (i == first_index)
+			set_extent_bits(io_tree, page_start, page_end,
+					EXTENT_BOUNDARY, GFP_NOFS);
+		btrfs_set_extent_delalloc(inode, page_start, page_end);
+
+		set_page_dirty(page);
+		total_dirty++;
+
+		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+		unlock_page(page);
+		page_cache_release(page);
+	}
+
+out_unlock:
+	kfree(ra);
+	mutex_unlock(&inode->i_mutex);
+	balance_dirty_pages_ratelimited_nr(inode->i_mapping, total_dirty);
+	return ret;
+}
+
+static noinline int relocate_data_extent(struct inode *reloc_inode,
+					 struct btrfs_key *extent_key,
+					 u64 offset)
+{
+	struct btrfs_root *root = BTRFS_I(reloc_inode)->root;
+	struct extent_map_tree *em_tree = &BTRFS_I(reloc_inode)->extent_tree;
+	struct extent_map *em;
+	u64 start = extent_key->objectid - offset;
+	u64 end = start + extent_key->offset - 1;
+
+	em = alloc_extent_map(GFP_NOFS);
+	BUG_ON(!em || IS_ERR(em));
+
+	em->start = start;
+	em->len = extent_key->offset;
+	em->block_len = extent_key->offset;
+	em->block_start = extent_key->objectid;
+	em->bdev = root->fs_info->fs_devices->latest_bdev;
+	set_bit(EXTENT_FLAG_PINNED, &em->flags);
+
+	/* setup extent map to cheat btrfs_readpage */
+	lock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS);
+	while (1) {
+		int ret;
+		spin_lock(&em_tree->lock);
+		ret = add_extent_mapping(em_tree, em);
+		spin_unlock(&em_tree->lock);
+		if (ret != -EEXIST) {
+			free_extent_map(em);
+			break;
+		}
+		btrfs_drop_extent_cache(reloc_inode, start, end, 0);
+	}
+	unlock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS);
+
+	return relocate_inode_pages(reloc_inode, start, extent_key->offset);
+}
+
+struct btrfs_ref_path {
+	u64 extent_start;
+	u64 nodes[BTRFS_MAX_LEVEL];
+	u64 root_objectid;
+	u64 root_generation;
+	u64 owner_objectid;
+	u32 num_refs;
+	int lowest_level;
+	int current_level;
+	int shared_level;
+
+	struct btrfs_key node_keys[BTRFS_MAX_LEVEL];
+	u64 new_nodes[BTRFS_MAX_LEVEL];
+};
+
+struct disk_extent {
+	u64 ram_bytes;
+	u64 disk_bytenr;
+	u64 disk_num_bytes;
+	u64 offset;
+	u64 num_bytes;
+	u8 compression;
+	u8 encryption;
+	u16 other_encoding;
+};
+
+static int is_cowonly_root(u64 root_objectid)
+{
+	if (root_objectid == BTRFS_ROOT_TREE_OBJECTID ||
+	    root_objectid == BTRFS_EXTENT_TREE_OBJECTID ||
+	    root_objectid == BTRFS_CHUNK_TREE_OBJECTID ||
+	    root_objectid == BTRFS_DEV_TREE_OBJECTID ||
+	    root_objectid == BTRFS_TREE_LOG_OBJECTID ||
+	    root_objectid == BTRFS_CSUM_TREE_OBJECTID)
+		return 1;
+	return 0;
+}
+
+static noinline int __next_ref_path(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *extent_root,
+				    struct btrfs_ref_path *ref_path,
+				    int first_time)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_path *path;
+	struct btrfs_extent_ref *ref;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	u64 bytenr;
+	u32 nritems;
+	int level;
+	int ret = 1;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	if (first_time) {
+		ref_path->lowest_level = -1;
+		ref_path->current_level = -1;
+		ref_path->shared_level = -1;
+		goto walk_up;
+	}
+walk_down:
+	level = ref_path->current_level - 1;
+	while (level >= -1) {
+		u64 parent;
+		if (level < ref_path->lowest_level)
+			break;
+
+		if (level >= 0)
+			bytenr = ref_path->nodes[level];
+		else
+			bytenr = ref_path->extent_start;
+		BUG_ON(bytenr == 0);
+
+		parent = ref_path->nodes[level + 1];
+		ref_path->nodes[level + 1] = 0;
+		ref_path->current_level = level;
+		BUG_ON(parent == 0);
+
+		key.objectid = bytenr;
+		key.offset = parent + 1;
+		key.type = BTRFS_EXTENT_REF_KEY;
+
+		ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+		BUG_ON(ret == 0);
+
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(extent_root, path);
+			if (ret < 0)
+				goto out;
+			if (ret > 0)
+				goto next;
+			leaf = path->nodes[0];
+		}
+
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (found_key.objectid == bytenr &&
+		    found_key.type == BTRFS_EXTENT_REF_KEY) {
+			if (level < ref_path->shared_level)
+				ref_path->shared_level = level;
+			goto found;
+		}
+next:
+		level--;
+		btrfs_release_path(extent_root, path);
+		cond_resched();
+	}
+	/* reached lowest level */
+	ret = 1;
+	goto out;
+walk_up:
+	level = ref_path->current_level;
+	while (level < BTRFS_MAX_LEVEL - 1) {
+		u64 ref_objectid;
+
+		if (level >= 0)
+			bytenr = ref_path->nodes[level];
+		else
+			bytenr = ref_path->extent_start;
+
+		BUG_ON(bytenr == 0);
+
+		key.objectid = bytenr;
+		key.offset = 0;
+		key.type = BTRFS_EXTENT_REF_KEY;
+
+		ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(extent_root, path);
+			if (ret < 0)
+				goto out;
+			if (ret > 0) {
+				/* the extent was freed by someone */
+				if (ref_path->lowest_level == level)
+					goto out;
+				btrfs_release_path(extent_root, path);
+				goto walk_down;
+			}
+			leaf = path->nodes[0];
+		}
+
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (found_key.objectid != bytenr ||
+				found_key.type != BTRFS_EXTENT_REF_KEY) {
+			/* the extent was freed by someone */
+			if (ref_path->lowest_level == level) {
+				ret = 1;
+				goto out;
+			}
+			btrfs_release_path(extent_root, path);
+			goto walk_down;
+		}
+found:
+		ref = btrfs_item_ptr(leaf, path->slots[0],
+				struct btrfs_extent_ref);
+		ref_objectid = btrfs_ref_objectid(leaf, ref);
+		if (ref_objectid < BTRFS_FIRST_FREE_OBJECTID) {
+			if (first_time) {
+				level = (int)ref_objectid;
+				BUG_ON(level >= BTRFS_MAX_LEVEL);
+				ref_path->lowest_level = level;
+				ref_path->current_level = level;
+				ref_path->nodes[level] = bytenr;
+			} else {
+				WARN_ON(ref_objectid != level);
+			}
+		} else {
+			WARN_ON(level != -1);
+		}
+		first_time = 0;
+
+		if (ref_path->lowest_level == level) {
+			ref_path->owner_objectid = ref_objectid;
+			ref_path->num_refs = btrfs_ref_num_refs(leaf, ref);
+		}
+
+		/*
+		 * the block is tree root or the block isn't in reference
+		 * counted tree.
+		 */
+		if (found_key.objectid == found_key.offset ||
+		    is_cowonly_root(btrfs_ref_root(leaf, ref))) {
+			ref_path->root_objectid = btrfs_ref_root(leaf, ref);
+			ref_path->root_generation =
+				btrfs_ref_generation(leaf, ref);
+			if (level < 0) {
+				/* special reference from the tree log */
+				ref_path->nodes[0] = found_key.offset;
+				ref_path->current_level = 0;
+			}
+			ret = 0;
+			goto out;
+		}
+
+		level++;
+		BUG_ON(ref_path->nodes[level] != 0);
+		ref_path->nodes[level] = found_key.offset;
+		ref_path->current_level = level;
+
+		/*
+		 * the reference was created in the running transaction,
+		 * no need to continue walking up.
+		 */
+		if (btrfs_ref_generation(leaf, ref) == trans->transid) {
+			ref_path->root_objectid = btrfs_ref_root(leaf, ref);
+			ref_path->root_generation =
+				btrfs_ref_generation(leaf, ref);
+			ret = 0;
+			goto out;
+		}
+
+		btrfs_release_path(extent_root, path);
+		cond_resched();
+	}
+	/* reached max tree level, but no tree root found. */
+	BUG();
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_first_ref_path(struct btrfs_trans_handle *trans,
+				struct btrfs_root *extent_root,
+				struct btrfs_ref_path *ref_path,
+				u64 extent_start)
+{
+	memset(ref_path, 0, sizeof(*ref_path));
+	ref_path->extent_start = extent_start;
+
+	return __next_ref_path(trans, extent_root, ref_path, 1);
+}
+
+static int btrfs_next_ref_path(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *extent_root,
+			       struct btrfs_ref_path *ref_path)
+{
+	return __next_ref_path(trans, extent_root, ref_path, 0);
+}
+
+static noinline int get_new_locations(struct inode *reloc_inode,
+				      struct btrfs_key *extent_key,
+				      u64 offset, int no_fragment,
+				      struct disk_extent **extents,
+				      int *nr_extents)
+{
+	struct btrfs_root *root = BTRFS_I(reloc_inode)->root;
+	struct btrfs_path *path;
+	struct btrfs_file_extent_item *fi;
+	struct extent_buffer *leaf;
+	struct disk_extent *exts = *extents;
+	struct btrfs_key found_key;
+	u64 cur_pos;
+	u64 last_byte;
+	u32 nritems;
+	int nr = 0;
+	int max = *nr_extents;
+	int ret;
+
+	WARN_ON(!no_fragment && *extents);
+	if (!exts) {
+		max = 1;
+		exts = kmalloc(sizeof(*exts) * max, GFP_NOFS);
+		if (!exts)
+			return -ENOMEM;
+	}
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	cur_pos = extent_key->objectid - offset;
+	last_byte = extent_key->objectid + extent_key->offset;
+	ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
+				       cur_pos, 0);
+	if (ret < 0)
+		goto out;
+	if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	while (1) {
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			if (ret > 0)
+				break;
+			leaf = path->nodes[0];
+		}
+
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (found_key.offset != cur_pos ||
+		    found_key.type != BTRFS_EXTENT_DATA_KEY ||
+		    found_key.objectid != reloc_inode->i_ino)
+			break;
+
+		fi = btrfs_item_ptr(leaf, path->slots[0],
+				    struct btrfs_file_extent_item);
+		if (btrfs_file_extent_type(leaf, fi) !=
+		    BTRFS_FILE_EXTENT_REG ||
+		    btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
+			break;
+
+		if (nr == max) {
+			struct disk_extent *old = exts;
+			max *= 2;
+			exts = kzalloc(sizeof(*exts) * max, GFP_NOFS);
+			memcpy(exts, old, sizeof(*exts) * nr);
+			if (old != *extents)
+				kfree(old);
+		}
+
+		exts[nr].disk_bytenr =
+			btrfs_file_extent_disk_bytenr(leaf, fi);
+		exts[nr].disk_num_bytes =
+			btrfs_file_extent_disk_num_bytes(leaf, fi);
+		exts[nr].offset = btrfs_file_extent_offset(leaf, fi);
+		exts[nr].num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
+		exts[nr].ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
+		exts[nr].compression = btrfs_file_extent_compression(leaf, fi);
+		exts[nr].encryption = btrfs_file_extent_encryption(leaf, fi);
+		exts[nr].other_encoding = btrfs_file_extent_other_encoding(leaf,
+									   fi);
+		BUG_ON(exts[nr].offset > 0);
+		BUG_ON(exts[nr].compression || exts[nr].encryption);
+		BUG_ON(exts[nr].num_bytes != exts[nr].disk_num_bytes);
+
+		cur_pos += exts[nr].num_bytes;
+		nr++;
+
+		if (cur_pos + offset >= last_byte)
+			break;
+
+		if (no_fragment) {
+			ret = 1;
+			goto out;
+		}
+		path->slots[0]++;
+	}
+
+	BUG_ON(cur_pos + offset > last_byte);
+	if (cur_pos + offset < last_byte) {
+		ret = -ENOENT;
+		goto out;
+	}
+	ret = 0;
+out:
+	btrfs_free_path(path);
+	if (ret) {
+		if (exts != *extents)
+			kfree(exts);
+	} else {
+		*extents = exts;
+		*nr_extents = nr;
+	}
+	return ret;
+}
+
+static noinline int replace_one_extent(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct btrfs_path *path,
+					struct btrfs_key *extent_key,
+					struct btrfs_key *leaf_key,
+					struct btrfs_ref_path *ref_path,
+					struct disk_extent *new_extents,
+					int nr_extents)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_file_extent_item *fi;
+	struct inode *inode = NULL;
+	struct btrfs_key key;
+	u64 lock_start = 0;
+	u64 lock_end = 0;
+	u64 num_bytes;
+	u64 ext_offset;
+	u64 first_pos;
+	u32 nritems;
+	int nr_scaned = 0;
+	int extent_locked = 0;
+	int extent_type;
+	int ret;
+
+	memcpy(&key, leaf_key, sizeof(key));
+	first_pos = INT_LIMIT(loff_t) - extent_key->offset;
+	if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
+		if (key.objectid < ref_path->owner_objectid ||
+		    (key.objectid == ref_path->owner_objectid &&
+		     key.type < BTRFS_EXTENT_DATA_KEY)) {
+			key.objectid = ref_path->owner_objectid;
+			key.type = BTRFS_EXTENT_DATA_KEY;
+			key.offset = 0;
+		}
+	}
+
+	while (1) {
+		ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+		if (ret < 0)
+			goto out;
+
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+next:
+		if (extent_locked && ret > 0) {
+			/*
+			 * the file extent item was modified by someone
+			 * before the extent got locked.
+			 */
+			unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
+				      lock_end, GFP_NOFS);
+			extent_locked = 0;
+		}
+
+		if (path->slots[0] >= nritems) {
+			if (++nr_scaned > 2)
+				break;
+
+			BUG_ON(extent_locked);
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			if (ret > 0)
+				break;
+			leaf = path->nodes[0];
+			nritems = btrfs_header_nritems(leaf);
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+
+		if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
+			if ((key.objectid > ref_path->owner_objectid) ||
+			    (key.objectid == ref_path->owner_objectid &&
+			     key.type > BTRFS_EXTENT_DATA_KEY) ||
+			    (key.offset >= first_pos + extent_key->offset))
+				break;
+		}
+
+		if (inode && key.objectid != inode->i_ino) {
+			BUG_ON(extent_locked);
+			btrfs_release_path(root, path);
+			mutex_unlock(&inode->i_mutex);
+			iput(inode);
+			inode = NULL;
+			continue;
+		}
+
+		if (key.type != BTRFS_EXTENT_DATA_KEY) {
+			path->slots[0]++;
+			ret = 1;
+			goto next;
+		}
+		fi = btrfs_item_ptr(leaf, path->slots[0],
+				    struct btrfs_file_extent_item);
+		extent_type = btrfs_file_extent_type(leaf, fi);
+		if ((extent_type != BTRFS_FILE_EXTENT_REG &&
+		     extent_type != BTRFS_FILE_EXTENT_PREALLOC) ||
+		    (btrfs_file_extent_disk_bytenr(leaf, fi) !=
+		     extent_key->objectid)) {
+			path->slots[0]++;
+			ret = 1;
+			goto next;
+		}
+
+		num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
+		ext_offset = btrfs_file_extent_offset(leaf, fi);
+
+		if (first_pos > key.offset - ext_offset)
+			first_pos = key.offset - ext_offset;
+
+		if (!extent_locked) {
+			lock_start = key.offset;
+			lock_end = lock_start + num_bytes - 1;
+		} else {
+			if (lock_start > key.offset ||
+			    lock_end + 1 < key.offset + num_bytes) {
+				unlock_extent(&BTRFS_I(inode)->io_tree,
+					      lock_start, lock_end, GFP_NOFS);
+				extent_locked = 0;
+			}
+		}
+
+		if (!inode) {
+			btrfs_release_path(root, path);
+
+			inode = btrfs_iget_locked(root->fs_info->sb,
+						  key.objectid, root);
+			if (inode->i_state & I_NEW) {
+				BTRFS_I(inode)->root = root;
+				BTRFS_I(inode)->location.objectid =
+					key.objectid;
+				BTRFS_I(inode)->location.type =
+					BTRFS_INODE_ITEM_KEY;
+				BTRFS_I(inode)->location.offset = 0;
+				btrfs_read_locked_inode(inode);
+				unlock_new_inode(inode);
+			}
+			/*
+			 * some code call btrfs_commit_transaction while
+			 * holding the i_mutex, so we can't use mutex_lock
+			 * here.
+			 */
+			if (is_bad_inode(inode) ||
+			    !mutex_trylock(&inode->i_mutex)) {
+				iput(inode);
+				inode = NULL;
+				key.offset = (u64)-1;
+				goto skip;
+			}
+		}
+
+		if (!extent_locked) {
+			struct btrfs_ordered_extent *ordered;
+
+			btrfs_release_path(root, path);
+
+			lock_extent(&BTRFS_I(inode)->io_tree, lock_start,
+				    lock_end, GFP_NOFS);
+			ordered = btrfs_lookup_first_ordered_extent(inode,
+								    lock_end);
+			if (ordered &&
+			    ordered->file_offset <= lock_end &&
+			    ordered->file_offset + ordered->len > lock_start) {
+				unlock_extent(&BTRFS_I(inode)->io_tree,
+					      lock_start, lock_end, GFP_NOFS);
+				btrfs_start_ordered_extent(inode, ordered, 1);
+				btrfs_put_ordered_extent(ordered);
+				key.offset += num_bytes;
+				goto skip;
+			}
+			if (ordered)
+				btrfs_put_ordered_extent(ordered);
+
+			extent_locked = 1;
+			continue;
+		}
+
+		if (nr_extents == 1) {
+			/* update extent pointer in place */
+			btrfs_set_file_extent_disk_bytenr(leaf, fi,
+						new_extents[0].disk_bytenr);
+			btrfs_set_file_extent_disk_num_bytes(leaf, fi,
+						new_extents[0].disk_num_bytes);
+			btrfs_mark_buffer_dirty(leaf);
+
+			btrfs_drop_extent_cache(inode, key.offset,
+						key.offset + num_bytes - 1, 0);
+
+			ret = btrfs_inc_extent_ref(trans, root,
+						new_extents[0].disk_bytenr,
+						new_extents[0].disk_num_bytes,
+						leaf->start,
+						root->root_key.objectid,
+						trans->transid,
+						key.objectid);
+			BUG_ON(ret);
+
+			ret = btrfs_free_extent(trans, root,
+						extent_key->objectid,
+						extent_key->offset,
+						leaf->start,
+						btrfs_header_owner(leaf),
+						btrfs_header_generation(leaf),
+						key.objectid, 0);
+			BUG_ON(ret);
+
+			btrfs_release_path(root, path);
+			key.offset += num_bytes;
+		} else {
+			BUG_ON(1);
+#if 0
+			u64 alloc_hint;
+			u64 extent_len;
+			int i;
+			/*
+			 * drop old extent pointer at first, then insert the
+			 * new pointers one bye one
+			 */
+			btrfs_release_path(root, path);
+			ret = btrfs_drop_extents(trans, root, inode, key.offset,
+						 key.offset + num_bytes,
+						 key.offset, &alloc_hint);
+			BUG_ON(ret);
+
+			for (i = 0; i < nr_extents; i++) {
+				if (ext_offset >= new_extents[i].num_bytes) {
+					ext_offset -= new_extents[i].num_bytes;
+					continue;
+				}
+				extent_len = min(new_extents[i].num_bytes -
+						 ext_offset, num_bytes);
+
+				ret = btrfs_insert_empty_item(trans, root,
+							      path, &key,
+							      sizeof(*fi));
+				BUG_ON(ret);
+
+				leaf = path->nodes[0];
+				fi = btrfs_item_ptr(leaf, path->slots[0],
+						struct btrfs_file_extent_item);
+				btrfs_set_file_extent_generation(leaf, fi,
+							trans->transid);
+				btrfs_set_file_extent_type(leaf, fi,
+							BTRFS_FILE_EXTENT_REG);
+				btrfs_set_file_extent_disk_bytenr(leaf, fi,
+						new_extents[i].disk_bytenr);
+				btrfs_set_file_extent_disk_num_bytes(leaf, fi,
+						new_extents[i].disk_num_bytes);
+				btrfs_set_file_extent_ram_bytes(leaf, fi,
+						new_extents[i].ram_bytes);
+
+				btrfs_set_file_extent_compression(leaf, fi,
+						new_extents[i].compression);
+				btrfs_set_file_extent_encryption(leaf, fi,
+						new_extents[i].encryption);
+				btrfs_set_file_extent_other_encoding(leaf, fi,
+						new_extents[i].other_encoding);
+
+				btrfs_set_file_extent_num_bytes(leaf, fi,
+							extent_len);
+				ext_offset += new_extents[i].offset;
+				btrfs_set_file_extent_offset(leaf, fi,
+							ext_offset);
+				btrfs_mark_buffer_dirty(leaf);
+
+				btrfs_drop_extent_cache(inode, key.offset,
+						key.offset + extent_len - 1, 0);
+
+				ret = btrfs_inc_extent_ref(trans, root,
+						new_extents[i].disk_bytenr,
+						new_extents[i].disk_num_bytes,
+						leaf->start,
+						root->root_key.objectid,
+						trans->transid, key.objectid);
+				BUG_ON(ret);
+				btrfs_release_path(root, path);
+
+				inode_add_bytes(inode, extent_len);
+
+				ext_offset = 0;
+				num_bytes -= extent_len;
+				key.offset += extent_len;
+
+				if (num_bytes == 0)
+					break;
+			}
+			BUG_ON(i >= nr_extents);
+#endif
+		}
+
+		if (extent_locked) {
+			unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
+				      lock_end, GFP_NOFS);
+			extent_locked = 0;
+		}
+skip:
+		if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS &&
+		    key.offset >= first_pos + extent_key->offset)
+			break;
+
+		cond_resched();
+	}
+	ret = 0;
+out:
+	btrfs_release_path(root, path);
+	if (inode) {
+		mutex_unlock(&inode->i_mutex);
+		if (extent_locked) {
+			unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
+				      lock_end, GFP_NOFS);
+		}
+		iput(inode);
+	}
+	return ret;
+}
+
+int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       struct extent_buffer *buf, u64 orig_start)
+{
+	int level;
+	int ret;
+
+	BUG_ON(btrfs_header_generation(buf) != trans->transid);
+	BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
+
+	level = btrfs_header_level(buf);
+	if (level == 0) {
+		struct btrfs_leaf_ref *ref;
+		struct btrfs_leaf_ref *orig_ref;
+
+		orig_ref = btrfs_lookup_leaf_ref(root, orig_start);
+		if (!orig_ref)
+			return -ENOENT;
+
+		ref = btrfs_alloc_leaf_ref(root, orig_ref->nritems);
+		if (!ref) {
+			btrfs_free_leaf_ref(root, orig_ref);
+			return -ENOMEM;
+		}
+
+		ref->nritems = orig_ref->nritems;
+		memcpy(ref->extents, orig_ref->extents,
+			sizeof(ref->extents[0]) * ref->nritems);
+
+		btrfs_free_leaf_ref(root, orig_ref);
+
+		ref->root_gen = trans->transid;
+		ref->bytenr = buf->start;
+		ref->owner = btrfs_header_owner(buf);
+		ref->generation = btrfs_header_generation(buf);
+		ret = btrfs_add_leaf_ref(root, ref, 0);
+		WARN_ON(ret);
+		btrfs_free_leaf_ref(root, ref);
+	}
+	return 0;
+}
+
+static noinline int invalidate_extent_cache(struct btrfs_root *root,
+					struct extent_buffer *leaf,
+					struct btrfs_block_group_cache *group,
+					struct btrfs_root *target_root)
+{
+	struct btrfs_key key;
+	struct inode *inode = NULL;
+	struct btrfs_file_extent_item *fi;
+	u64 num_bytes;
+	u64 skip_objectid = 0;
+	u32 nritems;
+	u32 i;
+
+	nritems = btrfs_header_nritems(leaf);
+	for (i = 0; i < nritems; i++) {
+		btrfs_item_key_to_cpu(leaf, &key, i);
+		if (key.objectid == skip_objectid ||
+		    key.type != BTRFS_EXTENT_DATA_KEY)
+			continue;
+		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
+		if (btrfs_file_extent_type(leaf, fi) ==
+		    BTRFS_FILE_EXTENT_INLINE)
+			continue;
+		if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
+			continue;
+		if (!inode || inode->i_ino != key.objectid) {
+			iput(inode);
+			inode = btrfs_ilookup(target_root->fs_info->sb,
+					      key.objectid, target_root, 1);
+		}
+		if (!inode) {
+			skip_objectid = key.objectid;
+			continue;
+		}
+		num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
+
+		lock_extent(&BTRFS_I(inode)->io_tree, key.offset,
+			    key.offset + num_bytes - 1, GFP_NOFS);
+		btrfs_drop_extent_cache(inode, key.offset,
+					key.offset + num_bytes - 1, 1);
+		unlock_extent(&BTRFS_I(inode)->io_tree, key.offset,
+			      key.offset + num_bytes - 1, GFP_NOFS);
+		cond_resched();
+	}
+	iput(inode);
+	return 0;
+}
+
+static noinline int replace_extents_in_leaf(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct extent_buffer *leaf,
+					struct btrfs_block_group_cache *group,
+					struct inode *reloc_inode)
+{
+	struct btrfs_key key;
+	struct btrfs_key extent_key;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_leaf_ref *ref;
+	struct disk_extent *new_extent;
+	u64 bytenr;
+	u64 num_bytes;
+	u32 nritems;
+	u32 i;
+	int ext_index;
+	int nr_extent;
+	int ret;
+
+	new_extent = kmalloc(sizeof(*new_extent), GFP_NOFS);
+	BUG_ON(!new_extent);
+
+	ref = btrfs_lookup_leaf_ref(root, leaf->start);
+	BUG_ON(!ref);
+
+	ext_index = -1;
+	nritems = btrfs_header_nritems(leaf);
+	for (i = 0; i < nritems; i++) {
+		btrfs_item_key_to_cpu(leaf, &key, i);
+		if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
+			continue;
+		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
+		if (btrfs_file_extent_type(leaf, fi) ==
+		    BTRFS_FILE_EXTENT_INLINE)
+			continue;
+		bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+		num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
+		if (bytenr == 0)
+			continue;
+
+		ext_index++;
+		if (bytenr >= group->key.objectid + group->key.offset ||
+		    bytenr + num_bytes <= group->key.objectid)
+			continue;
+
+		extent_key.objectid = bytenr;
+		extent_key.offset = num_bytes;
+		extent_key.type = BTRFS_EXTENT_ITEM_KEY;
+		nr_extent = 1;
+		ret = get_new_locations(reloc_inode, &extent_key,
+					group->key.objectid, 1,
+					&new_extent, &nr_extent);
+		if (ret > 0)
+			continue;
+		BUG_ON(ret < 0);
+
+		BUG_ON(ref->extents[ext_index].bytenr != bytenr);
+		BUG_ON(ref->extents[ext_index].num_bytes != num_bytes);
+		ref->extents[ext_index].bytenr = new_extent->disk_bytenr;
+		ref->extents[ext_index].num_bytes = new_extent->disk_num_bytes;
+
+		btrfs_set_file_extent_disk_bytenr(leaf, fi,
+						new_extent->disk_bytenr);
+		btrfs_set_file_extent_disk_num_bytes(leaf, fi,
+						new_extent->disk_num_bytes);
+		btrfs_mark_buffer_dirty(leaf);
+
+		ret = btrfs_inc_extent_ref(trans, root,
+					new_extent->disk_bytenr,
+					new_extent->disk_num_bytes,
+					leaf->start,
+					root->root_key.objectid,
+					trans->transid, key.objectid);
+		BUG_ON(ret);
+		ret = btrfs_free_extent(trans, root,
+					bytenr, num_bytes, leaf->start,
+					btrfs_header_owner(leaf),
+					btrfs_header_generation(leaf),
+					key.objectid, 0);
+		BUG_ON(ret);
+		cond_resched();
+	}
+	kfree(new_extent);
+	BUG_ON(ext_index + 1 != ref->nritems);
+	btrfs_free_leaf_ref(root, ref);
+	return 0;
+}
+
+int btrfs_free_reloc_root(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root)
+{
+	struct btrfs_root *reloc_root;
+	int ret;
+
+	if (root->reloc_root) {
+		reloc_root = root->reloc_root;
+		root->reloc_root = NULL;
+		list_add(&reloc_root->dead_list,
+			 &root->fs_info->dead_reloc_roots);
+
+		btrfs_set_root_bytenr(&reloc_root->root_item,
+				      reloc_root->node->start);
+		btrfs_set_root_level(&root->root_item,
+				     btrfs_header_level(reloc_root->node));
+		memset(&reloc_root->root_item.drop_progress, 0,
+			sizeof(struct btrfs_disk_key));
+		reloc_root->root_item.drop_level = 0;
+
+		ret = btrfs_update_root(trans, root->fs_info->tree_root,
+					&reloc_root->root_key,
+					&reloc_root->root_item);
+		BUG_ON(ret);
+	}
+	return 0;
+}
+
+int btrfs_drop_dead_reloc_roots(struct btrfs_root *root)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *reloc_root;
+	struct btrfs_root *prev_root = NULL;
+	struct list_head dead_roots;
+	int ret;
+	unsigned long nr;
+
+	INIT_LIST_HEAD(&dead_roots);
+	list_splice_init(&root->fs_info->dead_reloc_roots, &dead_roots);
+
+	while (!list_empty(&dead_roots)) {
+		reloc_root = list_entry(dead_roots.prev,
+					struct btrfs_root, dead_list);
+		list_del_init(&reloc_root->dead_list);
+
+		BUG_ON(reloc_root->commit_root != NULL);
+		while (1) {
+			trans = btrfs_join_transaction(root, 1);
+			BUG_ON(!trans);
+
+			mutex_lock(&root->fs_info->drop_mutex);
+			ret = btrfs_drop_snapshot(trans, reloc_root);
+			if (ret != -EAGAIN)
+				break;
+			mutex_unlock(&root->fs_info->drop_mutex);
+
+			nr = trans->blocks_used;
+			ret = btrfs_end_transaction(trans, root);
+			BUG_ON(ret);
+			btrfs_btree_balance_dirty(root, nr);
+		}
+
+		free_extent_buffer(reloc_root->node);
+
+		ret = btrfs_del_root(trans, root->fs_info->tree_root,
+				     &reloc_root->root_key);
+		BUG_ON(ret);
+		mutex_unlock(&root->fs_info->drop_mutex);
+
+		nr = trans->blocks_used;
+		ret = btrfs_end_transaction(trans, root);
+		BUG_ON(ret);
+		btrfs_btree_balance_dirty(root, nr);
+
+		kfree(prev_root);
+		prev_root = reloc_root;
+	}
+	if (prev_root) {
+		btrfs_remove_leaf_refs(prev_root, (u64)-1, 0);
+		kfree(prev_root);
+	}
+	return 0;
+}
+
+int btrfs_add_dead_reloc_root(struct btrfs_root *root)
+{
+	list_add(&root->dead_list, &root->fs_info->dead_reloc_roots);
+	return 0;
+}
+
+int btrfs_cleanup_reloc_trees(struct btrfs_root *root)
+{
+	struct btrfs_root *reloc_root;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_key location;
+	int found;
+	int ret;
+
+	mutex_lock(&root->fs_info->tree_reloc_mutex);
+	ret = btrfs_find_dead_roots(root, BTRFS_TREE_RELOC_OBJECTID, NULL);
+	BUG_ON(ret);
+	found = !list_empty(&root->fs_info->dead_reloc_roots);
+	mutex_unlock(&root->fs_info->tree_reloc_mutex);
+
+	if (found) {
+		trans = btrfs_start_transaction(root, 1);
+		BUG_ON(!trans);
+		ret = btrfs_commit_transaction(trans, root);
+		BUG_ON(ret);
+	}
+
+	location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
+	location.offset = (u64)-1;
+	location.type = BTRFS_ROOT_ITEM_KEY;
+
+	reloc_root = btrfs_read_fs_root_no_name(root->fs_info, &location);
+	BUG_ON(!reloc_root);
+	btrfs_orphan_cleanup(reloc_root);
+	return 0;
+}
+
+static noinline int init_reloc_tree(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root)
+{
+	struct btrfs_root *reloc_root;
+	struct extent_buffer *eb;
+	struct btrfs_root_item *root_item;
+	struct btrfs_key root_key;
+	int ret;
+
+	BUG_ON(!root->ref_cows);
+	if (root->reloc_root)
+		return 0;
+
+	root_item = kmalloc(sizeof(*root_item), GFP_NOFS);
+	BUG_ON(!root_item);
+
+	ret = btrfs_copy_root(trans, root, root->commit_root,
+			      &eb, BTRFS_TREE_RELOC_OBJECTID);
+	BUG_ON(ret);
+
+	root_key.objectid = BTRFS_TREE_RELOC_OBJECTID;
+	root_key.offset = root->root_key.objectid;
+	root_key.type = BTRFS_ROOT_ITEM_KEY;
+
+	memcpy(root_item, &root->root_item, sizeof(root_item));
+	btrfs_set_root_refs(root_item, 0);
+	btrfs_set_root_bytenr(root_item, eb->start);
+	btrfs_set_root_level(root_item, btrfs_header_level(eb));
+	btrfs_set_root_generation(root_item, trans->transid);
+
+	btrfs_tree_unlock(eb);
+	free_extent_buffer(eb);
+
+	ret = btrfs_insert_root(trans, root->fs_info->tree_root,
+				&root_key, root_item);
+	BUG_ON(ret);
+	kfree(root_item);
+
+	reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
+						 &root_key);
+	BUG_ON(!reloc_root);
+	reloc_root->last_trans = trans->transid;
+	reloc_root->commit_root = NULL;
+	reloc_root->ref_tree = &root->fs_info->reloc_ref_tree;
+
+	root->reloc_root = reloc_root;
+	return 0;
+}
+
+/*
+ * Core function of space balance.
+ *
+ * The idea is using reloc trees to relocate tree blocks in reference
+ * counted roots. There is one reloc tree for each subvol, and all
+ * reloc trees share same root key objectid. Reloc trees are snapshots
+ * of the latest committed roots of subvols (root->commit_root).
+ *
+ * To relocate a tree block referenced by a subvol, there are two steps.
+ * COW the block through subvol's reloc tree, then update block pointer
+ * in the subvol to point to the new block. Since all reloc trees share
+ * same root key objectid, doing special handing for tree blocks owned
+ * by them is easy. Once a tree block has been COWed in one reloc tree,
+ * we can use the resulting new block directly when the same block is
+ * required to COW again through other reloc trees. By this way, relocated
+ * tree blocks are shared between reloc trees, so they are also shared
+ * between subvols.
+ */
+static noinline int relocate_one_path(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      struct btrfs_key *first_key,
+				      struct btrfs_ref_path *ref_path,
+				      struct btrfs_block_group_cache *group,
+				      struct inode *reloc_inode)
+{
+	struct btrfs_root *reloc_root;
+	struct extent_buffer *eb = NULL;
+	struct btrfs_key *keys;
+	u64 *nodes;
+	int level;
+	int shared_level;
+	int lowest_level = 0;
+	int ret;
+
+	if (ref_path->owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
+		lowest_level = ref_path->owner_objectid;
+
+	if (!root->ref_cows) {
+		path->lowest_level = lowest_level;
+		ret = btrfs_search_slot(trans, root, first_key, path, 0, 1);
+		BUG_ON(ret < 0);
+		path->lowest_level = 0;
+		btrfs_release_path(root, path);
+		return 0;
+	}
+
+	mutex_lock(&root->fs_info->tree_reloc_mutex);
+	ret = init_reloc_tree(trans, root);
+	BUG_ON(ret);
+	reloc_root = root->reloc_root;
+
+	shared_level = ref_path->shared_level;
+	ref_path->shared_level = BTRFS_MAX_LEVEL - 1;
+
+	keys = ref_path->node_keys;
+	nodes = ref_path->new_nodes;
+	memset(&keys[shared_level + 1], 0,
+	       sizeof(*keys) * (BTRFS_MAX_LEVEL - shared_level - 1));
+	memset(&nodes[shared_level + 1], 0,
+	       sizeof(*nodes) * (BTRFS_MAX_LEVEL - shared_level - 1));
+
+	if (nodes[lowest_level] == 0) {
+		path->lowest_level = lowest_level;
+		ret = btrfs_search_slot(trans, reloc_root, first_key, path,
+					0, 1);
+		BUG_ON(ret);
+		for (level = lowest_level; level < BTRFS_MAX_LEVEL; level++) {
+			eb = path->nodes[level];
+			if (!eb || eb == reloc_root->node)
+				break;
+			nodes[level] = eb->start;
+			if (level == 0)
+				btrfs_item_key_to_cpu(eb, &keys[level], 0);
+			else
+				btrfs_node_key_to_cpu(eb, &keys[level], 0);
+		}
+		if (nodes[0] &&
+		    ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
+			eb = path->nodes[0];
+			ret = replace_extents_in_leaf(trans, reloc_root, eb,
+						      group, reloc_inode);
+			BUG_ON(ret);
+		}
+		btrfs_release_path(reloc_root, path);
+	} else {
+		ret = btrfs_merge_path(trans, reloc_root, keys, nodes,
+				       lowest_level);
+		BUG_ON(ret);
+	}
+
+	/*
+	 * replace tree blocks in the fs tree with tree blocks in
+	 * the reloc tree.
+	 */
+	ret = btrfs_merge_path(trans, root, keys, nodes, lowest_level);
+	BUG_ON(ret < 0);
+
+	if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
+		ret = btrfs_search_slot(trans, reloc_root, first_key, path,
+					0, 0);
+		BUG_ON(ret);
+		extent_buffer_get(path->nodes[0]);
+		eb = path->nodes[0];
+		btrfs_release_path(reloc_root, path);
+		ret = invalidate_extent_cache(reloc_root, eb, group, root);
+		BUG_ON(ret);
+		free_extent_buffer(eb);
+	}
+
+	mutex_unlock(&root->fs_info->tree_reloc_mutex);
+	path->lowest_level = 0;
+	return 0;
+}
+
+static noinline int relocate_tree_block(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct btrfs_path *path,
+					struct btrfs_key *first_key,
+					struct btrfs_ref_path *ref_path)
+{
+	int ret;
+
+	ret = relocate_one_path(trans, root, path, first_key,
+				ref_path, NULL, NULL);
+	BUG_ON(ret);
+
+	if (root == root->fs_info->extent_root)
+		btrfs_extent_post_op(trans, root);
+
+	return 0;
+}
+
+static noinline int del_extent_zero(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *extent_root,
+				    struct btrfs_path *path,
+				    struct btrfs_key *extent_key)
+{
+	int ret;
+
+	ret = btrfs_search_slot(trans, extent_root, extent_key, path, -1, 1);
+	if (ret)
+		goto out;
+	ret = btrfs_del_item(trans, extent_root, path);
+out:
+	btrfs_release_path(extent_root, path);
+	return ret;
+}
+
+static noinline struct btrfs_root *read_ref_root(struct btrfs_fs_info *fs_info,
+						struct btrfs_ref_path *ref_path)
+{
+	struct btrfs_key root_key;
+
+	root_key.objectid = ref_path->root_objectid;
+	root_key.type = BTRFS_ROOT_ITEM_KEY;
+	if (is_cowonly_root(ref_path->root_objectid))
+		root_key.offset = 0;
+	else
+		root_key.offset = (u64)-1;
+
+	return btrfs_read_fs_root_no_name(fs_info, &root_key);
+}
+
+static noinline int relocate_one_extent(struct btrfs_root *extent_root,
+					struct btrfs_path *path,
+					struct btrfs_key *extent_key,
+					struct btrfs_block_group_cache *group,
+					struct inode *reloc_inode, int pass)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *found_root;
+	struct btrfs_ref_path *ref_path = NULL;
+	struct disk_extent *new_extents = NULL;
+	int nr_extents = 0;
+	int loops;
+	int ret;
+	int level;
+	struct btrfs_key first_key;
+	u64 prev_block = 0;
+
+
+	trans = btrfs_start_transaction(extent_root, 1);
+	BUG_ON(!trans);
+
+	if (extent_key->objectid == 0) {
+		ret = del_extent_zero(trans, extent_root, path, extent_key);
+		goto out;
+	}
+
+	ref_path = kmalloc(sizeof(*ref_path), GFP_NOFS);
+	if (!ref_path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	for (loops = 0; ; loops++) {
+		if (loops == 0) {
+			ret = btrfs_first_ref_path(trans, extent_root, ref_path,
+						   extent_key->objectid);
+		} else {
+			ret = btrfs_next_ref_path(trans, extent_root, ref_path);
+		}
+		if (ret < 0)
+			goto out;
+		if (ret > 0)
+			break;
+
+		if (ref_path->root_objectid == BTRFS_TREE_LOG_OBJECTID ||
+		    ref_path->root_objectid == BTRFS_TREE_RELOC_OBJECTID)
+			continue;
+
+		found_root = read_ref_root(extent_root->fs_info, ref_path);
+		BUG_ON(!found_root);
+		/*
+		 * for reference counted tree, only process reference paths
+		 * rooted at the latest committed root.
+		 */
+		if (found_root->ref_cows &&
+		    ref_path->root_generation != found_root->root_key.offset)
+			continue;
+
+		if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
+			if (pass == 0) {
+				/*
+				 * copy data extents to new locations
+				 */
+				u64 group_start = group->key.objectid;
+				ret = relocate_data_extent(reloc_inode,
+							   extent_key,
+							   group_start);
+				if (ret < 0)
+					goto out;
+				break;
+			}
+			level = 0;
+		} else {
+			level = ref_path->owner_objectid;
+		}
+
+		if (prev_block != ref_path->nodes[level]) {
+			struct extent_buffer *eb;
+			u64 block_start = ref_path->nodes[level];
+			u64 block_size = btrfs_level_size(found_root, level);
+
+			eb = read_tree_block(found_root, block_start,
+					     block_size, 0);
+			btrfs_tree_lock(eb);
+			BUG_ON(level != btrfs_header_level(eb));
+
+			if (level == 0)
+				btrfs_item_key_to_cpu(eb, &first_key, 0);
+			else
+				btrfs_node_key_to_cpu(eb, &first_key, 0);
+
+			btrfs_tree_unlock(eb);
+			free_extent_buffer(eb);
+			prev_block = block_start;
+		}
+
+		btrfs_record_root_in_trans(found_root);
+		if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
+			/*
+			 * try to update data extent references while
+			 * keeping metadata shared between snapshots.
+			 */
+			if (pass == 1) {
+				ret = relocate_one_path(trans, found_root,
+						path, &first_key, ref_path,
+						group, reloc_inode);
+				if (ret < 0)
+					goto out;
+				continue;
+			}
+			/*
+			 * use fallback method to process the remaining
+			 * references.
+			 */
+			if (!new_extents) {
+				u64 group_start = group->key.objectid;
+				new_extents = kmalloc(sizeof(*new_extents),
+						      GFP_NOFS);
+				nr_extents = 1;
+				ret = get_new_locations(reloc_inode,
+							extent_key,
+							group_start, 1,
+							&new_extents,
+							&nr_extents);
+				if (ret)
+					goto out;
+			}
+			ret = replace_one_extent(trans, found_root,
+						path, extent_key,
+						&first_key, ref_path,
+						new_extents, nr_extents);
+		} else {
+			ret = relocate_tree_block(trans, found_root, path,
+						  &first_key, ref_path);
+		}
+		if (ret < 0)
+			goto out;
+	}
+	ret = 0;
+out:
+	btrfs_end_transaction(trans, extent_root);
+	kfree(new_extents);
+	kfree(ref_path);
+	return ret;
+}
+
+static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
+{
+	u64 num_devices;
+	u64 stripped = BTRFS_BLOCK_GROUP_RAID0 |
+		BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
+
+	num_devices = root->fs_info->fs_devices->rw_devices;
+	if (num_devices == 1) {
+		stripped |= BTRFS_BLOCK_GROUP_DUP;
+		stripped = flags & ~stripped;
+
+		/* turn raid0 into single device chunks */
+		if (flags & BTRFS_BLOCK_GROUP_RAID0)
+			return stripped;
+
+		/* turn mirroring into duplication */
+		if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
+			     BTRFS_BLOCK_GROUP_RAID10))
+			return stripped | BTRFS_BLOCK_GROUP_DUP;
+		return flags;
+	} else {
+		/* they already had raid on here, just return */
+		if (flags & stripped)
+			return flags;
+
+		stripped |= BTRFS_BLOCK_GROUP_DUP;
+		stripped = flags & ~stripped;
+
+		/* switch duplicated blocks with raid1 */
+		if (flags & BTRFS_BLOCK_GROUP_DUP)
+			return stripped | BTRFS_BLOCK_GROUP_RAID1;
+
+		/* turn single device chunks into raid0 */
+		return stripped | BTRFS_BLOCK_GROUP_RAID0;
+	}
+	return flags;
+}
+
+static int __alloc_chunk_for_shrink(struct btrfs_root *root,
+		     struct btrfs_block_group_cache *shrink_block_group,
+		     int force)
+{
+	struct btrfs_trans_handle *trans;
+	u64 new_alloc_flags;
+	u64 calc;
+
+	spin_lock(&shrink_block_group->lock);
+	if (btrfs_block_group_used(&shrink_block_group->item) > 0) {
+		spin_unlock(&shrink_block_group->lock);
+
+		trans = btrfs_start_transaction(root, 1);
+		spin_lock(&shrink_block_group->lock);
+
+		new_alloc_flags = update_block_group_flags(root,
+						   shrink_block_group->flags);
+		if (new_alloc_flags != shrink_block_group->flags) {
+			calc =
+			     btrfs_block_group_used(&shrink_block_group->item);
+		} else {
+			calc = shrink_block_group->key.offset;
+		}
+		spin_unlock(&shrink_block_group->lock);
+
+		do_chunk_alloc(trans, root->fs_info->extent_root,
+			       calc + 2 * 1024 * 1024, new_alloc_flags, force);
+
+		btrfs_end_transaction(trans, root);
+	} else
+		spin_unlock(&shrink_block_group->lock);
+	return 0;
+}
+
+static int __insert_orphan_inode(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 u64 objectid, u64 size)
+{
+	struct btrfs_path *path;
+	struct btrfs_inode_item *item;
+	struct extent_buffer *leaf;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_insert_empty_inode(trans, root, path, objectid);
+	if (ret)
+		goto out;
+
+	leaf = path->nodes[0];
+	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item);
+	memset_extent_buffer(leaf, 0, (unsigned long)item, sizeof(*item));
+	btrfs_set_inode_generation(leaf, item, 1);
+	btrfs_set_inode_size(leaf, item, size);
+	btrfs_set_inode_mode(leaf, item, S_IFREG | 0600);
+	btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS);
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_release_path(root, path);
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static noinline struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
+					struct btrfs_block_group_cache *group)
+{
+	struct inode *inode = NULL;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root;
+	struct btrfs_key root_key;
+	u64 objectid = BTRFS_FIRST_FREE_OBJECTID;
+	int err = 0;
+
+	root_key.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
+	root_key.type = BTRFS_ROOT_ITEM_KEY;
+	root_key.offset = (u64)-1;
+	root = btrfs_read_fs_root_no_name(fs_info, &root_key);
+	if (IS_ERR(root))
+		return ERR_CAST(root);
+
+	trans = btrfs_start_transaction(root, 1);
+	BUG_ON(!trans);
+
+	err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
+	if (err)
+		goto out;
+
+	err = __insert_orphan_inode(trans, root, objectid, group->key.offset);
+	BUG_ON(err);
+
+	err = btrfs_insert_file_extent(trans, root, objectid, 0, 0, 0,
+				       group->key.offset, 0, group->key.offset,
+				       0, 0, 0);
+	BUG_ON(err);
+
+	inode = btrfs_iget_locked(root->fs_info->sb, objectid, root);
+	if (inode->i_state & I_NEW) {
+		BTRFS_I(inode)->root = root;
+		BTRFS_I(inode)->location.objectid = objectid;
+		BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
+		BTRFS_I(inode)->location.offset = 0;
+		btrfs_read_locked_inode(inode);
+		unlock_new_inode(inode);
+		BUG_ON(is_bad_inode(inode));
+	} else {
+		BUG_ON(1);
+	}
+	BTRFS_I(inode)->index_cnt = group->key.objectid;
+
+	err = btrfs_orphan_add(trans, inode);
+out:
+	btrfs_end_transaction(trans, root);
+	if (err) {
+		if (inode)
+			iput(inode);
+		inode = ERR_PTR(err);
+	}
+	return inode;
+}
+
+int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
+{
+
+	struct btrfs_ordered_sum *sums;
+	struct btrfs_sector_sum *sector_sum;
+	struct btrfs_ordered_extent *ordered;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct list_head list;
+	size_t offset;
+	int ret;
+	u64 disk_bytenr;
+
+	INIT_LIST_HEAD(&list);
+
+	ordered = btrfs_lookup_ordered_extent(inode, file_pos);
+	BUG_ON(ordered->file_offset != file_pos || ordered->len != len);
+
+	disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt;
+	ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr,
+				       disk_bytenr + len - 1, &list);
+
+	while (!list_empty(&list)) {
+		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
+		list_del_init(&sums->list);
+
+		sector_sum = sums->sums;
+		sums->bytenr = ordered->start;
+
+		offset = 0;
+		while (offset < sums->len) {
+			sector_sum->bytenr += ordered->start - disk_bytenr;
+			sector_sum++;
+			offset += root->sectorsize;
+		}
+
+		btrfs_add_ordered_sum(inode, ordered, sums);
+	}
+	btrfs_put_ordered_extent(ordered);
+	return 0;
+}
+
+int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_path *path;
+	struct btrfs_fs_info *info = root->fs_info;
+	struct extent_buffer *leaf;
+	struct inode *reloc_inode;
+	struct btrfs_block_group_cache *block_group;
+	struct btrfs_key key;
+	u64 skipped;
+	u64 cur_byte;
+	u64 total_found;
+	u32 nritems;
+	int ret;
+	int progress;
+	int pass = 0;
+
+	root = root->fs_info->extent_root;
+
+	block_group = btrfs_lookup_block_group(info, group_start);
+	BUG_ON(!block_group);
+
+	printk(KERN_INFO "btrfs relocating block group %llu flags %llu\n",
+	       (unsigned long long)block_group->key.objectid,
+	       (unsigned long long)block_group->flags);
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	reloc_inode = create_reloc_inode(info, block_group);
+	BUG_ON(IS_ERR(reloc_inode));
+
+	__alloc_chunk_for_shrink(root, block_group, 1);
+	set_block_group_readonly(block_group);
+
+	btrfs_start_delalloc_inodes(info->tree_root);
+	btrfs_wait_ordered_extents(info->tree_root, 0);
+again:
+	skipped = 0;
+	total_found = 0;
+	progress = 0;
+	key.objectid = block_group->key.objectid;
+	key.offset = 0;
+	key.type = 0;
+	cur_byte = key.objectid;
+
+	trans = btrfs_start_transaction(info->tree_root, 1);
+	btrfs_commit_transaction(trans, info->tree_root);
+
+	mutex_lock(&root->fs_info->cleaner_mutex);
+	btrfs_clean_old_snapshots(info->tree_root);
+	btrfs_remove_leaf_refs(info->tree_root, (u64)-1, 1);
+	mutex_unlock(&root->fs_info->cleaner_mutex);
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+next:
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			if (ret == 1) {
+				ret = 0;
+				break;
+			}
+			leaf = path->nodes[0];
+			nritems = btrfs_header_nritems(leaf);
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+
+		if (key.objectid >= block_group->key.objectid +
+		    block_group->key.offset)
+			break;
+
+		if (progress && need_resched()) {
+			btrfs_release_path(root, path);
+			cond_resched();
+			progress = 0;
+			continue;
+		}
+		progress = 1;
+
+		if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY ||
+		    key.objectid + key.offset <= cur_byte) {
+			path->slots[0]++;
+			goto next;
+		}
+
+		total_found++;
+		cur_byte = key.objectid + key.offset;
+		btrfs_release_path(root, path);
+
+		__alloc_chunk_for_shrink(root, block_group, 0);
+		ret = relocate_one_extent(root, path, &key, block_group,
+					  reloc_inode, pass);
+		BUG_ON(ret < 0);
+		if (ret > 0)
+			skipped++;
+
+		key.objectid = cur_byte;
+		key.type = 0;
+		key.offset = 0;
+	}
+
+	btrfs_release_path(root, path);
+
+	if (pass == 0) {
+		btrfs_wait_ordered_range(reloc_inode, 0, (u64)-1);
+		invalidate_mapping_pages(reloc_inode->i_mapping, 0, -1);
+	}
+
+	if (total_found > 0) {
+		printk(KERN_INFO "btrfs found %llu extents in pass %d\n",
+		       (unsigned long long)total_found, pass);
+		pass++;
+		if (total_found == skipped && pass > 2) {
+			iput(reloc_inode);
+			reloc_inode = create_reloc_inode(info, block_group);
+			pass = 0;
+		}
+		goto again;
+	}
+
+	/* delete reloc_inode */
+	iput(reloc_inode);
+
+	/* unpin extents in this range */
+	trans = btrfs_start_transaction(info->tree_root, 1);
+	btrfs_commit_transaction(trans, info->tree_root);
+
+	spin_lock(&block_group->lock);
+	WARN_ON(block_group->pinned > 0);
+	WARN_ON(block_group->reserved > 0);
+	WARN_ON(btrfs_block_group_used(&block_group->item) > 0);
+	spin_unlock(&block_group->lock);
+	put_block_group(block_group);
+	ret = 0;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int find_first_block_group(struct btrfs_root *root,
+		struct btrfs_path *path, struct btrfs_key *key)
+{
+	int ret = 0;
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf;
+	int slot;
+
+	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+
+	while (1) {
+		slot = path->slots[0];
+		leaf = path->nodes[0];
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret == 0)
+				continue;
+			if (ret < 0)
+				goto out;
+			break;
+		}
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		if (found_key.objectid >= key->objectid &&
+		    found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
+			ret = 0;
+			goto out;
+		}
+		path->slots[0]++;
+	}
+	ret = -ENOENT;
+out:
+	return ret;
+}
+
+int btrfs_free_block_groups(struct btrfs_fs_info *info)
+{
+	struct btrfs_block_group_cache *block_group;
+	struct rb_node *n;
+
+	spin_lock(&info->block_group_cache_lock);
+	while ((n = rb_last(&info->block_group_cache_tree)) != NULL) {
+		block_group = rb_entry(n, struct btrfs_block_group_cache,
+				       cache_node);
+		rb_erase(&block_group->cache_node,
+			 &info->block_group_cache_tree);
+		spin_unlock(&info->block_group_cache_lock);
+
+		btrfs_remove_free_space_cache(block_group);
+		down_write(&block_group->space_info->groups_sem);
+		list_del(&block_group->list);
+		up_write(&block_group->space_info->groups_sem);
+
+		WARN_ON(atomic_read(&block_group->count) != 1);
+		kfree(block_group);
+
+		spin_lock(&info->block_group_cache_lock);
+	}
+	spin_unlock(&info->block_group_cache_lock);
+	return 0;
+}
+
+int btrfs_read_block_groups(struct btrfs_root *root)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct btrfs_block_group_cache *cache;
+	struct btrfs_fs_info *info = root->fs_info;
+	struct btrfs_space_info *space_info;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf;
+
+	root = info->extent_root;
+	key.objectid = 0;
+	key.offset = 0;
+	btrfs_set_key_type(&key, BTRFS_BLOCK_GROUP_ITEM_KEY);
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	while (1) {
+		ret = find_first_block_group(root, path, &key);
+		if (ret > 0) {
+			ret = 0;
+			goto error;
+		}
+		if (ret != 0)
+			goto error;
+
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		cache = kzalloc(sizeof(*cache), GFP_NOFS);
+		if (!cache) {
+			ret = -ENOMEM;
+			break;
+		}
+
+		atomic_set(&cache->count, 1);
+		spin_lock_init(&cache->lock);
+		mutex_init(&cache->alloc_mutex);
+		mutex_init(&cache->cache_mutex);
+		INIT_LIST_HEAD(&cache->list);
+		read_extent_buffer(leaf, &cache->item,
+				   btrfs_item_ptr_offset(leaf, path->slots[0]),
+				   sizeof(cache->item));
+		memcpy(&cache->key, &found_key, sizeof(found_key));
+
+		key.objectid = found_key.objectid + found_key.offset;
+		btrfs_release_path(root, path);
+		cache->flags = btrfs_block_group_flags(&cache->item);
+
+		ret = update_space_info(info, cache->flags, found_key.offset,
+					btrfs_block_group_used(&cache->item),
+					&space_info);
+		BUG_ON(ret);
+		cache->space_info = space_info;
+		down_write(&space_info->groups_sem);
+		list_add_tail(&cache->list, &space_info->block_groups);
+		up_write(&space_info->groups_sem);
+
+		ret = btrfs_add_block_group_cache(root->fs_info, cache);
+		BUG_ON(ret);
+
+		set_avail_alloc_bits(root->fs_info, cache->flags);
+		if (btrfs_chunk_readonly(root, cache->key.objectid))
+			set_block_group_readonly(cache);
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_make_block_group(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, u64 bytes_used,
+			   u64 type, u64 chunk_objectid, u64 chunk_offset,
+			   u64 size)
+{
+	int ret;
+	struct btrfs_root *extent_root;
+	struct btrfs_block_group_cache *cache;
+
+	extent_root = root->fs_info->extent_root;
+
+	root->fs_info->last_trans_new_blockgroup = trans->transid;
+
+	cache = kzalloc(sizeof(*cache), GFP_NOFS);
+	if (!cache)
+		return -ENOMEM;
+
+	cache->key.objectid = chunk_offset;
+	cache->key.offset = size;
+	cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
+	atomic_set(&cache->count, 1);
+	spin_lock_init(&cache->lock);
+	mutex_init(&cache->alloc_mutex);
+	mutex_init(&cache->cache_mutex);
+	INIT_LIST_HEAD(&cache->list);
+
+	btrfs_set_block_group_used(&cache->item, bytes_used);
+	btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid);
+	cache->flags = type;
+	btrfs_set_block_group_flags(&cache->item, type);
+
+	ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
+				&cache->space_info);
+	BUG_ON(ret);
+	down_write(&cache->space_info->groups_sem);
+	list_add_tail(&cache->list, &cache->space_info->block_groups);
+	up_write(&cache->space_info->groups_sem);
+
+	ret = btrfs_add_block_group_cache(root->fs_info, cache);
+	BUG_ON(ret);
+
+	ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item,
+				sizeof(cache->item));
+	BUG_ON(ret);
+
+	finish_current_insert(trans, extent_root, 0);
+	ret = del_pending_extents(trans, extent_root, 0);
+	BUG_ON(ret);
+	set_avail_alloc_bits(extent_root->fs_info, type);
+
+	return 0;
+}
+
+int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, u64 group_start)
+{
+	struct btrfs_path *path;
+	struct btrfs_block_group_cache *block_group;
+	struct btrfs_key key;
+	int ret;
+
+	root = root->fs_info->extent_root;
+
+	block_group = btrfs_lookup_block_group(root->fs_info, group_start);
+	BUG_ON(!block_group);
+	BUG_ON(!block_group->ro);
+
+	memcpy(&key, &block_group->key, sizeof(key));
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	btrfs_remove_free_space_cache(block_group);
+	rb_erase(&block_group->cache_node,
+		 &root->fs_info->block_group_cache_tree);
+	down_write(&block_group->space_info->groups_sem);
+	list_del(&block_group->list);
+	up_write(&block_group->space_info->groups_sem);
+
+	spin_lock(&block_group->space_info->lock);
+	block_group->space_info->total_bytes -= block_group->key.offset;
+	block_group->space_info->bytes_readonly -= block_group->key.offset;
+	spin_unlock(&block_group->space_info->lock);
+	block_group->space_info->full = 0;
+
+	put_block_group(block_group);
+	put_block_group(block_group);
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret > 0)
+		ret = -EIO;
+	if (ret < 0)
+		goto out;
+
+	ret = btrfs_del_item(trans, root, path);
+out:
+	btrfs_free_path(path);
+	return ret;
+}
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
new file mode 100644
index 00000000000..e086d407f1f
--- /dev/null
+++ b/fs/btrfs/extent_io.c
@@ -0,0 +1,3717 @@
+#include <linux/bitops.h>
+#include <linux/slab.h>
+#include <linux/bio.h>
+#include <linux/mm.h>
+#include <linux/gfp.h>
+#include <linux/pagemap.h>
+#include <linux/page-flags.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/blkdev.h>
+#include <linux/swap.h>
+#include <linux/version.h>
+#include <linux/writeback.h>
+#include <linux/pagevec.h>
+#include "extent_io.h"
+#include "extent_map.h"
+#include "compat.h"
+#include "ctree.h"
+#include "btrfs_inode.h"
+
+/* temporary define until extent_map moves out of btrfs */
+struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
+				       unsigned long extra_flags,
+				       void (*ctor)(void *, struct kmem_cache *,
+						    unsigned long));
+
+static struct kmem_cache *extent_state_cache;
+static struct kmem_cache *extent_buffer_cache;
+
+static LIST_HEAD(buffers);
+static LIST_HEAD(states);
+
+#define LEAK_DEBUG 0
+#ifdef LEAK_DEBUG
+static DEFINE_SPINLOCK(leak_lock);
+#endif
+
+#define BUFFER_LRU_MAX 64
+
+struct tree_entry {
+	u64 start;
+	u64 end;
+	struct rb_node rb_node;
+};
+
+struct extent_page_data {
+	struct bio *bio;
+	struct extent_io_tree *tree;
+	get_extent_t *get_extent;
+
+	/* tells writepage not to lock the state bits for this range
+	 * it still does the unlocking
+	 */
+	int extent_locked;
+};
+
+int __init extent_io_init(void)
+{
+	extent_state_cache = btrfs_cache_create("extent_state",
+					    sizeof(struct extent_state), 0,
+					    NULL);
+	if (!extent_state_cache)
+		return -ENOMEM;
+
+	extent_buffer_cache = btrfs_cache_create("extent_buffers",
+					    sizeof(struct extent_buffer), 0,
+					    NULL);
+	if (!extent_buffer_cache)
+		goto free_state_cache;
+	return 0;
+
+free_state_cache:
+	kmem_cache_destroy(extent_state_cache);
+	return -ENOMEM;
+}
+
+void extent_io_exit(void)
+{
+	struct extent_state *state;
+	struct extent_buffer *eb;
+
+	while (!list_empty(&states)) {
+		state = list_entry(states.next, struct extent_state, leak_list);
+		printk(KERN_ERR "btrfs state leak: start %llu end %llu "
+		       "state %lu in tree %p refs %d\n",
+		       (unsigned long long)state->start,
+		       (unsigned long long)state->end,
+		       state->state, state->tree, atomic_read(&state->refs));
+		list_del(&state->leak_list);
+		kmem_cache_free(extent_state_cache, state);
+
+	}
+
+	while (!list_empty(&buffers)) {
+		eb = list_entry(buffers.next, struct extent_buffer, leak_list);
+		printk(KERN_ERR "btrfs buffer leak start %llu len %lu "
+		       "refs %d\n", (unsigned long long)eb->start,
+		       eb->len, atomic_read(&eb->refs));
+		list_del(&eb->leak_list);
+		kmem_cache_free(extent_buffer_cache, eb);
+	}
+	if (extent_state_cache)
+		kmem_cache_destroy(extent_state_cache);
+	if (extent_buffer_cache)
+		kmem_cache_destroy(extent_buffer_cache);
+}
+
+void extent_io_tree_init(struct extent_io_tree *tree,
+			  struct address_space *mapping, gfp_t mask)
+{
+	tree->state.rb_node = NULL;
+	tree->buffer.rb_node = NULL;
+	tree->ops = NULL;
+	tree->dirty_bytes = 0;
+	spin_lock_init(&tree->lock);
+	spin_lock_init(&tree->buffer_lock);
+	tree->mapping = mapping;
+}
+
+static struct extent_state *alloc_extent_state(gfp_t mask)
+{
+	struct extent_state *state;
+#ifdef LEAK_DEBUG
+	unsigned long flags;
+#endif
+
+	state = kmem_cache_alloc(extent_state_cache, mask);
+	if (!state)
+		return state;
+	state->state = 0;
+	state->private = 0;
+	state->tree = NULL;
+#ifdef LEAK_DEBUG
+	spin_lock_irqsave(&leak_lock, flags);
+	list_add(&state->leak_list, &states);
+	spin_unlock_irqrestore(&leak_lock, flags);
+#endif
+	atomic_set(&state->refs, 1);
+	init_waitqueue_head(&state->wq);
+	return state;
+}
+
+static void free_extent_state(struct extent_state *state)
+{
+	if (!state)
+		return;
+	if (atomic_dec_and_test(&state->refs)) {
+#ifdef LEAK_DEBUG
+		unsigned long flags;
+#endif
+		WARN_ON(state->tree);
+#ifdef LEAK_DEBUG
+		spin_lock_irqsave(&leak_lock, flags);
+		list_del(&state->leak_list);
+		spin_unlock_irqrestore(&leak_lock, flags);
+#endif
+		kmem_cache_free(extent_state_cache, state);
+	}
+}
+
+static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
+				   struct rb_node *node)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct tree_entry *entry;
+
+	while (*p) {
+		parent = *p;
+		entry = rb_entry(parent, struct tree_entry, rb_node);
+
+		if (offset < entry->start)
+			p = &(*p)->rb_left;
+		else if (offset > entry->end)
+			p = &(*p)->rb_right;
+		else
+			return parent;
+	}
+
+	entry = rb_entry(node, struct tree_entry, rb_node);
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+	return NULL;
+}
+
+static struct rb_node *__etree_search(struct extent_io_tree *tree, u64 offset,
+				     struct rb_node **prev_ret,
+				     struct rb_node **next_ret)
+{
+	struct rb_root *root = &tree->state;
+	struct rb_node *n = root->rb_node;
+	struct rb_node *prev = NULL;
+	struct rb_node *orig_prev = NULL;
+	struct tree_entry *entry;
+	struct tree_entry *prev_entry = NULL;
+
+	while (n) {
+		entry = rb_entry(n, struct tree_entry, rb_node);
+		prev = n;
+		prev_entry = entry;
+
+		if (offset < entry->start)
+			n = n->rb_left;
+		else if (offset > entry->end)
+			n = n->rb_right;
+		else
+			return n;
+	}
+
+	if (prev_ret) {
+		orig_prev = prev;
+		while (prev && offset > prev_entry->end) {
+			prev = rb_next(prev);
+			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+		}
+		*prev_ret = prev;
+		prev = orig_prev;
+	}
+
+	if (next_ret) {
+		prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+		while (prev && offset < prev_entry->start) {
+			prev = rb_prev(prev);
+			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+		}
+		*next_ret = prev;
+	}
+	return NULL;
+}
+
+static inline struct rb_node *tree_search(struct extent_io_tree *tree,
+					  u64 offset)
+{
+	struct rb_node *prev = NULL;
+	struct rb_node *ret;
+
+	ret = __etree_search(tree, offset, &prev, NULL);
+	if (!ret)
+		return prev;
+	return ret;
+}
+
+static struct extent_buffer *buffer_tree_insert(struct extent_io_tree *tree,
+					  u64 offset, struct rb_node *node)
+{
+	struct rb_root *root = &tree->buffer;
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct extent_buffer *eb;
+
+	while (*p) {
+		parent = *p;
+		eb = rb_entry(parent, struct extent_buffer, rb_node);
+
+		if (offset < eb->start)
+			p = &(*p)->rb_left;
+		else if (offset > eb->start)
+			p = &(*p)->rb_right;
+		else
+			return eb;
+	}
+
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+	return NULL;
+}
+
+static struct extent_buffer *buffer_search(struct extent_io_tree *tree,
+					   u64 offset)
+{
+	struct rb_root *root = &tree->buffer;
+	struct rb_node *n = root->rb_node;
+	struct extent_buffer *eb;
+
+	while (n) {
+		eb = rb_entry(n, struct extent_buffer, rb_node);
+		if (offset < eb->start)
+			n = n->rb_left;
+		else if (offset > eb->start)
+			n = n->rb_right;
+		else
+			return eb;
+	}
+	return NULL;
+}
+
+/*
+ * utility function to look for merge candidates inside a given range.
+ * Any extents with matching state are merged together into a single
+ * extent in the tree.  Extents with EXTENT_IO in their state field
+ * are not merged because the end_io handlers need to be able to do
+ * operations on them without sleeping (or doing allocations/splits).
+ *
+ * This should be called with the tree lock held.
+ */
+static int merge_state(struct extent_io_tree *tree,
+		       struct extent_state *state)
+{
+	struct extent_state *other;
+	struct rb_node *other_node;
+
+	if (state->state & (EXTENT_IOBITS | EXTENT_BOUNDARY))
+		return 0;
+
+	other_node = rb_prev(&state->rb_node);
+	if (other_node) {
+		other = rb_entry(other_node, struct extent_state, rb_node);
+		if (other->end == state->start - 1 &&
+		    other->state == state->state) {
+			state->start = other->start;
+			other->tree = NULL;
+			rb_erase(&other->rb_node, &tree->state);
+			free_extent_state(other);
+		}
+	}
+	other_node = rb_next(&state->rb_node);
+	if (other_node) {
+		other = rb_entry(other_node, struct extent_state, rb_node);
+		if (other->start == state->end + 1 &&
+		    other->state == state->state) {
+			other->start = state->start;
+			state->tree = NULL;
+			rb_erase(&state->rb_node, &tree->state);
+			free_extent_state(state);
+		}
+	}
+	return 0;
+}
+
+static void set_state_cb(struct extent_io_tree *tree,
+			 struct extent_state *state,
+			 unsigned long bits)
+{
+	if (tree->ops && tree->ops->set_bit_hook) {
+		tree->ops->set_bit_hook(tree->mapping->host, state->start,
+					state->end, state->state, bits);
+	}
+}
+
+static void clear_state_cb(struct extent_io_tree *tree,
+			   struct extent_state *state,
+			   unsigned long bits)
+{
+	if (tree->ops && tree->ops->clear_bit_hook) {
+		tree->ops->clear_bit_hook(tree->mapping->host, state->start,
+					  state->end, state->state, bits);
+	}
+}
+
+/*
+ * insert an extent_state struct into the tree.  'bits' are set on the
+ * struct before it is inserted.
+ *
+ * This may return -EEXIST if the extent is already there, in which case the
+ * state struct is freed.
+ *
+ * The tree lock is not taken internally.  This is a utility function and
+ * probably isn't what you want to call (see set/clear_extent_bit).
+ */
+static int insert_state(struct extent_io_tree *tree,
+			struct extent_state *state, u64 start, u64 end,
+			int bits)
+{
+	struct rb_node *node;
+
+	if (end < start) {
+		printk(KERN_ERR "btrfs end < start %llu %llu\n",
+		       (unsigned long long)end,
+		       (unsigned long long)start);
+		WARN_ON(1);
+	}
+	if (bits & EXTENT_DIRTY)
+		tree->dirty_bytes += end - start + 1;
+	set_state_cb(tree, state, bits);
+	state->state |= bits;
+	state->start = start;
+	state->end = end;
+	node = tree_insert(&tree->state, end, &state->rb_node);
+	if (node) {
+		struct extent_state *found;
+		found = rb_entry(node, struct extent_state, rb_node);
+		printk(KERN_ERR "btrfs found node %llu %llu on insert of "
+		       "%llu %llu\n", (unsigned long long)found->start,
+		       (unsigned long long)found->end,
+		       (unsigned long long)start, (unsigned long long)end);
+		free_extent_state(state);
+		return -EEXIST;
+	}
+	state->tree = tree;
+	merge_state(tree, state);
+	return 0;
+}
+
+/*
+ * split a given extent state struct in two, inserting the preallocated
+ * struct 'prealloc' as the newly created second half.  'split' indicates an
+ * offset inside 'orig' where it should be split.
+ *
+ * Before calling,
+ * the tree has 'orig' at [orig->start, orig->end].  After calling, there
+ * are two extent state structs in the tree:
+ * prealloc: [orig->start, split - 1]
+ * orig: [ split, orig->end ]
+ *
+ * The tree locks are not taken by this function. They need to be held
+ * by the caller.
+ */
+static int split_state(struct extent_io_tree *tree, struct extent_state *orig,
+		       struct extent_state *prealloc, u64 split)
+{
+	struct rb_node *node;
+	prealloc->start = orig->start;
+	prealloc->end = split - 1;
+	prealloc->state = orig->state;
+	orig->start = split;
+
+	node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node);
+	if (node) {
+		struct extent_state *found;
+		found = rb_entry(node, struct extent_state, rb_node);
+		free_extent_state(prealloc);
+		return -EEXIST;
+	}
+	prealloc->tree = tree;
+	return 0;
+}
+
+/*
+ * utility function to clear some bits in an extent state struct.
+ * it will optionally wake up any one waiting on this state (wake == 1), or
+ * forcibly remove the state from the tree (delete == 1).
+ *
+ * If no bits are set on the state struct after clearing things, the
+ * struct is freed and removed from the tree
+ */
+static int clear_state_bit(struct extent_io_tree *tree,
+			    struct extent_state *state, int bits, int wake,
+			    int delete)
+{
+	int ret = state->state & bits;
+
+	if ((bits & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
+		u64 range = state->end - state->start + 1;
+		WARN_ON(range > tree->dirty_bytes);
+		tree->dirty_bytes -= range;
+	}
+	clear_state_cb(tree, state, bits);
+	state->state &= ~bits;
+	if (wake)
+		wake_up(&state->wq);
+	if (delete || state->state == 0) {
+		if (state->tree) {
+			clear_state_cb(tree, state, state->state);
+			rb_erase(&state->rb_node, &tree->state);
+			state->tree = NULL;
+			free_extent_state(state);
+		} else {
+			WARN_ON(1);
+		}
+	} else {
+		merge_state(tree, state);
+	}
+	return ret;
+}
+
+/*
+ * clear some bits on a range in the tree.  This may require splitting
+ * or inserting elements in the tree, so the gfp mask is used to
+ * indicate which allocations or sleeping are allowed.
+ *
+ * pass 'wake' == 1 to kick any sleepers, and 'delete' == 1 to remove
+ * the given range from the tree regardless of state (ie for truncate).
+ *
+ * the range [start, end] is inclusive.
+ *
+ * This takes the tree lock, and returns < 0 on error, > 0 if any of the
+ * bits were already set, or zero if none of the bits were already set.
+ */
+int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		     int bits, int wake, int delete, gfp_t mask)
+{
+	struct extent_state *state;
+	struct extent_state *prealloc = NULL;
+	struct rb_node *node;
+	int err;
+	int set = 0;
+
+again:
+	if (!prealloc && (mask & __GFP_WAIT)) {
+		prealloc = alloc_extent_state(mask);
+		if (!prealloc)
+			return -ENOMEM;
+	}
+
+	spin_lock(&tree->lock);
+	/*
+	 * this search will find the extents that end after
+	 * our range starts
+	 */
+	node = tree_search(tree, start);
+	if (!node)
+		goto out;
+	state = rb_entry(node, struct extent_state, rb_node);
+	if (state->start > end)
+		goto out;
+	WARN_ON(state->end < start);
+
+	/*
+	 *     | ---- desired range ---- |
+	 *  | state | or
+	 *  | ------------- state -------------- |
+	 *
+	 * We need to split the extent we found, and may flip
+	 * bits on second half.
+	 *
+	 * If the extent we found extends past our range, we
+	 * just split and search again.  It'll get split again
+	 * the next time though.
+	 *
+	 * If the extent we found is inside our range, we clear
+	 * the desired bit on it.
+	 */
+
+	if (state->start < start) {
+		if (!prealloc)
+			prealloc = alloc_extent_state(GFP_ATOMIC);
+		err = split_state(tree, state, prealloc, start);
+		BUG_ON(err == -EEXIST);
+		prealloc = NULL;
+		if (err)
+			goto out;
+		if (state->end <= end) {
+			start = state->end + 1;
+			set |= clear_state_bit(tree, state, bits,
+					wake, delete);
+		} else {
+			start = state->start;
+		}
+		goto search_again;
+	}
+	/*
+	 * | ---- desired range ---- |
+	 *                        | state |
+	 * We need to split the extent, and clear the bit
+	 * on the first half
+	 */
+	if (state->start <= end && state->end > end) {
+		if (!prealloc)
+			prealloc = alloc_extent_state(GFP_ATOMIC);
+		err = split_state(tree, state, prealloc, end + 1);
+		BUG_ON(err == -EEXIST);
+
+		if (wake)
+			wake_up(&state->wq);
+		set |= clear_state_bit(tree, prealloc, bits,
+				       wake, delete);
+		prealloc = NULL;
+		goto out;
+	}
+
+	start = state->end + 1;
+	set |= clear_state_bit(tree, state, bits, wake, delete);
+	goto search_again;
+
+out:
+	spin_unlock(&tree->lock);
+	if (prealloc)
+		free_extent_state(prealloc);
+
+	return set;
+
+search_again:
+	if (start > end)
+		goto out;
+	spin_unlock(&tree->lock);
+	if (mask & __GFP_WAIT)
+		cond_resched();
+	goto again;
+}
+
+static int wait_on_state(struct extent_io_tree *tree,
+			 struct extent_state *state)
+		__releases(tree->lock)
+		__acquires(tree->lock)
+{
+	DEFINE_WAIT(wait);
+	prepare_to_wait(&state->wq, &wait, TASK_UNINTERRUPTIBLE);
+	spin_unlock(&tree->lock);
+	schedule();
+	spin_lock(&tree->lock);
+	finish_wait(&state->wq, &wait);
+	return 0;
+}
+
+/*
+ * waits for one or more bits to clear on a range in the state tree.
+ * The range [start, end] is inclusive.
+ * The tree lock is taken by this function
+ */
+int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits)
+{
+	struct extent_state *state;
+	struct rb_node *node;
+
+	spin_lock(&tree->lock);
+again:
+	while (1) {
+		/*
+		 * this search will find all the extents that end after
+		 * our range starts
+		 */
+		node = tree_search(tree, start);
+		if (!node)
+			break;
+
+		state = rb_entry(node, struct extent_state, rb_node);
+
+		if (state->start > end)
+			goto out;
+
+		if (state->state & bits) {
+			start = state->start;
+			atomic_inc(&state->refs);
+			wait_on_state(tree, state);
+			free_extent_state(state);
+			goto again;
+		}
+		start = state->end + 1;
+
+		if (start > end)
+			break;
+
+		if (need_resched()) {
+			spin_unlock(&tree->lock);
+			cond_resched();
+			spin_lock(&tree->lock);
+		}
+	}
+out:
+	spin_unlock(&tree->lock);
+	return 0;
+}
+
+static void set_state_bits(struct extent_io_tree *tree,
+			   struct extent_state *state,
+			   int bits)
+{
+	if ((bits & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
+		u64 range = state->end - state->start + 1;
+		tree->dirty_bytes += range;
+	}
+	set_state_cb(tree, state, bits);
+	state->state |= bits;
+}
+
+/*
+ * set some bits on a range in the tree.  This may require allocations
+ * or sleeping, so the gfp mask is used to indicate what is allowed.
+ *
+ * If 'exclusive' == 1, this will fail with -EEXIST if some part of the
+ * range already has the desired bits set.  The start of the existing
+ * range is returned in failed_start in this case.
+ *
+ * [start, end] is inclusive
+ * This takes the tree lock.
+ */
+static int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+			  int bits, int exclusive, u64 *failed_start,
+			  gfp_t mask)
+{
+	struct extent_state *state;
+	struct extent_state *prealloc = NULL;
+	struct rb_node *node;
+	int err = 0;
+	int set;
+	u64 last_start;
+	u64 last_end;
+again:
+	if (!prealloc && (mask & __GFP_WAIT)) {
+		prealloc = alloc_extent_state(mask);
+		if (!prealloc)
+			return -ENOMEM;
+	}
+
+	spin_lock(&tree->lock);
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, start);
+	if (!node) {
+		err = insert_state(tree, prealloc, start, end, bits);
+		prealloc = NULL;
+		BUG_ON(err == -EEXIST);
+		goto out;
+	}
+
+	state = rb_entry(node, struct extent_state, rb_node);
+	last_start = state->start;
+	last_end = state->end;
+
+	/*
+	 * | ---- desired range ---- |
+	 * | state |
+	 *
+	 * Just lock what we found and keep going
+	 */
+	if (state->start == start && state->end <= end) {
+		set = state->state & bits;
+		if (set && exclusive) {
+			*failed_start = state->start;
+			err = -EEXIST;
+			goto out;
+		}
+		set_state_bits(tree, state, bits);
+		start = state->end + 1;
+		merge_state(tree, state);
+		goto search_again;
+	}
+
+	/*
+	 *     | ---- desired range ---- |
+	 * | state |
+	 *   or
+	 * | ------------- state -------------- |
+	 *
+	 * We need to split the extent we found, and may flip bits on
+	 * second half.
+	 *
+	 * If the extent we found extends past our
+	 * range, we just split and search again.  It'll get split
+	 * again the next time though.
+	 *
+	 * If the extent we found is inside our range, we set the
+	 * desired bit on it.
+	 */
+	if (state->start < start) {
+		set = state->state & bits;
+		if (exclusive && set) {
+			*failed_start = start;
+			err = -EEXIST;
+			goto out;
+		}
+		err = split_state(tree, state, prealloc, start);
+		BUG_ON(err == -EEXIST);
+		prealloc = NULL;
+		if (err)
+			goto out;
+		if (state->end <= end) {
+			set_state_bits(tree, state, bits);
+			start = state->end + 1;
+			merge_state(tree, state);
+		} else {
+			start = state->start;
+		}
+		goto search_again;
+	}
+	/*
+	 * | ---- desired range ---- |
+	 *     | state | or               | state |
+	 *
+	 * There's a hole, we need to insert something in it and
+	 * ignore the extent we found.
+	 */
+	if (state->start > start) {
+		u64 this_end;
+		if (end < last_start)
+			this_end = end;
+		else
+			this_end = last_start - 1;
+		err = insert_state(tree, prealloc, start, this_end,
+				   bits);
+		prealloc = NULL;
+		BUG_ON(err == -EEXIST);
+		if (err)
+			goto out;
+		start = this_end + 1;
+		goto search_again;
+	}
+	/*
+	 * | ---- desired range ---- |
+	 *                        | state |
+	 * We need to split the extent, and set the bit
+	 * on the first half
+	 */
+	if (state->start <= end && state->end > end) {
+		set = state->state & bits;
+		if (exclusive && set) {
+			*failed_start = start;
+			err = -EEXIST;
+			goto out;
+		}
+		err = split_state(tree, state, prealloc, end + 1);
+		BUG_ON(err == -EEXIST);
+
+		set_state_bits(tree, prealloc, bits);
+		merge_state(tree, prealloc);
+		prealloc = NULL;
+		goto out;
+	}
+
+	goto search_again;
+
+out:
+	spin_unlock(&tree->lock);
+	if (prealloc)
+		free_extent_state(prealloc);
+
+	return err;
+
+search_again:
+	if (start > end)
+		goto out;
+	spin_unlock(&tree->lock);
+	if (mask & __GFP_WAIT)
+		cond_resched();
+	goto again;
+}
+
+/* wrappers around set/clear extent bit */
+int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
+		     gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
+			      mask);
+}
+
+int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
+		       gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_ORDERED, 0, NULL, mask);
+}
+
+int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+		    int bits, gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, bits, 0, NULL,
+			      mask);
+}
+
+int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+		      int bits, gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, bits, 0, 0, mask);
+}
+
+int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
+		     gfp_t mask)
+{
+	return set_extent_bit(tree, start, end,
+			      EXTENT_DELALLOC | EXTENT_DIRTY,
+			      0, NULL, mask);
+}
+
+int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
+		       gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end,
+				EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, mask);
+}
+
+int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
+			 gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_ORDERED, 1, 0, mask);
+}
+
+int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
+		     gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL,
+			      mask);
+}
+
+static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
+		       gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, mask);
+}
+
+int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
+			gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
+			      mask);
+}
+
+static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
+				 u64 end, gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0, mask);
+}
+
+static int set_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end,
+			 gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_WRITEBACK,
+			      0, NULL, mask);
+}
+
+static int clear_extent_writeback(struct extent_io_tree *tree, u64 start,
+				  u64 end, gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_WRITEBACK, 1, 0, mask);
+}
+
+int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end)
+{
+	return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
+}
+
+/*
+ * either insert or lock state struct between start and end use mask to tell
+ * us if waiting is desired.
+ */
+int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
+{
+	int err;
+	u64 failed_start;
+	while (1) {
+		err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
+				     &failed_start, mask);
+		if (err == -EEXIST && (mask & __GFP_WAIT)) {
+			wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
+			start = failed_start;
+		} else {
+			break;
+		}
+		WARN_ON(start > end);
+	}
+	return err;
+}
+
+int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
+		    gfp_t mask)
+{
+	int err;
+	u64 failed_start;
+
+	err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
+			     &failed_start, mask);
+	if (err == -EEXIST) {
+		if (failed_start > start)
+			clear_extent_bit(tree, start, failed_start - 1,
+					 EXTENT_LOCKED, 1, 0, mask);
+		return 0;
+	}
+	return 1;
+}
+
+int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end,
+		  gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, mask);
+}
+
+/*
+ * helper function to set pages and extents in the tree dirty
+ */
+int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+
+	while (index <= end_index) {
+		page = find_get_page(tree->mapping, index);
+		BUG_ON(!page);
+		__set_page_dirty_nobuffers(page);
+		page_cache_release(page);
+		index++;
+	}
+	set_extent_dirty(tree, start, end, GFP_NOFS);
+	return 0;
+}
+
+/*
+ * helper function to set both pages and extents in the tree writeback
+ */
+static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+
+	while (index <= end_index) {
+		page = find_get_page(tree->mapping, index);
+		BUG_ON(!page);
+		set_page_writeback(page);
+		page_cache_release(page);
+		index++;
+	}
+	set_extent_writeback(tree, start, end, GFP_NOFS);
+	return 0;
+}
+
+/*
+ * find the first offset in the io tree with 'bits' set. zero is
+ * returned if we find something, and *start_ret and *end_ret are
+ * set to reflect the state struct that was found.
+ *
+ * If nothing was found, 1 is returned, < 0 on error
+ */
+int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
+			  u64 *start_ret, u64 *end_ret, int bits)
+{
+	struct rb_node *node;
+	struct extent_state *state;
+	int ret = 1;
+
+	spin_lock(&tree->lock);
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, start);
+	if (!node)
+		goto out;
+
+	while (1) {
+		state = rb_entry(node, struct extent_state, rb_node);
+		if (state->end >= start && (state->state & bits)) {
+			*start_ret = state->start;
+			*end_ret = state->end;
+			ret = 0;
+			break;
+		}
+		node = rb_next(node);
+		if (!node)
+			break;
+	}
+out:
+	spin_unlock(&tree->lock);
+	return ret;
+}
+
+/* find the first state struct with 'bits' set after 'start', and
+ * return it.  tree->lock must be held.  NULL will returned if
+ * nothing was found after 'start'
+ */
+struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree,
+						 u64 start, int bits)
+{
+	struct rb_node *node;
+	struct extent_state *state;
+
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, start);
+	if (!node)
+		goto out;
+
+	while (1) {
+		state = rb_entry(node, struct extent_state, rb_node);
+		if (state->end >= start && (state->state & bits))
+			return state;
+
+		node = rb_next(node);
+		if (!node)
+			break;
+	}
+out:
+	return NULL;
+}
+
+/*
+ * find a contiguous range of bytes in the file marked as delalloc, not
+ * more than 'max_bytes'.  start and end are used to return the range,
+ *
+ * 1 is returned if we find something, 0 if nothing was in the tree
+ */
+static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
+					u64 *start, u64 *end, u64 max_bytes)
+{
+	struct rb_node *node;
+	struct extent_state *state;
+	u64 cur_start = *start;
+	u64 found = 0;
+	u64 total_bytes = 0;
+
+	spin_lock(&tree->lock);
+
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, cur_start);
+	if (!node) {
+		if (!found)
+			*end = (u64)-1;
+		goto out;
+	}
+
+	while (1) {
+		state = rb_entry(node, struct extent_state, rb_node);
+		if (found && (state->start != cur_start ||
+			      (state->state & EXTENT_BOUNDARY))) {
+			goto out;
+		}
+		if (!(state->state & EXTENT_DELALLOC)) {
+			if (!found)
+				*end = state->end;
+			goto out;
+		}
+		if (!found)
+			*start = state->start;
+		found++;
+		*end = state->end;
+		cur_start = state->end + 1;
+		node = rb_next(node);
+		if (!node)
+			break;
+		total_bytes += state->end - state->start + 1;
+		if (total_bytes >= max_bytes)
+			break;
+	}
+out:
+	spin_unlock(&tree->lock);
+	return found;
+}
+
+static noinline int __unlock_for_delalloc(struct inode *inode,
+					  struct page *locked_page,
+					  u64 start, u64 end)
+{
+	int ret;
+	struct page *pages[16];
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	unsigned long nr_pages = end_index - index + 1;
+	int i;
+
+	if (index == locked_page->index && end_index == index)
+		return 0;
+
+	while (nr_pages > 0) {
+		ret = find_get_pages_contig(inode->i_mapping, index,
+				     min_t(unsigned long, nr_pages,
+				     ARRAY_SIZE(pages)), pages);
+		for (i = 0; i < ret; i++) {
+			if (pages[i] != locked_page)
+				unlock_page(pages[i]);
+			page_cache_release(pages[i]);
+		}
+		nr_pages -= ret;
+		index += ret;
+		cond_resched();
+	}
+	return 0;
+}
+
+static noinline int lock_delalloc_pages(struct inode *inode,
+					struct page *locked_page,
+					u64 delalloc_start,
+					u64 delalloc_end)
+{
+	unsigned long index = delalloc_start >> PAGE_CACHE_SHIFT;
+	unsigned long start_index = index;
+	unsigned long end_index = delalloc_end >> PAGE_CACHE_SHIFT;
+	unsigned long pages_locked = 0;
+	struct page *pages[16];
+	unsigned long nrpages;
+	int ret;
+	int i;
+
+	/* the caller is responsible for locking the start index */
+	if (index == locked_page->index && index == end_index)
+		return 0;
+
+	/* skip the page at the start index */
+	nrpages = end_index - index + 1;
+	while (nrpages > 0) {
+		ret = find_get_pages_contig(inode->i_mapping, index,
+				     min_t(unsigned long,
+				     nrpages, ARRAY_SIZE(pages)), pages);
+		if (ret == 0) {
+			ret = -EAGAIN;
+			goto done;
+		}
+		/* now we have an array of pages, lock them all */
+		for (i = 0; i < ret; i++) {
+			/*
+			 * the caller is taking responsibility for
+			 * locked_page
+			 */
+			if (pages[i] != locked_page) {
+				lock_page(pages[i]);
+				if (!PageDirty(pages[i]) ||
+				    pages[i]->mapping != inode->i_mapping) {
+					ret = -EAGAIN;
+					unlock_page(pages[i]);
+					page_cache_release(pages[i]);
+					goto done;
+				}
+			}
+			page_cache_release(pages[i]);
+			pages_locked++;
+		}
+		nrpages -= ret;
+		index += ret;
+		cond_resched();
+	}
+	ret = 0;
+done:
+	if (ret && pages_locked) {
+		__unlock_for_delalloc(inode, locked_page,
+			      delalloc_start,
+			      ((u64)(start_index + pages_locked - 1)) <<
+			      PAGE_CACHE_SHIFT);
+	}
+	return ret;
+}
+
+/*
+ * find a contiguous range of bytes in the file marked as delalloc, not
+ * more than 'max_bytes'.  start and end are used to return the range,
+ *
+ * 1 is returned if we find something, 0 if nothing was in the tree
+ */
+static noinline u64 find_lock_delalloc_range(struct inode *inode,
+					     struct extent_io_tree *tree,
+					     struct page *locked_page,
+					     u64 *start, u64 *end,
+					     u64 max_bytes)
+{
+	u64 delalloc_start;
+	u64 delalloc_end;
+	u64 found;
+	int ret;
+	int loops = 0;
+
+again:
+	/* step one, find a bunch of delalloc bytes starting at start */
+	delalloc_start = *start;
+	delalloc_end = 0;
+	found = find_delalloc_range(tree, &delalloc_start, &delalloc_end,
+				    max_bytes);
+	if (!found || delalloc_end <= *start) {
+		*start = delalloc_start;
+		*end = delalloc_end;
+		return found;
+	}
+
+	/*
+	 * start comes from the offset of locked_page.  We have to lock
+	 * pages in order, so we can't process delalloc bytes before
+	 * locked_page
+	 */
+	if (delalloc_start < *start)
+		delalloc_start = *start;
+
+	/*
+	 * make sure to limit the number of pages we try to lock down
+	 * if we're looping.
+	 */
+	if (delalloc_end + 1 - delalloc_start > max_bytes && loops)
+		delalloc_end = delalloc_start + PAGE_CACHE_SIZE - 1;
+
+	/* step two, lock all the pages after the page that has start */
+	ret = lock_delalloc_pages(inode, locked_page,
+				  delalloc_start, delalloc_end);
+	if (ret == -EAGAIN) {
+		/* some of the pages are gone, lets avoid looping by
+		 * shortening the size of the delalloc range we're searching
+		 */
+		if (!loops) {
+			unsigned long offset = (*start) & (PAGE_CACHE_SIZE - 1);
+			max_bytes = PAGE_CACHE_SIZE - offset;
+			loops = 1;
+			goto again;
+		} else {
+			found = 0;
+			goto out_failed;
+		}
+	}
+	BUG_ON(ret);
+
+	/* step three, lock the state bits for the whole range */
+	lock_extent(tree, delalloc_start, delalloc_end, GFP_NOFS);
+
+	/* then test to make sure it is all still delalloc */
+	ret = test_range_bit(tree, delalloc_start, delalloc_end,
+			     EXTENT_DELALLOC, 1);
+	if (!ret) {
+		unlock_extent(tree, delalloc_start, delalloc_end, GFP_NOFS);
+		__unlock_for_delalloc(inode, locked_page,
+			      delalloc_start, delalloc_end);
+		cond_resched();
+		goto again;
+	}
+	*start = delalloc_start;
+	*end = delalloc_end;
+out_failed:
+	return found;
+}
+
+int extent_clear_unlock_delalloc(struct inode *inode,
+				struct extent_io_tree *tree,
+				u64 start, u64 end, struct page *locked_page,
+				int unlock_pages,
+				int clear_unlock,
+				int clear_delalloc, int clear_dirty,
+				int set_writeback,
+				int end_writeback)
+{
+	int ret;
+	struct page *pages[16];
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	unsigned long nr_pages = end_index - index + 1;
+	int i;
+	int clear_bits = 0;
+
+	if (clear_unlock)
+		clear_bits |= EXTENT_LOCKED;
+	if (clear_dirty)
+		clear_bits |= EXTENT_DIRTY;
+
+	if (clear_delalloc)
+		clear_bits |= EXTENT_DELALLOC;
+
+	clear_extent_bit(tree, start, end, clear_bits, 1, 0, GFP_NOFS);
+	if (!(unlock_pages || clear_dirty || set_writeback || end_writeback))
+		return 0;
+
+	while (nr_pages > 0) {
+		ret = find_get_pages_contig(inode->i_mapping, index,
+				     min_t(unsigned long,
+				     nr_pages, ARRAY_SIZE(pages)), pages);
+		for (i = 0; i < ret; i++) {
+			if (pages[i] == locked_page) {
+				page_cache_release(pages[i]);
+				continue;
+			}
+			if (clear_dirty)
+				clear_page_dirty_for_io(pages[i]);
+			if (set_writeback)
+				set_page_writeback(pages[i]);
+			if (end_writeback)
+				end_page_writeback(pages[i]);
+			if (unlock_pages)
+				unlock_page(pages[i]);
+			page_cache_release(pages[i]);
+		}
+		nr_pages -= ret;
+		index += ret;
+		cond_resched();
+	}
+	return 0;
+}
+
+/*
+ * count the number of bytes in the tree that have a given bit(s)
+ * set.  This can be fairly slow, except for EXTENT_DIRTY which is
+ * cached.  The total number found is returned.
+ */
+u64 count_range_bits(struct extent_io_tree *tree,
+		     u64 *start, u64 search_end, u64 max_bytes,
+		     unsigned long bits)
+{
+	struct rb_node *node;
+	struct extent_state *state;
+	u64 cur_start = *start;
+	u64 total_bytes = 0;
+	int found = 0;
+
+	if (search_end <= cur_start) {
+		WARN_ON(1);
+		return 0;
+	}
+
+	spin_lock(&tree->lock);
+	if (cur_start == 0 && bits == EXTENT_DIRTY) {
+		total_bytes = tree->dirty_bytes;
+		goto out;
+	}
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, cur_start);
+	if (!node)
+		goto out;
+
+	while (1) {
+		state = rb_entry(node, struct extent_state, rb_node);
+		if (state->start > search_end)
+			break;
+		if (state->end >= cur_start && (state->state & bits)) {
+			total_bytes += min(search_end, state->end) + 1 -
+				       max(cur_start, state->start);
+			if (total_bytes >= max_bytes)
+				break;
+			if (!found) {
+				*start = state->start;
+				found = 1;
+			}
+		}
+		node = rb_next(node);
+		if (!node)
+			break;
+	}
+out:
+	spin_unlock(&tree->lock);
+	return total_bytes;
+}
+
+#if 0
+/*
+ * helper function to lock both pages and extents in the tree.
+ * pages must be locked first.
+ */
+static int lock_range(struct extent_io_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+	int err;
+
+	while (index <= end_index) {
+		page = grab_cache_page(tree->mapping, index);
+		if (!page) {
+			err = -ENOMEM;
+			goto failed;
+		}
+		if (IS_ERR(page)) {
+			err = PTR_ERR(page);
+			goto failed;
+		}
+		index++;
+	}
+	lock_extent(tree, start, end, GFP_NOFS);
+	return 0;
+
+failed:
+	/*
+	 * we failed above in getting the page at 'index', so we undo here
+	 * up to but not including the page at 'index'
+	 */
+	end_index = index;
+	index = start >> PAGE_CACHE_SHIFT;
+	while (index < end_index) {
+		page = find_get_page(tree->mapping, index);
+		unlock_page(page);
+		page_cache_release(page);
+		index++;
+	}
+	return err;
+}
+
+/*
+ * helper function to unlock both pages and extents in the tree.
+ */
+static int unlock_range(struct extent_io_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+
+	while (index <= end_index) {
+		page = find_get_page(tree->mapping, index);
+		unlock_page(page);
+		page_cache_release(page);
+		index++;
+	}
+	unlock_extent(tree, start, end, GFP_NOFS);
+	return 0;
+}
+#endif
+
+/*
+ * set the private field for a given byte offset in the tree.  If there isn't
+ * an extent_state there already, this does nothing.
+ */
+int set_state_private(struct extent_io_tree *tree, u64 start, u64 private)
+{
+	struct rb_node *node;
+	struct extent_state *state;
+	int ret = 0;
+
+	spin_lock(&tree->lock);
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, start);
+	if (!node) {
+		ret = -ENOENT;
+		goto out;
+	}
+	state = rb_entry(node, struct extent_state, rb_node);
+	if (state->start != start) {
+		ret = -ENOENT;
+		goto out;
+	}
+	state->private = private;
+out:
+	spin_unlock(&tree->lock);
+	return ret;
+}
+
+int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private)
+{
+	struct rb_node *node;
+	struct extent_state *state;
+	int ret = 0;
+
+	spin_lock(&tree->lock);
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(tree, start);
+	if (!node) {
+		ret = -ENOENT;
+		goto out;
+	}
+	state = rb_entry(node, struct extent_state, rb_node);
+	if (state->start != start) {
+		ret = -ENOENT;
+		goto out;
+	}
+	*private = state->private;
+out:
+	spin_unlock(&tree->lock);
+	return ret;
+}
+
+/*
+ * searches a range in the state tree for a given mask.
+ * If 'filled' == 1, this returns 1 only if every extent in the tree
+ * has the bits set.  Otherwise, 1 is returned if any bit in the
+ * range is found set.
+ */
+int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		   int bits, int filled)
+{
+	struct extent_state *state = NULL;
+	struct rb_node *node;
+	int bitset = 0;
+
+	spin_lock(&tree->lock);
+	node = tree_search(tree, start);
+	while (node && start <= end) {
+		state = rb_entry(node, struct extent_state, rb_node);
+
+		if (filled && state->start > start) {
+			bitset = 0;
+			break;
+		}
+
+		if (state->start > end)
+			break;
+
+		if (state->state & bits) {
+			bitset = 1;
+			if (!filled)
+				break;
+		} else if (filled) {
+			bitset = 0;
+			break;
+		}
+		start = state->end + 1;
+		if (start > end)
+			break;
+		node = rb_next(node);
+		if (!node) {
+			if (filled)
+				bitset = 0;
+			break;
+		}
+	}
+	spin_unlock(&tree->lock);
+	return bitset;
+}
+
+/*
+ * helper function to set a given page up to date if all the
+ * extents in the tree for that page are up to date
+ */
+static int check_page_uptodate(struct extent_io_tree *tree,
+			       struct page *page)
+{
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+	if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1))
+		SetPageUptodate(page);
+	return 0;
+}
+
+/*
+ * helper function to unlock a page if all the extents in the tree
+ * for that page are unlocked
+ */
+static int check_page_locked(struct extent_io_tree *tree,
+			     struct page *page)
+{
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+	if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0))
+		unlock_page(page);
+	return 0;
+}
+
+/*
+ * helper function to end page writeback if all the extents
+ * in the tree for that page are done with writeback
+ */
+static int check_page_writeback(struct extent_io_tree *tree,
+			     struct page *page)
+{
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+	if (!test_range_bit(tree, start, end, EXTENT_WRITEBACK, 0))
+		end_page_writeback(page);
+	return 0;
+}
+
+/* lots and lots of room for performance fixes in the end_bio funcs */
+
+/*
+ * after a writepage IO is done, we need to:
+ * clear the uptodate bits on error
+ * clear the writeback bits in the extent tree for this IO
+ * end_page_writeback if the page has no more pending IO
+ *
+ * Scheduling is not allowed, so the extent state tree is expected
+ * to have one and only one object corresponding to this IO.
+ */
+static void end_bio_extent_writepage(struct bio *bio, int err)
+{
+	int uptodate = err == 0;
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct extent_io_tree *tree;
+	u64 start;
+	u64 end;
+	int whole_page;
+	int ret;
+
+	do {
+		struct page *page = bvec->bv_page;
+		tree = &BTRFS_I(page->mapping->host)->io_tree;
+
+		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
+			 bvec->bv_offset;
+		end = start + bvec->bv_len - 1;
+
+		if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
+			whole_page = 1;
+		else
+			whole_page = 0;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+		if (tree->ops && tree->ops->writepage_end_io_hook) {
+			ret = tree->ops->writepage_end_io_hook(page, start,
+						       end, NULL, uptodate);
+			if (ret)
+				uptodate = 0;
+		}
+
+		if (!uptodate && tree->ops &&
+		    tree->ops->writepage_io_failed_hook) {
+			ret = tree->ops->writepage_io_failed_hook(bio, page,
+							 start, end, NULL);
+			if (ret == 0) {
+				uptodate = (err == 0);
+				continue;
+			}
+		}
+
+		if (!uptodate) {
+			clear_extent_uptodate(tree, start, end, GFP_ATOMIC);
+			ClearPageUptodate(page);
+			SetPageError(page);
+		}
+
+		clear_extent_writeback(tree, start, end, GFP_ATOMIC);
+
+		if (whole_page)
+			end_page_writeback(page);
+		else
+			check_page_writeback(tree, page);
+	} while (bvec >= bio->bi_io_vec);
+
+	bio_put(bio);
+}
+
+/*
+ * after a readpage IO is done, we need to:
+ * clear the uptodate bits on error
+ * set the uptodate bits if things worked
+ * set the page up to date if all extents in the tree are uptodate
+ * clear the lock bit in the extent tree
+ * unlock the page if there are no other extents locked for it
+ *
+ * Scheduling is not allowed, so the extent state tree is expected
+ * to have one and only one object corresponding to this IO.
+ */
+static void end_bio_extent_readpage(struct bio *bio, int err)
+{
+	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct extent_io_tree *tree;
+	u64 start;
+	u64 end;
+	int whole_page;
+	int ret;
+
+	if (err)
+		uptodate = 0;
+
+	do {
+		struct page *page = bvec->bv_page;
+		tree = &BTRFS_I(page->mapping->host)->io_tree;
+
+		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
+			bvec->bv_offset;
+		end = start + bvec->bv_len - 1;
+
+		if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
+			whole_page = 1;
+		else
+			whole_page = 0;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+
+		if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
+			ret = tree->ops->readpage_end_io_hook(page, start, end,
+							      NULL);
+			if (ret)
+				uptodate = 0;
+		}
+		if (!uptodate && tree->ops &&
+		    tree->ops->readpage_io_failed_hook) {
+			ret = tree->ops->readpage_io_failed_hook(bio, page,
+							 start, end, NULL);
+			if (ret == 0) {
+				uptodate =
+					test_bit(BIO_UPTODATE, &bio->bi_flags);
+				if (err)
+					uptodate = 0;
+				continue;
+			}
+		}
+
+		if (uptodate) {
+			set_extent_uptodate(tree, start, end,
+					    GFP_ATOMIC);
+		}
+		unlock_extent(tree, start, end, GFP_ATOMIC);
+
+		if (whole_page) {
+			if (uptodate) {
+				SetPageUptodate(page);
+			} else {
+				ClearPageUptodate(page);
+				SetPageError(page);
+			}
+			unlock_page(page);
+		} else {
+			if (uptodate) {
+				check_page_uptodate(tree, page);
+			} else {
+				ClearPageUptodate(page);
+				SetPageError(page);
+			}
+			check_page_locked(tree, page);
+		}
+	} while (bvec >= bio->bi_io_vec);
+
+	bio_put(bio);
+}
+
+/*
+ * IO done from prepare_write is pretty simple, we just unlock
+ * the structs in the extent tree when done, and set the uptodate bits
+ * as appropriate.
+ */
+static void end_bio_extent_preparewrite(struct bio *bio, int err)
+{
+	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct extent_io_tree *tree;
+	u64 start;
+	u64 end;
+
+	do {
+		struct page *page = bvec->bv_page;
+		tree = &BTRFS_I(page->mapping->host)->io_tree;
+
+		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
+			bvec->bv_offset;
+		end = start + bvec->bv_len - 1;
+
+		if (--bvec >= bio->bi_io_vec)
+			prefetchw(&bvec->bv_page->flags);
+
+		if (uptodate) {
+			set_extent_uptodate(tree, start, end, GFP_ATOMIC);
+		} else {
+			ClearPageUptodate(page);
+			SetPageError(page);
+		}
+
+		unlock_extent(tree, start, end, GFP_ATOMIC);
+
+	} while (bvec >= bio->bi_io_vec);
+
+	bio_put(bio);
+}
+
+static struct bio *
+extent_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
+		 gfp_t gfp_flags)
+{
+	struct bio *bio;
+
+	bio = bio_alloc(gfp_flags, nr_vecs);
+
+	if (bio == NULL && (current->flags & PF_MEMALLOC)) {
+		while (!bio && (nr_vecs /= 2))
+			bio = bio_alloc(gfp_flags, nr_vecs);
+	}
+
+	if (bio) {
+		bio->bi_size = 0;
+		bio->bi_bdev = bdev;
+		bio->bi_sector = first_sector;
+	}
+	return bio;
+}
+
+static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
+			  unsigned long bio_flags)
+{
+	int ret = 0;
+	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
+	struct page *page = bvec->bv_page;
+	struct extent_io_tree *tree = bio->bi_private;
+	u64 start;
+	u64 end;
+
+	start = ((u64)page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
+	end = start + bvec->bv_len - 1;
+
+	bio->bi_private = NULL;
+
+	bio_get(bio);
+
+	if (tree->ops && tree->ops->submit_bio_hook)
+		tree->ops->submit_bio_hook(page->mapping->host, rw, bio,
+					   mirror_num, bio_flags);
+	else
+		submit_bio(rw, bio);
+	if (bio_flagged(bio, BIO_EOPNOTSUPP))
+		ret = -EOPNOTSUPP;
+	bio_put(bio);
+	return ret;
+}
+
+static int submit_extent_page(int rw, struct extent_io_tree *tree,
+			      struct page *page, sector_t sector,
+			      size_t size, unsigned long offset,
+			      struct block_device *bdev,
+			      struct bio **bio_ret,
+			      unsigned long max_pages,
+			      bio_end_io_t end_io_func,
+			      int mirror_num,
+			      unsigned long prev_bio_flags,
+			      unsigned long bio_flags)
+{
+	int ret = 0;
+	struct bio *bio;
+	int nr;
+	int contig = 0;
+	int this_compressed = bio_flags & EXTENT_BIO_COMPRESSED;
+	int old_compressed = prev_bio_flags & EXTENT_BIO_COMPRESSED;
+	size_t page_size = min_t(size_t, size, PAGE_CACHE_SIZE);
+
+	if (bio_ret && *bio_ret) {
+		bio = *bio_ret;
+		if (old_compressed)
+			contig = bio->bi_sector == sector;
+		else
+			contig = bio->bi_sector + (bio->bi_size >> 9) ==
+				sector;
+
+		if (prev_bio_flags != bio_flags || !contig ||
+		    (tree->ops && tree->ops->merge_bio_hook &&
+		     tree->ops->merge_bio_hook(page, offset, page_size, bio,
+					       bio_flags)) ||
+		    bio_add_page(bio, page, page_size, offset) < page_size) {
+			ret = submit_one_bio(rw, bio, mirror_num,
+					     prev_bio_flags);
+			bio = NULL;
+		} else {
+			return 0;
+		}
+	}
+	if (this_compressed)
+		nr = BIO_MAX_PAGES;
+	else
+		nr = bio_get_nr_vecs(bdev);
+
+	bio = extent_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH);
+
+	bio_add_page(bio, page, page_size, offset);
+	bio->bi_end_io = end_io_func;
+	bio->bi_private = tree;
+
+	if (bio_ret)
+		*bio_ret = bio;
+	else
+		ret = submit_one_bio(rw, bio, mirror_num, bio_flags);
+
+	return ret;
+}
+
+void set_page_extent_mapped(struct page *page)
+{
+	if (!PagePrivate(page)) {
+		SetPagePrivate(page);
+		page_cache_get(page);
+		set_page_private(page, EXTENT_PAGE_PRIVATE);
+	}
+}
+
+static void set_page_extent_head(struct page *page, unsigned long len)
+{
+	set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2);
+}
+
+/*
+ * basic readpage implementation.  Locked extent state structs are inserted
+ * into the tree that are removed when the IO is done (by the end_io
+ * handlers)
+ */
+static int __extent_read_full_page(struct extent_io_tree *tree,
+				   struct page *page,
+				   get_extent_t *get_extent,
+				   struct bio **bio, int mirror_num,
+				   unsigned long *bio_flags)
+{
+	struct inode *inode = page->mapping->host;
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 page_end = start + PAGE_CACHE_SIZE - 1;
+	u64 end;
+	u64 cur = start;
+	u64 extent_offset;
+	u64 last_byte = i_size_read(inode);
+	u64 block_start;
+	u64 cur_end;
+	sector_t sector;
+	struct extent_map *em;
+	struct block_device *bdev;
+	int ret;
+	int nr = 0;
+	size_t page_offset = 0;
+	size_t iosize;
+	size_t disk_io_size;
+	size_t blocksize = inode->i_sb->s_blocksize;
+	unsigned long this_bio_flag = 0;
+
+	set_page_extent_mapped(page);
+
+	end = page_end;
+	lock_extent(tree, start, end, GFP_NOFS);
+
+	if (page->index == last_byte >> PAGE_CACHE_SHIFT) {
+		char *userpage;
+		size_t zero_offset = last_byte & (PAGE_CACHE_SIZE - 1);
+
+		if (zero_offset) {
+			iosize = PAGE_CACHE_SIZE - zero_offset;
+			userpage = kmap_atomic(page, KM_USER0);
+			memset(userpage + zero_offset, 0, iosize);
+			flush_dcache_page(page);
+			kunmap_atomic(userpage, KM_USER0);
+		}
+	}
+	while (cur <= end) {
+		if (cur >= last_byte) {
+			char *userpage;
+			iosize = PAGE_CACHE_SIZE - page_offset;
+			userpage = kmap_atomic(page, KM_USER0);
+			memset(userpage + page_offset, 0, iosize);
+			flush_dcache_page(page);
+			kunmap_atomic(userpage, KM_USER0);
+			set_extent_uptodate(tree, cur, cur + iosize - 1,
+					    GFP_NOFS);
+			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+			break;
+		}
+		em = get_extent(inode, page, page_offset, cur,
+				end - cur + 1, 0);
+		if (IS_ERR(em) || !em) {
+			SetPageError(page);
+			unlock_extent(tree, cur, end, GFP_NOFS);
+			break;
+		}
+		extent_offset = cur - em->start;
+		BUG_ON(extent_map_end(em) <= cur);
+		BUG_ON(end < cur);
+
+		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
+			this_bio_flag = EXTENT_BIO_COMPRESSED;
+
+		iosize = min(extent_map_end(em) - cur, end - cur + 1);
+		cur_end = min(extent_map_end(em) - 1, end);
+		iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
+		if (this_bio_flag & EXTENT_BIO_COMPRESSED) {
+			disk_io_size = em->block_len;
+			sector = em->block_start >> 9;
+		} else {
+			sector = (em->block_start + extent_offset) >> 9;
+			disk_io_size = iosize;
+		}
+		bdev = em->bdev;
+		block_start = em->block_start;
+		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
+			block_start = EXTENT_MAP_HOLE;
+		free_extent_map(em);
+		em = NULL;
+
+		/* we've found a hole, just zero and go on */
+		if (block_start == EXTENT_MAP_HOLE) {
+			char *userpage;
+			userpage = kmap_atomic(page, KM_USER0);
+			memset(userpage + page_offset, 0, iosize);
+			flush_dcache_page(page);
+			kunmap_atomic(userpage, KM_USER0);
+
+			set_extent_uptodate(tree, cur, cur + iosize - 1,
+					    GFP_NOFS);
+			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+			cur = cur + iosize;
+			page_offset += iosize;
+			continue;
+		}
+		/* the get_extent function already copied into the page */
+		if (test_range_bit(tree, cur, cur_end, EXTENT_UPTODATE, 1)) {
+			check_page_uptodate(tree, page);
+			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+			cur = cur + iosize;
+			page_offset += iosize;
+			continue;
+		}
+		/* we have an inline extent but it didn't get marked up
+		 * to date.  Error out
+		 */
+		if (block_start == EXTENT_MAP_INLINE) {
+			SetPageError(page);
+			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
+			cur = cur + iosize;
+			page_offset += iosize;
+			continue;
+		}
+
+		ret = 0;
+		if (tree->ops && tree->ops->readpage_io_hook) {
+			ret = tree->ops->readpage_io_hook(page, cur,
+							  cur + iosize - 1);
+		}
+		if (!ret) {
+			unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
+			pnr -= page->index;
+			ret = submit_extent_page(READ, tree, page,
+					 sector, disk_io_size, page_offset,
+					 bdev, bio, pnr,
+					 end_bio_extent_readpage, mirror_num,
+					 *bio_flags,
+					 this_bio_flag);
+			nr++;
+			*bio_flags = this_bio_flag;
+		}
+		if (ret)
+			SetPageError(page);
+		cur = cur + iosize;
+		page_offset += iosize;
+	}
+	if (!nr) {
+		if (!PageError(page))
+			SetPageUptodate(page);
+		unlock_page(page);
+	}
+	return 0;
+}
+
+int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
+			    get_extent_t *get_extent)
+{
+	struct bio *bio = NULL;
+	unsigned long bio_flags = 0;
+	int ret;
+
+	ret = __extent_read_full_page(tree, page, get_extent, &bio, 0,
+				      &bio_flags);
+	if (bio)
+		submit_one_bio(READ, bio, 0, bio_flags);
+	return ret;
+}
+
+/*
+ * the writepage semantics are similar to regular writepage.  extent
+ * records are inserted to lock ranges in the tree, and as dirty areas
+ * are found, they are marked writeback.  Then the lock bits are removed
+ * and the end_io handler clears the writeback ranges
+ */
+static int __extent_writepage(struct page *page, struct writeback_control *wbc,
+			      void *data)
+{
+	struct inode *inode = page->mapping->host;
+	struct extent_page_data *epd = data;
+	struct extent_io_tree *tree = epd->tree;
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 delalloc_start;
+	u64 page_end = start + PAGE_CACHE_SIZE - 1;
+	u64 end;
+	u64 cur = start;
+	u64 extent_offset;
+	u64 last_byte = i_size_read(inode);
+	u64 block_start;
+	u64 iosize;
+	u64 unlock_start;
+	sector_t sector;
+	struct extent_map *em;
+	struct block_device *bdev;
+	int ret;
+	int nr = 0;
+	size_t pg_offset = 0;
+	size_t blocksize;
+	loff_t i_size = i_size_read(inode);
+	unsigned long end_index = i_size >> PAGE_CACHE_SHIFT;
+	u64 nr_delalloc;
+	u64 delalloc_end;
+	int page_started;
+	int compressed;
+	unsigned long nr_written = 0;
+
+	WARN_ON(!PageLocked(page));
+	pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
+	if (page->index > end_index ||
+	   (page->index == end_index && !pg_offset)) {
+		page->mapping->a_ops->invalidatepage(page, 0);
+		unlock_page(page);
+		return 0;
+	}
+
+	if (page->index == end_index) {
+		char *userpage;
+
+		userpage = kmap_atomic(page, KM_USER0);
+		memset(userpage + pg_offset, 0,
+		       PAGE_CACHE_SIZE - pg_offset);
+		kunmap_atomic(userpage, KM_USER0);
+		flush_dcache_page(page);
+	}
+	pg_offset = 0;
+
+	set_page_extent_mapped(page);
+
+	delalloc_start = start;
+	delalloc_end = 0;
+	page_started = 0;
+	if (!epd->extent_locked) {
+		while (delalloc_end < page_end) {
+			nr_delalloc = find_lock_delalloc_range(inode, tree,
+						       page,
+						       &delalloc_start,
+						       &delalloc_end,
+						       128 * 1024 * 1024);
+			if (nr_delalloc == 0) {
+				delalloc_start = delalloc_end + 1;
+				continue;
+			}
+			tree->ops->fill_delalloc(inode, page, delalloc_start,
+						 delalloc_end, &page_started,
+						 &nr_written);
+			delalloc_start = delalloc_end + 1;
+		}
+
+		/* did the fill delalloc function already unlock and start
+		 * the IO?
+		 */
+		if (page_started) {
+			ret = 0;
+			goto update_nr_written;
+		}
+	}
+	lock_extent(tree, start, page_end, GFP_NOFS);
+
+	unlock_start = start;
+
+	if (tree->ops && tree->ops->writepage_start_hook) {
+		ret = tree->ops->writepage_start_hook(page, start,
+						      page_end);
+		if (ret == -EAGAIN) {
+			unlock_extent(tree, start, page_end, GFP_NOFS);
+			redirty_page_for_writepage(wbc, page);
+			unlock_page(page);
+			ret = 0;
+			goto update_nr_written;
+		}
+	}
+
+	nr_written++;
+
+	end = page_end;
+	if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0))
+		printk(KERN_ERR "btrfs delalloc bits after lock_extent\n");
+
+	if (last_byte <= start) {
+		clear_extent_dirty(tree, start, page_end, GFP_NOFS);
+		unlock_extent(tree, start, page_end, GFP_NOFS);
+		if (tree->ops && tree->ops->writepage_end_io_hook)
+			tree->ops->writepage_end_io_hook(page, start,
+							 page_end, NULL, 1);
+		unlock_start = page_end + 1;
+		goto done;
+	}
+
+	set_extent_uptodate(tree, start, page_end, GFP_NOFS);
+	blocksize = inode->i_sb->s_blocksize;
+
+	while (cur <= end) {
+		if (cur >= last_byte) {
+			clear_extent_dirty(tree, cur, page_end, GFP_NOFS);
+			unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
+			if (tree->ops && tree->ops->writepage_end_io_hook)
+				tree->ops->writepage_end_io_hook(page, cur,
+							 page_end, NULL, 1);
+			unlock_start = page_end + 1;
+			break;
+		}
+		em = epd->get_extent(inode, page, pg_offset, cur,
+				     end - cur + 1, 1);
+		if (IS_ERR(em) || !em) {
+			SetPageError(page);
+			break;
+		}
+
+		extent_offset = cur - em->start;
+		BUG_ON(extent_map_end(em) <= cur);
+		BUG_ON(end < cur);
+		iosize = min(extent_map_end(em) - cur, end - cur + 1);
+		iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
+		sector = (em->block_start + extent_offset) >> 9;
+		bdev = em->bdev;
+		block_start = em->block_start;
+		compressed = test_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+		free_extent_map(em);
+		em = NULL;
+
+		/*
+		 * compressed and inline extents are written through other
+		 * paths in the FS
+		 */
+		if (compressed || block_start == EXTENT_MAP_HOLE ||
+		    block_start == EXTENT_MAP_INLINE) {
+			clear_extent_dirty(tree, cur,
+					   cur + iosize - 1, GFP_NOFS);
+
+			unlock_extent(tree, unlock_start, cur + iosize - 1,
+				      GFP_NOFS);
+
+			/*
+			 * end_io notification does not happen here for
+			 * compressed extents
+			 */
+			if (!compressed && tree->ops &&
+			    tree->ops->writepage_end_io_hook)
+				tree->ops->writepage_end_io_hook(page, cur,
+							 cur + iosize - 1,
+							 NULL, 1);
+			else if (compressed) {
+				/* we don't want to end_page_writeback on
+				 * a compressed extent.  this happens
+				 * elsewhere
+				 */
+				nr++;
+			}
+
+			cur += iosize;
+			pg_offset += iosize;
+			unlock_start = cur;
+			continue;
+		}
+		/* leave this out until we have a page_mkwrite call */
+		if (0 && !test_range_bit(tree, cur, cur + iosize - 1,
+				   EXTENT_DIRTY, 0)) {
+			cur = cur + iosize;
+			pg_offset += iosize;
+			continue;
+		}
+
+		clear_extent_dirty(tree, cur, cur + iosize - 1, GFP_NOFS);
+		if (tree->ops && tree->ops->writepage_io_hook) {
+			ret = tree->ops->writepage_io_hook(page, cur,
+						cur + iosize - 1);
+		} else {
+			ret = 0;
+		}
+		if (ret) {
+			SetPageError(page);
+		} else {
+			unsigned long max_nr = end_index + 1;
+
+			set_range_writeback(tree, cur, cur + iosize - 1);
+			if (!PageWriteback(page)) {
+				printk(KERN_ERR "btrfs warning page %lu not "
+				       "writeback, cur %llu end %llu\n",
+				       page->index, (unsigned long long)cur,
+				       (unsigned long long)end);
+			}
+
+			ret = submit_extent_page(WRITE, tree, page, sector,
+						 iosize, pg_offset, bdev,
+						 &epd->bio, max_nr,
+						 end_bio_extent_writepage,
+						 0, 0, 0);
+			if (ret)
+				SetPageError(page);
+		}
+		cur = cur + iosize;
+		pg_offset += iosize;
+		nr++;
+	}
+done:
+	if (nr == 0) {
+		/* make sure the mapping tag for page dirty gets cleared */
+		set_page_writeback(page);
+		end_page_writeback(page);
+	}
+	if (unlock_start <= page_end)
+		unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
+	unlock_page(page);
+
+update_nr_written:
+	wbc->nr_to_write -= nr_written;
+	if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
+	    wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
+		page->mapping->writeback_index = page->index + nr_written;
+	return 0;
+}
+
+/**
+ * write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
+ * @mapping: address space structure to write
+ * @wbc: subtract the number of written pages from *@wbc->nr_to_write
+ * @writepage: function called for each page
+ * @data: data passed to writepage function
+ *
+ * If a page is already under I/O, write_cache_pages() skips it, even
+ * if it's dirty.  This is desirable behaviour for memory-cleaning writeback,
+ * but it is INCORRECT for data-integrity system calls such as fsync().  fsync()
+ * and msync() need to guarantee that all the data which was dirty at the time
+ * the call was made get new I/O started against them.  If wbc->sync_mode is
+ * WB_SYNC_ALL then we were called for data integrity and we must wait for
+ * existing IO to complete.
+ */
+static int extent_write_cache_pages(struct extent_io_tree *tree,
+			     struct address_space *mapping,
+			     struct writeback_control *wbc,
+			     writepage_t writepage, void *data,
+			     void (*flush_fn)(void *))
+{
+	struct backing_dev_info *bdi = mapping->backing_dev_info;
+	int ret = 0;
+	int done = 0;
+	struct pagevec pvec;
+	int nr_pages;
+	pgoff_t index;
+	pgoff_t end;		/* Inclusive */
+	int scanned = 0;
+	int range_whole = 0;
+
+	if (wbc->nonblocking && bdi_write_congested(bdi)) {
+		wbc->encountered_congestion = 1;
+		return 0;
+	}
+
+	pagevec_init(&pvec, 0);
+	if (wbc->range_cyclic) {
+		index = mapping->writeback_index; /* Start from prev offset */
+		end = -1;
+	} else {
+		index = wbc->range_start >> PAGE_CACHE_SHIFT;
+		end = wbc->range_end >> PAGE_CACHE_SHIFT;
+		if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
+			range_whole = 1;
+		scanned = 1;
+	}
+retry:
+	while (!done && (index <= end) &&
+	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+			      PAGECACHE_TAG_DIRTY, min(end - index,
+				  (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
+		unsigned i;
+
+		scanned = 1;
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+
+			/*
+			 * At this point we hold neither mapping->tree_lock nor
+			 * lock on the page itself: the page may be truncated or
+			 * invalidated (changing page->mapping to NULL), or even
+			 * swizzled back from swapper_space to tmpfs file
+			 * mapping
+			 */
+			if (tree->ops && tree->ops->write_cache_pages_lock_hook)
+				tree->ops->write_cache_pages_lock_hook(page);
+			else
+				lock_page(page);
+
+			if (unlikely(page->mapping != mapping)) {
+				unlock_page(page);
+				continue;
+			}
+
+			if (!wbc->range_cyclic && page->index > end) {
+				done = 1;
+				unlock_page(page);
+				continue;
+			}
+
+			if (wbc->sync_mode != WB_SYNC_NONE) {
+				if (PageWriteback(page))
+					flush_fn(data);
+				wait_on_page_writeback(page);
+			}
+
+			if (PageWriteback(page) ||
+			    !clear_page_dirty_for_io(page)) {
+				unlock_page(page);
+				continue;
+			}
+
+			ret = (*writepage)(page, wbc, data);
+
+			if (unlikely(ret == AOP_WRITEPAGE_ACTIVATE)) {
+				unlock_page(page);
+				ret = 0;
+			}
+			if (ret || wbc->nr_to_write <= 0)
+				done = 1;
+			if (wbc->nonblocking && bdi_write_congested(bdi)) {
+				wbc->encountered_congestion = 1;
+				done = 1;
+			}
+		}
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+	if (!scanned && !done) {
+		/*
+		 * We hit the last page and there is more work to be done: wrap
+		 * back to the start of the file
+		 */
+		scanned = 1;
+		index = 0;
+		goto retry;
+	}
+	return ret;
+}
+
+static noinline void flush_write_bio(void *data)
+{
+	struct extent_page_data *epd = data;
+	if (epd->bio) {
+		submit_one_bio(WRITE, epd->bio, 0, 0);
+		epd->bio = NULL;
+	}
+}
+
+int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
+			  get_extent_t *get_extent,
+			  struct writeback_control *wbc)
+{
+	int ret;
+	struct address_space *mapping = page->mapping;
+	struct extent_page_data epd = {
+		.bio = NULL,
+		.tree = tree,
+		.get_extent = get_extent,
+		.extent_locked = 0,
+	};
+	struct writeback_control wbc_writepages = {
+		.bdi		= wbc->bdi,
+		.sync_mode	= WB_SYNC_NONE,
+		.older_than_this = NULL,
+		.nr_to_write	= 64,
+		.range_start	= page_offset(page) + PAGE_CACHE_SIZE,
+		.range_end	= (loff_t)-1,
+	};
+
+
+	ret = __extent_writepage(page, wbc, &epd);
+
+	extent_write_cache_pages(tree, mapping, &wbc_writepages,
+				 __extent_writepage, &epd, flush_write_bio);
+	if (epd.bio)
+		submit_one_bio(WRITE, epd.bio, 0, 0);
+	return ret;
+}
+
+int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
+			      u64 start, u64 end, get_extent_t *get_extent,
+			      int mode)
+{
+	int ret = 0;
+	struct address_space *mapping = inode->i_mapping;
+	struct page *page;
+	unsigned long nr_pages = (end - start + PAGE_CACHE_SIZE) >>
+		PAGE_CACHE_SHIFT;
+
+	struct extent_page_data epd = {
+		.bio = NULL,
+		.tree = tree,
+		.get_extent = get_extent,
+		.extent_locked = 1,
+	};
+	struct writeback_control wbc_writepages = {
+		.bdi		= inode->i_mapping->backing_dev_info,
+		.sync_mode	= mode,
+		.older_than_this = NULL,
+		.nr_to_write	= nr_pages * 2,
+		.range_start	= start,
+		.range_end	= end + 1,
+	};
+
+	while (start <= end) {
+		page = find_get_page(mapping, start >> PAGE_CACHE_SHIFT);
+		if (clear_page_dirty_for_io(page))
+			ret = __extent_writepage(page, &wbc_writepages, &epd);
+		else {
+			if (tree->ops && tree->ops->writepage_end_io_hook)
+				tree->ops->writepage_end_io_hook(page, start,
+						 start + PAGE_CACHE_SIZE - 1,
+						 NULL, 1);
+			unlock_page(page);
+		}
+		page_cache_release(page);
+		start += PAGE_CACHE_SIZE;
+	}
+
+	if (epd.bio)
+		submit_one_bio(WRITE, epd.bio, 0, 0);
+	return ret;
+}
+
+int extent_writepages(struct extent_io_tree *tree,
+		      struct address_space *mapping,
+		      get_extent_t *get_extent,
+		      struct writeback_control *wbc)
+{
+	int ret = 0;
+	struct extent_page_data epd = {
+		.bio = NULL,
+		.tree = tree,
+		.get_extent = get_extent,
+		.extent_locked = 0,
+	};
+
+	ret = extent_write_cache_pages(tree, mapping, wbc,
+				       __extent_writepage, &epd,
+				       flush_write_bio);
+	if (epd.bio)
+		submit_one_bio(WRITE, epd.bio, 0, 0);
+	return ret;
+}
+
+int extent_readpages(struct extent_io_tree *tree,
+		     struct address_space *mapping,
+		     struct list_head *pages, unsigned nr_pages,
+		     get_extent_t get_extent)
+{
+	struct bio *bio = NULL;
+	unsigned page_idx;
+	struct pagevec pvec;
+	unsigned long bio_flags = 0;
+
+	pagevec_init(&pvec, 0);
+	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
+		struct page *page = list_entry(pages->prev, struct page, lru);
+
+		prefetchw(&page->flags);
+		list_del(&page->lru);
+		/*
+		 * what we want to do here is call add_to_page_cache_lru,
+		 * but that isn't exported, so we reproduce it here
+		 */
+		if (!add_to_page_cache(page, mapping,
+					page->index, GFP_KERNEL)) {
+
+			/* open coding of lru_cache_add, also not exported */
+			page_cache_get(page);
+			if (!pagevec_add(&pvec, page))
+				__pagevec_lru_add_file(&pvec);
+			__extent_read_full_page(tree, page, get_extent,
+						&bio, 0, &bio_flags);
+		}
+		page_cache_release(page);
+	}
+	if (pagevec_count(&pvec))
+		__pagevec_lru_add_file(&pvec);
+	BUG_ON(!list_empty(pages));
+	if (bio)
+		submit_one_bio(READ, bio, 0, bio_flags);
+	return 0;
+}
+
+/*
+ * basic invalidatepage code, this waits on any locked or writeback
+ * ranges corresponding to the page, and then deletes any extent state
+ * records from the tree
+ */
+int extent_invalidatepage(struct extent_io_tree *tree,
+			  struct page *page, unsigned long offset)
+{
+	u64 start = ((u64)page->index << PAGE_CACHE_SHIFT);
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+	size_t blocksize = page->mapping->host->i_sb->s_blocksize;
+
+	start += (offset + blocksize - 1) & ~(blocksize - 1);
+	if (start > end)
+		return 0;
+
+	lock_extent(tree, start, end, GFP_NOFS);
+	wait_on_extent_writeback(tree, start, end);
+	clear_extent_bit(tree, start, end,
+			 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC,
+			 1, 1, GFP_NOFS);
+	return 0;
+}
+
+/*
+ * simple commit_write call, set_range_dirty is used to mark both
+ * the pages and the extent records as dirty
+ */
+int extent_commit_write(struct extent_io_tree *tree,
+			struct inode *inode, struct page *page,
+			unsigned from, unsigned to)
+{
+	loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
+
+	set_page_extent_mapped(page);
+	set_page_dirty(page);
+
+	if (pos > inode->i_size) {
+		i_size_write(inode, pos);
+		mark_inode_dirty(inode);
+	}
+	return 0;
+}
+
+int extent_prepare_write(struct extent_io_tree *tree,
+			 struct inode *inode, struct page *page,
+			 unsigned from, unsigned to, get_extent_t *get_extent)
+{
+	u64 page_start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
+	u64 block_start;
+	u64 orig_block_start;
+	u64 block_end;
+	u64 cur_end;
+	struct extent_map *em;
+	unsigned blocksize = 1 << inode->i_blkbits;
+	size_t page_offset = 0;
+	size_t block_off_start;
+	size_t block_off_end;
+	int err = 0;
+	int iocount = 0;
+	int ret = 0;
+	int isnew;
+
+	set_page_extent_mapped(page);
+
+	block_start = (page_start + from) & ~((u64)blocksize - 1);
+	block_end = (page_start + to - 1) | (blocksize - 1);
+	orig_block_start = block_start;
+
+	lock_extent(tree, page_start, page_end, GFP_NOFS);
+	while (block_start <= block_end) {
+		em = get_extent(inode, page, page_offset, block_start,
+				block_end - block_start + 1, 1);
+		if (IS_ERR(em) || !em)
+			goto err;
+
+		cur_end = min(block_end, extent_map_end(em) - 1);
+		block_off_start = block_start & (PAGE_CACHE_SIZE - 1);
+		block_off_end = block_off_start + blocksize;
+		isnew = clear_extent_new(tree, block_start, cur_end, GFP_NOFS);
+
+		if (!PageUptodate(page) && isnew &&
+		    (block_off_end > to || block_off_start < from)) {
+			void *kaddr;
+
+			kaddr = kmap_atomic(page, KM_USER0);
+			if (block_off_end > to)
+				memset(kaddr + to, 0, block_off_end - to);
+			if (block_off_start < from)
+				memset(kaddr + block_off_start, 0,
+				       from - block_off_start);
+			flush_dcache_page(page);
+			kunmap_atomic(kaddr, KM_USER0);
+		}
+		if ((em->block_start != EXTENT_MAP_HOLE &&
+		     em->block_start != EXTENT_MAP_INLINE) &&
+		    !isnew && !PageUptodate(page) &&
+		    (block_off_end > to || block_off_start < from) &&
+		    !test_range_bit(tree, block_start, cur_end,
+				    EXTENT_UPTODATE, 1)) {
+			u64 sector;
+			u64 extent_offset = block_start - em->start;
+			size_t iosize;
+			sector = (em->block_start + extent_offset) >> 9;
+			iosize = (cur_end - block_start + blocksize) &
+				~((u64)blocksize - 1);
+			/*
+			 * we've already got the extent locked, but we
+			 * need to split the state such that our end_bio
+			 * handler can clear the lock.
+			 */
+			set_extent_bit(tree, block_start,
+				       block_start + iosize - 1,
+				       EXTENT_LOCKED, 0, NULL, GFP_NOFS);
+			ret = submit_extent_page(READ, tree, page,
+					 sector, iosize, page_offset, em->bdev,
+					 NULL, 1,
+					 end_bio_extent_preparewrite, 0,
+					 0, 0);
+			iocount++;
+			block_start = block_start + iosize;
+		} else {
+			set_extent_uptodate(tree, block_start, cur_end,
+					    GFP_NOFS);
+			unlock_extent(tree, block_start, cur_end, GFP_NOFS);
+			block_start = cur_end + 1;
+		}
+		page_offset = block_start & (PAGE_CACHE_SIZE - 1);
+		free_extent_map(em);
+	}
+	if (iocount) {
+		wait_extent_bit(tree, orig_block_start,
+				block_end, EXTENT_LOCKED);
+	}
+	check_page_uptodate(tree, page);
+err:
+	/* FIXME, zero out newly allocated blocks on error */
+	return err;
+}
+
+/*
+ * a helper for releasepage, this tests for areas of the page that
+ * are locked or under IO and drops the related state bits if it is safe
+ * to drop the page.
+ */
+int try_release_extent_state(struct extent_map_tree *map,
+			     struct extent_io_tree *tree, struct page *page,
+			     gfp_t mask)
+{
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+	int ret = 1;
+
+	if (test_range_bit(tree, start, end,
+			   EXTENT_IOBITS | EXTENT_ORDERED, 0))
+		ret = 0;
+	else {
+		if ((mask & GFP_NOFS) == GFP_NOFS)
+			mask = GFP_NOFS;
+		clear_extent_bit(tree, start, end, EXTENT_UPTODATE,
+				 1, 1, mask);
+	}
+	return ret;
+}
+
+/*
+ * a helper for releasepage.  As long as there are no locked extents
+ * in the range corresponding to the page, both state records and extent
+ * map records are removed
+ */
+int try_release_extent_mapping(struct extent_map_tree *map,
+			       struct extent_io_tree *tree, struct page *page,
+			       gfp_t mask)
+{
+	struct extent_map *em;
+	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+
+	if ((mask & __GFP_WAIT) &&
+	    page->mapping->host->i_size > 16 * 1024 * 1024) {
+		u64 len;
+		while (start <= end) {
+			len = end - start + 1;
+			spin_lock(&map->lock);
+			em = lookup_extent_mapping(map, start, len);
+			if (!em || IS_ERR(em)) {
+				spin_unlock(&map->lock);
+				break;
+			}
+			if (test_bit(EXTENT_FLAG_PINNED, &em->flags) ||
+			    em->start != start) {
+				spin_unlock(&map->lock);
+				free_extent_map(em);
+				break;
+			}
+			if (!test_range_bit(tree, em->start,
+					    extent_map_end(em) - 1,
+					    EXTENT_LOCKED | EXTENT_WRITEBACK |
+					    EXTENT_ORDERED,
+					    0)) {
+				remove_extent_mapping(map, em);
+				/* once for the rb tree */
+				free_extent_map(em);
+			}
+			start = extent_map_end(em);
+			spin_unlock(&map->lock);
+
+			/* once for us */
+			free_extent_map(em);
+		}
+	}
+	return try_release_extent_state(map, tree, page, mask);
+}
+
+sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
+		get_extent_t *get_extent)
+{
+	struct inode *inode = mapping->host;
+	u64 start = iblock << inode->i_blkbits;
+	sector_t sector = 0;
+	size_t blksize = (1 << inode->i_blkbits);
+	struct extent_map *em;
+
+	lock_extent(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
+		    GFP_NOFS);
+	em = get_extent(inode, NULL, 0, start, blksize, 0);
+	unlock_extent(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
+		      GFP_NOFS);
+	if (!em || IS_ERR(em))
+		return 0;
+
+	if (em->block_start > EXTENT_MAP_LAST_BYTE)
+		goto out;
+
+	sector = (em->block_start + start - em->start) >> inode->i_blkbits;
+out:
+	free_extent_map(em);
+	return sector;
+}
+
+static inline struct page *extent_buffer_page(struct extent_buffer *eb,
+					      unsigned long i)
+{
+	struct page *p;
+	struct address_space *mapping;
+
+	if (i == 0)
+		return eb->first_page;
+	i += eb->start >> PAGE_CACHE_SHIFT;
+	mapping = eb->first_page->mapping;
+	if (!mapping)
+		return NULL;
+
+	/*
+	 * extent_buffer_page is only called after pinning the page
+	 * by increasing the reference count.  So we know the page must
+	 * be in the radix tree.
+	 */
+	rcu_read_lock();
+	p = radix_tree_lookup(&mapping->page_tree, i);
+	rcu_read_unlock();
+
+	return p;
+}
+
+static inline unsigned long num_extent_pages(u64 start, u64 len)
+{
+	return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
+		(start >> PAGE_CACHE_SHIFT);
+}
+
+static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
+						   u64 start,
+						   unsigned long len,
+						   gfp_t mask)
+{
+	struct extent_buffer *eb = NULL;
+#ifdef LEAK_DEBUG
+	unsigned long flags;
+#endif
+
+	eb = kmem_cache_zalloc(extent_buffer_cache, mask);
+	eb->start = start;
+	eb->len = len;
+	mutex_init(&eb->mutex);
+#ifdef LEAK_DEBUG
+	spin_lock_irqsave(&leak_lock, flags);
+	list_add(&eb->leak_list, &buffers);
+	spin_unlock_irqrestore(&leak_lock, flags);
+#endif
+	atomic_set(&eb->refs, 1);
+
+	return eb;
+}
+
+static void __free_extent_buffer(struct extent_buffer *eb)
+{
+#ifdef LEAK_DEBUG
+	unsigned long flags;
+	spin_lock_irqsave(&leak_lock, flags);
+	list_del(&eb->leak_list);
+	spin_unlock_irqrestore(&leak_lock, flags);
+#endif
+	kmem_cache_free(extent_buffer_cache, eb);
+}
+
+struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
+					  u64 start, unsigned long len,
+					  struct page *page0,
+					  gfp_t mask)
+{
+	unsigned long num_pages = num_extent_pages(start, len);
+	unsigned long i;
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	struct extent_buffer *eb;
+	struct extent_buffer *exists = NULL;
+	struct page *p;
+	struct address_space *mapping = tree->mapping;
+	int uptodate = 1;
+
+	spin_lock(&tree->buffer_lock);
+	eb = buffer_search(tree, start);
+	if (eb) {
+		atomic_inc(&eb->refs);
+		spin_unlock(&tree->buffer_lock);
+		mark_page_accessed(eb->first_page);
+		return eb;
+	}
+	spin_unlock(&tree->buffer_lock);
+
+	eb = __alloc_extent_buffer(tree, start, len, mask);
+	if (!eb)
+		return NULL;
+
+	if (page0) {
+		eb->first_page = page0;
+		i = 1;
+		index++;
+		page_cache_get(page0);
+		mark_page_accessed(page0);
+		set_page_extent_mapped(page0);
+		set_page_extent_head(page0, len);
+		uptodate = PageUptodate(page0);
+	} else {
+		i = 0;
+	}
+	for (; i < num_pages; i++, index++) {
+		p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM);
+		if (!p) {
+			WARN_ON(1);
+			goto free_eb;
+		}
+		set_page_extent_mapped(p);
+		mark_page_accessed(p);
+		if (i == 0) {
+			eb->first_page = p;
+			set_page_extent_head(p, len);
+		} else {
+			set_page_private(p, EXTENT_PAGE_PRIVATE);
+		}
+		if (!PageUptodate(p))
+			uptodate = 0;
+		unlock_page(p);
+	}
+	if (uptodate)
+		eb->flags |= EXTENT_UPTODATE;
+	eb->flags |= EXTENT_BUFFER_FILLED;
+
+	spin_lock(&tree->buffer_lock);
+	exists = buffer_tree_insert(tree, start, &eb->rb_node);
+	if (exists) {
+		/* add one reference for the caller */
+		atomic_inc(&exists->refs);
+		spin_unlock(&tree->buffer_lock);
+		goto free_eb;
+	}
+	spin_unlock(&tree->buffer_lock);
+
+	/* add one reference for the tree */
+	atomic_inc(&eb->refs);
+	return eb;
+
+free_eb:
+	if (!atomic_dec_and_test(&eb->refs))
+		return exists;
+	for (index = 1; index < i; index++)
+		page_cache_release(extent_buffer_page(eb, index));
+	page_cache_release(extent_buffer_page(eb, 0));
+	__free_extent_buffer(eb);
+	return exists;
+}
+
+struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
+					 u64 start, unsigned long len,
+					  gfp_t mask)
+{
+	struct extent_buffer *eb;
+
+	spin_lock(&tree->buffer_lock);
+	eb = buffer_search(tree, start);
+	if (eb)
+		atomic_inc(&eb->refs);
+	spin_unlock(&tree->buffer_lock);
+
+	if (eb)
+		mark_page_accessed(eb->first_page);
+
+	return eb;
+}
+
+void free_extent_buffer(struct extent_buffer *eb)
+{
+	if (!eb)
+		return;
+
+	if (!atomic_dec_and_test(&eb->refs))
+		return;
+
+	WARN_ON(1);
+}
+
+int clear_extent_buffer_dirty(struct extent_io_tree *tree,
+			      struct extent_buffer *eb)
+{
+	int set;
+	unsigned long i;
+	unsigned long num_pages;
+	struct page *page;
+
+	u64 start = eb->start;
+	u64 end = start + eb->len - 1;
+
+	set = clear_extent_dirty(tree, start, end, GFP_NOFS);
+	num_pages = num_extent_pages(eb->start, eb->len);
+
+	for (i = 0; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		if (!set && !PageDirty(page))
+			continue;
+
+		lock_page(page);
+		if (i == 0)
+			set_page_extent_head(page, eb->len);
+		else
+			set_page_private(page, EXTENT_PAGE_PRIVATE);
+
+		/*
+		 * if we're on the last page or the first page and the
+		 * block isn't aligned on a page boundary, do extra checks
+		 * to make sure we don't clean page that is partially dirty
+		 */
+		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
+		    ((i == num_pages - 1) &&
+		     ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) {
+			start = (u64)page->index << PAGE_CACHE_SHIFT;
+			end  = start + PAGE_CACHE_SIZE - 1;
+			if (test_range_bit(tree, start, end,
+					   EXTENT_DIRTY, 0)) {
+				unlock_page(page);
+				continue;
+			}
+		}
+		clear_page_dirty_for_io(page);
+		spin_lock_irq(&page->mapping->tree_lock);
+		if (!PageDirty(page)) {
+			radix_tree_tag_clear(&page->mapping->page_tree,
+						page_index(page),
+						PAGECACHE_TAG_DIRTY);
+		}
+		spin_unlock_irq(&page->mapping->tree_lock);
+		unlock_page(page);
+	}
+	return 0;
+}
+
+int wait_on_extent_buffer_writeback(struct extent_io_tree *tree,
+				    struct extent_buffer *eb)
+{
+	return wait_on_extent_writeback(tree, eb->start,
+					eb->start + eb->len - 1);
+}
+
+int set_extent_buffer_dirty(struct extent_io_tree *tree,
+			     struct extent_buffer *eb)
+{
+	unsigned long i;
+	unsigned long num_pages;
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	for (i = 0; i < num_pages; i++) {
+		struct page *page = extent_buffer_page(eb, i);
+		/* writepage may need to do something special for the
+		 * first page, we have to make sure page->private is
+		 * properly set.  releasepage may drop page->private
+		 * on us if the page isn't already dirty.
+		 */
+		lock_page(page);
+		if (i == 0) {
+			set_page_extent_head(page, eb->len);
+		} else if (PagePrivate(page) &&
+			   page->private != EXTENT_PAGE_PRIVATE) {
+			set_page_extent_mapped(page);
+		}
+		__set_page_dirty_nobuffers(extent_buffer_page(eb, i));
+		set_extent_dirty(tree, page_offset(page),
+				 page_offset(page) + PAGE_CACHE_SIZE - 1,
+				 GFP_NOFS);
+		unlock_page(page);
+	}
+	return 0;
+}
+
+int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
+				struct extent_buffer *eb)
+{
+	unsigned long i;
+	struct page *page;
+	unsigned long num_pages;
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	eb->flags &= ~EXTENT_UPTODATE;
+
+	clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
+			      GFP_NOFS);
+	for (i = 0; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		if (page)
+			ClearPageUptodate(page);
+	}
+	return 0;
+}
+
+int set_extent_buffer_uptodate(struct extent_io_tree *tree,
+				struct extent_buffer *eb)
+{
+	unsigned long i;
+	struct page *page;
+	unsigned long num_pages;
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+
+	set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
+			    GFP_NOFS);
+	for (i = 0; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
+		    ((i == num_pages - 1) &&
+		     ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) {
+			check_page_uptodate(tree, page);
+			continue;
+		}
+		SetPageUptodate(page);
+	}
+	return 0;
+}
+
+int extent_range_uptodate(struct extent_io_tree *tree,
+			  u64 start, u64 end)
+{
+	struct page *page;
+	int ret;
+	int pg_uptodate = 1;
+	int uptodate;
+	unsigned long index;
+
+	ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1);
+	if (ret)
+		return 1;
+	while (start <= end) {
+		index = start >> PAGE_CACHE_SHIFT;
+		page = find_get_page(tree->mapping, index);
+		uptodate = PageUptodate(page);
+		page_cache_release(page);
+		if (!uptodate) {
+			pg_uptodate = 0;
+			break;
+		}
+		start += PAGE_CACHE_SIZE;
+	}
+	return pg_uptodate;
+}
+
+int extent_buffer_uptodate(struct extent_io_tree *tree,
+			   struct extent_buffer *eb)
+{
+	int ret = 0;
+	unsigned long num_pages;
+	unsigned long i;
+	struct page *page;
+	int pg_uptodate = 1;
+
+	if (eb->flags & EXTENT_UPTODATE)
+		return 1;
+
+	ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
+			   EXTENT_UPTODATE, 1);
+	if (ret)
+		return ret;
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	for (i = 0; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		if (!PageUptodate(page)) {
+			pg_uptodate = 0;
+			break;
+		}
+	}
+	return pg_uptodate;
+}
+
+int read_extent_buffer_pages(struct extent_io_tree *tree,
+			     struct extent_buffer *eb,
+			     u64 start, int wait,
+			     get_extent_t *get_extent, int mirror_num)
+{
+	unsigned long i;
+	unsigned long start_i;
+	struct page *page;
+	int err;
+	int ret = 0;
+	int locked_pages = 0;
+	int all_uptodate = 1;
+	int inc_all_pages = 0;
+	unsigned long num_pages;
+	struct bio *bio = NULL;
+	unsigned long bio_flags = 0;
+
+	if (eb->flags & EXTENT_UPTODATE)
+		return 0;
+
+	if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
+			   EXTENT_UPTODATE, 1)) {
+		return 0;
+	}
+
+	if (start) {
+		WARN_ON(start < eb->start);
+		start_i = (start >> PAGE_CACHE_SHIFT) -
+			(eb->start >> PAGE_CACHE_SHIFT);
+	} else {
+		start_i = 0;
+	}
+
+	num_pages = num_extent_pages(eb->start, eb->len);
+	for (i = start_i; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		if (!wait) {
+			if (!trylock_page(page))
+				goto unlock_exit;
+		} else {
+			lock_page(page);
+		}
+		locked_pages++;
+		if (!PageUptodate(page))
+			all_uptodate = 0;
+	}
+	if (all_uptodate) {
+		if (start_i == 0)
+			eb->flags |= EXTENT_UPTODATE;
+		goto unlock_exit;
+	}
+
+	for (i = start_i; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		if (inc_all_pages)
+			page_cache_get(page);
+		if (!PageUptodate(page)) {
+			if (start_i == 0)
+				inc_all_pages = 1;
+			ClearPageError(page);
+			err = __extent_read_full_page(tree, page,
+						      get_extent, &bio,
+						      mirror_num, &bio_flags);
+			if (err)
+				ret = err;
+		} else {
+			unlock_page(page);
+		}
+	}
+
+	if (bio)
+		submit_one_bio(READ, bio, mirror_num, bio_flags);
+
+	if (ret || !wait)
+		return ret;
+
+	for (i = start_i; i < num_pages; i++) {
+		page = extent_buffer_page(eb, i);
+		wait_on_page_locked(page);
+		if (!PageUptodate(page))
+			ret = -EIO;
+	}
+
+	if (!ret)
+		eb->flags |= EXTENT_UPTODATE;
+	return ret;
+
+unlock_exit:
+	i = start_i;
+	while (locked_pages > 0) {
+		page = extent_buffer_page(eb, i);
+		i++;
+		unlock_page(page);
+		locked_pages--;
+	}
+	return ret;
+}
+
+void read_extent_buffer(struct extent_buffer *eb, void *dstv,
+			unsigned long start,
+			unsigned long len)
+{
+	size_t cur;
+	size_t offset;
+	struct page *page;
+	char *kaddr;
+	char *dst = (char *)dstv;
+	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
+
+	WARN_ON(start > eb->len);
+	WARN_ON(start + len > eb->start + eb->len);
+
+	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+
+	while (len > 0) {
+		page = extent_buffer_page(eb, i);
+
+		cur = min(len, (PAGE_CACHE_SIZE - offset));
+		kaddr = kmap_atomic(page, KM_USER1);
+		memcpy(dst, kaddr + offset, cur);
+		kunmap_atomic(kaddr, KM_USER1);
+
+		dst += cur;
+		len -= cur;
+		offset = 0;
+		i++;
+	}
+}
+
+int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
+			       unsigned long min_len, char **token, char **map,
+			       unsigned long *map_start,
+			       unsigned long *map_len, int km)
+{
+	size_t offset = start & (PAGE_CACHE_SIZE - 1);
+	char *kaddr;
+	struct page *p;
+	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
+	unsigned long end_i = (start_offset + start + min_len - 1) >>
+		PAGE_CACHE_SHIFT;
+
+	if (i != end_i)
+		return -EINVAL;
+
+	if (i == 0) {
+		offset = start_offset;
+		*map_start = 0;
+	} else {
+		offset = 0;
+		*map_start = ((u64)i << PAGE_CACHE_SHIFT) - start_offset;
+	}
+
+	if (start + min_len > eb->len) {
+		printk(KERN_ERR "btrfs bad mapping eb start %llu len %lu, "
+		       "wanted %lu %lu\n", (unsigned long long)eb->start,
+		       eb->len, start, min_len);
+		WARN_ON(1);
+	}
+
+	p = extent_buffer_page(eb, i);
+	kaddr = kmap_atomic(p, km);
+	*token = kaddr;
+	*map = kaddr + offset;
+	*map_len = PAGE_CACHE_SIZE - offset;
+	return 0;
+}
+
+int map_extent_buffer(struct extent_buffer *eb, unsigned long start,
+		      unsigned long min_len,
+		      char **token, char **map,
+		      unsigned long *map_start,
+		      unsigned long *map_len, int km)
+{
+	int err;
+	int save = 0;
+	if (eb->map_token) {
+		unmap_extent_buffer(eb, eb->map_token, km);
+		eb->map_token = NULL;
+		save = 1;
+		WARN_ON(!mutex_is_locked(&eb->mutex));
+	}
+	err = map_private_extent_buffer(eb, start, min_len, token, map,
+				       map_start, map_len, km);
+	if (!err && save) {
+		eb->map_token = *token;
+		eb->kaddr = *map;
+		eb->map_start = *map_start;
+		eb->map_len = *map_len;
+	}
+	return err;
+}
+
+void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km)
+{
+	kunmap_atomic(token, km);
+}
+
+int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
+			  unsigned long start,
+			  unsigned long len)
+{
+	size_t cur;
+	size_t offset;
+	struct page *page;
+	char *kaddr;
+	char *ptr = (char *)ptrv;
+	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
+	int ret = 0;
+
+	WARN_ON(start > eb->len);
+	WARN_ON(start + len > eb->start + eb->len);
+
+	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+
+	while (len > 0) {
+		page = extent_buffer_page(eb, i);
+
+		cur = min(len, (PAGE_CACHE_SIZE - offset));
+
+		kaddr = kmap_atomic(page, KM_USER0);
+		ret = memcmp(ptr, kaddr + offset, cur);
+		kunmap_atomic(kaddr, KM_USER0);
+		if (ret)
+			break;
+
+		ptr += cur;
+		len -= cur;
+		offset = 0;
+		i++;
+	}
+	return ret;
+}
+
+void write_extent_buffer(struct extent_buffer *eb, const void *srcv,
+			 unsigned long start, unsigned long len)
+{
+	size_t cur;
+	size_t offset;
+	struct page *page;
+	char *kaddr;
+	char *src = (char *)srcv;
+	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
+
+	WARN_ON(start > eb->len);
+	WARN_ON(start + len > eb->start + eb->len);
+
+	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+
+	while (len > 0) {
+		page = extent_buffer_page(eb, i);
+		WARN_ON(!PageUptodate(page));
+
+		cur = min(len, PAGE_CACHE_SIZE - offset);
+		kaddr = kmap_atomic(page, KM_USER1);
+		memcpy(kaddr + offset, src, cur);
+		kunmap_atomic(kaddr, KM_USER1);
+
+		src += cur;
+		len -= cur;
+		offset = 0;
+		i++;
+	}
+}
+
+void memset_extent_buffer(struct extent_buffer *eb, char c,
+			  unsigned long start, unsigned long len)
+{
+	size_t cur;
+	size_t offset;
+	struct page *page;
+	char *kaddr;
+	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
+
+	WARN_ON(start > eb->len);
+	WARN_ON(start + len > eb->start + eb->len);
+
+	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
+
+	while (len > 0) {
+		page = extent_buffer_page(eb, i);
+		WARN_ON(!PageUptodate(page));
+
+		cur = min(len, PAGE_CACHE_SIZE - offset);
+		kaddr = kmap_atomic(page, KM_USER0);
+		memset(kaddr + offset, c, cur);
+		kunmap_atomic(kaddr, KM_USER0);
+
+		len -= cur;
+		offset = 0;
+		i++;
+	}
+}
+
+void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
+			unsigned long dst_offset, unsigned long src_offset,
+			unsigned long len)
+{
+	u64 dst_len = dst->len;
+	size_t cur;
+	size_t offset;
+	struct page *page;
+	char *kaddr;
+	size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;
+
+	WARN_ON(src->len != dst_len);
+
+	offset = (start_offset + dst_offset) &
+		((unsigned long)PAGE_CACHE_SIZE - 1);
+
+	while (len > 0) {
+		page = extent_buffer_page(dst, i);
+		WARN_ON(!PageUptodate(page));
+
+		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE - offset));
+
+		kaddr = kmap_atomic(page, KM_USER0);
+		read_extent_buffer(src, kaddr + offset, src_offset, cur);
+		kunmap_atomic(kaddr, KM_USER0);
+
+		src_offset += cur;
+		len -= cur;
+		offset = 0;
+		i++;
+	}
+}
+
+static void move_pages(struct page *dst_page, struct page *src_page,
+		       unsigned long dst_off, unsigned long src_off,
+		       unsigned long len)
+{
+	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
+	if (dst_page == src_page) {
+		memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len);
+	} else {
+		char *src_kaddr = kmap_atomic(src_page, KM_USER1);
+		char *p = dst_kaddr + dst_off + len;
+		char *s = src_kaddr + src_off + len;
+
+		while (len--)
+			*--p = *--s;
+
+		kunmap_atomic(src_kaddr, KM_USER1);
+	}
+	kunmap_atomic(dst_kaddr, KM_USER0);
+}
+
+static void copy_pages(struct page *dst_page, struct page *src_page,
+		       unsigned long dst_off, unsigned long src_off,
+		       unsigned long len)
+{
+	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
+	char *src_kaddr;
+
+	if (dst_page != src_page)
+		src_kaddr = kmap_atomic(src_page, KM_USER1);
+	else
+		src_kaddr = dst_kaddr;
+
+	memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
+	kunmap_atomic(dst_kaddr, KM_USER0);
+	if (dst_page != src_page)
+		kunmap_atomic(src_kaddr, KM_USER1);
+}
+
+void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
+			   unsigned long src_offset, unsigned long len)
+{
+	size_t cur;
+	size_t dst_off_in_page;
+	size_t src_off_in_page;
+	size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long dst_i;
+	unsigned long src_i;
+
+	if (src_offset + len > dst->len) {
+		printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
+		       "len %lu dst len %lu\n", src_offset, len, dst->len);
+		BUG_ON(1);
+	}
+	if (dst_offset + len > dst->len) {
+		printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
+		       "len %lu dst len %lu\n", dst_offset, len, dst->len);
+		BUG_ON(1);
+	}
+
+	while (len > 0) {
+		dst_off_in_page = (start_offset + dst_offset) &
+			((unsigned long)PAGE_CACHE_SIZE - 1);
+		src_off_in_page = (start_offset + src_offset) &
+			((unsigned long)PAGE_CACHE_SIZE - 1);
+
+		dst_i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;
+		src_i = (start_offset + src_offset) >> PAGE_CACHE_SHIFT;
+
+		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE -
+					       src_off_in_page));
+		cur = min_t(unsigned long, cur,
+			(unsigned long)(PAGE_CACHE_SIZE - dst_off_in_page));
+
+		copy_pages(extent_buffer_page(dst, dst_i),
+			   extent_buffer_page(dst, src_i),
+			   dst_off_in_page, src_off_in_page, cur);
+
+		src_offset += cur;
+		dst_offset += cur;
+		len -= cur;
+	}
+}
+
+void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
+			   unsigned long src_offset, unsigned long len)
+{
+	size_t cur;
+	size_t dst_off_in_page;
+	size_t src_off_in_page;
+	unsigned long dst_end = dst_offset + len - 1;
+	unsigned long src_end = src_offset + len - 1;
+	size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
+	unsigned long dst_i;
+	unsigned long src_i;
+
+	if (src_offset + len > dst->len) {
+		printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
+		       "len %lu len %lu\n", src_offset, len, dst->len);
+		BUG_ON(1);
+	}
+	if (dst_offset + len > dst->len) {
+		printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
+		       "len %lu len %lu\n", dst_offset, len, dst->len);
+		BUG_ON(1);
+	}
+	if (dst_offset < src_offset) {
+		memcpy_extent_buffer(dst, dst_offset, src_offset, len);
+		return;
+	}
+	while (len > 0) {
+		dst_i = (start_offset + dst_end) >> PAGE_CACHE_SHIFT;
+		src_i = (start_offset + src_end) >> PAGE_CACHE_SHIFT;
+
+		dst_off_in_page = (start_offset + dst_end) &
+			((unsigned long)PAGE_CACHE_SIZE - 1);
+		src_off_in_page = (start_offset + src_end) &
+			((unsigned long)PAGE_CACHE_SIZE - 1);
+
+		cur = min_t(unsigned long, len, src_off_in_page + 1);
+		cur = min(cur, dst_off_in_page + 1);
+		move_pages(extent_buffer_page(dst, dst_i),
+			   extent_buffer_page(dst, src_i),
+			   dst_off_in_page - cur + 1,
+			   src_off_in_page - cur + 1, cur);
+
+		dst_end -= cur;
+		src_end -= cur;
+		len -= cur;
+	}
+}
+
+int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page)
+{
+	u64 start = page_offset(page);
+	struct extent_buffer *eb;
+	int ret = 1;
+	unsigned long i;
+	unsigned long num_pages;
+
+	spin_lock(&tree->buffer_lock);
+	eb = buffer_search(tree, start);
+	if (!eb)
+		goto out;
+
+	if (atomic_read(&eb->refs) > 1) {
+		ret = 0;
+		goto out;
+	}
+	/* at this point we can safely release the extent buffer */
+	num_pages = num_extent_pages(eb->start, eb->len);
+	for (i = 0; i < num_pages; i++)
+		page_cache_release(extent_buffer_page(eb, i));
+	rb_erase(&eb->rb_node, &tree->buffer);
+	__free_extent_buffer(eb);
+out:
+	spin_unlock(&tree->buffer_lock);
+	return ret;
+}
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
new file mode 100644
index 00000000000..c5b483a7913
--- /dev/null
+++ b/fs/btrfs/extent_io.h
@@ -0,0 +1,269 @@
+#ifndef __EXTENTIO__
+#define __EXTENTIO__
+
+#include <linux/rbtree.h>
+
+/* bits for the extent state */
+#define EXTENT_DIRTY 1
+#define EXTENT_WRITEBACK (1 << 1)
+#define EXTENT_UPTODATE (1 << 2)
+#define EXTENT_LOCKED (1 << 3)
+#define EXTENT_NEW (1 << 4)
+#define EXTENT_DELALLOC (1 << 5)
+#define EXTENT_DEFRAG (1 << 6)
+#define EXTENT_DEFRAG_DONE (1 << 7)
+#define EXTENT_BUFFER_FILLED (1 << 8)
+#define EXTENT_ORDERED (1 << 9)
+#define EXTENT_ORDERED_METADATA (1 << 10)
+#define EXTENT_BOUNDARY (1 << 11)
+#define EXTENT_NODATASUM (1 << 12)
+#define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK)
+
+/* flags for bio submission */
+#define EXTENT_BIO_COMPRESSED 1
+
+/*
+ * page->private values.  Every page that is controlled by the extent
+ * map has page->private set to one.
+ */
+#define EXTENT_PAGE_PRIVATE 1
+#define EXTENT_PAGE_PRIVATE_FIRST_PAGE 3
+
+struct extent_state;
+
+typedef	int (extent_submit_bio_hook_t)(struct inode *inode, int rw,
+				       struct bio *bio, int mirror_num,
+				       unsigned long bio_flags);
+struct extent_io_ops {
+	int (*fill_delalloc)(struct inode *inode, struct page *locked_page,
+			     u64 start, u64 end, int *page_started,
+			     unsigned long *nr_written);
+	int (*writepage_start_hook)(struct page *page, u64 start, u64 end);
+	int (*writepage_io_hook)(struct page *page, u64 start, u64 end);
+	extent_submit_bio_hook_t *submit_bio_hook;
+	int (*merge_bio_hook)(struct page *page, unsigned long offset,
+			      size_t size, struct bio *bio,
+			      unsigned long bio_flags);
+	int (*readpage_io_hook)(struct page *page, u64 start, u64 end);
+	int (*readpage_io_failed_hook)(struct bio *bio, struct page *page,
+				       u64 start, u64 end,
+				       struct extent_state *state);
+	int (*writepage_io_failed_hook)(struct bio *bio, struct page *page,
+					u64 start, u64 end,
+				       struct extent_state *state);
+	int (*readpage_end_io_hook)(struct page *page, u64 start, u64 end,
+				    struct extent_state *state);
+	int (*writepage_end_io_hook)(struct page *page, u64 start, u64 end,
+				      struct extent_state *state, int uptodate);
+	int (*set_bit_hook)(struct inode *inode, u64 start, u64 end,
+			    unsigned long old, unsigned long bits);
+	int (*clear_bit_hook)(struct inode *inode, u64 start, u64 end,
+			    unsigned long old, unsigned long bits);
+	int (*write_cache_pages_lock_hook)(struct page *page);
+};
+
+struct extent_io_tree {
+	struct rb_root state;
+	struct rb_root buffer;
+	struct address_space *mapping;
+	u64 dirty_bytes;
+	spinlock_t lock;
+	spinlock_t buffer_lock;
+	struct extent_io_ops *ops;
+};
+
+struct extent_state {
+	u64 start;
+	u64 end; /* inclusive */
+	struct rb_node rb_node;
+	struct extent_io_tree *tree;
+	wait_queue_head_t wq;
+	atomic_t refs;
+	unsigned long state;
+
+	/* for use by the FS */
+	u64 private;
+
+	struct list_head leak_list;
+};
+
+struct extent_buffer {
+	u64 start;
+	unsigned long len;
+	char *map_token;
+	char *kaddr;
+	unsigned long map_start;
+	unsigned long map_len;
+	struct page *first_page;
+	atomic_t refs;
+	int flags;
+	struct list_head leak_list;
+	struct rb_node rb_node;
+	struct mutex mutex;
+};
+
+struct extent_map_tree;
+
+static inline struct extent_state *extent_state_next(struct extent_state *state)
+{
+	struct rb_node *node;
+	node = rb_next(&state->rb_node);
+	if (!node)
+		return NULL;
+	return rb_entry(node, struct extent_state, rb_node);
+}
+
+typedef struct extent_map *(get_extent_t)(struct inode *inode,
+					  struct page *page,
+					  size_t page_offset,
+					  u64 start, u64 len,
+					  int create);
+
+void extent_io_tree_init(struct extent_io_tree *tree,
+			  struct address_space *mapping, gfp_t mask);
+int try_release_extent_mapping(struct extent_map_tree *map,
+			       struct extent_io_tree *tree, struct page *page,
+			       gfp_t mask);
+int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page);
+int try_release_extent_state(struct extent_map_tree *map,
+			     struct extent_io_tree *tree, struct page *page,
+			     gfp_t mask);
+int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask);
+int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask);
+int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
+		    gfp_t mask);
+int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
+			  get_extent_t *get_extent);
+int __init extent_io_init(void);
+void extent_io_exit(void);
+
+u64 count_range_bits(struct extent_io_tree *tree,
+		     u64 *start, u64 search_end,
+		     u64 max_bytes, unsigned long bits);
+
+int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		   int bits, int filled);
+int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+		      int bits, gfp_t mask);
+int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		     int bits, int wake, int delete, gfp_t mask);
+int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+		    int bits, gfp_t mask);
+int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
+			gfp_t mask);
+int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
+		   gfp_t mask);
+int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
+		     gfp_t mask);
+int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
+		       gfp_t mask);
+int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
+		       gfp_t mask);
+int clear_extent_ordered_metadata(struct extent_io_tree *tree, u64 start,
+				  u64 end, gfp_t mask);
+int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
+		     gfp_t mask);
+int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
+		     gfp_t mask);
+int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
+			  u64 *start_ret, u64 *end_ret, int bits);
+struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree,
+						 u64 start, int bits);
+int extent_invalidatepage(struct extent_io_tree *tree,
+			  struct page *page, unsigned long offset);
+int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
+			  get_extent_t *get_extent,
+			  struct writeback_control *wbc);
+int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
+			      u64 start, u64 end, get_extent_t *get_extent,
+			      int mode);
+int extent_writepages(struct extent_io_tree *tree,
+		      struct address_space *mapping,
+		      get_extent_t *get_extent,
+		      struct writeback_control *wbc);
+int extent_readpages(struct extent_io_tree *tree,
+		     struct address_space *mapping,
+		     struct list_head *pages, unsigned nr_pages,
+		     get_extent_t get_extent);
+int extent_prepare_write(struct extent_io_tree *tree,
+			 struct inode *inode, struct page *page,
+			 unsigned from, unsigned to, get_extent_t *get_extent);
+int extent_commit_write(struct extent_io_tree *tree,
+			struct inode *inode, struct page *page,
+			unsigned from, unsigned to);
+sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
+		get_extent_t *get_extent);
+int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end);
+int set_state_private(struct extent_io_tree *tree, u64 start, u64 private);
+int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private);
+void set_page_extent_mapped(struct page *page);
+
+struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
+					  u64 start, unsigned long len,
+					  struct page *page0,
+					  gfp_t mask);
+struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
+					 u64 start, unsigned long len,
+					  gfp_t mask);
+void free_extent_buffer(struct extent_buffer *eb);
+int read_extent_buffer_pages(struct extent_io_tree *tree,
+			     struct extent_buffer *eb, u64 start, int wait,
+			     get_extent_t *get_extent, int mirror_num);
+
+static inline void extent_buffer_get(struct extent_buffer *eb)
+{
+	atomic_inc(&eb->refs);
+}
+
+int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
+			  unsigned long start,
+			  unsigned long len);
+void read_extent_buffer(struct extent_buffer *eb, void *dst,
+			unsigned long start,
+			unsigned long len);
+void write_extent_buffer(struct extent_buffer *eb, const void *src,
+			 unsigned long start, unsigned long len);
+void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
+			unsigned long dst_offset, unsigned long src_offset,
+			unsigned long len);
+void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
+			   unsigned long src_offset, unsigned long len);
+void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
+			   unsigned long src_offset, unsigned long len);
+void memset_extent_buffer(struct extent_buffer *eb, char c,
+			  unsigned long start, unsigned long len);
+int wait_on_extent_buffer_writeback(struct extent_io_tree *tree,
+				    struct extent_buffer *eb);
+int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end);
+int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits);
+int clear_extent_buffer_dirty(struct extent_io_tree *tree,
+			      struct extent_buffer *eb);
+int set_extent_buffer_dirty(struct extent_io_tree *tree,
+			     struct extent_buffer *eb);
+int set_extent_buffer_uptodate(struct extent_io_tree *tree,
+			       struct extent_buffer *eb);
+int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
+				struct extent_buffer *eb);
+int extent_buffer_uptodate(struct extent_io_tree *tree,
+			   struct extent_buffer *eb);
+int map_extent_buffer(struct extent_buffer *eb, unsigned long offset,
+		      unsigned long min_len, char **token, char **map,
+		      unsigned long *map_start,
+		      unsigned long *map_len, int km);
+int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset,
+		      unsigned long min_len, char **token, char **map,
+		      unsigned long *map_start,
+		      unsigned long *map_len, int km);
+void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km);
+int release_extent_buffer_tail_pages(struct extent_buffer *eb);
+int extent_range_uptodate(struct extent_io_tree *tree,
+			  u64 start, u64 end);
+int extent_clear_unlock_delalloc(struct inode *inode,
+				struct extent_io_tree *tree,
+				u64 start, u64 end, struct page *locked_page,
+				int unlock_page,
+				int clear_unlock,
+				int clear_delalloc, int clear_dirty,
+				int set_writeback,
+				int end_writeback);
+#endif
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
new file mode 100644
index 00000000000..4a83e33ada3
--- /dev/null
+++ b/fs/btrfs/extent_map.c
@@ -0,0 +1,351 @@
+#include <linux/err.h>
+#include <linux/gfp.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/version.h>
+#include <linux/hardirq.h>
+#include "extent_map.h"
+
+/* temporary define until extent_map moves out of btrfs */
+struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
+				       unsigned long extra_flags,
+				       void (*ctor)(void *, struct kmem_cache *,
+						    unsigned long));
+
+static struct kmem_cache *extent_map_cache;
+
+int __init extent_map_init(void)
+{
+	extent_map_cache = btrfs_cache_create("extent_map",
+					    sizeof(struct extent_map), 0,
+					    NULL);
+	if (!extent_map_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void extent_map_exit(void)
+{
+	if (extent_map_cache)
+		kmem_cache_destroy(extent_map_cache);
+}
+
+/**
+ * extent_map_tree_init - initialize extent map tree
+ * @tree:		tree to initialize
+ * @mask:		flags for memory allocations during tree operations
+ *
+ * Initialize the extent tree @tree.  Should be called for each new inode
+ * or other user of the extent_map interface.
+ */
+void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask)
+{
+	tree->map.rb_node = NULL;
+	spin_lock_init(&tree->lock);
+}
+EXPORT_SYMBOL(extent_map_tree_init);
+
+/**
+ * alloc_extent_map - allocate new extent map structure
+ * @mask:	memory allocation flags
+ *
+ * Allocate a new extent_map structure.  The new structure is
+ * returned with a reference count of one and needs to be
+ * freed using free_extent_map()
+ */
+struct extent_map *alloc_extent_map(gfp_t mask)
+{
+	struct extent_map *em;
+	em = kmem_cache_alloc(extent_map_cache, mask);
+	if (!em || IS_ERR(em))
+		return em;
+	em->in_tree = 0;
+	em->flags = 0;
+	atomic_set(&em->refs, 1);
+	return em;
+}
+EXPORT_SYMBOL(alloc_extent_map);
+
+/**
+ * free_extent_map - drop reference count of an extent_map
+ * @em:		extent map beeing releasead
+ *
+ * Drops the reference out on @em by one and free the structure
+ * if the reference count hits zero.
+ */
+void free_extent_map(struct extent_map *em)
+{
+	if (!em)
+		return;
+	WARN_ON(atomic_read(&em->refs) == 0);
+	if (atomic_dec_and_test(&em->refs)) {
+		WARN_ON(em->in_tree);
+		kmem_cache_free(extent_map_cache, em);
+	}
+}
+EXPORT_SYMBOL(free_extent_map);
+
+static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
+				   struct rb_node *node)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct extent_map *entry;
+
+	while (*p) {
+		parent = *p;
+		entry = rb_entry(parent, struct extent_map, rb_node);
+
+		WARN_ON(!entry->in_tree);
+
+		if (offset < entry->start)
+			p = &(*p)->rb_left;
+		else if (offset >= extent_map_end(entry))
+			p = &(*p)->rb_right;
+		else
+			return parent;
+	}
+
+	entry = rb_entry(node, struct extent_map, rb_node);
+	entry->in_tree = 1;
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+	return NULL;
+}
+
+/*
+ * search through the tree for an extent_map with a given offset.  If
+ * it can't be found, try to find some neighboring extents
+ */
+static struct rb_node *__tree_search(struct rb_root *root, u64 offset,
+				     struct rb_node **prev_ret,
+				     struct rb_node **next_ret)
+{
+	struct rb_node *n = root->rb_node;
+	struct rb_node *prev = NULL;
+	struct rb_node *orig_prev = NULL;
+	struct extent_map *entry;
+	struct extent_map *prev_entry = NULL;
+
+	while (n) {
+		entry = rb_entry(n, struct extent_map, rb_node);
+		prev = n;
+		prev_entry = entry;
+
+		WARN_ON(!entry->in_tree);
+
+		if (offset < entry->start)
+			n = n->rb_left;
+		else if (offset >= extent_map_end(entry))
+			n = n->rb_right;
+		else
+			return n;
+	}
+
+	if (prev_ret) {
+		orig_prev = prev;
+		while (prev && offset >= extent_map_end(prev_entry)) {
+			prev = rb_next(prev);
+			prev_entry = rb_entry(prev, struct extent_map, rb_node);
+		}
+		*prev_ret = prev;
+		prev = orig_prev;
+	}
+
+	if (next_ret) {
+		prev_entry = rb_entry(prev, struct extent_map, rb_node);
+		while (prev && offset < prev_entry->start) {
+			prev = rb_prev(prev);
+			prev_entry = rb_entry(prev, struct extent_map, rb_node);
+		}
+		*next_ret = prev;
+	}
+	return NULL;
+}
+
+/*
+ * look for an offset in the tree, and if it can't be found, return
+ * the first offset we can find smaller than 'offset'.
+ */
+static inline struct rb_node *tree_search(struct rb_root *root, u64 offset)
+{
+	struct rb_node *prev;
+	struct rb_node *ret;
+	ret = __tree_search(root, offset, &prev, NULL);
+	if (!ret)
+		return prev;
+	return ret;
+}
+
+/* check to see if two extent_map structs are adjacent and safe to merge */
+static int mergable_maps(struct extent_map *prev, struct extent_map *next)
+{
+	if (test_bit(EXTENT_FLAG_PINNED, &prev->flags))
+		return 0;
+
+	/*
+	 * don't merge compressed extents, we need to know their
+	 * actual size
+	 */
+	if (test_bit(EXTENT_FLAG_COMPRESSED, &prev->flags))
+		return 0;
+
+	if (extent_map_end(prev) == next->start &&
+	    prev->flags == next->flags &&
+	    prev->bdev == next->bdev &&
+	    ((next->block_start == EXTENT_MAP_HOLE &&
+	      prev->block_start == EXTENT_MAP_HOLE) ||
+	     (next->block_start == EXTENT_MAP_INLINE &&
+	      prev->block_start == EXTENT_MAP_INLINE) ||
+	     (next->block_start == EXTENT_MAP_DELALLOC &&
+	      prev->block_start == EXTENT_MAP_DELALLOC) ||
+	     (next->block_start < EXTENT_MAP_LAST_BYTE - 1 &&
+	      next->block_start == extent_map_block_end(prev)))) {
+		return 1;
+	}
+	return 0;
+}
+
+/**
+ * add_extent_mapping - add new extent map to the extent tree
+ * @tree:	tree to insert new map in
+ * @em:		map to insert
+ *
+ * Insert @em into @tree or perform a simple forward/backward merge with
+ * existing mappings.  The extent_map struct passed in will be inserted
+ * into the tree directly, with an additional reference taken, or a
+ * reference dropped if the merge attempt was sucessfull.
+ */
+int add_extent_mapping(struct extent_map_tree *tree,
+		       struct extent_map *em)
+{
+	int ret = 0;
+	struct extent_map *merge = NULL;
+	struct rb_node *rb;
+	struct extent_map *exist;
+
+	exist = lookup_extent_mapping(tree, em->start, em->len);
+	if (exist) {
+		free_extent_map(exist);
+		ret = -EEXIST;
+		goto out;
+	}
+	assert_spin_locked(&tree->lock);
+	rb = tree_insert(&tree->map, em->start, &em->rb_node);
+	if (rb) {
+		ret = -EEXIST;
+		free_extent_map(merge);
+		goto out;
+	}
+	atomic_inc(&em->refs);
+	if (em->start != 0) {
+		rb = rb_prev(&em->rb_node);
+		if (rb)
+			merge = rb_entry(rb, struct extent_map, rb_node);
+		if (rb && mergable_maps(merge, em)) {
+			em->start = merge->start;
+			em->len += merge->len;
+			em->block_len += merge->block_len;
+			em->block_start = merge->block_start;
+			merge->in_tree = 0;
+			rb_erase(&merge->rb_node, &tree->map);
+			free_extent_map(merge);
+		}
+	 }
+	rb = rb_next(&em->rb_node);
+	if (rb)
+		merge = rb_entry(rb, struct extent_map, rb_node);
+	if (rb && mergable_maps(em, merge)) {
+		em->len += merge->len;
+		em->block_len += merge->len;
+		rb_erase(&merge->rb_node, &tree->map);
+		merge->in_tree = 0;
+		free_extent_map(merge);
+	}
+out:
+	return ret;
+}
+EXPORT_SYMBOL(add_extent_mapping);
+
+/* simple helper to do math around the end of an extent, handling wrap */
+static u64 range_end(u64 start, u64 len)
+{
+	if (start + len < start)
+		return (u64)-1;
+	return start + len;
+}
+
+/**
+ * lookup_extent_mapping - lookup extent_map
+ * @tree:	tree to lookup in
+ * @start:	byte offset to start the search
+ * @len:	length of the lookup range
+ *
+ * Find and return the first extent_map struct in @tree that intersects the
+ * [start, len] range.  There may be additional objects in the tree that
+ * intersect, so check the object returned carefully to make sure that no
+ * additional lookups are needed.
+ */
+struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
+					 u64 start, u64 len)
+{
+	struct extent_map *em;
+	struct rb_node *rb_node;
+	struct rb_node *prev = NULL;
+	struct rb_node *next = NULL;
+	u64 end = range_end(start, len);
+
+	assert_spin_locked(&tree->lock);
+	rb_node = __tree_search(&tree->map, start, &prev, &next);
+	if (!rb_node && prev) {
+		em = rb_entry(prev, struct extent_map, rb_node);
+		if (end > em->start && start < extent_map_end(em))
+			goto found;
+	}
+	if (!rb_node && next) {
+		em = rb_entry(next, struct extent_map, rb_node);
+		if (end > em->start && start < extent_map_end(em))
+			goto found;
+	}
+	if (!rb_node) {
+		em = NULL;
+		goto out;
+	}
+	if (IS_ERR(rb_node)) {
+		em = ERR_PTR(PTR_ERR(rb_node));
+		goto out;
+	}
+	em = rb_entry(rb_node, struct extent_map, rb_node);
+	if (end > em->start && start < extent_map_end(em))
+		goto found;
+
+	em = NULL;
+	goto out;
+
+found:
+	atomic_inc(&em->refs);
+out:
+	return em;
+}
+EXPORT_SYMBOL(lookup_extent_mapping);
+
+/**
+ * remove_extent_mapping - removes an extent_map from the extent tree
+ * @tree:	extent tree to remove from
+ * @em:		extent map beeing removed
+ *
+ * Removes @em from @tree.  No reference counts are dropped, and no checks
+ * are done to see if the range is in use
+ */
+int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
+{
+	int ret = 0;
+
+	WARN_ON(test_bit(EXTENT_FLAG_PINNED, &em->flags));
+	assert_spin_locked(&tree->lock);
+	rb_erase(&em->rb_node, &tree->map);
+	em->in_tree = 0;
+	return ret;
+}
+EXPORT_SYMBOL(remove_extent_mapping);
diff --git a/fs/btrfs/extent_map.h b/fs/btrfs/extent_map.h
new file mode 100644
index 00000000000..fb6eeef06bb
--- /dev/null
+++ b/fs/btrfs/extent_map.h
@@ -0,0 +1,62 @@
+#ifndef __EXTENTMAP__
+#define __EXTENTMAP__
+
+#include <linux/rbtree.h>
+
+#define EXTENT_MAP_LAST_BYTE (u64)-4
+#define EXTENT_MAP_HOLE (u64)-3
+#define EXTENT_MAP_INLINE (u64)-2
+#define EXTENT_MAP_DELALLOC (u64)-1
+
+/* bits for the flags field */
+#define EXTENT_FLAG_PINNED 0 /* this entry not yet on disk, don't free it */
+#define EXTENT_FLAG_COMPRESSED 1
+#define EXTENT_FLAG_VACANCY 2 /* no file extent item found */
+#define EXTENT_FLAG_PREALLOC 3 /* pre-allocated extent */
+
+struct extent_map {
+	struct rb_node rb_node;
+
+	/* all of these are in bytes */
+	u64 start;
+	u64 len;
+	u64 orig_start;
+	u64 block_start;
+	u64 block_len;
+	unsigned long flags;
+	struct block_device *bdev;
+	atomic_t refs;
+	int in_tree;
+};
+
+struct extent_map_tree {
+	struct rb_root map;
+	spinlock_t lock;
+};
+
+static inline u64 extent_map_end(struct extent_map *em)
+{
+	if (em->start + em->len < em->start)
+		return (u64)-1;
+	return em->start + em->len;
+}
+
+static inline u64 extent_map_block_end(struct extent_map *em)
+{
+	if (em->block_start + em->block_len < em->block_start)
+		return (u64)-1;
+	return em->block_start + em->block_len;
+}
+
+void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask);
+struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
+					 u64 start, u64 len);
+int add_extent_mapping(struct extent_map_tree *tree,
+		       struct extent_map *em);
+int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em);
+
+struct extent_map *alloc_extent_map(gfp_t mask);
+void free_extent_map(struct extent_map *em);
+int __init extent_map_init(void);
+void extent_map_exit(void);
+#endif
diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c
new file mode 100644
index 00000000000..964652435fd
--- /dev/null
+++ b/fs/btrfs/file-item.c
@@ -0,0 +1,831 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/bio.h>
+#include <linux/pagemap.h>
+#include <linux/highmem.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "print-tree.h"
+
+#define MAX_CSUM_ITEMS(r, size) ((((BTRFS_LEAF_DATA_SIZE(r) - \
+				   sizeof(struct btrfs_item) * 2) / \
+				  size) - 1))
+
+#define MAX_ORDERED_SUM_BYTES(r) ((PAGE_SIZE - \
+				   sizeof(struct btrfs_ordered_sum)) / \
+				   sizeof(struct btrfs_sector_sum) * \
+				   (r)->sectorsize - (r)->sectorsize)
+
+int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     u64 objectid, u64 pos,
+			     u64 disk_offset, u64 disk_num_bytes,
+			     u64 num_bytes, u64 offset, u64 ram_bytes,
+			     u8 compression, u8 encryption, u16 other_encoding)
+{
+	int ret = 0;
+	struct btrfs_file_extent_item *item;
+	struct btrfs_key file_key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	file_key.objectid = objectid;
+	file_key.offset = pos;
+	btrfs_set_key_type(&file_key, BTRFS_EXTENT_DATA_KEY);
+
+	ret = btrfs_insert_empty_item(trans, root, path, &file_key,
+				      sizeof(*item));
+	if (ret < 0)
+		goto out;
+	BUG_ON(ret);
+	leaf = path->nodes[0];
+	item = btrfs_item_ptr(leaf, path->slots[0],
+			      struct btrfs_file_extent_item);
+	btrfs_set_file_extent_disk_bytenr(leaf, item, disk_offset);
+	btrfs_set_file_extent_disk_num_bytes(leaf, item, disk_num_bytes);
+	btrfs_set_file_extent_offset(leaf, item, offset);
+	btrfs_set_file_extent_num_bytes(leaf, item, num_bytes);
+	btrfs_set_file_extent_ram_bytes(leaf, item, ram_bytes);
+	btrfs_set_file_extent_generation(leaf, item, trans->transid);
+	btrfs_set_file_extent_type(leaf, item, BTRFS_FILE_EXTENT_REG);
+	btrfs_set_file_extent_compression(leaf, item, compression);
+	btrfs_set_file_extent_encryption(leaf, item, encryption);
+	btrfs_set_file_extent_other_encoding(leaf, item, other_encoding);
+
+	btrfs_mark_buffer_dirty(leaf);
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans,
+					  struct btrfs_root *root,
+					  struct btrfs_path *path,
+					  u64 bytenr, int cow)
+{
+	int ret;
+	struct btrfs_key file_key;
+	struct btrfs_key found_key;
+	struct btrfs_csum_item *item;
+	struct extent_buffer *leaf;
+	u64 csum_offset = 0;
+	u16 csum_size =
+		btrfs_super_csum_size(&root->fs_info->super_copy);
+	int csums_in_item;
+
+	file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
+	file_key.offset = bytenr;
+	btrfs_set_key_type(&file_key, BTRFS_EXTENT_CSUM_KEY);
+	ret = btrfs_search_slot(trans, root, &file_key, path, 0, cow);
+	if (ret < 0)
+		goto fail;
+	leaf = path->nodes[0];
+	if (ret > 0) {
+		ret = 1;
+		if (path->slots[0] == 0)
+			goto fail;
+		path->slots[0]--;
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (btrfs_key_type(&found_key) != BTRFS_EXTENT_CSUM_KEY)
+			goto fail;
+
+		csum_offset = (bytenr - found_key.offset) >>
+				root->fs_info->sb->s_blocksize_bits;
+		csums_in_item = btrfs_item_size_nr(leaf, path->slots[0]);
+		csums_in_item /= csum_size;
+
+		if (csum_offset >= csums_in_item) {
+			ret = -EFBIG;
+			goto fail;
+		}
+	}
+	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
+	item = (struct btrfs_csum_item *)((unsigned char *)item +
+					  csum_offset * csum_size);
+	return item;
+fail:
+	if (ret > 0)
+		ret = -ENOENT;
+	return ERR_PTR(ret);
+}
+
+
+int btrfs_lookup_file_extent(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct btrfs_path *path, u64 objectid,
+			     u64 offset, int mod)
+{
+	int ret;
+	struct btrfs_key file_key;
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+
+	file_key.objectid = objectid;
+	file_key.offset = offset;
+	btrfs_set_key_type(&file_key, BTRFS_EXTENT_DATA_KEY);
+	ret = btrfs_search_slot(trans, root, &file_key, path, ins_len, cow);
+	return ret;
+}
+
+
+int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode,
+			  struct bio *bio, u32 *dst)
+{
+	u32 sum;
+	struct bio_vec *bvec = bio->bi_io_vec;
+	int bio_index = 0;
+	u64 offset;
+	u64 item_start_offset = 0;
+	u64 item_last_offset = 0;
+	u64 disk_bytenr;
+	u32 diff;
+	u16 csum_size =
+		btrfs_super_csum_size(&root->fs_info->super_copy);
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_csum_item *item = NULL;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+
+	path = btrfs_alloc_path();
+	if (bio->bi_size > PAGE_CACHE_SIZE * 8)
+		path->reada = 2;
+
+	WARN_ON(bio->bi_vcnt <= 0);
+
+	disk_bytenr = (u64)bio->bi_sector << 9;
+	while (bio_index < bio->bi_vcnt) {
+		offset = page_offset(bvec->bv_page) + bvec->bv_offset;
+		ret = btrfs_find_ordered_sum(inode, offset, disk_bytenr, &sum);
+		if (ret == 0)
+			goto found;
+
+		if (!item || disk_bytenr < item_start_offset ||
+		    disk_bytenr >= item_last_offset) {
+			struct btrfs_key found_key;
+			u32 item_size;
+
+			if (item)
+				btrfs_release_path(root, path);
+			item = btrfs_lookup_csum(NULL, root->fs_info->csum_root,
+						 path, disk_bytenr, 0);
+			if (IS_ERR(item)) {
+				ret = PTR_ERR(item);
+				if (ret == -ENOENT || ret == -EFBIG)
+					ret = 0;
+				sum = 0;
+				if (BTRFS_I(inode)->root->root_key.objectid ==
+				    BTRFS_DATA_RELOC_TREE_OBJECTID) {
+					set_extent_bits(io_tree, offset,
+						offset + bvec->bv_len - 1,
+						EXTENT_NODATASUM, GFP_NOFS);
+				} else {
+					printk(KERN_INFO "btrfs no csum found "
+					       "for inode %lu start %llu\n",
+					       inode->i_ino,
+					       (unsigned long long)offset);
+				}
+				item = NULL;
+				btrfs_release_path(root, path);
+				goto found;
+			}
+			btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+					      path->slots[0]);
+
+			item_start_offset = found_key.offset;
+			item_size = btrfs_item_size_nr(path->nodes[0],
+						       path->slots[0]);
+			item_last_offset = item_start_offset +
+				(item_size / csum_size) *
+				root->sectorsize;
+			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+					      struct btrfs_csum_item);
+		}
+		/*
+		 * this byte range must be able to fit inside
+		 * a single leaf so it will also fit inside a u32
+		 */
+		diff = disk_bytenr - item_start_offset;
+		diff = diff / root->sectorsize;
+		diff = diff * csum_size;
+
+		read_extent_buffer(path->nodes[0], &sum,
+				   ((unsigned long)item) + diff,
+				   csum_size);
+found:
+		if (dst)
+			*dst++ = sum;
+		else
+			set_state_private(io_tree, offset, sum);
+		disk_bytenr += bvec->bv_len;
+		bio_index++;
+		bvec++;
+	}
+	btrfs_free_path(path);
+	return 0;
+}
+
+int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
+			     struct list_head *list)
+{
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_ordered_sum *sums;
+	struct btrfs_sector_sum *sector_sum;
+	struct btrfs_csum_item *item;
+	unsigned long offset;
+	int ret;
+	size_t size;
+	u64 csum_end;
+	u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy);
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
+	key.offset = start;
+	key.type = BTRFS_EXTENT_CSUM_KEY;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto fail;
+	if (ret > 0 && path->slots[0] > 0) {
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1);
+		if (key.objectid == BTRFS_EXTENT_CSUM_OBJECTID &&
+		    key.type == BTRFS_EXTENT_CSUM_KEY) {
+			offset = (start - key.offset) >>
+				 root->fs_info->sb->s_blocksize_bits;
+			if (offset * csum_size <
+			    btrfs_item_size_nr(leaf, path->slots[0] - 1))
+				path->slots[0]--;
+		}
+	}
+
+	while (start <= end) {
+		leaf = path->nodes[0];
+		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto fail;
+			if (ret > 0)
+				break;
+			leaf = path->nodes[0];
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+		if (key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
+		    key.type != BTRFS_EXTENT_CSUM_KEY)
+			break;
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+		if (key.offset > end)
+			break;
+
+		if (key.offset > start)
+			start = key.offset;
+
+		size = btrfs_item_size_nr(leaf, path->slots[0]);
+		csum_end = key.offset + (size / csum_size) * root->sectorsize;
+		if (csum_end <= start) {
+			path->slots[0]++;
+			continue;
+		}
+
+		csum_end = min(csum_end, end + 1);
+		item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				      struct btrfs_csum_item);
+		while (start < csum_end) {
+			size = min_t(size_t, csum_end - start,
+					MAX_ORDERED_SUM_BYTES(root));
+			sums = kzalloc(btrfs_ordered_sum_size(root, size),
+					GFP_NOFS);
+			BUG_ON(!sums);
+
+			sector_sum = sums->sums;
+			sums->bytenr = start;
+			sums->len = size;
+
+			offset = (start - key.offset) >>
+				root->fs_info->sb->s_blocksize_bits;
+			offset *= csum_size;
+
+			while (size > 0) {
+				read_extent_buffer(path->nodes[0],
+						&sector_sum->sum,
+						((unsigned long)item) +
+						offset, csum_size);
+				sector_sum->bytenr = start;
+
+				size -= root->sectorsize;
+				start += root->sectorsize;
+				offset += csum_size;
+				sector_sum++;
+			}
+			list_add_tail(&sums->list, list);
+		}
+		path->slots[0]++;
+	}
+	ret = 0;
+fail:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
+		       struct bio *bio, u64 file_start, int contig)
+{
+	struct btrfs_ordered_sum *sums;
+	struct btrfs_sector_sum *sector_sum;
+	struct btrfs_ordered_extent *ordered;
+	char *data;
+	struct bio_vec *bvec = bio->bi_io_vec;
+	int bio_index = 0;
+	unsigned long total_bytes = 0;
+	unsigned long this_sum_bytes = 0;
+	u64 offset;
+	u64 disk_bytenr;
+
+	WARN_ON(bio->bi_vcnt <= 0);
+	sums = kzalloc(btrfs_ordered_sum_size(root, bio->bi_size), GFP_NOFS);
+	if (!sums)
+		return -ENOMEM;
+
+	sector_sum = sums->sums;
+	disk_bytenr = (u64)bio->bi_sector << 9;
+	sums->len = bio->bi_size;
+	INIT_LIST_HEAD(&sums->list);
+
+	if (contig)
+		offset = file_start;
+	else
+		offset = page_offset(bvec->bv_page) + bvec->bv_offset;
+
+	ordered = btrfs_lookup_ordered_extent(inode, offset);
+	BUG_ON(!ordered);
+	sums->bytenr = ordered->start;
+
+	while (bio_index < bio->bi_vcnt) {
+		if (!contig)
+			offset = page_offset(bvec->bv_page) + bvec->bv_offset;
+
+		if (!contig && (offset >= ordered->file_offset + ordered->len ||
+		    offset < ordered->file_offset)) {
+			unsigned long bytes_left;
+			sums->len = this_sum_bytes;
+			this_sum_bytes = 0;
+			btrfs_add_ordered_sum(inode, ordered, sums);
+			btrfs_put_ordered_extent(ordered);
+
+			bytes_left = bio->bi_size - total_bytes;
+
+			sums = kzalloc(btrfs_ordered_sum_size(root, bytes_left),
+				       GFP_NOFS);
+			BUG_ON(!sums);
+			sector_sum = sums->sums;
+			sums->len = bytes_left;
+			ordered = btrfs_lookup_ordered_extent(inode, offset);
+			BUG_ON(!ordered);
+			sums->bytenr = ordered->start;
+		}
+
+		data = kmap_atomic(bvec->bv_page, KM_USER0);
+		sector_sum->sum = ~(u32)0;
+		sector_sum->sum = btrfs_csum_data(root,
+						  data + bvec->bv_offset,
+						  sector_sum->sum,
+						  bvec->bv_len);
+		kunmap_atomic(data, KM_USER0);
+		btrfs_csum_final(sector_sum->sum,
+				 (char *)&sector_sum->sum);
+		sector_sum->bytenr = disk_bytenr;
+
+		sector_sum++;
+		bio_index++;
+		total_bytes += bvec->bv_len;
+		this_sum_bytes += bvec->bv_len;
+		disk_bytenr += bvec->bv_len;
+		offset += bvec->bv_len;
+		bvec++;
+	}
+	this_sum_bytes = 0;
+	btrfs_add_ordered_sum(inode, ordered, sums);
+	btrfs_put_ordered_extent(ordered);
+	return 0;
+}
+
+/*
+ * helper function for csum removal, this expects the
+ * key to describe the csum pointed to by the path, and it expects
+ * the csum to overlap the range [bytenr, len]
+ *
+ * The csum should not be entirely contained in the range and the
+ * range should not be entirely contained in the csum.
+ *
+ * This calls btrfs_truncate_item with the correct args based on the
+ * overlap, and fixes up the key as required.
+ */
+static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      struct btrfs_key *key,
+				      u64 bytenr, u64 len)
+{
+	struct extent_buffer *leaf;
+	u16 csum_size =
+		btrfs_super_csum_size(&root->fs_info->super_copy);
+	u64 csum_end;
+	u64 end_byte = bytenr + len;
+	u32 blocksize_bits = root->fs_info->sb->s_blocksize_bits;
+	int ret;
+
+	leaf = path->nodes[0];
+	csum_end = btrfs_item_size_nr(leaf, path->slots[0]) / csum_size;
+	csum_end <<= root->fs_info->sb->s_blocksize_bits;
+	csum_end += key->offset;
+
+	if (key->offset < bytenr && csum_end <= end_byte) {
+		/*
+		 *         [ bytenr - len ]
+		 *         [   ]
+		 *   [csum     ]
+		 *   A simple truncate off the end of the item
+		 */
+		u32 new_size = (bytenr - key->offset) >> blocksize_bits;
+		new_size *= csum_size;
+		ret = btrfs_truncate_item(trans, root, path, new_size, 1);
+		BUG_ON(ret);
+	} else if (key->offset >= bytenr && csum_end > end_byte &&
+		   end_byte > key->offset) {
+		/*
+		 *         [ bytenr - len ]
+		 *                 [ ]
+		 *                 [csum     ]
+		 * we need to truncate from the beginning of the csum
+		 */
+		u32 new_size = (csum_end - end_byte) >> blocksize_bits;
+		new_size *= csum_size;
+
+		ret = btrfs_truncate_item(trans, root, path, new_size, 0);
+		BUG_ON(ret);
+
+		key->offset = end_byte;
+		ret = btrfs_set_item_key_safe(trans, root, path, key);
+		BUG_ON(ret);
+	} else {
+		BUG();
+	}
+	return 0;
+}
+
+/*
+ * deletes the csum items from the csum tree for a given
+ * range of bytes.
+ */
+int btrfs_del_csums(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, u64 bytenr, u64 len)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	u64 end_byte = bytenr + len;
+	u64 csum_end;
+	struct extent_buffer *leaf;
+	int ret;
+	u16 csum_size =
+		btrfs_super_csum_size(&root->fs_info->super_copy);
+	int blocksize_bits = root->fs_info->sb->s_blocksize_bits;
+
+	root = root->fs_info->csum_root;
+
+	path = btrfs_alloc_path();
+
+	while (1) {
+		key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
+		key.offset = end_byte - 1;
+		key.type = BTRFS_EXTENT_CSUM_KEY;
+
+		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+		if (ret > 0) {
+			if (path->slots[0] == 0)
+				goto out;
+			path->slots[0]--;
+		}
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+
+		if (key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
+		    key.type != BTRFS_EXTENT_CSUM_KEY) {
+			break;
+		}
+
+		if (key.offset >= end_byte)
+			break;
+
+		csum_end = btrfs_item_size_nr(leaf, path->slots[0]) / csum_size;
+		csum_end <<= blocksize_bits;
+		csum_end += key.offset;
+
+		/* this csum ends before we start, we're done */
+		if (csum_end <= bytenr)
+			break;
+
+		/* delete the entire item, it is inside our range */
+		if (key.offset >= bytenr && csum_end <= end_byte) {
+			ret = btrfs_del_item(trans, root, path);
+			BUG_ON(ret);
+			if (key.offset == bytenr)
+				break;
+		} else if (key.offset < bytenr && csum_end > end_byte) {
+			unsigned long offset;
+			unsigned long shift_len;
+			unsigned long item_offset;
+			/*
+			 *        [ bytenr - len ]
+			 *     [csum                ]
+			 *
+			 * Our bytes are in the middle of the csum,
+			 * we need to split this item and insert a new one.
+			 *
+			 * But we can't drop the path because the
+			 * csum could change, get removed, extended etc.
+			 *
+			 * The trick here is the max size of a csum item leaves
+			 * enough room in the tree block for a single
+			 * item header.  So, we split the item in place,
+			 * adding a new header pointing to the existing
+			 * bytes.  Then we loop around again and we have
+			 * a nicely formed csum item that we can neatly
+			 * truncate.
+			 */
+			offset = (bytenr - key.offset) >> blocksize_bits;
+			offset *= csum_size;
+
+			shift_len = (len >> blocksize_bits) * csum_size;
+
+			item_offset = btrfs_item_ptr_offset(leaf,
+							    path->slots[0]);
+
+			memset_extent_buffer(leaf, 0, item_offset + offset,
+					     shift_len);
+			key.offset = bytenr;
+
+			/*
+			 * btrfs_split_item returns -EAGAIN when the
+			 * item changed size or key
+			 */
+			ret = btrfs_split_item(trans, root, path, &key, offset);
+			BUG_ON(ret && ret != -EAGAIN);
+
+			key.offset = end_byte - 1;
+		} else {
+			ret = truncate_one_csum(trans, root, path,
+						&key, bytenr, len);
+			BUG_ON(ret);
+			if (key.offset < bytenr)
+				break;
+		}
+		btrfs_release_path(root, path);
+	}
+out:
+	btrfs_free_path(path);
+	return 0;
+}
+
+int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct btrfs_ordered_sum *sums)
+{
+	u64 bytenr;
+	int ret;
+	struct btrfs_key file_key;
+	struct btrfs_key found_key;
+	u64 next_offset;
+	u64 total_bytes = 0;
+	int found_next;
+	struct btrfs_path *path;
+	struct btrfs_csum_item *item;
+	struct btrfs_csum_item *item_end;
+	struct extent_buffer *leaf = NULL;
+	u64 csum_offset;
+	struct btrfs_sector_sum *sector_sum;
+	u32 nritems;
+	u32 ins_size;
+	char *eb_map;
+	char *eb_token;
+	unsigned long map_len;
+	unsigned long map_start;
+	u16 csum_size =
+		btrfs_super_csum_size(&root->fs_info->super_copy);
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	sector_sum = sums->sums;
+again:
+	next_offset = (u64)-1;
+	found_next = 0;
+	file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
+	file_key.offset = sector_sum->bytenr;
+	bytenr = sector_sum->bytenr;
+	btrfs_set_key_type(&file_key, BTRFS_EXTENT_CSUM_KEY);
+
+	item = btrfs_lookup_csum(trans, root, path, sector_sum->bytenr, 1);
+	if (!IS_ERR(item)) {
+		leaf = path->nodes[0];
+		ret = 0;
+		goto found;
+	}
+	ret = PTR_ERR(item);
+	if (ret == -EFBIG) {
+		u32 item_size;
+		/* we found one, but it isn't big enough yet */
+		leaf = path->nodes[0];
+		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+		if ((item_size / csum_size) >=
+		    MAX_CSUM_ITEMS(root, csum_size)) {
+			/* already at max size, make a new one */
+			goto insert;
+		}
+	} else {
+		int slot = path->slots[0] + 1;
+		/* we didn't find a csum item, insert one */
+		nritems = btrfs_header_nritems(path->nodes[0]);
+		if (path->slots[0] >= nritems - 1) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret == 1)
+				found_next = 1;
+			if (ret != 0)
+				goto insert;
+			slot = 0;
+		}
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key, slot);
+		if (found_key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
+		    found_key.type != BTRFS_EXTENT_CSUM_KEY) {
+			found_next = 1;
+			goto insert;
+		}
+		next_offset = found_key.offset;
+		found_next = 1;
+		goto insert;
+	}
+
+	/*
+	 * at this point, we know the tree has an item, but it isn't big
+	 * enough yet to put our csum in.  Grow it
+	 */
+	btrfs_release_path(root, path);
+	ret = btrfs_search_slot(trans, root, &file_key, path,
+				csum_size, 1);
+	if (ret < 0)
+		goto fail_unlock;
+
+	if (ret > 0) {
+		if (path->slots[0] == 0)
+			goto insert;
+		path->slots[0]--;
+	}
+
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+	csum_offset = (bytenr - found_key.offset) >>
+			root->fs_info->sb->s_blocksize_bits;
+
+	if (btrfs_key_type(&found_key) != BTRFS_EXTENT_CSUM_KEY ||
+	    found_key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
+	    csum_offset >= MAX_CSUM_ITEMS(root, csum_size)) {
+		goto insert;
+	}
+
+	if (csum_offset >= btrfs_item_size_nr(leaf, path->slots[0]) /
+	    csum_size) {
+		u32 diff = (csum_offset + 1) * csum_size;
+
+		/*
+		 * is the item big enough already?  we dropped our lock
+		 * before and need to recheck
+		 */
+		if (diff < btrfs_item_size_nr(leaf, path->slots[0]))
+			goto csum;
+
+		diff = diff - btrfs_item_size_nr(leaf, path->slots[0]);
+		if (diff != csum_size)
+			goto insert;
+
+		ret = btrfs_extend_item(trans, root, path, diff);
+		BUG_ON(ret);
+		goto csum;
+	}
+
+insert:
+	btrfs_release_path(root, path);
+	csum_offset = 0;
+	if (found_next) {
+		u64 tmp = total_bytes + root->sectorsize;
+		u64 next_sector = sector_sum->bytenr;
+		struct btrfs_sector_sum *next = sector_sum + 1;
+
+		while (tmp < sums->len) {
+			if (next_sector + root->sectorsize != next->bytenr)
+				break;
+			tmp += root->sectorsize;
+			next_sector = next->bytenr;
+			next++;
+		}
+		tmp = min(tmp, next_offset - file_key.offset);
+		tmp >>= root->fs_info->sb->s_blocksize_bits;
+		tmp = max((u64)1, tmp);
+		tmp = min(tmp, (u64)MAX_CSUM_ITEMS(root, csum_size));
+		ins_size = csum_size * tmp;
+	} else {
+		ins_size = csum_size;
+	}
+	ret = btrfs_insert_empty_item(trans, root, path, &file_key,
+				      ins_size);
+	if (ret < 0)
+		goto fail_unlock;
+	if (ret != 0) {
+		WARN_ON(1);
+		goto fail_unlock;
+	}
+csum:
+	leaf = path->nodes[0];
+	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
+	ret = 0;
+	item = (struct btrfs_csum_item *)((unsigned char *)item +
+					  csum_offset * csum_size);
+found:
+	item_end = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
+	item_end = (struct btrfs_csum_item *)((unsigned char *)item_end +
+				      btrfs_item_size_nr(leaf, path->slots[0]));
+	eb_token = NULL;
+	cond_resched();
+next_sector:
+
+	if (!eb_token ||
+	   (unsigned long)item + csum_size >= map_start + map_len) {
+		int err;
+
+		if (eb_token)
+			unmap_extent_buffer(leaf, eb_token, KM_USER1);
+		eb_token = NULL;
+		err = map_private_extent_buffer(leaf, (unsigned long)item,
+						csum_size,
+						&eb_token, &eb_map,
+						&map_start, &map_len, KM_USER1);
+		if (err)
+			eb_token = NULL;
+	}
+	if (eb_token) {
+		memcpy(eb_token + ((unsigned long)item & (PAGE_CACHE_SIZE - 1)),
+		       &sector_sum->sum, csum_size);
+	} else {
+		write_extent_buffer(leaf, &sector_sum->sum,
+				    (unsigned long)item, csum_size);
+	}
+
+	total_bytes += root->sectorsize;
+	sector_sum++;
+	if (total_bytes < sums->len) {
+		item = (struct btrfs_csum_item *)((char *)item +
+						  csum_size);
+		if (item < item_end && bytenr + PAGE_CACHE_SIZE ==
+		    sector_sum->bytenr) {
+			bytenr = sector_sum->bytenr;
+			goto next_sector;
+		}
+	}
+	if (eb_token) {
+		unmap_extent_buffer(leaf, eb_token, KM_USER1);
+		eb_token = NULL;
+	}
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	cond_resched();
+	if (total_bytes < sums->len) {
+		btrfs_release_path(root, path);
+		goto again;
+	}
+out:
+	btrfs_free_path(path);
+	return ret;
+
+fail_unlock:
+	goto out;
+}
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
new file mode 100644
index 00000000000..90268334145
--- /dev/null
+++ b/fs/btrfs/file.c
@@ -0,0 +1,1288 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/fs.h>
+#include <linux/pagemap.h>
+#include <linux/highmem.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/smp_lock.h>
+#include <linux/backing-dev.h>
+#include <linux/mpage.h>
+#include <linux/swap.h>
+#include <linux/writeback.h>
+#include <linux/statfs.h>
+#include <linux/compat.h>
+#include <linux/version.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "ioctl.h"
+#include "print-tree.h"
+#include "tree-log.h"
+#include "locking.h"
+#include "compat.h"
+
+
+/* simple helper to fault in pages and copy.  This should go away
+ * and be replaced with calls into generic code.
+ */
+static noinline int btrfs_copy_from_user(loff_t pos, int num_pages,
+					 int write_bytes,
+					 struct page **prepared_pages,
+					 const char __user *buf)
+{
+	long page_fault = 0;
+	int i;
+	int offset = pos & (PAGE_CACHE_SIZE - 1);
+
+	for (i = 0; i < num_pages && write_bytes > 0; i++, offset = 0) {
+		size_t count = min_t(size_t,
+				     PAGE_CACHE_SIZE - offset, write_bytes);
+		struct page *page = prepared_pages[i];
+		fault_in_pages_readable(buf, count);
+
+		/* Copy data from userspace to the current page */
+		kmap(page);
+		page_fault = __copy_from_user(page_address(page) + offset,
+					      buf, count);
+		/* Flush processor's dcache for this page */
+		flush_dcache_page(page);
+		kunmap(page);
+		buf += count;
+		write_bytes -= count;
+
+		if (page_fault)
+			break;
+	}
+	return page_fault ? -EFAULT : 0;
+}
+
+/*
+ * unlocks pages after btrfs_file_write is done with them
+ */
+static noinline void btrfs_drop_pages(struct page **pages, size_t num_pages)
+{
+	size_t i;
+	for (i = 0; i < num_pages; i++) {
+		if (!pages[i])
+			break;
+		/* page checked is some magic around finding pages that
+		 * have been modified without going through btrfs_set_page_dirty
+		 * clear it here
+		 */
+		ClearPageChecked(pages[i]);
+		unlock_page(pages[i]);
+		mark_page_accessed(pages[i]);
+		page_cache_release(pages[i]);
+	}
+}
+
+/*
+ * after copy_from_user, pages need to be dirtied and we need to make
+ * sure holes are created between the current EOF and the start of
+ * any next extents (if required).
+ *
+ * this also makes the decision about creating an inline extent vs
+ * doing real data extents, marking pages dirty and delalloc as required.
+ */
+static noinline int dirty_and_release_pages(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root,
+				   struct file *file,
+				   struct page **pages,
+				   size_t num_pages,
+				   loff_t pos,
+				   size_t write_bytes)
+{
+	int err = 0;
+	int i;
+	struct inode *inode = fdentry(file)->d_inode;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	u64 hint_byte;
+	u64 num_bytes;
+	u64 start_pos;
+	u64 end_of_last_block;
+	u64 end_pos = pos + write_bytes;
+	loff_t isize = i_size_read(inode);
+
+	start_pos = pos & ~((u64)root->sectorsize - 1);
+	num_bytes = (write_bytes + pos - start_pos +
+		    root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
+
+	end_of_last_block = start_pos + num_bytes - 1;
+
+	lock_extent(io_tree, start_pos, end_of_last_block, GFP_NOFS);
+	trans = btrfs_join_transaction(root, 1);
+	if (!trans) {
+		err = -ENOMEM;
+		goto out_unlock;
+	}
+	btrfs_set_trans_block_group(trans, inode);
+	hint_byte = 0;
+
+	set_extent_uptodate(io_tree, start_pos, end_of_last_block, GFP_NOFS);
+
+	/* check for reserved extents on each page, we don't want
+	 * to reset the delalloc bit on things that already have
+	 * extents reserved.
+	 */
+	btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block);
+	for (i = 0; i < num_pages; i++) {
+		struct page *p = pages[i];
+		SetPageUptodate(p);
+		ClearPageChecked(p);
+		set_page_dirty(p);
+	}
+	if (end_pos > isize) {
+		i_size_write(inode, end_pos);
+		btrfs_update_inode(trans, root, inode);
+	}
+	err = btrfs_end_transaction(trans, root);
+out_unlock:
+	unlock_extent(io_tree, start_pos, end_of_last_block, GFP_NOFS);
+	return err;
+}
+
+/*
+ * this drops all the extents in the cache that intersect the range
+ * [start, end].  Existing extents are split as required.
+ */
+int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
+			    int skip_pinned)
+{
+	struct extent_map *em;
+	struct extent_map *split = NULL;
+	struct extent_map *split2 = NULL;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	u64 len = end - start + 1;
+	int ret;
+	int testend = 1;
+	unsigned long flags;
+	int compressed = 0;
+
+	WARN_ON(end < start);
+	if (end == (u64)-1) {
+		len = (u64)-1;
+		testend = 0;
+	}
+	while (1) {
+		if (!split)
+			split = alloc_extent_map(GFP_NOFS);
+		if (!split2)
+			split2 = alloc_extent_map(GFP_NOFS);
+
+		spin_lock(&em_tree->lock);
+		em = lookup_extent_mapping(em_tree, start, len);
+		if (!em) {
+			spin_unlock(&em_tree->lock);
+			break;
+		}
+		flags = em->flags;
+		if (skip_pinned && test_bit(EXTENT_FLAG_PINNED, &em->flags)) {
+			spin_unlock(&em_tree->lock);
+			if (em->start <= start &&
+			    (!testend || em->start + em->len >= start + len)) {
+				free_extent_map(em);
+				break;
+			}
+			if (start < em->start) {
+				len = em->start - start;
+			} else {
+				len = start + len - (em->start + em->len);
+				start = em->start + em->len;
+			}
+			free_extent_map(em);
+			continue;
+		}
+		compressed = test_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+		clear_bit(EXTENT_FLAG_PINNED, &em->flags);
+		remove_extent_mapping(em_tree, em);
+
+		if (em->block_start < EXTENT_MAP_LAST_BYTE &&
+		    em->start < start) {
+			split->start = em->start;
+			split->len = start - em->start;
+			split->orig_start = em->orig_start;
+			split->block_start = em->block_start;
+
+			if (compressed)
+				split->block_len = em->block_len;
+			else
+				split->block_len = split->len;
+
+			split->bdev = em->bdev;
+			split->flags = flags;
+			ret = add_extent_mapping(em_tree, split);
+			BUG_ON(ret);
+			free_extent_map(split);
+			split = split2;
+			split2 = NULL;
+		}
+		if (em->block_start < EXTENT_MAP_LAST_BYTE &&
+		    testend && em->start + em->len > start + len) {
+			u64 diff = start + len - em->start;
+
+			split->start = start + len;
+			split->len = em->start + em->len - (start + len);
+			split->bdev = em->bdev;
+			split->flags = flags;
+
+			if (compressed) {
+				split->block_len = em->block_len;
+				split->block_start = em->block_start;
+				split->orig_start = em->orig_start;
+			} else {
+				split->block_len = split->len;
+				split->block_start = em->block_start + diff;
+				split->orig_start = split->start;
+			}
+
+			ret = add_extent_mapping(em_tree, split);
+			BUG_ON(ret);
+			free_extent_map(split);
+			split = NULL;
+		}
+		spin_unlock(&em_tree->lock);
+
+		/* once for us */
+		free_extent_map(em);
+		/* once for the tree*/
+		free_extent_map(em);
+	}
+	if (split)
+		free_extent_map(split);
+	if (split2)
+		free_extent_map(split2);
+	return 0;
+}
+
+int btrfs_check_file(struct btrfs_root *root, struct inode *inode)
+{
+	return 0;
+#if 0
+	struct btrfs_path *path;
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf;
+	struct btrfs_file_extent_item *extent;
+	u64 last_offset = 0;
+	int nritems;
+	int slot;
+	int found_type;
+	int ret;
+	int err = 0;
+	u64 extent_end = 0;
+
+	path = btrfs_alloc_path();
+	ret = btrfs_lookup_file_extent(NULL, root, path, inode->i_ino,
+				       last_offset, 0);
+	while (1) {
+		nritems = btrfs_header_nritems(path->nodes[0]);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret)
+				goto out;
+			nritems = btrfs_header_nritems(path->nodes[0]);
+		}
+		slot = path->slots[0];
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+		if (found_key.objectid != inode->i_ino)
+			break;
+		if (found_key.type != BTRFS_EXTENT_DATA_KEY)
+			goto out;
+
+		if (found_key.offset < last_offset) {
+			WARN_ON(1);
+			btrfs_print_leaf(root, leaf);
+			printk(KERN_ERR "inode %lu found offset %llu "
+			       "expected %llu\n", inode->i_ino,
+			       (unsigned long long)found_key.offset,
+			       (unsigned long long)last_offset);
+			err = 1;
+			goto out;
+		}
+		extent = btrfs_item_ptr(leaf, slot,
+					struct btrfs_file_extent_item);
+		found_type = btrfs_file_extent_type(leaf, extent);
+		if (found_type == BTRFS_FILE_EXTENT_REG) {
+			extent_end = found_key.offset +
+			     btrfs_file_extent_num_bytes(leaf, extent);
+		} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+			struct btrfs_item *item;
+			item = btrfs_item_nr(leaf, slot);
+			extent_end = found_key.offset +
+			     btrfs_file_extent_inline_len(leaf, extent);
+			extent_end = (extent_end + root->sectorsize - 1) &
+				~((u64)root->sectorsize - 1);
+		}
+		last_offset = extent_end;
+		path->slots[0]++;
+	}
+	if (0 && last_offset < inode->i_size) {
+		WARN_ON(1);
+		btrfs_print_leaf(root, leaf);
+		printk(KERN_ERR "inode %lu found offset %llu size %llu\n",
+		       inode->i_ino, (unsigned long long)last_offset,
+		       (unsigned long long)inode->i_size);
+		err = 1;
+
+	}
+out:
+	btrfs_free_path(path);
+	return err;
+#endif
+}
+
+/*
+ * this is very complex, but the basic idea is to drop all extents
+ * in the range start - end.  hint_block is filled in with a block number
+ * that would be a good hint to the block allocator for this file.
+ *
+ * If an extent intersects the range but is not entirely inside the range
+ * it is either truncated or split.  Anything entirely inside the range
+ * is deleted from the tree.
+ *
+ * inline_limit is used to tell this code which offsets in the file to keep
+ * if they contain inline extents.
+ */
+noinline int btrfs_drop_extents(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root, struct inode *inode,
+		       u64 start, u64 end, u64 inline_limit, u64 *hint_byte)
+{
+	u64 extent_end = 0;
+	u64 locked_end = end;
+	u64 search_start = start;
+	u64 leaf_start;
+	u64 ram_bytes = 0;
+	u64 orig_parent = 0;
+	u64 disk_bytenr = 0;
+	u8 compression;
+	u8 encryption;
+	u16 other_encoding = 0;
+	u64 root_gen;
+	u64 root_owner;
+	struct extent_buffer *leaf;
+	struct btrfs_file_extent_item *extent;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_file_extent_item old;
+	int keep;
+	int slot;
+	int bookend;
+	int found_type = 0;
+	int found_extent;
+	int found_inline;
+	int recow;
+	int ret;
+
+	inline_limit = 0;
+	btrfs_drop_extent_cache(inode, start, end - 1, 0);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	while (1) {
+		recow = 0;
+		btrfs_release_path(root, path);
+		ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+					       search_start, -1);
+		if (ret < 0)
+			goto out;
+		if (ret > 0) {
+			if (path->slots[0] == 0) {
+				ret = 0;
+				goto out;
+			}
+			path->slots[0]--;
+		}
+next_slot:
+		keep = 0;
+		bookend = 0;
+		found_extent = 0;
+		found_inline = 0;
+		leaf_start = 0;
+		root_gen = 0;
+		root_owner = 0;
+		compression = 0;
+		encryption = 0;
+		extent = NULL;
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		ret = 0;
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY &&
+		    key.offset >= end) {
+			goto out;
+		}
+		if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY ||
+		    key.objectid != inode->i_ino) {
+			goto out;
+		}
+		if (recow) {
+			search_start = max(key.offset, start);
+			continue;
+		}
+		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) {
+			extent = btrfs_item_ptr(leaf, slot,
+						struct btrfs_file_extent_item);
+			found_type = btrfs_file_extent_type(leaf, extent);
+			compression = btrfs_file_extent_compression(leaf,
+								    extent);
+			encryption = btrfs_file_extent_encryption(leaf,
+								  extent);
+			other_encoding = btrfs_file_extent_other_encoding(leaf,
+								  extent);
+			if (found_type == BTRFS_FILE_EXTENT_REG ||
+			    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
+				extent_end =
+				     btrfs_file_extent_disk_bytenr(leaf,
+								   extent);
+				if (extent_end)
+					*hint_byte = extent_end;
+
+				extent_end = key.offset +
+				     btrfs_file_extent_num_bytes(leaf, extent);
+				ram_bytes = btrfs_file_extent_ram_bytes(leaf,
+								extent);
+				found_extent = 1;
+			} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+				found_inline = 1;
+				extent_end = key.offset +
+				     btrfs_file_extent_inline_len(leaf, extent);
+			}
+		} else {
+			extent_end = search_start;
+		}
+
+		/* we found nothing we can drop */
+		if ((!found_extent && !found_inline) ||
+		    search_start >= extent_end) {
+			int nextret;
+			u32 nritems;
+			nritems = btrfs_header_nritems(leaf);
+			if (slot >= nritems - 1) {
+				nextret = btrfs_next_leaf(root, path);
+				if (nextret)
+					goto out;
+				recow = 1;
+			} else {
+				path->slots[0]++;
+			}
+			goto next_slot;
+		}
+
+		if (end <= extent_end && start >= key.offset && found_inline)
+			*hint_byte = EXTENT_MAP_INLINE;
+
+		if (found_extent) {
+			read_extent_buffer(leaf, &old, (unsigned long)extent,
+					   sizeof(old));
+			root_gen = btrfs_header_generation(leaf);
+			root_owner = btrfs_header_owner(leaf);
+			leaf_start = leaf->start;
+		}
+
+		if (end < extent_end && end >= key.offset) {
+			bookend = 1;
+			if (found_inline && start <= key.offset)
+				keep = 1;
+		}
+
+		if (bookend && found_extent) {
+			if (locked_end < extent_end) {
+				ret = try_lock_extent(&BTRFS_I(inode)->io_tree,
+						locked_end, extent_end - 1,
+						GFP_NOFS);
+				if (!ret) {
+					btrfs_release_path(root, path);
+					lock_extent(&BTRFS_I(inode)->io_tree,
+						locked_end, extent_end - 1,
+						GFP_NOFS);
+					locked_end = extent_end;
+					continue;
+				}
+				locked_end = extent_end;
+			}
+			orig_parent = path->nodes[0]->start;
+			disk_bytenr = le64_to_cpu(old.disk_bytenr);
+			if (disk_bytenr != 0) {
+				ret = btrfs_inc_extent_ref(trans, root,
+					   disk_bytenr,
+					   le64_to_cpu(old.disk_num_bytes),
+					   orig_parent, root->root_key.objectid,
+					   trans->transid, inode->i_ino);
+				BUG_ON(ret);
+			}
+		}
+
+		if (found_inline) {
+			u64 mask = root->sectorsize - 1;
+			search_start = (extent_end + mask) & ~mask;
+		} else
+			search_start = extent_end;
+
+		/* truncate existing extent */
+		if (start > key.offset) {
+			u64 new_num;
+			u64 old_num;
+			keep = 1;
+			WARN_ON(start & (root->sectorsize - 1));
+			if (found_extent) {
+				new_num = start - key.offset;
+				old_num = btrfs_file_extent_num_bytes(leaf,
+								      extent);
+				*hint_byte =
+					btrfs_file_extent_disk_bytenr(leaf,
+								      extent);
+				if (btrfs_file_extent_disk_bytenr(leaf,
+								  extent)) {
+					inode_sub_bytes(inode, old_num -
+							new_num);
+				}
+				btrfs_set_file_extent_num_bytes(leaf,
+							extent, new_num);
+				btrfs_mark_buffer_dirty(leaf);
+			} else if (key.offset < inline_limit &&
+				   (end > extent_end) &&
+				   (inline_limit < extent_end)) {
+				u32 new_size;
+				new_size = btrfs_file_extent_calc_inline_size(
+						   inline_limit - key.offset);
+				inode_sub_bytes(inode, extent_end -
+						inline_limit);
+				btrfs_set_file_extent_ram_bytes(leaf, extent,
+							new_size);
+				if (!compression && !encryption) {
+					btrfs_truncate_item(trans, root, path,
+							    new_size, 1);
+				}
+			}
+		}
+		/* delete the entire extent */
+		if (!keep) {
+			if (found_inline)
+				inode_sub_bytes(inode, extent_end -
+						key.offset);
+			ret = btrfs_del_item(trans, root, path);
+			/* TODO update progress marker and return */
+			BUG_ON(ret);
+			extent = NULL;
+			btrfs_release_path(root, path);
+			/* the extent will be freed later */
+		}
+		if (bookend && found_inline && start <= key.offset) {
+			u32 new_size;
+			new_size = btrfs_file_extent_calc_inline_size(
+						   extent_end - end);
+			inode_sub_bytes(inode, end - key.offset);
+			btrfs_set_file_extent_ram_bytes(leaf, extent,
+							new_size);
+			if (!compression && !encryption)
+				ret = btrfs_truncate_item(trans, root, path,
+							  new_size, 0);
+			BUG_ON(ret);
+		}
+		/* create bookend, splitting the extent in two */
+		if (bookend && found_extent) {
+			struct btrfs_key ins;
+			ins.objectid = inode->i_ino;
+			ins.offset = end;
+			btrfs_set_key_type(&ins, BTRFS_EXTENT_DATA_KEY);
+
+			btrfs_release_path(root, path);
+			ret = btrfs_insert_empty_item(trans, root, path, &ins,
+						      sizeof(*extent));
+			BUG_ON(ret);
+
+			leaf = path->nodes[0];
+			extent = btrfs_item_ptr(leaf, path->slots[0],
+						struct btrfs_file_extent_item);
+			write_extent_buffer(leaf, &old,
+					    (unsigned long)extent, sizeof(old));
+
+			btrfs_set_file_extent_compression(leaf, extent,
+							  compression);
+			btrfs_set_file_extent_encryption(leaf, extent,
+							 encryption);
+			btrfs_set_file_extent_other_encoding(leaf, extent,
+							     other_encoding);
+			btrfs_set_file_extent_offset(leaf, extent,
+				    le64_to_cpu(old.offset) + end - key.offset);
+			WARN_ON(le64_to_cpu(old.num_bytes) <
+				(extent_end - end));
+			btrfs_set_file_extent_num_bytes(leaf, extent,
+							extent_end - end);
+
+			/*
+			 * set the ram bytes to the size of the full extent
+			 * before splitting.  This is a worst case flag,
+			 * but its the best we can do because we don't know
+			 * how splitting affects compression
+			 */
+			btrfs_set_file_extent_ram_bytes(leaf, extent,
+							ram_bytes);
+			btrfs_set_file_extent_type(leaf, extent, found_type);
+
+			btrfs_mark_buffer_dirty(path->nodes[0]);
+
+			if (disk_bytenr != 0) {
+				ret = btrfs_update_extent_ref(trans, root,
+						disk_bytenr, orig_parent,
+						leaf->start,
+						root->root_key.objectid,
+						trans->transid, ins.objectid);
+
+				BUG_ON(ret);
+			}
+			btrfs_release_path(root, path);
+			if (disk_bytenr != 0)
+				inode_add_bytes(inode, extent_end - end);
+		}
+
+		if (found_extent && !keep) {
+			u64 old_disk_bytenr = le64_to_cpu(old.disk_bytenr);
+
+			if (old_disk_bytenr != 0) {
+				inode_sub_bytes(inode,
+						le64_to_cpu(old.num_bytes));
+				ret = btrfs_free_extent(trans, root,
+						old_disk_bytenr,
+						le64_to_cpu(old.disk_num_bytes),
+						leaf_start, root_owner,
+						root_gen, key.objectid, 0);
+				BUG_ON(ret);
+				*hint_byte = old_disk_bytenr;
+			}
+		}
+
+		if (search_start >= end) {
+			ret = 0;
+			goto out;
+		}
+	}
+out:
+	btrfs_free_path(path);
+	if (locked_end > end) {
+		unlock_extent(&BTRFS_I(inode)->io_tree, end, locked_end - 1,
+			      GFP_NOFS);
+	}
+	btrfs_check_file(root, inode);
+	return ret;
+}
+
+static int extent_mergeable(struct extent_buffer *leaf, int slot,
+			    u64 objectid, u64 bytenr, u64 *start, u64 *end)
+{
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_key key;
+	u64 extent_end;
+
+	if (slot < 0 || slot >= btrfs_header_nritems(leaf))
+		return 0;
+
+	btrfs_item_key_to_cpu(leaf, &key, slot);
+	if (key.objectid != objectid || key.type != BTRFS_EXTENT_DATA_KEY)
+		return 0;
+
+	fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+	if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG ||
+	    btrfs_file_extent_disk_bytenr(leaf, fi) != bytenr ||
+	    btrfs_file_extent_compression(leaf, fi) ||
+	    btrfs_file_extent_encryption(leaf, fi) ||
+	    btrfs_file_extent_other_encoding(leaf, fi))
+		return 0;
+
+	extent_end = key.offset + btrfs_file_extent_num_bytes(leaf, fi);
+	if ((*start && *start != key.offset) || (*end && *end != extent_end))
+		return 0;
+
+	*start = key.offset;
+	*end = extent_end;
+	return 1;
+}
+
+/*
+ * Mark extent in the range start - end as written.
+ *
+ * This changes extent type from 'pre-allocated' to 'regular'. If only
+ * part of extent is marked as written, the extent will be split into
+ * two or three.
+ */
+int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct inode *inode, u64 start, u64 end)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_path *path;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_key key;
+	u64 bytenr;
+	u64 num_bytes;
+	u64 extent_end;
+	u64 extent_offset;
+	u64 other_start;
+	u64 other_end;
+	u64 split = start;
+	u64 locked_end = end;
+	u64 orig_parent;
+	int extent_type;
+	int split_end = 1;
+	int ret;
+
+	btrfs_drop_extent_cache(inode, start, end - 1, 0);
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+again:
+	key.objectid = inode->i_ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	if (split == start)
+		key.offset = split;
+	else
+		key.offset = split - 1;
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret > 0 && path->slots[0] > 0)
+		path->slots[0]--;
+
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+	BUG_ON(key.objectid != inode->i_ino ||
+	       key.type != BTRFS_EXTENT_DATA_KEY);
+	fi = btrfs_item_ptr(leaf, path->slots[0],
+			    struct btrfs_file_extent_item);
+	extent_type = btrfs_file_extent_type(leaf, fi);
+	BUG_ON(extent_type != BTRFS_FILE_EXTENT_PREALLOC);
+	extent_end = key.offset + btrfs_file_extent_num_bytes(leaf, fi);
+	BUG_ON(key.offset > start || extent_end < end);
+
+	bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+	num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
+	extent_offset = btrfs_file_extent_offset(leaf, fi);
+
+	if (key.offset == start)
+		split = end;
+
+	if (key.offset == start && extent_end == end) {
+		int del_nr = 0;
+		int del_slot = 0;
+		u64 leaf_owner = btrfs_header_owner(leaf);
+		u64 leaf_gen = btrfs_header_generation(leaf);
+		other_start = end;
+		other_end = 0;
+		if (extent_mergeable(leaf, path->slots[0] + 1, inode->i_ino,
+				     bytenr, &other_start, &other_end)) {
+			extent_end = other_end;
+			del_slot = path->slots[0] + 1;
+			del_nr++;
+			ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
+						leaf->start, leaf_owner,
+						leaf_gen, inode->i_ino, 0);
+			BUG_ON(ret);
+		}
+		other_start = 0;
+		other_end = start;
+		if (extent_mergeable(leaf, path->slots[0] - 1, inode->i_ino,
+				     bytenr, &other_start, &other_end)) {
+			key.offset = other_start;
+			del_slot = path->slots[0];
+			del_nr++;
+			ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
+						leaf->start, leaf_owner,
+						leaf_gen, inode->i_ino, 0);
+			BUG_ON(ret);
+		}
+		split_end = 0;
+		if (del_nr == 0) {
+			btrfs_set_file_extent_type(leaf, fi,
+						   BTRFS_FILE_EXTENT_REG);
+			goto done;
+		}
+
+		fi = btrfs_item_ptr(leaf, del_slot - 1,
+				    struct btrfs_file_extent_item);
+		btrfs_set_file_extent_type(leaf, fi, BTRFS_FILE_EXTENT_REG);
+		btrfs_set_file_extent_num_bytes(leaf, fi,
+						extent_end - key.offset);
+		btrfs_mark_buffer_dirty(leaf);
+
+		ret = btrfs_del_items(trans, root, path, del_slot, del_nr);
+		BUG_ON(ret);
+		goto done;
+	} else if (split == start) {
+		if (locked_end < extent_end) {
+			ret = try_lock_extent(&BTRFS_I(inode)->io_tree,
+					locked_end, extent_end - 1, GFP_NOFS);
+			if (!ret) {
+				btrfs_release_path(root, path);
+				lock_extent(&BTRFS_I(inode)->io_tree,
+					locked_end, extent_end - 1, GFP_NOFS);
+				locked_end = extent_end;
+				goto again;
+			}
+			locked_end = extent_end;
+		}
+		btrfs_set_file_extent_num_bytes(leaf, fi, split - key.offset);
+		extent_offset += split - key.offset;
+	} else  {
+		BUG_ON(key.offset != start);
+		btrfs_set_file_extent_offset(leaf, fi, extent_offset +
+					     split - key.offset);
+		btrfs_set_file_extent_num_bytes(leaf, fi, extent_end - split);
+		key.offset = split;
+		btrfs_set_item_key_safe(trans, root, path, &key);
+		extent_end = split;
+	}
+
+	if (extent_end == end) {
+		split_end = 0;
+		extent_type = BTRFS_FILE_EXTENT_REG;
+	}
+	if (extent_end == end && split == start) {
+		other_start = end;
+		other_end = 0;
+		if (extent_mergeable(leaf, path->slots[0] + 1, inode->i_ino,
+				     bytenr, &other_start, &other_end)) {
+			path->slots[0]++;
+			fi = btrfs_item_ptr(leaf, path->slots[0],
+					    struct btrfs_file_extent_item);
+			key.offset = split;
+			btrfs_set_item_key_safe(trans, root, path, &key);
+			btrfs_set_file_extent_offset(leaf, fi, extent_offset);
+			btrfs_set_file_extent_num_bytes(leaf, fi,
+							other_end - split);
+			goto done;
+		}
+	}
+	if (extent_end == end && split == end) {
+		other_start = 0;
+		other_end = start;
+		if (extent_mergeable(leaf, path->slots[0] - 1 , inode->i_ino,
+				     bytenr, &other_start, &other_end)) {
+			path->slots[0]--;
+			fi = btrfs_item_ptr(leaf, path->slots[0],
+					    struct btrfs_file_extent_item);
+			btrfs_set_file_extent_num_bytes(leaf, fi, extent_end -
+							other_start);
+			goto done;
+		}
+	}
+
+	btrfs_mark_buffer_dirty(leaf);
+
+	orig_parent = leaf->start;
+	ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes,
+				   orig_parent, root->root_key.objectid,
+				   trans->transid, inode->i_ino);
+	BUG_ON(ret);
+	btrfs_release_path(root, path);
+
+	key.offset = start;
+	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*fi));
+	BUG_ON(ret);
+
+	leaf = path->nodes[0];
+	fi = btrfs_item_ptr(leaf, path->slots[0],
+			    struct btrfs_file_extent_item);
+	btrfs_set_file_extent_generation(leaf, fi, trans->transid);
+	btrfs_set_file_extent_type(leaf, fi, extent_type);
+	btrfs_set_file_extent_disk_bytenr(leaf, fi, bytenr);
+	btrfs_set_file_extent_disk_num_bytes(leaf, fi, num_bytes);
+	btrfs_set_file_extent_offset(leaf, fi, extent_offset);
+	btrfs_set_file_extent_num_bytes(leaf, fi, extent_end - key.offset);
+	btrfs_set_file_extent_ram_bytes(leaf, fi, num_bytes);
+	btrfs_set_file_extent_compression(leaf, fi, 0);
+	btrfs_set_file_extent_encryption(leaf, fi, 0);
+	btrfs_set_file_extent_other_encoding(leaf, fi, 0);
+
+	if (orig_parent != leaf->start) {
+		ret = btrfs_update_extent_ref(trans, root, bytenr,
+					      orig_parent, leaf->start,
+					      root->root_key.objectid,
+					      trans->transid, inode->i_ino);
+		BUG_ON(ret);
+	}
+done:
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_release_path(root, path);
+	if (split_end && split == start) {
+		split = end;
+		goto again;
+	}
+	if (locked_end > end) {
+		unlock_extent(&BTRFS_I(inode)->io_tree, end, locked_end - 1,
+			      GFP_NOFS);
+	}
+	btrfs_free_path(path);
+	return 0;
+}
+
+/*
+ * this gets pages into the page cache and locks them down, it also properly
+ * waits for data=ordered extents to finish before allowing the pages to be
+ * modified.
+ */
+static noinline int prepare_pages(struct btrfs_root *root, struct file *file,
+			 struct page **pages, size_t num_pages,
+			 loff_t pos, unsigned long first_index,
+			 unsigned long last_index, size_t write_bytes)
+{
+	int i;
+	unsigned long index = pos >> PAGE_CACHE_SHIFT;
+	struct inode *inode = fdentry(file)->d_inode;
+	int err = 0;
+	u64 start_pos;
+	u64 last_pos;
+
+	start_pos = pos & ~((u64)root->sectorsize - 1);
+	last_pos = ((u64)index + num_pages) << PAGE_CACHE_SHIFT;
+
+	if (start_pos > inode->i_size) {
+		err = btrfs_cont_expand(inode, start_pos);
+		if (err)
+			return err;
+	}
+
+	memset(pages, 0, num_pages * sizeof(struct page *));
+again:
+	for (i = 0; i < num_pages; i++) {
+		pages[i] = grab_cache_page(inode->i_mapping, index + i);
+		if (!pages[i]) {
+			err = -ENOMEM;
+			BUG_ON(1);
+		}
+		wait_on_page_writeback(pages[i]);
+	}
+	if (start_pos < inode->i_size) {
+		struct btrfs_ordered_extent *ordered;
+		lock_extent(&BTRFS_I(inode)->io_tree,
+			    start_pos, last_pos - 1, GFP_NOFS);
+		ordered = btrfs_lookup_first_ordered_extent(inode,
+							    last_pos - 1);
+		if (ordered &&
+		    ordered->file_offset + ordered->len > start_pos &&
+		    ordered->file_offset < last_pos) {
+			btrfs_put_ordered_extent(ordered);
+			unlock_extent(&BTRFS_I(inode)->io_tree,
+				      start_pos, last_pos - 1, GFP_NOFS);
+			for (i = 0; i < num_pages; i++) {
+				unlock_page(pages[i]);
+				page_cache_release(pages[i]);
+			}
+			btrfs_wait_ordered_range(inode, start_pos,
+						 last_pos - start_pos);
+			goto again;
+		}
+		if (ordered)
+			btrfs_put_ordered_extent(ordered);
+
+		clear_extent_bits(&BTRFS_I(inode)->io_tree, start_pos,
+				  last_pos - 1, EXTENT_DIRTY | EXTENT_DELALLOC,
+				  GFP_NOFS);
+		unlock_extent(&BTRFS_I(inode)->io_tree,
+			      start_pos, last_pos - 1, GFP_NOFS);
+	}
+	for (i = 0; i < num_pages; i++) {
+		clear_page_dirty_for_io(pages[i]);
+		set_page_extent_mapped(pages[i]);
+		WARN_ON(!PageLocked(pages[i]));
+	}
+	return 0;
+}
+
+static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
+				size_t count, loff_t *ppos)
+{
+	loff_t pos;
+	loff_t start_pos;
+	ssize_t num_written = 0;
+	ssize_t err = 0;
+	int ret = 0;
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct page **pages = NULL;
+	int nrptrs;
+	struct page *pinned[2];
+	unsigned long first_index;
+	unsigned long last_index;
+	int will_write;
+
+	will_write = ((file->f_flags & O_SYNC) || IS_SYNC(inode) ||
+		      (file->f_flags & O_DIRECT));
+
+	nrptrs = min((count + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE,
+		     PAGE_CACHE_SIZE / (sizeof(struct page *)));
+	pinned[0] = NULL;
+	pinned[1] = NULL;
+
+	pos = *ppos;
+	start_pos = pos;
+
+	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
+	current->backing_dev_info = inode->i_mapping->backing_dev_info;
+	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
+	if (err)
+		goto out_nolock;
+	if (count == 0)
+		goto out_nolock;
+
+	err = file_remove_suid(file);
+	if (err)
+		goto out_nolock;
+	file_update_time(file);
+
+	pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL);
+
+	mutex_lock(&inode->i_mutex);
+	BTRFS_I(inode)->sequence++;
+	first_index = pos >> PAGE_CACHE_SHIFT;
+	last_index = (pos + count) >> PAGE_CACHE_SHIFT;
+
+	/*
+	 * there are lots of better ways to do this, but this code
+	 * makes sure the first and last page in the file range are
+	 * up to date and ready for cow
+	 */
+	if ((pos & (PAGE_CACHE_SIZE - 1))) {
+		pinned[0] = grab_cache_page(inode->i_mapping, first_index);
+		if (!PageUptodate(pinned[0])) {
+			ret = btrfs_readpage(NULL, pinned[0]);
+			BUG_ON(ret);
+			wait_on_page_locked(pinned[0]);
+		} else {
+			unlock_page(pinned[0]);
+		}
+	}
+	if ((pos + count) & (PAGE_CACHE_SIZE - 1)) {
+		pinned[1] = grab_cache_page(inode->i_mapping, last_index);
+		if (!PageUptodate(pinned[1])) {
+			ret = btrfs_readpage(NULL, pinned[1]);
+			BUG_ON(ret);
+			wait_on_page_locked(pinned[1]);
+		} else {
+			unlock_page(pinned[1]);
+		}
+	}
+
+	while (count > 0) {
+		size_t offset = pos & (PAGE_CACHE_SIZE - 1);
+		size_t write_bytes = min(count, nrptrs *
+					(size_t)PAGE_CACHE_SIZE -
+					 offset);
+		size_t num_pages = (write_bytes + PAGE_CACHE_SIZE - 1) >>
+					PAGE_CACHE_SHIFT;
+
+		WARN_ON(num_pages > nrptrs);
+		memset(pages, 0, sizeof(struct page *) * nrptrs);
+
+		ret = btrfs_check_free_space(root, write_bytes, 0);
+		if (ret)
+			goto out;
+
+		ret = prepare_pages(root, file, pages, num_pages,
+				    pos, first_index, last_index,
+				    write_bytes);
+		if (ret)
+			goto out;
+
+		ret = btrfs_copy_from_user(pos, num_pages,
+					   write_bytes, pages, buf);
+		if (ret) {
+			btrfs_drop_pages(pages, num_pages);
+			goto out;
+		}
+
+		ret = dirty_and_release_pages(NULL, root, file, pages,
+					      num_pages, pos, write_bytes);
+		btrfs_drop_pages(pages, num_pages);
+		if (ret)
+			goto out;
+
+		if (will_write) {
+			btrfs_fdatawrite_range(inode->i_mapping, pos,
+					       pos + write_bytes - 1,
+					       WB_SYNC_NONE);
+		} else {
+			balance_dirty_pages_ratelimited_nr(inode->i_mapping,
+							   num_pages);
+			if (num_pages <
+			    (root->leafsize >> PAGE_CACHE_SHIFT) + 1)
+				btrfs_btree_balance_dirty(root, 1);
+			btrfs_throttle(root);
+		}
+
+		buf += write_bytes;
+		count -= write_bytes;
+		pos += write_bytes;
+		num_written += write_bytes;
+
+		cond_resched();
+	}
+out:
+	mutex_unlock(&inode->i_mutex);
+
+out_nolock:
+	kfree(pages);
+	if (pinned[0])
+		page_cache_release(pinned[0]);
+	if (pinned[1])
+		page_cache_release(pinned[1]);
+	*ppos = pos;
+
+	if (num_written > 0 && will_write) {
+		struct btrfs_trans_handle *trans;
+
+		err = btrfs_wait_ordered_range(inode, start_pos, num_written);
+		if (err)
+			num_written = err;
+
+		if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
+			trans = btrfs_start_transaction(root, 1);
+			ret = btrfs_log_dentry_safe(trans, root,
+						    file->f_dentry);
+			if (ret == 0) {
+				btrfs_sync_log(trans, root);
+				btrfs_end_transaction(trans, root);
+			} else {
+				btrfs_commit_transaction(trans, root);
+			}
+		}
+		if (file->f_flags & O_DIRECT) {
+			invalidate_mapping_pages(inode->i_mapping,
+			      start_pos >> PAGE_CACHE_SHIFT,
+			     (start_pos + num_written - 1) >> PAGE_CACHE_SHIFT);
+		}
+	}
+	current->backing_dev_info = NULL;
+	return num_written ? num_written : err;
+}
+
+int btrfs_release_file(struct inode *inode, struct file *filp)
+{
+	if (filp->private_data)
+		btrfs_ioctl_trans_end(filp);
+	return 0;
+}
+
+/*
+ * fsync call for both files and directories.  This logs the inode into
+ * the tree log instead of forcing full commits whenever possible.
+ *
+ * It needs to call filemap_fdatawait so that all ordered extent updates are
+ * in the metadata btree are up to date for copying to the log.
+ *
+ * It drops the inode mutex before doing the tree log commit.  This is an
+ * important optimization for directories because holding the mutex prevents
+ * new operations on the dir while we write to disk.
+ */
+int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
+{
+	struct inode *inode = dentry->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret = 0;
+	struct btrfs_trans_handle *trans;
+
+	/*
+	 * check the transaction that last modified this inode
+	 * and see if its already been committed
+	 */
+	if (!BTRFS_I(inode)->last_trans)
+		goto out;
+
+	mutex_lock(&root->fs_info->trans_mutex);
+	if (BTRFS_I(inode)->last_trans <=
+	    root->fs_info->last_trans_committed) {
+		BTRFS_I(inode)->last_trans = 0;
+		mutex_unlock(&root->fs_info->trans_mutex);
+		goto out;
+	}
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	root->fs_info->tree_log_batch++;
+	filemap_fdatawrite(inode->i_mapping);
+	btrfs_wait_ordered_range(inode, 0, (u64)-1);
+	root->fs_info->tree_log_batch++;
+
+	/*
+	 * ok we haven't committed the transaction yet, lets do a commit
+	 */
+	if (file->private_data)
+		btrfs_ioctl_trans_end(file);
+
+	trans = btrfs_start_transaction(root, 1);
+	if (!trans) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = btrfs_log_dentry_safe(trans, root, file->f_dentry);
+	if (ret < 0)
+		goto out;
+
+	/* we've logged all the items and now have a consistent
+	 * version of the file in the log.  It is possible that
+	 * someone will come in and modify the file, but that's
+	 * fine because the log is consistent on disk, and we
+	 * have references to all of the file's extents
+	 *
+	 * It is possible that someone will come in and log the
+	 * file again, but that will end up using the synchronization
+	 * inside btrfs_sync_log to keep things safe.
+	 */
+	mutex_unlock(&file->f_dentry->d_inode->i_mutex);
+
+	if (ret > 0) {
+		ret = btrfs_commit_transaction(trans, root);
+	} else {
+		btrfs_sync_log(trans, root);
+		ret = btrfs_end_transaction(trans, root);
+	}
+	mutex_lock(&file->f_dentry->d_inode->i_mutex);
+out:
+	return ret > 0 ? EIO : ret;
+}
+
+static struct vm_operations_struct btrfs_file_vm_ops = {
+	.fault		= filemap_fault,
+	.page_mkwrite	= btrfs_page_mkwrite,
+};
+
+static int btrfs_file_mmap(struct file	*filp, struct vm_area_struct *vma)
+{
+	vma->vm_ops = &btrfs_file_vm_ops;
+	file_accessed(filp);
+	return 0;
+}
+
+struct file_operations btrfs_file_operations = {
+	.llseek		= generic_file_llseek,
+	.read		= do_sync_read,
+	.aio_read       = generic_file_aio_read,
+	.splice_read	= generic_file_splice_read,
+	.write		= btrfs_file_write,
+	.mmap		= btrfs_file_mmap,
+	.open		= generic_file_open,
+	.release	= btrfs_release_file,
+	.fsync		= btrfs_sync_file,
+	.unlocked_ioctl	= btrfs_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= btrfs_ioctl,
+#endif
+};
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
new file mode 100644
index 00000000000..d1e5f0e84c5
--- /dev/null
+++ b/fs/btrfs/free-space-cache.c
@@ -0,0 +1,495 @@
+/*
+ * Copyright (C) 2008 Red Hat.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include "ctree.h"
+
+static int tree_insert_offset(struct rb_root *root, u64 offset,
+			      struct rb_node *node)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct btrfs_free_space *info;
+
+	while (*p) {
+		parent = *p;
+		info = rb_entry(parent, struct btrfs_free_space, offset_index);
+
+		if (offset < info->offset)
+			p = &(*p)->rb_left;
+		else if (offset > info->offset)
+			p = &(*p)->rb_right;
+		else
+			return -EEXIST;
+	}
+
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+
+	return 0;
+}
+
+static int tree_insert_bytes(struct rb_root *root, u64 bytes,
+			     struct rb_node *node)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct btrfs_free_space *info;
+
+	while (*p) {
+		parent = *p;
+		info = rb_entry(parent, struct btrfs_free_space, bytes_index);
+
+		if (bytes < info->bytes)
+			p = &(*p)->rb_left;
+		else
+			p = &(*p)->rb_right;
+	}
+
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+
+	return 0;
+}
+
+/*
+ * searches the tree for the given offset.  If contains is set we will return
+ * the free space that contains the given offset.  If contains is not set we
+ * will return the free space that starts at or after the given offset and is
+ * at least bytes long.
+ */
+static struct btrfs_free_space *tree_search_offset(struct rb_root *root,
+						   u64 offset, u64 bytes,
+						   int contains)
+{
+	struct rb_node *n = root->rb_node;
+	struct btrfs_free_space *entry, *ret = NULL;
+
+	while (n) {
+		entry = rb_entry(n, struct btrfs_free_space, offset_index);
+
+		if (offset < entry->offset) {
+			if (!contains &&
+			    (!ret || entry->offset < ret->offset) &&
+			    (bytes <= entry->bytes))
+				ret = entry;
+			n = n->rb_left;
+		} else if (offset > entry->offset) {
+			if ((entry->offset + entry->bytes - 1) >= offset &&
+			    bytes <= entry->bytes) {
+				ret = entry;
+				break;
+			}
+			n = n->rb_right;
+		} else {
+			if (bytes > entry->bytes) {
+				n = n->rb_right;
+				continue;
+			}
+			ret = entry;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+/*
+ * return a chunk at least bytes size, as close to offset that we can get.
+ */
+static struct btrfs_free_space *tree_search_bytes(struct rb_root *root,
+						  u64 offset, u64 bytes)
+{
+	struct rb_node *n = root->rb_node;
+	struct btrfs_free_space *entry, *ret = NULL;
+
+	while (n) {
+		entry = rb_entry(n, struct btrfs_free_space, bytes_index);
+
+		if (bytes < entry->bytes) {
+			/*
+			 * We prefer to get a hole size as close to the size we
+			 * are asking for so we don't take small slivers out of
+			 * huge holes, but we also want to get as close to the
+			 * offset as possible so we don't have a whole lot of
+			 * fragmentation.
+			 */
+			if (offset <= entry->offset) {
+				if (!ret)
+					ret = entry;
+				else if (entry->bytes < ret->bytes)
+					ret = entry;
+				else if (entry->offset < ret->offset)
+					ret = entry;
+			}
+			n = n->rb_left;
+		} else if (bytes > entry->bytes) {
+			n = n->rb_right;
+		} else {
+			/*
+			 * Ok we may have multiple chunks of the wanted size,
+			 * so we don't want to take the first one we find, we
+			 * want to take the one closest to our given offset, so
+			 * keep searching just in case theres a better match.
+			 */
+			n = n->rb_right;
+			if (offset > entry->offset)
+				continue;
+			else if (!ret || entry->offset < ret->offset)
+				ret = entry;
+		}
+	}
+
+	return ret;
+}
+
+static void unlink_free_space(struct btrfs_block_group_cache *block_group,
+			      struct btrfs_free_space *info)
+{
+	rb_erase(&info->offset_index, &block_group->free_space_offset);
+	rb_erase(&info->bytes_index, &block_group->free_space_bytes);
+}
+
+static int link_free_space(struct btrfs_block_group_cache *block_group,
+			   struct btrfs_free_space *info)
+{
+	int ret = 0;
+
+
+	ret = tree_insert_offset(&block_group->free_space_offset, info->offset,
+				 &info->offset_index);
+	if (ret)
+		return ret;
+
+	ret = tree_insert_bytes(&block_group->free_space_bytes, info->bytes,
+				&info->bytes_index);
+	if (ret)
+		return ret;
+
+	return ret;
+}
+
+static int __btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
+				  u64 offset, u64 bytes)
+{
+	struct btrfs_free_space *right_info;
+	struct btrfs_free_space *left_info;
+	struct btrfs_free_space *info = NULL;
+	struct btrfs_free_space *alloc_info;
+	int ret = 0;
+
+	alloc_info = kzalloc(sizeof(struct btrfs_free_space), GFP_NOFS);
+	if (!alloc_info)
+		return -ENOMEM;
+
+	/*
+	 * first we want to see if there is free space adjacent to the range we
+	 * are adding, if there is remove that struct and add a new one to
+	 * cover the entire range
+	 */
+	right_info = tree_search_offset(&block_group->free_space_offset,
+					offset+bytes, 0, 1);
+	left_info = tree_search_offset(&block_group->free_space_offset,
+				       offset-1, 0, 1);
+
+	if (right_info && right_info->offset == offset+bytes) {
+		unlink_free_space(block_group, right_info);
+		info = right_info;
+		info->offset = offset;
+		info->bytes += bytes;
+	} else if (right_info && right_info->offset != offset+bytes) {
+		printk(KERN_ERR "btrfs adding space in the middle of an "
+		       "existing free space area. existing: "
+		       "offset=%llu, bytes=%llu. new: offset=%llu, "
+		       "bytes=%llu\n", (unsigned long long)right_info->offset,
+		       (unsigned long long)right_info->bytes,
+		       (unsigned long long)offset,
+		       (unsigned long long)bytes);
+		BUG();
+	}
+
+	if (left_info) {
+		unlink_free_space(block_group, left_info);
+
+		if (unlikely((left_info->offset + left_info->bytes) !=
+			     offset)) {
+			printk(KERN_ERR "btrfs free space to the left "
+			       "of new free space isn't "
+			       "quite right. existing: offset=%llu, "
+			       "bytes=%llu. new: offset=%llu, bytes=%llu\n",
+			       (unsigned long long)left_info->offset,
+			       (unsigned long long)left_info->bytes,
+			       (unsigned long long)offset,
+			       (unsigned long long)bytes);
+			BUG();
+		}
+
+		if (info) {
+			info->offset = left_info->offset;
+			info->bytes += left_info->bytes;
+			kfree(left_info);
+		} else {
+			info = left_info;
+			info->bytes += bytes;
+		}
+	}
+
+	if (info) {
+		ret = link_free_space(block_group, info);
+		if (!ret)
+			info = NULL;
+		goto out;
+	}
+
+	info = alloc_info;
+	alloc_info = NULL;
+	info->offset = offset;
+	info->bytes = bytes;
+
+	ret = link_free_space(block_group, info);
+	if (ret)
+		kfree(info);
+out:
+	if (ret) {
+		printk(KERN_ERR "btrfs: unable to add free space :%d\n", ret);
+		if (ret == -EEXIST)
+			BUG();
+	}
+
+	kfree(alloc_info);
+
+	return ret;
+}
+
+static int
+__btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
+			  u64 offset, u64 bytes)
+{
+	struct btrfs_free_space *info;
+	int ret = 0;
+
+	info = tree_search_offset(&block_group->free_space_offset, offset, 0,
+				  1);
+
+	if (info && info->offset == offset) {
+		if (info->bytes < bytes) {
+			printk(KERN_ERR "Found free space at %llu, size %llu,"
+			       "trying to use %llu\n",
+			       (unsigned long long)info->offset,
+			       (unsigned long long)info->bytes,
+			       (unsigned long long)bytes);
+			WARN_ON(1);
+			ret = -EINVAL;
+			goto out;
+		}
+		unlink_free_space(block_group, info);
+
+		if (info->bytes == bytes) {
+			kfree(info);
+			goto out;
+		}
+
+		info->offset += bytes;
+		info->bytes -= bytes;
+
+		ret = link_free_space(block_group, info);
+		BUG_ON(ret);
+	} else if (info && info->offset < offset &&
+		   info->offset + info->bytes >= offset + bytes) {
+		u64 old_start = info->offset;
+		/*
+		 * we're freeing space in the middle of the info,
+		 * this can happen during tree log replay
+		 *
+		 * first unlink the old info and then
+		 * insert it again after the hole we're creating
+		 */
+		unlink_free_space(block_group, info);
+		if (offset + bytes < info->offset + info->bytes) {
+			u64 old_end = info->offset + info->bytes;
+
+			info->offset = offset + bytes;
+			info->bytes = old_end - info->offset;
+			ret = link_free_space(block_group, info);
+			BUG_ON(ret);
+		} else {
+			/* the hole we're creating ends at the end
+			 * of the info struct, just free the info
+			 */
+			kfree(info);
+		}
+
+		/* step two, insert a new info struct to cover anything
+		 * before the hole
+		 */
+		ret = __btrfs_add_free_space(block_group, old_start,
+					     offset - old_start);
+		BUG_ON(ret);
+	} else {
+		WARN_ON(1);
+	}
+out:
+	return ret;
+}
+
+int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
+			 u64 offset, u64 bytes)
+{
+	int ret;
+	struct btrfs_free_space *sp;
+
+	mutex_lock(&block_group->alloc_mutex);
+	ret = __btrfs_add_free_space(block_group, offset, bytes);
+	sp = tree_search_offset(&block_group->free_space_offset, offset, 0, 1);
+	BUG_ON(!sp);
+	mutex_unlock(&block_group->alloc_mutex);
+
+	return ret;
+}
+
+int btrfs_add_free_space_lock(struct btrfs_block_group_cache *block_group,
+			      u64 offset, u64 bytes)
+{
+	int ret;
+	struct btrfs_free_space *sp;
+
+	ret = __btrfs_add_free_space(block_group, offset, bytes);
+	sp = tree_search_offset(&block_group->free_space_offset, offset, 0, 1);
+	BUG_ON(!sp);
+
+	return ret;
+}
+
+int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
+			    u64 offset, u64 bytes)
+{
+	int ret = 0;
+
+	mutex_lock(&block_group->alloc_mutex);
+	ret = __btrfs_remove_free_space(block_group, offset, bytes);
+	mutex_unlock(&block_group->alloc_mutex);
+
+	return ret;
+}
+
+int btrfs_remove_free_space_lock(struct btrfs_block_group_cache *block_group,
+				 u64 offset, u64 bytes)
+{
+	int ret;
+
+	ret = __btrfs_remove_free_space(block_group, offset, bytes);
+
+	return ret;
+}
+
+void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
+			   u64 bytes)
+{
+	struct btrfs_free_space *info;
+	struct rb_node *n;
+	int count = 0;
+
+	for (n = rb_first(&block_group->free_space_offset); n; n = rb_next(n)) {
+		info = rb_entry(n, struct btrfs_free_space, offset_index);
+		if (info->bytes >= bytes)
+			count++;
+	}
+	printk(KERN_INFO "%d blocks of free space at or bigger than bytes is"
+	       "\n", count);
+}
+
+u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group)
+{
+	struct btrfs_free_space *info;
+	struct rb_node *n;
+	u64 ret = 0;
+
+	for (n = rb_first(&block_group->free_space_offset); n;
+	     n = rb_next(n)) {
+		info = rb_entry(n, struct btrfs_free_space, offset_index);
+		ret += info->bytes;
+	}
+
+	return ret;
+}
+
+void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
+{
+	struct btrfs_free_space *info;
+	struct rb_node *node;
+
+	mutex_lock(&block_group->alloc_mutex);
+	while ((node = rb_last(&block_group->free_space_bytes)) != NULL) {
+		info = rb_entry(node, struct btrfs_free_space, bytes_index);
+		unlink_free_space(block_group, info);
+		kfree(info);
+		if (need_resched()) {
+			mutex_unlock(&block_group->alloc_mutex);
+			cond_resched();
+			mutex_lock(&block_group->alloc_mutex);
+		}
+	}
+	mutex_unlock(&block_group->alloc_mutex);
+}
+
+#if 0
+static struct btrfs_free_space *btrfs_find_free_space_offset(struct
+						      btrfs_block_group_cache
+						      *block_group, u64 offset,
+						      u64 bytes)
+{
+	struct btrfs_free_space *ret;
+
+	mutex_lock(&block_group->alloc_mutex);
+	ret = tree_search_offset(&block_group->free_space_offset, offset,
+				 bytes, 0);
+	mutex_unlock(&block_group->alloc_mutex);
+
+	return ret;
+}
+
+static struct btrfs_free_space *btrfs_find_free_space_bytes(struct
+						     btrfs_block_group_cache
+						     *block_group, u64 offset,
+						     u64 bytes)
+{
+	struct btrfs_free_space *ret;
+
+	mutex_lock(&block_group->alloc_mutex);
+
+	ret = tree_search_bytes(&block_group->free_space_bytes, offset, bytes);
+	mutex_unlock(&block_group->alloc_mutex);
+
+	return ret;
+}
+#endif
+
+struct btrfs_free_space *btrfs_find_free_space(struct btrfs_block_group_cache
+					       *block_group, u64 offset,
+					       u64 bytes)
+{
+	struct btrfs_free_space *ret = NULL;
+
+	ret = tree_search_offset(&block_group->free_space_offset, offset,
+				 bytes, 0);
+	if (!ret)
+		ret = tree_search_bytes(&block_group->free_space_bytes,
+					offset, bytes);
+
+	return ret;
+}
diff --git a/fs/btrfs/hash.h b/fs/btrfs/hash.h
new file mode 100644
index 00000000000..2a020b27676
--- /dev/null
+++ b/fs/btrfs/hash.h
@@ -0,0 +1,27 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __HASH__
+#define __HASH__
+
+#include "crc32c.h"
+static inline u64 btrfs_name_hash(const char *name, int len)
+{
+	return btrfs_crc32c((u32)~1, name, len);
+}
+#endif
diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c
new file mode 100644
index 00000000000..3d46fa1f29a
--- /dev/null
+++ b/fs/btrfs/inode-item.c
@@ -0,0 +1,206 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+
+static int find_name_in_backref(struct btrfs_path *path, const char *name,
+			 int name_len, struct btrfs_inode_ref **ref_ret)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_inode_ref *ref;
+	unsigned long ptr;
+	unsigned long name_ptr;
+	u32 item_size;
+	u32 cur_offset = 0;
+	int len;
+
+	leaf = path->nodes[0];
+	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+	while (cur_offset < item_size) {
+		ref = (struct btrfs_inode_ref *)(ptr + cur_offset);
+		len = btrfs_inode_ref_name_len(leaf, ref);
+		name_ptr = (unsigned long)(ref + 1);
+		cur_offset += len + sizeof(*ref);
+		if (len != name_len)
+			continue;
+		if (memcmp_extent_buffer(leaf, name, name_ptr, name_len) == 0) {
+			*ref_ret = ref;
+			return 1;
+		}
+	}
+	return 0;
+}
+
+int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   const char *name, int name_len,
+			   u64 inode_objectid, u64 ref_objectid, u64 *index)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_inode_ref *ref;
+	struct extent_buffer *leaf;
+	unsigned long ptr;
+	unsigned long item_start;
+	u32 item_size;
+	u32 sub_item_len;
+	int ret;
+	int del_len = name_len + sizeof(*ref);
+
+	key.objectid = inode_objectid;
+	key.offset = ref_objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_REF_KEY);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	} else if (ret < 0) {
+		goto out;
+	}
+	if (!find_name_in_backref(path, name, name_len, &ref)) {
+		ret = -ENOENT;
+		goto out;
+	}
+	leaf = path->nodes[0];
+	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
+
+	if (index)
+		*index = btrfs_inode_ref_index(leaf, ref);
+
+	if (del_len == item_size) {
+		ret = btrfs_del_item(trans, root, path);
+		goto out;
+	}
+	ptr = (unsigned long)ref;
+	sub_item_len = name_len + sizeof(*ref);
+	item_start = btrfs_item_ptr_offset(leaf, path->slots[0]);
+	memmove_extent_buffer(leaf, ptr, ptr + sub_item_len,
+			      item_size - (ptr + sub_item_len - item_start));
+	ret = btrfs_truncate_item(trans, root, path,
+				  item_size - sub_item_len, 1);
+	BUG_ON(ret);
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   const char *name, int name_len,
+			   u64 inode_objectid, u64 ref_objectid, u64 index)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_inode_ref *ref;
+	unsigned long ptr;
+	int ret;
+	int ins_len = name_len + sizeof(*ref);
+
+	key.objectid = inode_objectid;
+	key.offset = ref_objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_REF_KEY);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &key,
+				      ins_len);
+	if (ret == -EEXIST) {
+		u32 old_size;
+
+		if (find_name_in_backref(path, name, name_len, &ref))
+			goto out;
+
+		old_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
+		ret = btrfs_extend_item(trans, root, path, ins_len);
+		BUG_ON(ret);
+		ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				     struct btrfs_inode_ref);
+		ref = (struct btrfs_inode_ref *)((unsigned long)ref + old_size);
+		btrfs_set_inode_ref_name_len(path->nodes[0], ref, name_len);
+		btrfs_set_inode_ref_index(path->nodes[0], ref, index);
+		ptr = (unsigned long)(ref + 1);
+		ret = 0;
+	} else if (ret < 0) {
+		goto out;
+	} else {
+		ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				     struct btrfs_inode_ref);
+		btrfs_set_inode_ref_name_len(path->nodes[0], ref, name_len);
+		btrfs_set_inode_ref_index(path->nodes[0], ref, index);
+		ptr = (unsigned long)(ref + 1);
+	}
+	write_extent_buffer(path->nodes[0], name, ptr, name_len);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_insert_empty_inode(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct btrfs_path *path, u64 objectid)
+{
+	struct btrfs_key key;
+	int ret;
+	key.objectid = objectid;
+	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+	key.offset = 0;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &key,
+				      sizeof(struct btrfs_inode_item));
+	if (ret == 0 && objectid > root->highest_inode)
+		root->highest_inode = objectid;
+	return ret;
+}
+
+int btrfs_lookup_inode(struct btrfs_trans_handle *trans, struct btrfs_root
+		       *root, struct btrfs_path *path,
+		       struct btrfs_key *location, int mod)
+{
+	int ins_len = mod < 0 ? -1 : 0;
+	int cow = mod != 0;
+	int ret;
+	int slot;
+	struct extent_buffer *leaf;
+	struct btrfs_key found_key;
+
+	ret = btrfs_search_slot(trans, root, location, path, ins_len, cow);
+	if (ret > 0 && btrfs_key_type(location) == BTRFS_ROOT_ITEM_KEY &&
+	    location->offset == (u64)-1 && path->slots[0] != 0) {
+		slot = path->slots[0] - 1;
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+		if (found_key.objectid == location->objectid &&
+		    btrfs_key_type(&found_key) == btrfs_key_type(location)) {
+			path->slots[0]--;
+			return 0;
+		}
+	}
+	return ret;
+}
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
new file mode 100644
index 00000000000..2aa79873eb4
--- /dev/null
+++ b/fs/btrfs/inode-map.c
@@ -0,0 +1,144 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+
+int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct extent_buffer *l;
+	struct btrfs_key search_key;
+	struct btrfs_key found_key;
+	int slot;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	search_key.objectid = BTRFS_LAST_FREE_OBJECTID;
+	search_key.type = -1;
+	search_key.offset = (u64)-1;
+	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
+	if (ret < 0)
+		goto error;
+	BUG_ON(ret == 0);
+	if (path->slots[0] > 0) {
+		slot = path->slots[0] - 1;
+		l = path->nodes[0];
+		btrfs_item_key_to_cpu(l, &found_key, slot);
+		*objectid = found_key.objectid;
+	} else {
+		*objectid = BTRFS_FIRST_FREE_OBJECTID;
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * walks the btree of allocated inodes and find a hole.
+ */
+int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     u64 dirid, u64 *objectid)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	int ret;
+	int slot = 0;
+	u64 last_ino = 0;
+	int start_found;
+	struct extent_buffer *l;
+	struct btrfs_key search_key;
+	u64 search_start = dirid;
+
+	mutex_lock(&root->objectid_mutex);
+	if (root->last_inode_alloc >= BTRFS_FIRST_FREE_OBJECTID &&
+	    root->last_inode_alloc < BTRFS_LAST_FREE_OBJECTID) {
+		*objectid = ++root->last_inode_alloc;
+		mutex_unlock(&root->objectid_mutex);
+		return 0;
+	}
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	search_start = max(search_start, BTRFS_FIRST_FREE_OBJECTID);
+	search_key.objectid = search_start;
+	search_key.type = 0;
+	search_key.offset = 0;
+
+	btrfs_init_path(path);
+	start_found = 0;
+	ret = btrfs_search_slot(trans, root, &search_key, path, 0, 0);
+	if (ret < 0)
+		goto error;
+
+	while (1) {
+		l = path->nodes[0];
+		slot = path->slots[0];
+		if (slot >= btrfs_header_nritems(l)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret == 0)
+				continue;
+			if (ret < 0)
+				goto error;
+			if (!start_found) {
+				*objectid = search_start;
+				start_found = 1;
+				goto found;
+			}
+			*objectid = last_ino > search_start ?
+				last_ino : search_start;
+			goto found;
+		}
+		btrfs_item_key_to_cpu(l, &key, slot);
+		if (key.objectid >= search_start) {
+			if (start_found) {
+				if (last_ino < search_start)
+					last_ino = search_start;
+				if (key.objectid > last_ino) {
+					*objectid = last_ino;
+					goto found;
+				}
+			} else if (key.objectid > search_start) {
+				*objectid = search_start;
+				goto found;
+			}
+		}
+		if (key.objectid >= BTRFS_LAST_FREE_OBJECTID)
+			break;
+
+		start_found = 1;
+		last_ino = key.objectid + 1;
+		path->slots[0]++;
+	}
+	BUG_ON(1);
+found:
+	btrfs_release_path(root, path);
+	btrfs_free_path(path);
+	BUG_ON(*objectid < search_start);
+	mutex_unlock(&root->objectid_mutex);
+	return 0;
+error:
+	btrfs_release_path(root, path);
+	btrfs_free_path(path);
+	mutex_unlock(&root->objectid_mutex);
+	return ret;
+}
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
new file mode 100644
index 00000000000..8adfe059ab4
--- /dev/null
+++ b/fs/btrfs/inode.c
@@ -0,0 +1,5035 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/kernel.h>
+#include <linux/bio.h>
+#include <linux/buffer_head.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/pagemap.h>
+#include <linux/highmem.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/smp_lock.h>
+#include <linux/backing-dev.h>
+#include <linux/mpage.h>
+#include <linux/swap.h>
+#include <linux/writeback.h>
+#include <linux/statfs.h>
+#include <linux/compat.h>
+#include <linux/bit_spinlock.h>
+#include <linux/version.h>
+#include <linux/xattr.h>
+#include <linux/posix_acl.h>
+#include <linux/falloc.h>
+#include "compat.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "ioctl.h"
+#include "print-tree.h"
+#include "volumes.h"
+#include "ordered-data.h"
+#include "xattr.h"
+#include "tree-log.h"
+#include "ref-cache.h"
+#include "compression.h"
+
+struct btrfs_iget_args {
+	u64 ino;
+	struct btrfs_root *root;
+};
+
+static struct inode_operations btrfs_dir_inode_operations;
+static struct inode_operations btrfs_symlink_inode_operations;
+static struct inode_operations btrfs_dir_ro_inode_operations;
+static struct inode_operations btrfs_special_inode_operations;
+static struct inode_operations btrfs_file_inode_operations;
+static struct address_space_operations btrfs_aops;
+static struct address_space_operations btrfs_symlink_aops;
+static struct file_operations btrfs_dir_file_operations;
+static struct extent_io_ops btrfs_extent_io_ops;
+
+static struct kmem_cache *btrfs_inode_cachep;
+struct kmem_cache *btrfs_trans_handle_cachep;
+struct kmem_cache *btrfs_transaction_cachep;
+struct kmem_cache *btrfs_bit_radix_cachep;
+struct kmem_cache *btrfs_path_cachep;
+
+#define S_SHIFT 12
+static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = {
+	[S_IFREG >> S_SHIFT]	= BTRFS_FT_REG_FILE,
+	[S_IFDIR >> S_SHIFT]	= BTRFS_FT_DIR,
+	[S_IFCHR >> S_SHIFT]	= BTRFS_FT_CHRDEV,
+	[S_IFBLK >> S_SHIFT]	= BTRFS_FT_BLKDEV,
+	[S_IFIFO >> S_SHIFT]	= BTRFS_FT_FIFO,
+	[S_IFSOCK >> S_SHIFT]	= BTRFS_FT_SOCK,
+	[S_IFLNK >> S_SHIFT]	= BTRFS_FT_SYMLINK,
+};
+
+static void btrfs_truncate(struct inode *inode);
+static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end);
+static noinline int cow_file_range(struct inode *inode,
+				   struct page *locked_page,
+				   u64 start, u64 end, int *page_started,
+				   unsigned long *nr_written, int unlock);
+
+/*
+ * a very lame attempt at stopping writes when the FS is 85% full.  There
+ * are countless ways this is incorrect, but it is better than nothing.
+ */
+int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
+			   int for_del)
+{
+	u64 total;
+	u64 used;
+	u64 thresh;
+	int ret = 0;
+
+	spin_lock(&root->fs_info->delalloc_lock);
+	total = btrfs_super_total_bytes(&root->fs_info->super_copy);
+	used = btrfs_super_bytes_used(&root->fs_info->super_copy);
+	if (for_del)
+		thresh = total * 90;
+	else
+		thresh = total * 85;
+
+	do_div(thresh, 100);
+
+	if (used + root->fs_info->delalloc_bytes + num_required > thresh)
+		ret = -ENOSPC;
+	spin_unlock(&root->fs_info->delalloc_lock);
+	return ret;
+}
+
+/*
+ * this does all the hard work for inserting an inline extent into
+ * the btree.  The caller should have done a btrfs_drop_extents so that
+ * no overlapping inline items exist in the btree
+ */
+static noinline int insert_inline_extent(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root, struct inode *inode,
+				u64 start, size_t size, size_t compressed_size,
+				struct page **compressed_pages)
+{
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct page *page = NULL;
+	char *kaddr;
+	unsigned long ptr;
+	struct btrfs_file_extent_item *ei;
+	int err = 0;
+	int ret;
+	size_t cur_size = size;
+	size_t datasize;
+	unsigned long offset;
+	int use_compress = 0;
+
+	if (compressed_size && compressed_pages) {
+		use_compress = 1;
+		cur_size = compressed_size;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	btrfs_set_trans_block_group(trans, inode);
+
+	key.objectid = inode->i_ino;
+	key.offset = start;
+	btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
+	datasize = btrfs_file_extent_calc_inline_size(cur_size);
+
+	inode_add_bytes(inode, size);
+	ret = btrfs_insert_empty_item(trans, root, path, &key,
+				      datasize);
+	BUG_ON(ret);
+	if (ret) {
+		err = ret;
+		goto fail;
+	}
+	leaf = path->nodes[0];
+	ei = btrfs_item_ptr(leaf, path->slots[0],
+			    struct btrfs_file_extent_item);
+	btrfs_set_file_extent_generation(leaf, ei, trans->transid);
+	btrfs_set_file_extent_type(leaf, ei, BTRFS_FILE_EXTENT_INLINE);
+	btrfs_set_file_extent_encryption(leaf, ei, 0);
+	btrfs_set_file_extent_other_encoding(leaf, ei, 0);
+	btrfs_set_file_extent_ram_bytes(leaf, ei, size);
+	ptr = btrfs_file_extent_inline_start(ei);
+
+	if (use_compress) {
+		struct page *cpage;
+		int i = 0;
+		while (compressed_size > 0) {
+			cpage = compressed_pages[i];
+			cur_size = min_t(unsigned long, compressed_size,
+				       PAGE_CACHE_SIZE);
+
+			kaddr = kmap(cpage);
+			write_extent_buffer(leaf, kaddr, ptr, cur_size);
+			kunmap(cpage);
+
+			i++;
+			ptr += cur_size;
+			compressed_size -= cur_size;
+		}
+		btrfs_set_file_extent_compression(leaf, ei,
+						  BTRFS_COMPRESS_ZLIB);
+	} else {
+		page = find_get_page(inode->i_mapping,
+				     start >> PAGE_CACHE_SHIFT);
+		btrfs_set_file_extent_compression(leaf, ei, 0);
+		kaddr = kmap_atomic(page, KM_USER0);
+		offset = start & (PAGE_CACHE_SIZE - 1);
+		write_extent_buffer(leaf, kaddr + offset, ptr, size);
+		kunmap_atomic(kaddr, KM_USER0);
+		page_cache_release(page);
+	}
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_free_path(path);
+
+	BTRFS_I(inode)->disk_i_size = inode->i_size;
+	btrfs_update_inode(trans, root, inode);
+	return 0;
+fail:
+	btrfs_free_path(path);
+	return err;
+}
+
+
+/*
+ * conditionally insert an inline extent into the file.  This
+ * does the checks required to make sure the data is small enough
+ * to fit as an inline extent.
+ */
+static int cow_file_range_inline(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct inode *inode, u64 start, u64 end,
+				 size_t compressed_size,
+				 struct page **compressed_pages)
+{
+	u64 isize = i_size_read(inode);
+	u64 actual_end = min(end + 1, isize);
+	u64 inline_len = actual_end - start;
+	u64 aligned_end = (end + root->sectorsize - 1) &
+			~((u64)root->sectorsize - 1);
+	u64 hint_byte;
+	u64 data_len = inline_len;
+	int ret;
+
+	if (compressed_size)
+		data_len = compressed_size;
+
+	if (start > 0 ||
+	    actual_end >= PAGE_CACHE_SIZE ||
+	    data_len >= BTRFS_MAX_INLINE_DATA_SIZE(root) ||
+	    (!compressed_size &&
+	    (actual_end & (root->sectorsize - 1)) == 0) ||
+	    end + 1 < isize ||
+	    data_len > root->fs_info->max_inline) {
+		return 1;
+	}
+
+	ret = btrfs_drop_extents(trans, root, inode, start,
+				 aligned_end, start, &hint_byte);
+	BUG_ON(ret);
+
+	if (isize > actual_end)
+		inline_len = min_t(u64, isize, actual_end);
+	ret = insert_inline_extent(trans, root, inode, start,
+				   inline_len, compressed_size,
+				   compressed_pages);
+	BUG_ON(ret);
+	btrfs_drop_extent_cache(inode, start, aligned_end, 0);
+	return 0;
+}
+
+struct async_extent {
+	u64 start;
+	u64 ram_size;
+	u64 compressed_size;
+	struct page **pages;
+	unsigned long nr_pages;
+	struct list_head list;
+};
+
+struct async_cow {
+	struct inode *inode;
+	struct btrfs_root *root;
+	struct page *locked_page;
+	u64 start;
+	u64 end;
+	struct list_head extents;
+	struct btrfs_work work;
+};
+
+static noinline int add_async_extent(struct async_cow *cow,
+				     u64 start, u64 ram_size,
+				     u64 compressed_size,
+				     struct page **pages,
+				     unsigned long nr_pages)
+{
+	struct async_extent *async_extent;
+
+	async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS);
+	async_extent->start = start;
+	async_extent->ram_size = ram_size;
+	async_extent->compressed_size = compressed_size;
+	async_extent->pages = pages;
+	async_extent->nr_pages = nr_pages;
+	list_add_tail(&async_extent->list, &cow->extents);
+	return 0;
+}
+
+/*
+ * we create compressed extents in two phases.  The first
+ * phase compresses a range of pages that have already been
+ * locked (both pages and state bits are locked).
+ *
+ * This is done inside an ordered work queue, and the compression
+ * is spread across many cpus.  The actual IO submission is step
+ * two, and the ordered work queue takes care of making sure that
+ * happens in the same order things were put onto the queue by
+ * writepages and friends.
+ *
+ * If this code finds it can't get good compression, it puts an
+ * entry onto the work queue to write the uncompressed bytes.  This
+ * makes sure that both compressed inodes and uncompressed inodes
+ * are written in the same order that pdflush sent them down.
+ */
+static noinline int compress_file_range(struct inode *inode,
+					struct page *locked_page,
+					u64 start, u64 end,
+					struct async_cow *async_cow,
+					int *num_added)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	u64 num_bytes;
+	u64 orig_start;
+	u64 disk_num_bytes;
+	u64 blocksize = root->sectorsize;
+	u64 actual_end;
+	u64 isize = i_size_read(inode);
+	int ret = 0;
+	struct page **pages = NULL;
+	unsigned long nr_pages;
+	unsigned long nr_pages_ret = 0;
+	unsigned long total_compressed = 0;
+	unsigned long total_in = 0;
+	unsigned long max_compressed = 128 * 1024;
+	unsigned long max_uncompressed = 128 * 1024;
+	int i;
+	int will_compress;
+
+	orig_start = start;
+
+	actual_end = min_t(u64, isize, end + 1);
+again:
+	will_compress = 0;
+	nr_pages = (end >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT) + 1;
+	nr_pages = min(nr_pages, (128 * 1024UL) / PAGE_CACHE_SIZE);
+
+	total_compressed = actual_end - start;
+
+	/* we want to make sure that amount of ram required to uncompress
+	 * an extent is reasonable, so we limit the total size in ram
+	 * of a compressed extent to 128k.  This is a crucial number
+	 * because it also controls how easily we can spread reads across
+	 * cpus for decompression.
+	 *
+	 * We also want to make sure the amount of IO required to do
+	 * a random read is reasonably small, so we limit the size of
+	 * a compressed extent to 128k.
+	 */
+	total_compressed = min(total_compressed, max_uncompressed);
+	num_bytes = (end - start + blocksize) & ~(blocksize - 1);
+	num_bytes = max(blocksize,  num_bytes);
+	disk_num_bytes = num_bytes;
+	total_in = 0;
+	ret = 0;
+
+	/*
+	 * we do compression for mount -o compress and when the
+	 * inode has not been flagged as nocompress.  This flag can
+	 * change at any time if we discover bad compression ratios.
+	 */
+	if (!btrfs_test_flag(inode, NOCOMPRESS) &&
+	    btrfs_test_opt(root, COMPRESS)) {
+		WARN_ON(pages);
+		pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS);
+
+		ret = btrfs_zlib_compress_pages(inode->i_mapping, start,
+						total_compressed, pages,
+						nr_pages, &nr_pages_ret,
+						&total_in,
+						&total_compressed,
+						max_compressed);
+
+		if (!ret) {
+			unsigned long offset = total_compressed &
+				(PAGE_CACHE_SIZE - 1);
+			struct page *page = pages[nr_pages_ret - 1];
+			char *kaddr;
+
+			/* zero the tail end of the last page, we might be
+			 * sending it down to disk
+			 */
+			if (offset) {
+				kaddr = kmap_atomic(page, KM_USER0);
+				memset(kaddr + offset, 0,
+				       PAGE_CACHE_SIZE - offset);
+				kunmap_atomic(kaddr, KM_USER0);
+			}
+			will_compress = 1;
+		}
+	}
+	if (start == 0) {
+		trans = btrfs_join_transaction(root, 1);
+		BUG_ON(!trans);
+		btrfs_set_trans_block_group(trans, inode);
+
+		/* lets try to make an inline extent */
+		if (ret || total_in < (actual_end - start)) {
+			/* we didn't compress the entire range, try
+			 * to make an uncompressed inline extent.
+			 */
+			ret = cow_file_range_inline(trans, root, inode,
+						    start, end, 0, NULL);
+		} else {
+			/* try making a compressed inline extent */
+			ret = cow_file_range_inline(trans, root, inode,
+						    start, end,
+						    total_compressed, pages);
+		}
+		btrfs_end_transaction(trans, root);
+		if (ret == 0) {
+			/*
+			 * inline extent creation worked, we don't need
+			 * to create any more async work items.  Unlock
+			 * and free up our temp pages.
+			 */
+			extent_clear_unlock_delalloc(inode,
+						     &BTRFS_I(inode)->io_tree,
+						     start, end, NULL, 1, 0,
+						     0, 1, 1, 1);
+			ret = 0;
+			goto free_pages_out;
+		}
+	}
+
+	if (will_compress) {
+		/*
+		 * we aren't doing an inline extent round the compressed size
+		 * up to a block size boundary so the allocator does sane
+		 * things
+		 */
+		total_compressed = (total_compressed + blocksize - 1) &
+			~(blocksize - 1);
+
+		/*
+		 * one last check to make sure the compression is really a
+		 * win, compare the page count read with the blocks on disk
+		 */
+		total_in = (total_in + PAGE_CACHE_SIZE - 1) &
+			~(PAGE_CACHE_SIZE - 1);
+		if (total_compressed >= total_in) {
+			will_compress = 0;
+		} else {
+			disk_num_bytes = total_compressed;
+			num_bytes = total_in;
+		}
+	}
+	if (!will_compress && pages) {
+		/*
+		 * the compression code ran but failed to make things smaller,
+		 * free any pages it allocated and our page pointer array
+		 */
+		for (i = 0; i < nr_pages_ret; i++) {
+			WARN_ON(pages[i]->mapping);
+			page_cache_release(pages[i]);
+		}
+		kfree(pages);
+		pages = NULL;
+		total_compressed = 0;
+		nr_pages_ret = 0;
+
+		/* flag the file so we don't compress in the future */
+		btrfs_set_flag(inode, NOCOMPRESS);
+	}
+	if (will_compress) {
+		*num_added += 1;
+
+		/* the async work queues will take care of doing actual
+		 * allocation on disk for these compressed pages,
+		 * and will submit them to the elevator.
+		 */
+		add_async_extent(async_cow, start, num_bytes,
+				 total_compressed, pages, nr_pages_ret);
+
+		if (start + num_bytes < end && start + num_bytes < actual_end) {
+			start += num_bytes;
+			pages = NULL;
+			cond_resched();
+			goto again;
+		}
+	} else {
+		/*
+		 * No compression, but we still need to write the pages in
+		 * the file we've been given so far.  redirty the locked
+		 * page if it corresponds to our extent and set things up
+		 * for the async work queue to run cow_file_range to do
+		 * the normal delalloc dance
+		 */
+		if (page_offset(locked_page) >= start &&
+		    page_offset(locked_page) <= end) {
+			__set_page_dirty_nobuffers(locked_page);
+			/* unlocked later on in the async handlers */
+		}
+		add_async_extent(async_cow, start, end - start + 1, 0, NULL, 0);
+		*num_added += 1;
+	}
+
+out:
+	return 0;
+
+free_pages_out:
+	for (i = 0; i < nr_pages_ret; i++) {
+		WARN_ON(pages[i]->mapping);
+		page_cache_release(pages[i]);
+	}
+	kfree(pages);
+
+	goto out;
+}
+
+/*
+ * phase two of compressed writeback.  This is the ordered portion
+ * of the code, which only gets called in the order the work was
+ * queued.  We walk all the async extents created by compress_file_range
+ * and send them down to the disk.
+ */
+static noinline int submit_compressed_extents(struct inode *inode,
+					      struct async_cow *async_cow)
+{
+	struct async_extent *async_extent;
+	u64 alloc_hint = 0;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_key ins;
+	struct extent_map *em;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct extent_io_tree *io_tree;
+	int ret;
+
+	if (list_empty(&async_cow->extents))
+		return 0;
+
+	trans = btrfs_join_transaction(root, 1);
+
+	while (!list_empty(&async_cow->extents)) {
+		async_extent = list_entry(async_cow->extents.next,
+					  struct async_extent, list);
+		list_del(&async_extent->list);
+
+		io_tree = &BTRFS_I(inode)->io_tree;
+
+		/* did the compression code fall back to uncompressed IO? */
+		if (!async_extent->pages) {
+			int page_started = 0;
+			unsigned long nr_written = 0;
+
+			lock_extent(io_tree, async_extent->start,
+				    async_extent->start +
+				    async_extent->ram_size - 1, GFP_NOFS);
+
+			/* allocate blocks */
+			cow_file_range(inode, async_cow->locked_page,
+				       async_extent->start,
+				       async_extent->start +
+				       async_extent->ram_size - 1,
+				       &page_started, &nr_written, 0);
+
+			/*
+			 * if page_started, cow_file_range inserted an
+			 * inline extent and took care of all the unlocking
+			 * and IO for us.  Otherwise, we need to submit
+			 * all those pages down to the drive.
+			 */
+			if (!page_started)
+				extent_write_locked_range(io_tree,
+						  inode, async_extent->start,
+						  async_extent->start +
+						  async_extent->ram_size - 1,
+						  btrfs_get_extent,
+						  WB_SYNC_ALL);
+			kfree(async_extent);
+			cond_resched();
+			continue;
+		}
+
+		lock_extent(io_tree, async_extent->start,
+			    async_extent->start + async_extent->ram_size - 1,
+			    GFP_NOFS);
+		/*
+		 * here we're doing allocation and writeback of the
+		 * compressed pages
+		 */
+		btrfs_drop_extent_cache(inode, async_extent->start,
+					async_extent->start +
+					async_extent->ram_size - 1, 0);
+
+		ret = btrfs_reserve_extent(trans, root,
+					   async_extent->compressed_size,
+					   async_extent->compressed_size,
+					   0, alloc_hint,
+					   (u64)-1, &ins, 1);
+		BUG_ON(ret);
+		em = alloc_extent_map(GFP_NOFS);
+		em->start = async_extent->start;
+		em->len = async_extent->ram_size;
+		em->orig_start = em->start;
+
+		em->block_start = ins.objectid;
+		em->block_len = ins.offset;
+		em->bdev = root->fs_info->fs_devices->latest_bdev;
+		set_bit(EXTENT_FLAG_PINNED, &em->flags);
+		set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+
+		while (1) {
+			spin_lock(&em_tree->lock);
+			ret = add_extent_mapping(em_tree, em);
+			spin_unlock(&em_tree->lock);
+			if (ret != -EEXIST) {
+				free_extent_map(em);
+				break;
+			}
+			btrfs_drop_extent_cache(inode, async_extent->start,
+						async_extent->start +
+						async_extent->ram_size - 1, 0);
+		}
+
+		ret = btrfs_add_ordered_extent(inode, async_extent->start,
+					       ins.objectid,
+					       async_extent->ram_size,
+					       ins.offset,
+					       BTRFS_ORDERED_COMPRESSED);
+		BUG_ON(ret);
+
+		btrfs_end_transaction(trans, root);
+
+		/*
+		 * clear dirty, set writeback and unlock the pages.
+		 */
+		extent_clear_unlock_delalloc(inode,
+					     &BTRFS_I(inode)->io_tree,
+					     async_extent->start,
+					     async_extent->start +
+					     async_extent->ram_size - 1,
+					     NULL, 1, 1, 0, 1, 1, 0);
+
+		ret = btrfs_submit_compressed_write(inode,
+				    async_extent->start,
+				    async_extent->ram_size,
+				    ins.objectid,
+				    ins.offset, async_extent->pages,
+				    async_extent->nr_pages);
+
+		BUG_ON(ret);
+		trans = btrfs_join_transaction(root, 1);
+		alloc_hint = ins.objectid + ins.offset;
+		kfree(async_extent);
+		cond_resched();
+	}
+
+	btrfs_end_transaction(trans, root);
+	return 0;
+}
+
+/*
+ * when extent_io.c finds a delayed allocation range in the file,
+ * the call backs end up in this code.  The basic idea is to
+ * allocate extents on disk for the range, and create ordered data structs
+ * in ram to track those extents.
+ *
+ * locked_page is the page that writepage had locked already.  We use
+ * it to make sure we don't do extra locks or unlocks.
+ *
+ * *page_started is set to one if we unlock locked_page and do everything
+ * required to start IO on it.  It may be clean and already done with
+ * IO when we return.
+ */
+static noinline int cow_file_range(struct inode *inode,
+				   struct page *locked_page,
+				   u64 start, u64 end, int *page_started,
+				   unsigned long *nr_written,
+				   int unlock)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	u64 alloc_hint = 0;
+	u64 num_bytes;
+	unsigned long ram_size;
+	u64 disk_num_bytes;
+	u64 cur_alloc_size;
+	u64 blocksize = root->sectorsize;
+	u64 actual_end;
+	u64 isize = i_size_read(inode);
+	struct btrfs_key ins;
+	struct extent_map *em;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	int ret = 0;
+
+	trans = btrfs_join_transaction(root, 1);
+	BUG_ON(!trans);
+	btrfs_set_trans_block_group(trans, inode);
+
+	actual_end = min_t(u64, isize, end + 1);
+
+	num_bytes = (end - start + blocksize) & ~(blocksize - 1);
+	num_bytes = max(blocksize,  num_bytes);
+	disk_num_bytes = num_bytes;
+	ret = 0;
+
+	if (start == 0) {
+		/* lets try to make an inline extent */
+		ret = cow_file_range_inline(trans, root, inode,
+					    start, end, 0, NULL);
+		if (ret == 0) {
+			extent_clear_unlock_delalloc(inode,
+						     &BTRFS_I(inode)->io_tree,
+						     start, end, NULL, 1, 1,
+						     1, 1, 1, 1);
+			*nr_written = *nr_written +
+			     (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE;
+			*page_started = 1;
+			ret = 0;
+			goto out;
+		}
+	}
+
+	BUG_ON(disk_num_bytes >
+	       btrfs_super_total_bytes(&root->fs_info->super_copy));
+
+	btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0);
+
+	while (disk_num_bytes > 0) {
+		cur_alloc_size = min(disk_num_bytes, root->fs_info->max_extent);
+		ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
+					   root->sectorsize, 0, alloc_hint,
+					   (u64)-1, &ins, 1);
+		BUG_ON(ret);
+
+		em = alloc_extent_map(GFP_NOFS);
+		em->start = start;
+		em->orig_start = em->start;
+
+		ram_size = ins.offset;
+		em->len = ins.offset;
+
+		em->block_start = ins.objectid;
+		em->block_len = ins.offset;
+		em->bdev = root->fs_info->fs_devices->latest_bdev;
+		set_bit(EXTENT_FLAG_PINNED, &em->flags);
+
+		while (1) {
+			spin_lock(&em_tree->lock);
+			ret = add_extent_mapping(em_tree, em);
+			spin_unlock(&em_tree->lock);
+			if (ret != -EEXIST) {
+				free_extent_map(em);
+				break;
+			}
+			btrfs_drop_extent_cache(inode, start,
+						start + ram_size - 1, 0);
+		}
+
+		cur_alloc_size = ins.offset;
+		ret = btrfs_add_ordered_extent(inode, start, ins.objectid,
+					       ram_size, cur_alloc_size, 0);
+		BUG_ON(ret);
+
+		if (root->root_key.objectid ==
+		    BTRFS_DATA_RELOC_TREE_OBJECTID) {
+			ret = btrfs_reloc_clone_csums(inode, start,
+						      cur_alloc_size);
+			BUG_ON(ret);
+		}
+
+		if (disk_num_bytes < cur_alloc_size)
+			break;
+
+		/* we're not doing compressed IO, don't unlock the first
+		 * page (which the caller expects to stay locked), don't
+		 * clear any dirty bits and don't set any writeback bits
+		 */
+		extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
+					     start, start + ram_size - 1,
+					     locked_page, unlock, 1,
+					     1, 0, 0, 0);
+		disk_num_bytes -= cur_alloc_size;
+		num_bytes -= cur_alloc_size;
+		alloc_hint = ins.objectid + ins.offset;
+		start += cur_alloc_size;
+	}
+out:
+	ret = 0;
+	btrfs_end_transaction(trans, root);
+
+	return ret;
+}
+
+/*
+ * work queue call back to started compression on a file and pages
+ */
+static noinline void async_cow_start(struct btrfs_work *work)
+{
+	struct async_cow *async_cow;
+	int num_added = 0;
+	async_cow = container_of(work, struct async_cow, work);
+
+	compress_file_range(async_cow->inode, async_cow->locked_page,
+			    async_cow->start, async_cow->end, async_cow,
+			    &num_added);
+	if (num_added == 0)
+		async_cow->inode = NULL;
+}
+
+/*
+ * work queue call back to submit previously compressed pages
+ */
+static noinline void async_cow_submit(struct btrfs_work *work)
+{
+	struct async_cow *async_cow;
+	struct btrfs_root *root;
+	unsigned long nr_pages;
+
+	async_cow = container_of(work, struct async_cow, work);
+
+	root = async_cow->root;
+	nr_pages = (async_cow->end - async_cow->start + PAGE_CACHE_SIZE) >>
+		PAGE_CACHE_SHIFT;
+
+	atomic_sub(nr_pages, &root->fs_info->async_delalloc_pages);
+
+	if (atomic_read(&root->fs_info->async_delalloc_pages) <
+	    5 * 1042 * 1024 &&
+	    waitqueue_active(&root->fs_info->async_submit_wait))
+		wake_up(&root->fs_info->async_submit_wait);
+
+	if (async_cow->inode)
+		submit_compressed_extents(async_cow->inode, async_cow);
+}
+
+static noinline void async_cow_free(struct btrfs_work *work)
+{
+	struct async_cow *async_cow;
+	async_cow = container_of(work, struct async_cow, work);
+	kfree(async_cow);
+}
+
+static int cow_file_range_async(struct inode *inode, struct page *locked_page,
+				u64 start, u64 end, int *page_started,
+				unsigned long *nr_written)
+{
+	struct async_cow *async_cow;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	unsigned long nr_pages;
+	u64 cur_end;
+	int limit = 10 * 1024 * 1042;
+
+	if (!btrfs_test_opt(root, COMPRESS)) {
+		return cow_file_range(inode, locked_page, start, end,
+				      page_started, nr_written, 1);
+	}
+
+	clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED |
+			 EXTENT_DELALLOC, 1, 0, GFP_NOFS);
+	while (start < end) {
+		async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS);
+		async_cow->inode = inode;
+		async_cow->root = root;
+		async_cow->locked_page = locked_page;
+		async_cow->start = start;
+
+		if (btrfs_test_flag(inode, NOCOMPRESS))
+			cur_end = end;
+		else
+			cur_end = min(end, start + 512 * 1024 - 1);
+
+		async_cow->end = cur_end;
+		INIT_LIST_HEAD(&async_cow->extents);
+
+		async_cow->work.func = async_cow_start;
+		async_cow->work.ordered_func = async_cow_submit;
+		async_cow->work.ordered_free = async_cow_free;
+		async_cow->work.flags = 0;
+
+		nr_pages = (cur_end - start + PAGE_CACHE_SIZE) >>
+			PAGE_CACHE_SHIFT;
+		atomic_add(nr_pages, &root->fs_info->async_delalloc_pages);
+
+		btrfs_queue_worker(&root->fs_info->delalloc_workers,
+				   &async_cow->work);
+
+		if (atomic_read(&root->fs_info->async_delalloc_pages) > limit) {
+			wait_event(root->fs_info->async_submit_wait,
+			   (atomic_read(&root->fs_info->async_delalloc_pages) <
+			    limit));
+		}
+
+		while (atomic_read(&root->fs_info->async_submit_draining) &&
+		      atomic_read(&root->fs_info->async_delalloc_pages)) {
+			wait_event(root->fs_info->async_submit_wait,
+			  (atomic_read(&root->fs_info->async_delalloc_pages) ==
+			   0));
+		}
+
+		*nr_written += nr_pages;
+		start = cur_end + 1;
+	}
+	*page_started = 1;
+	return 0;
+}
+
+static noinline int csum_exist_in_range(struct btrfs_root *root,
+					u64 bytenr, u64 num_bytes)
+{
+	int ret;
+	struct btrfs_ordered_sum *sums;
+	LIST_HEAD(list);
+
+	ret = btrfs_lookup_csums_range(root->fs_info->csum_root, bytenr,
+				       bytenr + num_bytes - 1, &list);
+	if (ret == 0 && list_empty(&list))
+		return 0;
+
+	while (!list_empty(&list)) {
+		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
+		list_del(&sums->list);
+		kfree(sums);
+	}
+	return 1;
+}
+
+/*
+ * when nowcow writeback call back.  This checks for snapshots or COW copies
+ * of the extents that exist in the file, and COWs the file as required.
+ *
+ * If no cow copies or snapshots exist, we write directly to the existing
+ * blocks on disk
+ */
+static int run_delalloc_nocow(struct inode *inode, struct page *locked_page,
+			      u64 start, u64 end, int *page_started, int force,
+			      unsigned long *nr_written)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	struct extent_buffer *leaf;
+	struct btrfs_path *path;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_key found_key;
+	u64 cow_start;
+	u64 cur_offset;
+	u64 extent_end;
+	u64 disk_bytenr;
+	u64 num_bytes;
+	int extent_type;
+	int ret;
+	int type;
+	int nocow;
+	int check_prev = 1;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	trans = btrfs_join_transaction(root, 1);
+	BUG_ON(!trans);
+
+	cow_start = (u64)-1;
+	cur_offset = start;
+	while (1) {
+		ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+					       cur_offset, 0);
+		BUG_ON(ret < 0);
+		if (ret > 0 && path->slots[0] > 0 && check_prev) {
+			leaf = path->nodes[0];
+			btrfs_item_key_to_cpu(leaf, &found_key,
+					      path->slots[0] - 1);
+			if (found_key.objectid == inode->i_ino &&
+			    found_key.type == BTRFS_EXTENT_DATA_KEY)
+				path->slots[0]--;
+		}
+		check_prev = 0;
+next_slot:
+		leaf = path->nodes[0];
+		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				BUG_ON(1);
+			if (ret > 0)
+				break;
+			leaf = path->nodes[0];
+		}
+
+		nocow = 0;
+		disk_bytenr = 0;
+		num_bytes = 0;
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+		if (found_key.objectid > inode->i_ino ||
+		    found_key.type > BTRFS_EXTENT_DATA_KEY ||
+		    found_key.offset > end)
+			break;
+
+		if (found_key.offset > cur_offset) {
+			extent_end = found_key.offset;
+			goto out_check;
+		}
+
+		fi = btrfs_item_ptr(leaf, path->slots[0],
+				    struct btrfs_file_extent_item);
+		extent_type = btrfs_file_extent_type(leaf, fi);
+
+		if (extent_type == BTRFS_FILE_EXTENT_REG ||
+		    extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
+			disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+			extent_end = found_key.offset +
+				btrfs_file_extent_num_bytes(leaf, fi);
+			if (extent_end <= start) {
+				path->slots[0]++;
+				goto next_slot;
+			}
+			if (disk_bytenr == 0)
+				goto out_check;
+			if (btrfs_file_extent_compression(leaf, fi) ||
+			    btrfs_file_extent_encryption(leaf, fi) ||
+			    btrfs_file_extent_other_encoding(leaf, fi))
+				goto out_check;
+			if (extent_type == BTRFS_FILE_EXTENT_REG && !force)
+				goto out_check;
+			if (btrfs_extent_readonly(root, disk_bytenr))
+				goto out_check;
+			if (btrfs_cross_ref_exist(trans, root, inode->i_ino,
+						  disk_bytenr))
+				goto out_check;
+			disk_bytenr += btrfs_file_extent_offset(leaf, fi);
+			disk_bytenr += cur_offset - found_key.offset;
+			num_bytes = min(end + 1, extent_end) - cur_offset;
+			/*
+			 * force cow if csum exists in the range.
+			 * this ensure that csum for a given extent are
+			 * either valid or do not exist.
+			 */
+			if (csum_exist_in_range(root, disk_bytenr, num_bytes))
+				goto out_check;
+			nocow = 1;
+		} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
+			extent_end = found_key.offset +
+				btrfs_file_extent_inline_len(leaf, fi);
+			extent_end = ALIGN(extent_end, root->sectorsize);
+		} else {
+			BUG_ON(1);
+		}
+out_check:
+		if (extent_end <= start) {
+			path->slots[0]++;
+			goto next_slot;
+		}
+		if (!nocow) {
+			if (cow_start == (u64)-1)
+				cow_start = cur_offset;
+			cur_offset = extent_end;
+			if (cur_offset > end)
+				break;
+			path->slots[0]++;
+			goto next_slot;
+		}
+
+		btrfs_release_path(root, path);
+		if (cow_start != (u64)-1) {
+			ret = cow_file_range(inode, locked_page, cow_start,
+					found_key.offset - 1, page_started,
+					nr_written, 1);
+			BUG_ON(ret);
+			cow_start = (u64)-1;
+		}
+
+		if (extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
+			struct extent_map *em;
+			struct extent_map_tree *em_tree;
+			em_tree = &BTRFS_I(inode)->extent_tree;
+			em = alloc_extent_map(GFP_NOFS);
+			em->start = cur_offset;
+			em->orig_start = em->start;
+			em->len = num_bytes;
+			em->block_len = num_bytes;
+			em->block_start = disk_bytenr;
+			em->bdev = root->fs_info->fs_devices->latest_bdev;
+			set_bit(EXTENT_FLAG_PINNED, &em->flags);
+			while (1) {
+				spin_lock(&em_tree->lock);
+				ret = add_extent_mapping(em_tree, em);
+				spin_unlock(&em_tree->lock);
+				if (ret != -EEXIST) {
+					free_extent_map(em);
+					break;
+				}
+				btrfs_drop_extent_cache(inode, em->start,
+						em->start + em->len - 1, 0);
+			}
+			type = BTRFS_ORDERED_PREALLOC;
+		} else {
+			type = BTRFS_ORDERED_NOCOW;
+		}
+
+		ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr,
+					       num_bytes, num_bytes, type);
+		BUG_ON(ret);
+
+		extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
+					cur_offset, cur_offset + num_bytes - 1,
+					locked_page, 1, 1, 1, 0, 0, 0);
+		cur_offset = extent_end;
+		if (cur_offset > end)
+			break;
+	}
+	btrfs_release_path(root, path);
+
+	if (cur_offset <= end && cow_start == (u64)-1)
+		cow_start = cur_offset;
+	if (cow_start != (u64)-1) {
+		ret = cow_file_range(inode, locked_page, cow_start, end,
+				     page_started, nr_written, 1);
+		BUG_ON(ret);
+	}
+
+	ret = btrfs_end_transaction(trans, root);
+	BUG_ON(ret);
+	btrfs_free_path(path);
+	return 0;
+}
+
+/*
+ * extent_io.c call back to do delayed allocation processing
+ */
+static int run_delalloc_range(struct inode *inode, struct page *locked_page,
+			      u64 start, u64 end, int *page_started,
+			      unsigned long *nr_written)
+{
+	int ret;
+
+	if (btrfs_test_flag(inode, NODATACOW))
+		ret = run_delalloc_nocow(inode, locked_page, start, end,
+					 page_started, 1, nr_written);
+	else if (btrfs_test_flag(inode, PREALLOC))
+		ret = run_delalloc_nocow(inode, locked_page, start, end,
+					 page_started, 0, nr_written);
+	else
+		ret = cow_file_range_async(inode, locked_page, start, end,
+					   page_started, nr_written);
+
+	return ret;
+}
+
+/*
+ * extent_io.c set_bit_hook, used to track delayed allocation
+ * bytes in this file, and to maintain the list of inodes that
+ * have pending delalloc work to be done.
+ */
+static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
+		       unsigned long old, unsigned long bits)
+{
+	/*
+	 * set_bit and clear bit hooks normally require _irqsave/restore
+	 * but in this case, we are only testeing for the DELALLOC
+	 * bit, which is only set or cleared with irqs on
+	 */
+	if (!(old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+		struct btrfs_root *root = BTRFS_I(inode)->root;
+		spin_lock(&root->fs_info->delalloc_lock);
+		BTRFS_I(inode)->delalloc_bytes += end - start + 1;
+		root->fs_info->delalloc_bytes += end - start + 1;
+		if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+			list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
+				      &root->fs_info->delalloc_inodes);
+		}
+		spin_unlock(&root->fs_info->delalloc_lock);
+	}
+	return 0;
+}
+
+/*
+ * extent_io.c clear_bit_hook, see set_bit_hook for why
+ */
+static int btrfs_clear_bit_hook(struct inode *inode, u64 start, u64 end,
+			 unsigned long old, unsigned long bits)
+{
+	/*
+	 * set_bit and clear bit hooks normally require _irqsave/restore
+	 * but in this case, we are only testeing for the DELALLOC
+	 * bit, which is only set or cleared with irqs on
+	 */
+	if ((old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+		struct btrfs_root *root = BTRFS_I(inode)->root;
+
+		spin_lock(&root->fs_info->delalloc_lock);
+		if (end - start + 1 > root->fs_info->delalloc_bytes) {
+			printk(KERN_INFO "btrfs warning: delalloc account "
+			       "%llu %llu\n",
+			       (unsigned long long)end - start + 1,
+			       (unsigned long long)
+			       root->fs_info->delalloc_bytes);
+			root->fs_info->delalloc_bytes = 0;
+			BTRFS_I(inode)->delalloc_bytes = 0;
+		} else {
+			root->fs_info->delalloc_bytes -= end - start + 1;
+			BTRFS_I(inode)->delalloc_bytes -= end - start + 1;
+		}
+		if (BTRFS_I(inode)->delalloc_bytes == 0 &&
+		    !list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+			list_del_init(&BTRFS_I(inode)->delalloc_inodes);
+		}
+		spin_unlock(&root->fs_info->delalloc_lock);
+	}
+	return 0;
+}
+
+/*
+ * extent_io.c merge_bio_hook, this must check the chunk tree to make sure
+ * we don't create bios that span stripes or chunks
+ */
+int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
+			 size_t size, struct bio *bio,
+			 unsigned long bio_flags)
+{
+	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
+	struct btrfs_mapping_tree *map_tree;
+	u64 logical = (u64)bio->bi_sector << 9;
+	u64 length = 0;
+	u64 map_length;
+	int ret;
+
+	if (bio_flags & EXTENT_BIO_COMPRESSED)
+		return 0;
+
+	length = bio->bi_size;
+	map_tree = &root->fs_info->mapping_tree;
+	map_length = length;
+	ret = btrfs_map_block(map_tree, READ, logical,
+			      &map_length, NULL, 0);
+
+	if (map_length < length + size)
+		return 1;
+	return 0;
+}
+
+/*
+ * in order to insert checksums into the metadata in large chunks,
+ * we wait until bio submission time.   All the pages in the bio are
+ * checksummed and sums are attached onto the ordered extent record.
+ *
+ * At IO completion time the cums attached on the ordered extent record
+ * are inserted into the btree
+ */
+static int __btrfs_submit_bio_start(struct inode *inode, int rw,
+				    struct bio *bio, int mirror_num,
+				    unsigned long bio_flags)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret = 0;
+
+	ret = btrfs_csum_one_bio(root, inode, bio, 0, 0);
+	BUG_ON(ret);
+	return 0;
+}
+
+/*
+ * in order to insert checksums into the metadata in large chunks,
+ * we wait until bio submission time.   All the pages in the bio are
+ * checksummed and sums are attached onto the ordered extent record.
+ *
+ * At IO completion time the cums attached on the ordered extent record
+ * are inserted into the btree
+ */
+static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
+			  int mirror_num, unsigned long bio_flags)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	return btrfs_map_bio(root, rw, bio, mirror_num, 1);
+}
+
+/*
+ * extent_io.c submission hook. This does the right thing for csum calculation
+ * on write, or reading the csums from the tree before a read
+ */
+static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
+			  int mirror_num, unsigned long bio_flags)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret = 0;
+	int skip_sum;
+
+	skip_sum = btrfs_test_flag(inode, NODATASUM);
+
+	ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
+	BUG_ON(ret);
+
+	if (!(rw & (1 << BIO_RW))) {
+		if (bio_flags & EXTENT_BIO_COMPRESSED) {
+			return btrfs_submit_compressed_read(inode, bio,
+						    mirror_num, bio_flags);
+		} else if (!skip_sum)
+			btrfs_lookup_bio_sums(root, inode, bio, NULL);
+		goto mapit;
+	} else if (!skip_sum) {
+		/* csum items have already been cloned */
+		if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID)
+			goto mapit;
+		/* we're doing a write, do the async checksumming */
+		return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
+				   inode, rw, bio, mirror_num,
+				   bio_flags, __btrfs_submit_bio_start,
+				   __btrfs_submit_bio_done);
+	}
+
+mapit:
+	return btrfs_map_bio(root, rw, bio, mirror_num, 0);
+}
+
+/*
+ * given a list of ordered sums record them in the inode.  This happens
+ * at IO completion time based on sums calculated at bio submission time.
+ */
+static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
+			     struct inode *inode, u64 file_offset,
+			     struct list_head *list)
+{
+	struct list_head *cur;
+	struct btrfs_ordered_sum *sum;
+
+	btrfs_set_trans_block_group(trans, inode);
+	list_for_each(cur, list) {
+		sum = list_entry(cur, struct btrfs_ordered_sum, list);
+		btrfs_csum_file_blocks(trans,
+		       BTRFS_I(inode)->root->fs_info->csum_root, sum);
+	}
+	return 0;
+}
+
+int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end)
+{
+	if ((end & (PAGE_CACHE_SIZE - 1)) == 0)
+		WARN_ON(1);
+	return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
+				   GFP_NOFS);
+}
+
+/* see btrfs_writepage_start_hook for details on why this is required */
+struct btrfs_writepage_fixup {
+	struct page *page;
+	struct btrfs_work work;
+};
+
+static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
+{
+	struct btrfs_writepage_fixup *fixup;
+	struct btrfs_ordered_extent *ordered;
+	struct page *page;
+	struct inode *inode;
+	u64 page_start;
+	u64 page_end;
+
+	fixup = container_of(work, struct btrfs_writepage_fixup, work);
+	page = fixup->page;
+again:
+	lock_page(page);
+	if (!page->mapping || !PageDirty(page) || !PageChecked(page)) {
+		ClearPageChecked(page);
+		goto out_page;
+	}
+
+	inode = page->mapping->host;
+	page_start = page_offset(page);
+	page_end = page_offset(page) + PAGE_CACHE_SIZE - 1;
+
+	lock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+
+	/* already ordered? We're done */
+	if (test_range_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
+			     EXTENT_ORDERED, 0)) {
+		goto out;
+	}
+
+	ordered = btrfs_lookup_ordered_extent(inode, page_start);
+	if (ordered) {
+		unlock_extent(&BTRFS_I(inode)->io_tree, page_start,
+			      page_end, GFP_NOFS);
+		unlock_page(page);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		goto again;
+	}
+
+	btrfs_set_extent_delalloc(inode, page_start, page_end);
+	ClearPageChecked(page);
+out:
+	unlock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+out_page:
+	unlock_page(page);
+	page_cache_release(page);
+}
+
+/*
+ * There are a few paths in the higher layers of the kernel that directly
+ * set the page dirty bit without asking the filesystem if it is a
+ * good idea.  This causes problems because we want to make sure COW
+ * properly happens and the data=ordered rules are followed.
+ *
+ * In our case any range that doesn't have the ORDERED bit set
+ * hasn't been properly setup for IO.  We kick off an async process
+ * to fix it up.  The async helper will wait for ordered extents, set
+ * the delalloc bit and make it safe to write the page.
+ */
+static int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end)
+{
+	struct inode *inode = page->mapping->host;
+	struct btrfs_writepage_fixup *fixup;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret;
+
+	ret = test_range_bit(&BTRFS_I(inode)->io_tree, start, end,
+			     EXTENT_ORDERED, 0);
+	if (ret)
+		return 0;
+
+	if (PageChecked(page))
+		return -EAGAIN;
+
+	fixup = kzalloc(sizeof(*fixup), GFP_NOFS);
+	if (!fixup)
+		return -EAGAIN;
+
+	SetPageChecked(page);
+	page_cache_get(page);
+	fixup->work.func = btrfs_writepage_fixup_worker;
+	fixup->page = page;
+	btrfs_queue_worker(&root->fs_info->fixup_workers, &fixup->work);
+	return -EAGAIN;
+}
+
+static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
+				       struct inode *inode, u64 file_pos,
+				       u64 disk_bytenr, u64 disk_num_bytes,
+				       u64 num_bytes, u64 ram_bytes,
+				       u8 compression, u8 encryption,
+				       u16 other_encoding, int extent_type)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_key ins;
+	u64 hint;
+	int ret;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	ret = btrfs_drop_extents(trans, root, inode, file_pos,
+				 file_pos + num_bytes, file_pos, &hint);
+	BUG_ON(ret);
+
+	ins.objectid = inode->i_ino;
+	ins.offset = file_pos;
+	ins.type = BTRFS_EXTENT_DATA_KEY;
+	ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi));
+	BUG_ON(ret);
+	leaf = path->nodes[0];
+	fi = btrfs_item_ptr(leaf, path->slots[0],
+			    struct btrfs_file_extent_item);
+	btrfs_set_file_extent_generation(leaf, fi, trans->transid);
+	btrfs_set_file_extent_type(leaf, fi, extent_type);
+	btrfs_set_file_extent_disk_bytenr(leaf, fi, disk_bytenr);
+	btrfs_set_file_extent_disk_num_bytes(leaf, fi, disk_num_bytes);
+	btrfs_set_file_extent_offset(leaf, fi, 0);
+	btrfs_set_file_extent_num_bytes(leaf, fi, num_bytes);
+	btrfs_set_file_extent_ram_bytes(leaf, fi, ram_bytes);
+	btrfs_set_file_extent_compression(leaf, fi, compression);
+	btrfs_set_file_extent_encryption(leaf, fi, encryption);
+	btrfs_set_file_extent_other_encoding(leaf, fi, other_encoding);
+	btrfs_mark_buffer_dirty(leaf);
+
+	inode_add_bytes(inode, num_bytes);
+	btrfs_drop_extent_cache(inode, file_pos, file_pos + num_bytes - 1, 0);
+
+	ins.objectid = disk_bytenr;
+	ins.offset = disk_num_bytes;
+	ins.type = BTRFS_EXTENT_ITEM_KEY;
+	ret = btrfs_alloc_reserved_extent(trans, root, leaf->start,
+					  root->root_key.objectid,
+					  trans->transid, inode->i_ino, &ins);
+	BUG_ON(ret);
+
+	btrfs_free_path(path);
+	return 0;
+}
+
+/* as ordered data IO finishes, this gets called so we can finish
+ * an ordered extent if the range of bytes in the file it covers are
+ * fully written.
+ */
+static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_ordered_extent *ordered_extent;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	int compressed = 0;
+	int ret;
+
+	ret = btrfs_dec_test_ordered_pending(inode, start, end - start + 1);
+	if (!ret)
+		return 0;
+
+	trans = btrfs_join_transaction(root, 1);
+
+	ordered_extent = btrfs_lookup_ordered_extent(inode, start);
+	BUG_ON(!ordered_extent);
+	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags))
+		goto nocow;
+
+	lock_extent(io_tree, ordered_extent->file_offset,
+		    ordered_extent->file_offset + ordered_extent->len - 1,
+		    GFP_NOFS);
+
+	if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
+		compressed = 1;
+	if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) {
+		BUG_ON(compressed);
+		ret = btrfs_mark_extent_written(trans, root, inode,
+						ordered_extent->file_offset,
+						ordered_extent->file_offset +
+						ordered_extent->len);
+		BUG_ON(ret);
+	} else {
+		ret = insert_reserved_file_extent(trans, inode,
+						ordered_extent->file_offset,
+						ordered_extent->start,
+						ordered_extent->disk_len,
+						ordered_extent->len,
+						ordered_extent->len,
+						compressed, 0, 0,
+						BTRFS_FILE_EXTENT_REG);
+		BUG_ON(ret);
+	}
+	unlock_extent(io_tree, ordered_extent->file_offset,
+		    ordered_extent->file_offset + ordered_extent->len - 1,
+		    GFP_NOFS);
+nocow:
+	add_pending_csums(trans, inode, ordered_extent->file_offset,
+			  &ordered_extent->list);
+
+	mutex_lock(&BTRFS_I(inode)->extent_mutex);
+	btrfs_ordered_update_i_size(inode, ordered_extent);
+	btrfs_update_inode(trans, root, inode);
+	btrfs_remove_ordered_extent(inode, ordered_extent);
+	mutex_unlock(&BTRFS_I(inode)->extent_mutex);
+
+	/* once for us */
+	btrfs_put_ordered_extent(ordered_extent);
+	/* once for the tree */
+	btrfs_put_ordered_extent(ordered_extent);
+
+	btrfs_end_transaction(trans, root);
+	return 0;
+}
+
+static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
+				struct extent_state *state, int uptodate)
+{
+	return btrfs_finish_ordered_io(page->mapping->host, start, end);
+}
+
+/*
+ * When IO fails, either with EIO or csum verification fails, we
+ * try other mirrors that might have a good copy of the data.  This
+ * io_failure_record is used to record state as we go through all the
+ * mirrors.  If another mirror has good data, the page is set up to date
+ * and things continue.  If a good mirror can't be found, the original
+ * bio end_io callback is called to indicate things have failed.
+ */
+struct io_failure_record {
+	struct page *page;
+	u64 start;
+	u64 len;
+	u64 logical;
+	unsigned long bio_flags;
+	int last_mirror;
+};
+
+static int btrfs_io_failed_hook(struct bio *failed_bio,
+			 struct page *page, u64 start, u64 end,
+			 struct extent_state *state)
+{
+	struct io_failure_record *failrec = NULL;
+	u64 private;
+	struct extent_map *em;
+	struct inode *inode = page->mapping->host;
+	struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct bio *bio;
+	int num_copies;
+	int ret;
+	int rw;
+	u64 logical;
+
+	ret = get_state_private(failure_tree, start, &private);
+	if (ret) {
+		failrec = kmalloc(sizeof(*failrec), GFP_NOFS);
+		if (!failrec)
+			return -ENOMEM;
+		failrec->start = start;
+		failrec->len = end - start + 1;
+		failrec->last_mirror = 0;
+		failrec->bio_flags = 0;
+
+		spin_lock(&em_tree->lock);
+		em = lookup_extent_mapping(em_tree, start, failrec->len);
+		if (em->start > start || em->start + em->len < start) {
+			free_extent_map(em);
+			em = NULL;
+		}
+		spin_unlock(&em_tree->lock);
+
+		if (!em || IS_ERR(em)) {
+			kfree(failrec);
+			return -EIO;
+		}
+		logical = start - em->start;
+		logical = em->block_start + logical;
+		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
+			logical = em->block_start;
+			failrec->bio_flags = EXTENT_BIO_COMPRESSED;
+		}
+		failrec->logical = logical;
+		free_extent_map(em);
+		set_extent_bits(failure_tree, start, end, EXTENT_LOCKED |
+				EXTENT_DIRTY, GFP_NOFS);
+		set_state_private(failure_tree, start,
+				 (u64)(unsigned long)failrec);
+	} else {
+		failrec = (struct io_failure_record *)(unsigned long)private;
+	}
+	num_copies = btrfs_num_copies(
+			      &BTRFS_I(inode)->root->fs_info->mapping_tree,
+			      failrec->logical, failrec->len);
+	failrec->last_mirror++;
+	if (!state) {
+		spin_lock(&BTRFS_I(inode)->io_tree.lock);
+		state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree,
+						    failrec->start,
+						    EXTENT_LOCKED);
+		if (state && state->start != failrec->start)
+			state = NULL;
+		spin_unlock(&BTRFS_I(inode)->io_tree.lock);
+	}
+	if (!state || failrec->last_mirror > num_copies) {
+		set_state_private(failure_tree, failrec->start, 0);
+		clear_extent_bits(failure_tree, failrec->start,
+				  failrec->start + failrec->len - 1,
+				  EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS);
+		kfree(failrec);
+		return -EIO;
+	}
+	bio = bio_alloc(GFP_NOFS, 1);
+	bio->bi_private = state;
+	bio->bi_end_io = failed_bio->bi_end_io;
+	bio->bi_sector = failrec->logical >> 9;
+	bio->bi_bdev = failed_bio->bi_bdev;
+	bio->bi_size = 0;
+
+	bio_add_page(bio, page, failrec->len, start - page_offset(page));
+	if (failed_bio->bi_rw & (1 << BIO_RW))
+		rw = WRITE;
+	else
+		rw = READ;
+
+	BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio,
+						      failrec->last_mirror,
+						      failrec->bio_flags);
+	return 0;
+}
+
+/*
+ * each time an IO finishes, we do a fast check in the IO failure tree
+ * to see if we need to process or clean up an io_failure_record
+ */
+static int btrfs_clean_io_failures(struct inode *inode, u64 start)
+{
+	u64 private;
+	u64 private_failure;
+	struct io_failure_record *failure;
+	int ret;
+
+	private = 0;
+	if (count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private,
+			     (u64)-1, 1, EXTENT_DIRTY)) {
+		ret = get_state_private(&BTRFS_I(inode)->io_failure_tree,
+					start, &private_failure);
+		if (ret == 0) {
+			failure = (struct io_failure_record *)(unsigned long)
+				   private_failure;
+			set_state_private(&BTRFS_I(inode)->io_failure_tree,
+					  failure->start, 0);
+			clear_extent_bits(&BTRFS_I(inode)->io_failure_tree,
+					  failure->start,
+					  failure->start + failure->len - 1,
+					  EXTENT_DIRTY | EXTENT_LOCKED,
+					  GFP_NOFS);
+			kfree(failure);
+		}
+	}
+	return 0;
+}
+
+/*
+ * when reads are done, we need to check csums to verify the data is correct
+ * if there's a match, we allow the bio to finish.  If not, we go through
+ * the io_failure_record routines to find good copies
+ */
+static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end,
+			       struct extent_state *state)
+{
+	size_t offset = start - ((u64)page->index << PAGE_CACHE_SHIFT);
+	struct inode *inode = page->mapping->host;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	char *kaddr;
+	u64 private = ~(u32)0;
+	int ret;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	u32 csum = ~(u32)0;
+
+	if (PageChecked(page)) {
+		ClearPageChecked(page);
+		goto good;
+	}
+	if (btrfs_test_flag(inode, NODATASUM))
+		return 0;
+
+	if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID &&
+	    test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1)) {
+		clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM,
+				  GFP_NOFS);
+		return 0;
+	}
+
+	if (state && state->start == start) {
+		private = state->private;
+		ret = 0;
+	} else {
+		ret = get_state_private(io_tree, start, &private);
+	}
+	kaddr = kmap_atomic(page, KM_USER0);
+	if (ret)
+		goto zeroit;
+
+	csum = btrfs_csum_data(root, kaddr + offset, csum,  end - start + 1);
+	btrfs_csum_final(csum, (char *)&csum);
+	if (csum != private)
+		goto zeroit;
+
+	kunmap_atomic(kaddr, KM_USER0);
+good:
+	/* if the io failure tree for this inode is non-empty,
+	 * check to see if we've recovered from a failed IO
+	 */
+	btrfs_clean_io_failures(inode, start);
+	return 0;
+
+zeroit:
+	printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
+	       "private %llu\n", page->mapping->host->i_ino,
+	       (unsigned long long)start, csum,
+	       (unsigned long long)private);
+	memset(kaddr + offset, 1, end - start + 1);
+	flush_dcache_page(page);
+	kunmap_atomic(kaddr, KM_USER0);
+	if (private == 0)
+		return 0;
+	return -EIO;
+}
+
+/*
+ * This creates an orphan entry for the given inode in case something goes
+ * wrong in the middle of an unlink/truncate.
+ */
+int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret = 0;
+
+	spin_lock(&root->list_lock);
+
+	/* already on the orphan list, we're good */
+	if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
+		spin_unlock(&root->list_lock);
+		return 0;
+	}
+
+	list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+
+	spin_unlock(&root->list_lock);
+
+	/*
+	 * insert an orphan item to track this unlinked/truncated file
+	 */
+	ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+
+	return ret;
+}
+
+/*
+ * We have done the truncate/delete so we can go ahead and remove the orphan
+ * item for this particular inode.
+ */
+int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret = 0;
+
+	spin_lock(&root->list_lock);
+
+	if (list_empty(&BTRFS_I(inode)->i_orphan)) {
+		spin_unlock(&root->list_lock);
+		return 0;
+	}
+
+	list_del_init(&BTRFS_I(inode)->i_orphan);
+	if (!trans) {
+		spin_unlock(&root->list_lock);
+		return 0;
+	}
+
+	spin_unlock(&root->list_lock);
+
+	ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+
+	return ret;
+}
+
+/*
+ * this cleans up any orphans that may be left on the list from the last use
+ * of this root.
+ */
+void btrfs_orphan_cleanup(struct btrfs_root *root)
+{
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	struct btrfs_key key, found_key;
+	struct btrfs_trans_handle *trans;
+	struct inode *inode;
+	int ret = 0, nr_unlink = 0, nr_truncate = 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return;
+	path->reada = -1;
+
+	key.objectid = BTRFS_ORPHAN_OBJECTID;
+	btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
+	key.offset = (u64)-1;
+
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0) {
+			printk(KERN_ERR "Error searching slot for orphan: %d"
+			       "\n", ret);
+			break;
+		}
+
+		/*
+		 * if ret == 0 means we found what we were searching for, which
+		 * is weird, but possible, so only screw with path if we didnt
+		 * find the key and see if we have stuff that matches
+		 */
+		if (ret > 0) {
+			if (path->slots[0] == 0)
+				break;
+			path->slots[0]--;
+		}
+
+		/* pull out the item */
+		leaf = path->nodes[0];
+		item = btrfs_item_nr(leaf, path->slots[0]);
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+		/* make sure the item matches what we want */
+		if (found_key.objectid != BTRFS_ORPHAN_OBJECTID)
+			break;
+		if (btrfs_key_type(&found_key) != BTRFS_ORPHAN_ITEM_KEY)
+			break;
+
+		/* release the path since we're done with it */
+		btrfs_release_path(root, path);
+
+		/*
+		 * this is where we are basically btrfs_lookup, without the
+		 * crossing root thing.  we store the inode number in the
+		 * offset of the orphan item.
+		 */
+		inode = btrfs_iget_locked(root->fs_info->sb,
+					  found_key.offset, root);
+		if (!inode)
+			break;
+
+		if (inode->i_state & I_NEW) {
+			BTRFS_I(inode)->root = root;
+
+			/* have to set the location manually */
+			BTRFS_I(inode)->location.objectid = inode->i_ino;
+			BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
+			BTRFS_I(inode)->location.offset = 0;
+
+			btrfs_read_locked_inode(inode);
+			unlock_new_inode(inode);
+		}
+
+		/*
+		 * add this inode to the orphan list so btrfs_orphan_del does
+		 * the proper thing when we hit it
+		 */
+		spin_lock(&root->list_lock);
+		list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+		spin_unlock(&root->list_lock);
+
+		/*
+		 * if this is a bad inode, means we actually succeeded in
+		 * removing the inode, but not the orphan record, which means
+		 * we need to manually delete the orphan since iput will just
+		 * do a destroy_inode
+		 */
+		if (is_bad_inode(inode)) {
+			trans = btrfs_start_transaction(root, 1);
+			btrfs_orphan_del(trans, inode);
+			btrfs_end_transaction(trans, root);
+			iput(inode);
+			continue;
+		}
+
+		/* if we have links, this was a truncate, lets do that */
+		if (inode->i_nlink) {
+			nr_truncate++;
+			btrfs_truncate(inode);
+		} else {
+			nr_unlink++;
+		}
+
+		/* this will do delete_inode and everything for us */
+		iput(inode);
+	}
+
+	if (nr_unlink)
+		printk(KERN_INFO "btrfs: unlinked %d orphans\n", nr_unlink);
+	if (nr_truncate)
+		printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate);
+
+	btrfs_free_path(path);
+}
+
+/*
+ * read an inode from the btree into the in-memory inode
+ */
+void btrfs_read_locked_inode(struct inode *inode)
+{
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_inode_item *inode_item;
+	struct btrfs_timespec *tspec;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_key location;
+	u64 alloc_group_block;
+	u32 rdev;
+	int ret;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	memcpy(&location, &BTRFS_I(inode)->location, sizeof(location));
+
+	ret = btrfs_lookup_inode(NULL, root, path, &location, 0);
+	if (ret)
+		goto make_bad;
+
+	leaf = path->nodes[0];
+	inode_item = btrfs_item_ptr(leaf, path->slots[0],
+				    struct btrfs_inode_item);
+
+	inode->i_mode = btrfs_inode_mode(leaf, inode_item);
+	inode->i_nlink = btrfs_inode_nlink(leaf, inode_item);
+	inode->i_uid = btrfs_inode_uid(leaf, inode_item);
+	inode->i_gid = btrfs_inode_gid(leaf, inode_item);
+	btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item));
+
+	tspec = btrfs_inode_atime(inode_item);
+	inode->i_atime.tv_sec = btrfs_timespec_sec(leaf, tspec);
+	inode->i_atime.tv_nsec = btrfs_timespec_nsec(leaf, tspec);
+
+	tspec = btrfs_inode_mtime(inode_item);
+	inode->i_mtime.tv_sec = btrfs_timespec_sec(leaf, tspec);
+	inode->i_mtime.tv_nsec = btrfs_timespec_nsec(leaf, tspec);
+
+	tspec = btrfs_inode_ctime(inode_item);
+	inode->i_ctime.tv_sec = btrfs_timespec_sec(leaf, tspec);
+	inode->i_ctime.tv_nsec = btrfs_timespec_nsec(leaf, tspec);
+
+	inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item));
+	BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
+	BTRFS_I(inode)->sequence = btrfs_inode_sequence(leaf, inode_item);
+	inode->i_generation = BTRFS_I(inode)->generation;
+	inode->i_rdev = 0;
+	rdev = btrfs_inode_rdev(leaf, inode_item);
+
+	BTRFS_I(inode)->index_cnt = (u64)-1;
+	BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item);
+
+	alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
+	BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0,
+						alloc_group_block, 0);
+	btrfs_free_path(path);
+	inode_item = NULL;
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFREG:
+		inode->i_mapping->a_ops = &btrfs_aops;
+		inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
+		BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
+		inode->i_fop = &btrfs_file_operations;
+		inode->i_op = &btrfs_file_inode_operations;
+		break;
+	case S_IFDIR:
+		inode->i_fop = &btrfs_dir_file_operations;
+		if (root == root->fs_info->tree_root)
+			inode->i_op = &btrfs_dir_ro_inode_operations;
+		else
+			inode->i_op = &btrfs_dir_inode_operations;
+		break;
+	case S_IFLNK:
+		inode->i_op = &btrfs_symlink_inode_operations;
+		inode->i_mapping->a_ops = &btrfs_symlink_aops;
+		inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
+		break;
+	default:
+		init_special_inode(inode, inode->i_mode, rdev);
+		break;
+	}
+	return;
+
+make_bad:
+	btrfs_free_path(path);
+	make_bad_inode(inode);
+}
+
+/*
+ * given a leaf and an inode, copy the inode fields into the leaf
+ */
+static void fill_inode_item(struct btrfs_trans_handle *trans,
+			    struct extent_buffer *leaf,
+			    struct btrfs_inode_item *item,
+			    struct inode *inode)
+{
+	btrfs_set_inode_uid(leaf, item, inode->i_uid);
+	btrfs_set_inode_gid(leaf, item, inode->i_gid);
+	btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size);
+	btrfs_set_inode_mode(leaf, item, inode->i_mode);
+	btrfs_set_inode_nlink(leaf, item, inode->i_nlink);
+
+	btrfs_set_timespec_sec(leaf, btrfs_inode_atime(item),
+			       inode->i_atime.tv_sec);
+	btrfs_set_timespec_nsec(leaf, btrfs_inode_atime(item),
+				inode->i_atime.tv_nsec);
+
+	btrfs_set_timespec_sec(leaf, btrfs_inode_mtime(item),
+			       inode->i_mtime.tv_sec);
+	btrfs_set_timespec_nsec(leaf, btrfs_inode_mtime(item),
+				inode->i_mtime.tv_nsec);
+
+	btrfs_set_timespec_sec(leaf, btrfs_inode_ctime(item),
+			       inode->i_ctime.tv_sec);
+	btrfs_set_timespec_nsec(leaf, btrfs_inode_ctime(item),
+				inode->i_ctime.tv_nsec);
+
+	btrfs_set_inode_nbytes(leaf, item, inode_get_bytes(inode));
+	btrfs_set_inode_generation(leaf, item, BTRFS_I(inode)->generation);
+	btrfs_set_inode_sequence(leaf, item, BTRFS_I(inode)->sequence);
+	btrfs_set_inode_transid(leaf, item, trans->transid);
+	btrfs_set_inode_rdev(leaf, item, inode->i_rdev);
+	btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags);
+	btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group);
+}
+
+/*
+ * copy everything in the in-memory inode into the btree.
+ */
+noinline int btrfs_update_inode(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root, struct inode *inode)
+{
+	struct btrfs_inode_item *inode_item;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	int ret;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	ret = btrfs_lookup_inode(trans, root, path,
+				 &BTRFS_I(inode)->location, 1);
+	if (ret) {
+		if (ret > 0)
+			ret = -ENOENT;
+		goto failed;
+	}
+
+	leaf = path->nodes[0];
+	inode_item = btrfs_item_ptr(leaf, path->slots[0],
+				  struct btrfs_inode_item);
+
+	fill_inode_item(trans, leaf, inode_item, inode);
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_set_inode_last_trans(trans, inode);
+	ret = 0;
+failed:
+	btrfs_free_path(path);
+	return ret;
+}
+
+
+/*
+ * unlink helper that gets used here in inode.c and in the tree logging
+ * recovery code.  It remove a link in a directory with a given name, and
+ * also drops the back refs in the inode to the directory
+ */
+int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *root,
+		       struct inode *dir, struct inode *inode,
+		       const char *name, int name_len)
+{
+	struct btrfs_path *path;
+	int ret = 0;
+	struct extent_buffer *leaf;
+	struct btrfs_dir_item *di;
+	struct btrfs_key key;
+	u64 index;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+				    name, name_len, -1);
+	if (IS_ERR(di)) {
+		ret = PTR_ERR(di);
+		goto err;
+	}
+	if (!di) {
+		ret = -ENOENT;
+		goto err;
+	}
+	leaf = path->nodes[0];
+	btrfs_dir_item_key_to_cpu(leaf, di, &key);
+	ret = btrfs_delete_one_dir_name(trans, root, path, di);
+	if (ret)
+		goto err;
+	btrfs_release_path(root, path);
+
+	ret = btrfs_del_inode_ref(trans, root, name, name_len,
+				  inode->i_ino,
+				  dir->i_ino, &index);
+	if (ret) {
+		printk(KERN_INFO "btrfs failed to delete reference to %.*s, "
+		       "inode %lu parent %lu\n", name_len, name,
+		       inode->i_ino, dir->i_ino);
+		goto err;
+	}
+
+	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+					 index, name, name_len, -1);
+	if (IS_ERR(di)) {
+		ret = PTR_ERR(di);
+		goto err;
+	}
+	if (!di) {
+		ret = -ENOENT;
+		goto err;
+	}
+	ret = btrfs_delete_one_dir_name(trans, root, path, di);
+	btrfs_release_path(root, path);
+
+	ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len,
+					 inode, dir->i_ino);
+	BUG_ON(ret != 0 && ret != -ENOENT);
+	if (ret != -ENOENT)
+		BTRFS_I(dir)->log_dirty_trans = trans->transid;
+
+	ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len,
+					   dir, index);
+	BUG_ON(ret);
+err:
+	btrfs_free_path(path);
+	if (ret)
+		goto out;
+
+	btrfs_i_size_write(dir, dir->i_size - name_len * 2);
+	inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME;
+	btrfs_update_inode(trans, root, dir);
+	btrfs_drop_nlink(inode);
+	ret = btrfs_update_inode(trans, root, inode);
+	dir->i_sb->s_dirt = 1;
+out:
+	return ret;
+}
+
+static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
+{
+	struct btrfs_root *root;
+	struct btrfs_trans_handle *trans;
+	struct inode *inode = dentry->d_inode;
+	int ret;
+	unsigned long nr = 0;
+
+	root = BTRFS_I(dir)->root;
+
+	ret = btrfs_check_free_space(root, 1, 1);
+	if (ret)
+		goto fail;
+
+	trans = btrfs_start_transaction(root, 1);
+
+	btrfs_set_trans_block_group(trans, dir);
+	ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
+				 dentry->d_name.name, dentry->d_name.len);
+
+	if (inode->i_nlink == 0)
+		ret = btrfs_orphan_add(trans, inode);
+
+	nr = trans->blocks_used;
+
+	btrfs_end_transaction_throttle(trans, root);
+fail:
+	btrfs_btree_balance_dirty(root, nr);
+	return ret;
+}
+
+static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
+{
+	struct inode *inode = dentry->d_inode;
+	int err = 0;
+	int ret;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_trans_handle *trans;
+	unsigned long nr = 0;
+
+	/*
+	 * the FIRST_FREE_OBJECTID check makes sure we don't try to rmdir
+	 * the root of a subvolume or snapshot
+	 */
+	if (inode->i_size > BTRFS_EMPTY_DIR_SIZE ||
+	    inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+		return -ENOTEMPTY;
+	}
+
+	ret = btrfs_check_free_space(root, 1, 1);
+	if (ret)
+		goto fail;
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, dir);
+
+	err = btrfs_orphan_add(trans, inode);
+	if (err)
+		goto fail_trans;
+
+	/* now the directory is empty */
+	err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
+				 dentry->d_name.name, dentry->d_name.len);
+	if (!err)
+		btrfs_i_size_write(inode, 0);
+
+fail_trans:
+	nr = trans->blocks_used;
+	ret = btrfs_end_transaction_throttle(trans, root);
+fail:
+	btrfs_btree_balance_dirty(root, nr);
+
+	if (ret && !err)
+		err = ret;
+	return err;
+}
+
+#if 0
+/*
+ * when truncating bytes in a file, it is possible to avoid reading
+ * the leaves that contain only checksum items.  This can be the
+ * majority of the IO required to delete a large file, but it must
+ * be done carefully.
+ *
+ * The keys in the level just above the leaves are checked to make sure
+ * the lowest key in a given leaf is a csum key, and starts at an offset
+ * after the new  size.
+ *
+ * Then the key for the next leaf is checked to make sure it also has
+ * a checksum item for the same file.  If it does, we know our target leaf
+ * contains only checksum items, and it can be safely freed without reading
+ * it.
+ *
+ * This is just an optimization targeted at large files.  It may do
+ * nothing.  It will return 0 unless things went badly.
+ */
+static noinline int drop_csum_leaves(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root,
+				     struct btrfs_path *path,
+				     struct inode *inode, u64 new_size)
+{
+	struct btrfs_key key;
+	int ret;
+	int nritems;
+	struct btrfs_key found_key;
+	struct btrfs_key other_key;
+	struct btrfs_leaf_ref *ref;
+	u64 leaf_gen;
+	u64 leaf_start;
+
+	path->lowest_level = 1;
+	key.objectid = inode->i_ino;
+	key.type = BTRFS_CSUM_ITEM_KEY;
+	key.offset = new_size;
+again:
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret < 0)
+		goto out;
+
+	if (path->nodes[1] == NULL) {
+		ret = 0;
+		goto out;
+	}
+	ret = 0;
+	btrfs_node_key_to_cpu(path->nodes[1], &found_key, path->slots[1]);
+	nritems = btrfs_header_nritems(path->nodes[1]);
+
+	if (!nritems)
+		goto out;
+
+	if (path->slots[1] >= nritems)
+		goto next_node;
+
+	/* did we find a key greater than anything we want to delete? */
+	if (found_key.objectid > inode->i_ino ||
+	   (found_key.objectid == inode->i_ino && found_key.type > key.type))
+		goto out;
+
+	/* we check the next key in the node to make sure the leave contains
+	 * only checksum items.  This comparison doesn't work if our
+	 * leaf is the last one in the node
+	 */
+	if (path->slots[1] + 1 >= nritems) {
+next_node:
+		/* search forward from the last key in the node, this
+		 * will bring us into the next node in the tree
+		 */
+		btrfs_node_key_to_cpu(path->nodes[1], &found_key, nritems - 1);
+
+		/* unlikely, but we inc below, so check to be safe */
+		if (found_key.offset == (u64)-1)
+			goto out;
+
+		/* search_forward needs a path with locks held, do the
+		 * search again for the original key.  It is possible
+		 * this will race with a balance and return a path that
+		 * we could modify, but this drop is just an optimization
+		 * and is allowed to miss some leaves.
+		 */
+		btrfs_release_path(root, path);
+		found_key.offset++;
+
+		/* setup a max key for search_forward */
+		other_key.offset = (u64)-1;
+		other_key.type = key.type;
+		other_key.objectid = key.objectid;
+
+		path->keep_locks = 1;
+		ret = btrfs_search_forward(root, &found_key, &other_key,
+					   path, 0, 0);
+		path->keep_locks = 0;
+		if (ret || found_key.objectid != key.objectid ||
+		    found_key.type != key.type) {
+			ret = 0;
+			goto out;
+		}
+
+		key.offset = found_key.offset;
+		btrfs_release_path(root, path);
+		cond_resched();
+		goto again;
+	}
+
+	/* we know there's one more slot after us in the tree,
+	 * read that key so we can verify it is also a checksum item
+	 */
+	btrfs_node_key_to_cpu(path->nodes[1], &other_key, path->slots[1] + 1);
+
+	if (found_key.objectid < inode->i_ino)
+		goto next_key;
+
+	if (found_key.type != key.type || found_key.offset < new_size)
+		goto next_key;
+
+	/*
+	 * if the key for the next leaf isn't a csum key from this objectid,
+	 * we can't be sure there aren't good items inside this leaf.
+	 * Bail out
+	 */
+	if (other_key.objectid != inode->i_ino || other_key.type != key.type)
+		goto out;
+
+	leaf_start = btrfs_node_blockptr(path->nodes[1], path->slots[1]);
+	leaf_gen = btrfs_node_ptr_generation(path->nodes[1], path->slots[1]);
+	/*
+	 * it is safe to delete this leaf, it contains only
+	 * csum items from this inode at an offset >= new_size
+	 */
+	ret = btrfs_del_leaf(trans, root, path, leaf_start);
+	BUG_ON(ret);
+
+	if (root->ref_cows && leaf_gen < trans->transid) {
+		ref = btrfs_alloc_leaf_ref(root, 0);
+		if (ref) {
+			ref->root_gen = root->root_key.offset;
+			ref->bytenr = leaf_start;
+			ref->owner = 0;
+			ref->generation = leaf_gen;
+			ref->nritems = 0;
+
+			ret = btrfs_add_leaf_ref(root, ref, 0);
+			WARN_ON(ret);
+			btrfs_free_leaf_ref(root, ref);
+		} else {
+			WARN_ON(1);
+		}
+	}
+next_key:
+	btrfs_release_path(root, path);
+
+	if (other_key.objectid == inode->i_ino &&
+	    other_key.type == key.type && other_key.offset > key.offset) {
+		key.offset = other_key.offset;
+		cond_resched();
+		goto again;
+	}
+	ret = 0;
+out:
+	/* fixup any changes we've made to the path */
+	path->lowest_level = 0;
+	path->keep_locks = 0;
+	btrfs_release_path(root, path);
+	return ret;
+}
+
+#endif
+
+/*
+ * this can truncate away extent items, csum items and directory items.
+ * It starts at a high offset and removes keys until it can't find
+ * any higher than new_size
+ *
+ * csum items that cross the new i_size are truncated to the new size
+ * as well.
+ *
+ * min_type is the minimum key type to truncate down to.  If set to 0, this
+ * will kill all the items on this inode, including the INODE_ITEM_KEY.
+ */
+noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct inode *inode,
+					u64 new_size, u32 min_type)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	u32 found_type;
+	struct extent_buffer *leaf;
+	struct btrfs_file_extent_item *fi;
+	u64 extent_start = 0;
+	u64 extent_num_bytes = 0;
+	u64 item_end = 0;
+	u64 root_gen = 0;
+	u64 root_owner = 0;
+	int found_extent;
+	int del_item;
+	int pending_del_nr = 0;
+	int pending_del_slot = 0;
+	int extent_type = -1;
+	int encoding;
+	u64 mask = root->sectorsize - 1;
+
+	if (root->ref_cows)
+		btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0);
+	path = btrfs_alloc_path();
+	path->reada = -1;
+	BUG_ON(!path);
+
+	/* FIXME, add redo link to tree so we don't leak on crash */
+	key.objectid = inode->i_ino;
+	key.offset = (u64)-1;
+	key.type = (u8)-1;
+
+	btrfs_init_path(path);
+
+search_again:
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret < 0)
+		goto error;
+
+	if (ret > 0) {
+		/* there are no items in the tree for us to truncate, we're
+		 * done
+		 */
+		if (path->slots[0] == 0) {
+			ret = 0;
+			goto error;
+		}
+		path->slots[0]--;
+	}
+
+	while (1) {
+		fi = NULL;
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		found_type = btrfs_key_type(&found_key);
+		encoding = 0;
+
+		if (found_key.objectid != inode->i_ino)
+			break;
+
+		if (found_type < min_type)
+			break;
+
+		item_end = found_key.offset;
+		if (found_type == BTRFS_EXTENT_DATA_KEY) {
+			fi = btrfs_item_ptr(leaf, path->slots[0],
+					    struct btrfs_file_extent_item);
+			extent_type = btrfs_file_extent_type(leaf, fi);
+			encoding = btrfs_file_extent_compression(leaf, fi);
+			encoding |= btrfs_file_extent_encryption(leaf, fi);
+			encoding |= btrfs_file_extent_other_encoding(leaf, fi);
+
+			if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
+				item_end +=
+				    btrfs_file_extent_num_bytes(leaf, fi);
+			} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
+				item_end += btrfs_file_extent_inline_len(leaf,
+									 fi);
+			}
+			item_end--;
+		}
+		if (item_end < new_size) {
+			if (found_type == BTRFS_DIR_ITEM_KEY)
+				found_type = BTRFS_INODE_ITEM_KEY;
+			else if (found_type == BTRFS_EXTENT_ITEM_KEY)
+				found_type = BTRFS_EXTENT_DATA_KEY;
+			else if (found_type == BTRFS_EXTENT_DATA_KEY)
+				found_type = BTRFS_XATTR_ITEM_KEY;
+			else if (found_type == BTRFS_XATTR_ITEM_KEY)
+				found_type = BTRFS_INODE_REF_KEY;
+			else if (found_type)
+				found_type--;
+			else
+				break;
+			btrfs_set_key_type(&key, found_type);
+			goto next;
+		}
+		if (found_key.offset >= new_size)
+			del_item = 1;
+		else
+			del_item = 0;
+		found_extent = 0;
+
+		/* FIXME, shrink the extent if the ref count is only 1 */
+		if (found_type != BTRFS_EXTENT_DATA_KEY)
+			goto delete;
+
+		if (extent_type != BTRFS_FILE_EXTENT_INLINE) {
+			u64 num_dec;
+			extent_start = btrfs_file_extent_disk_bytenr(leaf, fi);
+			if (!del_item && !encoding) {
+				u64 orig_num_bytes =
+					btrfs_file_extent_num_bytes(leaf, fi);
+				extent_num_bytes = new_size -
+					found_key.offset + root->sectorsize - 1;
+				extent_num_bytes = extent_num_bytes &
+					~((u64)root->sectorsize - 1);
+				btrfs_set_file_extent_num_bytes(leaf, fi,
+							 extent_num_bytes);
+				num_dec = (orig_num_bytes -
+					   extent_num_bytes);
+				if (root->ref_cows && extent_start != 0)
+					inode_sub_bytes(inode, num_dec);
+				btrfs_mark_buffer_dirty(leaf);
+			} else {
+				extent_num_bytes =
+					btrfs_file_extent_disk_num_bytes(leaf,
+									 fi);
+				/* FIXME blocksize != 4096 */
+				num_dec = btrfs_file_extent_num_bytes(leaf, fi);
+				if (extent_start != 0) {
+					found_extent = 1;
+					if (root->ref_cows)
+						inode_sub_bytes(inode, num_dec);
+				}
+				root_gen = btrfs_header_generation(leaf);
+				root_owner = btrfs_header_owner(leaf);
+			}
+		} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
+			/*
+			 * we can't truncate inline items that have had
+			 * special encodings
+			 */
+			if (!del_item &&
+			    btrfs_file_extent_compression(leaf, fi) == 0 &&
+			    btrfs_file_extent_encryption(leaf, fi) == 0 &&
+			    btrfs_file_extent_other_encoding(leaf, fi) == 0) {
+				u32 size = new_size - found_key.offset;
+
+				if (root->ref_cows) {
+					inode_sub_bytes(inode, item_end + 1 -
+							new_size);
+				}
+				size =
+				    btrfs_file_extent_calc_inline_size(size);
+				ret = btrfs_truncate_item(trans, root, path,
+							  size, 1);
+				BUG_ON(ret);
+			} else if (root->ref_cows) {
+				inode_sub_bytes(inode, item_end + 1 -
+						found_key.offset);
+			}
+		}
+delete:
+		if (del_item) {
+			if (!pending_del_nr) {
+				/* no pending yet, add ourselves */
+				pending_del_slot = path->slots[0];
+				pending_del_nr = 1;
+			} else if (pending_del_nr &&
+				   path->slots[0] + 1 == pending_del_slot) {
+				/* hop on the pending chunk */
+				pending_del_nr++;
+				pending_del_slot = path->slots[0];
+			} else {
+				BUG();
+			}
+		} else {
+			break;
+		}
+		if (found_extent) {
+			ret = btrfs_free_extent(trans, root, extent_start,
+						extent_num_bytes,
+						leaf->start, root_owner,
+						root_gen, inode->i_ino, 0);
+			BUG_ON(ret);
+		}
+next:
+		if (path->slots[0] == 0) {
+			if (pending_del_nr)
+				goto del_pending;
+			btrfs_release_path(root, path);
+			goto search_again;
+		}
+
+		path->slots[0]--;
+		if (pending_del_nr &&
+		    path->slots[0] + 1 != pending_del_slot) {
+			struct btrfs_key debug;
+del_pending:
+			btrfs_item_key_to_cpu(path->nodes[0], &debug,
+					      pending_del_slot);
+			ret = btrfs_del_items(trans, root, path,
+					      pending_del_slot,
+					      pending_del_nr);
+			BUG_ON(ret);
+			pending_del_nr = 0;
+			btrfs_release_path(root, path);
+			goto search_again;
+		}
+	}
+	ret = 0;
+error:
+	if (pending_del_nr) {
+		ret = btrfs_del_items(trans, root, path, pending_del_slot,
+				      pending_del_nr);
+	}
+	btrfs_free_path(path);
+	inode->i_sb->s_dirt = 1;
+	return ret;
+}
+
+/*
+ * taken from block_truncate_page, but does cow as it zeros out
+ * any bytes left in the last page in the file.
+ */
+static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
+{
+	struct inode *inode = mapping->host;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct btrfs_ordered_extent *ordered;
+	char *kaddr;
+	u32 blocksize = root->sectorsize;
+	pgoff_t index = from >> PAGE_CACHE_SHIFT;
+	unsigned offset = from & (PAGE_CACHE_SIZE-1);
+	struct page *page;
+	int ret = 0;
+	u64 page_start;
+	u64 page_end;
+
+	if ((offset & (blocksize - 1)) == 0)
+		goto out;
+
+	ret = -ENOMEM;
+again:
+	page = grab_cache_page(mapping, index);
+	if (!page)
+		goto out;
+
+	page_start = page_offset(page);
+	page_end = page_start + PAGE_CACHE_SIZE - 1;
+
+	if (!PageUptodate(page)) {
+		ret = btrfs_readpage(NULL, page);
+		lock_page(page);
+		if (page->mapping != mapping) {
+			unlock_page(page);
+			page_cache_release(page);
+			goto again;
+		}
+		if (!PageUptodate(page)) {
+			ret = -EIO;
+			goto out_unlock;
+		}
+	}
+	wait_on_page_writeback(page);
+
+	lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+	set_page_extent_mapped(page);
+
+	ordered = btrfs_lookup_ordered_extent(inode, page_start);
+	if (ordered) {
+		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+		unlock_page(page);
+		page_cache_release(page);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+		goto again;
+	}
+
+	btrfs_set_extent_delalloc(inode, page_start, page_end);
+	ret = 0;
+	if (offset != PAGE_CACHE_SIZE) {
+		kaddr = kmap(page);
+		memset(kaddr + offset, 0, PAGE_CACHE_SIZE - offset);
+		flush_dcache_page(page);
+		kunmap(page);
+	}
+	ClearPageChecked(page);
+	set_page_dirty(page);
+	unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+
+out_unlock:
+	unlock_page(page);
+	page_cache_release(page);
+out:
+	return ret;
+}
+
+int btrfs_cont_expand(struct inode *inode, loff_t size)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct extent_map *em;
+	u64 mask = root->sectorsize - 1;
+	u64 hole_start = (inode->i_size + mask) & ~mask;
+	u64 block_end = (size + mask) & ~mask;
+	u64 last_byte;
+	u64 cur_offset;
+	u64 hole_size;
+	int err;
+
+	if (size <= hole_start)
+		return 0;
+
+	err = btrfs_check_free_space(root, 1, 0);
+	if (err)
+		return err;
+
+	btrfs_truncate_page(inode->i_mapping, inode->i_size);
+
+	while (1) {
+		struct btrfs_ordered_extent *ordered;
+		btrfs_wait_ordered_range(inode, hole_start,
+					 block_end - hole_start);
+		lock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+		ordered = btrfs_lookup_ordered_extent(inode, hole_start);
+		if (!ordered)
+			break;
+		unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+		btrfs_put_ordered_extent(ordered);
+	}
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, inode);
+
+	cur_offset = hole_start;
+	while (1) {
+		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
+				block_end - cur_offset, 0);
+		BUG_ON(IS_ERR(em) || !em);
+		last_byte = min(extent_map_end(em), block_end);
+		last_byte = (last_byte + mask) & ~mask;
+		if (test_bit(EXTENT_FLAG_VACANCY, &em->flags)) {
+			u64 hint_byte = 0;
+			hole_size = last_byte - cur_offset;
+			err = btrfs_drop_extents(trans, root, inode,
+						 cur_offset,
+						 cur_offset + hole_size,
+						 cur_offset, &hint_byte);
+			if (err)
+				break;
+			err = btrfs_insert_file_extent(trans, root,
+					inode->i_ino, cur_offset, 0,
+					0, hole_size, 0, hole_size,
+					0, 0, 0);
+			btrfs_drop_extent_cache(inode, hole_start,
+					last_byte - 1, 0);
+		}
+		free_extent_map(em);
+		cur_offset = last_byte;
+		if (err || cur_offset >= block_end)
+			break;
+	}
+
+	btrfs_end_transaction(trans, root);
+	unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+	return err;
+}
+
+static int btrfs_setattr(struct dentry *dentry, struct iattr *attr)
+{
+	struct inode *inode = dentry->d_inode;
+	int err;
+
+	err = inode_change_ok(inode, attr);
+	if (err)
+		return err;
+
+	if (S_ISREG(inode->i_mode) &&
+	    attr->ia_valid & ATTR_SIZE && attr->ia_size > inode->i_size) {
+		err = btrfs_cont_expand(inode, attr->ia_size);
+		if (err)
+			return err;
+	}
+
+	err = inode_setattr(inode, attr);
+
+	if (!err && ((attr->ia_valid & ATTR_MODE)))
+		err = btrfs_acl_chmod(inode);
+	return err;
+}
+
+void btrfs_delete_inode(struct inode *inode)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	unsigned long nr;
+	int ret;
+
+	truncate_inode_pages(&inode->i_data, 0);
+	if (is_bad_inode(inode)) {
+		btrfs_orphan_del(NULL, inode);
+		goto no_delete;
+	}
+	btrfs_wait_ordered_range(inode, 0, (u64)-1);
+
+	btrfs_i_size_write(inode, 0);
+	trans = btrfs_join_transaction(root, 1);
+
+	btrfs_set_trans_block_group(trans, inode);
+	ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, 0);
+	if (ret) {
+		btrfs_orphan_del(NULL, inode);
+		goto no_delete_lock;
+	}
+
+	btrfs_orphan_del(trans, inode);
+
+	nr = trans->blocks_used;
+	clear_inode(inode);
+
+	btrfs_end_transaction(trans, root);
+	btrfs_btree_balance_dirty(root, nr);
+	return;
+
+no_delete_lock:
+	nr = trans->blocks_used;
+	btrfs_end_transaction(trans, root);
+	btrfs_btree_balance_dirty(root, nr);
+no_delete:
+	clear_inode(inode);
+}
+
+/*
+ * this returns the key found in the dir entry in the location pointer.
+ * If no dir entries were found, location->objectid is 0.
+ */
+static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry,
+			       struct btrfs_key *location)
+{
+	const char *name = dentry->d_name.name;
+	int namelen = dentry->d_name.len;
+	struct btrfs_dir_item *di;
+	struct btrfs_path *path;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	int ret = 0;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	di = btrfs_lookup_dir_item(NULL, root, path, dir->i_ino, name,
+				    namelen, 0);
+	if (IS_ERR(di))
+		ret = PTR_ERR(di);
+
+	if (!di || IS_ERR(di))
+		goto out_err;
+
+	btrfs_dir_item_key_to_cpu(path->nodes[0], di, location);
+out:
+	btrfs_free_path(path);
+	return ret;
+out_err:
+	location->objectid = 0;
+	goto out;
+}
+
+/*
+ * when we hit a tree root in a directory, the btrfs part of the inode
+ * needs to be changed to reflect the root directory of the tree root.  This
+ * is kind of like crossing a mount point.
+ */
+static int fixup_tree_root_location(struct btrfs_root *root,
+			     struct btrfs_key *location,
+			     struct btrfs_root **sub_root,
+			     struct dentry *dentry)
+{
+	struct btrfs_root_item *ri;
+
+	if (btrfs_key_type(location) != BTRFS_ROOT_ITEM_KEY)
+		return 0;
+	if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
+		return 0;
+
+	*sub_root = btrfs_read_fs_root(root->fs_info, location,
+					dentry->d_name.name,
+					dentry->d_name.len);
+	if (IS_ERR(*sub_root))
+		return PTR_ERR(*sub_root);
+
+	ri = &(*sub_root)->root_item;
+	location->objectid = btrfs_root_dirid(ri);
+	btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
+	location->offset = 0;
+
+	return 0;
+}
+
+static noinline void init_btrfs_i(struct inode *inode)
+{
+	struct btrfs_inode *bi = BTRFS_I(inode);
+
+	bi->i_acl = NULL;
+	bi->i_default_acl = NULL;
+
+	bi->generation = 0;
+	bi->sequence = 0;
+	bi->last_trans = 0;
+	bi->logged_trans = 0;
+	bi->delalloc_bytes = 0;
+	bi->disk_i_size = 0;
+	bi->flags = 0;
+	bi->index_cnt = (u64)-1;
+	bi->log_dirty_trans = 0;
+	extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
+	extent_io_tree_init(&BTRFS_I(inode)->io_tree,
+			     inode->i_mapping, GFP_NOFS);
+	extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
+			     inode->i_mapping, GFP_NOFS);
+	INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
+	btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
+	mutex_init(&BTRFS_I(inode)->extent_mutex);
+	mutex_init(&BTRFS_I(inode)->log_mutex);
+}
+
+static int btrfs_init_locked_inode(struct inode *inode, void *p)
+{
+	struct btrfs_iget_args *args = p;
+	inode->i_ino = args->ino;
+	init_btrfs_i(inode);
+	BTRFS_I(inode)->root = args->root;
+	return 0;
+}
+
+static int btrfs_find_actor(struct inode *inode, void *opaque)
+{
+	struct btrfs_iget_args *args = opaque;
+	return args->ino == inode->i_ino &&
+		args->root == BTRFS_I(inode)->root;
+}
+
+struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
+			    struct btrfs_root *root, int wait)
+{
+	struct inode *inode;
+	struct btrfs_iget_args args;
+	args.ino = objectid;
+	args.root = root;
+
+	if (wait) {
+		inode = ilookup5(s, objectid, btrfs_find_actor,
+				 (void *)&args);
+	} else {
+		inode = ilookup5_nowait(s, objectid, btrfs_find_actor,
+					(void *)&args);
+	}
+	return inode;
+}
+
+struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
+				struct btrfs_root *root)
+{
+	struct inode *inode;
+	struct btrfs_iget_args args;
+	args.ino = objectid;
+	args.root = root;
+
+	inode = iget5_locked(s, objectid, btrfs_find_actor,
+			     btrfs_init_locked_inode,
+			     (void *)&args);
+	return inode;
+}
+
+/* Get an inode object given its location and corresponding root.
+ * Returns in *is_new if the inode was read from disk
+ */
+struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
+			 struct btrfs_root *root, int *is_new)
+{
+	struct inode *inode;
+
+	inode = btrfs_iget_locked(s, location->objectid, root);
+	if (!inode)
+		return ERR_PTR(-EACCES);
+
+	if (inode->i_state & I_NEW) {
+		BTRFS_I(inode)->root = root;
+		memcpy(&BTRFS_I(inode)->location, location, sizeof(*location));
+		btrfs_read_locked_inode(inode);
+		unlock_new_inode(inode);
+		if (is_new)
+			*is_new = 1;
+	} else {
+		if (is_new)
+			*is_new = 0;
+	}
+
+	return inode;
+}
+
+struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
+{
+	struct inode *inode;
+	struct btrfs_inode *bi = BTRFS_I(dir);
+	struct btrfs_root *root = bi->root;
+	struct btrfs_root *sub_root = root;
+	struct btrfs_key location;
+	int ret, new;
+
+	if (dentry->d_name.len > BTRFS_NAME_LEN)
+		return ERR_PTR(-ENAMETOOLONG);
+
+	ret = btrfs_inode_by_name(dir, dentry, &location);
+
+	if (ret < 0)
+		return ERR_PTR(ret);
+
+	inode = NULL;
+	if (location.objectid) {
+		ret = fixup_tree_root_location(root, &location, &sub_root,
+						dentry);
+		if (ret < 0)
+			return ERR_PTR(ret);
+		if (ret > 0)
+			return ERR_PTR(-ENOENT);
+		inode = btrfs_iget(dir->i_sb, &location, sub_root, &new);
+		if (IS_ERR(inode))
+			return ERR_CAST(inode);
+	}
+	return inode;
+}
+
+static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
+				   struct nameidata *nd)
+{
+	struct inode *inode;
+
+	if (dentry->d_name.len > BTRFS_NAME_LEN)
+		return ERR_PTR(-ENAMETOOLONG);
+
+	inode = btrfs_lookup_dentry(dir, dentry);
+	if (IS_ERR(inode))
+		return ERR_CAST(inode);
+
+	return d_splice_alias(inode, dentry);
+}
+
+static unsigned char btrfs_filetype_table[] = {
+	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
+};
+
+static int btrfs_real_readdir(struct file *filp, void *dirent,
+			      filldir_t filldir)
+{
+	struct inode *inode = filp->f_dentry->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_item *item;
+	struct btrfs_dir_item *di;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_path *path;
+	int ret;
+	u32 nritems;
+	struct extent_buffer *leaf;
+	int slot;
+	int advance;
+	unsigned char d_type;
+	int over = 0;
+	u32 di_cur;
+	u32 di_total;
+	u32 di_len;
+	int key_type = BTRFS_DIR_INDEX_KEY;
+	char tmp_name[32];
+	char *name_ptr;
+	int name_len;
+
+	/* FIXME, use a real flag for deciding about the key type */
+	if (root->fs_info->tree_root == root)
+		key_type = BTRFS_DIR_ITEM_KEY;
+
+	/* special case for "." */
+	if (filp->f_pos == 0) {
+		over = filldir(dirent, ".", 1,
+			       1, inode->i_ino,
+			       DT_DIR);
+		if (over)
+			return 0;
+		filp->f_pos = 1;
+	}
+	/* special case for .., just use the back ref */
+	if (filp->f_pos == 1) {
+		u64 pino = parent_ino(filp->f_path.dentry);
+		over = filldir(dirent, "..", 2,
+			       2, pino, DT_DIR);
+		if (over)
+			return 0;
+		filp->f_pos = 2;
+	}
+	path = btrfs_alloc_path();
+	path->reada = 2;
+
+	btrfs_set_key_type(&key, key_type);
+	key.offset = filp->f_pos;
+	key.objectid = inode->i_ino;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto err;
+	advance = 0;
+
+	while (1) {
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		slot = path->slots[0];
+		if (advance || slot >= nritems) {
+			if (slot >= nritems - 1) {
+				ret = btrfs_next_leaf(root, path);
+				if (ret)
+					break;
+				leaf = path->nodes[0];
+				nritems = btrfs_header_nritems(leaf);
+				slot = path->slots[0];
+			} else {
+				slot++;
+				path->slots[0]++;
+			}
+		}
+
+		advance = 1;
+		item = btrfs_item_nr(leaf, slot);
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		if (found_key.objectid != key.objectid)
+			break;
+		if (btrfs_key_type(&found_key) != key_type)
+			break;
+		if (found_key.offset < filp->f_pos)
+			continue;
+
+		filp->f_pos = found_key.offset;
+
+		di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
+		di_cur = 0;
+		di_total = btrfs_item_size(leaf, item);
+
+		while (di_cur < di_total) {
+			struct btrfs_key location;
+
+			name_len = btrfs_dir_name_len(leaf, di);
+			if (name_len <= sizeof(tmp_name)) {
+				name_ptr = tmp_name;
+			} else {
+				name_ptr = kmalloc(name_len, GFP_NOFS);
+				if (!name_ptr) {
+					ret = -ENOMEM;
+					goto err;
+				}
+			}
+			read_extent_buffer(leaf, name_ptr,
+					   (unsigned long)(di + 1), name_len);
+
+			d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)];
+			btrfs_dir_item_key_to_cpu(leaf, di, &location);
+
+			/* is this a reference to our own snapshot? If so
+			 * skip it
+			 */
+			if (location.type == BTRFS_ROOT_ITEM_KEY &&
+			    location.objectid == root->root_key.objectid) {
+				over = 0;
+				goto skip;
+			}
+			over = filldir(dirent, name_ptr, name_len,
+				       found_key.offset, location.objectid,
+				       d_type);
+
+skip:
+			if (name_ptr != tmp_name)
+				kfree(name_ptr);
+
+			if (over)
+				goto nopos;
+			di_len = btrfs_dir_name_len(leaf, di) +
+				 btrfs_dir_data_len(leaf, di) + sizeof(*di);
+			di_cur += di_len;
+			di = (struct btrfs_dir_item *)((char *)di + di_len);
+		}
+	}
+
+	/* Reached end of directory/root. Bump pos past the last item. */
+	if (key_type == BTRFS_DIR_INDEX_KEY)
+		filp->f_pos = INT_LIMIT(typeof(filp->f_pos));
+	else
+		filp->f_pos++;
+nopos:
+	ret = 0;
+err:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_write_inode(struct inode *inode, int wait)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	int ret = 0;
+
+	if (root->fs_info->btree_inode == inode)
+		return 0;
+
+	if (wait) {
+		trans = btrfs_join_transaction(root, 1);
+		btrfs_set_trans_block_group(trans, inode);
+		ret = btrfs_commit_transaction(trans, root);
+	}
+	return ret;
+}
+
+/*
+ * This is somewhat expensive, updating the tree every time the
+ * inode changes.  But, it is most likely to find the inode in cache.
+ * FIXME, needs more benchmarking...there are no reasons other than performance
+ * to keep or drop this code.
+ */
+void btrfs_dirty_inode(struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+
+	trans = btrfs_join_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, inode);
+	btrfs_update_inode(trans, root, inode);
+	btrfs_end_transaction(trans, root);
+}
+
+/*
+ * find the highest existing sequence number in a directory
+ * and then set the in-memory index_cnt variable to reflect
+ * free sequence numbers
+ */
+static int btrfs_set_inode_index_count(struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_key key, found_key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	int ret;
+
+	key.objectid = inode->i_ino;
+	btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY);
+	key.offset = (u64)-1;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	/* FIXME: we should be able to handle this */
+	if (ret == 0)
+		goto out;
+	ret = 0;
+
+	/*
+	 * MAGIC NUMBER EXPLANATION:
+	 * since we search a directory based on f_pos we have to start at 2
+	 * since '.' and '..' have f_pos of 0 and 1 respectively, so everybody
+	 * else has to start at 2
+	 */
+	if (path->slots[0] == 0) {
+		BTRFS_I(inode)->index_cnt = 2;
+		goto out;
+	}
+
+	path->slots[0]--;
+
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+	if (found_key.objectid != inode->i_ino ||
+	    btrfs_key_type(&found_key) != BTRFS_DIR_INDEX_KEY) {
+		BTRFS_I(inode)->index_cnt = 2;
+		goto out;
+	}
+
+	BTRFS_I(inode)->index_cnt = found_key.offset + 1;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * helper to find a free sequence number in a given directory.  This current
+ * code is very simple, later versions will do smarter things in the btree
+ */
+int btrfs_set_inode_index(struct inode *dir, u64 *index)
+{
+	int ret = 0;
+
+	if (BTRFS_I(dir)->index_cnt == (u64)-1) {
+		ret = btrfs_set_inode_index_count(dir);
+		if (ret)
+			return ret;
+	}
+
+	*index = BTRFS_I(dir)->index_cnt;
+	BTRFS_I(dir)->index_cnt++;
+
+	return ret;
+}
+
+static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root,
+				     struct inode *dir,
+				     const char *name, int name_len,
+				     u64 ref_objectid, u64 objectid,
+				     u64 alloc_hint, int mode, u64 *index)
+{
+	struct inode *inode;
+	struct btrfs_inode_item *inode_item;
+	struct btrfs_key *location;
+	struct btrfs_path *path;
+	struct btrfs_inode_ref *ref;
+	struct btrfs_key key[2];
+	u32 sizes[2];
+	unsigned long ptr;
+	int ret;
+	int owner;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	inode = new_inode(root->fs_info->sb);
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+
+	if (dir) {
+		ret = btrfs_set_inode_index(dir, index);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+	/*
+	 * index_cnt is ignored for everything but a dir,
+	 * btrfs_get_inode_index_count has an explanation for the magic
+	 * number
+	 */
+	init_btrfs_i(inode);
+	BTRFS_I(inode)->index_cnt = 2;
+	BTRFS_I(inode)->root = root;
+	BTRFS_I(inode)->generation = trans->transid;
+
+	if (mode & S_IFDIR)
+		owner = 0;
+	else
+		owner = 1;
+	BTRFS_I(inode)->block_group =
+			btrfs_find_block_group(root, 0, alloc_hint, owner);
+	if ((mode & S_IFREG)) {
+		if (btrfs_test_opt(root, NODATASUM))
+			btrfs_set_flag(inode, NODATASUM);
+		if (btrfs_test_opt(root, NODATACOW))
+			btrfs_set_flag(inode, NODATACOW);
+	}
+
+	key[0].objectid = objectid;
+	btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY);
+	key[0].offset = 0;
+
+	key[1].objectid = objectid;
+	btrfs_set_key_type(&key[1], BTRFS_INODE_REF_KEY);
+	key[1].offset = ref_objectid;
+
+	sizes[0] = sizeof(struct btrfs_inode_item);
+	sizes[1] = name_len + sizeof(*ref);
+
+	ret = btrfs_insert_empty_items(trans, root, path, key, sizes, 2);
+	if (ret != 0)
+		goto fail;
+
+	if (objectid > root->highest_inode)
+		root->highest_inode = objectid;
+
+	inode->i_uid = current_fsuid();
+	inode->i_gid = current_fsgid();
+	inode->i_mode = mode;
+	inode->i_ino = objectid;
+	inode_set_bytes(inode, 0);
+	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
+	inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				  struct btrfs_inode_item);
+	fill_inode_item(trans, path->nodes[0], inode_item, inode);
+
+	ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1,
+			     struct btrfs_inode_ref);
+	btrfs_set_inode_ref_name_len(path->nodes[0], ref, name_len);
+	btrfs_set_inode_ref_index(path->nodes[0], ref, *index);
+	ptr = (unsigned long)(ref + 1);
+	write_extent_buffer(path->nodes[0], name, ptr, name_len);
+
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	btrfs_free_path(path);
+
+	location = &BTRFS_I(inode)->location;
+	location->objectid = objectid;
+	location->offset = 0;
+	btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
+
+	insert_inode_hash(inode);
+	return inode;
+fail:
+	if (dir)
+		BTRFS_I(dir)->index_cnt--;
+	btrfs_free_path(path);
+	return ERR_PTR(ret);
+}
+
+static inline u8 btrfs_inode_type(struct inode *inode)
+{
+	return btrfs_type_by_mode[(inode->i_mode & S_IFMT) >> S_SHIFT];
+}
+
+/*
+ * utility function to add 'inode' into 'parent_inode' with
+ * a give name and a given sequence number.
+ * if 'add_backref' is true, also insert a backref from the
+ * inode to the parent directory.
+ */
+int btrfs_add_link(struct btrfs_trans_handle *trans,
+		   struct inode *parent_inode, struct inode *inode,
+		   const char *name, int name_len, int add_backref, u64 index)
+{
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_root *root = BTRFS_I(parent_inode)->root;
+
+	key.objectid = inode->i_ino;
+	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+	key.offset = 0;
+
+	ret = btrfs_insert_dir_item(trans, root, name, name_len,
+				    parent_inode->i_ino,
+				    &key, btrfs_inode_type(inode),
+				    index);
+	if (ret == 0) {
+		if (add_backref) {
+			ret = btrfs_insert_inode_ref(trans, root,
+						     name, name_len,
+						     inode->i_ino,
+						     parent_inode->i_ino,
+						     index);
+		}
+		btrfs_i_size_write(parent_inode, parent_inode->i_size +
+				   name_len * 2);
+		parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
+		ret = btrfs_update_inode(trans, root, parent_inode);
+	}
+	return ret;
+}
+
+static int btrfs_add_nondir(struct btrfs_trans_handle *trans,
+			    struct dentry *dentry, struct inode *inode,
+			    int backref, u64 index)
+{
+	int err = btrfs_add_link(trans, dentry->d_parent->d_inode,
+				 inode, dentry->d_name.name,
+				 dentry->d_name.len, backref, index);
+	if (!err) {
+		d_instantiate(dentry, inode);
+		return 0;
+	}
+	if (err > 0)
+		err = -EEXIST;
+	return err;
+}
+
+static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
+			int mode, dev_t rdev)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct inode *inode = NULL;
+	int err;
+	int drop_inode = 0;
+	u64 objectid;
+	unsigned long nr = 0;
+	u64 index = 0;
+
+	if (!new_valid_dev(rdev))
+		return -EINVAL;
+
+	err = btrfs_check_free_space(root, 1, 0);
+	if (err)
+		goto fail;
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, dir);
+
+	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
+	if (err) {
+		err = -ENOSPC;
+		goto out_unlock;
+	}
+
+	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
+				dentry->d_name.len,
+				dentry->d_parent->d_inode->i_ino, objectid,
+				BTRFS_I(dir)->block_group, mode, &index);
+	err = PTR_ERR(inode);
+	if (IS_ERR(inode))
+		goto out_unlock;
+
+	err = btrfs_init_acl(inode, dir);
+	if (err) {
+		drop_inode = 1;
+		goto out_unlock;
+	}
+
+	btrfs_set_trans_block_group(trans, inode);
+	err = btrfs_add_nondir(trans, dentry, inode, 0, index);
+	if (err)
+		drop_inode = 1;
+	else {
+		inode->i_op = &btrfs_special_inode_operations;
+		init_special_inode(inode, inode->i_mode, rdev);
+		btrfs_update_inode(trans, root, inode);
+	}
+	dir->i_sb->s_dirt = 1;
+	btrfs_update_inode_block_group(trans, inode);
+	btrfs_update_inode_block_group(trans, dir);
+out_unlock:
+	nr = trans->blocks_used;
+	btrfs_end_transaction_throttle(trans, root);
+fail:
+	if (drop_inode) {
+		inode_dec_link_count(inode);
+		iput(inode);
+	}
+	btrfs_btree_balance_dirty(root, nr);
+	return err;
+}
+
+static int btrfs_create(struct inode *dir, struct dentry *dentry,
+			int mode, struct nameidata *nd)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct inode *inode = NULL;
+	int err;
+	int drop_inode = 0;
+	unsigned long nr = 0;
+	u64 objectid;
+	u64 index = 0;
+
+	err = btrfs_check_free_space(root, 1, 0);
+	if (err)
+		goto fail;
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, dir);
+
+	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
+	if (err) {
+		err = -ENOSPC;
+		goto out_unlock;
+	}
+
+	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
+				dentry->d_name.len,
+				dentry->d_parent->d_inode->i_ino,
+				objectid, BTRFS_I(dir)->block_group, mode,
+				&index);
+	err = PTR_ERR(inode);
+	if (IS_ERR(inode))
+		goto out_unlock;
+
+	err = btrfs_init_acl(inode, dir);
+	if (err) {
+		drop_inode = 1;
+		goto out_unlock;
+	}
+
+	btrfs_set_trans_block_group(trans, inode);
+	err = btrfs_add_nondir(trans, dentry, inode, 0, index);
+	if (err)
+		drop_inode = 1;
+	else {
+		inode->i_mapping->a_ops = &btrfs_aops;
+		inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
+		inode->i_fop = &btrfs_file_operations;
+		inode->i_op = &btrfs_file_inode_operations;
+		BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
+	}
+	dir->i_sb->s_dirt = 1;
+	btrfs_update_inode_block_group(trans, inode);
+	btrfs_update_inode_block_group(trans, dir);
+out_unlock:
+	nr = trans->blocks_used;
+	btrfs_end_transaction_throttle(trans, root);
+fail:
+	if (drop_inode) {
+		inode_dec_link_count(inode);
+		iput(inode);
+	}
+	btrfs_btree_balance_dirty(root, nr);
+	return err;
+}
+
+static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
+		      struct dentry *dentry)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct inode *inode = old_dentry->d_inode;
+	u64 index;
+	unsigned long nr = 0;
+	int err;
+	int drop_inode = 0;
+
+	if (inode->i_nlink == 0)
+		return -ENOENT;
+
+	btrfs_inc_nlink(inode);
+	err = btrfs_check_free_space(root, 1, 0);
+	if (err)
+		goto fail;
+	err = btrfs_set_inode_index(dir, &index);
+	if (err)
+		goto fail;
+
+	trans = btrfs_start_transaction(root, 1);
+
+	btrfs_set_trans_block_group(trans, dir);
+	atomic_inc(&inode->i_count);
+
+	err = btrfs_add_nondir(trans, dentry, inode, 1, index);
+
+	if (err)
+		drop_inode = 1;
+
+	dir->i_sb->s_dirt = 1;
+	btrfs_update_inode_block_group(trans, dir);
+	err = btrfs_update_inode(trans, root, inode);
+
+	if (err)
+		drop_inode = 1;
+
+	nr = trans->blocks_used;
+	btrfs_end_transaction_throttle(trans, root);
+fail:
+	if (drop_inode) {
+		inode_dec_link_count(inode);
+		iput(inode);
+	}
+	btrfs_btree_balance_dirty(root, nr);
+	return err;
+}
+
+static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
+{
+	struct inode *inode = NULL;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	int err = 0;
+	int drop_on_err = 0;
+	u64 objectid = 0;
+	u64 index = 0;
+	unsigned long nr = 1;
+
+	err = btrfs_check_free_space(root, 1, 0);
+	if (err)
+		goto out_unlock;
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, dir);
+
+	if (IS_ERR(trans)) {
+		err = PTR_ERR(trans);
+		goto out_unlock;
+	}
+
+	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
+	if (err) {
+		err = -ENOSPC;
+		goto out_unlock;
+	}
+
+	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
+				dentry->d_name.len,
+				dentry->d_parent->d_inode->i_ino, objectid,
+				BTRFS_I(dir)->block_group, S_IFDIR | mode,
+				&index);
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		goto out_fail;
+	}
+
+	drop_on_err = 1;
+
+	err = btrfs_init_acl(inode, dir);
+	if (err)
+		goto out_fail;
+
+	inode->i_op = &btrfs_dir_inode_operations;
+	inode->i_fop = &btrfs_dir_file_operations;
+	btrfs_set_trans_block_group(trans, inode);
+
+	btrfs_i_size_write(inode, 0);
+	err = btrfs_update_inode(trans, root, inode);
+	if (err)
+		goto out_fail;
+
+	err = btrfs_add_link(trans, dentry->d_parent->d_inode,
+				 inode, dentry->d_name.name,
+				 dentry->d_name.len, 0, index);
+	if (err)
+		goto out_fail;
+
+	d_instantiate(dentry, inode);
+	drop_on_err = 0;
+	dir->i_sb->s_dirt = 1;
+	btrfs_update_inode_block_group(trans, inode);
+	btrfs_update_inode_block_group(trans, dir);
+
+out_fail:
+	nr = trans->blocks_used;
+	btrfs_end_transaction_throttle(trans, root);
+
+out_unlock:
+	if (drop_on_err)
+		iput(inode);
+	btrfs_btree_balance_dirty(root, nr);
+	return err;
+}
+
+/* helper for btfs_get_extent.  Given an existing extent in the tree,
+ * and an extent that you want to insert, deal with overlap and insert
+ * the new extent into the tree.
+ */
+static int merge_extent_mapping(struct extent_map_tree *em_tree,
+				struct extent_map *existing,
+				struct extent_map *em,
+				u64 map_start, u64 map_len)
+{
+	u64 start_diff;
+
+	BUG_ON(map_start < em->start || map_start >= extent_map_end(em));
+	start_diff = map_start - em->start;
+	em->start = map_start;
+	em->len = map_len;
+	if (em->block_start < EXTENT_MAP_LAST_BYTE &&
+	    !test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
+		em->block_start += start_diff;
+		em->block_len -= start_diff;
+	}
+	return add_extent_mapping(em_tree, em);
+}
+
+static noinline int uncompress_inline(struct btrfs_path *path,
+				      struct inode *inode, struct page *page,
+				      size_t pg_offset, u64 extent_offset,
+				      struct btrfs_file_extent_item *item)
+{
+	int ret;
+	struct extent_buffer *leaf = path->nodes[0];
+	char *tmp;
+	size_t max_size;
+	unsigned long inline_size;
+	unsigned long ptr;
+
+	WARN_ON(pg_offset != 0);
+	max_size = btrfs_file_extent_ram_bytes(leaf, item);
+	inline_size = btrfs_file_extent_inline_item_len(leaf,
+					btrfs_item_nr(leaf, path->slots[0]));
+	tmp = kmalloc(inline_size, GFP_NOFS);
+	ptr = btrfs_file_extent_inline_start(item);
+
+	read_extent_buffer(leaf, tmp, ptr, inline_size);
+
+	max_size = min_t(unsigned long, PAGE_CACHE_SIZE, max_size);
+	ret = btrfs_zlib_decompress(tmp, page, extent_offset,
+				    inline_size, max_size);
+	if (ret) {
+		char *kaddr = kmap_atomic(page, KM_USER0);
+		unsigned long copy_size = min_t(u64,
+				  PAGE_CACHE_SIZE - pg_offset,
+				  max_size - extent_offset);
+		memset(kaddr + pg_offset, 0, copy_size);
+		kunmap_atomic(kaddr, KM_USER0);
+	}
+	kfree(tmp);
+	return 0;
+}
+
+/*
+ * a bit scary, this does extent mapping from logical file offset to the disk.
+ * the ugly parts come from merging extents from the disk with the in-ram
+ * representation.  This gets more complex because of the data=ordered code,
+ * where the in-ram extents might be locked pending data=ordered completion.
+ *
+ * This also copies inline extents directly into the page.
+ */
+
+struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
+				    size_t pg_offset, u64 start, u64 len,
+				    int create)
+{
+	int ret;
+	int err = 0;
+	u64 bytenr;
+	u64 extent_start = 0;
+	u64 extent_end = 0;
+	u64 objectid = inode->i_ino;
+	u32 found_type;
+	struct btrfs_path *path = NULL;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_file_extent_item *item;
+	struct extent_buffer *leaf;
+	struct btrfs_key found_key;
+	struct extent_map *em = NULL;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct btrfs_trans_handle *trans = NULL;
+	int compressed;
+
+again:
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, start, len);
+	if (em)
+		em->bdev = root->fs_info->fs_devices->latest_bdev;
+	spin_unlock(&em_tree->lock);
+
+	if (em) {
+		if (em->start > start || em->start + em->len <= start)
+			free_extent_map(em);
+		else if (em->block_start == EXTENT_MAP_INLINE && page)
+			free_extent_map(em);
+		else
+			goto out;
+	}
+	em = alloc_extent_map(GFP_NOFS);
+	if (!em) {
+		err = -ENOMEM;
+		goto out;
+	}
+	em->bdev = root->fs_info->fs_devices->latest_bdev;
+	em->start = EXTENT_MAP_HOLE;
+	em->orig_start = EXTENT_MAP_HOLE;
+	em->len = (u64)-1;
+	em->block_len = (u64)-1;
+
+	if (!path) {
+		path = btrfs_alloc_path();
+		BUG_ON(!path);
+	}
+
+	ret = btrfs_lookup_file_extent(trans, root, path,
+				       objectid, start, trans != NULL);
+	if (ret < 0) {
+		err = ret;
+		goto out;
+	}
+
+	if (ret != 0) {
+		if (path->slots[0] == 0)
+			goto not_found;
+		path->slots[0]--;
+	}
+
+	leaf = path->nodes[0];
+	item = btrfs_item_ptr(leaf, path->slots[0],
+			      struct btrfs_file_extent_item);
+	/* are we inside the extent that was found? */
+	btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+	found_type = btrfs_key_type(&found_key);
+	if (found_key.objectid != objectid ||
+	    found_type != BTRFS_EXTENT_DATA_KEY) {
+		goto not_found;
+	}
+
+	found_type = btrfs_file_extent_type(leaf, item);
+	extent_start = found_key.offset;
+	compressed = btrfs_file_extent_compression(leaf, item);
+	if (found_type == BTRFS_FILE_EXTENT_REG ||
+	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
+		extent_end = extent_start +
+		       btrfs_file_extent_num_bytes(leaf, item);
+	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+		size_t size;
+		size = btrfs_file_extent_inline_len(leaf, item);
+		extent_end = (extent_start + size + root->sectorsize - 1) &
+			~((u64)root->sectorsize - 1);
+	}
+
+	if (start >= extent_end) {
+		path->slots[0]++;
+		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0) {
+				err = ret;
+				goto out;
+			}
+			if (ret > 0)
+				goto not_found;
+			leaf = path->nodes[0];
+		}
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		if (found_key.objectid != objectid ||
+		    found_key.type != BTRFS_EXTENT_DATA_KEY)
+			goto not_found;
+		if (start + len <= found_key.offset)
+			goto not_found;
+		em->start = start;
+		em->len = found_key.offset - start;
+		goto not_found_em;
+	}
+
+	if (found_type == BTRFS_FILE_EXTENT_REG ||
+	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
+		em->start = extent_start;
+		em->len = extent_end - extent_start;
+		em->orig_start = extent_start -
+				 btrfs_file_extent_offset(leaf, item);
+		bytenr = btrfs_file_extent_disk_bytenr(leaf, item);
+		if (bytenr == 0) {
+			em->block_start = EXTENT_MAP_HOLE;
+			goto insert;
+		}
+		if (compressed) {
+			set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+			em->block_start = bytenr;
+			em->block_len = btrfs_file_extent_disk_num_bytes(leaf,
+									 item);
+		} else {
+			bytenr += btrfs_file_extent_offset(leaf, item);
+			em->block_start = bytenr;
+			em->block_len = em->len;
+			if (found_type == BTRFS_FILE_EXTENT_PREALLOC)
+				set_bit(EXTENT_FLAG_PREALLOC, &em->flags);
+		}
+		goto insert;
+	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+		unsigned long ptr;
+		char *map;
+		size_t size;
+		size_t extent_offset;
+		size_t copy_size;
+
+		em->block_start = EXTENT_MAP_INLINE;
+		if (!page || create) {
+			em->start = extent_start;
+			em->len = extent_end - extent_start;
+			goto out;
+		}
+
+		size = btrfs_file_extent_inline_len(leaf, item);
+		extent_offset = page_offset(page) + pg_offset - extent_start;
+		copy_size = min_t(u64, PAGE_CACHE_SIZE - pg_offset,
+				size - extent_offset);
+		em->start = extent_start + extent_offset;
+		em->len = (copy_size + root->sectorsize - 1) &
+			~((u64)root->sectorsize - 1);
+		em->orig_start = EXTENT_MAP_INLINE;
+		if (compressed)
+			set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
+		ptr = btrfs_file_extent_inline_start(item) + extent_offset;
+		if (create == 0 && !PageUptodate(page)) {
+			if (btrfs_file_extent_compression(leaf, item) ==
+			    BTRFS_COMPRESS_ZLIB) {
+				ret = uncompress_inline(path, inode, page,
+							pg_offset,
+							extent_offset, item);
+				BUG_ON(ret);
+			} else {
+				map = kmap(page);
+				read_extent_buffer(leaf, map + pg_offset, ptr,
+						   copy_size);
+				kunmap(page);
+			}
+			flush_dcache_page(page);
+		} else if (create && PageUptodate(page)) {
+			if (!trans) {
+				kunmap(page);
+				free_extent_map(em);
+				em = NULL;
+				btrfs_release_path(root, path);
+				trans = btrfs_join_transaction(root, 1);
+				goto again;
+			}
+			map = kmap(page);
+			write_extent_buffer(leaf, map + pg_offset, ptr,
+					    copy_size);
+			kunmap(page);
+			btrfs_mark_buffer_dirty(leaf);
+		}
+		set_extent_uptodate(io_tree, em->start,
+				    extent_map_end(em) - 1, GFP_NOFS);
+		goto insert;
+	} else {
+		printk(KERN_ERR "btrfs unknown found_type %d\n", found_type);
+		WARN_ON(1);
+	}
+not_found:
+	em->start = start;
+	em->len = len;
+not_found_em:
+	em->block_start = EXTENT_MAP_HOLE;
+	set_bit(EXTENT_FLAG_VACANCY, &em->flags);
+insert:
+	btrfs_release_path(root, path);
+	if (em->start > start || extent_map_end(em) <= start) {
+		printk(KERN_ERR "Btrfs: bad extent! em: [%llu %llu] passed "
+		       "[%llu %llu]\n", (unsigned long long)em->start,
+		       (unsigned long long)em->len,
+		       (unsigned long long)start,
+		       (unsigned long long)len);
+		err = -EIO;
+		goto out;
+	}
+
+	err = 0;
+	spin_lock(&em_tree->lock);
+	ret = add_extent_mapping(em_tree, em);
+	/* it is possible that someone inserted the extent into the tree
+	 * while we had the lock dropped.  It is also possible that
+	 * an overlapping map exists in the tree
+	 */
+	if (ret == -EEXIST) {
+		struct extent_map *existing;
+
+		ret = 0;
+
+		existing = lookup_extent_mapping(em_tree, start, len);
+		if (existing && (existing->start > start ||
+		    existing->start + existing->len <= start)) {
+			free_extent_map(existing);
+			existing = NULL;
+		}
+		if (!existing) {
+			existing = lookup_extent_mapping(em_tree, em->start,
+							 em->len);
+			if (existing) {
+				err = merge_extent_mapping(em_tree, existing,
+							   em, start,
+							   root->sectorsize);
+				free_extent_map(existing);
+				if (err) {
+					free_extent_map(em);
+					em = NULL;
+				}
+			} else {
+				err = -EIO;
+				free_extent_map(em);
+				em = NULL;
+			}
+		} else {
+			free_extent_map(em);
+			em = existing;
+			err = 0;
+		}
+	}
+	spin_unlock(&em_tree->lock);
+out:
+	if (path)
+		btrfs_free_path(path);
+	if (trans) {
+		ret = btrfs_end_transaction(trans, root);
+		if (!err)
+			err = ret;
+	}
+	if (err) {
+		free_extent_map(em);
+		WARN_ON(1);
+		return ERR_PTR(err);
+	}
+	return em;
+}
+
+static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
+			const struct iovec *iov, loff_t offset,
+			unsigned long nr_segs)
+{
+	return -EINVAL;
+}
+
+static sector_t btrfs_bmap(struct address_space *mapping, sector_t iblock)
+{
+	return extent_bmap(mapping, iblock, btrfs_get_extent);
+}
+
+int btrfs_readpage(struct file *file, struct page *page)
+{
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	return extent_read_full_page(tree, page, btrfs_get_extent);
+}
+
+static int btrfs_writepage(struct page *page, struct writeback_control *wbc)
+{
+	struct extent_io_tree *tree;
+
+
+	if (current->flags & PF_MEMALLOC) {
+		redirty_page_for_writepage(wbc, page);
+		unlock_page(page);
+		return 0;
+	}
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	return extent_write_full_page(tree, page, btrfs_get_extent, wbc);
+}
+
+int btrfs_writepages(struct address_space *mapping,
+		     struct writeback_control *wbc)
+{
+	struct extent_io_tree *tree;
+
+	tree = &BTRFS_I(mapping->host)->io_tree;
+	return extent_writepages(tree, mapping, btrfs_get_extent, wbc);
+}
+
+static int
+btrfs_readpages(struct file *file, struct address_space *mapping,
+		struct list_head *pages, unsigned nr_pages)
+{
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(mapping->host)->io_tree;
+	return extent_readpages(tree, mapping, pages, nr_pages,
+				btrfs_get_extent);
+}
+static int __btrfs_releasepage(struct page *page, gfp_t gfp_flags)
+{
+	struct extent_io_tree *tree;
+	struct extent_map_tree *map;
+	int ret;
+
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	map = &BTRFS_I(page->mapping->host)->extent_tree;
+	ret = try_release_extent_mapping(map, tree, page, gfp_flags);
+	if (ret == 1) {
+		ClearPagePrivate(page);
+		set_page_private(page, 0);
+		page_cache_release(page);
+	}
+	return ret;
+}
+
+static int btrfs_releasepage(struct page *page, gfp_t gfp_flags)
+{
+	if (PageWriteback(page) || PageDirty(page))
+		return 0;
+	return __btrfs_releasepage(page, gfp_flags);
+}
+
+static void btrfs_invalidatepage(struct page *page, unsigned long offset)
+{
+	struct extent_io_tree *tree;
+	struct btrfs_ordered_extent *ordered;
+	u64 page_start = page_offset(page);
+	u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
+
+	wait_on_page_writeback(page);
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	if (offset) {
+		btrfs_releasepage(page, GFP_NOFS);
+		return;
+	}
+
+	lock_extent(tree, page_start, page_end, GFP_NOFS);
+	ordered = btrfs_lookup_ordered_extent(page->mapping->host,
+					   page_offset(page));
+	if (ordered) {
+		/*
+		 * IO on this page will never be started, so we need
+		 * to account for any ordered extents now
+		 */
+		clear_extent_bit(tree, page_start, page_end,
+				 EXTENT_DIRTY | EXTENT_DELALLOC |
+				 EXTENT_LOCKED, 1, 0, GFP_NOFS);
+		btrfs_finish_ordered_io(page->mapping->host,
+					page_start, page_end);
+		btrfs_put_ordered_extent(ordered);
+		lock_extent(tree, page_start, page_end, GFP_NOFS);
+	}
+	clear_extent_bit(tree, page_start, page_end,
+		 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
+		 EXTENT_ORDERED,
+		 1, 1, GFP_NOFS);
+	__btrfs_releasepage(page, GFP_NOFS);
+
+	ClearPageChecked(page);
+	if (PagePrivate(page)) {
+		ClearPagePrivate(page);
+		set_page_private(page, 0);
+		page_cache_release(page);
+	}
+}
+
+/*
+ * btrfs_page_mkwrite() is not allowed to change the file size as it gets
+ * called from a page fault handler when a page is first dirtied. Hence we must
+ * be careful to check for EOF conditions here. We set the page up correctly
+ * for a written page which means we get ENOSPC checking when writing into
+ * holes and correct delalloc and unwritten extent mapping on filesystems that
+ * support these features.
+ *
+ * We are not allowed to take the i_mutex here so we have to play games to
+ * protect against truncate races as the page could now be beyond EOF.  Because
+ * vmtruncate() writes the inode size before removing pages, once we have the
+ * page lock we can determine safely if the page is beyond EOF. If it is not
+ * beyond EOF, then the page is guaranteed safe against truncation until we
+ * unlock the page.
+ */
+int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page)
+{
+	struct inode *inode = fdentry(vma->vm_file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct btrfs_ordered_extent *ordered;
+	char *kaddr;
+	unsigned long zero_start;
+	loff_t size;
+	int ret;
+	u64 page_start;
+	u64 page_end;
+
+	ret = btrfs_check_free_space(root, PAGE_CACHE_SIZE, 0);
+	if (ret)
+		goto out;
+
+	ret = -EINVAL;
+again:
+	lock_page(page);
+	size = i_size_read(inode);
+	page_start = page_offset(page);
+	page_end = page_start + PAGE_CACHE_SIZE - 1;
+
+	if ((page->mapping != inode->i_mapping) ||
+	    (page_start >= size)) {
+		/* page got truncated out from underneath us */
+		goto out_unlock;
+	}
+	wait_on_page_writeback(page);
+
+	lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+	set_page_extent_mapped(page);
+
+	/*
+	 * we can't set the delalloc bits if there are pending ordered
+	 * extents.  Drop our locks and wait for them to finish
+	 */
+	ordered = btrfs_lookup_ordered_extent(inode, page_start);
+	if (ordered) {
+		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+		unlock_page(page);
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		btrfs_put_ordered_extent(ordered);
+		goto again;
+	}
+
+	btrfs_set_extent_delalloc(inode, page_start, page_end);
+	ret = 0;
+
+	/* page is wholly or partially inside EOF */
+	if (page_start + PAGE_CACHE_SIZE > size)
+		zero_start = size & ~PAGE_CACHE_MASK;
+	else
+		zero_start = PAGE_CACHE_SIZE;
+
+	if (zero_start != PAGE_CACHE_SIZE) {
+		kaddr = kmap(page);
+		memset(kaddr + zero_start, 0, PAGE_CACHE_SIZE - zero_start);
+		flush_dcache_page(page);
+		kunmap(page);
+	}
+	ClearPageChecked(page);
+	set_page_dirty(page);
+	unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+
+out_unlock:
+	unlock_page(page);
+out:
+	return ret;
+}
+
+static void btrfs_truncate(struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret;
+	struct btrfs_trans_handle *trans;
+	unsigned long nr;
+	u64 mask = root->sectorsize - 1;
+
+	if (!S_ISREG(inode->i_mode))
+		return;
+	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+		return;
+
+	btrfs_truncate_page(inode->i_mapping, inode->i_size);
+	btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, inode);
+	btrfs_i_size_write(inode, inode->i_size);
+
+	ret = btrfs_orphan_add(trans, inode);
+	if (ret)
+		goto out;
+	/* FIXME, add redo link to tree so we don't leak on crash */
+	ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size,
+				      BTRFS_EXTENT_DATA_KEY);
+	btrfs_update_inode(trans, root, inode);
+
+	ret = btrfs_orphan_del(trans, inode);
+	BUG_ON(ret);
+
+out:
+	nr = trans->blocks_used;
+	ret = btrfs_end_transaction_throttle(trans, root);
+	BUG_ON(ret);
+	btrfs_btree_balance_dirty(root, nr);
+}
+
+/*
+ * create a new subvolume directory/inode (helper for the ioctl).
+ */
+int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *new_root, struct dentry *dentry,
+			     u64 new_dirid, u64 alloc_hint)
+{
+	struct inode *inode;
+	int error;
+	u64 index = 0;
+
+	inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, new_dirid,
+				new_dirid, alloc_hint, S_IFDIR | 0700, &index);
+	if (IS_ERR(inode))
+		return PTR_ERR(inode);
+	inode->i_op = &btrfs_dir_inode_operations;
+	inode->i_fop = &btrfs_dir_file_operations;
+
+	inode->i_nlink = 1;
+	btrfs_i_size_write(inode, 0);
+
+	error = btrfs_update_inode(trans, new_root, inode);
+	if (error)
+		return error;
+
+	d_instantiate(dentry, inode);
+	return 0;
+}
+
+/* helper function for file defrag and space balancing.  This
+ * forces readahead on a given range of bytes in an inode
+ */
+unsigned long btrfs_force_ra(struct address_space *mapping,
+			      struct file_ra_state *ra, struct file *file,
+			      pgoff_t offset, pgoff_t last_index)
+{
+	pgoff_t req_size = last_index - offset + 1;
+
+	page_cache_sync_readahead(mapping, ra, file, offset, req_size);
+	return offset + req_size;
+}
+
+struct inode *btrfs_alloc_inode(struct super_block *sb)
+{
+	struct btrfs_inode *ei;
+
+	ei = kmem_cache_alloc(btrfs_inode_cachep, GFP_NOFS);
+	if (!ei)
+		return NULL;
+	ei->last_trans = 0;
+	ei->logged_trans = 0;
+	btrfs_ordered_inode_tree_init(&ei->ordered_tree);
+	ei->i_acl = BTRFS_ACL_NOT_CACHED;
+	ei->i_default_acl = BTRFS_ACL_NOT_CACHED;
+	INIT_LIST_HEAD(&ei->i_orphan);
+	return &ei->vfs_inode;
+}
+
+void btrfs_destroy_inode(struct inode *inode)
+{
+	struct btrfs_ordered_extent *ordered;
+	WARN_ON(!list_empty(&inode->i_dentry));
+	WARN_ON(inode->i_data.nrpages);
+
+	if (BTRFS_I(inode)->i_acl &&
+	    BTRFS_I(inode)->i_acl != BTRFS_ACL_NOT_CACHED)
+		posix_acl_release(BTRFS_I(inode)->i_acl);
+	if (BTRFS_I(inode)->i_default_acl &&
+	    BTRFS_I(inode)->i_default_acl != BTRFS_ACL_NOT_CACHED)
+		posix_acl_release(BTRFS_I(inode)->i_default_acl);
+
+	spin_lock(&BTRFS_I(inode)->root->list_lock);
+	if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
+		printk(KERN_ERR "BTRFS: inode %lu: inode still on the orphan"
+		       " list\n", inode->i_ino);
+		dump_stack();
+	}
+	spin_unlock(&BTRFS_I(inode)->root->list_lock);
+
+	while (1) {
+		ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1);
+		if (!ordered)
+			break;
+		else {
+			printk(KERN_ERR "btrfs found ordered "
+			       "extent %llu %llu on inode cleanup\n",
+			       (unsigned long long)ordered->file_offset,
+			       (unsigned long long)ordered->len);
+			btrfs_remove_ordered_extent(inode, ordered);
+			btrfs_put_ordered_extent(ordered);
+			btrfs_put_ordered_extent(ordered);
+		}
+	}
+	btrfs_drop_extent_cache(inode, 0, (u64)-1, 0);
+	kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
+}
+
+static void init_once(void *foo)
+{
+	struct btrfs_inode *ei = (struct btrfs_inode *) foo;
+
+	inode_init_once(&ei->vfs_inode);
+}
+
+void btrfs_destroy_cachep(void)
+{
+	if (btrfs_inode_cachep)
+		kmem_cache_destroy(btrfs_inode_cachep);
+	if (btrfs_trans_handle_cachep)
+		kmem_cache_destroy(btrfs_trans_handle_cachep);
+	if (btrfs_transaction_cachep)
+		kmem_cache_destroy(btrfs_transaction_cachep);
+	if (btrfs_bit_radix_cachep)
+		kmem_cache_destroy(btrfs_bit_radix_cachep);
+	if (btrfs_path_cachep)
+		kmem_cache_destroy(btrfs_path_cachep);
+}
+
+struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
+				       unsigned long extra_flags,
+				       void (*ctor)(void *))
+{
+	return kmem_cache_create(name, size, 0, (SLAB_RECLAIM_ACCOUNT |
+				 SLAB_MEM_SPREAD | extra_flags), ctor);
+}
+
+int btrfs_init_cachep(void)
+{
+	btrfs_inode_cachep = btrfs_cache_create("btrfs_inode_cache",
+					  sizeof(struct btrfs_inode),
+					  0, init_once);
+	if (!btrfs_inode_cachep)
+		goto fail;
+	btrfs_trans_handle_cachep =
+			btrfs_cache_create("btrfs_trans_handle_cache",
+					   sizeof(struct btrfs_trans_handle),
+					   0, NULL);
+	if (!btrfs_trans_handle_cachep)
+		goto fail;
+	btrfs_transaction_cachep = btrfs_cache_create("btrfs_transaction_cache",
+					     sizeof(struct btrfs_transaction),
+					     0, NULL);
+	if (!btrfs_transaction_cachep)
+		goto fail;
+	btrfs_path_cachep = btrfs_cache_create("btrfs_path_cache",
+					 sizeof(struct btrfs_path),
+					 0, NULL);
+	if (!btrfs_path_cachep)
+		goto fail;
+	btrfs_bit_radix_cachep = btrfs_cache_create("btrfs_radix", 256,
+					      SLAB_DESTROY_BY_RCU, NULL);
+	if (!btrfs_bit_radix_cachep)
+		goto fail;
+	return 0;
+fail:
+	btrfs_destroy_cachep();
+	return -ENOMEM;
+}
+
+static int btrfs_getattr(struct vfsmount *mnt,
+			 struct dentry *dentry, struct kstat *stat)
+{
+	struct inode *inode = dentry->d_inode;
+	generic_fillattr(inode, stat);
+	stat->dev = BTRFS_I(inode)->root->anon_super.s_dev;
+	stat->blksize = PAGE_CACHE_SIZE;
+	stat->blocks = (inode_get_bytes(inode) +
+			BTRFS_I(inode)->delalloc_bytes) >> 9;
+	return 0;
+}
+
+static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry,
+			   struct inode *new_dir, struct dentry *new_dentry)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(old_dir)->root;
+	struct inode *new_inode = new_dentry->d_inode;
+	struct inode *old_inode = old_dentry->d_inode;
+	struct timespec ctime = CURRENT_TIME;
+	u64 index = 0;
+	int ret;
+
+	/* we're not allowed to rename between subvolumes */
+	if (BTRFS_I(old_inode)->root->root_key.objectid !=
+	    BTRFS_I(new_dir)->root->root_key.objectid)
+		return -EXDEV;
+
+	if (S_ISDIR(old_inode->i_mode) && new_inode &&
+	    new_inode->i_size > BTRFS_EMPTY_DIR_SIZE) {
+		return -ENOTEMPTY;
+	}
+
+	/* to rename a snapshot or subvolume, we need to juggle the
+	 * backrefs.  This isn't coded yet
+	 */
+	if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
+		return -EXDEV;
+
+	ret = btrfs_check_free_space(root, 1, 0);
+	if (ret)
+		goto out_unlock;
+
+	trans = btrfs_start_transaction(root, 1);
+
+	btrfs_set_trans_block_group(trans, new_dir);
+
+	btrfs_inc_nlink(old_dentry->d_inode);
+	old_dir->i_ctime = old_dir->i_mtime = ctime;
+	new_dir->i_ctime = new_dir->i_mtime = ctime;
+	old_inode->i_ctime = ctime;
+
+	ret = btrfs_unlink_inode(trans, root, old_dir, old_dentry->d_inode,
+				 old_dentry->d_name.name,
+				 old_dentry->d_name.len);
+	if (ret)
+		goto out_fail;
+
+	if (new_inode) {
+		new_inode->i_ctime = CURRENT_TIME;
+		ret = btrfs_unlink_inode(trans, root, new_dir,
+					 new_dentry->d_inode,
+					 new_dentry->d_name.name,
+					 new_dentry->d_name.len);
+		if (ret)
+			goto out_fail;
+		if (new_inode->i_nlink == 0) {
+			ret = btrfs_orphan_add(trans, new_dentry->d_inode);
+			if (ret)
+				goto out_fail;
+		}
+
+	}
+	ret = btrfs_set_inode_index(new_dir, &index);
+	if (ret)
+		goto out_fail;
+
+	ret = btrfs_add_link(trans, new_dentry->d_parent->d_inode,
+			     old_inode, new_dentry->d_name.name,
+			     new_dentry->d_name.len, 1, index);
+	if (ret)
+		goto out_fail;
+
+out_fail:
+	btrfs_end_transaction_throttle(trans, root);
+out_unlock:
+	return ret;
+}
+
+/*
+ * some fairly slow code that needs optimization. This walks the list
+ * of all the inodes with pending delalloc and forces them to disk.
+ */
+int btrfs_start_delalloc_inodes(struct btrfs_root *root)
+{
+	struct list_head *head = &root->fs_info->delalloc_inodes;
+	struct btrfs_inode *binode;
+	struct inode *inode;
+
+	if (root->fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	spin_lock(&root->fs_info->delalloc_lock);
+	while (!list_empty(head)) {
+		binode = list_entry(head->next, struct btrfs_inode,
+				    delalloc_inodes);
+		inode = igrab(&binode->vfs_inode);
+		if (!inode)
+			list_del_init(&binode->delalloc_inodes);
+		spin_unlock(&root->fs_info->delalloc_lock);
+		if (inode) {
+			filemap_flush(inode->i_mapping);
+			iput(inode);
+		}
+		cond_resched();
+		spin_lock(&root->fs_info->delalloc_lock);
+	}
+	spin_unlock(&root->fs_info->delalloc_lock);
+
+	/* the filemap_flush will queue IO into the worker threads, but
+	 * we have to make sure the IO is actually started and that
+	 * ordered extents get created before we return
+	 */
+	atomic_inc(&root->fs_info->async_submit_draining);
+	while (atomic_read(&root->fs_info->nr_async_submits) ||
+	      atomic_read(&root->fs_info->async_delalloc_pages)) {
+		wait_event(root->fs_info->async_submit_wait,
+		   (atomic_read(&root->fs_info->nr_async_submits) == 0 &&
+		    atomic_read(&root->fs_info->async_delalloc_pages) == 0));
+	}
+	atomic_dec(&root->fs_info->async_submit_draining);
+	return 0;
+}
+
+static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
+			 const char *symname)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(dir)->root;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct inode *inode = NULL;
+	int err;
+	int drop_inode = 0;
+	u64 objectid;
+	u64 index = 0 ;
+	int name_len;
+	int datasize;
+	unsigned long ptr;
+	struct btrfs_file_extent_item *ei;
+	struct extent_buffer *leaf;
+	unsigned long nr = 0;
+
+	name_len = strlen(symname) + 1;
+	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
+		return -ENAMETOOLONG;
+
+	err = btrfs_check_free_space(root, 1, 0);
+	if (err)
+		goto out_fail;
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, dir);
+
+	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
+	if (err) {
+		err = -ENOSPC;
+		goto out_unlock;
+	}
+
+	inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
+				dentry->d_name.len,
+				dentry->d_parent->d_inode->i_ino, objectid,
+				BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO,
+				&index);
+	err = PTR_ERR(inode);
+	if (IS_ERR(inode))
+		goto out_unlock;
+
+	err = btrfs_init_acl(inode, dir);
+	if (err) {
+		drop_inode = 1;
+		goto out_unlock;
+	}
+
+	btrfs_set_trans_block_group(trans, inode);
+	err = btrfs_add_nondir(trans, dentry, inode, 0, index);
+	if (err)
+		drop_inode = 1;
+	else {
+		inode->i_mapping->a_ops = &btrfs_aops;
+		inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
+		inode->i_fop = &btrfs_file_operations;
+		inode->i_op = &btrfs_file_inode_operations;
+		BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
+	}
+	dir->i_sb->s_dirt = 1;
+	btrfs_update_inode_block_group(trans, inode);
+	btrfs_update_inode_block_group(trans, dir);
+	if (drop_inode)
+		goto out_unlock;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	key.objectid = inode->i_ino;
+	key.offset = 0;
+	btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
+	datasize = btrfs_file_extent_calc_inline_size(name_len);
+	err = btrfs_insert_empty_item(trans, root, path, &key,
+				      datasize);
+	if (err) {
+		drop_inode = 1;
+		goto out_unlock;
+	}
+	leaf = path->nodes[0];
+	ei = btrfs_item_ptr(leaf, path->slots[0],
+			    struct btrfs_file_extent_item);
+	btrfs_set_file_extent_generation(leaf, ei, trans->transid);
+	btrfs_set_file_extent_type(leaf, ei,
+				   BTRFS_FILE_EXTENT_INLINE);
+	btrfs_set_file_extent_encryption(leaf, ei, 0);
+	btrfs_set_file_extent_compression(leaf, ei, 0);
+	btrfs_set_file_extent_other_encoding(leaf, ei, 0);
+	btrfs_set_file_extent_ram_bytes(leaf, ei, name_len);
+
+	ptr = btrfs_file_extent_inline_start(ei);
+	write_extent_buffer(leaf, symname, ptr, name_len);
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_free_path(path);
+
+	inode->i_op = &btrfs_symlink_inode_operations;
+	inode->i_mapping->a_ops = &btrfs_symlink_aops;
+	inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
+	inode_set_bytes(inode, name_len);
+	btrfs_i_size_write(inode, name_len - 1);
+	err = btrfs_update_inode(trans, root, inode);
+	if (err)
+		drop_inode = 1;
+
+out_unlock:
+	nr = trans->blocks_used;
+	btrfs_end_transaction_throttle(trans, root);
+out_fail:
+	if (drop_inode) {
+		inode_dec_link_count(inode);
+		iput(inode);
+	}
+	btrfs_btree_balance_dirty(root, nr);
+	return err;
+}
+
+static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
+			       u64 alloc_hint, int mode)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_key ins;
+	u64 alloc_size;
+	u64 cur_offset = start;
+	u64 num_bytes = end - start;
+	int ret = 0;
+
+	trans = btrfs_join_transaction(root, 1);
+	BUG_ON(!trans);
+	btrfs_set_trans_block_group(trans, inode);
+
+	while (num_bytes > 0) {
+		alloc_size = min(num_bytes, root->fs_info->max_extent);
+		ret = btrfs_reserve_extent(trans, root, alloc_size,
+					   root->sectorsize, 0, alloc_hint,
+					   (u64)-1, &ins, 1);
+		if (ret) {
+			WARN_ON(1);
+			goto out;
+		}
+		ret = insert_reserved_file_extent(trans, inode,
+						  cur_offset, ins.objectid,
+						  ins.offset, ins.offset,
+						  ins.offset, 0, 0, 0,
+						  BTRFS_FILE_EXTENT_PREALLOC);
+		BUG_ON(ret);
+		num_bytes -= ins.offset;
+		cur_offset += ins.offset;
+		alloc_hint = ins.objectid + ins.offset;
+	}
+out:
+	if (cur_offset > start) {
+		inode->i_ctime = CURRENT_TIME;
+		btrfs_set_flag(inode, PREALLOC);
+		if (!(mode & FALLOC_FL_KEEP_SIZE) &&
+		    cur_offset > i_size_read(inode))
+			btrfs_i_size_write(inode, cur_offset);
+		ret = btrfs_update_inode(trans, root, inode);
+		BUG_ON(ret);
+	}
+
+	btrfs_end_transaction(trans, root);
+	return ret;
+}
+
+static long btrfs_fallocate(struct inode *inode, int mode,
+			    loff_t offset, loff_t len)
+{
+	u64 cur_offset;
+	u64 last_byte;
+	u64 alloc_start;
+	u64 alloc_end;
+	u64 alloc_hint = 0;
+	u64 mask = BTRFS_I(inode)->root->sectorsize - 1;
+	struct extent_map *em;
+	int ret;
+
+	alloc_start = offset & ~mask;
+	alloc_end =  (offset + len + mask) & ~mask;
+
+	mutex_lock(&inode->i_mutex);
+	if (alloc_start > inode->i_size) {
+		ret = btrfs_cont_expand(inode, alloc_start);
+		if (ret)
+			goto out;
+	}
+
+	while (1) {
+		struct btrfs_ordered_extent *ordered;
+		lock_extent(&BTRFS_I(inode)->io_tree, alloc_start,
+			    alloc_end - 1, GFP_NOFS);
+		ordered = btrfs_lookup_first_ordered_extent(inode,
+							    alloc_end - 1);
+		if (ordered &&
+		    ordered->file_offset + ordered->len > alloc_start &&
+		    ordered->file_offset < alloc_end) {
+			btrfs_put_ordered_extent(ordered);
+			unlock_extent(&BTRFS_I(inode)->io_tree,
+				      alloc_start, alloc_end - 1, GFP_NOFS);
+			btrfs_wait_ordered_range(inode, alloc_start,
+						 alloc_end - alloc_start);
+		} else {
+			if (ordered)
+				btrfs_put_ordered_extent(ordered);
+			break;
+		}
+	}
+
+	cur_offset = alloc_start;
+	while (1) {
+		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
+				      alloc_end - cur_offset, 0);
+		BUG_ON(IS_ERR(em) || !em);
+		last_byte = min(extent_map_end(em), alloc_end);
+		last_byte = (last_byte + mask) & ~mask;
+		if (em->block_start == EXTENT_MAP_HOLE) {
+			ret = prealloc_file_range(inode, cur_offset,
+					last_byte, alloc_hint, mode);
+			if (ret < 0) {
+				free_extent_map(em);
+				break;
+			}
+		}
+		if (em->block_start <= EXTENT_MAP_LAST_BYTE)
+			alloc_hint = em->block_start;
+		free_extent_map(em);
+
+		cur_offset = last_byte;
+		if (cur_offset >= alloc_end) {
+			ret = 0;
+			break;
+		}
+	}
+	unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, alloc_end - 1,
+		      GFP_NOFS);
+out:
+	mutex_unlock(&inode->i_mutex);
+	return ret;
+}
+
+static int btrfs_set_page_dirty(struct page *page)
+{
+	return __set_page_dirty_nobuffers(page);
+}
+
+static int btrfs_permission(struct inode *inode, int mask)
+{
+	if (btrfs_test_flag(inode, READONLY) && (mask & MAY_WRITE))
+		return -EACCES;
+	return generic_permission(inode, mask, btrfs_check_acl);
+}
+
+static struct inode_operations btrfs_dir_inode_operations = {
+	.getattr	= btrfs_getattr,
+	.lookup		= btrfs_lookup,
+	.create		= btrfs_create,
+	.unlink		= btrfs_unlink,
+	.link		= btrfs_link,
+	.mkdir		= btrfs_mkdir,
+	.rmdir		= btrfs_rmdir,
+	.rename		= btrfs_rename,
+	.symlink	= btrfs_symlink,
+	.setattr	= btrfs_setattr,
+	.mknod		= btrfs_mknod,
+	.setxattr	= btrfs_setxattr,
+	.getxattr	= btrfs_getxattr,
+	.listxattr	= btrfs_listxattr,
+	.removexattr	= btrfs_removexattr,
+	.permission	= btrfs_permission,
+};
+static struct inode_operations btrfs_dir_ro_inode_operations = {
+	.lookup		= btrfs_lookup,
+	.permission	= btrfs_permission,
+};
+static struct file_operations btrfs_dir_file_operations = {
+	.llseek		= generic_file_llseek,
+	.read		= generic_read_dir,
+	.readdir	= btrfs_real_readdir,
+	.unlocked_ioctl	= btrfs_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= btrfs_ioctl,
+#endif
+	.release        = btrfs_release_file,
+	.fsync		= btrfs_sync_file,
+};
+
+static struct extent_io_ops btrfs_extent_io_ops = {
+	.fill_delalloc = run_delalloc_range,
+	.submit_bio_hook = btrfs_submit_bio_hook,
+	.merge_bio_hook = btrfs_merge_bio_hook,
+	.readpage_end_io_hook = btrfs_readpage_end_io_hook,
+	.writepage_end_io_hook = btrfs_writepage_end_io_hook,
+	.writepage_start_hook = btrfs_writepage_start_hook,
+	.readpage_io_failed_hook = btrfs_io_failed_hook,
+	.set_bit_hook = btrfs_set_bit_hook,
+	.clear_bit_hook = btrfs_clear_bit_hook,
+};
+
+static struct address_space_operations btrfs_aops = {
+	.readpage	= btrfs_readpage,
+	.writepage	= btrfs_writepage,
+	.writepages	= btrfs_writepages,
+	.readpages	= btrfs_readpages,
+	.sync_page	= block_sync_page,
+	.bmap		= btrfs_bmap,
+	.direct_IO	= btrfs_direct_IO,
+	.invalidatepage = btrfs_invalidatepage,
+	.releasepage	= btrfs_releasepage,
+	.set_page_dirty	= btrfs_set_page_dirty,
+};
+
+static struct address_space_operations btrfs_symlink_aops = {
+	.readpage	= btrfs_readpage,
+	.writepage	= btrfs_writepage,
+	.invalidatepage = btrfs_invalidatepage,
+	.releasepage	= btrfs_releasepage,
+};
+
+static struct inode_operations btrfs_file_inode_operations = {
+	.truncate	= btrfs_truncate,
+	.getattr	= btrfs_getattr,
+	.setattr	= btrfs_setattr,
+	.setxattr	= btrfs_setxattr,
+	.getxattr	= btrfs_getxattr,
+	.listxattr      = btrfs_listxattr,
+	.removexattr	= btrfs_removexattr,
+	.permission	= btrfs_permission,
+	.fallocate	= btrfs_fallocate,
+};
+static struct inode_operations btrfs_special_inode_operations = {
+	.getattr	= btrfs_getattr,
+	.setattr	= btrfs_setattr,
+	.permission	= btrfs_permission,
+	.setxattr	= btrfs_setxattr,
+	.getxattr	= btrfs_getxattr,
+	.listxattr	= btrfs_listxattr,
+	.removexattr	= btrfs_removexattr,
+};
+static struct inode_operations btrfs_symlink_inode_operations = {
+	.readlink	= generic_readlink,
+	.follow_link	= page_follow_link_light,
+	.put_link	= page_put_link,
+	.permission	= btrfs_permission,
+};
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
new file mode 100644
index 00000000000..c2aa33e3feb
--- /dev/null
+++ b/fs/btrfs/ioctl.c
@@ -0,0 +1,1132 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/kernel.h>
+#include <linux/bio.h>
+#include <linux/buffer_head.h>
+#include <linux/file.h>
+#include <linux/fs.h>
+#include <linux/fsnotify.h>
+#include <linux/pagemap.h>
+#include <linux/highmem.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/smp_lock.h>
+#include <linux/backing-dev.h>
+#include <linux/mount.h>
+#include <linux/mpage.h>
+#include <linux/namei.h>
+#include <linux/swap.h>
+#include <linux/writeback.h>
+#include <linux/statfs.h>
+#include <linux/compat.h>
+#include <linux/bit_spinlock.h>
+#include <linux/security.h>
+#include <linux/version.h>
+#include <linux/xattr.h>
+#include <linux/vmalloc.h>
+#include "compat.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "ioctl.h"
+#include "print-tree.h"
+#include "volumes.h"
+#include "locking.h"
+
+
+
+static noinline int create_subvol(struct btrfs_root *root,
+				  struct dentry *dentry,
+				  char *name, int namelen)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_key key;
+	struct btrfs_root_item root_item;
+	struct btrfs_inode_item *inode_item;
+	struct extent_buffer *leaf;
+	struct btrfs_root *new_root = root;
+	struct inode *dir;
+	int ret;
+	int err;
+	u64 objectid;
+	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
+	u64 index = 0;
+	unsigned long nr = 1;
+
+	ret = btrfs_check_free_space(root, 1, 0);
+	if (ret)
+		goto fail_commit;
+
+	trans = btrfs_start_transaction(root, 1);
+	BUG_ON(!trans);
+
+	ret = btrfs_find_free_objectid(trans, root->fs_info->tree_root,
+				       0, &objectid);
+	if (ret)
+		goto fail;
+
+	leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
+				      objectid, trans->transid, 0, 0, 0);
+	if (IS_ERR(leaf)) {
+		ret = PTR_ERR(leaf);
+		goto fail;
+	}
+
+	btrfs_set_header_nritems(leaf, 0);
+	btrfs_set_header_level(leaf, 0);
+	btrfs_set_header_bytenr(leaf, leaf->start);
+	btrfs_set_header_generation(leaf, trans->transid);
+	btrfs_set_header_owner(leaf, objectid);
+
+	write_extent_buffer(leaf, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(leaf),
+			    BTRFS_FSID_SIZE);
+	btrfs_mark_buffer_dirty(leaf);
+
+	inode_item = &root_item.inode;
+	memset(inode_item, 0, sizeof(*inode_item));
+	inode_item->generation = cpu_to_le64(1);
+	inode_item->size = cpu_to_le64(3);
+	inode_item->nlink = cpu_to_le32(1);
+	inode_item->nbytes = cpu_to_le64(root->leafsize);
+	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+
+	btrfs_set_root_bytenr(&root_item, leaf->start);
+	btrfs_set_root_generation(&root_item, trans->transid);
+	btrfs_set_root_level(&root_item, 0);
+	btrfs_set_root_refs(&root_item, 1);
+	btrfs_set_root_used(&root_item, 0);
+	btrfs_set_root_last_snapshot(&root_item, 0);
+
+	memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress));
+	root_item.drop_level = 0;
+
+	btrfs_tree_unlock(leaf);
+	free_extent_buffer(leaf);
+	leaf = NULL;
+
+	btrfs_set_root_dirid(&root_item, new_dirid);
+
+	key.objectid = objectid;
+	key.offset = 1;
+	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
+				&root_item);
+	if (ret)
+		goto fail;
+
+	/*
+	 * insert the directory item
+	 */
+	key.offset = (u64)-1;
+	dir = dentry->d_parent->d_inode;
+	ret = btrfs_set_inode_index(dir, &index);
+	BUG_ON(ret);
+
+	ret = btrfs_insert_dir_item(trans, root,
+				    name, namelen, dir->i_ino, &key,
+				    BTRFS_FT_DIR, index);
+	if (ret)
+		goto fail;
+
+	btrfs_i_size_write(dir, dir->i_size + namelen * 2);
+	ret = btrfs_update_inode(trans, root, dir);
+	BUG_ON(ret);
+
+	/* add the backref first */
+	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
+				 objectid, BTRFS_ROOT_BACKREF_KEY,
+				 root->root_key.objectid,
+				 dir->i_ino, index, name, namelen);
+
+	BUG_ON(ret);
+
+	/* now add the forward ref */
+	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
+				 root->root_key.objectid, BTRFS_ROOT_REF_KEY,
+				 objectid,
+				 dir->i_ino, index, name, namelen);
+
+	BUG_ON(ret);
+
+	ret = btrfs_commit_transaction(trans, root);
+	if (ret)
+		goto fail_commit;
+
+	new_root = btrfs_read_fs_root_no_name(root->fs_info, &key);
+	BUG_ON(!new_root);
+
+	trans = btrfs_start_transaction(new_root, 1);
+	BUG_ON(!trans);
+
+	ret = btrfs_create_subvol_root(trans, new_root, dentry, new_dirid,
+				       BTRFS_I(dir)->block_group);
+	if (ret)
+		goto fail;
+
+fail:
+	nr = trans->blocks_used;
+	err = btrfs_commit_transaction(trans, new_root);
+	if (err && !ret)
+		ret = err;
+fail_commit:
+	btrfs_btree_balance_dirty(root, nr);
+	return ret;
+}
+
+static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
+			   char *name, int namelen)
+{
+	struct btrfs_pending_snapshot *pending_snapshot;
+	struct btrfs_trans_handle *trans;
+	int ret = 0;
+	int err;
+	unsigned long nr = 0;
+
+	if (!root->ref_cows)
+		return -EINVAL;
+
+	ret = btrfs_check_free_space(root, 1, 0);
+	if (ret)
+		goto fail_unlock;
+
+	pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS);
+	if (!pending_snapshot) {
+		ret = -ENOMEM;
+		goto fail_unlock;
+	}
+	pending_snapshot->name = kmalloc(namelen + 1, GFP_NOFS);
+	if (!pending_snapshot->name) {
+		ret = -ENOMEM;
+		kfree(pending_snapshot);
+		goto fail_unlock;
+	}
+	memcpy(pending_snapshot->name, name, namelen);
+	pending_snapshot->name[namelen] = '\0';
+	pending_snapshot->dentry = dentry;
+	trans = btrfs_start_transaction(root, 1);
+	BUG_ON(!trans);
+	pending_snapshot->root = root;
+	list_add(&pending_snapshot->list,
+		 &trans->transaction->pending_snapshots);
+	err = btrfs_commit_transaction(trans, root);
+
+fail_unlock:
+	btrfs_btree_balance_dirty(root, nr);
+	return ret;
+}
+
+/* copy of may_create in fs/namei.c() */
+static inline int btrfs_may_create(struct inode *dir, struct dentry *child)
+{
+	if (child->d_inode)
+		return -EEXIST;
+	if (IS_DEADDIR(dir))
+		return -ENOENT;
+	return inode_permission(dir, MAY_WRITE | MAY_EXEC);
+}
+
+/*
+ * Create a new subvolume below @parent.  This is largely modeled after
+ * sys_mkdirat and vfs_mkdir, but we only do a single component lookup
+ * inside this filesystem so it's quite a bit simpler.
+ */
+static noinline int btrfs_mksubvol(struct path *parent, char *name,
+				   int mode, int namelen,
+				   struct btrfs_root *snap_src)
+{
+	struct dentry *dentry;
+	int error;
+
+	mutex_lock_nested(&parent->dentry->d_inode->i_mutex, I_MUTEX_PARENT);
+
+	dentry = lookup_one_len(name, parent->dentry, namelen);
+	error = PTR_ERR(dentry);
+	if (IS_ERR(dentry))
+		goto out_unlock;
+
+	error = -EEXIST;
+	if (dentry->d_inode)
+		goto out_dput;
+
+	if (!IS_POSIXACL(parent->dentry->d_inode))
+		mode &= ~current->fs->umask;
+
+	error = mnt_want_write(parent->mnt);
+	if (error)
+		goto out_dput;
+
+	error = btrfs_may_create(parent->dentry->d_inode, dentry);
+	if (error)
+		goto out_drop_write;
+
+	/*
+	 * Actually perform the low-level subvolume creation after all
+	 * this VFS fuzz.
+	 *
+	 * Eventually we want to pass in an inode under which we create this
+	 * subvolume, but for now all are under the filesystem root.
+	 *
+	 * Also we should pass on the mode eventually to allow creating new
+	 * subvolume with specific mode bits.
+	 */
+	if (snap_src) {
+		struct dentry *dir = dentry->d_parent;
+		struct dentry *test = dir->d_parent;
+		struct btrfs_path *path = btrfs_alloc_path();
+		int ret;
+		u64 test_oid;
+		u64 parent_oid = BTRFS_I(dir->d_inode)->root->root_key.objectid;
+
+		test_oid = snap_src->root_key.objectid;
+
+		ret = btrfs_find_root_ref(snap_src->fs_info->tree_root,
+					  path, parent_oid, test_oid);
+		if (ret == 0)
+			goto create;
+		btrfs_release_path(snap_src->fs_info->tree_root, path);
+
+		/* we need to make sure we aren't creating a directory loop
+		 * by taking a snapshot of something that has our current
+		 * subvol in its directory tree.  So, this loops through
+		 * the dentries and checks the forward refs for each subvolume
+		 * to see if is references the subvolume where we are
+		 * placing this new snapshot.
+		 */
+		while (1) {
+			if (!test ||
+			    dir == snap_src->fs_info->sb->s_root ||
+			    test == snap_src->fs_info->sb->s_root ||
+			    test->d_inode->i_sb != snap_src->fs_info->sb) {
+				break;
+			}
+			if (S_ISLNK(test->d_inode->i_mode)) {
+				printk(KERN_INFO "Btrfs symlink in snapshot "
+				       "path, failed\n");
+				error = -EMLINK;
+				btrfs_free_path(path);
+				goto out_drop_write;
+			}
+			test_oid =
+				BTRFS_I(test->d_inode)->root->root_key.objectid;
+			ret = btrfs_find_root_ref(snap_src->fs_info->tree_root,
+				  path, test_oid, parent_oid);
+			if (ret == 0) {
+				printk(KERN_INFO "Btrfs snapshot creation "
+				       "failed, looping\n");
+				error = -EMLINK;
+				btrfs_free_path(path);
+				goto out_drop_write;
+			}
+			btrfs_release_path(snap_src->fs_info->tree_root, path);
+			test = test->d_parent;
+		}
+create:
+		btrfs_free_path(path);
+		error = create_snapshot(snap_src, dentry, name, namelen);
+	} else {
+		error = create_subvol(BTRFS_I(parent->dentry->d_inode)->root,
+				      dentry, name, namelen);
+	}
+	if (error)
+		goto out_drop_write;
+
+	fsnotify_mkdir(parent->dentry->d_inode, dentry);
+out_drop_write:
+	mnt_drop_write(parent->mnt);
+out_dput:
+	dput(dentry);
+out_unlock:
+	mutex_unlock(&parent->dentry->d_inode->i_mutex);
+	return error;
+}
+
+
+static int btrfs_defrag_file(struct file *file)
+{
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct btrfs_ordered_extent *ordered;
+	struct page *page;
+	unsigned long last_index;
+	unsigned long ra_pages = root->fs_info->bdi.ra_pages;
+	unsigned long total_read = 0;
+	u64 page_start;
+	u64 page_end;
+	unsigned long i;
+	int ret;
+
+	ret = btrfs_check_free_space(root, inode->i_size, 0);
+	if (ret)
+		return -ENOSPC;
+
+	mutex_lock(&inode->i_mutex);
+	last_index = inode->i_size >> PAGE_CACHE_SHIFT;
+	for (i = 0; i <= last_index; i++) {
+		if (total_read % ra_pages == 0) {
+			btrfs_force_ra(inode->i_mapping, &file->f_ra, file, i,
+				       min(last_index, i + ra_pages - 1));
+		}
+		total_read++;
+again:
+		page = grab_cache_page(inode->i_mapping, i);
+		if (!page)
+			goto out_unlock;
+		if (!PageUptodate(page)) {
+			btrfs_readpage(NULL, page);
+			lock_page(page);
+			if (!PageUptodate(page)) {
+				unlock_page(page);
+				page_cache_release(page);
+				goto out_unlock;
+			}
+		}
+
+		wait_on_page_writeback(page);
+
+		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
+		page_end = page_start + PAGE_CACHE_SIZE - 1;
+		lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+
+		ordered = btrfs_lookup_ordered_extent(inode, page_start);
+		if (ordered) {
+			unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+			unlock_page(page);
+			page_cache_release(page);
+			btrfs_start_ordered_extent(inode, ordered, 1);
+			btrfs_put_ordered_extent(ordered);
+			goto again;
+		}
+		set_page_extent_mapped(page);
+
+		/*
+		 * this makes sure page_mkwrite is called on the
+		 * page if it is dirtied again later
+		 */
+		clear_page_dirty_for_io(page);
+
+		btrfs_set_extent_delalloc(inode, page_start, page_end);
+
+		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+		set_page_dirty(page);
+		unlock_page(page);
+		page_cache_release(page);
+		balance_dirty_pages_ratelimited_nr(inode->i_mapping, 1);
+	}
+
+out_unlock:
+	mutex_unlock(&inode->i_mutex);
+	return 0;
+}
+
+/*
+ * Called inside transaction, so use GFP_NOFS
+ */
+
+static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg)
+{
+	u64 new_size;
+	u64 old_size;
+	u64 devid = 1;
+	struct btrfs_ioctl_vol_args *vol_args;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_device *device = NULL;
+	char *sizestr;
+	char *devstr = NULL;
+	int ret = 0;
+	int namelen;
+	int mod = 0;
+
+	if (root->fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
+
+	if (!vol_args)
+		return -ENOMEM;
+
+	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	namelen = strlen(vol_args->name);
+
+	mutex_lock(&root->fs_info->volume_mutex);
+	sizestr = vol_args->name;
+	devstr = strchr(sizestr, ':');
+	if (devstr) {
+		char *end;
+		sizestr = devstr + 1;
+		*devstr = '\0';
+		devstr = vol_args->name;
+		devid = simple_strtoull(devstr, &end, 10);
+		printk(KERN_INFO "resizing devid %llu\n", devid);
+	}
+	device = btrfs_find_device(root, devid, NULL, NULL);
+	if (!device) {
+		printk(KERN_INFO "resizer unable to find device %llu\n", devid);
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+	if (!strcmp(sizestr, "max"))
+		new_size = device->bdev->bd_inode->i_size;
+	else {
+		if (sizestr[0] == '-') {
+			mod = -1;
+			sizestr++;
+		} else if (sizestr[0] == '+') {
+			mod = 1;
+			sizestr++;
+		}
+		new_size = btrfs_parse_size(sizestr);
+		if (new_size == 0) {
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+	}
+
+	old_size = device->total_bytes;
+
+	if (mod < 0) {
+		if (new_size > old_size) {
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+		new_size = old_size - new_size;
+	} else if (mod > 0) {
+		new_size = old_size + new_size;
+	}
+
+	if (new_size < 256 * 1024 * 1024) {
+		ret = -EINVAL;
+		goto out_unlock;
+	}
+	if (new_size > device->bdev->bd_inode->i_size) {
+		ret = -EFBIG;
+		goto out_unlock;
+	}
+
+	do_div(new_size, root->sectorsize);
+	new_size *= root->sectorsize;
+
+	printk(KERN_INFO "new size for %s is %llu\n",
+		device->name, (unsigned long long)new_size);
+
+	if (new_size > old_size) {
+		trans = btrfs_start_transaction(root, 1);
+		ret = btrfs_grow_device(trans, device, new_size);
+		btrfs_commit_transaction(trans, root);
+	} else {
+		ret = btrfs_shrink_device(device, new_size);
+	}
+
+out_unlock:
+	mutex_unlock(&root->fs_info->volume_mutex);
+out:
+	kfree(vol_args);
+	return ret;
+}
+
+static noinline int btrfs_ioctl_snap_create(struct file *file,
+					    void __user *arg, int subvol)
+{
+	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
+	struct btrfs_ioctl_vol_args *vol_args;
+	struct btrfs_dir_item *di;
+	struct btrfs_path *path;
+	struct file *src_file;
+	u64 root_dirid;
+	int namelen;
+	int ret = 0;
+
+	if (root->fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
+
+	if (!vol_args)
+		return -ENOMEM;
+
+	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	namelen = strlen(vol_args->name);
+	if (strchr(vol_args->name, '/')) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	root_dirid = root->fs_info->sb->s_root->d_inode->i_ino,
+	di = btrfs_lookup_dir_item(NULL, root->fs_info->tree_root,
+			    path, root_dirid,
+			    vol_args->name, namelen, 0);
+	btrfs_free_path(path);
+
+	if (di && !IS_ERR(di)) {
+		ret = -EEXIST;
+		goto out;
+	}
+
+	if (IS_ERR(di)) {
+		ret = PTR_ERR(di);
+		goto out;
+	}
+
+	if (subvol) {
+		ret = btrfs_mksubvol(&file->f_path, vol_args->name,
+				     file->f_path.dentry->d_inode->i_mode,
+				     namelen, NULL);
+	} else {
+		struct inode *src_inode;
+		src_file = fget(vol_args->fd);
+		if (!src_file) {
+			ret = -EINVAL;
+			goto out;
+		}
+
+		src_inode = src_file->f_path.dentry->d_inode;
+		if (src_inode->i_sb != file->f_path.dentry->d_inode->i_sb) {
+			printk(KERN_INFO "btrfs: Snapshot src from "
+			       "another FS\n");
+			ret = -EINVAL;
+			fput(src_file);
+			goto out;
+		}
+		ret = btrfs_mksubvol(&file->f_path, vol_args->name,
+			     file->f_path.dentry->d_inode->i_mode,
+			     namelen, BTRFS_I(src_inode)->root);
+		fput(src_file);
+	}
+
+out:
+	kfree(vol_args);
+	return ret;
+}
+
+static int btrfs_ioctl_defrag(struct file *file)
+{
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret;
+
+	ret = mnt_want_write(file->f_path.mnt);
+	if (ret)
+		return ret;
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFDIR:
+		if (!capable(CAP_SYS_ADMIN)) {
+			ret = -EPERM;
+			goto out;
+		}
+		btrfs_defrag_root(root, 0);
+		btrfs_defrag_root(root->fs_info->extent_root, 0);
+		break;
+	case S_IFREG:
+		if (!(file->f_mode & FMODE_WRITE)) {
+			ret = -EINVAL;
+			goto out;
+		}
+		btrfs_defrag_file(file);
+		break;
+	}
+out:
+	mnt_drop_write(file->f_path.mnt);
+	return ret;
+}
+
+static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg)
+{
+	struct btrfs_ioctl_vol_args *vol_args;
+	int ret;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
+
+	if (!vol_args)
+		return -ENOMEM;
+
+	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
+		ret = -EFAULT;
+		goto out;
+	}
+	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_init_new_device(root, vol_args->name);
+
+out:
+	kfree(vol_args);
+	return ret;
+}
+
+static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg)
+{
+	struct btrfs_ioctl_vol_args *vol_args;
+	int ret;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (root->fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS);
+
+	if (!vol_args)
+		return -ENOMEM;
+
+	if (copy_from_user(vol_args, arg, sizeof(*vol_args))) {
+		ret = -EFAULT;
+		goto out;
+	}
+	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
+	ret = btrfs_rm_device(root, vol_args->name);
+
+out:
+	kfree(vol_args);
+	return ret;
+}
+
+static long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
+		u64 off, u64 olen, u64 destoff)
+{
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct file *src_file;
+	struct inode *src;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	char *buf;
+	struct btrfs_key key;
+	u32 nritems;
+	int slot;
+	int ret;
+	u64 len = olen;
+	u64 bs = root->fs_info->sb->s_blocksize;
+	u64 hint_byte;
+
+	/*
+	 * TODO:
+	 * - split compressed inline extents.  annoying: we need to
+	 *   decompress into destination's address_space (the file offset
+	 *   may change, so source mapping won't do), then recompress (or
+	 *   otherwise reinsert) a subrange.
+	 * - allow ranges within the same file to be cloned (provided
+	 *   they don't overlap)?
+	 */
+
+	/* the destination must be opened for writing */
+	if (!(file->f_mode & FMODE_WRITE))
+		return -EINVAL;
+
+	ret = mnt_want_write(file->f_path.mnt);
+	if (ret)
+		return ret;
+
+	src_file = fget(srcfd);
+	if (!src_file) {
+		ret = -EBADF;
+		goto out_drop_write;
+	}
+	src = src_file->f_dentry->d_inode;
+
+	ret = -EINVAL;
+	if (src == inode)
+		goto out_fput;
+
+	ret = -EISDIR;
+	if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode))
+		goto out_fput;
+
+	ret = -EXDEV;
+	if (src->i_sb != inode->i_sb || BTRFS_I(src)->root != root)
+		goto out_fput;
+
+	ret = -ENOMEM;
+	buf = vmalloc(btrfs_level_size(root, 0));
+	if (!buf)
+		goto out_fput;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		vfree(buf);
+		goto out_fput;
+	}
+	path->reada = 2;
+
+	if (inode < src) {
+		mutex_lock(&inode->i_mutex);
+		mutex_lock(&src->i_mutex);
+	} else {
+		mutex_lock(&src->i_mutex);
+		mutex_lock(&inode->i_mutex);
+	}
+
+	/* determine range to clone */
+	ret = -EINVAL;
+	if (off >= src->i_size || off + len > src->i_size)
+		goto out_unlock;
+	if (len == 0)
+		olen = len = src->i_size - off;
+	/* if we extend to eof, continue to block boundary */
+	if (off + len == src->i_size)
+		len = ((src->i_size + bs-1) & ~(bs-1))
+			- off;
+
+	/* verify the end result is block aligned */
+	if ((off & (bs-1)) ||
+	    ((off + len) & (bs-1)))
+		goto out_unlock;
+
+	/* do any pending delalloc/csum calc on src, one way or
+	   another, and lock file content */
+	while (1) {
+		struct btrfs_ordered_extent *ordered;
+		lock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
+		ordered = btrfs_lookup_first_ordered_extent(inode, off+len);
+		if (BTRFS_I(src)->delalloc_bytes == 0 && !ordered)
+			break;
+		unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
+		if (ordered)
+			btrfs_put_ordered_extent(ordered);
+		btrfs_wait_ordered_range(src, off, off+len);
+	}
+
+	trans = btrfs_start_transaction(root, 1);
+	BUG_ON(!trans);
+
+	/* punch hole in destination first */
+	btrfs_drop_extents(trans, root, inode, off, off+len, 0, &hint_byte);
+
+	/* clone data */
+	key.objectid = src->i_ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = 0;
+
+	while (1) {
+		/*
+		 * note the key will change type as we walk through the
+		 * tree.
+		 */
+		ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+
+		nritems = btrfs_header_nritems(path->nodes[0]);
+		if (path->slots[0] >= nritems) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			if (ret > 0)
+				break;
+			nritems = btrfs_header_nritems(path->nodes[0]);
+		}
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY ||
+		    key.objectid != src->i_ino)
+			break;
+
+		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) {
+			struct btrfs_file_extent_item *extent;
+			int type;
+			u32 size;
+			struct btrfs_key new_key;
+			u64 disko = 0, diskl = 0;
+			u64 datao = 0, datal = 0;
+			u8 comp;
+
+			size = btrfs_item_size_nr(leaf, slot);
+			read_extent_buffer(leaf, buf,
+					   btrfs_item_ptr_offset(leaf, slot),
+					   size);
+
+			extent = btrfs_item_ptr(leaf, slot,
+						struct btrfs_file_extent_item);
+			comp = btrfs_file_extent_compression(leaf, extent);
+			type = btrfs_file_extent_type(leaf, extent);
+			if (type == BTRFS_FILE_EXTENT_REG) {
+				disko = btrfs_file_extent_disk_bytenr(leaf,
+								      extent);
+				diskl = btrfs_file_extent_disk_num_bytes(leaf,
+								 extent);
+				datao = btrfs_file_extent_offset(leaf, extent);
+				datal = btrfs_file_extent_num_bytes(leaf,
+								    extent);
+			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
+				/* take upper bound, may be compressed */
+				datal = btrfs_file_extent_ram_bytes(leaf,
+								    extent);
+			}
+			btrfs_release_path(root, path);
+
+			if (key.offset + datal < off ||
+			    key.offset >= off+len)
+				goto next;
+
+			memcpy(&new_key, &key, sizeof(new_key));
+			new_key.objectid = inode->i_ino;
+			new_key.offset = key.offset + destoff - off;
+
+			if (type == BTRFS_FILE_EXTENT_REG) {
+				ret = btrfs_insert_empty_item(trans, root, path,
+							      &new_key, size);
+				if (ret)
+					goto out;
+
+				leaf = path->nodes[0];
+				slot = path->slots[0];
+				write_extent_buffer(leaf, buf,
+					    btrfs_item_ptr_offset(leaf, slot),
+					    size);
+
+				extent = btrfs_item_ptr(leaf, slot,
+						struct btrfs_file_extent_item);
+
+				if (off > key.offset) {
+					datao += off - key.offset;
+					datal -= off - key.offset;
+				}
+				if (key.offset + datao + datal + key.offset >
+				    off + len)
+					datal = off + len - key.offset - datao;
+				/* disko == 0 means it's a hole */
+				if (!disko)
+					datao = 0;
+
+				btrfs_set_file_extent_offset(leaf, extent,
+							     datao);
+				btrfs_set_file_extent_num_bytes(leaf, extent,
+								datal);
+				if (disko) {
+					inode_add_bytes(inode, datal);
+					ret = btrfs_inc_extent_ref(trans, root,
+						   disko, diskl, leaf->start,
+						   root->root_key.objectid,
+						   trans->transid,
+						   inode->i_ino);
+					BUG_ON(ret);
+				}
+			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
+				u64 skip = 0;
+				u64 trim = 0;
+				if (off > key.offset) {
+					skip = off - key.offset;
+					new_key.offset += skip;
+				}
+
+				if (key.offset + datal > off+len)
+					trim = key.offset + datal - (off+len);
+
+				if (comp && (skip || trim)) {
+					ret = -EINVAL;
+					goto out;
+				}
+				size -= skip + trim;
+				datal -= skip + trim;
+				ret = btrfs_insert_empty_item(trans, root, path,
+							      &new_key, size);
+				if (ret)
+					goto out;
+
+				if (skip) {
+					u32 start =
+					  btrfs_file_extent_calc_inline_size(0);
+					memmove(buf+start, buf+start+skip,
+						datal);
+				}
+
+				leaf = path->nodes[0];
+				slot = path->slots[0];
+				write_extent_buffer(leaf, buf,
+					    btrfs_item_ptr_offset(leaf, slot),
+					    size);
+				inode_add_bytes(inode, datal);
+			}
+
+			btrfs_mark_buffer_dirty(leaf);
+		}
+
+next:
+		btrfs_release_path(root, path);
+		key.offset++;
+	}
+	ret = 0;
+out:
+	btrfs_release_path(root, path);
+	if (ret == 0) {
+		inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+		if (destoff + olen > inode->i_size)
+			btrfs_i_size_write(inode, destoff + olen);
+		BTRFS_I(inode)->flags = BTRFS_I(src)->flags;
+		ret = btrfs_update_inode(trans, root, inode);
+	}
+	btrfs_end_transaction(trans, root);
+	unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
+	if (ret)
+		vmtruncate(inode, 0);
+out_unlock:
+	mutex_unlock(&src->i_mutex);
+	mutex_unlock(&inode->i_mutex);
+	vfree(buf);
+	btrfs_free_path(path);
+out_fput:
+	fput(src_file);
+out_drop_write:
+	mnt_drop_write(file->f_path.mnt);
+	return ret;
+}
+
+static long btrfs_ioctl_clone_range(struct file *file, void __user *argp)
+{
+	struct btrfs_ioctl_clone_range_args args;
+
+	if (copy_from_user(&args, argp, sizeof(args)))
+		return -EFAULT;
+	return btrfs_ioctl_clone(file, args.src_fd, args.src_offset,
+				 args.src_length, args.dest_offset);
+}
+
+/*
+ * there are many ways the trans_start and trans_end ioctls can lead
+ * to deadlocks.  They should only be used by applications that
+ * basically own the machine, and have a very in depth understanding
+ * of all the possible deadlocks and enospc problems.
+ */
+static long btrfs_ioctl_trans_start(struct file *file)
+{
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	int ret = 0;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (file->private_data) {
+		ret = -EINPROGRESS;
+		goto out;
+	}
+
+	ret = mnt_want_write(file->f_path.mnt);
+	if (ret)
+		goto out;
+
+	mutex_lock(&root->fs_info->trans_mutex);
+	root->fs_info->open_ioctl_trans++;
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	trans = btrfs_start_ioctl_transaction(root, 0);
+	if (trans)
+		file->private_data = trans;
+	else
+		ret = -ENOMEM;
+	/*printk(KERN_INFO "btrfs_ioctl_trans_start on %p\n", file);*/
+out:
+	return ret;
+}
+
+/*
+ * there are many ways the trans_start and trans_end ioctls can lead
+ * to deadlocks.  They should only be used by applications that
+ * basically own the machine, and have a very in depth understanding
+ * of all the possible deadlocks and enospc problems.
+ */
+long btrfs_ioctl_trans_end(struct file *file)
+{
+	struct inode *inode = fdentry(file)->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	int ret = 0;
+
+	trans = file->private_data;
+	if (!trans) {
+		ret = -EINVAL;
+		goto out;
+	}
+	btrfs_end_transaction(trans, root);
+	file->private_data = NULL;
+
+	mutex_lock(&root->fs_info->trans_mutex);
+	root->fs_info->open_ioctl_trans--;
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	mnt_drop_write(file->f_path.mnt);
+
+out:
+	return ret;
+}
+
+long btrfs_ioctl(struct file *file, unsigned int
+		cmd, unsigned long arg)
+{
+	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
+	void __user *argp = (void __user *)arg;
+
+	switch (cmd) {
+	case BTRFS_IOC_SNAP_CREATE:
+		return btrfs_ioctl_snap_create(file, argp, 0);
+	case BTRFS_IOC_SUBVOL_CREATE:
+		return btrfs_ioctl_snap_create(file, argp, 1);
+	case BTRFS_IOC_DEFRAG:
+		return btrfs_ioctl_defrag(file);
+	case BTRFS_IOC_RESIZE:
+		return btrfs_ioctl_resize(root, argp);
+	case BTRFS_IOC_ADD_DEV:
+		return btrfs_ioctl_add_dev(root, argp);
+	case BTRFS_IOC_RM_DEV:
+		return btrfs_ioctl_rm_dev(root, argp);
+	case BTRFS_IOC_BALANCE:
+		return btrfs_balance(root->fs_info->dev_root);
+	case BTRFS_IOC_CLONE:
+		return btrfs_ioctl_clone(file, arg, 0, 0, 0);
+	case BTRFS_IOC_CLONE_RANGE:
+		return btrfs_ioctl_clone_range(file, argp);
+	case BTRFS_IOC_TRANS_START:
+		return btrfs_ioctl_trans_start(file);
+	case BTRFS_IOC_TRANS_END:
+		return btrfs_ioctl_trans_end(file);
+	case BTRFS_IOC_SYNC:
+		btrfs_sync_fs(file->f_dentry->d_sb, 1);
+		return 0;
+	}
+
+	return -ENOTTY;
+}
diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h
new file mode 100644
index 00000000000..78049ea208d
--- /dev/null
+++ b/fs/btrfs/ioctl.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __IOCTL_
+#define __IOCTL_
+#include <linux/ioctl.h>
+
+#define BTRFS_IOCTL_MAGIC 0x94
+#define BTRFS_VOL_NAME_MAX 255
+#define BTRFS_PATH_NAME_MAX 3072
+
+struct btrfs_ioctl_vol_args {
+	__s64 fd;
+	char name[BTRFS_PATH_NAME_MAX + 1];
+};
+
+#define BTRFS_IOC_SNAP_CREATE _IOW(BTRFS_IOCTL_MAGIC, 1, \
+				   struct btrfs_ioctl_vol_args)
+#define BTRFS_IOC_DEFRAG _IOW(BTRFS_IOCTL_MAGIC, 2, \
+				   struct btrfs_ioctl_vol_args)
+#define BTRFS_IOC_RESIZE _IOW(BTRFS_IOCTL_MAGIC, 3, \
+				   struct btrfs_ioctl_vol_args)
+#define BTRFS_IOC_SCAN_DEV _IOW(BTRFS_IOCTL_MAGIC, 4, \
+				   struct btrfs_ioctl_vol_args)
+/* trans start and trans end are dangerous, and only for
+ * use by applications that know how to avoid the
+ * resulting deadlocks
+ */
+#define BTRFS_IOC_TRANS_START  _IO(BTRFS_IOCTL_MAGIC, 6)
+#define BTRFS_IOC_TRANS_END    _IO(BTRFS_IOCTL_MAGIC, 7)
+#define BTRFS_IOC_SYNC         _IO(BTRFS_IOCTL_MAGIC, 8)
+
+#define BTRFS_IOC_CLONE        _IOW(BTRFS_IOCTL_MAGIC, 9, int)
+#define BTRFS_IOC_ADD_DEV _IOW(BTRFS_IOCTL_MAGIC, 10, \
+				   struct btrfs_ioctl_vol_args)
+#define BTRFS_IOC_RM_DEV _IOW(BTRFS_IOCTL_MAGIC, 11, \
+				   struct btrfs_ioctl_vol_args)
+#define BTRFS_IOC_BALANCE _IOW(BTRFS_IOCTL_MAGIC, 12, \
+				   struct btrfs_ioctl_vol_args)
+struct btrfs_ioctl_clone_range_args {
+  __s64 src_fd;
+  __u64 src_offset, src_length;
+  __u64 dest_offset;
+};
+
+#define BTRFS_IOC_CLONE_RANGE _IOW(BTRFS_IOCTL_MAGIC, 13, \
+				  struct btrfs_ioctl_clone_range_args)
+
+#define BTRFS_IOC_SUBVOL_CREATE _IOW(BTRFS_IOCTL_MAGIC, 14, \
+				   struct btrfs_ioctl_vol_args)
+
+#endif
diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c
new file mode 100644
index 00000000000..39bae7761db
--- /dev/null
+++ b/fs/btrfs/locking.c
@@ -0,0 +1,88 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+#include <linux/sched.h>
+#include <linux/gfp.h>
+#include <linux/pagemap.h>
+#include <linux/spinlock.h>
+#include <linux/page-flags.h>
+#include <asm/bug.h>
+#include "ctree.h"
+#include "extent_io.h"
+#include "locking.h"
+
+/*
+ * locks the per buffer mutex in an extent buffer.  This uses adaptive locks
+ * and the spin is not tuned very extensively.  The spinning does make a big
+ * difference in almost every workload, but spinning for the right amount of
+ * time needs some help.
+ *
+ * In general, we want to spin as long as the lock holder is doing btree
+ * searches, and we should give up if they are in more expensive code.
+ */
+
+int btrfs_tree_lock(struct extent_buffer *eb)
+{
+	int i;
+
+	if (mutex_trylock(&eb->mutex))
+		return 0;
+	for (i = 0; i < 512; i++) {
+		cpu_relax();
+		if (mutex_trylock(&eb->mutex))
+			return 0;
+	}
+	cpu_relax();
+	mutex_lock_nested(&eb->mutex, BTRFS_MAX_LEVEL - btrfs_header_level(eb));
+	return 0;
+}
+
+int btrfs_try_tree_lock(struct extent_buffer *eb)
+{
+	return mutex_trylock(&eb->mutex);
+}
+
+int btrfs_tree_unlock(struct extent_buffer *eb)
+{
+	mutex_unlock(&eb->mutex);
+	return 0;
+}
+
+int btrfs_tree_locked(struct extent_buffer *eb)
+{
+	return mutex_is_locked(&eb->mutex);
+}
+
+/*
+ * btrfs_search_slot uses this to decide if it should drop its locks
+ * before doing something expensive like allocating free blocks for cow.
+ */
+int btrfs_path_lock_waiting(struct btrfs_path *path, int level)
+{
+	int i;
+	struct extent_buffer *eb;
+	for (i = level; i <= level + 1 && i < BTRFS_MAX_LEVEL; i++) {
+		eb = path->nodes[i];
+		if (!eb)
+			break;
+		smp_mb();
+		if (!list_empty(&eb->mutex.wait_list))
+			return 1;
+	}
+	return 0;
+}
+
diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h
new file mode 100644
index 00000000000..bc1faef1251
--- /dev/null
+++ b/fs/btrfs/locking.h
@@ -0,0 +1,27 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_LOCKING_
+#define __BTRFS_LOCKING_
+
+int btrfs_tree_lock(struct extent_buffer *eb);
+int btrfs_tree_unlock(struct extent_buffer *eb);
+int btrfs_tree_locked(struct extent_buffer *eb);
+int btrfs_try_tree_lock(struct extent_buffer *eb);
+int btrfs_path_lock_waiting(struct btrfs_path *path, int level);
+#endif
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
new file mode 100644
index 00000000000..a2094017027
--- /dev/null
+++ b/fs/btrfs/ordered-data.c
@@ -0,0 +1,730 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/gfp.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/writeback.h>
+#include <linux/pagevec.h>
+#include "ctree.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "extent_io.h"
+
+static u64 entry_end(struct btrfs_ordered_extent *entry)
+{
+	if (entry->file_offset + entry->len < entry->file_offset)
+		return (u64)-1;
+	return entry->file_offset + entry->len;
+}
+
+/* returns NULL if the insertion worked, or it returns the node it did find
+ * in the tree
+ */
+static struct rb_node *tree_insert(struct rb_root *root, u64 file_offset,
+				   struct rb_node *node)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct btrfs_ordered_extent *entry;
+
+	while (*p) {
+		parent = *p;
+		entry = rb_entry(parent, struct btrfs_ordered_extent, rb_node);
+
+		if (file_offset < entry->file_offset)
+			p = &(*p)->rb_left;
+		else if (file_offset >= entry_end(entry))
+			p = &(*p)->rb_right;
+		else
+			return parent;
+	}
+
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+	return NULL;
+}
+
+/*
+ * look for a given offset in the tree, and if it can't be found return the
+ * first lesser offset
+ */
+static struct rb_node *__tree_search(struct rb_root *root, u64 file_offset,
+				     struct rb_node **prev_ret)
+{
+	struct rb_node *n = root->rb_node;
+	struct rb_node *prev = NULL;
+	struct rb_node *test;
+	struct btrfs_ordered_extent *entry;
+	struct btrfs_ordered_extent *prev_entry = NULL;
+
+	while (n) {
+		entry = rb_entry(n, struct btrfs_ordered_extent, rb_node);
+		prev = n;
+		prev_entry = entry;
+
+		if (file_offset < entry->file_offset)
+			n = n->rb_left;
+		else if (file_offset >= entry_end(entry))
+			n = n->rb_right;
+		else
+			return n;
+	}
+	if (!prev_ret)
+		return NULL;
+
+	while (prev && file_offset >= entry_end(prev_entry)) {
+		test = rb_next(prev);
+		if (!test)
+			break;
+		prev_entry = rb_entry(test, struct btrfs_ordered_extent,
+				      rb_node);
+		if (file_offset < entry_end(prev_entry))
+			break;
+
+		prev = test;
+	}
+	if (prev)
+		prev_entry = rb_entry(prev, struct btrfs_ordered_extent,
+				      rb_node);
+	while (prev && file_offset < entry_end(prev_entry)) {
+		test = rb_prev(prev);
+		if (!test)
+			break;
+		prev_entry = rb_entry(test, struct btrfs_ordered_extent,
+				      rb_node);
+		prev = test;
+	}
+	*prev_ret = prev;
+	return NULL;
+}
+
+/*
+ * helper to check if a given offset is inside a given entry
+ */
+static int offset_in_entry(struct btrfs_ordered_extent *entry, u64 file_offset)
+{
+	if (file_offset < entry->file_offset ||
+	    entry->file_offset + entry->len <= file_offset)
+		return 0;
+	return 1;
+}
+
+/*
+ * look find the first ordered struct that has this offset, otherwise
+ * the first one less than this offset
+ */
+static inline struct rb_node *tree_search(struct btrfs_ordered_inode_tree *tree,
+					  u64 file_offset)
+{
+	struct rb_root *root = &tree->tree;
+	struct rb_node *prev;
+	struct rb_node *ret;
+	struct btrfs_ordered_extent *entry;
+
+	if (tree->last) {
+		entry = rb_entry(tree->last, struct btrfs_ordered_extent,
+				 rb_node);
+		if (offset_in_entry(entry, file_offset))
+			return tree->last;
+	}
+	ret = __tree_search(root, file_offset, &prev);
+	if (!ret)
+		ret = prev;
+	if (ret)
+		tree->last = ret;
+	return ret;
+}
+
+/* allocate and add a new ordered_extent into the per-inode tree.
+ * file_offset is the logical offset in the file
+ *
+ * start is the disk block number of an extent already reserved in the
+ * extent allocation tree
+ *
+ * len is the length of the extent
+ *
+ * This also sets the EXTENT_ORDERED bit on the range in the inode.
+ *
+ * The tree is given a single reference on the ordered extent that was
+ * inserted.
+ */
+int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
+			     u64 start, u64 len, u64 disk_len, int type)
+{
+	struct btrfs_ordered_inode_tree *tree;
+	struct rb_node *node;
+	struct btrfs_ordered_extent *entry;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	entry = kzalloc(sizeof(*entry), GFP_NOFS);
+	if (!entry)
+		return -ENOMEM;
+
+	mutex_lock(&tree->mutex);
+	entry->file_offset = file_offset;
+	entry->start = start;
+	entry->len = len;
+	entry->disk_len = disk_len;
+	entry->inode = inode;
+	if (type != BTRFS_ORDERED_IO_DONE && type != BTRFS_ORDERED_COMPLETE)
+		set_bit(type, &entry->flags);
+
+	/* one ref for the tree */
+	atomic_set(&entry->refs, 1);
+	init_waitqueue_head(&entry->wait);
+	INIT_LIST_HEAD(&entry->list);
+	INIT_LIST_HEAD(&entry->root_extent_list);
+
+	node = tree_insert(&tree->tree, file_offset,
+			   &entry->rb_node);
+	BUG_ON(node);
+
+	set_extent_ordered(&BTRFS_I(inode)->io_tree, file_offset,
+			   entry_end(entry) - 1, GFP_NOFS);
+
+	spin_lock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
+	list_add_tail(&entry->root_extent_list,
+		      &BTRFS_I(inode)->root->fs_info->ordered_extents);
+	spin_unlock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
+
+	mutex_unlock(&tree->mutex);
+	BUG_ON(node);
+	return 0;
+}
+
+/*
+ * Add a struct btrfs_ordered_sum into the list of checksums to be inserted
+ * when an ordered extent is finished.  If the list covers more than one
+ * ordered extent, it is split across multiples.
+ */
+int btrfs_add_ordered_sum(struct inode *inode,
+			  struct btrfs_ordered_extent *entry,
+			  struct btrfs_ordered_sum *sum)
+{
+	struct btrfs_ordered_inode_tree *tree;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	mutex_lock(&tree->mutex);
+	list_add_tail(&sum->list, &entry->list);
+	mutex_unlock(&tree->mutex);
+	return 0;
+}
+
+/*
+ * this is used to account for finished IO across a given range
+ * of the file.  The IO should not span ordered extents.  If
+ * a given ordered_extent is completely done, 1 is returned, otherwise
+ * 0.
+ *
+ * test_and_set_bit on a flag in the struct btrfs_ordered_extent is used
+ * to make sure this function only returns 1 once for a given ordered extent.
+ */
+int btrfs_dec_test_ordered_pending(struct inode *inode,
+				   u64 file_offset, u64 io_size)
+{
+	struct btrfs_ordered_inode_tree *tree;
+	struct rb_node *node;
+	struct btrfs_ordered_extent *entry;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	int ret;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	mutex_lock(&tree->mutex);
+	clear_extent_ordered(io_tree, file_offset, file_offset + io_size - 1,
+			     GFP_NOFS);
+	node = tree_search(tree, file_offset);
+	if (!node) {
+		ret = 1;
+		goto out;
+	}
+
+	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
+	if (!offset_in_entry(entry, file_offset)) {
+		ret = 1;
+		goto out;
+	}
+
+	ret = test_range_bit(io_tree, entry->file_offset,
+			     entry->file_offset + entry->len - 1,
+			     EXTENT_ORDERED, 0);
+	if (ret == 0)
+		ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags);
+out:
+	mutex_unlock(&tree->mutex);
+	return ret == 0;
+}
+
+/*
+ * used to drop a reference on an ordered extent.  This will free
+ * the extent if the last reference is dropped
+ */
+int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
+{
+	struct list_head *cur;
+	struct btrfs_ordered_sum *sum;
+
+	if (atomic_dec_and_test(&entry->refs)) {
+		while (!list_empty(&entry->list)) {
+			cur = entry->list.next;
+			sum = list_entry(cur, struct btrfs_ordered_sum, list);
+			list_del(&sum->list);
+			kfree(sum);
+		}
+		kfree(entry);
+	}
+	return 0;
+}
+
+/*
+ * remove an ordered extent from the tree.  No references are dropped
+ * but, anyone waiting on this extent is woken up.
+ */
+int btrfs_remove_ordered_extent(struct inode *inode,
+				struct btrfs_ordered_extent *entry)
+{
+	struct btrfs_ordered_inode_tree *tree;
+	struct rb_node *node;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	mutex_lock(&tree->mutex);
+	node = &entry->rb_node;
+	rb_erase(node, &tree->tree);
+	tree->last = NULL;
+	set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags);
+
+	spin_lock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
+	list_del_init(&entry->root_extent_list);
+	spin_unlock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
+
+	mutex_unlock(&tree->mutex);
+	wake_up(&entry->wait);
+	return 0;
+}
+
+/*
+ * wait for all the ordered extents in a root.  This is done when balancing
+ * space between drives.
+ */
+int btrfs_wait_ordered_extents(struct btrfs_root *root, int nocow_only)
+{
+	struct list_head splice;
+	struct list_head *cur;
+	struct btrfs_ordered_extent *ordered;
+	struct inode *inode;
+
+	INIT_LIST_HEAD(&splice);
+
+	spin_lock(&root->fs_info->ordered_extent_lock);
+	list_splice_init(&root->fs_info->ordered_extents, &splice);
+	while (!list_empty(&splice)) {
+		cur = splice.next;
+		ordered = list_entry(cur, struct btrfs_ordered_extent,
+				     root_extent_list);
+		if (nocow_only &&
+		    !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags) &&
+		    !test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) {
+			list_move(&ordered->root_extent_list,
+				  &root->fs_info->ordered_extents);
+			cond_resched_lock(&root->fs_info->ordered_extent_lock);
+			continue;
+		}
+
+		list_del_init(&ordered->root_extent_list);
+		atomic_inc(&ordered->refs);
+
+		/*
+		 * the inode may be getting freed (in sys_unlink path).
+		 */
+		inode = igrab(ordered->inode);
+
+		spin_unlock(&root->fs_info->ordered_extent_lock);
+
+		if (inode) {
+			btrfs_start_ordered_extent(inode, ordered, 1);
+			btrfs_put_ordered_extent(ordered);
+			iput(inode);
+		} else {
+			btrfs_put_ordered_extent(ordered);
+		}
+
+		spin_lock(&root->fs_info->ordered_extent_lock);
+	}
+	spin_unlock(&root->fs_info->ordered_extent_lock);
+	return 0;
+}
+
+/*
+ * Used to start IO or wait for a given ordered extent to finish.
+ *
+ * If wait is one, this effectively waits on page writeback for all the pages
+ * in the extent, and it waits on the io completion code to insert
+ * metadata into the btree corresponding to the extent
+ */
+void btrfs_start_ordered_extent(struct inode *inode,
+				       struct btrfs_ordered_extent *entry,
+				       int wait)
+{
+	u64 start = entry->file_offset;
+	u64 end = start + entry->len - 1;
+
+	/*
+	 * pages in the range can be dirty, clean or writeback.  We
+	 * start IO on any dirty ones so the wait doesn't stall waiting
+	 * for pdflush to find them
+	 */
+	btrfs_fdatawrite_range(inode->i_mapping, start, end, WB_SYNC_ALL);
+	if (wait) {
+		wait_event(entry->wait, test_bit(BTRFS_ORDERED_COMPLETE,
+						 &entry->flags));
+	}
+}
+
+/*
+ * Used to wait on ordered extents across a large range of bytes.
+ */
+int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len)
+{
+	u64 end;
+	u64 orig_end;
+	u64 wait_end;
+	struct btrfs_ordered_extent *ordered;
+
+	if (start + len < start) {
+		orig_end = INT_LIMIT(loff_t);
+	} else {
+		orig_end = start + len - 1;
+		if (orig_end > INT_LIMIT(loff_t))
+			orig_end = INT_LIMIT(loff_t);
+	}
+	wait_end = orig_end;
+again:
+	/* start IO across the range first to instantiate any delalloc
+	 * extents
+	 */
+	btrfs_fdatawrite_range(inode->i_mapping, start, orig_end, WB_SYNC_NONE);
+
+	/* The compression code will leave pages locked but return from
+	 * writepage without setting the page writeback.  Starting again
+	 * with WB_SYNC_ALL will end up waiting for the IO to actually start.
+	 */
+	btrfs_fdatawrite_range(inode->i_mapping, start, orig_end, WB_SYNC_ALL);
+
+	btrfs_wait_on_page_writeback_range(inode->i_mapping,
+					   start >> PAGE_CACHE_SHIFT,
+					   orig_end >> PAGE_CACHE_SHIFT);
+
+	end = orig_end;
+	while (1) {
+		ordered = btrfs_lookup_first_ordered_extent(inode, end);
+		if (!ordered)
+			break;
+		if (ordered->file_offset > orig_end) {
+			btrfs_put_ordered_extent(ordered);
+			break;
+		}
+		if (ordered->file_offset + ordered->len < start) {
+			btrfs_put_ordered_extent(ordered);
+			break;
+		}
+		btrfs_start_ordered_extent(inode, ordered, 1);
+		end = ordered->file_offset;
+		btrfs_put_ordered_extent(ordered);
+		if (end == 0 || end == start)
+			break;
+		end--;
+	}
+	if (test_range_bit(&BTRFS_I(inode)->io_tree, start, orig_end,
+			   EXTENT_ORDERED | EXTENT_DELALLOC, 0)) {
+		schedule_timeout(1);
+		goto again;
+	}
+	return 0;
+}
+
+/*
+ * find an ordered extent corresponding to file_offset.  return NULL if
+ * nothing is found, otherwise take a reference on the extent and return it
+ */
+struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode,
+							 u64 file_offset)
+{
+	struct btrfs_ordered_inode_tree *tree;
+	struct rb_node *node;
+	struct btrfs_ordered_extent *entry = NULL;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	mutex_lock(&tree->mutex);
+	node = tree_search(tree, file_offset);
+	if (!node)
+		goto out;
+
+	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
+	if (!offset_in_entry(entry, file_offset))
+		entry = NULL;
+	if (entry)
+		atomic_inc(&entry->refs);
+out:
+	mutex_unlock(&tree->mutex);
+	return entry;
+}
+
+/*
+ * lookup and return any extent before 'file_offset'.  NULL is returned
+ * if none is found
+ */
+struct btrfs_ordered_extent *
+btrfs_lookup_first_ordered_extent(struct inode *inode, u64 file_offset)
+{
+	struct btrfs_ordered_inode_tree *tree;
+	struct rb_node *node;
+	struct btrfs_ordered_extent *entry = NULL;
+
+	tree = &BTRFS_I(inode)->ordered_tree;
+	mutex_lock(&tree->mutex);
+	node = tree_search(tree, file_offset);
+	if (!node)
+		goto out;
+
+	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
+	atomic_inc(&entry->refs);
+out:
+	mutex_unlock(&tree->mutex);
+	return entry;
+}
+
+/*
+ * After an extent is done, call this to conditionally update the on disk
+ * i_size.  i_size is updated to cover any fully written part of the file.
+ */
+int btrfs_ordered_update_i_size(struct inode *inode,
+				struct btrfs_ordered_extent *ordered)
+{
+	struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	u64 disk_i_size;
+	u64 new_i_size;
+	u64 i_size_test;
+	struct rb_node *node;
+	struct btrfs_ordered_extent *test;
+
+	mutex_lock(&tree->mutex);
+	disk_i_size = BTRFS_I(inode)->disk_i_size;
+
+	/*
+	 * if the disk i_size is already at the inode->i_size, or
+	 * this ordered extent is inside the disk i_size, we're done
+	 */
+	if (disk_i_size >= inode->i_size ||
+	    ordered->file_offset + ordered->len <= disk_i_size) {
+		goto out;
+	}
+
+	/*
+	 * we can't update the disk_isize if there are delalloc bytes
+	 * between disk_i_size and  this ordered extent
+	 */
+	if (test_range_bit(io_tree, disk_i_size,
+			   ordered->file_offset + ordered->len - 1,
+			   EXTENT_DELALLOC, 0)) {
+		goto out;
+	}
+	/*
+	 * walk backward from this ordered extent to disk_i_size.
+	 * if we find an ordered extent then we can't update disk i_size
+	 * yet
+	 */
+	node = &ordered->rb_node;
+	while (1) {
+		node = rb_prev(node);
+		if (!node)
+			break;
+		test = rb_entry(node, struct btrfs_ordered_extent, rb_node);
+		if (test->file_offset + test->len <= disk_i_size)
+			break;
+		if (test->file_offset >= inode->i_size)
+			break;
+		if (test->file_offset >= disk_i_size)
+			goto out;
+	}
+	new_i_size = min_t(u64, entry_end(ordered), i_size_read(inode));
+
+	/*
+	 * at this point, we know we can safely update i_size to at least
+	 * the offset from this ordered extent.  But, we need to
+	 * walk forward and see if ios from higher up in the file have
+	 * finished.
+	 */
+	node = rb_next(&ordered->rb_node);
+	i_size_test = 0;
+	if (node) {
+		/*
+		 * do we have an area where IO might have finished
+		 * between our ordered extent and the next one.
+		 */
+		test = rb_entry(node, struct btrfs_ordered_extent, rb_node);
+		if (test->file_offset > entry_end(ordered))
+			i_size_test = test->file_offset;
+	} else {
+		i_size_test = i_size_read(inode);
+	}
+
+	/*
+	 * i_size_test is the end of a region after this ordered
+	 * extent where there are no ordered extents.  As long as there
+	 * are no delalloc bytes in this area, it is safe to update
+	 * disk_i_size to the end of the region.
+	 */
+	if (i_size_test > entry_end(ordered) &&
+	    !test_range_bit(io_tree, entry_end(ordered), i_size_test - 1,
+			   EXTENT_DELALLOC, 0)) {
+		new_i_size = min_t(u64, i_size_test, i_size_read(inode));
+	}
+	BTRFS_I(inode)->disk_i_size = new_i_size;
+out:
+	mutex_unlock(&tree->mutex);
+	return 0;
+}
+
+/*
+ * search the ordered extents for one corresponding to 'offset' and
+ * try to find a checksum.  This is used because we allow pages to
+ * be reclaimed before their checksum is actually put into the btree
+ */
+int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
+			   u32 *sum)
+{
+	struct btrfs_ordered_sum *ordered_sum;
+	struct btrfs_sector_sum *sector_sums;
+	struct btrfs_ordered_extent *ordered;
+	struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree;
+	struct list_head *cur;
+	unsigned long num_sectors;
+	unsigned long i;
+	u32 sectorsize = BTRFS_I(inode)->root->sectorsize;
+	int ret = 1;
+
+	ordered = btrfs_lookup_ordered_extent(inode, offset);
+	if (!ordered)
+		return 1;
+
+	mutex_lock(&tree->mutex);
+	list_for_each_prev(cur, &ordered->list) {
+		ordered_sum = list_entry(cur, struct btrfs_ordered_sum, list);
+		if (disk_bytenr >= ordered_sum->bytenr) {
+			num_sectors = ordered_sum->len / sectorsize;
+			sector_sums = ordered_sum->sums;
+			for (i = 0; i < num_sectors; i++) {
+				if (sector_sums[i].bytenr == disk_bytenr) {
+					*sum = sector_sums[i].sum;
+					ret = 0;
+					goto out;
+				}
+			}
+		}
+	}
+out:
+	mutex_unlock(&tree->mutex);
+	btrfs_put_ordered_extent(ordered);
+	return ret;
+}
+
+
+/**
+ * taken from mm/filemap.c because it isn't exported
+ *
+ * __filemap_fdatawrite_range - start writeback on mapping dirty pages in range
+ * @mapping:	address space structure to write
+ * @start:	offset in bytes where the range starts
+ * @end:	offset in bytes where the range ends (inclusive)
+ * @sync_mode:	enable synchronous operation
+ *
+ * Start writeback against all of a mapping's dirty pages that lie
+ * within the byte offsets <start, end> inclusive.
+ *
+ * If sync_mode is WB_SYNC_ALL then this is a "data integrity" operation, as
+ * opposed to a regular memory cleansing writeback.  The difference between
+ * these two operations is that if a dirty page/buffer is encountered, it must
+ * be waited upon, and not just skipped over.
+ */
+int btrfs_fdatawrite_range(struct address_space *mapping, loff_t start,
+			   loff_t end, int sync_mode)
+{
+	struct writeback_control wbc = {
+		.sync_mode = sync_mode,
+		.nr_to_write = mapping->nrpages * 2,
+		.range_start = start,
+		.range_end = end,
+		.for_writepages = 1,
+	};
+	return btrfs_writepages(mapping, &wbc);
+}
+
+/**
+ * taken from mm/filemap.c because it isn't exported
+ *
+ * wait_on_page_writeback_range - wait for writeback to complete
+ * @mapping:	target address_space
+ * @start:	beginning page index
+ * @end:	ending page index
+ *
+ * Wait for writeback to complete against pages indexed by start->end
+ * inclusive
+ */
+int btrfs_wait_on_page_writeback_range(struct address_space *mapping,
+				       pgoff_t start, pgoff_t end)
+{
+	struct pagevec pvec;
+	int nr_pages;
+	int ret = 0;
+	pgoff_t index;
+
+	if (end < start)
+		return 0;
+
+	pagevec_init(&pvec, 0);
+	index = start;
+	while ((index <= end) &&
+			(nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+			PAGECACHE_TAG_WRITEBACK,
+			min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1)) != 0) {
+		unsigned i;
+
+		for (i = 0; i < nr_pages; i++) {
+			struct page *page = pvec.pages[i];
+
+			/* until radix tree lookup accepts end_index */
+			if (page->index > end)
+				continue;
+
+			wait_on_page_writeback(page);
+			if (PageError(page))
+				ret = -EIO;
+		}
+		pagevec_release(&pvec);
+		cond_resched();
+	}
+
+	/* Check for outstanding write errors */
+	if (test_and_clear_bit(AS_ENOSPC, &mapping->flags))
+		ret = -ENOSPC;
+	if (test_and_clear_bit(AS_EIO, &mapping->flags))
+		ret = -EIO;
+
+	return ret;
+}
diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h
new file mode 100644
index 00000000000..ab66d5e8d6d
--- /dev/null
+++ b/fs/btrfs/ordered-data.h
@@ -0,0 +1,158 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_ORDERED_DATA__
+#define __BTRFS_ORDERED_DATA__
+
+/* one of these per inode */
+struct btrfs_ordered_inode_tree {
+	struct mutex mutex;
+	struct rb_root tree;
+	struct rb_node *last;
+};
+
+/*
+ * these are used to collect checksums done just before bios submission.
+ * They are attached via a list into the ordered extent, and
+ * checksum items are inserted into the tree after all the blocks in
+ * the ordered extent are on disk
+ */
+struct btrfs_sector_sum {
+	/* bytenr on disk */
+	u64 bytenr;
+	u32 sum;
+};
+
+struct btrfs_ordered_sum {
+	/* bytenr is the start of this extent on disk */
+	u64 bytenr;
+
+	/*
+	 * this is the length in bytes covered by the sums array below.
+	 */
+	unsigned long len;
+	struct list_head list;
+	/* last field is a variable length array of btrfs_sector_sums */
+	struct btrfs_sector_sum sums[];
+};
+
+/*
+ * bits for the flags field:
+ *
+ * BTRFS_ORDERED_IO_DONE is set when all of the blocks are written.
+ * It is used to make sure metadata is inserted into the tree only once
+ * per extent.
+ *
+ * BTRFS_ORDERED_COMPLETE is set when the extent is removed from the
+ * rbtree, just before waking any waiters.  It is used to indicate the
+ * IO is done and any metadata is inserted into the tree.
+ */
+#define BTRFS_ORDERED_IO_DONE 0 /* set when all the pages are written */
+
+#define BTRFS_ORDERED_COMPLETE 1 /* set when removed from the tree */
+
+#define BTRFS_ORDERED_NOCOW 2 /* set when we want to write in place */
+
+#define BTRFS_ORDERED_COMPRESSED 3 /* writing a compressed extent */
+
+#define BTRFS_ORDERED_PREALLOC 4 /* set when writing to prealloced extent */
+
+struct btrfs_ordered_extent {
+	/* logical offset in the file */
+	u64 file_offset;
+
+	/* disk byte number */
+	u64 start;
+
+	/* ram length of the extent in bytes */
+	u64 len;
+
+	/* extent length on disk */
+	u64 disk_len;
+
+	/* flags (described above) */
+	unsigned long flags;
+
+	/* reference count */
+	atomic_t refs;
+
+	/* the inode we belong to */
+	struct inode *inode;
+
+	/* list of checksums for insertion when the extent io is done */
+	struct list_head list;
+
+	/* used to wait for the BTRFS_ORDERED_COMPLETE bit */
+	wait_queue_head_t wait;
+
+	/* our friendly rbtree entry */
+	struct rb_node rb_node;
+
+	/* a per root list of all the pending ordered extents */
+	struct list_head root_extent_list;
+};
+
+
+/*
+ * calculates the total size you need to allocate for an ordered sum
+ * structure spanning 'bytes' in the file
+ */
+static inline int btrfs_ordered_sum_size(struct btrfs_root *root,
+					 unsigned long bytes)
+{
+	unsigned long num_sectors = (bytes + root->sectorsize - 1) /
+		root->sectorsize;
+	num_sectors++;
+	return sizeof(struct btrfs_ordered_sum) +
+		num_sectors * sizeof(struct btrfs_sector_sum);
+}
+
+static inline void
+btrfs_ordered_inode_tree_init(struct btrfs_ordered_inode_tree *t)
+{
+	mutex_init(&t->mutex);
+	t->tree.rb_node = NULL;
+	t->last = NULL;
+}
+
+int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry);
+int btrfs_remove_ordered_extent(struct inode *inode,
+				struct btrfs_ordered_extent *entry);
+int btrfs_dec_test_ordered_pending(struct inode *inode,
+				       u64 file_offset, u64 io_size);
+int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
+			     u64 start, u64 len, u64 disk_len, int tyep);
+int btrfs_add_ordered_sum(struct inode *inode,
+			  struct btrfs_ordered_extent *entry,
+			  struct btrfs_ordered_sum *sum);
+struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode,
+							 u64 file_offset);
+void btrfs_start_ordered_extent(struct inode *inode,
+				struct btrfs_ordered_extent *entry, int wait);
+int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len);
+struct btrfs_ordered_extent *
+btrfs_lookup_first_ordered_extent(struct inode * inode, u64 file_offset);
+int btrfs_ordered_update_i_size(struct inode *inode,
+				struct btrfs_ordered_extent *ordered);
+int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum);
+int btrfs_wait_on_page_writeback_range(struct address_space *mapping,
+				       pgoff_t start, pgoff_t end);
+int btrfs_fdatawrite_range(struct address_space *mapping, loff_t start,
+			   loff_t end, int sync_mode);
+int btrfs_wait_ordered_extents(struct btrfs_root *root, int nocow_only);
+#endif
diff --git a/fs/btrfs/orphan.c b/fs/btrfs/orphan.c
new file mode 100644
index 00000000000..3c0d52af4f8
--- /dev/null
+++ b/fs/btrfs/orphan.c
@@ -0,0 +1,67 @@
+/*
+ * Copyright (C) 2008 Red Hat.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+
+int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, u64 offset)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	int ret = 0;
+
+	key.objectid = BTRFS_ORPHAN_OBJECTID;
+	btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
+	key.offset = offset;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, u64 offset)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	int ret = 0;
+
+	key.objectid = BTRFS_ORPHAN_OBJECTID;
+	btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
+	key.offset = offset;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret)
+		goto out;
+
+	ret = btrfs_del_item(trans, root, path);
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c
new file mode 100644
index 00000000000..5f8f218c100
--- /dev/null
+++ b/fs/btrfs/print-tree.c
@@ -0,0 +1,216 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "print-tree.h"
+
+static void print_chunk(struct extent_buffer *eb, struct btrfs_chunk *chunk)
+{
+	int num_stripes = btrfs_chunk_num_stripes(eb, chunk);
+	int i;
+	printk(KERN_INFO "\t\tchunk length %llu owner %llu type %llu "
+	       "num_stripes %d\n",
+	       (unsigned long long)btrfs_chunk_length(eb, chunk),
+	       (unsigned long long)btrfs_chunk_owner(eb, chunk),
+	       (unsigned long long)btrfs_chunk_type(eb, chunk),
+	       num_stripes);
+	for (i = 0 ; i < num_stripes ; i++) {
+		printk(KERN_INFO "\t\t\tstripe %d devid %llu offset %llu\n", i,
+		      (unsigned long long)btrfs_stripe_devid_nr(eb, chunk, i),
+		      (unsigned long long)btrfs_stripe_offset_nr(eb, chunk, i));
+	}
+}
+static void print_dev_item(struct extent_buffer *eb,
+			   struct btrfs_dev_item *dev_item)
+{
+	printk(KERN_INFO "\t\tdev item devid %llu "
+	       "total_bytes %llu bytes used %llu\n",
+	       (unsigned long long)btrfs_device_id(eb, dev_item),
+	       (unsigned long long)btrfs_device_total_bytes(eb, dev_item),
+	       (unsigned long long)btrfs_device_bytes_used(eb, dev_item));
+}
+void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
+{
+	int i;
+	u32 nr = btrfs_header_nritems(l);
+	struct btrfs_item *item;
+	struct btrfs_extent_item *ei;
+	struct btrfs_root_item *ri;
+	struct btrfs_dir_item *di;
+	struct btrfs_inode_item *ii;
+	struct btrfs_block_group_item *bi;
+	struct btrfs_file_extent_item *fi;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_extent_ref *ref;
+	struct btrfs_dev_extent *dev_extent;
+	u32 type;
+
+	printk(KERN_INFO "leaf %llu total ptrs %d free space %d\n",
+		(unsigned long long)btrfs_header_bytenr(l), nr,
+		btrfs_leaf_free_space(root, l));
+	for (i = 0 ; i < nr ; i++) {
+		item = btrfs_item_nr(l, i);
+		btrfs_item_key_to_cpu(l, &key, i);
+		type = btrfs_key_type(&key);
+		printk(KERN_INFO "\titem %d key (%llu %x %llu) itemoff %d "
+		       "itemsize %d\n",
+			i,
+			(unsigned long long)key.objectid, type,
+			(unsigned long long)key.offset,
+			btrfs_item_offset(l, item), btrfs_item_size(l, item));
+		switch (type) {
+		case BTRFS_INODE_ITEM_KEY:
+			ii = btrfs_item_ptr(l, i, struct btrfs_inode_item);
+			printk(KERN_INFO "\t\tinode generation %llu size %llu "
+			       "mode %o\n",
+			       (unsigned long long)
+			       btrfs_inode_generation(l, ii),
+			      (unsigned long long)btrfs_inode_size(l, ii),
+			       btrfs_inode_mode(l, ii));
+			break;
+		case BTRFS_DIR_ITEM_KEY:
+			di = btrfs_item_ptr(l, i, struct btrfs_dir_item);
+			btrfs_dir_item_key_to_cpu(l, di, &found_key);
+			printk(KERN_INFO "\t\tdir oid %llu type %u\n",
+				(unsigned long long)found_key.objectid,
+				btrfs_dir_type(l, di));
+			break;
+		case BTRFS_ROOT_ITEM_KEY:
+			ri = btrfs_item_ptr(l, i, struct btrfs_root_item);
+			printk(KERN_INFO "\t\troot data bytenr %llu refs %u\n",
+				(unsigned long long)
+				btrfs_disk_root_bytenr(l, ri),
+				btrfs_disk_root_refs(l, ri));
+			break;
+		case BTRFS_EXTENT_ITEM_KEY:
+			ei = btrfs_item_ptr(l, i, struct btrfs_extent_item);
+			printk(KERN_INFO "\t\textent data refs %u\n",
+				btrfs_extent_refs(l, ei));
+			break;
+		case BTRFS_EXTENT_REF_KEY:
+			ref = btrfs_item_ptr(l, i, struct btrfs_extent_ref);
+			printk(KERN_INFO "\t\textent back ref root %llu "
+			       "gen %llu owner %llu num_refs %lu\n",
+			       (unsigned long long)btrfs_ref_root(l, ref),
+			       (unsigned long long)btrfs_ref_generation(l, ref),
+			       (unsigned long long)btrfs_ref_objectid(l, ref),
+			       (unsigned long)btrfs_ref_num_refs(l, ref));
+			break;
+
+		case BTRFS_EXTENT_DATA_KEY:
+			fi = btrfs_item_ptr(l, i,
+					    struct btrfs_file_extent_item);
+			if (btrfs_file_extent_type(l, fi) ==
+			    BTRFS_FILE_EXTENT_INLINE) {
+				printk(KERN_INFO "\t\tinline extent data "
+				       "size %u\n",
+				       btrfs_file_extent_inline_len(l, fi));
+				break;
+			}
+			printk(KERN_INFO "\t\textent data disk bytenr %llu "
+			       "nr %llu\n",
+			       (unsigned long long)
+			       btrfs_file_extent_disk_bytenr(l, fi),
+			       (unsigned long long)
+			       btrfs_file_extent_disk_num_bytes(l, fi));
+			printk(KERN_INFO "\t\textent data offset %llu "
+			       "nr %llu ram %llu\n",
+			       (unsigned long long)
+			       btrfs_file_extent_offset(l, fi),
+			       (unsigned long long)
+			       btrfs_file_extent_num_bytes(l, fi),
+			       (unsigned long long)
+			       btrfs_file_extent_ram_bytes(l, fi));
+			break;
+		case BTRFS_BLOCK_GROUP_ITEM_KEY:
+			bi = btrfs_item_ptr(l, i,
+					    struct btrfs_block_group_item);
+			printk(KERN_INFO "\t\tblock group used %llu\n",
+			       (unsigned long long)
+			       btrfs_disk_block_group_used(l, bi));
+			break;
+		case BTRFS_CHUNK_ITEM_KEY:
+			print_chunk(l, btrfs_item_ptr(l, i,
+						      struct btrfs_chunk));
+			break;
+		case BTRFS_DEV_ITEM_KEY:
+			print_dev_item(l, btrfs_item_ptr(l, i,
+					struct btrfs_dev_item));
+			break;
+		case BTRFS_DEV_EXTENT_KEY:
+			dev_extent = btrfs_item_ptr(l, i,
+						    struct btrfs_dev_extent);
+			printk(KERN_INFO "\t\tdev extent chunk_tree %llu\n"
+			       "\t\tchunk objectid %llu chunk offset %llu "
+			       "length %llu\n",
+			       (unsigned long long)
+			       btrfs_dev_extent_chunk_tree(l, dev_extent),
+			       (unsigned long long)
+			       btrfs_dev_extent_chunk_objectid(l, dev_extent),
+			       (unsigned long long)
+			       btrfs_dev_extent_chunk_offset(l, dev_extent),
+			       (unsigned long long)
+			       btrfs_dev_extent_length(l, dev_extent));
+		};
+	}
+}
+
+void btrfs_print_tree(struct btrfs_root *root, struct extent_buffer *c)
+{
+	int i; u32 nr;
+	struct btrfs_key key;
+	int level;
+
+	if (!c)
+		return;
+	nr = btrfs_header_nritems(c);
+	level = btrfs_header_level(c);
+	if (level == 0) {
+		btrfs_print_leaf(root, c);
+		return;
+	}
+	printk(KERN_INFO "node %llu level %d total ptrs %d free spc %u\n",
+	       (unsigned long long)btrfs_header_bytenr(c),
+	       btrfs_header_level(c), nr,
+	       (u32)BTRFS_NODEPTRS_PER_BLOCK(root) - nr);
+	for (i = 0; i < nr; i++) {
+		btrfs_node_key_to_cpu(c, &key, i);
+		printk(KERN_INFO "\tkey %d (%llu %u %llu) block %llu\n",
+		       i,
+		       (unsigned long long)key.objectid,
+		       key.type,
+		       (unsigned long long)key.offset,
+		       (unsigned long long)btrfs_node_blockptr(c, i));
+	}
+	for (i = 0; i < nr; i++) {
+		struct extent_buffer *next = read_tree_block(root,
+					btrfs_node_blockptr(c, i),
+					btrfs_level_size(root, level - 1),
+					btrfs_node_ptr_generation(c, i));
+		if (btrfs_is_leaf(next) &&
+		    btrfs_header_level(c) != 1)
+			BUG();
+		if (btrfs_header_level(next) !=
+			btrfs_header_level(c) - 1)
+			BUG();
+		btrfs_print_tree(root, next);
+		free_extent_buffer(next);
+	}
+}
diff --git a/fs/btrfs/print-tree.h b/fs/btrfs/print-tree.h
new file mode 100644
index 00000000000..da75efe534d
--- /dev/null
+++ b/fs/btrfs/print-tree.h
@@ -0,0 +1,23 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __PRINT_TREE_
+#define __PRINT_TREE_
+void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l);
+void btrfs_print_tree(struct btrfs_root *root, struct extent_buffer *t);
+#endif
diff --git a/fs/btrfs/ref-cache.c b/fs/btrfs/ref-cache.c
new file mode 100644
index 00000000000..6f0acc4c9ea
--- /dev/null
+++ b/fs/btrfs/ref-cache.c
@@ -0,0 +1,230 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include "ctree.h"
+#include "ref-cache.h"
+#include "transaction.h"
+
+/*
+ * leaf refs are used to cache the information about which extents
+ * a given leaf has references on.  This allows us to process that leaf
+ * in btrfs_drop_snapshot without needing to read it back from disk.
+ */
+
+/*
+ * kmalloc a leaf reference struct and update the counters for the
+ * total ref cache size
+ */
+struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root,
+					    int nr_extents)
+{
+	struct btrfs_leaf_ref *ref;
+	size_t size = btrfs_leaf_ref_size(nr_extents);
+
+	ref = kmalloc(size, GFP_NOFS);
+	if (ref) {
+		spin_lock(&root->fs_info->ref_cache_lock);
+		root->fs_info->total_ref_cache_size += size;
+		spin_unlock(&root->fs_info->ref_cache_lock);
+
+		memset(ref, 0, sizeof(*ref));
+		atomic_set(&ref->usage, 1);
+		INIT_LIST_HEAD(&ref->list);
+	}
+	return ref;
+}
+
+/*
+ * free a leaf reference struct and update the counters for the
+ * total ref cache size
+ */
+void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref)
+{
+	if (!ref)
+		return;
+	WARN_ON(atomic_read(&ref->usage) == 0);
+	if (atomic_dec_and_test(&ref->usage)) {
+		size_t size = btrfs_leaf_ref_size(ref->nritems);
+
+		BUG_ON(ref->in_tree);
+		kfree(ref);
+
+		spin_lock(&root->fs_info->ref_cache_lock);
+		root->fs_info->total_ref_cache_size -= size;
+		spin_unlock(&root->fs_info->ref_cache_lock);
+	}
+}
+
+static struct rb_node *tree_insert(struct rb_root *root, u64 bytenr,
+				   struct rb_node *node)
+{
+	struct rb_node **p = &root->rb_node;
+	struct rb_node *parent = NULL;
+	struct btrfs_leaf_ref *entry;
+
+	while (*p) {
+		parent = *p;
+		entry = rb_entry(parent, struct btrfs_leaf_ref, rb_node);
+
+		if (bytenr < entry->bytenr)
+			p = &(*p)->rb_left;
+		else if (bytenr > entry->bytenr)
+			p = &(*p)->rb_right;
+		else
+			return parent;
+	}
+
+	entry = rb_entry(node, struct btrfs_leaf_ref, rb_node);
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+	return NULL;
+}
+
+static struct rb_node *tree_search(struct rb_root *root, u64 bytenr)
+{
+	struct rb_node *n = root->rb_node;
+	struct btrfs_leaf_ref *entry;
+
+	while (n) {
+		entry = rb_entry(n, struct btrfs_leaf_ref, rb_node);
+		WARN_ON(!entry->in_tree);
+
+		if (bytenr < entry->bytenr)
+			n = n->rb_left;
+		else if (bytenr > entry->bytenr)
+			n = n->rb_right;
+		else
+			return n;
+	}
+	return NULL;
+}
+
+int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen,
+			   int shared)
+{
+	struct btrfs_leaf_ref *ref = NULL;
+	struct btrfs_leaf_ref_tree *tree = root->ref_tree;
+
+	if (shared)
+		tree = &root->fs_info->shared_ref_tree;
+	if (!tree)
+		return 0;
+
+	spin_lock(&tree->lock);
+	while (!list_empty(&tree->list)) {
+		ref = list_entry(tree->list.next, struct btrfs_leaf_ref, list);
+		BUG_ON(ref->tree != tree);
+		if (ref->root_gen > max_root_gen)
+			break;
+		if (!xchg(&ref->in_tree, 0)) {
+			cond_resched_lock(&tree->lock);
+			continue;
+		}
+
+		rb_erase(&ref->rb_node, &tree->root);
+		list_del_init(&ref->list);
+
+		spin_unlock(&tree->lock);
+		btrfs_free_leaf_ref(root, ref);
+		cond_resched();
+		spin_lock(&tree->lock);
+	}
+	spin_unlock(&tree->lock);
+	return 0;
+}
+
+/*
+ * find the leaf ref for a given extent.  This returns the ref struct with
+ * a usage reference incremented
+ */
+struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root,
+					     u64 bytenr)
+{
+	struct rb_node *rb;
+	struct btrfs_leaf_ref *ref = NULL;
+	struct btrfs_leaf_ref_tree *tree = root->ref_tree;
+again:
+	if (tree) {
+		spin_lock(&tree->lock);
+		rb = tree_search(&tree->root, bytenr);
+		if (rb)
+			ref = rb_entry(rb, struct btrfs_leaf_ref, rb_node);
+		if (ref)
+			atomic_inc(&ref->usage);
+		spin_unlock(&tree->lock);
+		if (ref)
+			return ref;
+	}
+	if (tree != &root->fs_info->shared_ref_tree) {
+		tree = &root->fs_info->shared_ref_tree;
+		goto again;
+	}
+	return NULL;
+}
+
+/*
+ * add a fully filled in leaf ref struct
+ * remove all the refs older than a given root generation
+ */
+int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref,
+		       int shared)
+{
+	int ret = 0;
+	struct rb_node *rb;
+	struct btrfs_leaf_ref_tree *tree = root->ref_tree;
+
+	if (shared)
+		tree = &root->fs_info->shared_ref_tree;
+
+	spin_lock(&tree->lock);
+	rb = tree_insert(&tree->root, ref->bytenr, &ref->rb_node);
+	if (rb) {
+		ret = -EEXIST;
+	} else {
+		atomic_inc(&ref->usage);
+		ref->tree = tree;
+		ref->in_tree = 1;
+		list_add_tail(&ref->list, &tree->list);
+	}
+	spin_unlock(&tree->lock);
+	return ret;
+}
+
+/*
+ * remove a single leaf ref from the tree.  This drops the ref held by the tree
+ * only
+ */
+int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref)
+{
+	struct btrfs_leaf_ref_tree *tree;
+
+	if (!xchg(&ref->in_tree, 0))
+		return 0;
+
+	tree = ref->tree;
+	spin_lock(&tree->lock);
+
+	rb_erase(&ref->rb_node, &tree->root);
+	list_del_init(&ref->list);
+
+	spin_unlock(&tree->lock);
+
+	btrfs_free_leaf_ref(root, ref);
+	return 0;
+}
diff --git a/fs/btrfs/ref-cache.h b/fs/btrfs/ref-cache.h
new file mode 100644
index 00000000000..16f3183d7c5
--- /dev/null
+++ b/fs/btrfs/ref-cache.h
@@ -0,0 +1,77 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+#ifndef __REFCACHE__
+#define __REFCACHE__
+
+struct btrfs_extent_info {
+	/* bytenr and num_bytes find the extent in the extent allocation tree */
+	u64 bytenr;
+	u64 num_bytes;
+
+	/* objectid and offset find the back reference for the file */
+	u64 objectid;
+	u64 offset;
+};
+
+struct btrfs_leaf_ref {
+	struct rb_node rb_node;
+	struct btrfs_leaf_ref_tree *tree;
+	int in_tree;
+	atomic_t usage;
+
+	u64 root_gen;
+	u64 bytenr;
+	u64 owner;
+	u64 generation;
+	int nritems;
+
+	struct list_head list;
+	struct btrfs_extent_info extents[];
+};
+
+static inline size_t btrfs_leaf_ref_size(int nr_extents)
+{
+	return sizeof(struct btrfs_leaf_ref) +
+	       sizeof(struct btrfs_extent_info) * nr_extents;
+}
+
+static inline void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree)
+{
+	tree->root.rb_node = NULL;
+	INIT_LIST_HEAD(&tree->list);
+	spin_lock_init(&tree->lock);
+}
+
+static inline int btrfs_leaf_ref_tree_empty(struct btrfs_leaf_ref_tree *tree)
+{
+	return RB_EMPTY_ROOT(&tree->root);
+}
+
+void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree);
+struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root,
+					    int nr_extents);
+void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref);
+struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root,
+					     u64 bytenr);
+int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref,
+		       int shared);
+int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen,
+			   int shared);
+int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref);
+
+#endif
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
new file mode 100644
index 00000000000..b48650de447
--- /dev/null
+++ b/fs/btrfs/root-tree.c
@@ -0,0 +1,366 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include "ctree.h"
+#include "transaction.h"
+#include "disk-io.h"
+#include "print-tree.h"
+
+/*
+ *  search forward for a root, starting with objectid 'search_start'
+ *  if a root key is found, the objectid we find is filled into 'found_objectid'
+ *  and 0 is returned.  < 0 is returned on error, 1 if there is nothing
+ *  left in the tree.
+ */
+int btrfs_search_root(struct btrfs_root *root, u64 search_start,
+		      u64 *found_objectid)
+{
+	struct btrfs_path *path;
+	struct btrfs_key search_key;
+	int ret;
+
+	root = root->fs_info->tree_root;
+	search_key.objectid = search_start;
+	search_key.type = (u8)-1;
+	search_key.offset = (u64)-1;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+again:
+	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	if (ret == 0) {
+		ret = 1;
+		goto out;
+	}
+	if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
+		ret = btrfs_next_leaf(root, path);
+		if (ret)
+			goto out;
+	}
+	btrfs_item_key_to_cpu(path->nodes[0], &search_key, path->slots[0]);
+	if (search_key.type != BTRFS_ROOT_ITEM_KEY) {
+		search_key.offset++;
+		btrfs_release_path(root, path);
+		goto again;
+	}
+	ret = 0;
+	*found_objectid = search_key.objectid;
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * lookup the root with the highest offset for a given objectid.  The key we do
+ * find is copied into 'key'.  If we find something return 0, otherwise 1, < 0
+ * on error.
+ */
+int btrfs_find_last_root(struct btrfs_root *root, u64 objectid,
+			struct btrfs_root_item *item, struct btrfs_key *key)
+{
+	struct btrfs_path *path;
+	struct btrfs_key search_key;
+	struct btrfs_key found_key;
+	struct extent_buffer *l;
+	int ret;
+	int slot;
+
+	search_key.objectid = objectid;
+	search_key.type = BTRFS_ROOT_ITEM_KEY;
+	search_key.offset = (u64)-1;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+
+	BUG_ON(ret == 0);
+	l = path->nodes[0];
+	BUG_ON(path->slots[0] == 0);
+	slot = path->slots[0] - 1;
+	btrfs_item_key_to_cpu(l, &found_key, slot);
+	if (found_key.objectid != objectid) {
+		ret = 1;
+		goto out;
+	}
+	read_extent_buffer(l, item, btrfs_item_ptr_offset(l, slot),
+			   sizeof(*item));
+	memcpy(key, &found_key, sizeof(found_key));
+	ret = 0;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * copy the data in 'item' into the btree
+ */
+int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, struct btrfs_root_item
+		      *item)
+{
+	struct btrfs_path *path;
+	struct extent_buffer *l;
+	int ret;
+	int slot;
+	unsigned long ptr;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	ret = btrfs_search_slot(trans, root, key, path, 0, 1);
+	if (ret < 0)
+		goto out;
+
+	if (ret != 0) {
+		btrfs_print_leaf(root, path->nodes[0]);
+		printk(KERN_CRIT "unable to update root key %llu %u %llu\n",
+		       (unsigned long long)key->objectid, key->type,
+		       (unsigned long long)key->offset);
+		BUG_ON(1);
+	}
+
+	l = path->nodes[0];
+	slot = path->slots[0];
+	ptr = btrfs_item_ptr_offset(l, slot);
+	write_extent_buffer(l, item, ptr, sizeof(*item));
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+out:
+	btrfs_release_path(root, path);
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root
+		      *root, struct btrfs_key *key, struct btrfs_root_item
+		      *item)
+{
+	int ret;
+	ret = btrfs_insert_item(trans, root, key, item, sizeof(*item));
+	return ret;
+}
+
+/*
+ * at mount time we want to find all the old transaction snapshots that were in
+ * the process of being deleted if we crashed.  This is any root item with an
+ * offset lower than the latest root.  They need to be queued for deletion to
+ * finish what was happening when we crashed.
+ */
+int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid,
+			  struct btrfs_root *latest)
+{
+	struct btrfs_root *dead_root;
+	struct btrfs_item *item;
+	struct btrfs_root_item *ri;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_path *path;
+	int ret;
+	u32 nritems;
+	struct extent_buffer *leaf;
+	int slot;
+
+	key.objectid = objectid;
+	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+	key.offset = 0;
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+again:
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto err;
+	while (1) {
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		slot = path->slots[0];
+		if (slot >= nritems) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret)
+				break;
+			leaf = path->nodes[0];
+			nritems = btrfs_header_nritems(leaf);
+			slot = path->slots[0];
+		}
+		item = btrfs_item_nr(leaf, slot);
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (btrfs_key_type(&key) != BTRFS_ROOT_ITEM_KEY)
+			goto next;
+
+		if (key.objectid < objectid)
+			goto next;
+
+		if (key.objectid > objectid)
+			break;
+
+		ri = btrfs_item_ptr(leaf, slot, struct btrfs_root_item);
+		if (btrfs_disk_root_refs(leaf, ri) != 0)
+			goto next;
+
+		memcpy(&found_key, &key, sizeof(key));
+		key.offset++;
+		btrfs_release_path(root, path);
+		dead_root =
+			btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
+						    &found_key);
+		if (IS_ERR(dead_root)) {
+			ret = PTR_ERR(dead_root);
+			goto err;
+		}
+
+		if (objectid == BTRFS_TREE_RELOC_OBJECTID)
+			ret = btrfs_add_dead_reloc_root(dead_root);
+		else
+			ret = btrfs_add_dead_root(dead_root, latest);
+		if (ret)
+			goto err;
+		goto again;
+next:
+		slot++;
+		path->slots[0]++;
+	}
+	ret = 0;
+err:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/* drop the root item for 'key' from 'root' */
+int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+		   struct btrfs_key *key)
+{
+	struct btrfs_path *path;
+	int ret;
+	u32 refs;
+	struct btrfs_root_item *ri;
+	struct extent_buffer *leaf;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+	ret = btrfs_search_slot(trans, root, key, path, -1, 1);
+	if (ret < 0)
+		goto out;
+
+	BUG_ON(ret != 0);
+	leaf = path->nodes[0];
+	ri = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_item);
+
+	refs = btrfs_disk_root_refs(leaf, ri);
+	BUG_ON(refs != 0);
+	ret = btrfs_del_item(trans, root, path);
+out:
+	btrfs_release_path(root, path);
+	btrfs_free_path(path);
+	return ret;
+}
+
+#if 0 /* this will get used when snapshot deletion is implemented */
+int btrfs_del_root_ref(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *tree_root,
+		       u64 root_id, u8 type, u64 ref_id)
+{
+	struct btrfs_key key;
+	int ret;
+	struct btrfs_path *path;
+
+	path = btrfs_alloc_path();
+
+	key.objectid = root_id;
+	key.type = type;
+	key.offset = ref_id;
+
+	ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1);
+	BUG_ON(ret);
+
+	ret = btrfs_del_item(trans, tree_root, path);
+	BUG_ON(ret);
+
+	btrfs_free_path(path);
+	return ret;
+}
+#endif
+
+int btrfs_find_root_ref(struct btrfs_root *tree_root,
+		   struct btrfs_path *path,
+		   u64 root_id, u64 ref_id)
+{
+	struct btrfs_key key;
+	int ret;
+
+	key.objectid = root_id;
+	key.type = BTRFS_ROOT_REF_KEY;
+	key.offset = ref_id;
+
+	ret = btrfs_search_slot(NULL, tree_root, &key, path, 0, 0);
+	return ret;
+}
+
+
+/*
+ * add a btrfs_root_ref item.  type is either BTRFS_ROOT_REF_KEY
+ * or BTRFS_ROOT_BACKREF_KEY.
+ *
+ * The dirid, sequence, name and name_len refer to the directory entry
+ * that is referencing the root.
+ *
+ * For a forward ref, the root_id is the id of the tree referencing
+ * the root and ref_id is the id of the subvol  or snapshot.
+ *
+ * For a back ref the root_id is the id of the subvol or snapshot and
+ * ref_id is the id of the tree referencing it.
+ */
+int btrfs_add_root_ref(struct btrfs_trans_handle *trans,
+		       struct btrfs_root *tree_root,
+		       u64 root_id, u8 type, u64 ref_id,
+		       u64 dirid, u64 sequence,
+		       const char *name, int name_len)
+{
+	struct btrfs_key key;
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_root_ref *ref;
+	struct extent_buffer *leaf;
+	unsigned long ptr;
+
+
+	path = btrfs_alloc_path();
+
+	key.objectid = root_id;
+	key.type = type;
+	key.offset = ref_id;
+
+	ret = btrfs_insert_empty_item(trans, tree_root, path, &key,
+				      sizeof(*ref) + name_len);
+	BUG_ON(ret);
+
+	leaf = path->nodes[0];
+	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
+	btrfs_set_root_ref_dirid(leaf, ref, dirid);
+	btrfs_set_root_ref_sequence(leaf, ref, sequence);
+	btrfs_set_root_ref_name_len(leaf, ref, name_len);
+	ptr = (unsigned long)(ref + 1);
+	write_extent_buffer(leaf, name, ptr, name_len);
+	btrfs_mark_buffer_dirty(leaf);
+
+	btrfs_free_path(path);
+	return ret;
+}
diff --git a/fs/btrfs/struct-funcs.c b/fs/btrfs/struct-funcs.c
new file mode 100644
index 00000000000..c0f7ecaf1e7
--- /dev/null
+++ b/fs/btrfs/struct-funcs.c
@@ -0,0 +1,139 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/highmem.h>
+
+/* this is some deeply nasty code.  ctree.h has a different
+ * definition for this BTRFS_SETGET_FUNCS macro, behind a #ifndef
+ *
+ * The end result is that anyone who #includes ctree.h gets a
+ * declaration for the btrfs_set_foo functions and btrfs_foo functions
+ *
+ * This file declares the macros and then #includes ctree.h, which results
+ * in cpp creating the function here based on the template below.
+ *
+ * These setget functions do all the extent_buffer related mapping
+ * required to efficiently read and write specific fields in the extent
+ * buffers.  Every pointer to metadata items in btrfs is really just
+ * an unsigned long offset into the extent buffer which has been
+ * cast to a specific type.  This gives us all the gcc type checking.
+ *
+ * The extent buffer api is used to do all the kmapping and page
+ * spanning work required to get extent buffers in highmem and have
+ * a metadata blocksize different from the page size.
+ *
+ * The macro starts with a simple function prototype declaration so that
+ * sparse won't complain about it being static.
+ */
+
+#define BTRFS_SETGET_FUNCS(name, type, member, bits)			\
+u##bits btrfs_##name(struct extent_buffer *eb, type *s);		\
+void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val);	\
+u##bits btrfs_##name(struct extent_buffer *eb,				\
+				   type *s)				\
+{									\
+	unsigned long part_offset = (unsigned long)s;			\
+	unsigned long offset = part_offset + offsetof(type, member);	\
+	type *p;							\
+	/* ugly, but we want the fast path here */			\
+	if (eb->map_token && offset >= eb->map_start &&			\
+	    offset + sizeof(((type *)0)->member) <= eb->map_start +	\
+	    eb->map_len) {						\
+		p = (type *)(eb->kaddr + part_offset - eb->map_start);	\
+		return le##bits##_to_cpu(p->member);			\
+	}								\
+	{								\
+		int err;						\
+		char *map_token;					\
+		char *kaddr;						\
+		int unmap_on_exit = (eb->map_token == NULL);		\
+		unsigned long map_start;				\
+		unsigned long map_len;					\
+		u##bits res;						\
+		err = map_extent_buffer(eb, offset,			\
+				sizeof(((type *)0)->member),		\
+				&map_token, &kaddr,			\
+				&map_start, &map_len, KM_USER1);	\
+		if (err) {						\
+			__le##bits leres;				\
+			read_eb_member(eb, s, type, member, &leres);	\
+			return le##bits##_to_cpu(leres);		\
+		}							\
+		p = (type *)(kaddr + part_offset - map_start);		\
+		res = le##bits##_to_cpu(p->member);			\
+		if (unmap_on_exit)					\
+			unmap_extent_buffer(eb, map_token, KM_USER1);	\
+		return res;						\
+	}								\
+}									\
+void btrfs_set_##name(struct extent_buffer *eb,				\
+				    type *s, u##bits val)		\
+{									\
+	unsigned long part_offset = (unsigned long)s;			\
+	unsigned long offset = part_offset + offsetof(type, member);	\
+	type *p;							\
+	/* ugly, but we want the fast path here */			\
+	if (eb->map_token && offset >= eb->map_start &&			\
+	    offset + sizeof(((type *)0)->member) <= eb->map_start +	\
+	    eb->map_len) {						\
+		p = (type *)(eb->kaddr + part_offset - eb->map_start);	\
+		p->member = cpu_to_le##bits(val);			\
+		return;							\
+	}								\
+	{								\
+		int err;						\
+		char *map_token;					\
+		char *kaddr;						\
+		int unmap_on_exit = (eb->map_token == NULL);		\
+		unsigned long map_start;				\
+		unsigned long map_len;					\
+		err = map_extent_buffer(eb, offset,			\
+				sizeof(((type *)0)->member),		\
+				&map_token, &kaddr,			\
+				&map_start, &map_len, KM_USER1);	\
+		if (err) {						\
+			__le##bits val2;				\
+			val2 = cpu_to_le##bits(val);			\
+			write_eb_member(eb, s, type, member, &val2);	\
+			return;						\
+		}							\
+		p = (type *)(kaddr + part_offset - map_start);		\
+		p->member = cpu_to_le##bits(val);			\
+		if (unmap_on_exit)					\
+			unmap_extent_buffer(eb, map_token, KM_USER1);	\
+	}								\
+}
+
+#include "ctree.h"
+
+void btrfs_node_key(struct extent_buffer *eb,
+		    struct btrfs_disk_key *disk_key, int nr)
+{
+	unsigned long ptr = btrfs_node_key_ptr_offset(nr);
+	if (eb->map_token && ptr >= eb->map_start &&
+	    ptr + sizeof(*disk_key) <= eb->map_start + eb->map_len) {
+		memcpy(disk_key, eb->kaddr + ptr - eb->map_start,
+			sizeof(*disk_key));
+		return;
+	} else if (eb->map_token) {
+		unmap_extent_buffer(eb, eb->map_token, KM_USER1);
+		eb->map_token = NULL;
+	}
+	read_eb_member(eb, (struct btrfs_key_ptr *)ptr,
+		       struct btrfs_key_ptr, key, disk_key);
+}
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
new file mode 100644
index 00000000000..0a14b495532
--- /dev/null
+++ b/fs/btrfs/super.c
@@ -0,0 +1,722 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/blkdev.h>
+#include <linux/module.h>
+#include <linux/buffer_head.h>
+#include <linux/fs.h>
+#include <linux/pagemap.h>
+#include <linux/highmem.h>
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/smp_lock.h>
+#include <linux/backing-dev.h>
+#include <linux/mount.h>
+#include <linux/mpage.h>
+#include <linux/swap.h>
+#include <linux/writeback.h>
+#include <linux/statfs.h>
+#include <linux/compat.h>
+#include <linux/parser.h>
+#include <linux/ctype.h>
+#include <linux/namei.h>
+#include <linux/miscdevice.h>
+#include <linux/version.h>
+#include "compat.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "btrfs_inode.h"
+#include "ioctl.h"
+#include "print-tree.h"
+#include "xattr.h"
+#include "volumes.h"
+#include "version.h"
+#include "export.h"
+#include "compression.h"
+
+#define BTRFS_SUPER_MAGIC 0x9123683E
+
+static struct super_operations btrfs_super_ops;
+
+static void btrfs_put_super(struct super_block *sb)
+{
+	struct btrfs_root *root = btrfs_sb(sb);
+	int ret;
+
+	ret = close_ctree(root);
+	sb->s_fs_info = NULL;
+}
+
+enum {
+	Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow,
+	Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier,
+	Opt_ssd, Opt_thread_pool, Opt_noacl,  Opt_compress, Opt_err,
+};
+
+static match_table_t tokens = {
+	{Opt_degraded, "degraded"},
+	{Opt_subvol, "subvol=%s"},
+	{Opt_device, "device=%s"},
+	{Opt_nodatasum, "nodatasum"},
+	{Opt_nodatacow, "nodatacow"},
+	{Opt_nobarrier, "nobarrier"},
+	{Opt_max_extent, "max_extent=%s"},
+	{Opt_max_inline, "max_inline=%s"},
+	{Opt_alloc_start, "alloc_start=%s"},
+	{Opt_thread_pool, "thread_pool=%d"},
+	{Opt_compress, "compress"},
+	{Opt_ssd, "ssd"},
+	{Opt_noacl, "noacl"},
+	{Opt_err, NULL},
+};
+
+u64 btrfs_parse_size(char *str)
+{
+	u64 res;
+	int mult = 1;
+	char *end;
+	char last;
+
+	res = simple_strtoul(str, &end, 10);
+
+	last = end[0];
+	if (isalpha(last)) {
+		last = tolower(last);
+		switch (last) {
+		case 'g':
+			mult *= 1024;
+		case 'm':
+			mult *= 1024;
+		case 'k':
+			mult *= 1024;
+		}
+		res = res * mult;
+	}
+	return res;
+}
+
+/*
+ * Regular mount options parser.  Everything that is needed only when
+ * reading in a new superblock is parsed here.
+ */
+int btrfs_parse_options(struct btrfs_root *root, char *options)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	substring_t args[MAX_OPT_ARGS];
+	char *p, *num;
+	int intarg;
+
+	if (!options)
+		return 0;
+
+	/*
+	 * strsep changes the string, duplicate it because parse_options
+	 * gets called twice
+	 */
+	options = kstrdup(options, GFP_NOFS);
+	if (!options)
+		return -ENOMEM;
+
+
+	while ((p = strsep(&options, ",")) != NULL) {
+		int token;
+		if (!*p)
+			continue;
+
+		token = match_token(p, tokens, args);
+		switch (token) {
+		case Opt_degraded:
+			printk(KERN_INFO "btrfs: allowing degraded mounts\n");
+			btrfs_set_opt(info->mount_opt, DEGRADED);
+			break;
+		case Opt_subvol:
+		case Opt_device:
+			/*
+			 * These are parsed by btrfs_parse_early_options
+			 * and can be happily ignored here.
+			 */
+			break;
+		case Opt_nodatasum:
+			printk(KERN_INFO "btrfs: setting nodatacsum\n");
+			btrfs_set_opt(info->mount_opt, NODATASUM);
+			break;
+		case Opt_nodatacow:
+			printk(KERN_INFO "btrfs: setting nodatacow\n");
+			btrfs_set_opt(info->mount_opt, NODATACOW);
+			btrfs_set_opt(info->mount_opt, NODATASUM);
+			break;
+		case Opt_compress:
+			printk(KERN_INFO "btrfs: use compression\n");
+			btrfs_set_opt(info->mount_opt, COMPRESS);
+			break;
+		case Opt_ssd:
+			printk(KERN_INFO "btrfs: use ssd allocation scheme\n");
+			btrfs_set_opt(info->mount_opt, SSD);
+			break;
+		case Opt_nobarrier:
+			printk(KERN_INFO "btrfs: turning off barriers\n");
+			btrfs_set_opt(info->mount_opt, NOBARRIER);
+			break;
+		case Opt_thread_pool:
+			intarg = 0;
+			match_int(&args[0], &intarg);
+			if (intarg) {
+				info->thread_pool_size = intarg;
+				printk(KERN_INFO "btrfs: thread pool %d\n",
+				       info->thread_pool_size);
+			}
+			break;
+		case Opt_max_extent:
+			num = match_strdup(&args[0]);
+			if (num) {
+				info->max_extent = btrfs_parse_size(num);
+				kfree(num);
+
+				info->max_extent = max_t(u64,
+					info->max_extent, root->sectorsize);
+				printk(KERN_INFO "btrfs: max_extent at %llu\n",
+				       info->max_extent);
+			}
+			break;
+		case Opt_max_inline:
+			num = match_strdup(&args[0]);
+			if (num) {
+				info->max_inline = btrfs_parse_size(num);
+				kfree(num);
+
+				if (info->max_inline) {
+					info->max_inline = max_t(u64,
+						info->max_inline,
+						root->sectorsize);
+				}
+				printk(KERN_INFO "btrfs: max_inline at %llu\n",
+					info->max_inline);
+			}
+			break;
+		case Opt_alloc_start:
+			num = match_strdup(&args[0]);
+			if (num) {
+				info->alloc_start = btrfs_parse_size(num);
+				kfree(num);
+				printk(KERN_INFO
+					"btrfs: allocations start at %llu\n",
+					info->alloc_start);
+			}
+			break;
+		case Opt_noacl:
+			root->fs_info->sb->s_flags &= ~MS_POSIXACL;
+			break;
+		default:
+			break;
+		}
+	}
+	kfree(options);
+	return 0;
+}
+
+/*
+ * Parse mount options that are required early in the mount process.
+ *
+ * All other options will be parsed on much later in the mount process and
+ * only when we need to allocate a new super block.
+ */
+static int btrfs_parse_early_options(const char *options, fmode_t flags,
+		void *holder, char **subvol_name,
+		struct btrfs_fs_devices **fs_devices)
+{
+	substring_t args[MAX_OPT_ARGS];
+	char *opts, *p;
+	int error = 0;
+
+	if (!options)
+		goto out;
+
+	/*
+	 * strsep changes the string, duplicate it because parse_options
+	 * gets called twice
+	 */
+	opts = kstrdup(options, GFP_KERNEL);
+	if (!opts)
+		return -ENOMEM;
+
+	while ((p = strsep(&opts, ",")) != NULL) {
+		int token;
+		if (!*p)
+			continue;
+
+		token = match_token(p, tokens, args);
+		switch (token) {
+		case Opt_subvol:
+			*subvol_name = match_strdup(&args[0]);
+			break;
+		case Opt_device:
+			error = btrfs_scan_one_device(match_strdup(&args[0]),
+					flags, holder, fs_devices);
+			if (error)
+				goto out_free_opts;
+			break;
+		default:
+			break;
+		}
+	}
+
+ out_free_opts:
+	kfree(opts);
+ out:
+	/*
+	 * If no subvolume name is specified we use the default one.  Allocate
+	 * a copy of the string "." here so that code later in the
+	 * mount path doesn't care if it's the default volume or another one.
+	 */
+	if (!*subvol_name) {
+		*subvol_name = kstrdup(".", GFP_KERNEL);
+		if (!*subvol_name)
+			return -ENOMEM;
+	}
+	return error;
+}
+
+static int btrfs_fill_super(struct super_block *sb,
+			    struct btrfs_fs_devices *fs_devices,
+			    void *data, int silent)
+{
+	struct inode *inode;
+	struct dentry *root_dentry;
+	struct btrfs_super_block *disk_super;
+	struct btrfs_root *tree_root;
+	struct btrfs_inode *bi;
+	int err;
+
+	sb->s_maxbytes = MAX_LFS_FILESIZE;
+	sb->s_magic = BTRFS_SUPER_MAGIC;
+	sb->s_op = &btrfs_super_ops;
+	sb->s_export_op = &btrfs_export_ops;
+	sb->s_xattr = btrfs_xattr_handlers;
+	sb->s_time_gran = 1;
+	sb->s_flags |= MS_POSIXACL;
+
+	tree_root = open_ctree(sb, fs_devices, (char *)data);
+
+	if (IS_ERR(tree_root)) {
+		printk("btrfs: open_ctree failed\n");
+		return PTR_ERR(tree_root);
+	}
+	sb->s_fs_info = tree_root;
+	disk_super = &tree_root->fs_info->super_copy;
+	inode = btrfs_iget_locked(sb, BTRFS_FIRST_FREE_OBJECTID,
+				  tree_root->fs_info->fs_root);
+	bi = BTRFS_I(inode);
+	bi->location.objectid = inode->i_ino;
+	bi->location.offset = 0;
+	bi->root = tree_root->fs_info->fs_root;
+
+	btrfs_set_key_type(&bi->location, BTRFS_INODE_ITEM_KEY);
+
+	if (!inode) {
+		err = -ENOMEM;
+		goto fail_close;
+	}
+	if (inode->i_state & I_NEW) {
+		btrfs_read_locked_inode(inode);
+		unlock_new_inode(inode);
+	}
+
+	root_dentry = d_alloc_root(inode);
+	if (!root_dentry) {
+		iput(inode);
+		err = -ENOMEM;
+		goto fail_close;
+	}
+#if 0
+	/* this does the super kobj at the same time */
+	err = btrfs_sysfs_add_super(tree_root->fs_info);
+	if (err)
+		goto fail_close;
+#endif
+
+	sb->s_root = root_dentry;
+
+	save_mount_options(sb, data);
+	return 0;
+
+fail_close:
+	close_ctree(tree_root);
+	return err;
+}
+
+int btrfs_sync_fs(struct super_block *sb, int wait)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root;
+	int ret;
+	root = btrfs_sb(sb);
+
+	if (sb->s_flags & MS_RDONLY)
+		return 0;
+
+	sb->s_dirt = 0;
+	if (!wait) {
+		filemap_flush(root->fs_info->btree_inode->i_mapping);
+		return 0;
+	}
+
+	btrfs_start_delalloc_inodes(root);
+	btrfs_wait_ordered_extents(root, 0);
+
+	btrfs_clean_old_snapshots(root);
+	trans = btrfs_start_transaction(root, 1);
+	ret = btrfs_commit_transaction(trans, root);
+	sb->s_dirt = 0;
+	return ret;
+}
+
+static void btrfs_write_super(struct super_block *sb)
+{
+	sb->s_dirt = 0;
+}
+
+static int btrfs_test_super(struct super_block *s, void *data)
+{
+	struct btrfs_fs_devices *test_fs_devices = data;
+	struct btrfs_root *root = btrfs_sb(s);
+
+	return root->fs_info->fs_devices == test_fs_devices;
+}
+
+/*
+ * Find a superblock for the given device / mount point.
+ *
+ * Note:  This is based on get_sb_bdev from fs/super.c with a few additions
+ *	  for multiple device setup.  Make sure to keep it in sync.
+ */
+static int btrfs_get_sb(struct file_system_type *fs_type, int flags,
+		const char *dev_name, void *data, struct vfsmount *mnt)
+{
+	char *subvol_name = NULL;
+	struct block_device *bdev = NULL;
+	struct super_block *s;
+	struct dentry *root;
+	struct btrfs_fs_devices *fs_devices = NULL;
+	fmode_t mode = FMODE_READ;
+	int error = 0;
+
+	if (!(flags & MS_RDONLY))
+		mode |= FMODE_WRITE;
+
+	error = btrfs_parse_early_options(data, mode, fs_type,
+					  &subvol_name, &fs_devices);
+	if (error)
+		return error;
+
+	error = btrfs_scan_one_device(dev_name, mode, fs_type, &fs_devices);
+	if (error)
+		goto error_free_subvol_name;
+
+	error = btrfs_open_devices(fs_devices, mode, fs_type);
+	if (error)
+		goto error_free_subvol_name;
+
+	if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) {
+		error = -EACCES;
+		goto error_close_devices;
+	}
+
+	bdev = fs_devices->latest_bdev;
+	s = sget(fs_type, btrfs_test_super, set_anon_super, fs_devices);
+	if (IS_ERR(s))
+		goto error_s;
+
+	if (s->s_root) {
+		if ((flags ^ s->s_flags) & MS_RDONLY) {
+			up_write(&s->s_umount);
+			deactivate_super(s);
+			error = -EBUSY;
+			goto error_close_devices;
+		}
+
+		btrfs_close_devices(fs_devices);
+	} else {
+		char b[BDEVNAME_SIZE];
+
+		s->s_flags = flags;
+		strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
+		error = btrfs_fill_super(s, fs_devices, data,
+					 flags & MS_SILENT ? 1 : 0);
+		if (error) {
+			up_write(&s->s_umount);
+			deactivate_super(s);
+			goto error_free_subvol_name;
+		}
+
+		btrfs_sb(s)->fs_info->bdev_holder = fs_type;
+		s->s_flags |= MS_ACTIVE;
+	}
+
+	if (!strcmp(subvol_name, "."))
+		root = dget(s->s_root);
+	else {
+		mutex_lock(&s->s_root->d_inode->i_mutex);
+		root = lookup_one_len(subvol_name, s->s_root,
+				      strlen(subvol_name));
+		mutex_unlock(&s->s_root->d_inode->i_mutex);
+
+		if (IS_ERR(root)) {
+			up_write(&s->s_umount);
+			deactivate_super(s);
+			error = PTR_ERR(root);
+			goto error_free_subvol_name;
+		}
+		if (!root->d_inode) {
+			dput(root);
+			up_write(&s->s_umount);
+			deactivate_super(s);
+			error = -ENXIO;
+			goto error_free_subvol_name;
+		}
+	}
+
+	mnt->mnt_sb = s;
+	mnt->mnt_root = root;
+
+	kfree(subvol_name);
+	return 0;
+
+error_s:
+	error = PTR_ERR(s);
+error_close_devices:
+	btrfs_close_devices(fs_devices);
+error_free_subvol_name:
+	kfree(subvol_name);
+	return error;
+}
+
+static int btrfs_remount(struct super_block *sb, int *flags, char *data)
+{
+	struct btrfs_root *root = btrfs_sb(sb);
+	int ret;
+
+	if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY))
+		return 0;
+
+	if (*flags & MS_RDONLY) {
+		sb->s_flags |= MS_RDONLY;
+
+		ret =  btrfs_commit_super(root);
+		WARN_ON(ret);
+	} else {
+		if (root->fs_info->fs_devices->rw_devices == 0)
+			return -EACCES;
+
+		if (btrfs_super_log_root(&root->fs_info->super_copy) != 0)
+			return -EINVAL;
+
+		ret = btrfs_cleanup_reloc_trees(root);
+		WARN_ON(ret);
+
+		ret = btrfs_cleanup_fs_roots(root->fs_info);
+		WARN_ON(ret);
+
+		sb->s_flags &= ~MS_RDONLY;
+	}
+
+	return 0;
+}
+
+static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
+{
+	struct btrfs_root *root = btrfs_sb(dentry->d_sb);
+	struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
+	int bits = dentry->d_sb->s_blocksize_bits;
+	__be32 *fsid = (__be32 *)root->fs_info->fsid;
+
+	buf->f_namelen = BTRFS_NAME_LEN;
+	buf->f_blocks = btrfs_super_total_bytes(disk_super) >> bits;
+	buf->f_bfree = buf->f_blocks -
+		(btrfs_super_bytes_used(disk_super) >> bits);
+	buf->f_bavail = buf->f_bfree;
+	buf->f_bsize = dentry->d_sb->s_blocksize;
+	buf->f_type = BTRFS_SUPER_MAGIC;
+
+	/* We treat it as constant endianness (it doesn't matter _which_)
+	   because we want the fsid to come out the same whether mounted
+	   on a big-endian or little-endian host */
+	buf->f_fsid.val[0] = be32_to_cpu(fsid[0]) ^ be32_to_cpu(fsid[2]);
+	buf->f_fsid.val[1] = be32_to_cpu(fsid[1]) ^ be32_to_cpu(fsid[3]);
+	/* Mask in the root object ID too, to disambiguate subvols */
+	buf->f_fsid.val[0] ^= BTRFS_I(dentry->d_inode)->root->objectid >> 32;
+	buf->f_fsid.val[1] ^= BTRFS_I(dentry->d_inode)->root->objectid;
+
+	return 0;
+}
+
+static struct file_system_type btrfs_fs_type = {
+	.owner		= THIS_MODULE,
+	.name		= "btrfs",
+	.get_sb		= btrfs_get_sb,
+	.kill_sb	= kill_anon_super,
+	.fs_flags	= FS_REQUIRES_DEV,
+};
+
+/*
+ * used by btrfsctl to scan devices when no FS is mounted
+ */
+static long btrfs_control_ioctl(struct file *file, unsigned int cmd,
+				unsigned long arg)
+{
+	struct btrfs_ioctl_vol_args *vol;
+	struct btrfs_fs_devices *fs_devices;
+	int ret = 0;
+	int len;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	vol = kmalloc(sizeof(*vol), GFP_KERNEL);
+	if (copy_from_user(vol, (void __user *)arg, sizeof(*vol))) {
+		ret = -EFAULT;
+		goto out;
+	}
+	len = strnlen(vol->name, BTRFS_PATH_NAME_MAX);
+	switch (cmd) {
+	case BTRFS_IOC_SCAN_DEV:
+		ret = btrfs_scan_one_device(vol->name, FMODE_READ,
+					    &btrfs_fs_type, &fs_devices);
+		break;
+	}
+out:
+	kfree(vol);
+	return ret;
+}
+
+static int btrfs_freeze(struct super_block *sb)
+{
+	struct btrfs_root *root = btrfs_sb(sb);
+	mutex_lock(&root->fs_info->transaction_kthread_mutex);
+	mutex_lock(&root->fs_info->cleaner_mutex);
+	return 0;
+}
+
+static int btrfs_unfreeze(struct super_block *sb)
+{
+	struct btrfs_root *root = btrfs_sb(sb);
+	mutex_unlock(&root->fs_info->cleaner_mutex);
+	mutex_unlock(&root->fs_info->transaction_kthread_mutex);
+	return 0;
+}
+
+static struct super_operations btrfs_super_ops = {
+	.delete_inode	= btrfs_delete_inode,
+	.put_super	= btrfs_put_super,
+	.write_super	= btrfs_write_super,
+	.sync_fs	= btrfs_sync_fs,
+	.show_options	= generic_show_options,
+	.write_inode	= btrfs_write_inode,
+	.dirty_inode	= btrfs_dirty_inode,
+	.alloc_inode	= btrfs_alloc_inode,
+	.destroy_inode	= btrfs_destroy_inode,
+	.statfs		= btrfs_statfs,
+	.remount_fs	= btrfs_remount,
+	.freeze_fs	= btrfs_freeze,
+	.unfreeze_fs	= btrfs_unfreeze,
+};
+
+static const struct file_operations btrfs_ctl_fops = {
+	.unlocked_ioctl	 = btrfs_control_ioctl,
+	.compat_ioctl = btrfs_control_ioctl,
+	.owner	 = THIS_MODULE,
+};
+
+static struct miscdevice btrfs_misc = {
+	.minor		= MISC_DYNAMIC_MINOR,
+	.name		= "btrfs-control",
+	.fops		= &btrfs_ctl_fops
+};
+
+static int btrfs_interface_init(void)
+{
+	return misc_register(&btrfs_misc);
+}
+
+static void btrfs_interface_exit(void)
+{
+	if (misc_deregister(&btrfs_misc) < 0)
+		printk(KERN_INFO "misc_deregister failed for control device");
+}
+
+static int __init init_btrfs_fs(void)
+{
+	int err;
+
+	err = btrfs_init_sysfs();
+	if (err)
+		return err;
+
+	err = btrfs_init_cachep();
+	if (err)
+		goto free_sysfs;
+
+	err = extent_io_init();
+	if (err)
+		goto free_cachep;
+
+	err = extent_map_init();
+	if (err)
+		goto free_extent_io;
+
+	err = btrfs_interface_init();
+	if (err)
+		goto free_extent_map;
+
+	err = register_filesystem(&btrfs_fs_type);
+	if (err)
+		goto unregister_ioctl;
+
+	printk(KERN_INFO "%s loaded\n", BTRFS_BUILD_VERSION);
+	return 0;
+
+unregister_ioctl:
+	btrfs_interface_exit();
+free_extent_map:
+	extent_map_exit();
+free_extent_io:
+	extent_io_exit();
+free_cachep:
+	btrfs_destroy_cachep();
+free_sysfs:
+	btrfs_exit_sysfs();
+	return err;
+}
+
+static void __exit exit_btrfs_fs(void)
+{
+	btrfs_destroy_cachep();
+	extent_map_exit();
+	extent_io_exit();
+	btrfs_interface_exit();
+	unregister_filesystem(&btrfs_fs_type);
+	btrfs_exit_sysfs();
+	btrfs_cleanup_fs_uuids();
+	btrfs_zlib_exit();
+}
+
+module_init(init_btrfs_fs)
+module_exit(exit_btrfs_fs)
+
+MODULE_LICENSE("GPL");
diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c
new file mode 100644
index 00000000000..a240b6fa81d
--- /dev/null
+++ b/fs/btrfs/sysfs.c
@@ -0,0 +1,269 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/completion.h>
+#include <linux/buffer_head.h>
+#include <linux/module.h>
+#include <linux/kobject.h>
+
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+
+static ssize_t root_blocks_used_show(struct btrfs_root *root, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%llu\n",
+		(unsigned long long)btrfs_root_used(&root->root_item));
+}
+
+static ssize_t root_block_limit_show(struct btrfs_root *root, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%llu\n",
+		(unsigned long long)btrfs_root_limit(&root->root_item));
+}
+
+static ssize_t super_blocks_used_show(struct btrfs_fs_info *fs, char *buf)
+{
+
+	return snprintf(buf, PAGE_SIZE, "%llu\n",
+		(unsigned long long)btrfs_super_bytes_used(&fs->super_copy));
+}
+
+static ssize_t super_total_blocks_show(struct btrfs_fs_info *fs, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%llu\n",
+		(unsigned long long)btrfs_super_total_bytes(&fs->super_copy));
+}
+
+static ssize_t super_blocksize_show(struct btrfs_fs_info *fs, char *buf)
+{
+	return snprintf(buf, PAGE_SIZE, "%llu\n",
+		(unsigned long long)btrfs_super_sectorsize(&fs->super_copy));
+}
+
+/* this is for root attrs (subvols/snapshots) */
+struct btrfs_root_attr {
+	struct attribute attr;
+	ssize_t (*show)(struct btrfs_root *, char *);
+	ssize_t (*store)(struct btrfs_root *, const char *, size_t);
+};
+
+#define ROOT_ATTR(name, mode, show, store) \
+static struct btrfs_root_attr btrfs_root_attr_##name = __ATTR(name, mode, \
+							      show, store)
+
+ROOT_ATTR(blocks_used,	0444,	root_blocks_used_show,	NULL);
+ROOT_ATTR(block_limit,	0644,	root_block_limit_show,	NULL);
+
+static struct attribute *btrfs_root_attrs[] = {
+	&btrfs_root_attr_blocks_used.attr,
+	&btrfs_root_attr_block_limit.attr,
+	NULL,
+};
+
+/* this is for super attrs (actual full fs) */
+struct btrfs_super_attr {
+	struct attribute attr;
+	ssize_t (*show)(struct btrfs_fs_info *, char *);
+	ssize_t (*store)(struct btrfs_fs_info *, const char *, size_t);
+};
+
+#define SUPER_ATTR(name, mode, show, store) \
+static struct btrfs_super_attr btrfs_super_attr_##name = __ATTR(name, mode, \
+								show, store)
+
+SUPER_ATTR(blocks_used,		0444,	super_blocks_used_show,		NULL);
+SUPER_ATTR(total_blocks,	0444,	super_total_blocks_show,	NULL);
+SUPER_ATTR(blocksize,		0444,	super_blocksize_show,		NULL);
+
+static struct attribute *btrfs_super_attrs[] = {
+	&btrfs_super_attr_blocks_used.attr,
+	&btrfs_super_attr_total_blocks.attr,
+	&btrfs_super_attr_blocksize.attr,
+	NULL,
+};
+
+static ssize_t btrfs_super_attr_show(struct kobject *kobj,
+				    struct attribute *attr, char *buf)
+{
+	struct btrfs_fs_info *fs = container_of(kobj, struct btrfs_fs_info,
+						super_kobj);
+	struct btrfs_super_attr *a = container_of(attr,
+						  struct btrfs_super_attr,
+						  attr);
+
+	return a->show ? a->show(fs, buf) : 0;
+}
+
+static ssize_t btrfs_super_attr_store(struct kobject *kobj,
+				     struct attribute *attr,
+				     const char *buf, size_t len)
+{
+	struct btrfs_fs_info *fs = container_of(kobj, struct btrfs_fs_info,
+						super_kobj);
+	struct btrfs_super_attr *a = container_of(attr,
+						  struct btrfs_super_attr,
+						  attr);
+
+	return a->store ? a->store(fs, buf, len) : 0;
+}
+
+static ssize_t btrfs_root_attr_show(struct kobject *kobj,
+				    struct attribute *attr, char *buf)
+{
+	struct btrfs_root *root = container_of(kobj, struct btrfs_root,
+						root_kobj);
+	struct btrfs_root_attr *a = container_of(attr,
+						 struct btrfs_root_attr,
+						 attr);
+
+	return a->show ? a->show(root, buf) : 0;
+}
+
+static ssize_t btrfs_root_attr_store(struct kobject *kobj,
+				     struct attribute *attr,
+				     const char *buf, size_t len)
+{
+	struct btrfs_root *root = container_of(kobj, struct btrfs_root,
+						root_kobj);
+	struct btrfs_root_attr *a = container_of(attr,
+						 struct btrfs_root_attr,
+						 attr);
+	return a->store ? a->store(root, buf, len) : 0;
+}
+
+static void btrfs_super_release(struct kobject *kobj)
+{
+	struct btrfs_fs_info *fs = container_of(kobj, struct btrfs_fs_info,
+						super_kobj);
+	complete(&fs->kobj_unregister);
+}
+
+static void btrfs_root_release(struct kobject *kobj)
+{
+	struct btrfs_root *root = container_of(kobj, struct btrfs_root,
+						root_kobj);
+	complete(&root->kobj_unregister);
+}
+
+static struct sysfs_ops btrfs_super_attr_ops = {
+	.show	= btrfs_super_attr_show,
+	.store	= btrfs_super_attr_store,
+};
+
+static struct sysfs_ops btrfs_root_attr_ops = {
+	.show	= btrfs_root_attr_show,
+	.store	= btrfs_root_attr_store,
+};
+
+static struct kobj_type btrfs_root_ktype = {
+	.default_attrs	= btrfs_root_attrs,
+	.sysfs_ops	= &btrfs_root_attr_ops,
+	.release	= btrfs_root_release,
+};
+
+static struct kobj_type btrfs_super_ktype = {
+	.default_attrs	= btrfs_super_attrs,
+	.sysfs_ops	= &btrfs_super_attr_ops,
+	.release	= btrfs_super_release,
+};
+
+/* /sys/fs/btrfs/ entry */
+static struct kset *btrfs_kset;
+
+int btrfs_sysfs_add_super(struct btrfs_fs_info *fs)
+{
+	int error;
+	char *name;
+	char c;
+	int len = strlen(fs->sb->s_id) + 1;
+	int i;
+
+	name = kmalloc(len, GFP_NOFS);
+	if (!name) {
+		error = -ENOMEM;
+		goto fail;
+	}
+
+	for (i = 0; i < len; i++) {
+		c = fs->sb->s_id[i];
+		if (c == '/' || c == '\\')
+			c = '!';
+		name[i] = c;
+	}
+	name[len] = '\0';
+
+	fs->super_kobj.kset = btrfs_kset;
+	error = kobject_init_and_add(&fs->super_kobj, &btrfs_super_ktype,
+				     NULL, "%s", name);
+	kfree(name);
+	if (error)
+		goto fail;
+
+	return 0;
+
+fail:
+	printk(KERN_ERR "btrfs: sysfs creation for super failed\n");
+	return error;
+}
+
+int btrfs_sysfs_add_root(struct btrfs_root *root)
+{
+	int error;
+
+	error = kobject_init_and_add(&root->root_kobj, &btrfs_root_ktype,
+				     &root->fs_info->super_kobj,
+				     "%s", root->name);
+	if (error)
+		goto fail;
+
+	return 0;
+
+fail:
+	printk(KERN_ERR "btrfs: sysfs creation for root failed\n");
+	return error;
+}
+
+void btrfs_sysfs_del_root(struct btrfs_root *root)
+{
+	kobject_put(&root->root_kobj);
+	wait_for_completion(&root->kobj_unregister);
+}
+
+void btrfs_sysfs_del_super(struct btrfs_fs_info *fs)
+{
+	kobject_put(&fs->super_kobj);
+	wait_for_completion(&fs->kobj_unregister);
+}
+
+int btrfs_init_sysfs(void)
+{
+	btrfs_kset = kset_create_and_add("btrfs", NULL, fs_kobj);
+	if (!btrfs_kset)
+		return -ENOMEM;
+	return 0;
+}
+
+void btrfs_exit_sysfs(void)
+{
+	kset_unregister(btrfs_kset);
+}
+
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
new file mode 100644
index 00000000000..8a08f944334
--- /dev/null
+++ b/fs/btrfs/transaction.c
@@ -0,0 +1,1097 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/fs.h>
+#include <linux/sched.h>
+#include <linux/writeback.h>
+#include <linux/pagemap.h>
+#include <linux/blkdev.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "locking.h"
+#include "ref-cache.h"
+#include "tree-log.h"
+
+#define BTRFS_ROOT_TRANS_TAG 0
+
+static noinline void put_transaction(struct btrfs_transaction *transaction)
+{
+	WARN_ON(transaction->use_count == 0);
+	transaction->use_count--;
+	if (transaction->use_count == 0) {
+		list_del_init(&transaction->list);
+		memset(transaction, 0, sizeof(*transaction));
+		kmem_cache_free(btrfs_transaction_cachep, transaction);
+	}
+}
+
+/*
+ * either allocate a new transaction or hop into the existing one
+ */
+static noinline int join_transaction(struct btrfs_root *root)
+{
+	struct btrfs_transaction *cur_trans;
+	cur_trans = root->fs_info->running_transaction;
+	if (!cur_trans) {
+		cur_trans = kmem_cache_alloc(btrfs_transaction_cachep,
+					     GFP_NOFS);
+		BUG_ON(!cur_trans);
+		root->fs_info->generation++;
+		root->fs_info->last_alloc = 0;
+		root->fs_info->last_data_alloc = 0;
+		cur_trans->num_writers = 1;
+		cur_trans->num_joined = 0;
+		cur_trans->transid = root->fs_info->generation;
+		init_waitqueue_head(&cur_trans->writer_wait);
+		init_waitqueue_head(&cur_trans->commit_wait);
+		cur_trans->in_commit = 0;
+		cur_trans->blocked = 0;
+		cur_trans->use_count = 1;
+		cur_trans->commit_done = 0;
+		cur_trans->start_time = get_seconds();
+		INIT_LIST_HEAD(&cur_trans->pending_snapshots);
+		list_add_tail(&cur_trans->list, &root->fs_info->trans_list);
+		extent_io_tree_init(&cur_trans->dirty_pages,
+				     root->fs_info->btree_inode->i_mapping,
+				     GFP_NOFS);
+		spin_lock(&root->fs_info->new_trans_lock);
+		root->fs_info->running_transaction = cur_trans;
+		spin_unlock(&root->fs_info->new_trans_lock);
+	} else {
+		cur_trans->num_writers++;
+		cur_trans->num_joined++;
+	}
+
+	return 0;
+}
+
+/*
+ * this does all the record keeping required to make sure that a reference
+ * counted root is properly recorded in a given transaction.  This is required
+ * to make sure the old root from before we joined the transaction is deleted
+ * when the transaction commits
+ */
+noinline int btrfs_record_root_in_trans(struct btrfs_root *root)
+{
+	struct btrfs_dirty_root *dirty;
+	u64 running_trans_id = root->fs_info->running_transaction->transid;
+	if (root->ref_cows && root->last_trans < running_trans_id) {
+		WARN_ON(root == root->fs_info->extent_root);
+		if (root->root_item.refs != 0) {
+			radix_tree_tag_set(&root->fs_info->fs_roots_radix,
+				   (unsigned long)root->root_key.objectid,
+				   BTRFS_ROOT_TRANS_TAG);
+
+			dirty = kmalloc(sizeof(*dirty), GFP_NOFS);
+			BUG_ON(!dirty);
+			dirty->root = kmalloc(sizeof(*dirty->root), GFP_NOFS);
+			BUG_ON(!dirty->root);
+			dirty->latest_root = root;
+			INIT_LIST_HEAD(&dirty->list);
+
+			root->commit_root = btrfs_root_node(root);
+
+			memcpy(dirty->root, root, sizeof(*root));
+			spin_lock_init(&dirty->root->node_lock);
+			spin_lock_init(&dirty->root->list_lock);
+			mutex_init(&dirty->root->objectid_mutex);
+			mutex_init(&dirty->root->log_mutex);
+			INIT_LIST_HEAD(&dirty->root->dead_list);
+			dirty->root->node = root->commit_root;
+			dirty->root->commit_root = NULL;
+
+			spin_lock(&root->list_lock);
+			list_add(&dirty->root->dead_list, &root->dead_list);
+			spin_unlock(&root->list_lock);
+
+			root->dirty_root = dirty;
+		} else {
+			WARN_ON(1);
+		}
+		root->last_trans = running_trans_id;
+	}
+	return 0;
+}
+
+/* wait for commit against the current transaction to become unblocked
+ * when this is done, it is safe to start a new transaction, but the current
+ * transaction might not be fully on disk.
+ */
+static void wait_current_trans(struct btrfs_root *root)
+{
+	struct btrfs_transaction *cur_trans;
+
+	cur_trans = root->fs_info->running_transaction;
+	if (cur_trans && cur_trans->blocked) {
+		DEFINE_WAIT(wait);
+		cur_trans->use_count++;
+		while (1) {
+			prepare_to_wait(&root->fs_info->transaction_wait, &wait,
+					TASK_UNINTERRUPTIBLE);
+			if (cur_trans->blocked) {
+				mutex_unlock(&root->fs_info->trans_mutex);
+				schedule();
+				mutex_lock(&root->fs_info->trans_mutex);
+				finish_wait(&root->fs_info->transaction_wait,
+					    &wait);
+			} else {
+				finish_wait(&root->fs_info->transaction_wait,
+					    &wait);
+				break;
+			}
+		}
+		put_transaction(cur_trans);
+	}
+}
+
+static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
+					     int num_blocks, int wait)
+{
+	struct btrfs_trans_handle *h =
+		kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
+	int ret;
+
+	mutex_lock(&root->fs_info->trans_mutex);
+	if (!root->fs_info->log_root_recovering &&
+	    ((wait == 1 && !root->fs_info->open_ioctl_trans) || wait == 2))
+		wait_current_trans(root);
+	ret = join_transaction(root);
+	BUG_ON(ret);
+
+	btrfs_record_root_in_trans(root);
+	h->transid = root->fs_info->running_transaction->transid;
+	h->transaction = root->fs_info->running_transaction;
+	h->blocks_reserved = num_blocks;
+	h->blocks_used = 0;
+	h->block_group = 0;
+	h->alloc_exclude_nr = 0;
+	h->alloc_exclude_start = 0;
+	root->fs_info->running_transaction->use_count++;
+	mutex_unlock(&root->fs_info->trans_mutex);
+	return h;
+}
+
+struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
+						   int num_blocks)
+{
+	return start_transaction(root, num_blocks, 1);
+}
+struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
+						   int num_blocks)
+{
+	return start_transaction(root, num_blocks, 0);
+}
+
+struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
+							 int num_blocks)
+{
+	return start_transaction(r, num_blocks, 2);
+}
+
+/* wait for a transaction commit to be fully complete */
+static noinline int wait_for_commit(struct btrfs_root *root,
+				    struct btrfs_transaction *commit)
+{
+	DEFINE_WAIT(wait);
+	mutex_lock(&root->fs_info->trans_mutex);
+	while (!commit->commit_done) {
+		prepare_to_wait(&commit->commit_wait, &wait,
+				TASK_UNINTERRUPTIBLE);
+		if (commit->commit_done)
+			break;
+		mutex_unlock(&root->fs_info->trans_mutex);
+		schedule();
+		mutex_lock(&root->fs_info->trans_mutex);
+	}
+	mutex_unlock(&root->fs_info->trans_mutex);
+	finish_wait(&commit->commit_wait, &wait);
+	return 0;
+}
+
+/*
+ * rate limit against the drop_snapshot code.  This helps to slow down new
+ * operations if the drop_snapshot code isn't able to keep up.
+ */
+static void throttle_on_drops(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	int harder_count = 0;
+
+harder:
+	if (atomic_read(&info->throttles)) {
+		DEFINE_WAIT(wait);
+		int thr;
+		thr = atomic_read(&info->throttle_gen);
+
+		do {
+			prepare_to_wait(&info->transaction_throttle,
+					&wait, TASK_UNINTERRUPTIBLE);
+			if (!atomic_read(&info->throttles)) {
+				finish_wait(&info->transaction_throttle, &wait);
+				break;
+			}
+			schedule();
+			finish_wait(&info->transaction_throttle, &wait);
+		} while (thr == atomic_read(&info->throttle_gen));
+		harder_count++;
+
+		if (root->fs_info->total_ref_cache_size > 1 * 1024 * 1024 &&
+		    harder_count < 2)
+			goto harder;
+
+		if (root->fs_info->total_ref_cache_size > 5 * 1024 * 1024 &&
+		    harder_count < 10)
+			goto harder;
+
+		if (root->fs_info->total_ref_cache_size > 10 * 1024 * 1024 &&
+		    harder_count < 20)
+			goto harder;
+	}
+}
+
+void btrfs_throttle(struct btrfs_root *root)
+{
+	mutex_lock(&root->fs_info->trans_mutex);
+	if (!root->fs_info->open_ioctl_trans)
+		wait_current_trans(root);
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	throttle_on_drops(root);
+}
+
+static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, int throttle)
+{
+	struct btrfs_transaction *cur_trans;
+	struct btrfs_fs_info *info = root->fs_info;
+
+	mutex_lock(&info->trans_mutex);
+	cur_trans = info->running_transaction;
+	WARN_ON(cur_trans != trans->transaction);
+	WARN_ON(cur_trans->num_writers < 1);
+	cur_trans->num_writers--;
+
+	if (waitqueue_active(&cur_trans->writer_wait))
+		wake_up(&cur_trans->writer_wait);
+	put_transaction(cur_trans);
+	mutex_unlock(&info->trans_mutex);
+	memset(trans, 0, sizeof(*trans));
+	kmem_cache_free(btrfs_trans_handle_cachep, trans);
+
+	if (throttle)
+		throttle_on_drops(root);
+
+	return 0;
+}
+
+int btrfs_end_transaction(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root)
+{
+	return __btrfs_end_transaction(trans, root, 0);
+}
+
+int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root)
+{
+	return __btrfs_end_transaction(trans, root, 1);
+}
+
+/*
+ * when btree blocks are allocated, they have some corresponding bits set for
+ * them in one of two extent_io trees.  This is used to make sure all of
+ * those extents are on disk for transaction or log commit
+ */
+int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
+					struct extent_io_tree *dirty_pages)
+{
+	int ret;
+	int err = 0;
+	int werr = 0;
+	struct page *page;
+	struct inode *btree_inode = root->fs_info->btree_inode;
+	u64 start = 0;
+	u64 end;
+	unsigned long index;
+
+	while (1) {
+		ret = find_first_extent_bit(dirty_pages, start, &start, &end,
+					    EXTENT_DIRTY);
+		if (ret)
+			break;
+		while (start <= end) {
+			cond_resched();
+
+			index = start >> PAGE_CACHE_SHIFT;
+			start = (u64)(index + 1) << PAGE_CACHE_SHIFT;
+			page = find_get_page(btree_inode->i_mapping, index);
+			if (!page)
+				continue;
+
+			btree_lock_page_hook(page);
+			if (!page->mapping) {
+				unlock_page(page);
+				page_cache_release(page);
+				continue;
+			}
+
+			if (PageWriteback(page)) {
+				if (PageDirty(page))
+					wait_on_page_writeback(page);
+				else {
+					unlock_page(page);
+					page_cache_release(page);
+					continue;
+				}
+			}
+			err = write_one_page(page, 0);
+			if (err)
+				werr = err;
+			page_cache_release(page);
+		}
+	}
+	while (1) {
+		ret = find_first_extent_bit(dirty_pages, 0, &start, &end,
+					    EXTENT_DIRTY);
+		if (ret)
+			break;
+
+		clear_extent_dirty(dirty_pages, start, end, GFP_NOFS);
+		while (start <= end) {
+			index = start >> PAGE_CACHE_SHIFT;
+			start = (u64)(index + 1) << PAGE_CACHE_SHIFT;
+			page = find_get_page(btree_inode->i_mapping, index);
+			if (!page)
+				continue;
+			if (PageDirty(page)) {
+				btree_lock_page_hook(page);
+				wait_on_page_writeback(page);
+				err = write_one_page(page, 0);
+				if (err)
+					werr = err;
+			}
+			wait_on_page_writeback(page);
+			page_cache_release(page);
+			cond_resched();
+		}
+	}
+	if (err)
+		werr = err;
+	return werr;
+}
+
+int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root)
+{
+	if (!trans || !trans->transaction) {
+		struct inode *btree_inode;
+		btree_inode = root->fs_info->btree_inode;
+		return filemap_write_and_wait(btree_inode->i_mapping);
+	}
+	return btrfs_write_and_wait_marked_extents(root,
+					   &trans->transaction->dirty_pages);
+}
+
+/*
+ * this is used to update the root pointer in the tree of tree roots.
+ *
+ * But, in the case of the extent allocation tree, updating the root
+ * pointer may allocate blocks which may change the root of the extent
+ * allocation tree.
+ *
+ * So, this loops and repeats and makes sure the cowonly root didn't
+ * change while the root pointer was being updated in the metadata.
+ */
+static int update_cowonly_root(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root)
+{
+	int ret;
+	u64 old_root_bytenr;
+	struct btrfs_root *tree_root = root->fs_info->tree_root;
+
+	btrfs_extent_post_op(trans, root);
+	btrfs_write_dirty_block_groups(trans, root);
+	btrfs_extent_post_op(trans, root);
+
+	while (1) {
+		old_root_bytenr = btrfs_root_bytenr(&root->root_item);
+		if (old_root_bytenr == root->node->start)
+			break;
+		btrfs_set_root_bytenr(&root->root_item,
+				       root->node->start);
+		btrfs_set_root_level(&root->root_item,
+				     btrfs_header_level(root->node));
+		btrfs_set_root_generation(&root->root_item, trans->transid);
+
+		btrfs_extent_post_op(trans, root);
+
+		ret = btrfs_update_root(trans, tree_root,
+					&root->root_key,
+					&root->root_item);
+		BUG_ON(ret);
+		btrfs_write_dirty_block_groups(trans, root);
+		btrfs_extent_post_op(trans, root);
+	}
+	return 0;
+}
+
+/*
+ * update all the cowonly tree roots on disk
+ */
+int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct list_head *next;
+	struct extent_buffer *eb;
+
+	btrfs_extent_post_op(trans, fs_info->tree_root);
+
+	eb = btrfs_lock_root_node(fs_info->tree_root);
+	btrfs_cow_block(trans, fs_info->tree_root, eb, NULL, 0, &eb, 0);
+	btrfs_tree_unlock(eb);
+	free_extent_buffer(eb);
+
+	btrfs_extent_post_op(trans, fs_info->tree_root);
+
+	while (!list_empty(&fs_info->dirty_cowonly_roots)) {
+		next = fs_info->dirty_cowonly_roots.next;
+		list_del_init(next);
+		root = list_entry(next, struct btrfs_root, dirty_list);
+
+		update_cowonly_root(trans, root);
+	}
+	return 0;
+}
+
+/*
+ * dead roots are old snapshots that need to be deleted.  This allocates
+ * a dirty root struct and adds it into the list of dead roots that need to
+ * be deleted
+ */
+int btrfs_add_dead_root(struct btrfs_root *root, struct btrfs_root *latest)
+{
+	struct btrfs_dirty_root *dirty;
+
+	dirty = kmalloc(sizeof(*dirty), GFP_NOFS);
+	if (!dirty)
+		return -ENOMEM;
+	dirty->root = root;
+	dirty->latest_root = latest;
+
+	mutex_lock(&root->fs_info->trans_mutex);
+	list_add(&dirty->list, &latest->fs_info->dead_roots);
+	mutex_unlock(&root->fs_info->trans_mutex);
+	return 0;
+}
+
+/*
+ * at transaction commit time we need to schedule the old roots for
+ * deletion via btrfs_drop_snapshot.  This runs through all the
+ * reference counted roots that were modified in the current
+ * transaction and puts them into the drop list
+ */
+static noinline int add_dirty_roots(struct btrfs_trans_handle *trans,
+				    struct radix_tree_root *radix,
+				    struct list_head *list)
+{
+	struct btrfs_dirty_root *dirty;
+	struct btrfs_root *gang[8];
+	struct btrfs_root *root;
+	int i;
+	int ret;
+	int err = 0;
+	u32 refs;
+
+	while (1) {
+		ret = radix_tree_gang_lookup_tag(radix, (void **)gang, 0,
+						 ARRAY_SIZE(gang),
+						 BTRFS_ROOT_TRANS_TAG);
+		if (ret == 0)
+			break;
+		for (i = 0; i < ret; i++) {
+			root = gang[i];
+			radix_tree_tag_clear(radix,
+				     (unsigned long)root->root_key.objectid,
+				     BTRFS_ROOT_TRANS_TAG);
+
+			BUG_ON(!root->ref_tree);
+			dirty = root->dirty_root;
+
+			btrfs_free_log(trans, root);
+			btrfs_free_reloc_root(trans, root);
+
+			if (root->commit_root == root->node) {
+				WARN_ON(root->node->start !=
+					btrfs_root_bytenr(&root->root_item));
+
+				free_extent_buffer(root->commit_root);
+				root->commit_root = NULL;
+				root->dirty_root = NULL;
+
+				spin_lock(&root->list_lock);
+				list_del_init(&dirty->root->dead_list);
+				spin_unlock(&root->list_lock);
+
+				kfree(dirty->root);
+				kfree(dirty);
+
+				/* make sure to update the root on disk
+				 * so we get any updates to the block used
+				 * counts
+				 */
+				err = btrfs_update_root(trans,
+						root->fs_info->tree_root,
+						&root->root_key,
+						&root->root_item);
+				continue;
+			}
+
+			memset(&root->root_item.drop_progress, 0,
+			       sizeof(struct btrfs_disk_key));
+			root->root_item.drop_level = 0;
+			root->commit_root = NULL;
+			root->dirty_root = NULL;
+			root->root_key.offset = root->fs_info->generation;
+			btrfs_set_root_bytenr(&root->root_item,
+					      root->node->start);
+			btrfs_set_root_level(&root->root_item,
+					     btrfs_header_level(root->node));
+			btrfs_set_root_generation(&root->root_item,
+						  root->root_key.offset);
+
+			err = btrfs_insert_root(trans, root->fs_info->tree_root,
+						&root->root_key,
+						&root->root_item);
+			if (err)
+				break;
+
+			refs = btrfs_root_refs(&dirty->root->root_item);
+			btrfs_set_root_refs(&dirty->root->root_item, refs - 1);
+			err = btrfs_update_root(trans, root->fs_info->tree_root,
+						&dirty->root->root_key,
+						&dirty->root->root_item);
+
+			BUG_ON(err);
+			if (refs == 1) {
+				list_add(&dirty->list, list);
+			} else {
+				WARN_ON(1);
+				free_extent_buffer(dirty->root->node);
+				kfree(dirty->root);
+				kfree(dirty);
+			}
+		}
+	}
+	return err;
+}
+
+/*
+ * defrag a given btree.  If cacheonly == 1, this won't read from the disk,
+ * otherwise every leaf in the btree is read and defragged.
+ */
+int btrfs_defrag_root(struct btrfs_root *root, int cacheonly)
+{
+	struct btrfs_fs_info *info = root->fs_info;
+	int ret;
+	struct btrfs_trans_handle *trans;
+	unsigned long nr;
+
+	smp_mb();
+	if (root->defrag_running)
+		return 0;
+	trans = btrfs_start_transaction(root, 1);
+	while (1) {
+		root->defrag_running = 1;
+		ret = btrfs_defrag_leaves(trans, root, cacheonly);
+		nr = trans->blocks_used;
+		btrfs_end_transaction(trans, root);
+		btrfs_btree_balance_dirty(info->tree_root, nr);
+		cond_resched();
+
+		trans = btrfs_start_transaction(root, 1);
+		if (root->fs_info->closing || ret != -EAGAIN)
+			break;
+	}
+	root->defrag_running = 0;
+	smp_mb();
+	btrfs_end_transaction(trans, root);
+	return 0;
+}
+
+/*
+ * Given a list of roots that need to be deleted, call btrfs_drop_snapshot on
+ * all of them
+ */
+static noinline int drop_dirty_roots(struct btrfs_root *tree_root,
+				     struct list_head *list)
+{
+	struct btrfs_dirty_root *dirty;
+	struct btrfs_trans_handle *trans;
+	unsigned long nr;
+	u64 num_bytes;
+	u64 bytes_used;
+	u64 max_useless;
+	int ret = 0;
+	int err;
+
+	while (!list_empty(list)) {
+		struct btrfs_root *root;
+
+		dirty = list_entry(list->prev, struct btrfs_dirty_root, list);
+		list_del_init(&dirty->list);
+
+		num_bytes = btrfs_root_used(&dirty->root->root_item);
+		root = dirty->latest_root;
+		atomic_inc(&root->fs_info->throttles);
+
+		while (1) {
+			trans = btrfs_start_transaction(tree_root, 1);
+			mutex_lock(&root->fs_info->drop_mutex);
+			ret = btrfs_drop_snapshot(trans, dirty->root);
+			if (ret != -EAGAIN)
+				break;
+			mutex_unlock(&root->fs_info->drop_mutex);
+
+			err = btrfs_update_root(trans,
+					tree_root,
+					&dirty->root->root_key,
+					&dirty->root->root_item);
+			if (err)
+				ret = err;
+			nr = trans->blocks_used;
+			ret = btrfs_end_transaction(trans, tree_root);
+			BUG_ON(ret);
+
+			btrfs_btree_balance_dirty(tree_root, nr);
+			cond_resched();
+		}
+		BUG_ON(ret);
+		atomic_dec(&root->fs_info->throttles);
+		wake_up(&root->fs_info->transaction_throttle);
+
+		num_bytes -= btrfs_root_used(&dirty->root->root_item);
+		bytes_used = btrfs_root_used(&root->root_item);
+		if (num_bytes) {
+			btrfs_record_root_in_trans(root);
+			btrfs_set_root_used(&root->root_item,
+					    bytes_used - num_bytes);
+		}
+
+		ret = btrfs_del_root(trans, tree_root, &dirty->root->root_key);
+		if (ret) {
+			BUG();
+			break;
+		}
+		mutex_unlock(&root->fs_info->drop_mutex);
+
+		spin_lock(&root->list_lock);
+		list_del_init(&dirty->root->dead_list);
+		if (!list_empty(&root->dead_list)) {
+			struct btrfs_root *oldest;
+			oldest = list_entry(root->dead_list.prev,
+					    struct btrfs_root, dead_list);
+			max_useless = oldest->root_key.offset - 1;
+		} else {
+			max_useless = root->root_key.offset - 1;
+		}
+		spin_unlock(&root->list_lock);
+
+		nr = trans->blocks_used;
+		ret = btrfs_end_transaction(trans, tree_root);
+		BUG_ON(ret);
+
+		ret = btrfs_remove_leaf_refs(root, max_useless, 0);
+		BUG_ON(ret);
+
+		free_extent_buffer(dirty->root->node);
+		kfree(dirty->root);
+		kfree(dirty);
+
+		btrfs_btree_balance_dirty(tree_root, nr);
+		cond_resched();
+	}
+	return ret;
+}
+
+/*
+ * new snapshots need to be created at a very specific time in the
+ * transaction commit.  This does the actual creation
+ */
+static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
+				   struct btrfs_fs_info *fs_info,
+				   struct btrfs_pending_snapshot *pending)
+{
+	struct btrfs_key key;
+	struct btrfs_root_item *new_root_item;
+	struct btrfs_root *tree_root = fs_info->tree_root;
+	struct btrfs_root *root = pending->root;
+	struct extent_buffer *tmp;
+	struct extent_buffer *old;
+	int ret;
+	u64 objectid;
+
+	new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS);
+	if (!new_root_item) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+	ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid);
+	if (ret)
+		goto fail;
+
+	btrfs_record_root_in_trans(root);
+	btrfs_set_root_last_snapshot(&root->root_item, trans->transid);
+	memcpy(new_root_item, &root->root_item, sizeof(*new_root_item));
+
+	key.objectid = objectid;
+	key.offset = trans->transid;
+	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+
+	old = btrfs_lock_root_node(root);
+	btrfs_cow_block(trans, root, old, NULL, 0, &old, 0);
+
+	btrfs_copy_root(trans, root, old, &tmp, objectid);
+	btrfs_tree_unlock(old);
+	free_extent_buffer(old);
+
+	btrfs_set_root_bytenr(new_root_item, tmp->start);
+	btrfs_set_root_level(new_root_item, btrfs_header_level(tmp));
+	btrfs_set_root_generation(new_root_item, trans->transid);
+	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
+				new_root_item);
+	btrfs_tree_unlock(tmp);
+	free_extent_buffer(tmp);
+	if (ret)
+		goto fail;
+
+	key.offset = (u64)-1;
+	memcpy(&pending->root_key, &key, sizeof(key));
+fail:
+	kfree(new_root_item);
+	return ret;
+}
+
+static noinline int finish_pending_snapshot(struct btrfs_fs_info *fs_info,
+				   struct btrfs_pending_snapshot *pending)
+{
+	int ret;
+	int namelen;
+	u64 index = 0;
+	struct btrfs_trans_handle *trans;
+	struct inode *parent_inode;
+	struct inode *inode;
+	struct btrfs_root *parent_root;
+
+	parent_inode = pending->dentry->d_parent->d_inode;
+	parent_root = BTRFS_I(parent_inode)->root;
+	trans = btrfs_join_transaction(parent_root, 1);
+
+	/*
+	 * insert the directory item
+	 */
+	namelen = strlen(pending->name);
+	ret = btrfs_set_inode_index(parent_inode, &index);
+	ret = btrfs_insert_dir_item(trans, parent_root,
+			    pending->name, namelen,
+			    parent_inode->i_ino,
+			    &pending->root_key, BTRFS_FT_DIR, index);
+
+	if (ret)
+		goto fail;
+
+	btrfs_i_size_write(parent_inode, parent_inode->i_size + namelen * 2);
+	ret = btrfs_update_inode(trans, parent_root, parent_inode);
+	BUG_ON(ret);
+
+	/* add the backref first */
+	ret = btrfs_add_root_ref(trans, parent_root->fs_info->tree_root,
+				 pending->root_key.objectid,
+				 BTRFS_ROOT_BACKREF_KEY,
+				 parent_root->root_key.objectid,
+				 parent_inode->i_ino, index, pending->name,
+				 namelen);
+
+	BUG_ON(ret);
+
+	/* now add the forward ref */
+	ret = btrfs_add_root_ref(trans, parent_root->fs_info->tree_root,
+				 parent_root->root_key.objectid,
+				 BTRFS_ROOT_REF_KEY,
+				 pending->root_key.objectid,
+				 parent_inode->i_ino, index, pending->name,
+				 namelen);
+
+	inode = btrfs_lookup_dentry(parent_inode, pending->dentry);
+	d_instantiate(pending->dentry, inode);
+fail:
+	btrfs_end_transaction(trans, fs_info->fs_root);
+	return ret;
+}
+
+/*
+ * create all the snapshots we've scheduled for creation
+ */
+static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans,
+					     struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_pending_snapshot *pending;
+	struct list_head *head = &trans->transaction->pending_snapshots;
+	struct list_head *cur;
+	int ret;
+
+	list_for_each(cur, head) {
+		pending = list_entry(cur, struct btrfs_pending_snapshot, list);
+		ret = create_pending_snapshot(trans, fs_info, pending);
+		BUG_ON(ret);
+	}
+	return 0;
+}
+
+static noinline int finish_pending_snapshots(struct btrfs_trans_handle *trans,
+					     struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_pending_snapshot *pending;
+	struct list_head *head = &trans->transaction->pending_snapshots;
+	int ret;
+
+	while (!list_empty(head)) {
+		pending = list_entry(head->next,
+				     struct btrfs_pending_snapshot, list);
+		ret = finish_pending_snapshot(fs_info, pending);
+		BUG_ON(ret);
+		list_del(&pending->list);
+		kfree(pending->name);
+		kfree(pending);
+	}
+	return 0;
+}
+
+int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root)
+{
+	unsigned long joined = 0;
+	unsigned long timeout = 1;
+	struct btrfs_transaction *cur_trans;
+	struct btrfs_transaction *prev_trans = NULL;
+	struct btrfs_root *chunk_root = root->fs_info->chunk_root;
+	struct list_head dirty_fs_roots;
+	struct extent_io_tree *pinned_copy;
+	DEFINE_WAIT(wait);
+	int ret;
+
+	INIT_LIST_HEAD(&dirty_fs_roots);
+	mutex_lock(&root->fs_info->trans_mutex);
+	if (trans->transaction->in_commit) {
+		cur_trans = trans->transaction;
+		trans->transaction->use_count++;
+		mutex_unlock(&root->fs_info->trans_mutex);
+		btrfs_end_transaction(trans, root);
+
+		ret = wait_for_commit(root, cur_trans);
+		BUG_ON(ret);
+
+		mutex_lock(&root->fs_info->trans_mutex);
+		put_transaction(cur_trans);
+		mutex_unlock(&root->fs_info->trans_mutex);
+
+		return 0;
+	}
+
+	pinned_copy = kmalloc(sizeof(*pinned_copy), GFP_NOFS);
+	if (!pinned_copy)
+		return -ENOMEM;
+
+	extent_io_tree_init(pinned_copy,
+			     root->fs_info->btree_inode->i_mapping, GFP_NOFS);
+
+	trans->transaction->in_commit = 1;
+	trans->transaction->blocked = 1;
+	cur_trans = trans->transaction;
+	if (cur_trans->list.prev != &root->fs_info->trans_list) {
+		prev_trans = list_entry(cur_trans->list.prev,
+					struct btrfs_transaction, list);
+		if (!prev_trans->commit_done) {
+			prev_trans->use_count++;
+			mutex_unlock(&root->fs_info->trans_mutex);
+
+			wait_for_commit(root, prev_trans);
+
+			mutex_lock(&root->fs_info->trans_mutex);
+			put_transaction(prev_trans);
+		}
+	}
+
+	do {
+		int snap_pending = 0;
+		joined = cur_trans->num_joined;
+		if (!list_empty(&trans->transaction->pending_snapshots))
+			snap_pending = 1;
+
+		WARN_ON(cur_trans != trans->transaction);
+		prepare_to_wait(&cur_trans->writer_wait, &wait,
+				TASK_UNINTERRUPTIBLE);
+
+		if (cur_trans->num_writers > 1)
+			timeout = MAX_SCHEDULE_TIMEOUT;
+		else
+			timeout = 1;
+
+		mutex_unlock(&root->fs_info->trans_mutex);
+
+		if (snap_pending) {
+			ret = btrfs_wait_ordered_extents(root, 1);
+			BUG_ON(ret);
+		}
+
+		schedule_timeout(timeout);
+
+		mutex_lock(&root->fs_info->trans_mutex);
+		finish_wait(&cur_trans->writer_wait, &wait);
+	} while (cur_trans->num_writers > 1 ||
+		 (cur_trans->num_joined != joined));
+
+	ret = create_pending_snapshots(trans, root->fs_info);
+	BUG_ON(ret);
+
+	WARN_ON(cur_trans != trans->transaction);
+
+	/* btrfs_commit_tree_roots is responsible for getting the
+	 * various roots consistent with each other.  Every pointer
+	 * in the tree of tree roots has to point to the most up to date
+	 * root for every subvolume and other tree.  So, we have to keep
+	 * the tree logging code from jumping in and changing any
+	 * of the trees.
+	 *
+	 * At this point in the commit, there can't be any tree-log
+	 * writers, but a little lower down we drop the trans mutex
+	 * and let new people in.  By holding the tree_log_mutex
+	 * from now until after the super is written, we avoid races
+	 * with the tree-log code.
+	 */
+	mutex_lock(&root->fs_info->tree_log_mutex);
+	/*
+	 * keep tree reloc code from adding new reloc trees
+	 */
+	mutex_lock(&root->fs_info->tree_reloc_mutex);
+
+
+	ret = add_dirty_roots(trans, &root->fs_info->fs_roots_radix,
+			      &dirty_fs_roots);
+	BUG_ON(ret);
+
+	/* add_dirty_roots gets rid of all the tree log roots, it is now
+	 * safe to free the root of tree log roots
+	 */
+	btrfs_free_log_root_tree(trans, root->fs_info);
+
+	ret = btrfs_commit_tree_roots(trans, root);
+	BUG_ON(ret);
+
+	cur_trans = root->fs_info->running_transaction;
+	spin_lock(&root->fs_info->new_trans_lock);
+	root->fs_info->running_transaction = NULL;
+	spin_unlock(&root->fs_info->new_trans_lock);
+	btrfs_set_super_generation(&root->fs_info->super_copy,
+				   cur_trans->transid);
+	btrfs_set_super_root(&root->fs_info->super_copy,
+			     root->fs_info->tree_root->node->start);
+	btrfs_set_super_root_level(&root->fs_info->super_copy,
+			   btrfs_header_level(root->fs_info->tree_root->node));
+
+	btrfs_set_super_chunk_root(&root->fs_info->super_copy,
+				   chunk_root->node->start);
+	btrfs_set_super_chunk_root_level(&root->fs_info->super_copy,
+					 btrfs_header_level(chunk_root->node));
+	btrfs_set_super_chunk_root_generation(&root->fs_info->super_copy,
+				btrfs_header_generation(chunk_root->node));
+
+	if (!root->fs_info->log_root_recovering) {
+		btrfs_set_super_log_root(&root->fs_info->super_copy, 0);
+		btrfs_set_super_log_root_level(&root->fs_info->super_copy, 0);
+	}
+
+	memcpy(&root->fs_info->super_for_commit, &root->fs_info->super_copy,
+	       sizeof(root->fs_info->super_copy));
+
+	btrfs_copy_pinned(root, pinned_copy);
+
+	trans->transaction->blocked = 0;
+	wake_up(&root->fs_info->transaction_throttle);
+	wake_up(&root->fs_info->transaction_wait);
+
+	mutex_unlock(&root->fs_info->trans_mutex);
+	ret = btrfs_write_and_wait_transaction(trans, root);
+	BUG_ON(ret);
+	write_ctree_super(trans, root, 0);
+
+	/*
+	 * the super is written, we can safely allow the tree-loggers
+	 * to go about their business
+	 */
+	mutex_unlock(&root->fs_info->tree_log_mutex);
+
+	btrfs_finish_extent_commit(trans, root, pinned_copy);
+	kfree(pinned_copy);
+
+	btrfs_drop_dead_reloc_roots(root);
+	mutex_unlock(&root->fs_info->tree_reloc_mutex);
+
+	/* do the directory inserts of any pending snapshot creations */
+	finish_pending_snapshots(trans, root->fs_info);
+
+	mutex_lock(&root->fs_info->trans_mutex);
+
+	cur_trans->commit_done = 1;
+	root->fs_info->last_trans_committed = cur_trans->transid;
+	wake_up(&cur_trans->commit_wait);
+
+	put_transaction(cur_trans);
+	put_transaction(cur_trans);
+
+	list_splice_init(&dirty_fs_roots, &root->fs_info->dead_roots);
+	if (root->fs_info->closing)
+		list_splice_init(&root->fs_info->dead_roots, &dirty_fs_roots);
+
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	kmem_cache_free(btrfs_trans_handle_cachep, trans);
+
+	if (root->fs_info->closing)
+		drop_dirty_roots(root->fs_info->tree_root, &dirty_fs_roots);
+	return ret;
+}
+
+/*
+ * interface function to delete all the snapshots we have scheduled for deletion
+ */
+int btrfs_clean_old_snapshots(struct btrfs_root *root)
+{
+	struct list_head dirty_roots;
+	INIT_LIST_HEAD(&dirty_roots);
+again:
+	mutex_lock(&root->fs_info->trans_mutex);
+	list_splice_init(&root->fs_info->dead_roots, &dirty_roots);
+	mutex_unlock(&root->fs_info->trans_mutex);
+
+	if (!list_empty(&dirty_roots)) {
+		drop_dirty_roots(root, &dirty_roots);
+		goto again;
+	}
+	return 0;
+}
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
new file mode 100644
index 00000000000..ea292117f88
--- /dev/null
+++ b/fs/btrfs/transaction.h
@@ -0,0 +1,106 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_TRANSACTION__
+#define __BTRFS_TRANSACTION__
+#include "btrfs_inode.h"
+
+struct btrfs_transaction {
+	u64 transid;
+	unsigned long num_writers;
+	unsigned long num_joined;
+	int in_commit;
+	int use_count;
+	int commit_done;
+	int blocked;
+	struct list_head list;
+	struct extent_io_tree dirty_pages;
+	unsigned long start_time;
+	wait_queue_head_t writer_wait;
+	wait_queue_head_t commit_wait;
+	struct list_head pending_snapshots;
+};
+
+struct btrfs_trans_handle {
+	u64 transid;
+	unsigned long blocks_reserved;
+	unsigned long blocks_used;
+	struct btrfs_transaction *transaction;
+	u64 block_group;
+	u64 alloc_exclude_start;
+	u64 alloc_exclude_nr;
+};
+
+struct btrfs_pending_snapshot {
+	struct dentry *dentry;
+	struct btrfs_root *root;
+	char *name;
+	struct btrfs_key root_key;
+	struct list_head list;
+};
+
+struct btrfs_dirty_root {
+	struct list_head list;
+	struct btrfs_root *root;
+	struct btrfs_root *latest_root;
+};
+
+static inline void btrfs_set_trans_block_group(struct btrfs_trans_handle *trans,
+					       struct inode *inode)
+{
+	trans->block_group = BTRFS_I(inode)->block_group;
+}
+
+static inline void btrfs_update_inode_block_group(
+					  struct btrfs_trans_handle *trans,
+					  struct inode *inode)
+{
+	BTRFS_I(inode)->block_group = trans->block_group;
+}
+
+static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans,
+					      struct inode *inode)
+{
+	BTRFS_I(inode)->last_trans = trans->transaction->transid;
+}
+
+int btrfs_end_transaction(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root);
+struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
+						   int num_blocks);
+struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
+						   int num_blocks);
+struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
+						   int num_blocks);
+int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root);
+int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root);
+
+int btrfs_add_dead_root(struct btrfs_root *root, struct btrfs_root *latest);
+int btrfs_defrag_root(struct btrfs_root *root, int cacheonly);
+int btrfs_clean_old_snapshots(struct btrfs_root *root);
+int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root);
+int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root);
+void btrfs_throttle(struct btrfs_root *root);
+int btrfs_record_root_in_trans(struct btrfs_root *root);
+int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
+					struct extent_io_tree *dirty_pages);
+#endif
diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c
new file mode 100644
index 00000000000..3e8358c3616
--- /dev/null
+++ b/fs/btrfs/tree-defrag.c
@@ -0,0 +1,147 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include "ctree.h"
+#include "disk-io.h"
+#include "print-tree.h"
+#include "transaction.h"
+#include "locking.h"
+
+/* defrag all the leaves in a given btree.  If cache_only == 1, don't read
+ * things from disk, otherwise read all the leaves and try to get key order to
+ * better reflect disk order
+ */
+
+int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root, int cache_only)
+{
+	struct btrfs_path *path = NULL;
+	struct btrfs_key key;
+	int ret = 0;
+	int wret;
+	int level;
+	int orig_level;
+	int is_extent = 0;
+	int next_key_ret = 0;
+	u64 last_ret = 0;
+	u64 min_trans = 0;
+
+	if (cache_only)
+		goto out;
+
+	if (root->fs_info->extent_root == root) {
+		/*
+		 * there's recursion here right now in the tree locking,
+		 * we can't defrag the extent root without deadlock
+		 */
+		goto out;
+	}
+
+	if (root->ref_cows == 0 && !is_extent)
+		goto out;
+
+	if (btrfs_test_opt(root, SSD))
+		goto out;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	level = btrfs_header_level(root->node);
+	orig_level = level;
+
+	if (level == 0)
+		goto out;
+
+	if (root->defrag_progress.objectid == 0) {
+		struct extent_buffer *root_node;
+		u32 nritems;
+
+		root_node = btrfs_lock_root_node(root);
+		nritems = btrfs_header_nritems(root_node);
+		root->defrag_max.objectid = 0;
+		/* from above we know this is not a leaf */
+		btrfs_node_key_to_cpu(root_node, &root->defrag_max,
+				      nritems - 1);
+		btrfs_tree_unlock(root_node);
+		free_extent_buffer(root_node);
+		memset(&key, 0, sizeof(key));
+	} else {
+		memcpy(&key, &root->defrag_progress, sizeof(key));
+	}
+
+	path->keep_locks = 1;
+	if (cache_only)
+		min_trans = root->defrag_trans_start;
+
+	ret = btrfs_search_forward(root, &key, NULL, path,
+				   cache_only, min_trans);
+	if (ret < 0)
+		goto out;
+	if (ret > 0) {
+		ret = 0;
+		goto out;
+	}
+	btrfs_release_path(root, path);
+	wret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+
+	if (wret < 0) {
+		ret = wret;
+		goto out;
+	}
+	if (!path->nodes[1]) {
+		ret = 0;
+		goto out;
+	}
+	path->slots[1] = btrfs_header_nritems(path->nodes[1]);
+	next_key_ret = btrfs_find_next_key(root, path, &key, 1, cache_only,
+					   min_trans);
+	ret = btrfs_realloc_node(trans, root,
+				 path->nodes[1], 0,
+				 cache_only, &last_ret,
+				 &root->defrag_progress);
+	WARN_ON(ret && ret != -EAGAIN);
+	if (next_key_ret == 0) {
+		memcpy(&root->defrag_progress, &key, sizeof(key));
+		ret = -EAGAIN;
+	}
+
+	btrfs_release_path(root, path);
+	if (is_extent)
+		btrfs_extent_post_op(trans, root);
+out:
+	if (path)
+		btrfs_free_path(path);
+	if (ret == -EAGAIN) {
+		if (root->defrag_max.objectid > root->defrag_progress.objectid)
+			goto done;
+		if (root->defrag_max.type > root->defrag_progress.type)
+			goto done;
+		if (root->defrag_max.offset > root->defrag_progress.offset)
+			goto done;
+		ret = 0;
+	}
+done:
+	if (ret != -EAGAIN) {
+		memset(&root->defrag_progress, 0,
+		       sizeof(root->defrag_progress));
+		root->defrag_trans_start = trans->transid;
+	}
+	return ret;
+}
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
new file mode 100644
index 00000000000..d81cda2e077
--- /dev/null
+++ b/fs/btrfs/tree-log.c
@@ -0,0 +1,2898 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include "ctree.h"
+#include "transaction.h"
+#include "disk-io.h"
+#include "locking.h"
+#include "print-tree.h"
+#include "compat.h"
+#include "tree-log.h"
+
+/* magic values for the inode_only field in btrfs_log_inode:
+ *
+ * LOG_INODE_ALL means to log everything
+ * LOG_INODE_EXISTS means to log just enough to recreate the inode
+ * during log replay
+ */
+#define LOG_INODE_ALL 0
+#define LOG_INODE_EXISTS 1
+
+/*
+ * stages for the tree walking.  The first
+ * stage (0) is to only pin down the blocks we find
+ * the second stage (1) is to make sure that all the inodes
+ * we find in the log are created in the subvolume.
+ *
+ * The last stage is to deal with directories and links and extents
+ * and all the other fun semantics
+ */
+#define LOG_WALK_PIN_ONLY 0
+#define LOG_WALK_REPLAY_INODES 1
+#define LOG_WALK_REPLAY_ALL 2
+
+static int __btrfs_log_inode(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, struct inode *inode,
+			     int inode_only);
+static int link_to_fixup_dir(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root,
+			     struct btrfs_path *path, u64 objectid);
+
+/*
+ * tree logging is a special write ahead log used to make sure that
+ * fsyncs and O_SYNCs can happen without doing full tree commits.
+ *
+ * Full tree commits are expensive because they require commonly
+ * modified blocks to be recowed, creating many dirty pages in the
+ * extent tree an 4x-6x higher write load than ext3.
+ *
+ * Instead of doing a tree commit on every fsync, we use the
+ * key ranges and transaction ids to find items for a given file or directory
+ * that have changed in this transaction.  Those items are copied into
+ * a special tree (one per subvolume root), that tree is written to disk
+ * and then the fsync is considered complete.
+ *
+ * After a crash, items are copied out of the log-tree back into the
+ * subvolume tree.  Any file data extents found are recorded in the extent
+ * allocation tree, and the log-tree freed.
+ *
+ * The log tree is read three times, once to pin down all the extents it is
+ * using in ram and once, once to create all the inodes logged in the tree
+ * and once to do all the other items.
+ */
+
+/*
+ * btrfs_add_log_tree adds a new per-subvolume log tree into the
+ * tree of log tree roots.  This must be called with a tree log transaction
+ * running (see start_log_trans).
+ */
+static int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *root)
+{
+	struct btrfs_key key;
+	struct btrfs_root_item root_item;
+	struct btrfs_inode_item *inode_item;
+	struct extent_buffer *leaf;
+	struct btrfs_root *new_root = root;
+	int ret;
+	u64 objectid = root->root_key.objectid;
+
+	leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
+				      BTRFS_TREE_LOG_OBJECTID,
+				      trans->transid, 0, 0, 0);
+	if (IS_ERR(leaf)) {
+		ret = PTR_ERR(leaf);
+		return ret;
+	}
+
+	btrfs_set_header_nritems(leaf, 0);
+	btrfs_set_header_level(leaf, 0);
+	btrfs_set_header_bytenr(leaf, leaf->start);
+	btrfs_set_header_generation(leaf, trans->transid);
+	btrfs_set_header_owner(leaf, BTRFS_TREE_LOG_OBJECTID);
+
+	write_extent_buffer(leaf, root->fs_info->fsid,
+			    (unsigned long)btrfs_header_fsid(leaf),
+			    BTRFS_FSID_SIZE);
+	btrfs_mark_buffer_dirty(leaf);
+
+	inode_item = &root_item.inode;
+	memset(inode_item, 0, sizeof(*inode_item));
+	inode_item->generation = cpu_to_le64(1);
+	inode_item->size = cpu_to_le64(3);
+	inode_item->nlink = cpu_to_le32(1);
+	inode_item->nbytes = cpu_to_le64(root->leafsize);
+	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+
+	btrfs_set_root_bytenr(&root_item, leaf->start);
+	btrfs_set_root_generation(&root_item, trans->transid);
+	btrfs_set_root_level(&root_item, 0);
+	btrfs_set_root_refs(&root_item, 0);
+	btrfs_set_root_used(&root_item, 0);
+
+	memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress));
+	root_item.drop_level = 0;
+
+	btrfs_tree_unlock(leaf);
+	free_extent_buffer(leaf);
+	leaf = NULL;
+
+	btrfs_set_root_dirid(&root_item, 0);
+
+	key.objectid = BTRFS_TREE_LOG_OBJECTID;
+	key.offset = objectid;
+	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+	ret = btrfs_insert_root(trans, root->fs_info->log_root_tree, &key,
+				&root_item);
+	if (ret)
+		goto fail;
+
+	new_root = btrfs_read_fs_root_no_radix(root->fs_info->log_root_tree,
+					       &key);
+	BUG_ON(!new_root);
+
+	WARN_ON(root->log_root);
+	root->log_root = new_root;
+
+	/*
+	 * log trees do not get reference counted because they go away
+	 * before a real commit is actually done.  They do store pointers
+	 * to file data extents, and those reference counts still get
+	 * updated (along with back refs to the log tree).
+	 */
+	new_root->ref_cows = 0;
+	new_root->last_trans = trans->transid;
+
+	/*
+	 * we need to make sure the root block for this new tree
+	 * is marked as dirty in the dirty_log_pages tree.  This
+	 * is how it gets flushed down to disk at tree log commit time.
+	 *
+	 * the tree logging mutex keeps others from coming in and changing
+	 * the new_root->node, so we can safely access it here
+	 */
+	set_extent_dirty(&new_root->dirty_log_pages, new_root->node->start,
+			 new_root->node->start + new_root->node->len - 1,
+			 GFP_NOFS);
+
+fail:
+	return ret;
+}
+
+/*
+ * start a sub transaction and setup the log tree
+ * this increments the log tree writer count to make the people
+ * syncing the tree wait for us to finish
+ */
+static int start_log_trans(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root)
+{
+	int ret;
+	mutex_lock(&root->fs_info->tree_log_mutex);
+	if (!root->fs_info->log_root_tree) {
+		ret = btrfs_init_log_root_tree(trans, root->fs_info);
+		BUG_ON(ret);
+	}
+	if (!root->log_root) {
+		ret = btrfs_add_log_tree(trans, root);
+		BUG_ON(ret);
+	}
+	atomic_inc(&root->fs_info->tree_log_writers);
+	root->fs_info->tree_log_batch++;
+	mutex_unlock(&root->fs_info->tree_log_mutex);
+	return 0;
+}
+
+/*
+ * returns 0 if there was a log transaction running and we were able
+ * to join, or returns -ENOENT if there were not transactions
+ * in progress
+ */
+static int join_running_log_trans(struct btrfs_root *root)
+{
+	int ret = -ENOENT;
+
+	smp_mb();
+	if (!root->log_root)
+		return -ENOENT;
+
+	mutex_lock(&root->fs_info->tree_log_mutex);
+	if (root->log_root) {
+		ret = 0;
+		atomic_inc(&root->fs_info->tree_log_writers);
+		root->fs_info->tree_log_batch++;
+	}
+	mutex_unlock(&root->fs_info->tree_log_mutex);
+	return ret;
+}
+
+/*
+ * indicate we're done making changes to the log tree
+ * and wake up anyone waiting to do a sync
+ */
+static int end_log_trans(struct btrfs_root *root)
+{
+	atomic_dec(&root->fs_info->tree_log_writers);
+	smp_mb();
+	if (waitqueue_active(&root->fs_info->tree_log_wait))
+		wake_up(&root->fs_info->tree_log_wait);
+	return 0;
+}
+
+
+/*
+ * the walk control struct is used to pass state down the chain when
+ * processing the log tree.  The stage field tells us which part
+ * of the log tree processing we are currently doing.  The others
+ * are state fields used for that specific part
+ */
+struct walk_control {
+	/* should we free the extent on disk when done?  This is used
+	 * at transaction commit time while freeing a log tree
+	 */
+	int free;
+
+	/* should we write out the extent buffer?  This is used
+	 * while flushing the log tree to disk during a sync
+	 */
+	int write;
+
+	/* should we wait for the extent buffer io to finish?  Also used
+	 * while flushing the log tree to disk for a sync
+	 */
+	int wait;
+
+	/* pin only walk, we record which extents on disk belong to the
+	 * log trees
+	 */
+	int pin;
+
+	/* what stage of the replay code we're currently in */
+	int stage;
+
+	/* the root we are currently replaying */
+	struct btrfs_root *replay_dest;
+
+	/* the trans handle for the current replay */
+	struct btrfs_trans_handle *trans;
+
+	/* the function that gets used to process blocks we find in the
+	 * tree.  Note the extent_buffer might not be up to date when it is
+	 * passed in, and it must be checked or read if you need the data
+	 * inside it
+	 */
+	int (*process_func)(struct btrfs_root *log, struct extent_buffer *eb,
+			    struct walk_control *wc, u64 gen);
+};
+
+/*
+ * process_func used to pin down extents, write them or wait on them
+ */
+static int process_one_buffer(struct btrfs_root *log,
+			      struct extent_buffer *eb,
+			      struct walk_control *wc, u64 gen)
+{
+	if (wc->pin) {
+		mutex_lock(&log->fs_info->pinned_mutex);
+		btrfs_update_pinned_extents(log->fs_info->extent_root,
+					    eb->start, eb->len, 1);
+		mutex_unlock(&log->fs_info->pinned_mutex);
+	}
+
+	if (btrfs_buffer_uptodate(eb, gen)) {
+		if (wc->write)
+			btrfs_write_tree_block(eb);
+		if (wc->wait)
+			btrfs_wait_tree_block_writeback(eb);
+	}
+	return 0;
+}
+
+/*
+ * Item overwrite used by replay and tree logging.  eb, slot and key all refer
+ * to the src data we are copying out.
+ *
+ * root is the tree we are copying into, and path is a scratch
+ * path for use in this function (it should be released on entry and
+ * will be released on exit).
+ *
+ * If the key is already in the destination tree the existing item is
+ * overwritten.  If the existing item isn't big enough, it is extended.
+ * If it is too large, it is truncated.
+ *
+ * If the key isn't in the destination yet, a new item is inserted.
+ */
+static noinline int overwrite_item(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root,
+				   struct btrfs_path *path,
+				   struct extent_buffer *eb, int slot,
+				   struct btrfs_key *key)
+{
+	int ret;
+	u32 item_size;
+	u64 saved_i_size = 0;
+	int save_old_i_size = 0;
+	unsigned long src_ptr;
+	unsigned long dst_ptr;
+	int overwrite_root = 0;
+
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+		overwrite_root = 1;
+
+	item_size = btrfs_item_size_nr(eb, slot);
+	src_ptr = btrfs_item_ptr_offset(eb, slot);
+
+	/* look for the key in the destination tree */
+	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
+	if (ret == 0) {
+		char *src_copy;
+		char *dst_copy;
+		u32 dst_size = btrfs_item_size_nr(path->nodes[0],
+						  path->slots[0]);
+		if (dst_size != item_size)
+			goto insert;
+
+		if (item_size == 0) {
+			btrfs_release_path(root, path);
+			return 0;
+		}
+		dst_copy = kmalloc(item_size, GFP_NOFS);
+		src_copy = kmalloc(item_size, GFP_NOFS);
+
+		read_extent_buffer(eb, src_copy, src_ptr, item_size);
+
+		dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+		read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
+				   item_size);
+		ret = memcmp(dst_copy, src_copy, item_size);
+
+		kfree(dst_copy);
+		kfree(src_copy);
+		/*
+		 * they have the same contents, just return, this saves
+		 * us from cowing blocks in the destination tree and doing
+		 * extra writes that may not have been done by a previous
+		 * sync
+		 */
+		if (ret == 0) {
+			btrfs_release_path(root, path);
+			return 0;
+		}
+
+	}
+insert:
+	btrfs_release_path(root, path);
+	/* try to insert the key into the destination tree */
+	ret = btrfs_insert_empty_item(trans, root, path,
+				      key, item_size);
+
+	/* make sure any existing item is the correct size */
+	if (ret == -EEXIST) {
+		u32 found_size;
+		found_size = btrfs_item_size_nr(path->nodes[0],
+						path->slots[0]);
+		if (found_size > item_size) {
+			btrfs_truncate_item(trans, root, path, item_size, 1);
+		} else if (found_size < item_size) {
+			ret = btrfs_extend_item(trans, root, path,
+						item_size - found_size);
+			BUG_ON(ret);
+		}
+	} else if (ret) {
+		BUG();
+	}
+	dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
+					path->slots[0]);
+
+	/* don't overwrite an existing inode if the generation number
+	 * was logged as zero.  This is done when the tree logging code
+	 * is just logging an inode to make sure it exists after recovery.
+	 *
+	 * Also, don't overwrite i_size on directories during replay.
+	 * log replay inserts and removes directory items based on the
+	 * state of the tree found in the subvolume, and i_size is modified
+	 * as it goes
+	 */
+	if (key->type == BTRFS_INODE_ITEM_KEY && ret == -EEXIST) {
+		struct btrfs_inode_item *src_item;
+		struct btrfs_inode_item *dst_item;
+
+		src_item = (struct btrfs_inode_item *)src_ptr;
+		dst_item = (struct btrfs_inode_item *)dst_ptr;
+
+		if (btrfs_inode_generation(eb, src_item) == 0)
+			goto no_copy;
+
+		if (overwrite_root &&
+		    S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
+		    S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
+			save_old_i_size = 1;
+			saved_i_size = btrfs_inode_size(path->nodes[0],
+							dst_item);
+		}
+	}
+
+	copy_extent_buffer(path->nodes[0], eb, dst_ptr,
+			   src_ptr, item_size);
+
+	if (save_old_i_size) {
+		struct btrfs_inode_item *dst_item;
+		dst_item = (struct btrfs_inode_item *)dst_ptr;
+		btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
+	}
+
+	/* make sure the generation is filled in */
+	if (key->type == BTRFS_INODE_ITEM_KEY) {
+		struct btrfs_inode_item *dst_item;
+		dst_item = (struct btrfs_inode_item *)dst_ptr;
+		if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
+			btrfs_set_inode_generation(path->nodes[0], dst_item,
+						   trans->transid);
+		}
+	}
+no_copy:
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	btrfs_release_path(root, path);
+	return 0;
+}
+
+/*
+ * simple helper to read an inode off the disk from a given root
+ * This can only be called for subvolume roots and not for the log
+ */
+static noinline struct inode *read_one_inode(struct btrfs_root *root,
+					     u64 objectid)
+{
+	struct inode *inode;
+	inode = btrfs_iget_locked(root->fs_info->sb, objectid, root);
+	if (inode->i_state & I_NEW) {
+		BTRFS_I(inode)->root = root;
+		BTRFS_I(inode)->location.objectid = objectid;
+		BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
+		BTRFS_I(inode)->location.offset = 0;
+		btrfs_read_locked_inode(inode);
+		unlock_new_inode(inode);
+
+	}
+	if (is_bad_inode(inode)) {
+		iput(inode);
+		inode = NULL;
+	}
+	return inode;
+}
+
+/* replays a single extent in 'eb' at 'slot' with 'key' into the
+ * subvolume 'root'.  path is released on entry and should be released
+ * on exit.
+ *
+ * extents in the log tree have not been allocated out of the extent
+ * tree yet.  So, this completes the allocation, taking a reference
+ * as required if the extent already exists or creating a new extent
+ * if it isn't in the extent allocation tree yet.
+ *
+ * The extent is inserted into the file, dropping any existing extents
+ * from the file that overlap the new one.
+ */
+static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      struct extent_buffer *eb, int slot,
+				      struct btrfs_key *key)
+{
+	int found_type;
+	u64 mask = root->sectorsize - 1;
+	u64 extent_end;
+	u64 alloc_hint;
+	u64 start = key->offset;
+	u64 saved_nbytes;
+	struct btrfs_file_extent_item *item;
+	struct inode *inode = NULL;
+	unsigned long size;
+	int ret = 0;
+
+	item = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
+	found_type = btrfs_file_extent_type(eb, item);
+
+	if (found_type == BTRFS_FILE_EXTENT_REG ||
+	    found_type == BTRFS_FILE_EXTENT_PREALLOC)
+		extent_end = start + btrfs_file_extent_num_bytes(eb, item);
+	else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+		size = btrfs_file_extent_inline_len(eb, item);
+		extent_end = (start + size + mask) & ~mask;
+	} else {
+		ret = 0;
+		goto out;
+	}
+
+	inode = read_one_inode(root, key->objectid);
+	if (!inode) {
+		ret = -EIO;
+		goto out;
+	}
+
+	/*
+	 * first check to see if we already have this extent in the
+	 * file.  This must be done before the btrfs_drop_extents run
+	 * so we don't try to drop this extent.
+	 */
+	ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+				       start, 0);
+
+	if (ret == 0 &&
+	    (found_type == BTRFS_FILE_EXTENT_REG ||
+	     found_type == BTRFS_FILE_EXTENT_PREALLOC)) {
+		struct btrfs_file_extent_item cmp1;
+		struct btrfs_file_extent_item cmp2;
+		struct btrfs_file_extent_item *existing;
+		struct extent_buffer *leaf;
+
+		leaf = path->nodes[0];
+		existing = btrfs_item_ptr(leaf, path->slots[0],
+					  struct btrfs_file_extent_item);
+
+		read_extent_buffer(eb, &cmp1, (unsigned long)item,
+				   sizeof(cmp1));
+		read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
+				   sizeof(cmp2));
+
+		/*
+		 * we already have a pointer to this exact extent,
+		 * we don't have to do anything
+		 */
+		if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
+			btrfs_release_path(root, path);
+			goto out;
+		}
+	}
+	btrfs_release_path(root, path);
+
+	saved_nbytes = inode_get_bytes(inode);
+	/* drop any overlapping extents */
+	ret = btrfs_drop_extents(trans, root, inode,
+			 start, extent_end, start, &alloc_hint);
+	BUG_ON(ret);
+
+	if (found_type == BTRFS_FILE_EXTENT_REG ||
+	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
+		unsigned long dest_offset;
+		struct btrfs_key ins;
+
+		ret = btrfs_insert_empty_item(trans, root, path, key,
+					      sizeof(*item));
+		BUG_ON(ret);
+		dest_offset = btrfs_item_ptr_offset(path->nodes[0],
+						    path->slots[0]);
+		copy_extent_buffer(path->nodes[0], eb, dest_offset,
+				(unsigned long)item,  sizeof(*item));
+
+		ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
+		ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
+		ins.type = BTRFS_EXTENT_ITEM_KEY;
+
+		if (ins.objectid > 0) {
+			u64 csum_start;
+			u64 csum_end;
+			LIST_HEAD(ordered_sums);
+			/*
+			 * is this extent already allocated in the extent
+			 * allocation tree?  If so, just add a reference
+			 */
+			ret = btrfs_lookup_extent(root, ins.objectid,
+						ins.offset);
+			if (ret == 0) {
+				ret = btrfs_inc_extent_ref(trans, root,
+						ins.objectid, ins.offset,
+						path->nodes[0]->start,
+						root->root_key.objectid,
+						trans->transid, key->objectid);
+			} else {
+				/*
+				 * insert the extent pointer in the extent
+				 * allocation tree
+				 */
+				ret = btrfs_alloc_logged_extent(trans, root,
+						path->nodes[0]->start,
+						root->root_key.objectid,
+						trans->transid, key->objectid,
+						&ins);
+				BUG_ON(ret);
+			}
+			btrfs_release_path(root, path);
+
+			if (btrfs_file_extent_compression(eb, item)) {
+				csum_start = ins.objectid;
+				csum_end = csum_start + ins.offset;
+			} else {
+				csum_start = ins.objectid +
+					btrfs_file_extent_offset(eb, item);
+				csum_end = csum_start +
+					btrfs_file_extent_num_bytes(eb, item);
+			}
+
+			ret = btrfs_lookup_csums_range(root->log_root,
+						csum_start, csum_end - 1,
+						&ordered_sums);
+			BUG_ON(ret);
+			while (!list_empty(&ordered_sums)) {
+				struct btrfs_ordered_sum *sums;
+				sums = list_entry(ordered_sums.next,
+						struct btrfs_ordered_sum,
+						list);
+				ret = btrfs_csum_file_blocks(trans,
+						root->fs_info->csum_root,
+						sums);
+				BUG_ON(ret);
+				list_del(&sums->list);
+				kfree(sums);
+			}
+		} else {
+			btrfs_release_path(root, path);
+		}
+	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+		/* inline extents are easy, we just overwrite them */
+		ret = overwrite_item(trans, root, path, eb, slot, key);
+		BUG_ON(ret);
+	}
+
+	inode_set_bytes(inode, saved_nbytes);
+	btrfs_update_inode(trans, root, inode);
+out:
+	if (inode)
+		iput(inode);
+	return ret;
+}
+
+/*
+ * when cleaning up conflicts between the directory names in the
+ * subvolume, directory names in the log and directory names in the
+ * inode back references, we may have to unlink inodes from directories.
+ *
+ * This is a helper function to do the unlink of a specific directory
+ * item
+ */
+static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      struct inode *dir,
+				      struct btrfs_dir_item *di)
+{
+	struct inode *inode;
+	char *name;
+	int name_len;
+	struct extent_buffer *leaf;
+	struct btrfs_key location;
+	int ret;
+
+	leaf = path->nodes[0];
+
+	btrfs_dir_item_key_to_cpu(leaf, di, &location);
+	name_len = btrfs_dir_name_len(leaf, di);
+	name = kmalloc(name_len, GFP_NOFS);
+	read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
+	btrfs_release_path(root, path);
+
+	inode = read_one_inode(root, location.objectid);
+	BUG_ON(!inode);
+
+	ret = link_to_fixup_dir(trans, root, path, location.objectid);
+	BUG_ON(ret);
+	ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
+	BUG_ON(ret);
+	kfree(name);
+
+	iput(inode);
+	return ret;
+}
+
+/*
+ * helper function to see if a given name and sequence number found
+ * in an inode back reference are already in a directory and correctly
+ * point to this inode
+ */
+static noinline int inode_in_dir(struct btrfs_root *root,
+				 struct btrfs_path *path,
+				 u64 dirid, u64 objectid, u64 index,
+				 const char *name, int name_len)
+{
+	struct btrfs_dir_item *di;
+	struct btrfs_key location;
+	int match = 0;
+
+	di = btrfs_lookup_dir_index_item(NULL, root, path, dirid,
+					 index, name, name_len, 0);
+	if (di && !IS_ERR(di)) {
+		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
+		if (location.objectid != objectid)
+			goto out;
+	} else
+		goto out;
+	btrfs_release_path(root, path);
+
+	di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
+	if (di && !IS_ERR(di)) {
+		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
+		if (location.objectid != objectid)
+			goto out;
+	} else
+		goto out;
+	match = 1;
+out:
+	btrfs_release_path(root, path);
+	return match;
+}
+
+/*
+ * helper function to check a log tree for a named back reference in
+ * an inode.  This is used to decide if a back reference that is
+ * found in the subvolume conflicts with what we find in the log.
+ *
+ * inode backreferences may have multiple refs in a single item,
+ * during replay we process one reference at a time, and we don't
+ * want to delete valid links to a file from the subvolume if that
+ * link is also in the log.
+ */
+static noinline int backref_in_log(struct btrfs_root *log,
+				   struct btrfs_key *key,
+				   char *name, int namelen)
+{
+	struct btrfs_path *path;
+	struct btrfs_inode_ref *ref;
+	unsigned long ptr;
+	unsigned long ptr_end;
+	unsigned long name_ptr;
+	int found_name_len;
+	int item_size;
+	int ret;
+	int match = 0;
+
+	path = btrfs_alloc_path();
+	ret = btrfs_search_slot(NULL, log, key, path, 0, 0);
+	if (ret != 0)
+		goto out;
+
+	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
+	ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+	ptr_end = ptr + item_size;
+	while (ptr < ptr_end) {
+		ref = (struct btrfs_inode_ref *)ptr;
+		found_name_len = btrfs_inode_ref_name_len(path->nodes[0], ref);
+		if (found_name_len == namelen) {
+			name_ptr = (unsigned long)(ref + 1);
+			ret = memcmp_extent_buffer(path->nodes[0], name,
+						   name_ptr, namelen);
+			if (ret == 0) {
+				match = 1;
+				goto out;
+			}
+		}
+		ptr = (unsigned long)(ref + 1) + found_name_len;
+	}
+out:
+	btrfs_free_path(path);
+	return match;
+}
+
+
+/*
+ * replay one inode back reference item found in the log tree.
+ * eb, slot and key refer to the buffer and key found in the log tree.
+ * root is the destination we are replaying into, and path is for temp
+ * use by this function.  (it should be released on return).
+ */
+static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *root,
+				  struct btrfs_root *log,
+				  struct btrfs_path *path,
+				  struct extent_buffer *eb, int slot,
+				  struct btrfs_key *key)
+{
+	struct inode *dir;
+	int ret;
+	struct btrfs_key location;
+	struct btrfs_inode_ref *ref;
+	struct btrfs_dir_item *di;
+	struct inode *inode;
+	char *name;
+	int namelen;
+	unsigned long ref_ptr;
+	unsigned long ref_end;
+
+	location.objectid = key->objectid;
+	location.type = BTRFS_INODE_ITEM_KEY;
+	location.offset = 0;
+
+	/*
+	 * it is possible that we didn't log all the parent directories
+	 * for a given inode.  If we don't find the dir, just don't
+	 * copy the back ref in.  The link count fixup code will take
+	 * care of the rest
+	 */
+	dir = read_one_inode(root, key->offset);
+	if (!dir)
+		return -ENOENT;
+
+	inode = read_one_inode(root, key->objectid);
+	BUG_ON(!dir);
+
+	ref_ptr = btrfs_item_ptr_offset(eb, slot);
+	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
+
+again:
+	ref = (struct btrfs_inode_ref *)ref_ptr;
+
+	namelen = btrfs_inode_ref_name_len(eb, ref);
+	name = kmalloc(namelen, GFP_NOFS);
+	BUG_ON(!name);
+
+	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
+
+	/* if we already have a perfect match, we're done */
+	if (inode_in_dir(root, path, dir->i_ino, inode->i_ino,
+			 btrfs_inode_ref_index(eb, ref),
+			 name, namelen)) {
+		goto out;
+	}
+
+	/*
+	 * look for a conflicting back reference in the metadata.
+	 * if we find one we have to unlink that name of the file
+	 * before we add our new link.  Later on, we overwrite any
+	 * existing back reference, and we don't want to create
+	 * dangling pointers in the directory.
+	 */
+conflict_again:
+	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
+	if (ret == 0) {
+		char *victim_name;
+		int victim_name_len;
+		struct btrfs_inode_ref *victim_ref;
+		unsigned long ptr;
+		unsigned long ptr_end;
+		struct extent_buffer *leaf = path->nodes[0];
+
+		/* are we trying to overwrite a back ref for the root directory
+		 * if so, just jump out, we're done
+		 */
+		if (key->objectid == key->offset)
+			goto out_nowrite;
+
+		/* check all the names in this back reference to see
+		 * if they are in the log.  if so, we allow them to stay
+		 * otherwise they must be unlinked as a conflict
+		 */
+		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+		ptr_end = ptr + btrfs_item_size_nr(leaf, path->slots[0]);
+		while (ptr < ptr_end) {
+			victim_ref = (struct btrfs_inode_ref *)ptr;
+			victim_name_len = btrfs_inode_ref_name_len(leaf,
+								   victim_ref);
+			victim_name = kmalloc(victim_name_len, GFP_NOFS);
+			BUG_ON(!victim_name);
+
+			read_extent_buffer(leaf, victim_name,
+					   (unsigned long)(victim_ref + 1),
+					   victim_name_len);
+
+			if (!backref_in_log(log, key, victim_name,
+					    victim_name_len)) {
+				btrfs_inc_nlink(inode);
+				btrfs_release_path(root, path);
+				ret = btrfs_unlink_inode(trans, root, dir,
+							 inode, victim_name,
+							 victim_name_len);
+				kfree(victim_name);
+				btrfs_release_path(root, path);
+				goto conflict_again;
+			}
+			kfree(victim_name);
+			ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
+		}
+		BUG_ON(ret);
+	}
+	btrfs_release_path(root, path);
+
+	/* look for a conflicting sequence number */
+	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+					 btrfs_inode_ref_index(eb, ref),
+					 name, namelen, 0);
+	if (di && !IS_ERR(di)) {
+		ret = drop_one_dir_item(trans, root, path, dir, di);
+		BUG_ON(ret);
+	}
+	btrfs_release_path(root, path);
+
+
+	/* look for a conflicting name */
+	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+				   name, namelen, 0);
+	if (di && !IS_ERR(di)) {
+		ret = drop_one_dir_item(trans, root, path, dir, di);
+		BUG_ON(ret);
+	}
+	btrfs_release_path(root, path);
+
+	/* insert our name */
+	ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
+			     btrfs_inode_ref_index(eb, ref));
+	BUG_ON(ret);
+
+	btrfs_update_inode(trans, root, inode);
+
+out:
+	ref_ptr = (unsigned long)(ref + 1) + namelen;
+	kfree(name);
+	if (ref_ptr < ref_end)
+		goto again;
+
+	/* finally write the back reference in the inode */
+	ret = overwrite_item(trans, root, path, eb, slot, key);
+	BUG_ON(ret);
+
+out_nowrite:
+	btrfs_release_path(root, path);
+	iput(dir);
+	iput(inode);
+	return 0;
+}
+
+/*
+ * There are a few corners where the link count of the file can't
+ * be properly maintained during replay.  So, instead of adding
+ * lots of complexity to the log code, we just scan the backrefs
+ * for any file that has been through replay.
+ *
+ * The scan will update the link count on the inode to reflect the
+ * number of back refs found.  If it goes down to zero, the iput
+ * will free the inode.
+ */
+static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
+					   struct btrfs_root *root,
+					   struct inode *inode)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct btrfs_key key;
+	u64 nlink = 0;
+	unsigned long ptr;
+	unsigned long ptr_end;
+	int name_len;
+
+	key.objectid = inode->i_ino;
+	key.type = BTRFS_INODE_REF_KEY;
+	key.offset = (u64)-1;
+
+	path = btrfs_alloc_path();
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0)
+			break;
+		if (ret > 0) {
+			if (path->slots[0] == 0)
+				break;
+			path->slots[0]--;
+		}
+		btrfs_item_key_to_cpu(path->nodes[0], &key,
+				      path->slots[0]);
+		if (key.objectid != inode->i_ino ||
+		    key.type != BTRFS_INODE_REF_KEY)
+			break;
+		ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+		ptr_end = ptr + btrfs_item_size_nr(path->nodes[0],
+						   path->slots[0]);
+		while (ptr < ptr_end) {
+			struct btrfs_inode_ref *ref;
+
+			ref = (struct btrfs_inode_ref *)ptr;
+			name_len = btrfs_inode_ref_name_len(path->nodes[0],
+							    ref);
+			ptr = (unsigned long)(ref + 1) + name_len;
+			nlink++;
+		}
+
+		if (key.offset == 0)
+			break;
+		key.offset--;
+		btrfs_release_path(root, path);
+	}
+	btrfs_free_path(path);
+	if (nlink != inode->i_nlink) {
+		inode->i_nlink = nlink;
+		btrfs_update_inode(trans, root, inode);
+	}
+	BTRFS_I(inode)->index_cnt = (u64)-1;
+
+	return 0;
+}
+
+static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
+					    struct btrfs_root *root,
+					    struct btrfs_path *path)
+{
+	int ret;
+	struct btrfs_key key;
+	struct inode *inode;
+
+	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
+	key.type = BTRFS_ORPHAN_ITEM_KEY;
+	key.offset = (u64)-1;
+	while (1) {
+		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+		if (ret < 0)
+			break;
+
+		if (ret == 1) {
+			if (path->slots[0] == 0)
+				break;
+			path->slots[0]--;
+		}
+
+		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+		if (key.objectid != BTRFS_TREE_LOG_FIXUP_OBJECTID ||
+		    key.type != BTRFS_ORPHAN_ITEM_KEY)
+			break;
+
+		ret = btrfs_del_item(trans, root, path);
+		BUG_ON(ret);
+
+		btrfs_release_path(root, path);
+		inode = read_one_inode(root, key.offset);
+		BUG_ON(!inode);
+
+		ret = fixup_inode_link_count(trans, root, inode);
+		BUG_ON(ret);
+
+		iput(inode);
+
+		if (key.offset == 0)
+			break;
+		key.offset--;
+	}
+	btrfs_release_path(root, path);
+	return 0;
+}
+
+
+/*
+ * record a given inode in the fixup dir so we can check its link
+ * count when replay is done.  The link count is incremented here
+ * so the inode won't go away until we check it
+ */
+static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      u64 objectid)
+{
+	struct btrfs_key key;
+	int ret = 0;
+	struct inode *inode;
+
+	inode = read_one_inode(root, objectid);
+	BUG_ON(!inode);
+
+	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
+	btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
+	key.offset = objectid;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
+
+	btrfs_release_path(root, path);
+	if (ret == 0) {
+		btrfs_inc_nlink(inode);
+		btrfs_update_inode(trans, root, inode);
+	} else if (ret == -EEXIST) {
+		ret = 0;
+	} else {
+		BUG();
+	}
+	iput(inode);
+
+	return ret;
+}
+
+/*
+ * when replaying the log for a directory, we only insert names
+ * for inodes that actually exist.  This means an fsync on a directory
+ * does not implicitly fsync all the new files in it
+ */
+static noinline int insert_one_name(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root,
+				    struct btrfs_path *path,
+				    u64 dirid, u64 index,
+				    char *name, int name_len, u8 type,
+				    struct btrfs_key *location)
+{
+	struct inode *inode;
+	struct inode *dir;
+	int ret;
+
+	inode = read_one_inode(root, location->objectid);
+	if (!inode)
+		return -ENOENT;
+
+	dir = read_one_inode(root, dirid);
+	if (!dir) {
+		iput(inode);
+		return -EIO;
+	}
+	ret = btrfs_add_link(trans, dir, inode, name, name_len, 1, index);
+
+	/* FIXME, put inode into FIXUP list */
+
+	iput(inode);
+	iput(dir);
+	return ret;
+}
+
+/*
+ * take a single entry in a log directory item and replay it into
+ * the subvolume.
+ *
+ * if a conflicting item exists in the subdirectory already,
+ * the inode it points to is unlinked and put into the link count
+ * fix up tree.
+ *
+ * If a name from the log points to a file or directory that does
+ * not exist in the FS, it is skipped.  fsyncs on directories
+ * do not force down inodes inside that directory, just changes to the
+ * names or unlinks in a directory.
+ */
+static noinline int replay_one_name(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root,
+				    struct btrfs_path *path,
+				    struct extent_buffer *eb,
+				    struct btrfs_dir_item *di,
+				    struct btrfs_key *key)
+{
+	char *name;
+	int name_len;
+	struct btrfs_dir_item *dst_di;
+	struct btrfs_key found_key;
+	struct btrfs_key log_key;
+	struct inode *dir;
+	u8 log_type;
+	int exists;
+	int ret;
+
+	dir = read_one_inode(root, key->objectid);
+	BUG_ON(!dir);
+
+	name_len = btrfs_dir_name_len(eb, di);
+	name = kmalloc(name_len, GFP_NOFS);
+	log_type = btrfs_dir_type(eb, di);
+	read_extent_buffer(eb, name, (unsigned long)(di + 1),
+		   name_len);
+
+	btrfs_dir_item_key_to_cpu(eb, di, &log_key);
+	exists = btrfs_lookup_inode(trans, root, path, &log_key, 0);
+	if (exists == 0)
+		exists = 1;
+	else
+		exists = 0;
+	btrfs_release_path(root, path);
+
+	if (key->type == BTRFS_DIR_ITEM_KEY) {
+		dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
+				       name, name_len, 1);
+	} else if (key->type == BTRFS_DIR_INDEX_KEY) {
+		dst_di = btrfs_lookup_dir_index_item(trans, root, path,
+						     key->objectid,
+						     key->offset, name,
+						     name_len, 1);
+	} else {
+		BUG();
+	}
+	if (!dst_di || IS_ERR(dst_di)) {
+		/* we need a sequence number to insert, so we only
+		 * do inserts for the BTRFS_DIR_INDEX_KEY types
+		 */
+		if (key->type != BTRFS_DIR_INDEX_KEY)
+			goto out;
+		goto insert;
+	}
+
+	btrfs_dir_item_key_to_cpu(path->nodes[0], dst_di, &found_key);
+	/* the existing item matches the logged item */
+	if (found_key.objectid == log_key.objectid &&
+	    found_key.type == log_key.type &&
+	    found_key.offset == log_key.offset &&
+	    btrfs_dir_type(path->nodes[0], dst_di) == log_type) {
+		goto out;
+	}
+
+	/*
+	 * don't drop the conflicting directory entry if the inode
+	 * for the new entry doesn't exist
+	 */
+	if (!exists)
+		goto out;
+
+	ret = drop_one_dir_item(trans, root, path, dir, dst_di);
+	BUG_ON(ret);
+
+	if (key->type == BTRFS_DIR_INDEX_KEY)
+		goto insert;
+out:
+	btrfs_release_path(root, path);
+	kfree(name);
+	iput(dir);
+	return 0;
+
+insert:
+	btrfs_release_path(root, path);
+	ret = insert_one_name(trans, root, path, key->objectid, key->offset,
+			      name, name_len, log_type, &log_key);
+
+	if (ret && ret != -ENOENT)
+		BUG();
+	goto out;
+}
+
+/*
+ * find all the names in a directory item and reconcile them into
+ * the subvolume.  Only BTRFS_DIR_ITEM_KEY types will have more than
+ * one name in a directory item, but the same code gets used for
+ * both directory index types
+ */
+static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct btrfs_path *path,
+					struct extent_buffer *eb, int slot,
+					struct btrfs_key *key)
+{
+	int ret;
+	u32 item_size = btrfs_item_size_nr(eb, slot);
+	struct btrfs_dir_item *di;
+	int name_len;
+	unsigned long ptr;
+	unsigned long ptr_end;
+
+	ptr = btrfs_item_ptr_offset(eb, slot);
+	ptr_end = ptr + item_size;
+	while (ptr < ptr_end) {
+		di = (struct btrfs_dir_item *)ptr;
+		name_len = btrfs_dir_name_len(eb, di);
+		ret = replay_one_name(trans, root, path, eb, di, key);
+		BUG_ON(ret);
+		ptr = (unsigned long)(di + 1);
+		ptr += name_len;
+	}
+	return 0;
+}
+
+/*
+ * directory replay has two parts.  There are the standard directory
+ * items in the log copied from the subvolume, and range items
+ * created in the log while the subvolume was logged.
+ *
+ * The range items tell us which parts of the key space the log
+ * is authoritative for.  During replay, if a key in the subvolume
+ * directory is in a logged range item, but not actually in the log
+ * that means it was deleted from the directory before the fsync
+ * and should be removed.
+ */
+static noinline int find_dir_range(struct btrfs_root *root,
+				   struct btrfs_path *path,
+				   u64 dirid, int key_type,
+				   u64 *start_ret, u64 *end_ret)
+{
+	struct btrfs_key key;
+	u64 found_end;
+	struct btrfs_dir_log_item *item;
+	int ret;
+	int nritems;
+
+	if (*start_ret == (u64)-1)
+		return 1;
+
+	key.objectid = dirid;
+	key.type = key_type;
+	key.offset = *start_ret;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	if (ret > 0) {
+		if (path->slots[0] == 0)
+			goto out;
+		path->slots[0]--;
+	}
+	if (ret != 0)
+		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+
+	if (key.type != key_type || key.objectid != dirid) {
+		ret = 1;
+		goto next;
+	}
+	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			      struct btrfs_dir_log_item);
+	found_end = btrfs_dir_log_end(path->nodes[0], item);
+
+	if (*start_ret >= key.offset && *start_ret <= found_end) {
+		ret = 0;
+		*start_ret = key.offset;
+		*end_ret = found_end;
+		goto out;
+	}
+	ret = 1;
+next:
+	/* check the next slot in the tree to see if it is a valid item */
+	nritems = btrfs_header_nritems(path->nodes[0]);
+	if (path->slots[0] >= nritems) {
+		ret = btrfs_next_leaf(root, path);
+		if (ret)
+			goto out;
+	} else {
+		path->slots[0]++;
+	}
+
+	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+
+	if (key.type != key_type || key.objectid != dirid) {
+		ret = 1;
+		goto out;
+	}
+	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			      struct btrfs_dir_log_item);
+	found_end = btrfs_dir_log_end(path->nodes[0], item);
+	*start_ret = key.offset;
+	*end_ret = found_end;
+	ret = 0;
+out:
+	btrfs_release_path(root, path);
+	return ret;
+}
+
+/*
+ * this looks for a given directory item in the log.  If the directory
+ * item is not in the log, the item is removed and the inode it points
+ * to is unlinked
+ */
+static noinline int check_item_in_log(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_root *log,
+				      struct btrfs_path *path,
+				      struct btrfs_path *log_path,
+				      struct inode *dir,
+				      struct btrfs_key *dir_key)
+{
+	int ret;
+	struct extent_buffer *eb;
+	int slot;
+	u32 item_size;
+	struct btrfs_dir_item *di;
+	struct btrfs_dir_item *log_di;
+	int name_len;
+	unsigned long ptr;
+	unsigned long ptr_end;
+	char *name;
+	struct inode *inode;
+	struct btrfs_key location;
+
+again:
+	eb = path->nodes[0];
+	slot = path->slots[0];
+	item_size = btrfs_item_size_nr(eb, slot);
+	ptr = btrfs_item_ptr_offset(eb, slot);
+	ptr_end = ptr + item_size;
+	while (ptr < ptr_end) {
+		di = (struct btrfs_dir_item *)ptr;
+		name_len = btrfs_dir_name_len(eb, di);
+		name = kmalloc(name_len, GFP_NOFS);
+		if (!name) {
+			ret = -ENOMEM;
+			goto out;
+		}
+		read_extent_buffer(eb, name, (unsigned long)(di + 1),
+				  name_len);
+		log_di = NULL;
+		if (dir_key->type == BTRFS_DIR_ITEM_KEY) {
+			log_di = btrfs_lookup_dir_item(trans, log, log_path,
+						       dir_key->objectid,
+						       name, name_len, 0);
+		} else if (dir_key->type == BTRFS_DIR_INDEX_KEY) {
+			log_di = btrfs_lookup_dir_index_item(trans, log,
+						     log_path,
+						     dir_key->objectid,
+						     dir_key->offset,
+						     name, name_len, 0);
+		}
+		if (!log_di || IS_ERR(log_di)) {
+			btrfs_dir_item_key_to_cpu(eb, di, &location);
+			btrfs_release_path(root, path);
+			btrfs_release_path(log, log_path);
+			inode = read_one_inode(root, location.objectid);
+			BUG_ON(!inode);
+
+			ret = link_to_fixup_dir(trans, root,
+						path, location.objectid);
+			BUG_ON(ret);
+			btrfs_inc_nlink(inode);
+			ret = btrfs_unlink_inode(trans, root, dir, inode,
+						 name, name_len);
+			BUG_ON(ret);
+			kfree(name);
+			iput(inode);
+
+			/* there might still be more names under this key
+			 * check and repeat if required
+			 */
+			ret = btrfs_search_slot(NULL, root, dir_key, path,
+						0, 0);
+			if (ret == 0)
+				goto again;
+			ret = 0;
+			goto out;
+		}
+		btrfs_release_path(log, log_path);
+		kfree(name);
+
+		ptr = (unsigned long)(di + 1);
+		ptr += name_len;
+	}
+	ret = 0;
+out:
+	btrfs_release_path(root, path);
+	btrfs_release_path(log, log_path);
+	return ret;
+}
+
+/*
+ * deletion replay happens before we copy any new directory items
+ * out of the log or out of backreferences from inodes.  It
+ * scans the log to find ranges of keys that log is authoritative for,
+ * and then scans the directory to find items in those ranges that are
+ * not present in the log.
+ *
+ * Anything we don't find in the log is unlinked and removed from the
+ * directory.
+ */
+static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
+				       struct btrfs_root *root,
+				       struct btrfs_root *log,
+				       struct btrfs_path *path,
+				       u64 dirid)
+{
+	u64 range_start;
+	u64 range_end;
+	int key_type = BTRFS_DIR_LOG_ITEM_KEY;
+	int ret = 0;
+	struct btrfs_key dir_key;
+	struct btrfs_key found_key;
+	struct btrfs_path *log_path;
+	struct inode *dir;
+
+	dir_key.objectid = dirid;
+	dir_key.type = BTRFS_DIR_ITEM_KEY;
+	log_path = btrfs_alloc_path();
+	if (!log_path)
+		return -ENOMEM;
+
+	dir = read_one_inode(root, dirid);
+	/* it isn't an error if the inode isn't there, that can happen
+	 * because we replay the deletes before we copy in the inode item
+	 * from the log
+	 */
+	if (!dir) {
+		btrfs_free_path(log_path);
+		return 0;
+	}
+again:
+	range_start = 0;
+	range_end = 0;
+	while (1) {
+		ret = find_dir_range(log, path, dirid, key_type,
+				     &range_start, &range_end);
+		if (ret != 0)
+			break;
+
+		dir_key.offset = range_start;
+		while (1) {
+			int nritems;
+			ret = btrfs_search_slot(NULL, root, &dir_key, path,
+						0, 0);
+			if (ret < 0)
+				goto out;
+
+			nritems = btrfs_header_nritems(path->nodes[0]);
+			if (path->slots[0] >= nritems) {
+				ret = btrfs_next_leaf(root, path);
+				if (ret)
+					break;
+			}
+			btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+					      path->slots[0]);
+			if (found_key.objectid != dirid ||
+			    found_key.type != dir_key.type)
+				goto next_type;
+
+			if (found_key.offset > range_end)
+				break;
+
+			ret = check_item_in_log(trans, root, log, path,
+						log_path, dir, &found_key);
+			BUG_ON(ret);
+			if (found_key.offset == (u64)-1)
+				break;
+			dir_key.offset = found_key.offset + 1;
+		}
+		btrfs_release_path(root, path);
+		if (range_end == (u64)-1)
+			break;
+		range_start = range_end + 1;
+	}
+
+next_type:
+	ret = 0;
+	if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
+		key_type = BTRFS_DIR_LOG_INDEX_KEY;
+		dir_key.type = BTRFS_DIR_INDEX_KEY;
+		btrfs_release_path(root, path);
+		goto again;
+	}
+out:
+	btrfs_release_path(root, path);
+	btrfs_free_path(log_path);
+	iput(dir);
+	return ret;
+}
+
+/*
+ * the process_func used to replay items from the log tree.  This
+ * gets called in two different stages.  The first stage just looks
+ * for inodes and makes sure they are all copied into the subvolume.
+ *
+ * The second stage copies all the other item types from the log into
+ * the subvolume.  The two stage approach is slower, but gets rid of
+ * lots of complexity around inodes referencing other inodes that exist
+ * only in the log (references come from either directory items or inode
+ * back refs).
+ */
+static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
+			     struct walk_control *wc, u64 gen)
+{
+	int nritems;
+	struct btrfs_path *path;
+	struct btrfs_root *root = wc->replay_dest;
+	struct btrfs_key key;
+	u32 item_size;
+	int level;
+	int i;
+	int ret;
+
+	btrfs_read_buffer(eb, gen);
+
+	level = btrfs_header_level(eb);
+
+	if (level != 0)
+		return 0;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	nritems = btrfs_header_nritems(eb);
+	for (i = 0; i < nritems; i++) {
+		btrfs_item_key_to_cpu(eb, &key, i);
+		item_size = btrfs_item_size_nr(eb, i);
+
+		/* inode keys are done during the first stage */
+		if (key.type == BTRFS_INODE_ITEM_KEY &&
+		    wc->stage == LOG_WALK_REPLAY_INODES) {
+			struct inode *inode;
+			struct btrfs_inode_item *inode_item;
+			u32 mode;
+
+			inode_item = btrfs_item_ptr(eb, i,
+					    struct btrfs_inode_item);
+			mode = btrfs_inode_mode(eb, inode_item);
+			if (S_ISDIR(mode)) {
+				ret = replay_dir_deletes(wc->trans,
+					 root, log, path, key.objectid);
+				BUG_ON(ret);
+			}
+			ret = overwrite_item(wc->trans, root, path,
+					     eb, i, &key);
+			BUG_ON(ret);
+
+			/* for regular files, truncate away
+			 * extents past the new EOF
+			 */
+			if (S_ISREG(mode)) {
+				inode = read_one_inode(root,
+						       key.objectid);
+				BUG_ON(!inode);
+
+				ret = btrfs_truncate_inode_items(wc->trans,
+					root, inode, inode->i_size,
+					BTRFS_EXTENT_DATA_KEY);
+				BUG_ON(ret);
+				iput(inode);
+			}
+			ret = link_to_fixup_dir(wc->trans, root,
+						path, key.objectid);
+			BUG_ON(ret);
+		}
+		if (wc->stage < LOG_WALK_REPLAY_ALL)
+			continue;
+
+		/* these keys are simply copied */
+		if (key.type == BTRFS_XATTR_ITEM_KEY) {
+			ret = overwrite_item(wc->trans, root, path,
+					     eb, i, &key);
+			BUG_ON(ret);
+		} else if (key.type == BTRFS_INODE_REF_KEY) {
+			ret = add_inode_ref(wc->trans, root, log, path,
+					    eb, i, &key);
+			BUG_ON(ret && ret != -ENOENT);
+		} else if (key.type == BTRFS_EXTENT_DATA_KEY) {
+			ret = replay_one_extent(wc->trans, root, path,
+						eb, i, &key);
+			BUG_ON(ret);
+		} else if (key.type == BTRFS_DIR_ITEM_KEY ||
+			   key.type == BTRFS_DIR_INDEX_KEY) {
+			ret = replay_one_dir_item(wc->trans, root, path,
+						  eb, i, &key);
+			BUG_ON(ret);
+		}
+	}
+	btrfs_free_path(path);
+	return 0;
+}
+
+static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root,
+				   struct btrfs_path *path, int *level,
+				   struct walk_control *wc)
+{
+	u64 root_owner;
+	u64 root_gen;
+	u64 bytenr;
+	u64 ptr_gen;
+	struct extent_buffer *next;
+	struct extent_buffer *cur;
+	struct extent_buffer *parent;
+	u32 blocksize;
+	int ret = 0;
+
+	WARN_ON(*level < 0);
+	WARN_ON(*level >= BTRFS_MAX_LEVEL);
+
+	while (*level > 0) {
+		WARN_ON(*level < 0);
+		WARN_ON(*level >= BTRFS_MAX_LEVEL);
+		cur = path->nodes[*level];
+
+		if (btrfs_header_level(cur) != *level)
+			WARN_ON(1);
+
+		if (path->slots[*level] >=
+		    btrfs_header_nritems(cur))
+			break;
+
+		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
+		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+		blocksize = btrfs_level_size(root, *level - 1);
+
+		parent = path->nodes[*level];
+		root_owner = btrfs_header_owner(parent);
+		root_gen = btrfs_header_generation(parent);
+
+		next = btrfs_find_create_tree_block(root, bytenr, blocksize);
+
+		wc->process_func(root, next, wc, ptr_gen);
+
+		if (*level == 1) {
+			path->slots[*level]++;
+			if (wc->free) {
+				btrfs_read_buffer(next, ptr_gen);
+
+				btrfs_tree_lock(next);
+				clean_tree_block(trans, root, next);
+				btrfs_wait_tree_block_writeback(next);
+				btrfs_tree_unlock(next);
+
+				ret = btrfs_drop_leaf_ref(trans, root, next);
+				BUG_ON(ret);
+
+				WARN_ON(root_owner !=
+					BTRFS_TREE_LOG_OBJECTID);
+				ret = btrfs_free_reserved_extent(root,
+							 bytenr, blocksize);
+				BUG_ON(ret);
+			}
+			free_extent_buffer(next);
+			continue;
+		}
+		btrfs_read_buffer(next, ptr_gen);
+
+		WARN_ON(*level <= 0);
+		if (path->nodes[*level-1])
+			free_extent_buffer(path->nodes[*level-1]);
+		path->nodes[*level-1] = next;
+		*level = btrfs_header_level(next);
+		path->slots[*level] = 0;
+		cond_resched();
+	}
+	WARN_ON(*level < 0);
+	WARN_ON(*level >= BTRFS_MAX_LEVEL);
+
+	if (path->nodes[*level] == root->node)
+		parent = path->nodes[*level];
+	else
+		parent = path->nodes[*level + 1];
+
+	bytenr = path->nodes[*level]->start;
+
+	blocksize = btrfs_level_size(root, *level);
+	root_owner = btrfs_header_owner(parent);
+	root_gen = btrfs_header_generation(parent);
+
+	wc->process_func(root, path->nodes[*level], wc,
+			 btrfs_header_generation(path->nodes[*level]));
+
+	if (wc->free) {
+		next = path->nodes[*level];
+		btrfs_tree_lock(next);
+		clean_tree_block(trans, root, next);
+		btrfs_wait_tree_block_writeback(next);
+		btrfs_tree_unlock(next);
+
+		if (*level == 0) {
+			ret = btrfs_drop_leaf_ref(trans, root, next);
+			BUG_ON(ret);
+		}
+		WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
+		ret = btrfs_free_reserved_extent(root, bytenr, blocksize);
+		BUG_ON(ret);
+	}
+	free_extent_buffer(path->nodes[*level]);
+	path->nodes[*level] = NULL;
+	*level += 1;
+
+	cond_resched();
+	return 0;
+}
+
+static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 struct btrfs_path *path, int *level,
+				 struct walk_control *wc)
+{
+	u64 root_owner;
+	u64 root_gen;
+	int i;
+	int slot;
+	int ret;
+
+	for (i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
+		slot = path->slots[i];
+		if (slot < btrfs_header_nritems(path->nodes[i]) - 1) {
+			struct extent_buffer *node;
+			node = path->nodes[i];
+			path->slots[i]++;
+			*level = i;
+			WARN_ON(*level == 0);
+			return 0;
+		} else {
+			struct extent_buffer *parent;
+			if (path->nodes[*level] == root->node)
+				parent = path->nodes[*level];
+			else
+				parent = path->nodes[*level + 1];
+
+			root_owner = btrfs_header_owner(parent);
+			root_gen = btrfs_header_generation(parent);
+			wc->process_func(root, path->nodes[*level], wc,
+				 btrfs_header_generation(path->nodes[*level]));
+			if (wc->free) {
+				struct extent_buffer *next;
+
+				next = path->nodes[*level];
+
+				btrfs_tree_lock(next);
+				clean_tree_block(trans, root, next);
+				btrfs_wait_tree_block_writeback(next);
+				btrfs_tree_unlock(next);
+
+				if (*level == 0) {
+					ret = btrfs_drop_leaf_ref(trans, root,
+								  next);
+					BUG_ON(ret);
+				}
+
+				WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
+				ret = btrfs_free_reserved_extent(root,
+						path->nodes[*level]->start,
+						path->nodes[*level]->len);
+				BUG_ON(ret);
+			}
+			free_extent_buffer(path->nodes[*level]);
+			path->nodes[*level] = NULL;
+			*level = i + 1;
+		}
+	}
+	return 1;
+}
+
+/*
+ * drop the reference count on the tree rooted at 'snap'.  This traverses
+ * the tree freeing any blocks that have a ref count of zero after being
+ * decremented.
+ */
+static int walk_log_tree(struct btrfs_trans_handle *trans,
+			 struct btrfs_root *log, struct walk_control *wc)
+{
+	int ret = 0;
+	int wret;
+	int level;
+	struct btrfs_path *path;
+	int i;
+	int orig_level;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	level = btrfs_header_level(log->node);
+	orig_level = level;
+	path->nodes[level] = log->node;
+	extent_buffer_get(log->node);
+	path->slots[level] = 0;
+
+	while (1) {
+		wret = walk_down_log_tree(trans, log, path, &level, wc);
+		if (wret > 0)
+			break;
+		if (wret < 0)
+			ret = wret;
+
+		wret = walk_up_log_tree(trans, log, path, &level, wc);
+		if (wret > 0)
+			break;
+		if (wret < 0)
+			ret = wret;
+	}
+
+	/* was the root node processed? if not, catch it here */
+	if (path->nodes[orig_level]) {
+		wc->process_func(log, path->nodes[orig_level], wc,
+			 btrfs_header_generation(path->nodes[orig_level]));
+		if (wc->free) {
+			struct extent_buffer *next;
+
+			next = path->nodes[orig_level];
+
+			btrfs_tree_lock(next);
+			clean_tree_block(trans, log, next);
+			btrfs_wait_tree_block_writeback(next);
+			btrfs_tree_unlock(next);
+
+			if (orig_level == 0) {
+				ret = btrfs_drop_leaf_ref(trans, log,
+							  next);
+				BUG_ON(ret);
+			}
+			WARN_ON(log->root_key.objectid !=
+				BTRFS_TREE_LOG_OBJECTID);
+			ret = btrfs_free_reserved_extent(log, next->start,
+							 next->len);
+			BUG_ON(ret);
+		}
+	}
+
+	for (i = 0; i <= orig_level; i++) {
+		if (path->nodes[i]) {
+			free_extent_buffer(path->nodes[i]);
+			path->nodes[i] = NULL;
+		}
+	}
+	btrfs_free_path(path);
+	if (wc->free)
+		free_extent_buffer(log->node);
+	return ret;
+}
+
+static int wait_log_commit(struct btrfs_root *log)
+{
+	DEFINE_WAIT(wait);
+	u64 transid = log->fs_info->tree_log_transid;
+
+	do {
+		prepare_to_wait(&log->fs_info->tree_log_wait, &wait,
+				TASK_UNINTERRUPTIBLE);
+		mutex_unlock(&log->fs_info->tree_log_mutex);
+		if (atomic_read(&log->fs_info->tree_log_commit))
+			schedule();
+		finish_wait(&log->fs_info->tree_log_wait, &wait);
+		mutex_lock(&log->fs_info->tree_log_mutex);
+	} while (transid == log->fs_info->tree_log_transid &&
+		atomic_read(&log->fs_info->tree_log_commit));
+	return 0;
+}
+
+/*
+ * btrfs_sync_log does sends a given tree log down to the disk and
+ * updates the super blocks to record it.  When this call is done,
+ * you know that any inodes previously logged are safely on disk
+ */
+int btrfs_sync_log(struct btrfs_trans_handle *trans,
+		   struct btrfs_root *root)
+{
+	int ret;
+	unsigned long batch;
+	struct btrfs_root *log = root->log_root;
+
+	mutex_lock(&log->fs_info->tree_log_mutex);
+	if (atomic_read(&log->fs_info->tree_log_commit)) {
+		wait_log_commit(log);
+		goto out;
+	}
+	atomic_set(&log->fs_info->tree_log_commit, 1);
+
+	while (1) {
+		batch = log->fs_info->tree_log_batch;
+		mutex_unlock(&log->fs_info->tree_log_mutex);
+		schedule_timeout_uninterruptible(1);
+		mutex_lock(&log->fs_info->tree_log_mutex);
+
+		while (atomic_read(&log->fs_info->tree_log_writers)) {
+			DEFINE_WAIT(wait);
+			prepare_to_wait(&log->fs_info->tree_log_wait, &wait,
+					TASK_UNINTERRUPTIBLE);
+			mutex_unlock(&log->fs_info->tree_log_mutex);
+			if (atomic_read(&log->fs_info->tree_log_writers))
+				schedule();
+			mutex_lock(&log->fs_info->tree_log_mutex);
+			finish_wait(&log->fs_info->tree_log_wait, &wait);
+		}
+		if (batch == log->fs_info->tree_log_batch)
+			break;
+	}
+
+	ret = btrfs_write_and_wait_marked_extents(log, &log->dirty_log_pages);
+	BUG_ON(ret);
+	ret = btrfs_write_and_wait_marked_extents(root->fs_info->log_root_tree,
+			       &root->fs_info->log_root_tree->dirty_log_pages);
+	BUG_ON(ret);
+
+	btrfs_set_super_log_root(&root->fs_info->super_for_commit,
+				 log->fs_info->log_root_tree->node->start);
+	btrfs_set_super_log_root_level(&root->fs_info->super_for_commit,
+		       btrfs_header_level(log->fs_info->log_root_tree->node));
+
+	write_ctree_super(trans, log->fs_info->tree_root, 2);
+	log->fs_info->tree_log_transid++;
+	log->fs_info->tree_log_batch = 0;
+	atomic_set(&log->fs_info->tree_log_commit, 0);
+	smp_mb();
+	if (waitqueue_active(&log->fs_info->tree_log_wait))
+		wake_up(&log->fs_info->tree_log_wait);
+out:
+	mutex_unlock(&log->fs_info->tree_log_mutex);
+	return 0;
+}
+
+/* * free all the extents used by the tree log.  This should be called
+ * at commit time of the full transaction
+ */
+int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
+{
+	int ret;
+	struct btrfs_root *log;
+	struct key;
+	u64 start;
+	u64 end;
+	struct walk_control wc = {
+		.free = 1,
+		.process_func = process_one_buffer
+	};
+
+	if (!root->log_root || root->fs_info->log_root_recovering)
+		return 0;
+
+	log = root->log_root;
+	ret = walk_log_tree(trans, log, &wc);
+	BUG_ON(ret);
+
+	while (1) {
+		ret = find_first_extent_bit(&log->dirty_log_pages,
+				    0, &start, &end, EXTENT_DIRTY);
+		if (ret)
+			break;
+
+		clear_extent_dirty(&log->dirty_log_pages,
+				   start, end, GFP_NOFS);
+	}
+
+	log = root->log_root;
+	ret = btrfs_del_root(trans, root->fs_info->log_root_tree,
+			     &log->root_key);
+	BUG_ON(ret);
+	root->log_root = NULL;
+	kfree(root->log_root);
+	return 0;
+}
+
+/*
+ * helper function to update the item for a given subvolumes log root
+ * in the tree of log roots
+ */
+static int update_log_root(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *log)
+{
+	u64 bytenr = btrfs_root_bytenr(&log->root_item);
+	int ret;
+
+	if (log->node->start == bytenr)
+		return 0;
+
+	btrfs_set_root_bytenr(&log->root_item, log->node->start);
+	btrfs_set_root_generation(&log->root_item, trans->transid);
+	btrfs_set_root_level(&log->root_item, btrfs_header_level(log->node));
+	ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
+				&log->root_key, &log->root_item);
+	BUG_ON(ret);
+	return ret;
+}
+
+/*
+ * If both a file and directory are logged, and unlinks or renames are
+ * mixed in, we have a few interesting corners:
+ *
+ * create file X in dir Y
+ * link file X to X.link in dir Y
+ * fsync file X
+ * unlink file X but leave X.link
+ * fsync dir Y
+ *
+ * After a crash we would expect only X.link to exist.  But file X
+ * didn't get fsync'd again so the log has back refs for X and X.link.
+ *
+ * We solve this by removing directory entries and inode backrefs from the
+ * log when a file that was logged in the current transaction is
+ * unlinked.  Any later fsync will include the updated log entries, and
+ * we'll be able to reconstruct the proper directory items from backrefs.
+ *
+ * This optimizations allows us to avoid relogging the entire inode
+ * or the entire directory.
+ */
+int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 const char *name, int name_len,
+				 struct inode *dir, u64 index)
+{
+	struct btrfs_root *log;
+	struct btrfs_dir_item *di;
+	struct btrfs_path *path;
+	int ret;
+	int bytes_del = 0;
+
+	if (BTRFS_I(dir)->logged_trans < trans->transid)
+		return 0;
+
+	ret = join_running_log_trans(root);
+	if (ret)
+		return 0;
+
+	mutex_lock(&BTRFS_I(dir)->log_mutex);
+
+	log = root->log_root;
+	path = btrfs_alloc_path();
+	di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino,
+				   name, name_len, -1);
+	if (di && !IS_ERR(di)) {
+		ret = btrfs_delete_one_dir_name(trans, log, path, di);
+		bytes_del += name_len;
+		BUG_ON(ret);
+	}
+	btrfs_release_path(log, path);
+	di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino,
+					 index, name, name_len, -1);
+	if (di && !IS_ERR(di)) {
+		ret = btrfs_delete_one_dir_name(trans, log, path, di);
+		bytes_del += name_len;
+		BUG_ON(ret);
+	}
+
+	/* update the directory size in the log to reflect the names
+	 * we have removed
+	 */
+	if (bytes_del) {
+		struct btrfs_key key;
+
+		key.objectid = dir->i_ino;
+		key.offset = 0;
+		key.type = BTRFS_INODE_ITEM_KEY;
+		btrfs_release_path(log, path);
+
+		ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
+		if (ret == 0) {
+			struct btrfs_inode_item *item;
+			u64 i_size;
+
+			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+					      struct btrfs_inode_item);
+			i_size = btrfs_inode_size(path->nodes[0], item);
+			if (i_size > bytes_del)
+				i_size -= bytes_del;
+			else
+				i_size = 0;
+			btrfs_set_inode_size(path->nodes[0], item, i_size);
+			btrfs_mark_buffer_dirty(path->nodes[0]);
+		} else
+			ret = 0;
+		btrfs_release_path(log, path);
+	}
+
+	btrfs_free_path(path);
+	mutex_unlock(&BTRFS_I(dir)->log_mutex);
+	end_log_trans(root);
+
+	return 0;
+}
+
+/* see comments for btrfs_del_dir_entries_in_log */
+int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       const char *name, int name_len,
+			       struct inode *inode, u64 dirid)
+{
+	struct btrfs_root *log;
+	u64 index;
+	int ret;
+
+	if (BTRFS_I(inode)->logged_trans < trans->transid)
+		return 0;
+
+	ret = join_running_log_trans(root);
+	if (ret)
+		return 0;
+	log = root->log_root;
+	mutex_lock(&BTRFS_I(inode)->log_mutex);
+
+	ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino,
+				  dirid, &index);
+	mutex_unlock(&BTRFS_I(inode)->log_mutex);
+	end_log_trans(root);
+
+	return ret;
+}
+
+/*
+ * creates a range item in the log for 'dirid'.  first_offset and
+ * last_offset tell us which parts of the key space the log should
+ * be considered authoritative for.
+ */
+static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
+				       struct btrfs_root *log,
+				       struct btrfs_path *path,
+				       int key_type, u64 dirid,
+				       u64 first_offset, u64 last_offset)
+{
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_dir_log_item *item;
+
+	key.objectid = dirid;
+	key.offset = first_offset;
+	if (key_type == BTRFS_DIR_ITEM_KEY)
+		key.type = BTRFS_DIR_LOG_ITEM_KEY;
+	else
+		key.type = BTRFS_DIR_LOG_INDEX_KEY;
+	ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
+	BUG_ON(ret);
+
+	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+			      struct btrfs_dir_log_item);
+	btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+	btrfs_release_path(log, path);
+	return 0;
+}
+
+/*
+ * log all the items included in the current transaction for a given
+ * directory.  This also creates the range items in the log tree required
+ * to replay anything deleted before the fsync
+ */
+static noinline int log_dir_items(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct inode *inode,
+			  struct btrfs_path *path,
+			  struct btrfs_path *dst_path, int key_type,
+			  u64 min_offset, u64 *last_offset_ret)
+{
+	struct btrfs_key min_key;
+	struct btrfs_key max_key;
+	struct btrfs_root *log = root->log_root;
+	struct extent_buffer *src;
+	int ret;
+	int i;
+	int nritems;
+	u64 first_offset = min_offset;
+	u64 last_offset = (u64)-1;
+
+	log = root->log_root;
+	max_key.objectid = inode->i_ino;
+	max_key.offset = (u64)-1;
+	max_key.type = key_type;
+
+	min_key.objectid = inode->i_ino;
+	min_key.type = key_type;
+	min_key.offset = min_offset;
+
+	path->keep_locks = 1;
+
+	ret = btrfs_search_forward(root, &min_key, &max_key,
+				   path, 0, trans->transid);
+
+	/*
+	 * we didn't find anything from this transaction, see if there
+	 * is anything at all
+	 */
+	if (ret != 0 || min_key.objectid != inode->i_ino ||
+	    min_key.type != key_type) {
+		min_key.objectid = inode->i_ino;
+		min_key.type = key_type;
+		min_key.offset = (u64)-1;
+		btrfs_release_path(root, path);
+		ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
+		if (ret < 0) {
+			btrfs_release_path(root, path);
+			return ret;
+		}
+		ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+
+		/* if ret == 0 there are items for this type,
+		 * create a range to tell us the last key of this type.
+		 * otherwise, there are no items in this directory after
+		 * *min_offset, and we create a range to indicate that.
+		 */
+		if (ret == 0) {
+			struct btrfs_key tmp;
+			btrfs_item_key_to_cpu(path->nodes[0], &tmp,
+					      path->slots[0]);
+			if (key_type == tmp.type)
+				first_offset = max(min_offset, tmp.offset) + 1;
+		}
+		goto done;
+	}
+
+	/* go backward to find any previous key */
+	ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+	if (ret == 0) {
+		struct btrfs_key tmp;
+		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
+		if (key_type == tmp.type) {
+			first_offset = tmp.offset;
+			ret = overwrite_item(trans, log, dst_path,
+					     path->nodes[0], path->slots[0],
+					     &tmp);
+		}
+	}
+	btrfs_release_path(root, path);
+
+	/* find the first key from this transaction again */
+	ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
+	if (ret != 0) {
+		WARN_ON(1);
+		goto done;
+	}
+
+	/*
+	 * we have a block from this transaction, log every item in it
+	 * from our directory
+	 */
+	while (1) {
+		struct btrfs_key tmp;
+		src = path->nodes[0];
+		nritems = btrfs_header_nritems(src);
+		for (i = path->slots[0]; i < nritems; i++) {
+			btrfs_item_key_to_cpu(src, &min_key, i);
+
+			if (min_key.objectid != inode->i_ino ||
+			    min_key.type != key_type)
+				goto done;
+			ret = overwrite_item(trans, log, dst_path, src, i,
+					     &min_key);
+			BUG_ON(ret);
+		}
+		path->slots[0] = nritems;
+
+		/*
+		 * look ahead to the next item and see if it is also
+		 * from this directory and from this transaction
+		 */
+		ret = btrfs_next_leaf(root, path);
+		if (ret == 1) {
+			last_offset = (u64)-1;
+			goto done;
+		}
+		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
+		if (tmp.objectid != inode->i_ino || tmp.type != key_type) {
+			last_offset = (u64)-1;
+			goto done;
+		}
+		if (btrfs_header_generation(path->nodes[0]) != trans->transid) {
+			ret = overwrite_item(trans, log, dst_path,
+					     path->nodes[0], path->slots[0],
+					     &tmp);
+
+			BUG_ON(ret);
+			last_offset = tmp.offset;
+			goto done;
+		}
+	}
+done:
+	*last_offset_ret = last_offset;
+	btrfs_release_path(root, path);
+	btrfs_release_path(log, dst_path);
+
+	/* insert the log range keys to indicate where the log is valid */
+	ret = insert_dir_log_key(trans, log, path, key_type, inode->i_ino,
+				 first_offset, last_offset);
+	BUG_ON(ret);
+	return 0;
+}
+
+/*
+ * logging directories is very similar to logging inodes, We find all the items
+ * from the current transaction and write them to the log.
+ *
+ * The recovery code scans the directory in the subvolume, and if it finds a
+ * key in the range logged that is not present in the log tree, then it means
+ * that dir entry was unlinked during the transaction.
+ *
+ * In order for that scan to work, we must include one key smaller than
+ * the smallest logged by this transaction and one key larger than the largest
+ * key logged by this transaction.
+ */
+static noinline int log_directory_changes(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct inode *inode,
+			  struct btrfs_path *path,
+			  struct btrfs_path *dst_path)
+{
+	u64 min_key;
+	u64 max_key;
+	int ret;
+	int key_type = BTRFS_DIR_ITEM_KEY;
+
+again:
+	min_key = 0;
+	max_key = 0;
+	while (1) {
+		ret = log_dir_items(trans, root, inode, path,
+				    dst_path, key_type, min_key,
+				    &max_key);
+		BUG_ON(ret);
+		if (max_key == (u64)-1)
+			break;
+		min_key = max_key + 1;
+	}
+
+	if (key_type == BTRFS_DIR_ITEM_KEY) {
+		key_type = BTRFS_DIR_INDEX_KEY;
+		goto again;
+	}
+	return 0;
+}
+
+/*
+ * a helper function to drop items from the log before we relog an
+ * inode.  max_key_type indicates the highest item type to remove.
+ * This cannot be run for file data extents because it does not
+ * free the extents they point to.
+ */
+static int drop_objectid_items(struct btrfs_trans_handle *trans,
+				  struct btrfs_root *log,
+				  struct btrfs_path *path,
+				  u64 objectid, int max_key_type)
+{
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+
+	key.objectid = objectid;
+	key.type = max_key_type;
+	key.offset = (u64)-1;
+
+	while (1) {
+		ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
+
+		if (ret != 1)
+			break;
+
+		if (path->slots[0] == 0)
+			break;
+
+		path->slots[0]--;
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+				      path->slots[0]);
+
+		if (found_key.objectid != objectid)
+			break;
+
+		ret = btrfs_del_item(trans, log, path);
+		BUG_ON(ret);
+		btrfs_release_path(log, path);
+	}
+	btrfs_release_path(log, path);
+	return 0;
+}
+
+static noinline int copy_items(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *log,
+			       struct btrfs_path *dst_path,
+			       struct extent_buffer *src,
+			       int start_slot, int nr, int inode_only)
+{
+	unsigned long src_offset;
+	unsigned long dst_offset;
+	struct btrfs_file_extent_item *extent;
+	struct btrfs_inode_item *inode_item;
+	int ret;
+	struct btrfs_key *ins_keys;
+	u32 *ins_sizes;
+	char *ins_data;
+	int i;
+	struct list_head ordered_sums;
+
+	INIT_LIST_HEAD(&ordered_sums);
+
+	ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
+			   nr * sizeof(u32), GFP_NOFS);
+	ins_sizes = (u32 *)ins_data;
+	ins_keys = (struct btrfs_key *)(ins_data + nr * sizeof(u32));
+
+	for (i = 0; i < nr; i++) {
+		ins_sizes[i] = btrfs_item_size_nr(src, i + start_slot);
+		btrfs_item_key_to_cpu(src, ins_keys + i, i + start_slot);
+	}
+	ret = btrfs_insert_empty_items(trans, log, dst_path,
+				       ins_keys, ins_sizes, nr);
+	BUG_ON(ret);
+
+	for (i = 0; i < nr; i++) {
+		dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
+						   dst_path->slots[0]);
+
+		src_offset = btrfs_item_ptr_offset(src, start_slot + i);
+
+		copy_extent_buffer(dst_path->nodes[0], src, dst_offset,
+				   src_offset, ins_sizes[i]);
+
+		if (inode_only == LOG_INODE_EXISTS &&
+		    ins_keys[i].type == BTRFS_INODE_ITEM_KEY) {
+			inode_item = btrfs_item_ptr(dst_path->nodes[0],
+						    dst_path->slots[0],
+						    struct btrfs_inode_item);
+			btrfs_set_inode_size(dst_path->nodes[0], inode_item, 0);
+
+			/* set the generation to zero so the recover code
+			 * can tell the difference between an logging
+			 * just to say 'this inode exists' and a logging
+			 * to say 'update this inode with these values'
+			 */
+			btrfs_set_inode_generation(dst_path->nodes[0],
+						   inode_item, 0);
+		}
+		/* take a reference on file data extents so that truncates
+		 * or deletes of this inode don't have to relog the inode
+		 * again
+		 */
+		if (btrfs_key_type(ins_keys + i) == BTRFS_EXTENT_DATA_KEY) {
+			int found_type;
+			extent = btrfs_item_ptr(src, start_slot + i,
+						struct btrfs_file_extent_item);
+
+			found_type = btrfs_file_extent_type(src, extent);
+			if (found_type == BTRFS_FILE_EXTENT_REG ||
+			    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
+				u64 ds = btrfs_file_extent_disk_bytenr(src,
+								   extent);
+				u64 dl = btrfs_file_extent_disk_num_bytes(src,
+								      extent);
+				u64 cs = btrfs_file_extent_offset(src, extent);
+				u64 cl = btrfs_file_extent_num_bytes(src,
+								     extent);;
+				if (btrfs_file_extent_compression(src,
+								  extent)) {
+					cs = 0;
+					cl = dl;
+				}
+				/* ds == 0 is a hole */
+				if (ds != 0) {
+					ret = btrfs_inc_extent_ref(trans, log,
+						   ds, dl,
+						   dst_path->nodes[0]->start,
+						   BTRFS_TREE_LOG_OBJECTID,
+						   trans->transid,
+						   ins_keys[i].objectid);
+					BUG_ON(ret);
+					ret = btrfs_lookup_csums_range(
+						   log->fs_info->csum_root,
+						   ds + cs, ds + cs + cl - 1,
+						   &ordered_sums);
+					BUG_ON(ret);
+				}
+			}
+		}
+		dst_path->slots[0]++;
+	}
+
+	btrfs_mark_buffer_dirty(dst_path->nodes[0]);
+	btrfs_release_path(log, dst_path);
+	kfree(ins_data);
+
+	/*
+	 * we have to do this after the loop above to avoid changing the
+	 * log tree while trying to change the log tree.
+	 */
+	while (!list_empty(&ordered_sums)) {
+		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
+						   struct btrfs_ordered_sum,
+						   list);
+		ret = btrfs_csum_file_blocks(trans, log, sums);
+		BUG_ON(ret);
+		list_del(&sums->list);
+		kfree(sums);
+	}
+	return 0;
+}
+
+/* log a single inode in the tree log.
+ * At least one parent directory for this inode must exist in the tree
+ * or be logged already.
+ *
+ * Any items from this inode changed by the current transaction are copied
+ * to the log tree.  An extra reference is taken on any extents in this
+ * file, allowing us to avoid a whole pile of corner cases around logging
+ * blocks that have been removed from the tree.
+ *
+ * See LOG_INODE_ALL and related defines for a description of what inode_only
+ * does.
+ *
+ * This handles both files and directories.
+ */
+static int __btrfs_log_inode(struct btrfs_trans_handle *trans,
+			     struct btrfs_root *root, struct inode *inode,
+			     int inode_only)
+{
+	struct btrfs_path *path;
+	struct btrfs_path *dst_path;
+	struct btrfs_key min_key;
+	struct btrfs_key max_key;
+	struct btrfs_root *log = root->log_root;
+	struct extent_buffer *src = NULL;
+	u32 size;
+	int ret;
+	int nritems;
+	int ins_start_slot = 0;
+	int ins_nr;
+
+	log = root->log_root;
+
+	path = btrfs_alloc_path();
+	dst_path = btrfs_alloc_path();
+
+	min_key.objectid = inode->i_ino;
+	min_key.type = BTRFS_INODE_ITEM_KEY;
+	min_key.offset = 0;
+
+	max_key.objectid = inode->i_ino;
+	if (inode_only == LOG_INODE_EXISTS || S_ISDIR(inode->i_mode))
+		max_key.type = BTRFS_XATTR_ITEM_KEY;
+	else
+		max_key.type = (u8)-1;
+	max_key.offset = (u64)-1;
+
+	/*
+	 * if this inode has already been logged and we're in inode_only
+	 * mode, we don't want to delete the things that have already
+	 * been written to the log.
+	 *
+	 * But, if the inode has been through an inode_only log,
+	 * the logged_trans field is not set.  This allows us to catch
+	 * any new names for this inode in the backrefs by logging it
+	 * again
+	 */
+	if (inode_only == LOG_INODE_EXISTS &&
+	    BTRFS_I(inode)->logged_trans == trans->transid) {
+		btrfs_free_path(path);
+		btrfs_free_path(dst_path);
+		goto out;
+	}
+	mutex_lock(&BTRFS_I(inode)->log_mutex);
+
+	/*
+	 * a brute force approach to making sure we get the most uptodate
+	 * copies of everything.
+	 */
+	if (S_ISDIR(inode->i_mode)) {
+		int max_key_type = BTRFS_DIR_LOG_INDEX_KEY;
+
+		if (inode_only == LOG_INODE_EXISTS)
+			max_key_type = BTRFS_XATTR_ITEM_KEY;
+		ret = drop_objectid_items(trans, log, path,
+					  inode->i_ino, max_key_type);
+	} else {
+		ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0);
+	}
+	BUG_ON(ret);
+	path->keep_locks = 1;
+
+	while (1) {
+		ins_nr = 0;
+		ret = btrfs_search_forward(root, &min_key, &max_key,
+					   path, 0, trans->transid);
+		if (ret != 0)
+			break;
+again:
+		/* note, ins_nr might be > 0 here, cleanup outside the loop */
+		if (min_key.objectid != inode->i_ino)
+			break;
+		if (min_key.type > max_key.type)
+			break;
+
+		src = path->nodes[0];
+		size = btrfs_item_size_nr(src, path->slots[0]);
+		if (ins_nr && ins_start_slot + ins_nr == path->slots[0]) {
+			ins_nr++;
+			goto next_slot;
+		} else if (!ins_nr) {
+			ins_start_slot = path->slots[0];
+			ins_nr = 1;
+			goto next_slot;
+		}
+
+		ret = copy_items(trans, log, dst_path, src, ins_start_slot,
+				 ins_nr, inode_only);
+		BUG_ON(ret);
+		ins_nr = 1;
+		ins_start_slot = path->slots[0];
+next_slot:
+
+		nritems = btrfs_header_nritems(path->nodes[0]);
+		path->slots[0]++;
+		if (path->slots[0] < nritems) {
+			btrfs_item_key_to_cpu(path->nodes[0], &min_key,
+					      path->slots[0]);
+			goto again;
+		}
+		if (ins_nr) {
+			ret = copy_items(trans, log, dst_path, src,
+					 ins_start_slot,
+					 ins_nr, inode_only);
+			BUG_ON(ret);
+			ins_nr = 0;
+		}
+		btrfs_release_path(root, path);
+
+		if (min_key.offset < (u64)-1)
+			min_key.offset++;
+		else if (min_key.type < (u8)-1)
+			min_key.type++;
+		else if (min_key.objectid < (u64)-1)
+			min_key.objectid++;
+		else
+			break;
+	}
+	if (ins_nr) {
+		ret = copy_items(trans, log, dst_path, src,
+				 ins_start_slot,
+				 ins_nr, inode_only);
+		BUG_ON(ret);
+		ins_nr = 0;
+	}
+	WARN_ON(ins_nr);
+	if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
+		btrfs_release_path(root, path);
+		btrfs_release_path(log, dst_path);
+		BTRFS_I(inode)->log_dirty_trans = 0;
+		ret = log_directory_changes(trans, root, inode, path, dst_path);
+		BUG_ON(ret);
+	}
+	BTRFS_I(inode)->logged_trans = trans->transid;
+	mutex_unlock(&BTRFS_I(inode)->log_mutex);
+
+	btrfs_free_path(path);
+	btrfs_free_path(dst_path);
+
+	mutex_lock(&root->fs_info->tree_log_mutex);
+	ret = update_log_root(trans, log);
+	BUG_ON(ret);
+	mutex_unlock(&root->fs_info->tree_log_mutex);
+out:
+	return 0;
+}
+
+int btrfs_log_inode(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, struct inode *inode,
+		    int inode_only)
+{
+	int ret;
+
+	start_log_trans(trans, root);
+	ret = __btrfs_log_inode(trans, root, inode, inode_only);
+	end_log_trans(root);
+	return ret;
+}
+
+/*
+ * helper function around btrfs_log_inode to make sure newly created
+ * parent directories also end up in the log.  A minimal inode and backref
+ * only logging is done of any parent directories that are older than
+ * the last committed transaction
+ */
+int btrfs_log_dentry(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, struct dentry *dentry)
+{
+	int inode_only = LOG_INODE_ALL;
+	struct super_block *sb;
+	int ret;
+
+	start_log_trans(trans, root);
+	sb = dentry->d_inode->i_sb;
+	while (1) {
+		ret = __btrfs_log_inode(trans, root, dentry->d_inode,
+					inode_only);
+		BUG_ON(ret);
+		inode_only = LOG_INODE_EXISTS;
+
+		dentry = dentry->d_parent;
+		if (!dentry || !dentry->d_inode || sb != dentry->d_inode->i_sb)
+			break;
+
+		if (BTRFS_I(dentry->d_inode)->generation <=
+		    root->fs_info->last_trans_committed)
+			break;
+	}
+	end_log_trans(root);
+	return 0;
+}
+
+/*
+ * it is not safe to log dentry if the chunk root has added new
+ * chunks.  This returns 0 if the dentry was logged, and 1 otherwise.
+ * If this returns 1, you must commit the transaction to safely get your
+ * data on disk.
+ */
+int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct dentry *dentry)
+{
+	u64 gen;
+	gen = root->fs_info->last_trans_new_blockgroup;
+	if (gen > root->fs_info->last_trans_committed)
+		return 1;
+	else
+		return btrfs_log_dentry(trans, root, dentry);
+}
+
+/*
+ * should be called during mount to recover any replay any log trees
+ * from the FS
+ */
+int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_key tmp_key;
+	struct btrfs_root *log;
+	struct btrfs_fs_info *fs_info = log_root_tree->fs_info;
+	u64 highest_inode;
+	struct walk_control wc = {
+		.process_func = process_one_buffer,
+		.stage = 0,
+	};
+
+	fs_info->log_root_recovering = 1;
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	trans = btrfs_start_transaction(fs_info->tree_root, 1);
+
+	wc.trans = trans;
+	wc.pin = 1;
+
+	walk_log_tree(trans, log_root_tree, &wc);
+
+again:
+	key.objectid = BTRFS_TREE_LOG_OBJECTID;
+	key.offset = (u64)-1;
+	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
+		if (ret < 0)
+			break;
+		if (ret > 0) {
+			if (path->slots[0] == 0)
+				break;
+			path->slots[0]--;
+		}
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+				      path->slots[0]);
+		btrfs_release_path(log_root_tree, path);
+		if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+			break;
+
+		log = btrfs_read_fs_root_no_radix(log_root_tree,
+						  &found_key);
+		BUG_ON(!log);
+
+
+		tmp_key.objectid = found_key.offset;
+		tmp_key.type = BTRFS_ROOT_ITEM_KEY;
+		tmp_key.offset = (u64)-1;
+
+		wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key);
+		BUG_ON(!wc.replay_dest);
+
+		wc.replay_dest->log_root = log;
+		btrfs_record_root_in_trans(wc.replay_dest);
+		ret = walk_log_tree(trans, log, &wc);
+		BUG_ON(ret);
+
+		if (wc.stage == LOG_WALK_REPLAY_ALL) {
+			ret = fixup_inode_link_counts(trans, wc.replay_dest,
+						      path);
+			BUG_ON(ret);
+		}
+		ret = btrfs_find_highest_inode(wc.replay_dest, &highest_inode);
+		if (ret == 0) {
+			wc.replay_dest->highest_inode = highest_inode;
+			wc.replay_dest->last_inode_alloc = highest_inode;
+		}
+
+		key.offset = found_key.offset - 1;
+		wc.replay_dest->log_root = NULL;
+		free_extent_buffer(log->node);
+		kfree(log);
+
+		if (found_key.offset == 0)
+			break;
+	}
+	btrfs_release_path(log_root_tree, path);
+
+	/* step one is to pin it all, step two is to replay just inodes */
+	if (wc.pin) {
+		wc.pin = 0;
+		wc.process_func = replay_one_buffer;
+		wc.stage = LOG_WALK_REPLAY_INODES;
+		goto again;
+	}
+	/* step three is to replay everything */
+	if (wc.stage < LOG_WALK_REPLAY_ALL) {
+		wc.stage++;
+		goto again;
+	}
+
+	btrfs_free_path(path);
+
+	free_extent_buffer(log_root_tree->node);
+	log_root_tree->log_root = NULL;
+	fs_info->log_root_recovering = 0;
+
+	/* step 4: commit the transaction, which also unpins the blocks */
+	btrfs_commit_transaction(trans, fs_info->tree_root);
+
+	kfree(log_root_tree);
+	return 0;
+}
diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h
new file mode 100644
index 00000000000..b9409b32ed0
--- /dev/null
+++ b/fs/btrfs/tree-log.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __TREE_LOG_
+#define __TREE_LOG_
+
+int btrfs_sync_log(struct btrfs_trans_handle *trans,
+		   struct btrfs_root *root);
+int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root);
+int btrfs_log_dentry(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, struct dentry *dentry);
+int btrfs_recover_log_trees(struct btrfs_root *tree_root);
+int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct dentry *dentry);
+int btrfs_log_inode(struct btrfs_trans_handle *trans,
+		    struct btrfs_root *root, struct inode *inode,
+		    int inode_only);
+int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root,
+				 const char *name, int name_len,
+				 struct inode *dir, u64 index);
+int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root,
+			       const char *name, int name_len,
+			       struct inode *inode, u64 dirid);
+#endif
diff --git a/fs/btrfs/version.h b/fs/btrfs/version.h
new file mode 100644
index 00000000000..9bf3946d5ef
--- /dev/null
+++ b/fs/btrfs/version.h
@@ -0,0 +1,4 @@
+#ifndef __BTRFS_VERSION_H
+#define __BTRFS_VERSION_H
+#define BTRFS_BUILD_VERSION "Btrfs"
+#endif
diff --git a/fs/btrfs/version.sh b/fs/btrfs/version.sh
new file mode 100644
index 00000000000..1ca1952fd91
--- /dev/null
+++ b/fs/btrfs/version.sh
@@ -0,0 +1,43 @@
+#!/bin/bash
+#
+# determine-version -- report a useful version for releases
+#
+# Copyright 2008, Aron Griffis <agriffis@n01se.net>
+# Copyright 2008, Oracle
+# Released under the GNU GPLv2
+ 
+v="v0.16"
+
+which git &> /dev/null
+if [ $? == 0 ]; then
+    git branch >& /dev/null
+    if [ $? == 0 ]; then
+	    if head=`git rev-parse --verify HEAD 2>/dev/null`; then
+		if tag=`git describe --tags 2>/dev/null`; then
+		    v="$tag"
+		fi
+
+		# Are there uncommitted changes?
+		git update-index --refresh --unmerged > /dev/null
+		if git diff-index --name-only HEAD | \
+		    grep -v "^scripts/package" \
+		    | read dummy; then
+		    v="$v"-dirty
+		fi
+	    fi
+    fi
+fi
+ 
+echo "#ifndef __BUILD_VERSION" > .build-version.h
+echo "#define __BUILD_VERSION" >> .build-version.h
+echo "#define BTRFS_BUILD_VERSION \"Btrfs $v\"" >> .build-version.h
+echo "#endif" >> .build-version.h
+
+diff -q version.h .build-version.h >& /dev/null
+
+if [ $? == 0 ]; then
+    rm .build-version.h
+    exit 0
+fi
+
+mv .build-version.h version.h
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
new file mode 100644
index 00000000000..b187b537888
--- /dev/null
+++ b/fs/btrfs/volumes.c
@@ -0,0 +1,3218 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+#include <linux/sched.h>
+#include <linux/bio.h>
+#include <linux/buffer_head.h>
+#include <linux/blkdev.h>
+#include <linux/random.h>
+#include <linux/version.h>
+#include <asm/div64.h>
+#include "compat.h"
+#include "ctree.h"
+#include "extent_map.h"
+#include "disk-io.h"
+#include "transaction.h"
+#include "print-tree.h"
+#include "volumes.h"
+#include "async-thread.h"
+
+struct map_lookup {
+	u64 type;
+	int io_align;
+	int io_width;
+	int stripe_len;
+	int sector_size;
+	int num_stripes;
+	int sub_stripes;
+	struct btrfs_bio_stripe stripes[];
+};
+
+static int init_first_rw_device(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root,
+				struct btrfs_device *device);
+static int btrfs_relocate_sys_chunks(struct btrfs_root *root);
+
+#define map_lookup_size(n) (sizeof(struct map_lookup) + \
+			    (sizeof(struct btrfs_bio_stripe) * (n)))
+
+static DEFINE_MUTEX(uuid_mutex);
+static LIST_HEAD(fs_uuids);
+
+void btrfs_lock_volumes(void)
+{
+	mutex_lock(&uuid_mutex);
+}
+
+void btrfs_unlock_volumes(void)
+{
+	mutex_unlock(&uuid_mutex);
+}
+
+static void lock_chunks(struct btrfs_root *root)
+{
+	mutex_lock(&root->fs_info->chunk_mutex);
+}
+
+static void unlock_chunks(struct btrfs_root *root)
+{
+	mutex_unlock(&root->fs_info->chunk_mutex);
+}
+
+static void free_fs_devices(struct btrfs_fs_devices *fs_devices)
+{
+	struct btrfs_device *device;
+	WARN_ON(fs_devices->opened);
+	while (!list_empty(&fs_devices->devices)) {
+		device = list_entry(fs_devices->devices.next,
+				    struct btrfs_device, dev_list);
+		list_del(&device->dev_list);
+		kfree(device->name);
+		kfree(device);
+	}
+	kfree(fs_devices);
+}
+
+int btrfs_cleanup_fs_uuids(void)
+{
+	struct btrfs_fs_devices *fs_devices;
+
+	while (!list_empty(&fs_uuids)) {
+		fs_devices = list_entry(fs_uuids.next,
+					struct btrfs_fs_devices, list);
+		list_del(&fs_devices->list);
+		free_fs_devices(fs_devices);
+	}
+	return 0;
+}
+
+static noinline struct btrfs_device *__find_device(struct list_head *head,
+						   u64 devid, u8 *uuid)
+{
+	struct btrfs_device *dev;
+	struct list_head *cur;
+
+	list_for_each(cur, head) {
+		dev = list_entry(cur, struct btrfs_device, dev_list);
+		if (dev->devid == devid &&
+		    (!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
+			return dev;
+		}
+	}
+	return NULL;
+}
+
+static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid)
+{
+	struct list_head *cur;
+	struct btrfs_fs_devices *fs_devices;
+
+	list_for_each(cur, &fs_uuids) {
+		fs_devices = list_entry(cur, struct btrfs_fs_devices, list);
+		if (memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE) == 0)
+			return fs_devices;
+	}
+	return NULL;
+}
+
+/*
+ * we try to collect pending bios for a device so we don't get a large
+ * number of procs sending bios down to the same device.  This greatly
+ * improves the schedulers ability to collect and merge the bios.
+ *
+ * But, it also turns into a long list of bios to process and that is sure
+ * to eventually make the worker thread block.  The solution here is to
+ * make some progress and then put this work struct back at the end of
+ * the list if the block device is congested.  This way, multiple devices
+ * can make progress from a single worker thread.
+ */
+static noinline int run_scheduled_bios(struct btrfs_device *device)
+{
+	struct bio *pending;
+	struct backing_dev_info *bdi;
+	struct btrfs_fs_info *fs_info;
+	struct bio *tail;
+	struct bio *cur;
+	int again = 0;
+	unsigned long num_run = 0;
+	unsigned long limit;
+
+	bdi = device->bdev->bd_inode->i_mapping->backing_dev_info;
+	fs_info = device->dev_root->fs_info;
+	limit = btrfs_async_submit_limit(fs_info);
+	limit = limit * 2 / 3;
+
+loop:
+	spin_lock(&device->io_lock);
+
+	/* take all the bios off the list at once and process them
+	 * later on (without the lock held).  But, remember the
+	 * tail and other pointers so the bios can be properly reinserted
+	 * into the list if we hit congestion
+	 */
+	pending = device->pending_bios;
+	tail = device->pending_bio_tail;
+	WARN_ON(pending && !tail);
+	device->pending_bios = NULL;
+	device->pending_bio_tail = NULL;
+
+	/*
+	 * if pending was null this time around, no bios need processing
+	 * at all and we can stop.  Otherwise it'll loop back up again
+	 * and do an additional check so no bios are missed.
+	 *
+	 * device->running_pending is used to synchronize with the
+	 * schedule_bio code.
+	 */
+	if (pending) {
+		again = 1;
+		device->running_pending = 1;
+	} else {
+		again = 0;
+		device->running_pending = 0;
+	}
+	spin_unlock(&device->io_lock);
+
+	while (pending) {
+		cur = pending;
+		pending = pending->bi_next;
+		cur->bi_next = NULL;
+		atomic_dec(&fs_info->nr_async_bios);
+
+		if (atomic_read(&fs_info->nr_async_bios) < limit &&
+		    waitqueue_active(&fs_info->async_submit_wait))
+			wake_up(&fs_info->async_submit_wait);
+
+		BUG_ON(atomic_read(&cur->bi_cnt) == 0);
+		bio_get(cur);
+		submit_bio(cur->bi_rw, cur);
+		bio_put(cur);
+		num_run++;
+
+		/*
+		 * we made progress, there is more work to do and the bdi
+		 * is now congested.  Back off and let other work structs
+		 * run instead
+		 */
+		if (pending && bdi_write_congested(bdi) &&
+		    fs_info->fs_devices->open_devices > 1) {
+			struct bio *old_head;
+
+			spin_lock(&device->io_lock);
+
+			old_head = device->pending_bios;
+			device->pending_bios = pending;
+			if (device->pending_bio_tail)
+				tail->bi_next = old_head;
+			else
+				device->pending_bio_tail = tail;
+
+			spin_unlock(&device->io_lock);
+			btrfs_requeue_work(&device->work);
+			goto done;
+		}
+	}
+	if (again)
+		goto loop;
+done:
+	return 0;
+}
+
+static void pending_bios_fn(struct btrfs_work *work)
+{
+	struct btrfs_device *device;
+
+	device = container_of(work, struct btrfs_device, work);
+	run_scheduled_bios(device);
+}
+
+static noinline int device_list_add(const char *path,
+			   struct btrfs_super_block *disk_super,
+			   u64 devid, struct btrfs_fs_devices **fs_devices_ret)
+{
+	struct btrfs_device *device;
+	struct btrfs_fs_devices *fs_devices;
+	u64 found_transid = btrfs_super_generation(disk_super);
+
+	fs_devices = find_fsid(disk_super->fsid);
+	if (!fs_devices) {
+		fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
+		if (!fs_devices)
+			return -ENOMEM;
+		INIT_LIST_HEAD(&fs_devices->devices);
+		INIT_LIST_HEAD(&fs_devices->alloc_list);
+		list_add(&fs_devices->list, &fs_uuids);
+		memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
+		fs_devices->latest_devid = devid;
+		fs_devices->latest_trans = found_transid;
+		device = NULL;
+	} else {
+		device = __find_device(&fs_devices->devices, devid,
+				       disk_super->dev_item.uuid);
+	}
+	if (!device) {
+		if (fs_devices->opened)
+			return -EBUSY;
+
+		device = kzalloc(sizeof(*device), GFP_NOFS);
+		if (!device) {
+			/* we can safely leave the fs_devices entry around */
+			return -ENOMEM;
+		}
+		device->devid = devid;
+		device->work.func = pending_bios_fn;
+		memcpy(device->uuid, disk_super->dev_item.uuid,
+		       BTRFS_UUID_SIZE);
+		device->barriers = 1;
+		spin_lock_init(&device->io_lock);
+		device->name = kstrdup(path, GFP_NOFS);
+		if (!device->name) {
+			kfree(device);
+			return -ENOMEM;
+		}
+		INIT_LIST_HEAD(&device->dev_alloc_list);
+		list_add(&device->dev_list, &fs_devices->devices);
+		device->fs_devices = fs_devices;
+		fs_devices->num_devices++;
+	}
+
+	if (found_transid > fs_devices->latest_trans) {
+		fs_devices->latest_devid = devid;
+		fs_devices->latest_trans = found_transid;
+	}
+	*fs_devices_ret = fs_devices;
+	return 0;
+}
+
+static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
+{
+	struct btrfs_fs_devices *fs_devices;
+	struct btrfs_device *device;
+	struct btrfs_device *orig_dev;
+
+	fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
+	if (!fs_devices)
+		return ERR_PTR(-ENOMEM);
+
+	INIT_LIST_HEAD(&fs_devices->devices);
+	INIT_LIST_HEAD(&fs_devices->alloc_list);
+	INIT_LIST_HEAD(&fs_devices->list);
+	fs_devices->latest_devid = orig->latest_devid;
+	fs_devices->latest_trans = orig->latest_trans;
+	memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid));
+
+	list_for_each_entry(orig_dev, &orig->devices, dev_list) {
+		device = kzalloc(sizeof(*device), GFP_NOFS);
+		if (!device)
+			goto error;
+
+		device->name = kstrdup(orig_dev->name, GFP_NOFS);
+		if (!device->name)
+			goto error;
+
+		device->devid = orig_dev->devid;
+		device->work.func = pending_bios_fn;
+		memcpy(device->uuid, orig_dev->uuid, sizeof(device->uuid));
+		device->barriers = 1;
+		spin_lock_init(&device->io_lock);
+		INIT_LIST_HEAD(&device->dev_list);
+		INIT_LIST_HEAD(&device->dev_alloc_list);
+
+		list_add(&device->dev_list, &fs_devices->devices);
+		device->fs_devices = fs_devices;
+		fs_devices->num_devices++;
+	}
+	return fs_devices;
+error:
+	free_fs_devices(fs_devices);
+	return ERR_PTR(-ENOMEM);
+}
+
+int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
+{
+	struct list_head *tmp;
+	struct list_head *cur;
+	struct btrfs_device *device;
+
+	mutex_lock(&uuid_mutex);
+again:
+	list_for_each_safe(cur, tmp, &fs_devices->devices) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		if (device->in_fs_metadata)
+			continue;
+
+		if (device->bdev) {
+			close_bdev_exclusive(device->bdev, device->mode);
+			device->bdev = NULL;
+			fs_devices->open_devices--;
+		}
+		if (device->writeable) {
+			list_del_init(&device->dev_alloc_list);
+			device->writeable = 0;
+			fs_devices->rw_devices--;
+		}
+		list_del_init(&device->dev_list);
+		fs_devices->num_devices--;
+		kfree(device->name);
+		kfree(device);
+	}
+
+	if (fs_devices->seed) {
+		fs_devices = fs_devices->seed;
+		goto again;
+	}
+
+	mutex_unlock(&uuid_mutex);
+	return 0;
+}
+
+static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
+{
+	struct list_head *cur;
+	struct btrfs_device *device;
+
+	if (--fs_devices->opened > 0)
+		return 0;
+
+	list_for_each(cur, &fs_devices->devices) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		if (device->bdev) {
+			close_bdev_exclusive(device->bdev, device->mode);
+			fs_devices->open_devices--;
+		}
+		if (device->writeable) {
+			list_del_init(&device->dev_alloc_list);
+			fs_devices->rw_devices--;
+		}
+
+		device->bdev = NULL;
+		device->writeable = 0;
+		device->in_fs_metadata = 0;
+	}
+	WARN_ON(fs_devices->open_devices);
+	WARN_ON(fs_devices->rw_devices);
+	fs_devices->opened = 0;
+	fs_devices->seeding = 0;
+
+	return 0;
+}
+
+int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
+{
+	struct btrfs_fs_devices *seed_devices = NULL;
+	int ret;
+
+	mutex_lock(&uuid_mutex);
+	ret = __btrfs_close_devices(fs_devices);
+	if (!fs_devices->opened) {
+		seed_devices = fs_devices->seed;
+		fs_devices->seed = NULL;
+	}
+	mutex_unlock(&uuid_mutex);
+
+	while (seed_devices) {
+		fs_devices = seed_devices;
+		seed_devices = fs_devices->seed;
+		__btrfs_close_devices(fs_devices);
+		free_fs_devices(fs_devices);
+	}
+	return ret;
+}
+
+static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
+				fmode_t flags, void *holder)
+{
+	struct block_device *bdev;
+	struct list_head *head = &fs_devices->devices;
+	struct list_head *cur;
+	struct btrfs_device *device;
+	struct block_device *latest_bdev = NULL;
+	struct buffer_head *bh;
+	struct btrfs_super_block *disk_super;
+	u64 latest_devid = 0;
+	u64 latest_transid = 0;
+	u64 devid;
+	int seeding = 1;
+	int ret = 0;
+
+	list_for_each(cur, head) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		if (device->bdev)
+			continue;
+		if (!device->name)
+			continue;
+
+		bdev = open_bdev_exclusive(device->name, flags, holder);
+		if (IS_ERR(bdev)) {
+			printk(KERN_INFO "open %s failed\n", device->name);
+			goto error;
+		}
+		set_blocksize(bdev, 4096);
+
+		bh = btrfs_read_dev_super(bdev);
+		if (!bh)
+			goto error_close;
+
+		disk_super = (struct btrfs_super_block *)bh->b_data;
+		devid = le64_to_cpu(disk_super->dev_item.devid);
+		if (devid != device->devid)
+			goto error_brelse;
+
+		if (memcmp(device->uuid, disk_super->dev_item.uuid,
+			   BTRFS_UUID_SIZE))
+			goto error_brelse;
+
+		device->generation = btrfs_super_generation(disk_super);
+		if (!latest_transid || device->generation > latest_transid) {
+			latest_devid = devid;
+			latest_transid = device->generation;
+			latest_bdev = bdev;
+		}
+
+		if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING) {
+			device->writeable = 0;
+		} else {
+			device->writeable = !bdev_read_only(bdev);
+			seeding = 0;
+		}
+
+		device->bdev = bdev;
+		device->in_fs_metadata = 0;
+		device->mode = flags;
+
+		fs_devices->open_devices++;
+		if (device->writeable) {
+			fs_devices->rw_devices++;
+			list_add(&device->dev_alloc_list,
+				 &fs_devices->alloc_list);
+		}
+		continue;
+
+error_brelse:
+		brelse(bh);
+error_close:
+		close_bdev_exclusive(bdev, FMODE_READ);
+error:
+		continue;
+	}
+	if (fs_devices->open_devices == 0) {
+		ret = -EIO;
+		goto out;
+	}
+	fs_devices->seeding = seeding;
+	fs_devices->opened = 1;
+	fs_devices->latest_bdev = latest_bdev;
+	fs_devices->latest_devid = latest_devid;
+	fs_devices->latest_trans = latest_transid;
+	fs_devices->total_rw_bytes = 0;
+out:
+	return ret;
+}
+
+int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
+		       fmode_t flags, void *holder)
+{
+	int ret;
+
+	mutex_lock(&uuid_mutex);
+	if (fs_devices->opened) {
+		fs_devices->opened++;
+		ret = 0;
+	} else {
+		ret = __btrfs_open_devices(fs_devices, flags, holder);
+	}
+	mutex_unlock(&uuid_mutex);
+	return ret;
+}
+
+int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
+			  struct btrfs_fs_devices **fs_devices_ret)
+{
+	struct btrfs_super_block *disk_super;
+	struct block_device *bdev;
+	struct buffer_head *bh;
+	int ret;
+	u64 devid;
+	u64 transid;
+
+	mutex_lock(&uuid_mutex);
+
+	bdev = open_bdev_exclusive(path, flags, holder);
+
+	if (IS_ERR(bdev)) {
+		ret = PTR_ERR(bdev);
+		goto error;
+	}
+
+	ret = set_blocksize(bdev, 4096);
+	if (ret)
+		goto error_close;
+	bh = btrfs_read_dev_super(bdev);
+	if (!bh) {
+		ret = -EIO;
+		goto error_close;
+	}
+	disk_super = (struct btrfs_super_block *)bh->b_data;
+	devid = le64_to_cpu(disk_super->dev_item.devid);
+	transid = btrfs_super_generation(disk_super);
+	if (disk_super->label[0])
+		printk(KERN_INFO "device label %s ", disk_super->label);
+	else {
+		/* FIXME, make a readl uuid parser */
+		printk(KERN_INFO "device fsid %llx-%llx ",
+		       *(unsigned long long *)disk_super->fsid,
+		       *(unsigned long long *)(disk_super->fsid + 8));
+	}
+	printk(KERN_INFO "devid %llu transid %llu %s\n",
+	       (unsigned long long)devid, (unsigned long long)transid, path);
+	ret = device_list_add(path, disk_super, devid, fs_devices_ret);
+
+	brelse(bh);
+error_close:
+	close_bdev_exclusive(bdev, flags);
+error:
+	mutex_unlock(&uuid_mutex);
+	return ret;
+}
+
+/*
+ * this uses a pretty simple search, the expectation is that it is
+ * called very infrequently and that a given device has a small number
+ * of extents
+ */
+static noinline int find_free_dev_extent(struct btrfs_trans_handle *trans,
+					 struct btrfs_device *device,
+					 u64 num_bytes, u64 *start)
+{
+	struct btrfs_key key;
+	struct btrfs_root *root = device->dev_root;
+	struct btrfs_dev_extent *dev_extent = NULL;
+	struct btrfs_path *path;
+	u64 hole_size = 0;
+	u64 last_byte = 0;
+	u64 search_start = 0;
+	u64 search_end = device->total_bytes;
+	int ret;
+	int slot = 0;
+	int start_found;
+	struct extent_buffer *l;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	path->reada = 2;
+	start_found = 0;
+
+	/* FIXME use last free of some kind */
+
+	/* we don't want to overwrite the superblock on the drive,
+	 * so we make sure to start at an offset of at least 1MB
+	 */
+	search_start = max((u64)1024 * 1024, search_start);
+
+	if (root->fs_info->alloc_start + num_bytes <= device->total_bytes)
+		search_start = max(root->fs_info->alloc_start, search_start);
+
+	key.objectid = device->devid;
+	key.offset = search_start;
+	key.type = BTRFS_DEV_EXTENT_KEY;
+	ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto error;
+	ret = btrfs_previous_item(root, path, 0, key.type);
+	if (ret < 0)
+		goto error;
+	l = path->nodes[0];
+	btrfs_item_key_to_cpu(l, &key, path->slots[0]);
+	while (1) {
+		l = path->nodes[0];
+		slot = path->slots[0];
+		if (slot >= btrfs_header_nritems(l)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret == 0)
+				continue;
+			if (ret < 0)
+				goto error;
+no_more_items:
+			if (!start_found) {
+				if (search_start >= search_end) {
+					ret = -ENOSPC;
+					goto error;
+				}
+				*start = search_start;
+				start_found = 1;
+				goto check_pending;
+			}
+			*start = last_byte > search_start ?
+				last_byte : search_start;
+			if (search_end <= *start) {
+				ret = -ENOSPC;
+				goto error;
+			}
+			goto check_pending;
+		}
+		btrfs_item_key_to_cpu(l, &key, slot);
+
+		if (key.objectid < device->devid)
+			goto next;
+
+		if (key.objectid > device->devid)
+			goto no_more_items;
+
+		if (key.offset >= search_start && key.offset > last_byte &&
+		    start_found) {
+			if (last_byte < search_start)
+				last_byte = search_start;
+			hole_size = key.offset - last_byte;
+			if (key.offset > last_byte &&
+			    hole_size >= num_bytes) {
+				*start = last_byte;
+				goto check_pending;
+			}
+		}
+		if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
+			goto next;
+
+		start_found = 1;
+		dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
+		last_byte = key.offset + btrfs_dev_extent_length(l, dev_extent);
+next:
+		path->slots[0]++;
+		cond_resched();
+	}
+check_pending:
+	/* we have to make sure we didn't find an extent that has already
+	 * been allocated by the map tree or the original allocation
+	 */
+	BUG_ON(*start < search_start);
+
+	if (*start + num_bytes > search_end) {
+		ret = -ENOSPC;
+		goto error;
+	}
+	/* check for pending inserts here */
+	ret = 0;
+
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
+			  struct btrfs_device *device,
+			  u64 start)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_root *root = device->dev_root;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct extent_buffer *leaf = NULL;
+	struct btrfs_dev_extent *extent = NULL;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = device->devid;
+	key.offset = start;
+	key.type = BTRFS_DEV_EXTENT_KEY;
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret > 0) {
+		ret = btrfs_previous_item(root, path, key.objectid,
+					  BTRFS_DEV_EXTENT_KEY);
+		BUG_ON(ret);
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+		extent = btrfs_item_ptr(leaf, path->slots[0],
+					struct btrfs_dev_extent);
+		BUG_ON(found_key.offset > start || found_key.offset +
+		       btrfs_dev_extent_length(leaf, extent) < start);
+		ret = 0;
+	} else if (ret == 0) {
+		leaf = path->nodes[0];
+		extent = btrfs_item_ptr(leaf, path->slots[0],
+					struct btrfs_dev_extent);
+	}
+	BUG_ON(ret);
+
+	if (device->bytes_used > 0)
+		device->bytes_used -= btrfs_dev_extent_length(leaf, extent);
+	ret = btrfs_del_item(trans, root, path);
+	BUG_ON(ret);
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
+			   struct btrfs_device *device,
+			   u64 chunk_tree, u64 chunk_objectid,
+			   u64 chunk_offset, u64 start, u64 num_bytes)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_root *root = device->dev_root;
+	struct btrfs_dev_extent *extent;
+	struct extent_buffer *leaf;
+	struct btrfs_key key;
+
+	WARN_ON(!device->in_fs_metadata);
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = device->devid;
+	key.offset = start;
+	key.type = BTRFS_DEV_EXTENT_KEY;
+	ret = btrfs_insert_empty_item(trans, root, path, &key,
+				      sizeof(*extent));
+	BUG_ON(ret);
+
+	leaf = path->nodes[0];
+	extent = btrfs_item_ptr(leaf, path->slots[0],
+				struct btrfs_dev_extent);
+	btrfs_set_dev_extent_chunk_tree(leaf, extent, chunk_tree);
+	btrfs_set_dev_extent_chunk_objectid(leaf, extent, chunk_objectid);
+	btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);
+
+	write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
+		    (unsigned long)btrfs_dev_extent_chunk_tree_uuid(extent),
+		    BTRFS_UUID_SIZE);
+
+	btrfs_set_dev_extent_length(leaf, extent, num_bytes);
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_free_path(path);
+	return ret;
+}
+
+static noinline int find_next_chunk(struct btrfs_root *root,
+				    u64 objectid, u64 *offset)
+{
+	struct btrfs_path *path;
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_chunk *chunk;
+	struct btrfs_key found_key;
+
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	key.objectid = objectid;
+	key.offset = (u64)-1;
+	key.type = BTRFS_CHUNK_ITEM_KEY;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto error;
+
+	BUG_ON(ret == 0);
+
+	ret = btrfs_previous_item(root, path, 0, BTRFS_CHUNK_ITEM_KEY);
+	if (ret) {
+		*offset = 0;
+	} else {
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+				      path->slots[0]);
+		if (found_key.objectid != objectid)
+			*offset = 0;
+		else {
+			chunk = btrfs_item_ptr(path->nodes[0], path->slots[0],
+					       struct btrfs_chunk);
+			*offset = found_key.offset +
+				btrfs_chunk_length(path->nodes[0], chunk);
+		}
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static noinline int find_next_devid(struct btrfs_root *root, u64 *objectid)
+{
+	int ret;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_path *path;
+
+	root = root->fs_info->chunk_root;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+	key.type = BTRFS_DEV_ITEM_KEY;
+	key.offset = (u64)-1;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto error;
+
+	BUG_ON(ret == 0);
+
+	ret = btrfs_previous_item(root, path, BTRFS_DEV_ITEMS_OBJECTID,
+				  BTRFS_DEV_ITEM_KEY);
+	if (ret) {
+		*objectid = 1;
+	} else {
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+				      path->slots[0]);
+		*objectid = found_key.offset + 1;
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+/*
+ * the device information is stored in the chunk root
+ * the btrfs_device struct should be fully filled in
+ */
+int btrfs_add_device(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root,
+		     struct btrfs_device *device)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_dev_item *dev_item;
+	struct extent_buffer *leaf;
+	struct btrfs_key key;
+	unsigned long ptr;
+
+	root = root->fs_info->chunk_root;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+	key.type = BTRFS_DEV_ITEM_KEY;
+	key.offset = device->devid;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &key,
+				      sizeof(*dev_item));
+	if (ret)
+		goto out;
+
+	leaf = path->nodes[0];
+	dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);
+
+	btrfs_set_device_id(leaf, dev_item, device->devid);
+	btrfs_set_device_generation(leaf, dev_item, 0);
+	btrfs_set_device_type(leaf, dev_item, device->type);
+	btrfs_set_device_io_align(leaf, dev_item, device->io_align);
+	btrfs_set_device_io_width(leaf, dev_item, device->io_width);
+	btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
+	btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
+	btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
+	btrfs_set_device_group(leaf, dev_item, 0);
+	btrfs_set_device_seek_speed(leaf, dev_item, 0);
+	btrfs_set_device_bandwidth(leaf, dev_item, 0);
+	btrfs_set_device_start_offset(leaf, dev_item, 0);
+
+	ptr = (unsigned long)btrfs_device_uuid(dev_item);
+	write_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
+	ptr = (unsigned long)btrfs_device_fsid(dev_item);
+	write_extent_buffer(leaf, root->fs_info->fsid, ptr, BTRFS_UUID_SIZE);
+	btrfs_mark_buffer_dirty(leaf);
+
+	ret = 0;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_rm_dev_item(struct btrfs_root *root,
+			     struct btrfs_device *device)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_trans_handle *trans;
+
+	root = root->fs_info->chunk_root;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	trans = btrfs_start_transaction(root, 1);
+	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+	key.type = BTRFS_DEV_ITEM_KEY;
+	key.offset = device->devid;
+	lock_chunks(root);
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	if (ret < 0)
+		goto out;
+
+	if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	ret = btrfs_del_item(trans, root, path);
+	if (ret)
+		goto out;
+out:
+	btrfs_free_path(path);
+	unlock_chunks(root);
+	btrfs_commit_transaction(trans, root);
+	return ret;
+}
+
+int btrfs_rm_device(struct btrfs_root *root, char *device_path)
+{
+	struct btrfs_device *device;
+	struct btrfs_device *next_device;
+	struct block_device *bdev;
+	struct buffer_head *bh = NULL;
+	struct btrfs_super_block *disk_super;
+	u64 all_avail;
+	u64 devid;
+	u64 num_devices;
+	u8 *dev_uuid;
+	int ret = 0;
+
+	mutex_lock(&uuid_mutex);
+	mutex_lock(&root->fs_info->volume_mutex);
+
+	all_avail = root->fs_info->avail_data_alloc_bits |
+		root->fs_info->avail_system_alloc_bits |
+		root->fs_info->avail_metadata_alloc_bits;
+
+	if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) &&
+	    root->fs_info->fs_devices->rw_devices <= 4) {
+		printk(KERN_ERR "btrfs: unable to go below four devices "
+		       "on raid10\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) &&
+	    root->fs_info->fs_devices->rw_devices <= 2) {
+		printk(KERN_ERR "btrfs: unable to go below two "
+		       "devices on raid1\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
+	if (strcmp(device_path, "missing") == 0) {
+		struct list_head *cur;
+		struct list_head *devices;
+		struct btrfs_device *tmp;
+
+		device = NULL;
+		devices = &root->fs_info->fs_devices->devices;
+		list_for_each(cur, devices) {
+			tmp = list_entry(cur, struct btrfs_device, dev_list);
+			if (tmp->in_fs_metadata && !tmp->bdev) {
+				device = tmp;
+				break;
+			}
+		}
+		bdev = NULL;
+		bh = NULL;
+		disk_super = NULL;
+		if (!device) {
+			printk(KERN_ERR "btrfs: no missing devices found to "
+			       "remove\n");
+			goto out;
+		}
+	} else {
+		bdev = open_bdev_exclusive(device_path, FMODE_READ,
+				      root->fs_info->bdev_holder);
+		if (IS_ERR(bdev)) {
+			ret = PTR_ERR(bdev);
+			goto out;
+		}
+
+		set_blocksize(bdev, 4096);
+		bh = btrfs_read_dev_super(bdev);
+		if (!bh) {
+			ret = -EIO;
+			goto error_close;
+		}
+		disk_super = (struct btrfs_super_block *)bh->b_data;
+		devid = le64_to_cpu(disk_super->dev_item.devid);
+		dev_uuid = disk_super->dev_item.uuid;
+		device = btrfs_find_device(root, devid, dev_uuid,
+					   disk_super->fsid);
+		if (!device) {
+			ret = -ENOENT;
+			goto error_brelse;
+		}
+	}
+
+	if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) {
+		printk(KERN_ERR "btrfs: unable to remove the only writeable "
+		       "device\n");
+		ret = -EINVAL;
+		goto error_brelse;
+	}
+
+	if (device->writeable) {
+		list_del_init(&device->dev_alloc_list);
+		root->fs_info->fs_devices->rw_devices--;
+	}
+
+	ret = btrfs_shrink_device(device, 0);
+	if (ret)
+		goto error_brelse;
+
+	ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device);
+	if (ret)
+		goto error_brelse;
+
+	device->in_fs_metadata = 0;
+	list_del_init(&device->dev_list);
+	device->fs_devices->num_devices--;
+
+	next_device = list_entry(root->fs_info->fs_devices->devices.next,
+				 struct btrfs_device, dev_list);
+	if (device->bdev == root->fs_info->sb->s_bdev)
+		root->fs_info->sb->s_bdev = next_device->bdev;
+	if (device->bdev == root->fs_info->fs_devices->latest_bdev)
+		root->fs_info->fs_devices->latest_bdev = next_device->bdev;
+
+	if (device->bdev) {
+		close_bdev_exclusive(device->bdev, device->mode);
+		device->bdev = NULL;
+		device->fs_devices->open_devices--;
+	}
+
+	num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1;
+	btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices);
+
+	if (device->fs_devices->open_devices == 0) {
+		struct btrfs_fs_devices *fs_devices;
+		fs_devices = root->fs_info->fs_devices;
+		while (fs_devices) {
+			if (fs_devices->seed == device->fs_devices)
+				break;
+			fs_devices = fs_devices->seed;
+		}
+		fs_devices->seed = device->fs_devices->seed;
+		device->fs_devices->seed = NULL;
+		__btrfs_close_devices(device->fs_devices);
+		free_fs_devices(device->fs_devices);
+	}
+
+	/*
+	 * at this point, the device is zero sized.  We want to
+	 * remove it from the devices list and zero out the old super
+	 */
+	if (device->writeable) {
+		/* make sure this device isn't detected as part of
+		 * the FS anymore
+		 */
+		memset(&disk_super->magic, 0, sizeof(disk_super->magic));
+		set_buffer_dirty(bh);
+		sync_dirty_buffer(bh);
+	}
+
+	kfree(device->name);
+	kfree(device);
+	ret = 0;
+
+error_brelse:
+	brelse(bh);
+error_close:
+	if (bdev)
+		close_bdev_exclusive(bdev, FMODE_READ);
+out:
+	mutex_unlock(&root->fs_info->volume_mutex);
+	mutex_unlock(&uuid_mutex);
+	return ret;
+}
+
+/*
+ * does all the dirty work required for changing file system's UUID.
+ */
+static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root)
+{
+	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
+	struct btrfs_fs_devices *old_devices;
+	struct btrfs_fs_devices *seed_devices;
+	struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
+	struct btrfs_device *device;
+	u64 super_flags;
+
+	BUG_ON(!mutex_is_locked(&uuid_mutex));
+	if (!fs_devices->seeding)
+		return -EINVAL;
+
+	seed_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
+	if (!seed_devices)
+		return -ENOMEM;
+
+	old_devices = clone_fs_devices(fs_devices);
+	if (IS_ERR(old_devices)) {
+		kfree(seed_devices);
+		return PTR_ERR(old_devices);
+	}
+
+	list_add(&old_devices->list, &fs_uuids);
+
+	memcpy(seed_devices, fs_devices, sizeof(*seed_devices));
+	seed_devices->opened = 1;
+	INIT_LIST_HEAD(&seed_devices->devices);
+	INIT_LIST_HEAD(&seed_devices->alloc_list);
+	list_splice_init(&fs_devices->devices, &seed_devices->devices);
+	list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list);
+	list_for_each_entry(device, &seed_devices->devices, dev_list) {
+		device->fs_devices = seed_devices;
+	}
+
+	fs_devices->seeding = 0;
+	fs_devices->num_devices = 0;
+	fs_devices->open_devices = 0;
+	fs_devices->seed = seed_devices;
+
+	generate_random_uuid(fs_devices->fsid);
+	memcpy(root->fs_info->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
+	memcpy(disk_super->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
+	super_flags = btrfs_super_flags(disk_super) &
+		      ~BTRFS_SUPER_FLAG_SEEDING;
+	btrfs_set_super_flags(disk_super, super_flags);
+
+	return 0;
+}
+
+/*
+ * strore the expected generation for seed devices in device items.
+ */
+static int btrfs_finish_sprout(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root)
+{
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_dev_item *dev_item;
+	struct btrfs_device *device;
+	struct btrfs_key key;
+	u8 fs_uuid[BTRFS_UUID_SIZE];
+	u8 dev_uuid[BTRFS_UUID_SIZE];
+	u64 devid;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	root = root->fs_info->chunk_root;
+	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+	key.offset = 0;
+	key.type = BTRFS_DEV_ITEM_KEY;
+
+	while (1) {
+		ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+		if (ret < 0)
+			goto error;
+
+		leaf = path->nodes[0];
+next_slot:
+		if (path->slots[0] >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret > 0)
+				break;
+			if (ret < 0)
+				goto error;
+			leaf = path->nodes[0];
+			btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+			btrfs_release_path(root, path);
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+		if (key.objectid != BTRFS_DEV_ITEMS_OBJECTID ||
+		    key.type != BTRFS_DEV_ITEM_KEY)
+			break;
+
+		dev_item = btrfs_item_ptr(leaf, path->slots[0],
+					  struct btrfs_dev_item);
+		devid = btrfs_device_id(leaf, dev_item);
+		read_extent_buffer(leaf, dev_uuid,
+				   (unsigned long)btrfs_device_uuid(dev_item),
+				   BTRFS_UUID_SIZE);
+		read_extent_buffer(leaf, fs_uuid,
+				   (unsigned long)btrfs_device_fsid(dev_item),
+				   BTRFS_UUID_SIZE);
+		device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
+		BUG_ON(!device);
+
+		if (device->fs_devices->seeding) {
+			btrfs_set_device_generation(leaf, dev_item,
+						    device->generation);
+			btrfs_mark_buffer_dirty(leaf);
+		}
+
+		path->slots[0]++;
+		goto next_slot;
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_device *device;
+	struct block_device *bdev;
+	struct list_head *cur;
+	struct list_head *devices;
+	struct super_block *sb = root->fs_info->sb;
+	u64 total_bytes;
+	int seeding_dev = 0;
+	int ret = 0;
+
+	if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding)
+		return -EINVAL;
+
+	bdev = open_bdev_exclusive(device_path, 0, root->fs_info->bdev_holder);
+	if (!bdev)
+		return -EIO;
+
+	if (root->fs_info->fs_devices->seeding) {
+		seeding_dev = 1;
+		down_write(&sb->s_umount);
+		mutex_lock(&uuid_mutex);
+	}
+
+	filemap_write_and_wait(bdev->bd_inode->i_mapping);
+	mutex_lock(&root->fs_info->volume_mutex);
+
+	devices = &root->fs_info->fs_devices->devices;
+	list_for_each(cur, devices) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		if (device->bdev == bdev) {
+			ret = -EEXIST;
+			goto error;
+		}
+	}
+
+	device = kzalloc(sizeof(*device), GFP_NOFS);
+	if (!device) {
+		/* we can safely leave the fs_devices entry around */
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	device->name = kstrdup(device_path, GFP_NOFS);
+	if (!device->name) {
+		kfree(device);
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	ret = find_next_devid(root, &device->devid);
+	if (ret) {
+		kfree(device);
+		goto error;
+	}
+
+	trans = btrfs_start_transaction(root, 1);
+	lock_chunks(root);
+
+	device->barriers = 1;
+	device->writeable = 1;
+	device->work.func = pending_bios_fn;
+	generate_random_uuid(device->uuid);
+	spin_lock_init(&device->io_lock);
+	device->generation = trans->transid;
+	device->io_width = root->sectorsize;
+	device->io_align = root->sectorsize;
+	device->sector_size = root->sectorsize;
+	device->total_bytes = i_size_read(bdev->bd_inode);
+	device->dev_root = root->fs_info->dev_root;
+	device->bdev = bdev;
+	device->in_fs_metadata = 1;
+	device->mode = 0;
+	set_blocksize(device->bdev, 4096);
+
+	if (seeding_dev) {
+		sb->s_flags &= ~MS_RDONLY;
+		ret = btrfs_prepare_sprout(trans, root);
+		BUG_ON(ret);
+	}
+
+	device->fs_devices = root->fs_info->fs_devices;
+	list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
+	list_add(&device->dev_alloc_list,
+		 &root->fs_info->fs_devices->alloc_list);
+	root->fs_info->fs_devices->num_devices++;
+	root->fs_info->fs_devices->open_devices++;
+	root->fs_info->fs_devices->rw_devices++;
+	root->fs_info->fs_devices->total_rw_bytes += device->total_bytes;
+
+	total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
+	btrfs_set_super_total_bytes(&root->fs_info->super_copy,
+				    total_bytes + device->total_bytes);
+
+	total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
+	btrfs_set_super_num_devices(&root->fs_info->super_copy,
+				    total_bytes + 1);
+
+	if (seeding_dev) {
+		ret = init_first_rw_device(trans, root, device);
+		BUG_ON(ret);
+		ret = btrfs_finish_sprout(trans, root);
+		BUG_ON(ret);
+	} else {
+		ret = btrfs_add_device(trans, root, device);
+	}
+
+	unlock_chunks(root);
+	btrfs_commit_transaction(trans, root);
+
+	if (seeding_dev) {
+		mutex_unlock(&uuid_mutex);
+		up_write(&sb->s_umount);
+
+		ret = btrfs_relocate_sys_chunks(root);
+		BUG_ON(ret);
+	}
+out:
+	mutex_unlock(&root->fs_info->volume_mutex);
+	return ret;
+error:
+	close_bdev_exclusive(bdev, 0);
+	if (seeding_dev) {
+		mutex_unlock(&uuid_mutex);
+		up_write(&sb->s_umount);
+	}
+	goto out;
+}
+
+static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
+					struct btrfs_device *device)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_root *root;
+	struct btrfs_dev_item *dev_item;
+	struct extent_buffer *leaf;
+	struct btrfs_key key;
+
+	root = device->dev_root->fs_info->chunk_root;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+	key.type = BTRFS_DEV_ITEM_KEY;
+	key.offset = device->devid;
+
+	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
+	if (ret < 0)
+		goto out;
+
+	if (ret > 0) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);
+
+	btrfs_set_device_id(leaf, dev_item, device->devid);
+	btrfs_set_device_type(leaf, dev_item, device->type);
+	btrfs_set_device_io_align(leaf, dev_item, device->io_align);
+	btrfs_set_device_io_width(leaf, dev_item, device->io_width);
+	btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
+	btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
+	btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
+	btrfs_mark_buffer_dirty(leaf);
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
+		      struct btrfs_device *device, u64 new_size)
+{
+	struct btrfs_super_block *super_copy =
+		&device->dev_root->fs_info->super_copy;
+	u64 old_total = btrfs_super_total_bytes(super_copy);
+	u64 diff = new_size - device->total_bytes;
+
+	if (!device->writeable)
+		return -EACCES;
+	if (new_size <= device->total_bytes)
+		return -EINVAL;
+
+	btrfs_set_super_total_bytes(super_copy, old_total + diff);
+	device->fs_devices->total_rw_bytes += diff;
+
+	device->total_bytes = new_size;
+	return btrfs_update_device(trans, device);
+}
+
+int btrfs_grow_device(struct btrfs_trans_handle *trans,
+		      struct btrfs_device *device, u64 new_size)
+{
+	int ret;
+	lock_chunks(device->dev_root);
+	ret = __btrfs_grow_device(trans, device, new_size);
+	unlock_chunks(device->dev_root);
+	return ret;
+}
+
+static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root,
+			    u64 chunk_tree, u64 chunk_objectid,
+			    u64 chunk_offset)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+
+	root = root->fs_info->chunk_root;
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = chunk_objectid;
+	key.offset = chunk_offset;
+	key.type = BTRFS_CHUNK_ITEM_KEY;
+
+	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+	BUG_ON(ret);
+
+	ret = btrfs_del_item(trans, root, path);
+	BUG_ON(ret);
+
+	btrfs_free_path(path);
+	return 0;
+}
+
+static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
+			chunk_offset)
+{
+	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+	struct btrfs_disk_key *disk_key;
+	struct btrfs_chunk *chunk;
+	u8 *ptr;
+	int ret = 0;
+	u32 num_stripes;
+	u32 array_size;
+	u32 len = 0;
+	u32 cur;
+	struct btrfs_key key;
+
+	array_size = btrfs_super_sys_array_size(super_copy);
+
+	ptr = super_copy->sys_chunk_array;
+	cur = 0;
+
+	while (cur < array_size) {
+		disk_key = (struct btrfs_disk_key *)ptr;
+		btrfs_disk_key_to_cpu(&key, disk_key);
+
+		len = sizeof(*disk_key);
+
+		if (key.type == BTRFS_CHUNK_ITEM_KEY) {
+			chunk = (struct btrfs_chunk *)(ptr + len);
+			num_stripes = btrfs_stack_chunk_num_stripes(chunk);
+			len += btrfs_chunk_item_size(num_stripes);
+		} else {
+			ret = -EIO;
+			break;
+		}
+		if (key.objectid == chunk_objectid &&
+		    key.offset == chunk_offset) {
+			memmove(ptr, ptr + len, array_size - (cur + len));
+			array_size -= len;
+			btrfs_set_super_sys_array_size(super_copy, array_size);
+		} else {
+			ptr += len;
+			cur += len;
+		}
+	}
+	return ret;
+}
+
+static int btrfs_relocate_chunk(struct btrfs_root *root,
+			 u64 chunk_tree, u64 chunk_objectid,
+			 u64 chunk_offset)
+{
+	struct extent_map_tree *em_tree;
+	struct btrfs_root *extent_root;
+	struct btrfs_trans_handle *trans;
+	struct extent_map *em;
+	struct map_lookup *map;
+	int ret;
+	int i;
+
+	printk(KERN_INFO "btrfs relocating chunk %llu\n",
+	       (unsigned long long)chunk_offset);
+	root = root->fs_info->chunk_root;
+	extent_root = root->fs_info->extent_root;
+	em_tree = &root->fs_info->mapping_tree.map_tree;
+
+	/* step one, relocate all the extents inside this chunk */
+	ret = btrfs_relocate_block_group(extent_root, chunk_offset);
+	BUG_ON(ret);
+
+	trans = btrfs_start_transaction(root, 1);
+	BUG_ON(!trans);
+
+	lock_chunks(root);
+
+	/*
+	 * step two, delete the device extents and the
+	 * chunk tree entries
+	 */
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, chunk_offset, 1);
+	spin_unlock(&em_tree->lock);
+
+	BUG_ON(em->start > chunk_offset ||
+	       em->start + em->len < chunk_offset);
+	map = (struct map_lookup *)em->bdev;
+
+	for (i = 0; i < map->num_stripes; i++) {
+		ret = btrfs_free_dev_extent(trans, map->stripes[i].dev,
+					    map->stripes[i].physical);
+		BUG_ON(ret);
+
+		if (map->stripes[i].dev) {
+			ret = btrfs_update_device(trans, map->stripes[i].dev);
+			BUG_ON(ret);
+		}
+	}
+	ret = btrfs_free_chunk(trans, root, chunk_tree, chunk_objectid,
+			       chunk_offset);
+
+	BUG_ON(ret);
+
+	if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
+		ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset);
+		BUG_ON(ret);
+	}
+
+	ret = btrfs_remove_block_group(trans, extent_root, chunk_offset);
+	BUG_ON(ret);
+
+	spin_lock(&em_tree->lock);
+	remove_extent_mapping(em_tree, em);
+	spin_unlock(&em_tree->lock);
+
+	kfree(map);
+	em->bdev = NULL;
+
+	/* once for the tree */
+	free_extent_map(em);
+	/* once for us */
+	free_extent_map(em);
+
+	unlock_chunks(root);
+	btrfs_end_transaction(trans, root);
+	return 0;
+}
+
+static int btrfs_relocate_sys_chunks(struct btrfs_root *root)
+{
+	struct btrfs_root *chunk_root = root->fs_info->chunk_root;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_chunk *chunk;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	u64 chunk_tree = chunk_root->root_key.objectid;
+	u64 chunk_type;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+	key.offset = (u64)-1;
+	key.type = BTRFS_CHUNK_ITEM_KEY;
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
+		if (ret < 0)
+			goto error;
+		BUG_ON(ret == 0);
+
+		ret = btrfs_previous_item(chunk_root, path, key.objectid,
+					  key.type);
+		if (ret < 0)
+			goto error;
+		if (ret > 0)
+			break;
+
+		leaf = path->nodes[0];
+		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+		chunk = btrfs_item_ptr(leaf, path->slots[0],
+				       struct btrfs_chunk);
+		chunk_type = btrfs_chunk_type(leaf, chunk);
+		btrfs_release_path(chunk_root, path);
+
+		if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) {
+			ret = btrfs_relocate_chunk(chunk_root, chunk_tree,
+						   found_key.objectid,
+						   found_key.offset);
+			BUG_ON(ret);
+		}
+
+		if (found_key.offset == 0)
+			break;
+		key.offset = found_key.offset - 1;
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static u64 div_factor(u64 num, int factor)
+{
+	if (factor == 10)
+		return num;
+	num *= factor;
+	do_div(num, 10);
+	return num;
+}
+
+int btrfs_balance(struct btrfs_root *dev_root)
+{
+	int ret;
+	struct list_head *cur;
+	struct list_head *devices = &dev_root->fs_info->fs_devices->devices;
+	struct btrfs_device *device;
+	u64 old_size;
+	u64 size_to_free;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_chunk *chunk;
+	struct btrfs_root *chunk_root = dev_root->fs_info->chunk_root;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_key found_key;
+
+	if (dev_root->fs_info->sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	mutex_lock(&dev_root->fs_info->volume_mutex);
+	dev_root = dev_root->fs_info->dev_root;
+
+	/* step one make some room on all the devices */
+	list_for_each(cur, devices) {
+		device = list_entry(cur, struct btrfs_device, dev_list);
+		old_size = device->total_bytes;
+		size_to_free = div_factor(old_size, 1);
+		size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
+		if (!device->writeable ||
+		    device->total_bytes - device->bytes_used > size_to_free)
+			continue;
+
+		ret = btrfs_shrink_device(device, old_size - size_to_free);
+		BUG_ON(ret);
+
+		trans = btrfs_start_transaction(dev_root, 1);
+		BUG_ON(!trans);
+
+		ret = btrfs_grow_device(trans, device, old_size);
+		BUG_ON(ret);
+
+		btrfs_end_transaction(trans, dev_root);
+	}
+
+	/* step two, relocate all the chunks */
+	path = btrfs_alloc_path();
+	BUG_ON(!path);
+
+	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+	key.offset = (u64)-1;
+	key.type = BTRFS_CHUNK_ITEM_KEY;
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
+		if (ret < 0)
+			goto error;
+
+		/*
+		 * this shouldn't happen, it means the last relocate
+		 * failed
+		 */
+		if (ret == 0)
+			break;
+
+		ret = btrfs_previous_item(chunk_root, path, 0,
+					  BTRFS_CHUNK_ITEM_KEY);
+		if (ret)
+			break;
+
+		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+				      path->slots[0]);
+		if (found_key.objectid != key.objectid)
+			break;
+
+		chunk = btrfs_item_ptr(path->nodes[0],
+				       path->slots[0],
+				       struct btrfs_chunk);
+		key.offset = found_key.offset;
+		/* chunk zero is special */
+		if (key.offset == 0)
+			break;
+
+		btrfs_release_path(chunk_root, path);
+		ret = btrfs_relocate_chunk(chunk_root,
+					   chunk_root->root_key.objectid,
+					   found_key.objectid,
+					   found_key.offset);
+		BUG_ON(ret);
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	mutex_unlock(&dev_root->fs_info->volume_mutex);
+	return ret;
+}
+
+/*
+ * shrinking a device means finding all of the device extents past
+ * the new size, and then following the back refs to the chunks.
+ * The chunk relocation code actually frees the device extent
+ */
+int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
+{
+	struct btrfs_trans_handle *trans;
+	struct btrfs_root *root = device->dev_root;
+	struct btrfs_dev_extent *dev_extent = NULL;
+	struct btrfs_path *path;
+	u64 length;
+	u64 chunk_tree;
+	u64 chunk_objectid;
+	u64 chunk_offset;
+	int ret;
+	int slot;
+	struct extent_buffer *l;
+	struct btrfs_key key;
+	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+	u64 old_total = btrfs_super_total_bytes(super_copy);
+	u64 diff = device->total_bytes - new_size;
+
+	if (new_size >= device->total_bytes)
+		return -EINVAL;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	trans = btrfs_start_transaction(root, 1);
+	if (!trans) {
+		ret = -ENOMEM;
+		goto done;
+	}
+
+	path->reada = 2;
+
+	lock_chunks(root);
+
+	device->total_bytes = new_size;
+	if (device->writeable)
+		device->fs_devices->total_rw_bytes -= diff;
+	ret = btrfs_update_device(trans, device);
+	if (ret) {
+		unlock_chunks(root);
+		btrfs_end_transaction(trans, root);
+		goto done;
+	}
+	WARN_ON(diff > old_total);
+	btrfs_set_super_total_bytes(super_copy, old_total - diff);
+	unlock_chunks(root);
+	btrfs_end_transaction(trans, root);
+
+	key.objectid = device->devid;
+	key.offset = (u64)-1;
+	key.type = BTRFS_DEV_EXTENT_KEY;
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0)
+			goto done;
+
+		ret = btrfs_previous_item(root, path, 0, key.type);
+		if (ret < 0)
+			goto done;
+		if (ret) {
+			ret = 0;
+			goto done;
+		}
+
+		l = path->nodes[0];
+		slot = path->slots[0];
+		btrfs_item_key_to_cpu(l, &key, path->slots[0]);
+
+		if (key.objectid != device->devid)
+			goto done;
+
+		dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
+		length = btrfs_dev_extent_length(l, dev_extent);
+
+		if (key.offset + length <= new_size)
+			goto done;
+
+		chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
+		chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
+		chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
+		btrfs_release_path(root, path);
+
+		ret = btrfs_relocate_chunk(root, chunk_tree, chunk_objectid,
+					   chunk_offset);
+		if (ret)
+			goto done;
+	}
+
+done:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root,
+			   struct btrfs_key *key,
+			   struct btrfs_chunk *chunk, int item_size)
+{
+	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+	struct btrfs_disk_key disk_key;
+	u32 array_size;
+	u8 *ptr;
+
+	array_size = btrfs_super_sys_array_size(super_copy);
+	if (array_size + item_size > BTRFS_SYSTEM_CHUNK_ARRAY_SIZE)
+		return -EFBIG;
+
+	ptr = super_copy->sys_chunk_array + array_size;
+	btrfs_cpu_key_to_disk(&disk_key, key);
+	memcpy(ptr, &disk_key, sizeof(disk_key));
+	ptr += sizeof(disk_key);
+	memcpy(ptr, chunk, item_size);
+	item_size += sizeof(disk_key);
+	btrfs_set_super_sys_array_size(super_copy, array_size + item_size);
+	return 0;
+}
+
+static noinline u64 chunk_bytes_by_type(u64 type, u64 calc_size,
+					int num_stripes, int sub_stripes)
+{
+	if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP))
+		return calc_size;
+	else if (type & BTRFS_BLOCK_GROUP_RAID10)
+		return calc_size * (num_stripes / sub_stripes);
+	else
+		return calc_size * num_stripes;
+}
+
+static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *extent_root,
+			       struct map_lookup **map_ret,
+			       u64 *num_bytes, u64 *stripe_size,
+			       u64 start, u64 type)
+{
+	struct btrfs_fs_info *info = extent_root->fs_info;
+	struct btrfs_device *device = NULL;
+	struct btrfs_fs_devices *fs_devices = info->fs_devices;
+	struct list_head *cur;
+	struct map_lookup *map = NULL;
+	struct extent_map_tree *em_tree;
+	struct extent_map *em;
+	struct list_head private_devs;
+	int min_stripe_size = 1 * 1024 * 1024;
+	u64 calc_size = 1024 * 1024 * 1024;
+	u64 max_chunk_size = calc_size;
+	u64 min_free;
+	u64 avail;
+	u64 max_avail = 0;
+	u64 dev_offset;
+	int num_stripes = 1;
+	int min_stripes = 1;
+	int sub_stripes = 0;
+	int looped = 0;
+	int ret;
+	int index;
+	int stripe_len = 64 * 1024;
+
+	if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
+	    (type & BTRFS_BLOCK_GROUP_DUP)) {
+		WARN_ON(1);
+		type &= ~BTRFS_BLOCK_GROUP_DUP;
+	}
+	if (list_empty(&fs_devices->alloc_list))
+		return -ENOSPC;
+
+	if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
+		num_stripes = fs_devices->rw_devices;
+		min_stripes = 2;
+	}
+	if (type & (BTRFS_BLOCK_GROUP_DUP)) {
+		num_stripes = 2;
+		min_stripes = 2;
+	}
+	if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
+		num_stripes = min_t(u64, 2, fs_devices->rw_devices);
+		if (num_stripes < 2)
+			return -ENOSPC;
+		min_stripes = 2;
+	}
+	if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
+		num_stripes = fs_devices->rw_devices;
+		if (num_stripes < 4)
+			return -ENOSPC;
+		num_stripes &= ~(u32)1;
+		sub_stripes = 2;
+		min_stripes = 4;
+	}
+
+	if (type & BTRFS_BLOCK_GROUP_DATA) {
+		max_chunk_size = 10 * calc_size;
+		min_stripe_size = 64 * 1024 * 1024;
+	} else if (type & BTRFS_BLOCK_GROUP_METADATA) {
+		max_chunk_size = 4 * calc_size;
+		min_stripe_size = 32 * 1024 * 1024;
+	} else if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
+		calc_size = 8 * 1024 * 1024;
+		max_chunk_size = calc_size * 2;
+		min_stripe_size = 1 * 1024 * 1024;
+	}
+
+	/* we don't want a chunk larger than 10% of writeable space */
+	max_chunk_size = min(div_factor(fs_devices->total_rw_bytes, 1),
+			     max_chunk_size);
+
+again:
+	if (!map || map->num_stripes != num_stripes) {
+		kfree(map);
+		map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+		if (!map)
+			return -ENOMEM;
+		map->num_stripes = num_stripes;
+	}
+
+	if (calc_size * num_stripes > max_chunk_size) {
+		calc_size = max_chunk_size;
+		do_div(calc_size, num_stripes);
+		do_div(calc_size, stripe_len);
+		calc_size *= stripe_len;
+	}
+	/* we don't want tiny stripes */
+	calc_size = max_t(u64, min_stripe_size, calc_size);
+
+	do_div(calc_size, stripe_len);
+	calc_size *= stripe_len;
+
+	cur = fs_devices->alloc_list.next;
+	index = 0;
+
+	if (type & BTRFS_BLOCK_GROUP_DUP)
+		min_free = calc_size * 2;
+	else
+		min_free = calc_size;
+
+	/*
+	 * we add 1MB because we never use the first 1MB of the device, unless
+	 * we've looped, then we are likely allocating the maximum amount of
+	 * space left already
+	 */
+	if (!looped)
+		min_free += 1024 * 1024;
+
+	INIT_LIST_HEAD(&private_devs);
+	while (index < num_stripes) {
+		device = list_entry(cur, struct btrfs_device, dev_alloc_list);
+		BUG_ON(!device->writeable);
+		if (device->total_bytes > device->bytes_used)
+			avail = device->total_bytes - device->bytes_used;
+		else
+			avail = 0;
+		cur = cur->next;
+
+		if (device->in_fs_metadata && avail >= min_free) {
+			ret = find_free_dev_extent(trans, device,
+						   min_free, &dev_offset);
+			if (ret == 0) {
+				list_move_tail(&device->dev_alloc_list,
+					       &private_devs);
+				map->stripes[index].dev = device;
+				map->stripes[index].physical = dev_offset;
+				index++;
+				if (type & BTRFS_BLOCK_GROUP_DUP) {
+					map->stripes[index].dev = device;
+					map->stripes[index].physical =
+						dev_offset + calc_size;
+					index++;
+				}
+			}
+		} else if (device->in_fs_metadata && avail > max_avail)
+			max_avail = avail;
+		if (cur == &fs_devices->alloc_list)
+			break;
+	}
+	list_splice(&private_devs, &fs_devices->alloc_list);
+	if (index < num_stripes) {
+		if (index >= min_stripes) {
+			num_stripes = index;
+			if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
+				num_stripes /= sub_stripes;
+				num_stripes *= sub_stripes;
+			}
+			looped = 1;
+			goto again;
+		}
+		if (!looped && max_avail > 0) {
+			looped = 1;
+			calc_size = max_avail;
+			goto again;
+		}
+		kfree(map);
+		return -ENOSPC;
+	}
+	map->sector_size = extent_root->sectorsize;
+	map->stripe_len = stripe_len;
+	map->io_align = stripe_len;
+	map->io_width = stripe_len;
+	map->type = type;
+	map->num_stripes = num_stripes;
+	map->sub_stripes = sub_stripes;
+
+	*map_ret = map;
+	*stripe_size = calc_size;
+	*num_bytes = chunk_bytes_by_type(type, calc_size,
+					 num_stripes, sub_stripes);
+
+	em = alloc_extent_map(GFP_NOFS);
+	if (!em) {
+		kfree(map);
+		return -ENOMEM;
+	}
+	em->bdev = (struct block_device *)map;
+	em->start = start;
+	em->len = *num_bytes;
+	em->block_start = 0;
+	em->block_len = em->len;
+
+	em_tree = &extent_root->fs_info->mapping_tree.map_tree;
+	spin_lock(&em_tree->lock);
+	ret = add_extent_mapping(em_tree, em);
+	spin_unlock(&em_tree->lock);
+	BUG_ON(ret);
+	free_extent_map(em);
+
+	ret = btrfs_make_block_group(trans, extent_root, 0, type,
+				     BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+				     start, *num_bytes);
+	BUG_ON(ret);
+
+	index = 0;
+	while (index < map->num_stripes) {
+		device = map->stripes[index].dev;
+		dev_offset = map->stripes[index].physical;
+
+		ret = btrfs_alloc_dev_extent(trans, device,
+				info->chunk_root->root_key.objectid,
+				BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+				start, dev_offset, calc_size);
+		BUG_ON(ret);
+		index++;
+	}
+
+	return 0;
+}
+
+static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
+				struct btrfs_root *extent_root,
+				struct map_lookup *map, u64 chunk_offset,
+				u64 chunk_size, u64 stripe_size)
+{
+	u64 dev_offset;
+	struct btrfs_key key;
+	struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
+	struct btrfs_device *device;
+	struct btrfs_chunk *chunk;
+	struct btrfs_stripe *stripe;
+	size_t item_size = btrfs_chunk_item_size(map->num_stripes);
+	int index = 0;
+	int ret;
+
+	chunk = kzalloc(item_size, GFP_NOFS);
+	if (!chunk)
+		return -ENOMEM;
+
+	index = 0;
+	while (index < map->num_stripes) {
+		device = map->stripes[index].dev;
+		device->bytes_used += stripe_size;
+		ret = btrfs_update_device(trans, device);
+		BUG_ON(ret);
+		index++;
+	}
+
+	index = 0;
+	stripe = &chunk->stripe;
+	while (index < map->num_stripes) {
+		device = map->stripes[index].dev;
+		dev_offset = map->stripes[index].physical;
+
+		btrfs_set_stack_stripe_devid(stripe, device->devid);
+		btrfs_set_stack_stripe_offset(stripe, dev_offset);
+		memcpy(stripe->dev_uuid, device->uuid, BTRFS_UUID_SIZE);
+		stripe++;
+		index++;
+	}
+
+	btrfs_set_stack_chunk_length(chunk, chunk_size);
+	btrfs_set_stack_chunk_owner(chunk, extent_root->root_key.objectid);
+	btrfs_set_stack_chunk_stripe_len(chunk, map->stripe_len);
+	btrfs_set_stack_chunk_type(chunk, map->type);
+	btrfs_set_stack_chunk_num_stripes(chunk, map->num_stripes);
+	btrfs_set_stack_chunk_io_align(chunk, map->stripe_len);
+	btrfs_set_stack_chunk_io_width(chunk, map->stripe_len);
+	btrfs_set_stack_chunk_sector_size(chunk, extent_root->sectorsize);
+	btrfs_set_stack_chunk_sub_stripes(chunk, map->sub_stripes);
+
+	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+	key.type = BTRFS_CHUNK_ITEM_KEY;
+	key.offset = chunk_offset;
+
+	ret = btrfs_insert_item(trans, chunk_root, &key, chunk, item_size);
+	BUG_ON(ret);
+
+	if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
+		ret = btrfs_add_system_chunk(trans, chunk_root, &key, chunk,
+					     item_size);
+		BUG_ON(ret);
+	}
+	kfree(chunk);
+	return 0;
+}
+
+/*
+ * Chunk allocation falls into two parts. The first part does works
+ * that make the new allocated chunk useable, but not do any operation
+ * that modifies the chunk tree. The second part does the works that
+ * require modifying the chunk tree. This division is important for the
+ * bootstrap process of adding storage to a seed btrfs.
+ */
+int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *extent_root, u64 type)
+{
+	u64 chunk_offset;
+	u64 chunk_size;
+	u64 stripe_size;
+	struct map_lookup *map;
+	struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
+	int ret;
+
+	ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
+			      &chunk_offset);
+	if (ret)
+		return ret;
+
+	ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
+				  &stripe_size, chunk_offset, type);
+	if (ret)
+		return ret;
+
+	ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
+				   chunk_size, stripe_size);
+	BUG_ON(ret);
+	return 0;
+}
+
+static noinline int init_first_rw_device(struct btrfs_trans_handle *trans,
+					 struct btrfs_root *root,
+					 struct btrfs_device *device)
+{
+	u64 chunk_offset;
+	u64 sys_chunk_offset;
+	u64 chunk_size;
+	u64 sys_chunk_size;
+	u64 stripe_size;
+	u64 sys_stripe_size;
+	u64 alloc_profile;
+	struct map_lookup *map;
+	struct map_lookup *sys_map;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_root *extent_root = fs_info->extent_root;
+	int ret;
+
+	ret = find_next_chunk(fs_info->chunk_root,
+			      BTRFS_FIRST_CHUNK_TREE_OBJECTID, &chunk_offset);
+	BUG_ON(ret);
+
+	alloc_profile = BTRFS_BLOCK_GROUP_METADATA |
+			(fs_info->metadata_alloc_profile &
+			 fs_info->avail_metadata_alloc_bits);
+	alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile);
+
+	ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
+				  &stripe_size, chunk_offset, alloc_profile);
+	BUG_ON(ret);
+
+	sys_chunk_offset = chunk_offset + chunk_size;
+
+	alloc_profile = BTRFS_BLOCK_GROUP_SYSTEM |
+			(fs_info->system_alloc_profile &
+			 fs_info->avail_system_alloc_bits);
+	alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile);
+
+	ret = __btrfs_alloc_chunk(trans, extent_root, &sys_map,
+				  &sys_chunk_size, &sys_stripe_size,
+				  sys_chunk_offset, alloc_profile);
+	BUG_ON(ret);
+
+	ret = btrfs_add_device(trans, fs_info->chunk_root, device);
+	BUG_ON(ret);
+
+	/*
+	 * Modifying chunk tree needs allocating new blocks from both
+	 * system block group and metadata block group. So we only can
+	 * do operations require modifying the chunk tree after both
+	 * block groups were created.
+	 */
+	ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
+				   chunk_size, stripe_size);
+	BUG_ON(ret);
+
+	ret = __finish_chunk_alloc(trans, extent_root, sys_map,
+				   sys_chunk_offset, sys_chunk_size,
+				   sys_stripe_size);
+	BUG_ON(ret);
+	return 0;
+}
+
+int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset)
+{
+	struct extent_map *em;
+	struct map_lookup *map;
+	struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
+	int readonly = 0;
+	int i;
+
+	spin_lock(&map_tree->map_tree.lock);
+	em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1);
+	spin_unlock(&map_tree->map_tree.lock);
+	if (!em)
+		return 1;
+
+	map = (struct map_lookup *)em->bdev;
+	for (i = 0; i < map->num_stripes; i++) {
+		if (!map->stripes[i].dev->writeable) {
+			readonly = 1;
+			break;
+		}
+	}
+	free_extent_map(em);
+	return readonly;
+}
+
+void btrfs_mapping_init(struct btrfs_mapping_tree *tree)
+{
+	extent_map_tree_init(&tree->map_tree, GFP_NOFS);
+}
+
+void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree)
+{
+	struct extent_map *em;
+
+	while (1) {
+		spin_lock(&tree->map_tree.lock);
+		em = lookup_extent_mapping(&tree->map_tree, 0, (u64)-1);
+		if (em)
+			remove_extent_mapping(&tree->map_tree, em);
+		spin_unlock(&tree->map_tree.lock);
+		if (!em)
+			break;
+		kfree(em->bdev);
+		/* once for us */
+		free_extent_map(em);
+		/* once for the tree */
+		free_extent_map(em);
+	}
+}
+
+int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len)
+{
+	struct extent_map *em;
+	struct map_lookup *map;
+	struct extent_map_tree *em_tree = &map_tree->map_tree;
+	int ret;
+
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, logical, len);
+	spin_unlock(&em_tree->lock);
+	BUG_ON(!em);
+
+	BUG_ON(em->start > logical || em->start + em->len < logical);
+	map = (struct map_lookup *)em->bdev;
+	if (map->type & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1))
+		ret = map->num_stripes;
+	else if (map->type & BTRFS_BLOCK_GROUP_RAID10)
+		ret = map->sub_stripes;
+	else
+		ret = 1;
+	free_extent_map(em);
+	return ret;
+}
+
+static int find_live_mirror(struct map_lookup *map, int first, int num,
+			    int optimal)
+{
+	int i;
+	if (map->stripes[optimal].dev->bdev)
+		return optimal;
+	for (i = first; i < first + num; i++) {
+		if (map->stripes[i].dev->bdev)
+			return i;
+	}
+	/* we couldn't find one that doesn't fail.  Just return something
+	 * and the io error handling code will clean up eventually
+	 */
+	return optimal;
+}
+
+static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
+			     u64 logical, u64 *length,
+			     struct btrfs_multi_bio **multi_ret,
+			     int mirror_num, struct page *unplug_page)
+{
+	struct extent_map *em;
+	struct map_lookup *map;
+	struct extent_map_tree *em_tree = &map_tree->map_tree;
+	u64 offset;
+	u64 stripe_offset;
+	u64 stripe_nr;
+	int stripes_allocated = 8;
+	int stripes_required = 1;
+	int stripe_index;
+	int i;
+	int num_stripes;
+	int max_errors = 0;
+	struct btrfs_multi_bio *multi = NULL;
+
+	if (multi_ret && !(rw & (1 << BIO_RW)))
+		stripes_allocated = 1;
+again:
+	if (multi_ret) {
+		multi = kzalloc(btrfs_multi_bio_size(stripes_allocated),
+				GFP_NOFS);
+		if (!multi)
+			return -ENOMEM;
+
+		atomic_set(&multi->error, 0);
+	}
+
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, logical, *length);
+	spin_unlock(&em_tree->lock);
+
+	if (!em && unplug_page)
+		return 0;
+
+	if (!em) {
+		printk(KERN_CRIT "unable to find logical %llu len %llu\n",
+		       (unsigned long long)logical,
+		       (unsigned long long)*length);
+		BUG();
+	}
+
+	BUG_ON(em->start > logical || em->start + em->len < logical);
+	map = (struct map_lookup *)em->bdev;
+	offset = logical - em->start;
+
+	if (mirror_num > map->num_stripes)
+		mirror_num = 0;
+
+	/* if our multi bio struct is too small, back off and try again */
+	if (rw & (1 << BIO_RW)) {
+		if (map->type & (BTRFS_BLOCK_GROUP_RAID1 |
+				 BTRFS_BLOCK_GROUP_DUP)) {
+			stripes_required = map->num_stripes;
+			max_errors = 1;
+		} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
+			stripes_required = map->sub_stripes;
+			max_errors = 1;
+		}
+	}
+	if (multi_ret && rw == WRITE &&
+	    stripes_allocated < stripes_required) {
+		stripes_allocated = map->num_stripes;
+		free_extent_map(em);
+		kfree(multi);
+		goto again;
+	}
+	stripe_nr = offset;
+	/*
+	 * stripe_nr counts the total number of stripes we have to stride
+	 * to get to this block
+	 */
+	do_div(stripe_nr, map->stripe_len);
+
+	stripe_offset = stripe_nr * map->stripe_len;
+	BUG_ON(offset < stripe_offset);
+
+	/* stripe_offset is the offset of this block in its stripe*/
+	stripe_offset = offset - stripe_offset;
+
+	if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 |
+			 BTRFS_BLOCK_GROUP_RAID10 |
+			 BTRFS_BLOCK_GROUP_DUP)) {
+		/* we limit the length of each bio to what fits in a stripe */
+		*length = min_t(u64, em->len - offset,
+			      map->stripe_len - stripe_offset);
+	} else {
+		*length = em->len - offset;
+	}
+
+	if (!multi_ret && !unplug_page)
+		goto out;
+
+	num_stripes = 1;
+	stripe_index = 0;
+	if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
+		if (unplug_page || (rw & (1 << BIO_RW)))
+			num_stripes = map->num_stripes;
+		else if (mirror_num)
+			stripe_index = mirror_num - 1;
+		else {
+			stripe_index = find_live_mirror(map, 0,
+					    map->num_stripes,
+					    current->pid % map->num_stripes);
+		}
+
+	} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
+		if (rw & (1 << BIO_RW))
+			num_stripes = map->num_stripes;
+		else if (mirror_num)
+			stripe_index = mirror_num - 1;
+
+	} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
+		int factor = map->num_stripes / map->sub_stripes;
+
+		stripe_index = do_div(stripe_nr, factor);
+		stripe_index *= map->sub_stripes;
+
+		if (unplug_page || (rw & (1 << BIO_RW)))
+			num_stripes = map->sub_stripes;
+		else if (mirror_num)
+			stripe_index += mirror_num - 1;
+		else {
+			stripe_index = find_live_mirror(map, stripe_index,
+					      map->sub_stripes, stripe_index +
+					      current->pid % map->sub_stripes);
+		}
+	} else {
+		/*
+		 * after this do_div call, stripe_nr is the number of stripes
+		 * on this device we have to walk to find the data, and
+		 * stripe_index is the number of our device in the stripe array
+		 */
+		stripe_index = do_div(stripe_nr, map->num_stripes);
+	}
+	BUG_ON(stripe_index >= map->num_stripes);
+
+	for (i = 0; i < num_stripes; i++) {
+		if (unplug_page) {
+			struct btrfs_device *device;
+			struct backing_dev_info *bdi;
+
+			device = map->stripes[stripe_index].dev;
+			if (device->bdev) {
+				bdi = blk_get_backing_dev_info(device->bdev);
+				if (bdi->unplug_io_fn)
+					bdi->unplug_io_fn(bdi, unplug_page);
+			}
+		} else {
+			multi->stripes[i].physical =
+				map->stripes[stripe_index].physical +
+				stripe_offset + stripe_nr * map->stripe_len;
+			multi->stripes[i].dev = map->stripes[stripe_index].dev;
+		}
+		stripe_index++;
+	}
+	if (multi_ret) {
+		*multi_ret = multi;
+		multi->num_stripes = num_stripes;
+		multi->max_errors = max_errors;
+	}
+out:
+	free_extent_map(em);
+	return 0;
+}
+
+int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
+		      u64 logical, u64 *length,
+		      struct btrfs_multi_bio **multi_ret, int mirror_num)
+{
+	return __btrfs_map_block(map_tree, rw, logical, length, multi_ret,
+				 mirror_num, NULL);
+}
+
+int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
+		     u64 chunk_start, u64 physical, u64 devid,
+		     u64 **logical, int *naddrs, int *stripe_len)
+{
+	struct extent_map_tree *em_tree = &map_tree->map_tree;
+	struct extent_map *em;
+	struct map_lookup *map;
+	u64 *buf;
+	u64 bytenr;
+	u64 length;
+	u64 stripe_nr;
+	int i, j, nr = 0;
+
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, chunk_start, 1);
+	spin_unlock(&em_tree->lock);
+
+	BUG_ON(!em || em->start != chunk_start);
+	map = (struct map_lookup *)em->bdev;
+
+	length = em->len;
+	if (map->type & BTRFS_BLOCK_GROUP_RAID10)
+		do_div(length, map->num_stripes / map->sub_stripes);
+	else if (map->type & BTRFS_BLOCK_GROUP_RAID0)
+		do_div(length, map->num_stripes);
+
+	buf = kzalloc(sizeof(u64) * map->num_stripes, GFP_NOFS);
+	BUG_ON(!buf);
+
+	for (i = 0; i < map->num_stripes; i++) {
+		if (devid && map->stripes[i].dev->devid != devid)
+			continue;
+		if (map->stripes[i].physical > physical ||
+		    map->stripes[i].physical + length <= physical)
+			continue;
+
+		stripe_nr = physical - map->stripes[i].physical;
+		do_div(stripe_nr, map->stripe_len);
+
+		if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
+			stripe_nr = stripe_nr * map->num_stripes + i;
+			do_div(stripe_nr, map->sub_stripes);
+		} else if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
+			stripe_nr = stripe_nr * map->num_stripes + i;
+		}
+		bytenr = chunk_start + stripe_nr * map->stripe_len;
+		WARN_ON(nr >= map->num_stripes);
+		for (j = 0; j < nr; j++) {
+			if (buf[j] == bytenr)
+				break;
+		}
+		if (j == nr) {
+			WARN_ON(nr >= map->num_stripes);
+			buf[nr++] = bytenr;
+		}
+	}
+
+	for (i = 0; i > nr; i++) {
+		struct btrfs_multi_bio *multi;
+		struct btrfs_bio_stripe *stripe;
+		int ret;
+
+		length = 1;
+		ret = btrfs_map_block(map_tree, WRITE, buf[i],
+				      &length, &multi, 0);
+		BUG_ON(ret);
+
+		stripe = multi->stripes;
+		for (j = 0; j < multi->num_stripes; j++) {
+			if (stripe->physical >= physical &&
+			    physical < stripe->physical + length)
+				break;
+		}
+		BUG_ON(j >= multi->num_stripes);
+		kfree(multi);
+	}
+
+	*logical = buf;
+	*naddrs = nr;
+	*stripe_len = map->stripe_len;
+
+	free_extent_map(em);
+	return 0;
+}
+
+int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree,
+		      u64 logical, struct page *page)
+{
+	u64 length = PAGE_CACHE_SIZE;
+	return __btrfs_map_block(map_tree, READ, logical, &length,
+				 NULL, 0, page);
+}
+
+static void end_bio_multi_stripe(struct bio *bio, int err)
+{
+	struct btrfs_multi_bio *multi = bio->bi_private;
+	int is_orig_bio = 0;
+
+	if (err)
+		atomic_inc(&multi->error);
+
+	if (bio == multi->orig_bio)
+		is_orig_bio = 1;
+
+	if (atomic_dec_and_test(&multi->stripes_pending)) {
+		if (!is_orig_bio) {
+			bio_put(bio);
+			bio = multi->orig_bio;
+		}
+		bio->bi_private = multi->private;
+		bio->bi_end_io = multi->end_io;
+		/* only send an error to the higher layers if it is
+		 * beyond the tolerance of the multi-bio
+		 */
+		if (atomic_read(&multi->error) > multi->max_errors) {
+			err = -EIO;
+		} else if (err) {
+			/*
+			 * this bio is actually up to date, we didn't
+			 * go over the max number of errors
+			 */
+			set_bit(BIO_UPTODATE, &bio->bi_flags);
+			err = 0;
+		}
+		kfree(multi);
+
+		bio_endio(bio, err);
+	} else if (!is_orig_bio) {
+		bio_put(bio);
+	}
+}
+
+struct async_sched {
+	struct bio *bio;
+	int rw;
+	struct btrfs_fs_info *info;
+	struct btrfs_work work;
+};
+
+/*
+ * see run_scheduled_bios for a description of why bios are collected for
+ * async submit.
+ *
+ * This will add one bio to the pending list for a device and make sure
+ * the work struct is scheduled.
+ */
+static noinline int schedule_bio(struct btrfs_root *root,
+				 struct btrfs_device *device,
+				 int rw, struct bio *bio)
+{
+	int should_queue = 1;
+
+	/* don't bother with additional async steps for reads, right now */
+	if (!(rw & (1 << BIO_RW))) {
+		bio_get(bio);
+		submit_bio(rw, bio);
+		bio_put(bio);
+		return 0;
+	}
+
+	/*
+	 * nr_async_bios allows us to reliably return congestion to the
+	 * higher layers.  Otherwise, the async bio makes it appear we have
+	 * made progress against dirty pages when we've really just put it
+	 * on a queue for later
+	 */
+	atomic_inc(&root->fs_info->nr_async_bios);
+	WARN_ON(bio->bi_next);
+	bio->bi_next = NULL;
+	bio->bi_rw |= rw;
+
+	spin_lock(&device->io_lock);
+
+	if (device->pending_bio_tail)
+		device->pending_bio_tail->bi_next = bio;
+
+	device->pending_bio_tail = bio;
+	if (!device->pending_bios)
+		device->pending_bios = bio;
+	if (device->running_pending)
+		should_queue = 0;
+
+	spin_unlock(&device->io_lock);
+
+	if (should_queue)
+		btrfs_queue_worker(&root->fs_info->submit_workers,
+				   &device->work);
+	return 0;
+}
+
+int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
+		  int mirror_num, int async_submit)
+{
+	struct btrfs_mapping_tree *map_tree;
+	struct btrfs_device *dev;
+	struct bio *first_bio = bio;
+	u64 logical = (u64)bio->bi_sector << 9;
+	u64 length = 0;
+	u64 map_length;
+	struct btrfs_multi_bio *multi = NULL;
+	int ret;
+	int dev_nr = 0;
+	int total_devs = 1;
+
+	length = bio->bi_size;
+	map_tree = &root->fs_info->mapping_tree;
+	map_length = length;
+
+	ret = btrfs_map_block(map_tree, rw, logical, &map_length, &multi,
+			      mirror_num);
+	BUG_ON(ret);
+
+	total_devs = multi->num_stripes;
+	if (map_length < length) {
+		printk(KERN_CRIT "mapping failed logical %llu bio len %llu "
+		       "len %llu\n", (unsigned long long)logical,
+		       (unsigned long long)length,
+		       (unsigned long long)map_length);
+		BUG();
+	}
+	multi->end_io = first_bio->bi_end_io;
+	multi->private = first_bio->bi_private;
+	multi->orig_bio = first_bio;
+	atomic_set(&multi->stripes_pending, multi->num_stripes);
+
+	while (dev_nr < total_devs) {
+		if (total_devs > 1) {
+			if (dev_nr < total_devs - 1) {
+				bio = bio_clone(first_bio, GFP_NOFS);
+				BUG_ON(!bio);
+			} else {
+				bio = first_bio;
+			}
+			bio->bi_private = multi;
+			bio->bi_end_io = end_bio_multi_stripe;
+		}
+		bio->bi_sector = multi->stripes[dev_nr].physical >> 9;
+		dev = multi->stripes[dev_nr].dev;
+		BUG_ON(rw == WRITE && !dev->writeable);
+		if (dev && dev->bdev) {
+			bio->bi_bdev = dev->bdev;
+			if (async_submit)
+				schedule_bio(root, dev, rw, bio);
+			else
+				submit_bio(rw, bio);
+		} else {
+			bio->bi_bdev = root->fs_info->fs_devices->latest_bdev;
+			bio->bi_sector = logical >> 9;
+			bio_endio(bio, -EIO);
+		}
+		dev_nr++;
+	}
+	if (total_devs == 1)
+		kfree(multi);
+	return 0;
+}
+
+struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
+				       u8 *uuid, u8 *fsid)
+{
+	struct btrfs_device *device;
+	struct btrfs_fs_devices *cur_devices;
+
+	cur_devices = root->fs_info->fs_devices;
+	while (cur_devices) {
+		if (!fsid ||
+		    !memcmp(cur_devices->fsid, fsid, BTRFS_UUID_SIZE)) {
+			device = __find_device(&cur_devices->devices,
+					       devid, uuid);
+			if (device)
+				return device;
+		}
+		cur_devices = cur_devices->seed;
+	}
+	return NULL;
+}
+
+static struct btrfs_device *add_missing_dev(struct btrfs_root *root,
+					    u64 devid, u8 *dev_uuid)
+{
+	struct btrfs_device *device;
+	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
+
+	device = kzalloc(sizeof(*device), GFP_NOFS);
+	if (!device)
+		return NULL;
+	list_add(&device->dev_list,
+		 &fs_devices->devices);
+	device->barriers = 1;
+	device->dev_root = root->fs_info->dev_root;
+	device->devid = devid;
+	device->work.func = pending_bios_fn;
+	device->fs_devices = fs_devices;
+	fs_devices->num_devices++;
+	spin_lock_init(&device->io_lock);
+	INIT_LIST_HEAD(&device->dev_alloc_list);
+	memcpy(device->uuid, dev_uuid, BTRFS_UUID_SIZE);
+	return device;
+}
+
+static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
+			  struct extent_buffer *leaf,
+			  struct btrfs_chunk *chunk)
+{
+	struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
+	struct map_lookup *map;
+	struct extent_map *em;
+	u64 logical;
+	u64 length;
+	u64 devid;
+	u8 uuid[BTRFS_UUID_SIZE];
+	int num_stripes;
+	int ret;
+	int i;
+
+	logical = key->offset;
+	length = btrfs_chunk_length(leaf, chunk);
+
+	spin_lock(&map_tree->map_tree.lock);
+	em = lookup_extent_mapping(&map_tree->map_tree, logical, 1);
+	spin_unlock(&map_tree->map_tree.lock);
+
+	/* already mapped? */
+	if (em && em->start <= logical && em->start + em->len > logical) {
+		free_extent_map(em);
+		return 0;
+	} else if (em) {
+		free_extent_map(em);
+	}
+
+	map = kzalloc(sizeof(*map), GFP_NOFS);
+	if (!map)
+		return -ENOMEM;
+
+	em = alloc_extent_map(GFP_NOFS);
+	if (!em)
+		return -ENOMEM;
+	num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
+	map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+	if (!map) {
+		free_extent_map(em);
+		return -ENOMEM;
+	}
+
+	em->bdev = (struct block_device *)map;
+	em->start = logical;
+	em->len = length;
+	em->block_start = 0;
+	em->block_len = em->len;
+
+	map->num_stripes = num_stripes;
+	map->io_width = btrfs_chunk_io_width(leaf, chunk);
+	map->io_align = btrfs_chunk_io_align(leaf, chunk);
+	map->sector_size = btrfs_chunk_sector_size(leaf, chunk);
+	map->stripe_len = btrfs_chunk_stripe_len(leaf, chunk);
+	map->type = btrfs_chunk_type(leaf, chunk);
+	map->sub_stripes = btrfs_chunk_sub_stripes(leaf, chunk);
+	for (i = 0; i < num_stripes; i++) {
+		map->stripes[i].physical =
+			btrfs_stripe_offset_nr(leaf, chunk, i);
+		devid = btrfs_stripe_devid_nr(leaf, chunk, i);
+		read_extent_buffer(leaf, uuid, (unsigned long)
+				   btrfs_stripe_dev_uuid_nr(chunk, i),
+				   BTRFS_UUID_SIZE);
+		map->stripes[i].dev = btrfs_find_device(root, devid, uuid,
+							NULL);
+		if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) {
+			kfree(map);
+			free_extent_map(em);
+			return -EIO;
+		}
+		if (!map->stripes[i].dev) {
+			map->stripes[i].dev =
+				add_missing_dev(root, devid, uuid);
+			if (!map->stripes[i].dev) {
+				kfree(map);
+				free_extent_map(em);
+				return -EIO;
+			}
+		}
+		map->stripes[i].dev->in_fs_metadata = 1;
+	}
+
+	spin_lock(&map_tree->map_tree.lock);
+	ret = add_extent_mapping(&map_tree->map_tree, em);
+	spin_unlock(&map_tree->map_tree.lock);
+	BUG_ON(ret);
+	free_extent_map(em);
+
+	return 0;
+}
+
+static int fill_device_from_item(struct extent_buffer *leaf,
+				 struct btrfs_dev_item *dev_item,
+				 struct btrfs_device *device)
+{
+	unsigned long ptr;
+
+	device->devid = btrfs_device_id(leaf, dev_item);
+	device->total_bytes = btrfs_device_total_bytes(leaf, dev_item);
+	device->bytes_used = btrfs_device_bytes_used(leaf, dev_item);
+	device->type = btrfs_device_type(leaf, dev_item);
+	device->io_align = btrfs_device_io_align(leaf, dev_item);
+	device->io_width = btrfs_device_io_width(leaf, dev_item);
+	device->sector_size = btrfs_device_sector_size(leaf, dev_item);
+
+	ptr = (unsigned long)btrfs_device_uuid(dev_item);
+	read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
+
+	return 0;
+}
+
+static int open_seed_devices(struct btrfs_root *root, u8 *fsid)
+{
+	struct btrfs_fs_devices *fs_devices;
+	int ret;
+
+	mutex_lock(&uuid_mutex);
+
+	fs_devices = root->fs_info->fs_devices->seed;
+	while (fs_devices) {
+		if (!memcmp(fs_devices->fsid, fsid, BTRFS_UUID_SIZE)) {
+			ret = 0;
+			goto out;
+		}
+		fs_devices = fs_devices->seed;
+	}
+
+	fs_devices = find_fsid(fsid);
+	if (!fs_devices) {
+		ret = -ENOENT;
+		goto out;
+	}
+
+	fs_devices = clone_fs_devices(fs_devices);
+	if (IS_ERR(fs_devices)) {
+		ret = PTR_ERR(fs_devices);
+		goto out;
+	}
+
+	ret = __btrfs_open_devices(fs_devices, FMODE_READ,
+				   root->fs_info->bdev_holder);
+	if (ret)
+		goto out;
+
+	if (!fs_devices->seeding) {
+		__btrfs_close_devices(fs_devices);
+		free_fs_devices(fs_devices);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	fs_devices->seed = root->fs_info->fs_devices->seed;
+	root->fs_info->fs_devices->seed = fs_devices;
+out:
+	mutex_unlock(&uuid_mutex);
+	return ret;
+}
+
+static int read_one_dev(struct btrfs_root *root,
+			struct extent_buffer *leaf,
+			struct btrfs_dev_item *dev_item)
+{
+	struct btrfs_device *device;
+	u64 devid;
+	int ret;
+	u8 fs_uuid[BTRFS_UUID_SIZE];
+	u8 dev_uuid[BTRFS_UUID_SIZE];
+
+	devid = btrfs_device_id(leaf, dev_item);
+	read_extent_buffer(leaf, dev_uuid,
+			   (unsigned long)btrfs_device_uuid(dev_item),
+			   BTRFS_UUID_SIZE);
+	read_extent_buffer(leaf, fs_uuid,
+			   (unsigned long)btrfs_device_fsid(dev_item),
+			   BTRFS_UUID_SIZE);
+
+	if (memcmp(fs_uuid, root->fs_info->fsid, BTRFS_UUID_SIZE)) {
+		ret = open_seed_devices(root, fs_uuid);
+		if (ret && !btrfs_test_opt(root, DEGRADED))
+			return ret;
+	}
+
+	device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
+	if (!device || !device->bdev) {
+		if (!btrfs_test_opt(root, DEGRADED))
+			return -EIO;
+
+		if (!device) {
+			printk(KERN_WARNING "warning devid %llu missing\n",
+			       (unsigned long long)devid);
+			device = add_missing_dev(root, devid, dev_uuid);
+			if (!device)
+				return -ENOMEM;
+		}
+	}
+
+	if (device->fs_devices != root->fs_info->fs_devices) {
+		BUG_ON(device->writeable);
+		if (device->generation !=
+		    btrfs_device_generation(leaf, dev_item))
+			return -EINVAL;
+	}
+
+	fill_device_from_item(leaf, dev_item, device);
+	device->dev_root = root->fs_info->dev_root;
+	device->in_fs_metadata = 1;
+	if (device->writeable)
+		device->fs_devices->total_rw_bytes += device->total_bytes;
+	ret = 0;
+	return ret;
+}
+
+int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf)
+{
+	struct btrfs_dev_item *dev_item;
+
+	dev_item = (struct btrfs_dev_item *)offsetof(struct btrfs_super_block,
+						     dev_item);
+	return read_one_dev(root, buf, dev_item);
+}
+
+int btrfs_read_sys_array(struct btrfs_root *root)
+{
+	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+	struct extent_buffer *sb;
+	struct btrfs_disk_key *disk_key;
+	struct btrfs_chunk *chunk;
+	u8 *ptr;
+	unsigned long sb_ptr;
+	int ret = 0;
+	u32 num_stripes;
+	u32 array_size;
+	u32 len = 0;
+	u32 cur;
+	struct btrfs_key key;
+
+	sb = btrfs_find_create_tree_block(root, BTRFS_SUPER_INFO_OFFSET,
+					  BTRFS_SUPER_INFO_SIZE);
+	if (!sb)
+		return -ENOMEM;
+	btrfs_set_buffer_uptodate(sb);
+	write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE);
+	array_size = btrfs_super_sys_array_size(super_copy);
+
+	ptr = super_copy->sys_chunk_array;
+	sb_ptr = offsetof(struct btrfs_super_block, sys_chunk_array);
+	cur = 0;
+
+	while (cur < array_size) {
+		disk_key = (struct btrfs_disk_key *)ptr;
+		btrfs_disk_key_to_cpu(&key, disk_key);
+
+		len = sizeof(*disk_key); ptr += len;
+		sb_ptr += len;
+		cur += len;
+
+		if (key.type == BTRFS_CHUNK_ITEM_KEY) {
+			chunk = (struct btrfs_chunk *)sb_ptr;
+			ret = read_one_chunk(root, &key, sb, chunk);
+			if (ret)
+				break;
+			num_stripes = btrfs_chunk_num_stripes(sb, chunk);
+			len = btrfs_chunk_item_size(num_stripes);
+		} else {
+			ret = -EIO;
+			break;
+		}
+		ptr += len;
+		sb_ptr += len;
+		cur += len;
+	}
+	free_extent_buffer(sb);
+	return ret;
+}
+
+int btrfs_read_chunk_tree(struct btrfs_root *root)
+{
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	int ret;
+	int slot;
+
+	root = root->fs_info->chunk_root;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	/* first we search for all of the device items, and then we
+	 * read in all of the chunk items.  This way we can create chunk
+	 * mappings that reference all of the devices that are afound
+	 */
+	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
+	key.offset = 0;
+	key.type = 0;
+again:
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	while (1) {
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret == 0)
+				continue;
+			if (ret < 0)
+				goto error;
+			break;
+		}
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+		if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
+			if (found_key.objectid != BTRFS_DEV_ITEMS_OBJECTID)
+				break;
+			if (found_key.type == BTRFS_DEV_ITEM_KEY) {
+				struct btrfs_dev_item *dev_item;
+				dev_item = btrfs_item_ptr(leaf, slot,
+						  struct btrfs_dev_item);
+				ret = read_one_dev(root, leaf, dev_item);
+				if (ret)
+					goto error;
+			}
+		} else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) {
+			struct btrfs_chunk *chunk;
+			chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
+			ret = read_one_chunk(root, &found_key, leaf, chunk);
+			if (ret)
+				goto error;
+		}
+		path->slots[0]++;
+	}
+	if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
+		key.objectid = 0;
+		btrfs_release_path(root, path);
+		goto again;
+	}
+	ret = 0;
+error:
+	btrfs_free_path(path);
+	return ret;
+}
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
new file mode 100644
index 00000000000..86c44e9ae11
--- /dev/null
+++ b/fs/btrfs/volumes.h
@@ -0,0 +1,162 @@
+/*
+ * Copyright (C) 2007 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __BTRFS_VOLUMES_
+#define __BTRFS_VOLUMES_
+
+#include <linux/bio.h>
+#include "async-thread.h"
+
+struct buffer_head;
+struct btrfs_device {
+	struct list_head dev_list;
+	struct list_head dev_alloc_list;
+	struct btrfs_fs_devices *fs_devices;
+	struct btrfs_root *dev_root;
+	struct bio *pending_bios;
+	struct bio *pending_bio_tail;
+	int running_pending;
+	u64 generation;
+
+	int barriers;
+	int writeable;
+	int in_fs_metadata;
+
+	spinlock_t io_lock;
+
+	struct block_device *bdev;
+
+	/* the mode sent to open_bdev_exclusive */
+	fmode_t mode;
+
+	char *name;
+
+	/* the internal btrfs device id */
+	u64 devid;
+
+	/* size of the device */
+	u64 total_bytes;
+
+	/* bytes used */
+	u64 bytes_used;
+
+	/* optimal io alignment for this device */
+	u32 io_align;
+
+	/* optimal io width for this device */
+	u32 io_width;
+
+	/* minimal io size for this device */
+	u32 sector_size;
+
+	/* type and info about this device */
+	u64 type;
+
+	/* physical drive uuid (or lvm uuid) */
+	u8 uuid[BTRFS_UUID_SIZE];
+
+	struct btrfs_work work;
+};
+
+struct btrfs_fs_devices {
+	u8 fsid[BTRFS_FSID_SIZE]; /* FS specific uuid */
+
+	/* the device with this id has the most recent coyp of the super */
+	u64 latest_devid;
+	u64 latest_trans;
+	u64 num_devices;
+	u64 open_devices;
+	u64 rw_devices;
+	u64 total_rw_bytes;
+	struct block_device *latest_bdev;
+	/* all of the devices in the FS */
+	struct list_head devices;
+
+	/* devices not currently being allocated */
+	struct list_head alloc_list;
+	struct list_head list;
+
+	struct btrfs_fs_devices *seed;
+	int seeding;
+
+	int opened;
+};
+
+struct btrfs_bio_stripe {
+	struct btrfs_device *dev;
+	u64 physical;
+};
+
+struct btrfs_multi_bio {
+	atomic_t stripes_pending;
+	bio_end_io_t *end_io;
+	struct bio *orig_bio;
+	void *private;
+	atomic_t error;
+	int max_errors;
+	int num_stripes;
+	struct btrfs_bio_stripe stripes[];
+};
+
+#define btrfs_multi_bio_size(n) (sizeof(struct btrfs_multi_bio) + \
+			    (sizeof(struct btrfs_bio_stripe) * (n)))
+
+int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
+			   struct btrfs_device *device,
+			   u64 chunk_tree, u64 chunk_objectid,
+			   u64 chunk_offset, u64 start, u64 num_bytes);
+int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
+		    u64 logical, u64 *length,
+		    struct btrfs_multi_bio **multi_ret, int mirror_num);
+int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
+		     u64 chunk_start, u64 physical, u64 devid,
+		     u64 **logical, int *naddrs, int *stripe_len);
+int btrfs_read_sys_array(struct btrfs_root *root);
+int btrfs_read_chunk_tree(struct btrfs_root *root);
+int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
+		      struct btrfs_root *extent_root, u64 type);
+void btrfs_mapping_init(struct btrfs_mapping_tree *tree);
+void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree);
+int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
+		  int mirror_num, int async_submit);
+int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf);
+int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
+		       fmode_t flags, void *holder);
+int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
+			  struct btrfs_fs_devices **fs_devices_ret);
+int btrfs_close_devices(struct btrfs_fs_devices *fs_devices);
+int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices);
+int btrfs_add_device(struct btrfs_trans_handle *trans,
+		     struct btrfs_root *root,
+		     struct btrfs_device *device);
+int btrfs_rm_device(struct btrfs_root *root, char *device_path);
+int btrfs_cleanup_fs_uuids(void);
+int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len);
+int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree,
+		      u64 logical, struct page *page);
+int btrfs_grow_device(struct btrfs_trans_handle *trans,
+		      struct btrfs_device *device, u64 new_size);
+struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
+				       u8 *uuid, u8 *fsid);
+int btrfs_shrink_device(struct btrfs_device *device, u64 new_size);
+int btrfs_init_new_device(struct btrfs_root *root, char *path);
+int btrfs_balance(struct btrfs_root *dev_root);
+void btrfs_unlock_volumes(void);
+void btrfs_lock_volumes(void);
+int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset);
+#endif
diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
new file mode 100644
index 00000000000..7f332e27089
--- /dev/null
+++ b/fs/btrfs/xattr.c
@@ -0,0 +1,322 @@
+/*
+ * Copyright (C) 2007 Red Hat.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/slab.h>
+#include <linux/rwsem.h>
+#include <linux/xattr.h>
+#include "ctree.h"
+#include "btrfs_inode.h"
+#include "transaction.h"
+#include "xattr.h"
+#include "disk-io.h"
+
+
+ssize_t __btrfs_getxattr(struct inode *inode, const char *name,
+				void *buffer, size_t size)
+{
+	struct btrfs_dir_item *di;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	int ret = 0;
+	unsigned long data_ptr;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	/* lookup the xattr by name */
+	di = btrfs_lookup_xattr(NULL, root, path, inode->i_ino, name,
+				strlen(name), 0);
+	if (!di || IS_ERR(di)) {
+		ret = -ENODATA;
+		goto out;
+	}
+
+	leaf = path->nodes[0];
+	/* if size is 0, that means we want the size of the attr */
+	if (!size) {
+		ret = btrfs_dir_data_len(leaf, di);
+		goto out;
+	}
+
+	/* now get the data out of our dir_item */
+	if (btrfs_dir_data_len(leaf, di) > size) {
+		ret = -ERANGE;
+		goto out;
+	}
+	data_ptr = (unsigned long)((char *)(di + 1) +
+				   btrfs_dir_name_len(leaf, di));
+	read_extent_buffer(leaf, buffer, data_ptr,
+			   btrfs_dir_data_len(leaf, di));
+	ret = btrfs_dir_data_len(leaf, di);
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+int __btrfs_setxattr(struct inode *inode, const char *name,
+			    const void *value, size_t size, int flags)
+{
+	struct btrfs_dir_item *di;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_path *path;
+	int ret = 0, mod = 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	trans = btrfs_start_transaction(root, 1);
+	btrfs_set_trans_block_group(trans, inode);
+
+	/* first lets see if we already have this xattr */
+	di = btrfs_lookup_xattr(trans, root, path, inode->i_ino, name,
+				strlen(name), -1);
+	if (IS_ERR(di)) {
+		ret = PTR_ERR(di);
+		goto out;
+	}
+
+	/* ok we already have this xattr, lets remove it */
+	if (di) {
+		/* if we want create only exit */
+		if (flags & XATTR_CREATE) {
+			ret = -EEXIST;
+			goto out;
+		}
+
+		ret = btrfs_delete_one_dir_name(trans, root, path, di);
+		if (ret)
+			goto out;
+		btrfs_release_path(root, path);
+
+		/* if we don't have a value then we are removing the xattr */
+		if (!value) {
+			mod = 1;
+			goto out;
+		}
+	} else {
+		btrfs_release_path(root, path);
+
+		if (flags & XATTR_REPLACE) {
+			/* we couldn't find the attr to replace */
+			ret = -ENODATA;
+			goto out;
+		}
+	}
+
+	/* ok we have to create a completely new xattr */
+	ret = btrfs_insert_xattr_item(trans, root, name, strlen(name),
+				      value, size, inode->i_ino);
+	if (ret)
+		goto out;
+	mod = 1;
+
+out:
+	if (mod) {
+		inode->i_ctime = CURRENT_TIME;
+		ret = btrfs_update_inode(trans, root, inode);
+	}
+
+	btrfs_end_transaction(trans, root);
+	btrfs_free_path(path);
+	return ret;
+}
+
+ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
+{
+	struct btrfs_key key, found_key;
+	struct inode *inode = dentry->d_inode;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_path *path;
+	struct btrfs_item *item;
+	struct extent_buffer *leaf;
+	struct btrfs_dir_item *di;
+	int ret = 0, slot, advance;
+	size_t total_size = 0, size_left = size;
+	unsigned long name_ptr;
+	size_t name_len;
+	u32 nritems;
+
+	/*
+	 * ok we want all objects associated with this id.
+	 * NOTE: we set key.offset = 0; because we want to start with the
+	 * first xattr that we find and walk forward
+	 */
+	key.objectid = inode->i_ino;
+	btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY);
+	key.offset = 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	path->reada = 2;
+
+	/* search for our xattrs */
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto err;
+	ret = 0;
+	advance = 0;
+	while (1) {
+		leaf = path->nodes[0];
+		nritems = btrfs_header_nritems(leaf);
+		slot = path->slots[0];
+
+		/* this is where we start walking through the path */
+		if (advance || slot >= nritems) {
+			/*
+			 * if we've reached the last slot in this leaf we need
+			 * to go to the next leaf and reset everything
+			 */
+			if (slot >= nritems-1) {
+				ret = btrfs_next_leaf(root, path);
+				if (ret)
+					break;
+				leaf = path->nodes[0];
+				nritems = btrfs_header_nritems(leaf);
+				slot = path->slots[0];
+			} else {
+				/*
+				 * just walking through the slots on this leaf
+				 */
+				slot++;
+				path->slots[0]++;
+			}
+		}
+		advance = 1;
+
+		item = btrfs_item_nr(leaf, slot);
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		/* check to make sure this item is what we want */
+		if (found_key.objectid != key.objectid)
+			break;
+		if (btrfs_key_type(&found_key) != BTRFS_XATTR_ITEM_KEY)
+			break;
+
+		di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
+
+		name_len = btrfs_dir_name_len(leaf, di);
+		total_size += name_len + 1;
+
+		/* we are just looking for how big our buffer needs to be */
+		if (!size)
+			continue;
+
+		if (!buffer || (name_len + 1) > size_left) {
+			ret = -ERANGE;
+			goto err;
+		}
+
+		name_ptr = (unsigned long)(di + 1);
+		read_extent_buffer(leaf, buffer, name_ptr, name_len);
+		buffer[name_len] = '\0';
+
+		size_left -= name_len + 1;
+		buffer += name_len + 1;
+	}
+	ret = total_size;
+
+err:
+	btrfs_free_path(path);
+
+	return ret;
+}
+
+/*
+ * List of handlers for synthetic system.* attributes.  All real ondisk
+ * attributes are handled directly.
+ */
+struct xattr_handler *btrfs_xattr_handlers[] = {
+#ifdef CONFIG_FS_POSIX_ACL
+	&btrfs_xattr_acl_access_handler,
+	&btrfs_xattr_acl_default_handler,
+#endif
+	NULL,
+};
+
+/*
+ * Check if the attribute is in a supported namespace.
+ *
+ * This applied after the check for the synthetic attributes in the system
+ * namespace.
+ */
+static bool btrfs_is_valid_xattr(const char *name)
+{
+	return !strncmp(name, XATTR_SECURITY_PREFIX,
+			XATTR_SECURITY_PREFIX_LEN) ||
+	       !strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN) ||
+	       !strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN) ||
+	       !strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN);
+}
+
+ssize_t btrfs_getxattr(struct dentry *dentry, const char *name,
+		       void *buffer, size_t size)
+{
+	/*
+	 * If this is a request for a synthetic attribute in the system.*
+	 * namespace use the generic infrastructure to resolve a handler
+	 * for it via sb->s_xattr.
+	 */
+	if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN))
+		return generic_getxattr(dentry, name, buffer, size);
+
+	if (!btrfs_is_valid_xattr(name))
+		return -EOPNOTSUPP;
+	return __btrfs_getxattr(dentry->d_inode, name, buffer, size);
+}
+
+int btrfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		   size_t size, int flags)
+{
+	/*
+	 * If this is a request for a synthetic attribute in the system.*
+	 * namespace use the generic infrastructure to resolve a handler
+	 * for it via sb->s_xattr.
+	 */
+	if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN))
+		return generic_setxattr(dentry, name, value, size, flags);
+
+	if (!btrfs_is_valid_xattr(name))
+		return -EOPNOTSUPP;
+
+	if (size == 0)
+		value = "";  /* empty EA, do not remove */
+	return __btrfs_setxattr(dentry->d_inode, name, value, size, flags);
+}
+
+int btrfs_removexattr(struct dentry *dentry, const char *name)
+{
+	/*
+	 * If this is a request for a synthetic attribute in the system.*
+	 * namespace use the generic infrastructure to resolve a handler
+	 * for it via sb->s_xattr.
+	 */
+	if (!strncmp(name, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN))
+		return generic_removexattr(dentry, name);
+
+	if (!btrfs_is_valid_xattr(name))
+		return -EOPNOTSUPP;
+	return __btrfs_setxattr(dentry->d_inode, name, NULL, 0, XATTR_REPLACE);
+}
diff --git a/fs/btrfs/xattr.h b/fs/btrfs/xattr.h
new file mode 100644
index 00000000000..5b1d08f8e68
--- /dev/null
+++ b/fs/btrfs/xattr.h
@@ -0,0 +1,39 @@
+/*
+ * Copyright (C) 2007 Red Hat.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __XATTR__
+#define __XATTR__
+
+#include <linux/xattr.h>
+
+extern struct xattr_handler btrfs_xattr_acl_access_handler;
+extern struct xattr_handler btrfs_xattr_acl_default_handler;
+extern struct xattr_handler *btrfs_xattr_handlers[];
+
+extern ssize_t __btrfs_getxattr(struct inode *inode, const char *name,
+		void *buffer, size_t size);
+extern int __btrfs_setxattr(struct inode *inode, const char *name,
+		const void *value, size_t size, int flags);
+
+extern ssize_t btrfs_getxattr(struct dentry *dentry, const char *name,
+		void *buffer, size_t size);
+extern int btrfs_setxattr(struct dentry *dentry, const char *name,
+		const void *value, size_t size, int flags);
+extern int btrfs_removexattr(struct dentry *dentry, const char *name);
+
+#endif /* __XATTR__ */
diff --git a/fs/btrfs/zlib.c b/fs/btrfs/zlib.c
new file mode 100644
index 00000000000..ecfbce836d3
--- /dev/null
+++ b/fs/btrfs/zlib.c
@@ -0,0 +1,632 @@
+/*
+ * Copyright (C) 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ *
+ * Based on jffs2 zlib code:
+ * Copyright © 2001-2007 Red Hat, Inc.
+ * Created by David Woodhouse <dwmw2@infradead.org>
+ */
+
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/zlib.h>
+#include <linux/zutil.h>
+#include <linux/vmalloc.h>
+#include <linux/init.h>
+#include <linux/err.h>
+#include <linux/sched.h>
+#include <linux/pagemap.h>
+#include <linux/bio.h>
+#include "compression.h"
+
+/* Plan: call deflate() with avail_in == *sourcelen,
+	avail_out = *dstlen - 12 and flush == Z_FINISH.
+	If it doesn't manage to finish,	call it again with
+	avail_in == 0 and avail_out set to the remaining 12
+	bytes for it to clean up.
+   Q: Is 12 bytes sufficient?
+*/
+#define STREAM_END_SPACE 12
+
+struct workspace {
+	z_stream inf_strm;
+	z_stream def_strm;
+	char *buf;
+	struct list_head list;
+};
+
+static LIST_HEAD(idle_workspace);
+static DEFINE_SPINLOCK(workspace_lock);
+static unsigned long num_workspace;
+static atomic_t alloc_workspace = ATOMIC_INIT(0);
+static DECLARE_WAIT_QUEUE_HEAD(workspace_wait);
+
+/*
+ * this finds an available zlib workspace or allocates a new one
+ * NULL or an ERR_PTR is returned if things go bad.
+ */
+static struct workspace *find_zlib_workspace(void)
+{
+	struct workspace *workspace;
+	int ret;
+	int cpus = num_online_cpus();
+
+again:
+	spin_lock(&workspace_lock);
+	if (!list_empty(&idle_workspace)) {
+		workspace = list_entry(idle_workspace.next, struct workspace,
+				       list);
+		list_del(&workspace->list);
+		num_workspace--;
+		spin_unlock(&workspace_lock);
+		return workspace;
+
+	}
+	spin_unlock(&workspace_lock);
+	if (atomic_read(&alloc_workspace) > cpus) {
+		DEFINE_WAIT(wait);
+		prepare_to_wait(&workspace_wait, &wait, TASK_UNINTERRUPTIBLE);
+		if (atomic_read(&alloc_workspace) > cpus)
+			schedule();
+		finish_wait(&workspace_wait, &wait);
+		goto again;
+	}
+	atomic_inc(&alloc_workspace);
+	workspace = kzalloc(sizeof(*workspace), GFP_NOFS);
+	if (!workspace) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+
+	workspace->def_strm.workspace = vmalloc(zlib_deflate_workspacesize());
+	if (!workspace->def_strm.workspace) {
+		ret = -ENOMEM;
+		goto fail;
+	}
+	workspace->inf_strm.workspace = vmalloc(zlib_inflate_workspacesize());
+	if (!workspace->inf_strm.workspace) {
+		ret = -ENOMEM;
+		goto fail_inflate;
+	}
+	workspace->buf = kmalloc(PAGE_CACHE_SIZE, GFP_NOFS);
+	if (!workspace->buf) {
+		ret = -ENOMEM;
+		goto fail_kmalloc;
+	}
+	return workspace;
+
+fail_kmalloc:
+	vfree(workspace->inf_strm.workspace);
+fail_inflate:
+	vfree(workspace->def_strm.workspace);
+fail:
+	kfree(workspace);
+	atomic_dec(&alloc_workspace);
+	wake_up(&workspace_wait);
+	return ERR_PTR(ret);
+}
+
+/*
+ * put a workspace struct back on the list or free it if we have enough
+ * idle ones sitting around
+ */
+static int free_workspace(struct workspace *workspace)
+{
+	spin_lock(&workspace_lock);
+	if (num_workspace < num_online_cpus()) {
+		list_add_tail(&workspace->list, &idle_workspace);
+		num_workspace++;
+		spin_unlock(&workspace_lock);
+		if (waitqueue_active(&workspace_wait))
+			wake_up(&workspace_wait);
+		return 0;
+	}
+	spin_unlock(&workspace_lock);
+	vfree(workspace->def_strm.workspace);
+	vfree(workspace->inf_strm.workspace);
+	kfree(workspace->buf);
+	kfree(workspace);
+
+	atomic_dec(&alloc_workspace);
+	if (waitqueue_active(&workspace_wait))
+		wake_up(&workspace_wait);
+	return 0;
+}
+
+/*
+ * cleanup function for module exit
+ */
+static void free_workspaces(void)
+{
+	struct workspace *workspace;
+	while (!list_empty(&idle_workspace)) {
+		workspace = list_entry(idle_workspace.next, struct workspace,
+				       list);
+		list_del(&workspace->list);
+		vfree(workspace->def_strm.workspace);
+		vfree(workspace->inf_strm.workspace);
+		kfree(workspace->buf);
+		kfree(workspace);
+		atomic_dec(&alloc_workspace);
+	}
+}
+
+/*
+ * given an address space and start/len, compress the bytes.
+ *
+ * pages are allocated to hold the compressed result and stored
+ * in 'pages'
+ *
+ * out_pages is used to return the number of pages allocated.  There
+ * may be pages allocated even if we return an error
+ *
+ * total_in is used to return the number of bytes actually read.  It
+ * may be smaller then len if we had to exit early because we
+ * ran out of room in the pages array or because we cross the
+ * max_out threshold.
+ *
+ * total_out is used to return the total number of compressed bytes
+ *
+ * max_out tells us the max number of bytes that we're allowed to
+ * stuff into pages
+ */
+int btrfs_zlib_compress_pages(struct address_space *mapping,
+			      u64 start, unsigned long len,
+			      struct page **pages,
+			      unsigned long nr_dest_pages,
+			      unsigned long *out_pages,
+			      unsigned long *total_in,
+			      unsigned long *total_out,
+			      unsigned long max_out)
+{
+	int ret;
+	struct workspace *workspace;
+	char *data_in;
+	char *cpage_out;
+	int nr_pages = 0;
+	struct page *in_page = NULL;
+	struct page *out_page = NULL;
+	int out_written = 0;
+	int in_read = 0;
+	unsigned long bytes_left;
+
+	*out_pages = 0;
+	*total_out = 0;
+	*total_in = 0;
+
+	workspace = find_zlib_workspace();
+	if (!workspace)
+		return -1;
+
+	if (Z_OK != zlib_deflateInit(&workspace->def_strm, 3)) {
+		printk(KERN_WARNING "deflateInit failed\n");
+		ret = -1;
+		goto out;
+	}
+
+	workspace->def_strm.total_in = 0;
+	workspace->def_strm.total_out = 0;
+
+	in_page = find_get_page(mapping, start >> PAGE_CACHE_SHIFT);
+	data_in = kmap(in_page);
+
+	out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
+	cpage_out = kmap(out_page);
+	pages[0] = out_page;
+	nr_pages = 1;
+
+	workspace->def_strm.next_in = data_in;
+	workspace->def_strm.next_out = cpage_out;
+	workspace->def_strm.avail_out = PAGE_CACHE_SIZE;
+	workspace->def_strm.avail_in = min(len, PAGE_CACHE_SIZE);
+
+	out_written = 0;
+	in_read = 0;
+
+	while (workspace->def_strm.total_in < len) {
+		ret = zlib_deflate(&workspace->def_strm, Z_SYNC_FLUSH);
+		if (ret != Z_OK) {
+			printk(KERN_DEBUG "btrfs deflate in loop returned %d\n",
+			       ret);
+			zlib_deflateEnd(&workspace->def_strm);
+			ret = -1;
+			goto out;
+		}
+
+		/* we're making it bigger, give up */
+		if (workspace->def_strm.total_in > 8192 &&
+		    workspace->def_strm.total_in <
+		    workspace->def_strm.total_out) {
+			ret = -1;
+			goto out;
+		}
+		/* we need another page for writing out.  Test this
+		 * before the total_in so we will pull in a new page for
+		 * the stream end if required
+		 */
+		if (workspace->def_strm.avail_out == 0) {
+			kunmap(out_page);
+			if (nr_pages == nr_dest_pages) {
+				out_page = NULL;
+				ret = -1;
+				goto out;
+			}
+			out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
+			cpage_out = kmap(out_page);
+			pages[nr_pages] = out_page;
+			nr_pages++;
+			workspace->def_strm.avail_out = PAGE_CACHE_SIZE;
+			workspace->def_strm.next_out = cpage_out;
+		}
+		/* we're all done */
+		if (workspace->def_strm.total_in >= len)
+			break;
+
+		/* we've read in a full page, get a new one */
+		if (workspace->def_strm.avail_in == 0) {
+			if (workspace->def_strm.total_out > max_out)
+				break;
+
+			bytes_left = len - workspace->def_strm.total_in;
+			kunmap(in_page);
+			page_cache_release(in_page);
+
+			start += PAGE_CACHE_SIZE;
+			in_page = find_get_page(mapping,
+						start >> PAGE_CACHE_SHIFT);
+			data_in = kmap(in_page);
+			workspace->def_strm.avail_in = min(bytes_left,
+							   PAGE_CACHE_SIZE);
+			workspace->def_strm.next_in = data_in;
+		}
+	}
+	workspace->def_strm.avail_in = 0;
+	ret = zlib_deflate(&workspace->def_strm, Z_FINISH);
+	zlib_deflateEnd(&workspace->def_strm);
+
+	if (ret != Z_STREAM_END) {
+		ret = -1;
+		goto out;
+	}
+
+	if (workspace->def_strm.total_out >= workspace->def_strm.total_in) {
+		ret = -1;
+		goto out;
+	}
+
+	ret = 0;
+	*total_out = workspace->def_strm.total_out;
+	*total_in = workspace->def_strm.total_in;
+out:
+	*out_pages = nr_pages;
+	if (out_page)
+		kunmap(out_page);
+
+	if (in_page) {
+		kunmap(in_page);
+		page_cache_release(in_page);
+	}
+	free_workspace(workspace);
+	return ret;
+}
+
+/*
+ * pages_in is an array of pages with compressed data.
+ *
+ * disk_start is the starting logical offset of this array in the file
+ *
+ * bvec is a bio_vec of pages from the file that we want to decompress into
+ *
+ * vcnt is the count of pages in the biovec
+ *
+ * srclen is the number of bytes in pages_in
+ *
+ * The basic idea is that we have a bio that was created by readpages.
+ * The pages in the bio are for the uncompressed data, and they may not
+ * be contiguous.  They all correspond to the range of bytes covered by
+ * the compressed extent.
+ */
+int btrfs_zlib_decompress_biovec(struct page **pages_in,
+			      u64 disk_start,
+			      struct bio_vec *bvec,
+			      int vcnt,
+			      size_t srclen)
+{
+	int ret = 0;
+	int wbits = MAX_WBITS;
+	struct workspace *workspace;
+	char *data_in;
+	size_t total_out = 0;
+	unsigned long page_bytes_left;
+	unsigned long page_in_index = 0;
+	unsigned long page_out_index = 0;
+	struct page *page_out;
+	unsigned long total_pages_in = (srclen + PAGE_CACHE_SIZE - 1) /
+					PAGE_CACHE_SIZE;
+	unsigned long buf_start;
+	unsigned long buf_offset;
+	unsigned long bytes;
+	unsigned long working_bytes;
+	unsigned long pg_offset;
+	unsigned long start_byte;
+	unsigned long current_buf_start;
+	char *kaddr;
+
+	workspace = find_zlib_workspace();
+	if (!workspace)
+		return -ENOMEM;
+
+	data_in = kmap(pages_in[page_in_index]);
+	workspace->inf_strm.next_in = data_in;
+	workspace->inf_strm.avail_in = min_t(size_t, srclen, PAGE_CACHE_SIZE);
+	workspace->inf_strm.total_in = 0;
+
+	workspace->inf_strm.total_out = 0;
+	workspace->inf_strm.next_out = workspace->buf;
+	workspace->inf_strm.avail_out = PAGE_CACHE_SIZE;
+	page_out = bvec[page_out_index].bv_page;
+	page_bytes_left = PAGE_CACHE_SIZE;
+	pg_offset = 0;
+
+	/* If it's deflate, and it's got no preset dictionary, then
+	   we can tell zlib to skip the adler32 check. */
+	if (srclen > 2 && !(data_in[1] & PRESET_DICT) &&
+	    ((data_in[0] & 0x0f) == Z_DEFLATED) &&
+	    !(((data_in[0]<<8) + data_in[1]) % 31)) {
+
+		wbits = -((data_in[0] >> 4) + 8);
+		workspace->inf_strm.next_in += 2;
+		workspace->inf_strm.avail_in -= 2;
+	}
+
+	if (Z_OK != zlib_inflateInit2(&workspace->inf_strm, wbits)) {
+		printk(KERN_WARNING "inflateInit failed\n");
+		ret = -1;
+		goto out;
+	}
+	while (workspace->inf_strm.total_in < srclen) {
+		ret = zlib_inflate(&workspace->inf_strm, Z_NO_FLUSH);
+		if (ret != Z_OK && ret != Z_STREAM_END)
+			break;
+		/*
+		 * buf start is the byte offset we're of the start of
+		 * our workspace buffer
+		 */
+		buf_start = total_out;
+
+		/* total_out is the last byte of the workspace buffer */
+		total_out = workspace->inf_strm.total_out;
+
+		working_bytes = total_out - buf_start;
+
+		/*
+		 * start byte is the first byte of the page we're currently
+		 * copying into relative to the start of the compressed data.
+		 */
+		start_byte = page_offset(page_out) - disk_start;
+
+		if (working_bytes == 0) {
+			/* we didn't make progress in this inflate
+			 * call, we're done
+			 */
+			if (ret != Z_STREAM_END)
+				ret = -1;
+			break;
+		}
+
+		/* we haven't yet hit data corresponding to this page */
+		if (total_out <= start_byte)
+			goto next;
+
+		/*
+		 * the start of the data we care about is offset into
+		 * the middle of our working buffer
+		 */
+		if (total_out > start_byte && buf_start < start_byte) {
+			buf_offset = start_byte - buf_start;
+			working_bytes -= buf_offset;
+		} else {
+			buf_offset = 0;
+		}
+		current_buf_start = buf_start;
+
+		/* copy bytes from the working buffer into the pages */
+		while (working_bytes > 0) {
+			bytes = min(PAGE_CACHE_SIZE - pg_offset,
+				    PAGE_CACHE_SIZE - buf_offset);
+			bytes = min(bytes, working_bytes);
+			kaddr = kmap_atomic(page_out, KM_USER0);
+			memcpy(kaddr + pg_offset, workspace->buf + buf_offset,
+			       bytes);
+			kunmap_atomic(kaddr, KM_USER0);
+			flush_dcache_page(page_out);
+
+			pg_offset += bytes;
+			page_bytes_left -= bytes;
+			buf_offset += bytes;
+			working_bytes -= bytes;
+			current_buf_start += bytes;
+
+			/* check if we need to pick another page */
+			if (page_bytes_left == 0) {
+				page_out_index++;
+				if (page_out_index >= vcnt) {
+					ret = 0;
+					goto done;
+				}
+
+				page_out = bvec[page_out_index].bv_page;
+				pg_offset = 0;
+				page_bytes_left = PAGE_CACHE_SIZE;
+				start_byte = page_offset(page_out) - disk_start;
+
+				/*
+				 * make sure our new page is covered by this
+				 * working buffer
+				 */
+				if (total_out <= start_byte)
+					goto next;
+
+				/* the next page in the biovec might not
+				 * be adjacent to the last page, but it
+				 * might still be found inside this working
+				 * buffer.  bump our offset pointer
+				 */
+				if (total_out > start_byte &&
+				    current_buf_start < start_byte) {
+					buf_offset = start_byte - buf_start;
+					working_bytes = total_out - start_byte;
+					current_buf_start = buf_start +
+						buf_offset;
+				}
+			}
+		}
+next:
+		workspace->inf_strm.next_out = workspace->buf;
+		workspace->inf_strm.avail_out = PAGE_CACHE_SIZE;
+
+		if (workspace->inf_strm.avail_in == 0) {
+			unsigned long tmp;
+			kunmap(pages_in[page_in_index]);
+			page_in_index++;
+			if (page_in_index >= total_pages_in) {
+				data_in = NULL;
+				break;
+			}
+			data_in = kmap(pages_in[page_in_index]);
+			workspace->inf_strm.next_in = data_in;
+			tmp = srclen - workspace->inf_strm.total_in;
+			workspace->inf_strm.avail_in = min(tmp,
+							   PAGE_CACHE_SIZE);
+		}
+	}
+	if (ret != Z_STREAM_END)
+		ret = -1;
+	else
+		ret = 0;
+done:
+	zlib_inflateEnd(&workspace->inf_strm);
+	if (data_in)
+		kunmap(pages_in[page_in_index]);
+out:
+	free_workspace(workspace);
+	return ret;
+}
+
+/*
+ * a less complex decompression routine.  Our compressed data fits in a
+ * single page, and we want to read a single page out of it.
+ * start_byte tells us the offset into the compressed data we're interested in
+ */
+int btrfs_zlib_decompress(unsigned char *data_in,
+			  struct page *dest_page,
+			  unsigned long start_byte,
+			  size_t srclen, size_t destlen)
+{
+	int ret = 0;
+	int wbits = MAX_WBITS;
+	struct workspace *workspace;
+	unsigned long bytes_left = destlen;
+	unsigned long total_out = 0;
+	char *kaddr;
+
+	if (destlen > PAGE_CACHE_SIZE)
+		return -ENOMEM;
+
+	workspace = find_zlib_workspace();
+	if (!workspace)
+		return -ENOMEM;
+
+	workspace->inf_strm.next_in = data_in;
+	workspace->inf_strm.avail_in = srclen;
+	workspace->inf_strm.total_in = 0;
+
+	workspace->inf_strm.next_out = workspace->buf;
+	workspace->inf_strm.avail_out = PAGE_CACHE_SIZE;
+	workspace->inf_strm.total_out = 0;
+	/* If it's deflate, and it's got no preset dictionary, then
+	   we can tell zlib to skip the adler32 check. */
+	if (srclen > 2 && !(data_in[1] & PRESET_DICT) &&
+	    ((data_in[0] & 0x0f) == Z_DEFLATED) &&
+	    !(((data_in[0]<<8) + data_in[1]) % 31)) {
+
+		wbits = -((data_in[0] >> 4) + 8);
+		workspace->inf_strm.next_in += 2;
+		workspace->inf_strm.avail_in -= 2;
+	}
+
+	if (Z_OK != zlib_inflateInit2(&workspace->inf_strm, wbits)) {
+		printk(KERN_WARNING "inflateInit failed\n");
+		ret = -1;
+		goto out;
+	}
+
+	while (bytes_left > 0) {
+		unsigned long buf_start;
+		unsigned long buf_offset;
+		unsigned long bytes;
+		unsigned long pg_offset = 0;
+
+		ret = zlib_inflate(&workspace->inf_strm, Z_NO_FLUSH);
+		if (ret != Z_OK && ret != Z_STREAM_END)
+			break;
+
+		buf_start = total_out;
+		total_out = workspace->inf_strm.total_out;
+
+		if (total_out == buf_start) {
+			ret = -1;
+			break;
+		}
+
+		if (total_out <= start_byte)
+			goto next;
+
+		if (total_out > start_byte && buf_start < start_byte)
+			buf_offset = start_byte - buf_start;
+		else
+			buf_offset = 0;
+
+		bytes = min(PAGE_CACHE_SIZE - pg_offset,
+			    PAGE_CACHE_SIZE - buf_offset);
+		bytes = min(bytes, bytes_left);
+
+		kaddr = kmap_atomic(dest_page, KM_USER0);
+		memcpy(kaddr + pg_offset, workspace->buf + buf_offset, bytes);
+		kunmap_atomic(kaddr, KM_USER0);
+
+		pg_offset += bytes;
+		bytes_left -= bytes;
+next:
+		workspace->inf_strm.next_out = workspace->buf;
+		workspace->inf_strm.avail_out = PAGE_CACHE_SIZE;
+	}
+
+	if (ret != Z_STREAM_END && bytes_left != 0)
+		ret = -1;
+	else
+		ret = 0;
+
+	zlib_inflateEnd(&workspace->inf_strm);
+out:
+	free_workspace(workspace);
+	return ret;
+}
+
+void btrfs_zlib_exit(void)
+{
+    free_workspaces();
+}
diff --git a/fs/buffer.c b/fs/buffer.c
index 10179cfa115..b6e8b8632e2 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -99,10 +99,18 @@ __clear_page_buffers(struct page *page)
 	page_cache_release(page);
 }
 
+
+static int quiet_error(struct buffer_head *bh)
+{
+	if (!test_bit(BH_Quiet, &bh->b_state) && printk_ratelimit())
+		return 0;
+	return 1;
+}
+
+
 static void buffer_io_error(struct buffer_head *bh)
 {
 	char b[BDEVNAME_SIZE];
-
 	printk(KERN_ERR "Buffer I/O error on device %s, logical block %Lu\n",
 			bdevname(bh->b_bdev, b),
 			(unsigned long long)bh->b_blocknr);
@@ -144,7 +152,7 @@ void end_buffer_write_sync(struct buffer_head *bh, int uptodate)
 	if (uptodate) {
 		set_buffer_uptodate(bh);
 	} else {
-		if (!buffer_eopnotsupp(bh) && printk_ratelimit()) {
+		if (!buffer_eopnotsupp(bh) && !quiet_error(bh)) {
 			buffer_io_error(bh);
 			printk(KERN_WARNING "lost page write due to "
 					"I/O error on %s\n",
@@ -195,10 +203,25 @@ int fsync_bdev(struct block_device *bdev)
  * happen on bdev until thaw_bdev() is called.
  * If a superblock is found on this device, we take the s_umount semaphore
  * on it to make sure nobody unmounts until the snapshot creation is done.
+ * The reference counter (bd_fsfreeze_count) guarantees that only the last
+ * unfreeze process can unfreeze the frozen filesystem actually when multiple
+ * freeze requests arrive simultaneously. It counts up in freeze_bdev() and
+ * count down in thaw_bdev(). When it becomes 0, thaw_bdev() will unfreeze
+ * actually.
  */
 struct super_block *freeze_bdev(struct block_device *bdev)
 {
 	struct super_block *sb;
+	int error = 0;
+
+	mutex_lock(&bdev->bd_fsfreeze_mutex);
+	if (bdev->bd_fsfreeze_count > 0) {
+		bdev->bd_fsfreeze_count++;
+		sb = get_super(bdev);
+		mutex_unlock(&bdev->bd_fsfreeze_mutex);
+		return sb;
+	}
+	bdev->bd_fsfreeze_count++;
 
 	down(&bdev->bd_mount_sem);
 	sb = get_super(bdev);
@@ -213,11 +236,24 @@ struct super_block *freeze_bdev(struct block_device *bdev)
 
 		sync_blockdev(sb->s_bdev);
 
-		if (sb->s_op->write_super_lockfs)
-			sb->s_op->write_super_lockfs(sb);
+		if (sb->s_op->freeze_fs) {
+			error = sb->s_op->freeze_fs(sb);
+			if (error) {
+				printk(KERN_ERR
+					"VFS:Filesystem freeze failed\n");
+				sb->s_frozen = SB_UNFROZEN;
+				drop_super(sb);
+				up(&bdev->bd_mount_sem);
+				bdev->bd_fsfreeze_count--;
+				mutex_unlock(&bdev->bd_fsfreeze_mutex);
+				return ERR_PTR(error);
+			}
+		}
 	}
 
 	sync_blockdev(bdev);
+	mutex_unlock(&bdev->bd_fsfreeze_mutex);
+
 	return sb;	/* thaw_bdev releases s->s_umount and bd_mount_sem */
 }
 EXPORT_SYMBOL(freeze_bdev);
@@ -229,20 +265,48 @@ EXPORT_SYMBOL(freeze_bdev);
  *
  * Unlocks the filesystem and marks it writeable again after freeze_bdev().
  */
-void thaw_bdev(struct block_device *bdev, struct super_block *sb)
+int thaw_bdev(struct block_device *bdev, struct super_block *sb)
 {
+	int error = 0;
+
+	mutex_lock(&bdev->bd_fsfreeze_mutex);
+	if (!bdev->bd_fsfreeze_count) {
+		mutex_unlock(&bdev->bd_fsfreeze_mutex);
+		return -EINVAL;
+	}
+
+	bdev->bd_fsfreeze_count--;
+	if (bdev->bd_fsfreeze_count > 0) {
+		if (sb)
+			drop_super(sb);
+		mutex_unlock(&bdev->bd_fsfreeze_mutex);
+		return 0;
+	}
+
 	if (sb) {
 		BUG_ON(sb->s_bdev != bdev);
-
-		if (sb->s_op->unlockfs)
-			sb->s_op->unlockfs(sb);
-		sb->s_frozen = SB_UNFROZEN;
-		smp_wmb();
-		wake_up(&sb->s_wait_unfrozen);
+		if (!(sb->s_flags & MS_RDONLY)) {
+			if (sb->s_op->unfreeze_fs) {
+				error = sb->s_op->unfreeze_fs(sb);
+				if (error) {
+					printk(KERN_ERR
+						"VFS:Filesystem thaw failed\n");
+					sb->s_frozen = SB_FREEZE_TRANS;
+					bdev->bd_fsfreeze_count++;
+					mutex_unlock(&bdev->bd_fsfreeze_mutex);
+					return error;
+				}
+			}
+			sb->s_frozen = SB_UNFROZEN;
+			smp_wmb();
+			wake_up(&sb->s_wait_unfrozen);
+		}
 		drop_super(sb);
 	}
 
 	up(&bdev->bd_mount_sem);
+	mutex_unlock(&bdev->bd_fsfreeze_mutex);
+	return 0;
 }
 EXPORT_SYMBOL(thaw_bdev);
 
@@ -394,7 +458,7 @@ static void end_buffer_async_read(struct buffer_head *bh, int uptodate)
 		set_buffer_uptodate(bh);
 	} else {
 		clear_buffer_uptodate(bh);
-		if (printk_ratelimit())
+		if (!quiet_error(bh))
 			buffer_io_error(bh);
 		SetPageError(page);
 	}
@@ -455,7 +519,7 @@ static void end_buffer_async_write(struct buffer_head *bh, int uptodate)
 	if (uptodate) {
 		set_buffer_uptodate(bh);
 	} else {
-		if (printk_ratelimit()) {
+		if (!quiet_error(bh)) {
 			buffer_io_error(bh);
 			printk(KERN_WARNING "lost page write due to "
 					"I/O error on %s\n",
@@ -1988,7 +2052,7 @@ int block_write_begin(struct file *file, struct address_space *mapping,
 	page = *pagep;
 	if (page == NULL) {
 		ownpage = 1;
-		page = __grab_cache_page(mapping, index);
+		page = grab_cache_page_write_begin(mapping, index, flags);
 		if (!page) {
 			status = -ENOMEM;
 			goto out;
@@ -2014,7 +2078,6 @@ int block_write_begin(struct file *file, struct address_space *mapping,
 			if (pos + len > inode->i_size)
 				vmtruncate(inode, inode->i_size);
 		}
-		goto out;
 	}
 
 out:
@@ -2494,7 +2557,7 @@ int nobh_write_begin(struct file *file, struct address_space *mapping,
 	from = pos & (PAGE_CACHE_SIZE - 1);
 	to = from + len;
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
@@ -2913,6 +2976,9 @@ static void end_bio_bh_io_sync(struct bio *bio, int err)
 		set_bit(BH_Eopnotsupp, &bh->b_state);
 	}
 
+	if (unlikely (test_bit(BIO_QUIET,&bio->bi_flags)))
+		set_bit(BH_Quiet, &bh->b_state);
+
 	bh->b_end_io(bh, test_bit(BIO_UPTODATE, &bio->bi_flags));
 	bio_put(bio);
 }
diff --git a/fs/char_dev.c b/fs/char_dev.c
index 700697a7261..38f71222a55 100644
--- a/fs/char_dev.c
+++ b/fs/char_dev.c
@@ -120,7 +120,7 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor,
 	cd->major = major;
 	cd->baseminor = baseminor;
 	cd->minorct = minorct;
-	strncpy(cd->name,name, 64);
+	strlcpy(cd->name, name, sizeof(cd->name));
 
 	i = major_to_index(major);
 
diff --git a/fs/cifs/Makefile b/fs/cifs/Makefile
index 6ba43fb346f..9948c0030e8 100644
--- a/fs/cifs/Makefile
+++ b/fs/cifs/Makefile
@@ -5,7 +5,7 @@ obj-$(CONFIG_CIFS) += cifs.o
 
 cifs-y := cifsfs.o cifssmb.o cifs_debug.o connect.o dir.o file.o inode.o \
 	  link.o misc.o netmisc.o smbdes.o smbencrypt.o transport.o asn1.o \
-	  md4.o md5.o cifs_unicode.o nterr.o xattr.o cifsencrypt.o fcntl.o \
+	  md4.o md5.o cifs_unicode.o nterr.o xattr.o cifsencrypt.o \
 	  readdir.o ioctl.o sess.o export.o cifsacl.o
 
 cifs-$(CONFIG_CIFS_UPCALL) += cifs_spnego.o
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 0005a194a75..13ea53251dc 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -747,7 +747,6 @@ const struct file_operations cifs_file_ops = {
 #endif /* CONFIG_CIFS_POSIX */
 
 #ifdef CONFIG_CIFS_EXPERIMENTAL
-	.dir_notify = cifs_dir_notify,
 	.setlease = cifs_setlease,
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
@@ -768,7 +767,6 @@ const struct file_operations cifs_file_direct_ops = {
 #endif /* CONFIG_CIFS_POSIX */
 	.llseek = cifs_llseek,
 #ifdef CONFIG_CIFS_EXPERIMENTAL
-	.dir_notify = cifs_dir_notify,
 	.setlease = cifs_setlease,
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
@@ -789,7 +787,6 @@ const struct file_operations cifs_file_nobrl_ops = {
 #endif /* CONFIG_CIFS_POSIX */
 
 #ifdef CONFIG_CIFS_EXPERIMENTAL
-	.dir_notify = cifs_dir_notify,
 	.setlease = cifs_setlease,
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
@@ -809,7 +806,6 @@ const struct file_operations cifs_file_direct_nobrl_ops = {
 #endif /* CONFIG_CIFS_POSIX */
 	.llseek = cifs_llseek,
 #ifdef CONFIG_CIFS_EXPERIMENTAL
-	.dir_notify = cifs_dir_notify,
 	.setlease = cifs_setlease,
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
@@ -818,9 +814,6 @@ const struct file_operations cifs_dir_ops = {
 	.readdir = cifs_readdir,
 	.release = cifs_closedir,
 	.read    = generic_read_dir,
-#ifdef CONFIG_CIFS_EXPERIMENTAL
-	.dir_notify = cifs_dir_notify,
-#endif /* CONFIG_CIFS_EXPERIMENTAL */
 	.unlocked_ioctl  = cifs_ioctl,
 	.llseek = generic_file_llseek,
 };
diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index 2ce04c73d74..7ac481841f8 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -76,7 +76,6 @@ extern int cifs_file_mmap(struct file * , struct vm_area_struct *);
 extern const struct file_operations cifs_dir_ops;
 extern int cifs_dir_open(struct inode *inode, struct file *file);
 extern int cifs_readdir(struct file *file, void *direntry, filldir_t filldir);
-extern int cifs_dir_notify(struct file *, unsigned long arg);
 
 /* Functions related to dir entries */
 extern struct dentry_operations cifs_dentry_ops;
diff --git a/fs/cifs/fcntl.c b/fs/cifs/fcntl.c
deleted file mode 100644
index 5a57581eb4b..00000000000
--- a/fs/cifs/fcntl.c
+++ /dev/null
@@ -1,118 +0,0 @@
-/*
- *   fs/cifs/fcntl.c
- *
- *   vfs operations that deal with the file control API
- *
- *   Copyright (C) International Business Machines  Corp., 2003,2004
- *   Author(s): Steve French (sfrench@us.ibm.com)
- *
- *   This library is free software; you can redistribute it and/or modify
- *   it under the terms of the GNU Lesser General Public License as published
- *   by the Free Software Foundation; either version 2.1 of the License, or
- *   (at your option) any later version.
- *
- *   This library is distributed in the hope that it will be useful,
- *   but WITHOUT ANY WARRANTY; without even the implied warranty of
- *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
- *   the GNU Lesser General Public License for more details.
- *
- *   You should have received a copy of the GNU Lesser General Public License
- *   along with this library; if not, write to the Free Software
- *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
- */
-#include <linux/fs.h>
-#include <linux/stat.h>
-#include <linux/fcntl.h>
-#include "cifsglob.h"
-#include "cifsproto.h"
-#include "cifs_unicode.h"
-#include "cifs_debug.h"
-#include "cifsfs.h"
-
-static __u32 convert_to_cifs_notify_flags(unsigned long fcntl_notify_flags)
-{
-	__u32 cifs_ntfy_flags = 0;
-
-	/* No way on Linux VFS to ask to monitor xattr
-	changes (and no stream support either */
-	if (fcntl_notify_flags & DN_ACCESS)
-		cifs_ntfy_flags |= FILE_NOTIFY_CHANGE_LAST_ACCESS;
-	if (fcntl_notify_flags & DN_MODIFY) {
-		/* What does this mean on directories? */
-		cifs_ntfy_flags |= FILE_NOTIFY_CHANGE_LAST_WRITE |
-			FILE_NOTIFY_CHANGE_SIZE;
-	}
-	if (fcntl_notify_flags & DN_CREATE) {
-		cifs_ntfy_flags |= FILE_NOTIFY_CHANGE_CREATION |
-			FILE_NOTIFY_CHANGE_LAST_WRITE;
-	}
-	if (fcntl_notify_flags & DN_DELETE)
-		cifs_ntfy_flags |= FILE_NOTIFY_CHANGE_LAST_WRITE;
-	if (fcntl_notify_flags & DN_RENAME) {
-		/* BB review this - checking various server behaviors */
-		cifs_ntfy_flags |= FILE_NOTIFY_CHANGE_DIR_NAME |
-			FILE_NOTIFY_CHANGE_FILE_NAME;
-	}
-	if (fcntl_notify_flags & DN_ATTRIB) {
-		cifs_ntfy_flags |= FILE_NOTIFY_CHANGE_SECURITY |
-			FILE_NOTIFY_CHANGE_ATTRIBUTES;
-	}
-/*	if (fcntl_notify_flags & DN_MULTISHOT) {
-		cifs_ntfy_flags |= ;
-	} */ /* BB fixme - not sure how to handle this with CIFS yet */
-
-	return cifs_ntfy_flags;
-}
-
-int cifs_dir_notify(struct file *file, unsigned long arg)
-{
-	int xid;
-	int rc = -EINVAL;
-	int oplock = 0;
-	struct cifs_sb_info *cifs_sb;
-	struct cifsTconInfo *pTcon;
-	char *full_path = NULL;
-	__u32 filter = FILE_NOTIFY_CHANGE_NAME | FILE_NOTIFY_CHANGE_ATTRIBUTES;
-	__u16 netfid;
-
-	if (experimEnabled == 0)
-		return 0;
-
-	xid = GetXid();
-	cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
-	pTcon = cifs_sb->tcon;
-
-	full_path = build_path_from_dentry(file->f_path.dentry);
-
-	if (full_path == NULL) {
-		rc = -ENOMEM;
-	} else {
-		cFYI(1, ("dir notify on file %s Arg 0x%lx", full_path, arg));
-		rc = CIFSSMBOpen(xid, pTcon, full_path, FILE_OPEN,
-			GENERIC_READ | SYNCHRONIZE, 0 /* create options */,
-			&netfid, &oplock, NULL, cifs_sb->local_nls,
-			cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR);
-		/* BB fixme - add this handle to a notify handle list */
-		if (rc) {
-			cFYI(1, ("Could not open directory for notify"));
-		} else {
-			filter = convert_to_cifs_notify_flags(arg);
-			if (filter != 0) {
-				rc = CIFSSMBNotify(xid, pTcon,
-					0 /* no subdirs */, netfid,
-					filter, file, arg & DN_MULTISHOT,
-					cifs_sb->local_nls);
-			} else {
-				rc = -EINVAL;
-			}
-			/* BB add code to close file eventually (at unmount
-			it would close automatically but may be a way
-			to do it easily when inode freed or when
-			notify info is cleared/changed */
-			cFYI(1, ("notify rc %d", rc));
-		}
-	}
-
-	FreeXid(xid);
-	return rc;
-}
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index b1e1fc6a6e6..12bb656fbe7 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -2074,7 +2074,7 @@ static int cifs_write_begin(struct file *file, struct address_space *mapping,
 
 	cFYI(1, ("write_begin from %lld len %d", (long long)pos, len));
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
 		rc = -ENOMEM;
 		goto out;
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index f247da9f4ed..5ab9896fdcb 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -1641,7 +1641,7 @@ do_expand:
 	i_size_write(inode, offset);
 	spin_unlock(&inode->i_lock);
 out_truncate:
-	if (inode->i_op && inode->i_op->truncate)
+	if (inode->i_op->truncate)
 		inode->i_op->truncate(inode);
 	return 0;
 out_sig:
diff --git a/fs/coda/file.c b/fs/coda/file.c
index 466303db2df..6a347fbc998 100644
--- a/fs/coda/file.c
+++ b/fs/coda/file.c
@@ -201,8 +201,7 @@ int coda_release(struct inode *coda_inode, struct file *coda_file)
 int coda_fsync(struct file *coda_file, struct dentry *coda_dentry, int datasync)
 {
 	struct file *host_file;
-	struct dentry *host_dentry;
-	struct inode *host_inode, *coda_inode = coda_dentry->d_inode;
+	struct inode *coda_inode = coda_dentry->d_inode;
 	struct coda_file_info *cfi;
 	int err = 0;
 
@@ -214,14 +213,7 @@ int coda_fsync(struct file *coda_file, struct dentry *coda_dentry, int datasync)
 	BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC);
 	host_file = cfi->cfi_container;
 
-	if (host_file->f_op && host_file->f_op->fsync) {
-		host_dentry = host_file->f_path.dentry;
-		host_inode = host_dentry->d_inode;
-		mutex_lock(&host_inode->i_mutex);
-		err = host_file->f_op->fsync(host_file, host_dentry, datasync);
-		mutex_unlock(&host_inode->i_mutex);
-	}
-
+	err = vfs_fsync(host_file, host_file->f_path.dentry, datasync);
 	if ( !err && !datasync ) {
 		lock_kernel();
 		err = venus_fsync(coda_inode->i_sb, coda_i2f(coda_inode));
diff --git a/fs/coda/sysctl.c b/fs/coda/sysctl.c
index 81b7771c646..43c96ce2961 100644
--- a/fs/coda/sysctl.c
+++ b/fs/coda/sysctl.c
@@ -11,7 +11,9 @@
 
 #include "coda_int.h"
 
+#ifdef CONFIG_SYSCTL
 static struct ctl_table_header *fs_table_header;
+#endif
 
 static ctl_table coda_table[] = {
 	{
@@ -41,6 +43,7 @@ static ctl_table coda_table[] = {
 	{}
 };
 
+#ifdef CONFIG_SYSCTL
 static ctl_table fs_table[] = {
 	{
 		.ctl_name	= CTL_UNNUMBERED,
@@ -50,7 +53,7 @@ static ctl_table fs_table[] = {
 	},
 	{}
 };
-
+#endif
 
 void coda_sysctl_init(void)
 {
diff --git a/fs/compat.c b/fs/compat.c
index d1ece79b641..30f2faa22f5 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -1187,6 +1187,9 @@ compat_sys_readv(unsigned long fd, const struct compat_iovec __user *vec, unsign
 	ret = compat_do_readv_writev(READ, file, vec, vlen, &file->f_pos);
 
 out:
+	if (ret > 0)
+		add_rchar(current, ret);
+	inc_syscr(current);
 	fput(file);
 	return ret;
 }
@@ -1210,6 +1213,9 @@ compat_sys_writev(unsigned long fd, const struct compat_iovec __user *vec, unsig
 	ret = compat_do_readv_writev(WRITE, file, vec, vlen, &file->f_pos);
 
 out:
+	if (ret > 0)
+		add_wchar(current, ret);
+	inc_syscw(current);
 	fput(file);
 	return ret;
 }
diff --git a/fs/configfs/inode.c b/fs/configfs/inode.c
index 4803ccc9448..5d349d38e05 100644
--- a/fs/configfs/inode.c
+++ b/fs/configfs/inode.c
@@ -117,8 +117,6 @@ int configfs_setattr(struct dentry * dentry, struct iattr * iattr)
 static inline void set_default_inode_attr(struct inode * inode, mode_t mode)
 {
 	inode->i_mode = mode;
-	inode->i_uid = 0;
-	inode->i_gid = 0;
 	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 }
 
@@ -136,7 +134,6 @@ struct inode * configfs_new_inode(mode_t mode, struct configfs_dirent * sd)
 {
 	struct inode * inode = new_inode(configfs_sb);
 	if (inode) {
-		inode->i_blocks = 0;
 		inode->i_mapping->a_ops = &configfs_aops;
 		inode->i_mapping->backing_dev_info = &configfs_backing_dev_info;
 		inode->i_op = &configfs_inode_operations;
diff --git a/fs/cramfs/inode.c b/fs/cramfs/inode.c
index f40423eb1a1..a07338d2d14 100644
--- a/fs/cramfs/inode.c
+++ b/fs/cramfs/inode.c
@@ -83,8 +83,6 @@ static struct inode *get_cramfs_inode(struct super_block *sb,
 			inode->i_op = &page_symlink_inode_operations;
 			inode->i_data.a_ops = &cramfs_aops;
 		} else {
-			inode->i_size = 0;
-			inode->i_blocks = 0;
 			init_special_inode(inode, inode->i_mode,
 				old_decode_dev(cramfs_inode->size));
 		}
diff --git a/fs/dcache.c b/fs/dcache.c
index a1d86c7f3e6..4547f66884a 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -34,7 +34,6 @@
 #include <linux/bootmem.h>
 #include "internal.h"
 
-
 int sysctl_vfs_cache_pressure __read_mostly = 100;
 EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure);
 
@@ -948,9 +947,6 @@ struct dentry *d_alloc(struct dentry * parent, const struct qstr *name)
 	dentry->d_op = NULL;
 	dentry->d_fsdata = NULL;
 	dentry->d_mounted = 0;
-#ifdef CONFIG_PROFILING
-	dentry->d_cookie = NULL;
-#endif
 	INIT_HLIST_NODE(&dentry->d_hash);
 	INIT_LIST_HEAD(&dentry->d_lru);
 	INIT_LIST_HEAD(&dentry->d_subdirs);
@@ -1336,7 +1332,7 @@ err_out:
  *
  * Searches the children of the parent dentry for the name in question. If
  * the dentry is found its reference count is incremented and the dentry
- * is returned. The caller must use d_put to free the entry when it has
+ * is returned. The caller must use dput to free the entry when it has
  * finished using it. %NULL is returned on failure.
  *
  * __d_lookup is dcache_lock free. The hash list is protected using RCU.
@@ -1571,10 +1567,6 @@ void d_rehash(struct dentry * entry)
 	spin_unlock(&dcache_lock);
 }
 
-#define do_switch(x,y) do { \
-	__typeof__ (x) __tmp = x; \
-	x = y; y = __tmp; } while (0)
-
 /*
  * When switching names, the actual string doesn't strictly have to
  * be preserved in the target - because we're dropping the target
@@ -1593,7 +1585,7 @@ static void switch_names(struct dentry *dentry, struct dentry *target)
 			/*
 			 * Both external: swap the pointers
 			 */
-			do_switch(target->d_name.name, dentry->d_name.name);
+			swap(target->d_name.name, dentry->d_name.name);
 		} else {
 			/*
 			 * dentry:internal, target:external.  Steal target's
@@ -1620,8 +1612,11 @@ static void switch_names(struct dentry *dentry, struct dentry *target)
 			 */
 			memcpy(dentry->d_iname, target->d_name.name,
 					target->d_name.len + 1);
+			dentry->d_name.len = target->d_name.len;
+			return;
 		}
 	}
+	swap(dentry->d_name.len, target->d_name.len);
 }
 
 /*
@@ -1681,8 +1676,7 @@ already_unhashed:
 
 	/* Switch the names.. */
 	switch_names(dentry, target);
-	do_switch(dentry->d_name.len, target->d_name.len);
-	do_switch(dentry->d_name.hash, target->d_name.hash);
+	swap(dentry->d_name.hash, target->d_name.hash);
 
 	/* ... and switch the parents */
 	if (IS_ROOT(dentry)) {
@@ -1690,7 +1684,7 @@ already_unhashed:
 		target->d_parent = target;
 		INIT_LIST_HEAD(&target->d_u.d_child);
 	} else {
-		do_switch(dentry->d_parent, target->d_parent);
+		swap(dentry->d_parent, target->d_parent);
 
 		/* And add them back to the (new) parent lists */
 		list_add(&target->d_u.d_child, &target->d_parent->d_subdirs);
@@ -1791,8 +1785,7 @@ static void __d_materialise_dentry(struct dentry *dentry, struct dentry *anon)
 	struct dentry *dparent, *aparent;
 
 	switch_names(dentry, anon);
-	do_switch(dentry->d_name.len, anon->d_name.len);
-	do_switch(dentry->d_name.hash, anon->d_name.hash);
+	swap(dentry->d_name.hash, anon->d_name.hash);
 
 	dparent = dentry->d_parent;
 	aparent = anon->d_parent;
@@ -1911,7 +1904,8 @@ static int prepend_name(char **buffer, int *buflen, struct qstr *name)
  * Convert a dentry into an ASCII path name. If the entry has been deleted
  * the string " (deleted)" is appended. Note that this is ambiguous.
  *
- * Returns the buffer or an error code if the path was too long.
+ * Returns a pointer into the buffer or an error code if the
+ * path was too long.
  *
  * "buflen" should be positive. Caller holds the dcache_lock.
  *
@@ -1987,7 +1981,10 @@ Elong:
  * Convert a dentry into an ASCII path name. If the entry has been deleted
  * the string " (deleted)" is appended. Note that this is ambiguous.
  *
- * Returns the buffer or an error code if the path was too long.
+ * Returns a pointer into the buffer or an error code if the path was
+ * too long. Note: Callers should use the returned pointer, not the passed
+ * in buffer, to use the name! The implementation often starts at an offset
+ * into the buffer, and may leave 0 bytes at the start.
  *
  * "buflen" should be positive.
  */
@@ -2313,9 +2310,6 @@ static void __init dcache_init(void)
 /* SLAB cache for __getname() consumers */
 struct kmem_cache *names_cachep __read_mostly;
 
-/* SLAB cache for file structures */
-struct kmem_cache *filp_cachep __read_mostly;
-
 EXPORT_SYMBOL(d_genocide);
 
 void __init vfs_caches_init_early(void)
@@ -2337,9 +2331,6 @@ void __init vfs_caches_init(unsigned long mempages)
 	names_cachep = kmem_cache_create("names_cache", PATH_MAX, 0,
 			SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
 
-	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
-			SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
-
 	dcache_init();
 	inode_init();
 	files_init(mempages);
diff --git a/fs/dcookies.c b/fs/dcookies.c
index 855d4b1d619..180e9fec4ad 100644
--- a/fs/dcookies.c
+++ b/fs/dcookies.c
@@ -93,10 +93,15 @@ static struct dcookie_struct *alloc_dcookie(struct path *path)
 {
 	struct dcookie_struct *dcs = kmem_cache_alloc(dcookie_cache,
 							GFP_KERNEL);
+	struct dentry *d;
 	if (!dcs)
 		return NULL;
 
-	path->dentry->d_cookie = dcs;
+	d = path->dentry;
+	spin_lock(&d->d_lock);
+	d->d_flags |= DCACHE_COOKIE;
+	spin_unlock(&d->d_lock);
+
 	dcs->path = *path;
 	path_get(path);
 	hash_dcookie(dcs);
@@ -119,14 +124,14 @@ int get_dcookie(struct path *path, unsigned long *cookie)
 		goto out;
 	}
 
-	dcs = path->dentry->d_cookie;
-
-	if (!dcs)
+	if (path->dentry->d_flags & DCACHE_COOKIE) {
+		dcs = find_dcookie((unsigned long)path->dentry);
+	} else {
 		dcs = alloc_dcookie(path);
-
-	if (!dcs) {
-		err = -ENOMEM;
-		goto out;
+		if (!dcs) {
+			err = -ENOMEM;
+			goto out;
+		}
 	}
 
 	*cookie = dcookie_value(dcs);
@@ -251,7 +256,12 @@ out_kmem:
 
 static void free_dcookie(struct dcookie_struct * dcs)
 {
-	dcs->path.dentry->d_cookie = NULL;
+	struct dentry *d = dcs->path.dentry;
+
+	spin_lock(&d->d_lock);
+	d->d_flags &= ~DCACHE_COOKIE;
+	spin_unlock(&d->d_lock);
+
 	path_put(&dcs->path);
 	kmem_cache_free(dcookie_cache, dcs);
 }
diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c
index 159a5efd6a8..33a90120f6a 100644
--- a/fs/debugfs/file.c
+++ b/fs/debugfs/file.c
@@ -294,6 +294,38 @@ struct dentry *debugfs_create_x32(const char *name, mode_t mode,
 }
 EXPORT_SYMBOL_GPL(debugfs_create_x32);
 
+
+static int debugfs_size_t_set(void *data, u64 val)
+{
+	*(size_t *)data = val;
+	return 0;
+}
+static int debugfs_size_t_get(void *data, u64 *val)
+{
+	*val = *(size_t *)data;
+	return 0;
+}
+DEFINE_SIMPLE_ATTRIBUTE(fops_size_t, debugfs_size_t_get, debugfs_size_t_set,
+			"%llu\n");	/* %llu and %zu are more or less the same */
+
+/**
+ * debugfs_create_size_t - create a debugfs file that is used to read and write an size_t value
+ * @name: a pointer to a string containing the name of the file to create.
+ * @mode: the permission that the file should have
+ * @parent: a pointer to the parent dentry for this file.  This should be a
+ *          directory dentry if set.  If this parameter is %NULL, then the
+ *          file will be created in the root of the debugfs filesystem.
+ * @value: a pointer to the variable that the file should read to and write
+ *         from.
+ */
+struct dentry *debugfs_create_size_t(const char *name, mode_t mode,
+				     struct dentry *parent, size_t *value)
+{
+	return debugfs_create_file(name, mode, parent, value, &fops_size_t);
+}
+EXPORT_SYMBOL_GPL(debugfs_create_size_t);
+
+
 static ssize_t read_file_bool(struct file *file, char __user *user_buf,
 			      size_t count, loff_t *ppos)
 {
diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c
index 3dbe2169cf3..81ae9ea3c6e 100644
--- a/fs/debugfs/inode.c
+++ b/fs/debugfs/inode.c
@@ -37,9 +37,6 @@ static struct inode *debugfs_get_inode(struct super_block *sb, int mode, dev_t d
 
 	if (inode) {
 		inode->i_mode = mode;
-		inode->i_uid = 0;
-		inode->i_gid = 0;
-		inode->i_blocks = 0;
 		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 		switch (mode & S_IFMT) {
 		default:
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index 5d61b7c06e1..5f3231b9633 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -27,25 +27,32 @@
 #define DEVPTS_SUPER_MAGIC 0x1cd1
 
 #define DEVPTS_DEFAULT_MODE 0600
+/*
+ * ptmx is a new node in /dev/pts and will be unused in legacy (single-
+ * instance) mode. To prevent surprises in user space, set permissions of
+ * ptmx to 0. Use 'chmod' or remount with '-o ptmxmode' to set meaningful
+ * permissions.
+ */
+#define DEVPTS_DEFAULT_PTMX_MODE 0000
 #define PTMX_MINOR	2
 
 extern int pty_limit;			/* Config limit on Unix98 ptys */
-static DEFINE_IDA(allocated_ptys);
 static DEFINE_MUTEX(allocated_ptys_lock);
 
 static struct vfsmount *devpts_mnt;
-static struct dentry *devpts_root;
 
-static struct {
+struct pts_mount_opts {
 	int setuid;
 	int setgid;
 	uid_t   uid;
 	gid_t   gid;
 	umode_t mode;
-} config = {.mode = DEVPTS_DEFAULT_MODE};
+	umode_t ptmxmode;
+	int newinstance;
+};
 
 enum {
-	Opt_uid, Opt_gid, Opt_mode,
+	Opt_uid, Opt_gid, Opt_mode, Opt_ptmxmode, Opt_newinstance,
 	Opt_err
 };
 
@@ -53,18 +60,50 @@ static const match_table_t tokens = {
 	{Opt_uid, "uid=%u"},
 	{Opt_gid, "gid=%u"},
 	{Opt_mode, "mode=%o"},
+#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES
+	{Opt_ptmxmode, "ptmxmode=%o"},
+	{Opt_newinstance, "newinstance"},
+#endif
 	{Opt_err, NULL}
 };
 
-static int devpts_remount(struct super_block *sb, int *flags, char *data)
+struct pts_fs_info {
+	struct ida allocated_ptys;
+	struct pts_mount_opts mount_opts;
+	struct dentry *ptmx_dentry;
+};
+
+static inline struct pts_fs_info *DEVPTS_SB(struct super_block *sb)
+{
+	return sb->s_fs_info;
+}
+
+static inline struct super_block *pts_sb_from_inode(struct inode *inode)
+{
+#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES
+	if (inode->i_sb->s_magic == DEVPTS_SUPER_MAGIC)
+		return inode->i_sb;
+#endif
+	return devpts_mnt->mnt_sb;
+}
+
+#define PARSE_MOUNT	0
+#define PARSE_REMOUNT	1
+
+static int parse_mount_options(char *data, int op, struct pts_mount_opts *opts)
 {
 	char *p;
 
-	config.setuid  = 0;
-	config.setgid  = 0;
-	config.uid     = 0;
-	config.gid     = 0;
-	config.mode    = DEVPTS_DEFAULT_MODE;
+	opts->setuid  = 0;
+	opts->setgid  = 0;
+	opts->uid     = 0;
+	opts->gid     = 0;
+	opts->mode    = DEVPTS_DEFAULT_MODE;
+	opts->ptmxmode = DEVPTS_DEFAULT_PTMX_MODE;
+
+	/* newinstance makes sense only on initial mount */
+	if (op == PARSE_MOUNT)
+		opts->newinstance = 0;
 
 	while ((p = strsep(&data, ",")) != NULL) {
 		substring_t args[MAX_OPT_ARGS];
@@ -79,20 +118,32 @@ static int devpts_remount(struct super_block *sb, int *flags, char *data)
 		case Opt_uid:
 			if (match_int(&args[0], &option))
 				return -EINVAL;
-			config.uid = option;
-			config.setuid = 1;
+			opts->uid = option;
+			opts->setuid = 1;
 			break;
 		case Opt_gid:
 			if (match_int(&args[0], &option))
 				return -EINVAL;
-			config.gid = option;
-			config.setgid = 1;
+			opts->gid = option;
+			opts->setgid = 1;
 			break;
 		case Opt_mode:
 			if (match_octal(&args[0], &option))
 				return -EINVAL;
-			config.mode = option & S_IALLUGO;
+			opts->mode = option & S_IALLUGO;
+			break;
+#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES
+		case Opt_ptmxmode:
+			if (match_octal(&args[0], &option))
+				return -EINVAL;
+			opts->ptmxmode = option & S_IALLUGO;
+			break;
+		case Opt_newinstance:
+			/* newinstance makes sense only on initial mount */
+			if (op == PARSE_MOUNT)
+				opts->newinstance = 1;
 			break;
+#endif
 		default:
 			printk(KERN_ERR "devpts: called with bogus options\n");
 			return -EINVAL;
@@ -102,13 +153,106 @@ static int devpts_remount(struct super_block *sb, int *flags, char *data)
 	return 0;
 }
 
+#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES
+static int mknod_ptmx(struct super_block *sb)
+{
+	int mode;
+	int rc = -ENOMEM;
+	struct dentry *dentry;
+	struct inode *inode;
+	struct dentry *root = sb->s_root;
+	struct pts_fs_info *fsi = DEVPTS_SB(sb);
+	struct pts_mount_opts *opts = &fsi->mount_opts;
+
+	mutex_lock(&root->d_inode->i_mutex);
+
+	/* If we have already created ptmx node, return */
+	if (fsi->ptmx_dentry) {
+		rc = 0;
+		goto out;
+	}
+
+	dentry = d_alloc_name(root, "ptmx");
+	if (!dentry) {
+		printk(KERN_NOTICE "Unable to alloc dentry for ptmx node\n");
+		goto out;
+	}
+
+	/*
+	 * Create a new 'ptmx' node in this mount of devpts.
+	 */
+	inode = new_inode(sb);
+	if (!inode) {
+		printk(KERN_ERR "Unable to alloc inode for ptmx node\n");
+		dput(dentry);
+		goto out;
+	}
+
+	inode->i_ino = 2;
+	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
+
+	mode = S_IFCHR|opts->ptmxmode;
+	init_special_inode(inode, mode, MKDEV(TTYAUX_MAJOR, 2));
+
+	d_add(dentry, inode);
+
+	fsi->ptmx_dentry = dentry;
+	rc = 0;
+
+	printk(KERN_DEBUG "Created ptmx node in devpts ino %lu\n",
+			inode->i_ino);
+out:
+	mutex_unlock(&root->d_inode->i_mutex);
+	return rc;
+}
+
+static void update_ptmx_mode(struct pts_fs_info *fsi)
+{
+	struct inode *inode;
+	if (fsi->ptmx_dentry) {
+		inode = fsi->ptmx_dentry->d_inode;
+		inode->i_mode = S_IFCHR|fsi->mount_opts.ptmxmode;
+	}
+}
+#else
+static inline void update_ptmx_mode(struct pts_fs_info *fsi)
+{
+       return;
+}
+#endif
+
+static int devpts_remount(struct super_block *sb, int *flags, char *data)
+{
+	int err;
+	struct pts_fs_info *fsi = DEVPTS_SB(sb);
+	struct pts_mount_opts *opts = &fsi->mount_opts;
+
+	err = parse_mount_options(data, PARSE_REMOUNT, opts);
+
+	/*
+	 * parse_mount_options() restores options to default values
+	 * before parsing and may have changed ptmxmode. So, update the
+	 * mode in the inode too. Bogus options don't fail the remount,
+	 * so do this even on error return.
+	 */
+	update_ptmx_mode(fsi);
+
+	return err;
+}
+
 static int devpts_show_options(struct seq_file *seq, struct vfsmount *vfs)
 {
-	if (config.setuid)
-		seq_printf(seq, ",uid=%u", config.uid);
-	if (config.setgid)
-		seq_printf(seq, ",gid=%u", config.gid);
-	seq_printf(seq, ",mode=%03o", config.mode);
+	struct pts_fs_info *fsi = DEVPTS_SB(vfs->mnt_sb);
+	struct pts_mount_opts *opts = &fsi->mount_opts;
+
+	if (opts->setuid)
+		seq_printf(seq, ",uid=%u", opts->uid);
+	if (opts->setgid)
+		seq_printf(seq, ",gid=%u", opts->gid);
+	seq_printf(seq, ",mode=%03o", opts->mode);
+#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES
+	seq_printf(seq, ",ptmxmode=%03o", opts->ptmxmode);
+#endif
 
 	return 0;
 }
@@ -119,10 +263,25 @@ static const struct super_operations devpts_sops = {
 	.show_options	= devpts_show_options,
 };
 
+static void *new_pts_fs_info(void)
+{
+	struct pts_fs_info *fsi;
+
+	fsi = kzalloc(sizeof(struct pts_fs_info), GFP_KERNEL);
+	if (!fsi)
+		return NULL;
+
+	ida_init(&fsi->allocated_ptys);
+	fsi->mount_opts.mode = DEVPTS_DEFAULT_MODE;
+	fsi->mount_opts.ptmxmode = DEVPTS_DEFAULT_PTMX_MODE;
+
+	return fsi;
+}
+
 static int
 devpts_fill_super(struct super_block *s, void *data, int silent)
 {
-	struct inode * inode;
+	struct inode *inode;
 
 	s->s_blocksize = 1024;
 	s->s_blocksize_bits = 10;
@@ -130,39 +289,240 @@ devpts_fill_super(struct super_block *s, void *data, int silent)
 	s->s_op = &devpts_sops;
 	s->s_time_gran = 1;
 
+	s->s_fs_info = new_pts_fs_info();
+	if (!s->s_fs_info)
+		goto fail;
+
 	inode = new_inode(s);
 	if (!inode)
-		goto fail;
+		goto free_fsi;
 	inode->i_ino = 1;
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
-	inode->i_blocks = 0;
-	inode->i_uid = inode->i_gid = 0;
 	inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO | S_IWUSR;
 	inode->i_op = &simple_dir_inode_operations;
 	inode->i_fop = &simple_dir_operations;
 	inode->i_nlink = 2;
 
-	devpts_root = s->s_root = d_alloc_root(inode);
+	s->s_root = d_alloc_root(inode);
 	if (s->s_root)
 		return 0;
-	
-	printk("devpts: get root dentry failed\n");
+
+	printk(KERN_ERR "devpts: get root dentry failed\n");
 	iput(inode);
+
+free_fsi:
+	kfree(s->s_fs_info);
 fail:
 	return -ENOMEM;
 }
 
+#ifdef CONFIG_DEVPTS_MULTIPLE_INSTANCES
+static int compare_init_pts_sb(struct super_block *s, void *p)
+{
+	if (devpts_mnt)
+		return devpts_mnt->mnt_sb == s;
+	return 0;
+}
+
+/*
+ * Safely parse the mount options in @data and update @opts.
+ *
+ * devpts ends up parsing options two times during mount, due to the
+ * two modes of operation it supports. The first parse occurs in
+ * devpts_get_sb() when determining the mode (single-instance or
+ * multi-instance mode). The second parse happens in devpts_remount()
+ * or new_pts_mount() depending on the mode.
+ *
+ * Parsing of options modifies the @data making subsequent parsing
+ * incorrect. So make a local copy of @data and parse it.
+ *
+ * Return: 0 On success, -errno on error
+ */
+static int safe_parse_mount_options(void *data, struct pts_mount_opts *opts)
+{
+	int rc;
+	void *datacp;
+
+	if (!data)
+		return 0;
+
+	/* Use kstrdup() ?  */
+	datacp = kmalloc(PAGE_SIZE, GFP_KERNEL);
+	if (!datacp)
+		return -ENOMEM;
+
+	memcpy(datacp, data, PAGE_SIZE);
+	rc = parse_mount_options((char *)datacp, PARSE_MOUNT, opts);
+	kfree(datacp);
+
+	return rc;
+}
+
+/*
+ * Mount a new (private) instance of devpts.  PTYs created in this
+ * instance are independent of the PTYs in other devpts instances.
+ */
+static int new_pts_mount(struct file_system_type *fs_type, int flags,
+		void *data, struct vfsmount *mnt)
+{
+	int err;
+	struct pts_fs_info *fsi;
+	struct pts_mount_opts *opts;
+
+	printk(KERN_NOTICE "devpts: newinstance mount\n");
+
+	err = get_sb_nodev(fs_type, flags, data, devpts_fill_super, mnt);
+	if (err)
+		return err;
+
+	fsi = DEVPTS_SB(mnt->mnt_sb);
+	opts = &fsi->mount_opts;
+
+	err = parse_mount_options(data, PARSE_MOUNT, opts);
+	if (err)
+		goto fail;
+
+	err = mknod_ptmx(mnt->mnt_sb);
+	if (err)
+		goto fail;
+
+	return 0;
+
+fail:
+	dput(mnt->mnt_sb->s_root);
+	deactivate_super(mnt->mnt_sb);
+	return err;
+}
+
+/*
+ * Check if 'newinstance' mount option was specified in @data.
+ *
+ * Return: -errno  	on error (eg: invalid mount options specified)
+ * 	 : 1 		if 'newinstance' mount option was specified
+ * 	 : 0 		if 'newinstance' mount option was NOT specified
+ */
+static int is_new_instance_mount(void *data)
+{
+	int rc;
+	struct pts_mount_opts opts;
+
+	if (!data)
+		return 0;
+
+	rc = safe_parse_mount_options(data, &opts);
+	if (!rc)
+		rc = opts.newinstance;
+
+	return rc;
+}
+
+/*
+ * get_init_pts_sb()
+ *
+ *     This interface is needed to support multiple namespace semantics in
+ *     devpts while preserving backward compatibility of the current 'single-
+ *     namespace' semantics. i.e all mounts of devpts without the 'newinstance'
+ *     mount option should bind to the initial kernel mount, like
+ *     get_sb_single().
+ *
+ *     Mounts with 'newinstance' option create a new private namespace.
+ *
+ *     But for single-mount semantics, devpts cannot use get_sb_single(),
+ *     because get_sb_single()/sget() find and use the super-block from
+ *     the most recent mount of devpts. But that recent mount may be a
+ *     'newinstance' mount and get_sb_single() would pick the newinstance
+ *     super-block instead of the initial super-block.
+ *
+ *     This interface is identical to get_sb_single() except that it
+ *     consistently selects the 'single-namespace' superblock even in the
+ *     presence of the private namespace (i.e 'newinstance') super-blocks.
+ */
+static int get_init_pts_sb(struct file_system_type *fs_type, int flags,
+		void *data, struct vfsmount *mnt)
+{
+	struct super_block *s;
+	int error;
+
+	s = sget(fs_type, compare_init_pts_sb, set_anon_super, NULL);
+	if (IS_ERR(s))
+		return PTR_ERR(s);
+
+	if (!s->s_root) {
+		s->s_flags = flags;
+		error = devpts_fill_super(s, data, flags & MS_SILENT ? 1 : 0);
+		if (error) {
+			up_write(&s->s_umount);
+			deactivate_super(s);
+			return error;
+		}
+		s->s_flags |= MS_ACTIVE;
+	}
+	do_remount_sb(s, flags, data, 0);
+	return simple_set_mnt(mnt, s);
+}
+
+/*
+ * Mount or remount the initial kernel mount of devpts. This type of
+ * mount maintains the legacy, single-instance semantics, while the
+ * kernel still allows multiple-instances.
+ */
+static int init_pts_mount(struct file_system_type *fs_type, int flags,
+		void *data, struct vfsmount *mnt)
+{
+	int err;
+
+	err = get_init_pts_sb(fs_type, flags, data, mnt);
+	if (err)
+		return err;
+
+	err = mknod_ptmx(mnt->mnt_sb);
+	if (err) {
+		dput(mnt->mnt_sb->s_root);
+		deactivate_super(mnt->mnt_sb);
+	}
+
+	return err;
+}
+
 static int devpts_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *data, struct vfsmount *mnt)
 {
+	int new;
+
+	new = is_new_instance_mount(data);
+	if (new < 0)
+		return new;
+
+	if (new)
+		return new_pts_mount(fs_type, flags, data, mnt);
+
+	return init_pts_mount(fs_type, flags, data, mnt);
+}
+#else
+/*
+ * This supports only the legacy single-instance semantics (no
+ * multiple-instance semantics)
+ */
+static int devpts_get_sb(struct file_system_type *fs_type, int flags,
+		const char *dev_name, void *data, struct vfsmount *mnt)
+{
 	return get_sb_single(fs_type, flags, data, devpts_fill_super, mnt);
 }
+#endif
+
+static void devpts_kill_sb(struct super_block *sb)
+{
+	struct pts_fs_info *fsi = DEVPTS_SB(sb);
+
+	kfree(fsi);
+	kill_litter_super(sb);
+}
 
 static struct file_system_type devpts_fs_type = {
 	.owner		= THIS_MODULE,
 	.name		= "devpts",
 	.get_sb		= devpts_get_sb,
-	.kill_sb	= kill_anon_super,
+	.kill_sb	= devpts_kill_sb,
 };
 
 /*
@@ -172,16 +532,17 @@ static struct file_system_type devpts_fs_type = {
 
 int devpts_new_index(struct inode *ptmx_inode)
 {
+	struct super_block *sb = pts_sb_from_inode(ptmx_inode);
+	struct pts_fs_info *fsi = DEVPTS_SB(sb);
 	int index;
 	int ida_ret;
 
 retry:
-	if (!ida_pre_get(&allocated_ptys, GFP_KERNEL)) {
+	if (!ida_pre_get(&fsi->allocated_ptys, GFP_KERNEL))
 		return -ENOMEM;
-	}
 
 	mutex_lock(&allocated_ptys_lock);
-	ida_ret = ida_get_new(&allocated_ptys, &index);
+	ida_ret = ida_get_new(&fsi->allocated_ptys, &index);
 	if (ida_ret < 0) {
 		mutex_unlock(&allocated_ptys_lock);
 		if (ida_ret == -EAGAIN)
@@ -190,7 +551,7 @@ retry:
 	}
 
 	if (index >= pty_limit) {
-		ida_remove(&allocated_ptys, index);
+		ida_remove(&fsi->allocated_ptys, index);
 		mutex_unlock(&allocated_ptys_lock);
 		return -EIO;
 	}
@@ -200,18 +561,26 @@ retry:
 
 void devpts_kill_index(struct inode *ptmx_inode, int idx)
 {
+	struct super_block *sb = pts_sb_from_inode(ptmx_inode);
+	struct pts_fs_info *fsi = DEVPTS_SB(sb);
+
 	mutex_lock(&allocated_ptys_lock);
-	ida_remove(&allocated_ptys, idx);
+	ida_remove(&fsi->allocated_ptys, idx);
 	mutex_unlock(&allocated_ptys_lock);
 }
 
 int devpts_pty_new(struct inode *ptmx_inode, struct tty_struct *tty)
 {
-	int number = tty->index; /* tty layer puts index from devpts_new_index() in here */
+	/* tty layer puts index from devpts_new_index() in here */
+	int number = tty->index;
 	struct tty_driver *driver = tty->driver;
 	dev_t device = MKDEV(driver->major, driver->minor_start+number);
 	struct dentry *dentry;
-	struct inode *inode = new_inode(devpts_mnt->mnt_sb);
+	struct super_block *sb = pts_sb_from_inode(ptmx_inode);
+	struct inode *inode = new_inode(sb);
+	struct dentry *root = sb->s_root;
+	struct pts_fs_info *fsi = DEVPTS_SB(sb);
+	struct pts_mount_opts *opts = &fsi->mount_opts;
 	char s[12];
 
 	/* We're supposed to be given the slave end of a pty */
@@ -221,25 +590,25 @@ int devpts_pty_new(struct inode *ptmx_inode, struct tty_struct *tty)
 	if (!inode)
 		return -ENOMEM;
 
-	inode->i_ino = number+2;
-	inode->i_uid = config.setuid ? config.uid : current_fsuid();
-	inode->i_gid = config.setgid ? config.gid : current_fsgid();
+	inode->i_ino = number + 3;
+	inode->i_uid = opts->setuid ? opts->uid : current_fsuid();
+	inode->i_gid = opts->setgid ? opts->gid : current_fsgid();
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
-	init_special_inode(inode, S_IFCHR|config.mode, device);
+	init_special_inode(inode, S_IFCHR|opts->mode, device);
 	inode->i_private = tty;
 	tty->driver_data = inode;
 
 	sprintf(s, "%d", number);
 
-	mutex_lock(&devpts_root->d_inode->i_mutex);
+	mutex_lock(&root->d_inode->i_mutex);
 
-	dentry = d_alloc_name(devpts_root, s);
+	dentry = d_alloc_name(root, s);
 	if (!IS_ERR(dentry)) {
 		d_add(dentry, inode);
-		fsnotify_create(devpts_root->d_inode, dentry);
+		fsnotify_create(root->d_inode, dentry);
 	}
 
-	mutex_unlock(&devpts_root->d_inode->i_mutex);
+	mutex_unlock(&root->d_inode->i_mutex);
 
 	return 0;
 }
@@ -256,20 +625,27 @@ struct tty_struct *devpts_get_tty(struct inode *pts_inode, int number)
 void devpts_pty_kill(struct tty_struct *tty)
 {
 	struct inode *inode = tty->driver_data;
+	struct super_block *sb = pts_sb_from_inode(inode);
+	struct dentry *root = sb->s_root;
 	struct dentry *dentry;
 
 	BUG_ON(inode->i_rdev == MKDEV(TTYAUX_MAJOR, PTMX_MINOR));
 
-	mutex_lock(&devpts_root->d_inode->i_mutex);
+	mutex_lock(&root->d_inode->i_mutex);
 
 	dentry = d_find_alias(inode);
-	if (dentry && !IS_ERR(dentry)) {
+	if (IS_ERR(dentry))
+		goto out;
+
+	if (dentry) {
 		inode->i_nlink--;
 		d_delete(dentry);
-		dput(dentry);
+		dput(dentry);	/* d_alloc_name() in devpts_pty_new() */
 	}
 
-	mutex_unlock(&devpts_root->d_inode->i_mutex);
+	dput(dentry);		/* d_find_alias above */
+out:
+	mutex_unlock(&root->d_inode->i_mutex);
 }
 
 static int __init init_devpts_fs(void)
diff --git a/fs/direct-io.c b/fs/direct-io.c
index af0558dbe8b..b6d43908ff7 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -1209,6 +1209,19 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 	retval = direct_io_worker(rw, iocb, inode, iov, offset,
 				nr_segs, blkbits, get_block, end_io, dio);
 
+	/*
+	 * In case of error extending write may have instantiated a few
+	 * blocks outside i_size. Trim these off again for DIO_LOCKING.
+	 * NOTE: DIO_NO_LOCK/DIO_OWN_LOCK callers have to handle this by
+	 * it's own meaner.
+	 */
+	if (unlikely(retval < 0 && (rw & WRITE))) {
+		loff_t isize = i_size_read(inode);
+
+		if (end > isize && dio_lock_type == DIO_LOCKING)
+			vmtruncate(inode, isize);
+	}
+
 	if (rw == READ && dio_lock_type == DIO_LOCKING)
 		release_i_mutex = 0;
 
diff --git a/fs/dlm/ast.c b/fs/dlm/ast.c
index 8bf31e3fbf0..dc2ad6008b2 100644
--- a/fs/dlm/ast.c
+++ b/fs/dlm/ast.c
@@ -2,7 +2,7 @@
 *******************************************************************************
 **
 **  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
-**  Copyright (C) 2004-2005 Red Hat, Inc.  All rights reserved.
+**  Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
 **
 **  This copyrighted material is made available to anyone wishing to use,
 **  modify, copy, or redistribute it subject to the terms and conditions
@@ -33,10 +33,10 @@ void dlm_del_ast(struct dlm_lkb *lkb)
 	spin_unlock(&ast_queue_lock);
 }
 
-void dlm_add_ast(struct dlm_lkb *lkb, int type)
+void dlm_add_ast(struct dlm_lkb *lkb, int type, int bastmode)
 {
 	if (lkb->lkb_flags & DLM_IFL_USER) {
-		dlm_user_add_ast(lkb, type);
+		dlm_user_add_ast(lkb, type, bastmode);
 		return;
 	}
 
@@ -46,6 +46,8 @@ void dlm_add_ast(struct dlm_lkb *lkb, int type)
 		list_add_tail(&lkb->lkb_astqueue, &ast_queue);
 	}
 	lkb->lkb_ast_type |= type;
+	if (bastmode)
+		lkb->lkb_bastmode = bastmode;
 	spin_unlock(&ast_queue_lock);
 
 	set_bit(WAKE_ASTS, &astd_wakeflags);
@@ -59,50 +61,40 @@ static void process_asts(void)
 	struct dlm_lkb *lkb;
 	void (*cast) (void *astparam);
 	void (*bast) (void *astparam, int mode);
-	int type = 0, found, bmode;
-
-	for (;;) {
-		found = 0;
-		spin_lock(&ast_queue_lock);
-		list_for_each_entry(lkb, &ast_queue, lkb_astqueue) {
-			r = lkb->lkb_resource;
-			ls = r->res_ls;
-
-			if (dlm_locking_stopped(ls))
-				continue;
-
-			list_del(&lkb->lkb_astqueue);
-			type = lkb->lkb_ast_type;
-			lkb->lkb_ast_type = 0;
-			found = 1;
-			break;
-		}
-		spin_unlock(&ast_queue_lock);
+	int type = 0, bastmode;
+
+repeat:
+	spin_lock(&ast_queue_lock);
+	list_for_each_entry(lkb, &ast_queue, lkb_astqueue) {
+		r = lkb->lkb_resource;
+		ls = r->res_ls;
+
+		if (dlm_locking_stopped(ls))
+			continue;
 
-		if (!found)
-			break;
+		list_del(&lkb->lkb_astqueue);
+		type = lkb->lkb_ast_type;
+		lkb->lkb_ast_type = 0;
+		bastmode = lkb->lkb_bastmode;
 
+		spin_unlock(&ast_queue_lock);
 		cast = lkb->lkb_astfn;
 		bast = lkb->lkb_bastfn;
-		bmode = lkb->lkb_bastmode;
 
 		if ((type & AST_COMP) && cast)
 			cast(lkb->lkb_astparam);
 
-		/* FIXME: Is it safe to look at lkb_grmode here
-		   without doing a lock_rsb() ?
-		   Look at other checks in v1 to avoid basts. */
-
 		if ((type & AST_BAST) && bast)
-			if (!dlm_modes_compat(lkb->lkb_grmode, bmode))
-				bast(lkb->lkb_astparam, bmode);
+			bast(lkb->lkb_astparam, bastmode);
 
 		/* this removes the reference added by dlm_add_ast
 		   and may result in the lkb being freed */
 		dlm_put_lkb(lkb);
 
-		schedule();
+		cond_resched();
+		goto repeat;
 	}
+	spin_unlock(&ast_queue_lock);
 }
 
 static inline int no_asts(void)
diff --git a/fs/dlm/ast.h b/fs/dlm/ast.h
index 6ee276c74c5..1b5fc5f428f 100644
--- a/fs/dlm/ast.h
+++ b/fs/dlm/ast.h
@@ -1,7 +1,7 @@
 /******************************************************************************
 *******************************************************************************
 **
-**  Copyright (C) 2005 Red Hat, Inc.  All rights reserved.
+**  Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
 **
 **  This copyrighted material is made available to anyone wishing to use,
 **  modify, copy, or redistribute it subject to the terms and conditions
@@ -13,7 +13,7 @@
 #ifndef __ASTD_DOT_H__
 #define __ASTD_DOT_H__
 
-void dlm_add_ast(struct dlm_lkb *lkb, int type);
+void dlm_add_ast(struct dlm_lkb *lkb, int type, int bastmode);
 void dlm_del_ast(struct dlm_lkb *lkb);
 
 void dlm_astd_wake(void);
diff --git a/fs/dlm/debug_fs.c b/fs/dlm/debug_fs.c
index 8fc24f4507a..2f107d1a6a4 100644
--- a/fs/dlm/debug_fs.c
+++ b/fs/dlm/debug_fs.c
@@ -1,7 +1,7 @@
 /******************************************************************************
 *******************************************************************************
 **
-**  Copyright (C) 2005 Red Hat, Inc.  All rights reserved.
+**  Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
 **
 **  This copyrighted material is made available to anyone wishing to use,
 **  modify, copy, or redistribute it subject to the terms and conditions
@@ -27,7 +27,7 @@ static struct dentry *dlm_root;
 
 struct rsb_iter {
 	int entry;
-	int locks;
+	int format;
 	int header;
 	struct dlm_ls *ls;
 	struct list_head *next;
@@ -60,8 +60,8 @@ static char *print_lockmode(int mode)
 	}
 }
 
-static void print_resource_lock(struct seq_file *s, struct dlm_lkb *lkb,
-				struct dlm_rsb *res)
+static void print_format1_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			       struct dlm_rsb *res)
 {
 	seq_printf(s, "%08x %s", lkb->lkb_id, print_lockmode(lkb->lkb_grmode));
 
@@ -83,7 +83,7 @@ static void print_resource_lock(struct seq_file *s, struct dlm_lkb *lkb,
 	seq_printf(s, "\n");
 }
 
-static int print_resource(struct dlm_rsb *res, struct seq_file *s)
+static int print_format1(struct dlm_rsb *res, struct seq_file *s)
 {
 	struct dlm_lkb *lkb;
 	int i, lvblen = res->res_ls->ls_lvblen, recover_list, root_list;
@@ -134,15 +134,15 @@ static int print_resource(struct dlm_rsb *res, struct seq_file *s)
 	/* Print the locks attached to this resource */
 	seq_printf(s, "Granted Queue\n");
 	list_for_each_entry(lkb, &res->res_grantqueue, lkb_statequeue)
-		print_resource_lock(s, lkb, res);
+		print_format1_lock(s, lkb, res);
 
 	seq_printf(s, "Conversion Queue\n");
 	list_for_each_entry(lkb, &res->res_convertqueue, lkb_statequeue)
-		print_resource_lock(s, lkb, res);
+		print_format1_lock(s, lkb, res);
 
 	seq_printf(s, "Waiting Queue\n");
 	list_for_each_entry(lkb, &res->res_waitqueue, lkb_statequeue)
-		print_resource_lock(s, lkb, res);
+		print_format1_lock(s, lkb, res);
 
 	if (list_empty(&res->res_lookup))
 		goto out;
@@ -160,23 +160,24 @@ static int print_resource(struct dlm_rsb *res, struct seq_file *s)
 	return 0;
 }
 
-static void print_lock(struct seq_file *s, struct dlm_lkb *lkb, struct dlm_rsb *r)
+static void print_format2_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			       struct dlm_rsb *r)
 {
-	unsigned int waiting = 0;
-	uint64_t xid = 0;
+	u64 xid = 0;
+	u64 us;
 
 	if (lkb->lkb_flags & DLM_IFL_USER) {
 		if (lkb->lkb_ua)
 			xid = lkb->lkb_ua->xid;
 	}
 
-	if (lkb->lkb_timestamp)
-		waiting = jiffies_to_msecs(jiffies - lkb->lkb_timestamp);
+	/* microseconds since lkb was added to current queue */
+	us = ktime_to_us(ktime_sub(ktime_get(), lkb->lkb_timestamp));
 
-	/* id nodeid remid pid xid exflags flags sts grmode rqmode time_ms
+	/* id nodeid remid pid xid exflags flags sts grmode rqmode time_us
 	   r_nodeid r_len r_name */
 
-	seq_printf(s, "%x %d %x %u %llu %x %x %d %d %d %u %u %d \"%s\"\n",
+	seq_printf(s, "%x %d %x %u %llu %x %x %d %d %d %llu %u %d \"%s\"\n",
 		   lkb->lkb_id,
 		   lkb->lkb_nodeid,
 		   lkb->lkb_remid,
@@ -187,26 +188,114 @@ static void print_lock(struct seq_file *s, struct dlm_lkb *lkb, struct dlm_rsb *
 		   lkb->lkb_status,
 		   lkb->lkb_grmode,
 		   lkb->lkb_rqmode,
-		   waiting,
+		   (unsigned long long)us,
 		   r->res_nodeid,
 		   r->res_length,
 		   r->res_name);
 }
 
-static int print_locks(struct dlm_rsb *r, struct seq_file *s)
+static int print_format2(struct dlm_rsb *r, struct seq_file *s)
 {
 	struct dlm_lkb *lkb;
 
 	lock_rsb(r);
 
 	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue)
-		print_lock(s, lkb, r);
+		print_format2_lock(s, lkb, r);
 
 	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue)
-		print_lock(s, lkb, r);
+		print_format2_lock(s, lkb, r);
 
 	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue)
-		print_lock(s, lkb, r);
+		print_format2_lock(s, lkb, r);
+
+	unlock_rsb(r);
+	return 0;
+}
+
+static void print_format3_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			       int rsb_lookup)
+{
+	u64 xid = 0;
+
+	if (lkb->lkb_flags & DLM_IFL_USER) {
+		if (lkb->lkb_ua)
+			xid = lkb->lkb_ua->xid;
+	}
+
+	seq_printf(s, "lkb %x %d %x %u %llu %x %x %d %d %d %d %d %d %u %llu %llu\n",
+		   lkb->lkb_id,
+		   lkb->lkb_nodeid,
+		   lkb->lkb_remid,
+		   lkb->lkb_ownpid,
+		   (unsigned long long)xid,
+		   lkb->lkb_exflags,
+		   lkb->lkb_flags,
+		   lkb->lkb_status,
+		   lkb->lkb_grmode,
+		   lkb->lkb_rqmode,
+		   lkb->lkb_highbast,
+		   rsb_lookup,
+		   lkb->lkb_wait_type,
+		   lkb->lkb_lvbseq,
+		   (unsigned long long)ktime_to_ns(lkb->lkb_timestamp),
+		   (unsigned long long)ktime_to_ns(lkb->lkb_time_bast));
+}
+
+static int print_format3(struct dlm_rsb *r, struct seq_file *s)
+{
+	struct dlm_lkb *lkb;
+	int i, lvblen = r->res_ls->ls_lvblen;
+	int print_name = 1;
+
+	lock_rsb(r);
+
+	seq_printf(s, "rsb %p %d %x %lx %d %d %u %d ",
+		   r,
+		   r->res_nodeid,
+		   r->res_first_lkid,
+		   r->res_flags,
+		   !list_empty(&r->res_root_list),
+		   !list_empty(&r->res_recover_list),
+		   r->res_recover_locks_count,
+		   r->res_length);
+
+	for (i = 0; i < r->res_length; i++) {
+		if (!isascii(r->res_name[i]) || !isprint(r->res_name[i]))
+			print_name = 0;
+	}
+
+	seq_printf(s, "%s", print_name ? "str " : "hex");
+
+	for (i = 0; i < r->res_length; i++) {
+		if (print_name)
+			seq_printf(s, "%c", r->res_name[i]);
+		else
+			seq_printf(s, " %02x", (unsigned char)r->res_name[i]);
+	}
+	seq_printf(s, "\n");
+
+	if (!r->res_lvbptr)
+		goto do_locks;
+
+	seq_printf(s, "lvb %u %d", r->res_lvbseq, lvblen);
+
+	for (i = 0; i < lvblen; i++)
+		seq_printf(s, " %02x", (unsigned char)r->res_lvbptr[i]);
+	seq_printf(s, "\n");
+
+ do_locks:
+	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue)
+		print_format3_lock(s, lkb, 0);
+
+	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue)
+		print_format3_lock(s, lkb, 0);
+
+	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue)
+		print_format3_lock(s, lkb, 0);
+
+	list_for_each_entry(lkb, &r->res_lookup, lkb_rsb_lookup)
+		print_format3_lock(s, lkb, 1);
 
 	unlock_rsb(r);
 	return 0;
@@ -231,7 +320,7 @@ static int rsb_iter_next(struct rsb_iter *ri)
 				break;
 			}
 			read_unlock(&ls->ls_rsbtbl[i].lock);
-                }
+		}
 		ri->entry = i;
 
 		if (ri->entry >= ls->ls_rsbtbl_size)
@@ -248,7 +337,7 @@ static int rsb_iter_next(struct rsb_iter *ri)
 			read_unlock(&ls->ls_rsbtbl[i].lock);
 			dlm_put_rsb(old);
 			goto top;
-                }
+		}
 		ri->rsb = list_entry(ri->next, struct dlm_rsb, res_hashchain);
 		dlm_hold_rsb(ri->rsb);
 		read_unlock(&ls->ls_rsbtbl[i].lock);
@@ -274,6 +363,7 @@ static struct rsb_iter *rsb_iter_init(struct dlm_ls *ls)
 	ri->ls = ls;
 	ri->entry = 0;
 	ri->next = NULL;
+	ri->format = 1;
 
 	if (rsb_iter_next(ri)) {
 		rsb_iter_free(ri);
@@ -325,16 +415,26 @@ static int rsb_seq_show(struct seq_file *file, void *iter_ptr)
 {
 	struct rsb_iter *ri = iter_ptr;
 
-	if (ri->locks) {
+	switch (ri->format) {
+	case 1:
+		print_format1(ri->rsb, file);
+		break;
+	case 2:
 		if (ri->header) {
-			seq_printf(file, "id nodeid remid pid xid exflags flags "
-					 "sts grmode rqmode time_ms r_nodeid "
-					 "r_len r_name\n");
+			seq_printf(file, "id nodeid remid pid xid exflags "
+					 "flags sts grmode rqmode time_ms "
+					 "r_nodeid r_len r_name\n");
 			ri->header = 0;
 		}
-		print_locks(ri->rsb, file);
-	} else {
-		print_resource(ri->rsb, file);
+		print_format2(ri->rsb, file);
+		break;
+	case 3:
+		if (ri->header) {
+			seq_printf(file, "version rsb 1.1 lvb 1.1 lkb 1.1\n");
+			ri->header = 0;
+		}
+		print_format3(ri->rsb, file);
+		break;
 	}
 
 	return 0;
@@ -385,7 +485,7 @@ static struct rsb_iter *locks_iter_init(struct dlm_ls *ls, loff_t *pos)
 	ri->ls = ls;
 	ri->entry = 0;
 	ri->next = NULL;
-	ri->locks = 1;
+	ri->format = 2;
 
 	if (*pos == 0)
 		ri->header = 1;
@@ -448,6 +548,84 @@ static const struct file_operations locks_fops = {
 };
 
 /*
+ * Dump all rsb/lvb/lkb state in compact listing, more complete than _locks
+ * This can replace both formats 1 and 2 eventually.
+ */
+
+static struct rsb_iter *all_iter_init(struct dlm_ls *ls, loff_t *pos)
+{
+	struct rsb_iter *ri;
+
+	ri = kzalloc(sizeof *ri, GFP_KERNEL);
+	if (!ri)
+		return NULL;
+
+	ri->ls = ls;
+	ri->entry = 0;
+	ri->next = NULL;
+	ri->format = 3;
+
+	if (*pos == 0)
+		ri->header = 1;
+
+	if (rsb_iter_next(ri)) {
+		rsb_iter_free(ri);
+		return NULL;
+	}
+
+	return ri;
+}
+
+static void *all_seq_start(struct seq_file *file, loff_t *pos)
+{
+	struct rsb_iter *ri;
+	loff_t n = *pos;
+
+	ri = all_iter_init(file->private, pos);
+	if (!ri)
+		return NULL;
+
+	while (n--) {
+		if (rsb_iter_next(ri)) {
+			rsb_iter_free(ri);
+			return NULL;
+		}
+	}
+
+	return ri;
+}
+
+static struct seq_operations all_seq_ops = {
+	.start = all_seq_start,
+	.next  = rsb_seq_next,
+	.stop  = rsb_seq_stop,
+	.show  = rsb_seq_show,
+};
+
+static int all_open(struct inode *inode, struct file *file)
+{
+	struct seq_file *seq;
+	int ret;
+
+	ret = seq_open(file, &all_seq_ops);
+	if (ret)
+		return ret;
+
+	seq = file->private_data;
+	seq->private = inode->i_private;
+
+	return 0;
+}
+
+static const struct file_operations all_fops = {
+	.owner   = THIS_MODULE,
+	.open    = all_open,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+/*
  * dump lkb's on the ls_waiters list
  */
 
@@ -489,30 +667,33 @@ static const struct file_operations waiters_fops = {
 	.read    = waiters_read
 };
 
+void dlm_delete_debug_file(struct dlm_ls *ls)
+{
+	if (ls->ls_debug_rsb_dentry)
+		debugfs_remove(ls->ls_debug_rsb_dentry);
+	if (ls->ls_debug_waiters_dentry)
+		debugfs_remove(ls->ls_debug_waiters_dentry);
+	if (ls->ls_debug_locks_dentry)
+		debugfs_remove(ls->ls_debug_locks_dentry);
+	if (ls->ls_debug_all_dentry)
+		debugfs_remove(ls->ls_debug_all_dentry);
+}
+
 int dlm_create_debug_file(struct dlm_ls *ls)
 {
 	char name[DLM_LOCKSPACE_LEN+8];
 
+	/* format 1 */
+
 	ls->ls_debug_rsb_dentry = debugfs_create_file(ls->ls_name,
 						      S_IFREG | S_IRUGO,
 						      dlm_root,
 						      ls,
 						      &rsb_fops);
 	if (!ls->ls_debug_rsb_dentry)
-		return -ENOMEM;
+		goto fail;
 
-	memset(name, 0, sizeof(name));
-	snprintf(name, DLM_LOCKSPACE_LEN+8, "%s_waiters", ls->ls_name);
-
-	ls->ls_debug_waiters_dentry = debugfs_create_file(name,
-							  S_IFREG | S_IRUGO,
-							  dlm_root,
-							  ls,
-							  &waiters_fops);
-	if (!ls->ls_debug_waiters_dentry) {
-		debugfs_remove(ls->ls_debug_rsb_dentry);
-		return -ENOMEM;
-	}
+	/* format 2 */
 
 	memset(name, 0, sizeof(name));
 	snprintf(name, DLM_LOCKSPACE_LEN+8, "%s_locks", ls->ls_name);
@@ -522,23 +703,38 @@ int dlm_create_debug_file(struct dlm_ls *ls)
 							dlm_root,
 							ls,
 							&locks_fops);
-	if (!ls->ls_debug_locks_dentry) {
-		debugfs_remove(ls->ls_debug_waiters_dentry);
-		debugfs_remove(ls->ls_debug_rsb_dentry);
-		return -ENOMEM;
-	}
+	if (!ls->ls_debug_locks_dentry)
+		goto fail;
+
+	/* format 3 */
+
+	memset(name, 0, sizeof(name));
+	snprintf(name, DLM_LOCKSPACE_LEN+8, "%s_all", ls->ls_name);
+
+	ls->ls_debug_all_dentry = debugfs_create_file(name,
+						      S_IFREG | S_IRUGO,
+						      dlm_root,
+						      ls,
+						      &all_fops);
+	if (!ls->ls_debug_all_dentry)
+		goto fail;
+
+	memset(name, 0, sizeof(name));
+	snprintf(name, DLM_LOCKSPACE_LEN+8, "%s_waiters", ls->ls_name);
+
+	ls->ls_debug_waiters_dentry = debugfs_create_file(name,
+							  S_IFREG | S_IRUGO,
+							  dlm_root,
+							  ls,
+							  &waiters_fops);
+	if (!ls->ls_debug_waiters_dentry)
+		goto fail;
 
 	return 0;
-}
 
-void dlm_delete_debug_file(struct dlm_ls *ls)
-{
-	if (ls->ls_debug_rsb_dentry)
-		debugfs_remove(ls->ls_debug_rsb_dentry);
-	if (ls->ls_debug_waiters_dentry)
-		debugfs_remove(ls->ls_debug_waiters_dentry);
-	if (ls->ls_debug_locks_dentry)
-		debugfs_remove(ls->ls_debug_locks_dentry);
+ fail:
+	dlm_delete_debug_file(ls);
+	return -ENOMEM;
 }
 
 int __init dlm_register_debugfs(void)
diff --git a/fs/dlm/dir.c b/fs/dlm/dir.c
index 85defeb64df..92969f879a1 100644
--- a/fs/dlm/dir.c
+++ b/fs/dlm/dir.c
@@ -374,7 +374,7 @@ void dlm_copy_master_names(struct dlm_ls *ls, char *inbuf, int inlen,
 	struct list_head *list;
 	struct dlm_rsb *r;
 	int offset = 0, dir_nodeid;
-	uint16_t be_namelen;
+	__be16 be_namelen;
 
 	down_read(&ls->ls_root_sem);
 
@@ -410,15 +410,15 @@ void dlm_copy_master_names(struct dlm_ls *ls, char *inbuf, int inlen,
 
 		if (offset + sizeof(uint16_t)*2 + r->res_length > outlen) {
 			/* Write end-of-block record */
-			be_namelen = 0;
-			memcpy(outbuf + offset, &be_namelen, sizeof(uint16_t));
-			offset += sizeof(uint16_t);
+			be_namelen = cpu_to_be16(0);
+			memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
+			offset += sizeof(__be16);
 			goto out;
 		}
 
 		be_namelen = cpu_to_be16(r->res_length);
-		memcpy(outbuf + offset, &be_namelen, sizeof(uint16_t));
-		offset += sizeof(uint16_t);
+		memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
+		offset += sizeof(__be16);
 		memcpy(outbuf + offset, r->res_name, r->res_length);
 		offset += r->res_length;
 	}
@@ -430,9 +430,9 @@ void dlm_copy_master_names(struct dlm_ls *ls, char *inbuf, int inlen,
 
 	if ((list == &ls->ls_root_list) &&
 	    (offset + sizeof(uint16_t) <= outlen)) {
-		be_namelen = 0xFFFF;
-		memcpy(outbuf + offset, &be_namelen, sizeof(uint16_t));
-		offset += sizeof(uint16_t);
+		be_namelen = cpu_to_be16(0xFFFF);
+		memcpy(outbuf + offset, &be_namelen, sizeof(__be16));
+		offset += sizeof(__be16);
 	}
 
  out:
diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h
index 868e4c9ef12..ef2f1e35396 100644
--- a/fs/dlm/dlm_internal.h
+++ b/fs/dlm/dlm_internal.h
@@ -245,7 +245,8 @@ struct dlm_lkb {
 	struct list_head	lkb_astqueue;	/* need ast to be sent */
 	struct list_head	lkb_ownqueue;	/* list of locks for a process */
 	struct list_head	lkb_time_list;
-	unsigned long		lkb_timestamp;
+	ktime_t			lkb_time_bast;	/* for debugging */
+	ktime_t			lkb_timestamp;
 	unsigned long		lkb_timeout_cs;
 
 	char			*lkb_lvbptr;
@@ -481,6 +482,7 @@ struct dlm_ls {
 	struct dentry		*ls_debug_rsb_dentry; /* debugfs */
 	struct dentry		*ls_debug_waiters_dentry; /* debugfs */
 	struct dentry		*ls_debug_locks_dentry; /* debugfs */
+	struct dentry		*ls_debug_all_dentry; /* debugfs */
 
 	wait_queue_head_t	ls_uevent_wait;	/* user part of join/leave */
 	int			ls_uevent_result;
diff --git a/fs/dlm/lock.c b/fs/dlm/lock.c
index 724ddac9153..6cfe65bbf4a 100644
--- a/fs/dlm/lock.c
+++ b/fs/dlm/lock.c
@@ -307,7 +307,7 @@ static void queue_cast(struct dlm_rsb *r, struct dlm_lkb *lkb, int rv)
 	lkb->lkb_lksb->sb_status = rv;
 	lkb->lkb_lksb->sb_flags = lkb->lkb_sbflags;
 
-	dlm_add_ast(lkb, AST_COMP);
+	dlm_add_ast(lkb, AST_COMP, 0);
 }
 
 static inline void queue_cast_overlap(struct dlm_rsb *r, struct dlm_lkb *lkb)
@@ -318,12 +318,12 @@ static inline void queue_cast_overlap(struct dlm_rsb *r, struct dlm_lkb *lkb)
 
 static void queue_bast(struct dlm_rsb *r, struct dlm_lkb *lkb, int rqmode)
 {
+	lkb->lkb_time_bast = ktime_get();
+
 	if (is_master_copy(lkb))
 		send_bast(r, lkb, rqmode);
-	else {
-		lkb->lkb_bastmode = rqmode;
-		dlm_add_ast(lkb, AST_BAST);
-	}
+	else
+		dlm_add_ast(lkb, AST_BAST, rqmode);
 }
 
 /*
@@ -744,6 +744,8 @@ static void add_lkb(struct dlm_rsb *r, struct dlm_lkb *lkb, int status)
 
 	DLM_ASSERT(!lkb->lkb_status, dlm_print_lkb(lkb););
 
+	lkb->lkb_timestamp = ktime_get();
+
 	lkb->lkb_status = status;
 
 	switch (status) {
@@ -1013,10 +1015,8 @@ static void add_timeout(struct dlm_lkb *lkb)
 {
 	struct dlm_ls *ls = lkb->lkb_resource->res_ls;
 
-	if (is_master_copy(lkb)) {
-		lkb->lkb_timestamp = jiffies;
+	if (is_master_copy(lkb))
 		return;
-	}
 
 	if (test_bit(LSFL_TIMEWARN, &ls->ls_flags) &&
 	    !(lkb->lkb_exflags & DLM_LKF_NODLCKWT)) {
@@ -1031,7 +1031,6 @@ static void add_timeout(struct dlm_lkb *lkb)
 	DLM_ASSERT(list_empty(&lkb->lkb_time_list), dlm_print_lkb(lkb););
 	mutex_lock(&ls->ls_timeout_mutex);
 	hold_lkb(lkb);
-	lkb->lkb_timestamp = jiffies;
 	list_add_tail(&lkb->lkb_time_list, &ls->ls_timeout);
 	mutex_unlock(&ls->ls_timeout_mutex);
 }
@@ -1059,6 +1058,7 @@ void dlm_scan_timeout(struct dlm_ls *ls)
 	struct dlm_rsb *r;
 	struct dlm_lkb *lkb;
 	int do_cancel, do_warn;
+	s64 wait_us;
 
 	for (;;) {
 		if (dlm_locking_stopped(ls))
@@ -1069,14 +1069,15 @@ void dlm_scan_timeout(struct dlm_ls *ls)
 		mutex_lock(&ls->ls_timeout_mutex);
 		list_for_each_entry(lkb, &ls->ls_timeout, lkb_time_list) {
 
+			wait_us = ktime_to_us(ktime_sub(ktime_get(),
+					      		lkb->lkb_timestamp));
+
 			if ((lkb->lkb_exflags & DLM_LKF_TIMEOUT) &&
-			    time_after_eq(jiffies, lkb->lkb_timestamp +
-					  lkb->lkb_timeout_cs * HZ/100))
+			    wait_us >= (lkb->lkb_timeout_cs * 10000))
 				do_cancel = 1;
 
 			if ((lkb->lkb_flags & DLM_IFL_WATCH_TIMEWARN) &&
-			    time_after_eq(jiffies, lkb->lkb_timestamp +
-				   	   dlm_config.ci_timewarn_cs * HZ/100))
+			    wait_us >= dlm_config.ci_timewarn_cs * 10000)
 				do_warn = 1;
 
 			if (!do_cancel && !do_warn)
@@ -1122,12 +1123,12 @@ void dlm_scan_timeout(struct dlm_ls *ls)
 void dlm_adjust_timeouts(struct dlm_ls *ls)
 {
 	struct dlm_lkb *lkb;
-	long adj = jiffies - ls->ls_recover_begin;
+	u64 adj_us = jiffies_to_usecs(jiffies - ls->ls_recover_begin);
 
 	ls->ls_recover_begin = 0;
 	mutex_lock(&ls->ls_timeout_mutex);
 	list_for_each_entry(lkb, &ls->ls_timeout, lkb_time_list)
-		lkb->lkb_timestamp += adj;
+		lkb->lkb_timestamp = ktime_add_us(lkb->lkb_timestamp, adj_us);
 	mutex_unlock(&ls->ls_timeout_mutex);
 }
 
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 3962262f991..103a5ebd137 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -295,6 +295,7 @@ static int add_sock(struct socket *sock, struct connection *con)
 	con->sock->sk->sk_write_space = lowcomms_write_space;
 	con->sock->sk->sk_state_change = lowcomms_state_change;
 	con->sock->sk->sk_user_data = con;
+	con->sock->sk->sk_allocation = GFP_NOFS;
 	return 0;
 }
 
@@ -823,7 +824,6 @@ static void sctp_init_assoc(struct connection *con)
 	len = e->len;
 	offset = e->offset;
 	spin_unlock(&con->writequeue_lock);
-	kmap(e->page);
 
 	/* Send the first block off the write queue */
 	iov[0].iov_base = page_address(e->page)+offset;
@@ -854,7 +854,6 @@ static void sctp_init_assoc(struct connection *con)
 
 		if (e->len == 0 && e->users == 0) {
 			list_del(&e->list);
-			kunmap(e->page);
 			free_entry(e);
 		}
 		spin_unlock(&con->writequeue_lock);
@@ -1203,8 +1202,6 @@ void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc)
 
 	if (e) {
 	got_one:
-		if (users == 0)
-			kmap(e->page);
 		*ppc = page_address(e->page) + offset;
 		return e;
 	}
@@ -1233,7 +1230,6 @@ void dlm_lowcomms_commit_buffer(void *mh)
 	if (users)
 		goto out;
 	e->len = e->end - e->offset;
-	kunmap(e->page);
 	spin_unlock(&con->writequeue_lock);
 
 	if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags)) {
@@ -1272,7 +1268,6 @@ static void send_to_sock(struct connection *con)
 		offset = e->offset;
 		BUG_ON(len == 0 && e->users == 0);
 		spin_unlock(&con->writequeue_lock);
-		kmap(e->page);
 
 		ret = 0;
 		if (len) {
@@ -1294,7 +1289,6 @@ static void send_to_sock(struct connection *con)
 
 		if (e->len == 0 && e->users == 0) {
 			list_del(&e->list);
-			kunmap(e->page);
 			free_entry(e);
 			continue;
 		}
diff --git a/fs/dlm/memory.c b/fs/dlm/memory.c
index 54c14c6d06c..c1775b84eba 100644
--- a/fs/dlm/memory.c
+++ b/fs/dlm/memory.c
@@ -39,7 +39,7 @@ char *dlm_allocate_lvb(struct dlm_ls *ls)
 {
 	char *p;
 
-	p = kzalloc(ls->ls_lvblen, GFP_KERNEL);
+	p = kzalloc(ls->ls_lvblen, ls->ls_allocation);
 	return p;
 }
 
@@ -57,7 +57,7 @@ struct dlm_rsb *dlm_allocate_rsb(struct dlm_ls *ls, int namelen)
 
 	DLM_ASSERT(namelen <= DLM_RESNAME_MAXLEN,);
 
-	r = kzalloc(sizeof(*r) + namelen, GFP_KERNEL);
+	r = kzalloc(sizeof(*r) + namelen, ls->ls_allocation);
 	return r;
 }
 
@@ -72,7 +72,7 @@ struct dlm_lkb *dlm_allocate_lkb(struct dlm_ls *ls)
 {
 	struct dlm_lkb *lkb;
 
-	lkb = kmem_cache_zalloc(lkb_cache, GFP_KERNEL);
+	lkb = kmem_cache_zalloc(lkb_cache, ls->ls_allocation);
 	return lkb;
 }
 
diff --git a/fs/dlm/midcomms.c b/fs/dlm/midcomms.c
index 07ac709f3ed..f3396c622ae 100644
--- a/fs/dlm/midcomms.c
+++ b/fs/dlm/midcomms.c
@@ -112,7 +112,7 @@ int dlm_process_incoming_buffer(int nodeid, const void *base,
 		   ordinary messages). */
 
 		if (msglen > sizeof(__tmp) && p == &__tmp.p) {
-			p = kmalloc(dlm_config.ci_buffer_size, GFP_KERNEL);
+			p = kmalloc(dlm_config.ci_buffer_size, GFP_NOFS);
 			if (p == NULL)
 				return ret;
 		}
diff --git a/fs/dlm/netlink.c b/fs/dlm/netlink.c
index aa2a5775a02..ccc9d62c462 100644
--- a/fs/dlm/netlink.c
+++ b/fs/dlm/netlink.c
@@ -115,7 +115,6 @@ static void fill_data(struct dlm_lock_data *data, struct dlm_lkb *lkb)
 	data->status = lkb->lkb_status;
 	data->grmode = lkb->lkb_grmode;
 	data->rqmode = lkb->lkb_rqmode;
-	data->timestamp = lkb->lkb_timestamp;
 	if (lkb->lkb_ua)
 		data->xid = lkb->lkb_ua->xid;
 	if (r) {
diff --git a/fs/dlm/user.c b/fs/dlm/user.c
index b3832c67194..065149e84f4 100644
--- a/fs/dlm/user.c
+++ b/fs/dlm/user.c
@@ -175,7 +175,7 @@ static int lkb_is_endoflife(struct dlm_lkb *lkb, int sb_status, int type)
 /* we could possibly check if the cancel of an orphan has resulted in the lkb
    being removed and then remove that lkb from the orphans list and free it */
 
-void dlm_user_add_ast(struct dlm_lkb *lkb, int type)
+void dlm_user_add_ast(struct dlm_lkb *lkb, int type, int bastmode)
 {
 	struct dlm_ls *ls;
 	struct dlm_user_args *ua;
@@ -208,6 +208,8 @@ void dlm_user_add_ast(struct dlm_lkb *lkb, int type)
 
 	ast_type = lkb->lkb_ast_type;
 	lkb->lkb_ast_type |= type;
+	if (bastmode)
+		lkb->lkb_bastmode = bastmode;
 
 	if (!ast_type) {
 		kref_get(&lkb->lkb_ref);
diff --git a/fs/dlm/user.h b/fs/dlm/user.h
index 35eb6a13d61..1c968649228 100644
--- a/fs/dlm/user.h
+++ b/fs/dlm/user.h
@@ -9,7 +9,7 @@
 #ifndef __USER_DOT_H__
 #define __USER_DOT_H__
 
-void dlm_user_add_ast(struct dlm_lkb *lkb, int type);
+void dlm_user_add_ast(struct dlm_lkb *lkb, int type, int bastmode);
 int dlm_user_init(void);
 void dlm_user_exit(void);
 int dlm_device_deregister(struct dlm_ls *ls);
diff --git a/fs/dquot.c b/fs/dquot.c
index c237ccc8581..48c0571f831 100644
--- a/fs/dquot.c
+++ b/fs/dquot.c
@@ -211,8 +211,6 @@ static struct hlist_head *dquot_hash;
 
 struct dqstats dqstats;
 
-static void dqput(struct dquot *dquot);
-
 static inline unsigned int
 hashfn(const struct super_block *sb, unsigned int id, int type)
 {
@@ -415,6 +413,17 @@ out_dqlock:
 	return ret;
 }
 
+void dquot_destroy(struct dquot *dquot)
+{
+	kmem_cache_free(dquot_cachep, dquot);
+}
+EXPORT_SYMBOL(dquot_destroy);
+
+static inline void do_destroy_dquot(struct dquot *dquot)
+{
+	dquot->dq_sb->dq_op->destroy_dquot(dquot);
+}
+
 /* Invalidate all dquots on the list. Note that this function is called after
  * quota is disabled and pointers from inodes removed so there cannot be new
  * quota users. There can still be some users of quotas due to inodes being
@@ -463,9 +472,44 @@ restart:
 		remove_dquot_hash(dquot);
 		remove_free_dquot(dquot);
 		remove_inuse(dquot);
-		kmem_cache_free(dquot_cachep, dquot);
+		do_destroy_dquot(dquot);
+	}
+	spin_unlock(&dq_list_lock);
+}
+
+/* Call callback for every active dquot on given filesystem */
+int dquot_scan_active(struct super_block *sb,
+		      int (*fn)(struct dquot *dquot, unsigned long priv),
+		      unsigned long priv)
+{
+	struct dquot *dquot, *old_dquot = NULL;
+	int ret = 0;
+
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
+	spin_lock(&dq_list_lock);
+	list_for_each_entry(dquot, &inuse_list, dq_inuse) {
+		if (!test_bit(DQ_ACTIVE_B, &dquot->dq_flags))
+			continue;
+		if (dquot->dq_sb != sb)
+			continue;
+		/* Now we have active dquot so we can just increase use count */
+		atomic_inc(&dquot->dq_count);
+		dqstats.lookups++;
+		spin_unlock(&dq_list_lock);
+		dqput(old_dquot);
+		old_dquot = dquot;
+		ret = fn(dquot, priv);
+		if (ret < 0)
+			goto out;
+		spin_lock(&dq_list_lock);
+		/* We are safe to continue now because our dquot could not
+		 * be moved out of the inuse list while we hold the reference */
 	}
 	spin_unlock(&dq_list_lock);
+out:
+	dqput(old_dquot);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
+	return ret;
 }
 
 int vfs_quota_sync(struct super_block *sb, int type)
@@ -479,7 +523,7 @@ int vfs_quota_sync(struct super_block *sb, int type)
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (type != -1 && cnt != type)
 			continue;
-		if (!sb_has_quota_enabled(sb, cnt))
+		if (!sb_has_quota_active(sb, cnt))
 			continue;
 		spin_lock(&dq_list_lock);
 		dirty = &dqopt->info[cnt].dqi_dirty_list;
@@ -504,8 +548,8 @@ int vfs_quota_sync(struct super_block *sb, int type)
 	}
 
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
-		if ((cnt == type || type == -1) && sb_has_quota_enabled(sb, cnt)
-			&& info_dirty(&dqopt->info[cnt]))
+		if ((cnt == type || type == -1) && sb_has_quota_active(sb, cnt)
+		    && info_dirty(&dqopt->info[cnt]))
 			sb->dq_op->write_info(sb, cnt);
 	spin_lock(&dq_list_lock);
 	dqstats.syncs++;
@@ -527,7 +571,7 @@ static void prune_dqcache(int count)
 		remove_dquot_hash(dquot);
 		remove_free_dquot(dquot);
 		remove_inuse(dquot);
-		kmem_cache_free(dquot_cachep, dquot);
+		do_destroy_dquot(dquot);
 		count--;
 		head = free_dquots.prev;
 	}
@@ -558,7 +602,7 @@ static struct shrinker dqcache_shrinker = {
  * NOTE: If you change this function please check whether dqput_blocks() works right...
  * MUST be called with either dqptr_sem or dqonoff_mutex held
  */
-static void dqput(struct dquot *dquot)
+void dqput(struct dquot *dquot)
 {
 	int ret;
 
@@ -584,7 +628,7 @@ we_slept:
 		/* We have more than one user... nothing to do */
 		atomic_dec(&dquot->dq_count);
 		/* Releasing dquot during quotaoff phase? */
-		if (!sb_has_quota_enabled(dquot->dq_sb, dquot->dq_type) &&
+		if (!sb_has_quota_active(dquot->dq_sb, dquot->dq_type) &&
 		    atomic_read(&dquot->dq_count) == 1)
 			wake_up(&dquot->dq_wait_unused);
 		spin_unlock(&dq_list_lock);
@@ -625,11 +669,17 @@ we_slept:
 	spin_unlock(&dq_list_lock);
 }
 
+struct dquot *dquot_alloc(struct super_block *sb, int type)
+{
+	return kmem_cache_zalloc(dquot_cachep, GFP_NOFS);
+}
+EXPORT_SYMBOL(dquot_alloc);
+
 static struct dquot *get_empty_dquot(struct super_block *sb, int type)
 {
 	struct dquot *dquot;
 
-	dquot = kmem_cache_zalloc(dquot_cachep, GFP_NOFS);
+	dquot = sb->dq_op->alloc_dquot(sb, type);
 	if(!dquot)
 		return NODQUOT;
 
@@ -647,15 +697,33 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type)
 }
 
 /*
+ * Check whether dquot is in memory.
+ * MUST be called with either dqptr_sem or dqonoff_mutex held
+ */
+int dquot_is_cached(struct super_block *sb, unsigned int id, int type)
+{
+	unsigned int hashent = hashfn(sb, id, type);
+	int ret = 0;
+
+        if (!sb_has_quota_active(sb, type))
+		return 0;
+	spin_lock(&dq_list_lock);
+	if (find_dquot(hashent, sb, id, type) != NODQUOT)
+		ret = 1;
+	spin_unlock(&dq_list_lock);
+	return ret;
+}
+
+/*
  * Get reference to dquot
  * MUST be called with either dqptr_sem or dqonoff_mutex held
  */
-static struct dquot *dqget(struct super_block *sb, unsigned int id, int type)
+struct dquot *dqget(struct super_block *sb, unsigned int id, int type)
 {
 	unsigned int hashent = hashfn(sb, id, type);
 	struct dquot *dquot, *empty = NODQUOT;
 
-        if (!sb_has_quota_enabled(sb, type))
+        if (!sb_has_quota_active(sb, type))
 		return NODQUOT;
 we_slept:
 	spin_lock(&dq_list_lock);
@@ -682,7 +750,7 @@ we_slept:
 		dqstats.lookups++;
 		spin_unlock(&dq_list_lock);
 		if (empty)
-			kmem_cache_free(dquot_cachep, empty);
+			do_destroy_dquot(empty);
 	}
 	/* Wait for dq_lock - after this we know that either dquot_release() is already
 	 * finished or it will be canceled due to dq_count > 1 test */
@@ -820,7 +888,7 @@ static void drop_dquot_ref(struct super_block *sb, int type)
 	}
 }
 
-static inline void dquot_incr_inodes(struct dquot *dquot, unsigned long number)
+static inline void dquot_incr_inodes(struct dquot *dquot, qsize_t number)
 {
 	dquot->dq_dqb.dqb_curinodes += number;
 }
@@ -830,9 +898,10 @@ static inline void dquot_incr_space(struct dquot *dquot, qsize_t number)
 	dquot->dq_dqb.dqb_curspace += number;
 }
 
-static inline void dquot_decr_inodes(struct dquot *dquot, unsigned long number)
+static inline void dquot_decr_inodes(struct dquot *dquot, qsize_t number)
 {
-	if (dquot->dq_dqb.dqb_curinodes > number)
+	if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NEGATIVE_USAGE ||
+	    dquot->dq_dqb.dqb_curinodes >= number)
 		dquot->dq_dqb.dqb_curinodes -= number;
 	else
 		dquot->dq_dqb.dqb_curinodes = 0;
@@ -843,11 +912,12 @@ static inline void dquot_decr_inodes(struct dquot *dquot, unsigned long number)
 
 static inline void dquot_decr_space(struct dquot *dquot, qsize_t number)
 {
-	if (dquot->dq_dqb.dqb_curspace > number)
+	if (sb_dqopt(dquot->dq_sb)->flags & DQUOT_NEGATIVE_USAGE ||
+	    dquot->dq_dqb.dqb_curspace >= number)
 		dquot->dq_dqb.dqb_curspace -= number;
 	else
 		dquot->dq_dqb.dqb_curspace = 0;
-	if (toqb(dquot->dq_dqb.dqb_curspace) <= dquot->dq_dqb.dqb_bsoftlimit)
+	if (dquot->dq_dqb.dqb_curspace <= dquot->dq_dqb.dqb_bsoftlimit)
 		dquot->dq_dqb.dqb_btime = (time_t) 0;
 	clear_bit(DQ_BLKS_B, &dquot->dq_flags);
 }
@@ -1023,10 +1093,11 @@ static inline char ignore_hardlimit(struct dquot *dquot)
 }
 
 /* needs dq_data_lock */
-static int check_idq(struct dquot *dquot, ulong inodes, char *warntype)
+static int check_idq(struct dquot *dquot, qsize_t inodes, char *warntype)
 {
 	*warntype = QUOTA_NL_NOWARN;
-	if (inodes <= 0 || test_bit(DQ_FAKE_B, &dquot->dq_flags))
+	if (!sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type) ||
+	    test_bit(DQ_FAKE_B, &dquot->dq_flags))
 		return QUOTA_OK;
 
 	if (dquot->dq_dqb.dqb_ihardlimit &&
@@ -1058,11 +1129,12 @@ static int check_idq(struct dquot *dquot, ulong inodes, char *warntype)
 static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *warntype)
 {
 	*warntype = QUOTA_NL_NOWARN;
-	if (space <= 0 || test_bit(DQ_FAKE_B, &dquot->dq_flags))
+	if (!sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type) ||
+	    test_bit(DQ_FAKE_B, &dquot->dq_flags))
 		return QUOTA_OK;
 
 	if (dquot->dq_dqb.dqb_bhardlimit &&
-	   toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bhardlimit &&
+	    dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bhardlimit &&
             !ignore_hardlimit(dquot)) {
 		if (!prealloc)
 			*warntype = QUOTA_NL_BHARDWARN;
@@ -1070,7 +1142,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war
 	}
 
 	if (dquot->dq_dqb.dqb_bsoftlimit &&
-	   toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bsoftlimit &&
+	    dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bsoftlimit &&
 	    dquot->dq_dqb.dqb_btime && get_seconds() >= dquot->dq_dqb.dqb_btime &&
             !ignore_hardlimit(dquot)) {
 		if (!prealloc)
@@ -1079,7 +1151,7 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war
 	}
 
 	if (dquot->dq_dqb.dqb_bsoftlimit &&
-	   toqb(dquot->dq_dqb.dqb_curspace + space) > dquot->dq_dqb.dqb_bsoftlimit &&
+	    dquot->dq_dqb.dqb_curspace + space > dquot->dq_dqb.dqb_bsoftlimit &&
 	    dquot->dq_dqb.dqb_btime == 0) {
 		if (!prealloc) {
 			*warntype = QUOTA_NL_BSOFTWARN;
@@ -1096,10 +1168,11 @@ static int check_bdq(struct dquot *dquot, qsize_t space, int prealloc, char *war
 	return QUOTA_OK;
 }
 
-static int info_idq_free(struct dquot *dquot, ulong inodes)
+static int info_idq_free(struct dquot *dquot, qsize_t inodes)
 {
 	if (test_bit(DQ_FAKE_B, &dquot->dq_flags) ||
-	    dquot->dq_dqb.dqb_curinodes <= dquot->dq_dqb.dqb_isoftlimit)
+	    dquot->dq_dqb.dqb_curinodes <= dquot->dq_dqb.dqb_isoftlimit ||
+	    !sb_has_quota_limits_enabled(dquot->dq_sb, dquot->dq_type))
 		return QUOTA_NL_NOWARN;
 
 	if (dquot->dq_dqb.dqb_curinodes - inodes <= dquot->dq_dqb.dqb_isoftlimit)
@@ -1113,15 +1186,13 @@ static int info_idq_free(struct dquot *dquot, ulong inodes)
 static int info_bdq_free(struct dquot *dquot, qsize_t space)
 {
 	if (test_bit(DQ_FAKE_B, &dquot->dq_flags) ||
-	    toqb(dquot->dq_dqb.dqb_curspace) <= dquot->dq_dqb.dqb_bsoftlimit)
+	    dquot->dq_dqb.dqb_curspace <= dquot->dq_dqb.dqb_bsoftlimit)
 		return QUOTA_NL_NOWARN;
 
-	if (toqb(dquot->dq_dqb.dqb_curspace - space) <=
-	    dquot->dq_dqb.dqb_bsoftlimit)
+	if (dquot->dq_dqb.dqb_curspace - space <= dquot->dq_dqb.dqb_bsoftlimit)
 		return QUOTA_NL_BSOFTBELOW;
-	if (toqb(dquot->dq_dqb.dqb_curspace) >= dquot->dq_dqb.dqb_bhardlimit &&
-	    toqb(dquot->dq_dqb.dqb_curspace - space) <
-						dquot->dq_dqb.dqb_bhardlimit)
+	if (dquot->dq_dqb.dqb_curspace >= dquot->dq_dqb.dqb_bhardlimit &&
+	    dquot->dq_dqb.dqb_curspace - space < dquot->dq_dqb.dqb_bhardlimit)
 		return QUOTA_NL_BHARDBELOW;
 	return QUOTA_NL_NOWARN;
 }
@@ -1166,17 +1237,23 @@ out_err:
  * 	Release all quotas referenced by inode
  *	Transaction must be started at an entry
  */
-int dquot_drop(struct inode *inode)
+int dquot_drop_locked(struct inode *inode)
 {
 	int cnt;
 
-	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (inode->i_dquot[cnt] != NODQUOT) {
 			dqput(inode->i_dquot[cnt]);
 			inode->i_dquot[cnt] = NODQUOT;
 		}
 	}
+	return 0;
+}
+
+int dquot_drop(struct inode *inode)
+{
+	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
+	dquot_drop_locked(inode);
 	up_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
 	return 0;
 }
@@ -1264,7 +1341,7 @@ warn_put_all:
 /*
  * This operation can block, but only after everything is updated
  */
-int dquot_alloc_inode(const struct inode *inode, unsigned long number)
+int dquot_alloc_inode(const struct inode *inode, qsize_t number)
 {
 	int cnt, ret = NO_QUOTA;
 	char warntype[MAXQUOTAS];
@@ -1349,7 +1426,7 @@ out_sub:
 /*
  * This operation can block, but only after everything is updated
  */
-int dquot_free_inode(const struct inode *inode, unsigned long number)
+int dquot_free_inode(const struct inode *inode, qsize_t number)
 {
 	unsigned int cnt;
 	char warntype[MAXQUOTAS];
@@ -1495,7 +1572,7 @@ warn_put_all:
 /* Wrapper for transferring ownership of an inode */
 int vfs_dq_transfer(struct inode *inode, struct iattr *iattr)
 {
-	if (sb_any_quota_enabled(inode->i_sb) && !IS_NOQUOTA(inode)) {
+	if (sb_any_quota_active(inode->i_sb) && !IS_NOQUOTA(inode)) {
 		vfs_dq_init(inode);
 		if (inode->i_sb->dq_op->transfer(inode, iattr) == NO_QUOTA)
 			return 1;
@@ -1533,54 +1610,27 @@ struct dquot_operations dquot_operations = {
 	.acquire_dquot	= dquot_acquire,
 	.release_dquot	= dquot_release,
 	.mark_dirty	= dquot_mark_dquot_dirty,
-	.write_info	= dquot_commit_info
+	.write_info	= dquot_commit_info,
+	.alloc_dquot	= dquot_alloc,
+	.destroy_dquot	= dquot_destroy,
 };
 
-static inline void set_enable_flags(struct quota_info *dqopt, int type)
-{
-	switch (type) {
-		case USRQUOTA:
-			dqopt->flags |= DQUOT_USR_ENABLED;
-			dqopt->flags &= ~DQUOT_USR_SUSPENDED;
-			break;
-		case GRPQUOTA:
-			dqopt->flags |= DQUOT_GRP_ENABLED;
-			dqopt->flags &= ~DQUOT_GRP_SUSPENDED;
-			break;
-	}
-}
-
-static inline void reset_enable_flags(struct quota_info *dqopt, int type,
-				      int remount)
-{
-	switch (type) {
-		case USRQUOTA:
-			dqopt->flags &= ~DQUOT_USR_ENABLED;
-			if (remount)
-				dqopt->flags |= DQUOT_USR_SUSPENDED;
-			else
-				dqopt->flags &= ~DQUOT_USR_SUSPENDED;
-			break;
-		case GRPQUOTA:
-			dqopt->flags &= ~DQUOT_GRP_ENABLED;
-			if (remount)
-				dqopt->flags |= DQUOT_GRP_SUSPENDED;
-			else
-				dqopt->flags &= ~DQUOT_GRP_SUSPENDED;
-			break;
-	}
-}
-
-
 /*
  * Turn quota off on a device. type == -1 ==> quotaoff for all types (umount)
  */
-int vfs_quota_off(struct super_block *sb, int type, int remount)
+int vfs_quota_disable(struct super_block *sb, int type, unsigned int flags)
 {
 	int cnt, ret = 0;
 	struct quota_info *dqopt = sb_dqopt(sb);
 	struct inode *toputinode[MAXQUOTAS];
 
+	/* Cannot turn off usage accounting without turning off limits, or
+	 * suspend quotas and simultaneously turn quotas off. */
+	if ((flags & DQUOT_USAGE_ENABLED && !(flags & DQUOT_LIMITS_ENABLED))
+	    || (flags & DQUOT_SUSPENDED && flags & (DQUOT_LIMITS_ENABLED |
+	    DQUOT_USAGE_ENABLED)))
+		return -EINVAL;
+
 	/* We need to serialize quota_off() for device */
 	mutex_lock(&dqopt->dqonoff_mutex);
 
@@ -1589,7 +1639,7 @@ int vfs_quota_off(struct super_block *sb, int type, int remount)
 	 * sometimes we are called when fill_super() failed and calling
 	 * sync_fs() in such cases does no good.
 	 */
-	if (!sb_any_quota_enabled(sb) && !sb_any_quota_suspended(sb)) {
+	if (!sb_any_quota_loaded(sb)) {
 		mutex_unlock(&dqopt->dqonoff_mutex);
 		return 0;
 	}
@@ -1597,17 +1647,28 @@ int vfs_quota_off(struct super_block *sb, int type, int remount)
 		toputinode[cnt] = NULL;
 		if (type != -1 && cnt != type)
 			continue;
-		/* If we keep inodes of quota files after remount and quotaoff
-		 * is called, drop kept inodes. */
-		if (!remount && sb_has_quota_suspended(sb, cnt)) {
-			iput(dqopt->files[cnt]);
-			dqopt->files[cnt] = NULL;
-			reset_enable_flags(dqopt, cnt, 0);
+		if (!sb_has_quota_loaded(sb, cnt))
 			continue;
+
+		if (flags & DQUOT_SUSPENDED) {
+			dqopt->flags |=
+				dquot_state_flag(DQUOT_SUSPENDED, cnt);
+		} else {
+			dqopt->flags &= ~dquot_state_flag(flags, cnt);
+			/* Turning off suspended quotas? */
+			if (!sb_has_quota_loaded(sb, cnt) &&
+			    sb_has_quota_suspended(sb, cnt)) {
+				dqopt->flags &=	~dquot_state_flag(
+							DQUOT_SUSPENDED, cnt);
+				iput(dqopt->files[cnt]);
+				dqopt->files[cnt] = NULL;
+				continue;
+			}
 		}
-		if (!sb_has_quota_enabled(sb, cnt))
+
+		/* We still have to keep quota loaded? */
+		if (sb_has_quota_loaded(sb, cnt) && !(flags & DQUOT_SUSPENDED))
 			continue;
-		reset_enable_flags(dqopt, cnt, remount);
 
 		/* Note: these are blocking operations */
 		drop_dquot_ref(sb, cnt);
@@ -1623,7 +1684,7 @@ int vfs_quota_off(struct super_block *sb, int type, int remount)
 		put_quota_format(dqopt->info[cnt].dqi_format);
 
 		toputinode[cnt] = dqopt->files[cnt];
-		if (!remount)
+		if (!sb_has_quota_loaded(sb, cnt))
 			dqopt->files[cnt] = NULL;
 		dqopt->info[cnt].dqi_flags = 0;
 		dqopt->info[cnt].dqi_igrace = 0;
@@ -1631,6 +1692,11 @@ int vfs_quota_off(struct super_block *sb, int type, int remount)
 		dqopt->ops[cnt] = NULL;
 	}
 	mutex_unlock(&dqopt->dqonoff_mutex);
+
+	/* Skip syncing and setting flags if quota files are hidden */
+	if (dqopt->flags & DQUOT_QUOTA_SYS_FILE)
+		goto put_inodes;
+
 	/* Sync the superblock so that buffers with quota data are written to
 	 * disk (and so userspace sees correct data afterwards). */
 	if (sb->s_op->sync_fs)
@@ -1646,7 +1712,7 @@ int vfs_quota_off(struct super_block *sb, int type, int remount)
 			mutex_lock(&dqopt->dqonoff_mutex);
 			/* If quota was reenabled in the meantime, we have
 			 * nothing to do */
-			if (!sb_has_quota_enabled(sb, cnt)) {
+			if (!sb_has_quota_loaded(sb, cnt)) {
 				mutex_lock_nested(&toputinode[cnt]->i_mutex, I_MUTEX_QUOTA);
 				toputinode[cnt]->i_flags &= ~(S_IMMUTABLE |
 				  S_NOATIME | S_NOQUOTA);
@@ -1655,26 +1721,43 @@ int vfs_quota_off(struct super_block *sb, int type, int remount)
 				mark_inode_dirty(toputinode[cnt]);
 			}
 			mutex_unlock(&dqopt->dqonoff_mutex);
+		}
+	if (sb->s_bdev)
+		invalidate_bdev(sb->s_bdev);
+put_inodes:
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
+		if (toputinode[cnt]) {
 			/* On remount RO, we keep the inode pointer so that we
-			 * can reenable quota on the subsequent remount RW.
-			 * But we have better not keep inode pointer when there
-			 * is pending delete on the quota file... */
-			if (!remount)
+			 * can reenable quota on the subsequent remount RW. We
+			 * have to check 'flags' variable and not use sb_has_
+			 * function because another quotaon / quotaoff could
+			 * change global state before we got here. We refuse
+			 * to suspend quotas when there is pending delete on
+			 * the quota file... */
+			if (!(flags & DQUOT_SUSPENDED))
 				iput(toputinode[cnt]);
 			else if (!toputinode[cnt]->i_nlink)
 				ret = -EBUSY;
 		}
-	if (sb->s_bdev)
-		invalidate_bdev(sb->s_bdev);
 	return ret;
 }
 
+int vfs_quota_off(struct super_block *sb, int type, int remount)
+{
+	return vfs_quota_disable(sb, type, remount ? DQUOT_SUSPENDED :
+				 (DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED));
+}
+
 /*
  *	Turn quotas on on a device
  */
 
-/* Helper function when we already have the inode */
-static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
+/*
+ * Helper function to turn quotas on when we already have the inode of
+ * quota file and no quota information is loaded.
+ */
+static int vfs_load_quota_inode(struct inode *inode, int type, int format_id,
+	unsigned int flags)
 {
 	struct quota_format_type *fmt = find_quota_format(format_id);
 	struct super_block *sb = inode->i_sb;
@@ -1696,27 +1779,37 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
 		error = -EINVAL;
 		goto out_fmt;
 	}
+	/* Usage always has to be set... */
+	if (!(flags & DQUOT_USAGE_ENABLED)) {
+		error = -EINVAL;
+		goto out_fmt;
+	}
 
-	/* As we bypass the pagecache we must now flush the inode so that
-	 * we see all the changes from userspace... */
-	write_inode_now(inode, 1);
-	/* And now flush the block cache so that kernel sees the changes */
-	invalidate_bdev(sb->s_bdev);
+	if (!(dqopt->flags & DQUOT_QUOTA_SYS_FILE)) {
+		/* As we bypass the pagecache we must now flush the inode so
+		 * that we see all the changes from userspace... */
+		write_inode_now(inode, 1);
+		/* And now flush the block cache so that kernel sees the
+		 * changes */
+		invalidate_bdev(sb->s_bdev);
+	}
 	mutex_lock(&inode->i_mutex);
 	mutex_lock(&dqopt->dqonoff_mutex);
-	if (sb_has_quota_enabled(sb, type) ||
-			sb_has_quota_suspended(sb, type)) {
+	if (sb_has_quota_loaded(sb, type)) {
 		error = -EBUSY;
 		goto out_lock;
 	}
-	/* We don't want quota and atime on quota files (deadlocks possible)
-	 * Also nobody should write to the file - we use special IO operations
-	 * which ignore the immutable bit. */
-	down_write(&dqopt->dqptr_sem);
-	oldflags = inode->i_flags & (S_NOATIME | S_IMMUTABLE | S_NOQUOTA);
-	inode->i_flags |= S_NOQUOTA | S_NOATIME | S_IMMUTABLE;
-	up_write(&dqopt->dqptr_sem);
-	sb->dq_op->drop(inode);
+
+	if (!(dqopt->flags & DQUOT_QUOTA_SYS_FILE)) {
+		/* We don't want quota and atime on quota files (deadlocks
+		 * possible) Also nobody should write to the file - we use
+		 * special IO operations which ignore the immutable bit. */
+		down_write(&dqopt->dqptr_sem);
+		oldflags = inode->i_flags & (S_NOATIME | S_IMMUTABLE | S_NOQUOTA);
+		inode->i_flags |= S_NOQUOTA | S_NOATIME | S_IMMUTABLE;
+		up_write(&dqopt->dqptr_sem);
+		sb->dq_op->drop(inode);
+	}
 
 	error = -EIO;
 	dqopt->files[type] = igrab(inode);
@@ -1737,7 +1830,7 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
 	}
 	mutex_unlock(&dqopt->dqio_mutex);
 	mutex_unlock(&inode->i_mutex);
-	set_enable_flags(dqopt, type);
+	dqopt->flags |= dquot_state_flag(flags, type);
 
 	add_dquot_ref(sb, type);
 	mutex_unlock(&dqopt->dqonoff_mutex);
@@ -1770,20 +1863,23 @@ static int vfs_quota_on_remount(struct super_block *sb, int type)
 	struct quota_info *dqopt = sb_dqopt(sb);
 	struct inode *inode;
 	int ret;
+	unsigned int flags;
 
 	mutex_lock(&dqopt->dqonoff_mutex);
 	if (!sb_has_quota_suspended(sb, type)) {
 		mutex_unlock(&dqopt->dqonoff_mutex);
 		return 0;
 	}
-	BUG_ON(sb_has_quota_enabled(sb, type));
-
 	inode = dqopt->files[type];
 	dqopt->files[type] = NULL;
-	reset_enable_flags(dqopt, type, 0);
+	flags = dqopt->flags & dquot_state_flag(DQUOT_USAGE_ENABLED |
+						DQUOT_LIMITS_ENABLED, type);
+	dqopt->flags &= ~dquot_state_flag(DQUOT_STATE_FLAGS, type);
 	mutex_unlock(&dqopt->dqonoff_mutex);
 
-	ret = vfs_quota_on_inode(inode, type, dqopt->info[type].dqi_fmt_id);
+	flags = dquot_generic_flag(flags, type);
+	ret = vfs_load_quota_inode(inode, type, dqopt->info[type].dqi_fmt_id,
+				   flags);
 	iput(inode);
 
 	return ret;
@@ -1799,12 +1895,12 @@ int vfs_quota_on_path(struct super_block *sb, int type, int format_id,
 	if (path->mnt->mnt_sb != sb)
 		error = -EXDEV;
 	else
-		error = vfs_quota_on_inode(path->dentry->d_inode, type,
-					   format_id);
+		error = vfs_load_quota_inode(path->dentry->d_inode, type,
+					     format_id, DQUOT_USAGE_ENABLED |
+					     DQUOT_LIMITS_ENABLED);
 	return error;
 }
 
-/* Actual function called from quotactl() */
 int vfs_quota_on(struct super_block *sb, int type, int format_id, char *name,
 		 int remount)
 {
@@ -1823,6 +1919,50 @@ int vfs_quota_on(struct super_block *sb, int type, int format_id, char *name,
 }
 
 /*
+ * More powerful function for turning on quotas allowing setting
+ * of individual quota flags
+ */
+int vfs_quota_enable(struct inode *inode, int type, int format_id,
+		unsigned int flags)
+{
+	int ret = 0;
+	struct super_block *sb = inode->i_sb;
+	struct quota_info *dqopt = sb_dqopt(sb);
+
+	/* Just unsuspend quotas? */
+	if (flags & DQUOT_SUSPENDED)
+		return vfs_quota_on_remount(sb, type);
+	if (!flags)
+		return 0;
+	/* Just updating flags needed? */
+	if (sb_has_quota_loaded(sb, type)) {
+		mutex_lock(&dqopt->dqonoff_mutex);
+		/* Now do a reliable test... */
+		if (!sb_has_quota_loaded(sb, type)) {
+			mutex_unlock(&dqopt->dqonoff_mutex);
+			goto load_quota;
+		}
+		if (flags & DQUOT_USAGE_ENABLED &&
+		    sb_has_quota_usage_enabled(sb, type)) {
+			ret = -EBUSY;
+			goto out_lock;
+		}
+		if (flags & DQUOT_LIMITS_ENABLED &&
+		    sb_has_quota_limits_enabled(sb, type)) {
+			ret = -EBUSY;
+			goto out_lock;
+		}
+		sb_dqopt(sb)->flags |= dquot_state_flag(flags, type);
+out_lock:
+		mutex_unlock(&dqopt->dqonoff_mutex);
+		return ret;
+	}
+
+load_quota:
+	return vfs_load_quota_inode(inode, type, format_id, flags);
+}
+
+/*
  * This function is used when filesystem needs to initialize quotas
  * during mount time.
  */
@@ -1843,7 +1983,8 @@ int vfs_quota_on_mount(struct super_block *sb, char *qf_name,
 
 	error = security_quota_on(dentry);
 	if (!error)
-		error = vfs_quota_on_inode(dentry->d_inode, type, format_id);
+		error = vfs_load_quota_inode(dentry->d_inode, type, format_id,
+				DQUOT_USAGE_ENABLED | DQUOT_LIMITS_ENABLED);
 
 out:
 	dput(dentry);
@@ -1866,14 +2007,24 @@ int vfs_dq_quota_on_remount(struct super_block *sb)
 	return ret;
 }
 
+static inline qsize_t qbtos(qsize_t blocks)
+{
+	return blocks << QIF_DQBLKSIZE_BITS;
+}
+
+static inline qsize_t stoqb(qsize_t space)
+{
+	return (space + QIF_DQBLKSIZE - 1) >> QIF_DQBLKSIZE_BITS;
+}
+
 /* Generic routine for getting common part of quota structure */
 static void do_get_dqblk(struct dquot *dquot, struct if_dqblk *di)
 {
 	struct mem_dqblk *dm = &dquot->dq_dqb;
 
 	spin_lock(&dq_data_lock);
-	di->dqb_bhardlimit = dm->dqb_bhardlimit;
-	di->dqb_bsoftlimit = dm->dqb_bsoftlimit;
+	di->dqb_bhardlimit = stoqb(dm->dqb_bhardlimit);
+	di->dqb_bsoftlimit = stoqb(dm->dqb_bsoftlimit);
 	di->dqb_curspace = dm->dqb_curspace;
 	di->dqb_ihardlimit = dm->dqb_ihardlimit;
 	di->dqb_isoftlimit = dm->dqb_isoftlimit;
@@ -1918,28 +2069,38 @@ static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di)
 	if (di->dqb_valid & QIF_SPACE) {
 		dm->dqb_curspace = di->dqb_curspace;
 		check_blim = 1;
+		__set_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags);
 	}
 	if (di->dqb_valid & QIF_BLIMITS) {
-		dm->dqb_bsoftlimit = di->dqb_bsoftlimit;
-		dm->dqb_bhardlimit = di->dqb_bhardlimit;
+		dm->dqb_bsoftlimit = qbtos(di->dqb_bsoftlimit);
+		dm->dqb_bhardlimit = qbtos(di->dqb_bhardlimit);
 		check_blim = 1;
+		__set_bit(DQ_LASTSET_B + QIF_BLIMITS_B, &dquot->dq_flags);
 	}
 	if (di->dqb_valid & QIF_INODES) {
 		dm->dqb_curinodes = di->dqb_curinodes;
 		check_ilim = 1;
+		__set_bit(DQ_LASTSET_B + QIF_INODES_B, &dquot->dq_flags);
 	}
 	if (di->dqb_valid & QIF_ILIMITS) {
 		dm->dqb_isoftlimit = di->dqb_isoftlimit;
 		dm->dqb_ihardlimit = di->dqb_ihardlimit;
 		check_ilim = 1;
+		__set_bit(DQ_LASTSET_B + QIF_ILIMITS_B, &dquot->dq_flags);
 	}
-	if (di->dqb_valid & QIF_BTIME)
+	if (di->dqb_valid & QIF_BTIME) {
 		dm->dqb_btime = di->dqb_btime;
-	if (di->dqb_valid & QIF_ITIME)
+		check_blim = 1;
+		__set_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags);
+	}
+	if (di->dqb_valid & QIF_ITIME) {
 		dm->dqb_itime = di->dqb_itime;
+		check_ilim = 1;
+		__set_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags);
+	}
 
 	if (check_blim) {
-		if (!dm->dqb_bsoftlimit || toqb(dm->dqb_curspace) < dm->dqb_bsoftlimit) {
+		if (!dm->dqb_bsoftlimit || dm->dqb_curspace < dm->dqb_bsoftlimit) {
 			dm->dqb_btime = 0;
 			clear_bit(DQ_BLKS_B, &dquot->dq_flags);
 		}
@@ -1970,12 +2131,14 @@ int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d
 	int rc;
 
 	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
-	if (!(dquot = dqget(sb, id, type))) {
-		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
-		return -ESRCH;
+	dquot = dqget(sb, id, type);
+	if (!dquot) {
+		rc = -ESRCH;
+		goto out;
 	}
 	rc = do_set_dqblk(dquot, di);
 	dqput(dquot);
+out:
 	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return rc;
 }
@@ -1986,7 +2149,7 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	struct mem_dqinfo *mi;
   
 	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
-	if (!sb_has_quota_enabled(sb, type)) {
+	if (!sb_has_quota_active(sb, type)) {
 		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
@@ -2005,11 +2168,12 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
+	int err = 0;
 
 	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
-	if (!sb_has_quota_enabled(sb, type)) {
-		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
-		return -ESRCH;
+	if (!sb_has_quota_active(sb, type)) {
+		err = -ESRCH;
+		goto out;
 	}
 	mi = sb_dqopt(sb)->info + type;
 	spin_lock(&dq_data_lock);
@@ -2023,8 +2187,9 @@ int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	mark_info_dirty(sb, type);
 	/* Force write to disk */
 	sb->dq_op->write_info(sb, type);
+out:
 	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
-	return 0;
+	return err;
 }
 
 struct quotactl_ops vfs_quotactl_ops = {
@@ -2186,10 +2351,13 @@ EXPORT_SYMBOL(register_quota_format);
 EXPORT_SYMBOL(unregister_quota_format);
 EXPORT_SYMBOL(dqstats);
 EXPORT_SYMBOL(dq_data_lock);
+EXPORT_SYMBOL(vfs_quota_enable);
 EXPORT_SYMBOL(vfs_quota_on);
 EXPORT_SYMBOL(vfs_quota_on_path);
 EXPORT_SYMBOL(vfs_quota_on_mount);
+EXPORT_SYMBOL(vfs_quota_disable);
 EXPORT_SYMBOL(vfs_quota_off);
+EXPORT_SYMBOL(dquot_scan_active);
 EXPORT_SYMBOL(vfs_quota_sync);
 EXPORT_SYMBOL(vfs_get_dqinfo);
 EXPORT_SYMBOL(vfs_set_dqinfo);
@@ -2202,7 +2370,11 @@ EXPORT_SYMBOL(dquot_release);
 EXPORT_SYMBOL(dquot_mark_dquot_dirty);
 EXPORT_SYMBOL(dquot_initialize);
 EXPORT_SYMBOL(dquot_drop);
+EXPORT_SYMBOL(dquot_drop_locked);
 EXPORT_SYMBOL(vfs_dq_drop);
+EXPORT_SYMBOL(dqget);
+EXPORT_SYMBOL(dqput);
+EXPORT_SYMBOL(dquot_is_cached);
 EXPORT_SYMBOL(dquot_alloc_space);
 EXPORT_SYMBOL(dquot_alloc_inode);
 EXPORT_SYMBOL(dquot_free_space);
diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c
index 6046239465a..c01e043670e 100644
--- a/fs/ecryptfs/crypto.c
+++ b/fs/ecryptfs/crypto.c
@@ -175,8 +175,8 @@ out:
  *
  * Returns zero on success; non-zero on error.
  */
-static int ecryptfs_derive_iv(char *iv, struct ecryptfs_crypt_stat *crypt_stat,
-			      loff_t offset)
+int ecryptfs_derive_iv(char *iv, struct ecryptfs_crypt_stat *crypt_stat,
+		       loff_t offset)
 {
 	int rc = 0;
 	char dst[MD5_DIGEST_SIZE];
@@ -924,6 +924,15 @@ static void ecryptfs_copy_mount_wide_flags_to_inode_flags(
 		crypt_stat->flags |= ECRYPTFS_METADATA_IN_XATTR;
 	if (mount_crypt_stat->flags & ECRYPTFS_ENCRYPTED_VIEW_ENABLED)
 		crypt_stat->flags |= ECRYPTFS_VIEW_AS_ENCRYPTED;
+	if (mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES) {
+		crypt_stat->flags |= ECRYPTFS_ENCRYPT_FILENAMES;
+		if (mount_crypt_stat->flags
+		    & ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK)
+			crypt_stat->flags |= ECRYPTFS_ENCFN_USE_MOUNT_FNEK;
+		else if (mount_crypt_stat->flags
+			 & ECRYPTFS_GLOBAL_ENCFN_USE_FEK)
+			crypt_stat->flags |= ECRYPTFS_ENCFN_USE_FEK;
+	}
 }
 
 static int ecryptfs_copy_mount_wide_sigs_to_inode_sigs(
@@ -1060,7 +1069,8 @@ struct ecryptfs_flag_map_elem {
 static struct ecryptfs_flag_map_elem ecryptfs_flag_map[] = {
 	{0x00000001, ECRYPTFS_ENABLE_HMAC},
 	{0x00000002, ECRYPTFS_ENCRYPTED},
-	{0x00000004, ECRYPTFS_METADATA_IN_XATTR}
+	{0x00000004, ECRYPTFS_METADATA_IN_XATTR},
+	{0x00000008, ECRYPTFS_ENCRYPT_FILENAMES}
 };
 
 /**
@@ -1149,19 +1159,20 @@ ecryptfs_cipher_code_str_map[] = {
 
 /**
  * ecryptfs_code_for_cipher_string
- * @crypt_stat: The cryptographic context
+ * @cipher_name: The string alias for the cipher
+ * @key_bytes: Length of key in bytes; used for AES code selection
  *
  * Returns zero on no match, or the cipher code on match
  */
-u8 ecryptfs_code_for_cipher_string(struct ecryptfs_crypt_stat *crypt_stat)
+u8 ecryptfs_code_for_cipher_string(char *cipher_name, size_t key_bytes)
 {
 	int i;
 	u8 code = 0;
 	struct ecryptfs_cipher_code_str_map_elem *map =
 		ecryptfs_cipher_code_str_map;
 
-	if (strcmp(crypt_stat->cipher, "aes") == 0) {
-		switch (crypt_stat->key_size) {
+	if (strcmp(cipher_name, "aes") == 0) {
+		switch (key_bytes) {
 		case 16:
 			code = RFC2440_CIPHER_AES_128;
 			break;
@@ -1173,7 +1184,7 @@ u8 ecryptfs_code_for_cipher_string(struct ecryptfs_crypt_stat *crypt_stat)
 		}
 	} else {
 		for (i = 0; i < ARRAY_SIZE(ecryptfs_cipher_code_str_map); i++)
-			if (strcmp(crypt_stat->cipher, map[i].cipher_str) == 0){
+			if (strcmp(cipher_name, map[i].cipher_str) == 0) {
 				code = map[i].cipher_code;
 				break;
 			}
@@ -1212,6 +1223,8 @@ int ecryptfs_read_and_validate_header_region(char *data,
 		&(ecryptfs_inode_to_private(ecryptfs_inode)->crypt_stat);
 	int rc;
 
+	if (crypt_stat->extent_size == 0)
+		crypt_stat->extent_size = ECRYPTFS_DEFAULT_EXTENT_SIZE;
 	rc = ecryptfs_read_lower(data, 0, crypt_stat->extent_size,
 				 ecryptfs_inode);
 	if (rc) {
@@ -1221,7 +1234,6 @@ int ecryptfs_read_and_validate_header_region(char *data,
 	}
 	if (!contains_ecryptfs_marker(data + ECRYPTFS_FILE_SIZE_BYTES)) {
 		rc = -EINVAL;
-		ecryptfs_printk(KERN_DEBUG, "Valid marker not found\n");
 	}
 out:
 	return rc;
@@ -1628,95 +1640,95 @@ out:
 }
 
 /**
- * ecryptfs_encode_filename - converts a plaintext file name to cipher text
- * @crypt_stat: The crypt_stat struct associated with the file anem to encode
- * @name: The plaintext name
- * @length: The length of the plaintext
- * @encoded_name: The encypted name
+ * ecryptfs_encrypt_filename - encrypt filename
  *
- * Encrypts and encodes a filename into something that constitutes a
- * valid filename for a filesystem, with printable characters.
+ * CBC-encrypts the filename. We do not want to encrypt the same
+ * filename with the same key and IV, which may happen with hard
+ * links, so we prepend random bits to each filename.
  *
- * We assume that we have a properly initialized crypto context,
- * pointed to by crypt_stat->tfm.
- *
- * TODO: Implement filename decoding and decryption here, in place of
- * memcpy. We are keeping the framework around for now to (1)
- * facilitate testing of the components needed to implement filename
- * encryption and (2) to provide a code base from which other
- * developers in the community can easily implement this feature.
- *
- * Returns the length of encoded filename; negative if error
+ * Returns zero on success; non-zero otherwise
  */
-int
-ecryptfs_encode_filename(struct ecryptfs_crypt_stat *crypt_stat,
-			 const char *name, int length, char **encoded_name)
+static int
+ecryptfs_encrypt_filename(struct ecryptfs_filename *filename,
+			  struct ecryptfs_crypt_stat *crypt_stat,
+			  struct ecryptfs_mount_crypt_stat *mount_crypt_stat)
 {
-	int error = 0;
+	int rc = 0;
 
-	(*encoded_name) = kmalloc(length + 2, GFP_KERNEL);
-	if (!(*encoded_name)) {
-		error = -ENOMEM;
+	filename->encrypted_filename = NULL;
+	filename->encrypted_filename_size = 0;
+	if ((crypt_stat && (crypt_stat->flags & ECRYPTFS_ENCFN_USE_MOUNT_FNEK))
+	    || (mount_crypt_stat && (mount_crypt_stat->flags
+				     & ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK))) {
+		size_t packet_size;
+		size_t remaining_bytes;
+
+		rc = ecryptfs_write_tag_70_packet(
+			NULL, NULL,
+			&filename->encrypted_filename_size,
+			mount_crypt_stat, NULL,
+			filename->filename_size);
+		if (rc) {
+			printk(KERN_ERR "%s: Error attempting to get packet "
+			       "size for tag 72; rc = [%d]\n", __func__,
+			       rc);
+			filename->encrypted_filename_size = 0;
+			goto out;
+		}
+		filename->encrypted_filename =
+			kmalloc(filename->encrypted_filename_size, GFP_KERNEL);
+		if (!filename->encrypted_filename) {
+			printk(KERN_ERR "%s: Out of memory whilst attempting "
+			       "to kmalloc [%zd] bytes\n", __func__,
+			       filename->encrypted_filename_size);
+			rc = -ENOMEM;
+			goto out;
+		}
+		remaining_bytes = filename->encrypted_filename_size;
+		rc = ecryptfs_write_tag_70_packet(filename->encrypted_filename,
+						  &remaining_bytes,
+						  &packet_size,
+						  mount_crypt_stat,
+						  filename->filename,
+						  filename->filename_size);
+		if (rc) {
+			printk(KERN_ERR "%s: Error attempting to generate "
+			       "tag 70 packet; rc = [%d]\n", __func__,
+			       rc);
+			kfree(filename->encrypted_filename);
+			filename->encrypted_filename = NULL;
+			filename->encrypted_filename_size = 0;
+			goto out;
+		}
+		filename->encrypted_filename_size = packet_size;
+	} else {
+		printk(KERN_ERR "%s: No support for requested filename "
+		       "encryption method in this release\n", __func__);
+		rc = -ENOTSUPP;
 		goto out;
 	}
-	/* TODO: Filename encryption is a scheduled feature for a
-	 * future version of eCryptfs. This function is here only for
-	 * the purpose of providing a framework for other developers
-	 * to easily implement filename encryption. Hint: Replace this
-	 * memcpy() with a call to encrypt and encode the
-	 * filename, the set the length accordingly. */
-	memcpy((void *)(*encoded_name), (void *)name, length);
-	(*encoded_name)[length] = '\0';
-	error = length + 1;
 out:
-	return error;
+	return rc;
 }
 
-/**
- * ecryptfs_decode_filename - converts the cipher text name to plaintext
- * @crypt_stat: The crypt_stat struct associated with the file
- * @name: The filename in cipher text
- * @length: The length of the cipher text name
- * @decrypted_name: The plaintext name
- *
- * Decodes and decrypts the filename.
- *
- * We assume that we have a properly initialized crypto context,
- * pointed to by crypt_stat->tfm.
- *
- * TODO: Implement filename decoding and decryption here, in place of
- * memcpy. We are keeping the framework around for now to (1)
- * facilitate testing of the components needed to implement filename
- * encryption and (2) to provide a code base from which other
- * developers in the community can easily implement this feature.
- *
- * Returns the length of decoded filename; negative if error
- */
-int
-ecryptfs_decode_filename(struct ecryptfs_crypt_stat *crypt_stat,
-			 const char *name, int length, char **decrypted_name)
+static int ecryptfs_copy_filename(char **copied_name, size_t *copied_name_size,
+				  const char *name, size_t name_size)
 {
-	int error = 0;
+	int rc = 0;
 
-	(*decrypted_name) = kmalloc(length + 2, GFP_KERNEL);
-	if (!(*decrypted_name)) {
-		error = -ENOMEM;
+	(*copied_name) = kmalloc((name_size + 2), GFP_KERNEL);
+	if (!(*copied_name)) {
+		rc = -ENOMEM;
 		goto out;
 	}
-	/* TODO: Filename encryption is a scheduled feature for a
-	 * future version of eCryptfs. This function is here only for
-	 * the purpose of providing a framework for other developers
-	 * to easily implement filename encryption. Hint: Replace this
-	 * memcpy() with a call to decode and decrypt the
-	 * filename, the set the length accordingly. */
-	memcpy((void *)(*decrypted_name), (void *)name, length);
-	(*decrypted_name)[length + 1] = '\0';	/* Only for convenience
+	memcpy((void *)(*copied_name), (void *)name, name_size);
+	(*copied_name)[(name_size)] = '\0';	/* Only for convenience
 						 * in printing out the
 						 * string in debug
 						 * messages */
-	error = length;
+	(*copied_name_size) = (name_size + 1);
 out:
-	return error;
+	return rc;
 }
 
 /**
@@ -1740,7 +1752,7 @@ ecryptfs_process_key_cipher(struct crypto_blkcipher **key_tfm,
 	*key_tfm = NULL;
 	if (*key_size > ECRYPTFS_MAX_KEY_BYTES) {
 		rc = -EINVAL;
-		printk(KERN_ERR "Requested key size is [%Zd] bytes; maximum "
+		printk(KERN_ERR "Requested key size is [%zd] bytes; maximum "
 		      "allowable is [%d]\n", *key_size, ECRYPTFS_MAX_KEY_BYTES);
 		goto out;
 	}
@@ -1765,7 +1777,7 @@ ecryptfs_process_key_cipher(struct crypto_blkcipher **key_tfm,
 	get_random_bytes(dummy_key, *key_size);
 	rc = crypto_blkcipher_setkey(*key_tfm, dummy_key, *key_size);
 	if (rc) {
-		printk(KERN_ERR "Error attempting to set key of size [%Zd] for "
+		printk(KERN_ERR "Error attempting to set key of size [%zd] for "
 		       "cipher [%s]; rc = [%d]\n", *key_size, cipher_name, rc);
 		rc = -EINVAL;
 		goto out;
@@ -1910,3 +1922,341 @@ out:
 	mutex_unlock(&key_tfm_list_mutex);
 	return rc;
 }
+
+/* 64 characters forming a 6-bit target field */
+static unsigned char *portable_filename_chars = ("-.0123456789ABCD"
+						 "EFGHIJKLMNOPQRST"
+						 "UVWXYZabcdefghij"
+						 "klmnopqrstuvwxyz");
+
+/* We could either offset on every reverse map or just pad some 0x00's
+ * at the front here */
+static const unsigned char filename_rev_map[] = {
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 7 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 15 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 23 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 31 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 39 */
+	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, /* 47 */
+	0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, /* 55 */
+	0x0A, 0x0B, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 63 */
+	0x00, 0x0C, 0x0D, 0x0E, 0x0F, 0x10, 0x11, 0x12, /* 71 */
+	0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0x1A, /* 79 */
+	0x1B, 0x1C, 0x1D, 0x1E, 0x1F, 0x20, 0x21, 0x22, /* 87 */
+	0x23, 0x24, 0x25, 0x00, 0x00, 0x00, 0x00, 0x00, /* 95 */
+	0x00, 0x26, 0x27, 0x28, 0x29, 0x2A, 0x2B, 0x2C, /* 103 */
+	0x2D, 0x2E, 0x2F, 0x30, 0x31, 0x32, 0x33, 0x34, /* 111 */
+	0x35, 0x36, 0x37, 0x38, 0x39, 0x3A, 0x3B, 0x3C, /* 119 */
+	0x3D, 0x3E, 0x3F
+};
+
+/**
+ * ecryptfs_encode_for_filename
+ * @dst: Destination location for encoded filename
+ * @dst_size: Size of the encoded filename in bytes
+ * @src: Source location for the filename to encode
+ * @src_size: Size of the source in bytes
+ */
+void ecryptfs_encode_for_filename(unsigned char *dst, size_t *dst_size,
+				  unsigned char *src, size_t src_size)
+{
+	size_t num_blocks;
+	size_t block_num = 0;
+	size_t dst_offset = 0;
+	unsigned char last_block[3];
+
+	if (src_size == 0) {
+		(*dst_size) = 0;
+		goto out;
+	}
+	num_blocks = (src_size / 3);
+	if ((src_size % 3) == 0) {
+		memcpy(last_block, (&src[src_size - 3]), 3);
+	} else {
+		num_blocks++;
+		last_block[2] = 0x00;
+		switch (src_size % 3) {
+		case 1:
+			last_block[0] = src[src_size - 1];
+			last_block[1] = 0x00;
+			break;
+		case 2:
+			last_block[0] = src[src_size - 2];
+			last_block[1] = src[src_size - 1];
+		}
+	}
+	(*dst_size) = (num_blocks * 4);
+	if (!dst)
+		goto out;
+	while (block_num < num_blocks) {
+		unsigned char *src_block;
+		unsigned char dst_block[4];
+
+		if (block_num == (num_blocks - 1))
+			src_block = last_block;
+		else
+			src_block = &src[block_num * 3];
+		dst_block[0] = ((src_block[0] >> 2) & 0x3F);
+		dst_block[1] = (((src_block[0] << 4) & 0x30)
+				| ((src_block[1] >> 4) & 0x0F));
+		dst_block[2] = (((src_block[1] << 2) & 0x3C)
+				| ((src_block[2] >> 6) & 0x03));
+		dst_block[3] = (src_block[2] & 0x3F);
+		dst[dst_offset++] = portable_filename_chars[dst_block[0]];
+		dst[dst_offset++] = portable_filename_chars[dst_block[1]];
+		dst[dst_offset++] = portable_filename_chars[dst_block[2]];
+		dst[dst_offset++] = portable_filename_chars[dst_block[3]];
+		block_num++;
+	}
+out:
+	return;
+}
+
+/**
+ * ecryptfs_decode_from_filename
+ * @dst: If NULL, this function only sets @dst_size and returns. If
+ *       non-NULL, this function decodes the encoded octets in @src
+ *       into the memory that @dst points to.
+ * @dst_size: Set to the size of the decoded string.
+ * @src: The encoded set of octets to decode.
+ * @src_size: The size of the encoded set of octets to decode.
+ */
+static void
+ecryptfs_decode_from_filename(unsigned char *dst, size_t *dst_size,
+			      const unsigned char *src, size_t src_size)
+{
+	u8 current_bit_offset = 0;
+	size_t src_byte_offset = 0;
+	size_t dst_byte_offset = 0;
+
+	if (dst == NULL) {
+		/* Not exact; conservatively long. Every block of 4
+		 * encoded characters decodes into a block of 3
+		 * decoded characters. This segment of code provides
+		 * the caller with the maximum amount of allocated
+		 * space that @dst will need to point to in a
+		 * subsequent call. */
+		(*dst_size) = (((src_size + 1) * 3) / 4);
+		goto out;
+	}
+	while (src_byte_offset < src_size) {
+		unsigned char src_byte =
+				filename_rev_map[(int)src[src_byte_offset]];
+
+		switch (current_bit_offset) {
+		case 0:
+			dst[dst_byte_offset] = (src_byte << 2);
+			current_bit_offset = 6;
+			break;
+		case 6:
+			dst[dst_byte_offset++] |= (src_byte >> 4);
+			dst[dst_byte_offset] = ((src_byte & 0xF)
+						 << 4);
+			current_bit_offset = 4;
+			break;
+		case 4:
+			dst[dst_byte_offset++] |= (src_byte >> 2);
+			dst[dst_byte_offset] = (src_byte << 6);
+			current_bit_offset = 2;
+			break;
+		case 2:
+			dst[dst_byte_offset++] |= (src_byte);
+			dst[dst_byte_offset] = 0;
+			current_bit_offset = 0;
+			break;
+		}
+		src_byte_offset++;
+	}
+	(*dst_size) = dst_byte_offset;
+out:
+	return;
+}
+
+/**
+ * ecryptfs_encrypt_and_encode_filename - converts a plaintext file name to cipher text
+ * @crypt_stat: The crypt_stat struct associated with the file anem to encode
+ * @name: The plaintext name
+ * @length: The length of the plaintext
+ * @encoded_name: The encypted name
+ *
+ * Encrypts and encodes a filename into something that constitutes a
+ * valid filename for a filesystem, with printable characters.
+ *
+ * We assume that we have a properly initialized crypto context,
+ * pointed to by crypt_stat->tfm.
+ *
+ * Returns zero on success; non-zero on otherwise
+ */
+int ecryptfs_encrypt_and_encode_filename(
+	char **encoded_name,
+	size_t *encoded_name_size,
+	struct ecryptfs_crypt_stat *crypt_stat,
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+	const char *name, size_t name_size)
+{
+	size_t encoded_name_no_prefix_size;
+	int rc = 0;
+
+	(*encoded_name) = NULL;
+	(*encoded_name_size) = 0;
+	if ((crypt_stat && (crypt_stat->flags & ECRYPTFS_ENCRYPT_FILENAMES))
+	    || (mount_crypt_stat && (mount_crypt_stat->flags
+				     & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES))) {
+		struct ecryptfs_filename *filename;
+
+		filename = kzalloc(sizeof(*filename), GFP_KERNEL);
+		if (!filename) {
+			printk(KERN_ERR "%s: Out of memory whilst attempting "
+			       "to kzalloc [%zd] bytes\n", __func__,
+			       sizeof(*filename));
+			rc = -ENOMEM;
+			goto out;
+		}
+		filename->filename = (char *)name;
+		filename->filename_size = name_size;
+		rc = ecryptfs_encrypt_filename(filename, crypt_stat,
+					       mount_crypt_stat);
+		if (rc) {
+			printk(KERN_ERR "%s: Error attempting to encrypt "
+			       "filename; rc = [%d]\n", __func__, rc);
+			kfree(filename);
+			goto out;
+		}
+		ecryptfs_encode_for_filename(
+			NULL, &encoded_name_no_prefix_size,
+			filename->encrypted_filename,
+			filename->encrypted_filename_size);
+		if ((crypt_stat && (crypt_stat->flags
+				    & ECRYPTFS_ENCFN_USE_MOUNT_FNEK))
+		    || (mount_crypt_stat
+			&& (mount_crypt_stat->flags
+			    & ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK)))
+			(*encoded_name_size) =
+				(ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE
+				 + encoded_name_no_prefix_size);
+		else
+			(*encoded_name_size) =
+				(ECRYPTFS_FEK_ENCRYPTED_FILENAME_PREFIX_SIZE
+				 + encoded_name_no_prefix_size);
+		(*encoded_name) = kmalloc((*encoded_name_size) + 1, GFP_KERNEL);
+		if (!(*encoded_name)) {
+			printk(KERN_ERR "%s: Out of memory whilst attempting "
+			       "to kzalloc [%zd] bytes\n", __func__,
+			       (*encoded_name_size));
+			rc = -ENOMEM;
+			kfree(filename->encrypted_filename);
+			kfree(filename);
+			goto out;
+		}
+		if ((crypt_stat && (crypt_stat->flags
+				    & ECRYPTFS_ENCFN_USE_MOUNT_FNEK))
+		    || (mount_crypt_stat
+			&& (mount_crypt_stat->flags
+			    & ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK))) {
+			memcpy((*encoded_name),
+			       ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX,
+			       ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE);
+			ecryptfs_encode_for_filename(
+			    ((*encoded_name)
+			     + ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE),
+			    &encoded_name_no_prefix_size,
+			    filename->encrypted_filename,
+			    filename->encrypted_filename_size);
+			(*encoded_name_size) =
+				(ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE
+				 + encoded_name_no_prefix_size);
+			(*encoded_name)[(*encoded_name_size)] = '\0';
+			(*encoded_name_size)++;
+		} else {
+			rc = -ENOTSUPP;
+		}
+		if (rc) {
+			printk(KERN_ERR "%s: Error attempting to encode "
+			       "encrypted filename; rc = [%d]\n", __func__,
+			       rc);
+			kfree((*encoded_name));
+			(*encoded_name) = NULL;
+			(*encoded_name_size) = 0;
+		}
+		kfree(filename->encrypted_filename);
+		kfree(filename);
+	} else {
+		rc = ecryptfs_copy_filename(encoded_name,
+					    encoded_name_size,
+					    name, name_size);
+	}
+out:
+	return rc;
+}
+
+/**
+ * ecryptfs_decode_and_decrypt_filename - converts the encoded cipher text name to decoded plaintext
+ * @plaintext_name: The plaintext name
+ * @plaintext_name_size: The plaintext name size
+ * @ecryptfs_dir_dentry: eCryptfs directory dentry
+ * @name: The filename in cipher text
+ * @name_size: The cipher text name size
+ *
+ * Decrypts and decodes the filename.
+ *
+ * Returns zero on error; non-zero otherwise
+ */
+int ecryptfs_decode_and_decrypt_filename(char **plaintext_name,
+					 size_t *plaintext_name_size,
+					 struct dentry *ecryptfs_dir_dentry,
+					 const char *name, size_t name_size)
+{
+	char *decoded_name;
+	size_t decoded_name_size;
+	size_t packet_size;
+	int rc = 0;
+
+	if ((name_size > ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE)
+	    && (strncmp(name, ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX,
+			ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE) == 0)) {
+		struct ecryptfs_mount_crypt_stat *mount_crypt_stat =
+			&ecryptfs_superblock_to_private(
+				ecryptfs_dir_dentry->d_sb)->mount_crypt_stat;
+		const char *orig_name = name;
+		size_t orig_name_size = name_size;
+
+		name += ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE;
+		name_size -= ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE;
+		ecryptfs_decode_from_filename(NULL, &decoded_name_size,
+					      name, name_size);
+		decoded_name = kmalloc(decoded_name_size, GFP_KERNEL);
+		if (!decoded_name) {
+			printk(KERN_ERR "%s: Out of memory whilst attempting "
+			       "to kmalloc [%zd] bytes\n", __func__,
+			       decoded_name_size);
+			rc = -ENOMEM;
+			goto out;
+		}
+		ecryptfs_decode_from_filename(decoded_name, &decoded_name_size,
+					      name, name_size);
+		rc = ecryptfs_parse_tag_70_packet(plaintext_name,
+						  plaintext_name_size,
+						  &packet_size,
+						  mount_crypt_stat,
+						  decoded_name,
+						  decoded_name_size);
+		if (rc) {
+			printk(KERN_INFO "%s: Could not parse tag 70 packet "
+			       "from filename; copying through filename "
+			       "as-is\n", __func__);
+			rc = ecryptfs_copy_filename(plaintext_name,
+						    plaintext_name_size,
+						    orig_name, orig_name_size);
+			goto out_free;
+		}
+	} else {
+		rc = ecryptfs_copy_filename(plaintext_name,
+					    plaintext_name_size,
+					    name, name_size);
+		goto out;
+	}
+out_free:
+	kfree(decoded_name);
+out:
+	return rc;
+}
diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h
index a75026d35d1..c11fc95714a 100644
--- a/fs/ecryptfs/ecryptfs_kernel.h
+++ b/fs/ecryptfs/ecryptfs_kernel.h
@@ -51,12 +51,16 @@
 #define ECRYPTFS_VERSIONING_XATTR                 0x00000010
 #define ECRYPTFS_VERSIONING_MULTKEY               0x00000020
 #define ECRYPTFS_VERSIONING_DEVMISC               0x00000040
+#define ECRYPTFS_VERSIONING_HMAC                  0x00000080
+#define ECRYPTFS_VERSIONING_FILENAME_ENCRYPTION   0x00000100
+#define ECRYPTFS_VERSIONING_GCM                   0x00000200
 #define ECRYPTFS_VERSIONING_MASK (ECRYPTFS_VERSIONING_PASSPHRASE \
 				  | ECRYPTFS_VERSIONING_PLAINTEXT_PASSTHROUGH \
 				  | ECRYPTFS_VERSIONING_PUBKEY \
 				  | ECRYPTFS_VERSIONING_XATTR \
 				  | ECRYPTFS_VERSIONING_MULTKEY \
-				  | ECRYPTFS_VERSIONING_DEVMISC)
+				  | ECRYPTFS_VERSIONING_DEVMISC \
+				  | ECRYPTFS_VERSIONING_FILENAME_ENCRYPTION)
 #define ECRYPTFS_MAX_PASSWORD_LENGTH 64
 #define ECRYPTFS_MAX_PASSPHRASE_BYTES ECRYPTFS_MAX_PASSWORD_LENGTH
 #define ECRYPTFS_SALT_SIZE 8
@@ -199,6 +203,7 @@ ecryptfs_get_key_payload_data(struct key *key)
 #define ECRYPTFS_DEFAULT_CIPHER "aes"
 #define ECRYPTFS_DEFAULT_KEY_BYTES 16
 #define ECRYPTFS_DEFAULT_HASH "md5"
+#define ECRYPTFS_TAG_70_DIGEST ECRYPTFS_DEFAULT_HASH
 #define ECRYPTFS_TAG_1_PACKET_TYPE 0x01
 #define ECRYPTFS_TAG_3_PACKET_TYPE 0x8C
 #define ECRYPTFS_TAG_11_PACKET_TYPE 0xED
@@ -206,30 +211,64 @@ ecryptfs_get_key_payload_data(struct key *key)
 #define ECRYPTFS_TAG_65_PACKET_TYPE 0x41
 #define ECRYPTFS_TAG_66_PACKET_TYPE 0x42
 #define ECRYPTFS_TAG_67_PACKET_TYPE 0x43
+#define ECRYPTFS_TAG_70_PACKET_TYPE 0x46 /* FNEK-encrypted filename
+					  * as dentry name */
+#define ECRYPTFS_TAG_71_PACKET_TYPE 0x47 /* FNEK-encrypted filename in
+					  * metadata */
+#define ECRYPTFS_TAG_72_PACKET_TYPE 0x48 /* FEK-encrypted filename as
+					  * dentry name */
+#define ECRYPTFS_TAG_73_PACKET_TYPE 0x49 /* FEK-encrypted filename as
+					  * metadata */
+/* Constraint: ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES >=
+ * ECRYPTFS_MAX_IV_BYTES */
+#define ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES 16
+#define ECRYPTFS_NON_NULL 0x42 /* A reasonable substitute for NULL */
 #define MD5_DIGEST_SIZE 16
+#define ECRYPTFS_TAG_70_DIGEST_SIZE MD5_DIGEST_SIZE
+#define ECRYPTFS_FEK_ENCRYPTED_FILENAME_PREFIX "ECRYPTFS_FEK_ENCRYPTED."
+#define ECRYPTFS_FEK_ENCRYPTED_FILENAME_PREFIX_SIZE 23
+#define ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX "ECRYPTFS_FNEK_ENCRYPTED."
+#define ECRYPTFS_FNEK_ENCRYPTED_FILENAME_PREFIX_SIZE 24
+#define ECRYPTFS_ENCRYPTED_DENTRY_NAME_LEN (18 + 1 + 4 + 1 + 32)
 
 struct ecryptfs_key_sig {
 	struct list_head crypt_stat_list;
 	char keysig[ECRYPTFS_SIG_SIZE_HEX];
 };
 
+struct ecryptfs_filename {
+	struct list_head crypt_stat_list;
+#define ECRYPTFS_FILENAME_CONTAINS_DECRYPTED 0x00000001
+	u32 flags;
+	u32 seq_no;
+	char *filename;
+	char *encrypted_filename;
+	size_t filename_size;
+	size_t encrypted_filename_size;
+	char fnek_sig[ECRYPTFS_SIG_SIZE_HEX];
+	char dentry_name[ECRYPTFS_ENCRYPTED_DENTRY_NAME_LEN + 1];
+};
+
 /**
  * This is the primary struct associated with each encrypted file.
  *
  * TODO: cache align/pack?
  */
 struct ecryptfs_crypt_stat {
-#define ECRYPTFS_STRUCT_INITIALIZED 0x00000001
-#define ECRYPTFS_POLICY_APPLIED     0x00000002
-#define ECRYPTFS_NEW_FILE           0x00000004
-#define ECRYPTFS_ENCRYPTED          0x00000008
-#define ECRYPTFS_SECURITY_WARNING   0x00000010
-#define ECRYPTFS_ENABLE_HMAC        0x00000020
-#define ECRYPTFS_ENCRYPT_IV_PAGES   0x00000040
-#define ECRYPTFS_KEY_VALID          0x00000080
-#define ECRYPTFS_METADATA_IN_XATTR  0x00000100
-#define ECRYPTFS_VIEW_AS_ENCRYPTED  0x00000200
-#define ECRYPTFS_KEY_SET            0x00000400
+#define ECRYPTFS_STRUCT_INITIALIZED   0x00000001
+#define ECRYPTFS_POLICY_APPLIED       0x00000002
+#define ECRYPTFS_NEW_FILE             0x00000004
+#define ECRYPTFS_ENCRYPTED            0x00000008
+#define ECRYPTFS_SECURITY_WARNING     0x00000010
+#define ECRYPTFS_ENABLE_HMAC          0x00000020
+#define ECRYPTFS_ENCRYPT_IV_PAGES     0x00000040
+#define ECRYPTFS_KEY_VALID            0x00000080
+#define ECRYPTFS_METADATA_IN_XATTR    0x00000100
+#define ECRYPTFS_VIEW_AS_ENCRYPTED    0x00000200
+#define ECRYPTFS_KEY_SET              0x00000400
+#define ECRYPTFS_ENCRYPT_FILENAMES    0x00000800
+#define ECRYPTFS_ENCFN_USE_MOUNT_FNEK 0x00001000
+#define ECRYPTFS_ENCFN_USE_FEK        0x00002000
 	u32 flags;
 	unsigned int file_version;
 	size_t iv_bytes;
@@ -332,13 +371,20 @@ struct ecryptfs_mount_crypt_stat {
 #define ECRYPTFS_XATTR_METADATA_ENABLED        0x00000002
 #define ECRYPTFS_ENCRYPTED_VIEW_ENABLED        0x00000004
 #define ECRYPTFS_MOUNT_CRYPT_STAT_INITIALIZED  0x00000008
+#define ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES      0x00000010
+#define ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK   0x00000020
+#define ECRYPTFS_GLOBAL_ENCFN_USE_FEK          0x00000040
 	u32 flags;
 	struct list_head global_auth_tok_list;
 	struct mutex global_auth_tok_list_mutex;
 	size_t num_global_auth_toks;
 	size_t global_default_cipher_key_size;
+	size_t global_default_fn_cipher_key_bytes;
 	unsigned char global_default_cipher_name[ECRYPTFS_MAX_CIPHER_NAME_SIZE
 						 + 1];
+	unsigned char global_default_fn_cipher_name[
+		ECRYPTFS_MAX_CIPHER_NAME_SIZE + 1];
+	char global_default_fnek_sig[ECRYPTFS_SIG_SIZE_HEX + 1];
 };
 
 /* superblock private data. */
@@ -571,13 +617,22 @@ struct ecryptfs_open_req {
 int ecryptfs_interpose(struct dentry *hidden_dentry,
 		       struct dentry *this_dentry, struct super_block *sb,
 		       u32 flags);
+int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry,
+					struct dentry *lower_dentry,
+					struct ecryptfs_crypt_stat *crypt_stat,
+					struct inode *ecryptfs_dir_inode,
+					struct nameidata *ecryptfs_nd);
+int ecryptfs_decode_and_decrypt_filename(char **decrypted_name,
+					 size_t *decrypted_name_size,
+					 struct dentry *ecryptfs_dentry,
+					 const char *name, size_t name_size);
 int ecryptfs_fill_zeros(struct file *file, loff_t new_length);
-int ecryptfs_decode_filename(struct ecryptfs_crypt_stat *crypt_stat,
-			     const char *name, int length,
-			     char **decrypted_name);
-int ecryptfs_encode_filename(struct ecryptfs_crypt_stat *crypt_stat,
-			     const char *name, int length,
-			     char **encoded_name);
+int ecryptfs_encrypt_and_encode_filename(
+	char **encoded_name,
+	size_t *encoded_name_size,
+	struct ecryptfs_crypt_stat *crypt_stat,
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+	const char *name, size_t name_size);
 struct dentry *ecryptfs_lower_dentry(struct dentry *this_dentry);
 void ecryptfs_dump_hex(char *data, int bytes);
 int virt_to_scatterlist(const void *addr, int size, struct scatterlist *sg,
@@ -599,7 +654,7 @@ int ecryptfs_read_and_validate_header_region(char *data,
 					     struct inode *ecryptfs_inode);
 int ecryptfs_read_and_validate_xattr_region(char *page_virt,
 					    struct dentry *ecryptfs_dentry);
-u8 ecryptfs_code_for_cipher_string(struct ecryptfs_crypt_stat *crypt_stat);
+u8 ecryptfs_code_for_cipher_string(char *cipher_name, size_t key_bytes);
 int ecryptfs_cipher_code_to_string(char *str, u8 cipher_code);
 void ecryptfs_set_default_sizes(struct ecryptfs_crypt_stat *crypt_stat);
 int ecryptfs_generate_key_packet_set(char *dest_base,
@@ -694,5 +749,17 @@ int ecryptfs_privileged_open(struct file **lower_file,
 			     struct vfsmount *lower_mnt,
 			     const struct cred *cred);
 int ecryptfs_init_persistent_file(struct dentry *ecryptfs_dentry);
+int
+ecryptfs_write_tag_70_packet(char *dest, size_t *remaining_bytes,
+			     size_t *packet_size,
+			     struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+			     char *filename, size_t filename_size);
+int
+ecryptfs_parse_tag_70_packet(char **filename, size_t *filename_size,
+			     size_t *packet_size,
+			     struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+			     char *data, size_t max_packet_size);
+int ecryptfs_derive_iv(char *iv, struct ecryptfs_crypt_stat *crypt_stat,
+		       loff_t offset);
 
 #endif /* #ifndef ECRYPTFS_KERNEL_H */
diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c
index eb3dc4c7ac0..9e944057001 100644
--- a/fs/ecryptfs/file.c
+++ b/fs/ecryptfs/file.c
@@ -77,27 +77,27 @@ struct ecryptfs_getdents_callback {
 
 /* Inspired by generic filldir in fs/readdir.c */
 static int
-ecryptfs_filldir(void *dirent, const char *name, int namelen, loff_t offset,
-		 u64 ino, unsigned int d_type)
+ecryptfs_filldir(void *dirent, const char *lower_name, int lower_namelen,
+		 loff_t offset, u64 ino, unsigned int d_type)
 {
-	struct ecryptfs_crypt_stat *crypt_stat;
 	struct ecryptfs_getdents_callback *buf =
 	    (struct ecryptfs_getdents_callback *)dirent;
+	size_t name_size;
+	char *name;
 	int rc;
-	int decoded_length;
-	char *decoded_name;
 
-	crypt_stat = ecryptfs_dentry_to_private(buf->dentry)->crypt_stat;
 	buf->filldir_called++;
-	decoded_length = ecryptfs_decode_filename(crypt_stat, name, namelen,
-						  &decoded_name);
-	if (decoded_length < 0) {
-		rc = decoded_length;
+	rc = ecryptfs_decode_and_decrypt_filename(&name, &name_size,
+						  buf->dentry, lower_name,
+						  lower_namelen);
+	if (rc) {
+		printk(KERN_ERR "%s: Error attempting to decode and decrypt "
+		       "filename [%s]; rc = [%d]\n", __func__, lower_name,
+		       rc);
 		goto out;
 	}
-	rc = buf->filldir(buf->dirent, decoded_name, decoded_length, offset,
-			  ino, d_type);
-	kfree(decoded_name);
+	rc = buf->filldir(buf->dirent, name, name_size, offset, ino, d_type);
+	kfree(name);
 	if (rc >= 0)
 		buf->entries_written++;
 out:
@@ -106,8 +106,8 @@ out:
 
 /**
  * ecryptfs_readdir
- * @file: The ecryptfs file struct
- * @dirent: Directory entry
+ * @file: The eCryptfs directory file
+ * @dirent: Directory entry handle
  * @filldir: The filldir callback function
  */
 static int ecryptfs_readdir(struct file *file, void *dirent, filldir_t filldir)
@@ -275,18 +275,9 @@ static int ecryptfs_release(struct inode *inode, struct file *file)
 static int
 ecryptfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 {
-	struct file *lower_file = ecryptfs_file_to_lower(file);
-	struct dentry *lower_dentry = ecryptfs_dentry_to_lower(dentry);
-	struct inode *lower_inode = lower_dentry->d_inode;
-	int rc = -EINVAL;
-
-	if (lower_inode->i_fop->fsync) {
-		mutex_lock(&lower_inode->i_mutex);
-		rc = lower_inode->i_fop->fsync(lower_file, lower_dentry,
-					       datasync);
-		mutex_unlock(&lower_inode->i_mutex);
-	}
-	return rc;
+	return vfs_fsync(ecryptfs_file_to_lower(file),
+			 ecryptfs_dentry_to_lower(dentry),
+			 datasync);
 }
 
 static int ecryptfs_fasync(int fd, struct file *file, int flag)
diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c
index 89209f00f9c..5697899a168 100644
--- a/fs/ecryptfs/inode.c
+++ b/fs/ecryptfs/inode.c
@@ -52,8 +52,7 @@ static void unlock_dir(struct dentry *dir)
 /**
  * ecryptfs_create_underlying_file
  * @lower_dir_inode: inode of the parent in the lower fs of the new file
- * @lower_dentry: New file's dentry in the lower fs
- * @ecryptfs_dentry: New file's dentry in ecryptfs
+ * @dentry: New file's dentry
  * @mode: The mode of the new file
  * @nd: nameidata of ecryptfs' parent's dentry & vfsmount
  *
@@ -228,8 +227,7 @@ ecryptfs_create(struct inode *directory_inode, struct dentry *ecryptfs_dentry,
 {
 	int rc;
 
-	/* ecryptfs_do_create() calls ecryptfs_interpose(), which opens
-	 * the crypt_stat->lower_file (persistent file) */
+	/* ecryptfs_do_create() calls ecryptfs_interpose() */
 	rc = ecryptfs_do_create(directory_inode, ecryptfs_dentry, mode, nd);
 	if (unlikely(rc)) {
 		ecryptfs_printk(KERN_WARNING, "Failed to create file in"
@@ -244,141 +242,91 @@ out:
 }
 
 /**
- * ecryptfs_lookup
- * @dir: inode
- * @dentry: The dentry
- * @nd: nameidata, may be NULL
- *
- * Find a file on disk. If the file does not exist, then we'll add it to the
- * dentry cache and continue on to read it from the disk.
+ * ecryptfs_lookup_and_interpose_lower - Perform a lookup
  */
-static struct dentry *ecryptfs_lookup(struct inode *dir, struct dentry *dentry,
-				      struct nameidata *nd)
+int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry,
+					struct dentry *lower_dentry,
+					struct ecryptfs_crypt_stat *crypt_stat,
+					struct inode *ecryptfs_dir_inode,
+					struct nameidata *ecryptfs_nd)
 {
-	int rc = 0;
 	struct dentry *lower_dir_dentry;
-	struct dentry *lower_dentry;
 	struct vfsmount *lower_mnt;
-	char *encoded_name;
-	int encoded_namelen;
-	struct ecryptfs_crypt_stat *crypt_stat = NULL;
+	struct inode *lower_inode;
 	struct ecryptfs_mount_crypt_stat *mount_crypt_stat;
 	char *page_virt = NULL;
-	struct inode *lower_inode;
 	u64 file_size;
+	int rc = 0;
 
-	lower_dir_dentry = ecryptfs_dentry_to_lower(dentry->d_parent);
-	dentry->d_op = &ecryptfs_dops;
-	if ((dentry->d_name.len == 1 && !strcmp(dentry->d_name.name, "."))
-	    || (dentry->d_name.len == 2
-		&& !strcmp(dentry->d_name.name, ".."))) {
-		d_drop(dentry);
-		goto out;
-	}
-	encoded_namelen = ecryptfs_encode_filename(crypt_stat,
-						   dentry->d_name.name,
-						   dentry->d_name.len,
-						   &encoded_name);
-	if (encoded_namelen < 0) {
-		rc = encoded_namelen;
-		d_drop(dentry);
-		goto out;
-	}
-	ecryptfs_printk(KERN_DEBUG, "encoded_name = [%s]; encoded_namelen "
-			"= [%d]\n", encoded_name, encoded_namelen);
-	lower_dentry = lookup_one_len(encoded_name, lower_dir_dentry,
-				      encoded_namelen - 1);
-	kfree(encoded_name);
-	if (IS_ERR(lower_dentry)) {
-		ecryptfs_printk(KERN_ERR, "ERR from lower_dentry\n");
-		rc = PTR_ERR(lower_dentry);
-		d_drop(dentry);
-		goto out;
-	}
-	lower_mnt = mntget(ecryptfs_dentry_to_lower_mnt(dentry->d_parent));
-	ecryptfs_printk(KERN_DEBUG, "lower_dentry = [%p]; lower_dentry->"
-       		"d_name.name = [%s]\n", lower_dentry,
-		lower_dentry->d_name.name);
+	lower_dir_dentry = lower_dentry->d_parent;
+	lower_mnt = mntget(ecryptfs_dentry_to_lower_mnt(
+				   ecryptfs_dentry->d_parent));
 	lower_inode = lower_dentry->d_inode;
-	fsstack_copy_attr_atime(dir, lower_dir_dentry->d_inode);
+	fsstack_copy_attr_atime(ecryptfs_dir_inode, lower_dir_dentry->d_inode);
 	BUG_ON(!atomic_read(&lower_dentry->d_count));
-	ecryptfs_set_dentry_private(dentry,
+	ecryptfs_set_dentry_private(ecryptfs_dentry,
 				    kmem_cache_alloc(ecryptfs_dentry_info_cache,
 						     GFP_KERNEL));
-	if (!ecryptfs_dentry_to_private(dentry)) {
+	if (!ecryptfs_dentry_to_private(ecryptfs_dentry)) {
 		rc = -ENOMEM;
-		ecryptfs_printk(KERN_ERR, "Out of memory whilst attempting "
-				"to allocate ecryptfs_dentry_info struct\n");
+		printk(KERN_ERR "%s: Out of memory whilst attempting "
+		       "to allocate ecryptfs_dentry_info struct\n",
+			__func__);
 		goto out_dput;
 	}
-	ecryptfs_set_dentry_lower(dentry, lower_dentry);
-	ecryptfs_set_dentry_lower_mnt(dentry, lower_mnt);
+	ecryptfs_set_dentry_lower(ecryptfs_dentry, lower_dentry);
+	ecryptfs_set_dentry_lower_mnt(ecryptfs_dentry, lower_mnt);
 	if (!lower_dentry->d_inode) {
 		/* We want to add because we couldn't find in lower */
-		d_add(dentry, NULL);
+		d_add(ecryptfs_dentry, NULL);
 		goto out;
 	}
-	rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb,
-				ECRYPTFS_INTERPOSE_FLAG_D_ADD);
+	rc = ecryptfs_interpose(lower_dentry, ecryptfs_dentry,
+				ecryptfs_dir_inode->i_sb, 1);
 	if (rc) {
-		ecryptfs_printk(KERN_ERR, "Error interposing\n");
+		printk(KERN_ERR "%s: Error interposing; rc = [%d]\n",
+		       __func__, rc);
 		goto out;
 	}
-	if (S_ISDIR(lower_inode->i_mode)) {
-		ecryptfs_printk(KERN_DEBUG, "Is a directory; returning\n");
+	if (S_ISDIR(lower_inode->i_mode))
 		goto out;
-	}
-	if (S_ISLNK(lower_inode->i_mode)) {
-		ecryptfs_printk(KERN_DEBUG, "Is a symlink; returning\n");
+	if (S_ISLNK(lower_inode->i_mode))
 		goto out;
-	}
-	if (special_file(lower_inode->i_mode)) {
-		ecryptfs_printk(KERN_DEBUG, "Is a special file; returning\n");
+	if (special_file(lower_inode->i_mode))
 		goto out;
-	}
-	if (!nd) {
-		ecryptfs_printk(KERN_DEBUG, "We have a NULL nd, just leave"
-				"as we *think* we are about to unlink\n");
+	if (!ecryptfs_nd)
 		goto out;
-	}
 	/* Released in this function */
-	page_virt = kmem_cache_zalloc(ecryptfs_header_cache_2,
-				      GFP_USER);
+	page_virt = kmem_cache_zalloc(ecryptfs_header_cache_2, GFP_USER);
 	if (!page_virt) {
+		printk(KERN_ERR "%s: Cannot kmem_cache_zalloc() a page\n",
+		       __func__);
 		rc = -ENOMEM;
-		ecryptfs_printk(KERN_ERR,
-				"Cannot ecryptfs_kmalloc a page\n");
 		goto out;
 	}
-	crypt_stat = &ecryptfs_inode_to_private(dentry->d_inode)->crypt_stat;
-	if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED))
-		ecryptfs_set_default_sizes(crypt_stat);
-	if (!ecryptfs_inode_to_private(dentry->d_inode)->lower_file) {
-		rc = ecryptfs_init_persistent_file(dentry);
+	if (!ecryptfs_inode_to_private(ecryptfs_dentry->d_inode)->lower_file) {
+		rc = ecryptfs_init_persistent_file(ecryptfs_dentry);
 		if (rc) {
 			printk(KERN_ERR "%s: Error attempting to initialize "
 			       "the persistent file for the dentry with name "
 			       "[%s]; rc = [%d]\n", __func__,
-			       dentry->d_name.name, rc);
-			goto out;
+			       ecryptfs_dentry->d_name.name, rc);
+			goto out_free_kmem;
 		}
 	}
 	rc = ecryptfs_read_and_validate_header_region(page_virt,
-						      dentry->d_inode);
+						      ecryptfs_dentry->d_inode);
 	if (rc) {
-		rc = ecryptfs_read_and_validate_xattr_region(page_virt, dentry);
+		rc = ecryptfs_read_and_validate_xattr_region(page_virt,
+							     ecryptfs_dentry);
 		if (rc) {
-			printk(KERN_DEBUG "Valid metadata not found in header "
-			       "region or xattr region; treating file as "
-			       "unencrypted\n");
 			rc = 0;
-			kmem_cache_free(ecryptfs_header_cache_2, page_virt);
-			goto out;
+			goto out_free_kmem;
 		}
 		crypt_stat->flags |= ECRYPTFS_METADATA_IN_XATTR;
 	}
 	mount_crypt_stat = &ecryptfs_superblock_to_private(
-		dentry->d_sb)->mount_crypt_stat;
+		ecryptfs_dentry->d_sb)->mount_crypt_stat;
 	if (mount_crypt_stat->flags & ECRYPTFS_ENCRYPTED_VIEW_ENABLED) {
 		if (crypt_stat->flags & ECRYPTFS_METADATA_IN_XATTR)
 			file_size = (crypt_stat->num_header_bytes_at_front
@@ -388,14 +336,103 @@ static struct dentry *ecryptfs_lookup(struct inode *dir, struct dentry *dentry,
 	} else {
 		file_size = get_unaligned_be64(page_virt);
 	}
-	i_size_write(dentry->d_inode, (loff_t)file_size);
+	i_size_write(ecryptfs_dentry->d_inode, (loff_t)file_size);
+out_free_kmem:
 	kmem_cache_free(ecryptfs_header_cache_2, page_virt);
 	goto out;
-
 out_dput:
 	dput(lower_dentry);
-	d_drop(dentry);
+	d_drop(ecryptfs_dentry);
 out:
+	return rc;
+}
+
+/**
+ * ecryptfs_lookup
+ * @ecryptfs_dir_inode: The eCryptfs directory inode
+ * @ecryptfs_dentry: The eCryptfs dentry that we are looking up
+ * @ecryptfs_nd: nameidata; may be NULL
+ *
+ * Find a file on disk. If the file does not exist, then we'll add it to the
+ * dentry cache and continue on to read it from the disk.
+ */
+static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode,
+				      struct dentry *ecryptfs_dentry,
+				      struct nameidata *ecryptfs_nd)
+{
+	char *encrypted_and_encoded_name = NULL;
+	size_t encrypted_and_encoded_name_size;
+	struct ecryptfs_crypt_stat *crypt_stat = NULL;
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat = NULL;
+	struct ecryptfs_inode_info *inode_info;
+	struct dentry *lower_dir_dentry, *lower_dentry;
+	int rc = 0;
+
+	ecryptfs_dentry->d_op = &ecryptfs_dops;
+	if ((ecryptfs_dentry->d_name.len == 1
+	     && !strcmp(ecryptfs_dentry->d_name.name, "."))
+	    || (ecryptfs_dentry->d_name.len == 2
+		&& !strcmp(ecryptfs_dentry->d_name.name, ".."))) {
+		goto out_d_drop;
+	}
+	lower_dir_dentry = ecryptfs_dentry_to_lower(ecryptfs_dentry->d_parent);
+	lower_dentry = lookup_one_len(ecryptfs_dentry->d_name.name,
+				      lower_dir_dentry,
+				      ecryptfs_dentry->d_name.len);
+	if (IS_ERR(lower_dentry)) {
+		rc = PTR_ERR(lower_dentry);
+		printk(KERN_ERR "%s: lookup_one_len() returned [%d] on "
+		       "lower_dentry = [%s]\n", __func__, rc,
+		       ecryptfs_dentry->d_name.name);
+		goto out_d_drop;
+	}
+	if (lower_dentry->d_inode)
+		goto lookup_and_interpose;
+	inode_info =  ecryptfs_inode_to_private(ecryptfs_dentry->d_inode);
+	if (inode_info) {
+		crypt_stat = &inode_info->crypt_stat;
+		/* TODO: lock for crypt_stat comparison */
+		if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED))
+			ecryptfs_set_default_sizes(crypt_stat);
+	}
+	if (crypt_stat)
+		mount_crypt_stat = crypt_stat->mount_crypt_stat;
+	else
+		mount_crypt_stat = &ecryptfs_superblock_to_private(
+			ecryptfs_dentry->d_sb)->mount_crypt_stat;
+	if (!(crypt_stat && (crypt_stat->flags & ECRYPTFS_ENCRYPT_FILENAMES))
+	    && !(mount_crypt_stat && (mount_crypt_stat->flags
+				     & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)))
+		goto lookup_and_interpose;
+	dput(lower_dentry);
+	rc = ecryptfs_encrypt_and_encode_filename(
+		&encrypted_and_encoded_name, &encrypted_and_encoded_name_size,
+		crypt_stat, mount_crypt_stat, ecryptfs_dentry->d_name.name,
+		ecryptfs_dentry->d_name.len);
+	if (rc) {
+		printk(KERN_ERR "%s: Error attempting to encrypt and encode "
+		       "filename; rc = [%d]\n", __func__, rc);
+		goto out_d_drop;
+	}
+	lower_dentry = lookup_one_len(encrypted_and_encoded_name,
+				      lower_dir_dentry,
+				      encrypted_and_encoded_name_size - 1);
+	if (IS_ERR(lower_dentry)) {
+		rc = PTR_ERR(lower_dentry);
+		printk(KERN_ERR "%s: lookup_one_len() returned [%d] on "
+		       "lower_dentry = [%s]\n", __func__, rc,
+		       encrypted_and_encoded_name);
+		goto out_d_drop;
+	}
+lookup_and_interpose:
+	rc = ecryptfs_lookup_and_interpose_lower(ecryptfs_dentry, lower_dentry,
+						 crypt_stat, ecryptfs_dir_inode,
+						 ecryptfs_nd);
+	goto out;
+out_d_drop:
+	d_drop(ecryptfs_dentry);
+out:
+	kfree(encrypted_and_encoded_name);
 	return ERR_PTR(rc);
 }
 
@@ -466,19 +503,21 @@ static int ecryptfs_symlink(struct inode *dir, struct dentry *dentry,
 	struct dentry *lower_dentry;
 	struct dentry *lower_dir_dentry;
 	char *encoded_symname;
-	int encoded_symlen;
-	struct ecryptfs_crypt_stat *crypt_stat = NULL;
+	size_t encoded_symlen;
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat = NULL;
 
 	lower_dentry = ecryptfs_dentry_to_lower(dentry);
 	dget(lower_dentry);
 	lower_dir_dentry = lock_parent(lower_dentry);
-	encoded_symlen = ecryptfs_encode_filename(crypt_stat, symname,
-						  strlen(symname),
-						  &encoded_symname);
-	if (encoded_symlen < 0) {
-		rc = encoded_symlen;
+	mount_crypt_stat = &ecryptfs_superblock_to_private(
+		dir->i_sb)->mount_crypt_stat;
+	rc = ecryptfs_encrypt_and_encode_filename(&encoded_symname,
+						  &encoded_symlen,
+						  NULL,
+						  mount_crypt_stat, symname,
+						  strlen(symname));
+	if (rc)
 		goto out_lock;
-	}
 	rc = vfs_symlink(lower_dir_dentry->d_inode, lower_dentry,
 			 encoded_symname);
 	kfree(encoded_symname);
@@ -602,53 +641,54 @@ out_lock:
 }
 
 static int
-ecryptfs_readlink(struct dentry *dentry, char __user * buf, int bufsiz)
+ecryptfs_readlink(struct dentry *dentry, char __user *buf, int bufsiz)
 {
-	int rc;
-	struct dentry *lower_dentry;
-	char *decoded_name;
 	char *lower_buf;
-	mm_segment_t old_fs;
+	struct dentry *lower_dentry;
 	struct ecryptfs_crypt_stat *crypt_stat;
+	char *plaintext_name;
+	size_t plaintext_name_size;
+	mm_segment_t old_fs;
+	int rc;
 
 	lower_dentry = ecryptfs_dentry_to_lower(dentry);
-	if (!lower_dentry->d_inode->i_op ||
-	    !lower_dentry->d_inode->i_op->readlink) {
+	if (!lower_dentry->d_inode->i_op->readlink) {
 		rc = -EINVAL;
 		goto out;
 	}
+	crypt_stat = &ecryptfs_inode_to_private(dentry->d_inode)->crypt_stat;
 	/* Released in this function */
 	lower_buf = kmalloc(bufsiz, GFP_KERNEL);
 	if (lower_buf == NULL) {
-		ecryptfs_printk(KERN_ERR, "Out of memory\n");
+		printk(KERN_ERR "%s: Out of memory whilst attempting to "
+		       "kmalloc [%d] bytes\n", __func__, bufsiz);
 		rc = -ENOMEM;
 		goto out;
 	}
 	old_fs = get_fs();
 	set_fs(get_ds());
-	ecryptfs_printk(KERN_DEBUG, "Calling readlink w/ "
-			"lower_dentry->d_name.name = [%s]\n",
-			lower_dentry->d_name.name);
 	rc = lower_dentry->d_inode->i_op->readlink(lower_dentry,
 						   (char __user *)lower_buf,
 						   bufsiz);
 	set_fs(old_fs);
 	if (rc >= 0) {
-		crypt_stat = NULL;
-		rc = ecryptfs_decode_filename(crypt_stat, lower_buf, rc,
-					      &decoded_name);
-		if (rc == -ENOMEM)
+		rc = ecryptfs_decode_and_decrypt_filename(&plaintext_name,
+							  &plaintext_name_size,
+							  dentry, lower_buf,
+							  rc);
+		if (rc) {
+			printk(KERN_ERR "%s: Error attempting to decode and "
+			       "decrypt filename; rc = [%d]\n", __func__,
+				rc);
 			goto out_free_lower_buf;
-		if (rc > 0) {
-			ecryptfs_printk(KERN_DEBUG, "Copying [%d] bytes "
-					"to userspace: [%*s]\n", rc,
-					decoded_name);
-			if (copy_to_user(buf, decoded_name, rc))
-				rc = -EFAULT;
 		}
-		kfree(decoded_name);
-		fsstack_copy_attr_atime(dentry->d_inode,
-					lower_dentry->d_inode);
+		rc = copy_to_user(buf, plaintext_name, plaintext_name_size);
+		if (rc)
+			rc = -EFAULT;
+		else
+			rc = plaintext_name_size;
+		kfree(plaintext_name);
+		fsstack_copy_attr_atime(dentry->d_inode, lower_dentry->d_inode);
 	}
 out_free_lower_buf:
 	kfree(lower_buf);
@@ -670,13 +710,12 @@ static void *ecryptfs_follow_link(struct dentry *dentry, struct nameidata *nd)
 	}
 	old_fs = get_fs();
 	set_fs(get_ds());
-	ecryptfs_printk(KERN_DEBUG, "Calling readlink w/ "
-			"dentry->d_name.name = [%s]\n", dentry->d_name.name);
 	rc = dentry->d_inode->i_op->readlink(dentry, (char __user *)buf, len);
-	buf[rc] = '\0';
 	set_fs(old_fs);
 	if (rc < 0)
 		goto out_free;
+	else
+		buf[rc] = '\0';
 	rc = 0;
 	nd_set_link(nd, buf);
 	goto out;
diff --git a/fs/ecryptfs/keystore.c b/fs/ecryptfs/keystore.c
index 0d713b69194..ff539420cc6 100644
--- a/fs/ecryptfs/keystore.c
+++ b/fs/ecryptfs/keystore.c
@@ -358,7 +358,7 @@ parse_tag_67_packet(struct ecryptfs_key_record *key_rec,
 	/* verify that everything through the encrypted FEK size is present */
 	if (message_len < 4) {
 		rc = -EIO;
-		printk(KERN_ERR "%s: message_len is [%Zd]; minimum acceptable "
+		printk(KERN_ERR "%s: message_len is [%zd]; minimum acceptable "
 		       "message length is [%d]\n", __func__, message_len, 4);
 		goto out;
 	}
@@ -385,13 +385,13 @@ parse_tag_67_packet(struct ecryptfs_key_record *key_rec,
 	i += data_len;
 	if (message_len < (i + key_rec->enc_key_size)) {
 		rc = -EIO;
-		printk(KERN_ERR "%s: message_len [%Zd]; max len is [%Zd]\n",
+		printk(KERN_ERR "%s: message_len [%zd]; max len is [%zd]\n",
 		       __func__, message_len, (i + key_rec->enc_key_size));
 		goto out;
 	}
 	if (key_rec->enc_key_size > ECRYPTFS_MAX_ENCRYPTED_KEY_BYTES) {
 		rc = -EIO;
-		printk(KERN_ERR "%s: Encrypted key_size [%Zd] larger than "
+		printk(KERN_ERR "%s: Encrypted key_size [%zd] larger than "
 		       "the maximum key size [%d]\n", __func__,
 		       key_rec->enc_key_size,
 		       ECRYPTFS_MAX_ENCRYPTED_KEY_BYTES);
@@ -403,6 +403,580 @@ out:
 }
 
 static int
+ecryptfs_find_global_auth_tok_for_sig(
+	struct ecryptfs_global_auth_tok **global_auth_tok,
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat, char *sig)
+{
+	struct ecryptfs_global_auth_tok *walker;
+	int rc = 0;
+
+	(*global_auth_tok) = NULL;
+	mutex_lock(&mount_crypt_stat->global_auth_tok_list_mutex);
+	list_for_each_entry(walker,
+			    &mount_crypt_stat->global_auth_tok_list,
+			    mount_crypt_stat_list) {
+		if (memcmp(walker->sig, sig, ECRYPTFS_SIG_SIZE_HEX) == 0) {
+			(*global_auth_tok) = walker;
+			goto out;
+		}
+	}
+	rc = -EINVAL;
+out:
+	mutex_unlock(&mount_crypt_stat->global_auth_tok_list_mutex);
+	return rc;
+}
+
+/**
+ * ecryptfs_find_auth_tok_for_sig
+ * @auth_tok: Set to the matching auth_tok; NULL if not found
+ * @crypt_stat: inode crypt_stat crypto context
+ * @sig: Sig of auth_tok to find
+ *
+ * For now, this function simply looks at the registered auth_tok's
+ * linked off the mount_crypt_stat, so all the auth_toks that can be
+ * used must be registered at mount time. This function could
+ * potentially try a lot harder to find auth_tok's (e.g., by calling
+ * out to ecryptfsd to dynamically retrieve an auth_tok object) so
+ * that static registration of auth_tok's will no longer be necessary.
+ *
+ * Returns zero on no error; non-zero on error
+ */
+static int
+ecryptfs_find_auth_tok_for_sig(
+	struct ecryptfs_auth_tok **auth_tok,
+	struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+	char *sig)
+{
+	struct ecryptfs_global_auth_tok *global_auth_tok;
+	int rc = 0;
+
+	(*auth_tok) = NULL;
+	if (ecryptfs_find_global_auth_tok_for_sig(&global_auth_tok,
+						  mount_crypt_stat, sig)) {
+		struct key *auth_tok_key;
+
+		rc = ecryptfs_keyring_auth_tok_for_sig(&auth_tok_key, auth_tok,
+						       sig);
+	} else
+		(*auth_tok) = global_auth_tok->global_auth_tok;
+	return rc;
+}
+
+/**
+ * write_tag_70_packet can gobble a lot of stack space. We stuff most
+ * of the function's parameters in a kmalloc'd struct to help reduce
+ * eCryptfs' overall stack usage.
+ */
+struct ecryptfs_write_tag_70_packet_silly_stack {
+	u8 cipher_code;
+	size_t max_packet_size;
+	size_t packet_size_len;
+	size_t block_aligned_filename_size;
+	size_t block_size;
+	size_t i;
+	size_t j;
+	size_t num_rand_bytes;
+	struct mutex *tfm_mutex;
+	char *block_aligned_filename;
+	struct ecryptfs_auth_tok *auth_tok;
+	struct scatterlist src_sg;
+	struct scatterlist dst_sg;
+	struct blkcipher_desc desc;
+	char iv[ECRYPTFS_MAX_IV_BYTES];
+	char hash[ECRYPTFS_TAG_70_DIGEST_SIZE];
+	char tmp_hash[ECRYPTFS_TAG_70_DIGEST_SIZE];
+	struct hash_desc hash_desc;
+	struct scatterlist hash_sg;
+};
+
+/**
+ * write_tag_70_packet - Write encrypted filename (EFN) packet against FNEK
+ * @filename: NULL-terminated filename string
+ *
+ * This is the simplest mechanism for achieving filename encryption in
+ * eCryptfs. It encrypts the given filename with the mount-wide
+ * filename encryption key (FNEK) and stores it in a packet to @dest,
+ * which the callee will encode and write directly into the dentry
+ * name.
+ */
+int
+ecryptfs_write_tag_70_packet(char *dest, size_t *remaining_bytes,
+			     size_t *packet_size,
+			     struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+			     char *filename, size_t filename_size)
+{
+	struct ecryptfs_write_tag_70_packet_silly_stack *s;
+	int rc = 0;
+
+	s = kmalloc(sizeof(*s), GFP_KERNEL);
+	if (!s) {
+		printk(KERN_ERR "%s: Out of memory whilst trying to kmalloc "
+		       "[%zd] bytes of kernel memory\n", __func__, sizeof(*s));
+		goto out;
+	}
+	s->desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+	(*packet_size) = 0;
+	rc = ecryptfs_get_tfm_and_mutex_for_cipher_name(
+		&s->desc.tfm,
+		&s->tfm_mutex, mount_crypt_stat->global_default_fn_cipher_name);
+	if (unlikely(rc)) {
+		printk(KERN_ERR "Internal error whilst attempting to get "
+		       "tfm and mutex for cipher name [%s]; rc = [%d]\n",
+		       mount_crypt_stat->global_default_fn_cipher_name, rc);
+		goto out;
+	}
+	mutex_lock(s->tfm_mutex);
+	s->block_size = crypto_blkcipher_blocksize(s->desc.tfm);
+	/* Plus one for the \0 separator between the random prefix
+	 * and the plaintext filename */
+	s->num_rand_bytes = (ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES + 1);
+	s->block_aligned_filename_size = (s->num_rand_bytes + filename_size);
+	if ((s->block_aligned_filename_size % s->block_size) != 0) {
+		s->num_rand_bytes += (s->block_size
+				      - (s->block_aligned_filename_size
+					 % s->block_size));
+		s->block_aligned_filename_size = (s->num_rand_bytes
+						  + filename_size);
+	}
+	/* Octet 0: Tag 70 identifier
+	 * Octets 1-N1: Tag 70 packet size (includes cipher identifier
+	 *              and block-aligned encrypted filename size)
+	 * Octets N1-N2: FNEK sig (ECRYPTFS_SIG_SIZE)
+	 * Octet N2-N3: Cipher identifier (1 octet)
+	 * Octets N3-N4: Block-aligned encrypted filename
+	 *  - Consists of a minimum number of random characters, a \0
+	 *    separator, and then the filename */
+	s->max_packet_size = (1                   /* Tag 70 identifier */
+			      + 3                 /* Max Tag 70 packet size */
+			      + ECRYPTFS_SIG_SIZE /* FNEK sig */
+			      + 1                 /* Cipher identifier */
+			      + s->block_aligned_filename_size);
+	if (dest == NULL) {
+		(*packet_size) = s->max_packet_size;
+		goto out_unlock;
+	}
+	if (s->max_packet_size > (*remaining_bytes)) {
+		printk(KERN_WARNING "%s: Require [%zd] bytes to write; only "
+		       "[%zd] available\n", __func__, s->max_packet_size,
+		       (*remaining_bytes));
+		rc = -EINVAL;
+		goto out_unlock;
+	}
+	s->block_aligned_filename = kzalloc(s->block_aligned_filename_size,
+					    GFP_KERNEL);
+	if (!s->block_aligned_filename) {
+		printk(KERN_ERR "%s: Out of kernel memory whilst attempting to "
+		       "kzalloc [%zd] bytes\n", __func__,
+		       s->block_aligned_filename_size);
+		rc = -ENOMEM;
+		goto out_unlock;
+	}
+	s->i = 0;
+	dest[s->i++] = ECRYPTFS_TAG_70_PACKET_TYPE;
+	rc = ecryptfs_write_packet_length(&dest[s->i],
+					  (ECRYPTFS_SIG_SIZE
+					   + 1 /* Cipher code */
+					   + s->block_aligned_filename_size),
+					  &s->packet_size_len);
+	if (rc) {
+		printk(KERN_ERR "%s: Error generating tag 70 packet "
+		       "header; cannot generate packet length; rc = [%d]\n",
+		       __func__, rc);
+		goto out_free_unlock;
+	}
+	s->i += s->packet_size_len;
+	ecryptfs_from_hex(&dest[s->i],
+			  mount_crypt_stat->global_default_fnek_sig,
+			  ECRYPTFS_SIG_SIZE);
+	s->i += ECRYPTFS_SIG_SIZE;
+	s->cipher_code = ecryptfs_code_for_cipher_string(
+		mount_crypt_stat->global_default_fn_cipher_name,
+		mount_crypt_stat->global_default_fn_cipher_key_bytes);
+	if (s->cipher_code == 0) {
+		printk(KERN_WARNING "%s: Unable to generate code for "
+		       "cipher [%s] with key bytes [%zd]\n", __func__,
+		       mount_crypt_stat->global_default_fn_cipher_name,
+		       mount_crypt_stat->global_default_fn_cipher_key_bytes);
+		rc = -EINVAL;
+		goto out_free_unlock;
+	}
+	dest[s->i++] = s->cipher_code;
+	rc = ecryptfs_find_auth_tok_for_sig(
+		&s->auth_tok, mount_crypt_stat,
+		mount_crypt_stat->global_default_fnek_sig);
+	if (rc) {
+		printk(KERN_ERR "%s: Error attempting to find auth tok for "
+		       "fnek sig [%s]; rc = [%d]\n", __func__,
+		       mount_crypt_stat->global_default_fnek_sig, rc);
+		goto out_free_unlock;
+	}
+	/* TODO: Support other key modules than passphrase for
+	 * filename encryption */
+	BUG_ON(s->auth_tok->token_type != ECRYPTFS_PASSWORD);
+	sg_init_one(
+		&s->hash_sg,
+		(u8 *)s->auth_tok->token.password.session_key_encryption_key,
+		s->auth_tok->token.password.session_key_encryption_key_bytes);
+	s->hash_desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+	s->hash_desc.tfm = crypto_alloc_hash(ECRYPTFS_TAG_70_DIGEST, 0,
+					     CRYPTO_ALG_ASYNC);
+	if (IS_ERR(s->hash_desc.tfm)) {
+			rc = PTR_ERR(s->hash_desc.tfm);
+			printk(KERN_ERR "%s: Error attempting to "
+			       "allocate hash crypto context; rc = [%d]\n",
+			       __func__, rc);
+			goto out_free_unlock;
+	}
+	rc = crypto_hash_init(&s->hash_desc);
+	if (rc) {
+		printk(KERN_ERR
+		       "%s: Error initializing crypto hash; rc = [%d]\n",
+		       __func__, rc);
+		goto out_release_free_unlock;
+	}
+	rc = crypto_hash_update(
+		&s->hash_desc, &s->hash_sg,
+		s->auth_tok->token.password.session_key_encryption_key_bytes);
+	if (rc) {
+		printk(KERN_ERR
+		       "%s: Error updating crypto hash; rc = [%d]\n",
+		       __func__, rc);
+		goto out_release_free_unlock;
+	}
+	rc = crypto_hash_final(&s->hash_desc, s->hash);
+	if (rc) {
+		printk(KERN_ERR
+		       "%s: Error finalizing crypto hash; rc = [%d]\n",
+		       __func__, rc);
+		goto out_release_free_unlock;
+	}
+	for (s->j = 0; s->j < (s->num_rand_bytes - 1); s->j++) {
+		s->block_aligned_filename[s->j] =
+			s->hash[(s->j % ECRYPTFS_TAG_70_DIGEST_SIZE)];
+		if ((s->j % ECRYPTFS_TAG_70_DIGEST_SIZE)
+		    == (ECRYPTFS_TAG_70_DIGEST_SIZE - 1)) {
+			sg_init_one(&s->hash_sg, (u8 *)s->hash,
+				    ECRYPTFS_TAG_70_DIGEST_SIZE);
+			rc = crypto_hash_init(&s->hash_desc);
+			if (rc) {
+				printk(KERN_ERR
+				       "%s: Error initializing crypto hash; "
+				       "rc = [%d]\n", __func__, rc);
+				goto out_release_free_unlock;
+			}
+			rc = crypto_hash_update(&s->hash_desc, &s->hash_sg,
+						ECRYPTFS_TAG_70_DIGEST_SIZE);
+			if (rc) {
+				printk(KERN_ERR
+				       "%s: Error updating crypto hash; "
+				       "rc = [%d]\n", __func__, rc);
+				goto out_release_free_unlock;
+			}
+			rc = crypto_hash_final(&s->hash_desc, s->tmp_hash);
+			if (rc) {
+				printk(KERN_ERR
+				       "%s: Error finalizing crypto hash; "
+				       "rc = [%d]\n", __func__, rc);
+				goto out_release_free_unlock;
+			}
+			memcpy(s->hash, s->tmp_hash,
+			       ECRYPTFS_TAG_70_DIGEST_SIZE);
+		}
+		if (s->block_aligned_filename[s->j] == '\0')
+			s->block_aligned_filename[s->j] = ECRYPTFS_NON_NULL;
+	}
+	memcpy(&s->block_aligned_filename[s->num_rand_bytes], filename,
+	       filename_size);
+	rc = virt_to_scatterlist(s->block_aligned_filename,
+				 s->block_aligned_filename_size, &s->src_sg, 1);
+	if (rc != 1) {
+		printk(KERN_ERR "%s: Internal error whilst attempting to "
+		       "convert filename memory to scatterlist; "
+		       "expected rc = 1; got rc = [%d]. "
+		       "block_aligned_filename_size = [%zd]\n", __func__, rc,
+		       s->block_aligned_filename_size);
+		goto out_release_free_unlock;
+	}
+	rc = virt_to_scatterlist(&dest[s->i], s->block_aligned_filename_size,
+				 &s->dst_sg, 1);
+	if (rc != 1) {
+		printk(KERN_ERR "%s: Internal error whilst attempting to "
+		       "convert encrypted filename memory to scatterlist; "
+		       "expected rc = 1; got rc = [%d]. "
+		       "block_aligned_filename_size = [%zd]\n", __func__, rc,
+		       s->block_aligned_filename_size);
+		goto out_release_free_unlock;
+	}
+	/* The characters in the first block effectively do the job
+	 * of the IV here, so we just use 0's for the IV. Note the
+	 * constraint that ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES
+	 * >= ECRYPTFS_MAX_IV_BYTES. */
+	memset(s->iv, 0, ECRYPTFS_MAX_IV_BYTES);
+	s->desc.info = s->iv;
+	rc = crypto_blkcipher_setkey(
+		s->desc.tfm,
+		s->auth_tok->token.password.session_key_encryption_key,
+		mount_crypt_stat->global_default_fn_cipher_key_bytes);
+	if (rc < 0) {
+		printk(KERN_ERR "%s: Error setting key for crypto context; "
+		       "rc = [%d]. s->auth_tok->token.password.session_key_"
+		       "encryption_key = [0x%p]; mount_crypt_stat->"
+		       "global_default_fn_cipher_key_bytes = [%zd]\n", __func__,
+		       rc,
+		       s->auth_tok->token.password.session_key_encryption_key,
+		       mount_crypt_stat->global_default_fn_cipher_key_bytes);
+		goto out_release_free_unlock;
+	}
+	rc = crypto_blkcipher_encrypt_iv(&s->desc, &s->dst_sg, &s->src_sg,
+					 s->block_aligned_filename_size);
+	if (rc) {
+		printk(KERN_ERR "%s: Error attempting to encrypt filename; "
+		       "rc = [%d]\n", __func__, rc);
+		goto out_release_free_unlock;
+	}
+	s->i += s->block_aligned_filename_size;
+	(*packet_size) = s->i;
+	(*remaining_bytes) -= (*packet_size);
+out_release_free_unlock:
+	crypto_free_hash(s->hash_desc.tfm);
+out_free_unlock:
+	memset(s->block_aligned_filename, 0, s->block_aligned_filename_size);
+	kfree(s->block_aligned_filename);
+out_unlock:
+	mutex_unlock(s->tfm_mutex);
+out:
+	kfree(s);
+	return rc;
+}
+
+struct ecryptfs_parse_tag_70_packet_silly_stack {
+	u8 cipher_code;
+	size_t max_packet_size;
+	size_t packet_size_len;
+	size_t parsed_tag_70_packet_size;
+	size_t block_aligned_filename_size;
+	size_t block_size;
+	size_t i;
+	struct mutex *tfm_mutex;
+	char *decrypted_filename;
+	struct ecryptfs_auth_tok *auth_tok;
+	struct scatterlist src_sg;
+	struct scatterlist dst_sg;
+	struct blkcipher_desc desc;
+	char fnek_sig_hex[ECRYPTFS_SIG_SIZE_HEX + 1];
+	char iv[ECRYPTFS_MAX_IV_BYTES];
+	char cipher_string[ECRYPTFS_MAX_CIPHER_NAME_SIZE];
+};
+
+/**
+ * parse_tag_70_packet - Parse and process FNEK-encrypted passphrase packet
+ * @filename: This function kmalloc's the memory for the filename
+ * @filename_size: This function sets this to the amount of memory
+ *                 kmalloc'd for the filename
+ * @packet_size: This function sets this to the the number of octets
+ *               in the packet parsed
+ * @mount_crypt_stat: The mount-wide cryptographic context
+ * @data: The memory location containing the start of the tag 70
+ *        packet
+ * @max_packet_size: The maximum legal size of the packet to be parsed
+ *                   from @data
+ *
+ * Returns zero on success; non-zero otherwise
+ */
+int
+ecryptfs_parse_tag_70_packet(char **filename, size_t *filename_size,
+			     size_t *packet_size,
+			     struct ecryptfs_mount_crypt_stat *mount_crypt_stat,
+			     char *data, size_t max_packet_size)
+{
+	struct ecryptfs_parse_tag_70_packet_silly_stack *s;
+	int rc = 0;
+
+	(*packet_size) = 0;
+	(*filename_size) = 0;
+	(*filename) = NULL;
+	s = kmalloc(sizeof(*s), GFP_KERNEL);
+	if (!s) {
+		printk(KERN_ERR "%s: Out of memory whilst trying to kmalloc "
+		       "[%zd] bytes of kernel memory\n", __func__, sizeof(*s));
+		goto out;
+	}
+	s->desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+	if (max_packet_size < (1 + 1 + ECRYPTFS_SIG_SIZE + 1 + 1)) {
+		printk(KERN_WARNING "%s: max_packet_size is [%zd]; it must be "
+		       "at least [%d]\n", __func__, max_packet_size,
+			(1 + 1 + ECRYPTFS_SIG_SIZE + 1 + 1));
+		rc = -EINVAL;
+		goto out;
+	}
+	/* Octet 0: Tag 70 identifier
+	 * Octets 1-N1: Tag 70 packet size (includes cipher identifier
+	 *              and block-aligned encrypted filename size)
+	 * Octets N1-N2: FNEK sig (ECRYPTFS_SIG_SIZE)
+	 * Octet N2-N3: Cipher identifier (1 octet)
+	 * Octets N3-N4: Block-aligned encrypted filename
+	 *  - Consists of a minimum number of random numbers, a \0
+	 *    separator, and then the filename */
+	if (data[(*packet_size)++] != ECRYPTFS_TAG_70_PACKET_TYPE) {
+		printk(KERN_WARNING "%s: Invalid packet tag [0x%.2x]; must be "
+		       "tag [0x%.2x]\n", __func__,
+		       data[((*packet_size) - 1)], ECRYPTFS_TAG_70_PACKET_TYPE);
+		rc = -EINVAL;
+		goto out;
+	}
+	rc = ecryptfs_parse_packet_length(&data[(*packet_size)],
+					  &s->parsed_tag_70_packet_size,
+					  &s->packet_size_len);
+	if (rc) {
+		printk(KERN_WARNING "%s: Error parsing packet length; "
+		       "rc = [%d]\n", __func__, rc);
+		goto out;
+	}
+	s->block_aligned_filename_size = (s->parsed_tag_70_packet_size
+					  - ECRYPTFS_SIG_SIZE - 1);
+	if ((1 + s->packet_size_len + s->parsed_tag_70_packet_size)
+	    > max_packet_size) {
+		printk(KERN_WARNING "%s: max_packet_size is [%zd]; real packet "
+		       "size is [%zd]\n", __func__, max_packet_size,
+		       (1 + s->packet_size_len + 1
+			+ s->block_aligned_filename_size));
+		rc = -EINVAL;
+		goto out;
+	}
+	(*packet_size) += s->packet_size_len;
+	ecryptfs_to_hex(s->fnek_sig_hex, &data[(*packet_size)],
+			ECRYPTFS_SIG_SIZE);
+	s->fnek_sig_hex[ECRYPTFS_SIG_SIZE_HEX] = '\0';
+	(*packet_size) += ECRYPTFS_SIG_SIZE;
+	s->cipher_code = data[(*packet_size)++];
+	rc = ecryptfs_cipher_code_to_string(s->cipher_string, s->cipher_code);
+	if (rc) {
+		printk(KERN_WARNING "%s: Cipher code [%d] is invalid\n",
+		       __func__, s->cipher_code);
+		goto out;
+	}
+	rc = ecryptfs_get_tfm_and_mutex_for_cipher_name(&s->desc.tfm,
+							&s->tfm_mutex,
+							s->cipher_string);
+	if (unlikely(rc)) {
+		printk(KERN_ERR "Internal error whilst attempting to get "
+		       "tfm and mutex for cipher name [%s]; rc = [%d]\n",
+		       s->cipher_string, rc);
+		goto out;
+	}
+	mutex_lock(s->tfm_mutex);
+	rc = virt_to_scatterlist(&data[(*packet_size)],
+				 s->block_aligned_filename_size, &s->src_sg, 1);
+	if (rc != 1) {
+		printk(KERN_ERR "%s: Internal error whilst attempting to "
+		       "convert encrypted filename memory to scatterlist; "
+		       "expected rc = 1; got rc = [%d]. "
+		       "block_aligned_filename_size = [%zd]\n", __func__, rc,
+		       s->block_aligned_filename_size);
+		goto out_unlock;
+	}
+	(*packet_size) += s->block_aligned_filename_size;
+	s->decrypted_filename = kmalloc(s->block_aligned_filename_size,
+					GFP_KERNEL);
+	if (!s->decrypted_filename) {
+		printk(KERN_ERR "%s: Out of memory whilst attempting to "
+		       "kmalloc [%zd] bytes\n", __func__,
+		       s->block_aligned_filename_size);
+		rc = -ENOMEM;
+		goto out_unlock;
+	}
+	rc = virt_to_scatterlist(s->decrypted_filename,
+				 s->block_aligned_filename_size, &s->dst_sg, 1);
+	if (rc != 1) {
+		printk(KERN_ERR "%s: Internal error whilst attempting to "
+		       "convert decrypted filename memory to scatterlist; "
+		       "expected rc = 1; got rc = [%d]. "
+		       "block_aligned_filename_size = [%zd]\n", __func__, rc,
+		       s->block_aligned_filename_size);
+		goto out_free_unlock;
+	}
+	/* The characters in the first block effectively do the job of
+	 * the IV here, so we just use 0's for the IV. Note the
+	 * constraint that ECRYPTFS_FILENAME_MIN_RANDOM_PREPEND_BYTES
+	 * >= ECRYPTFS_MAX_IV_BYTES. */
+	memset(s->iv, 0, ECRYPTFS_MAX_IV_BYTES);
+	s->desc.info = s->iv;
+	rc = ecryptfs_find_auth_tok_for_sig(&s->auth_tok, mount_crypt_stat,
+					    s->fnek_sig_hex);
+	if (rc) {
+		printk(KERN_ERR "%s: Error attempting to find auth tok for "
+		       "fnek sig [%s]; rc = [%d]\n", __func__, s->fnek_sig_hex,
+		       rc);
+		goto out_free_unlock;
+	}
+	/* TODO: Support other key modules than passphrase for
+	 * filename encryption */
+	BUG_ON(s->auth_tok->token_type != ECRYPTFS_PASSWORD);
+	rc = crypto_blkcipher_setkey(
+		s->desc.tfm,
+		s->auth_tok->token.password.session_key_encryption_key,
+		mount_crypt_stat->global_default_fn_cipher_key_bytes);
+	if (rc < 0) {
+		printk(KERN_ERR "%s: Error setting key for crypto context; "
+		       "rc = [%d]. s->auth_tok->token.password.session_key_"
+		       "encryption_key = [0x%p]; mount_crypt_stat->"
+		       "global_default_fn_cipher_key_bytes = [%zd]\n", __func__,
+		       rc,
+		       s->auth_tok->token.password.session_key_encryption_key,
+		       mount_crypt_stat->global_default_fn_cipher_key_bytes);
+		goto out_free_unlock;
+	}
+	rc = crypto_blkcipher_decrypt_iv(&s->desc, &s->dst_sg, &s->src_sg,
+					 s->block_aligned_filename_size);
+	if (rc) {
+		printk(KERN_ERR "%s: Error attempting to decrypt filename; "
+		       "rc = [%d]\n", __func__, rc);
+		goto out_free_unlock;
+	}
+	s->i = 0;
+	while (s->decrypted_filename[s->i] != '\0'
+	       && s->i < s->block_aligned_filename_size)
+		s->i++;
+	if (s->i == s->block_aligned_filename_size) {
+		printk(KERN_WARNING "%s: Invalid tag 70 packet; could not "
+		       "find valid separator between random characters and "
+		       "the filename\n", __func__);
+		rc = -EINVAL;
+		goto out_free_unlock;
+	}
+	s->i++;
+	(*filename_size) = (s->block_aligned_filename_size - s->i);
+	if (!((*filename_size) > 0 && (*filename_size < PATH_MAX))) {
+		printk(KERN_WARNING "%s: Filename size is [%zd], which is "
+		       "invalid\n", __func__, (*filename_size));
+		rc = -EINVAL;
+		goto out_free_unlock;
+	}
+	(*filename) = kmalloc(((*filename_size) + 1), GFP_KERNEL);
+	if (!(*filename)) {
+		printk(KERN_ERR "%s: Out of memory whilst attempting to "
+		       "kmalloc [%zd] bytes\n", __func__,
+		       ((*filename_size) + 1));
+		rc = -ENOMEM;
+		goto out_free_unlock;
+	}
+	memcpy((*filename), &s->decrypted_filename[s->i], (*filename_size));
+	(*filename)[(*filename_size)] = '\0';
+out_free_unlock:
+	kfree(s->decrypted_filename);
+out_unlock:
+	mutex_unlock(s->tfm_mutex);
+out:
+	if (rc) {
+		(*packet_size) = 0;
+		(*filename_size) = 0;
+		(*filename) = NULL;
+	}
+	kfree(s);
+	return rc;
+}
+
+static int
 ecryptfs_get_auth_tok_sig(char **sig, struct ecryptfs_auth_tok *auth_tok)
 {
 	int rc = 0;
@@ -897,30 +1471,6 @@ out:
 	return rc;
 }
 
-static int
-ecryptfs_find_global_auth_tok_for_sig(
-	struct ecryptfs_global_auth_tok **global_auth_tok,
-	struct ecryptfs_mount_crypt_stat *mount_crypt_stat, char *sig)
-{
-	struct ecryptfs_global_auth_tok *walker;
-	int rc = 0;
-
-	(*global_auth_tok) = NULL;
-	mutex_lock(&mount_crypt_stat->global_auth_tok_list_mutex);
-	list_for_each_entry(walker,
-			    &mount_crypt_stat->global_auth_tok_list,
-			    mount_crypt_stat_list) {
-		if (memcmp(walker->sig, sig, ECRYPTFS_SIG_SIZE_HEX) == 0) {
-			(*global_auth_tok) = walker;
-			goto out;
-		}
-	}
-	rc = -EINVAL;
-out:
-	mutex_unlock(&mount_crypt_stat->global_auth_tok_list_mutex);
-	return rc;
-}
-
 /**
  * ecryptfs_verify_version
  * @version: The version number to confirm
@@ -990,43 +1540,6 @@ out:
 }
 
 /**
- * ecryptfs_find_auth_tok_for_sig
- * @auth_tok: Set to the matching auth_tok; NULL if not found
- * @crypt_stat: inode crypt_stat crypto context
- * @sig: Sig of auth_tok to find
- *
- * For now, this function simply looks at the registered auth_tok's
- * linked off the mount_crypt_stat, so all the auth_toks that can be
- * used must be registered at mount time. This function could
- * potentially try a lot harder to find auth_tok's (e.g., by calling
- * out to ecryptfsd to dynamically retrieve an auth_tok object) so
- * that static registration of auth_tok's will no longer be necessary.
- *
- * Returns zero on no error; non-zero on error
- */
-static int
-ecryptfs_find_auth_tok_for_sig(
-	struct ecryptfs_auth_tok **auth_tok,
-	struct ecryptfs_crypt_stat *crypt_stat, char *sig)
-{
-	struct ecryptfs_mount_crypt_stat *mount_crypt_stat =
-		crypt_stat->mount_crypt_stat;
-	struct ecryptfs_global_auth_tok *global_auth_tok;
-	int rc = 0;
-
-	(*auth_tok) = NULL;
-	if (ecryptfs_find_global_auth_tok_for_sig(&global_auth_tok,
-						  mount_crypt_stat, sig)) {
-		struct key *auth_tok_key;
-
-		rc = ecryptfs_keyring_auth_tok_for_sig(&auth_tok_key, auth_tok,
-						       sig);
-	} else
-		(*auth_tok) = global_auth_tok->global_auth_tok;
-	return rc;
-}
-
-/**
  * decrypt_passphrase_encrypted_session_key - Decrypt the session key with the given auth_tok.
  * @auth_tok: The passphrase authentication token to use to encrypt the FEK
  * @crypt_stat: The cryptographic context
@@ -1256,7 +1769,8 @@ find_next_matching_auth_tok:
 			rc = -EINVAL;
 			goto out_wipe_list;
 		}
-		ecryptfs_find_auth_tok_for_sig(&matching_auth_tok, crypt_stat,
+		ecryptfs_find_auth_tok_for_sig(&matching_auth_tok,
+					       crypt_stat->mount_crypt_stat,
 					       candidate_auth_tok_sig);
 		if (matching_auth_tok) {
 			found_auth_tok = 1;
@@ -1336,7 +1850,9 @@ pki_encrypt_session_key(struct ecryptfs_auth_tok *auth_tok,
 	int rc;
 
 	rc = write_tag_66_packet(auth_tok->token.private_key.signature,
-				 ecryptfs_code_for_cipher_string(crypt_stat),
+				 ecryptfs_code_for_cipher_string(
+					 crypt_stat->cipher,
+					 crypt_stat->key_size),
 				 crypt_stat, &payload, &payload_len);
 	if (rc) {
 		ecryptfs_printk(KERN_ERR, "Error generating tag 66 packet\n");
@@ -1696,7 +2212,8 @@ encrypted_session_key_set:
 	dest[(*packet_size)++] = 0x04; /* version 4 */
 	/* TODO: Break from RFC2440 so that arbitrary ciphers can be
 	 * specified with strings */
-	cipher_code = ecryptfs_code_for_cipher_string(crypt_stat);
+	cipher_code = ecryptfs_code_for_cipher_string(crypt_stat->cipher,
+						      crypt_stat->key_size);
 	if (cipher_code == 0) {
 		ecryptfs_printk(KERN_WARNING, "Unable to generate code for "
 				"cipher [%s]\n", crypt_stat->cipher);
diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c
index fd630713c5c..789cf2e1be1 100644
--- a/fs/ecryptfs/main.c
+++ b/fs/ecryptfs/main.c
@@ -206,7 +206,9 @@ enum { ecryptfs_opt_sig, ecryptfs_opt_ecryptfs_sig,
        ecryptfs_opt_cipher, ecryptfs_opt_ecryptfs_cipher,
        ecryptfs_opt_ecryptfs_key_bytes,
        ecryptfs_opt_passthrough, ecryptfs_opt_xattr_metadata,
-       ecryptfs_opt_encrypted_view, ecryptfs_opt_err };
+       ecryptfs_opt_encrypted_view, ecryptfs_opt_fnek_sig,
+       ecryptfs_opt_fn_cipher, ecryptfs_opt_fn_cipher_key_bytes,
+       ecryptfs_opt_err };
 
 static const match_table_t tokens = {
 	{ecryptfs_opt_sig, "sig=%s"},
@@ -217,6 +219,9 @@ static const match_table_t tokens = {
 	{ecryptfs_opt_passthrough, "ecryptfs_passthrough"},
 	{ecryptfs_opt_xattr_metadata, "ecryptfs_xattr_metadata"},
 	{ecryptfs_opt_encrypted_view, "ecryptfs_encrypted_view"},
+	{ecryptfs_opt_fnek_sig, "ecryptfs_fnek_sig=%s"},
+	{ecryptfs_opt_fn_cipher, "ecryptfs_fn_cipher=%s"},
+	{ecryptfs_opt_fn_cipher_key_bytes, "ecryptfs_fn_key_bytes=%u"},
 	{ecryptfs_opt_err, NULL}
 };
 
@@ -281,8 +286,11 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 	int rc = 0;
 	int sig_set = 0;
 	int cipher_name_set = 0;
+	int fn_cipher_name_set = 0;
 	int cipher_key_bytes;
 	int cipher_key_bytes_set = 0;
+	int fn_cipher_key_bytes;
+	int fn_cipher_key_bytes_set = 0;
 	struct ecryptfs_mount_crypt_stat *mount_crypt_stat =
 		&ecryptfs_superblock_to_private(sb)->mount_crypt_stat;
 	substring_t args[MAX_OPT_ARGS];
@@ -290,7 +298,12 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 	char *sig_src;
 	char *cipher_name_dst;
 	char *cipher_name_src;
+	char *fn_cipher_name_dst;
+	char *fn_cipher_name_src;
+	char *fnek_dst;
+	char *fnek_src;
 	char *cipher_key_bytes_src;
+	char *fn_cipher_key_bytes_src;
 
 	if (!options) {
 		rc = -EINVAL;
@@ -322,10 +335,7 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 				global_default_cipher_name;
 			strncpy(cipher_name_dst, cipher_name_src,
 				ECRYPTFS_MAX_CIPHER_NAME_SIZE);
-			ecryptfs_printk(KERN_DEBUG,
-					"The mount_crypt_stat "
-					"global_default_cipher_name set to: "
-					"[%s]\n", cipher_name_dst);
+			cipher_name_dst[ECRYPTFS_MAX_CIPHER_NAME_SIZE] = '\0';
 			cipher_name_set = 1;
 			break;
 		case ecryptfs_opt_ecryptfs_key_bytes:
@@ -335,11 +345,6 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 						   &cipher_key_bytes_src, 0);
 			mount_crypt_stat->global_default_cipher_key_size =
 				cipher_key_bytes;
-			ecryptfs_printk(KERN_DEBUG,
-					"The mount_crypt_stat "
-					"global_default_cipher_key_size "
-					"set to: [%d]\n", mount_crypt_stat->
-					global_default_cipher_key_size);
 			cipher_key_bytes_set = 1;
 			break;
 		case ecryptfs_opt_passthrough:
@@ -356,11 +361,51 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 			mount_crypt_stat->flags |=
 				ECRYPTFS_ENCRYPTED_VIEW_ENABLED;
 			break;
+		case ecryptfs_opt_fnek_sig:
+			fnek_src = args[0].from;
+			fnek_dst =
+				mount_crypt_stat->global_default_fnek_sig;
+			strncpy(fnek_dst, fnek_src, ECRYPTFS_SIG_SIZE_HEX);
+			mount_crypt_stat->global_default_fnek_sig[
+				ECRYPTFS_SIG_SIZE_HEX] = '\0';
+			rc = ecryptfs_add_global_auth_tok(
+				mount_crypt_stat,
+				mount_crypt_stat->global_default_fnek_sig);
+			if (rc) {
+				printk(KERN_ERR "Error attempting to register "
+				       "global fnek sig [%s]; rc = [%d]\n",
+				       mount_crypt_stat->global_default_fnek_sig,
+				       rc);
+				goto out;
+			}
+			mount_crypt_stat->flags |=
+				(ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES
+				 | ECRYPTFS_GLOBAL_ENCFN_USE_MOUNT_FNEK);
+			break;
+		case ecryptfs_opt_fn_cipher:
+			fn_cipher_name_src = args[0].from;
+			fn_cipher_name_dst =
+				mount_crypt_stat->global_default_fn_cipher_name;
+			strncpy(fn_cipher_name_dst, fn_cipher_name_src,
+				ECRYPTFS_MAX_CIPHER_NAME_SIZE);
+			mount_crypt_stat->global_default_fn_cipher_name[
+				ECRYPTFS_MAX_CIPHER_NAME_SIZE] = '\0';
+			fn_cipher_name_set = 1;
+			break;
+		case ecryptfs_opt_fn_cipher_key_bytes:
+			fn_cipher_key_bytes_src = args[0].from;
+			fn_cipher_key_bytes =
+				(int)simple_strtol(fn_cipher_key_bytes_src,
+						   &fn_cipher_key_bytes_src, 0);
+			mount_crypt_stat->global_default_fn_cipher_key_bytes =
+				fn_cipher_key_bytes;
+			fn_cipher_key_bytes_set = 1;
+			break;
 		case ecryptfs_opt_err:
 		default:
-			ecryptfs_printk(KERN_WARNING,
-					"eCryptfs: unrecognized option '%s'\n",
-					p);
+			printk(KERN_WARNING
+			       "%s: eCryptfs: unrecognized option [%s]\n",
+			       __func__, p);
 		}
 	}
 	if (!sig_set) {
@@ -374,33 +419,60 @@ static int ecryptfs_parse_options(struct super_block *sb, char *options)
 		int cipher_name_len = strlen(ECRYPTFS_DEFAULT_CIPHER);
 
 		BUG_ON(cipher_name_len >= ECRYPTFS_MAX_CIPHER_NAME_SIZE);
-
 		strcpy(mount_crypt_stat->global_default_cipher_name,
 		       ECRYPTFS_DEFAULT_CIPHER);
 	}
-	if (!cipher_key_bytes_set) {
+	if ((mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)
+	    && !fn_cipher_name_set)
+		strcpy(mount_crypt_stat->global_default_fn_cipher_name,
+		       mount_crypt_stat->global_default_cipher_name);
+	if (!cipher_key_bytes_set)
 		mount_crypt_stat->global_default_cipher_key_size = 0;
-	}
+	if ((mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)
+	    && !fn_cipher_key_bytes_set)
+		mount_crypt_stat->global_default_fn_cipher_key_bytes =
+			mount_crypt_stat->global_default_cipher_key_size;
 	mutex_lock(&key_tfm_list_mutex);
 	if (!ecryptfs_tfm_exists(mount_crypt_stat->global_default_cipher_name,
-				 NULL))
+				 NULL)) {
 		rc = ecryptfs_add_new_key_tfm(
 			NULL, mount_crypt_stat->global_default_cipher_name,
 			mount_crypt_stat->global_default_cipher_key_size);
-	mutex_unlock(&key_tfm_list_mutex);
-	if (rc) {
-		printk(KERN_ERR "Error attempting to initialize cipher with "
-		       "name = [%s] and key size = [%td]; rc = [%d]\n",
-		       mount_crypt_stat->global_default_cipher_name,
-		       mount_crypt_stat->global_default_cipher_key_size, rc);
-		rc = -EINVAL;
-		goto out;
+		if (rc) {
+			printk(KERN_ERR "Error attempting to initialize "
+			       "cipher with name = [%s] and key size = [%td]; "
+			       "rc = [%d]\n",
+			       mount_crypt_stat->global_default_cipher_name,
+			       mount_crypt_stat->global_default_cipher_key_size,
+			       rc);
+			rc = -EINVAL;
+			mutex_unlock(&key_tfm_list_mutex);
+			goto out;
+		}
 	}
+	if ((mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES)
+	    && !ecryptfs_tfm_exists(
+		    mount_crypt_stat->global_default_fn_cipher_name, NULL)) {
+		rc = ecryptfs_add_new_key_tfm(
+			NULL, mount_crypt_stat->global_default_fn_cipher_name,
+			mount_crypt_stat->global_default_fn_cipher_key_bytes);
+		if (rc) {
+			printk(KERN_ERR "Error attempting to initialize "
+			       "cipher with name = [%s] and key size = [%td]; "
+			       "rc = [%d]\n",
+			       mount_crypt_stat->global_default_fn_cipher_name,
+			       mount_crypt_stat->global_default_fn_cipher_key_bytes,
+			       rc);
+			rc = -EINVAL;
+			mutex_unlock(&key_tfm_list_mutex);
+			goto out;
+		}
+	}
+	mutex_unlock(&key_tfm_list_mutex);
 	rc = ecryptfs_init_global_auth_toks(mount_crypt_stat);
-	if (rc) {
+	if (rc)
 		printk(KERN_WARNING "One or more global auth toks could not "
 		       "properly register; rc = [%d]\n", rc);
-	}
 out:
 	return rc;
 }
diff --git a/fs/ecryptfs/messaging.c b/fs/ecryptfs/messaging.c
index 6913f727624..96ef51489e0 100644
--- a/fs/ecryptfs/messaging.c
+++ b/fs/ecryptfs/messaging.c
@@ -193,7 +193,7 @@ ecryptfs_spawn_daemon(struct ecryptfs_daemon **daemon, uid_t euid,
 	(*daemon) = kzalloc(sizeof(**daemon), GFP_KERNEL);
 	if (!(*daemon)) {
 		rc = -ENOMEM;
-		printk(KERN_ERR "%s: Failed to allocate [%Zd] bytes of "
+		printk(KERN_ERR "%s: Failed to allocate [%zd] bytes of "
 		       "GFP_KERNEL memory\n", __func__, sizeof(**daemon));
 		goto out;
 	}
@@ -435,7 +435,7 @@ int ecryptfs_process_response(struct ecryptfs_message *msg, uid_t euid,
 	msg_ctx->msg = kmalloc(msg_size, GFP_KERNEL);
 	if (!msg_ctx->msg) {
 		rc = -ENOMEM;
-		printk(KERN_ERR "%s: Failed to allocate [%Zd] bytes of "
+		printk(KERN_ERR "%s: Failed to allocate [%zd] bytes of "
 		       "GFP_KERNEL memory\n", __func__, msg_size);
 		goto unlock;
 	}
diff --git a/fs/ecryptfs/miscdev.c b/fs/ecryptfs/miscdev.c
index efd95a0ed1e..a67fea655f4 100644
--- a/fs/ecryptfs/miscdev.c
+++ b/fs/ecryptfs/miscdev.c
@@ -199,7 +199,7 @@ int ecryptfs_send_miscdev(char *data, size_t data_size,
 		if (!msg_ctx->msg) {
 			rc = -ENOMEM;
 			printk(KERN_ERR "%s: Out of memory whilst attempting "
-			       "to kmalloc(%Zd, GFP_KERNEL)\n", __func__,
+			       "to kmalloc(%zd, GFP_KERNEL)\n", __func__,
 			       (sizeof(*msg_ctx->msg) + data_size));
 			goto out_unlock;
 		}
@@ -322,7 +322,7 @@ check_list:
 	if (count < total_length) {
 		rc = 0;
 		printk(KERN_WARNING "%s: Only given user buffer of "
-		       "size [%Zd], but we need [%Zd] to read the "
+		       "size [%zd], but we need [%zd] to read the "
 		       "pending message\n", __func__, count, total_length);
 		goto out_unlock_msg_ctx;
 	}
@@ -376,7 +376,7 @@ static int ecryptfs_miscdev_response(char *data, size_t data_size,
 
 	if ((sizeof(*msg) + msg->data_len) != data_size) {
 		printk(KERN_WARNING "%s: (sizeof(*msg) + msg->data_len) = "
-		       "[%Zd]; data_size = [%Zd]. Invalid packet.\n", __func__,
+		       "[%zd]; data_size = [%zd]. Invalid packet.\n", __func__,
 		       (sizeof(*msg) + msg->data_len), data_size);
 		rc = -EINVAL;
 		goto out;
@@ -421,7 +421,7 @@ ecryptfs_miscdev_write(struct file *file, const char __user *buf,
 	data = kmalloc(count, GFP_KERNEL);
 	if (!data) {
 		printk(KERN_ERR "%s: Out of memory whilst attempting to "
-		       "kmalloc([%Zd], GFP_KERNEL)\n", __func__, count);
+		       "kmalloc([%zd], GFP_KERNEL)\n", __func__, count);
 		goto out;
 	}
 	rc = copy_from_user(data, buf, count);
@@ -436,8 +436,8 @@ ecryptfs_miscdev_write(struct file *file, const char __user *buf,
 	case ECRYPTFS_MSG_RESPONSE:
 		if (count < (1 + 4 + 1 + sizeof(struct ecryptfs_message))) {
 			printk(KERN_WARNING "%s: Minimum acceptable packet "
-			       "size is [%Zd], but amount of data written is "
-			       "only [%Zd]. Discarding response packet.\n",
+			       "size is [%zd], but amount of data written is "
+			       "only [%zd]. Discarding response packet.\n",
 			       __func__,
 			       (1 + 4 + 1 + sizeof(struct ecryptfs_message)),
 			       count);
@@ -455,9 +455,9 @@ ecryptfs_miscdev_write(struct file *file, const char __user *buf,
 		}
 		i += packet_size_length;
 		if ((1 + 4 + packet_size_length + packet_size) != count) {
-			printk(KERN_WARNING "%s: (1 + packet_size_length([%Zd])"
-			       " + packet_size([%Zd]))([%Zd]) != "
-			       "count([%Zd]). Invalid packet format.\n",
+			printk(KERN_WARNING "%s: (1 + packet_size_length([%zd])"
+			       " + packet_size([%zd]))([%zd]) != "
+			       "count([%zd]). Invalid packet format.\n",
 			       __func__, packet_size_length, packet_size,
 			       (1 + packet_size_length + packet_size), count);
 			goto out_free;
diff --git a/fs/ecryptfs/mmap.c b/fs/ecryptfs/mmap.c
index 04d7b3fa1ac..46cec2b6979 100644
--- a/fs/ecryptfs/mmap.c
+++ b/fs/ecryptfs/mmap.c
@@ -288,7 +288,7 @@ static int ecryptfs_write_begin(struct file *file,
 	loff_t prev_page_end_size;
 	int rc = 0;
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
diff --git a/fs/exec.c b/fs/exec.c
index 911dd0fd7e0..605be573fe8 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -52,17 +52,13 @@
 #include <linux/audit.h>
 #include <linux/tracehook.h>
 #include <linux/kmod.h>
+#include <linux/fsnotify.h>
 
 #include <asm/uaccess.h>
 #include <asm/mmu_context.h>
 #include <asm/tlb.h>
 #include "internal.h"
 
-#ifdef __alpha__
-/* for /sbin/loader handling in search_binary_handler() */
-#include <linux/a.out.h>
-#endif
-
 int core_uses_pid;
 char core_pattern[CORENAME_MAX_SIZE] = "core";
 int suid_dumpable = 0;
@@ -128,7 +124,8 @@ asmlinkage long sys_uselib(const char __user * library)
 	if (nd.path.mnt->mnt_flags & MNT_NOEXEC)
 		goto exit;
 
-	error = vfs_permission(&nd, MAY_READ | MAY_EXEC | MAY_OPEN);
+	error = inode_permission(nd.path.dentry->d_inode,
+				 MAY_READ | MAY_EXEC | MAY_OPEN);
 	if (error)
 		goto exit;
 
@@ -137,6 +134,8 @@ asmlinkage long sys_uselib(const char __user * library)
 	if (IS_ERR(file))
 		goto out;
 
+	fsnotify_open(file->f_path.dentry);
+
 	error = -ENOEXEC;
 	if(file->f_op) {
 		struct linux_binfmt * fmt;
@@ -234,13 +233,13 @@ static void flush_arg_page(struct linux_binprm *bprm, unsigned long pos,
 
 static int __bprm_mm_init(struct linux_binprm *bprm)
 {
-	int err = -ENOMEM;
+	int err;
 	struct vm_area_struct *vma = NULL;
 	struct mm_struct *mm = bprm->mm;
 
 	bprm->vma = vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
 	if (!vma)
-		goto err;
+		return -ENOMEM;
 
 	down_write(&mm->mmap_sem);
 	vma->vm_mm = mm;
@@ -253,28 +252,20 @@ static int __bprm_mm_init(struct linux_binprm *bprm)
 	 */
 	vma->vm_end = STACK_TOP_MAX;
 	vma->vm_start = vma->vm_end - PAGE_SIZE;
-
 	vma->vm_flags = VM_STACK_FLAGS;
 	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
 	err = insert_vm_struct(mm, vma);
-	if (err) {
-		up_write(&mm->mmap_sem);
+	if (err)
 		goto err;
-	}
 
 	mm->stack_vm = mm->total_vm = 1;
 	up_write(&mm->mmap_sem);
-
 	bprm->p = vma->vm_end - sizeof(void *);
-
 	return 0;
-
 err:
-	if (vma) {
-		bprm->vma = NULL;
-		kmem_cache_free(vm_area_cachep, vma);
-	}
-
+	up_write(&mm->mmap_sem);
+	bprm->vma = NULL;
+	kmem_cache_free(vm_area_cachep, vma);
 	return err;
 }
 
@@ -681,7 +672,7 @@ struct file *open_exec(const char *name)
 	if (nd.path.mnt->mnt_flags & MNT_NOEXEC)
 		goto out_path_put;
 
-	err = vfs_permission(&nd, MAY_EXEC | MAY_OPEN);
+	err = inode_permission(nd.path.dentry->d_inode, MAY_EXEC | MAY_OPEN);
 	if (err)
 		goto out_path_put;
 
@@ -689,6 +680,8 @@ struct file *open_exec(const char *name)
 	if (IS_ERR(file))
 		return file;
 
+	fsnotify_open(file->f_path.dentry);
+
 	err = deny_write_access(file);
 	if (err) {
 		fput(file);
@@ -774,7 +767,6 @@ static int de_thread(struct task_struct *tsk)
 	struct signal_struct *sig = tsk->signal;
 	struct sighand_struct *oldsighand = tsk->sighand;
 	spinlock_t *lock = &oldsighand->siglock;
-	struct task_struct *leader = NULL;
 	int count;
 
 	if (thread_group_empty(tsk))
@@ -812,7 +804,7 @@ static int de_thread(struct task_struct *tsk)
 	 * and to assume its PID:
 	 */
 	if (!thread_group_leader(tsk)) {
-		leader = tsk->group_leader;
+		struct task_struct *leader = tsk->group_leader;
 
 		sig->notify_count = -1;	/* for exit_notify() */
 		for (;;) {
@@ -864,8 +856,9 @@ static int de_thread(struct task_struct *tsk)
 
 		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
 		leader->exit_state = EXIT_DEAD;
-
 		write_unlock_irq(&tasklist_lock);
+
+		release_task(leader);
 	}
 
 	sig->group_exit_task = NULL;
@@ -874,8 +867,6 @@ static int de_thread(struct task_struct *tsk)
 no_thread_group:
 	exit_itimers(sig);
 	flush_itimer_signals();
-	if (leader)
-		release_task(leader);
 
 	if (atomic_read(&oldsighand->count) != 1) {
 		struct sighand_struct *newsighand;
@@ -1181,41 +1172,7 @@ int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs)
 	unsigned int depth = bprm->recursion_depth;
 	int try,retval;
 	struct linux_binfmt *fmt;
-#ifdef __alpha__
-	/* handle /sbin/loader.. */
-	{
-	    struct exec * eh = (struct exec *) bprm->buf;
-
-	    if (!bprm->loader && eh->fh.f_magic == 0x183 &&
-		(eh->fh.f_flags & 0x3000) == 0x3000)
-	    {
-		struct file * file;
-		unsigned long loader;
 
-		allow_write_access(bprm->file);
-		fput(bprm->file);
-		bprm->file = NULL;
-
-		loader = bprm->vma->vm_end - sizeof(void *);
-
-		file = open_exec("/sbin/loader");
-		retval = PTR_ERR(file);
-		if (IS_ERR(file))
-			return retval;
-
-		/* Remember if the application is TASO.  */
-		bprm->taso = eh->ah.entry < 0x100000000UL;
-
-		bprm->file = file;
-		bprm->loader = loader;
-		retval = prepare_binprm(bprm);
-		if (retval<0)
-			return retval;
-		/* should call search_binary_handler recursively here,
-		   but it does not matter */
-	    }
-	}
-#endif
 	retval = security_bprm_check(bprm);
 	if (retval)
 		return retval;
@@ -1737,7 +1694,7 @@ int get_dumpable(struct mm_struct *mm)
 	return (ret >= 2) ? 2 : ret;
 }
 
-int do_coredump(long signr, int exit_code, struct pt_regs * regs)
+void do_coredump(long signr, int exit_code, struct pt_regs *regs)
 {
 	struct core_state core_state;
 	char corename[CORENAME_MAX_SIZE + 1];
@@ -1821,6 +1778,11 @@ int do_coredump(long signr, int exit_code, struct pt_regs * regs)
 
  	if (ispipe) {
 		helper_argv = argv_split(GFP_KERNEL, corename+1, &helper_argc);
+		if (!helper_argv) {
+			printk(KERN_WARNING "%s failed to allocate memory\n",
+			       __func__);
+			goto fail_unlock;
+		}
 		/* Terminate the string before the first option */
 		delimit = strchr(corename, ' ');
 		if (delimit)
@@ -1888,5 +1850,5 @@ fail_unlock:
 	put_cred(cred);
 	coredump_finish(mm);
 fail:
-	return retval;
+	return;
 }
diff --git a/fs/ext2/ialloc.c b/fs/ext2/ialloc.c
index 8d0add62587..66321a877e7 100644
--- a/fs/ext2/ialloc.c
+++ b/fs/ext2/ialloc.c
@@ -565,12 +565,8 @@ got:
 	inode->i_blocks = 0;
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC;
 	memset(ei->i_data, 0, sizeof(ei->i_data));
-	ei->i_flags = EXT2_I(dir)->i_flags & ~EXT2_BTREE_FL;
-	if (S_ISLNK(mode))
-		ei->i_flags &= ~(EXT2_IMMUTABLE_FL|EXT2_APPEND_FL);
-	/* dirsync is only applied to directories */
-	if (!S_ISDIR(mode))
-		ei->i_flags &= ~EXT2_DIRSYNC_FL;
+	ei->i_flags =
+		ext2_mask_flags(mode, EXT2_I(dir)->i_flags & EXT2_FL_INHERITED);
 	ei->i_faddr = 0;
 	ei->i_frag_no = 0;
 	ei->i_frag_size = 0;
@@ -585,7 +581,10 @@ got:
 	spin_lock(&sbi->s_next_gen_lock);
 	inode->i_generation = sbi->s_next_generation++;
 	spin_unlock(&sbi->s_next_gen_lock);
-	insert_inode_hash(inode);
+	if (insert_inode_locked(inode) < 0) {
+		err = -EINVAL;
+		goto fail_drop;
+	}
 
 	if (DQUOT_ALLOC_INODE(inode)) {
 		err = -EDQUOT;
@@ -612,6 +611,7 @@ fail_drop:
 	DQUOT_DROP(inode);
 	inode->i_flags |= S_NOQUOTA;
 	inode->i_nlink = 0;
+	unlock_new_inode(inode);
 	iput(inode);
 	return ERR_PTR(err);
 
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index 7658b33e265..23fff2f8778 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -32,6 +32,7 @@
 #include <linux/buffer_head.h>
 #include <linux/mpage.h>
 #include <linux/fiemap.h>
+#include <linux/namei.h>
 #include "ext2.h"
 #include "acl.h"
 #include "xip.h"
@@ -497,8 +498,6 @@ static int ext2_alloc_branch(struct inode *inode,
  * ext2_splice_branch - splice the allocated branch onto inode.
  * @inode: owner
  * @block: (logical) number of block we are adding
- * @chain: chain of indirect blocks (with a missing link - see
- *	ext2_alloc_branch)
  * @where: location of missing link
  * @num:   number of indirect blocks we are adding
  * @blks:  number of direct blocks we are adding
@@ -1286,9 +1285,11 @@ struct inode *ext2_iget (struct super_block *sb, unsigned long ino)
 		else
 			inode->i_mapping->a_ops = &ext2_aops;
 	} else if (S_ISLNK(inode->i_mode)) {
-		if (ext2_inode_is_fast_symlink(inode))
+		if (ext2_inode_is_fast_symlink(inode)) {
 			inode->i_op = &ext2_fast_symlink_inode_operations;
-		else {
+			nd_terminate_link(ei->i_data, inode->i_size,
+				sizeof(ei->i_data) - 1);
+		} else {
 			inode->i_op = &ext2_symlink_inode_operations;
 			if (test_opt(inode->i_sb, NOBH))
 				inode->i_mapping->a_ops = &ext2_nobh_aops;
diff --git a/fs/ext2/ioctl.c b/fs/ext2/ioctl.c
index de876fa793e..7cb4badef92 100644
--- a/fs/ext2/ioctl.c
+++ b/fs/ext2/ioctl.c
@@ -50,8 +50,7 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			goto setflags_out;
 		}
 
-		if (!S_ISDIR(inode->i_mode))
-			flags &= ~EXT2_DIRSYNC_FL;
+		flags = ext2_mask_flags(inode->i_mode, flags);
 
 		mutex_lock(&inode->i_mutex);
 		/* Is it quota file? Do not allow user to mess with it */
diff --git a/fs/ext2/namei.c b/fs/ext2/namei.c
index 2a747252ec1..90ea17998a7 100644
--- a/fs/ext2/namei.c
+++ b/fs/ext2/namei.c
@@ -41,9 +41,11 @@ static inline int ext2_add_nondir(struct dentry *dentry, struct inode *inode)
 	int err = ext2_add_link(dentry, inode);
 	if (!err) {
 		d_instantiate(dentry, inode);
+		unlock_new_inode(inode);
 		return 0;
 	}
 	inode_dec_link_count(inode);
+	unlock_new_inode(inode);
 	iput(inode);
 	return err;
 }
@@ -170,6 +172,7 @@ out:
 
 out_fail:
 	inode_dec_link_count(inode);
+	unlock_new_inode(inode);
 	iput (inode);
 	goto out;
 }
@@ -178,6 +181,7 @@ static int ext2_link (struct dentry * old_dentry, struct inode * dir,
 	struct dentry *dentry)
 {
 	struct inode *inode = old_dentry->d_inode;
+	int err;
 
 	if (inode->i_nlink >= EXT2_LINK_MAX)
 		return -EMLINK;
@@ -186,7 +190,14 @@ static int ext2_link (struct dentry * old_dentry, struct inode * dir,
 	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
-	return ext2_add_nondir(dentry, inode);
+	err = ext2_add_link(dentry, inode);
+	if (!err) {
+		d_instantiate(dentry, inode);
+		return 0;
+	}
+	inode_dec_link_count(inode);
+	iput(inode);
+	return err;
 }
 
 static int ext2_mkdir(struct inode * dir, struct dentry * dentry, int mode)
@@ -222,12 +233,14 @@ static int ext2_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 		goto out_fail;
 
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 out:
 	return err;
 
 out_fail:
 	inode_dec_link_count(inode);
 	inode_dec_link_count(inode);
+	unlock_new_inode(inode);
 	iput(inode);
 out_dir:
 	inode_dec_link_count(dir);
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index 647cd888ac8..da8bdeaa2e6 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -132,6 +132,7 @@ static void ext2_put_super (struct super_block * sb)
 	percpu_counter_destroy(&sbi->s_dirs_counter);
 	brelse (sbi->s_sbh);
 	sb->s_fs_info = NULL;
+	kfree(sbi->s_blockgroup_lock);
 	kfree(sbi);
 
 	return;
@@ -756,6 +757,13 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent)
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
 		return -ENOMEM;
+
+	sbi->s_blockgroup_lock =
+		kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL);
+	if (!sbi->s_blockgroup_lock) {
+		kfree(sbi);
+		return -ENOMEM;
+	}
 	sb->s_fs_info = sbi;
 	sbi->s_sb_block = sb_block;
 
@@ -983,7 +991,7 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent)
 		printk ("EXT2-fs: not enough memory\n");
 		goto failed_mount;
 	}
-	bgl_lock_init(&sbi->s_blockgroup_lock);
+	bgl_lock_init(sbi->s_blockgroup_lock);
 	sbi->s_debts = kcalloc(sbi->s_groups_count, sizeof(*sbi->s_debts), GFP_KERNEL);
 	if (!sbi->s_debts) {
 		printk ("EXT2-fs: not enough memory\n");
diff --git a/fs/ext3/hash.c b/fs/ext3/hash.c
index c30e149fbd2..7d215b4d4f2 100644
--- a/fs/ext3/hash.c
+++ b/fs/ext3/hash.c
@@ -35,23 +35,71 @@ static void TEA_transform(__u32 buf[4], __u32 const in[])
 
 
 /* The old legacy hash */
-static __u32 dx_hack_hash (const char *name, int len)
+static __u32 dx_hack_hash_unsigned(const char *name, int len)
 {
-	__u32 hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9;
+	__u32 hash, hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9;
+	const unsigned char *ucp = (const unsigned char *) name;
+
+	while (len--) {
+		hash = hash1 + (hash0 ^ (((int) *ucp++) * 7152373));
+
+		if (hash & 0x80000000)
+			hash -= 0x7fffffff;
+		hash1 = hash0;
+		hash0 = hash;
+	}
+	return hash0 << 1;
+}
+
+static __u32 dx_hack_hash_signed(const char *name, int len)
+{
+	__u32 hash, hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9;
+	const signed char *scp = (const signed char *) name;
+
 	while (len--) {
-		__u32 hash = hash1 + (hash0 ^ (*name++ * 7152373));
+		hash = hash1 + (hash0 ^ (((int) *scp++) * 7152373));
 
-		if (hash & 0x80000000) hash -= 0x7fffffff;
+		if (hash & 0x80000000)
+			hash -= 0x7fffffff;
 		hash1 = hash0;
 		hash0 = hash;
 	}
-	return (hash0 << 1);
+	return hash0 << 1;
 }
 
-static void str2hashbuf(const char *msg, int len, __u32 *buf, int num)
+static void str2hashbuf_signed(const char *msg, int len, __u32 *buf, int num)
 {
 	__u32	pad, val;
 	int	i;
+	const signed char *scp = (const signed char *) msg;
+
+	pad = (__u32)len | ((__u32)len << 8);
+	pad |= pad << 16;
+
+	val = pad;
+	if (len > num*4)
+		len = num * 4;
+	for (i = 0; i < len; i++) {
+		if ((i % 4) == 0)
+			val = pad;
+		val = ((int) scp[i]) + (val << 8);
+		if ((i % 4) == 3) {
+			*buf++ = val;
+			val = pad;
+			num--;
+		}
+	}
+	if (--num >= 0)
+		*buf++ = val;
+	while (--num >= 0)
+		*buf++ = pad;
+}
+
+static void str2hashbuf_unsigned(const char *msg, int len, __u32 *buf, int num)
+{
+	__u32	pad, val;
+	int	i;
+	const unsigned char *ucp = (const unsigned char *) msg;
 
 	pad = (__u32)len | ((__u32)len << 8);
 	pad |= pad << 16;
@@ -62,7 +110,7 @@ static void str2hashbuf(const char *msg, int len, __u32 *buf, int num)
 	for (i=0; i < len; i++) {
 		if ((i % 4) == 0)
 			val = pad;
-		val = msg[i] + (val << 8);
+		val = ((int) ucp[i]) + (val << 8);
 		if ((i % 4) == 3) {
 			*buf++ = val;
 			val = pad;
@@ -95,6 +143,8 @@ int ext3fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo)
 	const char	*p;
 	int		i;
 	__u32		in[8], buf[4];
+	void		(*str2hashbuf)(const char *, int, __u32 *, int) =
+				str2hashbuf_signed;
 
 	/* Initialize the default seed for the hash checksum functions */
 	buf[0] = 0x67452301;
@@ -113,13 +163,18 @@ int ext3fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo)
 	}
 
 	switch (hinfo->hash_version) {
+	case DX_HASH_LEGACY_UNSIGNED:
+		hash = dx_hack_hash_unsigned(name, len);
+		break;
 	case DX_HASH_LEGACY:
-		hash = dx_hack_hash(name, len);
+		hash = dx_hack_hash_signed(name, len);
 		break;
+	case DX_HASH_HALF_MD4_UNSIGNED:
+		str2hashbuf = str2hashbuf_unsigned;
 	case DX_HASH_HALF_MD4:
 		p = name;
 		while (len > 0) {
-			str2hashbuf(p, len, in, 8);
+			(*str2hashbuf)(p, len, in, 8);
 			half_md4_transform(buf, in);
 			len -= 32;
 			p += 32;
@@ -127,10 +182,12 @@ int ext3fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo)
 		minor_hash = buf[2];
 		hash = buf[1];
 		break;
+	case DX_HASH_TEA_UNSIGNED:
+		str2hashbuf = str2hashbuf_unsigned;
 	case DX_HASH_TEA:
 		p = name;
 		while (len > 0) {
-			str2hashbuf(p, len, in, 4);
+			(*str2hashbuf)(p, len, in, 4);
 			TEA_transform(buf, in);
 			len -= 16;
 			p += 16;
diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c
index 490bd0ed789..8de6c720e51 100644
--- a/fs/ext3/ialloc.c
+++ b/fs/ext3/ialloc.c
@@ -559,12 +559,8 @@ got:
 	ei->i_dir_start_lookup = 0;
 	ei->i_disksize = 0;
 
-	ei->i_flags = EXT3_I(dir)->i_flags & ~EXT3_INDEX_FL;
-	if (S_ISLNK(mode))
-		ei->i_flags &= ~(EXT3_IMMUTABLE_FL|EXT3_APPEND_FL);
-	/* dirsync only applies to directories */
-	if (!S_ISDIR(mode))
-		ei->i_flags &= ~EXT3_DIRSYNC_FL;
+	ei->i_flags =
+		ext3_mask_flags(mode, EXT3_I(dir)->i_flags & EXT3_FL_INHERITED);
 #ifdef EXT3_FRAGMENTS
 	ei->i_faddr = 0;
 	ei->i_frag_no = 0;
@@ -579,7 +575,10 @@ got:
 	ext3_set_inode_flags(inode);
 	if (IS_DIRSYNC(inode))
 		handle->h_sync = 1;
-	insert_inode_hash(inode);
+	if (insert_inode_locked(inode) < 0) {
+		err = -EINVAL;
+		goto fail_drop;
+	}
 	spin_lock(&sbi->s_next_gen_lock);
 	inode->i_generation = sbi->s_next_generation++;
 	spin_unlock(&sbi->s_next_gen_lock);
@@ -627,6 +626,7 @@ fail_drop:
 	DQUOT_DROP(inode);
 	inode->i_flags |= S_NOQUOTA;
 	inode->i_nlink = 0;
+	unlock_new_inode(inode);
 	iput(inode);
 	brelse(bitmap_bh);
 	return ERR_PTR(err);
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index f8424ad8997..5fa453b49a6 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -37,6 +37,7 @@
 #include <linux/uio.h>
 #include <linux/bio.h>
 #include <linux/fiemap.h>
+#include <linux/namei.h>
 #include "xattr.h"
 #include "acl.h"
 
@@ -1160,7 +1161,7 @@ static int ext3_write_begin(struct file *file, struct address_space *mapping,
 	to = from + len;
 
 retry:
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
@@ -2817,9 +2818,11 @@ struct inode *ext3_iget(struct super_block *sb, unsigned long ino)
 		inode->i_op = &ext3_dir_inode_operations;
 		inode->i_fop = &ext3_dir_operations;
 	} else if (S_ISLNK(inode->i_mode)) {
-		if (ext3_inode_is_fast_symlink(inode))
+		if (ext3_inode_is_fast_symlink(inode)) {
 			inode->i_op = &ext3_fast_symlink_inode_operations;
-		else {
+			nd_terminate_link(ei->i_data, inode->i_size,
+				sizeof(ei->i_data) - 1);
+		} else {
 			inode->i_op = &ext3_symlink_inode_operations;
 			ext3_set_aops(inode);
 		}
diff --git a/fs/ext3/ioctl.c b/fs/ext3/ioctl.c
index b7394d05ee8..5e86ce9a86e 100644
--- a/fs/ext3/ioctl.c
+++ b/fs/ext3/ioctl.c
@@ -53,8 +53,7 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd,
 			goto flags_out;
 		}
 
-		if (!S_ISDIR(inode->i_mode))
-			flags &= ~EXT3_DIRSYNC_FL;
+		flags = ext3_mask_flags(inode->i_mode, flags);
 
 		mutex_lock(&inode->i_mutex);
 		/* Is it quota file? Do not allow user to mess with it */
diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
index 3e5edc92aa0..69a3d19ca9f 100644
--- a/fs/ext3/namei.c
+++ b/fs/ext3/namei.c
@@ -74,10 +74,6 @@ static struct buffer_head *ext3_append(handle_t *handle,
 #define assert(test) J_ASSERT(test)
 #endif
 
-#ifndef swap
-#define swap(x, y) do { typeof(x) z = x; x = y; y = z; } while (0)
-#endif
-
 #ifdef DX_DEBUG
 #define dxtrace(command) command
 #else
@@ -368,6 +364,8 @@ dx_probe(struct qstr *entry, struct inode *dir,
 		goto fail;
 	}
 	hinfo->hash_version = root->info.hash_version;
+	if (hinfo->hash_version <= DX_HASH_TEA)
+		hinfo->hash_version += EXT3_SB(dir->i_sb)->s_hash_unsigned;
 	hinfo->seed = EXT3_SB(dir->i_sb)->s_hash_seed;
 	if (entry)
 		ext3fs_dirhash(entry->name, entry->len, hinfo);
@@ -636,6 +634,9 @@ int ext3_htree_fill_tree(struct file *dir_file, __u32 start_hash,
 	dir = dir_file->f_path.dentry->d_inode;
 	if (!(EXT3_I(dir)->i_flags & EXT3_INDEX_FL)) {
 		hinfo.hash_version = EXT3_SB(dir->i_sb)->s_def_hash_version;
+		if (hinfo.hash_version <= DX_HASH_TEA)
+			hinfo.hash_version +=
+				EXT3_SB(dir->i_sb)->s_hash_unsigned;
 		hinfo.seed = EXT3_SB(dir->i_sb)->s_hash_seed;
 		count = htree_dirblock_to_tree(dir_file, dir, 0, &hinfo,
 					       start_hash, start_minor_hash);
@@ -1156,9 +1157,9 @@ static struct ext3_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
 	u32 hash2;
 	struct dx_map_entry *map;
 	char *data1 = (*bh)->b_data, *data2;
-	unsigned split, move, size, i;
+	unsigned split, move, size;
 	struct ext3_dir_entry_2 *de = NULL, *de2;
-	int	err = 0;
+	int	err = 0, i;
 
 	bh2 = ext3_append (handle, dir, &newblock, &err);
 	if (!(bh2)) {
@@ -1398,6 +1399,8 @@ static int make_indexed_dir(handle_t *handle, struct dentry *dentry,
 
 	/* Initialize as for dx_probe */
 	hinfo.hash_version = root->info.hash_version;
+	if (hinfo.hash_version <= DX_HASH_TEA)
+		hinfo.hash_version += EXT3_SB(dir->i_sb)->s_hash_unsigned;
 	hinfo.seed = EXT3_SB(dir->i_sb)->s_hash_seed;
 	ext3fs_dirhash(name, namelen, &hinfo);
 	frame = frames;
@@ -1652,9 +1655,11 @@ static int ext3_add_nondir(handle_t *handle,
 	if (!err) {
 		ext3_mark_inode_dirty(handle, inode);
 		d_instantiate(dentry, inode);
+		unlock_new_inode(inode);
 		return 0;
 	}
 	drop_nlink(inode);
+	unlock_new_inode(inode);
 	iput(inode);
 	return err;
 }
@@ -1765,6 +1770,7 @@ retry:
 	dir_block = ext3_bread (handle, inode, 0, 1, &err);
 	if (!dir_block) {
 		drop_nlink(inode); /* is this nlink == 0? */
+		unlock_new_inode(inode);
 		ext3_mark_inode_dirty(handle, inode);
 		iput (inode);
 		goto out_stop;
@@ -1792,6 +1798,7 @@ retry:
 	err = ext3_add_entry (handle, dentry, inode);
 	if (err) {
 		inode->i_nlink = 0;
+		unlock_new_inode(inode);
 		ext3_mark_inode_dirty(handle, inode);
 		iput (inode);
 		goto out_stop;
@@ -1800,6 +1807,7 @@ retry:
 	ext3_update_dx_flag(dir);
 	ext3_mark_inode_dirty(handle, dir);
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 out_stop:
 	ext3_journal_stop(handle);
 	if (err == -ENOSPC && ext3_should_retry_alloc(dir->i_sb, &retries))
@@ -2170,10 +2178,10 @@ retry:
 		 * We have a transaction open.  All is sweetness.  It also sets
 		 * i_size in generic_commit_write().
 		 */
-		err = __page_symlink(inode, symname, l,
-				mapping_gfp_mask(inode->i_mapping) & ~__GFP_FS);
+		err = __page_symlink(inode, symname, l, 1);
 		if (err) {
 			drop_nlink(inode);
+			unlock_new_inode(inode);
 			ext3_mark_inode_dirty(handle, inode);
 			iput (inode);
 			goto out_stop;
@@ -2221,7 +2229,14 @@ retry:
 	inc_nlink(inode);
 	atomic_inc(&inode->i_count);
 
-	err = ext3_add_nondir(handle, dentry, inode);
+	err = ext3_add_entry(handle, dentry, inode);
+	if (!err) {
+		ext3_mark_inode_dirty(handle, inode);
+		d_instantiate(dentry, inode);
+	} else {
+		drop_nlink(inode);
+		iput(inode);
+	}
 	ext3_journal_stop(handle);
 	if (err == -ENOSPC && ext3_should_retry_alloc(dir->i_sb, &retries))
 		goto retry;
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index f6c94f232ec..b70d90e08a3 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -48,8 +48,8 @@ static int ext3_load_journal(struct super_block *, struct ext3_super_block *,
 			     unsigned long journal_devnum);
 static int ext3_create_journal(struct super_block *, struct ext3_super_block *,
 			       unsigned int);
-static void ext3_commit_super (struct super_block * sb,
-			       struct ext3_super_block * es,
+static int ext3_commit_super(struct super_block *sb,
+			       struct ext3_super_block *es,
 			       int sync);
 static void ext3_mark_recovery_complete(struct super_block * sb,
 					struct ext3_super_block * es);
@@ -60,9 +60,9 @@ static const char *ext3_decode_error(struct super_block * sb, int errno,
 				     char nbuf[16]);
 static int ext3_remount (struct super_block * sb, int * flags, char * data);
 static int ext3_statfs (struct dentry * dentry, struct kstatfs * buf);
-static void ext3_unlockfs(struct super_block *sb);
+static int ext3_unfreeze(struct super_block *sb);
 static void ext3_write_super (struct super_block * sb);
-static void ext3_write_super_lockfs(struct super_block *sb);
+static int ext3_freeze(struct super_block *sb);
 
 /*
  * Wrappers for journal_start/end.
@@ -439,6 +439,7 @@ static void ext3_put_super (struct super_block * sb)
 		ext3_blkdev_remove(sbi);
 	}
 	sb->s_fs_info = NULL;
+	kfree(sbi->s_blockgroup_lock);
 	kfree(sbi);
 	return;
 }
@@ -682,6 +683,26 @@ static struct dentry *ext3_fh_to_parent(struct super_block *sb, struct fid *fid,
 				    ext3_nfs_get_inode);
 }
 
+/*
+ * Try to release metadata pages (indirect blocks, directories) which are
+ * mapped via the block device.  Since these pages could have journal heads
+ * which would prevent try_to_free_buffers() from freeing them, we must use
+ * jbd layer's try_to_free_buffers() function to release them.
+ */
+static int bdev_try_to_free_page(struct super_block *sb, struct page *page,
+				 gfp_t wait)
+{
+	journal_t *journal = EXT3_SB(sb)->s_journal;
+
+	WARN_ON(PageChecked(page));
+	if (!page_has_buffers(page))
+		return 0;
+	if (journal)
+		return journal_try_to_free_buffers(journal, page, 
+						   wait & ~__GFP_WAIT);
+	return try_to_free_buffers(page);
+}
+
 #ifdef CONFIG_QUOTA
 #define QTYPE2NAME(t) ((t)==USRQUOTA?"user":"group")
 #define QTYPE2MOPT(on, t) ((t)==USRQUOTA?((on)##USRJQUOTA):((on)##GRPJQUOTA))
@@ -713,7 +734,9 @@ static struct dquot_operations ext3_quota_operations = {
 	.acquire_dquot	= ext3_acquire_dquot,
 	.release_dquot	= ext3_release_dquot,
 	.mark_dirty	= ext3_mark_dquot_dirty,
-	.write_info	= ext3_write_info
+	.write_info	= ext3_write_info,
+	.alloc_dquot	= dquot_alloc,
+	.destroy_dquot	= dquot_destroy,
 };
 
 static struct quotactl_ops ext3_qctl_operations = {
@@ -736,8 +759,8 @@ static const struct super_operations ext3_sops = {
 	.put_super	= ext3_put_super,
 	.write_super	= ext3_write_super,
 	.sync_fs	= ext3_sync_fs,
-	.write_super_lockfs = ext3_write_super_lockfs,
-	.unlockfs	= ext3_unlockfs,
+	.freeze_fs	= ext3_freeze,
+	.unfreeze_fs	= ext3_unfreeze,
 	.statfs		= ext3_statfs,
 	.remount_fs	= ext3_remount,
 	.clear_inode	= ext3_clear_inode,
@@ -746,6 +769,7 @@ static const struct super_operations ext3_sops = {
 	.quota_read	= ext3_quota_read,
 	.quota_write	= ext3_quota_write,
 #endif
+	.bdev_try_to_free_page = bdev_try_to_free_page,
 };
 
 static const struct export_operations ext3_export_ops = {
@@ -1035,8 +1059,7 @@ static int parse_options (char *options, struct super_block *sb,
 		case Opt_grpjquota:
 			qtype = GRPQUOTA;
 set_qf_name:
-			if ((sb_any_quota_enabled(sb) ||
-			     sb_any_quota_suspended(sb)) &&
+			if (sb_any_quota_loaded(sb) &&
 			    !sbi->s_qf_names[qtype]) {
 				printk(KERN_ERR
 					"EXT3-fs: Cannot change journaled "
@@ -1075,8 +1098,7 @@ set_qf_name:
 		case Opt_offgrpjquota:
 			qtype = GRPQUOTA;
 clear_qf_name:
-			if ((sb_any_quota_enabled(sb) ||
-			     sb_any_quota_suspended(sb)) &&
+			if (sb_any_quota_loaded(sb) &&
 			    sbi->s_qf_names[qtype]) {
 				printk(KERN_ERR "EXT3-fs: Cannot change "
 					"journaled quota options when "
@@ -1095,8 +1117,7 @@ clear_qf_name:
 		case Opt_jqfmt_vfsv0:
 			qfmt = QFMT_VFS_V0;
 set_qf_format:
-			if ((sb_any_quota_enabled(sb) ||
-			     sb_any_quota_suspended(sb)) &&
+			if (sb_any_quota_loaded(sb) &&
 			    sbi->s_jquota_fmt != qfmt) {
 				printk(KERN_ERR "EXT3-fs: Cannot change "
 					"journaled quota options when "
@@ -1115,8 +1136,7 @@ set_qf_format:
 			set_opt(sbi->s_mount_opt, GRPQUOTA);
 			break;
 		case Opt_noquota:
-			if (sb_any_quota_enabled(sb) ||
-			    sb_any_quota_suspended(sb)) {
+			if (sb_any_quota_loaded(sb)) {
 				printk(KERN_ERR "EXT3-fs: Cannot change quota "
 					"options when quota turned on.\n");
 				return 0;
@@ -1548,6 +1568,13 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
 		return -ENOMEM;
+
+	sbi->s_blockgroup_lock =
+		kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL);
+	if (!sbi->s_blockgroup_lock) {
+		kfree(sbi);
+		return -ENOMEM;
+	}
 	sb->s_fs_info = sbi;
 	sbi->s_mount_opt = 0;
 	sbi->s_resuid = EXT3_DEF_RESUID;
@@ -1744,6 +1771,18 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	for (i=0; i < 4; i++)
 		sbi->s_hash_seed[i] = le32_to_cpu(es->s_hash_seed[i]);
 	sbi->s_def_hash_version = es->s_def_hash_version;
+	i = le32_to_cpu(es->s_flags);
+	if (i & EXT2_FLAGS_UNSIGNED_HASH)
+		sbi->s_hash_unsigned = 3;
+	else if ((i & EXT2_FLAGS_SIGNED_HASH) == 0) {
+#ifdef __CHAR_UNSIGNED__
+		es->s_flags |= cpu_to_le32(EXT2_FLAGS_UNSIGNED_HASH);
+		sbi->s_hash_unsigned = 3;
+#else
+		es->s_flags |= cpu_to_le32(EXT2_FLAGS_SIGNED_HASH);
+#endif
+		sb->s_dirt = 1;
+	}
 
 	if (sbi->s_blocks_per_group > blocksize * 8) {
 		printk (KERN_ERR
@@ -1788,7 +1827,7 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 		goto failed_mount;
 	}
 
-	bgl_lock_init(&sbi->s_blockgroup_lock);
+	bgl_lock_init(sbi->s_blockgroup_lock);
 
 	for (i = 0; i < db_count; i++) {
 		block = descriptor_loc(sb, logic_sb_block, i);
@@ -2272,21 +2311,23 @@ static int ext3_create_journal(struct super_block * sb,
 	return 0;
 }
 
-static void ext3_commit_super (struct super_block * sb,
-			       struct ext3_super_block * es,
+static int ext3_commit_super(struct super_block *sb,
+			       struct ext3_super_block *es,
 			       int sync)
 {
 	struct buffer_head *sbh = EXT3_SB(sb)->s_sbh;
+	int error = 0;
 
 	if (!sbh)
-		return;
+		return error;
 	es->s_wtime = cpu_to_le32(get_seconds());
 	es->s_free_blocks_count = cpu_to_le32(ext3_count_free_blocks(sb));
 	es->s_free_inodes_count = cpu_to_le32(ext3_count_free_inodes(sb));
 	BUFFER_TRACE(sbh, "marking dirty");
 	mark_buffer_dirty(sbh);
 	if (sync)
-		sync_dirty_buffer(sbh);
+		error = sync_dirty_buffer(sbh);
+	return error;
 }
 
 
@@ -2400,12 +2441,14 @@ static int ext3_sync_fs(struct super_block *sb, int wait)
  * LVM calls this function before a (read-only) snapshot is created.  This
  * gives us a chance to flush the journal completely and mark the fs clean.
  */
-static void ext3_write_super_lockfs(struct super_block *sb)
+static int ext3_freeze(struct super_block *sb)
 {
+	int error = 0;
+	journal_t *journal;
 	sb->s_dirt = 0;
 
 	if (!(sb->s_flags & MS_RDONLY)) {
-		journal_t *journal = EXT3_SB(sb)->s_journal;
+		journal = EXT3_SB(sb)->s_journal;
 
 		/* Now we set up the journal barrier. */
 		journal_lock_updates(journal);
@@ -2414,20 +2457,28 @@ static void ext3_write_super_lockfs(struct super_block *sb)
 		 * We don't want to clear needs_recovery flag when we failed
 		 * to flush the journal.
 		 */
-		if (journal_flush(journal) < 0)
-			return;
+		error = journal_flush(journal);
+		if (error < 0)
+			goto out;
 
 		/* Journal blocked and flushed, clear needs_recovery flag. */
 		EXT3_CLEAR_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER);
-		ext3_commit_super(sb, EXT3_SB(sb)->s_es, 1);
+		error = ext3_commit_super(sb, EXT3_SB(sb)->s_es, 1);
+		if (error)
+			goto out;
 	}
+	return 0;
+
+out:
+	journal_unlock_updates(journal);
+	return error;
 }
 
 /*
  * Called by LVM after the snapshot is done.  We need to reset the RECOVER
  * flag here, even though the filesystem is not technically dirty yet.
  */
-static void ext3_unlockfs(struct super_block *sb)
+static int ext3_unfreeze(struct super_block *sb)
 {
 	if (!(sb->s_flags & MS_RDONLY)) {
 		lock_super(sb);
@@ -2437,6 +2488,7 @@ static void ext3_unlockfs(struct super_block *sb)
 		unlock_super(sb);
 		journal_unlock_updates(EXT3_SB(sb)->s_journal);
 	}
+	return 0;
 }
 
 static int ext3_remount (struct super_block * sb, int * flags, char * data)
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 38b3acf5683..6bba06b09dd 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -20,6 +20,7 @@
 #include "ext4.h"
 #include "ext4_jbd2.h"
 #include "group.h"
+#include "mballoc.h"
 
 /*
  * balloc.c contains the blocks allocation and deallocation routines
@@ -100,10 +101,10 @@ unsigned ext4_init_block_bitmap(struct super_block *sb, struct buffer_head *bh,
 		 * essentially implementing a per-group read-only flag. */
 		if (!ext4_group_desc_csum_verify(sbi, block_group, gdp)) {
 			ext4_error(sb, __func__,
-				  "Checksum bad for group %lu\n", block_group);
-			gdp->bg_free_blocks_count = 0;
-			gdp->bg_free_inodes_count = 0;
-			gdp->bg_itable_unused = 0;
+				  "Checksum bad for group %u", block_group);
+			ext4_free_blks_set(sb, gdp, 0);
+			ext4_free_inodes_set(sb, gdp, 0);
+			ext4_itable_unused_set(sb, gdp, 0);
 			memset(bh->b_data, 0xff, sb->s_blocksize);
 			return 0;
 		}
@@ -205,15 +206,15 @@ struct ext4_group_desc * ext4_get_group_desc(struct super_block *sb,
 					     ext4_group_t block_group,
 					     struct buffer_head **bh)
 {
-	unsigned long group_desc;
-	unsigned long offset;
+	unsigned int group_desc;
+	unsigned int offset;
 	struct ext4_group_desc *desc;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 
 	if (block_group >= sbi->s_groups_count) {
 		ext4_error(sb, "ext4_get_group_desc",
 			   "block_group >= groups_count - "
-			   "block_group = %lu, groups_count = %lu",
+			   "block_group = %u, groups_count = %u",
 			   block_group, sbi->s_groups_count);
 
 		return NULL;
@@ -225,7 +226,7 @@ struct ext4_group_desc * ext4_get_group_desc(struct super_block *sb,
 	if (!sbi->s_group_desc[group_desc]) {
 		ext4_error(sb, "ext4_get_group_desc",
 			   "Group descriptor not loaded - "
-			   "block_group = %lu, group_desc = %lu, desc = %lu",
+			   "block_group = %u, group_desc = %u, desc = %u",
 			   block_group, group_desc, offset);
 		return NULL;
 	}
@@ -315,29 +316,50 @@ ext4_read_block_bitmap(struct super_block *sb, ext4_group_t block_group)
 	if (unlikely(!bh)) {
 		ext4_error(sb, __func__,
 			    "Cannot read block bitmap - "
-			    "block_group = %lu, block_bitmap = %llu",
+			    "block_group = %u, block_bitmap = %llu",
 			    block_group, bitmap_blk);
 		return NULL;
 	}
-	if (buffer_uptodate(bh) &&
-	    !(desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)))
+
+	if (bitmap_uptodate(bh))
 		return bh;
 
 	lock_buffer(bh);
+	if (bitmap_uptodate(bh)) {
+		unlock_buffer(bh);
+		return bh;
+	}
 	spin_lock(sb_bgl_lock(EXT4_SB(sb), block_group));
 	if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
 		ext4_init_block_bitmap(sb, bh, block_group, desc);
+		set_bitmap_uptodate(bh);
 		set_buffer_uptodate(bh);
-		unlock_buffer(bh);
 		spin_unlock(sb_bgl_lock(EXT4_SB(sb), block_group));
+		unlock_buffer(bh);
 		return bh;
 	}
 	spin_unlock(sb_bgl_lock(EXT4_SB(sb), block_group));
+	if (buffer_uptodate(bh)) {
+		/*
+		 * if not uninit if bh is uptodate,
+		 * bitmap is also uptodate
+		 */
+		set_bitmap_uptodate(bh);
+		unlock_buffer(bh);
+		return bh;
+	}
+	/*
+	 * submit the buffer_head for read. We can
+	 * safely mark the bitmap as uptodate now.
+	 * We do it here so the bitmap uptodate bit
+	 * get set with buffer lock held.
+	 */
+	set_bitmap_uptodate(bh);
 	if (bh_submit_read(bh) < 0) {
 		put_bh(bh);
 		ext4_error(sb, __func__,
 			    "Cannot read block bitmap - "
-			    "block_group = %lu, block_bitmap = %llu",
+			    "block_group = %u, block_bitmap = %llu",
 			    block_group, bitmap_blk);
 		return NULL;
 	}
@@ -350,62 +372,44 @@ ext4_read_block_bitmap(struct super_block *sb, ext4_group_t block_group)
 }
 
 /**
- * ext4_free_blocks_sb() -- Free given blocks and update quota
+ * ext4_add_groupblocks() -- Add given blocks to an existing group
  * @handle:			handle to this transaction
  * @sb:				super block
- * @block:			start physcial block to free
+ * @block:			start physcial block to add to the block group
  * @count:			number of blocks to free
- * @pdquot_freed_blocks:	pointer to quota
  *
- * XXX This function is only used by the on-line resizing code, which
- * should probably be fixed up to call the mballoc variant.  There
- * this needs to be cleaned up later; in fact, I'm not convinced this
- * is 100% correct in the face of the mballoc code.  The online resizing
- * code needs to be fixed up to more tightly (and correctly) interlock
- * with the mballoc code.
+ * This marks the blocks as free in the bitmap. We ask the
+ * mballoc to reload the buddy after this by setting group
+ * EXT4_GROUP_INFO_NEED_INIT_BIT flag
  */
-void ext4_free_blocks_sb(handle_t *handle, struct super_block *sb,
-			 ext4_fsblk_t block, unsigned long count,
-			 unsigned long *pdquot_freed_blocks)
+void ext4_add_groupblocks(handle_t *handle, struct super_block *sb,
+			 ext4_fsblk_t block, unsigned long count)
 {
 	struct buffer_head *bitmap_bh = NULL;
 	struct buffer_head *gd_bh;
 	ext4_group_t block_group;
 	ext4_grpblk_t bit;
-	unsigned long i;
-	unsigned long overflow;
+	unsigned int i;
 	struct ext4_group_desc *desc;
 	struct ext4_super_block *es;
 	struct ext4_sb_info *sbi;
-	int err = 0, ret;
-	ext4_grpblk_t group_freed;
+	int err = 0, ret, blk_free_count;
+	ext4_grpblk_t blocks_freed;
+	struct ext4_group_info *grp;
 
-	*pdquot_freed_blocks = 0;
 	sbi = EXT4_SB(sb);
 	es = sbi->s_es;
-	if (block < le32_to_cpu(es->s_first_data_block) ||
-	    block + count < block ||
-	    block + count > ext4_blocks_count(es)) {
-		ext4_error(sb, "ext4_free_blocks",
-			   "Freeing blocks not in datazone - "
-			   "block = %llu, count = %lu", block, count);
-		goto error_return;
-	}
-
-	ext4_debug("freeing block(s) %llu-%llu\n", block, block + count - 1);
+	ext4_debug("Adding block(s) %llu-%llu\n", block, block + count - 1);
 
-do_more:
-	overflow = 0;
 	ext4_get_group_no_and_offset(sb, block, &block_group, &bit);
+	grp = ext4_get_group_info(sb, block_group);
 	/*
 	 * Check to see if we are freeing blocks across a group
 	 * boundary.
 	 */
 	if (bit + count > EXT4_BLOCKS_PER_GROUP(sb)) {
-		overflow = bit + count - EXT4_BLOCKS_PER_GROUP(sb);
-		count -= overflow;
+		goto error_return;
 	}
-	brelse(bitmap_bh);
 	bitmap_bh = ext4_read_block_bitmap(sb, block_group);
 	if (!bitmap_bh)
 		goto error_return;
@@ -418,18 +422,17 @@ do_more:
 	    in_range(block, ext4_inode_table(sb, desc), sbi->s_itb_per_group) ||
 	    in_range(block + count - 1, ext4_inode_table(sb, desc),
 		     sbi->s_itb_per_group)) {
-		ext4_error(sb, "ext4_free_blocks",
-			   "Freeing blocks in system zones - "
+		ext4_error(sb, __func__,
+			   "Adding blocks in system zones - "
 			   "Block = %llu, count = %lu",
 			   block, count);
 		goto error_return;
 	}
 
 	/*
-	 * We are about to start releasing blocks in the bitmap,
+	 * We are about to add blocks to the bitmap,
 	 * so we need undo access.
 	 */
-	/* @@@ check errors */
 	BUFFER_TRACE(bitmap_bh, "getting undo access");
 	err = ext4_journal_get_undo_access(handle, bitmap_bh);
 	if (err)
@@ -444,107 +447,55 @@ do_more:
 	err = ext4_journal_get_write_access(handle, gd_bh);
 	if (err)
 		goto error_return;
-
-	jbd_lock_bh_state(bitmap_bh);
-
-	for (i = 0, group_freed = 0; i < count; i++) {
-		/*
-		 * An HJ special.  This is expensive...
-		 */
-#ifdef CONFIG_JBD2_DEBUG
-		jbd_unlock_bh_state(bitmap_bh);
-		{
-			struct buffer_head *debug_bh;
-			debug_bh = sb_find_get_block(sb, block + i);
-			if (debug_bh) {
-				BUFFER_TRACE(debug_bh, "Deleted!");
-				if (!bh2jh(bitmap_bh)->b_committed_data)
-					BUFFER_TRACE(debug_bh,
-						"No commited data in bitmap");
-				BUFFER_TRACE2(debug_bh, bitmap_bh, "bitmap");
-				__brelse(debug_bh);
-			}
-		}
-		jbd_lock_bh_state(bitmap_bh);
-#endif
-		if (need_resched()) {
-			jbd_unlock_bh_state(bitmap_bh);
-			cond_resched();
-			jbd_lock_bh_state(bitmap_bh);
-		}
-		/* @@@ This prevents newly-allocated data from being
-		 * freed and then reallocated within the same
-		 * transaction.
-		 *
-		 * Ideally we would want to allow that to happen, but to
-		 * do so requires making jbd2_journal_forget() capable of
-		 * revoking the queued write of a data block, which
-		 * implies blocking on the journal lock.  *forget()
-		 * cannot block due to truncate races.
-		 *
-		 * Eventually we can fix this by making jbd2_journal_forget()
-		 * return a status indicating whether or not it was able
-		 * to revoke the buffer.  On successful revoke, it is
-		 * safe not to set the allocation bit in the committed
-		 * bitmap, because we know that there is no outstanding
-		 * activity on the buffer any more and so it is safe to
-		 * reallocate it.
-		 */
-		BUFFER_TRACE(bitmap_bh, "set in b_committed_data");
-		J_ASSERT_BH(bitmap_bh,
-				bh2jh(bitmap_bh)->b_committed_data != NULL);
-		ext4_set_bit_atomic(sb_bgl_lock(sbi, block_group), bit + i,
-				bh2jh(bitmap_bh)->b_committed_data);
-
-		/*
-		 * We clear the bit in the bitmap after setting the committed
-		 * data bit, because this is the reverse order to that which
-		 * the allocator uses.
-		 */
+	/*
+	 * make sure we don't allow a parallel init on other groups in the
+	 * same buddy cache
+	 */
+	down_write(&grp->alloc_sem);
+	for (i = 0, blocks_freed = 0; i < count; i++) {
 		BUFFER_TRACE(bitmap_bh, "clear bit");
 		if (!ext4_clear_bit_atomic(sb_bgl_lock(sbi, block_group),
 						bit + i, bitmap_bh->b_data)) {
-			jbd_unlock_bh_state(bitmap_bh);
 			ext4_error(sb, __func__,
 				   "bit already cleared for block %llu",
 				   (ext4_fsblk_t)(block + i));
-			jbd_lock_bh_state(bitmap_bh);
 			BUFFER_TRACE(bitmap_bh, "bit already cleared");
 		} else {
-			group_freed++;
+			blocks_freed++;
 		}
 	}
-	jbd_unlock_bh_state(bitmap_bh);
-
 	spin_lock(sb_bgl_lock(sbi, block_group));
-	le16_add_cpu(&desc->bg_free_blocks_count, group_freed);
+	blk_free_count = blocks_freed + ext4_free_blks_count(sb, desc);
+	ext4_free_blks_set(sb, desc, blk_free_count);
 	desc->bg_checksum = ext4_group_desc_csum(sbi, block_group, desc);
 	spin_unlock(sb_bgl_lock(sbi, block_group));
-	percpu_counter_add(&sbi->s_freeblocks_counter, count);
+	percpu_counter_add(&sbi->s_freeblocks_counter, blocks_freed);
 
 	if (sbi->s_log_groups_per_flex) {
 		ext4_group_t flex_group = ext4_flex_group(sbi, block_group);
 		spin_lock(sb_bgl_lock(sbi, flex_group));
-		sbi->s_flex_groups[flex_group].free_blocks += count;
+		sbi->s_flex_groups[flex_group].free_blocks += blocks_freed;
 		spin_unlock(sb_bgl_lock(sbi, flex_group));
 	}
+	/*
+	 * request to reload the buddy with the
+	 * new bitmap information
+	 */
+	set_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state));
+	ext4_mb_update_group_info(grp, blocks_freed);
+	up_write(&grp->alloc_sem);
 
 	/* We dirtied the bitmap block */
 	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
-	err = ext4_journal_dirty_metadata(handle, bitmap_bh);
+	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
 
 	/* And the group descriptor block */
 	BUFFER_TRACE(gd_bh, "dirtied group descriptor block");
-	ret = ext4_journal_dirty_metadata(handle, gd_bh);
-	if (!err) err = ret;
-	*pdquot_freed_blocks += group_freed;
-
-	if (overflow && !err) {
-		block += count;
-		count = overflow;
-		goto do_more;
-	}
+	ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh);
+	if (!err)
+		err = ret;
 	sb->s_dirt = 1;
+
 error_return:
 	brelse(bitmap_bh);
 	ext4_std_error(sb, err);
@@ -614,7 +565,7 @@ int ext4_has_free_blocks(struct ext4_sb_info *sbi, s64 nblocks)
 		if (dirty_blocks < 0) {
 			printk(KERN_CRIT "Dirty block accounting "
 					"went wrong %lld\n",
-					dirty_blocks);
+					(long long)dirty_blocks);
 		}
 	}
 	/* Check whether we have space after
@@ -666,101 +617,45 @@ int ext4_should_retry_alloc(struct super_block *sb, int *retries)
 	return jbd2_journal_force_commit_nested(EXT4_SB(sb)->s_journal);
 }
 
-#define EXT4_META_BLOCK 0x1
-
-static ext4_fsblk_t do_blk_alloc(handle_t *handle, struct inode *inode,
-				ext4_lblk_t iblock, ext4_fsblk_t goal,
-				unsigned long *count, int *errp, int flags)
-{
-	struct ext4_allocation_request ar;
-	ext4_fsblk_t ret;
-
-	memset(&ar, 0, sizeof(ar));
-	/* Fill with neighbour allocated blocks */
-
-	ar.inode = inode;
-	ar.goal = goal;
-	ar.len = *count;
-	ar.logical = iblock;
-
-	if (S_ISREG(inode->i_mode) && !(flags & EXT4_META_BLOCK))
-		/* enable in-core preallocation for data block allocation */
-		ar.flags = EXT4_MB_HINT_DATA;
-	else
-		/* disable in-core preallocation for non-regular files */
-		ar.flags = 0;
-
-	ret = ext4_mb_new_blocks(handle, &ar, errp);
-	*count = ar.len;
-	return ret;
-}
-
 /*
  * ext4_new_meta_blocks() -- allocate block for meta data (indexing) blocks
  *
  * @handle:             handle to this transaction
  * @inode:              file inode
  * @goal:               given target block(filesystem wide)
- * @count:		total number of blocks need
+ * @count:		pointer to total number of blocks needed
  * @errp:               error code
  *
- * Return 1st allocated block numberon success, *count stores total account
+ * Return 1st allocated block number on success, *count stores total account
  * error stores in errp pointer
  */
 ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
 		ext4_fsblk_t goal, unsigned long *count, int *errp)
 {
+	struct ext4_allocation_request ar;
 	ext4_fsblk_t ret;
-	ret = do_blk_alloc(handle, inode, 0, goal,
-				count, errp, EXT4_META_BLOCK);
+
+	memset(&ar, 0, sizeof(ar));
+	/* Fill with neighbour allocated blocks */
+	ar.inode = inode;
+	ar.goal = goal;
+	ar.len = count ? *count : 1;
+
+	ret = ext4_mb_new_blocks(handle, &ar, errp);
+	if (count)
+		*count = ar.len;
+
 	/*
 	 * Account for the allocated meta blocks
 	 */
 	if (!(*errp) && EXT4_I(inode)->i_delalloc_reserved_flag) {
 		spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
-		EXT4_I(inode)->i_allocated_meta_blocks += *count;
+		EXT4_I(inode)->i_allocated_meta_blocks += ar.len;
 		spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
 	}
 	return ret;
 }
 
-/*
- * ext4_new_meta_block() -- allocate block for meta data (indexing) blocks
- *
- * @handle:             handle to this transaction
- * @inode:              file inode
- * @goal:               given target block(filesystem wide)
- * @errp:               error code
- *
- * Return allocated block number on success
- */
-ext4_fsblk_t ext4_new_meta_block(handle_t *handle, struct inode *inode,
-		ext4_fsblk_t goal, int *errp)
-{
-	unsigned long count = 1;
-	return ext4_new_meta_blocks(handle, inode, goal, &count, errp);
-}
-
-/*
- * ext4_new_blocks() -- allocate data blocks
- *
- * @handle:             handle to this transaction
- * @inode:              file inode
- * @goal:               given target block(filesystem wide)
- * @count:		total number of blocks need
- * @errp:               error code
- *
- * Return 1st allocated block numberon success, *count stores total account
- * error stores in errp pointer
- */
-
-ext4_fsblk_t ext4_new_blocks(handle_t *handle, struct inode *inode,
-				ext4_lblk_t iblock, ext4_fsblk_t goal,
-				unsigned long *count, int *errp)
-{
-	return do_blk_alloc(handle, inode, iblock, goal, count, errp, 0);
-}
-
 /**
  * ext4_count_free_blocks() -- count filesystem free blocks
  * @sb:		superblock
@@ -776,7 +671,7 @@ ext4_fsblk_t ext4_count_free_blocks(struct super_block *sb)
 #ifdef EXT4FS_DEBUG
 	struct ext4_super_block *es;
 	ext4_fsblk_t bitmap_count;
-	unsigned long x;
+	unsigned int x;
 	struct buffer_head *bitmap_bh = NULL;
 
 	es = EXT4_SB(sb)->s_es;
@@ -796,7 +691,7 @@ ext4_fsblk_t ext4_count_free_blocks(struct super_block *sb)
 			continue;
 
 		x = ext4_count_free(bitmap_bh, sb->s_blocksize);
-		printk(KERN_DEBUG "group %lu: stored = %d, counted = %lu\n",
+		printk(KERN_DEBUG "group %lu: stored = %d, counted = %u\n",
 			i, le16_to_cpu(gdp->bg_free_blocks_count), x);
 		bitmap_count += x;
 	}
@@ -812,7 +707,7 @@ ext4_fsblk_t ext4_count_free_blocks(struct super_block *sb)
 		gdp = ext4_get_group_desc(sb, i, NULL);
 		if (!gdp)
 			continue;
-		desc_count += le16_to_cpu(gdp->bg_free_blocks_count);
+		desc_count += ext4_free_blks_count(sb, gdp);
 	}
 
 	return desc_count;
diff --git a/fs/ext4/bitmap.c b/fs/ext4/bitmap.c
index 0a7a6663c19..fa3af81ac56 100644
--- a/fs/ext4/bitmap.c
+++ b/fs/ext4/bitmap.c
@@ -15,10 +15,9 @@
 
 static const int nibblemap[] = {4, 3, 3, 2, 3, 2, 2, 1, 3, 2, 2, 1, 2, 1, 1, 0};
 
-unsigned long ext4_count_free(struct buffer_head *map, unsigned int numchars)
+unsigned int ext4_count_free(struct buffer_head *map, unsigned int numchars)
 {
-	unsigned int i;
-	unsigned long sum = 0;
+	unsigned int i, sum = 0;
 
 	if (!map)
 		return 0;
diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c
index fed5b610df5..2df2e40b01a 100644
--- a/fs/ext4/dir.c
+++ b/fs/ext4/dir.c
@@ -64,7 +64,7 @@ static unsigned char get_dtype(struct super_block *sb, int filetype)
 int ext4_check_dir_entry(const char *function, struct inode *dir,
 			 struct ext4_dir_entry_2 *de,
 			 struct buffer_head *bh,
-			 unsigned long offset)
+			 unsigned int offset)
 {
 	const char *error_msg = NULL;
 	const int rlen = ext4_rec_len_from_disk(de->rec_len);
@@ -84,9 +84,9 @@ int ext4_check_dir_entry(const char *function, struct inode *dir,
 	if (error_msg != NULL)
 		ext4_error(dir->i_sb, function,
 			"bad entry in directory #%lu: %s - "
-			"offset=%lu, inode=%lu, rec_len=%d, name_len=%d",
+			"offset=%u, inode=%u, rec_len=%d, name_len=%d",
 			dir->i_ino, error_msg, offset,
-			(unsigned long) le32_to_cpu(de->inode),
+			le32_to_cpu(de->inode),
 			rlen, de->name_len);
 	return error_msg == NULL ? 1 : 0;
 }
@@ -95,7 +95,7 @@ static int ext4_readdir(struct file *filp,
 			 void *dirent, filldir_t filldir)
 {
 	int error = 0;
-	unsigned long offset;
+	unsigned int offset;
 	int i, stored;
 	struct ext4_dir_entry_2 *de;
 	struct super_block *sb;
@@ -405,7 +405,7 @@ static int call_filldir(struct file *filp, void *dirent,
 	sb = inode->i_sb;
 
 	if (!fname) {
-		printk(KERN_ERR "ext4: call_filldir: called with "
+		printk(KERN_ERR "EXT4-fs: call_filldir: called with "
 		       "null fname?!?\n");
 		return 0;
 	}
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index b0537c82702..c668e4377d7 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -19,6 +19,7 @@
 #include <linux/types.h>
 #include <linux/blkdev.h>
 #include <linux/magic.h>
+#include <linux/jbd2.h>
 #include "ext4_i.h"
 
 /*
@@ -94,9 +95,9 @@ struct ext4_allocation_request {
 	/* phys. block for ^^^ */
 	ext4_fsblk_t pright;
 	/* how many blocks we want to allocate */
-	unsigned long len;
+	unsigned int len;
 	/* flags. see above EXT4_MB_HINT_* */
-	unsigned long flags;
+	unsigned int flags;
 };
 
 /*
@@ -156,12 +157,12 @@ struct ext4_group_desc
 	__le32	bg_block_bitmap_lo;	/* Blocks bitmap block */
 	__le32	bg_inode_bitmap_lo;	/* Inodes bitmap block */
 	__le32	bg_inode_table_lo;	/* Inodes table block */
-	__le16	bg_free_blocks_count;	/* Free blocks count */
-	__le16	bg_free_inodes_count;	/* Free inodes count */
-	__le16	bg_used_dirs_count;	/* Directories count */
+	__le16	bg_free_blocks_count_lo;/* Free blocks count */
+	__le16	bg_free_inodes_count_lo;/* Free inodes count */
+	__le16	bg_used_dirs_count_lo;	/* Directories count */
 	__le16	bg_flags;		/* EXT4_BG_flags (INODE_UNINIT, etc) */
 	__u32	bg_reserved[2];		/* Likely block/inode bitmap checksum */
-	__le16  bg_itable_unused;	/* Unused inodes count */
+	__le16  bg_itable_unused_lo;	/* Unused inodes count */
 	__le16  bg_checksum;		/* crc16(sb_uuid+group+desc) */
 	__le32	bg_block_bitmap_hi;	/* Blocks bitmap block MSB */
 	__le32	bg_inode_bitmap_hi;	/* Inodes bitmap block MSB */
@@ -169,7 +170,7 @@ struct ext4_group_desc
 	__le16	bg_free_blocks_count_hi;/* Free blocks count MSB */
 	__le16	bg_free_inodes_count_hi;/* Free inodes count MSB */
 	__le16	bg_used_dirs_count_hi;	/* Directories count MSB */
-	__le16	bg_itable_unused_hi;	/* Unused inodes count MSB */
+	__le16  bg_itable_unused_hi;    /* Unused inodes count MSB */
 	__u32	bg_reserved2[3];
 };
 
@@ -328,6 +329,7 @@ struct ext4_mount_options {
 	uid_t s_resuid;
 	gid_t s_resgid;
 	unsigned long s_commit_interval;
+	u32 s_min_batch_time, s_max_batch_time;
 #ifdef CONFIG_QUOTA
 	int s_jquota_fmt;
 	char *s_qf_names[MAXQUOTAS];
@@ -534,7 +536,6 @@ do {									       \
 #define EXT4_MOUNT_QUOTA		0x80000 /* Some quota option set */
 #define EXT4_MOUNT_USRQUOTA		0x100000 /* "old" user quota */
 #define EXT4_MOUNT_GRPQUOTA		0x200000 /* "old" group quota */
-#define EXT4_MOUNT_EXTENTS		0x400000 /* Extents support */
 #define EXT4_MOUNT_JOURNAL_CHECKSUM	0x800000 /* Journal checksums */
 #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT	0x1000000 /* Journal Async Commit */
 #define EXT4_MOUNT_I_VERSION            0x2000000 /* i_version support */
@@ -726,11 +727,11 @@ static inline int ext4_valid_inum(struct super_block *sb, unsigned long ino)
  */
 
 #define EXT4_HAS_COMPAT_FEATURE(sb,mask)			\
-	(EXT4_SB(sb)->s_es->s_feature_compat & cpu_to_le32(mask))
+	((EXT4_SB(sb)->s_es->s_feature_compat & cpu_to_le32(mask)) != 0)
 #define EXT4_HAS_RO_COMPAT_FEATURE(sb,mask)			\
-	(EXT4_SB(sb)->s_es->s_feature_ro_compat & cpu_to_le32(mask))
+	((EXT4_SB(sb)->s_es->s_feature_ro_compat & cpu_to_le32(mask)) != 0)
 #define EXT4_HAS_INCOMPAT_FEATURE(sb,mask)			\
-	(EXT4_SB(sb)->s_es->s_feature_incompat & cpu_to_le32(mask))
+	((EXT4_SB(sb)->s_es->s_feature_incompat & cpu_to_le32(mask)) != 0)
 #define EXT4_SET_COMPAT_FEATURE(sb,mask)			\
 	EXT4_SB(sb)->s_es->s_feature_compat |= cpu_to_le32(mask)
 #define EXT4_SET_RO_COMPAT_FEATURE(sb,mask)			\
@@ -806,6 +807,12 @@ static inline int ext4_valid_inum(struct super_block *sb, unsigned long ino)
 #define EXT4_DEFM_JMODE_WBACK	0x0060
 
 /*
+ * Default journal batch times
+ */
+#define EXT4_DEF_MIN_BATCH_TIME	0
+#define EXT4_DEF_MAX_BATCH_TIME	15000 /* 15ms */
+
+/*
  * Structure of a directory entry
  */
 #define EXT4_NAME_LEN 255
@@ -891,6 +898,9 @@ static inline __le16 ext4_rec_len_to_disk(unsigned len)
 #define DX_HASH_LEGACY		0
 #define DX_HASH_HALF_MD4	1
 #define DX_HASH_TEA		2
+#define DX_HASH_LEGACY_UNSIGNED	3
+#define DX_HASH_HALF_MD4_UNSIGNED	4
+#define DX_HASH_TEA_UNSIGNED		5
 
 #ifdef __KERNEL__
 
@@ -955,7 +965,7 @@ ext4_group_first_block_no(struct super_block *sb, ext4_group_t group_no)
 #define ERR_BAD_DX_DIR	-75000
 
 void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr,
-			unsigned long *blockgrpp, ext4_grpblk_t *offsetp);
+			ext4_group_t *blockgrpp, ext4_grpblk_t *offsetp);
 
 extern struct proc_dir_entry *ext4_proc_root;
 
@@ -987,6 +997,9 @@ do {									\
 # define ATTRIB_NORET	__attribute__((noreturn))
 # define NORET_AND	noreturn,
 
+/* bitmap.c */
+extern unsigned int ext4_count_free(struct buffer_head *, unsigned);
+
 /* balloc.c */
 extern unsigned int ext4_block_group(struct super_block *sb,
 			ext4_fsblk_t blocknr);
@@ -995,20 +1008,14 @@ extern ext4_grpblk_t ext4_block_group_offset(struct super_block *sb,
 extern int ext4_bg_has_super(struct super_block *sb, ext4_group_t group);
 extern unsigned long ext4_bg_num_gdb(struct super_block *sb,
 			ext4_group_t group);
-extern ext4_fsblk_t ext4_new_meta_block(handle_t *handle, struct inode *inode,
-			ext4_fsblk_t goal, int *errp);
 extern ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
 			ext4_fsblk_t goal, unsigned long *count, int *errp);
-extern ext4_fsblk_t ext4_new_blocks(handle_t *handle, struct inode *inode,
-					ext4_lblk_t iblock, ext4_fsblk_t goal,
-					unsigned long *count, int *errp);
 extern int ext4_claim_free_blocks(struct ext4_sb_info *sbi, s64 nblocks);
 extern int ext4_has_free_blocks(struct ext4_sb_info *sbi, s64 nblocks);
 extern void ext4_free_blocks(handle_t *handle, struct inode *inode,
 			ext4_fsblk_t block, unsigned long count, int metadata);
-extern void ext4_free_blocks_sb(handle_t *handle, struct super_block *sb,
-				ext4_fsblk_t block, unsigned long count,
-				unsigned long *pdquot_freed_blocks);
+extern void ext4_add_groupblocks(handle_t *handle, struct super_block *sb,
+				ext4_fsblk_t block, unsigned long count);
 extern ext4_fsblk_t ext4_count_free_blocks(struct super_block *);
 extern void ext4_check_blocks_bitmap(struct super_block *);
 extern struct ext4_group_desc * ext4_get_group_desc(struct super_block * sb,
@@ -1019,7 +1026,7 @@ extern int ext4_should_retry_alloc(struct super_block *sb, int *retries);
 /* dir.c */
 extern int ext4_check_dir_entry(const char *, struct inode *,
 				struct ext4_dir_entry_2 *,
-				struct buffer_head *, unsigned long);
+				struct buffer_head *, unsigned int);
 extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
 				    __u32 minor_hash,
 				    struct ext4_dir_entry_2 *dirent);
@@ -1039,7 +1046,6 @@ extern struct inode * ext4_orphan_get(struct super_block *, unsigned long);
 extern unsigned long ext4_count_free_inodes(struct super_block *);
 extern unsigned long ext4_count_dirs(struct super_block *);
 extern void ext4_check_inodes_bitmap(struct super_block *);
-extern unsigned long ext4_count_free(struct buffer_head *, unsigned);
 
 /* mballoc.c */
 extern long ext4_mb_stats;
@@ -1054,12 +1060,13 @@ extern int __init init_ext4_mballoc(void);
 extern void exit_ext4_mballoc(void);
 extern void ext4_mb_free_blocks(handle_t *, struct inode *,
 		unsigned long, unsigned long, int, unsigned long *);
-extern int ext4_mb_add_more_groupinfo(struct super_block *sb,
+extern int ext4_mb_add_groupinfo(struct super_block *sb,
 		ext4_group_t i, struct ext4_group_desc *desc);
 extern void ext4_mb_update_group_info(struct ext4_group_info *grp,
 		ext4_grpblk_t add);
-
-
+extern int ext4_mb_get_buddy_cache_lock(struct super_block *, ext4_group_t);
+extern void ext4_mb_put_buddy_cache_lock(struct super_block *,
+						ext4_group_t, int);
 /* inode.c */
 int ext4_forget(handle_t *handle, int is_metadata, struct inode *inode,
 		struct buffer_head *bh, ext4_fsblk_t blocknr);
@@ -1069,10 +1076,6 @@ struct buffer_head *ext4_bread(handle_t *, struct inode *,
 						ext4_lblk_t, int, int *);
 int ext4_get_block(struct inode *inode, sector_t iblock,
 				struct buffer_head *bh_result, int create);
-int ext4_get_blocks_handle(handle_t *handle, struct inode *inode,
-				ext4_lblk_t iblock, unsigned long maxblocks,
-				struct buffer_head *bh_result,
-				int create, int extend_disksize);
 
 extern struct inode *ext4_iget(struct super_block *, unsigned long);
 extern int  ext4_write_inode(struct inode *, int);
@@ -1123,6 +1126,9 @@ extern void ext4_abort(struct super_block *, const char *, const char *, ...)
 	__attribute__ ((format (printf, 3, 4)));
 extern void ext4_warning(struct super_block *, const char *, const char *, ...)
 	__attribute__ ((format (printf, 3, 4)));
+extern void ext4_grp_locked_error(struct super_block *, ext4_group_t,
+				const char *, const char *, ...)
+	__attribute__ ((format (printf, 4, 5)));
 extern void ext4_update_dynamic_rev(struct super_block *sb);
 extern int ext4_update_compat_feature(handle_t *handle, struct super_block *sb,
 					__u32 compat);
@@ -1136,12 +1142,28 @@ extern ext4_fsblk_t ext4_inode_bitmap(struct super_block *sb,
 				      struct ext4_group_desc *bg);
 extern ext4_fsblk_t ext4_inode_table(struct super_block *sb,
 				     struct ext4_group_desc *bg);
+extern __u32 ext4_free_blks_count(struct super_block *sb,
+				struct ext4_group_desc *bg);
+extern __u32 ext4_free_inodes_count(struct super_block *sb,
+				 struct ext4_group_desc *bg);
+extern __u32 ext4_used_dirs_count(struct super_block *sb,
+				struct ext4_group_desc *bg);
+extern __u32 ext4_itable_unused_count(struct super_block *sb,
+				   struct ext4_group_desc *bg);
 extern void ext4_block_bitmap_set(struct super_block *sb,
 				  struct ext4_group_desc *bg, ext4_fsblk_t blk);
 extern void ext4_inode_bitmap_set(struct super_block *sb,
 				  struct ext4_group_desc *bg, ext4_fsblk_t blk);
 extern void ext4_inode_table_set(struct super_block *sb,
 				 struct ext4_group_desc *bg, ext4_fsblk_t blk);
+extern void ext4_free_blks_set(struct super_block *sb,
+			       struct ext4_group_desc *bg, __u32 count);
+extern void ext4_free_inodes_set(struct super_block *sb,
+				struct ext4_group_desc *bg, __u32 count);
+extern void ext4_used_dirs_set(struct super_block *sb,
+				struct ext4_group_desc *bg, __u32 count);
+extern void ext4_itable_unused_set(struct super_block *sb,
+				   struct ext4_group_desc *bg, __u32 count);
 
 static inline ext4_fsblk_t ext4_blocks_count(struct ext4_super_block *es)
 {
@@ -1225,11 +1247,11 @@ do {								\
 } while (0)
 
 #ifdef CONFIG_SMP
-/* Each CPU can accumulate FBC_BATCH blocks in their local
+/* Each CPU can accumulate percpu_counter_batch blocks in their local
  * counters. So we need to make sure we have free blocks more
- * than FBC_BATCH  * nr_cpu_ids. Also add a window of 4 times.
+ * than percpu_counter_batch  * nr_cpu_ids. Also add a window of 4 times.
  */
-#define EXT4_FREEBLOCKS_WATERMARK (4 * (FBC_BATCH * nr_cpu_ids))
+#define EXT4_FREEBLOCKS_WATERMARK (4 * (percpu_counter_batch * nr_cpu_ids))
 #else
 #define EXT4_FREEBLOCKS_WATERMARK 0
 #endif
@@ -1246,6 +1268,50 @@ static inline void ext4_update_i_disksize(struct inode *inode, loff_t newsize)
 	return ;
 }
 
+struct ext4_group_info {
+	unsigned long   bb_state;
+	struct rb_root  bb_free_root;
+	unsigned short  bb_first_free;
+	unsigned short  bb_free;
+	unsigned short  bb_fragments;
+	struct          list_head bb_prealloc_list;
+#ifdef DOUBLE_CHECK
+	void            *bb_bitmap;
+#endif
+	struct rw_semaphore alloc_sem;
+	unsigned short  bb_counters[];
+};
+
+#define EXT4_GROUP_INFO_NEED_INIT_BIT	0
+#define EXT4_GROUP_INFO_LOCKED_BIT	1
+
+#define EXT4_MB_GRP_NEED_INIT(grp)	\
+	(test_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &((grp)->bb_state)))
+
+static inline void ext4_lock_group(struct super_block *sb, ext4_group_t group)
+{
+	struct ext4_group_info *grinfo = ext4_get_group_info(sb, group);
+
+	bit_spin_lock(EXT4_GROUP_INFO_LOCKED_BIT, &(grinfo->bb_state));
+}
+
+static inline void ext4_unlock_group(struct super_block *sb,
+					ext4_group_t group)
+{
+	struct ext4_group_info *grinfo = ext4_get_group_info(sb, group);
+
+	bit_spin_unlock(EXT4_GROUP_INFO_LOCKED_BIT, &(grinfo->bb_state));
+}
+
+static inline int ext4_is_group_locked(struct super_block *sb,
+					ext4_group_t group)
+{
+	struct ext4_group_info *grinfo = ext4_get_group_info(sb, group);
+
+	return bit_spin_is_locked(EXT4_GROUP_INFO_LOCKED_BIT,
+						&(grinfo->bb_state));
+}
+
 /*
  * Inodes and files operations
  */
@@ -1271,18 +1337,38 @@ extern int ext4_ext_writepage_trans_blocks(struct inode *, int);
 extern int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks,
 				       int chunk);
 extern int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
-			ext4_lblk_t iblock,
-			unsigned long max_blocks, struct buffer_head *bh_result,
-			int create, int extend_disksize);
+			       ext4_lblk_t iblock, unsigned int max_blocks,
+			       struct buffer_head *bh_result,
+			       int create, int extend_disksize);
 extern void ext4_ext_truncate(struct inode *);
 extern void ext4_ext_init(struct super_block *);
 extern void ext4_ext_release(struct super_block *);
 extern long ext4_fallocate(struct inode *inode, int mode, loff_t offset,
 			  loff_t len);
 extern int ext4_get_blocks_wrap(handle_t *handle, struct inode *inode,
-			sector_t block, unsigned long max_blocks,
+			sector_t block, unsigned int max_blocks,
 			struct buffer_head *bh, int create,
 			int extend_disksize, int flag);
+extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+			__u64 start, __u64 len);
+
+/*
+ * Add new method to test wether block and inode bitmaps are properly
+ * initialized. With uninit_bg reading the block from disk is not enough
+ * to mark the bitmap uptodate. We need to also zero-out the bitmap
+ */
+#define BH_BITMAP_UPTODATE BH_JBDPrivateStart
+
+static inline int bitmap_uptodate(struct buffer_head *bh)
+{
+	return (buffer_uptodate(bh) &&
+			test_bit(BH_BITMAP_UPTODATE, &(bh)->b_state));
+}
+static inline void set_bitmap_uptodate(struct buffer_head *bh)
+{
+	set_bit(BH_BITMAP_UPTODATE, &(bh)->b_state);
+}
+
 #endif	/* __KERNEL__ */
 
 #endif	/* _EXT4_H */
diff --git a/fs/ext4/ext4_extents.h b/fs/ext4/ext4_extents.h
index bec7ce59fc0..18cb67b2cbb 100644
--- a/fs/ext4/ext4_extents.h
+++ b/fs/ext4/ext4_extents.h
@@ -194,11 +194,6 @@ static inline unsigned short ext_depth(struct inode *inode)
 	return le16_to_cpu(ext_inode_hdr(inode)->eh_depth);
 }
 
-static inline void ext4_ext_tree_changed(struct inode *inode)
-{
-	EXT4_I(inode)->i_ext_generation++;
-}
-
 static inline void
 ext4_ext_invalidate_cache(struct inode *inode)
 {
diff --git a/fs/ext4/ext4_i.h b/fs/ext4/ext4_i.h
index 5c124c0ac6d..e69acc16f5c 100644
--- a/fs/ext4/ext4_i.h
+++ b/fs/ext4/ext4_i.h
@@ -31,7 +31,7 @@ typedef unsigned long long ext4_fsblk_t;
 typedef __u32 ext4_lblk_t;
 
 /* data type for block group number */
-typedef unsigned long ext4_group_t;
+typedef unsigned int ext4_group_t;
 
 #define rsv_start rsv_window._rsv_start
 #define rsv_end rsv_window._rsv_end
@@ -100,9 +100,6 @@ struct ext4_inode_info {
 	 */
 	loff_t	i_disksize;
 
-	/* on-disk additional length */
-	__u16 i_extra_isize;
-
 	/*
 	 * i_data_sem is for serialising ext4_truncate() against
 	 * ext4_getblock().  In the 2.4 ext2 design, great chunks of inode's
@@ -117,7 +114,6 @@ struct ext4_inode_info {
 	struct inode vfs_inode;
 	struct jbd2_inode jinode;
 
-	unsigned long i_ext_generation;
 	struct ext4_ext_cache i_cached_extent;
 	/*
 	 * File creation time. Its function is same as that of
@@ -130,10 +126,14 @@ struct ext4_inode_info {
 	spinlock_t i_prealloc_lock;
 
 	/* allocation reservation info for delalloc */
-	unsigned long i_reserved_data_blocks;
-	unsigned long i_reserved_meta_blocks;
-	unsigned long i_allocated_meta_blocks;
+	unsigned int i_reserved_data_blocks;
+	unsigned int i_reserved_meta_blocks;
+	unsigned int i_allocated_meta_blocks;
 	unsigned short i_delalloc_reserved_flag;
+
+	/* on-disk additional length */
+	__u16 i_extra_isize;
+
 	spinlock_t i_block_reservation_lock;
 };
 
diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c
index c75384b34f2..ad13a84644e 100644
--- a/fs/ext4/ext4_jbd2.c
+++ b/fs/ext4/ext4_jbd2.c
@@ -7,53 +7,96 @@
 int __ext4_journal_get_undo_access(const char *where, handle_t *handle,
 				struct buffer_head *bh)
 {
-	int err = jbd2_journal_get_undo_access(handle, bh);
-	if (err)
-		ext4_journal_abort_handle(where, __func__, bh, handle, err);
+	int err = 0;
+
+	if (ext4_handle_valid(handle)) {
+		err = jbd2_journal_get_undo_access(handle, bh);
+		if (err)
+			ext4_journal_abort_handle(where, __func__, bh,
+						  handle, err);
+	}
 	return err;
 }
 
 int __ext4_journal_get_write_access(const char *where, handle_t *handle,
 				struct buffer_head *bh)
 {
-	int err = jbd2_journal_get_write_access(handle, bh);
-	if (err)
-		ext4_journal_abort_handle(where, __func__, bh, handle, err);
+	int err = 0;
+
+	if (ext4_handle_valid(handle)) {
+		err = jbd2_journal_get_write_access(handle, bh);
+		if (err)
+			ext4_journal_abort_handle(where, __func__, bh,
+						  handle, err);
+	}
 	return err;
 }
 
 int __ext4_journal_forget(const char *where, handle_t *handle,
 				struct buffer_head *bh)
 {
-	int err = jbd2_journal_forget(handle, bh);
-	if (err)
-		ext4_journal_abort_handle(where, __func__, bh, handle, err);
+	int err = 0;
+
+	if (ext4_handle_valid(handle)) {
+		err = jbd2_journal_forget(handle, bh);
+		if (err)
+			ext4_journal_abort_handle(where, __func__, bh,
+						  handle, err);
+	}
 	return err;
 }
 
 int __ext4_journal_revoke(const char *where, handle_t *handle,
 				ext4_fsblk_t blocknr, struct buffer_head *bh)
 {
-	int err = jbd2_journal_revoke(handle, blocknr, bh);
-	if (err)
-		ext4_journal_abort_handle(where, __func__, bh, handle, err);
+	int err = 0;
+
+	if (ext4_handle_valid(handle)) {
+		err = jbd2_journal_revoke(handle, blocknr, bh);
+		if (err)
+			ext4_journal_abort_handle(where, __func__, bh,
+						  handle, err);
+	}
 	return err;
 }
 
 int __ext4_journal_get_create_access(const char *where,
 				handle_t *handle, struct buffer_head *bh)
 {
-	int err = jbd2_journal_get_create_access(handle, bh);
-	if (err)
-		ext4_journal_abort_handle(where, __func__, bh, handle, err);
+	int err = 0;
+
+	if (ext4_handle_valid(handle)) {
+		err = jbd2_journal_get_create_access(handle, bh);
+		if (err)
+			ext4_journal_abort_handle(where, __func__, bh,
+						  handle, err);
+	}
 	return err;
 }
 
-int __ext4_journal_dirty_metadata(const char *where,
-				handle_t *handle, struct buffer_head *bh)
+int __ext4_handle_dirty_metadata(const char *where, handle_t *handle,
+				 struct inode *inode, struct buffer_head *bh)
 {
-	int err = jbd2_journal_dirty_metadata(handle, bh);
-	if (err)
-		ext4_journal_abort_handle(where, __func__, bh, handle, err);
+	int err = 0;
+
+	if (ext4_handle_valid(handle)) {
+		err = jbd2_journal_dirty_metadata(handle, bh);
+		if (err)
+			ext4_journal_abort_handle(where, __func__, bh,
+						  handle, err);
+	} else {
+		mark_buffer_dirty(bh);
+		if (inode && inode_needs_sync(inode)) {
+			sync_dirty_buffer(bh);
+			if (buffer_req(bh) && !buffer_uptodate(bh)) {
+				ext4_error(inode->i_sb, __func__,
+					   "IO error syncing inode, "
+					   "inode=%lu, block=%llu",
+					   inode->i_ino,
+					   (unsigned long long) bh->b_blocknr);
+				err = -EIO;
+			}
+		}
+	}
 	return err;
 }
diff --git a/fs/ext4/ext4_jbd2.h b/fs/ext4/ext4_jbd2.h
index b455c685a98..be2f426f680 100644
--- a/fs/ext4/ext4_jbd2.h
+++ b/fs/ext4/ext4_jbd2.h
@@ -32,8 +32,8 @@
  * 5 levels of tree + root which are stored in the inode. */
 
 #define EXT4_SINGLEDATA_TRANS_BLOCKS(sb)				\
-	(EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)	\
-		|| test_opt(sb, EXTENTS) ? 27U : 8U)
+	(EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)   \
+	 ? 27U : 8U)
 
 /* Extended attribute operations touch at most two data buffers,
  * two bitmap buffers, and two group summaries, in addition to the inode
@@ -122,12 +122,6 @@ int ext4_mark_inode_dirty(handle_t *handle, struct inode *inode);
  * been done yet.
  */
 
-static inline void ext4_journal_release_buffer(handle_t *handle,
-						struct buffer_head *bh)
-{
-	jbd2_journal_release_buffer(handle, bh);
-}
-
 void ext4_journal_abort_handle(const char *caller, const char *err_fn,
 		struct buffer_head *bh, handle_t *handle, int err);
 
@@ -146,8 +140,8 @@ int __ext4_journal_revoke(const char *where, handle_t *handle,
 int __ext4_journal_get_create_access(const char *where,
 				handle_t *handle, struct buffer_head *bh);
 
-int __ext4_journal_dirty_metadata(const char *where,
-				handle_t *handle, struct buffer_head *bh);
+int __ext4_handle_dirty_metadata(const char *where, handle_t *handle,
+				 struct inode *inode, struct buffer_head *bh);
 
 #define ext4_journal_get_undo_access(handle, bh) \
 	__ext4_journal_get_undo_access(__func__, (handle), (bh))
@@ -157,14 +151,57 @@ int __ext4_journal_dirty_metadata(const char *where,
 	__ext4_journal_revoke(__func__, (handle), (blocknr), (bh))
 #define ext4_journal_get_create_access(handle, bh) \
 	__ext4_journal_get_create_access(__func__, (handle), (bh))
-#define ext4_journal_dirty_metadata(handle, bh) \
-	__ext4_journal_dirty_metadata(__func__, (handle), (bh))
 #define ext4_journal_forget(handle, bh) \
 	__ext4_journal_forget(__func__, (handle), (bh))
+#define ext4_handle_dirty_metadata(handle, inode, bh) \
+	__ext4_handle_dirty_metadata(__func__, (handle), (inode), (bh))
 
 handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks);
 int __ext4_journal_stop(const char *where, handle_t *handle);
 
+#define EXT4_NOJOURNAL_HANDLE	((handle_t *) 0x1)
+
+static inline int ext4_handle_valid(handle_t *handle)
+{
+	if (handle == EXT4_NOJOURNAL_HANDLE)
+		return 0;
+	return 1;
+}
+
+static inline void ext4_handle_sync(handle_t *handle)
+{
+	if (ext4_handle_valid(handle))
+		handle->h_sync = 1;
+}
+
+static inline void ext4_handle_release_buffer(handle_t *handle,
+						struct buffer_head *bh)
+{
+	if (ext4_handle_valid(handle))
+		jbd2_journal_release_buffer(handle, bh);
+}
+
+static inline int ext4_handle_is_aborted(handle_t *handle)
+{
+	if (ext4_handle_valid(handle))
+		return is_handle_aborted(handle);
+	return 0;
+}
+
+static inline int ext4_handle_has_enough_credits(handle_t *handle, int needed)
+{
+	if (ext4_handle_valid(handle) && handle->h_buffer_credits < needed)
+		return 0;
+	return 1;
+}
+
+static inline void ext4_journal_release_buffer(handle_t *handle,
+						struct buffer_head *bh)
+{
+	if (ext4_handle_valid(handle))
+		jbd2_journal_release_buffer(handle, bh);
+}
+
 static inline handle_t *ext4_journal_start(struct inode *inode, int nblocks)
 {
 	return ext4_journal_start_sb(inode->i_sb, nblocks);
@@ -180,27 +217,37 @@ static inline handle_t *ext4_journal_current_handle(void)
 
 static inline int ext4_journal_extend(handle_t *handle, int nblocks)
 {
-	return jbd2_journal_extend(handle, nblocks);
+	if (ext4_handle_valid(handle))
+		return jbd2_journal_extend(handle, nblocks);
+	return 0;
 }
 
 static inline int ext4_journal_restart(handle_t *handle, int nblocks)
 {
-	return jbd2_journal_restart(handle, nblocks);
+	if (ext4_handle_valid(handle))
+		return jbd2_journal_restart(handle, nblocks);
+	return 0;
 }
 
 static inline int ext4_journal_blocks_per_page(struct inode *inode)
 {
-	return jbd2_journal_blocks_per_page(inode);
+	if (EXT4_JOURNAL(inode) != NULL)
+		return jbd2_journal_blocks_per_page(inode);
+	return 0;
 }
 
 static inline int ext4_journal_force_commit(journal_t *journal)
 {
-	return jbd2_journal_force_commit(journal);
+	if (journal)
+		return jbd2_journal_force_commit(journal);
+	return 0;
 }
 
 static inline int ext4_jbd2_file_inode(handle_t *handle, struct inode *inode)
 {
-	return jbd2_journal_file_inode(handle, &EXT4_I(inode)->jinode);
+	if (ext4_handle_valid(handle))
+		return jbd2_journal_file_inode(handle, &EXT4_I(inode)->jinode);
+	return 0;
 }
 
 /* super.c */
@@ -208,6 +255,8 @@ int ext4_force_commit(struct super_block *sb);
 
 static inline int ext4_should_journal_data(struct inode *inode)
 {
+	if (EXT4_JOURNAL(inode) == NULL)
+		return 0;
 	if (!S_ISREG(inode->i_mode))
 		return 1;
 	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA)
@@ -219,6 +268,8 @@ static inline int ext4_should_journal_data(struct inode *inode)
 
 static inline int ext4_should_order_data(struct inode *inode)
 {
+	if (EXT4_JOURNAL(inode) == NULL)
+		return 0;
 	if (!S_ISREG(inode->i_mode))
 		return 0;
 	if (EXT4_I(inode)->i_flags & EXT4_JOURNAL_DATA_FL)
@@ -230,6 +281,8 @@ static inline int ext4_should_order_data(struct inode *inode)
 
 static inline int ext4_should_writeback_data(struct inode *inode)
 {
+	if (EXT4_JOURNAL(inode) == NULL)
+		return 0;
 	if (!S_ISREG(inode->i_mode))
 		return 0;
 	if (EXT4_I(inode)->i_flags & EXT4_JOURNAL_DATA_FL)
diff --git a/fs/ext4/ext4_sb.h b/fs/ext4/ext4_sb.h
index 445fde603df..039b6ea1a04 100644
--- a/fs/ext4/ext4_sb.h
+++ b/fs/ext4/ext4_sb.h
@@ -57,6 +57,7 @@ struct ext4_sb_info {
 	u32 s_next_generation;
 	u32 s_hash_seed[4];
 	int s_def_hash_version;
+	int s_hash_unsigned;	/* 3 if hash should be signed, 0 if not */
 	struct percpu_counter s_freeblocks_counter;
 	struct percpu_counter s_freeinodes_counter;
 	struct percpu_counter s_dirs_counter;
@@ -73,6 +74,8 @@ struct ext4_sb_info {
 	struct journal_s *s_journal;
 	struct list_head s_orphan;
 	unsigned long s_commit_interval;
+	u32 s_max_batch_time;
+	u32 s_min_batch_time;
 	struct block_device *journal_bdev;
 #ifdef CONFIG_JBD2_DEBUG
 	struct timer_list turn_ro_timer;	/* For turning read-only (crash simulation) */
@@ -101,7 +104,8 @@ struct ext4_sb_info {
 	spinlock_t s_reserve_lock;
 	spinlock_t s_md_lock;
 	tid_t s_last_transaction;
-	unsigned short *s_mb_offsets, *s_mb_maxs;
+	unsigned short *s_mb_offsets;
+	unsigned int *s_mb_maxs;
 
 	/* tunables */
 	unsigned long s_stripe;
@@ -146,4 +150,10 @@ struct ext4_sb_info {
 	struct flex_groups *s_flex_groups;
 };
 
+static inline spinlock_t *
+sb_bgl_lock(struct ext4_sb_info *sbi, unsigned int block_group)
+{
+	return bgl_lock_ptr(&sbi->s_blockgroup_lock, block_group);
+}
+
 #endif	/* _EXT4_SB */
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index ea2ce3c0ae6..54bf0623a9a 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -97,6 +97,8 @@ static int ext4_ext_journal_restart(handle_t *handle, int needed)
 {
 	int err;
 
+	if (!ext4_handle_valid(handle))
+		return 0;
 	if (handle->h_buffer_credits > needed)
 		return 0;
 	err = ext4_journal_extend(handle, needed);
@@ -134,7 +136,7 @@ static int ext4_ext_dirty(handle_t *handle, struct inode *inode,
 	int err;
 	if (path->p_bh) {
 		/* path points to block */
-		err = ext4_journal_dirty_metadata(handle, path->p_bh);
+		err = ext4_handle_dirty_metadata(handle, inode, path->p_bh);
 	} else {
 		/* path points to leaf/index in inode body */
 		err = ext4_mark_inode_dirty(handle, inode);
@@ -191,7 +193,7 @@ ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
 	ext4_fsblk_t goal, newblock;
 
 	goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
-	newblock = ext4_new_meta_block(handle, inode, goal, err);
+	newblock = ext4_new_meta_blocks(handle, inode, goal, NULL, err);
 	return newblock;
 }
 
@@ -780,7 +782,7 @@ static int ext4_ext_split(handle_t *handle, struct inode *inode,
 	set_buffer_uptodate(bh);
 	unlock_buffer(bh);
 
-	err = ext4_journal_dirty_metadata(handle, bh);
+	err = ext4_handle_dirty_metadata(handle, inode, bh);
 	if (err)
 		goto cleanup;
 	brelse(bh);
@@ -859,7 +861,7 @@ static int ext4_ext_split(handle_t *handle, struct inode *inode,
 		set_buffer_uptodate(bh);
 		unlock_buffer(bh);
 
-		err = ext4_journal_dirty_metadata(handle, bh);
+		err = ext4_handle_dirty_metadata(handle, inode, bh);
 		if (err)
 			goto cleanup;
 		brelse(bh);
@@ -955,7 +957,7 @@ static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
 	set_buffer_uptodate(bh);
 	unlock_buffer(bh);
 
-	err = ext4_journal_dirty_metadata(handle, bh);
+	err = ext4_handle_dirty_metadata(handle, inode, bh);
 	if (err)
 		goto out;
 
@@ -1160,15 +1162,13 @@ ext4_ext_search_right(struct inode *inode, struct ext4_ext_path *path,
 	while (--depth >= 0) {
 		ix = path[depth].p_idx;
 		if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
-			break;
+			goto got_index;
 	}
 
-	if (depth < 0) {
-		/* we've gone up to the root and
-		 * found no index to the right */
-		return 0;
-	}
+	/* we've gone up to the root and found no index to the right */
+	return 0;
 
+got_index:
 	/* we've found index to the right, let's
 	 * follow it and find the closest allocated
 	 * block to the right */
@@ -1201,7 +1201,6 @@ ext4_ext_search_right(struct inode *inode, struct ext4_ext_path *path,
 	*phys = ext_pblock(ex);
 	put_bh(bh);
 	return 0;
-
 }
 
 /*
@@ -1622,7 +1621,6 @@ cleanup:
 		ext4_ext_drop_refs(npath);
 		kfree(npath);
 	}
-	ext4_ext_tree_changed(inode);
 	ext4_ext_invalidate_cache(inode);
 	return err;
 }
@@ -2233,7 +2231,6 @@ static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
 		}
 	}
 out:
-	ext4_ext_tree_changed(inode);
 	ext4_ext_drop_refs(path);
 	kfree(path);
 	ext4_journal_stop(handle);
@@ -2250,7 +2247,7 @@ void ext4_ext_init(struct super_block *sb)
 	 * possible initialization would be here
 	 */
 
-	if (test_opt(sb, EXTENTS)) {
+	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
 		printk(KERN_INFO "EXT4-fs: file extents enabled");
 #ifdef AGGRESSIVE_TEST
 		printk(", aggressive tests");
@@ -2275,7 +2272,7 @@ void ext4_ext_init(struct super_block *sb)
  */
 void ext4_ext_release(struct super_block *sb)
 {
-	if (!test_opt(sb, EXTENTS))
+	if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
 		return;
 
 #ifdef EXTENTS_STATS
@@ -2380,7 +2377,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 						struct inode *inode,
 						struct ext4_ext_path *path,
 						ext4_lblk_t iblock,
-						unsigned long max_blocks)
+						unsigned int max_blocks)
 {
 	struct ext4_extent *ex, newex, orig_ex;
 	struct ext4_extent *ex1 = NULL;
@@ -2536,7 +2533,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle,
 		 */
 		newdepth = ext_depth(inode);
 		/*
-		 * update the extent length after successfull insert of the
+		 * update the extent length after successful insert of the
 		 * split extent
 		 */
 		orig_ex.ee_len = cpu_to_le16(ee_len -
@@ -2678,26 +2675,26 @@ fix_extent_len:
  */
 int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 			ext4_lblk_t iblock,
-			unsigned long max_blocks, struct buffer_head *bh_result,
+			unsigned int max_blocks, struct buffer_head *bh_result,
 			int create, int extend_disksize)
 {
 	struct ext4_ext_path *path = NULL;
 	struct ext4_extent_header *eh;
 	struct ext4_extent newex, *ex;
-	ext4_fsblk_t goal, newblock;
-	int err = 0, depth, ret;
-	unsigned long allocated = 0;
+	ext4_fsblk_t newblock;
+	int err = 0, depth, ret, cache_type;
+	unsigned int allocated = 0;
 	struct ext4_allocation_request ar;
 	loff_t disksize;
 
 	__clear_bit(BH_New, &bh_result->b_state);
-	ext_debug("blocks %u/%lu requested for inode %u\n",
+	ext_debug("blocks %u/%u requested for inode %u\n",
 			iblock, max_blocks, inode->i_ino);
 
 	/* check in cache */
-	goal = ext4_ext_in_cache(inode, iblock, &newex);
-	if (goal) {
-		if (goal == EXT4_EXT_CACHE_GAP) {
+	cache_type = ext4_ext_in_cache(inode, iblock, &newex);
+	if (cache_type) {
+		if (cache_type == EXT4_EXT_CACHE_GAP) {
 			if (!create) {
 				/*
 				 * block isn't allocated yet and
@@ -2706,7 +2703,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 				goto out2;
 			}
 			/* we should allocate requested block */
-		} else if (goal == EXT4_EXT_CACHE_EXTENT) {
+		} else if (cache_type == EXT4_EXT_CACHE_EXTENT) {
 			/* block is already allocated */
 			newblock = iblock
 				   - le32_to_cpu(newex.ee_block)
@@ -2854,7 +2851,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	if (!newblock)
 		goto out2;
 	ext_debug("allocate new block: goal %llu, found %llu/%lu\n",
-			goal, newblock, allocated);
+		  ar.goal, newblock, allocated);
 
 	/* try to insert new extent into found leaf and return */
 	ext4_ext_store_pblock(&newex, newblock);
@@ -2950,7 +2947,7 @@ void ext4_ext_truncate(struct inode *inode)
 	 * transaction synchronous.
 	 */
 	if (IS_SYNC(inode))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 out_stop:
 	up_write(&EXT4_I(inode)->i_data_sem);
@@ -3004,7 +3001,7 @@ long ext4_fallocate(struct inode *inode, int mode, loff_t offset, loff_t len)
 	handle_t *handle;
 	ext4_lblk_t block;
 	loff_t new_size;
-	unsigned long max_blocks;
+	unsigned int max_blocks;
 	int ret = 0;
 	int ret2 = 0;
 	int retries = 0;
@@ -3083,7 +3080,7 @@ retry:
 /*
  * Callback function called for each extent to gather FIEMAP information.
  */
-int ext4_ext_fiemap_cb(struct inode *inode, struct ext4_ext_path *path,
+static int ext4_ext_fiemap_cb(struct inode *inode, struct ext4_ext_path *path,
 		       struct ext4_ext_cache *newex, struct ext4_extent *ex,
 		       void *data)
 {
@@ -3152,7 +3149,8 @@ int ext4_ext_fiemap_cb(struct inode *inode, struct ext4_ext_path *path,
 /* fiemap flags we can handle specified here */
 #define EXT4_FIEMAP_FLAGS	(FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
 
-int ext4_xattr_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo)
+static int ext4_xattr_fiemap(struct inode *inode,
+				struct fiemap_extent_info *fieinfo)
 {
 	__u64 physical = 0;
 	__u64 length;
diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 6bd11fba71f..f731cb545a0 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -140,9 +140,6 @@ static int ext4_file_mmap(struct file *file, struct vm_area_struct *vma)
 	return 0;
 }
 
-extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
-		__u64 start, __u64 len);
-
 const struct file_operations ext4_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
diff --git a/fs/ext4/hash.c b/fs/ext4/hash.c
index 556ca8eba3d..ac8f168c8ab 100644
--- a/fs/ext4/hash.c
+++ b/fs/ext4/hash.c
@@ -35,23 +35,71 @@ static void TEA_transform(__u32 buf[4], __u32 const in[])
 
 
 /* The old legacy hash */
-static __u32 dx_hack_hash(const char *name, int len)
+static __u32 dx_hack_hash_unsigned(const char *name, int len)
 {
-	__u32 hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9;
+	__u32 hash, hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9;
+	const unsigned char *ucp = (const unsigned char *) name;
+
+	while (len--) {
+		hash = hash1 + (hash0 ^ (((int) *ucp++) * 7152373));
+
+		if (hash & 0x80000000)
+			hash -= 0x7fffffff;
+		hash1 = hash0;
+		hash0 = hash;
+	}
+	return hash0 << 1;
+}
+
+static __u32 dx_hack_hash_signed(const char *name, int len)
+{
+	__u32 hash, hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9;
+	const signed char *scp = (const signed char *) name;
+
 	while (len--) {
-		__u32 hash = hash1 + (hash0 ^ (*name++ * 7152373));
+		hash = hash1 + (hash0 ^ (((int) *scp++) * 7152373));
 
-		if (hash & 0x80000000) hash -= 0x7fffffff;
+		if (hash & 0x80000000)
+			hash -= 0x7fffffff;
 		hash1 = hash0;
 		hash0 = hash;
 	}
-	return (hash0 << 1);
+	return hash0 << 1;
+}
+
+static void str2hashbuf_signed(const char *msg, int len, __u32 *buf, int num)
+{
+	__u32	pad, val;
+	int	i;
+	const signed char *scp = (const signed char *) msg;
+
+	pad = (__u32)len | ((__u32)len << 8);
+	pad |= pad << 16;
+
+	val = pad;
+	if (len > num*4)
+		len = num * 4;
+	for (i = 0; i < len; i++) {
+		if ((i % 4) == 0)
+			val = pad;
+		val = ((int) scp[i]) + (val << 8);
+		if ((i % 4) == 3) {
+			*buf++ = val;
+			val = pad;
+			num--;
+		}
+	}
+	if (--num >= 0)
+		*buf++ = val;
+	while (--num >= 0)
+		*buf++ = pad;
 }
 
-static void str2hashbuf(const char *msg, int len, __u32 *buf, int num)
+static void str2hashbuf_unsigned(const char *msg, int len, __u32 *buf, int num)
 {
 	__u32	pad, val;
 	int	i;
+	const unsigned char *ucp = (const unsigned char *) msg;
 
 	pad = (__u32)len | ((__u32)len << 8);
 	pad |= pad << 16;
@@ -62,7 +110,7 @@ static void str2hashbuf(const char *msg, int len, __u32 *buf, int num)
 	for (i = 0; i < len; i++) {
 		if ((i % 4) == 0)
 			val = pad;
-		val = msg[i] + (val << 8);
+		val = ((int) ucp[i]) + (val << 8);
 		if ((i % 4) == 3) {
 			*buf++ = val;
 			val = pad;
@@ -95,6 +143,8 @@ int ext4fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo)
 	const char	*p;
 	int		i;
 	__u32		in[8], buf[4];
+	void		(*str2hashbuf)(const char *, int, __u32 *, int) =
+				str2hashbuf_signed;
 
 	/* Initialize the default seed for the hash checksum functions */
 	buf[0] = 0x67452301;
@@ -113,13 +163,18 @@ int ext4fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo)
 	}
 
 	switch (hinfo->hash_version) {
+	case DX_HASH_LEGACY_UNSIGNED:
+		hash = dx_hack_hash_unsigned(name, len);
+		break;
 	case DX_HASH_LEGACY:
-		hash = dx_hack_hash(name, len);
+		hash = dx_hack_hash_signed(name, len);
 		break;
+	case DX_HASH_HALF_MD4_UNSIGNED:
+		str2hashbuf = str2hashbuf_unsigned;
 	case DX_HASH_HALF_MD4:
 		p = name;
 		while (len > 0) {
-			str2hashbuf(p, len, in, 8);
+			(*str2hashbuf)(p, len, in, 8);
 			half_md4_transform(buf, in);
 			len -= 32;
 			p += 32;
@@ -127,10 +182,12 @@ int ext4fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo)
 		minor_hash = buf[2];
 		hash = buf[1];
 		break;
+	case DX_HASH_TEA_UNSIGNED:
+		str2hashbuf = str2hashbuf_unsigned;
 	case DX_HASH_TEA:
 		p = name;
 		while (len > 0) {
-			str2hashbuf(p, len, in, 4);
+			(*str2hashbuf)(p, len, in, 4);
 			TEA_transform(buf, in);
 			len -= 16;
 			p += 16;
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 08cac9fcace..4fb86a0061d 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -74,17 +74,17 @@ unsigned ext4_init_inode_bitmap(struct super_block *sb, struct buffer_head *bh,
 	/* If checksum is bad mark all blocks and inodes use to prevent
 	 * allocation, essentially implementing a per-group read-only flag. */
 	if (!ext4_group_desc_csum_verify(sbi, block_group, gdp)) {
-		ext4_error(sb, __func__, "Checksum bad for group %lu\n",
+		ext4_error(sb, __func__, "Checksum bad for group %u",
 			   block_group);
-		gdp->bg_free_blocks_count = 0;
-		gdp->bg_free_inodes_count = 0;
-		gdp->bg_itable_unused = 0;
+		ext4_free_blks_set(sb, gdp, 0);
+		ext4_free_inodes_set(sb, gdp, 0);
+		ext4_itable_unused_set(sb, gdp, 0);
 		memset(bh->b_data, 0xff, sb->s_blocksize);
 		return 0;
 	}
 
 	memset(bh->b_data, 0, (EXT4_INODES_PER_GROUP(sb) + 7) / 8);
-	mark_bitmap_end(EXT4_INODES_PER_GROUP(sb), EXT4_BLOCKS_PER_GROUP(sb),
+	mark_bitmap_end(EXT4_INODES_PER_GROUP(sb), sb->s_blocksize * 8,
 			bh->b_data);
 
 	return EXT4_INODES_PER_GROUP(sb);
@@ -111,29 +111,49 @@ ext4_read_inode_bitmap(struct super_block *sb, ext4_group_t block_group)
 	if (unlikely(!bh)) {
 		ext4_error(sb, __func__,
 			    "Cannot read inode bitmap - "
-			    "block_group = %lu, inode_bitmap = %llu",
+			    "block_group = %u, inode_bitmap = %llu",
 			    block_group, bitmap_blk);
 		return NULL;
 	}
-	if (buffer_uptodate(bh) &&
-	    !(desc->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)))
+	if (bitmap_uptodate(bh))
 		return bh;
 
 	lock_buffer(bh);
+	if (bitmap_uptodate(bh)) {
+		unlock_buffer(bh);
+		return bh;
+	}
 	spin_lock(sb_bgl_lock(EXT4_SB(sb), block_group));
 	if (desc->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)) {
 		ext4_init_inode_bitmap(sb, bh, block_group, desc);
+		set_bitmap_uptodate(bh);
 		set_buffer_uptodate(bh);
-		unlock_buffer(bh);
 		spin_unlock(sb_bgl_lock(EXT4_SB(sb), block_group));
+		unlock_buffer(bh);
 		return bh;
 	}
 	spin_unlock(sb_bgl_lock(EXT4_SB(sb), block_group));
+	if (buffer_uptodate(bh)) {
+		/*
+		 * if not uninit if bh is uptodate,
+		 * bitmap is also uptodate
+		 */
+		set_bitmap_uptodate(bh);
+		unlock_buffer(bh);
+		return bh;
+	}
+	/*
+	 * submit the buffer_head for read. We can
+	 * safely mark the bitmap as uptodate now.
+	 * We do it here so the bitmap uptodate bit
+	 * get set with buffer lock held.
+	 */
+	set_bitmap_uptodate(bh);
 	if (bh_submit_read(bh) < 0) {
 		put_bh(bh);
 		ext4_error(sb, __func__,
 			    "Cannot read inode bitmap - "
-			    "block_group = %lu, inode_bitmap = %llu",
+			    "block_group = %u, inode_bitmap = %llu",
 			    block_group, bitmap_blk);
 		return NULL;
 	}
@@ -168,7 +188,7 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 	struct ext4_group_desc *gdp;
 	struct ext4_super_block *es;
 	struct ext4_sb_info *sbi;
-	int fatal = 0, err;
+	int fatal = 0, err, count;
 	ext4_group_t flex_group;
 
 	if (atomic_read(&inode->i_count) > 1) {
@@ -190,6 +210,11 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 
 	ino = inode->i_ino;
 	ext4_debug("freeing inode %lu\n", ino);
+	trace_mark(ext4_free_inode,
+		   "dev %s ino %lu mode %d uid %lu gid %lu bocks %llu",
+		   sb->s_id, inode->i_ino, inode->i_mode,
+		   (unsigned long) inode->i_uid, (unsigned long) inode->i_gid,
+		   (unsigned long long) inode->i_blocks);
 
 	/*
 	 * Note: we must free any quota before locking the superblock,
@@ -236,9 +261,12 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 
 		if (gdp) {
 			spin_lock(sb_bgl_lock(sbi, block_group));
-			le16_add_cpu(&gdp->bg_free_inodes_count, 1);
-			if (is_directory)
-				le16_add_cpu(&gdp->bg_used_dirs_count, -1);
+			count = ext4_free_inodes_count(sb, gdp) + 1;
+			ext4_free_inodes_set(sb, gdp, count);
+			if (is_directory) {
+				count = ext4_used_dirs_count(sb, gdp) - 1;
+				ext4_used_dirs_set(sb, gdp, count);
+			}
 			gdp->bg_checksum = ext4_group_desc_csum(sbi,
 							block_group, gdp);
 			spin_unlock(sb_bgl_lock(sbi, block_group));
@@ -253,12 +281,12 @@ void ext4_free_inode(handle_t *handle, struct inode *inode)
 				spin_unlock(sb_bgl_lock(sbi, flex_group));
 			}
 		}
-		BUFFER_TRACE(bh2, "call ext4_journal_dirty_metadata");
-		err = ext4_journal_dirty_metadata(handle, bh2);
+		BUFFER_TRACE(bh2, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, NULL, bh2);
 		if (!fatal) fatal = err;
 	}
-	BUFFER_TRACE(bitmap_bh, "call ext4_journal_dirty_metadata");
-	err = ext4_journal_dirty_metadata(handle, bitmap_bh);
+	BUFFER_TRACE(bitmap_bh, "call ext4_handle_dirty_metadata");
+	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
 	if (!fatal)
 		fatal = err;
 	sb->s_dirt = 1;
@@ -291,13 +319,13 @@ static int find_group_dir(struct super_block *sb, struct inode *parent,
 
 	for (group = 0; group < ngroups; group++) {
 		desc = ext4_get_group_desc(sb, group, NULL);
-		if (!desc || !desc->bg_free_inodes_count)
+		if (!desc || !ext4_free_inodes_count(sb, desc))
 			continue;
-		if (le16_to_cpu(desc->bg_free_inodes_count) < avefreei)
+		if (ext4_free_inodes_count(sb, desc) < avefreei)
 			continue;
 		if (!best_desc ||
-		    (le16_to_cpu(desc->bg_free_blocks_count) >
-		     le16_to_cpu(best_desc->bg_free_blocks_count))) {
+		    (ext4_free_blks_count(sb, desc) >
+		     ext4_free_blks_count(sb, best_desc))) {
 			*best_group = group;
 			best_desc = desc;
 			ret = 0;
@@ -369,7 +397,7 @@ found_flexbg:
 	for (i = best_flex * flex_size; i < ngroups &&
 		     i < (best_flex + 1) * flex_size; i++) {
 		desc = ext4_get_group_desc(sb, i, &bh);
-		if (le16_to_cpu(desc->bg_free_inodes_count)) {
+		if (ext4_free_inodes_count(sb, desc)) {
 			*best_group = i;
 			goto out;
 		}
@@ -443,17 +471,17 @@ static int find_group_orlov(struct super_block *sb, struct inode *parent,
 		for (i = 0; i < ngroups; i++) {
 			grp = (parent_group + i) % ngroups;
 			desc = ext4_get_group_desc(sb, grp, NULL);
-			if (!desc || !desc->bg_free_inodes_count)
+			if (!desc || !ext4_free_inodes_count(sb, desc))
 				continue;
-			if (le16_to_cpu(desc->bg_used_dirs_count) >= best_ndir)
+			if (ext4_used_dirs_count(sb, desc) >= best_ndir)
 				continue;
-			if (le16_to_cpu(desc->bg_free_inodes_count) < avefreei)
+			if (ext4_free_inodes_count(sb, desc) < avefreei)
 				continue;
-			if (le16_to_cpu(desc->bg_free_blocks_count) < avefreeb)
+			if (ext4_free_blks_count(sb, desc) < avefreeb)
 				continue;
 			*group = grp;
 			ret = 0;
-			best_ndir = le16_to_cpu(desc->bg_used_dirs_count);
+			best_ndir = ext4_used_dirs_count(sb, desc);
 		}
 		if (ret == 0)
 			return ret;
@@ -479,13 +507,13 @@ static int find_group_orlov(struct super_block *sb, struct inode *parent,
 	for (i = 0; i < ngroups; i++) {
 		*group = (parent_group + i) % ngroups;
 		desc = ext4_get_group_desc(sb, *group, NULL);
-		if (!desc || !desc->bg_free_inodes_count)
+		if (!desc || !ext4_free_inodes_count(sb, desc))
 			continue;
-		if (le16_to_cpu(desc->bg_used_dirs_count) >= max_dirs)
+		if (ext4_used_dirs_count(sb, desc) >= max_dirs)
 			continue;
-		if (le16_to_cpu(desc->bg_free_inodes_count) < min_inodes)
+		if (ext4_free_inodes_count(sb, desc) < min_inodes)
 			continue;
-		if (le16_to_cpu(desc->bg_free_blocks_count) < min_blocks)
+		if (ext4_free_blks_count(sb, desc) < min_blocks)
 			continue;
 		return 0;
 	}
@@ -494,8 +522,8 @@ fallback:
 	for (i = 0; i < ngroups; i++) {
 		*group = (parent_group + i) % ngroups;
 		desc = ext4_get_group_desc(sb, *group, NULL);
-		if (desc && desc->bg_free_inodes_count &&
-			le16_to_cpu(desc->bg_free_inodes_count) >= avefreei)
+		if (desc && ext4_free_inodes_count(sb, desc) &&
+			ext4_free_inodes_count(sb, desc) >= avefreei)
 			return 0;
 	}
 
@@ -524,8 +552,8 @@ static int find_group_other(struct super_block *sb, struct inode *parent,
 	 */
 	*group = parent_group;
 	desc = ext4_get_group_desc(sb, *group, NULL);
-	if (desc && le16_to_cpu(desc->bg_free_inodes_count) &&
-			le16_to_cpu(desc->bg_free_blocks_count))
+	if (desc && ext4_free_inodes_count(sb, desc) &&
+			ext4_free_blks_count(sb, desc))
 		return 0;
 
 	/*
@@ -548,8 +576,8 @@ static int find_group_other(struct super_block *sb, struct inode *parent,
 		if (*group >= ngroups)
 			*group -= ngroups;
 		desc = ext4_get_group_desc(sb, *group, NULL);
-		if (desc && le16_to_cpu(desc->bg_free_inodes_count) &&
-				le16_to_cpu(desc->bg_free_blocks_count))
+		if (desc && ext4_free_inodes_count(sb, desc) &&
+				ext4_free_blks_count(sb, desc))
 			return 0;
 	}
 
@@ -562,7 +590,7 @@ static int find_group_other(struct super_block *sb, struct inode *parent,
 		if (++*group >= ngroups)
 			*group = 0;
 		desc = ext4_get_group_desc(sb, *group, NULL);
-		if (desc && le16_to_cpu(desc->bg_free_inodes_count))
+		if (desc && ext4_free_inodes_count(sb, desc))
 			return 0;
 	}
 
@@ -570,6 +598,79 @@ static int find_group_other(struct super_block *sb, struct inode *parent,
 }
 
 /*
+ * claim the inode from the inode bitmap. If the group
+ * is uninit we need to take the groups's sb_bgl_lock
+ * and clear the uninit flag. The inode bitmap update
+ * and group desc uninit flag clear should be done
+ * after holding sb_bgl_lock so that ext4_read_inode_bitmap
+ * doesn't race with the ext4_claim_inode
+ */
+static int ext4_claim_inode(struct super_block *sb,
+			struct buffer_head *inode_bitmap_bh,
+			unsigned long ino, ext4_group_t group, int mode)
+{
+	int free = 0, retval = 0, count;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	struct ext4_group_desc *gdp = ext4_get_group_desc(sb, group, NULL);
+
+	spin_lock(sb_bgl_lock(sbi, group));
+	if (ext4_set_bit(ino, inode_bitmap_bh->b_data)) {
+		/* not a free inode */
+		retval = 1;
+		goto err_ret;
+	}
+	ino++;
+	if ((group == 0 && ino < EXT4_FIRST_INO(sb)) ||
+			ino > EXT4_INODES_PER_GROUP(sb)) {
+		spin_unlock(sb_bgl_lock(sbi, group));
+		ext4_error(sb, __func__,
+			   "reserved inode or inode > inodes count - "
+			   "block_group = %u, inode=%lu", group,
+			   ino + group * EXT4_INODES_PER_GROUP(sb));
+		return 1;
+	}
+	/* If we didn't allocate from within the initialized part of the inode
+	 * table then we need to initialize up to this inode. */
+	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) {
+
+		if (gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)) {
+			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_INODE_UNINIT);
+			/* When marking the block group with
+			 * ~EXT4_BG_INODE_UNINIT we don't want to depend
+			 * on the value of bg_itable_unused even though
+			 * mke2fs could have initialized the same for us.
+			 * Instead we calculated the value below
+			 */
+
+			free = 0;
+		} else {
+			free = EXT4_INODES_PER_GROUP(sb) -
+				ext4_itable_unused_count(sb, gdp);
+		}
+
+		/*
+		 * Check the relative inode number against the last used
+		 * relative inode number in this group. if it is greater
+		 * we need to  update the bg_itable_unused count
+		 *
+		 */
+		if (ino > free)
+			ext4_itable_unused_set(sb, gdp,
+					(EXT4_INODES_PER_GROUP(sb) - ino));
+	}
+	count = ext4_free_inodes_count(sb, gdp) - 1;
+	ext4_free_inodes_set(sb, gdp, count);
+	if (S_ISDIR(mode)) {
+		count = ext4_used_dirs_count(sb, gdp) + 1;
+		ext4_used_dirs_set(sb, gdp, count);
+	}
+	gdp->bg_checksum = ext4_group_desc_csum(sbi, group, gdp);
+err_ret:
+	spin_unlock(sb_bgl_lock(sbi, group));
+	return retval;
+}
+
+/*
  * There are two policies for allocating an inode.  If the new inode is
  * a directory, then a forward search is made for a block group with both
  * free space and a low directory-to-inode ratio; if that fails, then of
@@ -582,8 +683,8 @@ static int find_group_other(struct super_block *sb, struct inode *parent,
 struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode)
 {
 	struct super_block *sb;
-	struct buffer_head *bitmap_bh = NULL;
-	struct buffer_head *bh2;
+	struct buffer_head *inode_bitmap_bh = NULL;
+	struct buffer_head *group_desc_bh;
 	ext4_group_t group = 0;
 	unsigned long ino = 0;
 	struct inode *inode;
@@ -602,6 +703,8 @@ struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode)
 		return ERR_PTR(-EPERM);
 
 	sb = dir->i_sb;
+	trace_mark(ext4_request_inode, "dev %s dir %lu mode %d", sb->s_id,
+		   dir->i_ino, mode);
 	inode = new_inode(sb);
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
@@ -631,40 +734,52 @@ got_group:
 	for (i = 0; i < sbi->s_groups_count; i++) {
 		err = -EIO;
 
-		gdp = ext4_get_group_desc(sb, group, &bh2);
+		gdp = ext4_get_group_desc(sb, group, &group_desc_bh);
 		if (!gdp)
 			goto fail;
 
-		brelse(bitmap_bh);
-		bitmap_bh = ext4_read_inode_bitmap(sb, group);
-		if (!bitmap_bh)
+		brelse(inode_bitmap_bh);
+		inode_bitmap_bh = ext4_read_inode_bitmap(sb, group);
+		if (!inode_bitmap_bh)
 			goto fail;
 
 		ino = 0;
 
 repeat_in_this_group:
 		ino = ext4_find_next_zero_bit((unsigned long *)
-				bitmap_bh->b_data, EXT4_INODES_PER_GROUP(sb), ino);
+					      inode_bitmap_bh->b_data,
+					      EXT4_INODES_PER_GROUP(sb), ino);
+
 		if (ino < EXT4_INODES_PER_GROUP(sb)) {
 
-			BUFFER_TRACE(bitmap_bh, "get_write_access");
-			err = ext4_journal_get_write_access(handle, bitmap_bh);
+			BUFFER_TRACE(inode_bitmap_bh, "get_write_access");
+			err = ext4_journal_get_write_access(handle,
+							    inode_bitmap_bh);
 			if (err)
 				goto fail;
 
-			if (!ext4_set_bit_atomic(sb_bgl_lock(sbi, group),
-						ino, bitmap_bh->b_data)) {
+			BUFFER_TRACE(group_desc_bh, "get_write_access");
+			err = ext4_journal_get_write_access(handle,
+								group_desc_bh);
+			if (err)
+				goto fail;
+			if (!ext4_claim_inode(sb, inode_bitmap_bh,
+						ino, group, mode)) {
 				/* we won it */
-				BUFFER_TRACE(bitmap_bh,
-					"call ext4_journal_dirty_metadata");
-				err = ext4_journal_dirty_metadata(handle,
-								bitmap_bh);
+				BUFFER_TRACE(inode_bitmap_bh,
+					"call ext4_handle_dirty_metadata");
+				err = ext4_handle_dirty_metadata(handle,
+								 inode,
+							inode_bitmap_bh);
 				if (err)
 					goto fail;
+				/* zero bit is inode number 1*/
+				ino++;
 				goto got;
 			}
 			/* we lost it */
-			jbd2_journal_release_buffer(handle, bitmap_bh);
+			ext4_handle_release_buffer(handle, inode_bitmap_bh);
+			ext4_handle_release_buffer(handle, group_desc_bh);
 
 			if (++ino < EXT4_INODES_PER_GROUP(sb))
 				goto repeat_in_this_group;
@@ -684,30 +799,16 @@ repeat_in_this_group:
 	goto out;
 
 got:
-	ino++;
-	if ((group == 0 && ino < EXT4_FIRST_INO(sb)) ||
-	    ino > EXT4_INODES_PER_GROUP(sb)) {
-		ext4_error(sb, __func__,
-			   "reserved inode or inode > inodes count - "
-			   "block_group = %lu, inode=%lu", group,
-			   ino + group * EXT4_INODES_PER_GROUP(sb));
-		err = -EIO;
-		goto fail;
-	}
-
-	BUFFER_TRACE(bh2, "get_write_access");
-	err = ext4_journal_get_write_access(handle, bh2);
-	if (err) goto fail;
-
 	/* We may have to initialize the block bitmap if it isn't already */
 	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM) &&
 	    gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
-		struct buffer_head *block_bh = ext4_read_block_bitmap(sb, group);
+		struct buffer_head *block_bitmap_bh;
 
-		BUFFER_TRACE(block_bh, "get block bitmap access");
-		err = ext4_journal_get_write_access(handle, block_bh);
+		block_bitmap_bh = ext4_read_block_bitmap(sb, group);
+		BUFFER_TRACE(block_bitmap_bh, "get block bitmap access");
+		err = ext4_journal_get_write_access(handle, block_bitmap_bh);
 		if (err) {
-			brelse(block_bh);
+			brelse(block_bitmap_bh);
 			goto fail;
 		}
 
@@ -715,9 +816,9 @@ got:
 		spin_lock(sb_bgl_lock(sbi, group));
 		/* recheck and clear flag under lock if we still need to */
 		if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
-			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT);
 			free = ext4_free_blocks_after_init(sb, group, gdp);
-			gdp->bg_free_blocks_count = cpu_to_le16(free);
+			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT);
+			ext4_free_blks_set(sb, gdp, free);
 			gdp->bg_checksum = ext4_group_desc_csum(sbi, group,
 								gdp);
 		}
@@ -725,55 +826,19 @@ got:
 
 		/* Don't need to dirty bitmap block if we didn't change it */
 		if (free) {
-			BUFFER_TRACE(block_bh, "dirty block bitmap");
-			err = ext4_journal_dirty_metadata(handle, block_bh);
+			BUFFER_TRACE(block_bitmap_bh, "dirty block bitmap");
+			err = ext4_handle_dirty_metadata(handle,
+							NULL, block_bitmap_bh);
 		}
 
-		brelse(block_bh);
+		brelse(block_bitmap_bh);
 		if (err)
 			goto fail;
 	}
-
-	spin_lock(sb_bgl_lock(sbi, group));
-	/* If we didn't allocate from within the initialized part of the inode
-	 * table then we need to initialize up to this inode. */
-	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) {
-		if (gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)) {
-			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_INODE_UNINIT);
-
-			/* When marking the block group with
-			 * ~EXT4_BG_INODE_UNINIT we don't want to depend
-			 * on the value of bg_itable_unused even though
-			 * mke2fs could have initialized the same for us.
-			 * Instead we calculated the value below
-			 */
-
-			free = 0;
-		} else {
-			free = EXT4_INODES_PER_GROUP(sb) -
-				le16_to_cpu(gdp->bg_itable_unused);
-		}
-
-		/*
-		 * Check the relative inode number against the last used
-		 * relative inode number in this group. if it is greater
-		 * we need to  update the bg_itable_unused count
-		 *
-		 */
-		if (ino > free)
-			gdp->bg_itable_unused =
-				cpu_to_le16(EXT4_INODES_PER_GROUP(sb) - ino);
-	}
-
-	le16_add_cpu(&gdp->bg_free_inodes_count, -1);
-	if (S_ISDIR(mode)) {
-		le16_add_cpu(&gdp->bg_used_dirs_count, 1);
-	}
-	gdp->bg_checksum = ext4_group_desc_csum(sbi, group, gdp);
-	spin_unlock(sb_bgl_lock(sbi, group));
-	BUFFER_TRACE(bh2, "call ext4_journal_dirty_metadata");
-	err = ext4_journal_dirty_metadata(handle, bh2);
-	if (err) goto fail;
+	BUFFER_TRACE(group_desc_bh, "call ext4_handle_dirty_metadata");
+	err = ext4_handle_dirty_metadata(handle, NULL, group_desc_bh);
+	if (err)
+		goto fail;
 
 	percpu_counter_dec(&sbi->s_freeinodes_counter);
 	if (S_ISDIR(mode))
@@ -825,8 +890,11 @@ got:
 
 	ext4_set_inode_flags(inode);
 	if (IS_DIRSYNC(inode))
-		handle->h_sync = 1;
-	insert_inode_hash(inode);
+		ext4_handle_sync(handle);
+	if (insert_inode_locked(inode) < 0) {
+		err = -EINVAL;
+		goto fail_drop;
+	}
 	spin_lock(&sbi->s_next_gen_lock);
 	inode->i_generation = sbi->s_next_generation++;
 	spin_unlock(&sbi->s_next_gen_lock);
@@ -849,7 +917,7 @@ got:
 	if (err)
 		goto fail_free_drop;
 
-	if (test_opt(sb, EXTENTS)) {
+	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
 		/* set extent flag only for directory, file and normal symlink*/
 		if (S_ISDIR(mode) || S_ISREG(mode) || S_ISLNK(mode)) {
 			EXT4_I(inode)->i_flags |= EXT4_EXTENTS_FL;
@@ -864,6 +932,8 @@ got:
 	}
 
 	ext4_debug("allocating inode %lu\n", inode->i_ino);
+	trace_mark(ext4_allocate_inode, "dev %s ino %lu dir %lu mode %d",
+		   sb->s_id, inode->i_ino, dir->i_ino, mode);
 	goto really_out;
 fail:
 	ext4_std_error(sb, err);
@@ -871,7 +941,7 @@ out:
 	iput(inode);
 	ret = ERR_PTR(err);
 really_out:
-	brelse(bitmap_bh);
+	brelse(inode_bitmap_bh);
 	return ret;
 
 fail_free_drop:
@@ -881,8 +951,9 @@ fail_drop:
 	DQUOT_DROP(inode);
 	inode->i_flags |= S_NOQUOTA;
 	inode->i_nlink = 0;
+	unlock_new_inode(inode);
 	iput(inode);
-	brelse(bitmap_bh);
+	brelse(inode_bitmap_bh);
 	return ERR_PTR(err);
 }
 
@@ -981,7 +1052,7 @@ unsigned long ext4_count_free_inodes(struct super_block *sb)
 		gdp = ext4_get_group_desc(sb, i, NULL);
 		if (!gdp)
 			continue;
-		desc_count += le16_to_cpu(gdp->bg_free_inodes_count);
+		desc_count += ext4_free_inodes_count(sb, gdp);
 		brelse(bitmap_bh);
 		bitmap_bh = ext4_read_inode_bitmap(sb, i);
 		if (!bitmap_bh)
@@ -989,7 +1060,7 @@ unsigned long ext4_count_free_inodes(struct super_block *sb)
 
 		x = ext4_count_free(bitmap_bh, EXT4_INODES_PER_GROUP(sb) / 8);
 		printk(KERN_DEBUG "group %lu: stored = %d, counted = %lu\n",
-			i, le16_to_cpu(gdp->bg_free_inodes_count), x);
+			i, ext4_free_inodes_count(sb, gdp), x);
 		bitmap_count += x;
 	}
 	brelse(bitmap_bh);
@@ -1003,7 +1074,7 @@ unsigned long ext4_count_free_inodes(struct super_block *sb)
 		gdp = ext4_get_group_desc(sb, i, NULL);
 		if (!gdp)
 			continue;
-		desc_count += le16_to_cpu(gdp->bg_free_inodes_count);
+		desc_count += ext4_free_inodes_count(sb, gdp);
 		cond_resched();
 	}
 	return desc_count;
@@ -1020,8 +1091,7 @@ unsigned long ext4_count_dirs(struct super_block * sb)
 		struct ext4_group_desc *gdp = ext4_get_group_desc(sb, i, NULL);
 		if (!gdp)
 			continue;
-		count += le16_to_cpu(gdp->bg_used_dirs_count);
+		count += ext4_used_dirs_count(sb, gdp);
 	}
 	return count;
 }
-
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index be21a5ae33c..a6444cee0c7 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -34,6 +34,7 @@
 #include <linux/writeback.h>
 #include <linux/pagevec.h>
 #include <linux/mpage.h>
+#include <linux/namei.h>
 #include <linux/uio.h>
 #include <linux/bio.h>
 #include "ext4_jbd2.h"
@@ -71,12 +72,17 @@ static int ext4_inode_is_fast_symlink(struct inode *inode)
  * "bh" may be NULL: a metadata block may have been freed from memory
  * but there may still be a record of it in the journal, and that record
  * still needs to be revoked.
+ *
+ * If the handle isn't valid we're not journaling so there's nothing to do.
  */
 int ext4_forget(handle_t *handle, int is_metadata, struct inode *inode,
 			struct buffer_head *bh, ext4_fsblk_t blocknr)
 {
 	int err;
 
+	if (!ext4_handle_valid(handle))
+		return 0;
+
 	might_sleep();
 
 	BUFFER_TRACE(bh, "enter");
@@ -169,7 +175,9 @@ static handle_t *start_transaction(struct inode *inode)
  */
 static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
 {
-	if (handle->h_buffer_credits > EXT4_RESERVE_TRANS_BLOCKS)
+	if (!ext4_handle_valid(handle))
+		return 0;
+	if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
 		return 0;
 	if (!ext4_journal_extend(handle, blocks_for_truncate(inode)))
 		return 0;
@@ -183,6 +191,7 @@ static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
  */
 static int ext4_journal_test_restart(handle_t *handle, struct inode *inode)
 {
+	BUG_ON(EXT4_JOURNAL(inode) == NULL);
 	jbd_debug(2, "restarting handle %p\n", handle);
 	return ext4_journal_restart(handle, blocks_for_truncate(inode));
 }
@@ -215,7 +224,7 @@ void ext4_delete_inode(struct inode *inode)
 	}
 
 	if (IS_SYNC(inode))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 	inode->i_size = 0;
 	err = ext4_mark_inode_dirty(handle, inode);
 	if (err) {
@@ -232,7 +241,7 @@ void ext4_delete_inode(struct inode *inode)
 	 * enough credits left in the handle to remove the inode from
 	 * the orphan list and set the dtime field.
 	 */
-	if (handle->h_buffer_credits < 3) {
+	if (!ext4_handle_has_enough_credits(handle, 3)) {
 		err = ext4_journal_extend(handle, 3);
 		if (err > 0)
 			err = ext4_journal_restart(handle, 3);
@@ -505,10 +514,10 @@ static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block,
  *	return the total number of blocks to be allocate, including the
  *	direct and indirect blocks.
  */
-static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned long blks,
+static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned int blks,
 		int blocks_to_boundary)
 {
-	unsigned long count = 0;
+	unsigned int count = 0;
 
 	/*
 	 * Simple case, [t,d]Indirect block(s) has not allocated yet
@@ -546,6 +555,7 @@ static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
 				int indirect_blks, int blks,
 				ext4_fsblk_t new_blocks[4], int *err)
 {
+	struct ext4_allocation_request ar;
 	int target, i;
 	unsigned long count = 0, blk_allocated = 0;
 	int index = 0;
@@ -594,10 +604,17 @@ static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
 	if (!target)
 		goto allocated;
 	/* Now allocate data blocks */
-	count = target;
-	/* allocating blocks for data blocks */
-	current_block = ext4_new_blocks(handle, inode, iblock,
-						goal, &count, err);
+	memset(&ar, 0, sizeof(ar));
+	ar.inode = inode;
+	ar.goal = goal;
+	ar.len = target;
+	ar.logical = iblock;
+	if (S_ISREG(inode->i_mode))
+		/* enable in-core preallocation only for regular files */
+		ar.flags = EXT4_MB_HINT_DATA;
+
+	current_block = ext4_mb_new_blocks(handle, &ar, err);
+
 	if (*err && (target == blks)) {
 		/*
 		 * if the allocation failed and we didn't allocate
@@ -613,7 +630,7 @@ static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
 		 */
 			new_blocks[index] = current_block;
 		}
-		blk_allocated += count;
+		blk_allocated += ar.len;
 	}
 allocated:
 	/* total number of blocks allocated for direct blocks */
@@ -708,8 +725,8 @@ static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
 		set_buffer_uptodate(bh);
 		unlock_buffer(bh);
 
-		BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
-		err = ext4_journal_dirty_metadata(handle, bh);
+		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, inode, bh);
 		if (err)
 			goto failed;
 	}
@@ -791,8 +808,8 @@ static int ext4_splice_branch(handle_t *handle, struct inode *inode,
 		 * generic_commit_write->__mark_inode_dirty->ext4_dirty_inode.
 		 */
 		jbd_debug(5, "splicing indirect only\n");
-		BUFFER_TRACE(where->bh, "call ext4_journal_dirty_metadata");
-		err = ext4_journal_dirty_metadata(handle, where->bh);
+		BUFFER_TRACE(where->bh, "call ext4_handle_dirty_metadata");
+		err = ext4_handle_dirty_metadata(handle, inode, where->bh);
 		if (err)
 			goto err_out;
 	} else {
@@ -839,10 +856,10 @@ err_out:
  * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
  * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
  */
-int ext4_get_blocks_handle(handle_t *handle, struct inode *inode,
-		ext4_lblk_t iblock, unsigned long maxblocks,
-		struct buffer_head *bh_result,
-		int create, int extend_disksize)
+static int ext4_get_blocks_handle(handle_t *handle, struct inode *inode,
+				  ext4_lblk_t iblock, unsigned int maxblocks,
+				  struct buffer_head *bh_result,
+				  int create, int extend_disksize)
 {
 	int err = -EIO;
 	ext4_lblk_t offsets[4];
@@ -1044,7 +1061,7 @@ static void ext4_da_update_reserve_space(struct inode *inode, int used)
  * It returns the error in case of allocation failure.
  */
 int ext4_get_blocks_wrap(handle_t *handle, struct inode *inode, sector_t block,
-			unsigned long max_blocks, struct buffer_head *bh,
+			unsigned int max_blocks, struct buffer_head *bh,
 			int create, int extend_disksize, int flag)
 {
 	int retval;
@@ -1220,8 +1237,8 @@ struct buffer_head *ext4_getblk(handle_t *handle, struct inode *inode,
 				set_buffer_uptodate(bh);
 			}
 			unlock_buffer(bh);
-			BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
-			err = ext4_journal_dirty_metadata(handle, bh);
+			BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+			err = ext4_handle_dirty_metadata(handle, inode, bh);
 			if (!fatal)
 				fatal = err;
 		} else {
@@ -1334,6 +1351,10 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping,
  	pgoff_t index;
 	unsigned from, to;
 
+	trace_mark(ext4_write_begin,
+		   "dev %s ino %lu pos %llu len %u flags %u",
+		   inode->i_sb->s_id, inode->i_ino,
+		   (unsigned long long) pos, len, flags);
  	index = pos >> PAGE_CACHE_SHIFT;
 	from = pos & (PAGE_CACHE_SIZE - 1);
 	to = from + len;
@@ -1345,7 +1366,7 @@ retry:
 		goto out;
 	}
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
 		ext4_journal_stop(handle);
 		ret = -ENOMEM;
@@ -1386,7 +1407,7 @@ static int write_end_fn(handle_t *handle, struct buffer_head *bh)
 	if (!buffer_mapped(bh) || buffer_freed(bh))
 		return 0;
 	set_buffer_uptodate(bh);
-	return ext4_journal_dirty_metadata(handle, bh);
+	return ext4_handle_dirty_metadata(handle, NULL, bh);
 }
 
 /*
@@ -1405,6 +1426,10 @@ static int ext4_ordered_write_end(struct file *file,
 	struct inode *inode = mapping->host;
 	int ret = 0, ret2;
 
+	trace_mark(ext4_ordered_write_end,
+		   "dev %s ino %lu pos %llu len %u copied %u",
+		   inode->i_sb->s_id, inode->i_ino,
+		   (unsigned long long) pos, len, copied);
 	ret = ext4_jbd2_file_inode(handle, inode);
 
 	if (ret == 0) {
@@ -1443,6 +1468,10 @@ static int ext4_writeback_write_end(struct file *file,
 	int ret = 0, ret2;
 	loff_t new_i_size;
 
+	trace_mark(ext4_writeback_write_end,
+		   "dev %s ino %lu pos %llu len %u copied %u",
+		   inode->i_sb->s_id, inode->i_ino,
+		   (unsigned long long) pos, len, copied);
 	new_i_size = pos + copied;
 	if (new_i_size > EXT4_I(inode)->i_disksize) {
 		ext4_update_i_disksize(inode, new_i_size);
@@ -1478,6 +1507,10 @@ static int ext4_journalled_write_end(struct file *file,
 	unsigned from, to;
 	loff_t new_i_size;
 
+	trace_mark(ext4_journalled_write_end,
+		   "dev %s ino %lu pos %llu len %u copied %u",
+		   inode->i_sb->s_id, inode->i_ino,
+		   (unsigned long long) pos, len, copied);
 	from = pos & (PAGE_CACHE_SIZE - 1);
 	to = from + len;
 
@@ -1624,7 +1657,7 @@ struct mpage_da_data {
 	get_block_t *get_block;
 	struct writeback_control *wbc;
 	int io_done;
-	long pages_written;
+	int pages_written;
 	int retval;
 };
 
@@ -1644,35 +1677,39 @@ struct mpage_da_data {
  */
 static int mpage_da_submit_io(struct mpage_da_data *mpd)
 {
-	struct address_space *mapping = mpd->inode->i_mapping;
-	int ret = 0, err, nr_pages, i;
-	unsigned long index, end;
-	struct pagevec pvec;
 	long pages_skipped;
+	struct pagevec pvec;
+	unsigned long index, end;
+	int ret = 0, err, nr_pages, i;
+	struct inode *inode = mpd->inode;
+	struct address_space *mapping = inode->i_mapping;
 
 	BUG_ON(mpd->next_page <= mpd->first_page);
-	pagevec_init(&pvec, 0);
+	/*
+	 * We need to start from the first_page to the next_page - 1
+	 * to make sure we also write the mapped dirty buffer_heads.
+	 * If we look at mpd->lbh.b_blocknr we would only be looking
+	 * at the currently mapped buffer_heads.
+	 */
 	index = mpd->first_page;
 	end = mpd->next_page - 1;
 
+	pagevec_init(&pvec, 0);
 	while (index <= end) {
-		/*
-		 * We can use PAGECACHE_TAG_DIRTY lookup here because
-		 * even though we have cleared the dirty flag on the page
-		 * We still keep the page in the radix tree with tag
-		 * PAGECACHE_TAG_DIRTY. See clear_page_dirty_for_io.
-		 * The PAGECACHE_TAG_DIRTY is cleared in set_page_writeback
-		 * which is called via the below writepage callback.
-		 */
-		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
-					PAGECACHE_TAG_DIRTY,
-					min(end - index,
-					(pgoff_t)PAGEVEC_SIZE-1) + 1);
+		nr_pages = pagevec_lookup(&pvec, mapping, index, PAGEVEC_SIZE);
 		if (nr_pages == 0)
 			break;
 		for (i = 0; i < nr_pages; i++) {
 			struct page *page = pvec.pages[i];
 
+			index = page->index;
+			if (index > end)
+				break;
+			index++;
+
+			BUG_ON(!PageLocked(page));
+			BUG_ON(PageWriteback(page));
+
 			pages_skipped = mpd->wbc->pages_skipped;
 			err = mapping->a_ops->writepage(page, mpd->wbc);
 			if (!err && (pages_skipped == mpd->wbc->pages_skipped))
@@ -1830,13 +1867,13 @@ static void ext4_print_free_blocks(struct inode *inode)
 			ext4_count_free_blocks(inode->i_sb));
 	printk(KERN_EMERG "Free/Dirty block details\n");
 	printk(KERN_EMERG "free_blocks=%lld\n",
-			percpu_counter_sum(&sbi->s_freeblocks_counter));
+			(long long)percpu_counter_sum(&sbi->s_freeblocks_counter));
 	printk(KERN_EMERG "dirty_blocks=%lld\n",
-			percpu_counter_sum(&sbi->s_dirtyblocks_counter));
+			(long long)percpu_counter_sum(&sbi->s_dirtyblocks_counter));
 	printk(KERN_EMERG "Block reservation details\n");
-	printk(KERN_EMERG "i_reserved_data_blocks=%lu\n",
+	printk(KERN_EMERG "i_reserved_data_blocks=%u\n",
 			EXT4_I(inode)->i_reserved_data_blocks);
-	printk(KERN_EMERG "i_reserved_meta_blocks=%lu\n",
+	printk(KERN_EMERG "i_reserved_meta_blocks=%u\n",
 			EXT4_I(inode)->i_reserved_meta_blocks);
 	return;
 }
@@ -2086,11 +2123,29 @@ static int __mpage_da_writepage(struct page *page,
 		bh = head;
 		do {
 			BUG_ON(buffer_locked(bh));
+			/*
+			 * We need to try to allocate
+			 * unmapped blocks in the same page.
+			 * Otherwise we won't make progress
+			 * with the page in ext4_da_writepage
+			 */
 			if (buffer_dirty(bh) &&
 				(!buffer_mapped(bh) || buffer_delay(bh))) {
 				mpage_add_bh_to_extent(mpd, logical, bh);
 				if (mpd->io_done)
 					return MPAGE_DA_EXTENT_TAIL;
+			} else if (buffer_dirty(bh) && (buffer_mapped(bh))) {
+				/*
+				 * mapped dirty buffer. We need to update
+				 * the b_state because we look at
+				 * b_state in mpage_da_map_blocks. We don't
+				 * update b_size because if we find an
+				 * unmapped buffer_head later we need to
+				 * use the b_state flag of that buffer_head.
+				 */
+				if (mpd->lbh.b_size == 0)
+					mpd->lbh.b_state =
+						bh->b_state & BH_FLAGS;
 			}
 			logical++;
 		} while ((bh = bh->b_this_page) != head);
@@ -2268,10 +2323,13 @@ static int ext4_da_writepage(struct page *page,
 {
 	int ret = 0;
 	loff_t size;
-	unsigned long len;
+	unsigned int len;
 	struct buffer_head *page_bufs;
 	struct inode *inode = page->mapping->host;
 
+	trace_mark(ext4_da_writepage,
+		   "dev %s ino %lu page_index %lu",
+		   inode->i_sb->s_id, inode->i_ino, page->index);
 	size = i_size_read(inode);
 	if (page->index == size >> PAGE_CACHE_SHIFT)
 		len = size & ~PAGE_CACHE_MASK;
@@ -2377,10 +2435,25 @@ static int ext4_da_writepages(struct address_space *mapping,
 	struct mpage_da_data mpd;
 	struct inode *inode = mapping->host;
 	int no_nrwrite_index_update;
-	long pages_written = 0, pages_skipped;
+	int pages_written = 0;
+	long pages_skipped;
 	int needed_blocks, ret = 0, nr_to_writebump = 0;
 	struct ext4_sb_info *sbi = EXT4_SB(mapping->host->i_sb);
 
+	trace_mark(ext4_da_writepages,
+		   "dev %s ino %lu nr_t_write %ld "
+		   "pages_skipped %ld range_start %llu "
+		   "range_end %llu nonblocking %d "
+		   "for_kupdate %d for_reclaim %d "
+		   "for_writepages %d range_cyclic %d",
+		   inode->i_sb->s_id, inode->i_ino,
+		   wbc->nr_to_write, wbc->pages_skipped,
+		   (unsigned long long) wbc->range_start,
+		   (unsigned long long) wbc->range_end,
+		   wbc->nonblocking, wbc->for_kupdate,
+		   wbc->for_reclaim, wbc->for_writepages,
+		   wbc->range_cyclic);
+
 	/*
 	 * No pages to write? This is mainly a kludge to avoid starting
 	 * a transaction for special inodes like journal inode on last iput()
@@ -2388,6 +2461,20 @@ static int ext4_da_writepages(struct address_space *mapping,
 	 */
 	if (!mapping->nrpages || !mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
 		return 0;
+
+	/*
+	 * If the filesystem has aborted, it is read-only, so return
+	 * right away instead of dumping stack traces later on that
+	 * will obscure the real source of the problem.  We test
+	 * EXT4_MOUNT_ABORT instead of sb->s_flag's MS_RDONLY because
+	 * the latter could be true if the filesystem is mounted
+	 * read-only, and in that case, ext4_da_writepages should
+	 * *never* be called, so if that ever happens, we would want
+	 * the stack trace.
+	 */
+	if (unlikely(sbi->s_mount_opt & EXT4_MOUNT_ABORT))
+		return -EROFS;
+
 	/*
 	 * Make sure nr_to_write is >= sbi->s_mb_stream_request
 	 * This make sure small files blocks are allocated in
@@ -2432,7 +2519,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 		handle = ext4_journal_start(inode, needed_blocks);
 		if (IS_ERR(handle)) {
 			ret = PTR_ERR(handle);
-			printk(KERN_EMERG "%s: jbd2_start: "
+			printk(KERN_CRIT "%s: jbd2_start: "
 			       "%ld pages, ino %lu; err %d\n", __func__,
 				wbc->nr_to_write, inode->i_ino, ret);
 			dump_stack();
@@ -2485,6 +2572,14 @@ out_writepages:
 	if (!no_nrwrite_index_update)
 		wbc->no_nrwrite_index_update = 0;
 	wbc->nr_to_write -= nr_to_writebump;
+	trace_mark(ext4_da_writepage_result,
+		   "dev %s ino %lu ret %d pages_written %d "
+		   "pages_skipped %ld congestion %d "
+		   "more_io %d no_nrwrite_index_update %d",
+		   inode->i_sb->s_id, inode->i_ino, ret,
+		   pages_written, wbc->pages_skipped,
+		   wbc->encountered_congestion, wbc->more_io,
+		   wbc->no_nrwrite_index_update);
 	return ret;
 }
 
@@ -2497,7 +2592,7 @@ static int ext4_nonda_switch(struct super_block *sb)
 	/*
 	 * switch to non delalloc mode if we are running low
 	 * on free block. The free block accounting via percpu
-	 * counters can get slightly wrong with FBC_BATCH getting
+	 * counters can get slightly wrong with percpu_counter_batch getting
 	 * accumulated on each CPU without updating global counters
 	 * Delalloc need an accurate free block accounting. So switch
 	 * to non delalloc when we are near to error range.
@@ -2536,6 +2631,11 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping,
 					len, flags, pagep, fsdata);
 	}
 	*fsdata = (void *)0;
+
+	trace_mark(ext4_da_write_begin,
+		   "dev %s ino %lu pos %llu len %u flags %u",
+		   inode->i_sb->s_id, inode->i_ino,
+		   (unsigned long long) pos, len, flags);
 retry:
 	/*
 	 * With delayed allocation, we don't log the i_disksize update
@@ -2549,7 +2649,7 @@ retry:
 		goto out;
 	}
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page) {
 		ext4_journal_stop(handle);
 		ret = -ENOMEM;
@@ -2625,6 +2725,10 @@ static int ext4_da_write_end(struct file *file,
 		}
 	}
 
+	trace_mark(ext4_da_write_end,
+		   "dev %s ino %lu pos %llu len %u copied %u",
+		   inode->i_sb->s_id, inode->i_ino,
+		   (unsigned long long) pos, len, copied);
 	start = pos & (PAGE_CACHE_SIZE - 1);
 	end = start + copied - 1;
 
@@ -2717,7 +2821,10 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
 		filemap_write_and_wait(mapping);
 	}
 
-	if (EXT4_I(inode)->i_state & EXT4_STATE_JDATA) {
+	BUG_ON(!EXT4_JOURNAL(inode) &&
+	       EXT4_I(inode)->i_state & EXT4_STATE_JDATA);
+
+	if (EXT4_JOURNAL(inode) && EXT4_I(inode)->i_state & EXT4_STATE_JDATA) {
 		/*
 		 * This is a REALLY heavyweight approach, but the use of
 		 * bmap on dirty files is expected to be extremely rare:
@@ -2835,6 +2942,9 @@ static int ext4_normal_writepage(struct page *page,
 	loff_t size = i_size_read(inode);
 	loff_t len;
 
+	trace_mark(ext4_normal_writepage,
+		   "dev %s ino %lu page_index %lu",
+		   inode->i_sb->s_id, inode->i_ino, page->index);
 	J_ASSERT(PageLocked(page));
 	if (page->index == size >> PAGE_CACHE_SHIFT)
 		len = size & ~PAGE_CACHE_MASK;
@@ -2920,6 +3030,9 @@ static int ext4_journalled_writepage(struct page *page,
 	loff_t size = i_size_read(inode);
 	loff_t len;
 
+	trace_mark(ext4_journalled_writepage,
+		   "dev %s ino %lu page_index %lu",
+		   inode->i_sb->s_id, inode->i_ino, page->index);
 	J_ASSERT(PageLocked(page));
 	if (page->index == size >> PAGE_CACHE_SHIFT)
 		len = size & ~PAGE_CACHE_MASK;
@@ -2988,7 +3101,10 @@ static void ext4_invalidatepage(struct page *page, unsigned long offset)
 	if (offset == 0)
 		ClearPageChecked(page);
 
-	jbd2_journal_invalidatepage(journal, page, offset);
+	if (journal)
+		jbd2_journal_invalidatepage(journal, page, offset);
+	else
+		block_invalidatepage(page, offset);
 }
 
 static int ext4_releasepage(struct page *page, gfp_t wait)
@@ -2998,7 +3114,10 @@ static int ext4_releasepage(struct page *page, gfp_t wait)
 	WARN_ON(PageChecked(page));
 	if (!page_has_buffers(page))
 		return 0;
-	return jbd2_journal_try_to_free_buffers(journal, page, wait);
+	if (journal)
+		return jbd2_journal_try_to_free_buffers(journal, page, wait);
+	else
+		return try_to_free_buffers(page);
 }
 
 /*
@@ -3270,7 +3389,7 @@ int ext4_block_truncate_page(handle_t *handle,
 
 	err = 0;
 	if (ext4_should_journal_data(inode)) {
-		err = ext4_journal_dirty_metadata(handle, bh);
+		err = ext4_handle_dirty_metadata(handle, inode, bh);
 	} else {
 		if (ext4_should_order_data(inode))
 			err = ext4_jbd2_file_inode(handle, inode);
@@ -3394,8 +3513,8 @@ static void ext4_clear_blocks(handle_t *handle, struct inode *inode,
 	__le32 *p;
 	if (try_to_extend_transaction(handle, inode)) {
 		if (bh) {
-			BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
-			ext4_journal_dirty_metadata(handle, bh);
+			BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+			ext4_handle_dirty_metadata(handle, inode, bh);
 		}
 		ext4_mark_inode_dirty(handle, inode);
 		ext4_journal_test_restart(handle, inode);
@@ -3495,7 +3614,7 @@ static void ext4_free_data(handle_t *handle, struct inode *inode,
 				  count, block_to_free_p, p);
 
 	if (this_bh) {
-		BUFFER_TRACE(this_bh, "call ext4_journal_dirty_metadata");
+		BUFFER_TRACE(this_bh, "call ext4_handle_dirty_metadata");
 
 		/*
 		 * The buffer head should have an attached journal head at this
@@ -3504,7 +3623,7 @@ static void ext4_free_data(handle_t *handle, struct inode *inode,
 		 * the block was cleared. Check for this instead of OOPSing.
 		 */
 		if (bh2jh(this_bh))
-			ext4_journal_dirty_metadata(handle, this_bh);
+			ext4_handle_dirty_metadata(handle, inode, this_bh);
 		else
 			ext4_error(inode->i_sb, __func__,
 				   "circular indirect block detected, "
@@ -3534,7 +3653,7 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 	ext4_fsblk_t nr;
 	__le32 *p;
 
-	if (is_handle_aborted(handle))
+	if (ext4_handle_is_aborted(handle))
 		return;
 
 	if (depth--) {
@@ -3604,7 +3723,7 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 			 * will merely complain about releasing a free block,
 			 * rather than leaking blocks.
 			 */
-			if (is_handle_aborted(handle))
+			if (ext4_handle_is_aborted(handle))
 				return;
 			if (try_to_extend_transaction(handle, inode)) {
 				ext4_mark_inode_dirty(handle, inode);
@@ -3623,9 +3742,10 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 								   parent_bh)){
 					*p = 0;
 					BUFFER_TRACE(parent_bh,
-					"call ext4_journal_dirty_metadata");
-					ext4_journal_dirty_metadata(handle,
-								    parent_bh);
+					"call ext4_handle_dirty_metadata");
+					ext4_handle_dirty_metadata(handle,
+								   inode,
+								   parent_bh);
 				}
 			}
 		}
@@ -3813,7 +3933,7 @@ do_indirects:
 	 * synchronous
 	 */
 	if (IS_SYNC(inode))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 out_stop:
 	/*
 	 * If this was a simple ftruncate(), and the file will remain alive
@@ -3843,7 +3963,7 @@ static int __ext4_get_inode_loc(struct inode *inode,
 	ext4_fsblk_t		block;
 	int			inodes_per_block, inode_offset;
 
-	iloc->bh = 0;
+	iloc->bh = NULL;
 	if (!ext4_valid_inum(sb, inode->i_ino))
 		return -EIO;
 
@@ -3950,7 +4070,7 @@ make_io:
 			num = EXT4_INODES_PER_GROUP(sb);
 			if (EXT4_HAS_RO_COMPAT_FEATURE(sb,
 				       EXT4_FEATURE_RO_COMPAT_GDT_CSUM))
-				num -= le16_to_cpu(gdp->bg_itable_unused);
+				num -= ext4_itable_unused_count(sb, gdp);
 			table += num / inodes_per_block;
 			if (end > table)
 				end = table;
@@ -4164,9 +4284,11 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
 		inode->i_op = &ext4_dir_inode_operations;
 		inode->i_fop = &ext4_dir_operations;
 	} else if (S_ISLNK(inode->i_mode)) {
-		if (ext4_inode_is_fast_symlink(inode))
+		if (ext4_inode_is_fast_symlink(inode)) {
 			inode->i_op = &ext4_fast_symlink_inode_operations;
-		else {
+			nd_terminate_link(ei->i_data, inode->i_size,
+				sizeof(ei->i_data) - 1);
+		} else {
 			inode->i_op = &ext4_symlink_inode_operations;
 			ext4_set_aops(inode);
 		}
@@ -4310,8 +4432,8 @@ static int ext4_do_update_inode(handle_t *handle,
 			EXT4_SET_RO_COMPAT_FEATURE(sb,
 					EXT4_FEATURE_RO_COMPAT_LARGE_FILE);
 			sb->s_dirt = 1;
-			handle->h_sync = 1;
-			err = ext4_journal_dirty_metadata(handle,
+			ext4_handle_sync(handle);
+			err = ext4_handle_dirty_metadata(handle, inode,
 					EXT4_SB(sb)->s_sbh);
 		}
 	}
@@ -4338,9 +4460,8 @@ static int ext4_do_update_inode(handle_t *handle,
 		raw_inode->i_extra_isize = cpu_to_le16(ei->i_extra_isize);
 	}
 
-
-	BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
-	rc = ext4_journal_dirty_metadata(handle, bh);
+	BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+	rc = ext4_handle_dirty_metadata(handle, inode, bh);
 	if (!err)
 		err = rc;
 	ei->i_state &= ~EXT4_STATE_NEW;
@@ -4403,6 +4524,25 @@ int ext4_write_inode(struct inode *inode, int wait)
 	return ext4_force_commit(inode->i_sb);
 }
 
+int __ext4_write_dirty_metadata(struct inode *inode, struct buffer_head *bh)
+{
+	int err = 0;
+
+	mark_buffer_dirty(bh);
+	if (inode && inode_needs_sync(inode)) {
+		sync_dirty_buffer(bh);
+		if (buffer_req(bh) && !buffer_uptodate(bh)) {
+			ext4_error(inode->i_sb, __func__,
+				   "IO error syncing inode, "
+				   "inode=%lu, block=%llu",
+				   inode->i_ino,
+				   (unsigned long long)bh->b_blocknr);
+			err = -EIO;
+		}
+	}
+	return err;
+}
+
 /*
  * ext4_setattr()
  *
@@ -4707,16 +4847,15 @@ int
 ext4_reserve_inode_write(handle_t *handle, struct inode *inode,
 			 struct ext4_iloc *iloc)
 {
-	int err = 0;
-	if (handle) {
-		err = ext4_get_inode_loc(inode, iloc);
-		if (!err) {
-			BUFFER_TRACE(iloc->bh, "get_write_access");
-			err = ext4_journal_get_write_access(handle, iloc->bh);
-			if (err) {
-				brelse(iloc->bh);
-				iloc->bh = NULL;
-			}
+	int err;
+
+	err = ext4_get_inode_loc(inode, iloc);
+	if (!err) {
+		BUFFER_TRACE(iloc->bh, "get_write_access");
+		err = ext4_journal_get_write_access(handle, iloc->bh);
+		if (err) {
+			brelse(iloc->bh);
+			iloc->bh = NULL;
 		}
 	}
 	ext4_std_error(inode->i_sb, err);
@@ -4788,7 +4927,8 @@ int ext4_mark_inode_dirty(handle_t *handle, struct inode *inode)
 
 	might_sleep();
 	err = ext4_reserve_inode_write(handle, inode, &iloc);
-	if (EXT4_I(inode)->i_extra_isize < sbi->s_want_extra_isize &&
+	if (ext4_handle_valid(handle) &&
+	    EXT4_I(inode)->i_extra_isize < sbi->s_want_extra_isize &&
 	    !(EXT4_I(inode)->i_state & EXT4_STATE_NO_EXPAND)) {
 		/*
 		 * We need extra buffer credits since we may write into EA block
@@ -4840,6 +4980,11 @@ void ext4_dirty_inode(struct inode *inode)
 	handle_t *current_handle = ext4_journal_current_handle();
 	handle_t *handle;
 
+	if (!ext4_handle_valid(current_handle)) {
+		ext4_mark_inode_dirty(current_handle, inode);
+		return;
+	}
+
 	handle = ext4_journal_start(inode, 2);
 	if (IS_ERR(handle))
 		goto out;
@@ -4877,8 +5022,9 @@ static int ext4_pin_inode(handle_t *handle, struct inode *inode)
 			BUFFER_TRACE(iloc.bh, "get_write_access");
 			err = jbd2_journal_get_write_access(handle, iloc.bh);
 			if (!err)
-				err = ext4_journal_dirty_metadata(handle,
-								  iloc.bh);
+				err = ext4_handle_dirty_metadata(handle,
+								 inode,
+								 iloc.bh);
 			brelse(iloc.bh);
 		}
 	}
@@ -4904,6 +5050,8 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
 	 */
 
 	journal = EXT4_JOURNAL(inode);
+	if (!journal)
+		return 0;
 	if (is_journal_aborted(journal))
 		return -EROFS;
 
@@ -4933,7 +5081,7 @@ int ext4_change_inode_journal_flag(struct inode *inode, int val)
 		return PTR_ERR(handle);
 
 	err = ext4_mark_inode_dirty(handle, inode);
-	handle->h_sync = 1;
+	ext4_handle_sync(handle);
 	ext4_journal_stop(handle);
 	ext4_std_error(inode->i_sb, err);
 
diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index dc99b4776d5..42dc83fb247 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -99,7 +99,7 @@ long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			goto flags_out;
 		}
 		if (IS_SYNC(inode))
-			handle->h_sync = 1;
+			ext4_handle_sync(handle);
 		err = ext4_reserve_inode_write(handle, inode, &iloc);
 		if (err)
 			goto flags_err;
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 444ad998f72..918aec0c8a1 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -100,7 +100,7 @@
  * inode as:
  *
  *  {                        page                        }
- *  [ group 0 buddy][ group 0 bitmap] [group 1][ group 1]...
+ *  [ group 0 bitmap][ group 0 buddy] [group 1][ group 1]...
  *
  *
  * one block each for bitmap and buddy information.  So for each group we
@@ -330,6 +330,18 @@
  *        object
  *
  */
+static struct kmem_cache *ext4_pspace_cachep;
+static struct kmem_cache *ext4_ac_cachep;
+static struct kmem_cache *ext4_free_ext_cachep;
+static void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap,
+					ext4_group_t group);
+static void ext4_mb_generate_from_freelist(struct super_block *sb, void *bitmap,
+						ext4_group_t group);
+static int ext4_mb_init_per_dev_proc(struct super_block *sb);
+static int ext4_mb_destroy_per_dev_proc(struct super_block *sb);
+static void release_blocks_on_commit(journal_t *journal, transaction_t *txn);
+
+
 
 static inline void *mb_correct_addr_and_bit(int *bit, void *addr)
 {
@@ -445,9 +457,9 @@ static void mb_free_blocks_double(struct inode *inode, struct ext4_buddy *e4b,
 			blocknr += first + i;
 			blocknr +=
 			    le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
-
-			ext4_error(sb, __func__, "double-free of inode"
-				   " %lu's block %llu(bit %u in group %lu)\n",
+			ext4_grp_locked_error(sb, e4b->bd_group,
+				   __func__, "double-free of inode"
+				   " %lu's block %llu(bit %u in group %u)",
 				   inode ? inode->i_ino : 0, blocknr,
 				   first + i, e4b->bd_group);
 		}
@@ -477,7 +489,7 @@ static void mb_cmp_bitmaps(struct ext4_buddy *e4b, void *bitmap)
 		b2 = (unsigned char *) bitmap;
 		for (i = 0; i < e4b->bd_sb->s_blocksize; i++) {
 			if (b1[i] != b2[i]) {
-				printk(KERN_ERR "corruption in group %lu "
+				printk(KERN_ERR "corruption in group %u "
 				       "at byte %u(%u): %x in copy != %x "
 				       "on disk/prealloc\n",
 				       e4b->bd_group, i, i * 8, b1[i], b2[i]);
@@ -690,8 +702,8 @@ static void ext4_mb_generate_buddy(struct super_block *sb,
 	grp->bb_fragments = fragments;
 
 	if (free != grp->bb_free) {
-		ext4_error(sb, __func__,
-			"EXT4-fs: group %lu: %u blocks in bitmap, %u in gd\n",
+		ext4_grp_locked_error(sb, group,  __func__,
+			"EXT4-fs: group %u: %u blocks in bitmap, %u in gd",
 			group, free, grp->bb_free);
 		/*
 		 * If we intent to continue, we consider group descritor
@@ -716,7 +728,7 @@ static void ext4_mb_generate_buddy(struct super_block *sb,
  * stored in the inode as
  *
  * {                        page                        }
- * [ group 0 buddy][ group 0 bitmap] [group 1][ group 1]...
+ * [ group 0 bitmap][ group 0 buddy] [group 1][ group 1]...
  *
  *
  * one block each for bitmap and buddy information.
@@ -782,25 +794,45 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 		if (bh[i] == NULL)
 			goto out;
 
-		if (buffer_uptodate(bh[i]) &&
-		    !(desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)))
+		if (bitmap_uptodate(bh[i]))
 			continue;
 
 		lock_buffer(bh[i]);
+		if (bitmap_uptodate(bh[i])) {
+			unlock_buffer(bh[i]);
+			continue;
+		}
 		spin_lock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
 		if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
 			ext4_init_block_bitmap(sb, bh[i],
 						first_group + i, desc);
+			set_bitmap_uptodate(bh[i]);
 			set_buffer_uptodate(bh[i]);
-			unlock_buffer(bh[i]);
 			spin_unlock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
+			unlock_buffer(bh[i]);
 			continue;
 		}
 		spin_unlock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
+		if (buffer_uptodate(bh[i])) {
+			/*
+			 * if not uninit if bh is uptodate,
+			 * bitmap is also uptodate
+			 */
+			set_bitmap_uptodate(bh[i]);
+			unlock_buffer(bh[i]);
+			continue;
+		}
 		get_bh(bh[i]);
+		/*
+		 * submit the buffer_head for read. We can
+		 * safely mark the bitmap as uptodate now.
+		 * We do it here so the bitmap uptodate bit
+		 * get set with buffer lock held.
+		 */
+		set_bitmap_uptodate(bh[i]);
 		bh[i]->b_end_io = end_buffer_read_sync;
 		submit_bh(READ, bh[i]);
-		mb_debug("read bitmap for group %lu\n", first_group + i);
+		mb_debug("read bitmap for group %u\n", first_group + i);
 	}
 
 	/* wait for I/O completion */
@@ -814,6 +846,8 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 
 	err = 0;
 	first_block = page->index * blocks_per_page;
+	/* init the page  */
+	memset(page_address(page), 0xff, PAGE_CACHE_SIZE);
 	for (i = 0; i < blocks_per_page; i++) {
 		int group;
 		struct ext4_group_info *grinfo;
@@ -840,7 +874,6 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 			BUG_ON(incore == NULL);
 			mb_debug("put buddy for group %u in page %lu/%x\n",
 				group, page->index, i * blocksize);
-			memset(data, 0xff, blocksize);
 			grinfo = ext4_get_group_info(sb, group);
 			grinfo->bb_fragments = 0;
 			memset(grinfo->bb_counters, 0,
@@ -848,7 +881,9 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 			/*
 			 * incore got set to the group block bitmap below
 			 */
+			ext4_lock_group(sb, group);
 			ext4_mb_generate_buddy(sb, data, incore, group);
+			ext4_unlock_group(sb, group);
 			incore = NULL;
 		} else {
 			/* this is block of bitmap */
@@ -862,6 +897,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 
 			/* mark all preallocated blks used in in-core bitmap */
 			ext4_mb_generate_from_pa(sb, data, group);
+			ext4_mb_generate_from_freelist(sb, data, group);
 			ext4_unlock_group(sb, group);
 
 			/* set incore so that the buddy information can be
@@ -886,18 +922,20 @@ static noinline_for_stack int
 ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 					struct ext4_buddy *e4b)
 {
-	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	struct inode *inode = sbi->s_buddy_cache;
 	int blocks_per_page;
 	int block;
 	int pnum;
 	int poff;
 	struct page *page;
 	int ret;
+	struct ext4_group_info *grp;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	struct inode *inode = sbi->s_buddy_cache;
 
-	mb_debug("load group %lu\n", group);
+	mb_debug("load group %u\n", group);
 
 	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
+	grp = ext4_get_group_info(sb, group);
 
 	e4b->bd_blkbits = sb->s_blocksize_bits;
 	e4b->bd_info = ext4_get_group_info(sb, group);
@@ -905,6 +943,15 @@ ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 	e4b->bd_group = group;
 	e4b->bd_buddy_page = NULL;
 	e4b->bd_bitmap_page = NULL;
+	e4b->alloc_semp = &grp->alloc_sem;
+
+	/* Take the read lock on the group alloc
+	 * sem. This would make sure a parallel
+	 * ext4_mb_init_group happening on other
+	 * groups mapped by the page is blocked
+	 * till we are done with allocation
+	 */
+	down_read(e4b->alloc_semp);
 
 	/*
 	 * the buddy cache inode stores the block bitmap
@@ -920,6 +967,14 @@ ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 	page = find_get_page(inode->i_mapping, pnum);
 	if (page == NULL || !PageUptodate(page)) {
 		if (page)
+			/*
+			 * drop the page reference and try
+			 * to get the page with lock. If we
+			 * are not uptodate that implies
+			 * somebody just created the page but
+			 * is yet to initialize the same. So
+			 * wait for it to initialize.
+			 */
 			page_cache_release(page);
 		page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS);
 		if (page) {
@@ -985,6 +1040,9 @@ err:
 		page_cache_release(e4b->bd_buddy_page);
 	e4b->bd_buddy = NULL;
 	e4b->bd_bitmap = NULL;
+
+	/* Done with the buddy cache */
+	up_read(e4b->alloc_semp);
 	return ret;
 }
 
@@ -994,6 +1052,9 @@ static void ext4_mb_release_desc(struct ext4_buddy *e4b)
 		page_cache_release(e4b->bd_bitmap_page);
 	if (e4b->bd_buddy_page)
 		page_cache_release(e4b->bd_buddy_page);
+	/* Done with the buddy cache */
+	if (e4b->alloc_semp)
+		up_read(e4b->alloc_semp);
 }
 
 
@@ -1031,7 +1092,10 @@ static void mb_clear_bits(spinlock_t *lock, void *bm, int cur, int len)
 			cur += 32;
 			continue;
 		}
-		mb_clear_bit_atomic(lock, cur, bm);
+		if (lock)
+			mb_clear_bit_atomic(lock, cur, bm);
+		else
+			mb_clear_bit(cur, bm);
 		cur++;
 	}
 }
@@ -1049,7 +1113,10 @@ static void mb_set_bits(spinlock_t *lock, void *bm, int cur, int len)
 			cur += 32;
 			continue;
 		}
-		mb_set_bit_atomic(lock, cur, bm);
+		if (lock)
+			mb_set_bit_atomic(lock, cur, bm);
+		else
+			mb_set_bit(cur, bm);
 		cur++;
 	}
 }
@@ -1094,12 +1161,11 @@ static void mb_free_blocks(struct inode *inode, struct ext4_buddy *e4b,
 			blocknr += block;
 			blocknr +=
 			    le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
-			ext4_unlock_group(sb, e4b->bd_group);
-			ext4_error(sb, __func__, "double-free of inode"
-				   " %lu's block %llu(bit %u in group %lu)\n",
+			ext4_grp_locked_error(sb, e4b->bd_group,
+				   __func__, "double-free of inode"
+				   " %lu's block %llu(bit %u in group %u)",
 				   inode ? inode->i_ino : 0, blocknr, block,
 				   e4b->bd_group);
-			ext4_lock_group(sb, e4b->bd_group);
 		}
 		mb_clear_bit(block, EXT4_MB_BITMAP(e4b));
 		e4b->bd_info->bb_counters[order]++;
@@ -1296,13 +1362,20 @@ static void ext4_mb_use_best_found(struct ext4_allocation_context *ac,
 	ac->ac_tail = ret & 0xffff;
 	ac->ac_buddy = ret >> 16;
 
-	/* XXXXXXX: SUCH A HORRIBLE **CK */
-	/*FIXME!! Why ? */
+	/*
+	 * take the page reference. We want the page to be pinned
+	 * so that we don't get a ext4_mb_init_cache_call for this
+	 * group until we update the bitmap. That would mean we
+	 * double allocate blocks. The reference is dropped
+	 * in ext4_mb_release_context
+	 */
 	ac->ac_bitmap_page = e4b->bd_bitmap_page;
 	get_page(ac->ac_bitmap_page);
 	ac->ac_buddy_page = e4b->bd_buddy_page;
 	get_page(ac->ac_buddy_page);
-
+	/* on allocation we use ac to track the held semaphore */
+	ac->alloc_semp =  e4b->alloc_semp;
+	e4b->alloc_semp = NULL;
 	/* store last allocated for subsequent stream allocation */
 	if ((ac->ac_flags & EXT4_MB_HINT_DATA)) {
 		spin_lock(&sbi->s_md_lock);
@@ -1326,6 +1399,8 @@ static void ext4_mb_check_limits(struct ext4_allocation_context *ac,
 	struct ext4_free_extent ex;
 	int max;
 
+	if (ac->ac_status == AC_STATUS_FOUND)
+		return;
 	/*
 	 * We don't want to scan for a whole year
 	 */
@@ -1575,8 +1650,9 @@ static void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac,
 			 * free blocks even though group info says we
 			 * we have free blocks
 			 */
-			ext4_error(sb, __func__, "%d free blocks as per "
-					"group info. But bitmap says 0\n",
+			ext4_grp_locked_error(sb, e4b->bd_group,
+					__func__, "%d free blocks as per "
+					"group info. But bitmap says 0",
 					free);
 			break;
 		}
@@ -1584,8 +1660,9 @@ static void ext4_mb_complex_scan_group(struct ext4_allocation_context *ac,
 		mb_find_extent(e4b, 0, i, ac->ac_g_ex.fe_len, &ex);
 		BUG_ON(ex.fe_len <= 0);
 		if (free < ex.fe_len) {
-			ext4_error(sb, __func__, "%d free blocks as per "
-					"group info. But got %d blocks\n",
+			ext4_grp_locked_error(sb, e4b->bd_group,
+					__func__, "%d free blocks as per "
+					"group info. But got %d blocks",
 					free, ex.fe_len);
 			/*
 			 * The number of free blocks differs. This mostly
@@ -1692,6 +1769,173 @@ static int ext4_mb_good_group(struct ext4_allocation_context *ac,
 	return 0;
 }
 
+/*
+ * lock the group_info alloc_sem of all the groups
+ * belonging to the same buddy cache page. This
+ * make sure other parallel operation on the buddy
+ * cache doesn't happen  whild holding the buddy cache
+ * lock
+ */
+int ext4_mb_get_buddy_cache_lock(struct super_block *sb, ext4_group_t group)
+{
+	int i;
+	int block, pnum;
+	int blocks_per_page;
+	int groups_per_page;
+	ext4_group_t first_group;
+	struct ext4_group_info *grp;
+
+	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
+	/*
+	 * the buddy cache inode stores the block bitmap
+	 * and buddy information in consecutive blocks.
+	 * So for each group we need two blocks.
+	 */
+	block = group * 2;
+	pnum = block / blocks_per_page;
+	first_group = pnum * blocks_per_page / 2;
+
+	groups_per_page = blocks_per_page >> 1;
+	if (groups_per_page == 0)
+		groups_per_page = 1;
+	/* read all groups the page covers into the cache */
+	for (i = 0; i < groups_per_page; i++) {
+
+		if ((first_group + i) >= EXT4_SB(sb)->s_groups_count)
+			break;
+		grp = ext4_get_group_info(sb, first_group + i);
+		/* take all groups write allocation
+		 * semaphore. This make sure there is
+		 * no block allocation going on in any
+		 * of that groups
+		 */
+		down_write_nested(&grp->alloc_sem, i);
+	}
+	return i;
+}
+
+void ext4_mb_put_buddy_cache_lock(struct super_block *sb,
+					ext4_group_t group, int locked_group)
+{
+	int i;
+	int block, pnum;
+	int blocks_per_page;
+	ext4_group_t first_group;
+	struct ext4_group_info *grp;
+
+	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
+	/*
+	 * the buddy cache inode stores the block bitmap
+	 * and buddy information in consecutive blocks.
+	 * So for each group we need two blocks.
+	 */
+	block = group * 2;
+	pnum = block / blocks_per_page;
+	first_group = pnum * blocks_per_page / 2;
+	/* release locks on all the groups */
+	for (i = 0; i < locked_group; i++) {
+
+		grp = ext4_get_group_info(sb, first_group + i);
+		/* take all groups write allocation
+		 * semaphore. This make sure there is
+		 * no block allocation going on in any
+		 * of that groups
+		 */
+		up_write(&grp->alloc_sem);
+	}
+
+}
+
+static int ext4_mb_init_group(struct super_block *sb, ext4_group_t group)
+{
+
+	int ret;
+	void *bitmap;
+	int blocks_per_page;
+	int block, pnum, poff;
+	int num_grp_locked = 0;
+	struct ext4_group_info *this_grp;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	struct inode *inode = sbi->s_buddy_cache;
+	struct page *page = NULL, *bitmap_page = NULL;
+
+	mb_debug("init group %lu\n", group);
+	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
+	this_grp = ext4_get_group_info(sb, group);
+	/*
+	 * This ensures we don't add group
+	 * to this buddy cache via resize
+	 */
+	num_grp_locked =  ext4_mb_get_buddy_cache_lock(sb, group);
+	if (!EXT4_MB_GRP_NEED_INIT(this_grp)) {
+		/*
+		 * somebody initialized the group
+		 * return without doing anything
+		 */
+		ret = 0;
+		goto err;
+	}
+	/*
+	 * the buddy cache inode stores the block bitmap
+	 * and buddy information in consecutive blocks.
+	 * So for each group we need two blocks.
+	 */
+	block = group * 2;
+	pnum = block / blocks_per_page;
+	poff = block % blocks_per_page;
+	page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS);
+	if (page) {
+		BUG_ON(page->mapping != inode->i_mapping);
+		ret = ext4_mb_init_cache(page, NULL);
+		if (ret) {
+			unlock_page(page);
+			goto err;
+		}
+		unlock_page(page);
+	}
+	if (page == NULL || !PageUptodate(page)) {
+		ret = -EIO;
+		goto err;
+	}
+	mark_page_accessed(page);
+	bitmap_page = page;
+	bitmap = page_address(page) + (poff * sb->s_blocksize);
+
+	/* init buddy cache */
+	block++;
+	pnum = block / blocks_per_page;
+	poff = block % blocks_per_page;
+	page = find_or_create_page(inode->i_mapping, pnum, GFP_NOFS);
+	if (page == bitmap_page) {
+		/*
+		 * If both the bitmap and buddy are in
+		 * the same page we don't need to force
+		 * init the buddy
+		 */
+		unlock_page(page);
+	} else if (page) {
+		BUG_ON(page->mapping != inode->i_mapping);
+		ret = ext4_mb_init_cache(page, bitmap);
+		if (ret) {
+			unlock_page(page);
+			goto err;
+		}
+		unlock_page(page);
+	}
+	if (page == NULL || !PageUptodate(page)) {
+		ret = -EIO;
+		goto err;
+	}
+	mark_page_accessed(page);
+err:
+	ext4_mb_put_buddy_cache_lock(sb, group, num_grp_locked);
+	if (bitmap_page)
+		page_cache_release(bitmap_page);
+	if (page)
+		page_cache_release(page);
+	return ret;
+}
+
 static noinline_for_stack int
 ext4_mb_regular_allocator(struct ext4_allocation_context *ac)
 {
@@ -1775,7 +2019,7 @@ repeat:
 				group = 0;
 
 			/* quick check to skip empty groups */
-			grp = ext4_get_group_info(ac->ac_sb, group);
+			grp = ext4_get_group_info(sb, group);
 			if (grp->bb_free == 0)
 				continue;
 
@@ -1788,10 +2032,9 @@ repeat:
 				 * we need full data about the group
 				 * to make a good selection
 				 */
-				err = ext4_mb_load_buddy(sb, group, &e4b);
+				err = ext4_mb_init_group(sb, group);
 				if (err)
 					goto out;
-				ext4_mb_release_desc(&e4b);
 			}
 
 			/*
@@ -1932,13 +2175,13 @@ static int ext4_mb_seq_history_show(struct seq_file *seq, void *v)
 	if (hs->op == EXT4_MB_HISTORY_ALLOC) {
 		fmt = "%-5u %-8u %-23s %-23s %-23s %-5u %-5u %-2u "
 			"%-5u %-5s %-5u %-6u\n";
-		sprintf(buf2, "%lu/%d/%u@%u", hs->result.fe_group,
+		sprintf(buf2, "%u/%d/%u@%u", hs->result.fe_group,
 			hs->result.fe_start, hs->result.fe_len,
 			hs->result.fe_logical);
-		sprintf(buf, "%lu/%d/%u@%u", hs->orig.fe_group,
+		sprintf(buf, "%u/%d/%u@%u", hs->orig.fe_group,
 			hs->orig.fe_start, hs->orig.fe_len,
 			hs->orig.fe_logical);
-		sprintf(buf3, "%lu/%d/%u@%u", hs->goal.fe_group,
+		sprintf(buf3, "%u/%d/%u@%u", hs->goal.fe_group,
 			hs->goal.fe_start, hs->goal.fe_len,
 			hs->goal.fe_logical);
 		seq_printf(seq, fmt, hs->pid, hs->ino, buf, buf3, buf2,
@@ -1947,20 +2190,20 @@ static int ext4_mb_seq_history_show(struct seq_file *seq, void *v)
 				hs->buddy ? 1 << hs->buddy : 0);
 	} else if (hs->op == EXT4_MB_HISTORY_PREALLOC) {
 		fmt = "%-5u %-8u %-23s %-23s %-23s\n";
-		sprintf(buf2, "%lu/%d/%u@%u", hs->result.fe_group,
+		sprintf(buf2, "%u/%d/%u@%u", hs->result.fe_group,
 			hs->result.fe_start, hs->result.fe_len,
 			hs->result.fe_logical);
-		sprintf(buf, "%lu/%d/%u@%u", hs->orig.fe_group,
+		sprintf(buf, "%u/%d/%u@%u", hs->orig.fe_group,
 			hs->orig.fe_start, hs->orig.fe_len,
 			hs->orig.fe_logical);
 		seq_printf(seq, fmt, hs->pid, hs->ino, buf, "", buf2);
 	} else if (hs->op == EXT4_MB_HISTORY_DISCARD) {
-		sprintf(buf2, "%lu/%d/%u", hs->result.fe_group,
+		sprintf(buf2, "%u/%d/%u", hs->result.fe_group,
 			hs->result.fe_start, hs->result.fe_len);
 		seq_printf(seq, "%-5u %-8u %-23s discard\n",
 				hs->pid, hs->ino, buf2);
 	} else if (hs->op == EXT4_MB_HISTORY_FREE) {
-		sprintf(buf2, "%lu/%d/%u", hs->result.fe_group,
+		sprintf(buf2, "%u/%d/%u", hs->result.fe_group,
 			hs->result.fe_start, hs->result.fe_len);
 		seq_printf(seq, "%-5u %-8u %-23s free\n",
 				hs->pid, hs->ino, buf2);
@@ -2073,7 +2316,7 @@ static void *ext4_mb_seq_groups_start(struct seq_file *seq, loff_t *pos)
 		return NULL;
 
 	group = *pos + 1;
-	return (void *) group;
+	return (void *) ((unsigned long) group);
 }
 
 static void *ext4_mb_seq_groups_next(struct seq_file *seq, void *v, loff_t *pos)
@@ -2086,13 +2329,13 @@ static void *ext4_mb_seq_groups_next(struct seq_file *seq, void *v, loff_t *pos)
 	if (*pos < 0 || *pos >= sbi->s_groups_count)
 		return NULL;
 	group = *pos + 1;
-	return (void *) group;;
+	return (void *) ((unsigned long) group);
 }
 
 static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 {
 	struct super_block *sb = seq->private;
-	long group = (long) v;
+	ext4_group_t group = (ext4_group_t) ((unsigned long) v);
 	int i;
 	int err;
 	struct ext4_buddy e4b;
@@ -2114,7 +2357,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 		sizeof(struct ext4_group_info);
 	err = ext4_mb_load_buddy(sb, group, &e4b);
 	if (err) {
-		seq_printf(seq, "#%-5lu: I/O error\n", group);
+		seq_printf(seq, "#%-5u: I/O error\n", group);
 		return 0;
 	}
 	ext4_lock_group(sb, group);
@@ -2122,7 +2365,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 	ext4_unlock_group(sb, group);
 	ext4_mb_release_desc(&e4b);
 
-	seq_printf(seq, "#%-5lu: %-5u %-5u %-5u [", group, sg.info.bb_free,
+	seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free,
 			sg.info.bb_fragments, sg.info.bb_first_free);
 	for (i = 0; i <= 13; i++)
 		seq_printf(seq, " %-5u", i <= sb->s_blocksize_bits + 1 ?
@@ -2296,10 +2539,11 @@ int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 			ext4_free_blocks_after_init(sb, group, desc);
 	} else {
 		meta_group_info[i]->bb_free =
-			le16_to_cpu(desc->bg_free_blocks_count);
+			ext4_free_blks_count(sb, desc);
 	}
 
 	INIT_LIST_HEAD(&meta_group_info[i]->bb_prealloc_list);
+	init_rwsem(&meta_group_info[i]->alloc_sem);
 	meta_group_info[i]->bb_free_root.rb_node = NULL;;
 
 #ifdef DOUBLE_CHECK
@@ -2327,54 +2571,6 @@ exit_meta_group_info:
 } /* ext4_mb_add_groupinfo */
 
 /*
- * Add a group to the existing groups.
- * This function is used for online resize
- */
-int ext4_mb_add_more_groupinfo(struct super_block *sb, ext4_group_t group,
-			       struct ext4_group_desc *desc)
-{
-	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	struct inode *inode = sbi->s_buddy_cache;
-	int blocks_per_page;
-	int block;
-	int pnum;
-	struct page *page;
-	int err;
-
-	/* Add group based on group descriptor*/
-	err = ext4_mb_add_groupinfo(sb, group, desc);
-	if (err)
-		return err;
-
-	/*
-	 * Cache pages containing dynamic mb_alloc datas (buddy and bitmap
-	 * datas) are set not up to date so that they will be re-initilaized
-	 * during the next call to ext4_mb_load_buddy
-	 */
-
-	/* Set buddy page as not up to date */
-	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
-	block = group * 2;
-	pnum = block / blocks_per_page;
-	page = find_get_page(inode->i_mapping, pnum);
-	if (page != NULL) {
-		ClearPageUptodate(page);
-		page_cache_release(page);
-	}
-
-	/* Set bitmap page as not up to date */
-	block++;
-	pnum = block / blocks_per_page;
-	page = find_get_page(inode->i_mapping, pnum);
-	if (page != NULL) {
-		ClearPageUptodate(page);
-		page_cache_release(page);
-	}
-
-	return 0;
-}
-
-/*
  * Update an existing group.
  * This function is used for online resize
  */
@@ -2457,7 +2653,7 @@ static int ext4_mb_init_backend(struct super_block *sb)
 		desc = ext4_get_group_desc(sb, i, NULL);
 		if (desc == NULL) {
 			printk(KERN_ERR
-				"EXT4-fs: can't read descriptor %lu\n", i);
+				"EXT4-fs: can't read descriptor %u\n", i);
 			goto err_freebuddy;
 		}
 		if (ext4_mb_add_groupinfo(sb, i, desc) != 0)
@@ -2493,6 +2689,8 @@ int ext4_mb_init(struct super_block *sb, int needs_recovery)
 	if (sbi->s_mb_offsets == NULL) {
 		return -ENOMEM;
 	}
+
+	i = (sb->s_blocksize_bits + 2) * sizeof(unsigned int);
 	sbi->s_mb_maxs = kmalloc(i, GFP_KERNEL);
 	if (sbi->s_mb_maxs == NULL) {
 		kfree(sbi->s_mb_maxs);
@@ -2551,7 +2749,8 @@ int ext4_mb_init(struct super_block *sb, int needs_recovery)
 	ext4_mb_init_per_dev_proc(sb);
 	ext4_mb_history_init(sb);
 
-	sbi->s_journal->j_commit_callback = release_blocks_on_commit;
+	if (sbi->s_journal)
+		sbi->s_journal->j_commit_callback = release_blocks_on_commit;
 
 	printk(KERN_INFO "EXT4-fs: mballoc enabled\n");
 	return 0;
@@ -2652,7 +2851,7 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn)
 	list_for_each_safe(l, ltmp, &txn->t_private_list) {
 		entry = list_entry(l, struct ext4_free_data, list);
 
-		mb_debug("gonna free %u blocks in group %lu (0x%p):",
+		mb_debug("gonna free %u blocks in group %u (0x%p):",
 			 entry->count, entry->group, entry);
 
 		err = ext4_mb_load_buddy(sb, entry->group, &e4b);
@@ -2679,8 +2878,9 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn)
 		discard_block = (ext4_fsblk_t) entry->group * EXT4_BLOCKS_PER_GROUP(sb)
 			+ entry->start_blk
 			+ le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
-		trace_mark(ext4_discard_blocks, "dev %s blk %llu count %u", sb->s_id,
-			   (unsigned long long) discard_block, entry->count);
+		trace_mark(ext4_discard_blocks, "dev %s blk %llu count %u",
+			   sb->s_id, (unsigned long long) discard_block,
+			   entry->count);
 		sb_issue_discard(sb, discard_block, entry->count);
 
 		kmem_cache_free(ext4_free_ext_cachep, entry);
@@ -2791,7 +2991,7 @@ void exit_ext4_mballoc(void)
  */
 static noinline_for_stack int
 ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
-				handle_t *handle, unsigned long reserv_blks)
+				handle_t *handle, unsigned int reserv_blks)
 {
 	struct buffer_head *bitmap_bh = NULL;
 	struct ext4_super_block *es;
@@ -2824,7 +3024,7 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 	if (!gdp)
 		goto out_err;
 
-	ext4_debug("using block group %lu(%d)\n", ac->ac_b_ex.fe_group,
+	ext4_debug("using block group %u(%d)\n", ac->ac_b_ex.fe_group,
 			gdp->bg_free_blocks_count);
 
 	err = ext4_journal_get_write_access(handle, gdp_bh);
@@ -2843,8 +3043,8 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 	    in_range(block + len - 1, ext4_inode_table(sb, gdp),
 		     EXT4_SB(sb)->s_itb_per_group)) {
 		ext4_error(sb, __func__,
-			   "Allocating block in system zone - block = %llu",
-			   block);
+			   "Allocating block %llu in system zone of %d group\n",
+			   block, ac->ac_b_ex.fe_group);
 		/* File system mounted not to panic on error
 		 * Fix the bitmap and repeat the block allocation
 		 * We leak some of the blocks here.
@@ -2852,7 +3052,7 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 		mb_set_bits(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group),
 				bitmap_bh->b_data, ac->ac_b_ex.fe_start,
 				ac->ac_b_ex.fe_len);
-		err = ext4_journal_dirty_metadata(handle, bitmap_bh);
+		err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
 		if (!err)
 			err = -EAGAIN;
 		goto out_err;
@@ -2866,18 +3066,17 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 		}
 	}
 #endif
-	mb_set_bits(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group), bitmap_bh->b_data,
-				ac->ac_b_ex.fe_start, ac->ac_b_ex.fe_len);
-
 	spin_lock(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group));
+	mb_set_bits(NULL, bitmap_bh->b_data,
+				ac->ac_b_ex.fe_start, ac->ac_b_ex.fe_len);
 	if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
 		gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT);
-		gdp->bg_free_blocks_count =
-			cpu_to_le16(ext4_free_blocks_after_init(sb,
-						ac->ac_b_ex.fe_group,
-						gdp));
+		ext4_free_blks_set(sb, gdp,
+					ext4_free_blocks_after_init(sb,
+					ac->ac_b_ex.fe_group, gdp));
 	}
-	le16_add_cpu(&gdp->bg_free_blocks_count, -ac->ac_b_ex.fe_len);
+	len = ext4_free_blks_count(sb, gdp) - ac->ac_b_ex.fe_len;
+	ext4_free_blks_set(sb, gdp, len);
 	gdp->bg_checksum = ext4_group_desc_csum(sbi, ac->ac_b_ex.fe_group, gdp);
 	spin_unlock(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group));
 	percpu_counter_sub(&sbi->s_freeblocks_counter, ac->ac_b_ex.fe_len);
@@ -2899,10 +3098,10 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
 		spin_unlock(sb_bgl_lock(sbi, flex_group));
 	}
 
-	err = ext4_journal_dirty_metadata(handle, bitmap_bh);
+	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
 	if (err)
 		goto out_err;
-	err = ext4_journal_dirty_metadata(handle, gdp_bh);
+	err = ext4_handle_dirty_metadata(handle, NULL, gdp_bh);
 
 out_err:
 	sb->s_dirt = 1;
@@ -3031,7 +3230,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
 	/* check we don't cross already preallocated blocks */
 	rcu_read_lock();
 	list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) {
-		unsigned long pa_end;
+		ext4_lblk_t pa_end;
 
 		if (pa->pa_deleted)
 			continue;
@@ -3075,7 +3274,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
 	/* XXX: extra loop to check we really don't overlap preallocations */
 	rcu_read_lock();
 	list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) {
-		unsigned long pa_end;
+		ext4_lblk_t pa_end;
 		spin_lock(&pa->pa_lock);
 		if (pa->pa_deleted == 0) {
 			pa_end = pa->pa_lstart + pa->pa_len;
@@ -3307,6 +3506,32 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
 }
 
 /*
+ * the function goes through all block freed in the group
+ * but not yet committed and marks them used in in-core bitmap.
+ * buddy must be generated from this bitmap
+ * Need to be called with ext4 group lock (ext4_lock_group)
+ */
+static void ext4_mb_generate_from_freelist(struct super_block *sb, void *bitmap,
+						ext4_group_t group)
+{
+	struct rb_node *n;
+	struct ext4_group_info *grp;
+	struct ext4_free_data *entry;
+
+	grp = ext4_get_group_info(sb, group);
+	n = rb_first(&(grp->bb_free_root));
+
+	while (n) {
+		entry = rb_entry(n, struct ext4_free_data, node);
+		mb_set_bits(sb_bgl_lock(EXT4_SB(sb), group),
+				bitmap, entry->start_blk,
+				entry->count);
+		n = rb_next(n);
+	}
+	return;
+}
+
+/*
  * the function goes through all preallocation in this group and marks them
  * used in in-core bitmap. buddy must be generated from this bitmap
  * Need to be called with ext4 group lock (ext4_lock_group)
@@ -3346,7 +3571,7 @@ static void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap,
 		preallocated += len;
 		count++;
 	}
-	mb_debug("prellocated %u for group %lu\n", preallocated, group);
+	mb_debug("prellocated %u for group %u\n", preallocated, group);
 }
 
 static void ext4_mb_pa_callback(struct rcu_head *head)
@@ -3363,7 +3588,7 @@ static void ext4_mb_pa_callback(struct rcu_head *head)
 static void ext4_mb_put_pa(struct ext4_allocation_context *ac,
 			struct super_block *sb, struct ext4_prealloc_space *pa)
 {
-	unsigned long grp;
+	ext4_group_t grp;
 
 	if (!atomic_dec_and_test(&pa->pa_count) || pa->pa_free != 0)
 		return;
@@ -3473,6 +3698,10 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
 
 	mb_debug("new inode pa %p: %llu/%u for %u\n", pa,
 			pa->pa_pstart, pa->pa_len, pa->pa_lstart);
+	trace_mark(ext4_mb_new_inode_pa,
+		   "dev %s ino %lu pstart %llu len %u lstart %u",
+		   sb->s_id, ac->ac_inode->i_ino,
+		   pa->pa_pstart, pa->pa_len, pa->pa_lstart);
 
 	ext4_mb_use_inode_pa(ac, pa);
 	atomic_add(pa->pa_free, &EXT4_SB(sb)->s_mb_preallocated);
@@ -3530,7 +3759,9 @@ ext4_mb_new_group_pa(struct ext4_allocation_context *ac)
 	pa->pa_linear = 1;
 
 	mb_debug("new group pa %p: %llu/%u for %u\n", pa,
-			pa->pa_pstart, pa->pa_len, pa->pa_lstart);
+		 pa->pa_pstart, pa->pa_len, pa->pa_lstart);
+	trace_mark(ext4_mb_new_group_pa, "dev %s pstart %llu len %u lstart %u",
+		   sb->s_id, pa->pa_pstart, pa->pa_len, pa->pa_lstart);
 
 	ext4_mb_use_group_pa(ac, pa);
 	atomic_add(pa->pa_free, &EXT4_SB(sb)->s_mb_preallocated);
@@ -3579,16 +3810,18 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh,
 {
 	struct super_block *sb = e4b->bd_sb;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	unsigned long end;
-	unsigned long next;
+	unsigned int end;
+	unsigned int next;
 	ext4_group_t group;
 	ext4_grpblk_t bit;
+	unsigned long long grp_blk_start;
 	sector_t start;
 	int err = 0;
 	int free = 0;
 
 	BUG_ON(pa->pa_deleted == 0);
 	ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit);
+	grp_blk_start = pa->pa_pstart - bit;
 	BUG_ON(group != e4b->bd_group && pa->pa_len != 0);
 	end = bit + pa->pa_len;
 
@@ -3618,6 +3851,10 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh,
 			ext4_mb_store_history(ac);
 		}
 
+		trace_mark(ext4_mb_release_inode_pa,
+			   "dev %s ino %lu block %llu count %u",
+			   sb->s_id, pa->pa_inode->i_ino, grp_blk_start + bit,
+			   next - bit);
 		mb_free_blocks(pa->pa_inode, e4b, bit, next - bit);
 		bit = next + 1;
 	}
@@ -3626,8 +3863,9 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh,
 			pa, (unsigned long) pa->pa_lstart,
 			(unsigned long) pa->pa_pstart,
 			(unsigned long) pa->pa_len);
-		ext4_error(sb, __func__, "free %u, pa_free %u\n",
-						free, pa->pa_free);
+		ext4_grp_locked_error(sb, group,
+					__func__, "free %u, pa_free %u",
+					free, pa->pa_free);
 		/*
 		 * pa is already deleted so we use the value obtained
 		 * from the bitmap and continue.
@@ -3650,6 +3888,8 @@ ext4_mb_release_group_pa(struct ext4_buddy *e4b,
 	if (ac)
 		ac->ac_op = EXT4_MB_HISTORY_DISCARD;
 
+	trace_mark(ext4_mb_release_group_pa, "dev %s pstart %llu len %d",
+		   sb->s_id, pa->pa_pstart, pa->pa_len);
 	BUG_ON(pa->pa_deleted == 0);
 	ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit);
 	BUG_ON(group != e4b->bd_group && pa->pa_len != 0);
@@ -3692,7 +3932,7 @@ ext4_mb_discard_group_preallocations(struct super_block *sb,
 	int busy = 0;
 	int free = 0;
 
-	mb_debug("discard preallocation for group %lu\n", group);
+	mb_debug("discard preallocation for group %u\n", group);
 
 	if (list_empty(&grp->bb_prealloc_list))
 		return 0;
@@ -3700,14 +3940,14 @@ ext4_mb_discard_group_preallocations(struct super_block *sb,
 	bitmap_bh = ext4_read_block_bitmap(sb, group);
 	if (bitmap_bh == NULL) {
 		ext4_error(sb, __func__, "Error in reading block "
-				"bitmap for %lu\n", group);
+				"bitmap for %u", group);
 		return 0;
 	}
 
 	err = ext4_mb_load_buddy(sb, group, &e4b);
 	if (err) {
 		ext4_error(sb, __func__, "Error in loading buddy "
-				"information for %lu\n", group);
+				"information for %u", group);
 		put_bh(bitmap_bh);
 		return 0;
 	}
@@ -3815,6 +4055,8 @@ void ext4_discard_preallocations(struct inode *inode)
 	}
 
 	mb_debug("discard preallocation for inode %lu\n", inode->i_ino);
+	trace_mark(ext4_discard_preallocations, "dev %s ino %lu", sb->s_id,
+		   inode->i_ino);
 
 	INIT_LIST_HEAD(&list);
 
@@ -3874,14 +4116,14 @@ repeat:
 		err = ext4_mb_load_buddy(sb, group, &e4b);
 		if (err) {
 			ext4_error(sb, __func__, "Error in loading buddy "
-					"information for %lu\n", group);
+					"information for %u", group);
 			continue;
 		}
 
 		bitmap_bh = ext4_read_block_bitmap(sb, group);
 		if (bitmap_bh == NULL) {
 			ext4_error(sb, __func__, "Error in reading block "
-					"bitmap for %lu\n", group);
+					"bitmap for %u", group);
 			ext4_mb_release_desc(&e4b);
 			continue;
 		}
@@ -4024,8 +4266,8 @@ ext4_mb_initialize_context(struct ext4_allocation_context *ac,
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	struct ext4_super_block *es = sbi->s_es;
 	ext4_group_t group;
-	unsigned long len;
-	unsigned long goal;
+	unsigned int len;
+	ext4_fsblk_t goal;
 	ext4_grpblk_t block;
 
 	/* we can't allocate > group size */
@@ -4068,6 +4310,7 @@ ext4_mb_initialize_context(struct ext4_allocation_context *ac,
 	ac->ac_pa = NULL;
 	ac->ac_bitmap_page = NULL;
 	ac->ac_buddy_page = NULL;
+	ac->alloc_semp = NULL;
 	ac->ac_lg = NULL;
 
 	/* we have to define context: we'll we work with a file or
@@ -4146,7 +4389,7 @@ ext4_mb_discard_lg_preallocations(struct super_block *sb,
 		ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, NULL);
 		if (ext4_mb_load_buddy(sb, group, &e4b)) {
 			ext4_error(sb, __func__, "Error in loading buddy "
-					"information for %lu\n", group);
+					"information for %u", group);
 			continue;
 		}
 		ext4_lock_group(sb, group);
@@ -4248,6 +4491,8 @@ static int ext4_mb_release_context(struct ext4_allocation_context *ac)
 		}
 		ext4_mb_put_pa(ac, ac->ac_sb, pa);
 	}
+	if (ac->alloc_semp)
+		up_read(ac->alloc_semp);
 	if (ac->ac_bitmap_page)
 		page_cache_release(ac->ac_bitmap_page);
 	if (ac->ac_buddy_page)
@@ -4264,6 +4509,8 @@ static int ext4_mb_discard_preallocations(struct super_block *sb, int needed)
 	int ret;
 	int freed = 0;
 
+	trace_mark(ext4_mb_discard_preallocations, "dev %s needed %d",
+		   sb->s_id, needed);
 	for (i = 0; i < EXT4_SB(sb)->s_groups_count && needed > 0; i++) {
 		ret = ext4_mb_discard_group_preallocations(sb, i, needed);
 		freed += ret;
@@ -4286,12 +4533,24 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
 	struct ext4_sb_info *sbi;
 	struct super_block *sb;
 	ext4_fsblk_t block = 0;
-	unsigned long inquota;
-	unsigned long reserv_blks = 0;
+	unsigned int inquota;
+	unsigned int reserv_blks = 0;
 
 	sb = ar->inode->i_sb;
 	sbi = EXT4_SB(sb);
 
+	trace_mark(ext4_request_blocks, "dev %s flags %u len %u ino %lu "
+		   "lblk %llu goal %llu lleft %llu lright %llu "
+		   "pleft %llu pright %llu ",
+		   sb->s_id, ar->flags, ar->len,
+		   ar->inode ? ar->inode->i_ino : 0,
+		   (unsigned long long) ar->logical,
+		   (unsigned long long) ar->goal,
+		   (unsigned long long) ar->lleft,
+		   (unsigned long long) ar->lright,
+		   (unsigned long long) ar->pleft,
+		   (unsigned long long) ar->pright);
+
 	if (!EXT4_I(ar->inode)->i_delalloc_reserved_flag) {
 		/*
 		 * With delalloc we already reserved the blocks
@@ -4313,7 +4572,7 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
 	}
 	if (ar->len == 0) {
 		*errp = -EDQUOT;
-		return 0;
+		goto out3;
 	}
 	inquota = ar->len;
 
@@ -4348,10 +4607,14 @@ repeat:
 				ac->ac_o_ex.fe_len < ac->ac_b_ex.fe_len)
 			ext4_mb_new_preallocation(ac);
 	}
-
 	if (likely(ac->ac_status == AC_STATUS_FOUND)) {
 		*errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_blks);
 		if (*errp ==  -EAGAIN) {
+			/*
+			 * drop the reference that we took
+			 * in ext4_mb_use_best_found
+			 */
+			ext4_mb_release_context(ac);
 			ac->ac_b_ex.fe_group = 0;
 			ac->ac_b_ex.fe_start = 0;
 			ac->ac_b_ex.fe_len = 0;
@@ -4382,6 +4645,26 @@ out2:
 out1:
 	if (ar->len < inquota)
 		DQUOT_FREE_BLOCK(ar->inode, inquota - ar->len);
+out3:
+	if (!ar->len) {
+		if (!EXT4_I(ar->inode)->i_delalloc_reserved_flag)
+			/* release all the reserved blocks if non delalloc */
+			percpu_counter_sub(&sbi->s_dirtyblocks_counter,
+						reserv_blks);
+	}
+
+	trace_mark(ext4_allocate_blocks,
+		   "dev %s block %llu flags %u len %u ino %lu "
+		   "logical %llu goal %llu lleft %llu lright %llu "
+		   "pleft %llu pright %llu ",
+		   sb->s_id, (unsigned long long) block,
+		   ar->flags, ar->len, ar->inode ? ar->inode->i_ino : 0,
+		   (unsigned long long) ar->logical,
+		   (unsigned long long) ar->goal,
+		   (unsigned long long) ar->lleft,
+		   (unsigned long long) ar->lright,
+		   (unsigned long long) ar->pleft,
+		   (unsigned long long) ar->pright);
 
 	return block;
 }
@@ -4403,27 +4686,23 @@ static int can_merge(struct ext4_free_data *entry1,
 
 static noinline_for_stack int
 ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b,
-			  ext4_group_t group, ext4_grpblk_t block, int count)
+		      struct ext4_free_data *new_entry)
 {
+	ext4_grpblk_t block;
+	struct ext4_free_data *entry;
 	struct ext4_group_info *db = e4b->bd_info;
 	struct super_block *sb = e4b->bd_sb;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	struct ext4_free_data *entry, *new_entry;
 	struct rb_node **n = &db->bb_free_root.rb_node, *node;
 	struct rb_node *parent = NULL, *new_node;
 
-
+	BUG_ON(!ext4_handle_valid(handle));
 	BUG_ON(e4b->bd_bitmap_page == NULL);
 	BUG_ON(e4b->bd_buddy_page == NULL);
 
-	new_entry  = kmem_cache_alloc(ext4_free_ext_cachep, GFP_NOFS);
-	new_entry->start_blk = block;
-	new_entry->group  = group;
-	new_entry->count = count;
-	new_entry->t_tid = handle->h_transaction->t_tid;
 	new_node = &new_entry->node;
+	block = new_entry->start_blk;
 
-	ext4_lock_group(sb, group);
 	if (!*n) {
 		/* first free block exent. We need to
 		   protect buddy cache from being freed,
@@ -4441,10 +4720,9 @@ ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b,
 		else if (block >= (entry->start_blk + entry->count))
 			n = &(*n)->rb_right;
 		else {
-			ext4_unlock_group(sb, group);
-			ext4_error(sb, __func__,
-			    "Double free of blocks %d (%d %d)\n",
-			    block, entry->start_blk, entry->count);
+			ext4_grp_locked_error(sb, e4b->bd_group, __func__,
+					"Double free of blocks %d (%d %d)",
+					block, entry->start_blk, entry->count);
 			return 0;
 		}
 	}
@@ -4483,7 +4761,6 @@ ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b,
 	spin_lock(&sbi->s_md_lock);
 	list_add(&new_entry->list, &handle->h_transaction->t_private_list);
 	spin_unlock(&sbi->s_md_lock);
-	ext4_unlock_group(sb, group);
 	return 0;
 }
 
@@ -4499,7 +4776,7 @@ void ext4_mb_free_blocks(handle_t *handle, struct inode *inode,
 	struct ext4_allocation_context *ac = NULL;
 	struct ext4_group_desc *gdp;
 	struct ext4_super_block *es;
-	unsigned long overflow;
+	unsigned int overflow;
 	ext4_grpblk_t bit;
 	struct buffer_head *gd_bh;
 	ext4_group_t block_group;
@@ -4522,6 +4799,10 @@ void ext4_mb_free_blocks(handle_t *handle, struct inode *inode,
 	}
 
 	ext4_debug("freeing block %lu\n", block);
+	trace_mark(ext4_free_blocks,
+		   "dev %s block %llu count %lu metadata %d ino %lu",
+		   sb->s_id, (unsigned long long) block, count, metadata,
+		   inode ? inode->i_ino : 0);
 
 	ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS);
 	if (ac) {
@@ -4581,11 +4862,6 @@ do_more:
 	err = ext4_journal_get_write_access(handle, gd_bh);
 	if (err)
 		goto error_return;
-
-	err = ext4_mb_load_buddy(sb, block_group, &e4b);
-	if (err)
-		goto error_return;
-
 #ifdef AGGRESSIVE_CHECK
 	{
 		int i;
@@ -4593,13 +4869,6 @@ do_more:
 			BUG_ON(!mb_test_bit(bit + i, bitmap_bh->b_data));
 	}
 #endif
-	mb_clear_bits(sb_bgl_lock(sbi, block_group), bitmap_bh->b_data,
-			bit, count);
-
-	/* We dirtied the bitmap block */
-	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
-	err = ext4_journal_dirty_metadata(handle, bitmap_bh);
-
 	if (ac) {
 		ac->ac_b_ex.fe_group = block_group;
 		ac->ac_b_ex.fe_start = bit;
@@ -4607,19 +4876,41 @@ do_more:
 		ext4_mb_store_history(ac);
 	}
 
-	if (metadata) {
-		/* blocks being freed are metadata. these blocks shouldn't
-		 * be used until this transaction is committed */
-		ext4_mb_free_metadata(handle, &e4b, block_group, bit, count);
+	err = ext4_mb_load_buddy(sb, block_group, &e4b);
+	if (err)
+		goto error_return;
+	if (metadata && ext4_handle_valid(handle)) {
+		struct ext4_free_data *new_entry;
+		/*
+		 * blocks being freed are metadata. these blocks shouldn't
+		 * be used until this transaction is committed
+		 */
+		new_entry  = kmem_cache_alloc(ext4_free_ext_cachep, GFP_NOFS);
+		new_entry->start_blk = bit;
+		new_entry->group  = block_group;
+		new_entry->count = count;
+		new_entry->t_tid = handle->h_transaction->t_tid;
+		ext4_lock_group(sb, block_group);
+		mb_clear_bits(sb_bgl_lock(sbi, block_group), bitmap_bh->b_data,
+				bit, count);
+		ext4_mb_free_metadata(handle, &e4b, new_entry);
+		ext4_unlock_group(sb, block_group);
 	} else {
 		ext4_lock_group(sb, block_group);
+		/* need to update group_info->bb_free and bitmap
+		 * with group lock held. generate_buddy look at
+		 * them with group lock_held
+		 */
+		mb_clear_bits(sb_bgl_lock(sbi, block_group), bitmap_bh->b_data,
+				bit, count);
 		mb_free_blocks(inode, &e4b, bit, count);
 		ext4_mb_return_to_preallocation(inode, &e4b, block, count);
 		ext4_unlock_group(sb, block_group);
 	}
 
 	spin_lock(sb_bgl_lock(sbi, block_group));
-	le16_add_cpu(&gdp->bg_free_blocks_count, count);
+	ret = ext4_free_blks_count(sb, gdp) + count;
+	ext4_free_blks_set(sb, gdp, ret);
 	gdp->bg_checksum = ext4_group_desc_csum(sbi, block_group, gdp);
 	spin_unlock(sb_bgl_lock(sbi, block_group));
 	percpu_counter_add(&sbi->s_freeblocks_counter, count);
@@ -4635,9 +4926,13 @@ do_more:
 
 	*freed += count;
 
+	/* We dirtied the bitmap block */
+	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
+	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
+
 	/* And the group descriptor block */
 	BUFFER_TRACE(gd_bh, "dirtied group descriptor block");
-	ret = ext4_journal_dirty_metadata(handle, gd_bh);
+	ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh);
 	if (!err)
 		err = ret;
 
diff --git a/fs/ext4/mballoc.h b/fs/ext4/mballoc.h
index b5dff1fff1e..10a2921baf1 100644
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -20,6 +20,7 @@
 #include <linux/version.h>
 #include <linux/blkdev.h>
 #include <linux/marker.h>
+#include <linux/mutex.h>
 #include "ext4_jbd2.h"
 #include "ext4.h"
 #include "group.h"
@@ -98,9 +99,6 @@
  */
 #define MB_DEFAULT_GROUP_PREALLOC	512
 
-static struct kmem_cache *ext4_pspace_cachep;
-static struct kmem_cache *ext4_ac_cachep;
-static struct kmem_cache *ext4_free_ext_cachep;
 
 struct ext4_free_data {
 	/* this links the free block information from group_info */
@@ -120,26 +118,6 @@ struct ext4_free_data {
 	tid_t	t_tid;
 };
 
-struct ext4_group_info {
-	unsigned long	bb_state;
-	struct rb_root  bb_free_root;
-	unsigned short	bb_first_free;
-	unsigned short	bb_free;
-	unsigned short	bb_fragments;
-	struct		list_head bb_prealloc_list;
-#ifdef DOUBLE_CHECK
-	void		*bb_bitmap;
-#endif
-	unsigned short	bb_counters[];
-};
-
-#define EXT4_GROUP_INFO_NEED_INIT_BIT	0
-#define EXT4_GROUP_INFO_LOCKED_BIT	1
-
-#define EXT4_MB_GRP_NEED_INIT(grp)	\
-	(test_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &((grp)->bb_state)))
-
-
 struct ext4_prealloc_space {
 	struct list_head	pa_inode_list;
 	struct list_head	pa_group_list;
@@ -217,6 +195,11 @@ struct ext4_allocation_context {
 	__u8 ac_op;		/* operation, for history only */
 	struct page *ac_bitmap_page;
 	struct page *ac_buddy_page;
+	/*
+	 * pointer to the held semaphore upon successful
+	 * block allocation
+	 */
+	struct rw_semaphore *alloc_semp;
 	struct ext4_prealloc_space *ac_pa;
 	struct ext4_locality_group *ac_lg;
 };
@@ -250,6 +233,7 @@ struct ext4_buddy {
 	struct super_block *bd_sb;
 	__u16 bd_blkbits;
 	ext4_group_t bd_group;
+	struct rw_semaphore *alloc_semp;
 };
 #define EXT4_MB_BITMAP(e4b)	((e4b)->bd_bitmap)
 #define EXT4_MB_BUDDY(e4b)	((e4b)->bd_buddy)
@@ -259,51 +243,12 @@ static inline void ext4_mb_store_history(struct ext4_allocation_context *ac)
 {
 	return;
 }
-#else
-static void ext4_mb_store_history(struct ext4_allocation_context *ac);
 #endif
 
 #define in_range(b, first, len)	((b) >= (first) && (b) <= (first) + (len) - 1)
 
 struct buffer_head *read_block_bitmap(struct super_block *, ext4_group_t);
-
-static void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap,
-					ext4_group_t group);
-static void ext4_mb_return_to_preallocation(struct inode *inode,
-					struct ext4_buddy *e4b, sector_t block,
-					int count);
-static void ext4_mb_put_pa(struct ext4_allocation_context *,
-			struct super_block *, struct ext4_prealloc_space *pa);
-static int ext4_mb_init_per_dev_proc(struct super_block *sb);
-static int ext4_mb_destroy_per_dev_proc(struct super_block *sb);
-static void release_blocks_on_commit(journal_t *journal, transaction_t *txn);
-
-
-static inline void ext4_lock_group(struct super_block *sb, ext4_group_t group)
-{
-	struct ext4_group_info *grinfo = ext4_get_group_info(sb, group);
-
-	bit_spin_lock(EXT4_GROUP_INFO_LOCKED_BIT, &(grinfo->bb_state));
-}
-
-static inline void ext4_unlock_group(struct super_block *sb,
-					ext4_group_t group)
-{
-	struct ext4_group_info *grinfo = ext4_get_group_info(sb, group);
-
-	bit_spin_unlock(EXT4_GROUP_INFO_LOCKED_BIT, &(grinfo->bb_state));
-}
-
-static inline int ext4_is_group_locked(struct super_block *sb,
-					ext4_group_t group)
-{
-	struct ext4_group_info *grinfo = ext4_get_group_info(sb, group);
-
-	return bit_spin_is_locked(EXT4_GROUP_INFO_LOCKED_BIT,
-						&(grinfo->bb_state));
-}
-
-static ext4_fsblk_t ext4_grp_offs_to_block(struct super_block *sb,
+static inline ext4_fsblk_t ext4_grp_offs_to_block(struct super_block *sb,
 					struct ext4_free_extent *fex)
 {
 	ext4_fsblk_t block;
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
index f2a9cf498ec..734abca25e3 100644
--- a/fs/ext4/migrate.c
+++ b/fs/ext4/migrate.c
@@ -59,7 +59,8 @@ static int finish_range(handle_t *handle, struct inode *inode,
 	/*
 	 * Make sure the credit we accumalated is not really high
 	 */
-	if (needed && handle->h_buffer_credits >= EXT4_RESERVE_TRANS_BLOCKS) {
+	if (needed && ext4_handle_has_enough_credits(handle,
+						EXT4_RESERVE_TRANS_BLOCKS)) {
 		retval = ext4_journal_restart(handle, needed);
 		if (retval)
 			goto err_out;
@@ -229,7 +230,7 @@ static int extend_credit_for_blkdel(handle_t *handle, struct inode *inode)
 {
 	int retval = 0, needed;
 
-	if (handle->h_buffer_credits > EXT4_RESERVE_TRANS_BLOCKS)
+	if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
 		return 0;
 	/*
 	 * We are freeing a blocks. During this we touch
@@ -458,13 +459,13 @@ int ext4_ext_migrate(struct inode *inode)
 	struct list_blocks_struct lb;
 	unsigned long max_entries;
 
-	if (!test_opt(inode->i_sb, EXTENTS))
-		/*
-		 * if mounted with noextents we don't allow the migrate
-		 */
-		return -EINVAL;
-
-	if ((EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+	/*
+	 * If the filesystem does not support extents, or the inode
+	 * already is extent-based, error out.
+	 */
+	if (!EXT4_HAS_INCOMPAT_FEATURE(inode->i_sb,
+				       EXT4_FEATURE_INCOMPAT_EXTENTS) ||
+	    (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
 		return -EINVAL;
 
 	if (S_ISLNK(inode->i_mode) && inode->i_blocks == 0)
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 63adcb79298..fec0b4c2f5f 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -74,10 +74,6 @@ static struct buffer_head *ext4_append(handle_t *handle,
 #define assert(test) J_ASSERT(test)
 #endif
 
-#ifndef swap
-#define swap(x, y) do { typeof(x) z = x; x = y; y = z; } while (0)
-#endif
-
 #ifdef DX_DEBUG
 #define dxtrace(command) command
 #else
@@ -372,6 +368,8 @@ dx_probe(const struct qstr *d_name, struct inode *dir,
 		goto fail;
 	}
 	hinfo->hash_version = root->info.hash_version;
+	if (hinfo->hash_version <= DX_HASH_TEA)
+		hinfo->hash_version += EXT4_SB(dir->i_sb)->s_hash_unsigned;
 	hinfo->seed = EXT4_SB(dir->i_sb)->s_hash_seed;
 	if (d_name)
 		ext4fs_dirhash(d_name->name, d_name->len, hinfo);
@@ -641,6 +639,9 @@ int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash,
 	dir = dir_file->f_path.dentry->d_inode;
 	if (!(EXT4_I(dir)->i_flags & EXT4_INDEX_FL)) {
 		hinfo.hash_version = EXT4_SB(dir->i_sb)->s_def_hash_version;
+		if (hinfo.hash_version <= DX_HASH_TEA)
+			hinfo.hash_version +=
+				EXT4_SB(dir->i_sb)->s_hash_unsigned;
 		hinfo.seed = EXT4_SB(dir->i_sb)->s_hash_seed;
 		count = htree_dirblock_to_tree(dir_file, dir, 0, &hinfo,
 					       start_hash, start_minor_hash);
@@ -806,7 +807,7 @@ static inline int ext4_match (int len, const char * const name,
 static inline int search_dirblock(struct buffer_head *bh,
 				  struct inode *dir,
 				  const struct qstr *d_name,
-				  unsigned long offset,
+				  unsigned int offset,
 				  struct ext4_dir_entry_2 ** res_dir)
 {
 	struct ext4_dir_entry_2 * de;
@@ -1043,11 +1044,11 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, stru
 	bh = ext4_find_entry(dir, &dentry->d_name, &de);
 	inode = NULL;
 	if (bh) {
-		unsigned long ino = le32_to_cpu(de->inode);
+		__u32 ino = le32_to_cpu(de->inode);
 		brelse(bh);
 		if (!ext4_valid_inum(dir->i_sb, ino)) {
 			ext4_error(dir->i_sb, "ext4_lookup",
-				   "bad inode number: %lu", ino);
+				   "bad inode number: %u", ino);
 			return ERR_PTR(-EIO);
 		}
 		inode = ext4_iget(dir->i_sb, ino);
@@ -1060,7 +1061,7 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, stru
 
 struct dentry *ext4_get_parent(struct dentry *child)
 {
-	unsigned long ino;
+	__u32 ino;
 	struct inode *inode;
 	static const struct qstr dotdot = {
 		.name = "..",
@@ -1078,7 +1079,7 @@ struct dentry *ext4_get_parent(struct dentry *child)
 
 	if (!ext4_valid_inum(child->d_inode->i_sb, ino)) {
 		ext4_error(child->d_inode->i_sb, "ext4_get_parent",
-			   "bad inode number: %lu", ino);
+			   "bad inode number: %u", ino);
 		return ERR_PTR(-EIO);
 	}
 
@@ -1166,9 +1167,9 @@ static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
 	u32 hash2;
 	struct dx_map_entry *map;
 	char *data1 = (*bh)->b_data, *data2;
-	unsigned split, move, size, i;
+	unsigned split, move, size;
 	struct ext4_dir_entry_2 *de = NULL, *de2;
-	int	err = 0;
+	int	err = 0, i;
 
 	bh2 = ext4_append (handle, dir, &newblock, &err);
 	if (!(bh2)) {
@@ -1228,10 +1229,10 @@ static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
 		de = de2;
 	}
 	dx_insert_block(frame, hash2 + continued, newblock);
-	err = ext4_journal_dirty_metadata(handle, bh2);
+	err = ext4_handle_dirty_metadata(handle, dir, bh2);
 	if (err)
 		goto journal_error;
-	err = ext4_journal_dirty_metadata(handle, frame->bh);
+	err = ext4_handle_dirty_metadata(handle, dir, frame->bh);
 	if (err)
 		goto journal_error;
 	brelse(bh2);
@@ -1266,7 +1267,7 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
 	struct inode	*dir = dentry->d_parent->d_inode;
 	const char	*name = dentry->d_name.name;
 	int		namelen = dentry->d_name.len;
-	unsigned long	offset = 0;
+	unsigned int	offset = 0;
 	unsigned short	reclen;
 	int		nlen, rlen, err;
 	char		*top;
@@ -1335,8 +1336,8 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
 	ext4_update_dx_flag(dir);
 	dir->i_version++;
 	ext4_mark_inode_dirty(handle, dir);
-	BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
-	err = ext4_journal_dirty_metadata(handle, bh);
+	BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+	err = ext4_handle_dirty_metadata(handle, dir, bh);
 	if (err)
 		ext4_std_error(dir->i_sb, err);
 	brelse(bh);
@@ -1408,6 +1409,8 @@ static int make_indexed_dir(handle_t *handle, struct dentry *dentry,
 
 	/* Initialize as for dx_probe */
 	hinfo.hash_version = root->info.hash_version;
+	if (hinfo.hash_version <= DX_HASH_TEA)
+		hinfo.hash_version += EXT4_SB(dir->i_sb)->s_hash_unsigned;
 	hinfo.seed = EXT4_SB(dir->i_sb)->s_hash_seed;
 	ext4fs_dirhash(name, namelen, &hinfo);
 	frame = frames;
@@ -1437,7 +1440,6 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 			  struct inode *inode)
 {
 	struct inode *dir = dentry->d_parent->d_inode;
-	unsigned long offset;
 	struct buffer_head *bh;
 	struct ext4_dir_entry_2 *de;
 	struct super_block *sb;
@@ -1459,7 +1461,7 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 		ext4_mark_inode_dirty(handle, dir);
 	}
 	blocks = dir->i_size >> sb->s_blocksize_bits;
-	for (block = 0, offset = 0; block < blocks; block++) {
+	for (block = 0; block < blocks; block++) {
 		bh = ext4_bread(handle, dir, block, 0, &retval);
 		if(!bh)
 			return retval;
@@ -1574,7 +1576,7 @@ static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry,
 			dxtrace(dx_show_index("node", frames[1].entries));
 			dxtrace(dx_show_index("node",
 			       ((struct dx_node *) bh2->b_data)->entries));
-			err = ext4_journal_dirty_metadata(handle, bh2);
+			err = ext4_handle_dirty_metadata(handle, inode, bh2);
 			if (err)
 				goto journal_error;
 			brelse (bh2);
@@ -1600,7 +1602,7 @@ static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry,
 			if (err)
 				goto journal_error;
 		}
-		ext4_journal_dirty_metadata(handle, frames[0].bh);
+		ext4_handle_dirty_metadata(handle, inode, frames[0].bh);
 	}
 	de = do_split(handle, dir, &bh, frame, &hinfo, &err);
 	if (!de)
@@ -1646,8 +1648,8 @@ static int ext4_delete_entry(handle_t *handle,
 			else
 				de->inode = 0;
 			dir->i_version++;
-			BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
-			ext4_journal_dirty_metadata(handle, bh);
+			BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+			ext4_handle_dirty_metadata(handle, dir, bh);
 			return 0;
 		}
 		i += ext4_rec_len_from_disk(de->rec_len);
@@ -1693,9 +1695,11 @@ static int ext4_add_nondir(handle_t *handle,
 	if (!err) {
 		ext4_mark_inode_dirty(handle, inode);
 		d_instantiate(dentry, inode);
+		unlock_new_inode(inode);
 		return 0;
 	}
 	drop_nlink(inode);
+	unlock_new_inode(inode);
 	iput(inode);
 	return err;
 }
@@ -1723,7 +1727,7 @@ retry:
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	inode = ext4_new_inode (handle, dir, mode);
 	err = PTR_ERR(inode);
@@ -1757,7 +1761,7 @@ retry:
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	inode = ext4_new_inode(handle, dir, mode);
 	err = PTR_ERR(inode);
@@ -1793,7 +1797,7 @@ retry:
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	inode = ext4_new_inode(handle, dir, S_IFDIR | mode);
 	err = PTR_ERR(inode);
@@ -1822,14 +1826,15 @@ retry:
 	strcpy(de->name, "..");
 	ext4_set_de_type(dir->i_sb, de, S_IFDIR);
 	inode->i_nlink = 2;
-	BUFFER_TRACE(dir_block, "call ext4_journal_dirty_metadata");
-	ext4_journal_dirty_metadata(handle, dir_block);
+	BUFFER_TRACE(dir_block, "call ext4_handle_dirty_metadata");
+	ext4_handle_dirty_metadata(handle, dir, dir_block);
 	brelse(dir_block);
 	ext4_mark_inode_dirty(handle, inode);
 	err = ext4_add_entry(handle, dentry, inode);
 	if (err) {
 out_clear_inode:
 		clear_nlink(inode);
+		unlock_new_inode(inode);
 		ext4_mark_inode_dirty(handle, inode);
 		iput(inode);
 		goto out_stop;
@@ -1838,6 +1843,7 @@ out_clear_inode:
 	ext4_update_dx_flag(dir);
 	ext4_mark_inode_dirty(handle, dir);
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 out_stop:
 	ext4_journal_stop(handle);
 	if (err == -ENOSPC && ext4_should_retry_alloc(dir->i_sb, &retries))
@@ -1850,7 +1856,7 @@ out_stop:
  */
 static int empty_dir(struct inode *inode)
 {
-	unsigned long offset;
+	unsigned int offset;
 	struct buffer_head *bh;
 	struct ext4_dir_entry_2 *de, *de1;
 	struct super_block *sb;
@@ -1895,7 +1901,7 @@ static int empty_dir(struct inode *inode)
 				if (err)
 					ext4_error(sb, __func__,
 						   "error %d reading directory"
-						   " #%lu offset %lu",
+						   " #%lu offset %u",
 						   err, inode->i_ino, offset);
 				offset += sb->s_blocksize;
 				continue;
@@ -1933,6 +1939,9 @@ int ext4_orphan_add(handle_t *handle, struct inode *inode)
 	struct ext4_iloc iloc;
 	int err = 0, rc;
 
+	if (!ext4_handle_valid(handle))
+		return 0;
+
 	lock_super(sb);
 	if (!list_empty(&EXT4_I(inode)->i_orphan))
 		goto out_unlock;
@@ -1961,7 +1970,7 @@ int ext4_orphan_add(handle_t *handle, struct inode *inode)
 	/* Insert this inode at the head of the on-disk orphan list... */
 	NEXT_ORPHAN(inode) = le32_to_cpu(EXT4_SB(sb)->s_es->s_last_orphan);
 	EXT4_SB(sb)->s_es->s_last_orphan = cpu_to_le32(inode->i_ino);
-	err = ext4_journal_dirty_metadata(handle, EXT4_SB(sb)->s_sbh);
+	err = ext4_handle_dirty_metadata(handle, inode, EXT4_SB(sb)->s_sbh);
 	rc = ext4_mark_iloc_dirty(handle, inode, &iloc);
 	if (!err)
 		err = rc;
@@ -1995,10 +2004,13 @@ int ext4_orphan_del(handle_t *handle, struct inode *inode)
 	struct list_head *prev;
 	struct ext4_inode_info *ei = EXT4_I(inode);
 	struct ext4_sb_info *sbi;
-	unsigned long ino_next;
+	__u32 ino_next;
 	struct ext4_iloc iloc;
 	int err = 0;
 
+	if (!ext4_handle_valid(handle))
+		return 0;
+
 	lock_super(inode->i_sb);
 	if (list_empty(&ei->i_orphan)) {
 		unlock_super(inode->i_sb);
@@ -2017,7 +2029,7 @@ int ext4_orphan_del(handle_t *handle, struct inode *inode)
 	 * transaction handle with which to update the orphan list on
 	 * disk, but we still need to remove the inode from the linked
 	 * list in memory. */
-	if (!handle)
+	if (sbi->s_journal && !handle)
 		goto out;
 
 	err = ext4_reserve_inode_write(handle, inode, &iloc);
@@ -2025,19 +2037,19 @@ int ext4_orphan_del(handle_t *handle, struct inode *inode)
 		goto out_err;
 
 	if (prev == &sbi->s_orphan) {
-		jbd_debug(4, "superblock will point to %lu\n", ino_next);
+		jbd_debug(4, "superblock will point to %u\n", ino_next);
 		BUFFER_TRACE(sbi->s_sbh, "get_write_access");
 		err = ext4_journal_get_write_access(handle, sbi->s_sbh);
 		if (err)
 			goto out_brelse;
 		sbi->s_es->s_last_orphan = cpu_to_le32(ino_next);
-		err = ext4_journal_dirty_metadata(handle, sbi->s_sbh);
+		err = ext4_handle_dirty_metadata(handle, inode, sbi->s_sbh);
 	} else {
 		struct ext4_iloc iloc2;
 		struct inode *i_prev =
 			&list_entry(prev, struct ext4_inode_info, i_orphan)->vfs_inode;
 
-		jbd_debug(4, "orphan inode %lu will point to %lu\n",
+		jbd_debug(4, "orphan inode %lu will point to %u\n",
 			  i_prev->i_ino, ino_next);
 		err = ext4_reserve_inode_write(handle, i_prev, &iloc2);
 		if (err)
@@ -2082,7 +2094,7 @@ static int ext4_rmdir(struct inode *dir, struct dentry *dentry)
 		goto end_rmdir;
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	inode = dentry->d_inode;
 
@@ -2136,7 +2148,7 @@ static int ext4_unlink(struct inode *dir, struct dentry *dentry)
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	retval = -ENOENT;
 	bh = ext4_find_entry(dir, &dentry->d_name, &de);
@@ -2193,7 +2205,7 @@ retry:
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	inode = ext4_new_inode(handle, dir, S_IFLNK|S_IRWXUGO);
 	err = PTR_ERR(inode);
@@ -2208,10 +2220,10 @@ retry:
 		 * We have a transaction open.  All is sweetness.  It also sets
 		 * i_size in generic_commit_write().
 		 */
-		err = __page_symlink(inode, symname, l,
-				mapping_gfp_mask(inode->i_mapping) & ~__GFP_FS);
+		err = __page_symlink(inode, symname, l, 1);
 		if (err) {
 			clear_nlink(inode);
+			unlock_new_inode(inode);
 			ext4_mark_inode_dirty(handle, inode);
 			iput(inode);
 			goto out_stop;
@@ -2256,13 +2268,20 @@ retry:
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	inode->i_ctime = ext4_current_time(inode);
 	ext4_inc_count(handle, inode);
 	atomic_inc(&inode->i_count);
 
-	err = ext4_add_nondir(handle, dentry, inode);
+	err = ext4_add_entry(handle, dentry, inode);
+	if (!err) {
+		ext4_mark_inode_dirty(handle, inode);
+		d_instantiate(dentry, inode);
+	} else {
+		drop_nlink(inode);
+		iput(inode);
+	}
 	ext4_journal_stop(handle);
 	if (err == -ENOSPC && ext4_should_retry_alloc(dir->i_sb, &retries))
 		goto retry;
@@ -2298,7 +2317,7 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 		return PTR_ERR(handle);
 
 	if (IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir))
-		handle->h_sync = 1;
+		ext4_handle_sync(handle);
 
 	old_bh = ext4_find_entry(old_dir, &old_dentry->d_name, &old_de);
 	/*
@@ -2352,8 +2371,8 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 		new_dir->i_ctime = new_dir->i_mtime =
 					ext4_current_time(new_dir);
 		ext4_mark_inode_dirty(handle, new_dir);
-		BUFFER_TRACE(new_bh, "call ext4_journal_dirty_metadata");
-		ext4_journal_dirty_metadata(handle, new_bh);
+		BUFFER_TRACE(new_bh, "call ext4_handle_dirty_metadata");
+		ext4_handle_dirty_metadata(handle, new_dir, new_bh);
 		brelse(new_bh);
 		new_bh = NULL;
 	}
@@ -2403,8 +2422,8 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 		BUFFER_TRACE(dir_bh, "get_write_access");
 		ext4_journal_get_write_access(handle, dir_bh);
 		PARENT_INO(dir_bh->b_data) = cpu_to_le32(new_dir->i_ino);
-		BUFFER_TRACE(dir_bh, "call ext4_journal_dirty_metadata");
-		ext4_journal_dirty_metadata(handle, dir_bh);
+		BUFFER_TRACE(dir_bh, "call ext4_handle_dirty_metadata");
+		ext4_handle_dirty_metadata(handle, old_dir, dir_bh);
 		ext4_dec_count(handle, old_dir);
 		if (new_inode) {
 			/* checked empty_dir above, can't have another parent,
diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index b6ec1843a01..c328be5d688 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -50,7 +50,7 @@ static int verify_group_input(struct super_block *sb,
 	ext4_get_group_no_and_offset(sb, start, NULL, &offset);
 	if (group != sbi->s_groups_count)
 		ext4_warning(sb, __func__,
-			     "Cannot add at group %u (only %lu groups)",
+			     "Cannot add at group %u (only %u groups)",
 			     input->group, sbi->s_groups_count);
 	else if (offset != 0)
 			ext4_warning(sb, __func__, "Last group not full");
@@ -149,7 +149,7 @@ static int extend_or_restart_transaction(handle_t *handle, int thresh,
 {
 	int err;
 
-	if (handle->h_buffer_credits >= thresh)
+	if (ext4_handle_has_enough_credits(handle, thresh))
 		return 0;
 
 	err = ext4_journal_extend(handle, EXT4_MAX_TRANS_DATA);
@@ -232,7 +232,7 @@ static int setup_new_group_blocks(struct super_block *sb,
 		memcpy(gdb->b_data, sbi->s_group_desc[i]->b_data, gdb->b_size);
 		set_buffer_uptodate(gdb);
 		unlock_buffer(gdb);
-		ext4_journal_dirty_metadata(handle, gdb);
+		ext4_handle_dirty_metadata(handle, NULL, gdb);
 		ext4_set_bit(bit, bh->b_data);
 		brelse(gdb);
 	}
@@ -251,7 +251,7 @@ static int setup_new_group_blocks(struct super_block *sb,
 			err = PTR_ERR(bh);
 			goto exit_bh;
 		}
-		ext4_journal_dirty_metadata(handle, gdb);
+		ext4_handle_dirty_metadata(handle, NULL, gdb);
 		ext4_set_bit(bit, bh->b_data);
 		brelse(gdb);
 	}
@@ -276,7 +276,7 @@ static int setup_new_group_blocks(struct super_block *sb,
 			err = PTR_ERR(it);
 			goto exit_bh;
 		}
-		ext4_journal_dirty_metadata(handle, it);
+		ext4_handle_dirty_metadata(handle, NULL, it);
 		brelse(it);
 		ext4_set_bit(bit, bh->b_data);
 	}
@@ -284,11 +284,9 @@ static int setup_new_group_blocks(struct super_block *sb,
 	if ((err = extend_or_restart_transaction(handle, 2, bh)))
 		goto exit_bh;
 
-	mark_bitmap_end(input->blocks_count, EXT4_BLOCKS_PER_GROUP(sb),
-			bh->b_data);
-	ext4_journal_dirty_metadata(handle, bh);
+	mark_bitmap_end(input->blocks_count, sb->s_blocksize * 8, bh->b_data);
+	ext4_handle_dirty_metadata(handle, NULL, bh);
 	brelse(bh);
-
 	/* Mark unused entries in inode bitmap used */
 	ext4_debug("clear inode bitmap %#04llx (+%llu)\n",
 		   input->inode_bitmap, input->inode_bitmap - start);
@@ -297,9 +295,9 @@ static int setup_new_group_blocks(struct super_block *sb,
 		goto exit_journal;
 	}
 
-	mark_bitmap_end(EXT4_INODES_PER_GROUP(sb), EXT4_BLOCKS_PER_GROUP(sb),
+	mark_bitmap_end(EXT4_INODES_PER_GROUP(sb), sb->s_blocksize * 8,
 			bh->b_data);
-	ext4_journal_dirty_metadata(handle, bh);
+	ext4_handle_dirty_metadata(handle, NULL, bh);
 exit_bh:
 	brelse(bh);
 
@@ -486,12 +484,12 @@ static int add_new_gdb(handle_t *handle, struct inode *inode,
 	 * reserved inode, and will become GDT blocks (primary and backup).
 	 */
 	data[gdb_num % EXT4_ADDR_PER_BLOCK(sb)] = 0;
-	ext4_journal_dirty_metadata(handle, dind);
+	ext4_handle_dirty_metadata(handle, NULL, dind);
 	brelse(dind);
 	inode->i_blocks -= (gdbackups + 1) * sb->s_blocksize >> 9;
 	ext4_mark_iloc_dirty(handle, inode, &iloc);
 	memset((*primary)->b_data, 0, sb->s_blocksize);
-	ext4_journal_dirty_metadata(handle, *primary);
+	ext4_handle_dirty_metadata(handle, NULL, *primary);
 
 	o_group_desc = EXT4_SB(sb)->s_group_desc;
 	memcpy(n_group_desc, o_group_desc,
@@ -502,7 +500,7 @@ static int add_new_gdb(handle_t *handle, struct inode *inode,
 	kfree(o_group_desc);
 
 	le16_add_cpu(&es->s_reserved_gdt_blocks, -1);
-	ext4_journal_dirty_metadata(handle, EXT4_SB(sb)->s_sbh);
+	ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh);
 
 	return 0;
 
@@ -618,7 +616,7 @@ static int reserve_backup_gdb(handle_t *handle, struct inode *inode,
 		       primary[i]->b_blocknr, gdbackups,
 		       blk + primary[i]->b_blocknr); */
 		data[gdbackups] = cpu_to_le32(blk + primary[i]->b_blocknr);
-		err2 = ext4_journal_dirty_metadata(handle, primary[i]);
+		err2 = ext4_handle_dirty_metadata(handle, NULL, primary[i]);
 		if (!err)
 			err = err2;
 	}
@@ -676,7 +674,8 @@ static void update_backups(struct super_block *sb,
 		struct buffer_head *bh;
 
 		/* Out of journal space, and can't get more - abort - so sad */
-		if (handle->h_buffer_credits == 0 &&
+		if (ext4_handle_valid(handle) &&
+		    handle->h_buffer_credits == 0 &&
 		    ext4_journal_extend(handle, EXT4_MAX_TRANS_DATA) &&
 		    (err = ext4_journal_restart(handle, EXT4_MAX_TRANS_DATA)))
 			break;
@@ -696,7 +695,7 @@ static void update_backups(struct super_block *sb,
 			memset(bh->b_data + size, 0, rest);
 		set_buffer_uptodate(bh);
 		unlock_buffer(bh);
-		ext4_journal_dirty_metadata(handle, bh);
+		ext4_handle_dirty_metadata(handle, NULL, bh);
 		brelse(bh);
 	}
 	if ((err2 = ext4_journal_stop(handle)) && !err)
@@ -715,7 +714,7 @@ static void update_backups(struct super_block *sb,
 exit_err:
 	if (err) {
 		ext4_warning(sb, __func__,
-			     "can't update backup for group %lu (err %d), "
+			     "can't update backup for group %u (err %d), "
 			     "forcing fsck on next reboot", group, err);
 		sbi->s_mount_state &= ~EXT4_VALID_FS;
 		sbi->s_es->s_state &= cpu_to_le16(~EXT4_VALID_FS);
@@ -747,6 +746,7 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 	struct inode *inode = NULL;
 	handle_t *handle;
 	int gdb_off, gdb_num;
+	int num_grp_locked = 0;
 	int err, err2;
 
 	gdb_num = input->group / EXT4_DESC_PER_BLOCK(sb);
@@ -761,13 +761,13 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 
 	if (ext4_blocks_count(es) + input->blocks_count <
 	    ext4_blocks_count(es)) {
-		ext4_warning(sb, __func__, "blocks_count overflow\n");
+		ext4_warning(sb, __func__, "blocks_count overflow");
 		return -EINVAL;
 	}
 
 	if (le32_to_cpu(es->s_inodes_count) + EXT4_INODES_PER_GROUP(sb) <
 	    le32_to_cpu(es->s_inodes_count)) {
-		ext4_warning(sb, __func__, "inodes_count overflow\n");
+		ext4_warning(sb, __func__, "inodes_count overflow");
 		return -EINVAL;
 	}
 
@@ -787,6 +787,7 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 		}
 	}
 
+
 	if ((err = verify_group_input(sb, input)))
 		goto exit_put;
 
@@ -855,6 +856,7 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
          * using the new disk blocks.
          */
 
+	num_grp_locked = ext4_mb_get_buddy_cache_lock(sb, input->group);
 	/* Update group descriptor block for new group */
 	gdp = (struct ext4_group_desc *)((char *)primary->b_data +
 					 gdb_off * EXT4_DESC_SIZE(sb));
@@ -862,17 +864,20 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 	ext4_block_bitmap_set(sb, gdp, input->block_bitmap); /* LV FIXME */
 	ext4_inode_bitmap_set(sb, gdp, input->inode_bitmap); /* LV FIXME */
 	ext4_inode_table_set(sb, gdp, input->inode_table); /* LV FIXME */
-	gdp->bg_free_blocks_count = cpu_to_le16(input->free_blocks_count);
-	gdp->bg_free_inodes_count = cpu_to_le16(EXT4_INODES_PER_GROUP(sb));
+	ext4_free_blks_set(sb, gdp, input->free_blocks_count);
+	ext4_free_inodes_set(sb, gdp, EXT4_INODES_PER_GROUP(sb));
+	gdp->bg_flags |= cpu_to_le16(EXT4_BG_INODE_ZEROED);
 	gdp->bg_checksum = ext4_group_desc_csum(sbi, input->group, gdp);
 
 	/*
 	 * We can allocate memory for mb_alloc based on the new group
 	 * descriptor
 	 */
-	err = ext4_mb_add_more_groupinfo(sb, input->group, gdp);
-	if (err)
+	err = ext4_mb_add_groupinfo(sb, input->group, gdp);
+	if (err) {
+		ext4_mb_put_buddy_cache_lock(sb, input->group, num_grp_locked);
 		goto exit_journal;
+	}
 
 	/*
 	 * Make the new blocks and inodes valid next.  We do this before
@@ -914,8 +919,9 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 
 	/* Update the global fs size fields */
 	sbi->s_groups_count++;
+	ext4_mb_put_buddy_cache_lock(sb, input->group, num_grp_locked);
 
-	ext4_journal_dirty_metadata(handle, primary);
+	ext4_handle_dirty_metadata(handle, NULL, primary);
 
 	/* Update the reserved block counts only once the new group is
 	 * active. */
@@ -937,7 +943,7 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 			EXT4_INODES_PER_GROUP(sb);
 	}
 
-	ext4_journal_dirty_metadata(handle, sbi->s_sbh);
+	ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
 	sb->s_dirt = 1;
 
 exit_journal:
@@ -975,9 +981,7 @@ int ext4_group_extend(struct super_block *sb, struct ext4_super_block *es,
 	struct buffer_head *bh;
 	handle_t *handle;
 	int err;
-	unsigned long freed_blocks;
 	ext4_group_t group;
-	struct ext4_group_info *grp;
 
 	/* We don't need to worry about locking wrt other resizers just
 	 * yet: we're going to revalidate es->s_blocks_count after
@@ -997,8 +1001,7 @@ int ext4_group_extend(struct super_block *sb, struct ext4_super_block *es,
 			" too large to resize to %llu blocks safely\n",
 			sb->s_id, n_blocks_count);
 		if (sizeof(sector_t) < 8)
-			ext4_warning(sb, __func__,
-			"CONFIG_LBD not enabled\n");
+			ext4_warning(sb, __func__, "CONFIG_LBD not enabled");
 		return -EINVAL;
 	}
 
@@ -1071,62 +1074,18 @@ int ext4_group_extend(struct super_block *sb, struct ext4_super_block *es,
 		goto exit_put;
 	}
 	ext4_blocks_count_set(es, o_blocks_count + add);
-	ext4_journal_dirty_metadata(handle, EXT4_SB(sb)->s_sbh);
+	ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh);
 	sb->s_dirt = 1;
 	unlock_super(sb);
 	ext4_debug("freeing blocks %llu through %llu\n", o_blocks_count,
 		   o_blocks_count + add);
-	ext4_free_blocks_sb(handle, sb, o_blocks_count, add, &freed_blocks);
+	/* We add the blocks to the bitmap and set the group need init bit */
+	ext4_add_groupblocks(handle, sb, o_blocks_count, add);
 	ext4_debug("freed blocks %llu through %llu\n", o_blocks_count,
 		   o_blocks_count + add);
 	if ((err = ext4_journal_stop(handle)))
 		goto exit_put;
 
-	/*
-	 * Mark mballoc pages as not up to date so that they will be updated
-	 * next time they are loaded by ext4_mb_load_buddy.
-	 *
-	 * XXX Bad, Bad, BAD!!!  We should not be overloading the
-	 * Uptodate flag, particularly on thte bitmap bh, as way of
-	 * hinting to ext4_mb_load_buddy() that it needs to be
-	 * overloaded.  A user could take a LVM snapshot, then do an
-	 * on-line fsck, and clear the uptodate flag, and this would
-	 * not be a bug in userspace, but a bug in the kernel.  FIXME!!!
-	 */
-	{
-		struct ext4_sb_info *sbi = EXT4_SB(sb);
-		struct inode *inode = sbi->s_buddy_cache;
-		int blocks_per_page;
-		int block;
-		int pnum;
-		struct page *page;
-
-		/* Set buddy page as not up to date */
-		blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
-		block = group * 2;
-		pnum = block / blocks_per_page;
-		page = find_get_page(inode->i_mapping, pnum);
-		if (page != NULL) {
-			ClearPageUptodate(page);
-			page_cache_release(page);
-		}
-
-		/* Set bitmap page as not up to date */
-		block++;
-		pnum = block / blocks_per_page;
-		page = find_get_page(inode->i_mapping, pnum);
-		if (page != NULL) {
-			ClearPageUptodate(page);
-			page_cache_release(page);
-		}
-
-		/* Get the info on the last group */
-		grp = ext4_get_group_info(sb, group);
-
-		/* Update free blocks in group info */
-		ext4_mb_update_group_info(grp, add);
-	}
-
 	if (test_opt(sb, DEBUG))
 		printk(KERN_DEBUG "EXT4-fs: extended group to %llu blocks\n",
 		       ext4_blocks_count(es));
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index e4a241c65db..e5f06a5f045 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -51,9 +51,7 @@ struct proc_dir_entry *ext4_proc_root;
 
 static int ext4_load_journal(struct super_block *, struct ext4_super_block *,
 			     unsigned long journal_devnum);
-static int ext4_create_journal(struct super_block *, struct ext4_super_block *,
-			       unsigned int);
-static void ext4_commit_super(struct super_block *sb,
+static int ext4_commit_super(struct super_block *sb,
 			      struct ext4_super_block *es, int sync);
 static void ext4_mark_recovery_complete(struct super_block *sb,
 					struct ext4_super_block *es);
@@ -64,9 +62,9 @@ static const char *ext4_decode_error(struct super_block *sb, int errno,
 				     char nbuf[16]);
 static int ext4_remount(struct super_block *sb, int *flags, char *data);
 static int ext4_statfs(struct dentry *dentry, struct kstatfs *buf);
-static void ext4_unlockfs(struct super_block *sb);
+static int ext4_unfreeze(struct super_block *sb);
 static void ext4_write_super(struct super_block *sb);
-static void ext4_write_super_lockfs(struct super_block *sb);
+static int ext4_freeze(struct super_block *sb);
 
 
 ext4_fsblk_t ext4_block_bitmap(struct super_block *sb,
@@ -93,6 +91,38 @@ ext4_fsblk_t ext4_inode_table(struct super_block *sb,
 		(ext4_fsblk_t)le32_to_cpu(bg->bg_inode_table_hi) << 32 : 0);
 }
 
+__u32 ext4_free_blks_count(struct super_block *sb,
+			      struct ext4_group_desc *bg)
+{
+	return le16_to_cpu(bg->bg_free_blocks_count_lo) |
+		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
+		(__u32)le16_to_cpu(bg->bg_free_blocks_count_hi) << 16 : 0);
+}
+
+__u32 ext4_free_inodes_count(struct super_block *sb,
+			      struct ext4_group_desc *bg)
+{
+	return le16_to_cpu(bg->bg_free_inodes_count_lo) |
+		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
+		(__u32)le16_to_cpu(bg->bg_free_inodes_count_hi) << 16 : 0);
+}
+
+__u32 ext4_used_dirs_count(struct super_block *sb,
+			      struct ext4_group_desc *bg)
+{
+	return le16_to_cpu(bg->bg_used_dirs_count_lo) |
+		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
+		(__u32)le16_to_cpu(bg->bg_used_dirs_count_hi) << 16 : 0);
+}
+
+__u32 ext4_itable_unused_count(struct super_block *sb,
+			      struct ext4_group_desc *bg)
+{
+	return le16_to_cpu(bg->bg_itable_unused_lo) |
+		(EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT ?
+		(__u32)le16_to_cpu(bg->bg_itable_unused_hi) << 16 : 0);
+}
+
 void ext4_block_bitmap_set(struct super_block *sb,
 			   struct ext4_group_desc *bg, ext4_fsblk_t blk)
 {
@@ -117,6 +147,38 @@ void ext4_inode_table_set(struct super_block *sb,
 		bg->bg_inode_table_hi = cpu_to_le32(blk >> 32);
 }
 
+void ext4_free_blks_set(struct super_block *sb,
+			  struct ext4_group_desc *bg, __u32 count)
+{
+	bg->bg_free_blocks_count_lo = cpu_to_le16((__u16)count);
+	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
+		bg->bg_free_blocks_count_hi = cpu_to_le16(count >> 16);
+}
+
+void ext4_free_inodes_set(struct super_block *sb,
+			  struct ext4_group_desc *bg, __u32 count)
+{
+	bg->bg_free_inodes_count_lo = cpu_to_le16((__u16)count);
+	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
+		bg->bg_free_inodes_count_hi = cpu_to_le16(count >> 16);
+}
+
+void ext4_used_dirs_set(struct super_block *sb,
+			  struct ext4_group_desc *bg, __u32 count)
+{
+	bg->bg_used_dirs_count_lo = cpu_to_le16((__u16)count);
+	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
+		bg->bg_used_dirs_count_hi = cpu_to_le16(count >> 16);
+}
+
+void ext4_itable_unused_set(struct super_block *sb,
+			  struct ext4_group_desc *bg, __u32 count)
+{
+	bg->bg_itable_unused_lo = cpu_to_le16((__u16)count);
+	if (EXT4_DESC_SIZE(sb) >= EXT4_MIN_DESC_SIZE_64BIT)
+		bg->bg_itable_unused_hi = cpu_to_le16(count >> 16);
+}
+
 /*
  * Wrappers for jbd2_journal_start/end.
  *
@@ -136,13 +198,19 @@ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks)
 	 * backs (eg. EIO in the commit thread), then we still need to
 	 * take the FS itself readonly cleanly. */
 	journal = EXT4_SB(sb)->s_journal;
-	if (is_journal_aborted(journal)) {
-		ext4_abort(sb, __func__,
-			   "Detected aborted journal");
-		return ERR_PTR(-EROFS);
+	if (journal) {
+		if (is_journal_aborted(journal)) {
+			ext4_abort(sb, __func__,
+				   "Detected aborted journal");
+			return ERR_PTR(-EROFS);
+		}
+		return jbd2_journal_start(journal, nblocks);
 	}
-
-	return jbd2_journal_start(journal, nblocks);
+	/*
+	 * We're not journaling, return the appropriate indication.
+	 */
+	current->journal_info = EXT4_NOJOURNAL_HANDLE;
+	return current->journal_info;
 }
 
 /*
@@ -157,6 +225,14 @@ int __ext4_journal_stop(const char *where, handle_t *handle)
 	int err;
 	int rc;
 
+	if (!ext4_handle_valid(handle)) {
+		/*
+		 * Do this here since we don't call jbd2_journal_stop() in
+		 * no-journal mode.
+		 */
+		current->journal_info = NULL;
+		return 0;
+	}
 	sb = handle->h_transaction->t_journal->j_private;
 	err = handle->h_err;
 	rc = jbd2_journal_stop(handle);
@@ -174,6 +250,8 @@ void ext4_journal_abort_handle(const char *caller, const char *err_fn,
 	char nbuf[16];
 	const char *errstr = ext4_decode_error(NULL, err, nbuf);
 
+	BUG_ON(!ext4_handle_valid(handle));
+
 	if (bh)
 		BUFFER_TRACE(bh, "abort");
 
@@ -350,6 +428,44 @@ void ext4_warning(struct super_block *sb, const char *function,
 	va_end(args);
 }
 
+void ext4_grp_locked_error(struct super_block *sb, ext4_group_t grp,
+				const char *function, const char *fmt, ...)
+__releases(bitlock)
+__acquires(bitlock)
+{
+	va_list args;
+	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
+
+	va_start(args, fmt);
+	printk(KERN_CRIT "EXT4-fs error (device %s): %s: ", sb->s_id, function);
+	vprintk(fmt, args);
+	printk("\n");
+	va_end(args);
+
+	if (test_opt(sb, ERRORS_CONT)) {
+		EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
+		es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
+		ext4_commit_super(sb, es, 0);
+		return;
+	}
+	ext4_unlock_group(sb, grp);
+	ext4_handle_error(sb);
+	/*
+	 * We only get here in the ERRORS_RO case; relocking the group
+	 * may be dangerous, but nothing bad will happen since the
+	 * filesystem will have already been marked read/only and the
+	 * journal has been aborted.  We return 1 as a hint to callers
+	 * who might what to use the return value from
+	 * ext4_grp_locked_error() to distinguish beween the
+	 * ERRORS_CONT and ERRORS_RO case, and perhaps return more
+	 * aggressively from the ext4 function in question, with a
+	 * more appropriate error code.
+	 */
+	ext4_lock_group(sb, grp);
+	return;
+}
+
+
 void ext4_update_dynamic_rev(struct super_block *sb)
 {
 	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
@@ -389,7 +505,7 @@ static struct block_device *ext4_blkdev_get(dev_t dev)
 	return bdev;
 
 fail:
-	printk(KERN_ERR "EXT4: failed to open journal device %s: %ld\n",
+	printk(KERN_ERR "EXT4-fs: failed to open journal device %s: %ld\n",
 			__bdevname(dev, b), PTR_ERR(bdev));
 	return NULL;
 }
@@ -448,11 +564,13 @@ static void ext4_put_super(struct super_block *sb)
 	ext4_mb_release(sb);
 	ext4_ext_release(sb);
 	ext4_xattr_put_super(sb);
-	err = jbd2_journal_destroy(sbi->s_journal);
-	sbi->s_journal = NULL;
-	if (err < 0)
-		ext4_abort(sb, __func__, "Couldn't clean up the journal");
-
+	if (sbi->s_journal) {
+		err = jbd2_journal_destroy(sbi->s_journal);
+		sbi->s_journal = NULL;
+		if (err < 0)
+			ext4_abort(sb, __func__,
+				   "Couldn't clean up the journal");
+	}
 	if (!(sb->s_flags & MS_RDONLY)) {
 		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 		es->s_state = cpu_to_le16(sbi->s_mount_state);
@@ -522,6 +640,11 @@ static struct inode *ext4_alloc_inode(struct super_block *sb)
 	memset(&ei->i_cached_extent, 0, sizeof(struct ext4_ext_cache));
 	INIT_LIST_HEAD(&ei->i_prealloc_list);
 	spin_lock_init(&ei->i_prealloc_lock);
+	/*
+	 * Note:  We can be called before EXT4_SB(sb)->s_journal is set,
+	 * therefore it can be null here.  Don't check it, just initialize
+	 * jinode.
+	 */
 	jbd2_journal_init_jbd_inode(&ei->jinode, &ei->vfs_inode);
 	ei->i_reserved_data_blocks = 0;
 	ei->i_reserved_meta_blocks = 0;
@@ -588,7 +711,8 @@ static void ext4_clear_inode(struct inode *inode)
 	}
 #endif
 	ext4_discard_preallocations(inode);
-	jbd2_journal_release_jbd_inode(EXT4_SB(inode->i_sb)->s_journal,
+	if (EXT4_JOURNAL(inode))
+		jbd2_journal_release_jbd_inode(EXT4_SB(inode->i_sb)->s_journal,
 				       &EXT4_I(inode)->jinode);
 }
 
@@ -681,10 +805,19 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs)
 #endif
 	if (!test_opt(sb, RESERVATION))
 		seq_puts(seq, ",noreservation");
-	if (sbi->s_commit_interval) {
+	if (sbi->s_commit_interval != JBD2_DEFAULT_MAX_COMMIT_AGE*HZ) {
 		seq_printf(seq, ",commit=%u",
 			   (unsigned) (sbi->s_commit_interval / HZ));
 	}
+	if (sbi->s_min_batch_time != EXT4_DEF_MIN_BATCH_TIME) {
+		seq_printf(seq, ",min_batch_time=%u",
+			   (unsigned) sbi->s_min_batch_time);
+	}
+	if (sbi->s_max_batch_time != EXT4_DEF_MAX_BATCH_TIME) {
+		seq_printf(seq, ",max_batch_time=%u",
+			   (unsigned) sbi->s_min_batch_time);
+	}
+
 	/*
 	 * We're changing the default of barrier mount option, so
 	 * let's always display its mount state so it's clear what its
@@ -696,8 +829,6 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs)
 		seq_puts(seq, ",journal_async_commit");
 	if (test_opt(sb, NOBH))
 		seq_puts(seq, ",nobh");
-	if (!test_opt(sb, EXTENTS))
-		seq_puts(seq, ",noextents");
 	if (test_opt(sb, I_VERSION))
 		seq_puts(seq, ",i_version");
 	if (!test_opt(sb, DELALLOC))
@@ -772,6 +903,25 @@ static struct dentry *ext4_fh_to_parent(struct super_block *sb, struct fid *fid,
 				    ext4_nfs_get_inode);
 }
 
+/*
+ * Try to release metadata pages (indirect blocks, directories) which are
+ * mapped via the block device.  Since these pages could have journal heads
+ * which would prevent try_to_free_buffers() from freeing them, we must use
+ * jbd2 layer's try_to_free_buffers() function to release them.
+ */
+static int bdev_try_to_free_page(struct super_block *sb, struct page *page, gfp_t wait)
+{
+	journal_t *journal = EXT4_SB(sb)->s_journal;
+
+	WARN_ON(PageChecked(page));
+	if (!page_has_buffers(page))
+		return 0;
+	if (journal)
+		return jbd2_journal_try_to_free_buffers(journal, page,
+							wait & ~__GFP_WAIT);
+	return try_to_free_buffers(page);
+}
+
 #ifdef CONFIG_QUOTA
 #define QTYPE2NAME(t) ((t) == USRQUOTA ? "user" : "group")
 #define QTYPE2MOPT(on, t) ((t) == USRQUOTA?((on)##USRJQUOTA):((on)##GRPJQUOTA))
@@ -803,7 +953,9 @@ static struct dquot_operations ext4_quota_operations = {
 	.acquire_dquot	= ext4_acquire_dquot,
 	.release_dquot	= ext4_release_dquot,
 	.mark_dirty	= ext4_mark_dquot_dirty,
-	.write_info	= ext4_write_info
+	.write_info	= ext4_write_info,
+	.alloc_dquot	= dquot_alloc,
+	.destroy_dquot	= dquot_destroy,
 };
 
 static struct quotactl_ops ext4_qctl_operations = {
@@ -826,8 +978,8 @@ static const struct super_operations ext4_sops = {
 	.put_super	= ext4_put_super,
 	.write_super	= ext4_write_super,
 	.sync_fs	= ext4_sync_fs,
-	.write_super_lockfs = ext4_write_super_lockfs,
-	.unlockfs	= ext4_unlockfs,
+	.freeze_fs	= ext4_freeze,
+	.unfreeze_fs	= ext4_unfreeze,
 	.statfs		= ext4_statfs,
 	.remount_fs	= ext4_remount,
 	.clear_inode	= ext4_clear_inode,
@@ -836,6 +988,7 @@ static const struct super_operations ext4_sops = {
 	.quota_read	= ext4_quota_read,
 	.quota_write	= ext4_quota_write,
 #endif
+	.bdev_try_to_free_page = bdev_try_to_free_page,
 };
 
 static const struct export_operations ext4_export_ops = {
@@ -850,16 +1003,17 @@ enum {
 	Opt_nouid32, Opt_debug, Opt_oldalloc, Opt_orlov,
 	Opt_user_xattr, Opt_nouser_xattr, Opt_acl, Opt_noacl,
 	Opt_reservation, Opt_noreservation, Opt_noload, Opt_nobh, Opt_bh,
-	Opt_commit, Opt_journal_update, Opt_journal_inum, Opt_journal_dev,
+	Opt_commit, Opt_min_batch_time, Opt_max_batch_time,
+	Opt_journal_update, Opt_journal_dev,
 	Opt_journal_checksum, Opt_journal_async_commit,
 	Opt_abort, Opt_data_journal, Opt_data_ordered, Opt_data_writeback,
 	Opt_data_err_abort, Opt_data_err_ignore,
 	Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
 	Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_quota, Opt_noquota,
 	Opt_ignore, Opt_barrier, Opt_err, Opt_resize, Opt_usrquota,
-	Opt_grpquota, Opt_extents, Opt_noextents, Opt_i_version,
+	Opt_grpquota, Opt_i_version,
 	Opt_stripe, Opt_delalloc, Opt_nodelalloc,
-	Opt_inode_readahead_blks
+	Opt_inode_readahead_blks, Opt_journal_ioprio
 };
 
 static const match_table_t tokens = {
@@ -889,8 +1043,9 @@ static const match_table_t tokens = {
 	{Opt_nobh, "nobh"},
 	{Opt_bh, "bh"},
 	{Opt_commit, "commit=%u"},
+	{Opt_min_batch_time, "min_batch_time=%u"},
+	{Opt_max_batch_time, "max_batch_time=%u"},
 	{Opt_journal_update, "journal=update"},
-	{Opt_journal_inum, "journal=%u"},
 	{Opt_journal_dev, "journal_dev=%u"},
 	{Opt_journal_checksum, "journal_checksum"},
 	{Opt_journal_async_commit, "journal_async_commit"},
@@ -911,14 +1066,13 @@ static const match_table_t tokens = {
 	{Opt_quota, "quota"},
 	{Opt_usrquota, "usrquota"},
 	{Opt_barrier, "barrier=%u"},
-	{Opt_extents, "extents"},
-	{Opt_noextents, "noextents"},
 	{Opt_i_version, "i_version"},
 	{Opt_stripe, "stripe=%u"},
 	{Opt_resize, "resize"},
 	{Opt_delalloc, "delalloc"},
 	{Opt_nodelalloc, "nodelalloc"},
 	{Opt_inode_readahead_blks, "inode_readahead_blks=%u"},
+	{Opt_journal_ioprio, "journal_ioprio=%u"},
 	{Opt_err, NULL},
 };
 
@@ -943,8 +1097,11 @@ static ext4_fsblk_t get_sb_block(void **data)
 	return sb_block;
 }
 
+#define DEFAULT_JOURNAL_IOPRIO (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, 3))
+
 static int parse_options(char *options, struct super_block *sb,
-			 unsigned int *inum, unsigned long *journal_devnum,
+			 unsigned long *journal_devnum,
+			 unsigned int *journal_ioprio,
 			 ext4_fsblk_t *n_blocks_count, int is_remount)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
@@ -956,7 +1113,6 @@ static int parse_options(char *options, struct super_block *sb,
 	int qtype, qfmt;
 	char *qname;
 #endif
-	ext4_fsblk_t last_block;
 
 	if (!options)
 		return 1;
@@ -1068,16 +1224,6 @@ static int parse_options(char *options, struct super_block *sb,
 			}
 			set_opt(sbi->s_mount_opt, UPDATE_JOURNAL);
 			break;
-		case Opt_journal_inum:
-			if (is_remount) {
-				printk(KERN_ERR "EXT4-fs: cannot specify "
-				       "journal on remount\n");
-				return 0;
-			}
-			if (match_int(&args[0], &option))
-				return 0;
-			*inum = option;
-			break;
 		case Opt_journal_dev:
 			if (is_remount) {
 				printk(KERN_ERR "EXT4-fs: cannot specify "
@@ -1107,6 +1253,22 @@ static int parse_options(char *options, struct super_block *sb,
 				option = JBD2_DEFAULT_MAX_COMMIT_AGE;
 			sbi->s_commit_interval = HZ * option;
 			break;
+		case Opt_max_batch_time:
+			if (match_int(&args[0], &option))
+				return 0;
+			if (option < 0)
+				return 0;
+			if (option == 0)
+				option = EXT4_DEF_MAX_BATCH_TIME;
+			sbi->s_max_batch_time = option;
+			break;
+		case Opt_min_batch_time:
+			if (match_int(&args[0], &option))
+				return 0;
+			if (option < 0)
+				return 0;
+			sbi->s_min_batch_time = option;
+			break;
 		case Opt_data_journal:
 			data_opt = EXT4_MOUNT_JOURNAL_DATA;
 			goto datacheck;
@@ -1142,8 +1304,7 @@ static int parse_options(char *options, struct super_block *sb,
 		case Opt_grpjquota:
 			qtype = GRPQUOTA;
 set_qf_name:
-			if ((sb_any_quota_enabled(sb) ||
-			     sb_any_quota_suspended(sb)) &&
+			if (sb_any_quota_loaded(sb) &&
 			    !sbi->s_qf_names[qtype]) {
 				printk(KERN_ERR
 				       "EXT4-fs: Cannot change journaled "
@@ -1182,8 +1343,7 @@ set_qf_name:
 		case Opt_offgrpjquota:
 			qtype = GRPQUOTA;
 clear_qf_name:
-			if ((sb_any_quota_enabled(sb) ||
-			     sb_any_quota_suspended(sb)) &&
+			if (sb_any_quota_loaded(sb) &&
 			    sbi->s_qf_names[qtype]) {
 				printk(KERN_ERR "EXT4-fs: Cannot change "
 					"journaled quota options when "
@@ -1202,8 +1362,7 @@ clear_qf_name:
 		case Opt_jqfmt_vfsv0:
 			qfmt = QFMT_VFS_V0;
 set_qf_format:
-			if ((sb_any_quota_enabled(sb) ||
-			     sb_any_quota_suspended(sb)) &&
+			if (sb_any_quota_loaded(sb) &&
 			    sbi->s_jquota_fmt != qfmt) {
 				printk(KERN_ERR "EXT4-fs: Cannot change "
 					"journaled quota options when "
@@ -1222,7 +1381,7 @@ set_qf_format:
 			set_opt(sbi->s_mount_opt, GRPQUOTA);
 			break;
 		case Opt_noquota:
-			if (sb_any_quota_enabled(sb)) {
+			if (sb_any_quota_loaded(sb)) {
 				printk(KERN_ERR "EXT4-fs: Cannot change quota "
 					"options when quota turned on.\n");
 				return 0;
@@ -1280,33 +1439,6 @@ set_qf_format:
 		case Opt_bh:
 			clear_opt(sbi->s_mount_opt, NOBH);
 			break;
-		case Opt_extents:
-			if (!EXT4_HAS_INCOMPAT_FEATURE(sb,
-					EXT4_FEATURE_INCOMPAT_EXTENTS)) {
-				ext4_warning(sb, __func__,
-					"extents feature not enabled "
-					"on this filesystem, use tune2fs\n");
-				return 0;
-			}
-			set_opt(sbi->s_mount_opt, EXTENTS);
-			break;
-		case Opt_noextents:
-			/*
-			 * When e2fsprogs support resizing an already existing
-			 * ext3 file system to greater than 2**32 we need to
-			 * add support to block allocator to handle growing
-			 * already existing block  mapped inode so that blocks
-			 * allocated for them fall within 2**32
-			 */
-			last_block = ext4_blocks_count(sbi->s_es) - 1;
-			if (last_block  > 0xffffffffULL) {
-				printk(KERN_ERR "EXT4-fs: Filesystem too "
-						"large to mount with "
-						"-o noextents options\n");
-				return 0;
-			}
-			clear_opt(sbi->s_mount_opt, EXTENTS);
-			break;
 		case Opt_i_version:
 			set_opt(sbi->s_mount_opt, I_VERSION);
 			sb->s_flags |= MS_I_VERSION;
@@ -1331,6 +1463,14 @@ set_qf_format:
 				return 0;
 			sbi->s_inode_readahead_blks = option;
 			break;
+		case Opt_journal_ioprio:
+			if (match_int(&args[0], &option))
+				return 0;
+			if (option < 0 || option > 7)
+				break;
+			*journal_ioprio = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE,
+							    option);
+			break;
 		default:
 			printk(KERN_ERR
 			       "EXT4-fs: Unrecognized mount option \"%s\" "
@@ -1406,24 +1546,19 @@ static int ext4_setup_super(struct super_block *sb, struct ext4_super_block *es,
 		printk(KERN_WARNING
 		       "EXT4-fs warning: checktime reached, "
 		       "running e2fsck is recommended\n");
-#if 0
-		/* @@@ We _will_ want to clear the valid bit if we find
-		 * inconsistencies, to force a fsck at reboot.  But for
-		 * a plain journaled filesystem we can keep it set as
-		 * valid forever! :)
-		 */
-	es->s_state &= cpu_to_le16(~EXT4_VALID_FS);
-#endif
+	if (!sbi->s_journal) 
+		es->s_state &= cpu_to_le16(~EXT4_VALID_FS);
 	if (!(__s16) le16_to_cpu(es->s_max_mnt_count))
 		es->s_max_mnt_count = cpu_to_le16(EXT4_DFL_MAX_MNT_COUNT);
 	le16_add_cpu(&es->s_mnt_count, 1);
 	es->s_mtime = cpu_to_le32(get_seconds());
 	ext4_update_dynamic_rev(sb);
-	EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
+	if (sbi->s_journal)
+		EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 
 	ext4_commit_super(sb, es, 1);
 	if (test_opt(sb, DEBUG))
-		printk(KERN_INFO "[EXT4 FS bs=%lu, gc=%lu, "
+		printk(KERN_INFO "[EXT4 FS bs=%lu, gc=%u, "
 				"bpg=%lu, ipg=%lu, mo=%04lx]\n",
 			sb->s_blocksize,
 			sbi->s_groups_count,
@@ -1431,9 +1566,13 @@ static int ext4_setup_super(struct super_block *sb, struct ext4_super_block *es,
 			EXT4_INODES_PER_GROUP(sb),
 			sbi->s_mount_opt);
 
-	printk(KERN_INFO "EXT4 FS on %s, %s journal on %s\n",
-	       sb->s_id, EXT4_SB(sb)->s_journal->j_inode ? "internal" :
-	       "external", EXT4_SB(sb)->s_journal->j_devname);
+	if (EXT4_SB(sb)->s_journal) {
+		printk(KERN_INFO "EXT4 FS on %s, %s journal on %s\n",
+		       sb->s_id, EXT4_SB(sb)->s_journal->j_inode ? "internal" :
+		       "external", EXT4_SB(sb)->s_journal->j_devname);
+	} else {
+		printk(KERN_INFO "EXT4 FS on %s, no journal\n", sb->s_id);
+	}
 	return res;
 }
 
@@ -1445,7 +1584,6 @@ static int ext4_fill_flex_info(struct super_block *sb)
 	ext4_group_t flex_group_count;
 	ext4_group_t flex_group;
 	int groups_per_flex = 0;
-	__u64 block_bitmap = 0;
 	int i;
 
 	if (!sbi->s_es->s_log_groups_per_flex) {
@@ -1464,21 +1602,18 @@ static int ext4_fill_flex_info(struct super_block *sb)
 				     sizeof(struct flex_groups), GFP_KERNEL);
 	if (sbi->s_flex_groups == NULL) {
 		printk(KERN_ERR "EXT4-fs: not enough memory for "
-				"%lu flex groups\n", flex_group_count);
+				"%u flex groups\n", flex_group_count);
 		goto failed;
 	}
 
-	gdp = ext4_get_group_desc(sb, 1, &bh);
-	block_bitmap = ext4_block_bitmap(sb, gdp) - 1;
-
 	for (i = 0; i < sbi->s_groups_count; i++) {
 		gdp = ext4_get_group_desc(sb, i, &bh);
 
 		flex_group = ext4_flex_group(sbi, i);
 		sbi->s_flex_groups[flex_group].free_inodes +=
-			le16_to_cpu(gdp->bg_free_inodes_count);
+			ext4_free_inodes_count(sb, gdp);
 		sbi->s_flex_groups[flex_group].free_blocks +=
-			le16_to_cpu(gdp->bg_free_blocks_count);
+			ext4_free_blks_count(sb, gdp);
 	}
 
 	return 1;
@@ -1552,14 +1687,14 @@ static int ext4_check_descriptors(struct super_block *sb)
 		block_bitmap = ext4_block_bitmap(sb, gdp);
 		if (block_bitmap < first_block || block_bitmap > last_block) {
 			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
-			       "Block bitmap for group %lu not in group "
+			       "Block bitmap for group %u not in group "
 			       "(block %llu)!\n", i, block_bitmap);
 			return 0;
 		}
 		inode_bitmap = ext4_inode_bitmap(sb, gdp);
 		if (inode_bitmap < first_block || inode_bitmap > last_block) {
 			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
-			       "Inode bitmap for group %lu not in group "
+			       "Inode bitmap for group %u not in group "
 			       "(block %llu)!\n", i, inode_bitmap);
 			return 0;
 		}
@@ -1567,14 +1702,14 @@ static int ext4_check_descriptors(struct super_block *sb)
 		if (inode_table < first_block ||
 		    inode_table + sbi->s_itb_per_group - 1 > last_block) {
 			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
-			       "Inode table for group %lu not in group "
+			       "Inode table for group %u not in group "
 			       "(block %llu)!\n", i, inode_table);
 			return 0;
 		}
 		spin_lock(sb_bgl_lock(sbi, i));
 		if (!ext4_group_desc_csum_verify(sbi, i, gdp)) {
 			printk(KERN_ERR "EXT4-fs: ext4_check_descriptors: "
-			       "Checksum for group %lu failed (%u!=%u)\n",
+			       "Checksum for group %u failed (%u!=%u)\n",
 			       i, le16_to_cpu(ext4_group_desc_csum(sbi, i,
 			       gdp)), le16_to_cpu(gdp->bg_checksum));
 			if (!(sb->s_flags & MS_RDONLY)) {
@@ -1721,7 +1856,7 @@ static loff_t ext4_max_size(int blkbits, int has_huge_files)
 	/* small i_blocks in vfs inode? */
 	if (!has_huge_files || sizeof(blkcnt_t) < sizeof(u64)) {
 		/*
-		 * CONFIG_LSF is not enabled implies the inode
+		 * CONFIG_LBD is not enabled implies the inode
 		 * i_block represent total blocks in 512 bytes
 		 * 32 == size of vfs inode i_blocks * 8
 		 */
@@ -1764,7 +1899,7 @@ static loff_t ext4_max_bitmap_size(int bits, int has_huge_files)
 
 	if (!has_huge_files || sizeof(blkcnt_t) < sizeof(u64)) {
 		/*
-		 * !has_huge_files or CONFIG_LSF is not enabled
+		 * !has_huge_files or CONFIG_LBD is not enabled
 		 * implies the inode i_block represent total blocks in
 		 * 512 bytes 32 == size of vfs inode i_blocks * 8
 		 */
@@ -1866,19 +2001,20 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	ext4_fsblk_t sb_block = get_sb_block(&data);
 	ext4_fsblk_t logical_sb_block;
 	unsigned long offset = 0;
-	unsigned int journal_inum = 0;
 	unsigned long journal_devnum = 0;
 	unsigned long def_mount_opts;
 	struct inode *root;
 	char *cp;
+	const char *descr;
 	int ret = -EINVAL;
 	int blocksize;
-	int db_count;
-	int i;
+	unsigned int db_count;
+	unsigned int i;
 	int needs_recovery, has_huge_files;
-	__le32 features;
+	int features;
 	__u64 blocks_count;
 	int err;
+	unsigned int journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
 
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
 	if (!sbi)
@@ -1959,31 +2095,22 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 
 	sbi->s_resuid = le16_to_cpu(es->s_def_resuid);
 	sbi->s_resgid = le16_to_cpu(es->s_def_resgid);
+	sbi->s_commit_interval = JBD2_DEFAULT_MAX_COMMIT_AGE * HZ;
+	sbi->s_min_batch_time = EXT4_DEF_MIN_BATCH_TIME;
+	sbi->s_max_batch_time = EXT4_DEF_MAX_BATCH_TIME;
 
 	set_opt(sbi->s_mount_opt, RESERVATION);
 	set_opt(sbi->s_mount_opt, BARRIER);
 
 	/*
-	 * turn on extents feature by default in ext4 filesystem
-	 * only if feature flag already set by mkfs or tune2fs.
-	 * Use -o noextents to turn it off
-	 */
-	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
-		set_opt(sbi->s_mount_opt, EXTENTS);
-	else
-		ext4_warning(sb, __func__,
-			"extents feature not enabled on this filesystem, "
-			"use tune2fs.\n");
-
-	/*
 	 * enable delayed allocation by default
 	 * Use -o nodelalloc to turn it off
 	 */
 	set_opt(sbi->s_mount_opt, DELALLOC);
 
 
-	if (!parse_options((char *) data, sb, &journal_inum, &journal_devnum,
-			   NULL, 0))
+	if (!parse_options((char *) data, sb, &journal_devnum,
+			   &journal_ioprio, NULL, 0))
 		goto failed_mount;
 
 	sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
@@ -2005,15 +2132,17 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	features = EXT4_HAS_INCOMPAT_FEATURE(sb, ~EXT4_FEATURE_INCOMPAT_SUPP);
 	if (features) {
 		printk(KERN_ERR "EXT4-fs: %s: couldn't mount because of "
-		       "unsupported optional features (%x).\n",
-		       sb->s_id, le32_to_cpu(features));
+		       "unsupported optional features (%x).\n", sb->s_id,
+			(le32_to_cpu(EXT4_SB(sb)->s_es->s_feature_incompat) &
+			~EXT4_FEATURE_INCOMPAT_SUPP));
 		goto failed_mount;
 	}
 	features = EXT4_HAS_RO_COMPAT_FEATURE(sb, ~EXT4_FEATURE_RO_COMPAT_SUPP);
 	if (!(sb->s_flags & MS_RDONLY) && features) {
 		printk(KERN_ERR "EXT4-fs: %s: couldn't mount RDWR because of "
-		       "unsupported optional features (%x).\n",
-		       sb->s_id, le32_to_cpu(features));
+		       "unsupported optional features (%x).\n", sb->s_id,
+			(le32_to_cpu(EXT4_SB(sb)->s_es->s_feature_ro_compat) &
+			~EXT4_FEATURE_RO_COMPAT_SUPP));
 		goto failed_mount;
 	}
 	has_huge_files = EXT4_HAS_RO_COMPAT_FEATURE(sb,
@@ -2021,13 +2150,13 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	if (has_huge_files) {
 		/*
 		 * Large file size enabled file system can only be
-		 * mount if kernel is build with CONFIG_LSF
+		 * mount if kernel is build with CONFIG_LBD
 		 */
 		if (sizeof(root->i_blocks) < sizeof(u64) &&
 				!(sb->s_flags & MS_RDONLY)) {
 			printk(KERN_ERR "EXT4-fs: %s: Filesystem with huge "
 					"files cannot be mounted read-write "
-					"without CONFIG_LSF.\n", sb->s_id);
+					"without CONFIG_LBD.\n", sb->s_id);
 			goto failed_mount;
 		}
 	}
@@ -2118,6 +2247,18 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	for (i = 0; i < 4; i++)
 		sbi->s_hash_seed[i] = le32_to_cpu(es->s_hash_seed[i]);
 	sbi->s_def_hash_version = es->s_def_hash_version;
+	i = le32_to_cpu(es->s_flags);
+	if (i & EXT2_FLAGS_UNSIGNED_HASH)
+		sbi->s_hash_unsigned = 3;
+	else if ((i & EXT2_FLAGS_SIGNED_HASH) == 0) {
+#ifdef __CHAR_UNSIGNED__
+		es->s_flags |= cpu_to_le32(EXT2_FLAGS_UNSIGNED_HASH);
+		sbi->s_hash_unsigned = 3;
+#else
+		es->s_flags |= cpu_to_le32(EXT2_FLAGS_SIGNED_HASH);
+#endif
+		sb->s_dirt = 1;
+	}
 
 	if (sbi->s_blocks_per_group > blocksize * 8) {
 		printk(KERN_ERR
@@ -2145,20 +2286,30 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	if (EXT4_BLOCKS_PER_GROUP(sb) == 0)
 		goto cantfind_ext4;
 
-	/* ensure blocks_count calculation below doesn't sign-extend */
-	if (ext4_blocks_count(es) + EXT4_BLOCKS_PER_GROUP(sb) <
-	    le32_to_cpu(es->s_first_data_block) + 1) {
-		printk(KERN_WARNING "EXT4-fs: bad geometry: block count %llu, "
-		       "first data block %u, blocks per group %lu\n",
-			ext4_blocks_count(es),
-			le32_to_cpu(es->s_first_data_block),
-			EXT4_BLOCKS_PER_GROUP(sb));
+        /*
+         * It makes no sense for the first data block to be beyond the end
+         * of the filesystem.
+         */
+        if (le32_to_cpu(es->s_first_data_block) >= ext4_blocks_count(es)) {
+                printk(KERN_WARNING "EXT4-fs: bad geometry: first data"
+		       "block %u is beyond end of filesystem (%llu)\n",
+		       le32_to_cpu(es->s_first_data_block),
+		       ext4_blocks_count(es));
 		goto failed_mount;
 	}
 	blocks_count = (ext4_blocks_count(es) -
 			le32_to_cpu(es->s_first_data_block) +
 			EXT4_BLOCKS_PER_GROUP(sb) - 1);
 	do_div(blocks_count, EXT4_BLOCKS_PER_GROUP(sb));
+	if (blocks_count > ((uint64_t)1<<32) - EXT4_DESC_PER_BLOCK(sb)) {
+		printk(KERN_WARNING "EXT4-fs: groups count too large: %u "
+		       "(block count %llu, first data block %u, "
+		       "blocks per group %lu)\n", sbi->s_groups_count,
+		       ext4_blocks_count(es),
+		       le32_to_cpu(es->s_first_data_block),
+		       EXT4_BLOCKS_PER_GROUP(sb));
+		goto failed_mount;
+	}
 	sbi->s_groups_count = blocks_count;
 	db_count = (sbi->s_groups_count + EXT4_DESC_PER_BLOCK(sb) - 1) /
 		   EXT4_DESC_PER_BLOCK(sb);
@@ -2270,27 +2421,26 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 				EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
 				es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
 				ext4_commit_super(sb, es, 1);
-				printk(KERN_CRIT
-				       "EXT4-fs (device %s): mount failed\n",
-				      sb->s_id);
 				goto failed_mount4;
 			}
 		}
-	} else if (journal_inum) {
-		if (ext4_create_journal(sb, es, journal_inum))
-			goto failed_mount3;
+	} else if (test_opt(sb, NOLOAD) && !(sb->s_flags & MS_RDONLY) &&
+	      EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER)) {
+		printk(KERN_ERR "EXT4-fs: required journal recovery "
+		       "suppressed and not mounted read-only\n");
+		goto failed_mount4;
 	} else {
-		if (!silent)
-			printk(KERN_ERR
-			       "ext4: No journal on filesystem on %s\n",
-			       sb->s_id);
-		goto failed_mount3;
+		clear_opt(sbi->s_mount_opt, DATA_FLAGS);
+		set_opt(sbi->s_mount_opt, WRITEBACK_DATA);
+		sbi->s_journal = NULL;
+		needs_recovery = 0;
+		goto no_journal;
 	}
 
 	if (ext4_blocks_count(es) > 0xffffffffULL &&
 	    !jbd2_journal_set_features(EXT4_SB(sb)->s_journal, 0, 0,
 				       JBD2_FEATURE_INCOMPAT_64BIT)) {
-		printk(KERN_ERR "ext4: Failed to set 64-bit journal feature\n");
+		printk(KERN_ERR "EXT4-fs: Failed to set 64-bit journal feature\n");
 		goto failed_mount4;
 	}
 
@@ -2335,6 +2485,9 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	default:
 		break;
 	}
+	set_task_ioprio(sbi->s_journal->j_task, journal_ioprio);
+
+no_journal:
 
 	if (test_opt(sb, NOBH)) {
 		if (!(test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)) {
@@ -2420,13 +2573,22 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	EXT4_SB(sb)->s_mount_state |= EXT4_ORPHAN_FS;
 	ext4_orphan_cleanup(sb, es);
 	EXT4_SB(sb)->s_mount_state &= ~EXT4_ORPHAN_FS;
-	if (needs_recovery)
+	if (needs_recovery) {
 		printk(KERN_INFO "EXT4-fs: recovery complete.\n");
-	ext4_mark_recovery_complete(sb, es);
-	printk(KERN_INFO "EXT4-fs: mounted filesystem with %s data mode.\n",
-	       test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA ? "journal":
-	       test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA ? "ordered":
-	       "writeback");
+		ext4_mark_recovery_complete(sb, es);
+	}
+	if (EXT4_SB(sb)->s_journal) {
+		if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA)
+			descr = " journalled data mode";
+		else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA)
+			descr = " ordered data mode";
+		else
+			descr = " writeback data mode";
+	} else
+		descr = "out journal";
+
+	printk(KERN_INFO "EXT4-fs: mounted filesystem %s with%s\n",
+	       sb->s_id, descr);
 
 	lock_kernel();
 	return 0;
@@ -2438,8 +2600,11 @@ cantfind_ext4:
 	goto failed_mount;
 
 failed_mount4:
-	jbd2_journal_destroy(sbi->s_journal);
-	sbi->s_journal = NULL;
+	printk(KERN_ERR "EXT4-fs (device %s): mount failed\n", sb->s_id);
+	if (sbi->s_journal) {
+		jbd2_journal_destroy(sbi->s_journal);
+		sbi->s_journal = NULL;
+	}
 failed_mount3:
 	percpu_counter_destroy(&sbi->s_freeblocks_counter);
 	percpu_counter_destroy(&sbi->s_freeinodes_counter);
@@ -2476,11 +2641,9 @@ static void ext4_init_journal_params(struct super_block *sb, journal_t *journal)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 
-	if (sbi->s_commit_interval)
-		journal->j_commit_interval = sbi->s_commit_interval;
-	/* We could also set up an ext4-specific default for the commit
-	 * interval here, but for now we'll just fall back to the jbd
-	 * default. */
+	journal->j_commit_interval = sbi->s_commit_interval;
+	journal->j_min_batch_time = sbi->s_min_batch_time;
+	journal->j_max_batch_time = sbi->s_max_batch_time;
 
 	spin_lock(&journal->j_state_lock);
 	if (test_opt(sb, BARRIER))
@@ -2500,6 +2663,8 @@ static journal_t *ext4_get_journal(struct super_block *sb,
 	struct inode *journal_inode;
 	journal_t *journal;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	/* First, test for the existence of a valid inode on disk.  Bad
 	 * things happen if we iget() an unused inode, as the subsequent
 	 * iput() will try to delete it. */
@@ -2548,13 +2713,15 @@ static journal_t *ext4_get_dev_journal(struct super_block *sb,
 	struct ext4_super_block *es;
 	struct block_device *bdev;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	bdev = ext4_blkdev_get(j_dev);
 	if (bdev == NULL)
 		return NULL;
 
 	if (bd_claim(bdev, sb)) {
 		printk(KERN_ERR
-			"EXT4: failed to claim external journal device.\n");
+			"EXT4-fs: failed to claim external journal device.\n");
 		blkdev_put(bdev, FMODE_READ|FMODE_WRITE);
 		return NULL;
 	}
@@ -2635,6 +2802,8 @@ static int ext4_load_journal(struct super_block *sb,
 	int err = 0;
 	int really_read_only;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	if (journal_devnum &&
 	    journal_devnum != le32_to_cpu(es->s_journal_dev)) {
 		printk(KERN_INFO "EXT4-fs: external journal device major/minor "
@@ -2719,55 +2888,14 @@ static int ext4_load_journal(struct super_block *sb,
 	return 0;
 }
 
-static int ext4_create_journal(struct super_block *sb,
-			       struct ext4_super_block *es,
-			       unsigned int journal_inum)
-{
-	journal_t *journal;
-	int err;
-
-	if (sb->s_flags & MS_RDONLY) {
-		printk(KERN_ERR "EXT4-fs: readonly filesystem when trying to "
-				"create journal.\n");
-		return -EROFS;
-	}
-
-	journal = ext4_get_journal(sb, journal_inum);
-	if (!journal)
-		return -EINVAL;
-
-	printk(KERN_INFO "EXT4-fs: creating new journal on inode %u\n",
-	       journal_inum);
-
-	err = jbd2_journal_create(journal);
-	if (err) {
-		printk(KERN_ERR "EXT4-fs: error creating journal.\n");
-		jbd2_journal_destroy(journal);
-		return -EIO;
-	}
-
-	EXT4_SB(sb)->s_journal = journal;
-
-	ext4_update_dynamic_rev(sb);
-	EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
-	EXT4_SET_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL);
-
-	es->s_journal_inum = cpu_to_le32(journal_inum);
-	sb->s_dirt = 1;
-
-	/* Make sure we flush the recovery flag to disk. */
-	ext4_commit_super(sb, es, 1);
-
-	return 0;
-}
-
-static void ext4_commit_super(struct super_block *sb,
+static int ext4_commit_super(struct super_block *sb,
 			      struct ext4_super_block *es, int sync)
 {
 	struct buffer_head *sbh = EXT4_SB(sb)->s_sbh;
+	int error = 0;
 
 	if (!sbh)
-		return;
+		return error;
 	if (buffer_write_io_error(sbh)) {
 		/*
 		 * Oh, dear.  A previous attempt to write the
@@ -2777,25 +2905,33 @@ static void ext4_commit_super(struct super_block *sb,
 		 * be remapped.  Nothing we can do but to retry the
 		 * write and hope for the best.
 		 */
-		printk(KERN_ERR "ext4: previous I/O error to "
+		printk(KERN_ERR "EXT4-fs: previous I/O error to "
 		       "superblock detected for %s.\n", sb->s_id);
 		clear_buffer_write_io_error(sbh);
 		set_buffer_uptodate(sbh);
 	}
 	es->s_wtime = cpu_to_le32(get_seconds());
-	ext4_free_blocks_count_set(es, ext4_count_free_blocks(sb));
-	es->s_free_inodes_count = cpu_to_le32(ext4_count_free_inodes(sb));
+	ext4_free_blocks_count_set(es, percpu_counter_sum_positive(
+					&EXT4_SB(sb)->s_freeblocks_counter));
+	es->s_free_inodes_count = cpu_to_le32(percpu_counter_sum_positive(
+					&EXT4_SB(sb)->s_freeinodes_counter));
+
 	BUFFER_TRACE(sbh, "marking dirty");
 	mark_buffer_dirty(sbh);
 	if (sync) {
-		sync_dirty_buffer(sbh);
-		if (buffer_write_io_error(sbh)) {
-			printk(KERN_ERR "ext4: I/O error while writing "
+		error = sync_dirty_buffer(sbh);
+		if (error)
+			return error;
+
+		error = buffer_write_io_error(sbh);
+		if (error) {
+			printk(KERN_ERR "EXT4-fs: I/O error while writing "
 			       "superblock for %s.\n", sb->s_id);
 			clear_buffer_write_io_error(sbh);
 			set_buffer_uptodate(sbh);
 		}
 	}
+	return error;
 }
 
 
@@ -2809,6 +2945,10 @@ static void ext4_mark_recovery_complete(struct super_block *sb,
 {
 	journal_t *journal = EXT4_SB(sb)->s_journal;
 
+	if (!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL)) {
+		BUG_ON(journal != NULL);
+		return;
+	}
 	jbd2_journal_lock_updates(journal);
 	if (jbd2_journal_flush(journal) < 0)
 		goto out;
@@ -2838,6 +2978,8 @@ static void ext4_clear_journal_err(struct super_block *sb,
 	int j_errno;
 	const char *errstr;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	journal = EXT4_SB(sb)->s_journal;
 
 	/*
@@ -2870,14 +3012,17 @@ static void ext4_clear_journal_err(struct super_block *sb,
 int ext4_force_commit(struct super_block *sb)
 {
 	journal_t *journal;
-	int ret;
+	int ret = 0;
 
 	if (sb->s_flags & MS_RDONLY)
 		return 0;
 
 	journal = EXT4_SB(sb)->s_journal;
-	sb->s_dirt = 0;
-	ret = ext4_journal_force_commit(journal);
+	if (journal) {
+		sb->s_dirt = 0;
+		ret = ext4_journal_force_commit(journal);
+	}
+
 	return ret;
 }
 
@@ -2889,9 +3034,13 @@ int ext4_force_commit(struct super_block *sb)
  */
 static void ext4_write_super(struct super_block *sb)
 {
-	if (mutex_trylock(&sb->s_lock) != 0)
-		BUG();
-	sb->s_dirt = 0;
+	if (EXT4_SB(sb)->s_journal) {
+		if (mutex_trylock(&sb->s_lock) != 0)
+			BUG();
+		sb->s_dirt = 0;
+	} else {
+		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
+	}
 }
 
 static int ext4_sync_fs(struct super_block *sb, int wait)
@@ -2900,10 +3049,14 @@ static int ext4_sync_fs(struct super_block *sb, int wait)
 
 	trace_mark(ext4_sync_fs, "dev %s wait %d", sb->s_id, wait);
 	sb->s_dirt = 0;
-	if (wait)
-		ret = ext4_force_commit(sb);
-	else
-		jbd2_journal_start_commit(EXT4_SB(sb)->s_journal, NULL);
+	if (EXT4_SB(sb)->s_journal) {
+		if (wait)
+			ret = ext4_force_commit(sb);
+		else
+ 			jbd2_journal_start_commit(EXT4_SB(sb)->s_journal, NULL);
+	} else {
+		ext4_commit_super(sb, EXT4_SB(sb)->s_es, wait);
+	}
 	return ret;
 }
 
@@ -2911,36 +3064,48 @@ static int ext4_sync_fs(struct super_block *sb, int wait)
  * LVM calls this function before a (read-only) snapshot is created.  This
  * gives us a chance to flush the journal completely and mark the fs clean.
  */
-static void ext4_write_super_lockfs(struct super_block *sb)
+static int ext4_freeze(struct super_block *sb)
 {
+	int error = 0;
+	journal_t *journal;
 	sb->s_dirt = 0;
 
 	if (!(sb->s_flags & MS_RDONLY)) {
-		journal_t *journal = EXT4_SB(sb)->s_journal;
+		journal = EXT4_SB(sb)->s_journal;
 
-		/* Now we set up the journal barrier. */
-		jbd2_journal_lock_updates(journal);
+		if (journal) {
+			/* Now we set up the journal barrier. */
+			jbd2_journal_lock_updates(journal);
 
-		/*
-		 * We don't want to clear needs_recovery flag when we failed
-		 * to flush the journal.
-		 */
-		if (jbd2_journal_flush(journal) < 0)
-			return;
+			/*
+			 * We don't want to clear needs_recovery flag when we
+			 * failed to flush the journal.
+			 */
+			error = jbd2_journal_flush(journal);
+			if (error < 0)
+				goto out;
+		}
 
 		/* Journal blocked and flushed, clear needs_recovery flag. */
 		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
+		error = ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
+		if (error)
+			goto out;
 	}
+	return 0;
+out:
+	jbd2_journal_unlock_updates(journal);
+	return error;
 }
 
 /*
  * Called by LVM after the snapshot is done.  We need to reset the RECOVER
  * flag here, even though the filesystem is not technically dirty yet.
  */
-static void ext4_unlockfs(struct super_block *sb)
+static int ext4_unfreeze(struct super_block *sb)
 {
-	if (!(sb->s_flags & MS_RDONLY)) {
+	if (EXT4_SB(sb)->s_journal && !(sb->s_flags & MS_RDONLY)) {
 		lock_super(sb);
 		/* Reser the needs_recovery flag before the fs is unlocked. */
 		EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
@@ -2948,6 +3113,7 @@ static void ext4_unlockfs(struct super_block *sb)
 		unlock_super(sb);
 		jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal);
 	}
+	return 0;
 }
 
 static int ext4_remount(struct super_block *sb, int *flags, char *data)
@@ -2958,6 +3124,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 	unsigned long old_sb_flags;
 	struct ext4_mount_options old_opts;
 	ext4_group_t g;
+	unsigned int journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
 	int err;
 #ifdef CONFIG_QUOTA
 	int i;
@@ -2969,16 +3136,21 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 	old_opts.s_resuid = sbi->s_resuid;
 	old_opts.s_resgid = sbi->s_resgid;
 	old_opts.s_commit_interval = sbi->s_commit_interval;
+	old_opts.s_min_batch_time = sbi->s_min_batch_time;
+	old_opts.s_max_batch_time = sbi->s_max_batch_time;
 #ifdef CONFIG_QUOTA
 	old_opts.s_jquota_fmt = sbi->s_jquota_fmt;
 	for (i = 0; i < MAXQUOTAS; i++)
 		old_opts.s_qf_names[i] = sbi->s_qf_names[i];
 #endif
+	if (sbi->s_journal && sbi->s_journal->j_task->io_context)
+		journal_ioprio = sbi->s_journal->j_task->io_context->ioprio;
 
 	/*
 	 * Allow the "check" option to be passed as a remount option.
 	 */
-	if (!parse_options(data, sb, NULL, NULL, &n_blocks_count, 1)) {
+	if (!parse_options(data, sb, NULL, &journal_ioprio,
+			   &n_blocks_count, 1)) {
 		err = -EINVAL;
 		goto restore_opts;
 	}
@@ -2991,7 +3163,10 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 
 	es = sbi->s_es;
 
-	ext4_init_journal_params(sb, sbi->s_journal);
+	if (sbi->s_journal) {
+		ext4_init_journal_params(sb, sbi->s_journal);
+		set_task_ioprio(sbi->s_journal->j_task, journal_ioprio);
+	}
 
 	if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY) ||
 		n_blocks_count > ext4_blocks_count(es)) {
@@ -3020,17 +3195,20 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 			 * We have to unlock super so that we can wait for
 			 * transactions.
 			 */
-			unlock_super(sb);
-			ext4_mark_recovery_complete(sb, es);
-			lock_super(sb);
+			if (sbi->s_journal) {
+				unlock_super(sb);
+				ext4_mark_recovery_complete(sb, es);
+				lock_super(sb);
+			}
 		} else {
-			__le32 ret;
+			int ret;
 			if ((ret = EXT4_HAS_RO_COMPAT_FEATURE(sb,
 					~EXT4_FEATURE_RO_COMPAT_SUPP))) {
 				printk(KERN_WARNING "EXT4-fs: %s: couldn't "
 				       "remount RDWR because of unsupported "
-				       "optional features (%x).\n",
-				       sb->s_id, le32_to_cpu(ret));
+				       "optional features (%x).\n", sb->s_id,
+				(le32_to_cpu(sbi->s_es->s_feature_ro_compat) &
+					~EXT4_FEATURE_RO_COMPAT_SUPP));
 				err = -EROFS;
 				goto restore_opts;
 			}
@@ -3047,7 +3225,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 				if (!ext4_group_desc_csum_verify(sbi, g, gdp)) {
 					printk(KERN_ERR
 	       "EXT4-fs: ext4_remount: "
-		"Checksum for group %lu failed (%u!=%u)\n",
+		"Checksum for group %u failed (%u!=%u)\n",
 		g, le16_to_cpu(ext4_group_desc_csum(sbi, g, gdp)),
 					       le16_to_cpu(gdp->bg_checksum));
 					err = -EINVAL;
@@ -3076,7 +3254,8 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 			 * been changed by e2fsck since we originally mounted
 			 * the partition.)
 			 */
-			ext4_clear_journal_err(sb, es);
+			if (sbi->s_journal)
+				ext4_clear_journal_err(sb, es);
 			sbi->s_mount_state = le16_to_cpu(es->s_state);
 			if ((err = ext4_group_extend(sb, es, n_blocks_count)))
 				goto restore_opts;
@@ -3084,6 +3263,9 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 				sb->s_flags &= ~MS_RDONLY;
 		}
 	}
+	if (sbi->s_journal == NULL)
+		ext4_commit_super(sb, es, 1);
+
 #ifdef CONFIG_QUOTA
 	/* Release old quota file names */
 	for (i = 0; i < MAXQUOTAS; i++)
@@ -3098,6 +3280,8 @@ restore_opts:
 	sbi->s_resuid = old_opts.s_resuid;
 	sbi->s_resgid = old_opts.s_resgid;
 	sbi->s_commit_interval = old_opts.s_commit_interval;
+	sbi->s_min_batch_time = old_opts.s_min_batch_time;
+	sbi->s_max_batch_time = old_opts.s_max_batch_time;
 #ifdef CONFIG_QUOTA
 	sbi->s_jquota_fmt = old_opts.s_jquota_fmt;
 	for (i = 0; i < MAXQUOTAS; i++) {
@@ -3360,7 +3544,8 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id,
 	 * When we journal data on quota file, we have to flush journal to see
 	 * all updates to the file when we bypass pagecache...
 	 */
-	if (ext4_should_journal_data(path.dentry->d_inode)) {
+	if (EXT4_SB(sb)->s_journal &&
+	    ext4_should_journal_data(path.dentry->d_inode)) {
 		/*
 		 * We don't need to lock updates but journal_flush() could
 		 * otherwise be livelocked...
@@ -3434,7 +3619,7 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type,
 	struct buffer_head *bh;
 	handle_t *handle = journal_current_handle();
 
-	if (!handle) {
+	if (EXT4_SB(sb)->s_journal && !handle) {
 		printk(KERN_WARNING "EXT4-fs: Quota write (off=%llu, len=%llu)"
 			" cancelled because transaction is not started.\n",
 			(unsigned long long)off, (unsigned long long)len);
@@ -3459,7 +3644,7 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type,
 		flush_dcache_page(bh->b_page);
 		unlock_buffer(bh);
 		if (journal_quota)
-			err = ext4_journal_dirty_metadata(handle, bh);
+			err = ext4_handle_dirty_metadata(handle, NULL, bh);
 		else {
 			/* Always do at least ordered writes for quotas */
 			err = ext4_jbd2_file_inode(handle, inode);
@@ -3513,18 +3698,15 @@ static int ext4_ui_proc_open(struct inode *inode, struct file *file)
 static ssize_t ext4_ui_proc_write(struct file *file, const char __user *buf,
 			       size_t cnt, loff_t *ppos)
 {
-	unsigned int *p = PDE(file->f_path.dentry->d_inode)->data;
+	unsigned long *p = PDE(file->f_path.dentry->d_inode)->data;
 	char str[32];
-	unsigned long value;
 
 	if (cnt >= sizeof(str))
 		return -EINVAL;
 	if (copy_from_user(str, buf, cnt))
 		return -EFAULT;
-	value = simple_strtol(str, NULL, 0);
-	if (value < 0)
-		return -ERANGE;
-	*p = value;
+
+	*p = simple_strtoul(str, NULL, 0);
 	return cnt;
 }
 
@@ -3615,7 +3797,7 @@ static void __exit exit_ext4_fs(void)
 }
 
 MODULE_AUTHOR("Remy Card, Stephen Tweedie, Andrew Morton, Andreas Dilger, Theodore Ts'o and others");
-MODULE_DESCRIPTION("Fourth Extended Filesystem with extents");
+MODULE_DESCRIPTION("Fourth Extended Filesystem");
 MODULE_LICENSE("GPL");
 module_init(init_ext4_fs)
 module_exit(exit_ext4_fs)
diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index 80626d516fe..157ce6589c5 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -457,7 +457,7 @@ static void ext4_xattr_update_super_block(handle_t *handle,
 	if (ext4_journal_get_write_access(handle, EXT4_SB(sb)->s_sbh) == 0) {
 		EXT4_SET_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_EXT_ATTR);
 		sb->s_dirt = 1;
-		ext4_journal_dirty_metadata(handle, EXT4_SB(sb)->s_sbh);
+		ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh);
 	}
 }
 
@@ -487,9 +487,9 @@ ext4_xattr_release_block(handle_t *handle, struct inode *inode,
 		ext4_forget(handle, 1, inode, bh, bh->b_blocknr);
 	} else {
 		le32_add_cpu(&BHDR(bh)->h_refcount, -1);
-		error = ext4_journal_dirty_metadata(handle, bh);
+		error = ext4_handle_dirty_metadata(handle, inode, bh);
 		if (IS_SYNC(inode))
-			handle->h_sync = 1;
+			ext4_handle_sync(handle);
 		DQUOT_FREE_BLOCK(inode, 1);
 		ea_bdebug(bh, "refcount now=%d; releasing",
 			  le32_to_cpu(BHDR(bh)->h_refcount));
@@ -724,8 +724,9 @@ ext4_xattr_block_set(handle_t *handle, struct inode *inode,
 			if (error == -EIO)
 				goto bad_block;
 			if (!error)
-				error = ext4_journal_dirty_metadata(handle,
-								    bs->bh);
+				error = ext4_handle_dirty_metadata(handle,
+								   inode,
+								   bs->bh);
 			if (error)
 				goto cleanup;
 			goto inserted;
@@ -794,8 +795,9 @@ inserted:
 				ea_bdebug(new_bh, "reusing; refcount now=%d",
 					le32_to_cpu(BHDR(new_bh)->h_refcount));
 				unlock_buffer(new_bh);
-				error = ext4_journal_dirty_metadata(handle,
-								    new_bh);
+				error = ext4_handle_dirty_metadata(handle,
+								   inode,
+								   new_bh);
 				if (error)
 					goto cleanup_dquot;
 			}
@@ -810,8 +812,8 @@ inserted:
 			/* We need to allocate a new block */
 			ext4_fsblk_t goal = ext4_group_first_block_no(sb,
 						EXT4_I(inode)->i_block_group);
-			ext4_fsblk_t block = ext4_new_meta_block(handle, inode,
-							goal, &error);
+			ext4_fsblk_t block = ext4_new_meta_blocks(handle, inode,
+						  goal, NULL, &error);
 			if (error)
 				goto cleanup;
 			ea_idebug(inode, "creating block %d", block);
@@ -833,7 +835,8 @@ getblk_failed:
 			set_buffer_uptodate(new_bh);
 			unlock_buffer(new_bh);
 			ext4_xattr_cache_insert(new_bh);
-			error = ext4_journal_dirty_metadata(handle, new_bh);
+			error = ext4_handle_dirty_metadata(handle,
+							   inode, new_bh);
 			if (error)
 				goto cleanup;
 		}
@@ -1040,7 +1043,7 @@ ext4_xattr_set_handle(handle_t *handle, struct inode *inode, int name_index,
 		 */
 		is.iloc.bh = NULL;
 		if (IS_SYNC(inode))
-			handle->h_sync = 1;
+			ext4_handle_sync(handle);
 	}
 
 cleanup:
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index 67e05835709..3a7f603b698 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -841,7 +841,6 @@ const struct file_operations fat_dir_operations = {
 	.compat_ioctl	= fat_compat_dir_ioctl,
 #endif
 	.fsync		= file_fsync,
-	.llseek		= generic_file_llseek,
 };
 
 static int fat_get_short_entry(struct inode *dir, loff_t *pos,
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index d937aaf7737..6b74d09adbe 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -749,6 +749,8 @@ static struct dentry *fat_get_parent(struct dentry *child)
 	brelse(bh);
 
 	parent = d_obtain_alias(inode);
+	if (!IS_ERR(parent))
+		parent->d_op = sb->s_root->d_op;
 out:
 	unlock_super(sb);
 
diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c
index bf326d4356a..8ae32e37673 100644
--- a/fs/fat/namei_vfat.c
+++ b/fs/fat/namei_vfat.c
@@ -78,7 +78,7 @@ static int vfat_revalidate_ci(struct dentry *dentry, struct nameidata *nd)
 	 * for creation.
 	 */
 	if (!(nd->flags & (LOOKUP_CONTINUE | LOOKUP_PARENT))) {
-		if (nd->flags & LOOKUP_CREATE)
+		if (nd->flags & (LOOKUP_CREATE | LOOKUP_RENAME_TARGET))
 			return 0;
 	}
 
diff --git a/fs/file_table.c b/fs/file_table.c
index 0fbcacc3ea7..bbeeac6efa1 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -32,6 +32,9 @@ struct files_stat_struct files_stat = {
 /* public. Not pretty! */
 __cacheline_aligned_in_smp DEFINE_SPINLOCK(files_lock);
 
+/* SLAB cache for file structures */
+static struct kmem_cache *filp_cachep __read_mostly;
+
 static struct percpu_counter nr_files __cacheline_aligned_in_smp;
 
 static inline void file_free_rcu(struct rcu_head *head)
@@ -397,7 +400,12 @@ too_bad:
 void __init files_init(unsigned long mempages)
 { 
 	int n; 
-	/* One file with associated inode and dcache is very roughly 1K. 
+
+	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
+			SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL);
+
+	/*
+	 * One file with associated inode and dcache is very roughly 1K.
 	 * Per default don't use more than 10% of our memory for files. 
 	 */ 
 
diff --git a/fs/filesystems.c b/fs/filesystems.c
index d0e20ced62d..d488dcd7f2b 100644
--- a/fs/filesystems.c
+++ b/fs/filesystems.c
@@ -253,24 +253,27 @@ static int __init proc_filesystems_init(void)
 module_init(proc_filesystems_init);
 #endif
 
-struct file_system_type *get_fs_type(const char *name)
+static struct file_system_type *__get_fs_type(const char *name, int len)
 {
 	struct file_system_type *fs;
-	const char *dot = strchr(name, '.');
-	unsigned len = dot ? dot - name : strlen(name);
 
 	read_lock(&file_systems_lock);
 	fs = *(find_filesystem(name, len));
 	if (fs && !try_module_get(fs->owner))
 		fs = NULL;
 	read_unlock(&file_systems_lock);
-	if (!fs && (request_module("%.*s", len, name) == 0)) {
-		read_lock(&file_systems_lock);
-		fs = *(find_filesystem(name, len));
-		if (fs && !try_module_get(fs->owner))
-			fs = NULL;
-		read_unlock(&file_systems_lock);
-	}
+	return fs;
+}
+
+struct file_system_type *get_fs_type(const char *name)
+{
+	struct file_system_type *fs;
+	const char *dot = strchr(name, '.');
+	int len = dot ? dot - name : strlen(name);
+
+	fs = __get_fs_type(name, len);
+	if (!fs && (request_module("%.*s", len, name) == 0))
+		fs = __get_fs_type(name, len);
 
 	if (dot && fs && !(fs->fs_flags & FS_HAS_SUBTYPE)) {
 		put_filesystem(fs);
diff --git a/fs/freevxfs/vxfs_inode.c b/fs/freevxfs/vxfs_inode.c
index 9f3f2ceb73f..03a6ea5e99f 100644
--- a/fs/freevxfs/vxfs_inode.c
+++ b/fs/freevxfs/vxfs_inode.c
@@ -325,8 +325,10 @@ vxfs_iget(struct super_block *sbp, ino_t ino)
 		if (!VXFS_ISIMMED(vip)) {
 			ip->i_op = &page_symlink_inode_operations;
 			ip->i_mapping->a_ops = &vxfs_aops;
-		} else
+		} else {
 			ip->i_op = &vxfs_immed_symlink_iops;
+			vip->vii_immed.vi_immed[ip->i_size] = '\0';
+		}
 	} else
 		init_special_inode(ip, ip->i_mode, old_decode_dev(vip->vii_rdev));
 
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index d0ff0b8cf30..e5eaa62fd17 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -421,9 +421,6 @@ __writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
  * If we're a pdlfush thread, then implement pdflush collision avoidance
  * against the entire list.
  *
- * WB_SYNC_HOLD is a hack for sys_sync(): reattach the inode to sb->s_dirty so
- * that it can be located for waiting on in __writeback_single_inode().
- *
  * If `bdi' is non-zero then we're being asked to writeback a specific queue.
  * This function assumes that the blockdev superblock's inodes are backed by
  * a variety of queues, so all inodes are searched.  For other superblocks,
@@ -443,6 +440,7 @@ void generic_sync_sb_inodes(struct super_block *sb,
 				struct writeback_control *wbc)
 {
 	const unsigned long start = jiffies;	/* livelock avoidance */
+	int sync = wbc->sync_mode == WB_SYNC_ALL;
 
 	spin_lock(&inode_lock);
 	if (!wbc->for_kupdate || list_empty(&sb->s_io))
@@ -499,10 +497,6 @@ void generic_sync_sb_inodes(struct super_block *sb,
 		__iget(inode);
 		pages_skipped = wbc->pages_skipped;
 		__writeback_single_inode(inode, wbc);
-		if (wbc->sync_mode == WB_SYNC_HOLD) {
-			inode->dirtied_when = jiffies;
-			list_move(&inode->i_list, &sb->s_dirty);
-		}
 		if (current_is_pdflush())
 			writeback_release(bdi);
 		if (wbc->pages_skipped != pages_skipped) {
@@ -523,7 +517,49 @@ void generic_sync_sb_inodes(struct super_block *sb,
 		if (!list_empty(&sb->s_more_io))
 			wbc->more_io = 1;
 	}
-	spin_unlock(&inode_lock);
+
+	if (sync) {
+		struct inode *inode, *old_inode = NULL;
+
+		/*
+		 * Data integrity sync. Must wait for all pages under writeback,
+		 * because there may have been pages dirtied before our sync
+		 * call, but which had writeout started before we write it out.
+		 * In which case, the inode may not be on the dirty list, but
+		 * we still have to wait for that writeout.
+		 */
+		list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+			struct address_space *mapping;
+
+			if (inode->i_state & (I_FREEING|I_WILL_FREE))
+				continue;
+			mapping = inode->i_mapping;
+			if (mapping->nrpages == 0)
+				continue;
+			__iget(inode);
+			spin_unlock(&inode_lock);
+			/*
+			 * We hold a reference to 'inode' so it couldn't have
+			 * been removed from s_inodes list while we dropped the
+			 * inode_lock.  We cannot iput the inode now as we can
+			 * be holding the last reference and we cannot iput it
+			 * under inode_lock. So we keep the reference and iput
+			 * it later.
+			 */
+			iput(old_inode);
+			old_inode = inode;
+
+			filemap_fdatawait(mapping);
+
+			cond_resched();
+
+			spin_lock(&inode_lock);
+		}
+		spin_unlock(&inode_lock);
+		iput(old_inode);
+	} else
+		spin_unlock(&inode_lock);
+
 	return;		/* Leave any unwritten inodes on s_io */
 }
 EXPORT_SYMBOL_GPL(generic_sync_sb_inodes);
@@ -588,8 +624,7 @@ restart:
 
 /*
  * writeback and wait upon the filesystem's dirty inodes.  The caller will
- * do this in two passes - one to write, and one to wait.  WB_SYNC_HOLD is
- * used to park the written inodes on sb->s_dirty for the wait pass.
+ * do this in two passes - one to write, and one to wait.
  *
  * A finite limit is set on the number of pages which will be written.
  * To prevent infinite livelock of sys_sync().
@@ -600,30 +635,21 @@ restart:
 void sync_inodes_sb(struct super_block *sb, int wait)
 {
 	struct writeback_control wbc = {
-		.sync_mode	= wait ? WB_SYNC_ALL : WB_SYNC_HOLD,
+		.sync_mode	= wait ? WB_SYNC_ALL : WB_SYNC_NONE,
 		.range_start	= 0,
 		.range_end	= LLONG_MAX,
 	};
-	unsigned long nr_dirty = global_page_state(NR_FILE_DIRTY);
-	unsigned long nr_unstable = global_page_state(NR_UNSTABLE_NFS);
 
-	wbc.nr_to_write = nr_dirty + nr_unstable +
-			(inodes_stat.nr_inodes - inodes_stat.nr_unused) +
-			nr_dirty + nr_unstable;
-	wbc.nr_to_write += wbc.nr_to_write / 2;		/* Bit more for luck */
-	sync_sb_inodes(sb, &wbc);
-}
+	if (!wait) {
+		unsigned long nr_dirty = global_page_state(NR_FILE_DIRTY);
+		unsigned long nr_unstable = global_page_state(NR_UNSTABLE_NFS);
 
-/*
- * Rather lame livelock avoidance.
- */
-static void set_sb_syncing(int val)
-{
-	struct super_block *sb;
-	spin_lock(&sb_lock);
-	list_for_each_entry_reverse(sb, &super_blocks, s_list)
-		sb->s_syncing = val;
-	spin_unlock(&sb_lock);
+		wbc.nr_to_write = nr_dirty + nr_unstable +
+			(inodes_stat.nr_inodes - inodes_stat.nr_unused);
+	} else
+		wbc.nr_to_write = LONG_MAX; /* doesn't actually matter */
+
+	sync_sb_inodes(sb, &wbc);
 }
 
 /**
@@ -652,9 +678,6 @@ static void __sync_inodes(int wait)
 	spin_lock(&sb_lock);
 restart:
 	list_for_each_entry(sb, &super_blocks, s_list) {
-		if (sb->s_syncing)
-			continue;
-		sb->s_syncing = 1;
 		sb->s_count++;
 		spin_unlock(&sb_lock);
 		down_read(&sb->s_umount);
@@ -672,13 +695,10 @@ restart:
 
 void sync_inodes(int wait)
 {
-	set_sb_syncing(0);
 	__sync_inodes(0);
 
-	if (wait) {
-		set_sb_syncing(0);
+	if (wait)
 		__sync_inodes(1);
-	}
 }
 
 /**
diff --git a/fs/fuse/control.c b/fs/fuse/control.c
index 4f3cab32141..99c99dfb037 100644
--- a/fs/fuse/control.c
+++ b/fs/fuse/control.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -48,11 +48,13 @@ static ssize_t fuse_conn_waiting_read(struct file *file, char __user *buf,
 	size_t size;
 
 	if (!*ppos) {
+		long value;
 		struct fuse_conn *fc = fuse_ctl_file_conn_get(file);
 		if (!fc)
 			return 0;
 
-		file->private_data=(void *)(long)atomic_read(&fc->num_waiting);
+		value = atomic_read(&fc->num_waiting);
+		file->private_data = (void *)value;
 		fuse_conn_put(fc);
 	}
 	size = sprintf(tmp, "%ld\n", (long)file->private_data);
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index fba571648a8..e0c7ada08a1 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -269,7 +269,7 @@ static void flush_bg_queue(struct fuse_conn *fc)
  * Called with fc->lock, unlocks it
  */
 static void request_end(struct fuse_conn *fc, struct fuse_req *req)
-	__releases(fc->lock)
+__releases(&fc->lock)
 {
 	void (*end) (struct fuse_conn *, struct fuse_req *) = req->end;
 	req->end = NULL;
@@ -293,13 +293,13 @@ static void request_end(struct fuse_conn *fc, struct fuse_req *req)
 	wake_up(&req->waitq);
 	if (end)
 		end(fc, req);
-	else
-		fuse_put_request(fc, req);
+	fuse_put_request(fc, req);
 }
 
 static void wait_answer_interruptible(struct fuse_conn *fc,
 				      struct fuse_req *req)
-	__releases(fc->lock) __acquires(fc->lock)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	if (signal_pending(current))
 		return;
@@ -317,7 +317,8 @@ static void queue_interrupt(struct fuse_conn *fc, struct fuse_req *req)
 }
 
 static void request_wait_answer(struct fuse_conn *fc, struct fuse_req *req)
-	__releases(fc->lock) __acquires(fc->lock)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	if (!fc->no_interrupt) {
 		/* Any signal may interrupt this */
@@ -380,7 +381,7 @@ static void request_wait_answer(struct fuse_conn *fc, struct fuse_req *req)
 	}
 }
 
-void request_send(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send(struct fuse_conn *fc, struct fuse_req *req)
 {
 	req->isreply = 1;
 	spin_lock(&fc->lock);
@@ -399,8 +400,8 @@ void request_send(struct fuse_conn *fc, struct fuse_req *req)
 	spin_unlock(&fc->lock);
 }
 
-static void request_send_nowait_locked(struct fuse_conn *fc,
-				       struct fuse_req *req)
+static void fuse_request_send_nowait_locked(struct fuse_conn *fc,
+					    struct fuse_req *req)
 {
 	req->background = 1;
 	fc->num_background++;
@@ -414,11 +415,11 @@ static void request_send_nowait_locked(struct fuse_conn *fc,
 	flush_bg_queue(fc);
 }
 
-static void request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
+static void fuse_request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
 {
 	spin_lock(&fc->lock);
 	if (fc->connected) {
-		request_send_nowait_locked(fc, req);
+		fuse_request_send_nowait_locked(fc, req);
 		spin_unlock(&fc->lock);
 	} else {
 		req->out.h.error = -ENOTCONN;
@@ -426,16 +427,16 @@ static void request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
 	}
 }
 
-void request_send_noreply(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send_noreply(struct fuse_conn *fc, struct fuse_req *req)
 {
 	req->isreply = 0;
-	request_send_nowait(fc, req);
+	fuse_request_send_nowait(fc, req);
 }
 
-void request_send_background(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send_background(struct fuse_conn *fc, struct fuse_req *req)
 {
 	req->isreply = 1;
-	request_send_nowait(fc, req);
+	fuse_request_send_nowait(fc, req);
 }
 
 /*
@@ -443,10 +444,11 @@ void request_send_background(struct fuse_conn *fc, struct fuse_req *req)
  *
  * fc->connected must have been checked previously
  */
-void request_send_background_locked(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send_background_locked(struct fuse_conn *fc,
+					 struct fuse_req *req)
 {
 	req->isreply = 1;
-	request_send_nowait_locked(fc, req);
+	fuse_request_send_nowait_locked(fc, req);
 }
 
 /*
@@ -539,8 +541,8 @@ static int fuse_copy_fill(struct fuse_copy_state *cs)
 		BUG_ON(!cs->nr_segs);
 		cs->seglen = cs->iov[0].iov_len;
 		cs->addr = (unsigned long) cs->iov[0].iov_base;
-		cs->iov ++;
-		cs->nr_segs --;
+		cs->iov++;
+		cs->nr_segs--;
 	}
 	down_read(&current->mm->mmap_sem);
 	err = get_user_pages(current, current->mm, cs->addr, 1, cs->write, 0,
@@ -589,9 +591,11 @@ static int fuse_copy_page(struct fuse_copy_state *cs, struct page *page,
 		kunmap_atomic(mapaddr, KM_USER1);
 	}
 	while (count) {
-		int err;
-		if (!cs->len && (err = fuse_copy_fill(cs)))
-			return err;
+		if (!cs->len) {
+			int err = fuse_copy_fill(cs);
+			if (err)
+				return err;
+		}
 		if (page) {
 			void *mapaddr = kmap_atomic(page, KM_USER1);
 			void *buf = mapaddr + offset;
@@ -631,9 +635,11 @@ static int fuse_copy_pages(struct fuse_copy_state *cs, unsigned nbytes,
 static int fuse_copy_one(struct fuse_copy_state *cs, void *val, unsigned size)
 {
 	while (size) {
-		int err;
-		if (!cs->len && (err = fuse_copy_fill(cs)))
-			return err;
+		if (!cs->len) {
+			int err = fuse_copy_fill(cs);
+			if (err)
+				return err;
+		}
 		fuse_copy_do(cs, &val, &size);
 	}
 	return 0;
@@ -664,6 +670,8 @@ static int request_pending(struct fuse_conn *fc)
 
 /* Wait until a request is available on the pending list */
 static void request_wait(struct fuse_conn *fc)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	DECLARE_WAITQUEUE(wait, current);
 
@@ -691,7 +699,7 @@ static void request_wait(struct fuse_conn *fc)
  */
 static int fuse_read_interrupt(struct fuse_conn *fc, struct fuse_req *req,
 			       const struct iovec *iov, unsigned long nr_segs)
-	__releases(fc->lock)
+__releases(&fc->lock)
 {
 	struct fuse_copy_state cs;
 	struct fuse_in_header ih;
@@ -813,6 +821,34 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
 	return err;
 }
 
+static int fuse_notify_poll(struct fuse_conn *fc, unsigned int size,
+			    struct fuse_copy_state *cs)
+{
+	struct fuse_notify_poll_wakeup_out outarg;
+	int err;
+
+	if (size != sizeof(outarg))
+		return -EINVAL;
+
+	err = fuse_copy_one(cs, &outarg, sizeof(outarg));
+	if (err)
+		return err;
+
+	return fuse_notify_poll_wakeup(fc, &outarg);
+}
+
+static int fuse_notify(struct fuse_conn *fc, enum fuse_notify_code code,
+		       unsigned int size, struct fuse_copy_state *cs)
+{
+	switch (code) {
+	case FUSE_NOTIFY_POLL:
+		return fuse_notify_poll(fc, size, cs);
+
+	default:
+		return -EINVAL;
+	}
+}
+
 /* Look up request on processing list by unique ID */
 static struct fuse_req *request_find(struct fuse_conn *fc, u64 unique)
 {
@@ -876,9 +912,23 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 	err = fuse_copy_one(&cs, &oh, sizeof(oh));
 	if (err)
 		goto err_finish;
+
+	err = -EINVAL;
+	if (oh.len != nbytes)
+		goto err_finish;
+
+	/*
+	 * Zero oh.unique indicates unsolicited notification message
+	 * and error contains notification code.
+	 */
+	if (!oh.unique) {
+		err = fuse_notify(fc, oh.error, nbytes - sizeof(oh), &cs);
+		fuse_copy_finish(&cs);
+		return err ? err : nbytes;
+	}
+
 	err = -EINVAL;
-	if (!oh.unique || oh.error <= -1000 || oh.error > 0 ||
-	    oh.len != nbytes)
+	if (oh.error <= -1000 || oh.error > 0)
 		goto err_finish;
 
 	spin_lock(&fc->lock);
@@ -966,6 +1016,8 @@ static unsigned fuse_dev_poll(struct file *file, poll_table *wait)
  * This function releases and reacquires fc->lock
  */
 static void end_requests(struct fuse_conn *fc, struct list_head *head)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	while (!list_empty(head)) {
 		struct fuse_req *req;
@@ -988,7 +1040,8 @@ static void end_requests(struct fuse_conn *fc, struct list_head *head)
  * locked).
  */
 static void end_io_requests(struct fuse_conn *fc)
-	__releases(fc->lock) __acquires(fc->lock)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	while (!list_empty(&fc->io)) {
 		struct fuse_req *req =
@@ -1002,11 +1055,11 @@ static void end_io_requests(struct fuse_conn *fc)
 		wake_up(&req->waitq);
 		if (end) {
 			req->end = NULL;
-			/* The end function will consume this reference */
 			__fuse_get_request(req);
 			spin_unlock(&fc->lock);
 			wait_event(req->waitq, !req->locked);
 			end(fc, req);
+			fuse_put_request(fc, req);
 			spin_lock(&fc->lock);
 		}
 	}
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 95bc22bdd06..fdff346e96f 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -189,7 +189,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 		parent = dget_parent(entry);
 		fuse_lookup_init(fc, req, get_node_id(parent->d_inode),
 				 &entry->d_name, &outarg);
-		request_send(fc, req);
+		fuse_request_send(fc, req);
 		dput(parent);
 		err = req->out.h.error;
 		fuse_put_request(fc, req);
@@ -204,7 +204,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 				return 0;
 			}
 			spin_lock(&fc->lock);
-			fi->nlookup ++;
+			fi->nlookup++;
 			spin_unlock(&fc->lock);
 		}
 		fuse_put_request(fc, forget_req);
@@ -283,7 +283,7 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
 	attr_version = fuse_get_attr_version(fc);
 
 	fuse_lookup_init(fc, req, nodeid, name, outarg);
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	/* Zero nodeid is same as -ENOENT, but with valid timeout */
@@ -369,7 +369,7 @@ static void fuse_sync_release(struct fuse_conn *fc, struct fuse_file *ff,
 {
 	fuse_release_fill(ff, nodeid, flags, FUSE_RELEASE);
 	ff->reserved_req->force = 1;
-	request_send(fc, ff->reserved_req);
+	fuse_request_send(fc, ff->reserved_req);
 	fuse_put_request(fc, ff->reserved_req);
 	kfree(ff);
 }
@@ -408,7 +408,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
 		goto out_put_forget_req;
 
 	err = -ENOMEM;
-	ff = fuse_file_alloc();
+	ff = fuse_file_alloc(fc);
 	if (!ff)
 		goto out_put_request;
 
@@ -432,7 +432,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
 	req->out.args[0].value = &outentry;
 	req->out.args[1].size = sizeof(outopen);
 	req->out.args[1].value = &outopen;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	if (err) {
 		if (err == -ENOSYS)
@@ -502,7 +502,7 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
 	else
 		req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err)
@@ -631,15 +631,17 @@ static int fuse_unlink(struct inode *dir, struct dentry *entry)
 	req->in.numargs = 1;
 	req->in.args[0].size = entry->d_name.len + 1;
 	req->in.args[0].value = entry->d_name.name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
 		struct inode *inode = entry->d_inode;
 
-		/* Set nlink to zero so the inode can be cleared, if
-                   the inode does have more links this will be
-                   discovered at the next lookup/getattr */
+		/*
+		 * Set nlink to zero so the inode can be cleared, if the inode
+		 * does have more links this will be discovered at the next
+		 * lookup/getattr.
+		 */
 		clear_nlink(inode);
 		fuse_invalidate_attr(inode);
 		fuse_invalidate_attr(dir);
@@ -662,7 +664,7 @@ static int fuse_rmdir(struct inode *dir, struct dentry *entry)
 	req->in.numargs = 1;
 	req->in.args[0].size = entry->d_name.len + 1;
 	req->in.args[0].value = entry->d_name.name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
@@ -695,7 +697,7 @@ static int fuse_rename(struct inode *olddir, struct dentry *oldent,
 	req->in.args[1].value = oldent->d_name.name;
 	req->in.args[2].size = newent->d_name.len + 1;
 	req->in.args[2].value = newent->d_name.name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
@@ -811,7 +813,7 @@ static int fuse_do_getattr(struct inode *inode, struct kstat *stat,
 	else
 		req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
@@ -911,7 +913,7 @@ static int fuse_access(struct inode *inode, int mask)
 	req->in.numargs = 1;
 	req->in.args[0].size = sizeof(inarg);
 	req->in.args[0].value = &inarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
@@ -1033,7 +1035,7 @@ static int fuse_readdir(struct file *file, void *dstbuf, filldir_t filldir)
 	req->num_pages = 1;
 	req->pages[0] = page;
 	fuse_read_fill(req, file, inode, file->f_pos, PAGE_SIZE, FUSE_READDIR);
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	nbytes = req->out.args[0].size;
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
@@ -1067,7 +1069,7 @@ static char *read_link(struct dentry *dentry)
 	req->out.numargs = 1;
 	req->out.args[0].size = PAGE_SIZE - 1;
 	req->out.args[0].value = link;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	if (req->out.h.error) {
 		free_page((unsigned long) link);
 		link = ERR_PTR(req->out.h.error);
@@ -1273,7 +1275,7 @@ static int fuse_do_setattr(struct dentry *entry, struct iattr *attr,
 	else
 		req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err) {
@@ -1367,7 +1369,7 @@ static int fuse_setxattr(struct dentry *entry, const char *name,
 	req->in.args[1].value = name;
 	req->in.args[2].size = size;
 	req->in.args[2].value = value;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
@@ -1413,7 +1415,7 @@ static ssize_t fuse_getxattr(struct dentry *entry, const char *name,
 		req->out.args[0].size = sizeof(outarg);
 		req->out.args[0].value = &outarg;
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	ret = req->out.h.error;
 	if (!ret)
 		ret = size ? req->out.args[0].size : outarg.size;
@@ -1463,7 +1465,7 @@ static ssize_t fuse_listxattr(struct dentry *entry, char *list, size_t size)
 		req->out.args[0].size = sizeof(outarg);
 		req->out.args[0].value = &outarg;
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	ret = req->out.h.error;
 	if (!ret)
 		ret = size ? req->out.args[0].size : outarg.size;
@@ -1496,7 +1498,7 @@ static int fuse_removexattr(struct dentry *entry, const char *name)
 	req->in.numargs = 1;
 	req->in.args[0].size = strlen(name) + 1;
 	req->in.args[0].value = name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 34930a964b8..e8162646a9b 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -39,14 +39,14 @@ static int fuse_send_open(struct inode *inode, struct file *file, int isdir,
 	req->out.numargs = 1;
 	req->out.args[0].size = sizeof(*outargp);
 	req->out.args[0].value = outargp;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 
 	return err;
 }
 
-struct fuse_file *fuse_file_alloc(void)
+struct fuse_file *fuse_file_alloc(struct fuse_conn *fc)
 {
 	struct fuse_file *ff;
 	ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
@@ -58,7 +58,12 @@ struct fuse_file *fuse_file_alloc(void)
 		} else {
 			INIT_LIST_HEAD(&ff->write_entry);
 			atomic_set(&ff->count, 0);
+			spin_lock(&fc->lock);
+			ff->kh = ++fc->khctr;
+			spin_unlock(&fc->lock);
 		}
+		RB_CLEAR_NODE(&ff->polled_node);
+		init_waitqueue_head(&ff->poll_wait);
 	}
 	return ff;
 }
@@ -79,7 +84,6 @@ static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
 {
 	dput(req->misc.release.dentry);
 	mntput(req->misc.release.vfsmount);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_file_put(struct fuse_file *ff)
@@ -89,7 +93,7 @@ static void fuse_file_put(struct fuse_file *ff)
 		struct inode *inode = req->misc.release.dentry->d_inode;
 		struct fuse_conn *fc = get_fuse_conn(inode);
 		req->end = fuse_release_end;
-		request_send_background(fc, req);
+		fuse_request_send_background(fc, req);
 		kfree(ff);
 	}
 }
@@ -109,6 +113,7 @@ void fuse_finish_open(struct inode *inode, struct file *file,
 
 int fuse_open_common(struct inode *inode, struct file *file, int isdir)
 {
+	struct fuse_conn *fc = get_fuse_conn(inode);
 	struct fuse_open_out outarg;
 	struct fuse_file *ff;
 	int err;
@@ -121,7 +126,7 @@ int fuse_open_common(struct inode *inode, struct file *file, int isdir)
 	if (err)
 		return err;
 
-	ff = fuse_file_alloc();
+	ff = fuse_file_alloc(fc);
 	if (!ff)
 		return -ENOMEM;
 
@@ -167,7 +172,11 @@ int fuse_release_common(struct inode *inode, struct file *file, int isdir)
 
 		spin_lock(&fc->lock);
 		list_del(&ff->write_entry);
+		if (!RB_EMPTY_NODE(&ff->polled_node))
+			rb_erase(&ff->polled_node, &fc->polled_files);
 		spin_unlock(&fc->lock);
+
+		wake_up_interruptible_sync(&ff->poll_wait);
 		/*
 		 * Normally this will send the RELEASE request,
 		 * however if some asynchronous READ or WRITE requests
@@ -280,7 +289,7 @@ static int fuse_flush(struct file *file, fl_owner_t id)
 	req->in.args[0].size = sizeof(inarg);
 	req->in.args[0].value = &inarg;
 	req->force = 1;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
@@ -344,7 +353,7 @@ int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
 	req->in.numargs = 1;
 	req->in.args[0].size = sizeof(inarg);
 	req->in.args[0].value = &inarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
@@ -396,7 +405,7 @@ static size_t fuse_send_read(struct fuse_req *req, struct file *file,
 		inarg->read_flags |= FUSE_READ_LOCKOWNER;
 		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	return req->out.args[0].size;
 }
 
@@ -493,7 +502,6 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
 	}
 	if (req->ff)
 		fuse_file_put(req->ff);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_send_readpages(struct fuse_req *req, struct file *file,
@@ -509,10 +517,11 @@ static void fuse_send_readpages(struct fuse_req *req, struct file *file,
 		struct fuse_file *ff = file->private_data;
 		req->ff = fuse_file_get(ff);
 		req->end = fuse_readpages_end;
-		request_send_background(fc, req);
+		fuse_request_send_background(fc, req);
 	} else {
-		request_send(fc, req);
+		fuse_request_send(fc, req);
 		fuse_readpages_end(fc, req);
+		fuse_put_request(fc, req);
 	}
 }
 
@@ -543,7 +552,7 @@ static int fuse_readpages_fill(void *_data, struct page *page)
 		}
 	}
 	req->pages[req->num_pages] = page;
-	req->num_pages ++;
+	req->num_pages++;
 	return 0;
 }
 
@@ -636,7 +645,7 @@ static size_t fuse_send_write(struct fuse_req *req, struct file *file,
 		inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
 		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	return req->misc.write.out.size;
 }
 
@@ -646,7 +655,7 @@ static int fuse_write_begin(struct file *file, struct address_space *mapping,
 {
 	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
 
-	*pagep = __grab_cache_page(mapping, index);
+	*pagep = grab_cache_page_write_begin(mapping, index, flags);
 	if (!*pagep)
 		return -ENOMEM;
 	return 0;
@@ -779,7 +788,7 @@ static ssize_t fuse_fill_write_pages(struct fuse_req *req,
 			break;
 
 		err = -ENOMEM;
-		page = __grab_cache_page(mapping, index);
+		page = grab_cache_page_write_begin(mapping, index, 0);
 		if (!page)
 			break;
 
@@ -1042,7 +1051,6 @@ static void fuse_writepage_free(struct fuse_conn *fc, struct fuse_req *req)
 {
 	__free_page(req->pages[0]);
 	fuse_file_put(req->ff);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
@@ -1060,6 +1068,8 @@ static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
 
 /* Called under fc->lock, may release and reacquire it */
 static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	struct fuse_inode *fi = get_fuse_inode(req->inode);
 	loff_t size = i_size_read(req->inode);
@@ -1079,13 +1089,14 @@ static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
 
 	req->in.args[1].size = inarg->size;
 	fi->writectr++;
-	request_send_background_locked(fc, req);
+	fuse_request_send_background_locked(fc, req);
 	return;
 
  out_free:
 	fuse_writepage_finish(fc, req);
 	spin_unlock(&fc->lock);
 	fuse_writepage_free(fc, req);
+	fuse_put_request(fc, req);
 	spin_lock(&fc->lock);
 }
 
@@ -1096,6 +1107,8 @@ static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
  * Called with fc->lock
  */
 void fuse_flush_writepages(struct inode *inode)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	struct fuse_conn *fc = get_fuse_conn(inode);
 	struct fuse_inode *fi = get_fuse_inode(inode);
@@ -1325,7 +1338,7 @@ static int fuse_getlk(struct file *file, struct file_lock *fl)
 	req->out.numargs = 1;
 	req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err)
@@ -1357,7 +1370,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
 		return PTR_ERR(req);
 
 	fuse_lk_fill(req, file, fl, opcode, pid, flock);
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	/* locking is restartable */
 	if (err == -EINTR)
@@ -1433,7 +1446,7 @@ static sector_t fuse_bmap(struct address_space *mapping, sector_t block)
 	req->out.numargs = 1;
 	req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS)
@@ -1470,6 +1483,406 @@ static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin)
 	return retval;
 }
 
+static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
+			unsigned int nr_segs, size_t bytes, bool to_user)
+{
+	struct iov_iter ii;
+	int page_idx = 0;
+
+	if (!bytes)
+		return 0;
+
+	iov_iter_init(&ii, iov, nr_segs, bytes, 0);
+
+	while (iov_iter_count(&ii)) {
+		struct page *page = pages[page_idx++];
+		size_t todo = min_t(size_t, PAGE_SIZE, iov_iter_count(&ii));
+		void *kaddr, *map;
+
+		kaddr = map = kmap(page);
+
+		while (todo) {
+			char __user *uaddr = ii.iov->iov_base + ii.iov_offset;
+			size_t iov_len = ii.iov->iov_len - ii.iov_offset;
+			size_t copy = min(todo, iov_len);
+			size_t left;
+
+			if (!to_user)
+				left = copy_from_user(kaddr, uaddr, copy);
+			else
+				left = copy_to_user(uaddr, kaddr, copy);
+
+			if (unlikely(left))
+				return -EFAULT;
+
+			iov_iter_advance(&ii, copy);
+			todo -= copy;
+			kaddr += copy;
+		}
+
+		kunmap(map);
+	}
+
+	return 0;
+}
+
+/*
+ * For ioctls, there is no generic way to determine how much memory
+ * needs to be read and/or written.  Furthermore, ioctls are allowed
+ * to dereference the passed pointer, so the parameter requires deep
+ * copying but FUSE has no idea whatsoever about what to copy in or
+ * out.
+ *
+ * This is solved by allowing FUSE server to retry ioctl with
+ * necessary in/out iovecs.  Let's assume the ioctl implementation
+ * needs to read in the following structure.
+ *
+ * struct a {
+ *	char	*buf;
+ *	size_t	buflen;
+ * }
+ *
+ * On the first callout to FUSE server, inarg->in_size and
+ * inarg->out_size will be NULL; then, the server completes the ioctl
+ * with FUSE_IOCTL_RETRY set in out->flags, out->in_iovs set to 1 and
+ * the actual iov array to
+ *
+ * { { .iov_base = inarg.arg,	.iov_len = sizeof(struct a) } }
+ *
+ * which tells FUSE to copy in the requested area and retry the ioctl.
+ * On the second round, the server has access to the structure and
+ * from that it can tell what to look for next, so on the invocation,
+ * it sets FUSE_IOCTL_RETRY, out->in_iovs to 2 and iov array to
+ *
+ * { { .iov_base = inarg.arg,	.iov_len = sizeof(struct a)	},
+ *   { .iov_base = a.buf,	.iov_len = a.buflen		} }
+ *
+ * FUSE will copy both struct a and the pointed buffer from the
+ * process doing the ioctl and retry ioctl with both struct a and the
+ * buffer.
+ *
+ * This time, FUSE server has everything it needs and completes ioctl
+ * without FUSE_IOCTL_RETRY which finishes the ioctl call.
+ *
+ * Copying data out works the same way.
+ *
+ * Note that if FUSE_IOCTL_UNRESTRICTED is clear, the kernel
+ * automatically initializes in and out iovs by decoding @cmd with
+ * _IOC_* macros and the server is not allowed to request RETRY.  This
+ * limits ioctl data transfers to well-formed ioctls and is the forced
+ * behavior for all FUSE servers.
+ */
+static long fuse_file_do_ioctl(struct file *file, unsigned int cmd,
+			       unsigned long arg, unsigned int flags)
+{
+	struct inode *inode = file->f_dentry->d_inode;
+	struct fuse_file *ff = file->private_data;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_ioctl_in inarg = {
+		.fh = ff->fh,
+		.cmd = cmd,
+		.arg = arg,
+		.flags = flags
+	};
+	struct fuse_ioctl_out outarg;
+	struct fuse_req *req = NULL;
+	struct page **pages = NULL;
+	struct page *iov_page = NULL;
+	struct iovec *in_iov = NULL, *out_iov = NULL;
+	unsigned int in_iovs = 0, out_iovs = 0, num_pages = 0, max_pages;
+	size_t in_size, out_size, transferred;
+	int err;
+
+	/* assume all the iovs returned by client always fits in a page */
+	BUILD_BUG_ON(sizeof(struct iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE);
+
+	if (!fuse_allow_task(fc, current))
+		return -EACCES;
+
+	err = -EIO;
+	if (is_bad_inode(inode))
+		goto out;
+
+	err = -ENOMEM;
+	pages = kzalloc(sizeof(pages[0]) * FUSE_MAX_PAGES_PER_REQ, GFP_KERNEL);
+	iov_page = alloc_page(GFP_KERNEL);
+	if (!pages || !iov_page)
+		goto out;
+
+	/*
+	 * If restricted, initialize IO parameters as encoded in @cmd.
+	 * RETRY from server is not allowed.
+	 */
+	if (!(flags & FUSE_IOCTL_UNRESTRICTED)) {
+		struct iovec *iov = page_address(iov_page);
+
+		iov->iov_base = (void __user *)arg;
+		iov->iov_len = _IOC_SIZE(cmd);
+
+		if (_IOC_DIR(cmd) & _IOC_WRITE) {
+			in_iov = iov;
+			in_iovs = 1;
+		}
+
+		if (_IOC_DIR(cmd) & _IOC_READ) {
+			out_iov = iov;
+			out_iovs = 1;
+		}
+	}
+
+ retry:
+	inarg.in_size = in_size = iov_length(in_iov, in_iovs);
+	inarg.out_size = out_size = iov_length(out_iov, out_iovs);
+
+	/*
+	 * Out data can be used either for actual out data or iovs,
+	 * make sure there always is at least one page.
+	 */
+	out_size = max_t(size_t, out_size, PAGE_SIZE);
+	max_pages = DIV_ROUND_UP(max(in_size, out_size), PAGE_SIZE);
+
+	/* make sure there are enough buffer pages and init request with them */
+	err = -ENOMEM;
+	if (max_pages > FUSE_MAX_PAGES_PER_REQ)
+		goto out;
+	while (num_pages < max_pages) {
+		pages[num_pages] = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
+		if (!pages[num_pages])
+			goto out;
+		num_pages++;
+	}
+
+	req = fuse_get_req(fc);
+	if (IS_ERR(req)) {
+		err = PTR_ERR(req);
+		req = NULL;
+		goto out;
+	}
+	memcpy(req->pages, pages, sizeof(req->pages[0]) * num_pages);
+	req->num_pages = num_pages;
+
+	/* okay, let's send it to the client */
+	req->in.h.opcode = FUSE_IOCTL;
+	req->in.h.nodeid = get_node_id(inode);
+	req->in.numargs = 1;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	if (in_size) {
+		req->in.numargs++;
+		req->in.args[1].size = in_size;
+		req->in.argpages = 1;
+
+		err = fuse_ioctl_copy_user(pages, in_iov, in_iovs, in_size,
+					   false);
+		if (err)
+			goto out;
+	}
+
+	req->out.numargs = 2;
+	req->out.args[0].size = sizeof(outarg);
+	req->out.args[0].value = &outarg;
+	req->out.args[1].size = out_size;
+	req->out.argpages = 1;
+	req->out.argvar = 1;
+
+	fuse_request_send(fc, req);
+	err = req->out.h.error;
+	transferred = req->out.args[1].size;
+	fuse_put_request(fc, req);
+	req = NULL;
+	if (err)
+		goto out;
+
+	/* did it ask for retry? */
+	if (outarg.flags & FUSE_IOCTL_RETRY) {
+		char *vaddr;
+
+		/* no retry if in restricted mode */
+		err = -EIO;
+		if (!(flags & FUSE_IOCTL_UNRESTRICTED))
+			goto out;
+
+		in_iovs = outarg.in_iovs;
+		out_iovs = outarg.out_iovs;
+
+		/*
+		 * Make sure things are in boundary, separate checks
+		 * are to protect against overflow.
+		 */
+		err = -ENOMEM;
+		if (in_iovs > FUSE_IOCTL_MAX_IOV ||
+		    out_iovs > FUSE_IOCTL_MAX_IOV ||
+		    in_iovs + out_iovs > FUSE_IOCTL_MAX_IOV)
+			goto out;
+
+		err = -EIO;
+		if ((in_iovs + out_iovs) * sizeof(struct iovec) != transferred)
+			goto out;
+
+		/* okay, copy in iovs and retry */
+		vaddr = kmap_atomic(pages[0], KM_USER0);
+		memcpy(page_address(iov_page), vaddr, transferred);
+		kunmap_atomic(vaddr, KM_USER0);
+
+		in_iov = page_address(iov_page);
+		out_iov = in_iov + in_iovs;
+
+		goto retry;
+	}
+
+	err = -EIO;
+	if (transferred > inarg.out_size)
+		goto out;
+
+	err = fuse_ioctl_copy_user(pages, out_iov, out_iovs, transferred, true);
+ out:
+	if (req)
+		fuse_put_request(fc, req);
+	if (iov_page)
+		__free_page(iov_page);
+	while (num_pages)
+		__free_page(pages[--num_pages]);
+	kfree(pages);
+
+	return err ? err : outarg.result;
+}
+
+static long fuse_file_ioctl(struct file *file, unsigned int cmd,
+			    unsigned long arg)
+{
+	return fuse_file_do_ioctl(file, cmd, arg, 0);
+}
+
+static long fuse_file_compat_ioctl(struct file *file, unsigned int cmd,
+				   unsigned long arg)
+{
+	return fuse_file_do_ioctl(file, cmd, arg, FUSE_IOCTL_COMPAT);
+}
+
+/*
+ * All files which have been polled are linked to RB tree
+ * fuse_conn->polled_files which is indexed by kh.  Walk the tree and
+ * find the matching one.
+ */
+static struct rb_node **fuse_find_polled_node(struct fuse_conn *fc, u64 kh,
+					      struct rb_node **parent_out)
+{
+	struct rb_node **link = &fc->polled_files.rb_node;
+	struct rb_node *last = NULL;
+
+	while (*link) {
+		struct fuse_file *ff;
+
+		last = *link;
+		ff = rb_entry(last, struct fuse_file, polled_node);
+
+		if (kh < ff->kh)
+			link = &last->rb_left;
+		else if (kh > ff->kh)
+			link = &last->rb_right;
+		else
+			return link;
+	}
+
+	if (parent_out)
+		*parent_out = last;
+	return link;
+}
+
+/*
+ * The file is about to be polled.  Make sure it's on the polled_files
+ * RB tree.  Note that files once added to the polled_files tree are
+ * not removed before the file is released.  This is because a file
+ * polled once is likely to be polled again.
+ */
+static void fuse_register_polled_file(struct fuse_conn *fc,
+				      struct fuse_file *ff)
+{
+	spin_lock(&fc->lock);
+	if (RB_EMPTY_NODE(&ff->polled_node)) {
+		struct rb_node **link, *parent;
+
+		link = fuse_find_polled_node(fc, ff->kh, &parent);
+		BUG_ON(*link);
+		rb_link_node(&ff->polled_node, parent, link);
+		rb_insert_color(&ff->polled_node, &fc->polled_files);
+	}
+	spin_unlock(&fc->lock);
+}
+
+static unsigned fuse_file_poll(struct file *file, poll_table *wait)
+{
+	struct inode *inode = file->f_dentry->d_inode;
+	struct fuse_file *ff = file->private_data;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_poll_in inarg = { .fh = ff->fh, .kh = ff->kh };
+	struct fuse_poll_out outarg;
+	struct fuse_req *req;
+	int err;
+
+	if (fc->no_poll)
+		return DEFAULT_POLLMASK;
+
+	poll_wait(file, &ff->poll_wait, wait);
+
+	/*
+	 * Ask for notification iff there's someone waiting for it.
+	 * The client may ignore the flag and always notify.
+	 */
+	if (waitqueue_active(&ff->poll_wait)) {
+		inarg.flags |= FUSE_POLL_SCHEDULE_NOTIFY;
+		fuse_register_polled_file(fc, ff);
+	}
+
+	req = fuse_get_req(fc);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	req->in.h.opcode = FUSE_POLL;
+	req->in.h.nodeid = get_node_id(inode);
+	req->in.numargs = 1;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->out.numargs = 1;
+	req->out.args[0].size = sizeof(outarg);
+	req->out.args[0].value = &outarg;
+	fuse_request_send(fc, req);
+	err = req->out.h.error;
+	fuse_put_request(fc, req);
+
+	if (!err)
+		return outarg.revents;
+	if (err == -ENOSYS) {
+		fc->no_poll = 1;
+		return DEFAULT_POLLMASK;
+	}
+	return POLLERR;
+}
+
+/*
+ * This is called from fuse_handle_notify() on FUSE_NOTIFY_POLL and
+ * wakes up the poll waiters.
+ */
+int fuse_notify_poll_wakeup(struct fuse_conn *fc,
+			    struct fuse_notify_poll_wakeup_out *outarg)
+{
+	u64 kh = outarg->kh;
+	struct rb_node **link;
+
+	spin_lock(&fc->lock);
+
+	link = fuse_find_polled_node(fc, kh, NULL);
+	if (*link) {
+		struct fuse_file *ff;
+
+		ff = rb_entry(*link, struct fuse_file, polled_node);
+		wake_up_interruptible_sync(&ff->poll_wait);
+	}
+
+	spin_unlock(&fc->lock);
+	return 0;
+}
+
 static const struct file_operations fuse_file_operations = {
 	.llseek		= fuse_file_llseek,
 	.read		= do_sync_read,
@@ -1484,6 +1897,9 @@ static const struct file_operations fuse_file_operations = {
 	.lock		= fuse_file_lock,
 	.flock		= fuse_file_flock,
 	.splice_read	= generic_file_splice_read,
+	.unlocked_ioctl	= fuse_file_ioctl,
+	.compat_ioctl	= fuse_file_compat_ioctl,
+	.poll		= fuse_file_poll,
 };
 
 static const struct file_operations fuse_direct_io_file_operations = {
@@ -1496,6 +1912,9 @@ static const struct file_operations fuse_direct_io_file_operations = {
 	.fsync		= fuse_fsync,
 	.lock		= fuse_file_lock,
 	.flock		= fuse_file_flock,
+	.unlocked_ioctl	= fuse_file_ioctl,
+	.compat_ioctl	= fuse_file_compat_ioctl,
+	.poll		= fuse_file_poll,
 	/* no mmap and splice_read */
 };
 
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 35accfdd747..5e64b815a5a 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -19,6 +19,8 @@
 #include <linux/backing-dev.h>
 #include <linux/mutex.h>
 #include <linux/rwsem.h>
+#include <linux/rbtree.h>
+#include <linux/poll.h>
 
 /** Max number of pages that can be used in a single read request */
 #define FUSE_MAX_PAGES_PER_REQ 32
@@ -100,6 +102,9 @@ struct fuse_file {
 	/** Request reserved for flush and release */
 	struct fuse_req *reserved_req;
 
+	/** Kernel file handle guaranteed to be unique */
+	u64 kh;
+
 	/** File handle used by userspace */
 	u64 fh;
 
@@ -108,6 +113,12 @@ struct fuse_file {
 
 	/** Entry on inode's write_files list */
 	struct list_head write_entry;
+
+	/** RB node to be linked on fuse_conn->polled_files */
+	struct rb_node polled_node;
+
+	/** Wait queue head for poll */
+	wait_queue_head_t poll_wait;
 };
 
 /** One input argument of a request */
@@ -322,6 +333,12 @@ struct fuse_conn {
 	/** The list of requests under I/O */
 	struct list_head io;
 
+	/** The next unique kernel file handle */
+	u64 khctr;
+
+	/** rbtree of fuse_files waiting for poll events indexed by ph */
+	struct rb_root polled_files;
+
 	/** Number of requests currently in the background */
 	unsigned num_background;
 
@@ -355,19 +372,19 @@ struct fuse_conn {
 	/** Connection failed (version mismatch).  Cannot race with
 	    setting other bitfields since it is only set once in INIT
 	    reply, before any other request, and never cleared */
-	unsigned conn_error : 1;
+	unsigned conn_error:1;
 
 	/** Connection successful.  Only set in INIT */
-	unsigned conn_init : 1;
+	unsigned conn_init:1;
 
 	/** Do readpages asynchronously?  Only set in INIT */
-	unsigned async_read : 1;
+	unsigned async_read:1;
 
 	/** Do not send separate SETATTR request before open(O_TRUNC)  */
-	unsigned atomic_o_trunc : 1;
+	unsigned atomic_o_trunc:1;
 
 	/** Filesystem supports NFS exporting.  Only set in INIT */
-	unsigned export_support : 1;
+	unsigned export_support:1;
 
 	/*
 	 * The following bitfields are only for optimization purposes
@@ -375,43 +392,46 @@ struct fuse_conn {
 	 */
 
 	/** Is fsync not implemented by fs? */
-	unsigned no_fsync : 1;
+	unsigned no_fsync:1;
 
 	/** Is fsyncdir not implemented by fs? */
-	unsigned no_fsyncdir : 1;
+	unsigned no_fsyncdir:1;
 
 	/** Is flush not implemented by fs? */
-	unsigned no_flush : 1;
+	unsigned no_flush:1;
 
 	/** Is setxattr not implemented by fs? */
-	unsigned no_setxattr : 1;
+	unsigned no_setxattr:1;
 
 	/** Is getxattr not implemented by fs? */
-	unsigned no_getxattr : 1;
+	unsigned no_getxattr:1;
 
 	/** Is listxattr not implemented by fs? */
-	unsigned no_listxattr : 1;
+	unsigned no_listxattr:1;
 
 	/** Is removexattr not implemented by fs? */
-	unsigned no_removexattr : 1;
+	unsigned no_removexattr:1;
 
 	/** Are file locking primitives not implemented by fs? */
-	unsigned no_lock : 1;
+	unsigned no_lock:1;
 
 	/** Is access not implemented by fs? */
-	unsigned no_access : 1;
+	unsigned no_access:1;
 
 	/** Is create not implemented by fs? */
-	unsigned no_create : 1;
+	unsigned no_create:1;
 
 	/** Is interrupt not implemented by fs? */
-	unsigned no_interrupt : 1;
+	unsigned no_interrupt:1;
 
 	/** Is bmap not implemented by fs? */
-	unsigned no_bmap : 1;
+	unsigned no_bmap:1;
+
+	/** Is poll not implemented by fs? */
+	unsigned no_poll:1;
 
 	/** Do multi-page cached writes */
-	unsigned big_writes : 1;
+	unsigned big_writes:1;
 
 	/** The number of requests waiting for completion */
 	atomic_t num_waiting;
@@ -445,6 +465,9 @@ struct fuse_conn {
 
 	/** Version counter for attribute changes */
 	u64 attr_version;
+
+	/** Called on final put */
+	void (*release)(struct fuse_conn *);
 };
 
 static inline struct fuse_conn *get_fuse_conn_super(struct super_block *sb)
@@ -499,7 +522,7 @@ void fuse_read_fill(struct fuse_req *req, struct file *file,
  */
 int fuse_open_common(struct inode *inode, struct file *file, int isdir);
 
-struct fuse_file *fuse_file_alloc(void);
+struct fuse_file *fuse_file_alloc(struct fuse_conn *fc);
 void fuse_file_free(struct fuse_file *ff);
 void fuse_finish_open(struct inode *inode, struct file *file,
 		      struct fuse_file *ff, struct fuse_open_out *outarg);
@@ -519,6 +542,12 @@ int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
 		      int isdir);
 
 /**
+ * Notify poll wakeup
+ */
+int fuse_notify_poll_wakeup(struct fuse_conn *fc,
+			    struct fuse_notify_poll_wakeup_out *outarg);
+
+/**
  * Initialize file operations on a regular file
  */
 void fuse_init_file_inode(struct inode *inode);
@@ -593,19 +622,20 @@ void fuse_put_request(struct fuse_conn *fc, struct fuse_req *req);
 /**
  * Send a request (synchronous)
  */
-void request_send(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send(struct fuse_conn *fc, struct fuse_req *req);
 
 /**
  * Send a request with no reply
  */
-void request_send_noreply(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send_noreply(struct fuse_conn *fc, struct fuse_req *req);
 
 /**
  * Send a request in the background
  */
-void request_send_background(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send_background(struct fuse_conn *fc, struct fuse_req *req);
 
-void request_send_background_locked(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send_background_locked(struct fuse_conn *fc,
+					 struct fuse_req *req);
 
 /* Abort all requests */
 void fuse_abort_conn(struct fuse_conn *fc);
@@ -623,6 +653,11 @@ void fuse_invalidate_entry_cache(struct dentry *entry);
 struct fuse_conn *fuse_conn_get(struct fuse_conn *fc);
 
 /**
+ * Initialize fuse_conn
+ */
+int fuse_conn_init(struct fuse_conn *fc, struct super_block *sb);
+
+/**
  * Release reference to fuse_conn
  */
 void fuse_conn_put(struct fuse_conn *fc);
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 2e99f34b443..47c96fdca1a 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -37,10 +37,10 @@ struct fuse_mount_data {
 	unsigned rootmode;
 	unsigned user_id;
 	unsigned group_id;
-	unsigned fd_present : 1;
-	unsigned rootmode_present : 1;
-	unsigned user_id_present : 1;
-	unsigned group_id_present : 1;
+	unsigned fd_present:1;
+	unsigned rootmode_present:1;
+	unsigned user_id_present:1;
+	unsigned group_id_present:1;
 	unsigned flags;
 	unsigned max_read;
 	unsigned blksize;
@@ -94,7 +94,7 @@ void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
 	req->in.numargs = 1;
 	req->in.args[0].size = sizeof(struct fuse_forget_in);
 	req->in.args[0].value = inarg;
-	request_send_noreply(fc, req);
+	fuse_request_send_noreply(fc, req);
 }
 
 static void fuse_clear_inode(struct inode *inode)
@@ -250,7 +250,7 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid,
 
 	fi = get_fuse_inode(inode);
 	spin_lock(&fc->lock);
-	fi->nlookup ++;
+	fi->nlookup++;
 	spin_unlock(&fc->lock);
 	fuse_change_attributes(inode, attr, attr_valid, attr_version);
 
@@ -269,7 +269,7 @@ static void fuse_send_destroy(struct fuse_conn *fc)
 		fc->destroy_req = NULL;
 		req->in.h.opcode = FUSE_DESTROY;
 		req->force = 1;
-		request_send(fc, req);
+		fuse_request_send(fc, req);
 		fuse_put_request(fc, req);
 	}
 }
@@ -334,7 +334,7 @@ static int fuse_statfs(struct dentry *dentry, struct kstatfs *buf)
 	req->out.args[0].size =
 		fc->minor < 4 ? FUSE_COMPAT_STATFS_SIZE : sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	if (!err)
 		convert_fuse_statfs(buf, &outarg.st);
@@ -462,68 +462,69 @@ static int fuse_show_options(struct seq_file *m, struct vfsmount *mnt)
 	return 0;
 }
 
-static struct fuse_conn *new_conn(struct super_block *sb)
+int fuse_conn_init(struct fuse_conn *fc, struct super_block *sb)
 {
-	struct fuse_conn *fc;
 	int err;
 
-	fc = kzalloc(sizeof(*fc), GFP_KERNEL);
-	if (fc) {
-		spin_lock_init(&fc->lock);
-		mutex_init(&fc->inst_mutex);
-		atomic_set(&fc->count, 1);
-		init_waitqueue_head(&fc->waitq);
-		init_waitqueue_head(&fc->blocked_waitq);
-		init_waitqueue_head(&fc->reserved_req_waitq);
-		INIT_LIST_HEAD(&fc->pending);
-		INIT_LIST_HEAD(&fc->processing);
-		INIT_LIST_HEAD(&fc->io);
-		INIT_LIST_HEAD(&fc->interrupts);
-		INIT_LIST_HEAD(&fc->bg_queue);
-		atomic_set(&fc->num_waiting, 0);
-		fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
-		fc->bdi.unplug_io_fn = default_unplug_io_fn;
-		/* fuse does it's own writeback accounting */
-		fc->bdi.capabilities = BDI_CAP_NO_ACCT_WB;
-		fc->dev = sb->s_dev;
-		err = bdi_init(&fc->bdi);
-		if (err)
-			goto error_kfree;
-		if (sb->s_bdev) {
-			err = bdi_register(&fc->bdi, NULL, "%u:%u-fuseblk",
-					   MAJOR(fc->dev), MINOR(fc->dev));
-		} else {
-			err = bdi_register_dev(&fc->bdi, fc->dev);
-		}
-		if (err)
-			goto error_bdi_destroy;
-		/*
-		 * For a single fuse filesystem use max 1% of dirty +
-		 * writeback threshold.
-		 *
-		 * This gives about 1M of write buffer for memory maps on a
-		 * machine with 1G and 10% dirty_ratio, which should be more
-		 * than enough.
-		 *
-		 * Privileged users can raise it by writing to
-		 *
-		 *    /sys/class/bdi/<bdi>/max_ratio
-		 */
-		bdi_set_max_ratio(&fc->bdi, 1);
-		fc->reqctr = 0;
-		fc->blocked = 1;
-		fc->attr_version = 1;
-		get_random_bytes(&fc->scramble_key, sizeof(fc->scramble_key));
+	memset(fc, 0, sizeof(*fc));
+	spin_lock_init(&fc->lock);
+	mutex_init(&fc->inst_mutex);
+	atomic_set(&fc->count, 1);
+	init_waitqueue_head(&fc->waitq);
+	init_waitqueue_head(&fc->blocked_waitq);
+	init_waitqueue_head(&fc->reserved_req_waitq);
+	INIT_LIST_HEAD(&fc->pending);
+	INIT_LIST_HEAD(&fc->processing);
+	INIT_LIST_HEAD(&fc->io);
+	INIT_LIST_HEAD(&fc->interrupts);
+	INIT_LIST_HEAD(&fc->bg_queue);
+	INIT_LIST_HEAD(&fc->entry);
+	atomic_set(&fc->num_waiting, 0);
+	fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
+	fc->bdi.unplug_io_fn = default_unplug_io_fn;
+	/* fuse does it's own writeback accounting */
+	fc->bdi.capabilities = BDI_CAP_NO_ACCT_WB;
+	fc->khctr = 0;
+	fc->polled_files = RB_ROOT;
+	fc->dev = sb->s_dev;
+	err = bdi_init(&fc->bdi);
+	if (err)
+		goto error_mutex_destroy;
+	if (sb->s_bdev) {
+		err = bdi_register(&fc->bdi, NULL, "%u:%u-fuseblk",
+				   MAJOR(fc->dev), MINOR(fc->dev));
+	} else {
+		err = bdi_register_dev(&fc->bdi, fc->dev);
 	}
-	return fc;
+	if (err)
+		goto error_bdi_destroy;
+	/*
+	 * For a single fuse filesystem use max 1% of dirty +
+	 * writeback threshold.
+	 *
+	 * This gives about 1M of write buffer for memory maps on a
+	 * machine with 1G and 10% dirty_ratio, which should be more
+	 * than enough.
+	 *
+	 * Privileged users can raise it by writing to
+	 *
+	 *    /sys/class/bdi/<bdi>/max_ratio
+	 */
+	bdi_set_max_ratio(&fc->bdi, 1);
+	fc->reqctr = 0;
+	fc->blocked = 1;
+	fc->attr_version = 1;
+	get_random_bytes(&fc->scramble_key, sizeof(fc->scramble_key));
 
-error_bdi_destroy:
+	return 0;
+
+ error_bdi_destroy:
 	bdi_destroy(&fc->bdi);
-error_kfree:
+ error_mutex_destroy:
 	mutex_destroy(&fc->inst_mutex);
-	kfree(fc);
-	return NULL;
+	return err;
 }
+EXPORT_SYMBOL_GPL(fuse_conn_init);
 
 void fuse_conn_put(struct fuse_conn *fc)
 {
@@ -532,7 +533,7 @@ void fuse_conn_put(struct fuse_conn *fc)
 			fuse_request_free(fc->destroy_req);
 		mutex_destroy(&fc->inst_mutex);
 		bdi_destroy(&fc->bdi);
-		kfree(fc);
+		fc->release(fc);
 	}
 }
 
@@ -542,7 +543,7 @@ struct fuse_conn *fuse_conn_get(struct fuse_conn *fc)
 	return fc;
 }
 
-static struct inode *get_root_inode(struct super_block *sb, unsigned mode)
+static struct inode *fuse_get_root_inode(struct super_block *sb, unsigned mode)
 {
 	struct fuse_attr attr;
 	memset(&attr, 0, sizeof(attr));
@@ -553,8 +554,7 @@ static struct inode *get_root_inode(struct super_block *sb, unsigned mode)
 	return fuse_iget(sb, 1, 0, &attr, 0, 0);
 }
 
-struct fuse_inode_handle
-{
+struct fuse_inode_handle {
 	u64 nodeid;
 	u32 generation;
 };
@@ -761,7 +761,6 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
 		fc->max_write = max_t(unsigned, 4096, fc->max_write);
 		fc->conn_init = 1;
 	}
-	fuse_put_request(fc, req);
 	fc->blocked = 0;
 	wake_up_all(&fc->blocked_waitq);
 }
@@ -787,7 +786,12 @@ static void fuse_send_init(struct fuse_conn *fc, struct fuse_req *req)
 	req->out.args[0].size = sizeof(struct fuse_init_out);
 	req->out.args[0].value = &req->misc.init_out;
 	req->end = process_init_reply;
-	request_send_background(fc, req);
+	fuse_request_send_background(fc, req);
+}
+
+static void fuse_free_conn(struct fuse_conn *fc)
+{
+	kfree(fc);
 }
 
 static int fuse_fill_super(struct super_block *sb, void *data, int silent)
@@ -828,10 +832,17 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
 	if (file->f_op != &fuse_dev_operations)
 		return -EINVAL;
 
-	fc = new_conn(sb);
+	fc = kmalloc(sizeof(*fc), GFP_KERNEL);
 	if (!fc)
 		return -ENOMEM;
 
+	err = fuse_conn_init(fc, sb);
+	if (err) {
+		kfree(fc);
+		return err;
+	}
+
+	fc->release = fuse_free_conn;
 	fc->flags = d.flags;
 	fc->user_id = d.user_id;
 	fc->group_id = d.group_id;
@@ -841,7 +852,7 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
 	sb->s_fs_info = fc;
 
 	err = -ENOMEM;
-	root = get_root_inode(sb, d.rootmode);
+	root = fuse_get_root_inode(sb, d.rootmode);
 	if (!root)
 		goto err;
 
@@ -952,7 +963,7 @@ static inline void unregister_fuseblk(void)
 
 static void fuse_inode_init_once(void *foo)
 {
-	struct inode * inode = foo;
+	struct inode *inode = foo;
 
 	inode_init_once(inode);
 }
@@ -1031,7 +1042,7 @@ static int __init fuse_init(void)
 {
 	int res;
 
-	printk("fuse init (API version %i.%i)\n",
+	printk(KERN_INFO "fuse init (API version %i.%i)\n",
 	       FUSE_KERNEL_VERSION, FUSE_KERNEL_MINOR_VERSION);
 
 	INIT_LIST_HEAD(&fuse_conn_list);
diff --git a/fs/gfs2/Kconfig b/fs/gfs2/Kconfig
index ab2f57e3fb8..e563a644981 100644
--- a/fs/gfs2/Kconfig
+++ b/fs/gfs2/Kconfig
@@ -1,6 +1,6 @@
 config GFS2_FS
 	tristate "GFS2 file system support"
-	depends on EXPERIMENTAL && (64BIT || (LSF && LBD))
+	depends on EXPERIMENTAL && (64BIT || LBD)
 	select FS_POSIX_ACL
 	select CRC32
 	help
diff --git a/fs/gfs2/Makefile b/fs/gfs2/Makefile
index ec65851ec80..c1b4ec6a965 100644
--- a/fs/gfs2/Makefile
+++ b/fs/gfs2/Makefile
@@ -1,5 +1,5 @@
 obj-$(CONFIG_GFS2_FS) += gfs2.o
-gfs2-y := acl.o bmap.o daemon.o dir.o eaops.o eattr.o glock.o \
+gfs2-y := acl.o bmap.o dir.o eaops.o eattr.o glock.o \
 	glops.o inode.o log.o lops.o locking.o main.o meta_io.o \
 	mount.o ops_address.o ops_dentry.o ops_export.o ops_file.o \
 	ops_fstype.o ops_inode.o ops_super.o quota.o \
diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c
index 3e9bd46f27e..e335dceb6a4 100644
--- a/fs/gfs2/acl.c
+++ b/fs/gfs2/acl.c
@@ -91,7 +91,7 @@ static int acl_get(struct gfs2_inode *ip, int access, struct posix_acl **acl,
 	struct gfs2_ea_location el_this;
 	int error;
 
-	if (!ip->i_di.di_eattr)
+	if (!ip->i_eattr)
 		return 0;
 
 	memset(&er, 0, sizeof(struct gfs2_ea_request));
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index bec76b1c2bb..11ffc56f1f8 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -75,9 +75,9 @@ static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh,
 		void *kaddr = kmap(page);
 
 		memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode),
-		       ip->i_di.di_size);
-		memset(kaddr + ip->i_di.di_size, 0,
-		       PAGE_CACHE_SIZE - ip->i_di.di_size);
+		       ip->i_disksize);
+		memset(kaddr + ip->i_disksize, 0,
+		       PAGE_CACHE_SIZE - ip->i_disksize);
 		kunmap(page);
 
 		SetPageUptodate(page);
@@ -132,7 +132,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page)
 	if (error)
 		goto out;
 
-	if (ip->i_di.di_size) {
+	if (ip->i_disksize) {
 		/* Get a free block, fill it with the stuffed data,
 		   and write it out to disk */
 
@@ -159,7 +159,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page)
 	di = (struct gfs2_dinode *)dibh->b_data;
 	gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));
 
-	if (ip->i_di.di_size) {
+	if (ip->i_disksize) {
 		*(__be64 *)(di + 1) = cpu_to_be64(block);
 		gfs2_add_inode_blocks(&ip->i_inode, 1);
 		di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode));
@@ -926,7 +926,7 @@ static int do_grow(struct gfs2_inode *ip, u64 size)
 		}
 	}
 
-	ip->i_di.di_size = size;
+	ip->i_disksize = size;
 	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
 	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 	gfs2_dinode_out(ip, dibh->b_data);
@@ -1033,7 +1033,7 @@ static int trunc_start(struct gfs2_inode *ip, u64 size)
 		goto out;
 
 	if (gfs2_is_stuffed(ip)) {
-		ip->i_di.di_size = size;
+		ip->i_disksize = size;
 		ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
 		gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 		gfs2_dinode_out(ip, dibh->b_data);
@@ -1045,9 +1045,9 @@ static int trunc_start(struct gfs2_inode *ip, u64 size)
 			error = gfs2_block_truncate_page(ip->i_inode.i_mapping);
 
 		if (!error) {
-			ip->i_di.di_size = size;
+			ip->i_disksize = size;
 			ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
-			ip->i_di.di_flags |= GFS2_DIF_TRUNC_IN_PROG;
+			ip->i_diskflags |= GFS2_DIF_TRUNC_IN_PROG;
 			gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 			gfs2_dinode_out(ip, dibh->b_data);
 		}
@@ -1114,13 +1114,13 @@ static int trunc_end(struct gfs2_inode *ip)
 	if (error)
 		goto out;
 
-	if (!ip->i_di.di_size) {
+	if (!ip->i_disksize) {
 		ip->i_height = 0;
 		ip->i_goal = ip->i_no_addr;
 		gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode));
 	}
 	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
-	ip->i_di.di_flags &= ~GFS2_DIF_TRUNC_IN_PROG;
+	ip->i_diskflags &= ~GFS2_DIF_TRUNC_IN_PROG;
 
 	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 	gfs2_dinode_out(ip, dibh->b_data);
@@ -1205,9 +1205,9 @@ int gfs2_truncatei(struct gfs2_inode *ip, u64 size)
 	if (gfs2_assert_warn(GFS2_SB(&ip->i_inode), S_ISREG(ip->i_inode.i_mode)))
 		return -EINVAL;
 
-	if (size > ip->i_di.di_size)
+	if (size > ip->i_disksize)
 		error = do_grow(ip, size);
-	else if (size < ip->i_di.di_size)
+	else if (size < ip->i_disksize)
 		error = do_shrink(ip, size);
 	else
 		/* update time stamps */
@@ -1219,7 +1219,7 @@ int gfs2_truncatei(struct gfs2_inode *ip, u64 size)
 int gfs2_truncatei_resume(struct gfs2_inode *ip)
 {
 	int error;
-	error = trunc_dealloc(ip, ip->i_di.di_size);
+	error = trunc_dealloc(ip, ip->i_disksize);
 	if (!error)
 		error = trunc_end(ip);
 	return error;
@@ -1231,35 +1231,6 @@ int gfs2_file_dealloc(struct gfs2_inode *ip)
 }
 
 /**
- * gfs2_write_calc_reserv - calculate number of blocks needed to write to a file
- * @ip: the file
- * @len: the number of bytes to be written to the file
- * @data_blocks: returns the number of data blocks required
- * @ind_blocks: returns the number of indirect blocks required
- *
- */
-
-void gfs2_write_calc_reserv(struct gfs2_inode *ip, unsigned int len,
-			    unsigned int *data_blocks, unsigned int *ind_blocks)
-{
-	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
-	unsigned int tmp;
-
-	if (gfs2_is_dir(ip)) {
-		*data_blocks = DIV_ROUND_UP(len, sdp->sd_jbsize) + 2;
-		*ind_blocks = 3 * (sdp->sd_max_jheight - 1);
-	} else {
-		*data_blocks = (len >> sdp->sd_sb.sb_bsize_shift) + 3;
-		*ind_blocks = 3 * (sdp->sd_max_height - 1);
-	}
-
-	for (tmp = *data_blocks; tmp > sdp->sd_diptrs;) {
-		tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs);
-		*ind_blocks += tmp;
-	}
-}
-
-/**
  * gfs2_write_alloc_required - figure out if a write will require an allocation
  * @ip: the file being written to
  * @offset: the offset to write to
@@ -1276,6 +1247,7 @@ int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset,
 	struct buffer_head bh;
 	unsigned int shift;
 	u64 lblock, lblock_stop, size;
+	u64 end_of_file;
 
 	*alloc_required = 0;
 
@@ -1291,19 +1263,12 @@ int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset,
 
 	*alloc_required = 1;
 	shift = sdp->sd_sb.sb_bsize_shift;
-	if (gfs2_is_dir(ip)) {
-		unsigned int bsize = sdp->sd_jbsize;
-		lblock = offset;
-		do_div(lblock, bsize);
-		lblock_stop = offset + len + bsize - 1;
-		do_div(lblock_stop, bsize);
-	} else {
-		u64 end_of_file = (ip->i_di.di_size + sdp->sd_sb.sb_bsize - 1) >> shift;
-		lblock = offset >> shift;
-		lblock_stop = (offset + len + sdp->sd_sb.sb_bsize - 1) >> shift;
-		if (lblock_stop > end_of_file)
-			return 0;
-	}
+	BUG_ON(gfs2_is_dir(ip));
+	end_of_file = (ip->i_disksize + sdp->sd_sb.sb_bsize - 1) >> shift;
+	lblock = offset >> shift;
+	lblock_stop = (offset + len + sdp->sd_sb.sb_bsize - 1) >> shift;
+	if (lblock_stop > end_of_file)
+		return 0;
 
 	size = (lblock_stop - lblock) << shift;
 	do {
diff --git a/fs/gfs2/bmap.h b/fs/gfs2/bmap.h
index 4e6cde2943b..c983177e05a 100644
--- a/fs/gfs2/bmap.h
+++ b/fs/gfs2/bmap.h
@@ -10,10 +10,40 @@
 #ifndef __BMAP_DOT_H__
 #define __BMAP_DOT_H__
 
+#include "inode.h"
+
 struct inode;
 struct gfs2_inode;
 struct page;
 
+
+/**
+ * gfs2_write_calc_reserv - calculate number of blocks needed to write to a file
+ * @ip: the file
+ * @len: the number of bytes to be written to the file
+ * @data_blocks: returns the number of data blocks required
+ * @ind_blocks: returns the number of indirect blocks required
+ *
+ */
+
+static inline void gfs2_write_calc_reserv(const struct gfs2_inode *ip,
+					  unsigned int len,
+					  unsigned int *data_blocks,
+					  unsigned int *ind_blocks)
+{
+	const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
+	unsigned int tmp;
+
+	BUG_ON(gfs2_is_dir(ip));
+	*data_blocks = (len >> sdp->sd_sb.sb_bsize_shift) + 3;
+	*ind_blocks = 3 * (sdp->sd_max_height - 1);
+
+	for (tmp = *data_blocks; tmp > sdp->sd_diptrs;) {
+		tmp = DIV_ROUND_UP(tmp, sdp->sd_inptrs);
+		*ind_blocks += tmp;
+	}
+}
+
 int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page);
 int gfs2_block_map(struct inode *inode, sector_t lblock, struct buffer_head *bh, int create);
 int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsigned *extlen);
@@ -21,10 +51,6 @@ int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsi
 int gfs2_truncatei(struct gfs2_inode *ip, u64 size);
 int gfs2_truncatei_resume(struct gfs2_inode *ip);
 int gfs2_file_dealloc(struct gfs2_inode *ip);
-
-void gfs2_write_calc_reserv(struct gfs2_inode *ip, unsigned int len,
-			    unsigned int *data_blocks,
-			    unsigned int *ind_blocks);
 int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset,
 			      unsigned int len, int *alloc_required);
 
diff --git a/fs/gfs2/daemon.c b/fs/gfs2/daemon.c
deleted file mode 100644
index e51991947d2..00000000000
--- a/fs/gfs2/daemon.c
+++ /dev/null
@@ -1,136 +0,0 @@
-/*
- * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
- * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License version 2.
- */
-
-#include <linux/sched.h>
-#include <linux/slab.h>
-#include <linux/spinlock.h>
-#include <linux/completion.h>
-#include <linux/buffer_head.h>
-#include <linux/kthread.h>
-#include <linux/delay.h>
-#include <linux/gfs2_ondisk.h>
-#include <linux/lm_interface.h>
-#include <linux/freezer.h>
-
-#include "gfs2.h"
-#include "incore.h"
-#include "daemon.h"
-#include "glock.h"
-#include "log.h"
-#include "quota.h"
-#include "recovery.h"
-#include "super.h"
-#include "util.h"
-
-/* This uses schedule_timeout() instead of msleep() because it's good for
-   the daemons to wake up more often than the timeout when unmounting so
-   the user's unmount doesn't sit there forever.
-
-   The kthread functions used to start these daemons block and flush signals. */
-
-/**
- * gfs2_glockd - Reclaim unused glock structures
- * @sdp: Pointer to GFS2 superblock
- *
- * One or more of these daemons run, reclaiming glocks on sd_reclaim_list.
- * Number of daemons can be set by user, with num_glockd mount option.
- */
-
-int gfs2_glockd(void *data)
-{
-	struct gfs2_sbd *sdp = data;
-
-	while (!kthread_should_stop()) {
-		while (atomic_read(&sdp->sd_reclaim_count))
-			gfs2_reclaim_glock(sdp);
-
-		wait_event_interruptible(sdp->sd_reclaim_wq,
-					 (atomic_read(&sdp->sd_reclaim_count) ||
-					 kthread_should_stop()));
-		if (freezing(current))
-			refrigerator();
-	}
-
-	return 0;
-}
-
-/**
- * gfs2_recoverd - Recover dead machine's journals
- * @sdp: Pointer to GFS2 superblock
- *
- */
-
-int gfs2_recoverd(void *data)
-{
-	struct gfs2_sbd *sdp = data;
-	unsigned long t;
-
-	while (!kthread_should_stop()) {
-		gfs2_check_journals(sdp);
-		t = gfs2_tune_get(sdp,  gt_recoverd_secs) * HZ;
-		if (freezing(current))
-			refrigerator();
-		schedule_timeout_interruptible(t);
-	}
-
-	return 0;
-}
-
-/**
- * gfs2_quotad - Write cached quota changes into the quota file
- * @sdp: Pointer to GFS2 superblock
- *
- */
-
-int gfs2_quotad(void *data)
-{
-	struct gfs2_sbd *sdp = data;
-	unsigned long t;
-	int error;
-
-	while (!kthread_should_stop()) {
-		/* Update the master statfs file */
-
-		t = sdp->sd_statfs_sync_time +
-		    gfs2_tune_get(sdp, gt_statfs_quantum) * HZ;
-
-		if (time_after_eq(jiffies, t)) {
-			error = gfs2_statfs_sync(sdp);
-			if (error &&
-			    error != -EROFS &&
-			    !test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
-				fs_err(sdp, "quotad: (1) error=%d\n", error);
-			sdp->sd_statfs_sync_time = jiffies;
-		}
-
-		/* Update quota file */
-
-		t = sdp->sd_quota_sync_time +
-		    gfs2_tune_get(sdp, gt_quota_quantum) * HZ;
-
-		if (time_after_eq(jiffies, t)) {
-			error = gfs2_quota_sync(sdp);
-			if (error &&
-			    error != -EROFS &&
-			    !test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
-				fs_err(sdp, "quotad: (2) error=%d\n", error);
-			sdp->sd_quota_sync_time = jiffies;
-		}
-
-		gfs2_quota_scan(sdp);
-
-		t = gfs2_tune_get(sdp, gt_quotad_secs) * HZ;
-		if (freezing(current))
-			refrigerator();
-		schedule_timeout_interruptible(t);
-	}
-
-	return 0;
-}
-
diff --git a/fs/gfs2/daemon.h b/fs/gfs2/daemon.h
deleted file mode 100644
index 4be084fb6a6..00000000000
--- a/fs/gfs2/daemon.h
+++ /dev/null
@@ -1,17 +0,0 @@
-/*
- * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
- * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License version 2.
- */
-
-#ifndef __DAEMON_DOT_H__
-#define __DAEMON_DOT_H__
-
-int gfs2_glockd(void *data);
-int gfs2_recoverd(void *data);
-int gfs2_quotad(void *data);
-
-#endif /* __DAEMON_DOT_H__ */
diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c
index eed040d8ba3..b7c8e5c7079 100644
--- a/fs/gfs2/dir.c
+++ b/fs/gfs2/dir.c
@@ -36,7 +36,7 @@
  * the block.  In leaves, they begin at offset sizeof(struct gfs2_leaf) from the
  * beginning of the leaf block. The dirents reside in leaves when
  *
- * dip->i_di.di_flags & GFS2_DIF_EXHASH is true
+ * dip->i_diskflags & GFS2_DIF_EXHASH is true
  *
  * Otherwise, the dirents are "linear", within a single stuffed dinode block.
  *
@@ -128,8 +128,8 @@ static int gfs2_dir_write_stuffed(struct gfs2_inode *ip, const char *buf,
 
 	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 	memcpy(dibh->b_data + offset + sizeof(struct gfs2_dinode), buf, size);
-	if (ip->i_di.di_size < offset + size)
-		ip->i_di.di_size = offset + size;
+	if (ip->i_disksize < offset + size)
+		ip->i_disksize = offset + size;
 	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
 	gfs2_dinode_out(ip, dibh->b_data);
 
@@ -226,8 +226,8 @@ out:
 	if (error)
 		return error;
 
-	if (ip->i_di.di_size < offset + copied)
-		ip->i_di.di_size = offset + copied;
+	if (ip->i_disksize < offset + copied)
+		ip->i_disksize = offset + copied;
 	ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
 
 	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
@@ -277,11 +277,11 @@ static int gfs2_dir_read_data(struct gfs2_inode *ip, char *buf, u64 offset,
 	int copied = 0;
 	int error = 0;
 
-	if (offset >= ip->i_di.di_size)
+	if (offset >= ip->i_disksize)
 		return 0;
 
-	if (offset + size > ip->i_di.di_size)
-		size = ip->i_di.di_size - offset;
+	if (offset + size > ip->i_disksize)
+		size = ip->i_disksize - offset;
 
 	if (!size)
 		return 0;
@@ -755,12 +755,12 @@ static struct gfs2_dirent *gfs2_dirent_search(struct inode *inode,
 	struct gfs2_inode *ip = GFS2_I(inode);
 	int error;
 
-	if (ip->i_di.di_flags & GFS2_DIF_EXHASH) {
+	if (ip->i_diskflags & GFS2_DIF_EXHASH) {
 		struct gfs2_leaf *leaf;
 		unsigned hsize = 1 << ip->i_depth;
 		unsigned index;
 		u64 ln;
-		if (hsize * sizeof(u64) != ip->i_di.di_size) {
+		if (hsize * sizeof(u64) != ip->i_disksize) {
 			gfs2_consist_inode(ip);
 			return ERR_PTR(-EIO);
 		}
@@ -858,8 +858,8 @@ static int dir_make_exhash(struct inode *inode)
 		return -ENOSPC;
 	bn = bh->b_blocknr;
 
-	gfs2_assert(sdp, dip->i_di.di_entries < (1 << 16));
-	leaf->lf_entries = cpu_to_be16(dip->i_di.di_entries);
+	gfs2_assert(sdp, dip->i_entries < (1 << 16));
+	leaf->lf_entries = cpu_to_be16(dip->i_entries);
 
 	/*  Copy dirents  */
 
@@ -905,9 +905,9 @@ static int dir_make_exhash(struct inode *inode)
 	for (x = sdp->sd_hash_ptrs; x--; lp++)
 		*lp = cpu_to_be64(bn);
 
-	dip->i_di.di_size = sdp->sd_sb.sb_bsize / 2;
+	dip->i_disksize = sdp->sd_sb.sb_bsize / 2;
 	gfs2_add_inode_blocks(&dip->i_inode, 1);
-	dip->i_di.di_flags |= GFS2_DIF_EXHASH;
+	dip->i_diskflags |= GFS2_DIF_EXHASH;
 
 	for (x = sdp->sd_hash_ptrs, y = -1; x; x >>= 1, y++) ;
 	dip->i_depth = y;
@@ -1082,7 +1082,7 @@ static int dir_double_exhash(struct gfs2_inode *dip)
 	int error = 0;
 
 	hsize = 1 << dip->i_depth;
-	if (hsize * sizeof(u64) != dip->i_di.di_size) {
+	if (hsize * sizeof(u64) != dip->i_disksize) {
 		gfs2_consist_inode(dip);
 		return -EIO;
 	}
@@ -1091,7 +1091,7 @@ static int dir_double_exhash(struct gfs2_inode *dip)
 
 	buf = kcalloc(3, sdp->sd_hash_bsize, GFP_NOFS | __GFP_NOFAIL);
 
-	for (block = dip->i_di.di_size >> sdp->sd_hash_bsize_shift; block--;) {
+	for (block = dip->i_disksize >> sdp->sd_hash_bsize_shift; block--;) {
 		error = gfs2_dir_read_data(dip, (char *)buf,
 					    block * sdp->sd_hash_bsize,
 					    sdp->sd_hash_bsize, 1);
@@ -1370,7 +1370,7 @@ static int dir_e_read(struct inode *inode, u64 *offset, void *opaque,
 	unsigned depth = 0;
 
 	hsize = 1 << dip->i_depth;
-	if (hsize * sizeof(u64) != dip->i_di.di_size) {
+	if (hsize * sizeof(u64) != dip->i_disksize) {
 		gfs2_consist_inode(dip);
 		return -EIO;
 	}
@@ -1426,10 +1426,10 @@ int gfs2_dir_read(struct inode *inode, u64 *offset, void *opaque,
 	int copied = 0;
 	int error;
 
-	if (!dip->i_di.di_entries)
+	if (!dip->i_entries)
 		return 0;
 
-	if (dip->i_di.di_flags & GFS2_DIF_EXHASH)
+	if (dip->i_diskflags & GFS2_DIF_EXHASH)
 		return dir_e_read(inode, offset, opaque, filldir);
 
 	if (!gfs2_is_stuffed(dip)) {
@@ -1453,17 +1453,17 @@ int gfs2_dir_read(struct inode *inode, u64 *offset, void *opaque,
 			error = PTR_ERR(dent);
 			goto out;
 		}
-		if (dip->i_di.di_entries != g.offset) {
+		if (dip->i_entries != g.offset) {
 			fs_warn(sdp, "Number of entries corrupt in dir %llu, "
-				"ip->i_di.di_entries (%u) != g.offset (%u)\n",
+				"ip->i_entries (%u) != g.offset (%u)\n",
 				(unsigned long long)dip->i_no_addr,
-				dip->i_di.di_entries,
+				dip->i_entries,
 				g.offset);
 			error = -EIO;
 			goto out;
 		}
 		error = do_filldir_main(dip, offset, opaque, filldir, darr,
-					dip->i_di.di_entries, &copied);
+					dip->i_entries, &copied);
 out:
 		kfree(darr);
 	}
@@ -1612,7 +1612,7 @@ int gfs2_dir_add(struct inode *inode, const struct qstr *name,
 			dent = gfs2_init_dirent(inode, dent, name, bh);
 			gfs2_inum_out(nip, dent);
 			dent->de_type = cpu_to_be16(type);
-			if (ip->i_di.di_flags & GFS2_DIF_EXHASH) {
+			if (ip->i_diskflags & GFS2_DIF_EXHASH) {
 				leaf = (struct gfs2_leaf *)bh->b_data;
 				be16_add_cpu(&leaf->lf_entries, 1);
 			}
@@ -1621,14 +1621,14 @@ int gfs2_dir_add(struct inode *inode, const struct qstr *name,
 			if (error)
 				break;
 			gfs2_trans_add_bh(ip->i_gl, bh, 1);
-			ip->i_di.di_entries++;
+			ip->i_entries++;
 			ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME;
 			gfs2_dinode_out(ip, bh->b_data);
 			brelse(bh);
 			error = 0;
 			break;
 		}
-		if (!(ip->i_di.di_flags & GFS2_DIF_EXHASH)) {
+		if (!(ip->i_diskflags & GFS2_DIF_EXHASH)) {
 			error = dir_make_exhash(inode);
 			if (error)
 				break;
@@ -1691,7 +1691,7 @@ int gfs2_dir_del(struct gfs2_inode *dip, const struct qstr *name)
 	}
 
 	dirent_del(dip, bh, prev, dent);
-	if (dip->i_di.di_flags & GFS2_DIF_EXHASH) {
+	if (dip->i_diskflags & GFS2_DIF_EXHASH) {
 		struct gfs2_leaf *leaf = (struct gfs2_leaf *)bh->b_data;
 		u16 entries = be16_to_cpu(leaf->lf_entries);
 		if (!entries)
@@ -1704,10 +1704,10 @@ int gfs2_dir_del(struct gfs2_inode *dip, const struct qstr *name)
 	if (error)
 		return error;
 
-	if (!dip->i_di.di_entries)
+	if (!dip->i_entries)
 		gfs2_consist_inode(dip);
 	gfs2_trans_add_bh(dip->i_gl, bh, 1);
-	dip->i_di.di_entries--;
+	dip->i_entries--;
 	dip->i_inode.i_mtime = dip->i_inode.i_ctime = CURRENT_TIME;
 	gfs2_dinode_out(dip, bh->b_data);
 	brelse(bh);
@@ -1748,7 +1748,7 @@ int gfs2_dir_mvino(struct gfs2_inode *dip, const struct qstr *filename,
 	gfs2_inum_out(nip, dent);
 	dent->de_type = cpu_to_be16(new_type);
 
-	if (dip->i_di.di_flags & GFS2_DIF_EXHASH) {
+	if (dip->i_diskflags & GFS2_DIF_EXHASH) {
 		brelse(bh);
 		error = gfs2_meta_inode_buffer(dip, &bh);
 		if (error)
@@ -1784,7 +1784,7 @@ static int foreach_leaf(struct gfs2_inode *dip, leaf_call_t lc, void *data)
 	int error = 0;
 
 	hsize = 1 << dip->i_depth;
-	if (hsize * sizeof(u64) != dip->i_di.di_size) {
+	if (hsize * sizeof(u64) != dip->i_disksize) {
 		gfs2_consist_inode(dip);
 		return -EIO;
 	}
diff --git a/fs/gfs2/dir.h b/fs/gfs2/dir.h
index 8a468cac932..4f919440c3b 100644
--- a/fs/gfs2/dir.h
+++ b/fs/gfs2/dir.h
@@ -11,6 +11,7 @@
 #define __DIR_DOT_H__
 
 #include <linux/dcache.h>
+#include <linux/crc32.h>
 
 struct inode;
 struct gfs2_inode;
diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c
index e3f76f451b0..0d1c76d906a 100644
--- a/fs/gfs2/eattr.c
+++ b/fs/gfs2/eattr.c
@@ -114,11 +114,11 @@ static int ea_foreach(struct gfs2_inode *ip, ea_call_t ea_call, void *data)
 	__be64 *eablk, *end;
 	int error;
 
-	error = gfs2_meta_read(ip->i_gl, ip->i_di.di_eattr, DIO_WAIT, &bh);
+	error = gfs2_meta_read(ip->i_gl, ip->i_eattr, DIO_WAIT, &bh);
 	if (error)
 		return error;
 
-	if (!(ip->i_di.di_flags & GFS2_DIF_EA_INDIRECT)) {
+	if (!(ip->i_diskflags & GFS2_DIF_EA_INDIRECT)) {
 		error = ea_foreach_i(ip, bh, ea_call, data);
 		goto out;
 	}
@@ -414,7 +414,7 @@ int gfs2_ea_list(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 	if (error)
 		return error;
 
-	if (ip->i_di.di_eattr) {
+	if (ip->i_eattr) {
 		struct ea_list ei = { .ei_er = er, .ei_size = 0 };
 
 		error = ea_foreach(ip, ea_list_i, &ei);
@@ -514,7 +514,7 @@ int gfs2_ea_get_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 	struct gfs2_ea_location el;
 	int error;
 
-	if (!ip->i_di.di_eattr)
+	if (!ip->i_eattr)
 		return -ENODATA;
 
 	error = gfs2_ea_find(ip, er, &el);
@@ -741,7 +741,7 @@ static int ea_init_i(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 	if (error)
 		return error;
 
-	ip->i_di.di_eattr = bh->b_blocknr;
+	ip->i_eattr = bh->b_blocknr;
 	error = ea_write(ip, GFS2_EA_BH2FIRST(bh), er);
 
 	brelse(bh);
@@ -935,10 +935,10 @@ static int ea_set_block(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 	int error;
 	int mh_size = sizeof(struct gfs2_meta_header);
 
-	if (ip->i_di.di_flags & GFS2_DIF_EA_INDIRECT) {
+	if (ip->i_diskflags & GFS2_DIF_EA_INDIRECT) {
 		__be64 *end;
 
-		error = gfs2_meta_read(ip->i_gl, ip->i_di.di_eattr, DIO_WAIT,
+		error = gfs2_meta_read(ip->i_gl, ip->i_eattr, DIO_WAIT,
 				       &indbh);
 		if (error)
 			return error;
@@ -972,9 +972,9 @@ static int ea_set_block(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 		gfs2_buffer_clear_tail(indbh, mh_size);
 
 		eablk = (__be64 *)(indbh->b_data + mh_size);
-		*eablk = cpu_to_be64(ip->i_di.di_eattr);
-		ip->i_di.di_eattr = blk;
-		ip->i_di.di_flags |= GFS2_DIF_EA_INDIRECT;
+		*eablk = cpu_to_be64(ip->i_eattr);
+		ip->i_eattr = blk;
+		ip->i_diskflags |= GFS2_DIF_EA_INDIRECT;
 		gfs2_add_inode_blocks(&ip->i_inode, 1);
 
 		eablk++;
@@ -1015,7 +1015,7 @@ static int ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 	if (error)
 		return error;
 
-	if (!(ip->i_di.di_flags & GFS2_DIF_EA_INDIRECT))
+	if (!(ip->i_diskflags & GFS2_DIF_EA_INDIRECT))
 		blks++;
 	if (GFS2_EAREQ_SIZE_STUFFED(er) > GFS2_SB(&ip->i_inode)->sd_jbsize)
 		blks += DIV_ROUND_UP(er->er_data_len, GFS2_SB(&ip->i_inode)->sd_jbsize);
@@ -1040,7 +1040,7 @@ int gfs2_ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 	struct gfs2_ea_location el;
 	int error;
 
-	if (!ip->i_di.di_eattr) {
+	if (!ip->i_eattr) {
 		if (er->er_flags & XATTR_REPLACE)
 			return -ENODATA;
 		return ea_init(ip, er);
@@ -1051,7 +1051,7 @@ int gfs2_ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 		return error;
 
 	if (el.el_ea) {
-		if (ip->i_di.di_flags & GFS2_DIF_APPENDONLY) {
+		if (ip->i_diskflags & GFS2_DIF_APPENDONLY) {
 			brelse(el.el_bh);
 			return -EPERM;
 		}
@@ -1145,7 +1145,7 @@ int gfs2_ea_remove_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 	struct gfs2_ea_location el;
 	int error;
 
-	if (!ip->i_di.di_eattr)
+	if (!ip->i_eattr)
 		return -ENODATA;
 
 	error = gfs2_ea_find(ip, er, &el);
@@ -1309,7 +1309,7 @@ static int ea_dealloc_indirect(struct gfs2_inode *ip)
 
 	memset(&rlist, 0, sizeof(struct gfs2_rgrp_list));
 
-	error = gfs2_meta_read(ip->i_gl, ip->i_di.di_eattr, DIO_WAIT, &indbh);
+	error = gfs2_meta_read(ip->i_gl, ip->i_eattr, DIO_WAIT, &indbh);
 	if (error)
 		return error;
 
@@ -1388,7 +1388,7 @@ static int ea_dealloc_indirect(struct gfs2_inode *ip)
 	if (bstart)
 		gfs2_free_meta(ip, bstart, blen);
 
-	ip->i_di.di_flags &= ~GFS2_DIF_EA_INDIRECT;
+	ip->i_diskflags &= ~GFS2_DIF_EA_INDIRECT;
 
 	error = gfs2_meta_inode_buffer(ip, &dibh);
 	if (!error) {
@@ -1416,7 +1416,7 @@ static int ea_dealloc_block(struct gfs2_inode *ip)
 	struct buffer_head *dibh;
 	int error;
 
-	rgd = gfs2_blk2rgrpd(sdp, ip->i_di.di_eattr);
+	rgd = gfs2_blk2rgrpd(sdp, ip->i_eattr);
 	if (!rgd) {
 		gfs2_consist_inode(ip);
 		return -EIO;
@@ -1432,9 +1432,9 @@ static int ea_dealloc_block(struct gfs2_inode *ip)
 	if (error)
 		goto out_gunlock;
 
-	gfs2_free_meta(ip, ip->i_di.di_eattr, 1);
+	gfs2_free_meta(ip, ip->i_eattr, 1);
 
-	ip->i_di.di_eattr = 0;
+	ip->i_eattr = 0;
 	gfs2_add_inode_blocks(&ip->i_inode, -1);
 
 	error = gfs2_meta_inode_buffer(ip, &dibh);
@@ -1479,7 +1479,7 @@ int gfs2_ea_dealloc(struct gfs2_inode *ip)
 	if (error)
 		goto out_rindex;
 
-	if (ip->i_di.di_flags & GFS2_DIF_EA_INDIRECT) {
+	if (ip->i_diskflags & GFS2_DIF_EA_INDIRECT) {
 		error = ea_dealloc_indirect(ip);
 		if (error)
 			goto out_rindex;
diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c
index c962283d4e7..6b983aef785 100644
--- a/fs/gfs2/glock.c
+++ b/fs/gfs2/glock.c
@@ -40,6 +40,7 @@
 #include "quota.h"
 #include "super.h"
 #include "util.h"
+#include "bmap.h"
 
 struct gfs2_gl_hash_bucket {
         struct hlist_head hb_list;
@@ -61,9 +62,10 @@ static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int
 
 static DECLARE_RWSEM(gfs2_umount_flush_sem);
 static struct dentry *gfs2_root;
-static struct task_struct *scand_process;
-static unsigned int scand_secs = 5;
 static struct workqueue_struct *glock_workqueue;
+static LIST_HEAD(lru_list);
+static atomic_t lru_count = ATOMIC_INIT(0);
+static DEFINE_SPINLOCK(lru_lock);
 
 #define GFS2_GL_HASH_SHIFT      15
 #define GFS2_GL_HASH_SIZE       (1 << GFS2_GL_HASH_SHIFT)
@@ -174,6 +176,22 @@ static void gfs2_glock_hold(struct gfs2_glock *gl)
 }
 
 /**
+ * gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
+ * @gl: the glock
+ *
+ */
+
+static void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
+{
+	spin_lock(&lru_lock);
+	if (list_empty(&gl->gl_lru) && gl->gl_state != LM_ST_UNLOCKED) {
+		list_add_tail(&gl->gl_lru, &lru_list);
+		atomic_inc(&lru_count);
+	}
+	spin_unlock(&lru_lock);
+}
+
+/**
  * gfs2_glock_put() - Decrement reference count on glock
  * @gl: The glock to put
  *
@@ -187,14 +205,23 @@ int gfs2_glock_put(struct gfs2_glock *gl)
 	if (atomic_dec_and_test(&gl->gl_ref)) {
 		hlist_del(&gl->gl_list);
 		write_unlock(gl_lock_addr(gl->gl_hash));
+		spin_lock(&lru_lock);
+		if (!list_empty(&gl->gl_lru)) {
+			list_del_init(&gl->gl_lru);
+			atomic_dec(&lru_count);
+		}
+		spin_unlock(&lru_lock);
 		GLOCK_BUG_ON(gl, gl->gl_state != LM_ST_UNLOCKED);
-		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_reclaim));
+		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_lru));
 		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_holders));
 		glock_free(gl);
 		rv = 1;
 		goto out;
 	}
 	write_unlock(gl_lock_addr(gl->gl_hash));
+	/* 1 for being hashed, 1 for having state != LM_ST_UNLOCKED */
+	if (atomic_read(&gl->gl_ref) == 2)
+		gfs2_glock_schedule_for_reclaim(gl);
 out:
 	return rv;
 }
@@ -289,10 +316,13 @@ static void gfs2_holder_wake(struct gfs2_holder *gh)
  * do_promote - promote as many requests as possible on the current queue
  * @gl: The glock
  * 
- * Returns: true if there is a blocked holder at the head of the list
+ * Returns: 1 if there is a blocked holder at the head of the list, or 2
+ *          if a type specific operation is underway.
  */
 
 static int do_promote(struct gfs2_glock *gl)
+__releases(&gl->gl_spin)
+__acquires(&gl->gl_spin)
 {
 	const struct gfs2_glock_operations *glops = gl->gl_ops;
 	struct gfs2_holder *gh, *tmp;
@@ -310,6 +340,8 @@ restart:
 				ret = glops->go_lock(gh);
 				spin_lock(&gl->gl_spin);
 				if (ret) {
+					if (ret == 1)
+						return 2;
 					gh->gh_error = ret;
 					list_del_init(&gh->gh_list);
 					gfs2_holder_wake(gh);
@@ -414,6 +446,7 @@ static void finish_xmote(struct gfs2_glock *gl, unsigned int ret)
 	const struct gfs2_glock_operations *glops = gl->gl_ops;
 	struct gfs2_holder *gh;
 	unsigned state = ret & LM_OUT_ST_MASK;
+	int rv;
 
 	spin_lock(&gl->gl_spin);
 	state_change(gl, state);
@@ -468,7 +501,6 @@ retry:
 		gfs2_demote_wake(gl);
 	if (state != LM_ST_UNLOCKED) {
 		if (glops->go_xmote_bh) {
-			int rv;
 			spin_unlock(&gl->gl_spin);
 			rv = glops->go_xmote_bh(gl, gh);
 			if (rv == -EAGAIN)
@@ -479,10 +511,13 @@ retry:
 				goto out;
 			}
 		}
-		do_promote(gl);
+		rv = do_promote(gl);
+		if (rv == 2)
+			goto out_locked;
 	}
 out:
 	clear_bit(GLF_LOCK, &gl->gl_flags);
+out_locked:
 	spin_unlock(&gl->gl_spin);
 	gfs2_glock_put(gl);
 }
@@ -511,6 +546,8 @@ static unsigned int gfs2_lm_lock(struct gfs2_sbd *sdp, void *lock,
  */
 
 static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int target)
+__releases(&gl->gl_spin)
+__acquires(&gl->gl_spin)
 {
 	const struct gfs2_glock_operations *glops = gl->gl_ops;
 	struct gfs2_sbd *sdp = gl->gl_sbd;
@@ -576,8 +613,11 @@ static inline struct gfs2_holder *find_first_holder(const struct gfs2_glock *gl)
  */
 
 static void run_queue(struct gfs2_glock *gl, const int nonblock)
+__releases(&gl->gl_spin)
+__acquires(&gl->gl_spin)
 {
 	struct gfs2_holder *gh = NULL;
+	int ret;
 
 	if (test_and_set_bit(GLF_LOCK, &gl->gl_flags))
 		return;
@@ -596,8 +636,11 @@ static void run_queue(struct gfs2_glock *gl, const int nonblock)
 	} else {
 		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
 			gfs2_demote_wake(gl);
-		if (do_promote(gl) == 0)
+		ret = do_promote(gl);
+		if (ret == 0)
 			goto out;
+		if (ret == 2)
+			return;
 		gh = find_first_waiter(gl);
 		gl->gl_target = gh->gh_state;
 		if (!(gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)))
@@ -820,7 +863,7 @@ static void wait_on_demote(struct gfs2_glock *gl)
  */
 
 static void handle_callback(struct gfs2_glock *gl, unsigned int state,
-			    int remote, unsigned long delay)
+			    unsigned long delay)
 {
 	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
 
@@ -828,9 +871,6 @@ static void handle_callback(struct gfs2_glock *gl, unsigned int state,
 	if (gl->gl_demote_state == LM_ST_EXCLUSIVE) {
 		gl->gl_demote_state = state;
 		gl->gl_demote_time = jiffies;
-		if (remote && gl->gl_ops->go_type == LM_TYPE_IOPEN &&
-		    gl->gl_object)
-			gfs2_glock_schedule_for_reclaim(gl);
 	} else if (gl->gl_demote_state != LM_ST_UNLOCKED &&
 			gl->gl_demote_state != state) {
 		gl->gl_demote_state = LM_ST_UNLOCKED;
@@ -877,6 +917,8 @@ void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
  */
 
 static inline void add_to_queue(struct gfs2_holder *gh)
+__releases(&gl->gl_spin)
+__acquires(&gl->gl_spin)
 {
 	struct gfs2_glock *gl = gh->gh_gl;
 	struct gfs2_sbd *sdp = gl->gl_sbd;
@@ -998,7 +1040,7 @@ void gfs2_glock_dq(struct gfs2_holder *gh)
 
 	spin_lock(&gl->gl_spin);
 	if (gh->gh_flags & GL_NOCACHE)
-		handle_callback(gl, LM_ST_UNLOCKED, 0, 0);
+		handle_callback(gl, LM_ST_UNLOCKED, 0);
 
 	list_del_init(&gh->gh_list);
 	if (find_first_holder(gl) == NULL) {
@@ -1269,12 +1311,26 @@ static void blocking_cb(struct gfs2_sbd *sdp, struct lm_lockname *name,
 		delay = gl->gl_ops->go_min_hold_time;
 
 	spin_lock(&gl->gl_spin);
-	handle_callback(gl, state, 1, delay);
+	handle_callback(gl, state, delay);
 	spin_unlock(&gl->gl_spin);
 	if (queue_delayed_work(glock_workqueue, &gl->gl_work, delay) == 0)
 		gfs2_glock_put(gl);
 }
 
+static void gfs2_jdesc_make_dirty(struct gfs2_sbd *sdp, unsigned int jid)
+{
+	struct gfs2_jdesc *jd;
+
+	spin_lock(&sdp->sd_jindex_spin);
+	list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
+		if (jd->jd_jid != jid)
+			continue;
+		jd->jd_dirty = 1;
+		break;
+	}
+	spin_unlock(&sdp->sd_jindex_spin);
+}
+
 /**
  * gfs2_glock_cb - Callback used by locking module
  * @sdp: Pointer to the superblock
@@ -1338,80 +1394,83 @@ void gfs2_glock_cb(void *cb_data, unsigned int type, void *data)
  * Returns: 1 if it's ok
  */
 
-static int demote_ok(struct gfs2_glock *gl)
+static int demote_ok(const struct gfs2_glock *gl)
 {
 	const struct gfs2_glock_operations *glops = gl->gl_ops;
-	int demote = 1;
-
-	if (test_bit(GLF_STICKY, &gl->gl_flags))
-		demote = 0;
-	else if (glops->go_demote_ok)
-		demote = glops->go_demote_ok(gl);
-
-	return demote;
-}
-
-/**
- * gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
- * @gl: the glock
- *
- */
-
-void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
-{
-	struct gfs2_sbd *sdp = gl->gl_sbd;
 
-	spin_lock(&sdp->sd_reclaim_lock);
-	if (list_empty(&gl->gl_reclaim)) {
-		gfs2_glock_hold(gl);
-		list_add(&gl->gl_reclaim, &sdp->sd_reclaim_list);
-		atomic_inc(&sdp->sd_reclaim_count);
-		spin_unlock(&sdp->sd_reclaim_lock);
-		wake_up(&sdp->sd_reclaim_wq);
-	} else
-		spin_unlock(&sdp->sd_reclaim_lock);
+	if (gl->gl_state == LM_ST_UNLOCKED)
+		return 0;
+	if (!list_empty(&gl->gl_holders))
+		return 0;
+	if (glops->go_demote_ok)
+		return glops->go_demote_ok(gl);
+	return 1;
 }
 
-/**
- * gfs2_reclaim_glock - process the next glock on the filesystem's reclaim list
- * @sdp: the filesystem
- *
- * Called from gfs2_glockd() glock reclaim daemon, or when promoting a
- * different glock and we notice that there are a lot of glocks in the
- * reclaim list.
- *
- */
 
-void gfs2_reclaim_glock(struct gfs2_sbd *sdp)
+static int gfs2_shrink_glock_memory(int nr, gfp_t gfp_mask)
 {
 	struct gfs2_glock *gl;
-	int done_callback = 0;
+	int may_demote;
+	int nr_skipped = 0;
+	int got_ref = 0;
+	LIST_HEAD(skipped);
 
-	spin_lock(&sdp->sd_reclaim_lock);
-	if (list_empty(&sdp->sd_reclaim_list)) {
-		spin_unlock(&sdp->sd_reclaim_lock);
-		return;
-	}
-	gl = list_entry(sdp->sd_reclaim_list.next,
-			struct gfs2_glock, gl_reclaim);
-	list_del_init(&gl->gl_reclaim);
-	spin_unlock(&sdp->sd_reclaim_lock);
+	if (nr == 0)
+		goto out;
 
-	atomic_dec(&sdp->sd_reclaim_count);
-	atomic_inc(&sdp->sd_reclaimed);
+	if (!(gfp_mask & __GFP_FS))
+		return -1;
 
-	spin_lock(&gl->gl_spin);
-	if (find_first_holder(gl) == NULL &&
-	    gl->gl_state != LM_ST_UNLOCKED && demote_ok(gl)) {
-		handle_callback(gl, LM_ST_UNLOCKED, 0, 0);
-		done_callback = 1;
+	spin_lock(&lru_lock);
+	while(nr && !list_empty(&lru_list)) {
+		gl = list_entry(lru_list.next, struct gfs2_glock, gl_lru);
+		list_del_init(&gl->gl_lru);
+		atomic_dec(&lru_count);
+
+		/* Test for being demotable */
+		if (!test_and_set_bit(GLF_LOCK, &gl->gl_flags)) {
+			gfs2_glock_hold(gl);
+			got_ref = 1;
+			spin_unlock(&lru_lock);
+			spin_lock(&gl->gl_spin);
+			may_demote = demote_ok(gl);
+			spin_unlock(&gl->gl_spin);
+			clear_bit(GLF_LOCK, &gl->gl_flags);
+			if (may_demote) {
+				handle_callback(gl, LM_ST_UNLOCKED, 0);
+				nr--;
+				if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
+					gfs2_glock_put(gl);
+			}
+			spin_lock(&lru_lock);
+			if (may_demote)
+				continue;
+		}
+		if (list_empty(&gl->gl_lru) &&
+		    (atomic_read(&gl->gl_ref) <= (2 + got_ref))) {
+			nr_skipped++;
+			list_add(&gl->gl_lru, &skipped);
+		}
+		if (got_ref) {
+			spin_unlock(&lru_lock);
+			gfs2_glock_put(gl);
+			spin_lock(&lru_lock);
+			got_ref = 0;
+		}
 	}
-	spin_unlock(&gl->gl_spin);
-	if (!done_callback ||
-	    queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
-		gfs2_glock_put(gl);
+	list_splice(&skipped, &lru_list);
+	atomic_add(nr_skipped, &lru_count);
+	spin_unlock(&lru_lock);
+out:
+	return (atomic_read(&lru_count) / 100) * sysctl_vfs_cache_pressure;
 }
 
+static struct shrinker glock_shrinker = {
+	.shrink = gfs2_shrink_glock_memory,
+	.seeks = DEFAULT_SEEKS,
+};
+
 /**
  * examine_bucket - Call a function for glock in a hash bucket
  * @examiner: the function
@@ -1457,26 +1516,6 @@ out:
 }
 
 /**
- * scan_glock - look at a glock and see if we can reclaim it
- * @gl: the glock to look at
- *
- */
-
-static void scan_glock(struct gfs2_glock *gl)
-{
-	if (gl->gl_ops == &gfs2_inode_glops && gl->gl_object)
-		return;
-	if (test_bit(GLF_LOCK, &gl->gl_flags))
-		return;
-
-	spin_lock(&gl->gl_spin);
-	if (find_first_holder(gl) == NULL &&
-	    gl->gl_state != LM_ST_UNLOCKED && demote_ok(gl))
-		gfs2_glock_schedule_for_reclaim(gl);
-	spin_unlock(&gl->gl_spin);
-}
-
-/**
  * clear_glock - look at a glock and see if we can free it from glock cache
  * @gl: the glock to look at
  *
@@ -1484,23 +1523,16 @@ static void scan_glock(struct gfs2_glock *gl)
 
 static void clear_glock(struct gfs2_glock *gl)
 {
-	struct gfs2_sbd *sdp = gl->gl_sbd;
-	int released;
-
-	spin_lock(&sdp->sd_reclaim_lock);
-	if (!list_empty(&gl->gl_reclaim)) {
-		list_del_init(&gl->gl_reclaim);
-		atomic_dec(&sdp->sd_reclaim_count);
-		spin_unlock(&sdp->sd_reclaim_lock);
-		released = gfs2_glock_put(gl);
-		gfs2_assert(sdp, !released);
-	} else {
-		spin_unlock(&sdp->sd_reclaim_lock);
+	spin_lock(&lru_lock);
+	if (!list_empty(&gl->gl_lru)) {
+		list_del_init(&gl->gl_lru);
+		atomic_dec(&lru_count);
 	}
+	spin_unlock(&lru_lock);
 
 	spin_lock(&gl->gl_spin);
 	if (find_first_holder(gl) == NULL && gl->gl_state != LM_ST_UNLOCKED)
-		handle_callback(gl, LM_ST_UNLOCKED, 0, 0);
+		handle_callback(gl, LM_ST_UNLOCKED, 0);
 	spin_unlock(&gl->gl_spin);
 	gfs2_glock_hold(gl);
 	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
@@ -1548,6 +1580,20 @@ void gfs2_gl_hash_clear(struct gfs2_sbd *sdp)
 	}
 }
 
+void gfs2_glock_finish_truncate(struct gfs2_inode *ip)
+{
+	struct gfs2_glock *gl = ip->i_gl;
+	int ret;
+
+	ret = gfs2_truncatei_resume(ip);
+	gfs2_assert_withdraw(gl->gl_sbd, ret == 0);
+
+	spin_lock(&gl->gl_spin);
+	clear_bit(GLF_LOCK, &gl->gl_flags);
+	run_queue(gl, 1);
+	spin_unlock(&gl->gl_spin);
+}
+
 static const char *state2str(unsigned state)
 {
 	switch(state) {
@@ -1623,8 +1669,6 @@ static const char *gflags2str(char *buf, const unsigned long *gflags)
 	char *p = buf;
 	if (test_bit(GLF_LOCK, gflags))
 		*p++ = 'l';
-	if (test_bit(GLF_STICKY, gflags))
-		*p++ = 's';
 	if (test_bit(GLF_DEMOTE, gflags))
 		*p++ = 'D';
 	if (test_bit(GLF_PENDING_DEMOTE, gflags))
@@ -1743,34 +1787,6 @@ static int gfs2_dump_lockstate(struct gfs2_sbd *sdp)
 	return error;
 }
 
-/**
- * gfs2_scand - Look for cached glocks and inodes to toss from memory
- * @sdp: Pointer to GFS2 superblock
- *
- * One of these daemons runs, finding candidates to add to sd_reclaim_list.
- * See gfs2_glockd()
- */
-
-static int gfs2_scand(void *data)
-{
-	unsigned x;
-	unsigned delay;
-
-	while (!kthread_should_stop()) {
-		for (x = 0; x < GFS2_GL_HASH_SIZE; x++)
-			examine_bucket(scan_glock, NULL, x);
-		if (freezing(current))
-			refrigerator();
-		delay = scand_secs;
-		if (delay < 1)
-			delay = 1;
-		schedule_timeout_interruptible(delay * HZ);
-	}
-
-	return 0;
-}
-
-
 
 int __init gfs2_glock_init(void)
 {
@@ -1784,28 +1800,21 @@ int __init gfs2_glock_init(void)
 	}
 #endif
 
-	scand_process = kthread_run(gfs2_scand, NULL, "gfs2_scand");
-	if (IS_ERR(scand_process))
-		return PTR_ERR(scand_process);
-
 	glock_workqueue = create_workqueue("glock_workqueue");
-	if (IS_ERR(glock_workqueue)) {
-		kthread_stop(scand_process);
+	if (IS_ERR(glock_workqueue))
 		return PTR_ERR(glock_workqueue);
-	}
+
+	register_shrinker(&glock_shrinker);
 
 	return 0;
 }
 
 void gfs2_glock_exit(void)
 {
+	unregister_shrinker(&glock_shrinker);
 	destroy_workqueue(glock_workqueue);
-	kthread_stop(scand_process);
 }
 
-module_param(scand_secs, uint, S_IRUGO|S_IWUSR);
-MODULE_PARM_DESC(scand_secs, "The number of seconds between scand runs");
-
 static int gfs2_glock_iter_next(struct gfs2_glock_iter *gi)
 {
 	struct gfs2_glock *gl;
diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h
index 695c6b19361..543ec7ecfbd 100644
--- a/fs/gfs2/glock.h
+++ b/fs/gfs2/glock.h
@@ -129,9 +129,9 @@ int gfs2_lvb_hold(struct gfs2_glock *gl);
 void gfs2_lvb_unhold(struct gfs2_glock *gl);
 
 void gfs2_glock_cb(void *cb_data, unsigned int type, void *data);
-void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl);
 void gfs2_reclaim_glock(struct gfs2_sbd *sdp);
 void gfs2_gl_hash_clear(struct gfs2_sbd *sdp);
+void gfs2_glock_finish_truncate(struct gfs2_inode *ip);
 
 int __init gfs2_glock_init(void);
 void gfs2_glock_exit(void);
diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c
index c6c318c2a0f..8522d3aa64f 100644
--- a/fs/gfs2/glops.c
+++ b/fs/gfs2/glops.c
@@ -201,19 +201,12 @@ static void inode_go_inval(struct gfs2_glock *gl, int flags)
  * Returns: 1 if it's ok
  */
 
-static int inode_go_demote_ok(struct gfs2_glock *gl)
+static int inode_go_demote_ok(const struct gfs2_glock *gl)
 {
 	struct gfs2_sbd *sdp = gl->gl_sbd;
-	int demote = 0;
-
-	if (!gl->gl_object && !gl->gl_aspace->i_mapping->nrpages)
-		demote = 1;
-	else if (!sdp->sd_args.ar_localcaching &&
-		 time_after_eq(jiffies, gl->gl_stamp +
-			       gfs2_tune_get(sdp, gt_demote_secs) * HZ))
-		demote = 1;
-
-	return demote;
+	if (sdp->sd_jindex == gl->gl_object || sdp->sd_rindex == gl->gl_object)
+		return 0;
+	return 1;
 }
 
 /**
@@ -227,6 +220,7 @@ static int inode_go_demote_ok(struct gfs2_glock *gl)
 static int inode_go_lock(struct gfs2_holder *gh)
 {
 	struct gfs2_glock *gl = gh->gh_gl;
+	struct gfs2_sbd *sdp = gl->gl_sbd;
 	struct gfs2_inode *ip = gl->gl_object;
 	int error = 0;
 
@@ -239,10 +233,16 @@ static int inode_go_lock(struct gfs2_holder *gh)
 			return error;
 	}
 
-	if ((ip->i_di.di_flags & GFS2_DIF_TRUNC_IN_PROG) &&
+	if ((ip->i_diskflags & GFS2_DIF_TRUNC_IN_PROG) &&
 	    (gl->gl_state == LM_ST_EXCLUSIVE) &&
-	    (gh->gh_state == LM_ST_EXCLUSIVE))
-		error = gfs2_truncatei_resume(ip);
+	    (gh->gh_state == LM_ST_EXCLUSIVE)) {
+		spin_lock(&sdp->sd_trunc_lock);
+		if (list_empty(&ip->i_trunc_list))
+			list_add(&sdp->sd_trunc_list, &ip->i_trunc_list);
+		spin_unlock(&sdp->sd_trunc_lock);
+		wake_up(&sdp->sd_quota_wait);
+		return 1;
+	}
 
 	return error;
 }
@@ -260,10 +260,13 @@ static int inode_go_dump(struct seq_file *seq, const struct gfs2_glock *gl)
 	const struct gfs2_inode *ip = gl->gl_object;
 	if (ip == NULL)
 		return 0;
-	gfs2_print_dbg(seq, " I: n:%llu/%llu t:%u f:0x%08lx\n",
+	gfs2_print_dbg(seq, " I: n:%llu/%llu t:%u f:0x%02lx d:0x%08x s:%llu/%llu\n",
 		  (unsigned long long)ip->i_no_formal_ino,
 		  (unsigned long long)ip->i_no_addr,
-		  IF2DT(ip->i_inode.i_mode), ip->i_flags);
+		  IF2DT(ip->i_inode.i_mode), ip->i_flags,
+		  (unsigned int)ip->i_diskflags,
+		  (unsigned long long)ip->i_inode.i_size,
+		  (unsigned long long)ip->i_disksize);
 	return 0;
 }
 
@@ -274,7 +277,7 @@ static int inode_go_dump(struct seq_file *seq, const struct gfs2_glock *gl)
  * Returns: 1 if it's ok
  */
 
-static int rgrp_go_demote_ok(struct gfs2_glock *gl)
+static int rgrp_go_demote_ok(const struct gfs2_glock *gl)
 {
 	return !gl->gl_aspace->i_mapping->nrpages;
 }
@@ -318,7 +321,9 @@ static int rgrp_go_dump(struct seq_file *seq, const struct gfs2_glock *gl)
 	const struct gfs2_rgrpd *rgd = gl->gl_object;
 	if (rgd == NULL)
 		return 0;
-	gfs2_print_dbg(seq, " R: n:%llu\n", (unsigned long long)rgd->rd_addr);
+	gfs2_print_dbg(seq, " R: n:%llu f:%02x b:%u/%u i:%u\n",
+		       (unsigned long long)rgd->rd_addr, rgd->rd_flags,
+		       rgd->rd_free, rgd->rd_free_clone, rgd->rd_dinodes);
 	return 0;
 }
 
@@ -374,13 +379,25 @@ static int trans_go_xmote_bh(struct gfs2_glock *gl, struct gfs2_holder *gh)
 }
 
 /**
+ * trans_go_demote_ok
+ * @gl: the glock
+ *
+ * Always returns 0
+ */
+
+static int trans_go_demote_ok(const struct gfs2_glock *gl)
+{
+	return 0;
+}
+
+/**
  * quota_go_demote_ok - Check to see if it's ok to unlock a quota glock
  * @gl: the glock
  *
  * Returns: 1 if it's ok
  */
 
-static int quota_go_demote_ok(struct gfs2_glock *gl)
+static int quota_go_demote_ok(const struct gfs2_glock *gl)
 {
 	return !atomic_read(&gl->gl_lvb_count);
 }
@@ -414,6 +431,7 @@ const struct gfs2_glock_operations gfs2_rgrp_glops = {
 const struct gfs2_glock_operations gfs2_trans_glops = {
 	.go_xmote_th = trans_go_sync,
 	.go_xmote_bh = trans_go_xmote_bh,
+	.go_demote_ok = trans_go_demote_ok,
 	.go_type = LM_TYPE_NONDISK,
 };
 
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index f566ec1b4e8..608849d0002 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -68,12 +68,6 @@ struct gfs2_bitmap {
 	u32 bi_len;
 };
 
-struct gfs2_rgrp_host {
-	u32 rg_free;
-	u32 rg_dinodes;
-	u64 rg_igeneration;
-};
-
 struct gfs2_rgrpd {
 	struct list_head rd_list;	/* Link with superblock */
 	struct list_head rd_list_mru;
@@ -83,14 +77,16 @@ struct gfs2_rgrpd {
 	u32 rd_length;			/* length of rgrp header in fs blocks */
 	u32 rd_data;			/* num of data blocks in rgrp */
 	u32 rd_bitbytes;		/* number of bytes in data bitmaps */
-	struct gfs2_rgrp_host rd_rg;
+	u32 rd_free;
+	u32 rd_free_clone;
+	u32 rd_dinodes;
+	u64 rd_igeneration;
 	struct gfs2_bitmap *rd_bits;
-	unsigned int rd_bh_count;
 	struct mutex rd_mutex;
-	u32 rd_free_clone;
 	struct gfs2_log_element rd_le;
-	u32 rd_last_alloc;
 	struct gfs2_sbd *rd_sbd;
+	unsigned int rd_bh_count;
+	u32 rd_last_alloc;
 	unsigned char rd_flags;
 #define GFS2_RDF_CHECK        0x01      /* Need to check for unlinked inodes */
 #define GFS2_RDF_NOALLOC      0x02      /* rg prohibits allocation */
@@ -129,7 +125,7 @@ struct gfs2_glock_operations {
 	void (*go_xmote_th) (struct gfs2_glock *gl);
 	int (*go_xmote_bh) (struct gfs2_glock *gl, struct gfs2_holder *gh);
 	void (*go_inval) (struct gfs2_glock *gl, int flags);
-	int (*go_demote_ok) (struct gfs2_glock *gl);
+	int (*go_demote_ok) (const struct gfs2_glock *gl);
 	int (*go_lock) (struct gfs2_holder *gh);
 	void (*go_unlock) (struct gfs2_holder *gh);
 	int (*go_dump)(struct seq_file *seq, const struct gfs2_glock *gl);
@@ -159,7 +155,6 @@ struct gfs2_holder {
 
 enum {
 	GLF_LOCK			= 1,
-	GLF_STICKY			= 2,
 	GLF_DEMOTE			= 3,
 	GLF_PENDING_DEMOTE		= 4,
 	GLF_DEMOTE_IN_PROGRESS		= 5,
@@ -194,7 +189,7 @@ struct gfs2_glock {
 	unsigned long gl_tchange;
 	void *gl_object;
 
-	struct list_head gl_reclaim;
+	struct list_head gl_lru;
 
 	struct gfs2_sbd *gl_sbd;
 
@@ -233,29 +228,24 @@ enum {
 	GIF_USER                = 4, /* user inode, not metadata addr space */
 };
 
-struct gfs2_dinode_host {
-	u64 di_size;		/* number of bytes in file */
-	u64 di_generation;	/* generation number for NFS */
-	u32 di_flags;		/* GFS2_DIF_... */
-	/* These only apply to directories  */
-	u32 di_entries;		/* The number of entries in the directory */
-	u64 di_eattr;		/* extended attribute block number */
-};
 
 struct gfs2_inode {
 	struct inode i_inode;
 	u64 i_no_addr;
 	u64 i_no_formal_ino;
+	u64 i_generation;
+	u64 i_eattr;
+	loff_t i_disksize;
 	unsigned long i_flags;		/* GIF_... */
-
-	struct gfs2_dinode_host i_di; /* To be replaced by ref to block */
-
 	struct gfs2_glock *i_gl; /* Move into i_gh? */
 	struct gfs2_holder i_iopen_gh;
 	struct gfs2_holder i_gh; /* for prepare/commit_write only */
 	struct gfs2_alloc *i_alloc;
 	u64 i_goal;	/* goal block for allocations */
 	struct rw_semaphore i_rw_mutex;
+	struct list_head i_trunc_list;
+	u32 i_entries;
+	u32 i_diskflags;
 	u8 i_height;
 	u8 i_depth;
 };
@@ -406,13 +396,11 @@ struct gfs2_args {
 struct gfs2_tune {
 	spinlock_t gt_spin;
 
-	unsigned int gt_demote_secs; /* Cache retention for unheld glock */
 	unsigned int gt_incore_log_blocks;
 	unsigned int gt_log_flush_secs;
 
 	unsigned int gt_recoverd_secs;
 	unsigned int gt_logd_secs;
-	unsigned int gt_quotad_secs;
 
 	unsigned int gt_quota_simul_sync; /* Max quotavals to sync at once */
 	unsigned int gt_quota_warn_period; /* Secs between quota warn msgs */
@@ -488,10 +476,6 @@ struct gfs2_sbd {
 	/* Lock Stuff */
 
 	struct lm_lockstruct sd_lockstruct;
-	struct list_head sd_reclaim_list;
-	spinlock_t sd_reclaim_lock;
-	wait_queue_head_t sd_reclaim_wq;
-	atomic_t sd_reclaim_count;
 	struct gfs2_holder sd_live_gh;
 	struct gfs2_glock *sd_rename_gl;
 	struct gfs2_glock *sd_trans_gl;
@@ -519,7 +503,6 @@ struct gfs2_sbd {
 	spinlock_t sd_statfs_spin;
 	struct gfs2_statfs_change_host sd_statfs_master;
 	struct gfs2_statfs_change_host sd_statfs_local;
-	unsigned long sd_statfs_sync_time;
 
 	/* Resource group stuff */
 
@@ -552,8 +535,6 @@ struct gfs2_sbd {
 	struct task_struct *sd_recoverd_process;
 	struct task_struct *sd_logd_process;
 	struct task_struct *sd_quotad_process;
-	struct task_struct *sd_glockd_process[GFS2_GLOCKD_MAX];
-	unsigned int sd_glockd_num;
 
 	/* Quota stuff */
 
@@ -561,13 +542,15 @@ struct gfs2_sbd {
 	atomic_t sd_quota_count;
 	spinlock_t sd_quota_spin;
 	struct mutex sd_quota_mutex;
+	wait_queue_head_t sd_quota_wait;
+	struct list_head sd_trunc_list;
+	spinlock_t sd_trunc_lock;
 
 	unsigned int sd_quota_slots;
 	unsigned int sd_quota_chunks;
 	unsigned char **sd_quota_bitmap;
 
 	u64 sd_quota_sync_gen;
-	unsigned long sd_quota_sync_time;
 
 	/* Log stuff */
 
@@ -624,10 +607,6 @@ struct gfs2_sbd {
 	struct mutex sd_freeze_lock;
 	unsigned int sd_freeze_count;
 
-	/* Counters */
-
-	atomic_t sd_reclaimed;
-
 	char sd_fsname[GFS2_FSNAME_LEN];
 	char sd_table_name[GFS2_FSNAME_LEN];
 	char sd_proto_name[GFS2_FSNAME_LEN];
diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index d57616840e8..3b87c188da4 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -32,7 +32,6 @@
 #include "log.h"
 #include "meta_io.h"
 #include "ops_address.h"
-#include "ops_inode.h"
 #include "quota.h"
 #include "rgrp.h"
 #include "trans.h"
@@ -248,7 +247,6 @@ fail:
 
 static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
 {
-	struct gfs2_dinode_host *di = &ip->i_di;
 	const struct gfs2_dinode *str = buf;
 	struct timespec atime;
 	u16 height, depth;
@@ -274,8 +272,8 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
 	 * to do that.
 	 */
 	ip->i_inode.i_nlink = be32_to_cpu(str->di_nlink);
-	di->di_size = be64_to_cpu(str->di_size);
-	i_size_write(&ip->i_inode, di->di_size);
+	ip->i_disksize = be64_to_cpu(str->di_size);
+	i_size_write(&ip->i_inode, ip->i_disksize);
 	gfs2_set_inode_blocks(&ip->i_inode, be64_to_cpu(str->di_blocks));
 	atime.tv_sec = be64_to_cpu(str->di_atime);
 	atime.tv_nsec = be32_to_cpu(str->di_atime_nsec);
@@ -287,9 +285,9 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
 	ip->i_inode.i_ctime.tv_nsec = be32_to_cpu(str->di_ctime_nsec);
 
 	ip->i_goal = be64_to_cpu(str->di_goal_meta);
-	di->di_generation = be64_to_cpu(str->di_generation);
+	ip->i_generation = be64_to_cpu(str->di_generation);
 
-	di->di_flags = be32_to_cpu(str->di_flags);
+	ip->i_diskflags = be32_to_cpu(str->di_flags);
 	gfs2_set_inode_flags(&ip->i_inode);
 	height = be16_to_cpu(str->di_height);
 	if (unlikely(height > GFS2_MAX_META_HEIGHT))
@@ -300,9 +298,9 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf)
 	if (unlikely(depth > GFS2_DIR_MAX_DEPTH))
 		goto corrupt;
 	ip->i_depth = (u8)depth;
-	di->di_entries = be32_to_cpu(str->di_entries);
+	ip->i_entries = be32_to_cpu(str->di_entries);
 
-	di->di_eattr = be64_to_cpu(str->di_eattr);
+	ip->i_eattr = be64_to_cpu(str->di_eattr);
 	if (S_ISREG(ip->i_inode.i_mode))
 		gfs2_set_aops(&ip->i_inode);
 
@@ -388,7 +386,6 @@ int gfs2_dinode_dealloc(struct gfs2_inode *ip)
 	gfs2_free_di(rgd, ip);
 
 	gfs2_trans_end(sdp);
-	clear_bit(GLF_STICKY, &ip->i_gl->gl_flags);
 
 out_rg_gunlock:
 	gfs2_glock_dq_uninit(&al->al_rgd_gh);
@@ -690,7 +687,7 @@ static int create_ok(struct gfs2_inode *dip, const struct qstr *name,
 		return error;
 	}
 
-	if (dip->i_di.di_entries == (u32)-1)
+	if (dip->i_entries == (u32)-1)
 		return -EFBIG;
 	if (S_ISDIR(mode) && dip->i_inode.i_nlink == (u32)-1)
 		return -EMLINK;
@@ -790,11 +787,11 @@ static void init_dinode(struct gfs2_inode *dip, struct gfs2_glock *gl,
 	di->di_flags = 0;
 
 	if (S_ISREG(mode)) {
-		if ((dip->i_di.di_flags & GFS2_DIF_INHERIT_JDATA) ||
+		if ((dip->i_diskflags & GFS2_DIF_INHERIT_JDATA) ||
 		    gfs2_tune_get(sdp, gt_new_files_jdata))
 			di->di_flags |= cpu_to_be32(GFS2_DIF_JDATA);
 	} else if (S_ISDIR(mode)) {
-		di->di_flags |= cpu_to_be32(dip->i_di.di_flags &
+		di->di_flags |= cpu_to_be32(dip->i_diskflags &
 					    GFS2_DIF_INHERIT_JDATA);
 	}
 
@@ -1068,7 +1065,7 @@ int gfs2_rmdiri(struct gfs2_inode *dip, const struct qstr *name,
 	struct qstr dotname;
 	int error;
 
-	if (ip->i_di.di_entries != 2) {
+	if (ip->i_entries != 2) {
 		if (gfs2_consist_inode(ip))
 			gfs2_dinode_print(ip);
 		return -EIO;
@@ -1168,7 +1165,7 @@ int gfs2_readlinki(struct gfs2_inode *ip, char **buf, unsigned int *len)
 		return error;
 	}
 
-	if (!ip->i_di.di_size) {
+	if (!ip->i_disksize) {
 		gfs2_consist_inode(ip);
 		error = -EIO;
 		goto out;
@@ -1178,7 +1175,7 @@ int gfs2_readlinki(struct gfs2_inode *ip, char **buf, unsigned int *len)
 	if (error)
 		goto out;
 
-	x = ip->i_di.di_size + 1;
+	x = ip->i_disksize + 1;
 	if (x > *len) {
 		*buf = kmalloc(x, GFP_NOFS);
 		if (!*buf) {
@@ -1242,7 +1239,6 @@ int gfs2_setattr_simple(struct gfs2_inode *ip, struct iattr *attr)
 
 void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf)
 {
-	const struct gfs2_dinode_host *di = &ip->i_di;
 	struct gfs2_dinode *str = buf;
 
 	str->di_header.mh_magic = cpu_to_be32(GFS2_MAGIC);
@@ -1256,7 +1252,7 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf)
 	str->di_uid = cpu_to_be32(ip->i_inode.i_uid);
 	str->di_gid = cpu_to_be32(ip->i_inode.i_gid);
 	str->di_nlink = cpu_to_be32(ip->i_inode.i_nlink);
-	str->di_size = cpu_to_be64(di->di_size);
+	str->di_size = cpu_to_be64(ip->i_disksize);
 	str->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode));
 	str->di_atime = cpu_to_be64(ip->i_inode.i_atime.tv_sec);
 	str->di_mtime = cpu_to_be64(ip->i_inode.i_mtime.tv_sec);
@@ -1264,17 +1260,17 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf)
 
 	str->di_goal_meta = cpu_to_be64(ip->i_goal);
 	str->di_goal_data = cpu_to_be64(ip->i_goal);
-	str->di_generation = cpu_to_be64(di->di_generation);
+	str->di_generation = cpu_to_be64(ip->i_generation);
 
-	str->di_flags = cpu_to_be32(di->di_flags);
+	str->di_flags = cpu_to_be32(ip->i_diskflags);
 	str->di_height = cpu_to_be16(ip->i_height);
 	str->di_payload_format = cpu_to_be32(S_ISDIR(ip->i_inode.i_mode) &&
-					     !(ip->i_di.di_flags & GFS2_DIF_EXHASH) ?
+					     !(ip->i_diskflags & GFS2_DIF_EXHASH) ?
 					     GFS2_FORMAT_DE : 0);
 	str->di_depth = cpu_to_be16(ip->i_depth);
-	str->di_entries = cpu_to_be32(di->di_entries);
+	str->di_entries = cpu_to_be32(ip->i_entries);
 
-	str->di_eattr = cpu_to_be64(di->di_eattr);
+	str->di_eattr = cpu_to_be64(ip->i_eattr);
 	str->di_atime_nsec = cpu_to_be32(ip->i_inode.i_atime.tv_nsec);
 	str->di_mtime_nsec = cpu_to_be32(ip->i_inode.i_mtime.tv_nsec);
 	str->di_ctime_nsec = cpu_to_be32(ip->i_inode.i_ctime.tv_nsec);
@@ -1282,22 +1278,21 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf)
 
 void gfs2_dinode_print(const struct gfs2_inode *ip)
 {
-	const struct gfs2_dinode_host *di = &ip->i_di;
-
 	printk(KERN_INFO "  no_formal_ino = %llu\n",
 	       (unsigned long long)ip->i_no_formal_ino);
 	printk(KERN_INFO "  no_addr = %llu\n",
 	       (unsigned long long)ip->i_no_addr);
-	printk(KERN_INFO "  di_size = %llu\n", (unsigned long long)di->di_size);
+	printk(KERN_INFO "  i_disksize = %llu\n",
+	       (unsigned long long)ip->i_disksize);
 	printk(KERN_INFO "  blocks = %llu\n",
 	       (unsigned long long)gfs2_get_inode_blocks(&ip->i_inode));
 	printk(KERN_INFO "  i_goal = %llu\n",
 	       (unsigned long long)ip->i_goal);
-	printk(KERN_INFO "  di_flags = 0x%.8X\n", di->di_flags);
+	printk(KERN_INFO "  i_diskflags = 0x%.8X\n", ip->i_diskflags);
 	printk(KERN_INFO "  i_height = %u\n", ip->i_height);
 	printk(KERN_INFO "  i_depth = %u\n", ip->i_depth);
-	printk(KERN_INFO "  di_entries = %u\n", di->di_entries);
-	printk(KERN_INFO "  di_eattr = %llu\n",
-	       (unsigned long long)di->di_eattr);
+	printk(KERN_INFO "  i_entries = %u\n", ip->i_entries);
+	printk(KERN_INFO "  i_eattr = %llu\n",
+	       (unsigned long long)ip->i_eattr);
 }
 
diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h
index 2d43f69610a..d5329364cdf 100644
--- a/fs/gfs2/inode.h
+++ b/fs/gfs2/inode.h
@@ -10,6 +10,7 @@
 #ifndef __INODE_DOT_H__
 #define __INODE_DOT_H__
 
+#include <linux/fs.h>
 #include "util.h"
 
 static inline int gfs2_is_stuffed(const struct gfs2_inode *ip)
@@ -19,7 +20,7 @@ static inline int gfs2_is_stuffed(const struct gfs2_inode *ip)
 
 static inline int gfs2_is_jdata(const struct gfs2_inode *ip)
 {
-	return ip->i_di.di_flags & GFS2_DIF_JDATA;
+	return ip->i_diskflags & GFS2_DIF_JDATA;
 }
 
 static inline int gfs2_is_writeback(const struct gfs2_inode *ip)
@@ -97,5 +98,15 @@ struct inode *gfs2_lookup_simple(struct inode *dip, const char *name);
 void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf);
 void gfs2_dinode_print(const struct gfs2_inode *ip);
 
+extern const struct inode_operations gfs2_file_iops;
+extern const struct inode_operations gfs2_dir_iops;
+extern const struct inode_operations gfs2_symlink_iops;
+extern const struct file_operations gfs2_file_fops;
+extern const struct file_operations gfs2_dir_fops;
+extern const struct file_operations gfs2_file_fops_nolock;
+extern const struct file_operations gfs2_dir_fops_nolock;
+
+extern void gfs2_set_inode_flags(struct inode *inode);
+
 #endif /* __INODE_DOT_H__ */
 
diff --git a/fs/gfs2/locking/dlm/mount.c b/fs/gfs2/locking/dlm/mount.c
index 0c4cbe6c828..1aa7eb6a022 100644
--- a/fs/gfs2/locking/dlm/mount.c
+++ b/fs/gfs2/locking/dlm/mount.c
@@ -194,17 +194,25 @@ out:
 static void gdlm_recovery_done(void *lockspace, unsigned int jid,
                                unsigned int message)
 {
+	char env_jid[20];
+	char env_status[20];
+	char *envp[] = { env_jid, env_status, NULL };
 	struct gdlm_ls *ls = lockspace;
 	ls->recover_jid_done = jid;
 	ls->recover_jid_status = message;
-	kobject_uevent(&ls->kobj, KOBJ_CHANGE);
+	sprintf(env_jid, "JID=%d", jid);
+	sprintf(env_status, "RECOVERY=%s",
+		message == LM_RD_SUCCESS ? "Done" : "Failed");
+	kobject_uevent_env(&ls->kobj, KOBJ_CHANGE, envp);
 }
 
 static void gdlm_others_may_mount(void *lockspace)
 {
+	char *message = "FIRSTMOUNT=Done";
+	char *envp[] = { message, NULL };
 	struct gdlm_ls *ls = lockspace;
 	ls->first_done = 1;
-	kobject_uevent(&ls->kobj, KOBJ_CHANGE);
+	kobject_uevent_env(&ls->kobj, KOBJ_CHANGE, envp);
 }
 
 /* Userspace gets the offline uevent, blocks new gfs locks on
diff --git a/fs/gfs2/locking/dlm/sysfs.c b/fs/gfs2/locking/dlm/sysfs.c
index 4ec571c3d8a..9b7edcf7bd4 100644
--- a/fs/gfs2/locking/dlm/sysfs.c
+++ b/fs/gfs2/locking/dlm/sysfs.c
@@ -195,9 +195,23 @@ void gdlm_kobject_release(struct gdlm_ls *ls)
 	kobject_put(&ls->kobj);
 }
 
+static int gdlm_uevent(struct kset *kset, struct kobject *kobj,
+		       struct kobj_uevent_env *env)
+{
+        struct gdlm_ls *ls = container_of(kobj, struct gdlm_ls, kobj);
+        add_uevent_var(env, "LOCKTABLE=%s:%s", ls->clustername, ls->fsname);
+        add_uevent_var(env, "LOCKPROTO=lock_dlm");
+        return 0;
+}
+
+static struct kset_uevent_ops gdlm_uevent_ops = {
+	.uevent = gdlm_uevent,
+};
+
+
 int gdlm_sysfs_init(void)
 {
-	gdlm_kset = kset_create_and_add("lock_dlm", NULL, kernel_kobj);
+	gdlm_kset = kset_create_and_add("lock_dlm", &gdlm_uevent_ops, kernel_kobj);
 	if (!gdlm_kset) {
 		printk(KERN_WARNING "%s: can not create kset\n", __func__);
 		return -ENOMEM;
diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c
index bb2cc303ac2..7cacfde3219 100644
--- a/fs/gfs2/main.c
+++ b/fs/gfs2/main.c
@@ -19,7 +19,7 @@
 
 #include "gfs2.h"
 #include "incore.h"
-#include "ops_fstype.h"
+#include "super.h"
 #include "sys.h"
 #include "util.h"
 #include "glock.h"
@@ -30,6 +30,7 @@ static void gfs2_init_inode_once(void *foo)
 
 	inode_init_once(&ip->i_inode);
 	init_rwsem(&ip->i_rw_mutex);
+	INIT_LIST_HEAD(&ip->i_trunc_list);
 	ip->i_alloc = NULL;
 }
 
@@ -42,7 +43,7 @@ static void gfs2_init_glock_once(void *foo)
 	INIT_LIST_HEAD(&gl->gl_holders);
 	gl->gl_lvb = NULL;
 	atomic_set(&gl->gl_lvb_count, 0);
-	INIT_LIST_HEAD(&gl->gl_reclaim);
+	INIT_LIST_HEAD(&gl->gl_lru);
 	INIT_LIST_HEAD(&gl->gl_ail_list);
 	atomic_set(&gl->gl_ail_count, 0);
 }
@@ -93,6 +94,12 @@ static int __init init_gfs2_fs(void)
 	if (!gfs2_rgrpd_cachep)
 		goto fail;
 
+	gfs2_quotad_cachep = kmem_cache_create("gfs2_quotad",
+					       sizeof(struct gfs2_quota_data),
+					       0, 0, NULL);
+	if (!gfs2_quotad_cachep)
+		goto fail;
+
 	error = register_filesystem(&gfs2_fs_type);
 	if (error)
 		goto fail;
@@ -112,6 +119,9 @@ fail_unregister:
 fail:
 	gfs2_glock_exit();
 
+	if (gfs2_quotad_cachep)
+		kmem_cache_destroy(gfs2_quotad_cachep);
+
 	if (gfs2_rgrpd_cachep)
 		kmem_cache_destroy(gfs2_rgrpd_cachep);
 
@@ -140,6 +150,7 @@ static void __exit exit_gfs2_fs(void)
 	unregister_filesystem(&gfs2_fs_type);
 	unregister_filesystem(&gfs2meta_fs_type);
 
+	kmem_cache_destroy(gfs2_quotad_cachep);
 	kmem_cache_destroy(gfs2_rgrpd_cachep);
 	kmem_cache_destroy(gfs2_bufdata_cachep);
 	kmem_cache_destroy(gfs2_inode_cachep);
diff --git a/fs/gfs2/mount.c b/fs/gfs2/mount.c
index f96eb90a2cf..3cb0a44ba02 100644
--- a/fs/gfs2/mount.c
+++ b/fs/gfs2/mount.c
@@ -32,7 +32,6 @@ enum {
 	Opt_debug,
 	Opt_nodebug,
 	Opt_upgrade,
-	Opt_num_glockd,
 	Opt_acl,
 	Opt_noacl,
 	Opt_quota_off,
@@ -57,7 +56,6 @@ static const match_table_t tokens = {
 	{Opt_debug, "debug"},
 	{Opt_nodebug, "nodebug"},
 	{Opt_upgrade, "upgrade"},
-	{Opt_num_glockd, "num_glockd=%d"},
 	{Opt_acl, "acl"},
 	{Opt_noacl, "noacl"},
 	{Opt_quota_off, "quota=off"},
@@ -87,16 +85,7 @@ int gfs2_mount_args(struct gfs2_sbd *sdp, char *data_arg, int remount)
 	int error = 0;
 
 	if (!remount) {
-		/*  If someone preloaded options, use those instead  */
-		spin_lock(&gfs2_sys_margs_lock);
-		if (gfs2_sys_margs) {
-			data = gfs2_sys_margs;
-			gfs2_sys_margs = NULL;
-		}
-		spin_unlock(&gfs2_sys_margs_lock);
-
 		/*  Set some defaults  */
-		args->ar_num_glockd = GFS2_GLOCKD_DEFAULT;
 		args->ar_quota = GFS2_QUOTA_DEFAULT;
 		args->ar_data = GFS2_DATA_DEFAULT;
 	}
@@ -105,7 +94,7 @@ int gfs2_mount_args(struct gfs2_sbd *sdp, char *data_arg, int remount)
 	   process them */
 
 	for (options = data; (o = strsep(&options, ",")); ) {
-		int token, option;
+		int token;
 		substring_t tmp[MAX_OPT_ARGS];
 
 		if (!*o)
@@ -196,22 +185,6 @@ int gfs2_mount_args(struct gfs2_sbd *sdp, char *data_arg, int remount)
 				goto cant_remount;
 			args->ar_upgrade = 1;
 			break;
-		case Opt_num_glockd:
-			if ((error = match_int(&tmp[0], &option))) {
-				fs_info(sdp, "problem getting num_glockd\n");
-				goto out_error;
-			}
-
-			if (remount && option != args->ar_num_glockd)
-				goto cant_remount;
-			if (!option || option > GFS2_GLOCKD_MAX) {
-				fs_info(sdp, "0 < num_glockd <= %u  (not %u)\n",
-				        GFS2_GLOCKD_MAX, option);
-				error = -EINVAL;
-				goto out_error;
-			}
-			args->ar_num_glockd = option;
-			break;
 		case Opt_acl:
 			args->ar_posix_acl = 1;
 			sdp->sd_vfs->s_flags |= MS_POSIXACL;
diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c
index 27563816e1c..4ddab67867e 100644
--- a/fs/gfs2/ops_address.c
+++ b/fs/gfs2/ops_address.c
@@ -210,25 +210,23 @@ static int gfs2_jdata_writepage(struct page *page, struct writeback_control *wbc
 {
 	struct inode *inode = page->mapping->host;
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
-	int error;
+	int ret;
 	int done_trans = 0;
 
-	error = gfs2_writepage_common(page, wbc);
-	if (error <= 0)
-		return error;
-
 	if (PageChecked(page)) {
 		if (wbc->sync_mode != WB_SYNC_ALL)
 			goto out_ignore;
-		error = gfs2_trans_begin(sdp, RES_DINODE + 1, 0);
-		if (error)
+		ret = gfs2_trans_begin(sdp, RES_DINODE + 1, 0);
+		if (ret)
 			goto out_ignore;
 		done_trans = 1;
 	}
-	error = __gfs2_jdata_writepage(page, wbc);
+	ret = gfs2_writepage_common(page, wbc);
+	if (ret > 0)
+		ret = __gfs2_jdata_writepage(page, wbc);
 	if (done_trans)
 		gfs2_trans_end(sdp);
-	return error;
+	return ret;
 
 out_ignore:
 	redirty_page_for_writepage(wbc, page);
@@ -453,8 +451,8 @@ static int stuffed_readpage(struct gfs2_inode *ip, struct page *page)
 
 	kaddr = kmap_atomic(page, KM_USER0);
 	memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode),
-	       ip->i_di.di_size);
-	memset(kaddr + ip->i_di.di_size, 0, PAGE_CACHE_SIZE - ip->i_di.di_size);
+	       ip->i_disksize);
+	memset(kaddr + ip->i_disksize, 0, PAGE_CACHE_SIZE - ip->i_disksize);
 	kunmap_atomic(kaddr, KM_USER0);
 	flush_dcache_page(page);
 	brelse(dibh);
@@ -627,7 +625,7 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 {
 	struct gfs2_inode *ip = GFS2_I(mapping->host);
 	struct gfs2_sbd *sdp = GFS2_SB(mapping->host);
-	unsigned int data_blocks, ind_blocks, rblocks;
+	unsigned int data_blocks = 0, ind_blocks = 0, rblocks;
 	int alloc_required;
 	int error = 0;
 	struct gfs2_alloc *al;
@@ -641,11 +639,13 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 	if (unlikely(error))
 		goto out_uninit;
 
-	gfs2_write_calc_reserv(ip, len, &data_blocks, &ind_blocks);
 	error = gfs2_write_alloc_required(ip, pos, len, &alloc_required);
 	if (error)
 		goto out_unlock;
 
+	if (alloc_required || gfs2_is_jdata(ip))
+		gfs2_write_calc_reserv(ip, len, &data_blocks, &ind_blocks);
+
 	if (alloc_required) {
 		al = gfs2_alloc_get(ip);
 		if (!al) {
@@ -675,7 +675,8 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping,
 		goto out_trans_fail;
 
 	error = -ENOMEM;
-	page = __grab_cache_page(mapping, index);
+	flags |= AOP_FLAG_NOFS;
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	*pagep = page;
 	if (unlikely(!page))
 		goto out_endtrans;
@@ -782,7 +783,7 @@ static int gfs2_stuffed_write_end(struct inode *inode, struct buffer_head *dibh,
 
 	if (inode->i_size < to) {
 		i_size_write(inode, to);
-		ip->i_di.di_size = inode->i_size;
+		ip->i_disksize = inode->i_size;
 		di->di_size = cpu_to_be64(inode->i_size);
 		mark_inode_dirty(inode);
 	}
@@ -847,9 +848,9 @@ static int gfs2_write_end(struct file *file, struct address_space *mapping,
 
 	ret = generic_write_end(file, mapping, pos, len, copied, page, fsdata);
 
-	if (likely(ret >= 0) && (inode->i_size > ip->i_di.di_size)) {
+	if (likely(ret >= 0) && (inode->i_size > ip->i_disksize)) {
 		di = (struct gfs2_dinode *)dibh->b_data;
-		ip->i_di.di_size = inode->i_size;
+		ip->i_disksize = inode->i_size;
 		di->di_size = cpu_to_be64(inode->i_size);
 		mark_inode_dirty(inode);
 	}
diff --git a/fs/gfs2/ops_dentry.c b/fs/gfs2/ops_dentry.c
index 4a5e676b442..c2ad36330ca 100644
--- a/fs/gfs2/ops_dentry.c
+++ b/fs/gfs2/ops_dentry.c
@@ -19,7 +19,7 @@
 #include "incore.h"
 #include "dir.h"
 #include "glock.h"
-#include "ops_dentry.h"
+#include "super.h"
 #include "util.h"
 #include "inode.h"
 
diff --git a/fs/gfs2/ops_dentry.h b/fs/gfs2/ops_dentry.h
deleted file mode 100644
index 5caa3db4d3f..00000000000
--- a/fs/gfs2/ops_dentry.h
+++ /dev/null
@@ -1,17 +0,0 @@
-/*
- * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
- * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License version 2.
- */
-
-#ifndef __OPS_DENTRY_DOT_H__
-#define __OPS_DENTRY_DOT_H__
-
-#include <linux/dcache.h>
-
-extern struct dentry_operations gfs2_dops;
-
-#endif /* __OPS_DENTRY_DOT_H__ */
diff --git a/fs/gfs2/ops_export.c b/fs/gfs2/ops_export.c
index bbb8c36403a..7fdeb14ddd1 100644
--- a/fs/gfs2/ops_export.c
+++ b/fs/gfs2/ops_export.c
@@ -22,8 +22,7 @@
 #include "glock.h"
 #include "glops.h"
 #include "inode.h"
-#include "ops_dentry.h"
-#include "ops_fstype.h"
+#include "super.h"
 #include "rgrp.h"
 #include "util.h"
 
@@ -214,7 +213,7 @@ static struct dentry *gfs2_get_dentry(struct super_block *sb,
 	}
 
 	error = -EIO;
-	if (GFS2_I(inode)->i_di.di_flags & GFS2_DIF_SYSTEM) {
+	if (GFS2_I(inode)->i_diskflags & GFS2_DIF_SYSTEM) {
 		iput(inode);
 		goto fail;
 	}
diff --git a/fs/gfs2/ops_file.c b/fs/gfs2/ops_file.c
index 3a747f8e218..93fe41b67f9 100644
--- a/fs/gfs2/ops_file.c
+++ b/fs/gfs2/ops_file.c
@@ -39,7 +39,6 @@
 #include "util.h"
 #include "eaops.h"
 #include "ops_address.h"
-#include "ops_inode.h"
 
 /**
  * gfs2_llseek - seek to a location in a file
@@ -158,8 +157,8 @@ static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
 	if (error)
 		return error;
 
-	fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_di.di_flags);
-	if (!S_ISDIR(inode->i_mode) && ip->i_di.di_flags & GFS2_DIF_JDATA)
+	fsflags = fsflags_cvt(gfs2_to_fsflags, ip->i_diskflags);
+	if (!S_ISDIR(inode->i_mode) && ip->i_diskflags & GFS2_DIF_JDATA)
 		fsflags |= FS_JOURNAL_DATA_FL;
 	if (put_user(fsflags, ptr))
 		error = -EFAULT;
@@ -172,17 +171,16 @@ static int gfs2_get_flags(struct file *filp, u32 __user *ptr)
 void gfs2_set_inode_flags(struct inode *inode)
 {
 	struct gfs2_inode *ip = GFS2_I(inode);
-	struct gfs2_dinode_host *di = &ip->i_di;
 	unsigned int flags = inode->i_flags;
 
 	flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
-	if (di->di_flags & GFS2_DIF_IMMUTABLE)
+	if (ip->i_diskflags & GFS2_DIF_IMMUTABLE)
 		flags |= S_IMMUTABLE;
-	if (di->di_flags & GFS2_DIF_APPENDONLY)
+	if (ip->i_diskflags & GFS2_DIF_APPENDONLY)
 		flags |= S_APPEND;
-	if (di->di_flags & GFS2_DIF_NOATIME)
+	if (ip->i_diskflags & GFS2_DIF_NOATIME)
 		flags |= S_NOATIME;
-	if (di->di_flags & GFS2_DIF_SYNC)
+	if (ip->i_diskflags & GFS2_DIF_SYNC)
 		flags |= S_SYNC;
 	inode->i_flags = flags;
 }
@@ -221,7 +219,7 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 	if (error)
 		goto out_drop_write;
 
-	flags = ip->i_di.di_flags;
+	flags = ip->i_diskflags;
 	new_flags = (flags & ~mask) | (reqflags & mask);
 	if ((new_flags ^ flags) == 0)
 		goto out;
@@ -260,7 +258,7 @@ static int do_gfs2_set_flags(struct file *filp, u32 reqflags, u32 mask)
 	if (error)
 		goto out_trans_end;
 	gfs2_trans_add_bh(ip->i_gl, bh, 1);
-	ip->i_di.di_flags = new_flags;
+	ip->i_diskflags = new_flags;
 	gfs2_dinode_out(ip, bh->b_data);
 	brelse(bh);
 	gfs2_set_inode_flags(inode);
@@ -344,7 +342,7 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct page *page)
 	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	unsigned long last_index;
-	u64 pos = page->index << (PAGE_CACHE_SIZE - inode->i_blkbits);
+	u64 pos = page->index << PAGE_CACHE_SHIFT;
 	unsigned int data_blocks, ind_blocks, rblocks;
 	int alloc_required = 0;
 	struct gfs2_holder gh;
@@ -357,7 +355,6 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct page *page)
 		goto out;
 
 	set_bit(GIF_SW_PAGED, &ip->i_flags);
-	gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
 	ret = gfs2_write_alloc_required(ip, pos, PAGE_CACHE_SIZE, &alloc_required);
 	if (ret || !alloc_required)
 		goto out_unlock;
@@ -369,6 +366,7 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct page *page)
 	ret = gfs2_quota_lock_check(ip);
 	if (ret)
 		goto out_alloc_put;
+	gfs2_write_calc_reserv(ip, PAGE_CACHE_SIZE, &data_blocks, &ind_blocks);
 	al->al_requested = data_blocks + ind_blocks;
 	ret = gfs2_inplace_reserve(ip);
 	if (ret)
@@ -479,7 +477,7 @@ static int gfs2_open(struct inode *inode, struct file *file)
 			goto fail;
 
 		if (!(file->f_flags & O_LARGEFILE) &&
-		    ip->i_di.di_size > MAX_NON_LFS) {
+		    ip->i_disksize > MAX_NON_LFS) {
 			error = -EOVERFLOW;
 			goto fail_gunlock;
 		}
diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index b117fcf2c4f..f91eebdde58 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -22,20 +22,18 @@
 #include "gfs2.h"
 #include "incore.h"
 #include "bmap.h"
-#include "daemon.h"
 #include "glock.h"
 #include "glops.h"
 #include "inode.h"
 #include "mount.h"
-#include "ops_fstype.h"
-#include "ops_dentry.h"
-#include "ops_super.h"
 #include "recovery.h"
 #include "rgrp.h"
 #include "super.h"
 #include "sys.h"
 #include "util.h"
 #include "log.h"
+#include "quota.h"
+#include "dir.h"
 
 #define DO 0
 #define UNDO 1
@@ -58,12 +56,10 @@ static void gfs2_tune_init(struct gfs2_tune *gt)
 {
 	spin_lock_init(&gt->gt_spin);
 
-	gt->gt_demote_secs = 300;
 	gt->gt_incore_log_blocks = 1024;
 	gt->gt_log_flush_secs = 60;
 	gt->gt_recoverd_secs = 60;
 	gt->gt_logd_secs = 1;
-	gt->gt_quotad_secs = 5;
 	gt->gt_quota_simul_sync = 64;
 	gt->gt_quota_warn_period = 10;
 	gt->gt_quota_scale_num = 1;
@@ -91,10 +87,6 @@ static struct gfs2_sbd *init_sbd(struct super_block *sb)
 
 	gfs2_tune_init(&sdp->sd_tune);
 
-	INIT_LIST_HEAD(&sdp->sd_reclaim_list);
-	spin_lock_init(&sdp->sd_reclaim_lock);
-	init_waitqueue_head(&sdp->sd_reclaim_wq);
-
 	mutex_init(&sdp->sd_inum_mutex);
 	spin_lock_init(&sdp->sd_statfs_spin);
 
@@ -110,6 +102,9 @@ static struct gfs2_sbd *init_sbd(struct super_block *sb)
 	INIT_LIST_HEAD(&sdp->sd_quota_list);
 	spin_lock_init(&sdp->sd_quota_spin);
 	mutex_init(&sdp->sd_quota_mutex);
+	init_waitqueue_head(&sdp->sd_quota_wait);
+	INIT_LIST_HEAD(&sdp->sd_trunc_list);
+	spin_lock_init(&sdp->sd_trunc_lock);
 
 	spin_lock_init(&sdp->sd_log_lock);
 
@@ -443,24 +438,11 @@ out:
 static int init_locking(struct gfs2_sbd *sdp, struct gfs2_holder *mount_gh,
 			int undo)
 {
-	struct task_struct *p;
 	int error = 0;
 
 	if (undo)
 		goto fail_trans;
 
-	for (sdp->sd_glockd_num = 0;
-	     sdp->sd_glockd_num < sdp->sd_args.ar_num_glockd;
-	     sdp->sd_glockd_num++) {
-		p = kthread_run(gfs2_glockd, sdp, "gfs2_glockd");
-		error = IS_ERR(p);
-		if (error) {
-			fs_err(sdp, "can't start glockd thread: %d\n", error);
-			goto fail;
-		}
-		sdp->sd_glockd_process[sdp->sd_glockd_num] = p;
-	}
-
 	error = gfs2_glock_nq_num(sdp,
 				  GFS2_MOUNT_LOCK, &gfs2_nondisk_glops,
 				  LM_ST_EXCLUSIVE, LM_FLAG_NOEXP | GL_NOCACHE,
@@ -493,7 +475,6 @@ static int init_locking(struct gfs2_sbd *sdp, struct gfs2_holder *mount_gh,
 		fs_err(sdp, "can't create transaction glock: %d\n", error);
 		goto fail_rename;
 	}
-	set_bit(GLF_STICKY, &sdp->sd_trans_gl->gl_flags);
 
 	return 0;
 
@@ -506,9 +487,6 @@ fail_live:
 fail_mount:
 	gfs2_glock_dq_uninit(mount_gh);
 fail:
-	while (sdp->sd_glockd_num--)
-		kthread_stop(sdp->sd_glockd_process[sdp->sd_glockd_num]);
-
 	return error;
 }
 
@@ -620,7 +598,7 @@ static int map_journal_extents(struct gfs2_sbd *sdp)
 
 	prev_db = 0;
 
-	for (lb = 0; lb < ip->i_di.di_size >> sdp->sd_sb.sb_bsize_shift; lb++) {
+	for (lb = 0; lb < ip->i_disksize >> sdp->sd_sb.sb_bsize_shift; lb++) {
 		bh.b_state = 0;
 		bh.b_blocknr = 0;
 		bh.b_size = 1 << ip->i_inode.i_blkbits;
@@ -661,6 +639,72 @@ static void gfs2_lm_others_may_mount(struct gfs2_sbd *sdp)
 					sdp->sd_lockstruct.ls_lockspace);
 }
 
+/**
+ * gfs2_jindex_hold - Grab a lock on the jindex
+ * @sdp: The GFS2 superblock
+ * @ji_gh: the holder for the jindex glock
+ *
+ * Returns: errno
+ */
+
+static int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh)
+{
+	struct gfs2_inode *dip = GFS2_I(sdp->sd_jindex);
+	struct qstr name;
+	char buf[20];
+	struct gfs2_jdesc *jd;
+	int error;
+
+	name.name = buf;
+
+	mutex_lock(&sdp->sd_jindex_mutex);
+
+	for (;;) {
+		error = gfs2_glock_nq_init(dip->i_gl, LM_ST_SHARED, 0, ji_gh);
+		if (error)
+			break;
+
+		name.len = sprintf(buf, "journal%u", sdp->sd_journals);
+		name.hash = gfs2_disk_hash(name.name, name.len);
+
+		error = gfs2_dir_check(sdp->sd_jindex, &name, NULL);
+		if (error == -ENOENT) {
+			error = 0;
+			break;
+		}
+
+		gfs2_glock_dq_uninit(ji_gh);
+
+		if (error)
+			break;
+
+		error = -ENOMEM;
+		jd = kzalloc(sizeof(struct gfs2_jdesc), GFP_KERNEL);
+		if (!jd)
+			break;
+
+		INIT_LIST_HEAD(&jd->extent_list);
+		jd->jd_inode = gfs2_lookupi(sdp->sd_jindex, &name, 1);
+		if (!jd->jd_inode || IS_ERR(jd->jd_inode)) {
+			if (!jd->jd_inode)
+				error = -ENOENT;
+			else
+				error = PTR_ERR(jd->jd_inode);
+			kfree(jd);
+			break;
+		}
+
+		spin_lock(&sdp->sd_jindex_spin);
+		jd->jd_jid = sdp->sd_journals++;
+		list_add_tail(&jd->jd_list, &sdp->sd_jindex_list);
+		spin_unlock(&sdp->sd_jindex_spin);
+	}
+
+	mutex_unlock(&sdp->sd_jindex_mutex);
+
+	return error;
+}
+
 static int init_journal(struct gfs2_sbd *sdp, int undo)
 {
 	struct inode *master = sdp->sd_master_dir->d_inode;
@@ -681,7 +725,6 @@ static int init_journal(struct gfs2_sbd *sdp, int undo)
 		return PTR_ERR(sdp->sd_jindex);
 	}
 	ip = GFS2_I(sdp->sd_jindex);
-	set_bit(GLF_STICKY, &ip->i_gl->gl_flags);
 
 	/* Load in the journal index special file */
 
@@ -832,7 +875,6 @@ static int init_inodes(struct gfs2_sbd *sdp, int undo)
 		goto fail_statfs;
 	}
 	ip = GFS2_I(sdp->sd_rindex);
-	set_bit(GLF_STICKY, &ip->i_gl->gl_flags);
 	sdp->sd_rindex_uptodate = 0;
 
 	/* Read in the quota inode */
@@ -973,9 +1015,6 @@ static int init_threads(struct gfs2_sbd *sdp, int undo)
 	}
 	sdp->sd_logd_process = p;
 
-	sdp->sd_statfs_sync_time = jiffies;
-	sdp->sd_quota_sync_time = jiffies;
-
 	p = kthread_run(gfs2_quotad, sdp, "gfs2_quotad");
 	error = IS_ERR(p);
 	if (error) {
@@ -1224,17 +1263,21 @@ static int gfs2_get_sb_meta(struct file_system_type *fs_type, int flags,
 static void gfs2_kill_sb(struct super_block *sb)
 {
 	struct gfs2_sbd *sdp = sb->s_fs_info;
-	if (sdp) {
-		gfs2_meta_syncfs(sdp);
-		dput(sdp->sd_root_dir);
-		dput(sdp->sd_master_dir);
-		sdp->sd_root_dir = NULL;
-		sdp->sd_master_dir = NULL;
+
+	if (sdp == NULL) {
+		kill_block_super(sb);
+		return;
 	}
+
+	gfs2_meta_syncfs(sdp);
+	dput(sdp->sd_root_dir);
+	dput(sdp->sd_master_dir);
+	sdp->sd_root_dir = NULL;
+	sdp->sd_master_dir = NULL;
 	shrink_dcache_sb(sb);
 	kill_block_super(sb);
-	if (sdp)
-		gfs2_delete_debugfs_file(sdp);
+	gfs2_delete_debugfs_file(sdp);
+	kfree(sdp);
 }
 
 struct file_system_type gfs2_fs_type = {
diff --git a/fs/gfs2/ops_fstype.h b/fs/gfs2/ops_fstype.h
deleted file mode 100644
index da849051183..00000000000
--- a/fs/gfs2/ops_fstype.h
+++ /dev/null
@@ -1,19 +0,0 @@
-/*
- * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
- * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License version 2.
- */
-
-#ifndef __OPS_FSTYPE_DOT_H__
-#define __OPS_FSTYPE_DOT_H__
-
-#include <linux/fs.h>
-
-extern struct file_system_type gfs2_fs_type;
-extern struct file_system_type gfs2meta_fs_type;
-extern const struct export_operations gfs2_export_ops;
-
-#endif /* __OPS_FSTYPE_DOT_H__ */
diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c
index d232991b904..49877546beb 100644
--- a/fs/gfs2/ops_inode.c
+++ b/fs/gfs2/ops_inode.c
@@ -19,6 +19,7 @@
 #include <linux/gfs2_ondisk.h>
 #include <linux/crc32.h>
 #include <linux/lm_interface.h>
+#include <linux/fiemap.h>
 #include <asm/uaccess.h>
 
 #include "gfs2.h"
@@ -31,12 +32,11 @@
 #include "glock.h"
 #include "inode.h"
 #include "meta_io.h"
-#include "ops_dentry.h"
-#include "ops_inode.h"
 #include "quota.h"
 #include "rgrp.h"
 #include "trans.h"
 #include "util.h"
+#include "super.h"
 
 /**
  * gfs2_create - Create a file
@@ -185,7 +185,7 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir,
 	if (!dip->i_inode.i_nlink)
 		goto out_gunlock;
 	error = -EFBIG;
-	if (dip->i_di.di_entries == (u32)-1)
+	if (dip->i_entries == (u32)-1)
 		goto out_gunlock;
 	error = -EPERM;
 	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
@@ -371,7 +371,7 @@ static int gfs2_symlink(struct inode *dir, struct dentry *dentry,
 
 	ip = ghs[1].gh_gl->gl_object;
 
-	ip->i_di.di_size = size;
+	ip->i_disksize = size;
 
 	error = gfs2_meta_inode_buffer(ip, &dibh);
 
@@ -425,9 +425,9 @@ static int gfs2_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	ip = ghs[1].gh_gl->gl_object;
 
 	ip->i_inode.i_nlink = 2;
-	ip->i_di.di_size = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode);
-	ip->i_di.di_flags |= GFS2_DIF_JDATA;
-	ip->i_di.di_entries = 2;
+	ip->i_disksize = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode);
+	ip->i_diskflags |= GFS2_DIF_JDATA;
+	ip->i_entries = 2;
 
 	error = gfs2_meta_inode_buffer(ip, &dibh);
 
@@ -517,13 +517,13 @@ static int gfs2_rmdir(struct inode *dir, struct dentry *dentry)
 	if (error)
 		goto out_gunlock;
 
-	if (ip->i_di.di_entries < 2) {
+	if (ip->i_entries < 2) {
 		if (gfs2_consist_inode(ip))
 			gfs2_dinode_print(ip);
 		error = -EIO;
 		goto out_gunlock;
 	}
-	if (ip->i_di.di_entries > 2) {
+	if (ip->i_entries > 2) {
 		error = -ENOTEMPTY;
 		goto out_gunlock;
 	}
@@ -726,13 +726,13 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry,
 			goto out_gunlock;
 
 		if (S_ISDIR(nip->i_inode.i_mode)) {
-			if (nip->i_di.di_entries < 2) {
+			if (nip->i_entries < 2) {
 				if (gfs2_consist_inode(nip))
 					gfs2_dinode_print(nip);
 				error = -EIO;
 				goto out_gunlock;
 			}
-			if (nip->i_di.di_entries > 2) {
+			if (nip->i_entries > 2) {
 				error = -ENOTEMPTY;
 				goto out_gunlock;
 			}
@@ -758,7 +758,7 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry,
 				error = -EINVAL;
 				goto out_gunlock;
 			}
-			if (ndip->i_di.di_entries == (u32)-1) {
+			if (ndip->i_entries == (u32)-1) {
 				error = -EFBIG;
 				goto out_gunlock;
 			}
@@ -990,7 +990,7 @@ static int setattr_size(struct inode *inode, struct iattr *attr)
 	struct gfs2_sbd *sdp = GFS2_SB(inode);
 	int error;
 
-	if (attr->ia_size != ip->i_di.di_size) {
+	if (attr->ia_size != ip->i_disksize) {
 		error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
 		if (error)
 			return error;
@@ -1001,8 +1001,8 @@ static int setattr_size(struct inode *inode, struct iattr *attr)
 	}
 
 	error = gfs2_truncatei(ip, attr->ia_size);
-	if (error && (inode->i_size != ip->i_di.di_size))
-		i_size_write(inode, ip->i_di.di_size);
+	if (error && (inode->i_size != ip->i_disksize))
+		i_size_write(inode, ip->i_disksize);
 
 	return error;
 }
@@ -1212,6 +1212,48 @@ static int gfs2_removexattr(struct dentry *dentry, const char *name)
 	return gfs2_ea_remove(GFS2_I(dentry->d_inode), &er);
 }
 
+static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+		       u64 start, u64 len)
+{
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_holder gh;
+	int ret;
+
+	ret = fiemap_check_flags(fieinfo, FIEMAP_FLAG_SYNC);
+	if (ret)
+		return ret;
+
+	mutex_lock(&inode->i_mutex);
+
+	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &gh);
+	if (ret)
+		goto out;
+
+	if (gfs2_is_stuffed(ip)) {
+		u64 phys = ip->i_no_addr << inode->i_blkbits;
+		u64 size = i_size_read(inode);
+		u32 flags = FIEMAP_EXTENT_LAST|FIEMAP_EXTENT_NOT_ALIGNED|
+			    FIEMAP_EXTENT_DATA_INLINE;
+		phys += sizeof(struct gfs2_dinode);
+		phys += start;
+		if (start + len > size)
+			len = size - start;
+		if (start < size)
+			ret = fiemap_fill_next_extent(fieinfo, start, phys,
+						      len, flags);
+		if (ret == 1)
+			ret = 0;
+	} else {
+		ret = __generic_block_fiemap(inode, fieinfo, start, len,
+					     gfs2_block_map);
+	}
+
+	gfs2_glock_dq_uninit(&gh);
+out:
+	mutex_unlock(&inode->i_mutex);
+	return ret;
+}
+
 const struct inode_operations gfs2_file_iops = {
 	.permission = gfs2_permission,
 	.setattr = gfs2_setattr,
@@ -1220,6 +1262,7 @@ const struct inode_operations gfs2_file_iops = {
 	.getxattr = gfs2_getxattr,
 	.listxattr = gfs2_listxattr,
 	.removexattr = gfs2_removexattr,
+	.fiemap = gfs2_fiemap,
 };
 
 const struct inode_operations gfs2_dir_iops = {
@@ -1239,6 +1282,7 @@ const struct inode_operations gfs2_dir_iops = {
 	.getxattr = gfs2_getxattr,
 	.listxattr = gfs2_listxattr,
 	.removexattr = gfs2_removexattr,
+	.fiemap = gfs2_fiemap,
 };
 
 const struct inode_operations gfs2_symlink_iops = {
@@ -1251,5 +1295,6 @@ const struct inode_operations gfs2_symlink_iops = {
 	.getxattr = gfs2_getxattr,
 	.listxattr = gfs2_listxattr,
 	.removexattr = gfs2_removexattr,
+	.fiemap = gfs2_fiemap,
 };
 
diff --git a/fs/gfs2/ops_inode.h b/fs/gfs2/ops_inode.h
deleted file mode 100644
index 14b4b797622..00000000000
--- a/fs/gfs2/ops_inode.h
+++ /dev/null
@@ -1,25 +0,0 @@
-/*
- * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
- * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License version 2.
- */
-
-#ifndef __OPS_INODE_DOT_H__
-#define __OPS_INODE_DOT_H__
-
-#include <linux/fs.h>
-
-extern const struct inode_operations gfs2_file_iops;
-extern const struct inode_operations gfs2_dir_iops;
-extern const struct inode_operations gfs2_symlink_iops;
-extern const struct file_operations gfs2_file_fops;
-extern const struct file_operations gfs2_dir_fops;
-extern const struct file_operations gfs2_file_fops_nolock;
-extern const struct file_operations gfs2_dir_fops_nolock;
-
-extern void gfs2_set_inode_flags(struct inode *inode);
-
-#endif /* __OPS_INODE_DOT_H__ */
diff --git a/fs/gfs2/ops_super.c b/fs/gfs2/ops_super.c
index d5355d9b592..320323d0347 100644
--- a/fs/gfs2/ops_super.c
+++ b/fs/gfs2/ops_super.c
@@ -28,7 +28,6 @@
 #include "inode.h"
 #include "log.h"
 #include "mount.h"
-#include "ops_super.h"
 #include "quota.h"
 #include "recovery.h"
 #include "rgrp.h"
@@ -143,8 +142,6 @@ static void gfs2_put_super(struct super_block *sb)
 	kthread_stop(sdp->sd_quotad_process);
 	kthread_stop(sdp->sd_logd_process);
 	kthread_stop(sdp->sd_recoverd_process);
-	while (sdp->sd_glockd_num--)
-		kthread_stop(sdp->sd_glockd_process[sdp->sd_glockd_num]);
 
 	if (!(sb->s_flags & MS_RDONLY)) {
 		error = gfs2_make_fs_ro(sdp);
@@ -185,7 +182,6 @@ static void gfs2_put_super(struct super_block *sb)
 
 	/*  At this point, we're through participating in the lockspace  */
 	gfs2_sys_fs_del(sdp);
-	kfree(sdp);
 }
 
 /**
@@ -215,18 +211,18 @@ static int gfs2_sync_fs(struct super_block *sb, int wait)
 }
 
 /**
- * gfs2_write_super_lockfs - prevent further writes to the filesystem
+ * gfs2_freeze - prevent further writes to the filesystem
  * @sb: the VFS structure for the filesystem
  *
  */
 
-static void gfs2_write_super_lockfs(struct super_block *sb)
+static int gfs2_freeze(struct super_block *sb)
 {
 	struct gfs2_sbd *sdp = sb->s_fs_info;
 	int error;
 
 	if (test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
-		return;
+		return -EINVAL;
 
 	for (;;) {
 		error = gfs2_freeze_fs(sdp);
@@ -246,17 +242,150 @@ static void gfs2_write_super_lockfs(struct super_block *sb)
 		fs_err(sdp, "retrying...\n");
 		msleep(1000);
 	}
+	return 0;
 }
 
 /**
- * gfs2_unlockfs - reallow writes to the filesystem
+ * gfs2_unfreeze - reallow writes to the filesystem
  * @sb: the VFS structure for the filesystem
  *
  */
 
-static void gfs2_unlockfs(struct super_block *sb)
+static int gfs2_unfreeze(struct super_block *sb)
 {
 	gfs2_unfreeze_fs(sb->s_fs_info);
+	return 0;
+}
+
+/**
+ * statfs_fill - fill in the sg for a given RG
+ * @rgd: the RG
+ * @sc: the sc structure
+ *
+ * Returns: 0 on success, -ESTALE if the LVB is invalid
+ */
+
+static int statfs_slow_fill(struct gfs2_rgrpd *rgd,
+			    struct gfs2_statfs_change_host *sc)
+{
+	gfs2_rgrp_verify(rgd);
+	sc->sc_total += rgd->rd_data;
+	sc->sc_free += rgd->rd_free;
+	sc->sc_dinodes += rgd->rd_dinodes;
+	return 0;
+}
+
+/**
+ * gfs2_statfs_slow - Stat a filesystem using asynchronous locking
+ * @sdp: the filesystem
+ * @sc: the sc info that will be returned
+ *
+ * Any error (other than a signal) will cause this routine to fall back
+ * to the synchronous version.
+ *
+ * FIXME: This really shouldn't busy wait like this.
+ *
+ * Returns: errno
+ */
+
+static int gfs2_statfs_slow(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc)
+{
+	struct gfs2_holder ri_gh;
+	struct gfs2_rgrpd *rgd_next;
+	struct gfs2_holder *gha, *gh;
+	unsigned int slots = 64;
+	unsigned int x;
+	int done;
+	int error = 0, err;
+
+	memset(sc, 0, sizeof(struct gfs2_statfs_change_host));
+	gha = kcalloc(slots, sizeof(struct gfs2_holder), GFP_KERNEL);
+	if (!gha)
+		return -ENOMEM;
+
+	error = gfs2_rindex_hold(sdp, &ri_gh);
+	if (error)
+		goto out;
+
+	rgd_next = gfs2_rgrpd_get_first(sdp);
+
+	for (;;) {
+		done = 1;
+
+		for (x = 0; x < slots; x++) {
+			gh = gha + x;
+
+			if (gh->gh_gl && gfs2_glock_poll(gh)) {
+				err = gfs2_glock_wait(gh);
+				if (err) {
+					gfs2_holder_uninit(gh);
+					error = err;
+				} else {
+					if (!error)
+						error = statfs_slow_fill(
+							gh->gh_gl->gl_object, sc);
+					gfs2_glock_dq_uninit(gh);
+				}
+			}
+
+			if (gh->gh_gl)
+				done = 0;
+			else if (rgd_next && !error) {
+				error = gfs2_glock_nq_init(rgd_next->rd_gl,
+							   LM_ST_SHARED,
+							   GL_ASYNC,
+							   gh);
+				rgd_next = gfs2_rgrpd_get_next(rgd_next);
+				done = 0;
+			}
+
+			if (signal_pending(current))
+				error = -ERESTARTSYS;
+		}
+
+		if (done)
+			break;
+
+		yield();
+	}
+
+	gfs2_glock_dq_uninit(&ri_gh);
+
+out:
+	kfree(gha);
+	return error;
+}
+
+/**
+ * gfs2_statfs_i - Do a statfs
+ * @sdp: the filesystem
+ * @sg: the sg structure
+ *
+ * Returns: errno
+ */
+
+static int gfs2_statfs_i(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc)
+{
+	struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
+	struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
+
+	spin_lock(&sdp->sd_statfs_spin);
+
+	*sc = *m_sc;
+	sc->sc_total += l_sc->sc_total;
+	sc->sc_free += l_sc->sc_free;
+	sc->sc_dinodes += l_sc->sc_dinodes;
+
+	spin_unlock(&sdp->sd_statfs_spin);
+
+	if (sc->sc_free < 0)
+		sc->sc_free = 0;
+	if (sc->sc_free > sc->sc_total)
+		sc->sc_free = sc->sc_total;
+	if (sc->sc_dinodes < 0)
+		sc->sc_dinodes = 0;
+
+	return 0;
 }
 
 /**
@@ -370,7 +499,6 @@ static void gfs2_clear_inode(struct inode *inode)
 	 */
 	if (test_bit(GIF_USER, &ip->i_flags)) {
 		ip->i_gl->gl_object = NULL;
-		gfs2_glock_schedule_for_reclaim(ip->i_gl);
 		gfs2_glock_put(ip->i_gl);
 		ip->i_gl = NULL;
 		if (ip->i_iopen_gh.gh_gl) {
@@ -423,8 +551,6 @@ static int gfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 		seq_printf(s, ",debug");
 	if (args->ar_upgrade)
 		seq_printf(s, ",upgrade");
-	if (args->ar_num_glockd != GFS2_GLOCKD_DEFAULT)
-		seq_printf(s, ",num_glockd=%u", args->ar_num_glockd);
 	if (args->ar_posix_acl)
 		seq_printf(s, ",acl");
 	if (args->ar_quota != GFS2_QUOTA_DEFAULT) {
@@ -494,16 +620,16 @@ static void gfs2_delete_inode(struct inode *inode)
 	gfs2_holder_reinit(LM_ST_EXCLUSIVE, LM_FLAG_TRY_1CB | GL_NOCACHE, &ip->i_iopen_gh);
 	error = gfs2_glock_nq(&ip->i_iopen_gh);
 	if (error)
-		goto out_uninit;
+		goto out_truncate;
 
 	if (S_ISDIR(inode->i_mode) &&
-	    (ip->i_di.di_flags & GFS2_DIF_EXHASH)) {
+	    (ip->i_diskflags & GFS2_DIF_EXHASH)) {
 		error = gfs2_dir_exhash_dealloc(ip);
 		if (error)
 			goto out_unlock;
 	}
 
-	if (ip->i_di.di_eattr) {
+	if (ip->i_eattr) {
 		error = gfs2_ea_dealloc(ip);
 		if (error)
 			goto out_unlock;
@@ -519,6 +645,7 @@ static void gfs2_delete_inode(struct inode *inode)
 	if (error)
 		goto out_unlock;
 
+out_truncate:
 	error = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
 	if (error)
 		goto out_unlock;
@@ -527,8 +654,8 @@ static void gfs2_delete_inode(struct inode *inode)
 	gfs2_trans_end(sdp);
 
 out_unlock:
-	gfs2_glock_dq(&ip->i_iopen_gh);
-out_uninit:
+	if (test_bit(HIF_HOLDER, &ip->i_iopen_gh.gh_iflags))
+		gfs2_glock_dq(&ip->i_iopen_gh);
 	gfs2_holder_uninit(&ip->i_iopen_gh);
 	gfs2_glock_dq_uninit(&gh);
 	if (error && error != GLR_TRYFAILED)
@@ -563,8 +690,8 @@ const struct super_operations gfs2_super_ops = {
 	.put_super		= gfs2_put_super,
 	.write_super		= gfs2_write_super,
 	.sync_fs		= gfs2_sync_fs,
-	.write_super_lockfs 	= gfs2_write_super_lockfs,
-	.unlockfs		= gfs2_unlockfs,
+	.freeze_fs 		= gfs2_freeze,
+	.unfreeze_fs		= gfs2_unfreeze,
 	.statfs			= gfs2_statfs,
 	.remount_fs		= gfs2_remount_fs,
 	.clear_inode		= gfs2_clear_inode,
diff --git a/fs/gfs2/ops_super.h b/fs/gfs2/ops_super.h
deleted file mode 100644
index 442a274c627..00000000000
--- a/fs/gfs2/ops_super.h
+++ /dev/null
@@ -1,17 +0,0 @@
-/*
- * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
- * Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License version 2.
- */
-
-#ifndef __OPS_SUPER_DOT_H__
-#define __OPS_SUPER_DOT_H__
-
-#include <linux/fs.h>
-
-extern const struct super_operations gfs2_super_ops;
-
-#endif /* __OPS_SUPER_DOT_H__ */
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index 3e073f5144f..b08d09696b3 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -46,6 +46,8 @@
 #include <linux/bio.h>
 #include <linux/gfs2_ondisk.h>
 #include <linux/lm_interface.h>
+#include <linux/kthread.h>
+#include <linux/freezer.h>
 
 #include "gfs2.h"
 #include "incore.h"
@@ -94,7 +96,7 @@ static int qd_alloc(struct gfs2_sbd *sdp, int user, u32 id,
 	struct gfs2_quota_data *qd;
 	int error;
 
-	qd = kzalloc(sizeof(struct gfs2_quota_data), GFP_NOFS);
+	qd = kmem_cache_zalloc(gfs2_quotad_cachep, GFP_NOFS);
 	if (!qd)
 		return -ENOMEM;
 
@@ -119,7 +121,7 @@ static int qd_alloc(struct gfs2_sbd *sdp, int user, u32 id,
 	return 0;
 
 fail:
-	kfree(qd);
+	kmem_cache_free(gfs2_quotad_cachep, qd);
 	return error;
 }
 
@@ -158,7 +160,7 @@ static int qd_get(struct gfs2_sbd *sdp, int user, u32 id, int create,
 		if (qd || !create) {
 			if (new_qd) {
 				gfs2_lvb_unhold(new_qd->qd_gl);
-				kfree(new_qd);
+				kmem_cache_free(gfs2_quotad_cachep, new_qd);
 			}
 			*qdp = qd;
 			return 0;
@@ -1013,7 +1015,7 @@ void gfs2_quota_change(struct gfs2_inode *ip, s64 change,
 
 	if (gfs2_assert_warn(GFS2_SB(&ip->i_inode), change))
 		return;
-	if (ip->i_di.di_flags & GFS2_DIF_SYSTEM)
+	if (ip->i_diskflags & GFS2_DIF_SYSTEM)
 		return;
 
 	for (x = 0; x < al->al_qd_num; x++) {
@@ -1100,15 +1102,15 @@ static void gfs2_quota_change_in(struct gfs2_quota_change_host *qc, const void *
 int gfs2_quota_init(struct gfs2_sbd *sdp)
 {
 	struct gfs2_inode *ip = GFS2_I(sdp->sd_qc_inode);
-	unsigned int blocks = ip->i_di.di_size >> sdp->sd_sb.sb_bsize_shift;
+	unsigned int blocks = ip->i_disksize >> sdp->sd_sb.sb_bsize_shift;
 	unsigned int x, slot = 0;
 	unsigned int found = 0;
 	u64 dblock;
 	u32 extlen = 0;
 	int error;
 
-	if (!ip->i_di.di_size || ip->i_di.di_size > (64 << 20) ||
-	    ip->i_di.di_size & (sdp->sd_sb.sb_bsize - 1)) {
+	if (!ip->i_disksize || ip->i_disksize > (64 << 20) ||
+	    ip->i_disksize & (sdp->sd_sb.sb_bsize - 1)) {
 		gfs2_consist_inode(ip);
 		return -EIO;
 	}
@@ -1195,7 +1197,7 @@ fail:
 	return error;
 }
 
-void gfs2_quota_scan(struct gfs2_sbd *sdp)
+static void gfs2_quota_scan(struct gfs2_sbd *sdp)
 {
 	struct gfs2_quota_data *qd, *safe;
 	LIST_HEAD(dead);
@@ -1222,7 +1224,7 @@ void gfs2_quota_scan(struct gfs2_sbd *sdp)
 		gfs2_assert_warn(sdp, !qd->qd_bh_count);
 
 		gfs2_lvb_unhold(qd->qd_gl);
-		kfree(qd);
+		kmem_cache_free(gfs2_quotad_cachep, qd);
 	}
 }
 
@@ -1257,7 +1259,7 @@ void gfs2_quota_cleanup(struct gfs2_sbd *sdp)
 		gfs2_assert_warn(sdp, !qd->qd_bh_count);
 
 		gfs2_lvb_unhold(qd->qd_gl);
-		kfree(qd);
+		kmem_cache_free(gfs2_quotad_cachep, qd);
 
 		spin_lock(&sdp->sd_quota_spin);
 	}
@@ -1272,3 +1274,94 @@ void gfs2_quota_cleanup(struct gfs2_sbd *sdp)
 	}
 }
 
+static void quotad_error(struct gfs2_sbd *sdp, const char *msg, int error)
+{
+	if (error == 0 || error == -EROFS)
+		return;
+	if (!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))
+		fs_err(sdp, "gfs2_quotad: %s error %d\n", msg, error);
+}
+
+static void quotad_check_timeo(struct gfs2_sbd *sdp, const char *msg,
+			       int (*fxn)(struct gfs2_sbd *sdp),
+			       unsigned long t, unsigned long *timeo,
+			       unsigned int *new_timeo)
+{
+	if (t >= *timeo) {
+		int error = fxn(sdp);
+		quotad_error(sdp, msg, error);
+		*timeo = gfs2_tune_get_i(&sdp->sd_tune, new_timeo) * HZ;
+	} else {
+		*timeo -= t;
+	}
+}
+
+static void quotad_check_trunc_list(struct gfs2_sbd *sdp)
+{
+	struct gfs2_inode *ip;
+
+	while(1) {
+		ip = NULL;
+		spin_lock(&sdp->sd_trunc_lock);
+		if (!list_empty(&sdp->sd_trunc_list)) {
+			ip = list_entry(sdp->sd_trunc_list.next,
+					struct gfs2_inode, i_trunc_list);
+			list_del_init(&ip->i_trunc_list);
+		}
+		spin_unlock(&sdp->sd_trunc_lock);
+		if (ip == NULL)
+			return;
+		gfs2_glock_finish_truncate(ip);
+	}
+}
+
+/**
+ * gfs2_quotad - Write cached quota changes into the quota file
+ * @sdp: Pointer to GFS2 superblock
+ *
+ */
+
+int gfs2_quotad(void *data)
+{
+	struct gfs2_sbd *sdp = data;
+	struct gfs2_tune *tune = &sdp->sd_tune;
+	unsigned long statfs_timeo = 0;
+	unsigned long quotad_timeo = 0;
+	unsigned long t = 0;
+	DEFINE_WAIT(wait);
+	int empty;
+
+	while (!kthread_should_stop()) {
+
+		/* Update the master statfs file */
+		quotad_check_timeo(sdp, "statfs", gfs2_statfs_sync, t,
+				   &statfs_timeo, &tune->gt_statfs_quantum);
+
+		/* Update quota file */
+		quotad_check_timeo(sdp, "sync", gfs2_quota_sync, t,
+				   &quotad_timeo, &tune->gt_quota_quantum);
+
+		/* FIXME: This should be turned into a shrinker */
+		gfs2_quota_scan(sdp);
+
+		/* Check for & recover partially truncated inodes */
+		quotad_check_trunc_list(sdp);
+
+		if (freezing(current))
+			refrigerator();
+		t = min(quotad_timeo, statfs_timeo);
+
+		prepare_to_wait(&sdp->sd_quota_wait, &wait, TASK_UNINTERRUPTIBLE);
+		spin_lock(&sdp->sd_trunc_lock);
+		empty = list_empty(&sdp->sd_trunc_list);
+		spin_unlock(&sdp->sd_trunc_lock);
+		if (empty)
+			t -= schedule_timeout(t);
+		else
+			t = 0;
+		finish_wait(&sdp->sd_quota_wait, &wait);
+	}
+
+	return 0;
+}
+
diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h
index 3b7f4b0e5df..cec9032be97 100644
--- a/fs/gfs2/quota.h
+++ b/fs/gfs2/quota.h
@@ -15,22 +15,22 @@ struct gfs2_sbd;
 
 #define NO_QUOTA_CHANGE ((u32)-1)
 
-int gfs2_quota_hold(struct gfs2_inode *ip, u32 uid, u32 gid);
-void gfs2_quota_unhold(struct gfs2_inode *ip);
+extern int gfs2_quota_hold(struct gfs2_inode *ip, u32 uid, u32 gid);
+extern void gfs2_quota_unhold(struct gfs2_inode *ip);
 
-int gfs2_quota_lock(struct gfs2_inode *ip, u32 uid, u32 gid);
-void gfs2_quota_unlock(struct gfs2_inode *ip);
+extern int gfs2_quota_lock(struct gfs2_inode *ip, u32 uid, u32 gid);
+extern void gfs2_quota_unlock(struct gfs2_inode *ip);
 
-int gfs2_quota_check(struct gfs2_inode *ip, u32 uid, u32 gid);
-void gfs2_quota_change(struct gfs2_inode *ip, s64 change,
-		       u32 uid, u32 gid);
+extern int gfs2_quota_check(struct gfs2_inode *ip, u32 uid, u32 gid);
+extern void gfs2_quota_change(struct gfs2_inode *ip, s64 change,
+			      u32 uid, u32 gid);
 
-int gfs2_quota_sync(struct gfs2_sbd *sdp);
-int gfs2_quota_refresh(struct gfs2_sbd *sdp, int user, u32 id);
+extern int gfs2_quota_sync(struct gfs2_sbd *sdp);
+extern int gfs2_quota_refresh(struct gfs2_sbd *sdp, int user, u32 id);
 
-int gfs2_quota_init(struct gfs2_sbd *sdp);
-void gfs2_quota_scan(struct gfs2_sbd *sdp);
-void gfs2_quota_cleanup(struct gfs2_sbd *sdp);
+extern int gfs2_quota_init(struct gfs2_sbd *sdp);
+extern void gfs2_quota_cleanup(struct gfs2_sbd *sdp);
+extern int gfs2_quotad(void *data);
 
 static inline int gfs2_quota_lock_check(struct gfs2_inode *ip)
 {
diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c
index d5e91f4f6a0..efd09c3d2b2 100644
--- a/fs/gfs2/recovery.c
+++ b/fs/gfs2/recovery.c
@@ -14,6 +14,8 @@
 #include <linux/gfs2_ondisk.h>
 #include <linux/crc32.h>
 #include <linux/lm_interface.h>
+#include <linux/kthread.h>
+#include <linux/freezer.h>
 
 #include "gfs2.h"
 #include "incore.h"
@@ -583,13 +585,35 @@ fail:
 	return error;
 }
 
+static struct gfs2_jdesc *gfs2_jdesc_find_dirty(struct gfs2_sbd *sdp)
+{
+	struct gfs2_jdesc *jd;
+	int found = 0;
+
+	spin_lock(&sdp->sd_jindex_spin);
+
+	list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
+		if (jd->jd_dirty) {
+			jd->jd_dirty = 0;
+			found = 1;
+			break;
+		}
+	}
+	spin_unlock(&sdp->sd_jindex_spin);
+
+	if (!found)
+		jd = NULL;
+
+	return jd;
+}
+
 /**
  * gfs2_check_journals - Recover any dirty journals
  * @sdp: the filesystem
  *
  */
 
-void gfs2_check_journals(struct gfs2_sbd *sdp)
+static void gfs2_check_journals(struct gfs2_sbd *sdp)
 {
 	struct gfs2_jdesc *jd;
 
@@ -603,3 +627,25 @@ void gfs2_check_journals(struct gfs2_sbd *sdp)
 	}
 }
 
+/**
+ * gfs2_recoverd - Recover dead machine's journals
+ * @sdp: Pointer to GFS2 superblock
+ *
+ */
+
+int gfs2_recoverd(void *data)
+{
+	struct gfs2_sbd *sdp = data;
+	unsigned long t;
+
+	while (!kthread_should_stop()) {
+		gfs2_check_journals(sdp);
+		t = gfs2_tune_get(sdp,  gt_recoverd_secs) * HZ;
+		if (freezing(current))
+			refrigerator();
+		schedule_timeout_interruptible(t);
+	}
+
+	return 0;
+}
+
diff --git a/fs/gfs2/recovery.h b/fs/gfs2/recovery.h
index f7235e61c72..a8218ea15b5 100644
--- a/fs/gfs2/recovery.h
+++ b/fs/gfs2/recovery.h
@@ -18,17 +18,17 @@ static inline void gfs2_replay_incr_blk(struct gfs2_sbd *sdp, unsigned int *blk)
 	        *blk = 0;
 }
 
-int gfs2_replay_read_block(struct gfs2_jdesc *jd, unsigned int blk,
+extern int gfs2_replay_read_block(struct gfs2_jdesc *jd, unsigned int blk,
 			   struct buffer_head **bh);
 
-int gfs2_revoke_add(struct gfs2_sbd *sdp, u64 blkno, unsigned int where);
-int gfs2_revoke_check(struct gfs2_sbd *sdp, u64 blkno, unsigned int where);
-void gfs2_revoke_clean(struct gfs2_sbd *sdp);
+extern int gfs2_revoke_add(struct gfs2_sbd *sdp, u64 blkno, unsigned int where);
+extern int gfs2_revoke_check(struct gfs2_sbd *sdp, u64 blkno, unsigned int where);
+extern void gfs2_revoke_clean(struct gfs2_sbd *sdp);
 
-int gfs2_find_jhead(struct gfs2_jdesc *jd,
+extern int gfs2_find_jhead(struct gfs2_jdesc *jd,
 		    struct gfs2_log_header_host *head);
-int gfs2_recover_journal(struct gfs2_jdesc *gfs2_jd);
-void gfs2_check_journals(struct gfs2_sbd *sdp);
+extern int gfs2_recover_journal(struct gfs2_jdesc *gfs2_jd);
+extern int gfs2_recoverd(void *data);
 
 #endif /* __RECOVERY_DOT_H__ */
 
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 2d90fb25350..8b01c635d92 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -269,16 +269,14 @@ void gfs2_rgrp_verify(struct gfs2_rgrpd *rgd)
 						  bi->bi_len, x);
 	}
 
-	if (count[0] != rgd->rd_rg.rg_free) {
+	if (count[0] != rgd->rd_free) {
 		if (gfs2_consist_rgrpd(rgd))
 			fs_err(sdp, "free data mismatch:  %u != %u\n",
-			       count[0], rgd->rd_rg.rg_free);
+			       count[0], rgd->rd_free);
 		return;
 	}
 
-	tmp = rgd->rd_data -
-		rgd->rd_rg.rg_free -
-		rgd->rd_rg.rg_dinodes;
+	tmp = rgd->rd_data - rgd->rd_free - rgd->rd_dinodes;
 	if (count[1] + count[2] != tmp) {
 		if (gfs2_consist_rgrpd(rgd))
 			fs_err(sdp, "used data mismatch:  %u != %u\n",
@@ -286,10 +284,10 @@ void gfs2_rgrp_verify(struct gfs2_rgrpd *rgd)
 		return;
 	}
 
-	if (count[3] != rgd->rd_rg.rg_dinodes) {
+	if (count[3] != rgd->rd_dinodes) {
 		if (gfs2_consist_rgrpd(rgd))
 			fs_err(sdp, "used metadata mismatch:  %u != %u\n",
-			       count[3], rgd->rd_rg.rg_dinodes);
+			       count[3], rgd->rd_dinodes);
 		return;
 	}
 
@@ -501,7 +499,7 @@ u64 gfs2_ri_total(struct gfs2_sbd *sdp)
 	for (rgrps = 0;; rgrps++) {
 		loff_t pos = rgrps * sizeof(struct gfs2_rindex);
 
-		if (pos + sizeof(struct gfs2_rindex) >= ip->i_di.di_size)
+		if (pos + sizeof(struct gfs2_rindex) >= ip->i_disksize)
 			break;
 		error = gfs2_internal_read(ip, &ra_state, buf, &pos,
 					   sizeof(struct gfs2_rindex));
@@ -590,7 +588,7 @@ static int gfs2_ri_update(struct gfs2_inode *ip)
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
 	struct inode *inode = &ip->i_inode;
 	struct file_ra_state ra_state;
-	u64 rgrp_count = ip->i_di.di_size;
+	u64 rgrp_count = ip->i_disksize;
 	int error;
 
 	if (do_div(rgrp_count, sizeof(struct gfs2_rindex))) {
@@ -634,7 +632,7 @@ static int gfs2_ri_update_special(struct gfs2_inode *ip)
 	for (sdp->sd_rgrps = 0;; sdp->sd_rgrps++) {
 		/* Ignore partials */
 		if ((sdp->sd_rgrps + 1) * sizeof(struct gfs2_rindex) >
-		    ip->i_di.di_size)
+		    ip->i_disksize)
 			break;
 		error = read_rindex_entry(ip, &ra_state);
 		if (error) {
@@ -692,7 +690,6 @@ int gfs2_rindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ri_gh)
 static void gfs2_rgrp_in(struct gfs2_rgrpd *rgd, const void *buf)
 {
 	const struct gfs2_rgrp *str = buf;
-	struct gfs2_rgrp_host *rg = &rgd->rd_rg;
 	u32 rg_flags;
 
 	rg_flags = be32_to_cpu(str->rg_flags);
@@ -700,24 +697,23 @@ static void gfs2_rgrp_in(struct gfs2_rgrpd *rgd, const void *buf)
 		rgd->rd_flags |= GFS2_RDF_NOALLOC;
 	else
 		rgd->rd_flags &= ~GFS2_RDF_NOALLOC;
-	rg->rg_free = be32_to_cpu(str->rg_free);
-	rg->rg_dinodes = be32_to_cpu(str->rg_dinodes);
-	rg->rg_igeneration = be64_to_cpu(str->rg_igeneration);
+	rgd->rd_free = be32_to_cpu(str->rg_free);
+	rgd->rd_dinodes = be32_to_cpu(str->rg_dinodes);
+	rgd->rd_igeneration = be64_to_cpu(str->rg_igeneration);
 }
 
 static void gfs2_rgrp_out(struct gfs2_rgrpd *rgd, void *buf)
 {
 	struct gfs2_rgrp *str = buf;
-	struct gfs2_rgrp_host *rg = &rgd->rd_rg;
 	u32 rg_flags = 0;
 
 	if (rgd->rd_flags & GFS2_RDF_NOALLOC)
 		rg_flags |= GFS2_RGF_NOALLOC;
 	str->rg_flags = cpu_to_be32(rg_flags);
-	str->rg_free = cpu_to_be32(rg->rg_free);
-	str->rg_dinodes = cpu_to_be32(rg->rg_dinodes);
+	str->rg_free = cpu_to_be32(rgd->rd_free);
+	str->rg_dinodes = cpu_to_be32(rgd->rd_dinodes);
 	str->__pad = cpu_to_be32(0);
-	str->rg_igeneration = cpu_to_be64(rg->rg_igeneration);
+	str->rg_igeneration = cpu_to_be64(rgd->rd_igeneration);
 	memset(&str->rg_reserved, 0, sizeof(str->rg_reserved));
 }
 
@@ -776,7 +772,7 @@ int gfs2_rgrp_bh_get(struct gfs2_rgrpd *rgd)
 	}
 
 	spin_lock(&sdp->sd_rindex_spin);
-	rgd->rd_free_clone = rgd->rd_rg.rg_free;
+	rgd->rd_free_clone = rgd->rd_free;
 	rgd->rd_bh_count++;
 	spin_unlock(&sdp->sd_rindex_spin);
 
@@ -850,7 +846,7 @@ void gfs2_rgrp_repolish_clones(struct gfs2_rgrpd *rgd)
 	}
 
 	spin_lock(&sdp->sd_rindex_spin);
-	rgd->rd_free_clone = rgd->rd_rg.rg_free;
+	rgd->rd_free_clone = rgd->rd_free;
 	spin_unlock(&sdp->sd_rindex_spin);
 }
 
@@ -1403,8 +1399,8 @@ u64 gfs2_alloc_block(struct gfs2_inode *ip, unsigned int *n)
 	block = rgd->rd_data0 + blk;
 	ip->i_goal = block;
 
-	gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free >= *n);
-	rgd->rd_rg.rg_free -= *n;
+	gfs2_assert_withdraw(sdp, rgd->rd_free >= *n);
+	rgd->rd_free -= *n;
 
 	gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
@@ -1445,10 +1441,10 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation)
 
 	block = rgd->rd_data0 + blk;
 
-	gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free);
-	rgd->rd_rg.rg_free--;
-	rgd->rd_rg.rg_dinodes++;
-	*generation = rgd->rd_rg.rg_igeneration++;
+	gfs2_assert_withdraw(sdp, rgd->rd_free);
+	rgd->rd_free--;
+	rgd->rd_dinodes++;
+	*generation = rgd->rd_igeneration++;
 	gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
 
@@ -1481,7 +1477,7 @@ void gfs2_free_data(struct gfs2_inode *ip, u64 bstart, u32 blen)
 	if (!rgd)
 		return;
 
-	rgd->rd_rg.rg_free += blen;
+	rgd->rd_free += blen;
 
 	gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
@@ -1509,7 +1505,7 @@ void gfs2_free_meta(struct gfs2_inode *ip, u64 bstart, u32 blen)
 	if (!rgd)
 		return;
 
-	rgd->rd_rg.rg_free += blen;
+	rgd->rd_free += blen;
 
 	gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
@@ -1546,10 +1542,10 @@ static void gfs2_free_uninit_di(struct gfs2_rgrpd *rgd, u64 blkno)
 		return;
 	gfs2_assert_withdraw(sdp, rgd == tmp_rgd);
 
-	if (!rgd->rd_rg.rg_dinodes)
+	if (!rgd->rd_dinodes)
 		gfs2_consist_rgrpd(rgd);
-	rgd->rd_rg.rg_dinodes--;
-	rgd->rd_rg.rg_free++;
+	rgd->rd_dinodes--;
+	rgd->rd_free++;
 
 	gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index c3ba3d9d0aa..141b781f2fc 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -34,76 +34,6 @@
 #include "util.h"
 
 /**
- * gfs2_jindex_hold - Grab a lock on the jindex
- * @sdp: The GFS2 superblock
- * @ji_gh: the holder for the jindex glock
- *
- * This is very similar to the gfs2_rindex_hold() function, except that
- * in general we hold the jindex lock for longer periods of time and
- * we grab it far less frequently (in general) then the rgrp lock.
- *
- * Returns: errno
- */
-
-int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh)
-{
-	struct gfs2_inode *dip = GFS2_I(sdp->sd_jindex);
-	struct qstr name;
-	char buf[20];
-	struct gfs2_jdesc *jd;
-	int error;
-
-	name.name = buf;
-
-	mutex_lock(&sdp->sd_jindex_mutex);
-
-	for (;;) {
-		error = gfs2_glock_nq_init(dip->i_gl, LM_ST_SHARED, 0, ji_gh);
-		if (error)
-			break;
-
-		name.len = sprintf(buf, "journal%u", sdp->sd_journals);
-		name.hash = gfs2_disk_hash(name.name, name.len);
-
-		error = gfs2_dir_check(sdp->sd_jindex, &name, NULL);
-		if (error == -ENOENT) {
-			error = 0;
-			break;
-		}
-
-		gfs2_glock_dq_uninit(ji_gh);
-
-		if (error)
-			break;
-
-		error = -ENOMEM;
-		jd = kzalloc(sizeof(struct gfs2_jdesc), GFP_KERNEL);
-		if (!jd)
-			break;
-
-		INIT_LIST_HEAD(&jd->extent_list);
-		jd->jd_inode = gfs2_lookupi(sdp->sd_jindex, &name, 1);
-		if (!jd->jd_inode || IS_ERR(jd->jd_inode)) {
-			if (!jd->jd_inode)
-				error = -ENOENT;
-			else
-				error = PTR_ERR(jd->jd_inode);
-			kfree(jd);
-			break;
-		}
-
-		spin_lock(&sdp->sd_jindex_spin);
-		jd->jd_jid = sdp->sd_journals++;
-		list_add_tail(&jd->jd_list, &sdp->sd_jindex_list);
-		spin_unlock(&sdp->sd_jindex_spin);
-	}
-
-	mutex_unlock(&sdp->sd_jindex_mutex);
-
-	return error;
-}
-
-/**
  * gfs2_jindex_free - Clear all the journal index information
  * @sdp: The GFS2 superblock
  *
@@ -166,39 +96,6 @@ struct gfs2_jdesc *gfs2_jdesc_find(struct gfs2_sbd *sdp, unsigned int jid)
 	return jd;
 }
 
-void gfs2_jdesc_make_dirty(struct gfs2_sbd *sdp, unsigned int jid)
-{
-	struct gfs2_jdesc *jd;
-
-	spin_lock(&sdp->sd_jindex_spin);
-	jd = jdesc_find_i(&sdp->sd_jindex_list, jid);
-	if (jd)
-		jd->jd_dirty = 1;
-	spin_unlock(&sdp->sd_jindex_spin);
-}
-
-struct gfs2_jdesc *gfs2_jdesc_find_dirty(struct gfs2_sbd *sdp)
-{
-	struct gfs2_jdesc *jd;
-	int found = 0;
-
-	spin_lock(&sdp->sd_jindex_spin);
-
-	list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
-		if (jd->jd_dirty) {
-			jd->jd_dirty = 0;
-			found = 1;
-			break;
-		}
-	}
-	spin_unlock(&sdp->sd_jindex_spin);
-
-	if (!found)
-		jd = NULL;
-
-	return jd;
-}
-
 int gfs2_jdesc_check(struct gfs2_jdesc *jd)
 {
 	struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
@@ -206,14 +103,14 @@ int gfs2_jdesc_check(struct gfs2_jdesc *jd)
 	int ar;
 	int error;
 
-	if (ip->i_di.di_size < (8 << 20) || ip->i_di.di_size > (1 << 30) ||
-	    (ip->i_di.di_size & (sdp->sd_sb.sb_bsize - 1))) {
+	if (ip->i_disksize < (8 << 20) || ip->i_disksize > (1 << 30) ||
+	    (ip->i_disksize & (sdp->sd_sb.sb_bsize - 1))) {
 		gfs2_consist_inode(ip);
 		return -EIO;
 	}
-	jd->jd_blocks = ip->i_di.di_size >> sdp->sd_sb.sb_bsize_shift;
+	jd->jd_blocks = ip->i_disksize >> sdp->sd_sb.sb_bsize_shift;
 
-	error = gfs2_write_alloc_required(ip, 0, ip->i_di.di_size, &ar);
+	error = gfs2_write_alloc_required(ip, 0, ip->i_disksize, &ar);
 	if (!error && ar) {
 		gfs2_consist_inode(ip);
 		error = -EIO;
@@ -423,137 +320,6 @@ out:
 	return error;
 }
 
-/**
- * gfs2_statfs_i - Do a statfs
- * @sdp: the filesystem
- * @sg: the sg structure
- *
- * Returns: errno
- */
-
-int gfs2_statfs_i(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc)
-{
-	struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
-	struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
-
-	spin_lock(&sdp->sd_statfs_spin);
-
-	*sc = *m_sc;
-	sc->sc_total += l_sc->sc_total;
-	sc->sc_free += l_sc->sc_free;
-	sc->sc_dinodes += l_sc->sc_dinodes;
-
-	spin_unlock(&sdp->sd_statfs_spin);
-
-	if (sc->sc_free < 0)
-		sc->sc_free = 0;
-	if (sc->sc_free > sc->sc_total)
-		sc->sc_free = sc->sc_total;
-	if (sc->sc_dinodes < 0)
-		sc->sc_dinodes = 0;
-
-	return 0;
-}
-
-/**
- * statfs_fill - fill in the sg for a given RG
- * @rgd: the RG
- * @sc: the sc structure
- *
- * Returns: 0 on success, -ESTALE if the LVB is invalid
- */
-
-static int statfs_slow_fill(struct gfs2_rgrpd *rgd,
-			    struct gfs2_statfs_change_host *sc)
-{
-	gfs2_rgrp_verify(rgd);
-	sc->sc_total += rgd->rd_data;
-	sc->sc_free += rgd->rd_rg.rg_free;
-	sc->sc_dinodes += rgd->rd_rg.rg_dinodes;
-	return 0;
-}
-
-/**
- * gfs2_statfs_slow - Stat a filesystem using asynchronous locking
- * @sdp: the filesystem
- * @sc: the sc info that will be returned
- *
- * Any error (other than a signal) will cause this routine to fall back
- * to the synchronous version.
- *
- * FIXME: This really shouldn't busy wait like this.
- *
- * Returns: errno
- */
-
-int gfs2_statfs_slow(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc)
-{
-	struct gfs2_holder ri_gh;
-	struct gfs2_rgrpd *rgd_next;
-	struct gfs2_holder *gha, *gh;
-	unsigned int slots = 64;
-	unsigned int x;
-	int done;
-	int error = 0, err;
-
-	memset(sc, 0, sizeof(struct gfs2_statfs_change_host));
-	gha = kcalloc(slots, sizeof(struct gfs2_holder), GFP_KERNEL);
-	if (!gha)
-		return -ENOMEM;
-
-	error = gfs2_rindex_hold(sdp, &ri_gh);
-	if (error)
-		goto out;
-
-	rgd_next = gfs2_rgrpd_get_first(sdp);
-
-	for (;;) {
-		done = 1;
-
-		for (x = 0; x < slots; x++) {
-			gh = gha + x;
-
-			if (gh->gh_gl && gfs2_glock_poll(gh)) {
-				err = gfs2_glock_wait(gh);
-				if (err) {
-					gfs2_holder_uninit(gh);
-					error = err;
-				} else {
-					if (!error)
-						error = statfs_slow_fill(
-							gh->gh_gl->gl_object, sc);
-					gfs2_glock_dq_uninit(gh);
-				}
-			}
-
-			if (gh->gh_gl)
-				done = 0;
-			else if (rgd_next && !error) {
-				error = gfs2_glock_nq_init(rgd_next->rd_gl,
-							   LM_ST_SHARED,
-							   GL_ASYNC,
-							   gh);
-				rgd_next = gfs2_rgrpd_get_next(rgd_next);
-				done = 0;
-			}
-
-			if (signal_pending(current))
-				error = -ERESTARTSYS;
-		}
-
-		if (done)
-			break;
-
-		yield();
-	}
-
-	gfs2_glock_dq_uninit(&ri_gh);
-
-out:
-	kfree(gha);
-	return error;
-}
-
 struct lfcc {
 	struct list_head list;
 	struct gfs2_holder gh;
@@ -580,10 +346,6 @@ static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp,
 	struct gfs2_log_header_host lh;
 	int error;
 
-	error = gfs2_jindex_hold(sdp, &ji_gh);
-	if (error)
-		return error;
-
 	list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
 		lfcc = kmalloc(sizeof(struct lfcc), GFP_KERNEL);
 		if (!lfcc) {
diff --git a/fs/gfs2/super.h b/fs/gfs2/super.h
index 50a4c9b1215..f6b8b00ad88 100644
--- a/fs/gfs2/super.h
+++ b/fs/gfs2/super.h
@@ -10,6 +10,8 @@
 #ifndef __SUPER_DOT_H__
 #define __SUPER_DOT_H__
 
+#include <linux/fs.h>
+#include <linux/dcache.h>
 #include "incore.h"
 
 void gfs2_lm_unmount(struct gfs2_sbd *sdp);
@@ -23,12 +25,9 @@ static inline unsigned int gfs2_jindex_size(struct gfs2_sbd *sdp)
 	return x;
 }
 
-int gfs2_jindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ji_gh);
 void gfs2_jindex_free(struct gfs2_sbd *sdp);
 
 struct gfs2_jdesc *gfs2_jdesc_find(struct gfs2_sbd *sdp, unsigned int jid);
-void gfs2_jdesc_make_dirty(struct gfs2_sbd *sdp, unsigned int jid);
-struct gfs2_jdesc *gfs2_jdesc_find_dirty(struct gfs2_sbd *sdp);
 int gfs2_jdesc_check(struct gfs2_jdesc *jd);
 
 int gfs2_lookup_in_master_dir(struct gfs2_sbd *sdp, char *filename,
@@ -40,11 +39,15 @@ int gfs2_statfs_init(struct gfs2_sbd *sdp);
 void gfs2_statfs_change(struct gfs2_sbd *sdp,
 			s64 total, s64 free, s64 dinodes);
 int gfs2_statfs_sync(struct gfs2_sbd *sdp);
-int gfs2_statfs_i(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc);
-int gfs2_statfs_slow(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc);
 
 int gfs2_freeze_fs(struct gfs2_sbd *sdp);
 void gfs2_unfreeze_fs(struct gfs2_sbd *sdp);
 
+extern struct file_system_type gfs2_fs_type;
+extern struct file_system_type gfs2meta_fs_type;
+extern const struct export_operations gfs2_export_ops;
+extern const struct super_operations gfs2_super_ops;
+extern struct dentry_operations gfs2_dops;
+
 #endif /* __SUPER_DOT_H__ */
 
diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c
index 7e1879f1a02..26c1fa777a9 100644
--- a/fs/gfs2/sys.c
+++ b/fs/gfs2/sys.c
@@ -26,9 +26,6 @@
 #include "quota.h"
 #include "util.h"
 
-char *gfs2_sys_margs;
-spinlock_t gfs2_sys_margs_lock;
-
 static ssize_t id_show(struct gfs2_sbd *sdp, char *buf)
 {
 	return snprintf(buf, PAGE_SIZE, "%u:%u\n",
@@ -263,7 +260,6 @@ ARGS_ATTR(localcaching,    "%d\n");
 ARGS_ATTR(localflocks,     "%d\n");
 ARGS_ATTR(debug,           "%d\n");
 ARGS_ATTR(upgrade,         "%d\n");
-ARGS_ATTR(num_glockd,      "%u\n");
 ARGS_ATTR(posix_acl,       "%d\n");
 ARGS_ATTR(quota,           "%u\n");
 ARGS_ATTR(suiddir,         "%d\n");
@@ -279,7 +275,6 @@ static struct attribute *args_attrs[] = {
 	&args_attr_localflocks.attr,
 	&args_attr_debug.attr,
 	&args_attr_upgrade.attr,
-	&args_attr_num_glockd.attr,
 	&args_attr_posix_acl.attr,
 	&args_attr_quota.attr,
 	&args_attr_suiddir.attr,
@@ -288,30 +283,6 @@ static struct attribute *args_attrs[] = {
 };
 
 /*
- * display counters from superblock
- */
-
-struct counters_attr {
-	struct attribute attr;
-	ssize_t (*show)(struct gfs2_sbd *, char *);
-};
-
-#define COUNTERS_ATTR(name, fmt)                                            \
-static ssize_t name##_show(struct gfs2_sbd *sdp, char *buf)                 \
-{                                                                           \
-	return snprintf(buf, PAGE_SIZE, fmt,                                \
-			(unsigned int)atomic_read(&sdp->sd_##name));        \
-}                                                                           \
-static struct counters_attr counters_attr_##name = __ATTR_RO(name)
-
-COUNTERS_ATTR(reclaimed,        "%u\n");
-
-static struct attribute *counters_attrs[] = {
-	&counters_attr_reclaimed.attr,
-	NULL,
-};
-
-/*
  * get and set struct gfs2_tune fields
  */
 
@@ -393,7 +364,6 @@ static ssize_t name##_store(struct gfs2_sbd *sdp, const char *buf, size_t len)\
 }                                                                             \
 TUNE_ATTR_2(name, name##_store)
 
-TUNE_ATTR(demote_secs, 0);
 TUNE_ATTR(incore_log_blocks, 0);
 TUNE_ATTR(log_flush_secs, 0);
 TUNE_ATTR(quota_warn_period, 0);
@@ -408,11 +378,9 @@ TUNE_ATTR(stall_secs, 1);
 TUNE_ATTR(statfs_quantum, 1);
 TUNE_ATTR_DAEMON(recoverd_secs, recoverd_process);
 TUNE_ATTR_DAEMON(logd_secs, logd_process);
-TUNE_ATTR_DAEMON(quotad_secs, quotad_process);
 TUNE_ATTR_3(quota_scale, quota_scale_show, quota_scale_store);
 
 static struct attribute *tune_attrs[] = {
-	&tune_attr_demote_secs.attr,
 	&tune_attr_incore_log_blocks.attr,
 	&tune_attr_log_flush_secs.attr,
 	&tune_attr_quota_warn_period.attr,
@@ -426,7 +394,6 @@ static struct attribute *tune_attrs[] = {
 	&tune_attr_statfs_quantum.attr,
 	&tune_attr_recoverd_secs.attr,
 	&tune_attr_logd_secs.attr,
-	&tune_attr_quotad_secs.attr,
 	&tune_attr_quota_scale.attr,
 	&tune_attr_new_files_jdata.attr,
 	NULL,
@@ -437,11 +404,6 @@ static struct attribute_group lockstruct_group = {
 	.attrs = lockstruct_attrs,
 };
 
-static struct attribute_group counters_group = {
-	.name = "counters",
-	.attrs = counters_attrs,
-};
-
 static struct attribute_group args_group = {
 	.name = "args",
 	.attrs = args_attrs,
@@ -466,13 +428,9 @@ int gfs2_sys_fs_add(struct gfs2_sbd *sdp)
 	if (error)
 		goto fail_reg;
 
-	error = sysfs_create_group(&sdp->sd_kobj, &counters_group);
-	if (error)
-		goto fail_lockstruct;
-
 	error = sysfs_create_group(&sdp->sd_kobj, &args_group);
 	if (error)
-		goto fail_counters;
+		goto fail_lockstruct;
 
 	error = sysfs_create_group(&sdp->sd_kobj, &tune_group);
 	if (error)
@@ -483,8 +441,6 @@ int gfs2_sys_fs_add(struct gfs2_sbd *sdp)
 
 fail_args:
 	sysfs_remove_group(&sdp->sd_kobj, &args_group);
-fail_counters:
-	sysfs_remove_group(&sdp->sd_kobj, &counters_group);
 fail_lockstruct:
 	sysfs_remove_group(&sdp->sd_kobj, &lockstruct_group);
 fail_reg:
@@ -498,16 +454,27 @@ void gfs2_sys_fs_del(struct gfs2_sbd *sdp)
 {
 	sysfs_remove_group(&sdp->sd_kobj, &tune_group);
 	sysfs_remove_group(&sdp->sd_kobj, &args_group);
-	sysfs_remove_group(&sdp->sd_kobj, &counters_group);
 	sysfs_remove_group(&sdp->sd_kobj, &lockstruct_group);
 	kobject_put(&sdp->sd_kobj);
 }
 
+static int gfs2_uevent(struct kset *kset, struct kobject *kobj,
+		       struct kobj_uevent_env *env)
+{
+	struct gfs2_sbd *sdp = container_of(kobj, struct gfs2_sbd, sd_kobj);
+	add_uevent_var(env, "LOCKTABLE=%s", sdp->sd_table_name);
+	add_uevent_var(env, "LOCKPROTO=%s", sdp->sd_proto_name);
+	return 0;
+}
+
+static struct kset_uevent_ops gfs2_uevent_ops = {
+	.uevent = gfs2_uevent,
+};
+
+
 int gfs2_sys_init(void)
 {
-	gfs2_sys_margs = NULL;
-	spin_lock_init(&gfs2_sys_margs_lock);
-	gfs2_kset = kset_create_and_add("gfs2", NULL, fs_kobj);
+	gfs2_kset = kset_create_and_add("gfs2", &gfs2_uevent_ops, fs_kobj);
 	if (!gfs2_kset)
 		return -ENOMEM;
 	return 0;
@@ -515,7 +482,6 @@ int gfs2_sys_init(void)
 
 void gfs2_sys_uninit(void)
 {
-	kfree(gfs2_sys_margs);
 	kset_unregister(gfs2_kset);
 }
 
diff --git a/fs/gfs2/sys.h b/fs/gfs2/sys.h
index 1ca8cdac530..e94560e836d 100644
--- a/fs/gfs2/sys.h
+++ b/fs/gfs2/sys.h
@@ -13,10 +13,6 @@
 #include <linux/spinlock.h>
 struct gfs2_sbd;
 
-/* Allow args to be passed to GFS2 when using an initial ram disk */
-extern char *gfs2_sys_margs;
-extern spinlock_t gfs2_sys_margs_lock;
-
 int gfs2_sys_fs_add(struct gfs2_sbd *sdp);
 void gfs2_sys_fs_del(struct gfs2_sbd *sdp);
 
diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c
index d31e355c61f..374f50e9549 100644
--- a/fs/gfs2/util.c
+++ b/fs/gfs2/util.c
@@ -25,6 +25,7 @@ struct kmem_cache *gfs2_glock_cachep __read_mostly;
 struct kmem_cache *gfs2_inode_cachep __read_mostly;
 struct kmem_cache *gfs2_bufdata_cachep __read_mostly;
 struct kmem_cache *gfs2_rgrpd_cachep __read_mostly;
+struct kmem_cache *gfs2_quotad_cachep __read_mostly;
 
 void gfs2_assert_i(struct gfs2_sbd *sdp)
 {
diff --git a/fs/gfs2/util.h b/fs/gfs2/util.h
index 7f48576289c..33e96b0ce9a 100644
--- a/fs/gfs2/util.h
+++ b/fs/gfs2/util.h
@@ -148,6 +148,7 @@ extern struct kmem_cache *gfs2_glock_cachep;
 extern struct kmem_cache *gfs2_inode_cachep;
 extern struct kmem_cache *gfs2_bufdata_cachep;
 extern struct kmem_cache *gfs2_rgrpd_cachep;
+extern struct kmem_cache *gfs2_quotad_cachep;
 
 static inline unsigned int gfs2_tune_get_i(struct gfs2_tune *gt,
 					   unsigned int *p)
diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c
index 3a31451ac17..5c538e0ec14 100644
--- a/fs/hostfs/hostfs_kern.c
+++ b/fs/hostfs/hostfs_kern.c
@@ -501,7 +501,7 @@ int hostfs_write_begin(struct file *file, struct address_space *mapping,
 {
 	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
 
-	*pagep = __grab_cache_page(mapping, index);
+	*pagep = grab_cache_page_write_begin(mapping, index, flags);
 	if (!*pagep)
 		return -ENOMEM;
 	return 0;
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index 7d479ce3ace..6903d37af03 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -252,6 +252,7 @@ static ssize_t hugetlbfs_read(struct file *filp, char __user *buf,
 	for (;;) {
 		struct page *page;
 		unsigned long nr, ret;
+		int ra;
 
 		/* nr is the maximum number of bytes to copy from this page */
 		nr = huge_page_size(h);
@@ -274,16 +275,19 @@ static ssize_t hugetlbfs_read(struct file *filp, char __user *buf,
 			 */
 			ret = len < nr ? len : nr;
 			if (clear_user(buf, ret))
-				ret = -EFAULT;
+				ra = -EFAULT;
+			else
+				ra = 0;
 		} else {
 			/*
 			 * We have the page, copy it to user space buffer.
 			 */
-			ret = hugetlbfs_read_actor(page, offset, buf, len, nr);
+			ra = hugetlbfs_read_actor(page, offset, buf, len, nr);
+			ret = ra;
 		}
-		if (ret < 0) {
+		if (ra < 0) {
 			if (retval == 0)
-				retval = ret;
+				retval = ra;
 			if (page)
 				page_cache_release(page);
 			goto out;
@@ -506,7 +510,6 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb, uid_t uid,
 		inode->i_mode = mode;
 		inode->i_uid = uid;
 		inode->i_gid = gid;
-		inode->i_blocks = 0;
 		inode->i_mapping->a_ops = &hugetlbfs_aops;
 		inode->i_mapping->backing_dev_info =&hugetlbfs_backing_dev_info;
 		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
diff --git a/fs/inode.c b/fs/inode.c
index 0487ddba139..913ab2d9a5d 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -22,6 +22,7 @@
 #include <linux/bootmem.h>
 #include <linux/inotify.h>
 #include <linux/mount.h>
+#include <linux/async.h>
 
 /*
  * This is needed for the following functions:
@@ -108,84 +109,102 @@ static void wake_up_inode(struct inode *inode)
 	wake_up_bit(&inode->i_state, __I_LOCK);
 }
 
-static struct inode *alloc_inode(struct super_block *sb)
+/**
+ * inode_init_always - perform inode structure intialisation
+ * @sb: superblock inode belongs to
+ * @inode: inode to initialise
+ *
+ * These are initializations that need to be done on every inode
+ * allocation as the fields are not initialised by slab allocation.
+ */
+struct inode *inode_init_always(struct super_block *sb, struct inode *inode)
 {
 	static const struct address_space_operations empty_aops;
 	static struct inode_operations empty_iops;
 	static const struct file_operations empty_fops;
-	struct inode *inode;
 
-	if (sb->s_op->alloc_inode)
-		inode = sb->s_op->alloc_inode(sb);
-	else
-		inode = (struct inode *) kmem_cache_alloc(inode_cachep, GFP_KERNEL);
-
-	if (inode) {
-		struct address_space * const mapping = &inode->i_data;
-
-		inode->i_sb = sb;
-		inode->i_blkbits = sb->s_blocksize_bits;
-		inode->i_flags = 0;
-		atomic_set(&inode->i_count, 1);
-		inode->i_op = &empty_iops;
-		inode->i_fop = &empty_fops;
-		inode->i_nlink = 1;
-		atomic_set(&inode->i_writecount, 0);
-		inode->i_size = 0;
-		inode->i_blocks = 0;
-		inode->i_bytes = 0;
-		inode->i_generation = 0;
+	struct address_space * const mapping = &inode->i_data;
+
+	inode->i_sb = sb;
+	inode->i_blkbits = sb->s_blocksize_bits;
+	inode->i_flags = 0;
+	atomic_set(&inode->i_count, 1);
+	inode->i_op = &empty_iops;
+	inode->i_fop = &empty_fops;
+	inode->i_nlink = 1;
+	inode->i_uid = 0;
+	inode->i_gid = 0;
+	atomic_set(&inode->i_writecount, 0);
+	inode->i_size = 0;
+	inode->i_blocks = 0;
+	inode->i_bytes = 0;
+	inode->i_generation = 0;
 #ifdef CONFIG_QUOTA
-		memset(&inode->i_dquot, 0, sizeof(inode->i_dquot));
+	memset(&inode->i_dquot, 0, sizeof(inode->i_dquot));
 #endif
-		inode->i_pipe = NULL;
-		inode->i_bdev = NULL;
-		inode->i_cdev = NULL;
-		inode->i_rdev = 0;
-		inode->dirtied_when = 0;
-		if (security_inode_alloc(inode)) {
-			if (inode->i_sb->s_op->destroy_inode)
-				inode->i_sb->s_op->destroy_inode(inode);
-			else
-				kmem_cache_free(inode_cachep, (inode));
-			return NULL;
-		}
+	inode->i_pipe = NULL;
+	inode->i_bdev = NULL;
+	inode->i_cdev = NULL;
+	inode->i_rdev = 0;
+	inode->dirtied_when = 0;
+	if (security_inode_alloc(inode)) {
+		if (inode->i_sb->s_op->destroy_inode)
+			inode->i_sb->s_op->destroy_inode(inode);
+		else
+			kmem_cache_free(inode_cachep, (inode));
+		return NULL;
+	}
 
-		spin_lock_init(&inode->i_lock);
-		lockdep_set_class(&inode->i_lock, &sb->s_type->i_lock_key);
+	spin_lock_init(&inode->i_lock);
+	lockdep_set_class(&inode->i_lock, &sb->s_type->i_lock_key);
 
-		mutex_init(&inode->i_mutex);
-		lockdep_set_class(&inode->i_mutex, &sb->s_type->i_mutex_key);
+	mutex_init(&inode->i_mutex);
+	lockdep_set_class(&inode->i_mutex, &sb->s_type->i_mutex_key);
 
-		init_rwsem(&inode->i_alloc_sem);
-		lockdep_set_class(&inode->i_alloc_sem, &sb->s_type->i_alloc_sem_key);
+	init_rwsem(&inode->i_alloc_sem);
+	lockdep_set_class(&inode->i_alloc_sem, &sb->s_type->i_alloc_sem_key);
 
-		mapping->a_ops = &empty_aops;
- 		mapping->host = inode;
-		mapping->flags = 0;
-		mapping_set_gfp_mask(mapping, GFP_HIGHUSER_PAGECACHE);
-		mapping->assoc_mapping = NULL;
-		mapping->backing_dev_info = &default_backing_dev_info;
-		mapping->writeback_index = 0;
+	mapping->a_ops = &empty_aops;
+	mapping->host = inode;
+	mapping->flags = 0;
+	mapping_set_gfp_mask(mapping, GFP_HIGHUSER_MOVABLE);
+	mapping->assoc_mapping = NULL;
+	mapping->backing_dev_info = &default_backing_dev_info;
+	mapping->writeback_index = 0;
 
-		/*
-		 * If the block_device provides a backing_dev_info for client
-		 * inodes then use that.  Otherwise the inode share the bdev's
-		 * backing_dev_info.
-		 */
-		if (sb->s_bdev) {
-			struct backing_dev_info *bdi;
+	/*
+	 * If the block_device provides a backing_dev_info for client
+	 * inodes then use that.  Otherwise the inode share the bdev's
+	 * backing_dev_info.
+	 */
+	if (sb->s_bdev) {
+		struct backing_dev_info *bdi;
 
-			bdi = sb->s_bdev->bd_inode_backing_dev_info;
-			if (!bdi)
-				bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
-			mapping->backing_dev_info = bdi;
-		}
-		inode->i_private = NULL;
-		inode->i_mapping = mapping;
+		bdi = sb->s_bdev->bd_inode_backing_dev_info;
+		if (!bdi)
+			bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
+		mapping->backing_dev_info = bdi;
 	}
+	inode->i_private = NULL;
+	inode->i_mapping = mapping;
+
 	return inode;
 }
+EXPORT_SYMBOL(inode_init_always);
+
+static struct inode *alloc_inode(struct super_block *sb)
+{
+	struct inode *inode;
+
+	if (sb->s_op->alloc_inode)
+		inode = sb->s_op->alloc_inode(sb);
+	else
+		inode = kmem_cache_alloc(inode_cachep, GFP_KERNEL);
+
+	if (inode)
+		return inode_init_always(sb, inode);
+	return NULL;
+}
 
 void destroy_inode(struct inode *inode) 
 {
@@ -196,6 +215,7 @@ void destroy_inode(struct inode *inode)
 	else
 		kmem_cache_free(inode_cachep, (inode));
 }
+EXPORT_SYMBOL(destroy_inode);
 
 
 /*
@@ -534,12 +554,55 @@ repeat:
 	return node ? inode : NULL;
 }
 
+static unsigned long hash(struct super_block *sb, unsigned long hashval)
+{
+	unsigned long tmp;
+
+	tmp = (hashval * (unsigned long)sb) ^ (GOLDEN_RATIO_PRIME + hashval) /
+			L1_CACHE_BYTES;
+	tmp = tmp ^ ((tmp ^ GOLDEN_RATIO_PRIME) >> I_HASHBITS);
+	return tmp & I_HASHMASK;
+}
+
+static inline void
+__inode_add_to_lists(struct super_block *sb, struct hlist_head *head,
+			struct inode *inode)
+{
+	inodes_stat.nr_inodes++;
+	list_add(&inode->i_list, &inode_in_use);
+	list_add(&inode->i_sb_list, &sb->s_inodes);
+	if (head)
+		hlist_add_head(&inode->i_hash, head);
+}
+
+/**
+ * inode_add_to_lists - add a new inode to relevant lists
+ * @sb: superblock inode belongs to
+ * @inode: inode to mark in use
+ *
+ * When an inode is allocated it needs to be accounted for, added to the in use
+ * list, the owning superblock and the inode hash. This needs to be done under
+ * the inode_lock, so export a function to do this rather than the inode lock
+ * itself. We calculate the hash list to add to here so it is all internal
+ * which requires the caller to have already set up the inode number in the
+ * inode to add.
+ */
+void inode_add_to_lists(struct super_block *sb, struct inode *inode)
+{
+	struct hlist_head *head = inode_hashtable + hash(sb, inode->i_ino);
+
+	spin_lock(&inode_lock);
+	__inode_add_to_lists(sb, head, inode);
+	spin_unlock(&inode_lock);
+}
+EXPORT_SYMBOL_GPL(inode_add_to_lists);
+
 /**
  *	new_inode 	- obtain an inode
  *	@sb: superblock
  *
  *	Allocates a new inode for given superblock. The default gfp_mask
- *	for allocations related to inode->i_mapping is GFP_HIGHUSER_PAGECACHE.
+ *	for allocations related to inode->i_mapping is GFP_HIGHUSER_MOVABLE.
  *	If HIGHMEM pages are unsuitable or it is known that pages allocated
  *	for the page cache are not reclaimable or migratable,
  *	mapping_set_gfp_mask() must be called with suitable flags on the
@@ -561,9 +624,7 @@ struct inode *new_inode(struct super_block *sb)
 	inode = alloc_inode(sb);
 	if (inode) {
 		spin_lock(&inode_lock);
-		inodes_stat.nr_inodes++;
-		list_add(&inode->i_list, &inode_in_use);
-		list_add(&inode->i_sb_list, &sb->s_inodes);
+		__inode_add_to_lists(sb, NULL, inode);
 		inode->i_ino = ++last_ino;
 		inode->i_state = 0;
 		spin_unlock(&inode_lock);
@@ -622,10 +683,7 @@ static struct inode * get_new_inode(struct super_block *sb, struct hlist_head *h
 			if (set(inode, data))
 				goto set_failed;
 
-			inodes_stat.nr_inodes++;
-			list_add(&inode->i_list, &inode_in_use);
-			list_add(&inode->i_sb_list, &sb->s_inodes);
-			hlist_add_head(&inode->i_hash, head);
+			__inode_add_to_lists(sb, head, inode);
 			inode->i_state = I_LOCK|I_NEW;
 			spin_unlock(&inode_lock);
 
@@ -671,10 +729,7 @@ static struct inode * get_new_inode_fast(struct super_block *sb, struct hlist_he
 		old = find_inode_fast(sb, head, ino);
 		if (!old) {
 			inode->i_ino = ino;
-			inodes_stat.nr_inodes++;
-			list_add(&inode->i_list, &inode_in_use);
-			list_add(&inode->i_sb_list, &sb->s_inodes);
-			hlist_add_head(&inode->i_hash, head);
+			__inode_add_to_lists(sb, head, inode);
 			inode->i_state = I_LOCK|I_NEW;
 			spin_unlock(&inode_lock);
 
@@ -698,16 +753,6 @@ static struct inode * get_new_inode_fast(struct super_block *sb, struct hlist_he
 	return inode;
 }
 
-static unsigned long hash(struct super_block *sb, unsigned long hashval)
-{
-	unsigned long tmp;
-
-	tmp = (hashval * (unsigned long)sb) ^ (GOLDEN_RATIO_PRIME + hashval) /
-			L1_CACHE_BYTES;
-	tmp = tmp ^ ((tmp ^ GOLDEN_RATIO_PRIME) >> I_HASHBITS);
-	return tmp & I_HASHMASK;
-}
-
 /**
  *	iunique - get a unique inode number
  *	@sb: superblock
@@ -990,6 +1035,65 @@ struct inode *iget_locked(struct super_block *sb, unsigned long ino)
 
 EXPORT_SYMBOL(iget_locked);
 
+int insert_inode_locked(struct inode *inode)
+{
+	struct super_block *sb = inode->i_sb;
+	ino_t ino = inode->i_ino;
+	struct hlist_head *head = inode_hashtable + hash(sb, ino);
+	struct inode *old;
+
+	inode->i_state |= I_LOCK|I_NEW;
+	while (1) {
+		spin_lock(&inode_lock);
+		old = find_inode_fast(sb, head, ino);
+		if (likely(!old)) {
+			hlist_add_head(&inode->i_hash, head);
+			spin_unlock(&inode_lock);
+			return 0;
+		}
+		__iget(old);
+		spin_unlock(&inode_lock);
+		wait_on_inode(old);
+		if (unlikely(!hlist_unhashed(&old->i_hash))) {
+			iput(old);
+			return -EBUSY;
+		}
+		iput(old);
+	}
+}
+
+EXPORT_SYMBOL(insert_inode_locked);
+
+int insert_inode_locked4(struct inode *inode, unsigned long hashval,
+		int (*test)(struct inode *, void *), void *data)
+{
+	struct super_block *sb = inode->i_sb;
+	struct hlist_head *head = inode_hashtable + hash(sb, hashval);
+	struct inode *old;
+
+	inode->i_state |= I_LOCK|I_NEW;
+
+	while (1) {
+		spin_lock(&inode_lock);
+		old = find_inode(sb, head, test, data);
+		if (likely(!old)) {
+			hlist_add_head(&inode->i_hash, head);
+			spin_unlock(&inode_lock);
+			return 0;
+		}
+		__iget(old);
+		spin_unlock(&inode_lock);
+		wait_on_inode(old);
+		if (unlikely(!hlist_unhashed(&old->i_hash))) {
+			iput(old);
+			return -EBUSY;
+		}
+		iput(old);
+	}
+}
+
+EXPORT_SYMBOL(insert_inode_locked4);
+
 /**
  *	__insert_inode_hash - hash an inode
  *	@inode: unhashed inode
@@ -1292,6 +1396,7 @@ int inode_wait(void *word)
 	schedule();
 	return 0;
 }
+EXPORT_SYMBOL(inode_wait);
 
 /*
  * If we try to find an inode in the inode hash while it is being
diff --git a/fs/ioctl.c b/fs/ioctl.c
index 43e8b2c0664..20b0a8a24c6 100644
--- a/fs/ioctl.c
+++ b/fs/ioctl.c
@@ -231,7 +231,8 @@ static int ioctl_fiemap(struct file *filp, unsigned long arg)
 #define blk_to_logical(inode, blk) (blk << (inode)->i_blkbits)
 #define logical_to_blk(inode, offset) (offset >> (inode)->i_blkbits);
 
-/*
+/**
+ * __generic_block_fiemap - FIEMAP for block based inodes (no locking)
  * @inode - the inode to map
  * @arg - the pointer to userspace where we copy everything to
  * @get_block - the fs's get_block function
@@ -242,11 +243,15 @@ static int ioctl_fiemap(struct file *filp, unsigned long arg)
  *
  * If it is possible to have data blocks beyond a hole past @inode->i_size, then
  * please do not use this function, it will stop at the first unmapped block
- * beyond i_size
+ * beyond i_size.
+ *
+ * If you use this function directly, you need to do your own locking. Use
+ * generic_block_fiemap if you want the locking done for you.
  */
-int generic_block_fiemap(struct inode *inode,
-			 struct fiemap_extent_info *fieinfo, u64 start,
-			 u64 len, get_block_t *get_block)
+
+int __generic_block_fiemap(struct inode *inode,
+			   struct fiemap_extent_info *fieinfo, u64 start,
+			   u64 len, get_block_t *get_block)
 {
 	struct buffer_head tmp;
 	unsigned int start_blk;
@@ -260,9 +265,6 @@ int generic_block_fiemap(struct inode *inode,
 
 	start_blk = logical_to_blk(inode, start);
 
-	/* guard against change */
-	mutex_lock(&inode->i_mutex);
-
 	length = (long long)min_t(u64, len, i_size_read(inode));
 	map_len = length;
 
@@ -334,14 +336,36 @@ int generic_block_fiemap(struct inode *inode,
 		cond_resched();
 	} while (1);
 
-	mutex_unlock(&inode->i_mutex);
-
 	/* if ret is 1 then we just hit the end of the extent array */
 	if (ret == 1)
 		ret = 0;
 
 	return ret;
 }
+EXPORT_SYMBOL(__generic_block_fiemap);
+
+/**
+ * generic_block_fiemap - FIEMAP for block based inodes
+ * @inode: The inode to map
+ * @fieinfo: The mapping information
+ * @start: The initial block to map
+ * @len: The length of the extect to attempt to map
+ * @get_block: The block mapping function for the fs
+ *
+ * Calls __generic_block_fiemap to map the inode, after taking
+ * the inode's mutex lock.
+ */
+
+int generic_block_fiemap(struct inode *inode,
+			 struct fiemap_extent_info *fieinfo, u64 start,
+			 u64 len, get_block_t *get_block)
+{
+	int ret;
+	mutex_lock(&inode->i_mutex);
+	ret = __generic_block_fiemap(inode, fieinfo, start, len, get_block);
+	mutex_unlock(&inode->i_mutex);
+	return ret;
+}
 EXPORT_SYMBOL(generic_block_fiemap);
 
 #endif  /*  CONFIG_BLOCK  */
@@ -415,6 +439,43 @@ static int ioctl_fioasync(unsigned int fd, struct file *filp,
 	return error;
 }
 
+static int ioctl_fsfreeze(struct file *filp)
+{
+	struct super_block *sb = filp->f_path.dentry->d_inode->i_sb;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	/* If filesystem doesn't support freeze feature, return. */
+	if (sb->s_op->freeze_fs == NULL)
+		return -EOPNOTSUPP;
+
+	/* If a blockdevice-backed filesystem isn't specified, return. */
+	if (sb->s_bdev == NULL)
+		return -EINVAL;
+
+	/* Freeze */
+	sb = freeze_bdev(sb->s_bdev);
+	if (IS_ERR(sb))
+		return PTR_ERR(sb);
+	return 0;
+}
+
+static int ioctl_fsthaw(struct file *filp)
+{
+	struct super_block *sb = filp->f_path.dentry->d_inode->i_sb;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	/* If a blockdevice-backed filesystem isn't specified, return EINVAL. */
+	if (sb->s_bdev == NULL)
+		return -EINVAL;
+
+	/* Thaw */
+	return thaw_bdev(sb->s_bdev, sb);
+}
+
 /*
  * When you add any new common ioctls to the switches above and below
  * please update compat_sys_ioctl() too.
@@ -462,6 +523,15 @@ int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
 		} else
 			error = -ENOTTY;
 		break;
+
+	case FIFREEZE:
+		error = ioctl_fsfreeze(filp);
+		break;
+
+	case FITHAW:
+		error = ioctl_fsthaw(filp);
+		break;
+
 	default:
 		if (S_ISREG(filp->f_path.dentry->d_inode->i_mode))
 			error = file_ioctl(filp, cmd, arg);
diff --git a/fs/ioprio.c b/fs/ioprio.c
index 3569e0ad86a..1a39ac37094 100644
--- a/fs/ioprio.c
+++ b/fs/ioprio.c
@@ -27,7 +27,7 @@
 #include <linux/security.h>
 #include <linux/pid_namespace.h>
 
-static int set_task_ioprio(struct task_struct *task, int ioprio)
+int set_task_ioprio(struct task_struct *task, int ioprio)
 {
 	int err;
 	struct io_context *ioc;
@@ -70,6 +70,7 @@ static int set_task_ioprio(struct task_struct *task, int ioprio)
 	task_unlock(task);
 	return err;
 }
+EXPORT_SYMBOL_GPL(set_task_ioprio);
 
 asmlinkage long sys_ioprio_set(int which, int who, int ioprio)
 {
diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c
index 3f8af0f1505..6147ec3643a 100644
--- a/fs/isofs/inode.c
+++ b/fs/isofs/inode.c
@@ -855,10 +855,6 @@ root_found:
 	}
 	sbi->s_joliet_level = joliet_level;
 
-	/* check the root inode */
-	if (!inode->i_op)
-		goto out_bad_root;
-
 	/* Make sure the root inode is a directory */
 	if (!S_ISDIR(inode->i_mode)) {
 		printk(KERN_WARNING
@@ -886,8 +882,6 @@ root_found:
 	/*
 	 * Display error messages and free resources.
 	 */
-out_bad_root:
-	printk(KERN_WARNING "%s: root inode not initialized\n", __func__);
 out_iput:
 	iput(inode);
 	goto out_no_inode;
diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c
index 25719d902c5..3fbffb1ea71 100644
--- a/fs/jbd/commit.c
+++ b/fs/jbd/commit.c
@@ -306,6 +306,8 @@ void journal_commit_transaction(journal_t *journal)
 	int flags;
 	int err;
 	unsigned long blocknr;
+	ktime_t start_time;
+	u64 commit_time;
 	char *tagp = NULL;
 	journal_header_t *header;
 	journal_block_tag_t *tag = NULL;
@@ -418,6 +420,7 @@ void journal_commit_transaction(journal_t *journal)
 	commit_transaction->t_state = T_FLUSH;
 	journal->j_committing_transaction = commit_transaction;
 	journal->j_running_transaction = NULL;
+	start_time = ktime_get();
 	commit_transaction->t_log_start = journal->j_head;
 	wake_up(&journal->j_wait_transaction_locked);
 	spin_unlock(&journal->j_state_lock);
@@ -913,6 +916,18 @@ restart_loop:
 	J_ASSERT(commit_transaction == journal->j_committing_transaction);
 	journal->j_commit_sequence = commit_transaction->t_tid;
 	journal->j_committing_transaction = NULL;
+	commit_time = ktime_to_ns(ktime_sub(ktime_get(), start_time));
+
+	/*
+	 * weight the commit time higher than the average time so we don't
+	 * react too strongly to vast changes in commit time
+	 */
+	if (likely(journal->j_average_commit_time))
+		journal->j_average_commit_time = (commit_time*3 +
+				journal->j_average_commit_time) / 4;
+	else
+		journal->j_average_commit_time = commit_time;
+
 	spin_unlock(&journal->j_state_lock);
 
 	if (commit_transaction->t_checkpoint_list == NULL &&
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index 60d4c32c880..e6a11743127 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -25,6 +25,7 @@
 #include <linux/timer.h>
 #include <linux/mm.h>
 #include <linux/highmem.h>
+#include <linux/hrtimer.h>
 
 static void __journal_temp_unlink_buffer(struct journal_head *jh);
 
@@ -49,6 +50,7 @@ get_transaction(journal_t *journal, transaction_t *transaction)
 {
 	transaction->t_journal = journal;
 	transaction->t_state = T_RUNNING;
+	transaction->t_start_time = ktime_get();
 	transaction->t_tid = journal->j_transaction_sequence++;
 	transaction->t_expires = jiffies + journal->j_commit_interval;
 	spin_lock_init(&transaction->t_handle_lock);
@@ -752,7 +754,6 @@ out:
  * int journal_get_write_access() - notify intent to modify a buffer for metadata (not data) update.
  * @handle: transaction to add buffer modifications to
  * @bh:     bh to be used for metadata writes
- * @credits: variable that will receive credits for the buffer
  *
  * Returns an error code or 0 on success.
  *
@@ -1370,7 +1371,7 @@ int journal_stop(handle_t *handle)
 {
 	transaction_t *transaction = handle->h_transaction;
 	journal_t *journal = transaction->t_journal;
-	int old_handle_count, err;
+	int err;
 	pid_t pid;
 
 	J_ASSERT(journal_current_handle() == handle);
@@ -1399,6 +1400,17 @@ int journal_stop(handle_t *handle)
 	 * on IO anyway.  Speeds up many-threaded, many-dir operations
 	 * by 30x or more...
 	 *
+	 * We try and optimize the sleep time against what the underlying disk
+	 * can do, instead of having a static sleep time.  This is usefull for
+	 * the case where our storage is so fast that it is more optimal to go
+	 * ahead and force a flush and wait for the transaction to be committed
+	 * than it is to wait for an arbitrary amount of time for new writers to
+	 * join the transaction.  We acheive this by measuring how long it takes
+	 * to commit a transaction, and compare it with how long this
+	 * transaction has been running, and if run time < commit time then we
+	 * sleep for the delta and commit.  This greatly helps super fast disks
+	 * that would see slowdowns as more threads started doing fsyncs.
+	 *
 	 * But don't do this if this process was the most recent one to
 	 * perform a synchronous write.  We do this to detect the case where a
 	 * single process is doing a stream of sync writes.  No point in waiting
@@ -1406,11 +1418,26 @@ int journal_stop(handle_t *handle)
 	 */
 	pid = current->pid;
 	if (handle->h_sync && journal->j_last_sync_writer != pid) {
+		u64 commit_time, trans_time;
+
 		journal->j_last_sync_writer = pid;
-		do {
-			old_handle_count = transaction->t_handle_count;
-			schedule_timeout_uninterruptible(1);
-		} while (old_handle_count != transaction->t_handle_count);
+
+		spin_lock(&journal->j_state_lock);
+		commit_time = journal->j_average_commit_time;
+		spin_unlock(&journal->j_state_lock);
+
+		trans_time = ktime_to_ns(ktime_sub(ktime_get(),
+						   transaction->t_start_time));
+
+		commit_time = min_t(u64, commit_time,
+				    1000*jiffies_to_usecs(1));
+
+		if (trans_time < commit_time) {
+			ktime_t expires = ktime_add_ns(ktime_get(),
+						       commit_time);
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			schedule_hrtimeout(&expires, HRTIMER_MODE_ABS);
+		}
 	}
 
 	current->journal_info = NULL;
diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index 9497718fe92..17159cacbd9 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -249,16 +249,14 @@ restart:
 	return ret;
 }
 
-#define NR_BATCH	64
-
 static void
-__flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count)
+__flush_batch(journal_t *journal, int *batch_count)
 {
 	int i;
 
-	ll_rw_block(SWRITE, *batch_count, bhs);
+	ll_rw_block(SWRITE, *batch_count, journal->j_chkpt_bhs);
 	for (i = 0; i < *batch_count; i++) {
-		struct buffer_head *bh = bhs[i];
+		struct buffer_head *bh = journal->j_chkpt_bhs[i];
 		clear_buffer_jwrite(bh);
 		BUFFER_TRACE(bh, "brelse");
 		__brelse(bh);
@@ -277,8 +275,7 @@ __flush_batch(journal_t *journal, struct buffer_head **bhs, int *batch_count)
  * Called under jbd_lock_bh_state(jh2bh(jh)), and drops it
  */
 static int __process_buffer(journal_t *journal, struct journal_head *jh,
-			struct buffer_head **bhs, int *batch_count,
-			transaction_t *transaction)
+			    int *batch_count, transaction_t *transaction)
 {
 	struct buffer_head *bh = jh2bh(jh);
 	int ret = 0;
@@ -325,14 +322,14 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh,
 		get_bh(bh);
 		J_ASSERT_BH(bh, !buffer_jwrite(bh));
 		set_buffer_jwrite(bh);
-		bhs[*batch_count] = bh;
+		journal->j_chkpt_bhs[*batch_count] = bh;
 		__buffer_relink_io(jh);
 		jbd_unlock_bh_state(bh);
 		transaction->t_chp_stats.cs_written++;
 		(*batch_count)++;
-		if (*batch_count == NR_BATCH) {
+		if (*batch_count == JBD2_NR_BATCH) {
 			spin_unlock(&journal->j_list_lock);
-			__flush_batch(journal, bhs, batch_count);
+			__flush_batch(journal, batch_count);
 			ret = 1;
 		}
 	}
@@ -388,7 +385,6 @@ restart:
 	if (journal->j_checkpoint_transactions == transaction &&
 			transaction->t_tid == this_tid) {
 		int batch_count = 0;
-		struct buffer_head *bhs[NR_BATCH];
 		struct journal_head *jh;
 		int retry = 0, err;
 
@@ -402,7 +398,7 @@ restart:
 				retry = 1;
 				break;
 			}
-			retry = __process_buffer(journal, jh, bhs, &batch_count,
+			retry = __process_buffer(journal, jh, &batch_count,
 						 transaction);
 			if (retry < 0 && !result)
 				result = retry;
@@ -419,7 +415,7 @@ restart:
 				spin_unlock(&journal->j_list_lock);
 				retry = 1;
 			}
-			__flush_batch(journal, bhs, &batch_count);
+			__flush_batch(journal, &batch_count);
 		}
 
 		if (retry) {
@@ -686,6 +682,7 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh)
 	   safely remove this transaction from the log */
 
 	__jbd2_journal_drop_transaction(journal, transaction);
+	kfree(transaction);
 
 	/* Just in case anybody was waiting for more transactions to be
            checkpointed... */
@@ -760,5 +757,4 @@ void __jbd2_journal_drop_transaction(journal_t *journal, transaction_t *transact
 	J_ASSERT(journal->j_running_transaction != transaction);
 
 	jbd_debug(1, "Dropping transaction %d, all done\n", transaction->t_tid);
-	kfree(transaction);
 }
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index ebc667bc54a..62804e57a44 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -25,6 +25,7 @@
 #include <linux/crc32.h>
 #include <linux/writeback.h>
 #include <linux/backing-dev.h>
+#include <linux/bio.h>
 
 /*
  * Default IO end handler for temporary BJ_IO buffer_heads.
@@ -137,7 +138,7 @@ static int journal_submit_commit_record(journal_t *journal,
 		set_buffer_ordered(bh);
 		barrier_done = 1;
 	}
-	ret = submit_bh(WRITE, bh);
+	ret = submit_bh(WRITE_SYNC, bh);
 	if (barrier_done)
 		clear_buffer_ordered(bh);
 
@@ -158,7 +159,7 @@ static int journal_submit_commit_record(journal_t *journal,
 		lock_buffer(bh);
 		set_buffer_uptodate(bh);
 		clear_buffer_dirty(bh);
-		ret = submit_bh(WRITE, bh);
+		ret = submit_bh(WRITE_SYNC, bh);
 	}
 	*cbh = bh;
 	return ret;
@@ -168,12 +169,34 @@ static int journal_submit_commit_record(journal_t *journal,
  * This function along with journal_submit_commit_record
  * allows to write the commit record asynchronously.
  */
-static int journal_wait_on_commit_record(struct buffer_head *bh)
+static int journal_wait_on_commit_record(journal_t *journal,
+					 struct buffer_head *bh)
 {
 	int ret = 0;
 
+retry:
 	clear_buffer_dirty(bh);
 	wait_on_buffer(bh);
+	if (buffer_eopnotsupp(bh) && (journal->j_flags & JBD2_BARRIER)) {
+		printk(KERN_WARNING
+		       "JBD2: wait_on_commit_record: sync failed on %s - "
+		       "disabling barriers\n", journal->j_devname);
+		spin_lock(&journal->j_state_lock);
+		journal->j_flags &= ~JBD2_BARRIER;
+		spin_unlock(&journal->j_state_lock);
+
+		lock_buffer(bh);
+		clear_buffer_dirty(bh);
+		set_buffer_uptodate(bh);
+		bh->b_end_io = journal_end_buffer_io_sync;
+
+		ret = submit_bh(WRITE_SYNC, bh);
+		if (ret) {
+			unlock_buffer(bh);
+			return ret;
+		}
+		goto retry;
+	}
 
 	if (unlikely(!buffer_uptodate(bh)))
 		ret = -EIO;
@@ -332,13 +355,15 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	int flags;
 	int err;
 	unsigned long long blocknr;
+	ktime_t start_time;
+	u64 commit_time;
 	char *tagp = NULL;
 	journal_header_t *header;
 	journal_block_tag_t *tag = NULL;
 	int space_left = 0;
 	int first_tag = 0;
 	int tag_flag;
-	int i;
+	int i, to_free = 0;
 	int tag_bytes = journal_tag_bytes(journal);
 	struct buffer_head *cbh = NULL; /* For transactional checksums */
 	__u32 crc32_sum = ~0;
@@ -458,6 +483,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	commit_transaction->t_state = T_FLUSH;
 	journal->j_committing_transaction = commit_transaction;
 	journal->j_running_transaction = NULL;
+	start_time = ktime_get();
 	commit_transaction->t_log_start = journal->j_head;
 	wake_up(&journal->j_wait_transaction_locked);
 	spin_unlock(&journal->j_state_lock);
@@ -509,6 +535,10 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 		if (is_journal_aborted(journal)) {
 			clear_buffer_jbddirty(jh2bh(jh));
 			JBUFFER_TRACE(jh, "journal is aborting: refile");
+			jbd2_buffer_abort_trigger(jh,
+						  jh->b_frozen_data ?
+						  jh->b_frozen_triggers :
+						  jh->b_triggers);
 			jbd2_journal_refile_buffer(journal, jh);
 			/* If that was the last one, we need to clean up
 			 * any descriptor buffers which may have been
@@ -799,7 +829,7 @@ wait_for_iobuf:
 			__jbd2_journal_abort_hard(journal);
 	}
 	if (!err && !is_journal_aborted(journal))
-		err = journal_wait_on_commit_record(cbh);
+		err = journal_wait_on_commit_record(journal, cbh);
 
 	if (err)
 		jbd2_journal_abort(journal, err);
@@ -844,6 +874,9 @@ restart_loop:
 		 * data.
 		 *
 		 * Otherwise, we can just throw away the frozen data now.
+		 *
+		 * We also know that the frozen data has already fired
+		 * its triggers if they exist, so we can clear that too.
 		 */
 		if (jh->b_committed_data) {
 			jbd2_free(jh->b_committed_data, bh->b_size);
@@ -851,10 +884,12 @@ restart_loop:
 			if (jh->b_frozen_data) {
 				jh->b_committed_data = jh->b_frozen_data;
 				jh->b_frozen_data = NULL;
+				jh->b_frozen_triggers = NULL;
 			}
 		} else if (jh->b_frozen_data) {
 			jbd2_free(jh->b_frozen_data, bh->b_size);
 			jh->b_frozen_data = NULL;
+			jh->b_frozen_triggers = NULL;
 		}
 
 		spin_lock(&journal->j_list_lock);
@@ -972,14 +1007,23 @@ restart_loop:
 	J_ASSERT(commit_transaction == journal->j_committing_transaction);
 	journal->j_commit_sequence = commit_transaction->t_tid;
 	journal->j_committing_transaction = NULL;
-	spin_unlock(&journal->j_state_lock);
+	commit_time = ktime_to_ns(ktime_sub(ktime_get(), start_time));
 
-	if (journal->j_commit_callback)
-		journal->j_commit_callback(journal, commit_transaction);
+	/*
+	 * weight the commit time higher than the average time so we don't
+	 * react too strongly to vast changes in the commit time
+	 */
+	if (likely(journal->j_average_commit_time))
+		journal->j_average_commit_time = (commit_time +
+				journal->j_average_commit_time*3) / 4;
+	else
+		journal->j_average_commit_time = commit_time;
+	spin_unlock(&journal->j_state_lock);
 
 	if (commit_transaction->t_checkpoint_list == NULL &&
 	    commit_transaction->t_checkpoint_io_list == NULL) {
 		__jbd2_journal_drop_transaction(journal, commit_transaction);
+		to_free = 1;
 	} else {
 		if (journal->j_checkpoint_transactions == NULL) {
 			journal->j_checkpoint_transactions = commit_transaction;
@@ -998,11 +1042,16 @@ restart_loop:
 	}
 	spin_unlock(&journal->j_list_lock);
 
+	if (journal->j_commit_callback)
+		journal->j_commit_callback(journal, commit_transaction);
+
 	trace_mark(jbd2_end_commit, "dev %s transaction %d head %d",
-		   journal->j_devname, journal->j_commit_sequence,
+		   journal->j_devname, commit_transaction->t_tid,
 		   journal->j_tail_sequence);
 	jbd_debug(1, "JBD: commit %d complete, head %d\n",
 		  journal->j_commit_sequence, journal->j_tail_sequence);
+	if (to_free)
+		kfree(commit_transaction);
 
 	wake_up(&journal->j_wait_done_commit);
 }
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index e70d657a19f..56675306ed8 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -40,6 +40,7 @@
 
 #include <asm/uaccess.h>
 #include <asm/page.h>
+#include <asm/div64.h>
 
 EXPORT_SYMBOL(jbd2_journal_start);
 EXPORT_SYMBOL(jbd2_journal_restart);
@@ -50,6 +51,7 @@ EXPORT_SYMBOL(jbd2_journal_unlock_updates);
 EXPORT_SYMBOL(jbd2_journal_get_write_access);
 EXPORT_SYMBOL(jbd2_journal_get_create_access);
 EXPORT_SYMBOL(jbd2_journal_get_undo_access);
+EXPORT_SYMBOL(jbd2_journal_set_triggers);
 EXPORT_SYMBOL(jbd2_journal_dirty_metadata);
 EXPORT_SYMBOL(jbd2_journal_release_buffer);
 EXPORT_SYMBOL(jbd2_journal_forget);
@@ -65,7 +67,6 @@ EXPORT_SYMBOL(jbd2_journal_update_format);
 EXPORT_SYMBOL(jbd2_journal_check_used_features);
 EXPORT_SYMBOL(jbd2_journal_check_available_features);
 EXPORT_SYMBOL(jbd2_journal_set_features);
-EXPORT_SYMBOL(jbd2_journal_create);
 EXPORT_SYMBOL(jbd2_journal_load);
 EXPORT_SYMBOL(jbd2_journal_destroy);
 EXPORT_SYMBOL(jbd2_journal_abort);
@@ -131,8 +132,9 @@ static int kjournald2(void *arg)
 	journal->j_task = current;
 	wake_up(&journal->j_wait_done_commit);
 
-	printk(KERN_INFO "kjournald2 starting.  Commit interval %ld seconds\n",
-			journal->j_commit_interval / HZ);
+	printk(KERN_INFO "kjournald2 starting: pid %d, dev %s, "
+	       "commit interval %ld seconds\n", current->pid,
+	       journal->j_devname, journal->j_commit_interval / HZ);
 
 	/*
 	 * And now, wait forever for commit wakeup events.
@@ -290,6 +292,7 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 	struct page *new_page;
 	unsigned int new_offset;
 	struct buffer_head *bh_in = jh2bh(jh_in);
+	struct jbd2_buffer_trigger_type *triggers;
 
 	/*
 	 * The buffer really shouldn't be locked: only the current committing
@@ -314,13 +317,23 @@ repeat:
 		done_copy_out = 1;
 		new_page = virt_to_page(jh_in->b_frozen_data);
 		new_offset = offset_in_page(jh_in->b_frozen_data);
+		triggers = jh_in->b_frozen_triggers;
 	} else {
 		new_page = jh2bh(jh_in)->b_page;
 		new_offset = offset_in_page(jh2bh(jh_in)->b_data);
+		triggers = jh_in->b_triggers;
 	}
 
 	mapped_data = kmap_atomic(new_page, KM_USER0);
 	/*
+	 * Fire any commit trigger.  Do this before checking for escaping,
+	 * as the trigger may modify the magic offset.  If a copy-out
+	 * happens afterwards, it will have the correct data in the buffer.
+	 */
+	jbd2_buffer_commit_trigger(jh_in, mapped_data + new_offset,
+				   triggers);
+
+	/*
 	 * Check for escaping
 	 */
 	if (*((__be32 *)(mapped_data + new_offset)) ==
@@ -352,6 +365,13 @@ repeat:
 		new_page = virt_to_page(tmp);
 		new_offset = offset_in_page(tmp);
 		done_copy_out = 1;
+
+		/*
+		 * This isn't strictly necessary, as we're using frozen
+		 * data for the escaping, but it keeps consistency with
+		 * b_frozen_data usage.
+		 */
+		jh_in->b_frozen_triggers = jh_in->b_triggers;
 	}
 
 	/*
@@ -631,6 +651,8 @@ struct journal_head *jbd2_journal_get_descriptor_buffer(journal_t *journal)
 		return NULL;
 
 	bh = __getblk(journal->j_dev, blocknr, journal->j_blocksize);
+	if (!bh)
+		return NULL;
 	lock_buffer(bh);
 	memset(bh->b_data, 0, journal->j_blocksize);
 	set_buffer_uptodate(bh);
@@ -824,6 +846,8 @@ static int jbd2_seq_info_show(struct seq_file *seq, void *v)
 	    jiffies_to_msecs(s->stats->u.run.rs_flushing / s->stats->ts_tid));
 	seq_printf(seq, "  %ums logging transaction\n",
 	    jiffies_to_msecs(s->stats->u.run.rs_logging / s->stats->ts_tid));
+	seq_printf(seq, "  %luus average transaction commit time\n",
+		   do_div(s->journal->j_average_commit_time, 1000));
 	seq_printf(seq, "  %lu handles per transaction\n",
 	    s->stats->u.run.rs_handle_count / s->stats->ts_tid);
 	seq_printf(seq, "  %lu blocks per transaction\n",
@@ -961,6 +985,8 @@ static journal_t * journal_init_common (void)
 	spin_lock_init(&journal->j_state_lock);
 
 	journal->j_commit_interval = (HZ * JBD2_DEFAULT_MAX_COMMIT_AGE);
+	journal->j_min_batch_time = 0;
+	journal->j_max_batch_time = 15000; /* 15ms */
 
 	/* The journal is marked for error until we succeed with recovery! */
 	journal->j_flags = JBD2_ABORT;
@@ -1016,15 +1042,14 @@ journal_t * jbd2_journal_init_dev(struct block_device *bdev,
 
 	/* journal descriptor can store up to n blocks -bzzz */
 	journal->j_blocksize = blocksize;
+	jbd2_stats_proc_init(journal);
 	n = journal->j_blocksize / sizeof(journal_block_tag_t);
 	journal->j_wbufsize = n;
 	journal->j_wbuf = kmalloc(n * sizeof(struct buffer_head*), GFP_KERNEL);
 	if (!journal->j_wbuf) {
 		printk(KERN_ERR "%s: Cant allocate bhs for commit thread\n",
 			__func__);
-		kfree(journal);
-		journal = NULL;
-		goto out;
+		goto out_err;
 	}
 	journal->j_dev = bdev;
 	journal->j_fs_dev = fs_dev;
@@ -1034,14 +1059,22 @@ journal_t * jbd2_journal_init_dev(struct block_device *bdev,
 	p = journal->j_devname;
 	while ((p = strchr(p, '/')))
 		*p = '!';
-	jbd2_stats_proc_init(journal);
 
 	bh = __getblk(journal->j_dev, start, journal->j_blocksize);
-	J_ASSERT(bh != NULL);
+	if (!bh) {
+		printk(KERN_ERR
+		       "%s: Cannot get buffer for journal superblock\n",
+		       __func__);
+		goto out_err;
+	}
 	journal->j_sb_buffer = bh;
 	journal->j_superblock = (journal_superblock_t *)bh->b_data;
-out:
+
 	return journal;
+out_err:
+	jbd2_stats_proc_exit(journal);
+	kfree(journal);
+	return NULL;
 }
 
 /**
@@ -1089,9 +1122,7 @@ journal_t * jbd2_journal_init_inode (struct inode *inode)
 	if (!journal->j_wbuf) {
 		printk(KERN_ERR "%s: Cant allocate bhs for commit thread\n",
 			__func__);
-		jbd2_stats_proc_exit(journal);
-		kfree(journal);
-		return NULL;
+		goto out_err;
 	}
 
 	err = jbd2_journal_bmap(journal, 0, &blocknr);
@@ -1099,17 +1130,24 @@ journal_t * jbd2_journal_init_inode (struct inode *inode)
 	if (err) {
 		printk(KERN_ERR "%s: Cannnot locate journal superblock\n",
 		       __func__);
-		jbd2_stats_proc_exit(journal);
-		kfree(journal);
-		return NULL;
+		goto out_err;
 	}
 
 	bh = __getblk(journal->j_dev, blocknr, journal->j_blocksize);
-	J_ASSERT(bh != NULL);
+	if (!bh) {
+		printk(KERN_ERR
+		       "%s: Cannot get buffer for journal superblock\n",
+		       __func__);
+		goto out_err;
+	}
 	journal->j_sb_buffer = bh;
 	journal->j_superblock = (journal_superblock_t *)bh->b_data;
 
 	return journal;
+out_err:
+	jbd2_stats_proc_exit(journal);
+	kfree(journal);
+	return NULL;
 }
 
 /*
@@ -1158,77 +1196,6 @@ static int journal_reset(journal_t *journal)
 }
 
 /**
- * int jbd2_journal_create() - Initialise the new journal file
- * @journal: Journal to create. This structure must have been initialised
- *
- * Given a journal_t structure which tells us which disk blocks we can
- * use, create a new journal superblock and initialise all of the
- * journal fields from scratch.
- **/
-int jbd2_journal_create(journal_t *journal)
-{
-	unsigned long long blocknr;
-	struct buffer_head *bh;
-	journal_superblock_t *sb;
-	int i, err;
-
-	if (journal->j_maxlen < JBD2_MIN_JOURNAL_BLOCKS) {
-		printk (KERN_ERR "Journal length (%d blocks) too short.\n",
-			journal->j_maxlen);
-		journal_fail_superblock(journal);
-		return -EINVAL;
-	}
-
-	if (journal->j_inode == NULL) {
-		/*
-		 * We don't know what block to start at!
-		 */
-		printk(KERN_EMERG
-		       "%s: creation of journal on external device!\n",
-		       __func__);
-		BUG();
-	}
-
-	/* Zero out the entire journal on disk.  We cannot afford to
-	   have any blocks on disk beginning with JBD2_MAGIC_NUMBER. */
-	jbd_debug(1, "JBD: Zeroing out journal blocks...\n");
-	for (i = 0; i < journal->j_maxlen; i++) {
-		err = jbd2_journal_bmap(journal, i, &blocknr);
-		if (err)
-			return err;
-		bh = __getblk(journal->j_dev, blocknr, journal->j_blocksize);
-		lock_buffer(bh);
-		memset (bh->b_data, 0, journal->j_blocksize);
-		BUFFER_TRACE(bh, "marking dirty");
-		mark_buffer_dirty(bh);
-		BUFFER_TRACE(bh, "marking uptodate");
-		set_buffer_uptodate(bh);
-		unlock_buffer(bh);
-		__brelse(bh);
-	}
-
-	sync_blockdev(journal->j_dev);
-	jbd_debug(1, "JBD: journal cleared.\n");
-
-	/* OK, fill in the initial static fields in the new superblock */
-	sb = journal->j_superblock;
-
-	sb->s_header.h_magic	 = cpu_to_be32(JBD2_MAGIC_NUMBER);
-	sb->s_header.h_blocktype = cpu_to_be32(JBD2_SUPERBLOCK_V2);
-
-	sb->s_blocksize	= cpu_to_be32(journal->j_blocksize);
-	sb->s_maxlen	= cpu_to_be32(journal->j_maxlen);
-	sb->s_first	= cpu_to_be32(1);
-
-	journal->j_transaction_sequence = 1;
-
-	journal->j_flags &= ~JBD2_ABORT;
-	journal->j_format_version = 2;
-
-	return journal_reset(journal);
-}
-
-/**
  * void jbd2_journal_update_superblock() - Update journal sb on disk.
  * @journal: The journal to update.
  * @wait: Set to '0' if you don't want to wait for IO completion.
@@ -1472,7 +1439,9 @@ int jbd2_journal_destroy(journal_t *journal)
 	spin_lock(&journal->j_list_lock);
 	while (journal->j_checkpoint_transactions != NULL) {
 		spin_unlock(&journal->j_list_lock);
+		mutex_lock(&journal->j_checkpoint_mutex);
 		jbd2_log_do_checkpoint(journal);
+		mutex_unlock(&journal->j_checkpoint_mutex);
 		spin_lock(&journal->j_list_lock);
 	}
 
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index 39b7805a599..46b4e347ed7 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -25,6 +25,7 @@
 #include <linux/timer.h>
 #include <linux/mm.h>
 #include <linux/highmem.h>
+#include <linux/hrtimer.h>
 
 static void __jbd2_journal_temp_unlink_buffer(struct journal_head *jh);
 
@@ -48,6 +49,7 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction)
 {
 	transaction->t_journal = journal;
 	transaction->t_state = T_RUNNING;
+	transaction->t_start_time = ktime_get();
 	transaction->t_tid = journal->j_transaction_sequence++;
 	transaction->t_expires = jiffies + journal->j_commit_interval;
 	spin_lock_init(&transaction->t_handle_lock);
@@ -741,6 +743,12 @@ done:
 		source = kmap_atomic(page, KM_USER0);
 		memcpy(jh->b_frozen_data, source+offset, jh2bh(jh)->b_size);
 		kunmap_atomic(source, KM_USER0);
+
+		/*
+		 * Now that the frozen data is saved off, we need to store
+		 * any matching triggers.
+		 */
+		jh->b_frozen_triggers = jh->b_triggers;
 	}
 	jbd_unlock_bh_state(bh);
 
@@ -944,6 +952,47 @@ out:
 }
 
 /**
+ * void jbd2_journal_set_triggers() - Add triggers for commit writeout
+ * @bh: buffer to trigger on
+ * @type: struct jbd2_buffer_trigger_type containing the trigger(s).
+ *
+ * Set any triggers on this journal_head.  This is always safe, because
+ * triggers for a committing buffer will be saved off, and triggers for
+ * a running transaction will match the buffer in that transaction.
+ *
+ * Call with NULL to clear the triggers.
+ */
+void jbd2_journal_set_triggers(struct buffer_head *bh,
+			       struct jbd2_buffer_trigger_type *type)
+{
+	struct journal_head *jh = bh2jh(bh);
+
+	jh->b_triggers = type;
+}
+
+void jbd2_buffer_commit_trigger(struct journal_head *jh, void *mapped_data,
+				struct jbd2_buffer_trigger_type *triggers)
+{
+	struct buffer_head *bh = jh2bh(jh);
+
+	if (!triggers || !triggers->t_commit)
+		return;
+
+	triggers->t_commit(triggers, bh, mapped_data, bh->b_size);
+}
+
+void jbd2_buffer_abort_trigger(struct journal_head *jh,
+			       struct jbd2_buffer_trigger_type *triggers)
+{
+	if (!triggers || !triggers->t_abort)
+		return;
+
+	triggers->t_abort(triggers, jh2bh(jh));
+}
+
+
+
+/**
  * int jbd2_journal_dirty_metadata() -  mark a buffer as containing dirty metadata
  * @handle: transaction to add buffer to.
  * @bh: buffer to mark
@@ -1193,7 +1242,7 @@ int jbd2_journal_stop(handle_t *handle)
 {
 	transaction_t *transaction = handle->h_transaction;
 	journal_t *journal = transaction->t_journal;
-	int old_handle_count, err;
+	int err;
 	pid_t pid;
 
 	J_ASSERT(journal_current_handle() == handle);
@@ -1216,24 +1265,54 @@ int jbd2_journal_stop(handle_t *handle)
 	/*
 	 * Implement synchronous transaction batching.  If the handle
 	 * was synchronous, don't force a commit immediately.  Let's
-	 * yield and let another thread piggyback onto this transaction.
-	 * Keep doing that while new threads continue to arrive.
-	 * It doesn't cost much - we're about to run a commit and sleep
-	 * on IO anyway.  Speeds up many-threaded, many-dir operations
-	 * by 30x or more...
+	 * yield and let another thread piggyback onto this
+	 * transaction.  Keep doing that while new threads continue to
+	 * arrive.  It doesn't cost much - we're about to run a commit
+	 * and sleep on IO anyway.  Speeds up many-threaded, many-dir
+	 * operations by 30x or more...
 	 *
-	 * But don't do this if this process was the most recent one to
-	 * perform a synchronous write.  We do this to detect the case where a
-	 * single process is doing a stream of sync writes.  No point in waiting
-	 * for joiners in that case.
+	 * We try and optimize the sleep time against what the
+	 * underlying disk can do, instead of having a static sleep
+	 * time.  This is useful for the case where our storage is so
+	 * fast that it is more optimal to go ahead and force a flush
+	 * and wait for the transaction to be committed than it is to
+	 * wait for an arbitrary amount of time for new writers to
+	 * join the transaction.  We achieve this by measuring how
+	 * long it takes to commit a transaction, and compare it with
+	 * how long this transaction has been running, and if run time
+	 * < commit time then we sleep for the delta and commit.  This
+	 * greatly helps super fast disks that would see slowdowns as
+	 * more threads started doing fsyncs.
+	 *
+	 * But don't do this if this process was the most recent one
+	 * to perform a synchronous write.  We do this to detect the
+	 * case where a single process is doing a stream of sync
+	 * writes.  No point in waiting for joiners in that case.
 	 */
 	pid = current->pid;
 	if (handle->h_sync && journal->j_last_sync_writer != pid) {
+		u64 commit_time, trans_time;
+
 		journal->j_last_sync_writer = pid;
-		do {
-			old_handle_count = transaction->t_handle_count;
-			schedule_timeout_uninterruptible(1);
-		} while (old_handle_count != transaction->t_handle_count);
+
+		spin_lock(&journal->j_state_lock);
+		commit_time = journal->j_average_commit_time;
+		spin_unlock(&journal->j_state_lock);
+
+		trans_time = ktime_to_ns(ktime_sub(ktime_get(),
+						   transaction->t_start_time));
+
+		commit_time = max_t(u64, commit_time,
+				    1000*journal->j_min_batch_time);
+		commit_time = min_t(u64, commit_time,
+				    1000*journal->j_max_batch_time);
+
+		if (trans_time < commit_time) {
+			ktime_t expires = ktime_add_ns(ktime_get(),
+						       commit_time);
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			schedule_hrtimeout(&expires, HRTIMER_MODE_ABS);
+		}
 	}
 
 	current->journal_info = NULL;
diff --git a/fs/jffs2/compr_rubin.c b/fs/jffs2/compr_rubin.c
index c73fa89b5f8..170d289ac78 100644
--- a/fs/jffs2/compr_rubin.c
+++ b/fs/jffs2/compr_rubin.c
@@ -22,9 +22,7 @@
 
 
 #define BIT_DIVIDER_MIPS 1043
-static int bits_mips[8] = { 277,249,290,267,229,341,212,241}; /* mips32 */
-
-#include <linux/errno.h>
+static int bits_mips[8] = { 277, 249, 290, 267, 229, 341, 212, 241};
 
 struct pushpull {
 	unsigned char *buf;
@@ -43,7 +41,9 @@ struct rubin_state {
 	int bits[8];
 };
 
-static inline void init_pushpull(struct pushpull *pp, char *buf, unsigned buflen, unsigned ofs, unsigned reserve)
+static inline void init_pushpull(struct pushpull *pp, char *buf,
+				 unsigned buflen, unsigned ofs,
+				 unsigned reserve)
 {
 	pp->buf = buf;
 	pp->buflen = buflen;
@@ -53,16 +53,14 @@ static inline void init_pushpull(struct pushpull *pp, char *buf, unsigned buflen
 
 static inline int pushbit(struct pushpull *pp, int bit, int use_reserved)
 {
-	if (pp->ofs >= pp->buflen - (use_reserved?0:pp->reserve)) {
+	if (pp->ofs >= pp->buflen - (use_reserved?0:pp->reserve))
 		return -ENOSPC;
-	}
 
-	if (bit) {
-		pp->buf[pp->ofs >> 3] |= (1<<(7-(pp->ofs &7)));
-	}
-	else {
-		pp->buf[pp->ofs >> 3] &= ~(1<<(7-(pp->ofs &7)));
-	}
+	if (bit)
+		pp->buf[pp->ofs >> 3] |= (1<<(7-(pp->ofs & 7)));
+	else
+		pp->buf[pp->ofs >> 3] &= ~(1<<(7-(pp->ofs & 7)));
+
 	pp->ofs++;
 
 	return 0;
@@ -97,6 +95,7 @@ static void init_rubin(struct rubin_state *rs, int div, int *bits)
 	rs->p = (long) (2 * UPPER_BIT_RUBIN);
 	rs->bit_number = (long) 0;
 	rs->bit_divider = div;
+
 	for (c=0; c<8; c++)
 		rs->bits[c] = bits[c];
 }
@@ -108,7 +107,8 @@ static int encode(struct rubin_state *rs, long A, long B, int symbol)
 	long i0, i1;
 	int ret;
 
-	while ((rs->q >= UPPER_BIT_RUBIN) || ((rs->p + rs->q) <= UPPER_BIT_RUBIN)) {
+	while ((rs->q >= UPPER_BIT_RUBIN) ||
+	       ((rs->p + rs->q) <= UPPER_BIT_RUBIN)) {
 		rs->bit_number++;
 
 		ret = pushbit(&rs->pp, (rs->q & UPPER_BIT_RUBIN) ? 1 : 0, 0);
@@ -119,12 +119,12 @@ static int encode(struct rubin_state *rs, long A, long B, int symbol)
 		rs->p <<= 1;
 	}
 	i0 = A * rs->p / (A + B);
-	if (i0 <= 0) {
+	if (i0 <= 0)
 		i0 = 1;
-	}
-	if (i0 >= rs->p) {
+
+	if (i0 >= rs->p)
 		i0 = rs->p - 1;
-	}
+
 	i1 = rs->p - i0;
 
 	if (symbol == 0)
@@ -157,11 +157,13 @@ static void init_decode(struct rubin_state *rs, int div, int *bits)
 	/* behalve lower */
 	rs->rec_q = 0;
 
-	for (rs->bit_number = 0; rs->bit_number++ < RUBIN_REG_SIZE; rs->rec_q = rs->rec_q * 2 + (long) (pullbit(&rs->pp)))
+	for (rs->bit_number = 0; rs->bit_number++ < RUBIN_REG_SIZE;
+	     rs->rec_q = rs->rec_q * 2 + (long) (pullbit(&rs->pp)))
 		;
 }
 
-static void __do_decode(struct rubin_state *rs, unsigned long p, unsigned long q)
+static void __do_decode(struct rubin_state *rs, unsigned long p,
+			unsigned long q)
 {
 	register unsigned long lower_bits_rubin = LOWER_BITS_RUBIN;
 	unsigned long rec_q;
@@ -207,12 +209,11 @@ static int decode(struct rubin_state *rs, long A, long B)
 		__do_decode(rs, p, q);
 
 	i0 = A * rs->p / (A + B);
-	if (i0 <= 0) {
+	if (i0 <= 0)
 		i0 = 1;
-	}
-	if (i0 >= rs->p) {
+
+	if (i0 >= rs->p)
 		i0 = rs->p - 1;
-	}
 
 	threshold = rs->q + i0;
 	symbol = rs->rec_q >= threshold;
@@ -234,14 +235,15 @@ static int out_byte(struct rubin_state *rs, unsigned char byte)
 	struct rubin_state rs_copy;
 	rs_copy = *rs;
 
-	for (i=0;i<8;i++) {
-		ret = encode(rs, rs->bit_divider-rs->bits[i],rs->bits[i],byte&1);
+	for (i=0; i<8; i++) {
+		ret = encode(rs, rs->bit_divider-rs->bits[i],
+			     rs->bits[i], byte & 1);
 		if (ret) {
 			/* Failed. Restore old state */
 			*rs = rs_copy;
 			return ret;
 		}
-		byte=byte>>1;
+		byte >>= 1 ;
 	}
 	return 0;
 }
@@ -251,7 +253,8 @@ static int in_byte(struct rubin_state *rs)
 	int i, result = 0, bit_divider = rs->bit_divider;
 
 	for (i = 0; i < 8; i++)
-		result |= decode(rs, bit_divider - rs->bits[i], rs->bits[i]) << i;
+		result |= decode(rs, bit_divider - rs->bits[i],
+				 rs->bits[i]) << i;
 
 	return result;
 }
@@ -259,7 +262,8 @@ static int in_byte(struct rubin_state *rs)
 
 
 static int rubin_do_compress(int bit_divider, int *bits, unsigned char *data_in,
-		      unsigned char *cpage_out, uint32_t *sourcelen, uint32_t *dstlen)
+			     unsigned char *cpage_out, uint32_t *sourcelen,
+			     uint32_t *dstlen)
 	{
 	int outpos = 0;
 	int pos=0;
@@ -295,7 +299,8 @@ static int rubin_do_compress(int bit_divider, int *bits, unsigned char *data_in,
 int jffs2_rubinmips_compress(unsigned char *data_in, unsigned char *cpage_out,
 		   uint32_t *sourcelen, uint32_t *dstlen, void *model)
 {
-	return rubin_do_compress(BIT_DIVIDER_MIPS, bits_mips, data_in, cpage_out, sourcelen, dstlen);
+	return rubin_do_compress(BIT_DIVIDER_MIPS, bits_mips, data_in,
+				 cpage_out, sourcelen, dstlen);
 }
 #endif
 static int jffs2_dynrubin_compress(unsigned char *data_in,
@@ -316,9 +321,8 @@ static int jffs2_dynrubin_compress(unsigned char *data_in,
 		return -1;
 
 	memset(histo, 0, 256);
-	for (i=0; i<mysrclen; i++) {
+	for (i=0; i<mysrclen; i++)
 		histo[data_in[i]]++;
-	}
 	memset(bits, 0, sizeof(int)*8);
 	for (i=0; i<256; i++) {
 		if (i&128)
@@ -346,7 +350,8 @@ static int jffs2_dynrubin_compress(unsigned char *data_in,
 		cpage_out[i] = bits[i];
 	}
 
-	ret = rubin_do_compress(256, bits, data_in, cpage_out+8, &mysrclen, &mydstlen);
+	ret = rubin_do_compress(256, bits, data_in, cpage_out+8, &mysrclen,
+				&mydstlen);
 	if (ret)
 		return ret;
 
@@ -363,8 +368,10 @@ static int jffs2_dynrubin_compress(unsigned char *data_in,
 	return 0;
 }
 
-static void rubin_do_decompress(int bit_divider, int *bits, unsigned char *cdata_in,
-			 unsigned char *page_out, uint32_t srclen, uint32_t destlen)
+static void rubin_do_decompress(int bit_divider, int *bits,
+				unsigned char *cdata_in, 
+				unsigned char *page_out, uint32_t srclen,
+				uint32_t destlen)
 {
 	int outpos = 0;
 	struct rubin_state rs;
@@ -372,9 +379,8 @@ static void rubin_do_decompress(int bit_divider, int *bits, unsigned char *cdata
 	init_pushpull(&rs.pp, cdata_in, srclen, 0, 0);
 	init_decode(&rs, bit_divider, bits);
 
-	while (outpos < destlen) {
+	while (outpos < destlen)
 		page_out[outpos++] = in_byte(&rs);
-	}
 }
 
 
@@ -383,7 +389,8 @@ static int jffs2_rubinmips_decompress(unsigned char *data_in,
 				      uint32_t sourcelen, uint32_t dstlen,
 				      void *model)
 {
-	rubin_do_decompress(BIT_DIVIDER_MIPS, bits_mips, data_in, cpage_out, sourcelen, dstlen);
+	rubin_do_decompress(BIT_DIVIDER_MIPS, bits_mips, data_in,
+			    cpage_out, sourcelen, dstlen);
 	return 0;
 }
 
@@ -398,52 +405,53 @@ static int jffs2_dynrubin_decompress(unsigned char *data_in,
 	for (c=0; c<8; c++)
 		bits[c] = data_in[c];
 
-	rubin_do_decompress(256, bits, data_in+8, cpage_out, sourcelen-8, dstlen);
+	rubin_do_decompress(256, bits, data_in+8, cpage_out, sourcelen-8,
+			    dstlen);
 	return 0;
 }
 
 static struct jffs2_compressor jffs2_rubinmips_comp = {
-    .priority = JFFS2_RUBINMIPS_PRIORITY,
-    .name = "rubinmips",
-    .compr = JFFS2_COMPR_DYNRUBIN,
-    .compress = NULL, /*&jffs2_rubinmips_compress,*/
-    .decompress = &jffs2_rubinmips_decompress,
+	.priority = JFFS2_RUBINMIPS_PRIORITY,
+	.name = "rubinmips",
+	.compr = JFFS2_COMPR_DYNRUBIN,
+	.compress = NULL, /*&jffs2_rubinmips_compress,*/
+	.decompress = &jffs2_rubinmips_decompress,
 #ifdef JFFS2_RUBINMIPS_DISABLED
-    .disabled = 1,
+	.disabled = 1,
 #else
-    .disabled = 0,
+	.disabled = 0,
 #endif
 };
 
 int jffs2_rubinmips_init(void)
 {
-    return jffs2_register_compressor(&jffs2_rubinmips_comp);
+	return jffs2_register_compressor(&jffs2_rubinmips_comp);
 }
 
 void jffs2_rubinmips_exit(void)
 {
-    jffs2_unregister_compressor(&jffs2_rubinmips_comp);
+	jffs2_unregister_compressor(&jffs2_rubinmips_comp);
 }
 
 static struct jffs2_compressor jffs2_dynrubin_comp = {
-    .priority = JFFS2_DYNRUBIN_PRIORITY,
-    .name = "dynrubin",
-    .compr = JFFS2_COMPR_RUBINMIPS,
-    .compress = jffs2_dynrubin_compress,
-    .decompress = &jffs2_dynrubin_decompress,
+	.priority = JFFS2_DYNRUBIN_PRIORITY,
+	.name = "dynrubin",
+	.compr = JFFS2_COMPR_RUBINMIPS,
+	.compress = jffs2_dynrubin_compress,
+	.decompress = &jffs2_dynrubin_decompress,
 #ifdef JFFS2_DYNRUBIN_DISABLED
-    .disabled = 1,
+	.disabled = 1,
 #else
-    .disabled = 0,
+	.disabled = 0,
 #endif
 };
 
 int jffs2_dynrubin_init(void)
 {
-    return jffs2_register_compressor(&jffs2_dynrubin_comp);
+	return jffs2_register_compressor(&jffs2_dynrubin_comp);
 }
 
 void jffs2_dynrubin_exit(void)
 {
-    jffs2_unregister_compressor(&jffs2_dynrubin_comp);
+	jffs2_unregister_compressor(&jffs2_dynrubin_comp);
 }
diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c
index 259461b910a..c32b4a1ad6c 100644
--- a/fs/jffs2/erase.c
+++ b/fs/jffs2/erase.c
@@ -175,7 +175,7 @@ static void jffs2_erase_failed(struct jffs2_sb_info *c, struct jffs2_eraseblock
 {
 	/* For NAND, if the failure did not occur at the device level for a
 	   specific physical page, don't bother updating the bad block table. */
-	if (jffs2_cleanmarker_oob(c) && (bad_offset != MTD_FAIL_ADDR_UNKNOWN)) {
+	if (jffs2_cleanmarker_oob(c) && (bad_offset != (uint32_t)MTD_FAIL_ADDR_UNKNOWN)) {
 		/* We had a device-level failure to erase.  Let's see if we've
 		   failed too many times. */
 		if (!jffs2_write_nand_badblock(c, jeb, bad_offset)) {
@@ -209,7 +209,8 @@ static void jffs2_erase_callback(struct erase_info *instr)
 	struct erase_priv_struct *priv = (void *)instr->priv;
 
 	if(instr->state != MTD_ERASE_DONE) {
-		printk(KERN_WARNING "Erase at 0x%08x finished, but state != MTD_ERASE_DONE. State is 0x%x instead.\n", instr->addr, instr->state);
+		printk(KERN_WARNING "Erase at 0x%08llx finished, but state != MTD_ERASE_DONE. State is 0x%x instead.\n",
+			(unsigned long long)instr->addr, instr->state);
 		jffs2_erase_failed(priv->c, priv->jeb, instr->fail_addr);
 	} else {
 		jffs2_erase_succeeded(priv->c, priv->jeb);
diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index 5a98aa87c85..5edc2bf2058 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -132,7 +132,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping,
 	uint32_t pageofs = index << PAGE_CACHE_SHIFT;
 	int ret = 0;
 
-	pg = __grab_cache_page(mapping, index);
+	pg = grab_cache_page_write_begin(mapping, index, flags);
 	if (!pg)
 		return -ENOMEM;
 	*pagep = pg;
diff --git a/fs/jffs2/nodelist.h b/fs/jffs2/nodelist.h
index 1750445556c..507ed6ec184 100644
--- a/fs/jffs2/nodelist.h
+++ b/fs/jffs2/nodelist.h
@@ -366,9 +366,6 @@ void jffs2_free_ino_caches(struct jffs2_sb_info *c);
 void jffs2_free_raw_node_refs(struct jffs2_sb_info *c);
 struct jffs2_node_frag *jffs2_lookup_node_frag(struct rb_root *fragtree, uint32_t offset);
 void jffs2_kill_fragtree(struct rb_root *root, struct jffs2_sb_info *c_delete);
-struct rb_node *rb_next(struct rb_node *);
-struct rb_node *rb_prev(struct rb_node *);
-void rb_replace_node(struct rb_node *victim, struct rb_node *new, struct rb_root *root);
 int jffs2_add_full_dnode_to_inode(struct jffs2_sb_info *c, struct jffs2_inode_info *f, struct jffs2_full_dnode *fn);
 uint32_t jffs2_truncate_fragtree (struct jffs2_sb_info *c, struct rb_root *list, uint32_t size);
 struct jffs2_raw_node_ref *jffs2_link_node_ref(struct jffs2_sb_info *c,
diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c
index 210339784b5..b00ee9f05a0 100644
--- a/fs/jfs/inode.c
+++ b/fs/jfs/inode.c
@@ -59,8 +59,14 @@ struct inode *jfs_iget(struct super_block *sb, unsigned long ino)
 		if (inode->i_size >= IDATASIZE) {
 			inode->i_op = &page_symlink_inode_operations;
 			inode->i_mapping->a_ops = &jfs_aops;
-		} else
+		} else {
 			inode->i_op = &jfs_symlink_inode_operations;
+			/*
+			 * The inline data should be null-terminated, but
+			 * don't let on-disk corruption crash the kernel
+			 */
+			JFS_IP(inode)->i_inline[inode->i_size] = '\0';
+		}
 	} else {
 		inode->i_op = &jfs_file_inode_operations;
 		init_special_inode(inode, inode->i_mode, inode->i_rdev);
diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
index d6363d8309d..0f94381ca6d 100644
--- a/fs/jfs/jfs_imap.c
+++ b/fs/jfs/jfs_imap.c
@@ -58,9 +58,9 @@
 
 /*
  * __mark_inode_dirty expects inodes to be hashed.  Since we don't want
- * special inodes in the fileset inode space, we hash them to a dummy head
+ * special inodes in the fileset inode space, we make them appear hashed,
+ * but do not put on any lists.
  */
-static HLIST_HEAD(aggregate_hash);
 
 /*
  * imap locks
@@ -496,7 +496,11 @@ struct inode *diReadSpecial(struct super_block *sb, ino_t inum, int secondary)
 	/* release the page */
 	release_metapage(mp);
 
-	hlist_add_head(&ip->i_hash, &aggregate_hash);
+	/*
+	 * that will look hashed, but won't be on any list; hlist_del()
+	 * will work fine and require no locking.
+	 */
+	ip->i_hash.pprev = &ip->i_hash.next;
 
 	return (ip);
 }
diff --git a/fs/jfs/jfs_inode.c b/fs/jfs/jfs_inode.c
index 70022fd1c53..d4d142c2edd 100644
--- a/fs/jfs/jfs_inode.c
+++ b/fs/jfs/jfs_inode.c
@@ -79,7 +79,8 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 	inode = new_inode(sb);
 	if (!inode) {
 		jfs_warn("ialloc: new_inode returned NULL!");
-		return ERR_PTR(-ENOMEM);
+		rc = -ENOMEM;
+		goto fail;
 	}
 
 	jfs_inode = JFS_IP(inode);
@@ -89,8 +90,12 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 		jfs_warn("ialloc: diAlloc returned %d!", rc);
 		if (rc == -EIO)
 			make_bad_inode(inode);
-		iput(inode);
-		return ERR_PTR(rc);
+		goto fail_put;
+	}
+
+	if (insert_inode_locked(inode) < 0) {
+		rc = -EINVAL;
+		goto fail_unlock;
 	}
 
 	inode->i_uid = current_fsuid();
@@ -112,11 +117,8 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 	 * Allocate inode to quota.
 	 */
 	if (DQUOT_ALLOC_INODE(inode)) {
-		DQUOT_DROP(inode);
-		inode->i_flags |= S_NOQUOTA;
-		inode->i_nlink = 0;
-		iput(inode);
-		return ERR_PTR(-EDQUOT);
+		rc = -EDQUOT;
+		goto fail_drop;
 	}
 
 	inode->i_mode = mode;
@@ -158,4 +160,15 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 	jfs_info("ialloc returns inode = 0x%p\n", inode);
 
 	return inode;
+
+fail_drop:
+	DQUOT_DROP(inode);
+	inode->i_flags |= S_NOQUOTA;
+fail_unlock:
+	inode->i_nlink = 0;
+	unlock_new_inode(inode);
+fail_put:
+	iput(inode);
+fail:
+	return ERR_PTR(rc);
 }
diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c
index cc3cedffbfa..b4de56b851e 100644
--- a/fs/jfs/namei.c
+++ b/fs/jfs/namei.c
@@ -155,7 +155,6 @@ static int jfs_create(struct inode *dip, struct dentry *dentry, int mode,
 	ip->i_fop = &jfs_file_operations;
 	ip->i_mapping->a_ops = &jfs_aops;
 
-	insert_inode_hash(ip);
 	mark_inode_dirty(ip);
 
 	dip->i_ctime = dip->i_mtime = CURRENT_TIME;
@@ -171,9 +170,12 @@ static int jfs_create(struct inode *dip, struct dentry *dentry, int mode,
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
+		unlock_new_inode(ip);
 		iput(ip);
-	} else
+	} else {
 		d_instantiate(dentry, ip);
+		unlock_new_inode(ip);
+	}
 
       out2:
 	free_UCSname(&dname);
@@ -289,7 +291,6 @@ static int jfs_mkdir(struct inode *dip, struct dentry *dentry, int mode)
 	ip->i_op = &jfs_dir_inode_operations;
 	ip->i_fop = &jfs_dir_operations;
 
-	insert_inode_hash(ip);
 	mark_inode_dirty(ip);
 
 	/* update parent directory inode */
@@ -306,9 +307,12 @@ static int jfs_mkdir(struct inode *dip, struct dentry *dentry, int mode)
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
+		unlock_new_inode(ip);
 		iput(ip);
-	} else
+	} else {
 		d_instantiate(dentry, ip);
+		unlock_new_inode(ip);
+	}
 
       out2:
 	free_UCSname(&dname);
@@ -1019,7 +1023,6 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry,
 		goto out3;
 	}
 
-	insert_inode_hash(ip);
 	mark_inode_dirty(ip);
 
 	dip->i_ctime = dip->i_mtime = CURRENT_TIME;
@@ -1039,9 +1042,12 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry,
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
+		unlock_new_inode(ip);
 		iput(ip);
-	} else
+	} else {
 		d_instantiate(dentry, ip);
+		unlock_new_inode(ip);
+	}
 
       out2:
 	free_UCSname(&dname);
@@ -1399,7 +1405,6 @@ static int jfs_mknod(struct inode *dir, struct dentry *dentry,
 	jfs_ip->dev = new_encode_dev(rdev);
 	init_special_inode(ip, ip->i_mode, rdev);
 
-	insert_inode_hash(ip);
 	mark_inode_dirty(ip);
 
 	dir->i_ctime = dir->i_mtime = CURRENT_TIME;
@@ -1417,9 +1422,12 @@ static int jfs_mknod(struct inode *dir, struct dentry *dentry,
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
+		unlock_new_inode(ip);
 		iput(ip);
-	} else
+	} else {
 		d_instantiate(dentry, ip);
+		unlock_new_inode(ip);
+	}
 
       out1:
 	free_UCSname(&dname);
diff --git a/fs/jfs/super.c b/fs/jfs/super.c
index 0dae345e481..b37d1f78b85 100644
--- a/fs/jfs/super.c
+++ b/fs/jfs/super.c
@@ -543,7 +543,7 @@ out_kfree:
 	return ret;
 }
 
-static void jfs_write_super_lockfs(struct super_block *sb)
+static int jfs_freeze(struct super_block *sb)
 {
 	struct jfs_sb_info *sbi = JFS_SBI(sb);
 	struct jfs_log *log = sbi->log;
@@ -553,9 +553,10 @@ static void jfs_write_super_lockfs(struct super_block *sb)
 		lmLogShutdown(log);
 		updateSuper(sb, FM_CLEAN);
 	}
+	return 0;
 }
 
-static void jfs_unlockfs(struct super_block *sb)
+static int jfs_unfreeze(struct super_block *sb)
 {
 	struct jfs_sb_info *sbi = JFS_SBI(sb);
 	struct jfs_log *log = sbi->log;
@@ -568,6 +569,7 @@ static void jfs_unlockfs(struct super_block *sb)
 		else
 			txResume(sb);
 	}
+	return 0;
 }
 
 static int jfs_get_sb(struct file_system_type *fs_type,
@@ -735,8 +737,8 @@ static const struct super_operations jfs_super_operations = {
 	.delete_inode	= jfs_delete_inode,
 	.put_super	= jfs_put_super,
 	.sync_fs	= jfs_sync_fs,
-	.write_super_lockfs = jfs_write_super_lockfs,
-	.unlockfs       = jfs_unlockfs,
+	.freeze_fs	= jfs_freeze,
+	.unfreeze_fs	= jfs_unfreeze,
 	.statfs		= jfs_statfs,
 	.remount_fs	= jfs_remount,
 	.show_options	= jfs_show_options,
diff --git a/fs/libfs.c b/fs/libfs.c
index e960a832190..49b44099dab 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -231,7 +231,6 @@ int get_sb_pseudo(struct file_system_type *fs_type, char *name,
 	 */
 	root->i_ino = 1;
 	root->i_mode = S_IFDIR | S_IRUSR | S_IWUSR;
-	root->i_uid = root->i_gid = 0;
 	root->i_atime = root->i_mtime = root->i_ctime = CURRENT_TIME;
 	dentry = d_alloc(NULL, &d_name);
 	if (!dentry) {
@@ -360,7 +359,7 @@ int simple_write_begin(struct file *file, struct address_space *mapping,
 	index = pos >> PAGE_CACHE_SHIFT;
 	from = pos & (PAGE_CACHE_SIZE - 1);
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 
@@ -436,8 +435,6 @@ int simple_fill_super(struct super_block *s, int magic, struct tree_descr *files
 	 */
 	inode->i_ino = 1;
 	inode->i_mode = S_IFDIR | 0755;
-	inode->i_uid = inode->i_gid = 0;
-	inode->i_blocks = 0;
 	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 	inode->i_op = &simple_dir_inode_operations;
 	inode->i_fop = &simple_dir_operations;
@@ -464,8 +461,6 @@ int simple_fill_super(struct super_block *s, int magic, struct tree_descr *files
 		if (!inode)
 			goto out;
 		inode->i_mode = S_IFREG | files->mode;
-		inode->i_uid = inode->i_gid = 0;
-		inode->i_blocks = 0;
 		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 		inode->i_fop = files->ops;
 		inode->i_ino = i;
diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c
index 8307dd64bf4..1f3b0fc0d35 100644
--- a/fs/lockd/clntlock.c
+++ b/fs/lockd/clntlock.c
@@ -14,6 +14,7 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/lockd/lockd.h>
 #include <linux/smp_lock.h>
+#include <linux/kthread.h>
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 
@@ -60,7 +61,7 @@ struct nlm_host *nlmclnt_init(const struct nlmclnt_initdata *nlm_init)
 
 	host = nlmclnt_lookup_host(nlm_init->address, nlm_init->addrlen,
 				   nlm_init->protocol, nlm_version,
-				   nlm_init->hostname);
+				   nlm_init->hostname, nlm_init->noresvport);
 	if (host == NULL) {
 		lockd_down();
 		return ERR_PTR(-ENOLCK);
@@ -191,11 +192,15 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 void
 nlmclnt_recovery(struct nlm_host *host)
 {
+	struct task_struct *task;
+
 	if (!host->h_reclaiming++) {
 		nlm_get_host(host);
-		__module_get(THIS_MODULE);
-		if (kernel_thread(reclaimer, host, CLONE_FS | CLONE_FILES) < 0)
-			module_put(THIS_MODULE);
+		task = kthread_run(reclaimer, host, "%s-reclaim", host->h_name);
+		if (IS_ERR(task))
+			printk(KERN_ERR "lockd: unable to spawn reclaimer "
+				"thread. Locks for %s won't be reclaimed! "
+				"(%ld)\n", host->h_name, PTR_ERR(task));
 	}
 }
 
@@ -207,7 +212,6 @@ reclaimer(void *ptr)
 	struct file_lock *fl, *next;
 	u32 nsmstate;
 
-	daemonize("%s-reclaim", host->h_name);
 	allow_signal(SIGKILL);
 
 	down_write(&host->h_rwsem);
@@ -233,7 +237,12 @@ restart:
 	list_for_each_entry_safe(fl, next, &host->h_reclaim, fl_u.nfs_fl.list) {
 		list_del_init(&fl->fl_u.nfs_fl.list);
 
-		/* Why are we leaking memory here? --okir */
+		/*
+		 * sending this thread a SIGKILL will result in any unreclaimed
+		 * locks being removed from the h_granted list. This means that
+		 * the kernel will not attempt to reclaim them again if a new
+		 * reclaimer thread is spawned for this host.
+		 */
 		if (signalled())
 			continue;
 		if (nlmclnt_reclaim(host, fl) != 0)
@@ -261,5 +270,5 @@ restart:
 	nlm_release_host(host);
 	lockd_down();
 	unlock_kernel();
-	module_put_and_exit(0);
+	return 0;
 }
diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
index 31668b690e0..dd7957064a8 100644
--- a/fs/lockd/clntproc.c
+++ b/fs/lockd/clntproc.c
@@ -16,7 +16,6 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/svc.h>
 #include <linux/lockd/lockd.h>
-#include <linux/lockd/sm_inter.h>
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 #define NLMCLNT_GRACE_WAIT	(5*HZ)
@@ -518,11 +517,9 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 	unsigned char fl_type;
 	int status = -ENOLCK;
 
-	if (nsm_monitor(host) < 0) {
-		printk(KERN_NOTICE "lockd: failed to monitor %s\n",
-					host->h_name);
+	if (nsm_monitor(host) < 0)
 		goto out;
-	}
+
 	fl->fl_flags |= FL_ACCESS;
 	status = do_vfs_lock(fl);
 	fl->fl_flags = fl_flags;
diff --git a/fs/lockd/host.c b/fs/lockd/host.c
index e05d0441603..99d737bd432 100644
--- a/fs/lockd/host.c
+++ b/fs/lockd/host.c
@@ -15,7 +15,6 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/svc.h>
 #include <linux/lockd/lockd.h>
-#include <linux/lockd/sm_inter.h>
 #include <linux/mutex.h>
 
 #include <net/ipv6.h>
@@ -32,11 +31,6 @@ static int			nrhosts;
 static DEFINE_MUTEX(nlm_host_mutex);
 
 static void			nlm_gc_hosts(void);
-static struct nsm_handle	*nsm_find(const struct sockaddr *sap,
-						const size_t salen,
-						const char *hostname,
-						const size_t hostname_len,
-						const int create);
 
 struct nlm_lookup_host_info {
 	const int		server;		/* search for server|client */
@@ -48,6 +42,7 @@ struct nlm_lookup_host_info {
 	const size_t		hostname_len;	/* it's length */
 	const struct sockaddr	*src_sap;	/* our address (optional) */
 	const size_t		src_len;	/* it's length */
+	const int		noresvport;	/* use non-priv port */
 };
 
 /*
@@ -104,32 +99,6 @@ static void nlm_clear_port(struct sockaddr *sap)
 	}
 }
 
-static void nlm_display_address(const struct sockaddr *sap,
-				char *buf, const size_t len)
-{
-	const struct sockaddr_in *sin = (struct sockaddr_in *)sap;
-	const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap;
-
-	switch (sap->sa_family) {
-	case AF_UNSPEC:
-		snprintf(buf, len, "unspecified");
-		break;
-	case AF_INET:
-		snprintf(buf, len, "%pI4", &sin->sin_addr.s_addr);
-		break;
-	case AF_INET6:
-		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
-			snprintf(buf, len, "%pI4",
-				 &sin6->sin6_addr.s6_addr32[3]);
-		else
-			snprintf(buf, len, "%pI6", &sin6->sin6_addr);
-		break;
-	default:
-		snprintf(buf, len, "unsupported address family");
-		break;
-	}
-}
-
 /*
  * Common host lookup routine for server & client
  */
@@ -189,8 +158,8 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 		atomic_inc(&nsm->sm_count);
 	else {
 		host = NULL;
-		nsm = nsm_find(ni->sap, ni->salen,
-				ni->hostname, ni->hostname_len, 1);
+		nsm = nsm_get_handle(ni->sap, ni->salen,
+					ni->hostname, ni->hostname_len);
 		if (!nsm) {
 			dprintk("lockd: nlm_lookup_host failed; "
 				"no nsm handle\n");
@@ -205,6 +174,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 		goto out;
 	}
 	host->h_name	   = nsm->sm_name;
+	host->h_addrbuf    = nsm->sm_addrbuf;
 	memcpy(nlm_addr(host), ni->sap, ni->salen);
 	host->h_addrlen = ni->salen;
 	nlm_clear_port(nlm_addr(host));
@@ -222,6 +192,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 	host->h_nsmstate   = 0;			/* real NSM state */
 	host->h_nsmhandle  = nsm;
 	host->h_server	   = ni->server;
+	host->h_noresvport = ni->noresvport;
 	hlist_add_head(&host->h_hash, chain);
 	INIT_LIST_HEAD(&host->h_lockowners);
 	spin_lock_init(&host->h_lock);
@@ -230,11 +201,6 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 
 	nrhosts++;
 
-	nlm_display_address((struct sockaddr *)&host->h_addr,
-				host->h_addrbuf, sizeof(host->h_addrbuf));
-	nlm_display_address((struct sockaddr *)&host->h_srcaddr,
-				host->h_srcaddrbuf, sizeof(host->h_srcaddrbuf));
-
 	dprintk("lockd: nlm_lookup_host created host %s\n",
 			host->h_name);
 
@@ -254,10 +220,8 @@ nlm_destroy_host(struct nlm_host *host)
 	BUG_ON(!list_empty(&host->h_lockowners));
 	BUG_ON(atomic_read(&host->h_count));
 
-	/*
-	 * Release NSM handle and unmonitor host.
-	 */
 	nsm_unmonitor(host);
+	nsm_release(host->h_nsmhandle);
 
 	clnt = host->h_rpcclnt;
 	if (clnt != NULL)
@@ -272,6 +236,7 @@ nlm_destroy_host(struct nlm_host *host)
  * @protocol: transport protocol to use
  * @version: NLM protocol version
  * @hostname: '\0'-terminated hostname of server
+ * @noresvport: 1 if non-privileged port should be used
  *
  * Returns an nlm_host structure that matches the passed-in
  * [server address, transport protocol, NLM version, server hostname].
@@ -281,7 +246,9 @@ nlm_destroy_host(struct nlm_host *host)
 struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap,
 				     const size_t salen,
 				     const unsigned short protocol,
-				     const u32 version, const char *hostname)
+				     const u32 version,
+				     const char *hostname,
+				     int noresvport)
 {
 	const struct sockaddr source = {
 		.sa_family	= AF_UNSPEC,
@@ -296,6 +263,7 @@ struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap,
 		.hostname_len	= strlen(hostname),
 		.src_sap	= &source,
 		.src_len	= sizeof(source),
+		.noresvport	= noresvport,
 	};
 
 	dprintk("lockd: %s(host='%s', vers=%u, proto=%s)\n", __func__,
@@ -372,8 +340,8 @@ nlm_bind_host(struct nlm_host *host)
 {
 	struct rpc_clnt	*clnt;
 
-	dprintk("lockd: nlm_bind_host %s (%s), my addr=%s\n",
-			host->h_name, host->h_addrbuf, host->h_srcaddrbuf);
+	dprintk("lockd: nlm_bind_host %s (%s)\n",
+			host->h_name, host->h_addrbuf);
 
 	/* Lock host handle */
 	mutex_lock(&host->h_mutex);
@@ -417,6 +385,8 @@ nlm_bind_host(struct nlm_host *host)
 		 */
 		if (!host->h_server)
 			args.flags |= RPC_CLNT_CREATE_HARDRTRY;
+		if (host->h_noresvport)
+			args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
 
 		clnt = rpc_create(&args);
 		if (!IS_ERR(clnt))
@@ -473,35 +443,23 @@ void nlm_release_host(struct nlm_host *host)
 	}
 }
 
-/*
- * We were notified that the host indicated by address &sin
- * has rebooted.
- * Release all resources held by that peer.
+/**
+ * nlm_host_rebooted - Release all resources held by rebooted host
+ * @info: pointer to decoded results of NLM_SM_NOTIFY call
+ *
+ * We were notified that the specified host has rebooted.  Release
+ * all resources held by that peer.
  */
-void nlm_host_rebooted(const struct sockaddr_in *sin,
-				const char *hostname,
-				unsigned int hostname_len,
-				u32 new_state)
+void nlm_host_rebooted(const struct nlm_reboot *info)
 {
 	struct hlist_head *chain;
 	struct hlist_node *pos;
 	struct nsm_handle *nsm;
 	struct nlm_host	*host;
 
-	nsm = nsm_find((struct sockaddr *)sin, sizeof(*sin),
-			hostname, hostname_len, 0);
-	if (nsm == NULL) {
-		dprintk("lockd: never saw rebooted peer '%.*s' before\n",
-				hostname_len, hostname);
+	nsm = nsm_reboot_lookup(info);
+	if (unlikely(nsm == NULL))
 		return;
-	}
-
-	dprintk("lockd: nlm_host_rebooted(%.*s, %s)\n",
-			hostname_len, hostname, nsm->sm_addrbuf);
-
-	/* When reclaiming locks on this peer, make sure that
-	 * we set up a new notification */
-	nsm->sm_monitored = 0;
 
 	/* Mark all hosts tied to this NSM state as having rebooted.
 	 * We run the loop repeatedly, because we drop the host table
@@ -512,8 +470,8 @@ again:	mutex_lock(&nlm_host_mutex);
 	for (chain = nlm_hosts; chain < nlm_hosts + NLM_HOST_NRHASH; ++chain) {
 		hlist_for_each_entry(host, pos, chain, h_hash) {
 			if (host->h_nsmhandle == nsm
-			 && host->h_nsmstate != new_state) {
-				host->h_nsmstate = new_state;
+			 && host->h_nsmstate != info->state) {
+				host->h_nsmstate = info->state;
 				host->h_state++;
 
 				nlm_get_host(host);
@@ -621,89 +579,3 @@ nlm_gc_hosts(void)
 
 	next_gc = jiffies + NLM_HOST_COLLECT;
 }
-
-
-/*
- * Manage NSM handles
- */
-static LIST_HEAD(nsm_handles);
-static DEFINE_SPINLOCK(nsm_lock);
-
-static struct nsm_handle *nsm_find(const struct sockaddr *sap,
-				   const size_t salen,
-				   const char *hostname,
-				   const size_t hostname_len,
-				   const int create)
-{
-	struct nsm_handle *nsm = NULL;
-	struct nsm_handle *pos;
-
-	if (!sap)
-		return NULL;
-
-	if (hostname && memchr(hostname, '/', hostname_len) != NULL) {
-		if (printk_ratelimit()) {
-			printk(KERN_WARNING "Invalid hostname \"%.*s\" "
-					    "in NFS lock request\n",
-				(int)hostname_len, hostname);
-		}
-		return NULL;
-	}
-
-retry:
-	spin_lock(&nsm_lock);
-	list_for_each_entry(pos, &nsm_handles, sm_link) {
-
-		if (hostname && nsm_use_hostnames) {
-			if (strlen(pos->sm_name) != hostname_len
-			 || memcmp(pos->sm_name, hostname, hostname_len))
-				continue;
-		} else if (!nlm_cmp_addr(nsm_addr(pos), sap))
-			continue;
-		atomic_inc(&pos->sm_count);
-		kfree(nsm);
-		nsm = pos;
-		goto found;
-	}
-	if (nsm) {
-		list_add(&nsm->sm_link, &nsm_handles);
-		goto found;
-	}
-	spin_unlock(&nsm_lock);
-
-	if (!create)
-		return NULL;
-
-	nsm = kzalloc(sizeof(*nsm) + hostname_len + 1, GFP_KERNEL);
-	if (nsm == NULL)
-		return NULL;
-
-	memcpy(nsm_addr(nsm), sap, salen);
-	nsm->sm_addrlen = salen;
-	nsm->sm_name = (char *) (nsm + 1);
-	memcpy(nsm->sm_name, hostname, hostname_len);
-	nsm->sm_name[hostname_len] = '\0';
-	nlm_display_address((struct sockaddr *)&nsm->sm_addr,
-				nsm->sm_addrbuf, sizeof(nsm->sm_addrbuf));
-	atomic_set(&nsm->sm_count, 1);
-	goto retry;
-
-found:
-	spin_unlock(&nsm_lock);
-	return nsm;
-}
-
-/*
- * Release an NSM handle
- */
-void
-nsm_release(struct nsm_handle *nsm)
-{
-	if (!nsm)
-		return;
-	if (atomic_dec_and_lock(&nsm->sm_count, &nsm_lock)) {
-		list_del(&nsm->sm_link);
-		spin_unlock(&nsm_lock);
-		kfree(nsm);
-	}
-}
diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c
index ffd3461f75e..5e2c4d5ac82 100644
--- a/fs/lockd/mon.c
+++ b/fs/lockd/mon.c
@@ -9,35 +9,123 @@
 #include <linux/types.h>
 #include <linux/utsname.h>
 #include <linux/kernel.h>
+#include <linux/ktime.h>
+
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/xprtsock.h>
 #include <linux/sunrpc/svc.h>
 #include <linux/lockd/lockd.h>
-#include <linux/lockd/sm_inter.h>
-
 
 #define NLMDBG_FACILITY		NLMDBG_MONITOR
+#define NSM_PROGRAM		100024
+#define NSM_VERSION		1
+
+enum {
+	NSMPROC_NULL,
+	NSMPROC_STAT,
+	NSMPROC_MON,
+	NSMPROC_UNMON,
+	NSMPROC_UNMON_ALL,
+	NSMPROC_SIMU_CRASH,
+	NSMPROC_NOTIFY,
+};
+
+struct nsm_args {
+	struct nsm_private	*priv;
+	u32			prog;		/* RPC callback info */
+	u32			vers;
+	u32			proc;
 
-#define XDR_ADDRBUF_LEN		(20)
+	char			*mon_name;
+};
 
-static struct rpc_clnt *	nsm_create(void);
+struct nsm_res {
+	u32			status;
+	u32			state;
+};
 
 static struct rpc_program	nsm_program;
+static				LIST_HEAD(nsm_handles);
+static				DEFINE_SPINLOCK(nsm_lock);
 
 /*
  * Local NSM state
  */
-int				nsm_local_state;
+int	__read_mostly		nsm_local_state;
+int	__read_mostly		nsm_use_hostnames;
 
-/*
- * Common procedure for SM_MON/SM_UNMON calls
- */
-static int
-nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res)
+static inline struct sockaddr *nsm_addr(const struct nsm_handle *nsm)
+{
+	return (struct sockaddr *)&nsm->sm_addr;
+}
+
+static void nsm_display_ipv4_address(const struct sockaddr *sap, char *buf,
+				     const size_t len)
+{
+	const struct sockaddr_in *sin = (struct sockaddr_in *)sap;
+	snprintf(buf, len, "%pI4", &sin->sin_addr.s_addr);
+}
+
+static void nsm_display_ipv6_address(const struct sockaddr *sap, char *buf,
+				     const size_t len)
+{
+	const struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap;
+
+	if (ipv6_addr_v4mapped(&sin6->sin6_addr))
+		snprintf(buf, len, "%pI4", &sin6->sin6_addr.s6_addr32[3]);
+	else if (sin6->sin6_scope_id != 0)
+		snprintf(buf, len, "%pI6%%%u", &sin6->sin6_addr,
+				sin6->sin6_scope_id);
+	else
+		snprintf(buf, len, "%pI6", &sin6->sin6_addr);
+}
+
+static void nsm_display_address(const struct sockaddr *sap,
+				char *buf, const size_t len)
+{
+	switch (sap->sa_family) {
+	case AF_INET:
+		nsm_display_ipv4_address(sap, buf, len);
+		break;
+	case AF_INET6:
+		nsm_display_ipv6_address(sap, buf, len);
+		break;
+	default:
+		snprintf(buf, len, "unsupported address family");
+		break;
+	}
+}
+
+static struct rpc_clnt *nsm_create(void)
+{
+	struct sockaddr_in sin = {
+		.sin_family		= AF_INET,
+		.sin_addr.s_addr	= htonl(INADDR_LOOPBACK),
+	};
+	struct rpc_create_args args = {
+		.protocol		= XPRT_TRANSPORT_UDP,
+		.address		= (struct sockaddr *)&sin,
+		.addrsize		= sizeof(sin),
+		.servername		= "rpc.statd",
+		.program		= &nsm_program,
+		.version		= NSM_VERSION,
+		.authflavor		= RPC_AUTH_NULL,
+	};
+
+	return rpc_create(&args);
+}
+
+static int nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res)
 {
 	struct rpc_clnt	*clnt;
 	int		status;
-	struct nsm_args	args;
+	struct nsm_args args = {
+		.priv		= &nsm->sm_priv,
+		.prog		= NLM_PROGRAM,
+		.vers		= 3,
+		.proc		= NLMPROC_NSM_NOTIFY,
+		.mon_name	= nsm->sm_mon_name,
+	};
 	struct rpc_message msg = {
 		.rpc_argp	= &args,
 		.rpc_resp	= res,
@@ -46,22 +134,18 @@ nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res)
 	clnt = nsm_create();
 	if (IS_ERR(clnt)) {
 		status = PTR_ERR(clnt);
+		dprintk("lockd: failed to create NSM upcall transport, "
+				"status=%d\n", status);
 		goto out;
 	}
 
-	memset(&args, 0, sizeof(args));
-	args.mon_name = nsm->sm_name;
-	args.addr = nsm_addr_in(nsm)->sin_addr.s_addr;
-	args.prog = NLM_PROGRAM;
-	args.vers = 3;
-	args.proc = NLMPROC_NSM_NOTIFY;
 	memset(res, 0, sizeof(*res));
 
 	msg.rpc_proc = &clnt->cl_procinfo[proc];
 	status = rpc_call_sync(clnt, &msg, 0);
 	if (status < 0)
-		printk(KERN_DEBUG "nsm_mon_unmon: rpc failed, status=%d\n",
-			status);
+		dprintk("lockd: NSM upcall RPC failed, status=%d\n",
+				status);
 	else
 		status = 0;
 	rpc_shutdown_client(clnt);
@@ -69,82 +153,272 @@ nsm_mon_unmon(struct nsm_handle *nsm, u32 proc, struct nsm_res *res)
 	return status;
 }
 
-/*
- * Set up monitoring of a remote host
+/**
+ * nsm_monitor - Notify a peer in case we reboot
+ * @host: pointer to nlm_host of peer to notify
+ *
+ * If this peer is not already monitored, this function sends an
+ * upcall to the local rpc.statd to record the name/address of
+ * the peer to notify in case we reboot.
+ *
+ * Returns zero if the peer is monitored by the local rpc.statd;
+ * otherwise a negative errno value is returned.
  */
-int
-nsm_monitor(struct nlm_host *host)
+int nsm_monitor(const struct nlm_host *host)
 {
 	struct nsm_handle *nsm = host->h_nsmhandle;
 	struct nsm_res	res;
 	int		status;
 
-	dprintk("lockd: nsm_monitor(%s)\n", host->h_name);
-	BUG_ON(nsm == NULL);
+	dprintk("lockd: nsm_monitor(%s)\n", nsm->sm_name);
 
 	if (nsm->sm_monitored)
 		return 0;
 
-	status = nsm_mon_unmon(nsm, SM_MON, &res);
+	/*
+	 * Choose whether to record the caller_name or IP address of
+	 * this peer in the local rpc.statd's database.
+	 */
+	nsm->sm_mon_name = nsm_use_hostnames ? nsm->sm_name : nsm->sm_addrbuf;
 
-	if (status < 0 || res.status != 0)
-		printk(KERN_NOTICE "lockd: cannot monitor %s\n", host->h_name);
+	status = nsm_mon_unmon(nsm, NSMPROC_MON, &res);
+	if (res.status != 0)
+		status = -EIO;
+	if (status < 0)
+		printk(KERN_NOTICE "lockd: cannot monitor %s\n", nsm->sm_name);
 	else
 		nsm->sm_monitored = 1;
 	return status;
 }
 
-/*
- * Cease to monitor remote host
+/**
+ * nsm_unmonitor - Unregister peer notification
+ * @host: pointer to nlm_host of peer to stop monitoring
+ *
+ * If this peer is monitored, this function sends an upcall to
+ * tell the local rpc.statd not to send this peer a notification
+ * when we reboot.
  */
-int
-nsm_unmonitor(struct nlm_host *host)
+void nsm_unmonitor(const struct nlm_host *host)
 {
 	struct nsm_handle *nsm = host->h_nsmhandle;
 	struct nsm_res	res;
-	int		status = 0;
-
-	if (nsm == NULL)
-		return 0;
-	host->h_nsmhandle = NULL;
+	int status;
 
 	if (atomic_read(&nsm->sm_count) == 1
 	 && nsm->sm_monitored && !nsm->sm_sticky) {
-		dprintk("lockd: nsm_unmonitor(%s)\n", host->h_name);
+		dprintk("lockd: nsm_unmonitor(%s)\n", nsm->sm_name);
 
-		status = nsm_mon_unmon(nsm, SM_UNMON, &res);
+		status = nsm_mon_unmon(nsm, NSMPROC_UNMON, &res);
+		if (res.status != 0)
+			status = -EIO;
 		if (status < 0)
 			printk(KERN_NOTICE "lockd: cannot unmonitor %s\n",
-					host->h_name);
+					nsm->sm_name);
 		else
 			nsm->sm_monitored = 0;
 	}
-	nsm_release(nsm);
-	return status;
+}
+
+static struct nsm_handle *nsm_lookup_hostname(const char *hostname,
+					      const size_t len)
+{
+	struct nsm_handle *nsm;
+
+	list_for_each_entry(nsm, &nsm_handles, sm_link)
+		if (strlen(nsm->sm_name) == len &&
+		    memcmp(nsm->sm_name, hostname, len) == 0)
+			return nsm;
+	return NULL;
+}
+
+static struct nsm_handle *nsm_lookup_addr(const struct sockaddr *sap)
+{
+	struct nsm_handle *nsm;
+
+	list_for_each_entry(nsm, &nsm_handles, sm_link)
+		if (nlm_cmp_addr(nsm_addr(nsm), sap))
+			return nsm;
+	return NULL;
+}
+
+static struct nsm_handle *nsm_lookup_priv(const struct nsm_private *priv)
+{
+	struct nsm_handle *nsm;
+
+	list_for_each_entry(nsm, &nsm_handles, sm_link)
+		if (memcmp(nsm->sm_priv.data, priv->data,
+					sizeof(priv->data)) == 0)
+			return nsm;
+	return NULL;
 }
 
 /*
- * Create NSM client for the local host
+ * Construct a unique cookie to match this nsm_handle to this monitored
+ * host.  It is passed to the local rpc.statd via NSMPROC_MON, and
+ * returned via NLMPROC_SM_NOTIFY, in the "priv" field of these
+ * requests.
+ *
+ * The NSM protocol requires that these cookies be unique while the
+ * system is running.  We prefer a stronger requirement of making them
+ * unique across reboots.  If user space bugs cause a stale cookie to
+ * be sent to the kernel, it could cause the wrong host to lose its
+ * lock state if cookies were not unique across reboots.
+ *
+ * The cookies are exposed only to local user space via loopback.  They
+ * do not appear on the physical network.  If we want greater security
+ * for some reason, nsm_init_private() could perform a one-way hash to
+ * obscure the contents of the cookie.
  */
-static struct rpc_clnt *
-nsm_create(void)
+static void nsm_init_private(struct nsm_handle *nsm)
 {
-	struct sockaddr_in	sin = {
-		.sin_family	= AF_INET,
-		.sin_addr.s_addr = htonl(INADDR_LOOPBACK),
-		.sin_port	= 0,
-	};
-	struct rpc_create_args args = {
-		.protocol	= XPRT_TRANSPORT_UDP,
-		.address	= (struct sockaddr *)&sin,
-		.addrsize	= sizeof(sin),
-		.servername	= "localhost",
-		.program	= &nsm_program,
-		.version	= SM_VERSION,
-		.authflavor	= RPC_AUTH_NULL,
-	};
+	u64 *p = (u64 *)&nsm->sm_priv.data;
+	struct timespec ts;
 
-	return rpc_create(&args);
+	ktime_get_ts(&ts);
+	*p++ = timespec_to_ns(&ts);
+	*p = (unsigned long)nsm;
+}
+
+static struct nsm_handle *nsm_create_handle(const struct sockaddr *sap,
+					    const size_t salen,
+					    const char *hostname,
+					    const size_t hostname_len)
+{
+	struct nsm_handle *new;
+
+	new = kzalloc(sizeof(*new) + hostname_len + 1, GFP_KERNEL);
+	if (unlikely(new == NULL))
+		return NULL;
+
+	atomic_set(&new->sm_count, 1);
+	new->sm_name = (char *)(new + 1);
+	memcpy(nsm_addr(new), sap, salen);
+	new->sm_addrlen = salen;
+	nsm_init_private(new);
+	nsm_display_address((const struct sockaddr *)&new->sm_addr,
+				new->sm_addrbuf, sizeof(new->sm_addrbuf));
+	memcpy(new->sm_name, hostname, hostname_len);
+	new->sm_name[hostname_len] = '\0';
+
+	return new;
+}
+
+/**
+ * nsm_get_handle - Find or create a cached nsm_handle
+ * @sap: pointer to socket address of handle to find
+ * @salen: length of socket address
+ * @hostname: pointer to C string containing hostname to find
+ * @hostname_len: length of C string
+ *
+ * Behavior is modulated by the global nsm_use_hostnames variable.
+ *
+ * Returns a cached nsm_handle after bumping its ref count, or
+ * returns a fresh nsm_handle if a handle that matches @sap and/or
+ * @hostname cannot be found in the handle cache.  Returns NULL if
+ * an error occurs.
+ */
+struct nsm_handle *nsm_get_handle(const struct sockaddr *sap,
+				  const size_t salen, const char *hostname,
+				  const size_t hostname_len)
+{
+	struct nsm_handle *cached, *new = NULL;
+
+	if (hostname && memchr(hostname, '/', hostname_len) != NULL) {
+		if (printk_ratelimit()) {
+			printk(KERN_WARNING "Invalid hostname \"%.*s\" "
+					    "in NFS lock request\n",
+				(int)hostname_len, hostname);
+		}
+		return NULL;
+	}
+
+retry:
+	spin_lock(&nsm_lock);
+
+	if (nsm_use_hostnames && hostname != NULL)
+		cached = nsm_lookup_hostname(hostname, hostname_len);
+	else
+		cached = nsm_lookup_addr(sap);
+
+	if (cached != NULL) {
+		atomic_inc(&cached->sm_count);
+		spin_unlock(&nsm_lock);
+		kfree(new);
+		dprintk("lockd: found nsm_handle for %s (%s), "
+				"cnt %d\n", cached->sm_name,
+				cached->sm_addrbuf,
+				atomic_read(&cached->sm_count));
+		return cached;
+	}
+
+	if (new != NULL) {
+		list_add(&new->sm_link, &nsm_handles);
+		spin_unlock(&nsm_lock);
+		dprintk("lockd: created nsm_handle for %s (%s)\n",
+				new->sm_name, new->sm_addrbuf);
+		return new;
+	}
+
+	spin_unlock(&nsm_lock);
+
+	new = nsm_create_handle(sap, salen, hostname, hostname_len);
+	if (unlikely(new == NULL))
+		return NULL;
+	goto retry;
+}
+
+/**
+ * nsm_reboot_lookup - match NLMPROC_SM_NOTIFY arguments to an nsm_handle
+ * @info: pointer to NLMPROC_SM_NOTIFY arguments
+ *
+ * Returns a matching nsm_handle if found in the nsm cache; the returned
+ * nsm_handle's reference count is bumped and sm_monitored is cleared.
+ * Otherwise returns NULL if some error occurred.
+ */
+struct nsm_handle *nsm_reboot_lookup(const struct nlm_reboot *info)
+{
+	struct nsm_handle *cached;
+
+	spin_lock(&nsm_lock);
+
+	cached = nsm_lookup_priv(&info->priv);
+	if (unlikely(cached == NULL)) {
+		spin_unlock(&nsm_lock);
+		dprintk("lockd: never saw rebooted peer '%.*s' before\n",
+				info->len, info->mon);
+		return cached;
+	}
+
+	atomic_inc(&cached->sm_count);
+	spin_unlock(&nsm_lock);
+
+	/*
+	 * During subsequent lock activity, force a fresh
+	 * notification to be set up for this host.
+	 */
+	cached->sm_monitored = 0;
+
+	dprintk("lockd: host %s (%s) rebooted, cnt %d\n",
+			cached->sm_name, cached->sm_addrbuf,
+			atomic_read(&cached->sm_count));
+	return cached;
+}
+
+/**
+ * nsm_release - Release an NSM handle
+ * @nsm: pointer to handle to be released
+ *
+ */
+void nsm_release(struct nsm_handle *nsm)
+{
+	if (atomic_dec_and_lock(&nsm->sm_count, &nsm_lock)) {
+		list_del(&nsm->sm_link);
+		spin_unlock(&nsm_lock);
+		dprintk("lockd: destroyed nsm_handle for %s (%s)\n",
+				nsm->sm_name, nsm->sm_addrbuf);
+		kfree(nsm);
+	}
 }
 
 /*
@@ -154,127 +428,132 @@ nsm_create(void)
  * Status Monitor wire protocol.
  */
 
-static __be32 *xdr_encode_nsm_string(__be32 *p, char *string)
+static int encode_nsm_string(struct xdr_stream *xdr, const char *string)
 {
-	size_t len = strlen(string);
-
-	if (len > SM_MAXSTRLEN)
-		len = SM_MAXSTRLEN;
-	return xdr_encode_opaque(p, string, len);
+	const u32 len = strlen(string);
+	__be32 *p;
+
+	if (unlikely(len > SM_MAXSTRLEN))
+		return -EIO;
+	p = xdr_reserve_space(xdr, sizeof(u32) + len);
+	if (unlikely(p == NULL))
+		return -EIO;
+	xdr_encode_opaque(p, string, len);
+	return 0;
 }
 
 /*
  * "mon_name" specifies the host to be monitored.
- *
- * Linux uses a text version of the IP address of the remote
- * host as the host identifier (the "mon_name" argument).
- *
- * Linux statd always looks up the canonical hostname first for
- * whatever remote hostname it receives, so this works alright.
  */
-static __be32 *xdr_encode_mon_name(__be32 *p, struct nsm_args *argp)
+static int encode_mon_name(struct xdr_stream *xdr, const struct nsm_args *argp)
 {
-	char	buffer[XDR_ADDRBUF_LEN + 1];
-	char	*name = argp->mon_name;
-
-	if (!nsm_use_hostnames) {
-		snprintf(buffer, XDR_ADDRBUF_LEN,
-			 "%pI4", &argp->addr);
-		name = buffer;
-	}
-
-	return xdr_encode_nsm_string(p, name);
+	return encode_nsm_string(xdr, argp->mon_name);
 }
 
 /*
  * The "my_id" argument specifies the hostname and RPC procedure
  * to be called when the status manager receives notification
- * (via the SM_NOTIFY call) that the state of host "mon_name"
+ * (via the NLMPROC_SM_NOTIFY call) that the state of host "mon_name"
  * has changed.
  */
-static __be32 *xdr_encode_my_id(__be32 *p, struct nsm_args *argp)
+static int encode_my_id(struct xdr_stream *xdr, const struct nsm_args *argp)
 {
-	p = xdr_encode_nsm_string(p, utsname()->nodename);
-	if (!p)
-		return ERR_PTR(-EIO);
-
+	int status;
+	__be32 *p;
+
+	status = encode_nsm_string(xdr, utsname()->nodename);
+	if (unlikely(status != 0))
+		return status;
+	p = xdr_reserve_space(xdr, 3 * sizeof(u32));
+	if (unlikely(p == NULL))
+		return -EIO;
 	*p++ = htonl(argp->prog);
 	*p++ = htonl(argp->vers);
 	*p++ = htonl(argp->proc);
-
-	return p;
+	return 0;
 }
 
 /*
  * The "mon_id" argument specifies the non-private arguments
- * of an SM_MON or SM_UNMON call.
+ * of an NSMPROC_MON or NSMPROC_UNMON call.
  */
-static __be32 *xdr_encode_mon_id(__be32 *p, struct nsm_args *argp)
+static int encode_mon_id(struct xdr_stream *xdr, const struct nsm_args *argp)
 {
-	p = xdr_encode_mon_name(p, argp);
-	if (!p)
-		return ERR_PTR(-EIO);
+	int status;
 
-	return xdr_encode_my_id(p, argp);
+	status = encode_mon_name(xdr, argp);
+	if (unlikely(status != 0))
+		return status;
+	return encode_my_id(xdr, argp);
 }
 
 /*
  * The "priv" argument may contain private information required
- * by the SM_MON call. This information will be supplied in the
- * SM_NOTIFY call.
- *
- * Linux provides the raw IP address of the monitored host,
- * left in network byte order.
+ * by the NSMPROC_MON call. This information will be supplied in the
+ * NLMPROC_SM_NOTIFY call.
  */
-static __be32 *xdr_encode_priv(__be32 *p, struct nsm_args *argp)
+static int encode_priv(struct xdr_stream *xdr, const struct nsm_args *argp)
 {
-	*p++ = argp->addr;
-	*p++ = 0;
-	*p++ = 0;
-	*p++ = 0;
+	__be32 *p;
 
-	return p;
+	p = xdr_reserve_space(xdr, SM_PRIV_SIZE);
+	if (unlikely(p == NULL))
+		return -EIO;
+	xdr_encode_opaque_fixed(p, argp->priv->data, SM_PRIV_SIZE);
+	return 0;
 }
 
-static int
-xdr_encode_mon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp)
+static int xdr_enc_mon(struct rpc_rqst *req, __be32 *p,
+		       const struct nsm_args *argp)
 {
-	p = xdr_encode_mon_id(p, argp);
-	if (IS_ERR(p))
-		return PTR_ERR(p);
-
-	p = xdr_encode_priv(p, argp);
-	if (IS_ERR(p))
-		return PTR_ERR(p);
-
-	rqstp->rq_slen = xdr_adjust_iovec(rqstp->rq_svec, p);
-	return 0;
+	struct xdr_stream xdr;
+	int status;
+
+	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+	status = encode_mon_id(&xdr, argp);
+	if (unlikely(status))
+		return status;
+	return encode_priv(&xdr, argp);
 }
 
-static int
-xdr_encode_unmon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp)
+static int xdr_enc_unmon(struct rpc_rqst *req, __be32 *p,
+			 const struct nsm_args *argp)
 {
-	p = xdr_encode_mon_id(p, argp);
-	if (IS_ERR(p))
-		return PTR_ERR(p);
-	rqstp->rq_slen = xdr_adjust_iovec(rqstp->rq_svec, p);
-	return 0;
+	struct xdr_stream xdr;
+
+	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+	return encode_mon_id(&xdr, argp);
 }
 
-static int
-xdr_decode_stat_res(struct rpc_rqst *rqstp, __be32 *p, struct nsm_res *resp)
+static int xdr_dec_stat_res(struct rpc_rqst *rqstp, __be32 *p,
+			    struct nsm_res *resp)
 {
+	struct xdr_stream xdr;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	p = xdr_inline_decode(&xdr, 2 * sizeof(u32));
+	if (unlikely(p == NULL))
+		return -EIO;
 	resp->status = ntohl(*p++);
-	resp->state = ntohl(*p++);
-	dprintk("nsm: xdr_decode_stat_res status %d state %d\n",
+	resp->state = ntohl(*p);
+
+	dprintk("lockd: xdr_dec_stat_res status %d state %d\n",
 			resp->status, resp->state);
 	return 0;
 }
 
-static int
-xdr_decode_stat(struct rpc_rqst *rqstp, __be32 *p, struct nsm_res *resp)
+static int xdr_dec_stat(struct rpc_rqst *rqstp, __be32 *p,
+			struct nsm_res *resp)
 {
-	resp->state = ntohl(*p++);
+	struct xdr_stream xdr;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	p = xdr_inline_decode(&xdr, sizeof(u32));
+	if (unlikely(p == NULL))
+		return -EIO;
+	resp->state = ntohl(*p);
+
+	dprintk("lockd: xdr_dec_stat state %d\n", resp->state);
 	return 0;
 }
 
@@ -288,22 +567,22 @@ xdr_decode_stat(struct rpc_rqst *rqstp, __be32 *p, struct nsm_res *resp)
 #define SM_unmonres_sz	1
 
 static struct rpc_procinfo	nsm_procedures[] = {
-[SM_MON] = {
-		.p_proc		= SM_MON,
-		.p_encode	= (kxdrproc_t) xdr_encode_mon,
-		.p_decode	= (kxdrproc_t) xdr_decode_stat_res,
+[NSMPROC_MON] = {
+		.p_proc		= NSMPROC_MON,
+		.p_encode	= (kxdrproc_t)xdr_enc_mon,
+		.p_decode	= (kxdrproc_t)xdr_dec_stat_res,
 		.p_arglen	= SM_mon_sz,
 		.p_replen	= SM_monres_sz,
-		.p_statidx	= SM_MON,
+		.p_statidx	= NSMPROC_MON,
 		.p_name		= "MONITOR",
 	},
-[SM_UNMON] = {
-		.p_proc		= SM_UNMON,
-		.p_encode	= (kxdrproc_t) xdr_encode_unmon,
-		.p_decode	= (kxdrproc_t) xdr_decode_stat,
+[NSMPROC_UNMON] = {
+		.p_proc		= NSMPROC_UNMON,
+		.p_encode	= (kxdrproc_t)xdr_enc_unmon,
+		.p_decode	= (kxdrproc_t)xdr_dec_stat,
 		.p_arglen	= SM_mon_id_sz,
 		.p_replen	= SM_unmonres_sz,
-		.p_statidx	= SM_UNMON,
+		.p_statidx	= NSMPROC_UNMON,
 		.p_name		= "UNMONITOR",
 	},
 };
@@ -322,7 +601,7 @@ static struct rpc_stat		nsm_stats;
 
 static struct rpc_program	nsm_program = {
 		.name		= "statd",
-		.number		= SM_PROGRAM,
+		.number		= NSM_PROGRAM,
 		.nrvers		= ARRAY_SIZE(nsm_version),
 		.version	= nsm_version,
 		.stats		= &nsm_stats
diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 56b076736b5..64f1c31b585 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -35,7 +35,6 @@
 #include <linux/sunrpc/svcsock.h>
 #include <net/ip.h>
 #include <linux/lockd/lockd.h>
-#include <linux/lockd/sm_inter.h>
 #include <linux/nfs.h>
 
 #define NLMDBG_FACILITY		NLMDBG_SVC
@@ -45,7 +44,7 @@
 static struct svc_program	nlmsvc_program;
 
 struct nlmsvc_binding *		nlmsvc_ops;
-EXPORT_SYMBOL(nlmsvc_ops);
+EXPORT_SYMBOL_GPL(nlmsvc_ops);
 
 static DEFINE_MUTEX(nlmsvc_mutex);
 static unsigned int		nlmsvc_users;
@@ -54,13 +53,26 @@ static struct svc_rqst		*nlmsvc_rqst;
 unsigned long			nlmsvc_timeout;
 
 /*
+ * If the kernel has IPv6 support available, always listen for
+ * both AF_INET and AF_INET6 requests.
+ */
+#if (defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)) && \
+	defined(CONFIG_SUNRPC_REGISTER_V4)
+static const sa_family_t	nlmsvc_family = AF_INET6;
+#else	/* (CONFIG_IPV6 || CONFIG_IPV6_MODULE) && CONFIG_SUNRPC_REGISTER_V4 */
+static const sa_family_t	nlmsvc_family = AF_INET;
+#endif	/* (CONFIG_IPV6 || CONFIG_IPV6_MODULE) && CONFIG_SUNRPC_REGISTER_V4 */
+
+/*
  * These can be set at insmod time (useful for NFS as root filesystem),
  * and also changed through the sysctl interface.  -- Jamie Lokier, Aug 2003
  */
 static unsigned long		nlm_grace_period;
 static unsigned long		nlm_timeout = LOCKD_DFLT_TIMEO;
 static int			nlm_udpport, nlm_tcpport;
-int				nsm_use_hostnames = 0;
+
+/* RLIM_NOFILE defaults to 1024. That seems like a reasonable default here. */
+static unsigned int		nlm_max_connections = 1024;
 
 /*
  * Constants needed for the sysctl interface.
@@ -143,6 +155,9 @@ lockd(void *vrqstp)
 		long timeout = MAX_SCHEDULE_TIMEOUT;
 		RPC_IFDEBUG(char buf[RPC_MAX_ADDRBUFLEN]);
 
+		/* update sv_maxconn if it has changed */
+		rqstp->rq_server->sv_maxconn = nlm_max_connections;
+
 		if (signalled()) {
 			flush_signals(current);
 			if (nlmsvc_ops) {
@@ -189,6 +204,19 @@ lockd(void *vrqstp)
 	return 0;
 }
 
+static int create_lockd_listener(struct svc_serv *serv, char *name,
+				 unsigned short port)
+{
+	struct svc_xprt *xprt;
+
+	xprt = svc_find_xprt(serv, name, 0, 0);
+	if (xprt == NULL)
+		return svc_create_xprt(serv, name, port, SVC_SOCK_DEFAULTS);
+
+	svc_xprt_put(xprt);
+	return 0;
+}
+
 /*
  * Ensure there are active UDP and TCP listeners for lockd.
  *
@@ -202,29 +230,23 @@ lockd(void *vrqstp)
 static int make_socks(struct svc_serv *serv)
 {
 	static int warned;
-	struct svc_xprt *xprt;
-	int err = 0;
+	int err;
 
-	xprt = svc_find_xprt(serv, "udp", 0, 0);
-	if (!xprt)
-		err = svc_create_xprt(serv, "udp", nlm_udpport,
-				      SVC_SOCK_DEFAULTS);
-	else
-		svc_xprt_put(xprt);
-	if (err >= 0) {
-		xprt = svc_find_xprt(serv, "tcp", 0, 0);
-		if (!xprt)
-			err = svc_create_xprt(serv, "tcp", nlm_tcpport,
-					      SVC_SOCK_DEFAULTS);
-		else
-			svc_xprt_put(xprt);
-	}
-	if (err >= 0) {
-		warned = 0;
-		err = 0;
-	} else if (warned++ == 0)
+	err = create_lockd_listener(serv, "udp", nlm_udpport);
+	if (err < 0)
+		goto out_err;
+
+	err = create_lockd_listener(serv, "tcp", nlm_tcpport);
+	if (err < 0)
+		goto out_err;
+
+	warned = 0;
+	return 0;
+
+out_err:
+	if (warned++ == 0)
 		printk(KERN_WARNING
-		       "lockd_up: makesock failed, error=%d\n", err);
+			"lockd_up: makesock failed, error=%d\n", err);
 	return err;
 }
 
@@ -252,7 +274,7 @@ int lockd_up(void)
 			"lockd_up: no pid, %d users??\n", nlmsvc_users);
 
 	error = -ENOMEM;
-	serv = svc_create(&nlmsvc_program, LOCKD_BUFSIZE, AF_INET, NULL);
+	serv = svc_create(&nlmsvc_program, LOCKD_BUFSIZE, nlmsvc_family, NULL);
 	if (!serv) {
 		printk(KERN_WARNING "lockd_up: create service failed\n");
 		goto out;
@@ -276,6 +298,7 @@ int lockd_up(void)
 	}
 
 	svc_sock_update_bufs(serv);
+	serv->sv_maxconn = nlm_max_connections;
 
 	nlmsvc_task = kthread_run(lockd, nlmsvc_rqst, serv->sv_name);
 	if (IS_ERR(nlmsvc_task)) {
@@ -300,7 +323,7 @@ out:
 	mutex_unlock(&nlmsvc_mutex);
 	return error;
 }
-EXPORT_SYMBOL(lockd_up);
+EXPORT_SYMBOL_GPL(lockd_up);
 
 /*
  * Decrement the user count and bring down lockd if we're the last.
@@ -329,7 +352,7 @@ lockd_down(void)
 out:
 	mutex_unlock(&nlmsvc_mutex);
 }
-EXPORT_SYMBOL(lockd_down);
+EXPORT_SYMBOL_GPL(lockd_down);
 
 #ifdef CONFIG_SYSCTL
 
@@ -485,6 +508,7 @@ module_param_call(nlm_udpport, param_set_port, param_get_int,
 module_param_call(nlm_tcpport, param_set_port, param_get_int,
 		  &nlm_tcpport, 0644);
 module_param(nsm_use_hostnames, bool, 0644);
+module_param(nlm_max_connections, uint, 0644);
 
 /*
  * Initialising and terminating the module.
diff --git a/fs/lockd/svc4proc.c b/fs/lockd/svc4proc.c
index 4dfdcbc6bf6..1725037374c 100644
--- a/fs/lockd/svc4proc.c
+++ b/fs/lockd/svc4proc.c
@@ -16,8 +16,6 @@
 #include <linux/nfsd/nfsd.h>
 #include <linux/lockd/lockd.h>
 #include <linux/lockd/share.h>
-#include <linux/lockd/sm_inter.h>
-
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 
@@ -419,8 +417,6 @@ static __be32
 nlm4svc_proc_sm_notify(struct svc_rqst *rqstp, struct nlm_reboot *argp,
 					      void	        *resp)
 {
-	struct sockaddr_in	saddr;
-
 	dprintk("lockd: SM_NOTIFY     called\n");
 
 	if (!nlm_privileged_requester(rqstp)) {
@@ -430,14 +426,7 @@ nlm4svc_proc_sm_notify(struct svc_rqst *rqstp, struct nlm_reboot *argp,
 		return rpc_system_err;
 	}
 
-	/* Obtain the host pointer for this NFS server and try to
-	 * reclaim all locks we hold on this server.
-	 */
-	memset(&saddr, 0, sizeof(saddr));
-	saddr.sin_family = AF_INET;
-	saddr.sin_addr.s_addr = argp->addr;
-	nlm_host_rebooted(&saddr, argp->mon, argp->len, argp->state);
-
+	nlm_host_rebooted(argp);
 	return rpc_success;
 }
 
diff --git a/fs/lockd/svcproc.c b/fs/lockd/svcproc.c
index 3ca89e2a938..3688e55901f 100644
--- a/fs/lockd/svcproc.c
+++ b/fs/lockd/svcproc.c
@@ -16,8 +16,6 @@
 #include <linux/nfsd/nfsd.h>
 #include <linux/lockd/lockd.h>
 #include <linux/lockd/share.h>
-#include <linux/lockd/sm_inter.h>
-
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 
@@ -451,8 +449,6 @@ static __be32
 nlmsvc_proc_sm_notify(struct svc_rqst *rqstp, struct nlm_reboot *argp,
 					      void	        *resp)
 {
-	struct sockaddr_in	saddr;
-
 	dprintk("lockd: SM_NOTIFY     called\n");
 
 	if (!nlm_privileged_requester(rqstp)) {
@@ -462,14 +458,7 @@ nlmsvc_proc_sm_notify(struct svc_rqst *rqstp, struct nlm_reboot *argp,
 		return rpc_system_err;
 	}
 
-	/* Obtain the host pointer for this NFS server and try to
-	 * reclaim all locks we hold on this server.
-	 */
-	memset(&saddr, 0, sizeof(saddr));
-	saddr.sin_family = AF_INET;
-	saddr.sin_addr.s_addr = argp->addr;
-	nlm_host_rebooted(&saddr, argp->mon, argp->len, argp->state);
-
+	nlm_host_rebooted(argp);
 	return rpc_success;
 }
 
diff --git a/fs/lockd/svcsubs.c b/fs/lockd/svcsubs.c
index 34c2766e27c..9e4d6aab611 100644
--- a/fs/lockd/svcsubs.c
+++ b/fs/lockd/svcsubs.c
@@ -17,7 +17,6 @@
 #include <linux/nfsd/export.h>
 #include <linux/lockd/lockd.h>
 #include <linux/lockd/share.h>
-#include <linux/lockd/sm_inter.h>
 #include <linux/module.h>
 #include <linux/mount.h>
 
diff --git a/fs/lockd/xdr.c b/fs/lockd/xdr.c
index 1f226290c67..0336f2beacd 100644
--- a/fs/lockd/xdr.c
+++ b/fs/lockd/xdr.c
@@ -16,7 +16,6 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/stats.h>
 #include <linux/lockd/lockd.h>
-#include <linux/lockd/sm_inter.h>
 
 #define NLMDBG_FACILITY		NLMDBG_XDR
 
@@ -349,8 +348,8 @@ nlmsvc_decode_reboot(struct svc_rqst *rqstp, __be32 *p, struct nlm_reboot *argp)
 	if (!(p = xdr_decode_string_inplace(p, &argp->mon, &argp->len, SM_MAXSTRLEN)))
 		return 0;
 	argp->state = ntohl(*p++);
-	/* Preserve the address in network byte order */
-	argp->addr = *p++;
+	memcpy(&argp->priv.data, p, sizeof(argp->priv.data));
+	p += XDR_QUADLEN(SM_PRIV_SIZE);
 	return xdr_argsize_check(rqstp, p);
 }
 
diff --git a/fs/lockd/xdr4.c b/fs/lockd/xdr4.c
index 50c493a8ad8..e1d52865319 100644
--- a/fs/lockd/xdr4.c
+++ b/fs/lockd/xdr4.c
@@ -17,7 +17,6 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/stats.h>
 #include <linux/lockd/lockd.h>
-#include <linux/lockd/sm_inter.h>
 
 #define NLMDBG_FACILITY		NLMDBG_XDR
 
@@ -356,8 +355,8 @@ nlm4svc_decode_reboot(struct svc_rqst *rqstp, __be32 *p, struct nlm_reboot *argp
 	if (!(p = xdr_decode_string_inplace(p, &argp->mon, &argp->len, SM_MAXSTRLEN)))
 		return 0;
 	argp->state = ntohl(*p++);
-	/* Preserve the address in network byte order */
-	argp->addr  = *p++;
+	memcpy(&argp->priv.data, p, sizeof(argp->priv.data));
+	p += XDR_QUADLEN(SM_PRIV_SIZE);
 	return xdr_argsize_check(rqstp, p);
 }
 
diff --git a/fs/minix/dir.c b/fs/minix/dir.c
index f70433816a3..d4946c4c90e 100644
--- a/fs/minix/dir.c
+++ b/fs/minix/dir.c
@@ -280,7 +280,7 @@ int minix_add_link(struct dentry *dentry, struct inode *inode)
 	return -EINVAL;
 
 got_it:
-	pos = (page->index >> PAGE_CACHE_SHIFT) + p - (char*)page_address(page);
+	pos = page_offset(page) + p - (char *)page_address(page);
 	err = __minix_write_begin(NULL, page->mapping, pos, sbi->s_dirsize,
 					AOP_FLAG_UNINTERRUPTIBLE, &page, NULL);
 	if (err)
diff --git a/fs/mpage.c b/fs/mpage.c
index 552b80b3fac..16c3ef37eae 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -241,7 +241,6 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 				first_hole = page_block;
 			page_block++;
 			block_in_file++;
-			clear_buffer_mapped(map_bh);
 			continue;
 		}
 
@@ -308,7 +307,10 @@ alloc_new:
 		goto alloc_new;
 	}
 
-	if (buffer_boundary(map_bh) || (first_hole != blocks_per_page))
+	relative_block = block_in_file - *first_logical_block;
+	nblocks = map_bh->b_size >> blkbits;
+	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
+	    (first_hole != blocks_per_page))
 		bio = mpage_bio_submit(READ, bio);
 	else
 		*last_block_in_bio = blocks[blocks_per_page - 1];
diff --git a/fs/namei.c b/fs/namei.c
index af3783fff1d..f05bed24242 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -226,6 +226,16 @@ int generic_permission(struct inode *inode, int mask,
 	return -EACCES;
 }
 
+/**
+ * inode_permission  -  check for access rights to a given inode
+ * @inode:	inode to check permission on
+ * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
+ *
+ * Used to check for read/write/execute permissions on an inode.
+ * We use "fsuid" for this, letting us set arbitrary permissions
+ * for filesystem access without changing the "normal" uids which
+ * are used for other things.
+ */
 int inode_permission(struct inode *inode, int mask)
 {
 	int retval;
@@ -247,8 +257,7 @@ int inode_permission(struct inode *inode, int mask)
 			return -EACCES;
 	}
 
-	/* Ordinary permission routines do not understand MAY_APPEND. */
-	if (inode->i_op && inode->i_op->permission)
+	if (inode->i_op->permission)
 		retval = inode->i_op->permission(inode, mask);
 	else
 		retval = generic_permission(inode, mask, NULL);
@@ -265,21 +274,6 @@ int inode_permission(struct inode *inode, int mask)
 }
 
 /**
- * vfs_permission  -  check for access rights to a given path
- * @nd:		lookup result that describes the path
- * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
- *
- * Used to check for read/write/execute permissions on a path.
- * We use "fsuid" for this, letting us set arbitrary permissions
- * for filesystem access without changing the "normal" uids which
- * are used for other things.
- */
-int vfs_permission(struct nameidata *nd, int mask)
-{
-	return inode_permission(nd->path.dentry->d_inode, mask);
-}
-
-/**
  * file_permission  -  check for additional access rights to a given file
  * @file:	file to check access rights for
  * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
@@ -289,7 +283,7 @@ int vfs_permission(struct nameidata *nd, int mask)
  *
  * Note:
  *	Do not use this function in new code.  All access checks should
- *	be done using vfs_permission().
+ *	be done using inode_permission().
  */
 int file_permission(struct file *file, int mask)
 {
@@ -438,7 +432,7 @@ static int exec_permission_lite(struct inode *inode)
 {
 	umode_t	mode = inode->i_mode;
 
-	if (inode->i_op && inode->i_op->permission)
+	if (inode->i_op->permission)
 		return -EAGAIN;
 
 	if (current_fsuid() == inode->i_uid)
@@ -527,18 +521,6 @@ out_unlock:
 	return result;
 }
 
-/* SMP-safe */
-static __always_inline void
-walk_init_root(const char *name, struct nameidata *nd)
-{
-	struct fs_struct *fs = current->fs;
-
-	read_lock(&fs->lock);
-	nd->path = fs->root;
-	path_get(&fs->root);
-	read_unlock(&fs->lock);
-}
-
 /*
  * Wrapper to retry pathname resolution whenever the underlying
  * file system returns an ESTALE.
@@ -576,9 +558,16 @@ static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *l
 		goto fail;
 
 	if (*link == '/') {
+		struct fs_struct *fs = current->fs;
+
 		path_put(&nd->path);
-		walk_init_root(link, nd);
+
+		read_lock(&fs->lock);
+		nd->path = fs->root;
+		path_get(&fs->root);
+		read_unlock(&fs->lock);
 	}
+
 	res = link_path_walk(link, nd);
 	if (nd->depth || res || nd->last_type!=LAST_NORM)
 		return res;
@@ -859,7 +848,8 @@ static int __link_path_walk(const char *name, struct nameidata *nd)
 		nd->flags |= LOOKUP_CONTINUE;
 		err = exec_permission_lite(inode);
 		if (err == -EAGAIN)
-			err = vfs_permission(nd, MAY_EXEC);
+			err = inode_permission(nd->path.dentry->d_inode,
+					       MAY_EXEC);
  		if (err)
 			break;
 
@@ -918,9 +908,6 @@ static int __link_path_walk(const char *name, struct nameidata *nd)
 		inode = next.dentry->d_inode;
 		if (!inode)
 			goto out_dput;
-		err = -ENOTDIR; 
-		if (!inode->i_op)
-			goto out_dput;
 
 		if (inode->i_op->follow_link) {
 			err = do_follow_link(&next, nd);
@@ -930,9 +917,6 @@ static int __link_path_walk(const char *name, struct nameidata *nd)
 			inode = nd->path.dentry->d_inode;
 			if (!inode)
 				break;
-			err = -ENOTDIR; 
-			if (!inode->i_op)
-				break;
 		} else
 			path_to_nameidata(&next, nd);
 		err = -ENOTDIR; 
@@ -971,7 +955,7 @@ last_component:
 			break;
 		inode = next.dentry->d_inode;
 		if ((lookup_flags & LOOKUP_FOLLOW)
-		    && inode && inode->i_op && inode->i_op->follow_link) {
+		    && inode && inode->i_op->follow_link) {
 			err = do_follow_link(&next, nd);
 			if (err)
 				goto return_err;
@@ -983,7 +967,7 @@ last_component:
 			break;
 		if (lookup_flags & LOOKUP_DIRECTORY) {
 			err = -ENOTDIR; 
-			if (!inode->i_op || !inode->i_op->lookup)
+			if (!inode->i_op->lookup)
 				break;
 		}
 		goto return_base;
@@ -1479,7 +1463,7 @@ int vfs_create(struct inode *dir, struct dentry *dentry, int mode,
 	if (error)
 		return error;
 
-	if (!dir->i_op || !dir->i_op->create)
+	if (!dir->i_op->create)
 		return -EACCES;	/* shouldn't it be ENOSYS? */
 	mode &= S_IALLUGO;
 	mode |= S_IFREG;
@@ -1493,9 +1477,9 @@ int vfs_create(struct inode *dir, struct dentry *dentry, int mode,
 	return error;
 }
 
-int may_open(struct nameidata *nd, int acc_mode, int flag)
+int may_open(struct path *path, int acc_mode, int flag)
 {
-	struct dentry *dentry = nd->path.dentry;
+	struct dentry *dentry = path->dentry;
 	struct inode *inode = dentry->d_inode;
 	int error;
 
@@ -1516,13 +1500,13 @@ int may_open(struct nameidata *nd, int acc_mode, int flag)
 	if (S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode)) {
 	    	flag &= ~O_TRUNC;
 	} else if (S_ISBLK(inode->i_mode) || S_ISCHR(inode->i_mode)) {
-		if (nd->path.mnt->mnt_flags & MNT_NODEV)
+		if (path->mnt->mnt_flags & MNT_NODEV)
 			return -EACCES;
 
 		flag &= ~O_TRUNC;
 	}
 
-	error = vfs_permission(nd, acc_mode);
+	error = inode_permission(inode, acc_mode);
 	if (error)
 		return error;
 	/*
@@ -1556,6 +1540,9 @@ int may_open(struct nameidata *nd, int acc_mode, int flag)
 		 * Refuse to truncate files with mandatory locks held on them.
 		 */
 		error = locks_verify_locked(inode);
+		if (!error)
+			error = security_path_truncate(path, 0,
+					       ATTR_MTIME|ATTR_CTIME|ATTR_OPEN);
 		if (!error) {
 			DQUOT_INIT(inode);
 
@@ -1586,14 +1573,18 @@ static int __open_namei_create(struct nameidata *nd, struct path *path,
 
 	if (!IS_POSIXACL(dir->d_inode))
 		mode &= ~current->fs->umask;
+	error = security_path_mknod(&nd->path, path->dentry, mode, 0);
+	if (error)
+		goto out_unlock;
 	error = vfs_create(dir->d_inode, path->dentry, mode, nd);
+out_unlock:
 	mutex_unlock(&dir->d_inode->i_mutex);
 	dput(nd->path.dentry);
 	nd->path.dentry = path->dentry;
 	if (error)
 		return error;
 	/* Don't check for write permission, don't truncate */
-	return may_open(nd, 0, flag & ~O_TRUNC);
+	return may_open(&nd->path, 0, flag & ~O_TRUNC);
 }
 
 /*
@@ -1755,7 +1746,7 @@ do_last:
 	error = -ENOENT;
 	if (!path.dentry->d_inode)
 		goto exit_dput;
-	if (path.dentry->d_inode->i_op && path.dentry->d_inode->i_op->follow_link)
+	if (path.dentry->d_inode->i_op->follow_link)
 		goto do_link;
 
 	path_to_nameidata(&path, &nd);
@@ -1779,7 +1770,7 @@ ok:
 		if (error)
 			goto exit;
 	}
-	error = may_open(&nd, acc_mode, flag);
+	error = may_open(&nd.path, acc_mode, flag);
 	if (error) {
 		if (will_write)
 			mnt_drop_write(nd.path.mnt);
@@ -1936,7 +1927,7 @@ int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev)
 	if ((S_ISCHR(mode) || S_ISBLK(mode)) && !capable(CAP_MKNOD))
 		return -EPERM;
 
-	if (!dir->i_op || !dir->i_op->mknod)
+	if (!dir->i_op->mknod)
 		return -EPERM;
 
 	error = devcgroup_inode_mknod(mode, dev);
@@ -1999,6 +1990,9 @@ asmlinkage long sys_mknodat(int dfd, const char __user *filename, int mode,
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto out_dput;
+	error = security_path_mknod(&nd.path, dentry, mode, dev);
+	if (error)
+		goto out_drop_write;
 	switch (mode & S_IFMT) {
 		case 0: case S_IFREG:
 			error = vfs_create(nd.path.dentry->d_inode,dentry,mode,&nd);
@@ -2011,6 +2005,7 @@ asmlinkage long sys_mknodat(int dfd, const char __user *filename, int mode,
 			error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,0);
 			break;
 	}
+out_drop_write:
 	mnt_drop_write(nd.path.mnt);
 out_dput:
 	dput(dentry);
@@ -2034,7 +2029,7 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	if (error)
 		return error;
 
-	if (!dir->i_op || !dir->i_op->mkdir)
+	if (!dir->i_op->mkdir)
 		return -EPERM;
 
 	mode &= (S_IRWXUGO|S_ISVTX);
@@ -2070,7 +2065,11 @@ asmlinkage long sys_mkdirat(int dfd, const char __user *pathname, int mode)
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto out_dput;
+	error = security_path_mkdir(&nd.path, dentry, mode);
+	if (error)
+		goto out_drop_write;
 	error = vfs_mkdir(nd.path.dentry->d_inode, dentry, mode);
+out_drop_write:
 	mnt_drop_write(nd.path.mnt);
 out_dput:
 	dput(dentry);
@@ -2121,7 +2120,7 @@ int vfs_rmdir(struct inode *dir, struct dentry *dentry)
 	if (error)
 		return error;
 
-	if (!dir->i_op || !dir->i_op->rmdir)
+	if (!dir->i_op->rmdir)
 		return -EPERM;
 
 	DQUOT_INIT(dir);
@@ -2180,7 +2179,11 @@ static long do_rmdir(int dfd, const char __user *pathname)
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto exit3;
+	error = security_path_rmdir(&nd.path, dentry);
+	if (error)
+		goto exit4;
 	error = vfs_rmdir(nd.path.dentry->d_inode, dentry);
+exit4:
 	mnt_drop_write(nd.path.mnt);
 exit3:
 	dput(dentry);
@@ -2204,7 +2207,7 @@ int vfs_unlink(struct inode *dir, struct dentry *dentry)
 	if (error)
 		return error;
 
-	if (!dir->i_op || !dir->i_op->unlink)
+	if (!dir->i_op->unlink)
 		return -EPERM;
 
 	DQUOT_INIT(dir);
@@ -2265,7 +2268,11 @@ static long do_unlinkat(int dfd, const char __user *pathname)
 		error = mnt_want_write(nd.path.mnt);
 		if (error)
 			goto exit2;
+		error = security_path_unlink(&nd.path, dentry);
+		if (error)
+			goto exit3;
 		error = vfs_unlink(nd.path.dentry->d_inode, dentry);
+exit3:
 		mnt_drop_write(nd.path.mnt);
 	exit2:
 		dput(dentry);
@@ -2307,7 +2314,7 @@ int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname)
 	if (error)
 		return error;
 
-	if (!dir->i_op || !dir->i_op->symlink)
+	if (!dir->i_op->symlink)
 		return -EPERM;
 
 	error = security_inode_symlink(dir, dentry, oldname);
@@ -2346,7 +2353,11 @@ asmlinkage long sys_symlinkat(const char __user *oldname,
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto out_dput;
+	error = security_path_symlink(&nd.path, dentry, from);
+	if (error)
+		goto out_drop_write;
 	error = vfs_symlink(nd.path.dentry->d_inode, dentry, from);
+out_drop_write:
 	mnt_drop_write(nd.path.mnt);
 out_dput:
 	dput(dentry);
@@ -2384,7 +2395,7 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de
 	 */
 	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
 		return -EPERM;
-	if (!dir->i_op || !dir->i_op->link)
+	if (!dir->i_op->link)
 		return -EPERM;
 	if (S_ISDIR(inode->i_mode))
 		return -EPERM;
@@ -2443,7 +2454,11 @@ asmlinkage long sys_linkat(int olddfd, const char __user *oldname,
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto out_dput;
+	error = security_path_link(old_path.dentry, &nd.path, new_dentry);
+	if (error)
+		goto out_drop_write;
 	error = vfs_link(old_path.dentry, nd.path.dentry->d_inode, new_dentry);
+out_drop_write:
 	mnt_drop_write(nd.path.mnt);
 out_dput:
 	dput(new_dentry);
@@ -2587,7 +2602,7 @@ int vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (error)
 		return error;
 
-	if (!old_dir->i_op || !old_dir->i_op->rename)
+	if (!old_dir->i_op->rename)
 		return -EPERM;
 
 	DQUOT_INIT(old_dir);
@@ -2679,8 +2694,13 @@ asmlinkage long sys_renameat(int olddfd, const char __user *oldname,
 	error = mnt_want_write(oldnd.path.mnt);
 	if (error)
 		goto exit5;
+	error = security_path_rename(&oldnd.path, old_dentry,
+				     &newnd.path, new_dentry);
+	if (error)
+		goto exit6;
 	error = vfs_rename(old_dir->d_inode, old_dentry,
 				   new_dir->d_inode, new_dentry);
+exit6:
 	mnt_drop_write(oldnd.path.mnt);
 exit5:
 	dput(new_dentry);
@@ -2750,13 +2770,16 @@ int vfs_follow_link(struct nameidata *nd, const char *link)
 /* get the link contents into pagecache */
 static char *page_getlink(struct dentry * dentry, struct page **ppage)
 {
-	struct page * page;
+	char *kaddr;
+	struct page *page;
 	struct address_space *mapping = dentry->d_inode->i_mapping;
 	page = read_mapping_page(mapping, 0, NULL);
 	if (IS_ERR(page))
 		return (char*)page;
 	*ppage = page;
-	return kmap(page);
+	kaddr = kmap(page);
+	nd_terminate_link(kaddr, dentry->d_inode->i_size, PAGE_SIZE - 1);
+	return kaddr;
 }
 
 int page_readlink(struct dentry *dentry, char __user *buffer, int buflen)
@@ -2788,18 +2811,23 @@ void page_put_link(struct dentry *dentry, struct nameidata *nd, void *cookie)
 	}
 }
 
-int __page_symlink(struct inode *inode, const char *symname, int len,
-		gfp_t gfp_mask)
+/*
+ * The nofs argument instructs pagecache_write_begin to pass AOP_FLAG_NOFS
+ */
+int __page_symlink(struct inode *inode, const char *symname, int len, int nofs)
 {
 	struct address_space *mapping = inode->i_mapping;
 	struct page *page;
 	void *fsdata;
 	int err;
 	char *kaddr;
+	unsigned int flags = AOP_FLAG_UNINTERRUPTIBLE;
+	if (nofs)
+		flags |= AOP_FLAG_NOFS;
 
 retry:
 	err = pagecache_write_begin(NULL, mapping, 0, len-1,
-				AOP_FLAG_UNINTERRUPTIBLE, &page, &fsdata);
+				flags, &page, &fsdata);
 	if (err)
 		goto fail;
 
@@ -2823,7 +2851,7 @@ fail:
 int page_symlink(struct inode *inode, const char *symname, int len)
 {
 	return __page_symlink(inode, symname, len,
-			mapping_gfp_mask(inode->i_mapping));
+			!(mapping_gfp_mask(inode->i_mapping) & __GFP_FS));
 }
 
 const struct inode_operations page_symlink_inode_operations = {
@@ -2849,7 +2877,6 @@ EXPORT_SYMBOL(path_lookup);
 EXPORT_SYMBOL(kern_path);
 EXPORT_SYMBOL(vfs_path_lookup);
 EXPORT_SYMBOL(inode_permission);
-EXPORT_SYMBOL(vfs_permission);
 EXPORT_SYMBOL(file_permission);
 EXPORT_SYMBOL(unlock_rename);
 EXPORT_SYMBOL(vfs_create);
@@ -2865,3 +2892,10 @@ EXPORT_SYMBOL(vfs_symlink);
 EXPORT_SYMBOL(vfs_unlink);
 EXPORT_SYMBOL(dentry_unhash);
 EXPORT_SYMBOL(generic_readlink);
+
+/* to be mentioned only in INIT_TASK */
+struct fs_struct init_fs = {
+	.count		= ATOMIC_INIT(1),
+	.lock		= __RW_LOCK_UNLOCKED(init_fs.lock),
+	.umask		= 0022,
+};
diff --git a/fs/namespace.c b/fs/namespace.c
index 1c09cab8f7c..a40685d800a 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1990,7 +1990,7 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns,
 	if (!new_ns->root) {
 		up_write(&namespace_sem);
 		kfree(new_ns);
-		return ERR_PTR(-ENOMEM);;
+		return ERR_PTR(-ENOMEM);
 	}
 	spin_lock(&vfsmount_lock);
 	list_add_tail(&new_ns->list, &new_ns->root->mnt_list);
diff --git a/fs/ncpfs/getopt.c b/fs/ncpfs/getopt.c
index 335b003dddf..0af3349de85 100644
--- a/fs/ncpfs/getopt.c
+++ b/fs/ncpfs/getopt.c
@@ -16,7 +16,6 @@
  *	@opts: an array of &struct option entries controlling parser operations
  *	@optopt: output; will contain the current option
  *	@optarg: output; will contain the value (if one exists)
- *	@flag: output; may be NULL; should point to a long for or'ing flags
  *	@value: output; may be NULL; will be overwritten with the integer value
  *		of the current argument.
  *
diff --git a/fs/ncpfs/ioctl.c b/fs/ncpfs/ioctl.c
index 6d04e050c74..f54360f50a9 100644
--- a/fs/ncpfs/ioctl.c
+++ b/fs/ncpfs/ioctl.c
@@ -98,7 +98,7 @@ struct compat_ncp_objectname_ioctl
 {
 	s32		auth_type;
 	u32		object_name_len;
-	compat_caddr_t	object_name;	/* an userspace data, in most cases user name */
+	compat_caddr_t	object_name;	/* a userspace data, in most cases user name */
 };
 
 struct compat_ncp_fs_info_v2 {
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index c2e9cfd9e5a..3e634f2a108 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -16,6 +16,7 @@
 #include <linux/mutex.h>
 #include <linux/freezer.h>
 #include <linux/kthread.h>
+#include <linux/sunrpc/svcauth_gss.h>
 
 #include <net/inet_sock.h>
 
@@ -182,10 +183,34 @@ void nfs_callback_down(void)
 	mutex_unlock(&nfs_callback_mutex);
 }
 
+static int check_gss_callback_principal(struct nfs_client *clp,
+					struct svc_rqst *rqstp)
+{
+	struct rpc_clnt *r = clp->cl_rpcclient;
+	char *p = svc_gss_principal(rqstp);
+
+	/*
+	 * It might just be a normal user principal, in which case
+	 * userspace won't bother to tell us the name at all.
+	 */
+	if (p == NULL)
+		return SVC_DENIED;
+
+	/* Expect a GSS_C_NT_HOSTBASED_NAME like "nfs@serverhostname" */
+
+	if (memcmp(p, "nfs@", 4) != 0)
+		return SVC_DENIED;
+	p += 4;
+	if (strcmp(p, r->cl_server) != 0)
+		return SVC_DENIED;
+	return SVC_OK;
+}
+
 static int nfs_callback_authenticate(struct svc_rqst *rqstp)
 {
 	struct nfs_client *clp;
 	RPC_IFDEBUG(char buf[RPC_MAX_ADDRBUFLEN]);
+	int ret = SVC_OK;
 
 	/* Don't talk to strangers */
 	clp = nfs_find_client(svc_addr(rqstp), 4);
@@ -194,21 +219,22 @@ static int nfs_callback_authenticate(struct svc_rqst *rqstp)
 
 	dprintk("%s: %s NFSv4 callback!\n", __func__,
 			svc_print_addr(rqstp, buf, sizeof(buf)));
-	nfs_put_client(clp);
 
 	switch (rqstp->rq_authop->flavour) {
 		case RPC_AUTH_NULL:
 			if (rqstp->rq_proc != CB_NULL)
-				return SVC_DENIED;
+				ret = SVC_DENIED;
 			break;
 		case RPC_AUTH_UNIX:
 			break;
 		case RPC_AUTH_GSS:
-			/* FIXME: RPCSEC_GSS handling? */
+			ret = check_gss_callback_principal(clp, rqstp);
+			break;
 		default:
-			return SVC_DENIED;
+			ret = SVC_DENIED;
 	}
-	return SVC_OK;
+	nfs_put_client(clp);
+	return ret;
 }
 
 /*
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 7547600b617..9b728f3565a 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -143,7 +143,6 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
 	clp->cl_proto = cl_init->proto;
 
 #ifdef CONFIG_NFS_V4
-	init_rwsem(&clp->cl_sem);
 	INIT_LIST_HEAD(&clp->cl_delegations);
 	spin_lock_init(&clp->cl_lock);
 	INIT_DELAYED_WORK(&clp->cl_renewd, nfs4_renew_state);
@@ -224,31 +223,54 @@ void nfs_put_client(struct nfs_client *clp)
 	}
 }
 
-static int nfs_sockaddr_match_ipaddr4(const struct sockaddr_in *sa1,
-				 const struct sockaddr_in *sa2)
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+static const struct in6_addr *nfs_map_ipv4_addr(const struct sockaddr *sa, struct in6_addr *addr_mapped)
 {
-	return sa1->sin_addr.s_addr == sa2->sin_addr.s_addr;
+	switch (sa->sa_family) {
+		default:
+			return NULL;
+		case AF_INET6:
+			return &((const struct sockaddr_in6 *)sa)->sin6_addr;
+			break;
+		case AF_INET:
+			ipv6_addr_set_v4mapped(((const struct sockaddr_in *)sa)->sin_addr.s_addr,
+					addr_mapped);
+			return addr_mapped;
+	}
 }
 
-static int nfs_sockaddr_match_ipaddr6(const struct sockaddr_in6 *sa1,
-				 const struct sockaddr_in6 *sa2)
+static int nfs_sockaddr_match_ipaddr(const struct sockaddr *sa1,
+		const struct sockaddr *sa2)
+{
+	const struct in6_addr *addr1;
+	const struct in6_addr *addr2;
+	struct in6_addr addr1_mapped;
+	struct in6_addr addr2_mapped;
+
+	addr1 = nfs_map_ipv4_addr(sa1, &addr1_mapped);
+	if (likely(addr1 != NULL)) {
+		addr2 = nfs_map_ipv4_addr(sa2, &addr2_mapped);
+		if (likely(addr2 != NULL))
+			return ipv6_addr_equal(addr1, addr2);
+	}
+	return 0;
+}
+#else
+static int nfs_sockaddr_match_ipaddr4(const struct sockaddr_in *sa1,
+				 const struct sockaddr_in *sa2)
 {
-	return ipv6_addr_equal(&sa1->sin6_addr, &sa2->sin6_addr);
+	return sa1->sin_addr.s_addr == sa2->sin_addr.s_addr;
 }
 
 static int nfs_sockaddr_match_ipaddr(const struct sockaddr *sa1,
 				 const struct sockaddr *sa2)
 {
-	switch (sa1->sa_family) {
-	case AF_INET:
-		return nfs_sockaddr_match_ipaddr4((const struct sockaddr_in *)sa1,
-				(const struct sockaddr_in *)sa2);
-	case AF_INET6:
-		return nfs_sockaddr_match_ipaddr6((const struct sockaddr_in6 *)sa1,
-				(const struct sockaddr_in6 *)sa2);
-	}
-	BUG();
+	if (unlikely(sa1->sa_family != AF_INET || sa2->sa_family != AF_INET))
+		return 0;
+	return nfs_sockaddr_match_ipaddr4((const struct sockaddr_in *)sa1,
+			(const struct sockaddr_in *)sa2);
 }
+#endif
 
 /*
  * Find a client by IP address and protocol version
@@ -270,8 +292,6 @@ struct nfs_client *nfs_find_client(const struct sockaddr *addr, u32 nfsversion)
 		if (clp->rpc_ops->version != nfsversion)
 			continue;
 
-		if (addr->sa_family != clap->sa_family)
-			continue;
 		/* Match only the IP address, not the port number */
 		if (!nfs_sockaddr_match_ipaddr(addr, clap))
 			continue;
@@ -305,8 +325,6 @@ struct nfs_client *nfs_find_client_next(struct nfs_client *clp)
 		if (clp->rpc_ops->version != nfsvers)
 			continue;
 
-		if (sap->sa_family != clap->sa_family)
-			continue;
 		/* Match only the IP address, not the port number */
 		if (!nfs_sockaddr_match_ipaddr(sap, clap))
 			continue;
@@ -470,7 +488,7 @@ static void nfs_init_timeout_values(struct rpc_timeout *to, int proto,
 static int nfs_create_rpc_client(struct nfs_client *clp,
 				 const struct rpc_timeout *timeparms,
 				 rpc_authflavor_t flavor,
-				 int flags)
+				 int discrtry, int noresvport)
 {
 	struct rpc_clnt		*clnt = NULL;
 	struct rpc_create_args args = {
@@ -482,9 +500,13 @@ static int nfs_create_rpc_client(struct nfs_client *clp,
 		.program	= &nfs_program,
 		.version	= clp->rpc_ops->version,
 		.authflavor	= flavor,
-		.flags		= flags,
 	};
 
+	if (discrtry)
+		args.flags |= RPC_CLNT_CREATE_DISCRTRY;
+	if (noresvport)
+		args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
+
 	if (!IS_ERR(clp->cl_rpcclient))
 		return 0;
 
@@ -522,6 +544,8 @@ static int nfs_start_lockd(struct nfs_server *server)
 		.protocol	= server->flags & NFS_MOUNT_TCP ?
 						IPPROTO_TCP : IPPROTO_UDP,
 		.nfs_version	= clp->rpc_ops->version,
+		.noresvport	= server->flags & NFS_MOUNT_NORESVPORT ?
+					1 : 0,
 	};
 
 	if (nlm_init.nfs_version > 3)
@@ -623,7 +647,8 @@ static int nfs_init_client(struct nfs_client *clp,
 	 * Create a client RPC handle for doing FSSTAT with UNIX auth only
 	 * - RFC 2623, sec 2.3.2
 	 */
-	error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX, 0);
+	error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX,
+				      0, data->flags & NFS_MOUNT_NORESVPORT);
 	if (error < 0)
 		goto error;
 	nfs_mark_client_ready(clp, NFS_CS_READY);
@@ -965,7 +990,8 @@ error:
 static int nfs4_init_client(struct nfs_client *clp,
 		const struct rpc_timeout *timeparms,
 		const char *ip_addr,
-		rpc_authflavor_t authflavour)
+		rpc_authflavor_t authflavour,
+		int flags)
 {
 	int error;
 
@@ -979,7 +1005,7 @@ static int nfs4_init_client(struct nfs_client *clp,
 	clp->rpc_ops = &nfs_v4_clientops;
 
 	error = nfs_create_rpc_client(clp, timeparms, authflavour,
-					RPC_CLNT_CREATE_DISCRTRY);
+				      1, flags & NFS_MOUNT_NORESVPORT);
 	if (error < 0)
 		goto error;
 	memcpy(clp->cl_ipaddr, ip_addr, sizeof(clp->cl_ipaddr));
@@ -1030,7 +1056,8 @@ static int nfs4_set_client(struct nfs_server *server,
 		error = PTR_ERR(clp);
 		goto error;
 	}
-	error = nfs4_init_client(clp, timeparms, ip_addr, authflavour);
+	error = nfs4_init_client(clp, timeparms, ip_addr, authflavour,
+					server->flags);
 	if (error < 0)
 		goto error_put;
 
@@ -1059,6 +1086,10 @@ static int nfs4_init_server(struct nfs_server *server,
 	nfs_init_timeout_values(&timeparms, data->nfs_server.protocol,
 			data->timeo, data->retrans);
 
+	/* Initialise the client representation from the mount data */
+	server->flags = data->flags;
+	server->caps |= NFS_CAP_ATOMIC_OPEN;
+
 	/* Get a client record */
 	error = nfs4_set_client(server,
 			data->nfs_server.hostname,
@@ -1071,10 +1102,6 @@ static int nfs4_init_server(struct nfs_server *server,
 	if (error < 0)
 		goto error;
 
-	/* Initialise the client representation from the mount data */
-	server->flags = data->flags;
-	server->caps |= NFS_CAP_ATOMIC_OPEN;
-
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);
 	if (data->wsize)
@@ -1177,6 +1204,10 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 	parent_server = NFS_SB(data->sb);
 	parent_client = parent_server->nfs_client;
 
+	/* Initialise the client representation from the parent server */
+	nfs_server_copy_userdata(server, parent_server);
+	server->caps |= NFS_CAP_ATOMIC_OPEN;
+
 	/* Get a client representation.
 	 * Note: NFSv4 always uses TCP, */
 	error = nfs4_set_client(server, data->hostname,
@@ -1189,10 +1220,6 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 	if (error < 0)
 		goto error;
 
-	/* Initialise the client representation from the parent server */
-	nfs_server_copy_userdata(server, parent_server);
-	server->caps |= NFS_CAP_ATOMIC_OPEN;
-
 	error = nfs_init_server_rpcclient(server, parent_server->client->cl_timeout, data->authflavor);
 	if (error < 0)
 		goto error;
diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
index cc563cfa694..968225a8801 100644
--- a/fs/nfs/delegation.c
+++ b/fs/nfs/delegation.c
@@ -43,6 +43,27 @@ static void nfs_free_delegation(struct nfs_delegation *delegation)
 		put_rpccred(cred);
 }
 
+void nfs_mark_delegation_referenced(struct nfs_delegation *delegation)
+{
+	set_bit(NFS_DELEGATION_REFERENCED, &delegation->flags);
+}
+
+int nfs_have_delegation(struct inode *inode, fmode_t flags)
+{
+	struct nfs_delegation *delegation;
+	int ret = 0;
+
+	flags &= FMODE_READ|FMODE_WRITE;
+	rcu_read_lock();
+	delegation = rcu_dereference(NFS_I(inode)->delegation);
+	if (delegation != NULL && (delegation->type & flags) == flags) {
+		nfs_mark_delegation_referenced(delegation);
+		ret = 1;
+	}
+	rcu_read_unlock();
+	return ret;
+}
+
 static int nfs_delegation_claim_locks(struct nfs_open_context *ctx, struct nfs4_state *state)
 {
 	struct inode *inode = state->inode;
@@ -119,7 +140,7 @@ void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, st
 	delegation->maxsize = res->maxsize;
 	oldcred = delegation->cred;
 	delegation->cred = get_rpccred(cred);
-	delegation->flags &= ~NFS_DELEGATION_NEED_RECLAIM;
+	clear_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags);
 	NFS_I(inode)->delegation_state = delegation->type;
 	smp_wmb();
 	put_rpccred(oldcred);
@@ -134,19 +155,35 @@ static int nfs_do_return_delegation(struct inode *inode, struct nfs_delegation *
 	return res;
 }
 
+static struct inode *nfs_delegation_grab_inode(struct nfs_delegation *delegation)
+{
+	struct inode *inode = NULL;
+
+	spin_lock(&delegation->lock);
+	if (delegation->inode != NULL)
+		inode = igrab(delegation->inode);
+	spin_unlock(&delegation->lock);
+	return inode;
+}
+
 static struct nfs_delegation *nfs_detach_delegation_locked(struct nfs_inode *nfsi, const nfs4_stateid *stateid)
 {
 	struct nfs_delegation *delegation = rcu_dereference(nfsi->delegation);
 
 	if (delegation == NULL)
 		goto nomatch;
+	spin_lock(&delegation->lock);
 	if (stateid != NULL && memcmp(delegation->stateid.data, stateid->data,
 				sizeof(delegation->stateid.data)) != 0)
-		goto nomatch;
+		goto nomatch_unlock;
 	list_del_rcu(&delegation->super_list);
+	delegation->inode = NULL;
 	nfsi->delegation_state = 0;
 	rcu_assign_pointer(nfsi->delegation, NULL);
+	spin_unlock(&delegation->lock);
 	return delegation;
+nomatch_unlock:
+	spin_unlock(&delegation->lock);
 nomatch:
 	return NULL;
 }
@@ -172,6 +209,8 @@ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct
 	delegation->change_attr = nfsi->change_attr;
 	delegation->cred = get_rpccred(cred);
 	delegation->inode = inode;
+	delegation->flags = 1<<NFS_DELEGATION_REFERENCED;
+	spin_lock_init(&delegation->lock);
 
 	spin_lock(&clp->cl_lock);
 	if (rcu_dereference(nfsi->delegation) != NULL) {
@@ -226,22 +265,47 @@ static void nfs_msync_inode(struct inode *inode)
  */
 static int __nfs_inode_return_delegation(struct inode *inode, struct nfs_delegation *delegation)
 {
-	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	nfs_msync_inode(inode);
-	down_read(&clp->cl_sem);
 	/* Guard against new delegated open calls */
 	down_write(&nfsi->rwsem);
 	nfs_delegation_claim_opens(inode, &delegation->stateid);
 	up_write(&nfsi->rwsem);
-	up_read(&clp->cl_sem);
 	nfs_msync_inode(inode);
 
 	return nfs_do_return_delegation(inode, delegation, 1);
 }
 
 /*
+ * Return all delegations that have been marked for return
+ */
+void nfs_client_return_marked_delegations(struct nfs_client *clp)
+{
+	struct nfs_delegation *delegation;
+	struct inode *inode;
+
+restart:
+	rcu_read_lock();
+	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
+		if (!test_and_clear_bit(NFS_DELEGATION_RETURN, &delegation->flags))
+			continue;
+		inode = nfs_delegation_grab_inode(delegation);
+		if (inode == NULL)
+			continue;
+		spin_lock(&clp->cl_lock);
+		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
+		spin_unlock(&clp->cl_lock);
+		rcu_read_unlock();
+		if (delegation != NULL)
+			__nfs_inode_return_delegation(inode, delegation);
+		iput(inode);
+		goto restart;
+	}
+	rcu_read_unlock();
+}
+
+/*
  * This function returns the delegation without reclaiming opens
  * or protecting against delegation reclaims.
  * It is therefore really only safe to be called from
@@ -279,83 +343,55 @@ int nfs_inode_return_delegation(struct inode *inode)
 	return err;
 }
 
+static void nfs_mark_return_delegation(struct nfs_client *clp, struct nfs_delegation *delegation)
+{
+	set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+	set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
+}
+
 /*
  * Return all delegations associated to a super block
  */
-void nfs_return_all_delegations(struct super_block *sb)
+void nfs_super_return_all_delegations(struct super_block *sb)
 {
 	struct nfs_client *clp = NFS_SB(sb)->nfs_client;
 	struct nfs_delegation *delegation;
-	struct inode *inode;
 
 	if (clp == NULL)
 		return;
-restart:
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		if (delegation->inode->i_sb != sb)
-			continue;
-		inode = igrab(delegation->inode);
-		if (inode == NULL)
-			continue;
-		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
-		spin_unlock(&clp->cl_lock);
-		rcu_read_unlock();
-		if (delegation != NULL)
-			__nfs_inode_return_delegation(inode, delegation);
-		iput(inode);
-		goto restart;
+		spin_lock(&delegation->lock);
+		if (delegation->inode != NULL && delegation->inode->i_sb == sb)
+			set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+		spin_unlock(&delegation->lock);
 	}
 	rcu_read_unlock();
+	nfs_client_return_marked_delegations(clp);
 }
 
-static int nfs_do_expire_all_delegations(void *ptr)
+static void nfs_client_mark_return_all_delegations(struct nfs_client *clp)
 {
-	struct nfs_client *clp = ptr;
 	struct nfs_delegation *delegation;
-	struct inode *inode;
 
-	allow_signal(SIGKILL);
-restart:
-	if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) != 0)
-		goto out;
-	if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) == 0)
-		goto out;
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		inode = igrab(delegation->inode);
-		if (inode == NULL)
-			continue;
-		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
-		spin_unlock(&clp->cl_lock);
-		rcu_read_unlock();
-		if (delegation)
-			__nfs_inode_return_delegation(inode, delegation);
-		iput(inode);
-		goto restart;
+		set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+		set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
 	}
 	rcu_read_unlock();
-out:
-	nfs_put_client(clp);
-	module_put_and_exit(0);
+}
+
+static void nfs_delegation_run_state_manager(struct nfs_client *clp)
+{
+	if (test_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state))
+		nfs4_schedule_state_manager(clp);
 }
 
 void nfs_expire_all_delegations(struct nfs_client *clp)
 {
-	struct task_struct *task;
-
-	__module_get(THIS_MODULE);
-	atomic_inc(&clp->cl_count);
-	task = kthread_run(nfs_do_expire_all_delegations, clp,
-				"%s-delegreturn",
-				rpc_peeraddr2str(clp->cl_rpcclient,
-							RPC_DISPLAY_ADDR));
-	if (!IS_ERR(task))
-		return;
-	nfs_put_client(clp);
-	module_put(THIS_MODULE);
+	nfs_client_mark_return_all_delegations(clp);
+	nfs_delegation_run_state_manager(clp);
 }
 
 /*
@@ -363,68 +399,29 @@ void nfs_expire_all_delegations(struct nfs_client *clp)
  */
 void nfs_handle_cb_pathdown(struct nfs_client *clp)
 {
-	struct nfs_delegation *delegation;
-	struct inode *inode;
-
 	if (clp == NULL)
 		return;
-restart:
+	nfs_client_mark_return_all_delegations(clp);
+}
+
+static void nfs_client_mark_return_unreferenced_delegations(struct nfs_client *clp)
+{
+	struct nfs_delegation *delegation;
+
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		inode = igrab(delegation->inode);
-		if (inode == NULL)
+		if (test_and_clear_bit(NFS_DELEGATION_REFERENCED, &delegation->flags))
 			continue;
-		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
-		spin_unlock(&clp->cl_lock);
-		rcu_read_unlock();
-		if (delegation != NULL)
-			__nfs_inode_return_delegation(inode, delegation);
-		iput(inode);
-		goto restart;
+		set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+		set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
 	}
 	rcu_read_unlock();
 }
 
-struct recall_threadargs {
-	struct inode *inode;
-	struct nfs_client *clp;
-	const nfs4_stateid *stateid;
-
-	struct completion started;
-	int result;
-};
-
-static int recall_thread(void *data)
+void nfs_expire_unreferenced_delegations(struct nfs_client *clp)
 {
-	struct recall_threadargs *args = (struct recall_threadargs *)data;
-	struct inode *inode = igrab(args->inode);
-	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
-	struct nfs_inode *nfsi = NFS_I(inode);
-	struct nfs_delegation *delegation;
-
-	daemonize("nfsv4-delegreturn");
-
-	nfs_msync_inode(inode);
-	down_read(&clp->cl_sem);
-	down_write(&nfsi->rwsem);
-	spin_lock(&clp->cl_lock);
-	delegation = nfs_detach_delegation_locked(nfsi, args->stateid);
-	if (delegation != NULL)
-		args->result = 0;
-	else
-		args->result = -ENOENT;
-	spin_unlock(&clp->cl_lock);
-	complete(&args->started);
-	nfs_delegation_claim_opens(inode, args->stateid);
-	up_write(&nfsi->rwsem);
-	up_read(&clp->cl_sem);
-	nfs_msync_inode(inode);
-
-	if (delegation != NULL)
-		nfs_do_return_delegation(inode, delegation, 1);
-	iput(inode);
-	module_put_and_exit(0);
+	nfs_client_mark_return_unreferenced_delegations(clp);
+	nfs_delegation_run_state_manager(clp);
 }
 
 /*
@@ -432,22 +429,20 @@ static int recall_thread(void *data)
  */
 int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *stateid)
 {
-	struct recall_threadargs data = {
-		.inode = inode,
-		.stateid = stateid,
-	};
-	int status;
+	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
+	struct nfs_delegation *delegation;
 
-	init_completion(&data.started);
-	__module_get(THIS_MODULE);
-	status = kernel_thread(recall_thread, &data, CLONE_KERNEL);
-	if (status < 0)
-		goto out_module_put;
-	wait_for_completion(&data.started);
-	return data.result;
-out_module_put:
-	module_put(THIS_MODULE);
-	return status;
+	rcu_read_lock();
+	delegation = rcu_dereference(NFS_I(inode)->delegation);
+	if (delegation == NULL || memcmp(delegation->stateid.data, stateid->data,
+				sizeof(delegation->stateid.data)) != 0) {
+		rcu_read_unlock();
+		return -ENOENT;
+	}
+	nfs_mark_return_delegation(clp, delegation);
+	rcu_read_unlock();
+	nfs_delegation_run_state_manager(clp);
+	return 0;
 }
 
 /*
@@ -459,10 +454,14 @@ struct inode *nfs_delegation_find_inode(struct nfs_client *clp, const struct nfs
 	struct inode *res = NULL;
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		if (nfs_compare_fh(fhandle, &NFS_I(delegation->inode)->fh) == 0) {
+		spin_lock(&delegation->lock);
+		if (delegation->inode != NULL &&
+		    nfs_compare_fh(fhandle, &NFS_I(delegation->inode)->fh) == 0) {
 			res = igrab(delegation->inode);
-			break;
 		}
+		spin_unlock(&delegation->lock);
+		if (res != NULL)
+			break;
 	}
 	rcu_read_unlock();
 	return res;
@@ -476,7 +475,7 @@ void nfs_delegation_mark_reclaim(struct nfs_client *clp)
 	struct nfs_delegation *delegation;
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list)
-		delegation->flags |= NFS_DELEGATION_NEED_RECLAIM;
+		set_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags);
 	rcu_read_unlock();
 }
 
@@ -486,17 +485,22 @@ void nfs_delegation_mark_reclaim(struct nfs_client *clp)
 void nfs_delegation_reap_unclaimed(struct nfs_client *clp)
 {
 	struct nfs_delegation *delegation;
+	struct inode *inode;
 restart:
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		if ((delegation->flags & NFS_DELEGATION_NEED_RECLAIM) == 0)
+		if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) == 0)
+			continue;
+		inode = nfs_delegation_grab_inode(delegation);
+		if (inode == NULL)
 			continue;
 		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(delegation->inode), NULL);
+		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
 		spin_unlock(&clp->cl_lock);
 		rcu_read_unlock();
 		if (delegation != NULL)
 			nfs_free_delegation(delegation);
+		iput(inode);
 		goto restart;
 	}
 	rcu_read_unlock();
diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h
index f1c5e2a5d88..09f38379517 100644
--- a/fs/nfs/delegation.h
+++ b/fs/nfs/delegation.h
@@ -17,14 +17,20 @@ struct nfs_delegation {
 	struct rpc_cred *cred;
 	struct inode *inode;
 	nfs4_stateid stateid;
-	int type;
-#define NFS_DELEGATION_NEED_RECLAIM 1
-	long flags;
+	fmode_t type;
 	loff_t maxsize;
 	__u64 change_attr;
+	unsigned long flags;
+	spinlock_t lock;
 	struct rcu_head rcu;
 };
 
+enum {
+	NFS_DELEGATION_NEED_RECLAIM = 0,
+	NFS_DELEGATION_RETURN,
+	NFS_DELEGATION_REFERENCED,
+};
+
 int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res);
 void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res);
 int nfs_inode_return_delegation(struct inode *inode);
@@ -32,9 +38,11 @@ int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *s
 void nfs_inode_return_delegation_noreclaim(struct inode *inode);
 
 struct inode *nfs_delegation_find_inode(struct nfs_client *clp, const struct nfs_fh *fhandle);
-void nfs_return_all_delegations(struct super_block *sb);
+void nfs_super_return_all_delegations(struct super_block *sb);
 void nfs_expire_all_delegations(struct nfs_client *clp);
+void nfs_expire_unreferenced_delegations(struct nfs_client *clp);
 void nfs_handle_cb_pathdown(struct nfs_client *clp);
+void nfs_client_return_marked_delegations(struct nfs_client *clp);
 
 void nfs_delegation_mark_reclaim(struct nfs_client *clp);
 void nfs_delegation_reap_unclaimed(struct nfs_client *clp);
@@ -45,22 +53,11 @@ int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state
 int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl);
 int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode);
 
-static inline int nfs_have_delegation(struct inode *inode, int flags)
-{
-	struct nfs_delegation *delegation;
-	int ret = 0;
-
-	flags &= FMODE_READ|FMODE_WRITE;
-	rcu_read_lock();
-	delegation = rcu_dereference(NFS_I(inode)->delegation);
-	if (delegation != NULL && (delegation->type & flags) == flags)
-		ret = 1;
-	rcu_read_unlock();
-	return ret;
-}
+void nfs_mark_delegation_referenced(struct nfs_delegation *delegation);
+int nfs_have_delegation(struct inode *inode, fmode_t flags);
 
 #else
-static inline int nfs_have_delegation(struct inode *inode, int flags)
+static inline int nfs_have_delegation(struct inode *inode, fmode_t flags)
 {
 	return 0;
 }
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 3e64b98f3a9..e35c8199f82 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -799,6 +799,9 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
 		goto out_bad;
 	}
 
+	if (nfs_have_delegation(inode, FMODE_READ))
+		goto out_set_verifier;
+
 	/* Force a full look up iff the parent directory has changed */
 	if (!nfs_is_exclusive_create(dir, nd) && nfs_check_verifier(dir, dentry)) {
 		if (nfs_lookup_verify_inode(inode, nd))
@@ -817,6 +820,7 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
 	if ((error = nfs_refresh_inode(inode, &fattr)) != 0)
 		goto out_bad;
 
+out_set_verifier:
 	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
  out_valid:
 	dput(parent);
@@ -973,7 +977,7 @@ struct dentry_operations nfs4_dentry_operations = {
  * Use intent information to determine whether we need to substitute
  * the NFSv4-style stateful OPEN for the LOOKUP call
  */
-static int is_atomic_open(struct inode *dir, struct nameidata *nd)
+static int is_atomic_open(struct nameidata *nd)
 {
 	if (nd == NULL || nfs_lookup_check_intent(nd, LOOKUP_OPEN) == 0)
 		return 0;
@@ -996,7 +1000,7 @@ static struct dentry *nfs_atomic_lookup(struct inode *dir, struct dentry *dentry
 			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
 
 	/* Check that we are indeed trying to open this file */
-	if (!is_atomic_open(dir, nd))
+	if (!is_atomic_open(nd))
 		goto no_open;
 
 	if (dentry->d_name.len > NFS_SERVER(dir)->namelen) {
@@ -1047,10 +1051,10 @@ static int nfs_open_revalidate(struct dentry *dentry, struct nameidata *nd)
 	struct inode *dir;
 	int openflags, ret = 0;
 
+	if (!is_atomic_open(nd))
+		goto no_open;
 	parent = dget_parent(dentry);
 	dir = parent->d_inode;
-	if (!is_atomic_open(dir, nd))
-		goto no_open;
 	/* We can't create new files in nfs_open_revalidate(), so we
 	 * optimize away revalidation of negative dentries.
 	 */
@@ -1062,11 +1066,11 @@ static int nfs_open_revalidate(struct dentry *dentry, struct nameidata *nd)
 
 	/* NFS only supports OPEN on regular files */
 	if (!S_ISREG(inode->i_mode))
-		goto no_open;
+		goto no_open_dput;
 	openflags = nd->intent.open.flags;
 	/* We cannot do exclusive creation on a positive dentry */
 	if ((openflags & (O_CREAT|O_EXCL)) == (O_CREAT|O_EXCL))
-		goto no_open;
+		goto no_open_dput;
 	/* We can't create new files, or truncate existing ones here */
 	openflags &= ~(O_CREAT|O_TRUNC);
 
@@ -1081,10 +1085,9 @@ out:
 	if (!ret)
 		d_drop(dentry);
 	return ret;
-no_open:
+no_open_dput:
 	dput(parent);
-	if (inode != NULL && nfs_have_delegation(inode, FMODE_READ))
-		return 1;
+no_open:
 	return nfs_lookup_revalidate(dentry, nd);
 }
 #endif /* CONFIG_NFSV4 */
@@ -1794,7 +1797,8 @@ static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, str
 	cache = nfs_access_search_rbtree(inode, cred);
 	if (cache == NULL)
 		goto out;
-	if (!time_in_range(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo))
+	if (!nfs_have_delegation(inode, FMODE_READ) &&
+	    !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo))
 		goto out_stale;
 	res->jiffies = cache->jiffies;
 	res->cred = cache->cred;
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index d319b49f8f0..90f292b520d 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -354,7 +354,7 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping,
 		file->f_path.dentry->d_name.name,
 		mapping->host->i_ino, len, (long long) pos);
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index d22eb383e1c..0c381686171 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -592,7 +592,7 @@ static void nfs_file_set_open_context(struct file *filp, struct nfs_open_context
 /*
  * Given an inode, search for an open context with the desired characteristics
  */
-struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode)
+struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, fmode_t mode)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	struct nfs_open_context *pos, *ctx = NULL;
@@ -712,14 +712,7 @@ int nfs_attribute_timeout(struct inode *inode)
 
 	if (nfs_have_delegation(inode, FMODE_READ))
 		return 0;
-	/*
-	 * Special case: if the attribute timeout is set to 0, then always
-	 * 		 treat the cache as having expired (unless holding
-	 * 		 a delegation).
-	 */
-	if (nfsi->attrtimeo == 0)
-		return 1;
-	return !time_in_range(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
+	return !time_in_range_open(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
 }
 
 /**
@@ -1182,7 +1175,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		nfsi->attrtimeo_timestamp = now;
 		nfsi->attr_gencount = nfs_inc_attr_generation_counter();
 	} else {
-		if (!time_in_range(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
+		if (!time_in_range_open(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
 			if ((nfsi->attrtimeo <<= 1) > NFS_MAXATTRTIMEO(inode))
 				nfsi->attrtimeo = NFS_MAXATTRTIMEO(inode);
 			nfsi->attrtimeo_timestamp = now;
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index d212ee41caf..340ede8f608 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -63,6 +63,20 @@ struct nfs_parsed_mount_data {
 	struct security_mnt_opts lsm_opts;
 };
 
+/* mount_clnt.c */
+struct nfs_mount_request {
+	struct sockaddr		*sap;
+	size_t			salen;
+	char			*hostname;
+	char			*dirpath;
+	u32			version;
+	unsigned short		protocol;
+	struct nfs_fh		*fh;
+	int			noresvport;
+};
+
+extern int nfs_mount(struct nfs_mount_request *info);
+
 /* client.c */
 extern struct rpc_program nfs_program;
 
diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c
index 086a6830d78..ca905a5bb1b 100644
--- a/fs/nfs/mount_clnt.c
+++ b/fs/nfs/mount_clnt.c
@@ -29,47 +29,43 @@ struct mnt_fhstatus {
 
 /**
  * nfs_mount - Obtain an NFS file handle for the given host and path
- * @addr: pointer to server's address
- * @len: size of server's address
- * @hostname: name of server host, or NULL
- * @path: pointer to string containing export path to mount
- * @version: mount version to use for this request
- * @protocol: transport protocol to use for thie request
- * @fh: pointer to location to place returned file handle
+ * @info: pointer to mount request arguments
  *
  * Uses default timeout parameters specified by underlying transport.
  */
-int nfs_mount(struct sockaddr *addr, size_t len, char *hostname, char *path,
-	      int version, int protocol, struct nfs_fh *fh)
+int nfs_mount(struct nfs_mount_request *info)
 {
 	struct mnt_fhstatus	result = {
-		.fh		= fh
+		.fh		= info->fh
 	};
 	struct rpc_message msg	= {
-		.rpc_argp	= path,
+		.rpc_argp	= info->dirpath,
 		.rpc_resp	= &result,
 	};
 	struct rpc_create_args args = {
-		.protocol	= protocol,
-		.address	= addr,
-		.addrsize	= len,
-		.servername	= hostname,
+		.protocol	= info->protocol,
+		.address	= info->sap,
+		.addrsize	= info->salen,
+		.servername	= info->hostname,
 		.program	= &mnt_program,
-		.version	= version,
+		.version	= info->version,
 		.authflavor	= RPC_AUTH_UNIX,
-		.flags		= 0,
 	};
 	struct rpc_clnt		*mnt_clnt;
 	int			status;
 
 	dprintk("NFS: sending MNT request for %s:%s\n",
-		(hostname ? hostname : "server"), path);
+		(info->hostname ? info->hostname : "server"),
+			info->dirpath);
+
+	if (info->noresvport)
+		args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
 
 	mnt_clnt = rpc_create(&args);
 	if (IS_ERR(mnt_clnt))
 		goto out_clnt_err;
 
-	if (version == NFS_MNT3_VERSION)
+	if (info->version == NFS_MNT3_VERSION)
 		msg.rpc_proc = &mnt_clnt->cl_procinfo[MOUNTPROC3_MNT];
 	else
 		msg.rpc_proc = &mnt_clnt->cl_procinfo[MNTPROC_MNT];
diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index ea790645fda..4e4d3320437 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -38,8 +38,12 @@ struct idmap;
  ((err) != NFSERR_NOFILEHANDLE))
 
 enum nfs4_client_state {
-	NFS4CLNT_STATE_RECOVER  = 0,
+	NFS4CLNT_MANAGER_RUNNING  = 0,
+	NFS4CLNT_CHECK_LEASE,
 	NFS4CLNT_LEASE_EXPIRED,
+	NFS4CLNT_RECLAIM_REBOOT,
+	NFS4CLNT_RECLAIM_NOGRACE,
+	NFS4CLNT_DELEGRETURN,
 };
 
 /*
@@ -90,12 +94,18 @@ struct nfs4_state_owner {
 
 	spinlock_t	     so_lock;
 	atomic_t	     so_count;
+	unsigned long	     so_flags;
 	struct list_head     so_states;
 	struct list_head     so_delegations;
 	struct nfs_seqid_counter so_seqid;
 	struct rpc_sequence  so_sequence;
 };
 
+enum {
+	NFS_OWNER_RECLAIM_REBOOT,
+	NFS_OWNER_RECLAIM_NOGRACE
+};
+
 /*
  * struct nfs4_state maintains the client-side state for a given
  * (state_owner,inode) tuple (OPEN) or state_owner (LOCK).
@@ -128,6 +138,8 @@ enum {
 	NFS_O_RDONLY_STATE,		/* OPEN stateid has read-only state */
 	NFS_O_WRONLY_STATE,		/* OPEN stateid has write-only state */
 	NFS_O_RDWR_STATE,		/* OPEN stateid has read/write state */
+	NFS_STATE_RECLAIM_REBOOT,	/* OPEN stateid server rebooted */
+	NFS_STATE_RECLAIM_NOGRACE,	/* OPEN stateid needs to recover state */
 };
 
 struct nfs4_state {
@@ -149,7 +161,7 @@ struct nfs4_state {
 	unsigned int n_rdonly;		/* Number of read-only references */
 	unsigned int n_wronly;		/* Number of write-only references */
 	unsigned int n_rdwr;		/* Number of read/write references */
-	int state;			/* State on the server (R,W, or RW) */
+	fmode_t state;			/* State on the server (R,W, or RW) */
 	atomic_t count;
 };
 
@@ -157,9 +169,12 @@ struct nfs4_state {
 struct nfs4_exception {
 	long timeout;
 	int retry;
+	struct nfs4_state *state;
 };
 
 struct nfs4_state_recovery_ops {
+	int owner_flag_bit;
+	int state_flag_bit;
 	int (*recover_open)(struct nfs4_state_owner *, struct nfs4_state *);
 	int (*recover_lock)(struct nfs4_state *, struct file_lock *);
 };
@@ -174,7 +189,6 @@ extern ssize_t nfs4_listxattr(struct dentry *, char *, size_t);
 
 
 /* nfs4proc.c */
-extern int nfs4_map_errors(int err);
 extern int nfs4_proc_setclientid(struct nfs_client *, u32, unsigned short, struct rpc_cred *);
 extern int nfs4_proc_setclientid_confirm(struct nfs_client *, struct rpc_cred *);
 extern int nfs4_proc_async_renew(struct nfs_client *, struct rpc_cred *);
@@ -187,7 +201,7 @@ extern int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
 		struct nfs4_fs_locations *fs_locations, struct page *page);
 
 extern struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops;
-extern struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops;
+extern struct nfs4_state_recovery_ops nfs4_nograce_recovery_ops;
 
 extern const u32 nfs4_fattr_bitmap[2];
 extern const u32 nfs4_statfs_bitmap[2];
@@ -202,16 +216,18 @@ extern void nfs4_kill_renewd(struct nfs_client *);
 extern void nfs4_renew_state(struct work_struct *);
 
 /* nfs4state.c */
-struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp);
+struct rpc_cred *nfs4_get_renew_cred_locked(struct nfs_client *clp);
 
 extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *);
 extern void nfs4_put_state_owner(struct nfs4_state_owner *);
 extern struct nfs4_state * nfs4_get_open_state(struct inode *, struct nfs4_state_owner *);
 extern void nfs4_put_open_state(struct nfs4_state *);
-extern void nfs4_close_state(struct path *, struct nfs4_state *, mode_t);
-extern void nfs4_close_sync(struct path *, struct nfs4_state *, mode_t);
-extern void nfs4_state_set_mode_locked(struct nfs4_state *, mode_t);
+extern void nfs4_close_state(struct path *, struct nfs4_state *, fmode_t);
+extern void nfs4_close_sync(struct path *, struct nfs4_state *, fmode_t);
+extern void nfs4_state_set_mode_locked(struct nfs4_state *, fmode_t);
 extern void nfs4_schedule_state_recovery(struct nfs_client *);
+extern void nfs4_schedule_state_manager(struct nfs_client *);
+extern int nfs4_state_mark_reclaim_nograce(struct nfs_client *clp, struct nfs4_state *state);
 extern void nfs4_put_lock_state(struct nfs4_lock_state *lsp);
 extern int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl);
 extern void nfs4_copy_stateid(nfs4_stateid *, struct nfs4_state *, fl_owner_t);
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 83e700a2b0c..8dde84b988d 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -62,14 +62,12 @@
 struct nfs4_opendata;
 static int _nfs4_proc_open(struct nfs4_opendata *data);
 static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
-static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *);
-static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception);
-static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp);
+static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *, struct nfs4_state *);
 static int _nfs4_proc_lookup(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
 static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
 
 /* Prevent leaks of NFSv4 errors into userland */
-int nfs4_map_errors(int err)
+static int nfs4_map_errors(int err)
 {
 	if (err < -1000) {
 		dprintk("%s could not handle NFSv4 error %d\n",
@@ -195,6 +193,83 @@ static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dent
 	kunmap_atomic(start, KM_USER0);
 }
 
+static int nfs4_wait_bit_killable(void *word)
+{
+	if (fatal_signal_pending(current))
+		return -ERESTARTSYS;
+	schedule();
+	return 0;
+}
+
+static int nfs4_wait_clnt_recover(struct nfs_client *clp)
+{
+	int res;
+
+	might_sleep();
+
+	res = wait_on_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING,
+			nfs4_wait_bit_killable, TASK_KILLABLE);
+	return res;
+}
+
+static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
+{
+	int res = 0;
+
+	might_sleep();
+
+	if (*timeout <= 0)
+		*timeout = NFS4_POLL_RETRY_MIN;
+	if (*timeout > NFS4_POLL_RETRY_MAX)
+		*timeout = NFS4_POLL_RETRY_MAX;
+	schedule_timeout_killable(*timeout);
+	if (fatal_signal_pending(current))
+		res = -ERESTARTSYS;
+	*timeout <<= 1;
+	return res;
+}
+
+/* This is the error handling routine for processes that are allowed
+ * to sleep.
+ */
+static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
+{
+	struct nfs_client *clp = server->nfs_client;
+	struct nfs4_state *state = exception->state;
+	int ret = errorcode;
+
+	exception->retry = 0;
+	switch(errorcode) {
+		case 0:
+			return 0;
+		case -NFS4ERR_ADMIN_REVOKED:
+		case -NFS4ERR_BAD_STATEID:
+		case -NFS4ERR_OPENMODE:
+			if (state == NULL)
+				break;
+			nfs4_state_mark_reclaim_nograce(clp, state);
+		case -NFS4ERR_STALE_CLIENTID:
+		case -NFS4ERR_STALE_STATEID:
+		case -NFS4ERR_EXPIRED:
+			nfs4_schedule_state_recovery(clp);
+			ret = nfs4_wait_clnt_recover(clp);
+			if (ret == 0)
+				exception->retry = 1;
+			break;
+		case -NFS4ERR_FILE_OPEN:
+		case -NFS4ERR_GRACE:
+		case -NFS4ERR_DELAY:
+			ret = nfs4_delay(server->client, &exception->timeout);
+			if (ret != 0)
+				break;
+		case -NFS4ERR_OLD_STATEID:
+			exception->retry = 1;
+	}
+	/* We failed to handle the error */
+	return nfs4_map_errors(ret);
+}
+
+
 static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
 {
 	struct nfs_client *clp = server->nfs_client;
@@ -248,7 +323,7 @@ static void nfs4_init_opendata_res(struct nfs4_opendata *p)
 }
 
 static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
-		struct nfs4_state_owner *sp, int flags,
+		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
 		const struct iattr *attrs)
 {
 	struct dentry *parent = dget_parent(path->dentry);
@@ -268,7 +343,8 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
 	p->owner = sp;
 	atomic_inc(&sp->so_count);
 	p->o_arg.fh = NFS_FH(dir);
-	p->o_arg.open_flags = flags,
+	p->o_arg.open_flags = flags;
+	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
 	p->o_arg.clientid = server->nfs_client->cl_clientid;
 	p->o_arg.id = sp->so_owner_id.id;
 	p->o_arg.name = &p->path.dentry->d_name;
@@ -324,10 +400,13 @@ static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
 	return ret;
 }
 
-static int can_open_cached(struct nfs4_state *state, int mode)
+static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
 {
 	int ret = 0;
-	switch (mode & (FMODE_READ|FMODE_WRITE|O_EXCL)) {
+
+	if (open_mode & O_EXCL)
+		goto out;
+	switch (mode & (FMODE_READ|FMODE_WRITE)) {
 		case FMODE_READ:
 			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0;
 			break;
@@ -337,21 +416,23 @@ static int can_open_cached(struct nfs4_state *state, int mode)
 		case FMODE_READ|FMODE_WRITE:
 			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0;
 	}
+out:
 	return ret;
 }
 
-static int can_open_delegated(struct nfs_delegation *delegation, mode_t open_flags)
+static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
 {
-	if ((delegation->type & open_flags) != open_flags)
+	if ((delegation->type & fmode) != fmode)
 		return 0;
-	if (delegation->flags & NFS_DELEGATION_NEED_RECLAIM)
+	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
 		return 0;
+	nfs_mark_delegation_referenced(delegation);
 	return 1;
 }
 
-static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
+static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
 {
-	switch (open_flags) {
+	switch (fmode) {
 		case FMODE_WRITE:
 			state->n_wronly++;
 			break;
@@ -361,15 +442,15 @@ static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
 		case FMODE_READ|FMODE_WRITE:
 			state->n_rdwr++;
 	}
-	nfs4_state_set_mode_locked(state, state->state | open_flags);
+	nfs4_state_set_mode_locked(state, state->state | fmode);
 }
 
-static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
+static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
 {
 	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
 		memcpy(state->stateid.data, stateid->data, sizeof(state->stateid.data));
 	memcpy(state->open_stateid.data, stateid->data, sizeof(state->open_stateid.data));
-	switch (open_flags) {
+	switch (fmode) {
 		case FMODE_READ:
 			set_bit(NFS_O_RDONLY_STATE, &state->flags);
 			break;
@@ -381,16 +462,15 @@ static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *
 	}
 }
 
-static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
+static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
 {
 	write_seqlock(&state->seqlock);
-	nfs_set_open_stateid_locked(state, stateid, open_flags);
+	nfs_set_open_stateid_locked(state, stateid, fmode);
 	write_sequnlock(&state->seqlock);
 }
 
-static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *deleg_stateid, int open_flags)
+static void __update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, const nfs4_stateid *deleg_stateid, fmode_t fmode)
 {
-	open_flags &= (FMODE_READ|FMODE_WRITE);
 	/*
 	 * Protect the call to nfs4_state_set_mode_locked and
 	 * serialise the stateid update
@@ -401,20 +481,60 @@ static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_sta
 		set_bit(NFS_DELEGATED_STATE, &state->flags);
 	}
 	if (open_stateid != NULL)
-		nfs_set_open_stateid_locked(state, open_stateid, open_flags);
+		nfs_set_open_stateid_locked(state, open_stateid, fmode);
 	write_sequnlock(&state->seqlock);
 	spin_lock(&state->owner->so_lock);
-	update_open_stateflags(state, open_flags);
+	update_open_stateflags(state, fmode);
 	spin_unlock(&state->owner->so_lock);
 }
 
-static void nfs4_return_incompatible_delegation(struct inode *inode, mode_t open_flags)
+static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
+{
+	struct nfs_inode *nfsi = NFS_I(state->inode);
+	struct nfs_delegation *deleg_cur;
+	int ret = 0;
+
+	fmode &= (FMODE_READ|FMODE_WRITE);
+
+	rcu_read_lock();
+	deleg_cur = rcu_dereference(nfsi->delegation);
+	if (deleg_cur == NULL)
+		goto no_delegation;
+
+	spin_lock(&deleg_cur->lock);
+	if (nfsi->delegation != deleg_cur ||
+	    (deleg_cur->type & fmode) != fmode)
+		goto no_delegation_unlock;
+
+	if (delegation == NULL)
+		delegation = &deleg_cur->stateid;
+	else if (memcmp(deleg_cur->stateid.data, delegation->data, NFS4_STATEID_SIZE) != 0)
+		goto no_delegation_unlock;
+
+	nfs_mark_delegation_referenced(deleg_cur);
+	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
+	ret = 1;
+no_delegation_unlock:
+	spin_unlock(&deleg_cur->lock);
+no_delegation:
+	rcu_read_unlock();
+
+	if (!ret && open_stateid != NULL) {
+		__update_open_stateid(state, open_stateid, NULL, fmode);
+		ret = 1;
+	}
+
+	return ret;
+}
+
+
+static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
 {
 	struct nfs_delegation *delegation;
 
 	rcu_read_lock();
 	delegation = rcu_dereference(NFS_I(inode)->delegation);
-	if (delegation == NULL || (delegation->type & open_flags) == open_flags) {
+	if (delegation == NULL || (delegation->type & fmode) == fmode) {
 		rcu_read_unlock();
 		return;
 	}
@@ -427,27 +547,28 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
 	struct nfs4_state *state = opendata->state;
 	struct nfs_inode *nfsi = NFS_I(state->inode);
 	struct nfs_delegation *delegation;
-	int open_mode = opendata->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL);
+	int open_mode = opendata->o_arg.open_flags & O_EXCL;
+	fmode_t fmode = opendata->o_arg.fmode;
 	nfs4_stateid stateid;
 	int ret = -EAGAIN;
 
-	rcu_read_lock();
-	delegation = rcu_dereference(nfsi->delegation);
 	for (;;) {
-		if (can_open_cached(state, open_mode)) {
+		if (can_open_cached(state, fmode, open_mode)) {
 			spin_lock(&state->owner->so_lock);
-			if (can_open_cached(state, open_mode)) {
-				update_open_stateflags(state, open_mode);
+			if (can_open_cached(state, fmode, open_mode)) {
+				update_open_stateflags(state, fmode);
 				spin_unlock(&state->owner->so_lock);
-				rcu_read_unlock();
 				goto out_return_state;
 			}
 			spin_unlock(&state->owner->so_lock);
 		}
-		if (delegation == NULL)
-			break;
-		if (!can_open_delegated(delegation, open_mode))
+		rcu_read_lock();
+		delegation = rcu_dereference(nfsi->delegation);
+		if (delegation == NULL ||
+		    !can_open_delegated(delegation, fmode)) {
+			rcu_read_unlock();
 			break;
+		}
 		/* Save the delegation */
 		memcpy(stateid.data, delegation->stateid.data, sizeof(stateid.data));
 		rcu_read_unlock();
@@ -455,19 +576,11 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
 		if (ret != 0)
 			goto out;
 		ret = -EAGAIN;
-		rcu_read_lock();
-		delegation = rcu_dereference(nfsi->delegation);
-		/* If no delegation, try a cached open */
-		if (delegation == NULL)
-			continue;
-		/* Is the delegation still valid? */
-		if (memcmp(stateid.data, delegation->stateid.data, sizeof(stateid.data)) != 0)
-			continue;
-		rcu_read_unlock();
-		update_open_stateid(state, NULL, &stateid, open_mode);
-		goto out_return_state;
+
+		/* Try to update the stateid using the delegation */
+		if (update_open_stateid(state, NULL, &stateid, fmode))
+			goto out_return_state;
 	}
-	rcu_read_unlock();
 out:
 	return ERR_PTR(ret);
 out_return_state:
@@ -480,7 +593,6 @@ static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data
 	struct inode *inode;
 	struct nfs4_state *state = NULL;
 	struct nfs_delegation *delegation;
-	nfs4_stateid *deleg_stateid = NULL;
 	int ret;
 
 	if (!data->rpc_done) {
@@ -507,7 +619,7 @@ static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data
 		if (delegation)
 			delegation_flags = delegation->flags;
 		rcu_read_unlock();
-		if (!(delegation_flags & NFS_DELEGATION_NEED_RECLAIM))
+		if ((delegation_flags & 1UL<<NFS_DELEGATION_NEED_RECLAIM) == 0)
 			nfs_inode_set_delegation(state->inode,
 					data->owner->so_cred,
 					&data->o_res);
@@ -516,12 +628,9 @@ static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data
 					data->owner->so_cred,
 					&data->o_res);
 	}
-	rcu_read_lock();
-	delegation = rcu_dereference(NFS_I(inode)->delegation);
-	if (delegation != NULL)
-		deleg_stateid = &delegation->stateid;
-	update_open_stateid(state, &data->o_res.stateid, deleg_stateid, data->o_arg.open_flags);
-	rcu_read_unlock();
+
+	update_open_stateid(state, &data->o_res.stateid, NULL,
+			data->o_arg.fmode);
 	iput(inode);
 out:
 	return state;
@@ -552,7 +661,7 @@ static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context
 {
 	struct nfs4_opendata *opendata;
 
-	opendata = nfs4_opendata_alloc(&ctx->path, state->owner, 0, NULL);
+	opendata = nfs4_opendata_alloc(&ctx->path, state->owner, 0, 0, NULL);
 	if (opendata == NULL)
 		return ERR_PTR(-ENOMEM);
 	opendata->state = state;
@@ -560,12 +669,13 @@ static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context
 	return opendata;
 }
 
-static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openflags, struct nfs4_state **res)
+static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
 {
 	struct nfs4_state *newstate;
 	int ret;
 
-	opendata->o_arg.open_flags = openflags;
+	opendata->o_arg.open_flags = 0;
+	opendata->o_arg.fmode = fmode;
 	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
 	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
 	nfs4_init_opendata_res(opendata);
@@ -575,7 +685,7 @@ static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openf
 	newstate = nfs4_opendata_to_nfs4_state(opendata);
 	if (IS_ERR(newstate))
 		return PTR_ERR(newstate);
-	nfs4_close_state(&opendata->path, newstate, openflags);
+	nfs4_close_state(&opendata->path, newstate, fmode);
 	*res = newstate;
 	return 0;
 }
@@ -631,7 +741,7 @@ static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state
 {
 	struct nfs_delegation *delegation;
 	struct nfs4_opendata *opendata;
-	int delegation_type = 0;
+	fmode_t delegation_type = 0;
 	int status;
 
 	opendata = nfs4_open_recoverdata_alloc(ctx, state);
@@ -641,7 +751,7 @@ static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state
 	opendata->o_arg.fh = NFS_FH(state->inode);
 	rcu_read_lock();
 	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
-	if (delegation != NULL && (delegation->flags & NFS_DELEGATION_NEED_RECLAIM) != 0)
+	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
 		delegation_type = delegation->type;
 	rcu_read_unlock();
 	opendata->o_arg.u.delegation_type = delegation_type;
@@ -744,7 +854,7 @@ static void nfs4_open_confirm_release(void *calldata)
 		goto out_free;
 	state = nfs4_opendata_to_nfs4_state(data);
 	if (!IS_ERR(state))
-		nfs4_close_state(&data->path, state, data->o_arg.open_flags);
+		nfs4_close_state(&data->path, state, data->o_arg.fmode);
 out_free:
 	nfs4_opendata_put(data);
 }
@@ -808,12 +918,12 @@ static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
 	if (data->state != NULL) {
 		struct nfs_delegation *delegation;
 
-		if (can_open_cached(data->state, data->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL)))
+		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
 			goto out_no_action;
 		rcu_read_lock();
 		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
 		if (delegation != NULL &&
-		   (delegation->flags & NFS_DELEGATION_NEED_RECLAIM) == 0) {
+		    test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) == 0) {
 			rcu_read_unlock();
 			goto out_no_action;
 		}
@@ -877,7 +987,7 @@ static void nfs4_open_release(void *calldata)
 		goto out_free;
 	state = nfs4_opendata_to_nfs4_state(data);
 	if (!IS_ERR(state))
-		nfs4_close_state(&data->path, state, data->o_arg.open_flags);
+		nfs4_close_state(&data->path, state, data->o_arg.fmode);
 out_free:
 	nfs4_opendata_put(data);
 }
@@ -955,10 +1065,11 @@ static int nfs4_recover_expired_lease(struct nfs_server *server)
 	int ret;
 
 	for (;;) {
-		ret = nfs4_wait_clnt_recover(server->client, clp);
+		ret = nfs4_wait_clnt_recover(clp);
 		if (ret != 0)
 			return ret;
-		if (!test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
+		if (!test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) &&
+		    !test_bit(NFS4CLNT_CHECK_LEASE,&clp->cl_state))
 			break;
 		nfs4_schedule_state_recovery(clp);
 	}
@@ -993,8 +1104,9 @@ static inline int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4
 
 	do {
 		err = _nfs4_open_expired(ctx, state);
-		if (err == -NFS4ERR_DELAY)
-			nfs4_handle_exception(server, err, &exception);
+		if (err != -NFS4ERR_DELAY)
+			break;
+		nfs4_handle_exception(server, err, &exception);
 	} while (exception.retry);
 	return err;
 }
@@ -1031,12 +1143,11 @@ static inline void nfs4_exclusive_attrset(struct nfs4_opendata *opendata, struct
 /*
  * Returns a referenced nfs4_state
  */
-static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res)
+static int _nfs4_do_open(struct inode *dir, struct path *path, fmode_t fmode, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res)
 {
 	struct nfs4_state_owner  *sp;
 	struct nfs4_state     *state = NULL;
 	struct nfs_server       *server = NFS_SERVER(dir);
-	struct nfs_client *clp = server->nfs_client;
 	struct nfs4_opendata *opendata;
 	int status;
 
@@ -1050,12 +1161,11 @@ static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct
 	if (status != 0)
 		goto err_put_state_owner;
 	if (path->dentry->d_inode != NULL)
-		nfs4_return_incompatible_delegation(path->dentry->d_inode, flags & (FMODE_READ|FMODE_WRITE));
-	down_read(&clp->cl_sem);
+		nfs4_return_incompatible_delegation(path->dentry->d_inode, fmode);
 	status = -ENOMEM;
-	opendata = nfs4_opendata_alloc(path, sp, flags, sattr);
+	opendata = nfs4_opendata_alloc(path, sp, fmode, flags, sattr);
 	if (opendata == NULL)
-		goto err_release_rwsem;
+		goto err_put_state_owner;
 
 	if (path->dentry->d_inode != NULL)
 		opendata->state = nfs4_get_open_state(path->dentry->d_inode, sp);
@@ -1073,13 +1183,10 @@ static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct
 		goto err_opendata_put;
 	nfs4_opendata_put(opendata);
 	nfs4_put_state_owner(sp);
-	up_read(&clp->cl_sem);
 	*res = state;
 	return 0;
 err_opendata_put:
 	nfs4_opendata_put(opendata);
-err_release_rwsem:
-	up_read(&clp->cl_sem);
 err_put_state_owner:
 	nfs4_put_state_owner(sp);
 out_err:
@@ -1088,14 +1195,14 @@ out_err:
 }
 
 
-static struct nfs4_state *nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred)
+static struct nfs4_state *nfs4_do_open(struct inode *dir, struct path *path, fmode_t fmode, int flags, struct iattr *sattr, struct rpc_cred *cred)
 {
 	struct nfs4_exception exception = { };
 	struct nfs4_state *res;
 	int status;
 
 	do {
-		status = _nfs4_do_open(dir, path, flags, sattr, cred, &res);
+		status = _nfs4_do_open(dir, path, fmode, flags, sattr, cred, &res);
 		if (status == 0)
 			break;
 		/* NOTE: BAD_SEQID means the server and client disagree about the
@@ -1230,10 +1337,13 @@ static void nfs4_close_done(struct rpc_task *task, void *data)
 			renew_lease(server, calldata->timestamp);
 			break;
 		case -NFS4ERR_STALE_STATEID:
+		case -NFS4ERR_OLD_STATEID:
+		case -NFS4ERR_BAD_STATEID:
 		case -NFS4ERR_EXPIRED:
-			break;
+			if (calldata->arg.fmode == 0)
+				break;
 		default:
-			if (nfs4_async_handle_error(task, server) == -EAGAIN) {
+			if (nfs4_async_handle_error(task, server, state) == -EAGAIN) {
 				rpc_restart_call(task);
 				return;
 			}
@@ -1272,10 +1382,10 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data)
 	nfs_fattr_init(calldata->res.fattr);
 	if (test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0) {
 		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
-		calldata->arg.open_flags = FMODE_READ;
+		calldata->arg.fmode = FMODE_READ;
 	} else if (test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0) {
 		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
-		calldata->arg.open_flags = FMODE_WRITE;
+		calldata->arg.fmode = FMODE_WRITE;
 	}
 	calldata->timestamp = jiffies;
 	rpc_call_start(task);
@@ -1328,6 +1438,7 @@ int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait)
 	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid);
 	if (calldata->arg.seqid == NULL)
 		goto out_free_calldata;
+	calldata->arg.fmode = 0;
 	calldata->arg.bitmask = server->attr_bitmask;
 	calldata->res.fattr = &calldata->fattr;
 	calldata->res.seqid = calldata->arg.seqid;
@@ -1354,13 +1465,13 @@ out:
 	return status;
 }
 
-static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state)
+static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state, fmode_t fmode)
 {
 	struct file *filp;
 	int ret;
 
 	/* If the open_intent is for execute, we have an extra check to make */
-	if (nd->intent.open.flags & FMODE_EXEC) {
+	if (fmode & FMODE_EXEC) {
 		ret = nfs_may_open(state->inode,
 				state->owner->so_cred,
 				nd->intent.open.flags);
@@ -1376,7 +1487,7 @@ static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct
 	}
 	ret = PTR_ERR(filp);
 out_close:
-	nfs4_close_sync(path, state, nd->intent.open.flags);
+	nfs4_close_sync(path, state, fmode & (FMODE_READ|FMODE_WRITE));
 	return ret;
 }
 
@@ -1392,6 +1503,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 	struct rpc_cred *cred;
 	struct nfs4_state *state;
 	struct dentry *res;
+	fmode_t fmode = nd->intent.open.flags & (FMODE_READ | FMODE_WRITE | FMODE_EXEC);
 
 	if (nd->flags & LOOKUP_CREATE) {
 		attr.ia_mode = nd->intent.open.create_mode;
@@ -1409,7 +1521,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 	parent = dentry->d_parent;
 	/* Protect against concurrent sillydeletes */
 	nfs_block_sillyrename(parent);
-	state = nfs4_do_open(dir, &path, nd->intent.open.flags, &attr, cred);
+	state = nfs4_do_open(dir, &path, fmode, nd->intent.open.flags, &attr, cred);
 	put_rpccred(cred);
 	if (IS_ERR(state)) {
 		if (PTR_ERR(state) == -ENOENT) {
@@ -1424,7 +1536,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 		path.dentry = res;
 	nfs_set_verifier(path.dentry, nfs_save_change_attribute(dir));
 	nfs_unblock_sillyrename(parent);
-	nfs4_intent_set_file(nd, &path, state);
+	nfs4_intent_set_file(nd, &path, state, fmode);
 	return res;
 }
 
@@ -1437,11 +1549,12 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st
 	};
 	struct rpc_cred *cred;
 	struct nfs4_state *state;
+	fmode_t fmode = openflags & (FMODE_READ | FMODE_WRITE);
 
 	cred = rpc_lookup_cred();
 	if (IS_ERR(cred))
 		return PTR_ERR(cred);
-	state = nfs4_do_open(dir, &path, openflags, NULL, cred);
+	state = nfs4_do_open(dir, &path, fmode, openflags, NULL, cred);
 	put_rpccred(cred);
 	if (IS_ERR(state)) {
 		switch (PTR_ERR(state)) {
@@ -1458,10 +1571,10 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st
 	}
 	if (state->inode == dentry->d_inode) {
 		nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
-		nfs4_intent_set_file(nd, &path, state);
+		nfs4_intent_set_file(nd, &path, state, fmode);
 		return 1;
 	}
-	nfs4_close_sync(&path, state, openflags);
+	nfs4_close_sync(&path, state, fmode);
 out_drop:
 	d_drop(dentry);
 	return 0;
@@ -1887,6 +2000,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 	};
 	struct nfs4_state *state;
 	struct rpc_cred *cred;
+	fmode_t fmode = flags & (FMODE_READ | FMODE_WRITE);
 	int status = 0;
 
 	cred = rpc_lookup_cred();
@@ -1894,7 +2008,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		status = PTR_ERR(cred);
 		goto out;
 	}
-	state = nfs4_do_open(dir, &path, flags, sattr, cred);
+	state = nfs4_do_open(dir, &path, fmode, flags, sattr, cred);
 	d_drop(dentry);
 	if (IS_ERR(state)) {
 		status = PTR_ERR(state);
@@ -1910,9 +2024,9 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		nfs_post_op_update_inode(state->inode, &fattr);
 	}
 	if (status == 0 && (nd->flags & LOOKUP_OPEN) != 0)
-		status = nfs4_intent_set_file(nd, &path, state);
+		status = nfs4_intent_set_file(nd, &path, state, fmode);
 	else
-		nfs4_close_sync(&path, state, flags);
+		nfs4_close_sync(&path, state, fmode);
 out_putcred:
 	put_rpccred(cred);
 out:
@@ -1974,7 +2088,7 @@ static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
 {
 	struct nfs_removeres *res = task->tk_msg.rpc_resp;
 
-	if (nfs4_async_handle_error(task, res->server) == -EAGAIN)
+	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
 		return 0;
 	update_changeattr(dir, &res->cinfo);
 	nfs_post_op_update_inode(dir, &res->dir_attr);
@@ -2402,7 +2516,7 @@ static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
 {
 	struct nfs_server *server = NFS_SERVER(data->inode);
 
-	if (nfs4_async_handle_error(task, server) == -EAGAIN) {
+	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
 		rpc_restart_call(task);
 		return -EAGAIN;
 	}
@@ -2423,7 +2537,7 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
 {
 	struct inode *inode = data->inode;
 	
-	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
+	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
 		rpc_restart_call(task);
 		return -EAGAIN;
 	}
@@ -2449,7 +2563,7 @@ static int nfs4_commit_done(struct rpc_task *task, struct nfs_write_data *data)
 {
 	struct inode *inode = data->inode;
 	
-	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
+	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
 		rpc_restart_call(task);
 		return -EAGAIN;
 	}
@@ -2742,19 +2856,25 @@ static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen
 }
 
 static int
-nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
+nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
 {
 	struct nfs_client *clp = server->nfs_client;
 
 	if (!clp || task->tk_status >= 0)
 		return 0;
 	switch(task->tk_status) {
+		case -NFS4ERR_ADMIN_REVOKED:
+		case -NFS4ERR_BAD_STATEID:
+		case -NFS4ERR_OPENMODE:
+			if (state == NULL)
+				break;
+			nfs4_state_mark_reclaim_nograce(clp, state);
 		case -NFS4ERR_STALE_CLIENTID:
 		case -NFS4ERR_STALE_STATEID:
 		case -NFS4ERR_EXPIRED:
 			rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL);
 			nfs4_schedule_state_recovery(clp);
-			if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
+			if (test_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) == 0)
 				rpc_wake_up_queued_task(&clp->cl_rpcwaitq, task);
 			task->tk_status = 0;
 			return -EAGAIN;
@@ -2772,79 +2892,6 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
 	return 0;
 }
 
-static int nfs4_wait_bit_killable(void *word)
-{
-	if (fatal_signal_pending(current))
-		return -ERESTARTSYS;
-	schedule();
-	return 0;
-}
-
-static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp)
-{
-	int res;
-
-	might_sleep();
-
-	rwsem_acquire(&clp->cl_sem.dep_map, 0, 0, _RET_IP_);
-
-	res = wait_on_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER,
-			nfs4_wait_bit_killable, TASK_KILLABLE);
-
-	rwsem_release(&clp->cl_sem.dep_map, 1, _RET_IP_);
-	return res;
-}
-
-static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
-{
-	int res = 0;
-
-	might_sleep();
-
-	if (*timeout <= 0)
-		*timeout = NFS4_POLL_RETRY_MIN;
-	if (*timeout > NFS4_POLL_RETRY_MAX)
-		*timeout = NFS4_POLL_RETRY_MAX;
-	schedule_timeout_killable(*timeout);
-	if (fatal_signal_pending(current))
-		res = -ERESTARTSYS;
-	*timeout <<= 1;
-	return res;
-}
-
-/* This is the error handling routine for processes that are allowed
- * to sleep.
- */
-static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
-{
-	struct nfs_client *clp = server->nfs_client;
-	int ret = errorcode;
-
-	exception->retry = 0;
-	switch(errorcode) {
-		case 0:
-			return 0;
-		case -NFS4ERR_STALE_CLIENTID:
-		case -NFS4ERR_STALE_STATEID:
-		case -NFS4ERR_EXPIRED:
-			nfs4_schedule_state_recovery(clp);
-			ret = nfs4_wait_clnt_recover(server->client, clp);
-			if (ret == 0)
-				exception->retry = 1;
-			break;
-		case -NFS4ERR_FILE_OPEN:
-		case -NFS4ERR_GRACE:
-		case -NFS4ERR_DELAY:
-			ret = nfs4_delay(server->client, &exception->timeout);
-			if (ret != 0)
-				break;
-		case -NFS4ERR_OLD_STATEID:
-			exception->retry = 1;
-	}
-	/* We failed to handle the error */
-	return nfs4_map_errors(ret);
-}
-
 int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short port, struct rpc_cred *cred)
 {
 	nfs4_verifier sc_verifier;
@@ -2916,7 +2963,6 @@ static int _nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cre
 		spin_lock(&clp->cl_lock);
 		clp->cl_lease_time = fsinfo.lease_time * HZ;
 		clp->cl_last_renewal = now;
-		clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
 		spin_unlock(&clp->cl_lock);
 	}
 	return status;
@@ -3074,7 +3120,6 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock
 	struct nfs4_lock_state *lsp;
 	int status;
 
-	down_read(&clp->cl_sem);
 	arg.lock_owner.clientid = clp->cl_clientid;
 	status = nfs4_set_lock_state(state, request);
 	if (status != 0)
@@ -3091,7 +3136,6 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock
 	}
 	request->fl_ops->fl_release_private(request);
 out:
-	up_read(&clp->cl_sem);
 	return status;
 }
 
@@ -3181,11 +3225,13 @@ static void nfs4_locku_done(struct rpc_task *task, void *data)
 					sizeof(calldata->lsp->ls_stateid.data));
 			renew_lease(calldata->server, calldata->timestamp);
 			break;
+		case -NFS4ERR_BAD_STATEID:
+		case -NFS4ERR_OLD_STATEID:
 		case -NFS4ERR_STALE_STATEID:
 		case -NFS4ERR_EXPIRED:
 			break;
 		default:
-			if (nfs4_async_handle_error(task, calldata->server) == -EAGAIN)
+			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
 				rpc_restart_call(task);
 	}
 }
@@ -3248,6 +3294,7 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
 
 static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
 {
+	struct nfs_inode *nfsi = NFS_I(state->inode);
 	struct nfs_seqid *seqid;
 	struct nfs4_lock_state *lsp;
 	struct rpc_task *task;
@@ -3257,8 +3304,12 @@ static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *
 	status = nfs4_set_lock_state(state, request);
 	/* Unlock _before_ we do the RPC call */
 	request->fl_flags |= FL_EXISTS;
-	if (do_vfs_lock(request->fl_file, request) == -ENOENT)
+	down_read(&nfsi->rwsem);
+	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
+		up_read(&nfsi->rwsem);
 		goto out;
+	}
+	up_read(&nfsi->rwsem);
 	if (status != 0)
 		goto out;
 	/* Is this a delegated lock? */
@@ -3484,7 +3535,7 @@ static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request
 
 static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
 {
-	struct nfs_client *clp = state->owner->so_client;
+	struct nfs_inode *nfsi = NFS_I(state->inode);
 	unsigned char fl_flags = request->fl_flags;
 	int status;
 
@@ -3496,19 +3547,13 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock
 	status = do_vfs_lock(request->fl_file, request);
 	if (status < 0)
 		goto out;
-	down_read(&clp->cl_sem);
+	down_read(&nfsi->rwsem);
 	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
-		struct nfs_inode *nfsi = NFS_I(state->inode);
 		/* Yes: cache locks! */
-		down_read(&nfsi->rwsem);
 		/* ...but avoid races with delegation recall... */
-		if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
-			request->fl_flags = fl_flags & ~FL_SLEEP;
-			status = do_vfs_lock(request->fl_file, request);
-			up_read(&nfsi->rwsem);
-			goto out_unlock;
-		}
-		up_read(&nfsi->rwsem);
+		request->fl_flags = fl_flags & ~FL_SLEEP;
+		status = do_vfs_lock(request->fl_file, request);
+		goto out_unlock;
 	}
 	status = _nfs4_do_setlk(state, cmd, request, 0);
 	if (status != 0)
@@ -3518,7 +3563,7 @@ static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock
 	if (do_vfs_lock(request->fl_file, request) < 0)
 		printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __func__);
 out_unlock:
-	up_read(&clp->cl_sem);
+	up_read(&nfsi->rwsem);
 out:
 	request->fl_flags = fl_flags;
 	return status;
@@ -3664,11 +3709,15 @@ int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
 }
 
 struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops = {
+	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
+	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
 	.recover_open	= nfs4_open_reclaim,
 	.recover_lock	= nfs4_lock_reclaim,
 };
 
-struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops = {
+struct nfs4_state_recovery_ops nfs4_nograce_recovery_ops = {
+	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
+	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
 	.recover_open	= nfs4_open_expired,
 	.recover_lock	= nfs4_lock_expired,
 };
diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c
index 3305acbbe2a..f524e932ff7 100644
--- a/fs/nfs/nfs4renewd.c
+++ b/fs/nfs/nfs4renewd.c
@@ -65,7 +65,6 @@ nfs4_renew_state(struct work_struct *work)
 	long lease, timeout;
 	unsigned long last, now;
 
-	down_read(&clp->cl_sem);
 	dprintk("%s: start\n", __func__);
 	/* Are there any active superblocks? */
 	if (list_empty(&clp->cl_superblocks))
@@ -77,17 +76,19 @@ nfs4_renew_state(struct work_struct *work)
 	timeout = (2 * lease) / 3 + (long)last - (long)now;
 	/* Are we close to a lease timeout? */
 	if (time_after(now, last + lease/3)) {
-		cred = nfs4_get_renew_cred(clp);
+		cred = nfs4_get_renew_cred_locked(clp);
+		spin_unlock(&clp->cl_lock);
 		if (cred == NULL) {
-			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
-			spin_unlock(&clp->cl_lock);
+			if (list_empty(&clp->cl_delegations)) {
+				set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+				goto out;
+			}
 			nfs_expire_all_delegations(clp);
-			goto out;
+		} else {
+			/* Queue an asynchronous RENEW. */
+			nfs4_proc_async_renew(clp, cred);
+			put_rpccred(cred);
 		}
-		spin_unlock(&clp->cl_lock);
-		/* Queue an asynchronous RENEW. */
-		nfs4_proc_async_renew(clp, cred);
-		put_rpccred(cred);
 		timeout = (2 * lease) / 3;
 		spin_lock(&clp->cl_lock);
 	} else
@@ -100,12 +101,11 @@ nfs4_renew_state(struct work_struct *work)
 	cancel_delayed_work(&clp->cl_renewd);
 	schedule_delayed_work(&clp->cl_renewd, timeout);
 	spin_unlock(&clp->cl_lock);
+	nfs_expire_unreferenced_delegations(clp);
 out:
-	up_read(&clp->cl_sem);
 	dprintk("%s: done\n", __func__);
 }
 
-/* Must be called with clp->cl_sem locked for writes */
 void
 nfs4_schedule_state_renewal(struct nfs_client *clp)
 {
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 401ef8b28f9..2022fe47966 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -71,14 +71,12 @@ static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred)
 	return status;
 }
 
-static struct rpc_cred *nfs4_get_machine_cred(struct nfs_client *clp)
+static struct rpc_cred *nfs4_get_machine_cred_locked(struct nfs_client *clp)
 {
 	struct rpc_cred *cred = NULL;
 
-	spin_lock(&clp->cl_lock);
 	if (clp->cl_machine_cred != NULL)
 		cred = get_rpccred(clp->cl_machine_cred);
-	spin_unlock(&clp->cl_lock);
 	return cred;
 }
 
@@ -94,7 +92,7 @@ static void nfs4_clear_machine_cred(struct nfs_client *clp)
 		put_rpccred(cred);
 }
 
-struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
+struct rpc_cred *nfs4_get_renew_cred_locked(struct nfs_client *clp)
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
@@ -110,13 +108,24 @@ struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
 	return cred;
 }
 
+static struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
+{
+	struct rpc_cred *cred;
+
+	spin_lock(&clp->cl_lock);
+	cred = nfs4_get_renew_cred_locked(clp);
+	spin_unlock(&clp->cl_lock);
+	return cred;
+}
+
 static struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp)
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
 	struct rpc_cred *cred;
 
-	cred = nfs4_get_machine_cred(clp);
+	spin_lock(&clp->cl_lock);
+	cred = nfs4_get_machine_cred_locked(clp);
 	if (cred != NULL)
 		goto out;
 	pos = rb_first(&clp->cl_state_owners);
@@ -125,6 +134,7 @@ static struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp)
 		cred = get_rpccred(sp->so_cred);
 	}
 out:
+	spin_unlock(&clp->cl_lock);
 	return cred;
 }
 
@@ -295,10 +305,6 @@ nfs4_drop_state_owner(struct nfs4_state_owner *sp)
 	}
 }
 
-/*
- * Note: must be called with clp->cl_sem held in order to prevent races
- *       with reboot recovery!
- */
 struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct rpc_cred *cred)
 {
 	struct nfs_client *clp = server->nfs_client;
@@ -327,10 +333,6 @@ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct
 	return sp;
 }
 
-/*
- * Must be called with clp->cl_sem held in order to avoid races
- * with state recovery...
- */
 void nfs4_put_state_owner(struct nfs4_state_owner *sp)
 {
 	struct nfs_client *clp = sp->so_client;
@@ -361,18 +363,18 @@ nfs4_alloc_open_state(void)
 }
 
 void
-nfs4_state_set_mode_locked(struct nfs4_state *state, mode_t mode)
+nfs4_state_set_mode_locked(struct nfs4_state *state, fmode_t fmode)
 {
-	if (state->state == mode)
+	if (state->state == fmode)
 		return;
 	/* NB! List reordering - see the reclaim code for why.  */
-	if ((mode & FMODE_WRITE) != (state->state & FMODE_WRITE)) {
-		if (mode & FMODE_WRITE)
+	if ((fmode & FMODE_WRITE) != (state->state & FMODE_WRITE)) {
+		if (fmode & FMODE_WRITE)
 			list_move(&state->open_states, &state->owner->so_states);
 		else
 			list_move_tail(&state->open_states, &state->owner->so_states);
 	}
-	state->state = mode;
+	state->state = fmode;
 }
 
 static struct nfs4_state *
@@ -432,10 +434,6 @@ out:
 	return state;
 }
 
-/*
- * Beware! Caller must be holding exactly one
- * reference to clp->cl_sem!
- */
 void nfs4_put_open_state(struct nfs4_state *state)
 {
 	struct inode *inode = state->inode;
@@ -456,16 +454,16 @@ void nfs4_put_open_state(struct nfs4_state *state)
 /*
  * Close the current file.
  */
-static void __nfs4_close(struct path *path, struct nfs4_state *state, mode_t mode, int wait)
+static void __nfs4_close(struct path *path, struct nfs4_state *state, fmode_t fmode, int wait)
 {
 	struct nfs4_state_owner *owner = state->owner;
 	int call_close = 0;
-	int newstate;
+	fmode_t newstate;
 
 	atomic_inc(&owner->so_count);
 	/* Protect against nfs4_find_state() */
 	spin_lock(&owner->so_lock);
-	switch (mode & (FMODE_READ | FMODE_WRITE)) {
+	switch (fmode & (FMODE_READ | FMODE_WRITE)) {
 		case FMODE_READ:
 			state->n_rdonly--;
 			break;
@@ -500,14 +498,14 @@ static void __nfs4_close(struct path *path, struct nfs4_state *state, mode_t mod
 		nfs4_do_close(path, state, wait);
 }
 
-void nfs4_close_state(struct path *path, struct nfs4_state *state, mode_t mode)
+void nfs4_close_state(struct path *path, struct nfs4_state *state, fmode_t fmode)
 {
-	__nfs4_close(path, state, mode, 0);
+	__nfs4_close(path, state, fmode, 0);
 }
 
-void nfs4_close_sync(struct path *path, struct nfs4_state *state, mode_t mode)
+void nfs4_close_sync(struct path *path, struct nfs4_state *state, fmode_t fmode)
 {
-	__nfs4_close(path, state, mode, 1);
+	__nfs4_close(path, state, fmode, 1);
 }
 
 /*
@@ -568,7 +566,6 @@ static void nfs4_free_lock_state(struct nfs4_lock_state *lsp)
  * Return a compatible lock_state. If no initialized lock_state structure
  * exists, return an uninitialized one.
  *
- * The caller must be holding clp->cl_sem
  */
 static struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_owner_t owner)
 {
@@ -770,32 +767,34 @@ unlock:
 	return status;
 }
 
-static int reclaimer(void *);
+static int nfs4_run_state_manager(void *);
 
-static inline void nfs4_clear_recover_bit(struct nfs_client *clp)
+static void nfs4_clear_state_manager_bit(struct nfs_client *clp)
 {
 	smp_mb__before_clear_bit();
-	clear_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state);
+	clear_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state);
 	smp_mb__after_clear_bit();
-	wake_up_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER);
+	wake_up_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING);
 	rpc_wake_up(&clp->cl_rpcwaitq);
 }
 
 /*
- * State recovery routine
+ * Schedule the nfs_client asynchronous state management routine
  */
-static void nfs4_recover_state(struct nfs_client *clp)
+void nfs4_schedule_state_manager(struct nfs_client *clp)
 {
 	struct task_struct *task;
 
+	if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
+		return;
 	__module_get(THIS_MODULE);
 	atomic_inc(&clp->cl_count);
-	task = kthread_run(reclaimer, clp, "%s-reclaim",
+	task = kthread_run(nfs4_run_state_manager, clp, "%s-manager",
 				rpc_peeraddr2str(clp->cl_rpcclient,
 							RPC_DISPLAY_ADDR));
 	if (!IS_ERR(task))
 		return;
-	nfs4_clear_recover_bit(clp);
+	nfs4_clear_state_manager_bit(clp);
 	nfs_put_client(clp);
 	module_put(THIS_MODULE);
 }
@@ -807,16 +806,42 @@ void nfs4_schedule_state_recovery(struct nfs_client *clp)
 {
 	if (!clp)
 		return;
-	if (test_and_set_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
-		nfs4_recover_state(clp);
+	if (!test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
+		set_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state);
+	nfs4_schedule_state_manager(clp);
 }
 
-static int nfs4_reclaim_locks(struct nfs4_state_recovery_ops *ops, struct nfs4_state *state)
+static int nfs4_state_mark_reclaim_reboot(struct nfs_client *clp, struct nfs4_state *state)
+{
+
+	set_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags);
+	/* Don't recover state that expired before the reboot */
+	if (test_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags)) {
+		clear_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags);
+		return 0;
+	}
+	set_bit(NFS_OWNER_RECLAIM_REBOOT, &state->owner->so_flags);
+	set_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state);
+	return 1;
+}
+
+int nfs4_state_mark_reclaim_nograce(struct nfs_client *clp, struct nfs4_state *state)
+{
+	set_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags);
+	clear_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags);
+	set_bit(NFS_OWNER_RECLAIM_NOGRACE, &state->owner->so_flags);
+	set_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
+	return 1;
+}
+
+static int nfs4_reclaim_locks(struct nfs4_state *state, const struct nfs4_state_recovery_ops *ops)
 {
 	struct inode *inode = state->inode;
+	struct nfs_inode *nfsi = NFS_I(inode);
 	struct file_lock *fl;
 	int status = 0;
 
+	down_write(&nfsi->rwsem);
 	for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
 		if (!(fl->fl_flags & (FL_POSIX|FL_FLOCK)))
 			continue;
@@ -839,12 +864,14 @@ static int nfs4_reclaim_locks(struct nfs4_state_recovery_ops *ops, struct nfs4_s
 				goto out_err;
 		}
 	}
+	up_write(&nfsi->rwsem);
 	return 0;
 out_err:
+	up_write(&nfsi->rwsem);
 	return status;
 }
 
-static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct nfs4_state_owner *sp)
+static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp, const struct nfs4_state_recovery_ops *ops)
 {
 	struct nfs4_state *state;
 	struct nfs4_lock_state *lock;
@@ -858,28 +885,34 @@ static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct n
 	 * recovering after a network partition or a reboot from a
 	 * server that doesn't support a grace period.
 	 */
+restart:
+	spin_lock(&sp->so_lock);
 	list_for_each_entry(state, &sp->so_states, open_states) {
+		if (!test_and_clear_bit(ops->state_flag_bit, &state->flags))
+			continue;
 		if (state->state == 0)
 			continue;
+		atomic_inc(&state->count);
+		spin_unlock(&sp->so_lock);
 		status = ops->recover_open(sp, state);
 		if (status >= 0) {
-			status = nfs4_reclaim_locks(ops, state);
-			if (status < 0)
-				goto out_err;
-			list_for_each_entry(lock, &state->lock_states, ls_locks) {
-				if (!(lock->ls_flags & NFS_LOCK_INITIALIZED))
-					printk("%s: Lock reclaim failed!\n",
+			status = nfs4_reclaim_locks(state, ops);
+			if (status >= 0) {
+				list_for_each_entry(lock, &state->lock_states, ls_locks) {
+					if (!(lock->ls_flags & NFS_LOCK_INITIALIZED))
+						printk("%s: Lock reclaim failed!\n",
 							__func__);
+				}
+				nfs4_put_open_state(state);
+				goto restart;
 			}
-			continue;
 		}
 		switch (status) {
 			default:
 				printk(KERN_ERR "%s: unhandled error %d. Zeroing state\n",
 						__func__, status);
 			case -ENOENT:
-			case -NFS4ERR_RECLAIM_BAD:
-			case -NFS4ERR_RECLAIM_CONFLICT:
+			case -ESTALE:
 				/*
 				 * Open state on this file cannot be recovered
 				 * All we can do is revert to using the zero stateid.
@@ -889,84 +922,176 @@ static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct n
 				/* Mark the file as being 'closed' */
 				state->state = 0;
 				break;
+			case -NFS4ERR_RECLAIM_BAD:
+			case -NFS4ERR_RECLAIM_CONFLICT:
+				nfs4_state_mark_reclaim_nograce(sp->so_client, state);
+				break;
 			case -NFS4ERR_EXPIRED:
 			case -NFS4ERR_NO_GRACE:
+				nfs4_state_mark_reclaim_nograce(sp->so_client, state);
 			case -NFS4ERR_STALE_CLIENTID:
 				goto out_err;
 		}
+		nfs4_put_open_state(state);
+		goto restart;
 	}
+	spin_unlock(&sp->so_lock);
 	return 0;
 out_err:
+	nfs4_put_open_state(state);
 	return status;
 }
 
-static void nfs4_state_mark_reclaim(struct nfs_client *clp)
+static void nfs4_clear_open_state(struct nfs4_state *state)
+{
+	struct nfs4_lock_state *lock;
+
+	clear_bit(NFS_DELEGATED_STATE, &state->flags);
+	clear_bit(NFS_O_RDONLY_STATE, &state->flags);
+	clear_bit(NFS_O_WRONLY_STATE, &state->flags);
+	clear_bit(NFS_O_RDWR_STATE, &state->flags);
+	list_for_each_entry(lock, &state->lock_states, ls_locks) {
+		lock->ls_seqid.flags = 0;
+		lock->ls_flags &= ~NFS_LOCK_INITIALIZED;
+	}
+}
+
+static void nfs4_state_mark_reclaim_helper(struct nfs_client *clp, int (*mark_reclaim)(struct nfs_client *clp, struct nfs4_state *state))
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
 	struct nfs4_state *state;
-	struct nfs4_lock_state *lock;
 
 	/* Reset all sequence ids to zero */
 	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
 		sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
-		sp->so_seqid.counter = 0;
 		sp->so_seqid.flags = 0;
 		spin_lock(&sp->so_lock);
 		list_for_each_entry(state, &sp->so_states, open_states) {
-			clear_bit(NFS_DELEGATED_STATE, &state->flags);
-			clear_bit(NFS_O_RDONLY_STATE, &state->flags);
-			clear_bit(NFS_O_WRONLY_STATE, &state->flags);
-			clear_bit(NFS_O_RDWR_STATE, &state->flags);
-			list_for_each_entry(lock, &state->lock_states, ls_locks) {
-				lock->ls_seqid.counter = 0;
-				lock->ls_seqid.flags = 0;
-				lock->ls_flags &= ~NFS_LOCK_INITIALIZED;
-			}
+			if (mark_reclaim(clp, state))
+				nfs4_clear_open_state(state);
 		}
 		spin_unlock(&sp->so_lock);
 	}
 }
 
-static int reclaimer(void *ptr)
+static void nfs4_state_start_reclaim_reboot(struct nfs_client *clp)
+{
+	/* Mark all delegations for reclaim */
+	nfs_delegation_mark_reclaim(clp);
+	nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_reboot);
+}
+
+static void nfs4_state_end_reclaim_reboot(struct nfs_client *clp)
 {
-	struct nfs_client *clp = ptr;
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
-	struct nfs4_state_recovery_ops *ops;
-	struct rpc_cred *cred;
+	struct nfs4_state *state;
+
+	if (!test_and_clear_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state))
+		return;
+
+	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
+		sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
+		spin_lock(&sp->so_lock);
+		list_for_each_entry(state, &sp->so_states, open_states) {
+			if (!test_and_clear_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags))
+				continue;
+			nfs4_state_mark_reclaim_nograce(clp, state);
+		}
+		spin_unlock(&sp->so_lock);
+	}
+
+	nfs_delegation_reap_unclaimed(clp);
+}
+
+static void nfs_delegation_clear_all(struct nfs_client *clp)
+{
+	nfs_delegation_mark_reclaim(clp);
+	nfs_delegation_reap_unclaimed(clp);
+}
+
+static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
+{
+	nfs_delegation_clear_all(clp);
+	nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
+}
+
+static void nfs4_state_end_reclaim_nograce(struct nfs_client *clp)
+{
+	clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
+}
+
+static void nfs4_recovery_handle_error(struct nfs_client *clp, int error)
+{
+	switch (error) {
+		case -NFS4ERR_CB_PATH_DOWN:
+			nfs_handle_cb_pathdown(clp);
+			break;
+		case -NFS4ERR_STALE_CLIENTID:
+		case -NFS4ERR_LEASE_MOVED:
+			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+			nfs4_state_start_reclaim_reboot(clp);
+			break;
+		case -NFS4ERR_EXPIRED:
+			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+			nfs4_state_start_reclaim_nograce(clp);
+	}
+}
+
+static int nfs4_do_reclaim(struct nfs_client *clp, const struct nfs4_state_recovery_ops *ops)
+{
+	struct rb_node *pos;
 	int status = 0;
 
-	allow_signal(SIGKILL);
+restart:
+	spin_lock(&clp->cl_lock);
+	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
+		struct nfs4_state_owner *sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
+		if (!test_and_clear_bit(ops->owner_flag_bit, &sp->so_flags))
+			continue;
+		atomic_inc(&sp->so_count);
+		spin_unlock(&clp->cl_lock);
+		status = nfs4_reclaim_open_state(sp, ops);
+		if (status < 0) {
+			set_bit(ops->owner_flag_bit, &sp->so_flags);
+			nfs4_put_state_owner(sp);
+			nfs4_recovery_handle_error(clp, status);
+			return status;
+		}
+		nfs4_put_state_owner(sp);
+		goto restart;
+	}
+	spin_unlock(&clp->cl_lock);
+	return status;
+}
 
-	/* Ensure exclusive access to NFSv4 state */
-	down_write(&clp->cl_sem);
-	/* Are there any NFS mounts out there? */
-	if (list_empty(&clp->cl_superblocks))
-		goto out;
-restart_loop:
-	ops = &nfs4_network_partition_recovery_ops;
-	/* Are there any open files on this volume? */
+static int nfs4_check_lease(struct nfs_client *clp)
+{
+	struct rpc_cred *cred;
+	int status = -NFS4ERR_EXPIRED;
+
+	/* Is the client already known to have an expired lease? */
+	if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
+		return 0;
 	cred = nfs4_get_renew_cred(clp);
-	if (cred != NULL) {
-		/* Yes there are: try to renew the old lease */
-		status = nfs4_proc_renew(clp, cred);
-		put_rpccred(cred);
-		switch (status) {
-			case 0:
-			case -NFS4ERR_CB_PATH_DOWN:
-				goto out;
-			case -NFS4ERR_STALE_CLIENTID:
-			case -NFS4ERR_LEASE_MOVED:
-				ops = &nfs4_reboot_recovery_ops;
-		}
-	} else {
-		/* "reboot" to ensure we clear all state on the server */
-		clp->cl_boot_time = CURRENT_TIME;
+	if (cred == NULL) {
+		cred = nfs4_get_setclientid_cred(clp);
+		if (cred == NULL)
+			goto out;
 	}
-	/* We're going to have to re-establish a clientid */
-	nfs4_state_mark_reclaim(clp);
-	status = -ENOENT;
+	status = nfs4_proc_renew(clp, cred);
+	put_rpccred(cred);
+out:
+	nfs4_recovery_handle_error(clp, status);
+	return status;
+}
+
+static int nfs4_reclaim_lease(struct nfs_client *clp)
+{
+	struct rpc_cred *cred;
+	int status = -ENOENT;
+
 	cred = nfs4_get_setclientid_cred(clp);
 	if (cred != NULL) {
 		status = nfs4_init_client(clp, cred);
@@ -974,42 +1099,90 @@ restart_loop:
 		/* Handle case where the user hasn't set up machine creds */
 		if (status == -EACCES && cred == clp->cl_machine_cred) {
 			nfs4_clear_machine_cred(clp);
-			goto restart_loop;
+			status = -EAGAIN;
 		}
 	}
-	if (status)
-		goto out_error;
-	/* Mark all delegations for reclaim */
-	nfs_delegation_mark_reclaim(clp);
-	/* Note: list is protected by exclusive lock on cl->cl_sem */
-	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
-		sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
-		status = nfs4_reclaim_open_state(ops, sp);
-		if (status < 0) {
-			if (status == -NFS4ERR_NO_GRACE) {
-				ops = &nfs4_network_partition_recovery_ops;
-				status = nfs4_reclaim_open_state(ops, sp);
+	return status;
+}
+
+static void nfs4_state_manager(struct nfs_client *clp)
+{
+	int status = 0;
+
+	/* Ensure exclusive access to NFSv4 state */
+	for(;;) {
+		if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) {
+			/* We're going to have to re-establish a clientid */
+			status = nfs4_reclaim_lease(clp);
+			if (status) {
+				set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+				if (status == -EAGAIN)
+					continue;
+				goto out_error;
 			}
+			clear_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state);
+		}
+
+		if (test_and_clear_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state)) {
+			status = nfs4_check_lease(clp);
+			if (status != 0)
+				continue;
+		}
+
+		/* First recover reboot state... */
+		if (test_and_clear_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state)) {
+			status = nfs4_do_reclaim(clp, &nfs4_reboot_recovery_ops);
 			if (status == -NFS4ERR_STALE_CLIENTID)
-				goto restart_loop;
-			if (status == -NFS4ERR_EXPIRED)
-				goto restart_loop;
+				continue;
+			nfs4_state_end_reclaim_reboot(clp);
+			continue;
+		}
+
+		/* Now recover expired state... */
+		if (test_and_clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state)) {
+			status = nfs4_do_reclaim(clp, &nfs4_nograce_recovery_ops);
+			if (status < 0) {
+				set_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
+				if (status == -NFS4ERR_STALE_CLIENTID)
+					continue;
+				if (status == -NFS4ERR_EXPIRED)
+					continue;
+				goto out_error;
+			} else
+				nfs4_state_end_reclaim_nograce(clp);
+			continue;
 		}
+
+		if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
+			nfs_client_return_marked_delegations(clp);
+			continue;
+		}
+
+		nfs4_clear_state_manager_bit(clp);
+		/* Did we race with an attempt to give us more work? */
+		if (clp->cl_state == 0)
+			break;
+		if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
+			break;
 	}
-	nfs_delegation_reap_unclaimed(clp);
-out:
-	up_write(&clp->cl_sem);
-	if (status == -NFS4ERR_CB_PATH_DOWN)
-		nfs_handle_cb_pathdown(clp);
-	nfs4_clear_recover_bit(clp);
+	return;
+out_error:
+	printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
+			" with error %d\n", clp->cl_hostname, -status);
+	if (test_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state))
+		nfs4_state_end_reclaim_reboot(clp);
+	nfs4_clear_state_manager_bit(clp);
+}
+
+static int nfs4_run_state_manager(void *ptr)
+{
+	struct nfs_client *clp = ptr;
+
+	allow_signal(SIGKILL);
+	nfs4_state_manager(clp);
 	nfs_put_client(clp);
 	module_put_and_exit(0);
 	return 0;
-out_error:
-	printk(KERN_WARNING "Error: state recovery failed on NFSv4 server %s"
-			" with error %d\n", clp->cl_hostname, -status);
-	set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
-	goto out;
 }
 
 /*
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index b916297d233..d1e4c8f8a0a 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -8,7 +8,7 @@
  *
  *  Kendrick Smith <kmsmith@umich.edu>
  *  Andy Adamson   <andros@umich.edu>
- * 
+ *
  *  Redistribution and use in source and binary forms, with or without
  *  modification, are permitted provided that the following conditions
  *  are met:
@@ -67,7 +67,7 @@ static int nfs4_stat_to_errno(int);
 #define NFS4_MAXTAGLEN		0
 #endif
 
-/* lock,open owner id: 
+/* lock,open owner id:
  * we currently use size 2 (u64) out of (NFS4_OPAQUE_LIMIT  >> 2)
  */
 #define open_owner_id_maxsz	(1 + 4)
@@ -541,6 +541,7 @@ static struct {
 struct compound_hdr {
 	int32_t		status;
 	uint32_t	nops;
+	__be32 *	nops_p;
 	uint32_t	taglen;
 	char *		tag;
 };
@@ -578,7 +579,7 @@ static void encode_string(struct xdr_stream *xdr, unsigned int len, const char *
 	xdr_encode_opaque(p, str, len);
 }
 
-static int encode_compound_hdr(struct xdr_stream *xdr, struct compound_hdr *hdr)
+static void encode_compound_hdr(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -588,8 +589,13 @@ static int encode_compound_hdr(struct xdr_stream *xdr, struct compound_hdr *hdr)
 	WRITE32(hdr->taglen);
 	WRITEMEM(hdr->tag, hdr->taglen);
 	WRITE32(NFS4_MINOR_VERSION);
+	hdr->nops_p = p;
 	WRITE32(hdr->nops);
-	return 0;
+}
+
+static void encode_nops(struct compound_hdr *hdr)
+{
+	*hdr->nops_p = htonl(hdr->nops);
 }
 
 static void encode_nfs4_verifier(struct xdr_stream *xdr, const nfs4_verifier *verf)
@@ -601,7 +607,7 @@ static void encode_nfs4_verifier(struct xdr_stream *xdr, const nfs4_verifier *ve
 	xdr_encode_opaque_fixed(p, verf->data, NFS4_VERIFIER_SIZE);
 }
 
-static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const struct nfs_server *server)
+static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const struct nfs_server *server)
 {
 	char owner_name[IDMAP_NAMESZ];
 	char owner_group[IDMAP_NAMESZ];
@@ -612,7 +618,6 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s
 	int len;
 	uint32_t bmval0 = 0;
 	uint32_t bmval1 = 0;
-	int status;
 
 	/*
 	 * We reserve enough space to write the entire attribute buffer at once.
@@ -709,7 +714,7 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s
 		bmval1 |= FATTR4_WORD1_TIME_MODIFY_SET;
 		WRITE32(NFS4_SET_TO_SERVER_TIME);
 	}
-	
+
 	/*
 	 * Now we backfill the bitmap and the attribute buffer length.
 	 */
@@ -723,23 +728,20 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s
 	*q++ = htonl(bmval1);
 	*q++ = htonl(len);
 
-	status = 0;
 /* out: */
-	return status;
 }
 
-static int encode_access(struct xdr_stream *xdr, u32 access)
+static void encode_access(struct xdr_stream *xdr, u32 access, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(8);
 	WRITE32(OP_ACCESS);
 	WRITE32(access);
-	
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_close(struct xdr_stream *xdr, const struct nfs_closeargs *arg)
+static void encode_close(struct xdr_stream *xdr, const struct nfs_closeargs *arg, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -747,26 +749,24 @@ static int encode_close(struct xdr_stream *xdr, const struct nfs_closeargs *arg)
 	WRITE32(OP_CLOSE);
 	WRITE32(arg->seqid->sequence->counter);
 	WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE);
-	
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_commit(struct xdr_stream *xdr, const struct nfs_writeargs *args)
+static void encode_commit(struct xdr_stream *xdr, const struct nfs_writeargs *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
-        
-        RESERVE_SPACE(16);
-        WRITE32(OP_COMMIT);
-        WRITE64(args->offset);
-        WRITE32(args->count);
 
-        return 0;
+	RESERVE_SPACE(16);
+	WRITE32(OP_COMMIT);
+	WRITE64(args->offset);
+	WRITE32(args->count);
+	hdr->nops++;
 }
 
-static int encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *create)
+static void encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *create, struct compound_hdr *hdr)
 {
 	__be32 *p;
-	
+
 	RESERVE_SPACE(8);
 	WRITE32(OP_CREATE);
 	WRITE32(create->ftype);
@@ -791,64 +791,62 @@ static int encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *c
 	RESERVE_SPACE(4 + create->name->len);
 	WRITE32(create->name->len);
 	WRITEMEM(create->name->name, create->name->len);
+	hdr->nops++;
 
-	return encode_attrs(xdr, create->attrs, create->server);
+	encode_attrs(xdr, create->attrs, create->server);
 }
 
-static int encode_getattr_one(struct xdr_stream *xdr, uint32_t bitmap)
+static void encode_getattr_one(struct xdr_stream *xdr, uint32_t bitmap, struct compound_hdr *hdr)
 {
-        __be32 *p;
+	__be32 *p;
 
-        RESERVE_SPACE(12);
-        WRITE32(OP_GETATTR);
-        WRITE32(1);
-        WRITE32(bitmap);
-        return 0;
+	RESERVE_SPACE(12);
+	WRITE32(OP_GETATTR);
+	WRITE32(1);
+	WRITE32(bitmap);
+	hdr->nops++;
 }
 
-static int encode_getattr_two(struct xdr_stream *xdr, uint32_t bm0, uint32_t bm1)
+static void encode_getattr_two(struct xdr_stream *xdr, uint32_t bm0, uint32_t bm1, struct compound_hdr *hdr)
 {
-        __be32 *p;
+	__be32 *p;
 
-        RESERVE_SPACE(16);
-        WRITE32(OP_GETATTR);
-        WRITE32(2);
-        WRITE32(bm0);
-        WRITE32(bm1);
-        return 0;
+	RESERVE_SPACE(16);
+	WRITE32(OP_GETATTR);
+	WRITE32(2);
+	WRITE32(bm0);
+	WRITE32(bm1);
+	hdr->nops++;
 }
 
-static int encode_getfattr(struct xdr_stream *xdr, const u32* bitmask)
+static void encode_getfattr(struct xdr_stream *xdr, const u32* bitmask, struct compound_hdr *hdr)
 {
-	return encode_getattr_two(xdr,
-			bitmask[0] & nfs4_fattr_bitmap[0],
-			bitmask[1] & nfs4_fattr_bitmap[1]);
+	encode_getattr_two(xdr, bitmask[0] & nfs4_fattr_bitmap[0],
+			   bitmask[1] & nfs4_fattr_bitmap[1], hdr);
 }
 
-static int encode_fsinfo(struct xdr_stream *xdr, const u32* bitmask)
+static void encode_fsinfo(struct xdr_stream *xdr, const u32* bitmask, struct compound_hdr *hdr)
 {
-	return encode_getattr_two(xdr, bitmask[0] & nfs4_fsinfo_bitmap[0],
-			bitmask[1] & nfs4_fsinfo_bitmap[1]);
+	encode_getattr_two(xdr, bitmask[0] & nfs4_fsinfo_bitmap[0],
+			   bitmask[1] & nfs4_fsinfo_bitmap[1], hdr);
 }
 
-static int encode_fs_locations(struct xdr_stream *xdr, const u32* bitmask)
+static void encode_fs_locations(struct xdr_stream *xdr, const u32* bitmask, struct compound_hdr *hdr)
 {
-	return encode_getattr_two(xdr,
-				  bitmask[0] & nfs4_fs_locations_bitmap[0],
-				  bitmask[1] & nfs4_fs_locations_bitmap[1]);
+	encode_getattr_two(xdr, bitmask[0] & nfs4_fs_locations_bitmap[0],
+			   bitmask[1] & nfs4_fs_locations_bitmap[1], hdr);
 }
 
-static int encode_getfh(struct xdr_stream *xdr)
+static void encode_getfh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(4);
 	WRITE32(OP_GETFH);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_link(struct xdr_stream *xdr, const struct qstr *name)
+static void encode_link(struct xdr_stream *xdr, const struct qstr *name, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -856,8 +854,7 @@ static int encode_link(struct xdr_stream *xdr, const struct qstr *name)
 	WRITE32(OP_LINK);
 	WRITE32(name->len);
 	WRITEMEM(name->name, name->len);
-	
-	return 0;
+	hdr->nops++;
 }
 
 static inline int nfs4_lock_type(struct file_lock *fl, int block)
@@ -878,7 +875,7 @@ static inline uint64_t nfs4_lock_length(struct file_lock *fl)
  * opcode,type,reclaim,offset,length,new_lock_owner = 32
  * open_seqid,open_stateid,lock_seqid,lock_owner.clientid, lock_owner.id = 40
  */
-static int encode_lock(struct xdr_stream *xdr, const struct nfs_lock_args *args)
+static void encode_lock(struct xdr_stream *xdr, const struct nfs_lock_args *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -904,11 +901,10 @@ static int encode_lock(struct xdr_stream *xdr, const struct nfs_lock_args *args)
 		WRITEMEM(args->lock_stateid->data, NFS4_STATEID_SIZE);
 		WRITE32(args->lock_seqid->sequence->counter);
 	}
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_lockt(struct xdr_stream *xdr, const struct nfs_lockt_args *args)
+static void encode_lockt(struct xdr_stream *xdr, const struct nfs_lockt_args *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -921,11 +917,10 @@ static int encode_lockt(struct xdr_stream *xdr, const struct nfs_lockt_args *arg
 	WRITE32(16);
 	WRITEMEM("lock id:", 8);
 	WRITE64(args->lock_owner.id);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_locku(struct xdr_stream *xdr, const struct nfs_locku_args *args)
+static void encode_locku(struct xdr_stream *xdr, const struct nfs_locku_args *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -936,11 +931,10 @@ static int encode_locku(struct xdr_stream *xdr, const struct nfs_locku_args *arg
 	WRITEMEM(args->stateid->data, NFS4_STATEID_SIZE);
 	WRITE64(args->fl->fl_start);
 	WRITE64(nfs4_lock_length(args->fl));
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_lookup(struct xdr_stream *xdr, const struct qstr *name)
+static void encode_lookup(struct xdr_stream *xdr, const struct qstr *name, struct compound_hdr *hdr)
 {
 	int len = name->len;
 	__be32 *p;
@@ -949,27 +943,26 @@ static int encode_lookup(struct xdr_stream *xdr, const struct qstr *name)
 	WRITE32(OP_LOOKUP);
 	WRITE32(len);
 	WRITEMEM(name->name, len);
-
-	return 0;
+	hdr->nops++;
 }
 
-static void encode_share_access(struct xdr_stream *xdr, int open_flags)
+static void encode_share_access(struct xdr_stream *xdr, fmode_t fmode)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(8);
-	switch (open_flags & (FMODE_READ|FMODE_WRITE)) {
-		case FMODE_READ:
-			WRITE32(NFS4_SHARE_ACCESS_READ);
-			break;
-		case FMODE_WRITE:
-			WRITE32(NFS4_SHARE_ACCESS_WRITE);
-			break;
-		case FMODE_READ|FMODE_WRITE:
-			WRITE32(NFS4_SHARE_ACCESS_BOTH);
-			break;
-		default:
-			BUG();
+	switch (fmode & (FMODE_READ|FMODE_WRITE)) {
+	case FMODE_READ:
+		WRITE32(NFS4_SHARE_ACCESS_READ);
+		break;
+	case FMODE_WRITE:
+		WRITE32(NFS4_SHARE_ACCESS_WRITE);
+		break;
+	case FMODE_READ|FMODE_WRITE:
+		WRITE32(NFS4_SHARE_ACCESS_BOTH);
+		break;
+	default:
+		WRITE32(0);
 	}
 	WRITE32(0);		/* for linux, share_deny = 0 always */
 }
@@ -984,7 +977,7 @@ static inline void encode_openhdr(struct xdr_stream *xdr, const struct nfs_opena
 	RESERVE_SPACE(8);
 	WRITE32(OP_OPEN);
 	WRITE32(arg->seqid->sequence->counter);
-	encode_share_access(xdr, arg->open_flags);
+	encode_share_access(xdr, arg->fmode);
 	RESERVE_SPACE(28);
 	WRITE64(arg->clientid);
 	WRITE32(16);
@@ -998,13 +991,13 @@ static inline void encode_createmode(struct xdr_stream *xdr, const struct nfs_op
 
 	RESERVE_SPACE(4);
 	switch(arg->open_flags & O_EXCL) {
-		case 0:
-			WRITE32(NFS4_CREATE_UNCHECKED);
-			encode_attrs(xdr, arg->u.attrs, arg->server);
-			break;
-		default:
-			WRITE32(NFS4_CREATE_EXCLUSIVE);
-			encode_nfs4_verifier(xdr, &arg->u.verifier);
+	case 0:
+		WRITE32(NFS4_CREATE_UNCHECKED);
+		encode_attrs(xdr, arg->u.attrs, arg->server);
+		break;
+	default:
+		WRITE32(NFS4_CREATE_EXCLUSIVE);
+		encode_nfs4_verifier(xdr, &arg->u.verifier);
 	}
 }
 
@@ -1014,33 +1007,33 @@ static void encode_opentype(struct xdr_stream *xdr, const struct nfs_openargs *a
 
 	RESERVE_SPACE(4);
 	switch (arg->open_flags & O_CREAT) {
-		case 0:
-			WRITE32(NFS4_OPEN_NOCREATE);
-			break;
-		default:
-			BUG_ON(arg->claim != NFS4_OPEN_CLAIM_NULL);
-			WRITE32(NFS4_OPEN_CREATE);
-			encode_createmode(xdr, arg);
+	case 0:
+		WRITE32(NFS4_OPEN_NOCREATE);
+		break;
+	default:
+		BUG_ON(arg->claim != NFS4_OPEN_CLAIM_NULL);
+		WRITE32(NFS4_OPEN_CREATE);
+		encode_createmode(xdr, arg);
 	}
 }
 
-static inline void encode_delegation_type(struct xdr_stream *xdr, int delegation_type)
+static inline void encode_delegation_type(struct xdr_stream *xdr, fmode_t delegation_type)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(4);
 	switch (delegation_type) {
-		case 0:
-			WRITE32(NFS4_OPEN_DELEGATE_NONE);
-			break;
-		case FMODE_READ:
-			WRITE32(NFS4_OPEN_DELEGATE_READ);
-			break;
-		case FMODE_WRITE|FMODE_READ:
-			WRITE32(NFS4_OPEN_DELEGATE_WRITE);
-			break;
-		default:
-			BUG();
+	case 0:
+		WRITE32(NFS4_OPEN_DELEGATE_NONE);
+		break;
+	case FMODE_READ:
+		WRITE32(NFS4_OPEN_DELEGATE_READ);
+		break;
+	case FMODE_WRITE|FMODE_READ:
+		WRITE32(NFS4_OPEN_DELEGATE_WRITE);
+		break;
+	default:
+		BUG();
 	}
 }
 
@@ -1053,7 +1046,7 @@ static inline void encode_claim_null(struct xdr_stream *xdr, const struct qstr *
 	encode_string(xdr, name->len, name->name);
 }
 
-static inline void encode_claim_previous(struct xdr_stream *xdr, int type)
+static inline void encode_claim_previous(struct xdr_stream *xdr, fmode_t type)
 {
 	__be32 *p;
 
@@ -1072,27 +1065,27 @@ static inline void encode_claim_delegate_cur(struct xdr_stream *xdr, const struc
 	encode_string(xdr, name->len, name->name);
 }
 
-static int encode_open(struct xdr_stream *xdr, const struct nfs_openargs *arg)
+static void encode_open(struct xdr_stream *xdr, const struct nfs_openargs *arg, struct compound_hdr *hdr)
 {
 	encode_openhdr(xdr, arg);
 	encode_opentype(xdr, arg);
 	switch (arg->claim) {
-		case NFS4_OPEN_CLAIM_NULL:
-			encode_claim_null(xdr, arg->name);
-			break;
-		case NFS4_OPEN_CLAIM_PREVIOUS:
-			encode_claim_previous(xdr, arg->u.delegation_type);
-			break;
-		case NFS4_OPEN_CLAIM_DELEGATE_CUR:
-			encode_claim_delegate_cur(xdr, arg->name, &arg->u.delegation);
-			break;
-		default:
-			BUG();
+	case NFS4_OPEN_CLAIM_NULL:
+		encode_claim_null(xdr, arg->name);
+		break;
+	case NFS4_OPEN_CLAIM_PREVIOUS:
+		encode_claim_previous(xdr, arg->u.delegation_type);
+		break;
+	case NFS4_OPEN_CLAIM_DELEGATE_CUR:
+		encode_claim_delegate_cur(xdr, arg->name, &arg->u.delegation);
+		break;
+	default:
+		BUG();
 	}
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_open_confirm(struct xdr_stream *xdr, const struct nfs_open_confirmargs *arg)
+static void encode_open_confirm(struct xdr_stream *xdr, const struct nfs_open_confirmargs *arg, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1100,11 +1093,10 @@ static int encode_open_confirm(struct xdr_stream *xdr, const struct nfs_open_con
 	WRITE32(OP_OPEN_CONFIRM);
 	WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE);
 	WRITE32(arg->seqid->sequence->counter);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_open_downgrade(struct xdr_stream *xdr, const struct nfs_closeargs *arg)
+static void encode_open_downgrade(struct xdr_stream *xdr, const struct nfs_closeargs *arg, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1112,12 +1104,12 @@ static int encode_open_downgrade(struct xdr_stream *xdr, const struct nfs_closea
 	WRITE32(OP_OPEN_DOWNGRADE);
 	WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE);
 	WRITE32(arg->seqid->sequence->counter);
-	encode_share_access(xdr, arg->open_flags);
-	return 0;
+	encode_share_access(xdr, arg->fmode);
+	hdr->nops++;
 }
 
-static int
-encode_putfh(struct xdr_stream *xdr, const struct nfs_fh *fh)
+static void
+encode_putfh(struct xdr_stream *xdr, const struct nfs_fh *fh, struct compound_hdr *hdr)
 {
 	int len = fh->size;
 	__be32 *p;
@@ -1126,18 +1118,16 @@ encode_putfh(struct xdr_stream *xdr, const struct nfs_fh *fh)
 	WRITE32(OP_PUTFH);
 	WRITE32(len);
 	WRITEMEM(fh->data, len);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_putrootfh(struct xdr_stream *xdr)
+static void encode_putrootfh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
-        __be32 *p;
-        
-        RESERVE_SPACE(4);
-        WRITE32(OP_PUTROOTFH);
+	__be32 *p;
 
-        return 0;
+	RESERVE_SPACE(4);
+	WRITE32(OP_PUTROOTFH);
+	hdr->nops++;
 }
 
 static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context *ctx)
@@ -1153,7 +1143,7 @@ static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context
 		WRITEMEM(zero_stateid.data, NFS4_STATEID_SIZE);
 }
 
-static int encode_read(struct xdr_stream *xdr, const struct nfs_readargs *args)
+static void encode_read(struct xdr_stream *xdr, const struct nfs_readargs *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1165,11 +1155,10 @@ static int encode_read(struct xdr_stream *xdr, const struct nfs_readargs *args)
 	RESERVE_SPACE(12);
 	WRITE64(args->offset);
 	WRITE32(args->count);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg *readdir, struct rpc_rqst *req)
+static void encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg *readdir, struct rpc_rqst *req, struct compound_hdr *hdr)
 {
 	uint32_t attrs[2] = {
 		FATTR4_WORD0_RDATTR_ERROR|FATTR4_WORD0_FILEID,
@@ -1191,6 +1180,7 @@ static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg
 		attrs[1] &= ~FATTR4_WORD1_MOUNTED_ON_FILEID;
 	WRITE32(attrs[0] & readdir->bitmask[0]);
 	WRITE32(attrs[1] & readdir->bitmask[1]);
+	hdr->nops++;
 	dprintk("%s: cookie = %Lu, verifier = %08x:%08x, bitmap = %08x:%08x\n",
 			__func__,
 			(unsigned long long)readdir->cookie,
@@ -1198,21 +1188,18 @@ static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg
 			((u32 *)readdir->verifier.data)[1],
 			attrs[0] & readdir->bitmask[0],
 			attrs[1] & readdir->bitmask[1]);
-
-	return 0;
 }
 
-static int encode_readlink(struct xdr_stream *xdr, const struct nfs4_readlink *readlink, struct rpc_rqst *req)
+static void encode_readlink(struct xdr_stream *xdr, const struct nfs4_readlink *readlink, struct rpc_rqst *req, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(4);
 	WRITE32(OP_READLINK);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_remove(struct xdr_stream *xdr, const struct qstr *name)
+static void encode_remove(struct xdr_stream *xdr, const struct qstr *name, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1220,11 +1207,10 @@ static int encode_remove(struct xdr_stream *xdr, const struct qstr *name)
 	WRITE32(OP_REMOVE);
 	WRITE32(name->len);
 	WRITEMEM(name->name, name->len);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_rename(struct xdr_stream *xdr, const struct qstr *oldname, const struct qstr *newname)
+static void encode_rename(struct xdr_stream *xdr, const struct qstr *oldname, const struct qstr *newname, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1232,38 +1218,35 @@ static int encode_rename(struct xdr_stream *xdr, const struct qstr *oldname, con
 	WRITE32(OP_RENAME);
 	WRITE32(oldname->len);
 	WRITEMEM(oldname->name, oldname->len);
-	
+
 	RESERVE_SPACE(4 + newname->len);
 	WRITE32(newname->len);
 	WRITEMEM(newname->name, newname->len);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_renew(struct xdr_stream *xdr, const struct nfs_client *client_stateid)
+static void encode_renew(struct xdr_stream *xdr, const struct nfs_client *client_stateid, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(12);
 	WRITE32(OP_RENEW);
 	WRITE64(client_stateid->cl_clientid);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int
-encode_restorefh(struct xdr_stream *xdr)
+static void
+encode_restorefh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(4);
 	WRITE32(OP_RESTOREFH);
-
-	return 0;
+	hdr->nops++;
 }
 
 static int
-encode_setacl(struct xdr_stream *xdr, struct nfs_setaclargs *arg)
+encode_setacl(struct xdr_stream *xdr, struct nfs_setaclargs *arg, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1278,36 +1261,32 @@ encode_setacl(struct xdr_stream *xdr, struct nfs_setaclargs *arg)
 	RESERVE_SPACE(4);
 	WRITE32(arg->acl_len);
 	xdr_write_pages(xdr, arg->acl_pages, arg->acl_pgbase, arg->acl_len);
+	hdr->nops++;
 	return 0;
 }
 
-static int
-encode_savefh(struct xdr_stream *xdr)
+static void
+encode_savefh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
 	RESERVE_SPACE(4);
 	WRITE32(OP_SAVEFH);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_setattr(struct xdr_stream *xdr, const struct nfs_setattrargs *arg, const struct nfs_server *server)
+static void encode_setattr(struct xdr_stream *xdr, const struct nfs_setattrargs *arg, const struct nfs_server *server, struct compound_hdr *hdr)
 {
-	int status;
 	__be32 *p;
-	
-        RESERVE_SPACE(4+NFS4_STATEID_SIZE);
-        WRITE32(OP_SETATTR);
-	WRITEMEM(arg->stateid.data, NFS4_STATEID_SIZE);
 
-        if ((status = encode_attrs(xdr, arg->iap, server)))
-		return status;
-
-        return 0;
+	RESERVE_SPACE(4+NFS4_STATEID_SIZE);
+	WRITE32(OP_SETATTR);
+	WRITEMEM(arg->stateid.data, NFS4_STATEID_SIZE);
+	hdr->nops++;
+	encode_attrs(xdr, arg->iap, server);
 }
 
-static int encode_setclientid(struct xdr_stream *xdr, const struct nfs4_setclientid *setclientid)
+static void encode_setclientid(struct xdr_stream *xdr, const struct nfs4_setclientid *setclientid, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1322,23 +1301,21 @@ static int encode_setclientid(struct xdr_stream *xdr, const struct nfs4_setclien
 	encode_string(xdr, setclientid->sc_uaddr_len, setclientid->sc_uaddr);
 	RESERVE_SPACE(4);
 	WRITE32(setclientid->sc_cb_ident);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_setclientid_confirm(struct xdr_stream *xdr, const struct nfs_client *client_state)
+static void encode_setclientid_confirm(struct xdr_stream *xdr, const struct nfs_client *client_state, struct compound_hdr *hdr)
 {
-        __be32 *p;
-
-        RESERVE_SPACE(12 + NFS4_VERIFIER_SIZE);
-        WRITE32(OP_SETCLIENTID_CONFIRM);
-        WRITE64(client_state->cl_clientid);
-        WRITEMEM(client_state->cl_confirm.data, NFS4_VERIFIER_SIZE);
+	__be32 *p;
 
-        return 0;
+	RESERVE_SPACE(12 + NFS4_VERIFIER_SIZE);
+	WRITE32(OP_SETCLIENTID_CONFIRM);
+	WRITE64(client_state->cl_clientid);
+	WRITEMEM(client_state->cl_confirm.data, NFS4_VERIFIER_SIZE);
+	hdr->nops++;
 }
 
-static int encode_write(struct xdr_stream *xdr, const struct nfs_writeargs *args)
+static void encode_write(struct xdr_stream *xdr, const struct nfs_writeargs *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1353,11 +1330,10 @@ static int encode_write(struct xdr_stream *xdr, const struct nfs_writeargs *args
 	WRITE32(args->count);
 
 	xdr_write_pages(xdr, args->pages, args->pgbase, args->count);
-
-	return 0;
+	hdr->nops++;
 }
 
-static int encode_delegreturn(struct xdr_stream *xdr, const nfs4_stateid *stateid)
+static void encode_delegreturn(struct xdr_stream *xdr, const nfs4_stateid *stateid, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
@@ -1365,8 +1341,7 @@ static int encode_delegreturn(struct xdr_stream *xdr, const nfs4_stateid *statei
 
 	WRITE32(OP_DELEGRETURN);
 	WRITEMEM(stateid->data, NFS4_STATEID_SIZE);
-	return 0;
-
+	hdr->nops++;
 }
 /*
  * END OF "GENERIC" ENCODE ROUTINES.
@@ -1379,21 +1354,16 @@ static int nfs4_xdr_enc_access(struct rpc_rqst *req, __be32 *p, const struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status != 0)
-		goto out;
-	status = encode_access(&xdr, args->access);
-	if (status != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_access(&xdr, args->access, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1403,21 +1373,17 @@ static int nfs4_xdr_enc_lookup(struct rpc_rqst *req, __be32 *p, const struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 4,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->dir_fh)) != 0)
-		goto out;
-	if ((status = encode_lookup(&xdr, args->name)) != 0)
-		goto out;
-	if ((status = encode_getfh(&xdr)) != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->dir_fh, &hdr);
+	encode_lookup(&xdr, args->name, &hdr);
+	encode_getfh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1427,18 +1393,16 @@ static int nfs4_xdr_enc_lookup_root(struct rpc_rqst *req, __be32 *p, const struc
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putrootfh(&xdr)) != 0)
-		goto out;
-	if ((status = encode_getfh(&xdr)) == 0)
-		status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putrootfh(&xdr, &hdr);
+	encode_getfh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1448,19 +1412,16 @@ static int nfs4_xdr_enc_remove(struct rpc_rqst *req, __be32 *p, const struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->fh)) != 0)
-		goto out;
-	if ((status = encode_remove(&xdr, &args->name)) != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_remove(&xdr, &args->name, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1470,27 +1431,20 @@ static int nfs4_xdr_enc_rename(struct rpc_rqst *req, __be32 *p, const struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 7,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->old_dir)) != 0)
-		goto out;
-	if ((status = encode_savefh(&xdr)) != 0)
-		goto out;
-	if ((status = encode_putfh(&xdr, args->new_dir)) != 0)
-		goto out;
-	if ((status = encode_rename(&xdr, args->old_name, args->new_name)) != 0)
-		goto out;
-	if ((status = encode_getfattr(&xdr, args->bitmask)) != 0)
-		goto out;
-	if ((status = encode_restorefh(&xdr)) != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->old_dir, &hdr);
+	encode_savefh(&xdr, &hdr);
+	encode_putfh(&xdr, args->new_dir, &hdr);
+	encode_rename(&xdr, args->old_name, args->new_name, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_restorefh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1500,27 +1454,20 @@ static int nfs4_xdr_enc_link(struct rpc_rqst *req, __be32 *p, const struct nfs4_
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 7,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->fh)) != 0)
-		goto out;
-	if ((status = encode_savefh(&xdr)) != 0)
-		goto out;
-	if ((status = encode_putfh(&xdr, args->dir_fh)) != 0)
-		goto out;
-	if ((status = encode_link(&xdr, args->name)) != 0)
-		goto out;
-	if ((status = encode_getfattr(&xdr, args->bitmask)) != 0)
-		goto out;
-	if ((status = encode_restorefh(&xdr)) != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_savefh(&xdr, &hdr);
+	encode_putfh(&xdr, args->dir_fh, &hdr);
+	encode_link(&xdr, args->name, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_restorefh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1530,27 +1477,20 @@ static int nfs4_xdr_enc_create(struct rpc_rqst *req, __be32 *p, const struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 7,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->dir_fh)) != 0)
-		goto out;
-	if ((status = encode_savefh(&xdr)) != 0)
-		goto out;
-	if ((status = encode_create(&xdr, args)) != 0)
-		goto out;
-	if ((status = encode_getfh(&xdr)) != 0)
-		goto out;
-	if ((status = encode_getfattr(&xdr, args->bitmask)) != 0)
-		goto out;
-	if ((status = encode_restorefh(&xdr)) != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->dir_fh, &hdr);
+	encode_savefh(&xdr, &hdr);
+	encode_create(&xdr, args, &hdr);
+	encode_getfh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_restorefh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1568,15 +1508,15 @@ static int nfs4_xdr_enc_getattr(struct rpc_rqst *req, __be32 *p, const struct nf
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->fh)) == 0)
-		status = encode_getfattr(&xdr, args->bitmask);
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1584,23 +1524,18 @@ static int nfs4_xdr_enc_getattr(struct rpc_rqst *req, __be32 *p, const struct nf
  */
 static int nfs4_xdr_enc_close(struct rpc_rqst *req, __be32 *p, struct nfs_closeargs *args)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr = {
-                .nops   = 3,
-        };
-        int status;
-
-        xdr_init_encode(&xdr, &req->rq_snd_buf, p);
-        encode_compound_hdr(&xdr, &hdr);
-        status = encode_putfh(&xdr, args->fh);
-        if(status)
-                goto out;
-        status = encode_close(&xdr, args);
-	if (status != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-        return status;
+	struct xdr_stream xdr;
+	struct compound_hdr hdr = {
+		.nops   = 0,
+	};
+
+	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+	encode_compound_hdr(&xdr, &hdr);
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_close(&xdr, args, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1610,33 +1545,20 @@ static int nfs4_xdr_enc_open(struct rpc_rqst *req, __be32 *p, struct nfs_openarg
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 7,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_savefh(&xdr);
-	if (status)
-		goto out;
-	status = encode_open(&xdr, args);
-	if (status)
-		goto out;
-	status = encode_getfh(&xdr);
-	if (status)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-	if (status)
-		goto out;
-	status = encode_restorefh(&xdr);
-	if (status)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_savefh(&xdr, &hdr);
+	encode_open(&xdr, args, &hdr);
+	encode_getfh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_restorefh(&xdr, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1646,18 +1568,15 @@ static int nfs4_xdr_enc_open_confirm(struct rpc_rqst *req, __be32 *p, struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops   = 2,
+		.nops   = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if(status)
-		goto out;
-	status = encode_open_confirm(&xdr, args);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_open_confirm(&xdr, args, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1667,21 +1586,16 @@ static int nfs4_xdr_enc_open_noattr(struct rpc_rqst *req, __be32 *p, struct nfs_
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops   = 3,
+		.nops   = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_open(&xdr, args);
-	if (status)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_open(&xdr, args, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1691,21 +1605,16 @@ static int nfs4_xdr_enc_open_downgrade(struct rpc_rqst *req, __be32 *p, struct n
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops	= 3,
+		.nops	= 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_open_downgrade(&xdr, args);
-	if (status != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_open_downgrade(&xdr, args, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1715,18 +1624,15 @@ static int nfs4_xdr_enc_lock(struct rpc_rqst *req, __be32 *p, struct nfs_lock_ar
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops   = 2,
+		.nops   = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if(status)
-		goto out;
-	status = encode_lock(&xdr, args);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_lock(&xdr, args, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1736,18 +1642,15 @@ static int nfs4_xdr_enc_lockt(struct rpc_rqst *req, __be32 *p, struct nfs_lockt_
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops   = 2,
+		.nops   = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if(status)
-		goto out;
-	status = encode_lockt(&xdr, args);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_lockt(&xdr, args, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1757,18 +1660,15 @@ static int nfs4_xdr_enc_locku(struct rpc_rqst *req, __be32 *p, struct nfs_locku_
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops   = 2,
+		.nops   = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if(status)
-		goto out;
-	status = encode_locku(&xdr, args);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_locku(&xdr, args, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1778,18 +1678,15 @@ static int nfs4_xdr_enc_readlink(struct rpc_rqst *req, __be32 *p, const struct n
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
 	struct rpc_auth *auth = req->rq_task->tk_msg.rpc_cred->cr_auth;
 	unsigned int replen;
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if(status)
-		goto out;
-	status = encode_readlink(&xdr, args, req);
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_readlink(&xdr, args, req, &hdr);
 
 	/* set up reply kvec
 	 *    toplevel_status + taglen + rescount + OP_PUTFH + status
@@ -1798,9 +1695,8 @@ static int nfs4_xdr_enc_readlink(struct rpc_rqst *req, __be32 *p, const struct n
 	replen = (RPC_REPHDRSIZE + auth->au_rslack + NFS4_dec_readlink_sz) << 2;
 	xdr_inline_pages(&req->rq_rcv_buf, replen, args->pages,
 			args->pgbase, args->pglen);
-
-out:
-	return status;
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1810,18 +1706,15 @@ static int nfs4_xdr_enc_readdir(struct rpc_rqst *req, __be32 *p, const struct nf
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
 	struct rpc_auth *auth = req->rq_task->tk_msg.rpc_cred->cr_auth;
 	int replen;
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if(status)
-		goto out;
-	status = encode_readdir(&xdr, args, req);
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_readdir(&xdr, args, req, &hdr);
 
 	/* set up reply kvec
 	 *    toplevel_status + taglen + rescount + OP_PUTFH + status
@@ -1833,9 +1726,8 @@ static int nfs4_xdr_enc_readdir(struct rpc_rqst *req, __be32 *p, const struct nf
 	dprintk("%s: inlined page args = (%u, %p, %u, %u)\n",
 			__func__, replen, args->pages,
 			args->pgbase, args->count);
-
-out:
-	return status;
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1846,18 +1738,14 @@ static int nfs4_xdr_enc_read(struct rpc_rqst *req, __be32 *p, struct nfs_readarg
 	struct rpc_auth *auth = req->rq_task->tk_msg.rpc_cred->cr_auth;
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
-	int replen, status;
+	int replen;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_read(&xdr, args);
-	if (status)
-		goto out;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_read(&xdr, args, &hdr);
 
 	/* set up reply kvec
 	 *    toplevel status + taglen=0 + rescount + OP_PUTFH + status
@@ -1867,33 +1755,27 @@ static int nfs4_xdr_enc_read(struct rpc_rqst *req, __be32 *p, struct nfs_readarg
 	xdr_inline_pages(&req->rq_rcv_buf, replen,
 			 args->pages, args->pgbase, args->count);
 	req->rq_rcv_buf.flags |= XDRBUF_READ;
-out:
-	return status;
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
  * Encode an SETATTR request
  */
 static int nfs4_xdr_enc_setattr(struct rpc_rqst *req, __be32 *p, struct nfs_setattrargs *args)
-
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr = {
-                .nops   = 3,
-        };
-        int status;
-
-        xdr_init_encode(&xdr, &req->rq_snd_buf, p);
-        encode_compound_hdr(&xdr, &hdr);
-        status = encode_putfh(&xdr, args->fh);
-        if(status)
-                goto out;
-        status = encode_setattr(&xdr, args, args->server);
-        if(status)
-                goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-        return status;
+	struct xdr_stream xdr;
+	struct compound_hdr hdr = {
+		.nops   = 0,
+	};
+
+	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+	encode_compound_hdr(&xdr, &hdr);
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_setattr(&xdr, args, args->server, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1906,22 +1788,21 @@ nfs4_xdr_enc_getacl(struct rpc_rqst *req, __be32 *p,
 	struct xdr_stream xdr;
 	struct rpc_auth *auth = req->rq_task->tk_msg.rpc_cred->cr_auth;
 	struct compound_hdr hdr = {
-		.nops   = 2,
+		.nops   = 0,
 	};
-	int replen, status;
+	int replen;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_getattr_two(&xdr, FATTR4_WORD0_ACL, 0);
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_getattr_two(&xdr, FATTR4_WORD0_ACL, 0, &hdr);
+
 	/* set up reply buffer: */
 	replen = (RPC_REPHDRSIZE + auth->au_rslack + NFS4_dec_getacl_sz) << 2;
 	xdr_inline_pages(&req->rq_rcv_buf, replen,
 		args->acl_pages, args->acl_pgbase, args->acl_len);
-out:
-	return status;
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1931,22 +1812,17 @@ static int nfs4_xdr_enc_write(struct rpc_rqst *req, __be32 *p, struct nfs_writea
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_write(&xdr, args);
-	if (status)
-		goto out;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_write(&xdr, args, &hdr);
 	req->rq_snd_buf.flags |= XDRBUF_WRITE;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1956,21 +1832,16 @@ static int nfs4_xdr_enc_commit(struct rpc_rqst *req, __be32 *p, struct nfs_write
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status)
-		goto out;
-	status = encode_commit(&xdr, args);
-	if (status)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_commit(&xdr, args, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1980,16 +1851,15 @@ static int nfs4_xdr_enc_fsinfo(struct rpc_rqst *req, __be32 *p, struct nfs4_fsin
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops	= 2,
+		.nops	= 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (!status)
-		status = encode_fsinfo(&xdr, args->bitmask);
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_fsinfo(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -1999,17 +1869,16 @@ static int nfs4_xdr_enc_pathconf(struct rpc_rqst *req, __be32 *p, const struct n
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (!status)
-		status = encode_getattr_one(&xdr,
-				args->bitmask[0] & nfs4_pathconf_bitmap[0]);
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_getattr_one(&xdr, args->bitmask[0] & nfs4_pathconf_bitmap[0],
+			   &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2019,18 +1888,16 @@ static int nfs4_xdr_enc_statfs(struct rpc_rqst *req, __be32 *p, const struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fh);
-	if (status == 0)
-		status = encode_getattr_two(&xdr,
-				args->bitmask[0] & nfs4_statfs_bitmap[0],
-				args->bitmask[1] & nfs4_statfs_bitmap[1]);
-	return status;
+	encode_putfh(&xdr, args->fh, &hdr);
+	encode_getattr_two(&xdr, args->bitmask[0] & nfs4_statfs_bitmap[0],
+			   args->bitmask[1] & nfs4_statfs_bitmap[1], &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2040,19 +1907,18 @@ static int nfs4_xdr_enc_server_caps(struct rpc_rqst *req, __be32 *p, const struc
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, fhandle);
-	if (status == 0)
-		status = encode_getattr_one(&xdr, FATTR4_WORD0_SUPPORTED_ATTRS|
-				FATTR4_WORD0_LINK_SUPPORT|
-				FATTR4_WORD0_SYMLINK_SUPPORT|
-				FATTR4_WORD0_ACLSUPPORT);
-	return status;
+	encode_putfh(&xdr, fhandle, &hdr);
+	encode_getattr_one(&xdr, FATTR4_WORD0_SUPPORTED_ATTRS|
+			   FATTR4_WORD0_LINK_SUPPORT|
+			   FATTR4_WORD0_SYMLINK_SUPPORT|
+			   FATTR4_WORD0_ACLSUPPORT, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2062,12 +1928,14 @@ static int nfs4_xdr_enc_renew(struct rpc_rqst *req, __be32 *p, struct nfs_client
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops	= 1,
+		.nops	= 0,
 	};
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	return encode_renew(&xdr, clp);
+	encode_renew(&xdr, clp, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2077,12 +1945,14 @@ static int nfs4_xdr_enc_setclientid(struct rpc_rqst *req, __be32 *p, struct nfs4
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops	= 1,
+		.nops	= 0,
 	};
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	return encode_setclientid(&xdr, sc);
+	encode_setclientid(&xdr, sc, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2092,19 +1962,17 @@ static int nfs4_xdr_enc_setclientid_confirm(struct rpc_rqst *req, __be32 *p, str
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops	= 3,
+		.nops	= 0,
 	};
 	const u32 lease_bitmap[2] = { FATTR4_WORD0_LEASE_TIME, 0 };
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_setclientid_confirm(&xdr, clp);
-	if (!status)
-		status = encode_putrootfh(&xdr);
-	if (!status)
-		status = encode_fsinfo(&xdr, lease_bitmap);
-	return status;
+	encode_setclientid_confirm(&xdr, clp, &hdr);
+	encode_putrootfh(&xdr, &hdr);
+	encode_fsinfo(&xdr, lease_bitmap, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2114,21 +1982,16 @@ static int nfs4_xdr_enc_delegreturn(struct rpc_rqst *req, __be32 *p, const struc
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	status = encode_putfh(&xdr, args->fhandle);
-	if (status != 0)
-		goto out;
-	status = encode_delegreturn(&xdr, args->stateid);
-	if (status != 0)
-		goto out;
-	status = encode_getfattr(&xdr, args->bitmask);
-out:
-	return status;
+	encode_putfh(&xdr, args->fhandle, &hdr);
+	encode_delegreturn(&xdr, args->stateid, &hdr);
+	encode_getfattr(&xdr, args->bitmask, &hdr);
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2138,20 +2001,17 @@ static int nfs4_xdr_enc_fs_locations(struct rpc_rqst *req, __be32 *p, struct nfs
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 3,
+		.nops = 0,
 	};
 	struct rpc_auth *auth = req->rq_task->tk_msg.rpc_cred->cr_auth;
 	int replen;
-	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->dir_fh)) != 0)
-		goto out;
-	if ((status = encode_lookup(&xdr, args->name)) != 0)
-		goto out;
-	if ((status = encode_fs_locations(&xdr, args->bitmask)) != 0)
-		goto out;
+	encode_putfh(&xdr, args->dir_fh, &hdr);
+	encode_lookup(&xdr, args->name, &hdr);
+	encode_fs_locations(&xdr, args->bitmask, &hdr);
+
 	/* set up reply
 	 *   toplevel_status + OP_PUTFH + status
 	 *   + OP_LOOKUP + status + OP_GETATTR + status = 7
@@ -2159,8 +2019,8 @@ static int nfs4_xdr_enc_fs_locations(struct rpc_rqst *req, __be32 *p, struct nfs
 	replen = (RPC_REPHDRSIZE + auth->au_rslack + 7) << 2;
 	xdr_inline_pages(&req->rq_rcv_buf, replen, &args->page,
 			0, PAGE_SIZE);
-out:
-	return status;
+	encode_nops(&hdr);
+	return 0;
 }
 
 /*
@@ -2217,11 +2077,13 @@ static int decode_compound_hdr(struct xdr_stream *xdr, struct compound_hdr *hdr)
 	READ_BUF(8);
 	READ32(hdr->status);
 	READ32(hdr->taglen);
-	
+
 	READ_BUF(hdr->taglen + 4);
 	hdr->tag = (char *)p;
 	p += XDR_QUADLEN(hdr->taglen);
 	READ32(hdr->nops);
+	if (unlikely(hdr->nops < 1))
+		return nfs4_stat_to_errno(hdr->status);
 	return 0;
 }
 
@@ -3047,8 +2909,7 @@ static int decode_create(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
 static int decode_server_caps(struct xdr_stream *xdr, struct nfs4_server_caps_res *res)
 {
 	__be32 *savep;
-	uint32_t attrlen, 
-		 bitmap[2] = {0};
+	uint32_t attrlen, bitmap[2] = {0};
 	int status;
 
 	if ((status = decode_op_hdr(xdr, OP_GETATTR)) != 0)
@@ -3070,14 +2931,13 @@ xdr_error:
 	dprintk("%s: xdr returned %d!\n", __func__, -status);
 	return status;
 }
-	
+
 static int decode_statfs(struct xdr_stream *xdr, struct nfs_fsstat *fsstat)
 {
 	__be32 *savep;
-	uint32_t attrlen, 
-		 bitmap[2] = {0};
+	uint32_t attrlen, bitmap[2] = {0};
 	int status;
-	
+
 	if ((status = decode_op_hdr(xdr, OP_GETATTR)) != 0)
 		goto xdr_error;
 	if ((status = decode_attr_bitmap(xdr, bitmap)) != 0)
@@ -3107,10 +2967,9 @@ xdr_error:
 static int decode_pathconf(struct xdr_stream *xdr, struct nfs_pathconf *pathconf)
 {
 	__be32 *savep;
-	uint32_t attrlen, 
-		 bitmap[2] = {0};
+	uint32_t attrlen, bitmap[2] = {0};
 	int status;
-	
+
 	if ((status = decode_op_hdr(xdr, OP_GETATTR)) != 0)
 		goto xdr_error;
 	if ((status = decode_attr_bitmap(xdr, bitmap)) != 0)
@@ -3256,7 +3115,7 @@ static int decode_getfh(struct xdr_stream *xdr, struct nfs_fh *fh)
 static int decode_link(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
 {
 	int status;
-	
+
 	status = decode_op_hdr(xdr, OP_LINK);
 	if (status)
 		return status;
@@ -3344,27 +3203,27 @@ static int decode_lookup(struct xdr_stream *xdr)
 /* This is too sick! */
 static int decode_space_limit(struct xdr_stream *xdr, u64 *maxsize)
 {
-        __be32 *p;
+	__be32 *p;
 	uint32_t limit_type, nblocks, blocksize;
 
 	READ_BUF(12);
 	READ32(limit_type);
 	switch (limit_type) {
-		case 1:
-			READ64(*maxsize);
-			break;
-		case 2:
-			READ32(nblocks);
-			READ32(blocksize);
-			*maxsize = (uint64_t)nblocks * (uint64_t)blocksize;
+	case 1:
+		READ64(*maxsize);
+		break;
+	case 2:
+		READ32(nblocks);
+		READ32(blocksize);
+		*maxsize = (uint64_t)nblocks * (uint64_t)blocksize;
 	}
 	return 0;
 }
 
 static int decode_delegation(struct xdr_stream *xdr, struct nfs_openres *res)
 {
-        __be32 *p;
-        uint32_t delegation_type;
+	__be32 *p;
+	uint32_t delegation_type;
 
 	READ_BUF(4);
 	READ32(delegation_type);
@@ -3375,13 +3234,14 @@ static int decode_delegation(struct xdr_stream *xdr, struct nfs_openres *res)
 	READ_BUF(NFS4_STATEID_SIZE+4);
 	COPYMEM(res->delegation.data, NFS4_STATEID_SIZE);
 	READ32(res->do_recall);
+
 	switch (delegation_type) {
-		case NFS4_OPEN_DELEGATE_READ:
-			res->delegation_type = FMODE_READ;
-			break;
-		case NFS4_OPEN_DELEGATE_WRITE:
-			res->delegation_type = FMODE_WRITE|FMODE_READ;
-			if (decode_space_limit(xdr, &res->maxsize) < 0)
+	case NFS4_OPEN_DELEGATE_READ:
+		res->delegation_type = FMODE_READ;
+		break;
+	case NFS4_OPEN_DELEGATE_WRITE:
+		res->delegation_type = FMODE_WRITE|FMODE_READ;
+		if (decode_space_limit(xdr, &res->maxsize) < 0)
 				return -EIO;
 	}
 	return decode_ace(xdr, NULL, res->server->nfs_client);
@@ -3389,27 +3249,27 @@ static int decode_delegation(struct xdr_stream *xdr, struct nfs_openres *res)
 
 static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res)
 {
-        __be32 *p;
+	__be32 *p;
 	uint32_t savewords, bmlen, i;
-        int status;
+	int status;
 
-        status = decode_op_hdr(xdr, OP_OPEN);
+	status = decode_op_hdr(xdr, OP_OPEN);
 	if (status != -EIO)
 		nfs_increment_open_seqid(status, res->seqid);
-        if (status)
-                return status;
-        READ_BUF(NFS4_STATEID_SIZE);
-        COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
+	if (status)
+		return status;
+	READ_BUF(NFS4_STATEID_SIZE);
+	COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
 
-        decode_change_info(xdr, &res->cinfo);
+	decode_change_info(xdr, &res->cinfo);
 
-        READ_BUF(8);
-        READ32(res->rflags);
-        READ32(bmlen);
-        if (bmlen > 10)
-                goto xdr_error;
+	READ_BUF(8);
+	READ32(res->rflags);
+	READ32(bmlen);
+	if (bmlen > 10)
+		goto xdr_error;
 
-        READ_BUF(bmlen << 2);
+	READ_BUF(bmlen << 2);
 	savewords = min_t(uint32_t, bmlen, NFS4_BITMAP_SIZE);
 	for (i = 0; i < savewords; ++i)
 		READ32(res->attrset[i]);
@@ -3424,17 +3284,17 @@ xdr_error:
 
 static int decode_open_confirm(struct xdr_stream *xdr, struct nfs_open_confirmres *res)
 {
-        __be32 *p;
+	__be32 *p;
 	int status;
 
-        status = decode_op_hdr(xdr, OP_OPEN_CONFIRM);
+	status = decode_op_hdr(xdr, OP_OPEN_CONFIRM);
 	if (status != -EIO)
 		nfs_increment_open_seqid(status, res->seqid);
-        if (status)
-                return status;
-        READ_BUF(NFS4_STATEID_SIZE);
-        COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
-        return 0;
+	if (status)
+		return status;
+	READ_BUF(NFS4_STATEID_SIZE);
+	COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
+	return 0;
 }
 
 static int decode_open_downgrade(struct xdr_stream *xdr, struct nfs_closeres *res)
@@ -3562,7 +3422,7 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 		dprintk("NFS: readdir reply truncated!\n");
 		entry[1] = 1;
 	}
-out:	
+out:
 	kunmap_atomic(kaddr, KM_USER0);
 	return 0;
 short_pkt:
@@ -3718,7 +3578,6 @@ static int decode_setattr(struct xdr_stream *xdr, struct nfs_setattrres *res)
 	uint32_t bmlen;
 	int status;
 
-        
 	status = decode_op_hdr(xdr, OP_SETATTR);
 	if (status)
 		return status;
@@ -3738,7 +3597,7 @@ static int decode_setclientid(struct xdr_stream *xdr, struct nfs_client *clp)
 	READ32(opnum);
 	if (opnum != OP_SETCLIENTID) {
 		dprintk("nfs: decode_setclientid: Server returned operation"
-			       	" %d\n", opnum);
+			" %d\n", opnum);
 		return -EIO;
 	}
 	READ32(nfserr);
@@ -3792,34 +3651,34 @@ static int decode_delegreturn(struct xdr_stream *xdr)
 }
 
 /*
+ * END OF "GENERIC" DECODE ROUTINES.
+ */
+
+/*
  * Decode OPEN_DOWNGRADE response
  */
 static int nfs4_xdr_dec_open_downgrade(struct rpc_rqst *rqstp, __be32 *p, struct nfs_closeres *res)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr;
-        int status;
-
-        xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
-        status = decode_compound_hdr(&xdr, &hdr);
-        if (status)
-                goto out;
-        status = decode_putfh(&xdr);
-        if (status)
-                goto out;
-        status = decode_open_downgrade(&xdr, res);
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_open_downgrade(&xdr, res);
 	if (status != 0)
 		goto out;
 	decode_getfattr(&xdr, res->fattr, res->server);
 out:
-        return status;
+	return status;
 }
 
 /*
- * END OF "GENERIC" DECODE ROUTINES.
- */
-
-/*
  * Decode ACCESS response
  */
 static int nfs4_xdr_dec_access(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_accessres *res)
@@ -3827,7 +3686,7 @@ static int nfs4_xdr_dec_access(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_ac
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -3850,7 +3709,7 @@ static int nfs4_xdr_dec_lookup(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_lo
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -3873,7 +3732,7 @@ static int nfs4_xdr_dec_lookup_root(struct rpc_rqst *rqstp, __be32 *p, struct nf
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -3893,7 +3752,7 @@ static int nfs4_xdr_dec_remove(struct rpc_rqst *rqstp, __be32 *p, struct nfs_rem
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -3914,7 +3773,7 @@ static int nfs4_xdr_dec_rename(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_re
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -3944,7 +3803,7 @@ static int nfs4_xdr_dec_link(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_link
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -3977,7 +3836,7 @@ static int nfs4_xdr_dec_create(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_cr
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	if ((status = decode_compound_hdr(&xdr, &hdr)) != 0)
 		goto out;
@@ -4014,7 +3873,7 @@ static int nfs4_xdr_dec_getattr(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_g
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
 	int status;
-	
+
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	status = decode_compound_hdr(&xdr, &hdr);
 	if (status)
@@ -4025,7 +3884,6 @@ static int nfs4_xdr_dec_getattr(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_g
 	status = decode_getfattr(&xdr, res->fattr, res->server);
 out:
 	return status;
-
 }
 
 /*
@@ -4034,21 +3892,20 @@ out:
 static int
 nfs4_xdr_enc_setacl(struct rpc_rqst *req, __be32 *p, struct nfs_setaclargs *args)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr = {
-                .nops   = 2,
-        };
-        int status;
-
-        xdr_init_encode(&xdr, &req->rq_snd_buf, p);
-        encode_compound_hdr(&xdr, &hdr);
-        status = encode_putfh(&xdr, args->fh);
-        if (status)
-                goto out;
-        status = encode_setacl(&xdr, args);
-out:
-        return status;
+	struct xdr_stream xdr;
+	struct compound_hdr hdr = {
+		.nops   = 0,
+	};
+	int status;
+
+	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
+	encode_compound_hdr(&xdr, &hdr);
+	encode_putfh(&xdr, args->fh, &hdr);
+	status = encode_setacl(&xdr, args, &hdr);
+	encode_nops(&hdr);
+	return status;
 }
+
 /*
  * Decode SETACL response
  */
@@ -4099,18 +3956,18 @@ out:
  */
 static int nfs4_xdr_dec_close(struct rpc_rqst *rqstp, __be32 *p, struct nfs_closeres *res)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr;
-        int status;
-
-        xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
-        status = decode_compound_hdr(&xdr, &hdr);
-        if (status)
-                goto out;
-        status = decode_putfh(&xdr);
-        if (status)
-                goto out;
-        status = decode_close(&xdr, res);
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_close(&xdr, res);
 	if (status != 0)
 		goto out;
 	/*
@@ -4121,7 +3978,7 @@ static int nfs4_xdr_dec_close(struct rpc_rqst *rqstp, __be32 *p, struct nfs_clos
 	 */
 	decode_getfattr(&xdr, res->fattr, res->server);
 out:
-        return status;
+	return status;
 }
 
 /*
@@ -4129,23 +3986,23 @@ out:
  */
 static int nfs4_xdr_dec_open(struct rpc_rqst *rqstp, __be32 *p, struct nfs_openres *res)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr;
-        int status;
-
-        xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
-        status = decode_compound_hdr(&xdr, &hdr);
-        if (status)
-                goto out;
-        status = decode_putfh(&xdr);
-        if (status)
-                goto out;
-        status = decode_savefh(&xdr);
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_savefh(&xdr);
+	if (status)
+		goto out;
+	status = decode_open(&xdr, res);
 	if (status)
 		goto out;
-        status = decode_open(&xdr, res);
-        if (status)
-                goto out;
 	if (decode_getfh(&xdr, &res->fh) != 0)
 		goto out;
 	if (decode_getfattr(&xdr, res->f_attr, res->server) != 0)
@@ -4154,7 +4011,7 @@ static int nfs4_xdr_dec_open(struct rpc_rqst *rqstp, __be32 *p, struct nfs_openr
 		goto out;
 	decode_getfattr(&xdr, res->dir_attr, res->server);
 out:
-        return status;
+	return status;
 }
 
 /*
@@ -4162,20 +4019,20 @@ out:
  */
 static int nfs4_xdr_dec_open_confirm(struct rpc_rqst *rqstp, __be32 *p, struct nfs_open_confirmres *res)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr;
-        int status;
-
-        xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
-        status = decode_compound_hdr(&xdr, &hdr);
-        if (status)
-                goto out;
-        status = decode_putfh(&xdr);
-        if (status)
-                goto out;
-        status = decode_open_confirm(&xdr, res);
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_open_confirm(&xdr, res);
 out:
-        return status;
+	return status;
 }
 
 /*
@@ -4183,23 +4040,23 @@ out:
  */
 static int nfs4_xdr_dec_open_noattr(struct rpc_rqst *rqstp, __be32 *p, struct nfs_openres *res)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr;
-        int status;
-
-        xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
-        status = decode_compound_hdr(&xdr, &hdr);
-        if (status)
-                goto out;
-        status = decode_putfh(&xdr);
-        if (status)
-                goto out;
-        status = decode_open(&xdr, res);
-        if (status)
-                goto out;
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_open(&xdr, res);
+	if (status)
+		goto out;
 	decode_getfattr(&xdr, res->f_attr, res->server);
 out:
-        return status;
+	return status;
 }
 
 /*
@@ -4207,25 +4064,25 @@ out:
  */
 static int nfs4_xdr_dec_setattr(struct rpc_rqst *rqstp, __be32 *p, struct nfs_setattrres *res)
 {
-        struct xdr_stream xdr;
-        struct compound_hdr hdr;
-        int status;
-
-        xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
-        status = decode_compound_hdr(&xdr, &hdr);
-        if (status)
-                goto out;
-        status = decode_putfh(&xdr);
-        if (status)
-                goto out;
-        status = decode_setattr(&xdr, res);
-        if (status)
-                goto out;
+	struct xdr_stream xdr;
+	struct compound_hdr hdr;
+	int status;
+
+	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
+	status = decode_compound_hdr(&xdr, &hdr);
+	if (status)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status)
+		goto out;
+	status = decode_setattr(&xdr, res);
+	if (status)
+		goto out;
 	status = decode_getfattr(&xdr, res->fattr, res->server);
 	if (status == NFS4ERR_DELAY)
 		status = 0;
 out:
-        return status;
+	return status;
 }
 
 /*
@@ -4421,8 +4278,6 @@ static int nfs4_xdr_dec_fsinfo(struct rpc_rqst *req, __be32 *p, struct nfs_fsinf
 		status = decode_putfh(&xdr);
 	if (!status)
 		status = decode_fsinfo(&xdr, fsinfo);
-	if (!status)
-		status = nfs4_stat_to_errno(hdr.status);
 	return status;
 }
 
@@ -4511,8 +4366,6 @@ static int nfs4_xdr_dec_setclientid(struct rpc_rqst *req, __be32 *p,
 	status = decode_compound_hdr(&xdr, &hdr);
 	if (!status)
 		status = decode_setclientid(&xdr, clp);
-	if (!status)
-		status = nfs4_stat_to_errno(hdr.status);
 	return status;
 }
 
@@ -4533,8 +4386,6 @@ static int nfs4_xdr_dec_setclientid_confirm(struct rpc_rqst *req, __be32 *p, str
 		status = decode_putrootfh(&xdr);
 	if (!status)
 		status = decode_fsinfo(&xdr, fsinfo);
-	if (!status)
-		status = nfs4_stat_to_errno(hdr.status);
 	return status;
 }
 
@@ -4715,7 +4566,7 @@ nfs4_stat_to_errno(int stat)
 	.p_replen = NFS4_##restype##_sz,			\
 	.p_statidx = NFSPROC4_CLNT_##proc,			\
 	.p_name   = #proc,					\
-    }
+}
 
 struct rpc_procinfo	nfs4_procedures[] = {
   PROC(READ,		enc_read,	dec_read),
diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c
index d74d16ce0d4..d9ef602fbc5 100644
--- a/fs/nfs/nfsroot.c
+++ b/fs/nfs/nfsroot.c
@@ -86,6 +86,8 @@
 #include <net/ipconfig.h>
 #include <linux/parser.h>
 
+#include "internal.h"
+
 /* Define this to allow debugging output */
 #undef NFSROOT_DEBUG
 #define NFSDBG_FACILITY NFSDBG_ROOT
@@ -100,7 +102,7 @@ static char nfs_root_name[256] __initdata = "";
 static __be32 servaddr __initdata = 0;
 
 /* Name of directory to mount */
-static char nfs_path[NFS_MAXPATHLEN] __initdata = { 0, };
+static char nfs_export_path[NFS_MAXPATHLEN] __initdata = { 0, };
 
 /* NFS-related data */
 static struct nfs_mount_data nfs_data __initdata = { 0, };/* NFS mount info */
@@ -312,7 +314,7 @@ static int __init root_nfs_name(char *name)
 		printk(KERN_ERR "Root-NFS: Pathname for remote directory too long.\n");
 		return -1;
 	}
-	sprintf(nfs_path, buf, cp);
+	sprintf(nfs_export_path, buf, cp);
 
 	return 1;
 }
@@ -340,7 +342,7 @@ static int __init root_nfs_addr(void)
 static void __init root_nfs_print(void)
 {
 	printk(KERN_NOTICE "Root-NFS: Mounting %s on server %s as root\n",
-		nfs_path, nfs_data.hostname);
+		nfs_export_path, nfs_data.hostname);
 	printk(KERN_NOTICE "Root-NFS:     rsize = %d, wsize = %d, timeo = %d, retrans = %d\n",
 		nfs_data.rsize, nfs_data.wsize, nfs_data.timeo, nfs_data.retrans);
 	printk(KERN_NOTICE "Root-NFS:     acreg (min,max) = (%d,%d), acdir (min,max) = (%d,%d)\n",
@@ -485,18 +487,23 @@ static int __init root_nfs_get_handle(void)
 {
 	struct nfs_fh fh;
 	struct sockaddr_in sin;
+	struct nfs_mount_request request = {
+		.sap		= (struct sockaddr *)&sin,
+		.salen		= sizeof(sin),
+		.dirpath	= nfs_export_path,
+		.version	= (nfs_data.flags & NFS_MOUNT_VER3) ?
+					NFS_MNT3_VERSION : NFS_MNT_VERSION,
+		.protocol	= (nfs_data.flags & NFS_MOUNT_TCP) ?
+					XPRT_TRANSPORT_TCP : XPRT_TRANSPORT_UDP,
+		.fh		= &fh,
+	};
 	int status;
-	int protocol = (nfs_data.flags & NFS_MOUNT_TCP) ?
-					XPRT_TRANSPORT_TCP : XPRT_TRANSPORT_UDP;
-	int version = (nfs_data.flags & NFS_MOUNT_VER3) ?
-					NFS_MNT3_VERSION : NFS_MNT_VERSION;
 
 	set_sockaddr(&sin, servaddr, htons(mount_port));
-	status = nfs_mount((struct sockaddr *) &sin, sizeof(sin), NULL,
-			   nfs_path, version, protocol, &fh);
+	status = nfs_mount(&request);
 	if (status < 0)
 		printk(KERN_ERR "Root-NFS: Server returned error %d "
-				"while mounting %s\n", status, nfs_path);
+				"while mounting %s\n", status, nfs_export_path);
 	else {
 		nfs_data.root.size = fh.size;
 		memcpy(nfs_data.root.data, fh.data, fh.size);
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 40d17987d0e..f856004bb7f 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -533,12 +533,6 @@ readpage_async_filler(void *data, struct page *page)
 	unsigned int len;
 	int error;
 
-	error = nfs_wb_page(inode, page);
-	if (error)
-		goto out_unlock;
-	if (PageUptodate(page))
-		goto out_unlock;
-
 	len = nfs_page_length(page);
 	if (len == 0)
 		return nfs_return_empty_page(page);
diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index bb0313ac9e1..d6686f4786d 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -75,6 +75,7 @@ enum {
 	Opt_acl, Opt_noacl,
 	Opt_rdirplus, Opt_nordirplus,
 	Opt_sharecache, Opt_nosharecache,
+	Opt_resvport, Opt_noresvport,
 
 	/* Mount options that take integer arguments */
 	Opt_port,
@@ -129,6 +130,8 @@ static const match_table_t nfs_mount_option_tokens = {
 	{ Opt_nordirplus, "nordirplus" },
 	{ Opt_sharecache, "sharecache" },
 	{ Opt_nosharecache, "nosharecache" },
+	{ Opt_resvport, "resvport" },
+	{ Opt_noresvport, "noresvport" },
 
 	{ Opt_port, "port=%u" },
 	{ Opt_rsize, "rsize=%u" },
@@ -512,7 +515,8 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss,
 		{ NFS_MOUNT_NONLM, ",nolock", "" },
 		{ NFS_MOUNT_NOACL, ",noacl", "" },
 		{ NFS_MOUNT_NORDIRPLUS, ",nordirplus", "" },
-		{ NFS_MOUNT_UNSHARED, ",nosharecache", ""},
+		{ NFS_MOUNT_UNSHARED, ",nosharecache", "" },
+		{ NFS_MOUNT_NORESVPORT, ",noresvport", "" },
 		{ 0, NULL, NULL }
 	};
 	const struct proc_nfs_info *nfs_infop;
@@ -1033,6 +1037,12 @@ static int nfs_parse_mount_options(char *raw,
 		case Opt_nosharecache:
 			mnt->flags |= NFS_MOUNT_UNSHARED;
 			break;
+		case Opt_resvport:
+			mnt->flags &= ~NFS_MOUNT_NORESVPORT;
+			break;
+		case Opt_noresvport:
+			mnt->flags |= NFS_MOUNT_NORESVPORT;
+			break;
 
 		/*
 		 * options that take numeric values
@@ -1327,8 +1337,14 @@ out_security_failure:
 static int nfs_try_mount(struct nfs_parsed_mount_data *args,
 			 struct nfs_fh *root_fh)
 {
-	struct sockaddr *sap = (struct sockaddr *)&args->mount_server.address;
-	char *hostname;
+	struct nfs_mount_request request = {
+		.sap		= (struct sockaddr *)
+						&args->mount_server.address,
+		.dirpath	= args->nfs_server.export_path,
+		.protocol	= args->mount_server.protocol,
+		.fh		= root_fh,
+		.noresvport	= args->flags & NFS_MOUNT_NORESVPORT,
+	};
 	int status;
 
 	if (args->mount_server.version == 0) {
@@ -1337,42 +1353,38 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args,
 		else
 			args->mount_server.version = NFS_MNT_VERSION;
 	}
+	request.version = args->mount_server.version;
 
 	if (args->mount_server.hostname)
-		hostname = args->mount_server.hostname;
+		request.hostname = args->mount_server.hostname;
 	else
-		hostname = args->nfs_server.hostname;
+		request.hostname = args->nfs_server.hostname;
 
 	/*
 	 * Construct the mount server's address.
 	 */
 	if (args->mount_server.address.ss_family == AF_UNSPEC) {
-		memcpy(sap, &args->nfs_server.address,
+		memcpy(request.sap, &args->nfs_server.address,
 		       args->nfs_server.addrlen);
 		args->mount_server.addrlen = args->nfs_server.addrlen;
 	}
+	request.salen = args->mount_server.addrlen;
 
 	/*
 	 * autobind will be used if mount_server.port == 0
 	 */
-	nfs_set_port(sap, args->mount_server.port);
+	nfs_set_port(request.sap, args->mount_server.port);
 
 	/*
 	 * Now ask the mount server to map our export path
 	 * to a file handle.
 	 */
-	status = nfs_mount(sap,
-			   args->mount_server.addrlen,
-			   hostname,
-			   args->nfs_server.export_path,
-			   args->mount_server.version,
-			   args->mount_server.protocol,
-			   root_fh);
+	status = nfs_mount(&request);
 	if (status == 0)
 		return 0;
 
 	dfprintk(MOUNT, "NFS: unable to mount server %s, error %d\n",
-			hostname, status);
+			request.hostname, status);
 	return status;
 }
 
@@ -2419,7 +2431,7 @@ static void nfs4_kill_super(struct super_block *sb)
 {
 	struct nfs_server *server = NFS_SB(sb);
 
-	nfs_return_all_delegations(sb);
+	nfs_super_return_all_delegations(sb);
 	kill_anon_super(sb);
 
 	nfs4_renewd_prepare_shutdown(server);
diff --git a/fs/nfs_common/nfsacl.c b/fs/nfs_common/nfsacl.c
index c11f5375d7c..04133aacb1e 100644
--- a/fs/nfs_common/nfsacl.c
+++ b/fs/nfs_common/nfsacl.c
@@ -29,8 +29,8 @@
 
 MODULE_LICENSE("GPL");
 
-EXPORT_SYMBOL(nfsacl_encode);
-EXPORT_SYMBOL(nfsacl_decode);
+EXPORT_SYMBOL_GPL(nfsacl_encode);
+EXPORT_SYMBOL_GPL(nfsacl_decode);
 
 struct nfsacl_encode_desc {
 	struct xdr_array2_desc desc;
diff --git a/fs/nfsctl.c b/fs/nfsctl.c
index b1acbd6ab6f..b27451909df 100644
--- a/fs/nfsctl.c
+++ b/fs/nfsctl.c
@@ -38,9 +38,10 @@ static struct file *do_open(char *name, int flags)
 		return ERR_PTR(error);
 
 	if (flags == O_RDWR)
-		error = may_open(&nd,MAY_READ|MAY_WRITE,FMODE_READ|FMODE_WRITE);
+		error = may_open(&nd.path, MAY_READ|MAY_WRITE,
+					   FMODE_READ|FMODE_WRITE);
 	else
-		error = may_open(&nd, MAY_WRITE, FMODE_WRITE);
+		error = may_open(&nd.path, MAY_WRITE, FMODE_WRITE);
 
 	if (!error)
 		return dentry_open(nd.path.dentry, nd.path.mnt, flags,
diff --git a/fs/nfsd/auth.c b/fs/nfsd/auth.c
index 0184fe9b514..c903e04aa21 100644
--- a/fs/nfsd/auth.c
+++ b/fs/nfsd/auth.c
@@ -76,10 +76,10 @@ int nfsd_setuser(struct svc_rqst *rqstp, struct svc_export *exp)
 
 	ret = set_groups(new, gi);
 	put_group_info(gi);
-	if (!ret)
+	if (ret < 0)
 		goto error;
 
-	if (new->uid)
+	if (new->fsuid)
 		new->cap_effective = cap_drop_nfsd_set(new->cap_effective);
 	else
 		new->cap_effective = cap_raise_nfsd_set(new->cap_effective,
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 094747a1227..c464181b599 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -53,9 +53,6 @@
 #define NFSPROC4_CB_NULL 0
 #define NFSPROC4_CB_COMPOUND 1
 
-/* declarations */
-static const struct rpc_call_ops nfs4_cb_null_ops;
-
 /* Index of predefined Linux callback client operations */
 
 enum {
@@ -358,6 +355,7 @@ static struct rpc_program cb_program = {
 		.nrvers		= ARRAY_SIZE(nfs_cb_version),
 		.version	= nfs_cb_version,
 		.stats		= &cb_stats,
+		.pipe_dir_name  = "/nfsd4_cb",
 };
 
 /* Reference counting, callback cleanup, etc., all look racy as heck.
@@ -382,8 +380,9 @@ static int do_probe_callback(void *data)
 		.program	= &cb_program,
 		.prognumber	= cb->cb_prog,
 		.version	= nfs_cb_version[1]->number,
-		.authflavor	= RPC_AUTH_UNIX, /* XXX: need AUTH_GSS... */
+		.authflavor	= clp->cl_flavor,
 		.flags		= (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_QUIET),
+		.client_name    = clp->cl_principal,
 	};
 	struct rpc_message msg = {
 		.rpc_proc       = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL],
@@ -392,6 +391,11 @@ static int do_probe_callback(void *data)
 	struct rpc_clnt *client;
 	int status;
 
+	if (!clp->cl_principal && (clp->cl_flavor >= RPC_AUTH_GSS_KRB5)) {
+		status = nfserr_cb_path_down;
+		goto out_err;
+	}
+
 	/* Initialize address */
 	memset(&addr, 0, sizeof(addr));
 	addr.sin_family = AF_INET;
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 669461e291a..9fa60a3ad48 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -946,6 +946,11 @@ encode_op:
 			nfsd4_encode_operation(resp, op);
 			status = op->status;
 		}
+
+		dprintk("nfsv4 compound op %p opcnt %d #%d: %d: status %d\n",
+			args->ops, args->opcnt, resp->opcnt, op->opnum,
+			be32_to_cpu(status));
+
 		if (cstate->replay_owner) {
 			nfs4_put_stateowner(cstate->replay_owner);
 			cstate->replay_owner = NULL;
diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c
index 0f9d6efaa62..74f7b67567f 100644
--- a/fs/nfsd/nfs4recover.c
+++ b/fs/nfsd/nfs4recover.c
@@ -116,9 +116,9 @@ nfs4_make_rec_clidname(char *dname, struct xdr_netobj *clname)
 
 	md5_to_hex(dname, cksum.data);
 
-	kfree(cksum.data);
 	status = nfs_ok;
 out:
+	kfree(cksum.data);
 	crypto_free_hash(desc.tfm);
 out_no_tfm:
 	return status;
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index bf4cd46a5a1..88db7d3ec12 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -54,6 +54,7 @@
 #include <linux/mutex.h>
 #include <linux/lockd/bind.h>
 #include <linux/module.h>
+#include <linux/sunrpc/svcauth_gss.h>
 
 #define NFSDDBG_FACILITY                NFSDDBG_PROC
 
@@ -377,6 +378,7 @@ free_client(struct nfs4_client *clp)
 	shutdown_callback_client(clp);
 	if (clp->cl_cred.cr_group_info)
 		put_group_info(clp->cl_cred.cr_group_info);
+	kfree(clp->cl_principal);
 	kfree(clp->cl_name.data);
 	kfree(clp);
 }
@@ -696,6 +698,7 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	unsigned int 		strhashval;
 	struct nfs4_client	*conf, *unconf, *new;
 	__be32 			status;
+	char			*princ;
 	char                    dname[HEXDIR_LEN];
 	
 	if (!check_name(clname))
@@ -783,6 +786,15 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	}
 	copy_verf(new, &clverifier);
 	new->cl_addr = sin->sin_addr.s_addr;
+	new->cl_flavor = rqstp->rq_flavor;
+	princ = svc_gss_principal(rqstp);
+	if (princ) {
+		new->cl_principal = kstrdup(princ, GFP_KERNEL);
+		if (new->cl_principal == NULL) {
+			free_client(new);
+			goto out;
+		}
+	}
 	copy_cred(&new->cl_cred, &rqstp->rq_cred);
 	gen_confirm(new);
 	gen_callback(new, setclid);
@@ -2404,6 +2416,26 @@ out:
 #define LOCK_HASH_SIZE             (1 << LOCK_HASH_BITS)
 #define LOCK_HASH_MASK             (LOCK_HASH_SIZE - 1)
 
+static inline u64
+end_offset(u64 start, u64 len)
+{
+	u64 end;
+
+	end = start + len;
+	return end >= start ? end: NFS4_MAX_UINT64;
+}
+
+/* last octet in a range */
+static inline u64
+last_byte_offset(u64 start, u64 len)
+{
+	u64 end;
+
+	BUG_ON(!len);
+	end = start + len;
+	return end > start ? end - 1: NFS4_MAX_UINT64;
+}
+
 #define lockownerid_hashval(id) \
         ((id) & LOCK_HASH_MASK)
 
@@ -2423,13 +2455,13 @@ static struct list_head lockstateid_hashtbl[STATEID_HASH_SIZE];
 static struct nfs4_stateid *
 find_stateid(stateid_t *stid, int flags)
 {
-	struct nfs4_stateid *local = NULL;
+	struct nfs4_stateid *local;
 	u32 st_id = stid->si_stateownerid;
 	u32 f_id = stid->si_fileid;
 	unsigned int hashval;
 
 	dprintk("NFSD: find_stateid flags 0x%x\n",flags);
-	if ((flags & LOCK_STATE) || (flags & RD_STATE) || (flags & WR_STATE)) {
+	if (flags & (LOCK_STATE | RD_STATE | WR_STATE)) {
 		hashval = stateid_hashval(st_id, f_id);
 		list_for_each_entry(local, &lockstateid_hashtbl[hashval], st_hash) {
 			if ((local->st_stateid.si_stateownerid == st_id) &&
@@ -2437,7 +2469,8 @@ find_stateid(stateid_t *stid, int flags)
 				return local;
 		}
 	} 
-	if ((flags & OPEN_STATE) || (flags & RD_STATE) || (flags & WR_STATE)) {
+
+	if (flags & (OPEN_STATE | RD_STATE | WR_STATE)) {
 		hashval = stateid_hashval(st_id, f_id);
 		list_for_each_entry(local, &stateid_hashtbl[hashval], st_hash) {
 			if ((local->st_stateid.si_stateownerid == st_id) &&
@@ -2506,8 +2539,8 @@ nfs4_set_lock_denied(struct file_lock *fl, struct nfsd4_lock_denied *deny)
 		deny->ld_clientid.cl_id = 0;
 	}
 	deny->ld_start = fl->fl_start;
-	deny->ld_length = ~(u64)0;
-	if (fl->fl_end != ~(u64)0)
+	deny->ld_length = NFS4_MAX_UINT64;
+	if (fl->fl_end != NFS4_MAX_UINT64)
 		deny->ld_length = fl->fl_end - fl->fl_start + 1;        
 	deny->ld_type = NFS4_READ_LT;
 	if (fl->fl_type != F_RDLCK)
@@ -2604,7 +2637,7 @@ out:
 static int
 check_lock_length(u64 offset, u64 length)
 {
-	return ((length == 0)  || ((length != ~(u64)0) &&
+	return ((length == 0)  || ((length != NFS4_MAX_UINT64) &&
 	     LOFF_OVERFLOW(offset, length)));
 }
 
@@ -2724,11 +2757,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	file_lock.fl_lmops = &nfsd_posix_mng_ops;
 
 	file_lock.fl_start = lock->lk_offset;
-	if ((lock->lk_length == ~(u64)0) || 
-			LOFF_OVERFLOW(lock->lk_offset, lock->lk_length))
-		file_lock.fl_end = ~(u64)0;
-	else
-		file_lock.fl_end = lock->lk_offset + lock->lk_length - 1;
+	file_lock.fl_end = last_byte_offset(lock->lk_offset, lock->lk_length);
 	nfs4_transform_lock_offset(&file_lock);
 
 	/*
@@ -2769,6 +2798,25 @@ out:
 }
 
 /*
+ * The NFSv4 spec allows a client to do a LOCKT without holding an OPEN,
+ * so we do a temporary open here just to get an open file to pass to
+ * vfs_test_lock.  (Arguably perhaps test_lock should be done with an
+ * inode operation.)
+ */
+static int nfsd_test_lock(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file_lock *lock)
+{
+	struct file *file;
+	int err;
+
+	err = nfsd_open(rqstp, fhp, S_IFREG, NFSD_MAY_READ, &file);
+	if (err)
+		return err;
+	err = vfs_test_lock(file, lock);
+	nfsd_close(file);
+	return err;
+}
+
+/*
  * LOCKT operation
  */
 __be32
@@ -2776,7 +2824,6 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	    struct nfsd4_lockt *lockt)
 {
 	struct inode *inode;
-	struct file file;
 	struct file_lock file_lock;
 	int error;
 	__be32 status;
@@ -2827,23 +2874,12 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	file_lock.fl_lmops = &nfsd_posix_mng_ops;
 
 	file_lock.fl_start = lockt->lt_offset;
-	if ((lockt->lt_length == ~(u64)0) || LOFF_OVERFLOW(lockt->lt_offset, lockt->lt_length))
-		file_lock.fl_end = ~(u64)0;
-	else
-		file_lock.fl_end = lockt->lt_offset + lockt->lt_length - 1;
+	file_lock.fl_end = last_byte_offset(lockt->lt_offset, lockt->lt_length);
 
 	nfs4_transform_lock_offset(&file_lock);
 
-	/* vfs_test_lock uses the struct file _only_ to resolve the inode.
-	 * since LOCKT doesn't require an OPEN, and therefore a struct
-	 * file may not exist, pass vfs_test_lock a struct file with
-	 * only the dentry:inode set.
-	 */
-	memset(&file, 0, sizeof (struct file));
-	file.f_path.dentry = cstate->current_fh.fh_dentry;
-
 	status = nfs_ok;
-	error = vfs_test_lock(&file, &file_lock);
+	error = nfsd_test_lock(rqstp, &cstate->current_fh, &file_lock);
 	if (error) {
 		status = nfserrno(error);
 		goto out;
@@ -2894,10 +2930,7 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	file_lock.fl_lmops = &nfsd_posix_mng_ops;
 	file_lock.fl_start = locku->lu_offset;
 
-	if ((locku->lu_length == ~(u64)0) || LOFF_OVERFLOW(locku->lu_offset, locku->lu_length))
-		file_lock.fl_end = ~(u64)0;
-	else
-		file_lock.fl_end = locku->lu_offset + locku->lu_length - 1;
+	file_lock.fl_end = last_byte_offset(locku->lu_offset, locku->lu_length);
 	nfs4_transform_lock_offset(&file_lock);
 
 	/*
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index afcdf4b7684..f65953be39c 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -1,6 +1,4 @@
 /*
- *  fs/nfs/nfs4xdr.c
- *
  *  Server-side XDR for NFSv4
  *
  *  Copyright (c) 2002 The Regents of the University of Michigan.
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index 77d7b8c531a..3d93b2064ce 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -84,6 +84,8 @@ static ssize_t write_unexport(struct file *file, char *buf, size_t size);
 static ssize_t write_getfd(struct file *file, char *buf, size_t size);
 static ssize_t write_getfs(struct file *file, char *buf, size_t size);
 static ssize_t write_filehandle(struct file *file, char *buf, size_t size);
+static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size);
+static ssize_t write_unlock_fs(struct file *file, char *buf, size_t size);
 static ssize_t write_threads(struct file *file, char *buf, size_t size);
 static ssize_t write_pool_threads(struct file *file, char *buf, size_t size);
 static ssize_t write_versions(struct file *file, char *buf, size_t size);
@@ -94,9 +96,6 @@ static ssize_t write_leasetime(struct file *file, char *buf, size_t size);
 static ssize_t write_recoverydir(struct file *file, char *buf, size_t size);
 #endif
 
-static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size);
-static ssize_t failover_unlock_fs(struct file *file, char *buf, size_t size);
-
 static ssize_t (*write_op[])(struct file *, char *, size_t) = {
 	[NFSD_Svc] = write_svc,
 	[NFSD_Add] = write_add,
@@ -106,8 +105,8 @@ static ssize_t (*write_op[])(struct file *, char *, size_t) = {
 	[NFSD_Getfd] = write_getfd,
 	[NFSD_Getfs] = write_getfs,
 	[NFSD_Fh] = write_filehandle,
-	[NFSD_FO_UnlockIP] = failover_unlock_ip,
-	[NFSD_FO_UnlockFS] = failover_unlock_fs,
+	[NFSD_FO_UnlockIP] = write_unlock_ip,
+	[NFSD_FO_UnlockFS] = write_unlock_fs,
 	[NFSD_Threads] = write_threads,
 	[NFSD_Pool_Threads] = write_pool_threads,
 	[NFSD_Versions] = write_versions,
@@ -176,10 +175,24 @@ static const struct file_operations exports_operations = {
 /*----------------------------------------------------------------------------*/
 /*
  * payload - write methods
- * If the method has a response, the response should be put in buf,
- * and the length returned.  Otherwise return 0 or and -error.
  */
 
+/**
+ * write_svc - Start kernel's NFSD server
+ *
+ * Deprecated.  /proc/fs/nfsd/threads is preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_svc
+ *				svc_port:	port number of this
+ *						server's listener
+ *				svc_nthreads:	number of threads to start
+ *			size:	size in bytes of passed in nfsctl_svc
+ * Output:
+ *	On success:	returns zero
+ *	On error:	return code is negative errno value
+ */
 static ssize_t write_svc(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_svc *data;
@@ -189,6 +202,30 @@ static ssize_t write_svc(struct file *file, char *buf, size_t size)
 	return nfsd_svc(data->svc_port, data->svc_nthreads);
 }
 
+/**
+ * write_add - Add or modify client entry in auth unix cache
+ *
+ * Deprecated.  /proc/net/rpc/auth.unix.ip is preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_client
+ *				cl_ident:	'\0'-terminated C string
+ *						containing domain name
+ *						of client
+ *				cl_naddr:	no. of items in cl_addrlist
+ *				cl_addrlist:	array of client addresses
+ *				cl_fhkeytype:	ignored
+ *				cl_fhkeylen:	ignored
+ *				cl_fhkey:	ignored
+ *			size:	size in bytes of passed in nfsctl_client
+ * Output:
+ *	On success:	returns zero
+ *	On error:	return code is negative errno value
+ *
+ * Note: Only AF_INET client addresses are passed in, since
+ * nfsctl_client.cl_addrlist contains only in_addr fields for addresses.
+ */
 static ssize_t write_add(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_client *data;
@@ -198,6 +235,30 @@ static ssize_t write_add(struct file *file, char *buf, size_t size)
 	return exp_addclient(data);
 }
 
+/**
+ * write_del - Remove client from auth unix cache
+ *
+ * Deprecated.  /proc/net/rpc/auth.unix.ip is preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_client
+ *				cl_ident:	'\0'-terminated C string
+ *						containing domain name
+ *						of client
+ *				cl_naddr:	ignored
+ *				cl_addrlist:	ignored
+ *				cl_fhkeytype:	ignored
+ *				cl_fhkeylen:	ignored
+ *				cl_fhkey:	ignored
+ *			size:	size in bytes of passed in nfsctl_client
+ * Output:
+ *	On success:	returns zero
+ *	On error:	return code is negative errno value
+ *
+ * Note: Only AF_INET client addresses are passed in, since
+ * nfsctl_client.cl_addrlist contains only in_addr fields for addresses.
+ */
 static ssize_t write_del(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_client *data;
@@ -207,6 +268,33 @@ static ssize_t write_del(struct file *file, char *buf, size_t size)
 	return exp_delclient(data);
 }
 
+/**
+ * write_export - Export part or all of a local file system
+ *
+ * Deprecated.  /proc/net/rpc/{nfsd.export,nfsd.fh} are preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_export
+ *				ex_client:	'\0'-terminated C string
+ *						containing domain name
+ *						of client allowed to access
+ *						this export
+ *				ex_path:	'\0'-terminated C string
+ *						containing pathname of
+ *						directory in local file system
+ *				ex_dev:		fsid to use for this export
+ *				ex_ino:		ignored
+ *				ex_flags:	export flags for this export
+ *				ex_anon_uid:	UID to use for anonymous
+ *						requests
+ *				ex_anon_gid:	GID to use for anonymous
+ *						requests
+ *			size:	size in bytes of passed in nfsctl_export
+ * Output:
+ *	On success:	returns zero
+ *	On error:	return code is negative errno value
+ */
 static ssize_t write_export(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_export *data;
@@ -216,6 +304,31 @@ static ssize_t write_export(struct file *file, char *buf, size_t size)
 	return exp_export(data);
 }
 
+/**
+ * write_unexport - Unexport a previously exported file system
+ *
+ * Deprecated.  /proc/net/rpc/{nfsd.export,nfsd.fh} are preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_export
+ *				ex_client:	'\0'-terminated C string
+ *						containing domain name
+ *						of client no longer allowed
+ *						to access this export
+ *				ex_path:	'\0'-terminated C string
+ *						containing pathname of
+ *						directory in local file system
+ *				ex_dev:		ignored
+ *				ex_ino:		ignored
+ *				ex_flags:	ignored
+ *				ex_anon_uid:	ignored
+ *				ex_anon_gid:	ignored
+ *			size:	size in bytes of passed in nfsctl_export
+ * Output:
+ *	On success:	returns zero
+ *	On error:	return code is negative errno value
+ */
 static ssize_t write_unexport(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_export *data;
@@ -226,6 +339,30 @@ static ssize_t write_unexport(struct file *file, char *buf, size_t size)
 	return exp_unexport(data);
 }
 
+/**
+ * write_getfs - Get a variable-length NFS file handle by path
+ *
+ * Deprecated.  /proc/fs/nfsd/filehandle is preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_fsparm
+ *				gd_addr:	socket address of client
+ *				gd_path:	'\0'-terminated C string
+ *						containing pathname of
+ *						directory in local file system
+ *				gd_maxlen:	maximum size of returned file
+ *						handle
+ *			size:	size in bytes of passed in nfsctl_fsparm
+ * Output:
+ *	On success:	passed-in buffer filled with a knfsd_fh structure
+ *			(a variable-length raw NFS file handle);
+ *			return code is the size in bytes of the file handle
+ *	On error:	return code is negative errno value
+ *
+ * Note: Only AF_INET client addresses are passed in, since gd_addr
+ * is the same size as a struct sockaddr_in.
+ */
 static ssize_t write_getfs(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_fsparm *data;
@@ -265,6 +402,29 @@ static ssize_t write_getfs(struct file *file, char *buf, size_t size)
 	return err;
 }
 
+/**
+ * write_getfd - Get a fixed-length NFS file handle by path (used by mountd)
+ *
+ * Deprecated.  /proc/fs/nfsd/filehandle is preferred.
+ * Function remains to support old versions of nfs-utils.
+ *
+ * Input:
+ *			buf:	struct nfsctl_fdparm
+ *				gd_addr:	socket address of client
+ *				gd_path:	'\0'-terminated C string
+ *						containing pathname of
+ *						directory in local file system
+ *				gd_version:	fdparm structure version
+ *			size:	size in bytes of passed in nfsctl_fdparm
+ * Output:
+ *	On success:	passed-in buffer filled with nfsctl_res
+ *			(a fixed-length raw NFS file handle);
+ *			return code is the size in bytes of the file handle
+ *	On error:	return code is negative errno value
+ *
+ * Note: Only AF_INET client addresses are passed in, since gd_addr
+ * is the same size as a struct sockaddr_in.
+ */
 static ssize_t write_getfd(struct file *file, char *buf, size_t size)
 {
 	struct nfsctl_fdparm *data;
@@ -309,7 +469,23 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size)
 	return err;
 }
 
-static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size)
+/**
+ * write_unlock_ip - Release all locks used by a client
+ *
+ * Experimental.
+ *
+ * Input:
+ *			buf:	'\n'-terminated C string containing a
+ *				presentation format IPv4 address
+ *			size:	length of C string in @buf
+ * Output:
+ *	On success:	returns zero if all specified locks were released;
+ *			returns one if one or more locks were not released
+ *	On error:	return code is negative errno value
+ *
+ * Note: Only AF_INET client addresses are passed in
+ */
+static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size)
 {
 	struct sockaddr_in sin = {
 		.sin_family	= AF_INET,
@@ -339,7 +515,21 @@ static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size)
 	return nlmsvc_unlock_all_by_ip((struct sockaddr *)&sin);
 }
 
-static ssize_t failover_unlock_fs(struct file *file, char *buf, size_t size)
+/**
+ * write_unlock_fs - Release all locks on a local file system
+ *
+ * Experimental.
+ *
+ * Input:
+ *			buf:	'\n'-terminated C string containing the
+ *				absolute pathname of a local file system
+ *			size:	length of C string in @buf
+ * Output:
+ *	On success:	returns zero if all specified locks were released;
+ *			returns one if one or more locks were not released
+ *	On error:	return code is negative errno value
+ */
+static ssize_t write_unlock_fs(struct file *file, char *buf, size_t size)
 {
 	struct path path;
 	char *fo_path;
@@ -360,21 +550,44 @@ static ssize_t failover_unlock_fs(struct file *file, char *buf, size_t size)
 	if (error)
 		return error;
 
+	/*
+	 * XXX: Needs better sanity checking.  Otherwise we could end up
+	 * releasing locks on the wrong file system.
+	 *
+	 * For example:
+	 * 1.  Does the path refer to a directory?
+	 * 2.  Is that directory a mount point, or
+	 * 3.  Is that directory the root of an exported file system?
+	 */
 	error = nlmsvc_unlock_all_by_sb(path.mnt->mnt_sb);
 
 	path_put(&path);
 	return error;
 }
 
+/**
+ * write_filehandle - Get a variable-length NFS file handle by path
+ *
+ * On input, the buffer contains a '\n'-terminated C string comprised of
+ * three alphanumeric words separated by whitespace.  The string may
+ * contain escape sequences.
+ *
+ * Input:
+ *			buf:
+ *				domain:		client domain name
+ *				path:		export pathname
+ *				maxsize:	numeric maximum size of
+ *						@buf
+ *			size:	length of C string in @buf
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C
+ *			string containing a ASCII hex text version
+ *			of the NFS file handle;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is negative errno value
+ */
 static ssize_t write_filehandle(struct file *file, char *buf, size_t size)
 {
-	/* request is:
-	 *   domain path maxsize
-	 * response is
-	 *   filehandle
-	 *
-	 * qword quoting is used, so filehandle will be \x....
-	 */
 	char *dname, *path;
 	int uninitialized_var(maxsize);
 	char *mesg = buf;
@@ -391,11 +604,13 @@ static ssize_t write_filehandle(struct file *file, char *buf, size_t size)
 
 	dname = mesg;
 	len = qword_get(&mesg, dname, size);
-	if (len <= 0) return -EINVAL;
+	if (len <= 0)
+		return -EINVAL;
 	
 	path = dname+len+1;
 	len = qword_get(&mesg, path, size);
-	if (len <= 0) return -EINVAL;
+	if (len <= 0)
+		return -EINVAL;
 
 	len = get_int(&mesg, &maxsize);
 	if (len)
@@ -419,17 +634,43 @@ static ssize_t write_filehandle(struct file *file, char *buf, size_t size)
 	if (len)
 		return len;
 	
-	mesg = buf; len = SIMPLE_TRANSACTION_LIMIT;
+	mesg = buf;
+	len = SIMPLE_TRANSACTION_LIMIT;
 	qword_addhex(&mesg, &len, (char*)&fh.fh_base, fh.fh_size);
 	mesg[-1] = '\n';
 	return mesg - buf;	
 }
 
+/**
+ * write_threads - Start NFSD, or report the current number of running threads
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C
+ *			string numeric value representing the number of
+ *			running NFSD threads;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing an unsigned
+ *					integer value representing the
+ *					number of NFSD threads to start
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	NFS service is started;
+ *			passed-in buffer filled with '\n'-terminated C
+ *			string numeric value representing the number of
+ *			running NFSD threads;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ */
 static ssize_t write_threads(struct file *file, char *buf, size_t size)
 {
-	/* if size > 0, look for a number of threads and call nfsd_svc
-	 * then write out number of threads as reply
-	 */
 	char *mesg = buf;
 	int rv;
 	if (size > 0) {
@@ -437,9 +678,9 @@ static ssize_t write_threads(struct file *file, char *buf, size_t size)
 		rv = get_int(&mesg, &newthreads);
 		if (rv)
 			return rv;
-		if (newthreads <0)
+		if (newthreads < 0)
 			return -EINVAL;
-		rv = nfsd_svc(2049, newthreads);
+		rv = nfsd_svc(NFS_PORT, newthreads);
 		if (rv)
 			return rv;
 	}
@@ -447,6 +688,28 @@ static ssize_t write_threads(struct file *file, char *buf, size_t size)
 	return strlen(buf);
 }
 
+/**
+ * write_pool_threads - Set or report the current number of threads per pool
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ *
+ * OR
+ *
+ * Input:
+ * 			buf:		C string containing whitespace-
+ * 					separated unsigned integer values
+ *					representing the number of NFSD
+ *					threads to start in each pool
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C
+ *			string containing integer values representing the
+ *			number of NFSD threads in each pool;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ */
 static ssize_t write_pool_threads(struct file *file, char *buf, size_t size)
 {
 	/* if size > 0, look for an array of number of threads per node
@@ -517,10 +780,6 @@ out_free:
 
 static ssize_t __write_versions(struct file *file, char *buf, size_t size)
 {
-	/*
-	 * Format:
-	 *   [-/+]vers [-/+]vers ...
-	 */
 	char *mesg = buf;
 	char *vers, sign;
 	int len, num;
@@ -578,6 +837,38 @@ static ssize_t __write_versions(struct file *file, char *buf, size_t size)
 	return len;
 }
 
+/**
+ * write_versions - Set or report the available NFS protocol versions
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C
+ *			string containing positive or negative integer
+ *			values representing the current status of each
+ *			protocol version;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ *
+ * OR
+ *
+ * Input:
+ * 			buf:		C string containing whitespace-
+ * 					separated positive or negative
+ * 					integer values representing NFS
+ * 					protocol versions to enable ("+n")
+ * 					or disable ("-n")
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	status of zero or more protocol versions has
+ *			been updated; passed-in buffer filled with
+ *			'\n'-terminated C string containing positive
+ *			or negative integer values representing the
+ *			current status of each protocol version;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ */
 static ssize_t write_versions(struct file *file, char *buf, size_t size)
 {
 	ssize_t rv;
@@ -687,6 +978,75 @@ static ssize_t __write_ports(struct file *file, char *buf, size_t size)
 	return -EINVAL;
 }
 
+/**
+ * write_ports - Pass a socket file descriptor or transport name to listen on
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ * Output:
+ *	On success:	passed-in buffer filled with a '\n'-terminated C
+ *			string containing a whitespace-separated list of
+ *			named NFSD listeners;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing an unsigned
+ *					integer value representing a bound
+ *					but unconnected socket that is to be
+ *					used as an NFSD listener
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	NFS service is started;
+ *			passed-in buffer filled with a '\n'-terminated C
+ *			string containing a unique alphanumeric name of
+ *			the listener;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is a negative errno value
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing a "-" followed
+ *					by an integer value representing a
+ *					previously passed in socket file
+ *					descriptor
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	NFS service no longer listens on that socket;
+ *			passed-in buffer filled with a '\n'-terminated C
+ *			string containing a unique name of the listener;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is a negative errno value
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing a transport
+ *					name and an unsigned integer value
+ *					representing the port to listen on,
+ *					separated by whitespace
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	returns zero; NFS service is started
+ *	On error:	return code is a negative errno value
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing a "-" followed
+ *					by a transport name and an unsigned
+ *					integer value representing the port
+ *					to listen on, separated by whitespace
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	returns zero; NFS service no longer listens
+ *			on that transport
+ *	On error:	return code is a negative errno value
+ */
 static ssize_t write_ports(struct file *file, char *buf, size_t size)
 {
 	ssize_t rv;
@@ -700,6 +1060,27 @@ static ssize_t write_ports(struct file *file, char *buf, size_t size)
 
 int nfsd_max_blksize;
 
+/**
+ * write_maxblksize - Set or report the current NFS blksize
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ *
+ * OR
+ *
+ * Input:
+ * 			buf:		C string containing an unsigned
+ * 					integer value representing the new
+ * 					NFS blksize
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C string
+ *			containing numeric value of the current NFS blksize
+ *			setting;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ */
 static ssize_t write_maxblksize(struct file *file, char *buf, size_t size)
 {
 	char *mesg = buf;
@@ -752,6 +1133,27 @@ static ssize_t __write_leasetime(struct file *file, char *buf, size_t size)
 	return strlen(buf);
 }
 
+/**
+ * write_leasetime - Set or report the current NFSv4 lease time
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing an unsigned
+ *					integer value representing the new
+ *					NFSv4 lease expiry time
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C
+ *			string containing unsigned integer value of the
+ *			current lease expiry time;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ */
 static ssize_t write_leasetime(struct file *file, char *buf, size_t size)
 {
 	ssize_t rv;
@@ -788,6 +1190,27 @@ static ssize_t __write_recoverydir(struct file *file, char *buf, size_t size)
 	return strlen(buf);
 }
 
+/**
+ * write_recoverydir - Set or report the pathname of the recovery directory
+ *
+ * Input:
+ *			buf:		ignored
+ *			size:		zero
+ *
+ * OR
+ *
+ * Input:
+ *			buf:		C string containing the pathname
+ *					of the directory on a local file
+ *					system containing permanent NFSv4
+ *					recovery data
+ *			size:		non-zero length of C string in @buf
+ * Output:
+ *	On success:	passed-in buffer filled with '\n'-terminated C string
+ *			containing the current recovery pathname setting;
+ *			return code is the size in bytes of the string
+ *	On error:	return code is zero or a negative errno value
+ */
 static ssize_t write_recoverydir(struct file *file, char *buf, size_t size)
 {
 	ssize_t rv;
diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
index f0da7d9c3a9..9f1ca17293d 100644
--- a/fs/nfsd/nfsfh.c
+++ b/fs/nfsd/nfsfh.c
@@ -258,14 +258,32 @@ out:
 	return error;
 }
 
-/*
- * Perform sanity checks on the dentry in a client's file handle.
+/**
+ * fh_verify - filehandle lookup and access checking
+ * @rqstp: pointer to current rpc request
+ * @fhp: filehandle to be verified
+ * @type: expected type of object pointed to by filehandle
+ * @access: type of access needed to object
+ *
+ * Look up a dentry from the on-the-wire filehandle, check the client's
+ * access to the export, and set the current task's credentials.
+ *
+ * Regardless of success or failure of fh_verify(), fh_put() should be
+ * called on @fhp when the caller is finished with the filehandle.
  *
- * Note that the file handle dentry may need to be freed even after
- * an error return.
+ * fh_verify() may be called multiple times on a given filehandle, for
+ * example, when processing an NFSv4 compound.  The first call will look
+ * up a dentry using the on-the-wire filehandle.  Subsequent calls will
+ * skip the lookup and just perform the other checks and possibly change
+ * the current task's credentials.
  *
- * This is only called at the start of an nfsproc call, so fhp points to
- * a svc_fh which is all 0 except for the over-the-wire file handle.
+ * @type specifies the type of object expected using one of the S_IF*
+ * constants defined in include/linux/stat.h.  The caller may use zero
+ * to indicate that it doesn't care, or a negative integer to indicate
+ * that it expects something not of the given type.
+ *
+ * @access is formed from the NFSD_MAY_* constants defined in
+ * include/linux/nfsd/nfsd.h.
  */
 __be32
 fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
@@ -466,6 +484,8 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
 				goto retry;
 			break;
 		}
+	} else if (exp->ex_flags & NFSEXP_FSID) {
+		fsid_type = FSID_NUM;
 	} else if (exp->ex_uuid) {
 		if (fhp->fh_maxsize >= 64) {
 			if (root_export)
@@ -478,9 +498,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
 			else
 				fsid_type = FSID_UUID4_INUM;
 		}
-	} else if (exp->ex_flags & NFSEXP_FSID)
-		fsid_type = FSID_NUM;
-	else if (!old_valid_dev(ex_dev))
+	} else if (!old_valid_dev(ex_dev))
 		/* for newer device numbers, we must use a newer fsid format */
 		fsid_type = FSID_ENCODE_DEV;
 	else
diff --git a/fs/nfsd/nfsproc.c b/fs/nfsd/nfsproc.c
index 5cffeca7ace..6f7f2635122 100644
--- a/fs/nfsd/nfsproc.c
+++ b/fs/nfsd/nfsproc.c
@@ -622,6 +622,7 @@ nfserrno (int errno)
 		{ nfserr_badname, -ESRCH },
 		{ nfserr_io, -ETXTBSY },
 		{ nfserr_notsupp, -EOPNOTSUPP },
+		{ nfserr_toosmall, -ETOOSMALL },
 	};
 	int	i;
 
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index d1c5f787b36..6e50aaa56ca 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -764,7 +764,6 @@ static inline int nfsd_dosync(struct file *filp, struct dentry *dp,
 
 	return err;
 }
-	
 
 static int
 nfsd_sync(struct file *filp)
@@ -1211,7 +1210,7 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	dirp = dentry->d_inode;
 
 	err = nfserr_notdir;
-	if(!dirp->i_op || !dirp->i_op->lookup)
+	if (!dirp->i_op->lookup)
 		goto out;
 	/*
 	 * Check whether the response file handle has been verified yet.
@@ -1347,7 +1346,7 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	/* Get all the sanity checks out of the way before
 	 * we lock the parent. */
 	err = nfserr_notdir;
-	if(!dirp->i_op || !dirp->i_op->lookup)
+	if (!dirp->i_op->lookup)
 		goto out;
 	fh_lock_nested(fhp, I_MUTEX_PARENT);
 
@@ -1482,7 +1481,7 @@ nfsd_readlink(struct svc_rqst *rqstp, struct svc_fh *fhp, char *buf, int *lenp)
 	inode = dentry->d_inode;
 
 	err = nfserr_inval;
-	if (!inode->i_op || !inode->i_op->readlink)
+	if (!inode->i_op->readlink)
 		goto out;
 
 	touch_atime(fhp->fh_export->ex_path.mnt, dentry);
@@ -2162,7 +2161,7 @@ nfsd_set_posix_acl(struct svc_fh *fhp, int type, struct posix_acl *acl)
 	size_t size;
 	int error;
 
-	if (!IS_POSIXACL(inode) || !inode->i_op ||
+	if (!IS_POSIXACL(inode) ||
 	    !inode->i_op->setxattr || !inode->i_op->removexattr)
 		return -EOPNOTSUPP;
 	switch(type) {
diff --git a/fs/notify/Kconfig b/fs/notify/Kconfig
new file mode 100644
index 00000000000..50914d7303c
--- /dev/null
+++ b/fs/notify/Kconfig
@@ -0,0 +1,2 @@
+source "fs/notify/dnotify/Kconfig"
+source "fs/notify/inotify/Kconfig"
diff --git a/fs/notify/Makefile b/fs/notify/Makefile
new file mode 100644
index 00000000000..5a95b6010ce
--- /dev/null
+++ b/fs/notify/Makefile
@@ -0,0 +1,2 @@
+obj-y			+= dnotify/
+obj-y			+= inotify/
diff --git a/fs/notify/dnotify/Kconfig b/fs/notify/dnotify/Kconfig
new file mode 100644
index 00000000000..26adf5dfa64
--- /dev/null
+++ b/fs/notify/dnotify/Kconfig
@@ -0,0 +1,10 @@
+config DNOTIFY
+	bool "Dnotify support"
+	default y
+	help
+	  Dnotify is a directory-based per-fd file change notification system
+	  that uses signals to communicate events to user-space.  There exist
+	  superior alternatives, but some applications may still rely on
+	  dnotify.
+
+	  If unsure, say Y.
diff --git a/fs/notify/dnotify/Makefile b/fs/notify/dnotify/Makefile
new file mode 100644
index 00000000000..f145251dcad
--- /dev/null
+++ b/fs/notify/dnotify/Makefile
@@ -0,0 +1 @@
+obj-$(CONFIG_DNOTIFY)		+= dnotify.o
diff --git a/fs/dnotify.c b/fs/notify/dnotify/dnotify.c
index 676073b8dda..b0aa2cde80b 100644
--- a/fs/dnotify.c
+++ b/fs/notify/dnotify/dnotify.c
@@ -115,9 +115,6 @@ int fcntl_dirnotify(int fd, struct file *filp, unsigned long arg)
 	dn->dn_next = inode->i_dnotify;
 	inode->i_dnotify = dn;
 	spin_unlock(&inode->i_lock);
-
-	if (filp->f_op && filp->f_op->dir_notify)
-		return filp->f_op->dir_notify(filp, arg);
 	return 0;
 
 out_free:
diff --git a/fs/notify/inotify/Kconfig b/fs/notify/inotify/Kconfig
new file mode 100644
index 00000000000..44679284102
--- /dev/null
+++ b/fs/notify/inotify/Kconfig
@@ -0,0 +1,27 @@
+config INOTIFY
+	bool "Inotify file change notification support"
+	default y
+	---help---
+	  Say Y here to enable inotify support.  Inotify is a file change
+	  notification system and a replacement for dnotify.  Inotify fixes
+	  numerous shortcomings in dnotify and introduces several new features
+	  including multiple file events, one-shot support, and unmount
+	  notification.
+
+	  For more information, see <file:Documentation/filesystems/inotify.txt>
+
+	  If unsure, say Y.
+
+config INOTIFY_USER
+	bool "Inotify support for userspace"
+	depends on INOTIFY
+	default y
+	---help---
+	  Say Y here to enable inotify support for userspace, including the
+	  associated system calls.  Inotify allows monitoring of both files and
+	  directories via a single open fd.  Events are read from the file
+	  descriptor, which is also select()- and poll()-able.
+
+	  For more information, see <file:Documentation/filesystems/inotify.txt>
+
+	  If unsure, say Y.
diff --git a/fs/notify/inotify/Makefile b/fs/notify/inotify/Makefile
new file mode 100644
index 00000000000..e290f3bb9d8
--- /dev/null
+++ b/fs/notify/inotify/Makefile
@@ -0,0 +1,2 @@
+obj-$(CONFIG_INOTIFY)		+= inotify.o
+obj-$(CONFIG_INOTIFY_USER)	+= inotify_user.o
diff --git a/fs/inotify.c b/fs/notify/inotify/inotify.c
index dae3f28f30d..dae3f28f30d 100644
--- a/fs/inotify.c
+++ b/fs/notify/inotify/inotify.c
diff --git a/fs/inotify_user.c b/fs/notify/inotify/inotify_user.c
index e2425bbd871..81b8644b013 100644
--- a/fs/inotify_user.c
+++ b/fs/notify/inotify/inotify_user.c
@@ -76,10 +76,10 @@ struct inotify_device {
 	struct mutex		ev_mutex;	/* protects event queue */
 	struct mutex		up_mutex;	/* synchronizes watch updates */
 	struct list_head 	events;		/* list of queued events */
-	atomic_t		count;		/* reference count */
 	struct user_struct	*user;		/* user who opened this dev */
 	struct inotify_handle	*ih;		/* inotify handle */
 	struct fasync_struct    *fa;            /* async notification */
+	atomic_t		count;		/* reference count */
 	unsigned int		queue_size;	/* size of the queue (bytes) */
 	unsigned int		event_count;	/* number of pending events */
 	unsigned int		max_events;	/* maximum number of events */
@@ -704,7 +704,7 @@ fput_and_out:
 	return ret;
 }
 
-asmlinkage long sys_inotify_rm_watch(int fd, u32 wd)
+asmlinkage long sys_inotify_rm_watch(int fd, __s32 wd)
 {
 	struct file *filp;
 	struct inotify_device *dev;
diff --git a/fs/ntfs/inode.c b/fs/ntfs/inode.c
index e9da092e277..86bef156cf0 100644
--- a/fs/ntfs/inode.c
+++ b/fs/ntfs/inode.c
@@ -1406,9 +1406,6 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 		ni->allocated_size = sle64_to_cpu(
 				a->data.non_resident.allocated_size);
 	}
-	/* Setup the operations for this attribute inode. */
-	vi->i_op = NULL;
-	vi->i_fop = NULL;
 	if (NInoMstProtected(ni))
 		vi->i_mapping->a_ops = &ntfs_mst_aops;
 	else
diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile
index 589dcdfdfe3..01596079dd6 100644
--- a/fs/ocfs2/Makefile
+++ b/fs/ocfs2/Makefile
@@ -12,6 +12,7 @@ obj-$(CONFIG_OCFS2_FS_USERSPACE_CLUSTER) += ocfs2_stack_user.o
 ocfs2-objs := \
 	alloc.o 		\
 	aops.o 			\
+	blockcheck.o		\
 	buffer_head_io.o	\
 	dcache.o 		\
 	dir.o 			\
@@ -35,8 +36,14 @@ ocfs2-objs := \
 	sysfile.o 		\
 	uptodate.o		\
 	ver.o			\
+	quota_local.o		\
+	quota_global.o		\
 	xattr.o
 
+ifeq ($(CONFIG_OCFS2_FS_POSIX_ACL),y)
+ocfs2-objs += acl.o
+endif
+
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
 ocfs2_stack_user-objs := stack_user.o
diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c
new file mode 100644
index 00000000000..12dfb44c22e
--- /dev/null
+++ b/fs/ocfs2/acl.c
@@ -0,0 +1,479 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * acl.c
+ *
+ * Copyright (C) 2004, 2008 Oracle.  All rights reserved.
+ *
+ * CREDITS:
+ * Lots of code in this file is copy from linux/fs/ext3/acl.c.
+ * Copyright (C) 2001-2003 Andreas Gruenbacher, <agruen@suse.de>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/string.h>
+
+#define MLOG_MASK_PREFIX ML_INODE
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+#include "alloc.h"
+#include "dlmglue.h"
+#include "file.h"
+#include "ocfs2_fs.h"
+
+#include "xattr.h"
+#include "acl.h"
+
+/*
+ * Convert from xattr value to acl struct.
+ */
+static struct posix_acl *ocfs2_acl_from_xattr(const void *value, size_t size)
+{
+	int n, count;
+	struct posix_acl *acl;
+
+	if (!value)
+		return NULL;
+	if (size < sizeof(struct posix_acl_entry))
+		return ERR_PTR(-EINVAL);
+
+	count = size / sizeof(struct posix_acl_entry);
+	if (count < 0)
+		return ERR_PTR(-EINVAL);
+	if (count == 0)
+		return NULL;
+
+	acl = posix_acl_alloc(count, GFP_NOFS);
+	if (!acl)
+		return ERR_PTR(-ENOMEM);
+	for (n = 0; n < count; n++) {
+		struct ocfs2_acl_entry *entry =
+			(struct ocfs2_acl_entry *)value;
+
+		acl->a_entries[n].e_tag  = le16_to_cpu(entry->e_tag);
+		acl->a_entries[n].e_perm = le16_to_cpu(entry->e_perm);
+		acl->a_entries[n].e_id   = le32_to_cpu(entry->e_id);
+		value += sizeof(struct posix_acl_entry);
+
+	}
+	return acl;
+}
+
+/*
+ * Convert acl struct to xattr value.
+ */
+static void *ocfs2_acl_to_xattr(const struct posix_acl *acl, size_t *size)
+{
+	struct ocfs2_acl_entry *entry = NULL;
+	char *ocfs2_acl;
+	size_t n;
+
+	*size = acl->a_count * sizeof(struct posix_acl_entry);
+
+	ocfs2_acl = kmalloc(*size, GFP_NOFS);
+	if (!ocfs2_acl)
+		return ERR_PTR(-ENOMEM);
+
+	entry = (struct ocfs2_acl_entry *)ocfs2_acl;
+	for (n = 0; n < acl->a_count; n++, entry++) {
+		entry->e_tag  = cpu_to_le16(acl->a_entries[n].e_tag);
+		entry->e_perm = cpu_to_le16(acl->a_entries[n].e_perm);
+		entry->e_id   = cpu_to_le32(acl->a_entries[n].e_id);
+	}
+	return ocfs2_acl;
+}
+
+static struct posix_acl *ocfs2_get_acl_nolock(struct inode *inode,
+					      int type,
+					      struct buffer_head *di_bh)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	int name_index;
+	char *value = NULL;
+	struct posix_acl *acl;
+	int retval;
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return NULL;
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		name_index = OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS;
+		break;
+	case ACL_TYPE_DEFAULT:
+		name_index = OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT;
+		break;
+	default:
+		return ERR_PTR(-EINVAL);
+	}
+
+	retval = ocfs2_xattr_get_nolock(inode, di_bh, name_index, "", NULL, 0);
+	if (retval > 0) {
+		value = kmalloc(retval, GFP_NOFS);
+		if (!value)
+			return ERR_PTR(-ENOMEM);
+		retval = ocfs2_xattr_get_nolock(inode, di_bh, name_index,
+						"", value, retval);
+	}
+
+	if (retval > 0)
+		acl = ocfs2_acl_from_xattr(value, retval);
+	else if (retval == -ENODATA || retval == 0)
+		acl = NULL;
+	else
+		acl = ERR_PTR(retval);
+
+	kfree(value);
+
+	return acl;
+}
+
+
+/*
+ * Get posix acl.
+ */
+static struct posix_acl *ocfs2_get_acl(struct inode *inode, int type)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct buffer_head *di_bh = NULL;
+	struct posix_acl *acl;
+	int ret;
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return NULL;
+
+	ret = ocfs2_inode_lock(inode, &di_bh, 0);
+	if (ret < 0) {
+		mlog_errno(ret);
+		acl = ERR_PTR(ret);
+		return acl;
+	}
+
+	acl = ocfs2_get_acl_nolock(inode, type, di_bh);
+
+	ocfs2_inode_unlock(inode, 0);
+
+	brelse(di_bh);
+
+	return acl;
+}
+
+/*
+ * Set the access or default ACL of an inode.
+ */
+static int ocfs2_set_acl(handle_t *handle,
+			 struct inode *inode,
+			 struct buffer_head *di_bh,
+			 int type,
+			 struct posix_acl *acl,
+			 struct ocfs2_alloc_context *meta_ac,
+			 struct ocfs2_alloc_context *data_ac)
+{
+	int name_index;
+	void *value = NULL;
+	size_t size = 0;
+	int ret;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case ACL_TYPE_ACCESS:
+		name_index = OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS;
+		if (acl) {
+			mode_t mode = inode->i_mode;
+			ret = posix_acl_equiv_mode(acl, &mode);
+			if (ret < 0)
+				return ret;
+			else {
+				inode->i_mode = mode;
+				if (ret == 0)
+					acl = NULL;
+			}
+		}
+		break;
+	case ACL_TYPE_DEFAULT:
+		name_index = OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT;
+		if (!S_ISDIR(inode->i_mode))
+			return acl ? -EACCES : 0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (acl) {
+		value = ocfs2_acl_to_xattr(acl, &size);
+		if (IS_ERR(value))
+			return (int)PTR_ERR(value);
+	}
+
+	if (handle)
+		ret = ocfs2_xattr_set_handle(handle, inode, di_bh, name_index,
+					     "", value, size, 0,
+					     meta_ac, data_ac);
+	else
+		ret = ocfs2_xattr_set(inode, name_index, "", value, size, 0);
+
+	kfree(value);
+
+	return ret;
+}
+
+int ocfs2_check_acl(struct inode *inode, int mask)
+{
+	struct posix_acl *acl = ocfs2_get_acl(inode, ACL_TYPE_ACCESS);
+
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl) {
+		int ret = posix_acl_permission(inode, acl, mask);
+		posix_acl_release(acl);
+		return ret;
+	}
+
+	return -EAGAIN;
+}
+
+int ocfs2_acl_chmod(struct inode *inode)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct posix_acl *acl, *clone;
+	int ret;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return 0;
+
+	acl = ocfs2_get_acl(inode, ACL_TYPE_ACCESS);
+	if (IS_ERR(acl) || !acl)
+		return PTR_ERR(acl);
+	clone = posix_acl_clone(acl, GFP_KERNEL);
+	posix_acl_release(acl);
+	if (!clone)
+		return -ENOMEM;
+	ret = posix_acl_chmod_masq(clone, inode->i_mode);
+	if (!ret)
+		ret = ocfs2_set_acl(NULL, inode, NULL, ACL_TYPE_ACCESS,
+				    clone, NULL, NULL);
+	posix_acl_release(clone);
+	return ret;
+}
+
+/*
+ * Initialize the ACLs of a new inode. If parent directory has default ACL,
+ * then clone to new inode. Called from ocfs2_mknod.
+ */
+int ocfs2_init_acl(handle_t *handle,
+		   struct inode *inode,
+		   struct inode *dir,
+		   struct buffer_head *di_bh,
+		   struct buffer_head *dir_bh,
+		   struct ocfs2_alloc_context *meta_ac,
+		   struct ocfs2_alloc_context *data_ac)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct posix_acl *acl = NULL;
+	int ret = 0;
+
+	if (!S_ISLNK(inode->i_mode)) {
+		if (osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) {
+			acl = ocfs2_get_acl_nolock(dir, ACL_TYPE_DEFAULT,
+						   dir_bh);
+			if (IS_ERR(acl))
+				return PTR_ERR(acl);
+		}
+		if (!acl)
+			inode->i_mode &= ~current->fs->umask;
+	}
+	if ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) && acl) {
+		struct posix_acl *clone;
+		mode_t mode;
+
+		if (S_ISDIR(inode->i_mode)) {
+			ret = ocfs2_set_acl(handle, inode, di_bh,
+					    ACL_TYPE_DEFAULT, acl,
+					    meta_ac, data_ac);
+			if (ret)
+				goto cleanup;
+		}
+		clone = posix_acl_clone(acl, GFP_NOFS);
+		ret = -ENOMEM;
+		if (!clone)
+			goto cleanup;
+
+		mode = inode->i_mode;
+		ret = posix_acl_create_masq(clone, &mode);
+		if (ret >= 0) {
+			inode->i_mode = mode;
+			if (ret > 0) {
+				ret = ocfs2_set_acl(handle, inode,
+						    di_bh, ACL_TYPE_ACCESS,
+						    clone, meta_ac, data_ac);
+			}
+		}
+		posix_acl_release(clone);
+	}
+cleanup:
+	posix_acl_release(acl);
+	return ret;
+}
+
+static size_t ocfs2_xattr_list_acl_access(struct inode *inode,
+					  char *list,
+					  size_t list_len,
+					  const char *name,
+					  size_t name_len)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	const size_t size = sizeof(POSIX_ACL_XATTR_ACCESS);
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return 0;
+
+	if (list && size <= list_len)
+		memcpy(list, POSIX_ACL_XATTR_ACCESS, size);
+	return size;
+}
+
+static size_t ocfs2_xattr_list_acl_default(struct inode *inode,
+					   char *list,
+					   size_t list_len,
+					   const char *name,
+					   size_t name_len)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	const size_t size = sizeof(POSIX_ACL_XATTR_DEFAULT);
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return 0;
+
+	if (list && size <= list_len)
+		memcpy(list, POSIX_ACL_XATTR_DEFAULT, size);
+	return size;
+}
+
+static int ocfs2_xattr_get_acl(struct inode *inode,
+			       int type,
+			       void *buffer,
+			       size_t size)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct posix_acl *acl;
+	int ret;
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return -EOPNOTSUPP;
+
+	acl = ocfs2_get_acl(inode, type);
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl == NULL)
+		return -ENODATA;
+	ret = posix_acl_to_xattr(acl, buffer, size);
+	posix_acl_release(acl);
+
+	return ret;
+}
+
+static int ocfs2_xattr_get_acl_access(struct inode *inode,
+				      const char *name,
+				      void *buffer,
+				      size_t size)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ocfs2_xattr_get_acl(inode, ACL_TYPE_ACCESS, buffer, size);
+}
+
+static int ocfs2_xattr_get_acl_default(struct inode *inode,
+				       const char *name,
+				       void *buffer,
+				       size_t size)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ocfs2_xattr_get_acl(inode, ACL_TYPE_DEFAULT, buffer, size);
+}
+
+static int ocfs2_xattr_set_acl(struct inode *inode,
+			       int type,
+			       const void *value,
+			       size_t size)
+{
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct posix_acl *acl;
+	int ret = 0;
+
+	if (!(osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL))
+		return -EOPNOTSUPP;
+
+	if (!is_owner_or_cap(inode))
+		return -EPERM;
+
+	if (value) {
+		acl = posix_acl_from_xattr(value, size);
+		if (IS_ERR(acl))
+			return PTR_ERR(acl);
+		else if (acl) {
+			ret = posix_acl_valid(acl);
+			if (ret)
+				goto cleanup;
+		}
+	} else
+		acl = NULL;
+
+	ret = ocfs2_set_acl(NULL, inode, NULL, type, acl, NULL, NULL);
+
+cleanup:
+	posix_acl_release(acl);
+	return ret;
+}
+
+static int ocfs2_xattr_set_acl_access(struct inode *inode,
+				      const char *name,
+				      const void *value,
+				      size_t size,
+				      int flags)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ocfs2_xattr_set_acl(inode, ACL_TYPE_ACCESS, value, size);
+}
+
+static int ocfs2_xattr_set_acl_default(struct inode *inode,
+				       const char *name,
+				       const void *value,
+				       size_t size,
+				       int flags)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ocfs2_xattr_set_acl(inode, ACL_TYPE_DEFAULT, value, size);
+}
+
+struct xattr_handler ocfs2_xattr_acl_access_handler = {
+	.prefix	= POSIX_ACL_XATTR_ACCESS,
+	.list	= ocfs2_xattr_list_acl_access,
+	.get	= ocfs2_xattr_get_acl_access,
+	.set	= ocfs2_xattr_set_acl_access,
+};
+
+struct xattr_handler ocfs2_xattr_acl_default_handler = {
+	.prefix	= POSIX_ACL_XATTR_DEFAULT,
+	.list	= ocfs2_xattr_list_acl_default,
+	.get	= ocfs2_xattr_get_acl_default,
+	.set	= ocfs2_xattr_set_acl_default,
+};
diff --git a/fs/ocfs2/acl.h b/fs/ocfs2/acl.h
new file mode 100644
index 00000000000..8f6389ed4da
--- /dev/null
+++ b/fs/ocfs2/acl.h
@@ -0,0 +1,58 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * acl.h
+ *
+ * Copyright (C) 2004, 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#ifndef OCFS2_ACL_H
+#define OCFS2_ACL_H
+
+#include <linux/posix_acl_xattr.h>
+
+struct ocfs2_acl_entry {
+	__le16 e_tag;
+	__le16 e_perm;
+	__le32 e_id;
+};
+
+#ifdef CONFIG_OCFS2_FS_POSIX_ACL
+
+extern int ocfs2_check_acl(struct inode *, int);
+extern int ocfs2_acl_chmod(struct inode *);
+extern int ocfs2_init_acl(handle_t *, struct inode *, struct inode *,
+			  struct buffer_head *, struct buffer_head *,
+			  struct ocfs2_alloc_context *,
+			  struct ocfs2_alloc_context *);
+
+#else /* CONFIG_OCFS2_FS_POSIX_ACL*/
+
+#define ocfs2_check_acl NULL
+static inline int ocfs2_acl_chmod(struct inode *inode)
+{
+	return 0;
+}
+static inline int ocfs2_init_acl(handle_t *handle,
+				 struct inode *inode,
+				 struct inode *dir,
+				 struct buffer_head *di_bh,
+				 struct buffer_head *dir_bh,
+				 struct ocfs2_alloc_context *meta_ac,
+				 struct ocfs2_alloc_context *data_ac)
+{
+	return 0;
+}
+
+#endif /* CONFIG_OCFS2_FS_POSIX_ACL*/
+
+#endif /* OCFS2_ACL_H */
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 0cc2deb9394..d861096c9d8 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -28,6 +28,7 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/swap.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_DISK_ALLOC
 #include <cluster/masklog.h>
@@ -36,6 +37,7 @@
 
 #include "alloc.h"
 #include "aops.h"
+#include "blockcheck.h"
 #include "dlmglue.h"
 #include "extent_map.h"
 #include "inode.h"
@@ -46,6 +48,7 @@
 #include "file.h"
 #include "super.h"
 #include "uptodate.h"
+#include "xattr.h"
 
 #include "buffer_head_io.h"
 
@@ -187,20 +190,12 @@ static int ocfs2_dinode_insert_check(struct inode *inode,
 static int ocfs2_dinode_sanity_check(struct inode *inode,
 				     struct ocfs2_extent_tree *et)
 {
-	int ret = 0;
-	struct ocfs2_dinode *di;
+	struct ocfs2_dinode *di = et->et_object;
 
 	BUG_ON(et->et_ops != &ocfs2_dinode_et_ops);
+	BUG_ON(!OCFS2_IS_VALID_DINODE(di));
 
-	di = et->et_object;
-	if (!OCFS2_IS_VALID_DINODE(di)) {
-		ret = -EIO;
-		ocfs2_error(inode->i_sb,
-			"Inode %llu has invalid path root",
-			(unsigned long long)OCFS2_I(inode)->ip_blkno);
-	}
-
-	return ret;
+	return 0;
 }
 
 static void ocfs2_dinode_fill_root_el(struct ocfs2_extent_tree *et)
@@ -213,36 +208,33 @@ static void ocfs2_dinode_fill_root_el(struct ocfs2_extent_tree *et)
 
 static void ocfs2_xattr_value_fill_root_el(struct ocfs2_extent_tree *et)
 {
-	struct ocfs2_xattr_value_root *xv = et->et_object;
+	struct ocfs2_xattr_value_buf *vb = et->et_object;
 
-	et->et_root_el = &xv->xr_list;
+	et->et_root_el = &vb->vb_xv->xr_list;
 }
 
 static void ocfs2_xattr_value_set_last_eb_blk(struct ocfs2_extent_tree *et,
 					      u64 blkno)
 {
-	struct ocfs2_xattr_value_root *xv =
-		(struct ocfs2_xattr_value_root *)et->et_object;
+	struct ocfs2_xattr_value_buf *vb = et->et_object;
 
-	xv->xr_last_eb_blk = cpu_to_le64(blkno);
+	vb->vb_xv->xr_last_eb_blk = cpu_to_le64(blkno);
 }
 
 static u64 ocfs2_xattr_value_get_last_eb_blk(struct ocfs2_extent_tree *et)
 {
-	struct ocfs2_xattr_value_root *xv =
-		(struct ocfs2_xattr_value_root *) et->et_object;
+	struct ocfs2_xattr_value_buf *vb = et->et_object;
 
-	return le64_to_cpu(xv->xr_last_eb_blk);
+	return le64_to_cpu(vb->vb_xv->xr_last_eb_blk);
 }
 
 static void ocfs2_xattr_value_update_clusters(struct inode *inode,
 					      struct ocfs2_extent_tree *et,
 					      u32 clusters)
 {
-	struct ocfs2_xattr_value_root *xv =
-		(struct ocfs2_xattr_value_root *)et->et_object;
+	struct ocfs2_xattr_value_buf *vb = et->et_object;
 
-	le32_add_cpu(&xv->xr_clusters, clusters);
+	le32_add_cpu(&vb->vb_xv->xr_clusters, clusters);
 }
 
 static struct ocfs2_extent_tree_operations ocfs2_xattr_value_et_ops = {
@@ -304,11 +296,13 @@ static struct ocfs2_extent_tree_operations ocfs2_xattr_tree_et_ops = {
 static void __ocfs2_init_extent_tree(struct ocfs2_extent_tree *et,
 				     struct inode *inode,
 				     struct buffer_head *bh,
+				     ocfs2_journal_access_func access,
 				     void *obj,
 				     struct ocfs2_extent_tree_operations *ops)
 {
 	et->et_ops = ops;
 	et->et_root_bh = bh;
+	et->et_root_journal_access = access;
 	if (!obj)
 		obj = (void *)bh->b_data;
 	et->et_object = obj;
@@ -324,23 +318,23 @@ void ocfs2_init_dinode_extent_tree(struct ocfs2_extent_tree *et,
 				   struct inode *inode,
 				   struct buffer_head *bh)
 {
-	__ocfs2_init_extent_tree(et, inode, bh, NULL, &ocfs2_dinode_et_ops);
+	__ocfs2_init_extent_tree(et, inode, bh, ocfs2_journal_access_di,
+				 NULL, &ocfs2_dinode_et_ops);
 }
 
 void ocfs2_init_xattr_tree_extent_tree(struct ocfs2_extent_tree *et,
 				       struct inode *inode,
 				       struct buffer_head *bh)
 {
-	__ocfs2_init_extent_tree(et, inode, bh, NULL,
-				 &ocfs2_xattr_tree_et_ops);
+	__ocfs2_init_extent_tree(et, inode, bh, ocfs2_journal_access_xb,
+				 NULL, &ocfs2_xattr_tree_et_ops);
 }
 
 void ocfs2_init_xattr_value_extent_tree(struct ocfs2_extent_tree *et,
 					struct inode *inode,
-					struct buffer_head *bh,
-					struct ocfs2_xattr_value_root *xv)
+					struct ocfs2_xattr_value_buf *vb)
 {
-	__ocfs2_init_extent_tree(et, inode, bh, xv,
+	__ocfs2_init_extent_tree(et, inode, vb->vb_bh, vb->vb_access, vb,
 				 &ocfs2_xattr_value_et_ops);
 }
 
@@ -362,6 +356,15 @@ static inline void ocfs2_et_update_clusters(struct inode *inode,
 	et->et_ops->eo_update_clusters(inode, et, clusters);
 }
 
+static inline int ocfs2_et_root_journal_access(handle_t *handle,
+					       struct inode *inode,
+					       struct ocfs2_extent_tree *et,
+					       int type)
+{
+	return et->et_root_journal_access(handle, inode, et->et_root_bh,
+					  type);
+}
+
 static inline int ocfs2_et_insert_check(struct inode *inode,
 					struct ocfs2_extent_tree *et,
 					struct ocfs2_extent_rec *rec)
@@ -402,12 +405,14 @@ struct ocfs2_path_item {
 #define OCFS2_MAX_PATH_DEPTH	5
 
 struct ocfs2_path {
-	int			p_tree_depth;
-	struct ocfs2_path_item	p_node[OCFS2_MAX_PATH_DEPTH];
+	int				p_tree_depth;
+	ocfs2_journal_access_func	p_root_access;
+	struct ocfs2_path_item		p_node[OCFS2_MAX_PATH_DEPTH];
 };
 
 #define path_root_bh(_path) ((_path)->p_node[0].bh)
 #define path_root_el(_path) ((_path)->p_node[0].el)
+#define path_root_access(_path)((_path)->p_root_access)
 #define path_leaf_bh(_path) ((_path)->p_node[(_path)->p_tree_depth].bh)
 #define path_leaf_el(_path) ((_path)->p_node[(_path)->p_tree_depth].el)
 #define path_num_items(_path) ((_path)->p_tree_depth + 1)
@@ -440,6 +445,8 @@ static void ocfs2_reinit_path(struct ocfs2_path *path, int keep_root)
 	 */
 	if (keep_root)
 		depth = le16_to_cpu(path_root_el(path)->l_tree_depth);
+	else
+		path_root_access(path) = NULL;
 
 	path->p_tree_depth = depth;
 }
@@ -465,6 +472,7 @@ static void ocfs2_cp_path(struct ocfs2_path *dest, struct ocfs2_path *src)
 
 	BUG_ON(path_root_bh(dest) != path_root_bh(src));
 	BUG_ON(path_root_el(dest) != path_root_el(src));
+	BUG_ON(path_root_access(dest) != path_root_access(src));
 
 	ocfs2_reinit_path(dest, 1);
 
@@ -486,6 +494,7 @@ static void ocfs2_mv_path(struct ocfs2_path *dest, struct ocfs2_path *src)
 	int i;
 
 	BUG_ON(path_root_bh(dest) != path_root_bh(src));
+	BUG_ON(path_root_access(dest) != path_root_access(src));
 
 	for(i = 1; i < OCFS2_MAX_PATH_DEPTH; i++) {
 		brelse(dest->p_node[i].bh);
@@ -521,7 +530,8 @@ static inline void ocfs2_path_insert_eb(struct ocfs2_path *path, int index,
 }
 
 static struct ocfs2_path *ocfs2_new_path(struct buffer_head *root_bh,
-					 struct ocfs2_extent_list *root_el)
+					 struct ocfs2_extent_list *root_el,
+					 ocfs2_journal_access_func access)
 {
 	struct ocfs2_path *path;
 
@@ -533,11 +543,48 @@ static struct ocfs2_path *ocfs2_new_path(struct buffer_head *root_bh,
 		get_bh(root_bh);
 		path_root_bh(path) = root_bh;
 		path_root_el(path) = root_el;
+		path_root_access(path) = access;
 	}
 
 	return path;
 }
 
+static struct ocfs2_path *ocfs2_new_path_from_path(struct ocfs2_path *path)
+{
+	return ocfs2_new_path(path_root_bh(path), path_root_el(path),
+			      path_root_access(path));
+}
+
+static struct ocfs2_path *ocfs2_new_path_from_et(struct ocfs2_extent_tree *et)
+{
+	return ocfs2_new_path(et->et_root_bh, et->et_root_el,
+			      et->et_root_journal_access);
+}
+
+/*
+ * Journal the buffer at depth idx.  All idx>0 are extent_blocks,
+ * otherwise it's the root_access function.
+ *
+ * I don't like the way this function's name looks next to
+ * ocfs2_journal_access_path(), but I don't have a better one.
+ */
+static int ocfs2_path_bh_journal_access(handle_t *handle,
+					struct inode *inode,
+					struct ocfs2_path *path,
+					int idx)
+{
+	ocfs2_journal_access_func access = path_root_access(path);
+
+	if (!access)
+		access = ocfs2_journal_access;
+
+	if (idx)
+		access = ocfs2_journal_access_eb;
+
+	return access(handle, inode, path->p_node[idx].bh,
+		      OCFS2_JOURNAL_ACCESS_WRITE);
+}
+
 /*
  * Convenience function to journal all components in a path.
  */
@@ -550,8 +597,7 @@ static int ocfs2_journal_access_path(struct inode *inode, handle_t *handle,
 		goto out;
 
 	for(i = 0; i < path_num_items(path); i++) {
-		ret = ocfs2_journal_access(handle, inode, path->p_node[i].bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode, path, i);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -686,6 +732,80 @@ struct ocfs2_merge_ctxt {
 	int			c_split_covers_rec;
 };
 
+static int ocfs2_validate_extent_block(struct super_block *sb,
+				       struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_extent_block *eb =
+		(struct ocfs2_extent_block *)bh->b_data;
+
+	mlog(0, "Validating extent block %llu\n",
+	     (unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 */
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &eb->h_check);
+	if (rc) {
+		mlog(ML_ERROR, "Checksum failed for extent block %llu\n",
+		     (unsigned long long)bh->b_blocknr);
+		return rc;
+	}
+
+	/*
+	 * Errors after here are fatal.
+	 */
+
+	if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
+		ocfs2_error(sb,
+			    "Extent block #%llu has bad signature %.*s",
+			    (unsigned long long)bh->b_blocknr, 7,
+			    eb->h_signature);
+		return -EINVAL;
+	}
+
+	if (le64_to_cpu(eb->h_blkno) != bh->b_blocknr) {
+		ocfs2_error(sb,
+			    "Extent block #%llu has an invalid h_blkno "
+			    "of %llu",
+			    (unsigned long long)bh->b_blocknr,
+			    (unsigned long long)le64_to_cpu(eb->h_blkno));
+		return -EINVAL;
+	}
+
+	if (le32_to_cpu(eb->h_fs_generation) != OCFS2_SB(sb)->fs_generation) {
+		ocfs2_error(sb,
+			    "Extent block #%llu has an invalid "
+			    "h_fs_generation of #%u",
+			    (unsigned long long)bh->b_blocknr,
+			    le32_to_cpu(eb->h_fs_generation));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int ocfs2_read_extent_block(struct inode *inode, u64 eb_blkno,
+			    struct buffer_head **bh)
+{
+	int rc;
+	struct buffer_head *tmp = *bh;
+
+	rc = ocfs2_read_block(inode, eb_blkno, &tmp,
+			      ocfs2_validate_extent_block);
+
+	/* If ocfs2_read_block() got us a new bh, pass it up. */
+	if (!rc && !*bh)
+		*bh = tmp;
+
+	return rc;
+}
+
+
 /*
  * How many free extents have we got before we need more meta data?
  */
@@ -705,8 +825,7 @@ int ocfs2_num_free_extents(struct ocfs2_super *osb,
 	last_eb_blk = ocfs2_et_get_last_eb_blk(et);
 
 	if (last_eb_blk) {
-		retval = ocfs2_read_block(inode, last_eb_blk,
-					  &eb_bh);
+		retval = ocfs2_read_extent_block(inode, last_eb_blk, &eb_bh);
 		if (retval < 0) {
 			mlog_errno(retval);
 			goto bail;
@@ -768,8 +887,8 @@ static int ocfs2_create_new_meta_bhs(struct ocfs2_super *osb,
 			}
 			ocfs2_set_new_buffer_uptodate(inode, bhs[i]);
 
-			status = ocfs2_journal_access(handle, inode, bhs[i],
-						      OCFS2_JOURNAL_ACCESS_CREATE);
+			status = ocfs2_journal_access_eb(handle, inode, bhs[i],
+							 OCFS2_JOURNAL_ACCESS_CREATE);
 			if (status < 0) {
 				mlog_errno(status);
 				goto bail;
@@ -908,15 +1027,12 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 	for(i = 0; i < new_blocks; i++) {
 		bh = new_eb_bhs[i];
 		eb = (struct ocfs2_extent_block *) bh->b_data;
-		if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-			OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-			status = -EIO;
-			goto bail;
-		}
+		/* ocfs2_create_new_meta_bhs() should create it right! */
+		BUG_ON(!OCFS2_IS_VALID_EXTENT_BLOCK(eb));
 		eb_el = &eb->h_list;
 
-		status = ocfs2_journal_access(handle, inode, bh,
-					      OCFS2_JOURNAL_ACCESS_CREATE);
+		status = ocfs2_journal_access_eb(handle, inode, bh,
+						 OCFS2_JOURNAL_ACCESS_CREATE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -955,21 +1071,21 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 	 * journal_dirty erroring as it won't unless we've aborted the
 	 * handle (in which case we would never be here) so reserving
 	 * the write with journal_access is all we need to do. */
-	status = ocfs2_journal_access(handle, inode, *last_eb_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_eb(handle, inode, *last_eb_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
-	status = ocfs2_journal_access(handle, inode, et->et_root_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_et_root_journal_access(handle, inode, et,
+					      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 	if (eb_bh) {
-		status = ocfs2_journal_access(handle, inode, eb_bh,
-					      OCFS2_JOURNAL_ACCESS_WRITE);
+		status = ocfs2_journal_access_eb(handle, inode, eb_bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -1052,17 +1168,14 @@ static int ocfs2_shift_tree_depth(struct ocfs2_super *osb,
 	}
 
 	eb = (struct ocfs2_extent_block *) new_eb_bh->b_data;
-	if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-		OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-		status = -EIO;
-		goto bail;
-	}
+	/* ocfs2_create_new_meta_bhs() should create it right! */
+	BUG_ON(!OCFS2_IS_VALID_EXTENT_BLOCK(eb));
 
 	eb_el = &eb->h_list;
 	root_el = et->et_root_el;
 
-	status = ocfs2_journal_access(handle, inode, new_eb_bh,
-				      OCFS2_JOURNAL_ACCESS_CREATE);
+	status = ocfs2_journal_access_eb(handle, inode, new_eb_bh,
+					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1080,8 +1193,8 @@ static int ocfs2_shift_tree_depth(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_journal_access(handle, inode, et->et_root_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_et_root_journal_access(handle, inode, et,
+					      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1176,18 +1289,13 @@ static int ocfs2_find_branch_target(struct ocfs2_super *osb,
 		brelse(bh);
 		bh = NULL;
 
-		status = ocfs2_read_block(inode, blkno, &bh);
+		status = ocfs2_read_extent_block(inode, blkno, &bh);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
 		}
 
 		eb = (struct ocfs2_extent_block *) bh->b_data;
-		if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-			OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-			status = -EIO;
-			goto bail;
-		}
 		el = &eb->h_list;
 
 		if (le16_to_cpu(el->l_next_free_rec) <
@@ -1540,7 +1648,7 @@ static int __ocfs2_find_path(struct inode *inode,
 
 		brelse(bh);
 		bh = NULL;
-		ret = ocfs2_read_block(inode, blkno, &bh);
+		ret = ocfs2_read_extent_block(inode, blkno, &bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -1548,11 +1656,6 @@ static int __ocfs2_find_path(struct inode *inode,
 
 		eb = (struct ocfs2_extent_block *) bh->b_data;
 		el = &eb->h_list;
-		if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-			OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-			ret = -EIO;
-			goto out;
-		}
 
 		if (le16_to_cpu(el->l_next_free_rec) >
 		    le16_to_cpu(el->l_count)) {
@@ -1860,25 +1963,23 @@ static int ocfs2_rotate_subtree_right(struct inode *inode,
 	root_bh = left_path->p_node[subtree_index].bh;
 	BUG_ON(root_bh != right_path->p_node[subtree_index].bh);
 
-	ret = ocfs2_journal_access(handle, inode, root_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+					   subtree_index);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
 	for(i = subtree_index + 1; i < path_num_items(right_path); i++) {
-		ret = ocfs2_journal_access(handle, inode,
-					   right_path->p_node[i].bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode,
+						   right_path, i);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_journal_access(handle, inode,
-					   left_path->p_node[i].bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode,
+						   left_path, i);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2102,8 +2203,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 
 	*ret_left_path = NULL;
 
-	left_path = ocfs2_new_path(path_root_bh(right_path),
-				   path_root_el(right_path));
+	left_path = ocfs2_new_path_from_path(right_path);
 	if (!left_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -2398,9 +2498,9 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 			return -EAGAIN;
 
 		if (le16_to_cpu(right_leaf_el->l_next_free_rec) > 1) {
-			ret = ocfs2_journal_access(handle, inode,
-						   path_leaf_bh(right_path),
-						   OCFS2_JOURNAL_ACCESS_WRITE);
+			ret = ocfs2_journal_access_eb(handle, inode,
+						      path_leaf_bh(right_path),
+						      OCFS2_JOURNAL_ACCESS_WRITE);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -2417,8 +2517,8 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 		 * We have to update i_last_eb_blk during the meta
 		 * data delete.
 		 */
-		ret = ocfs2_journal_access(handle, inode, et_root_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_et_root_journal_access(handle, inode, et,
+						   OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2433,25 +2533,23 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 	 */
 	BUG_ON(right_has_empty && !del_right_subtree);
 
-	ret = ocfs2_journal_access(handle, inode, root_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+					   subtree_index);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
 	for(i = subtree_index + 1; i < path_num_items(right_path); i++) {
-		ret = ocfs2_journal_access(handle, inode,
-					   right_path->p_node[i].bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode,
+						   right_path, i);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_journal_access(handle, inode,
-					   left_path->p_node[i].bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode,
+						   left_path, i);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2596,16 +2694,17 @@ out:
 
 static int ocfs2_rotate_rightmost_leaf_left(struct inode *inode,
 					    handle_t *handle,
-					    struct buffer_head *bh,
-					    struct ocfs2_extent_list *el)
+					    struct ocfs2_path *path)
 {
 	int ret;
+	struct buffer_head *bh = path_leaf_bh(path);
+	struct ocfs2_extent_list *el = path_leaf_el(path);
 
 	if (!ocfs2_is_empty_extent(&el->l_recs[0]))
 		return 0;
 
-	ret = ocfs2_journal_access(handle, inode, bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_path_bh_journal_access(handle, inode, path,
+					   path_num_items(path) - 1);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -2644,8 +2743,7 @@ static int __ocfs2_rotate_tree_left(struct inode *inode,
 		goto out;
 	}
 
-	left_path = ocfs2_new_path(path_root_bh(path),
-				   path_root_el(path));
+	left_path = ocfs2_new_path_from_path(path);
 	if (!left_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -2654,8 +2752,7 @@ static int __ocfs2_rotate_tree_left(struct inode *inode,
 
 	ocfs2_cp_path(left_path, path);
 
-	right_path = ocfs2_new_path(path_root_bh(path),
-				    path_root_el(path));
+	right_path = ocfs2_new_path_from_path(path);
 	if (!right_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -2689,9 +2786,8 @@ static int __ocfs2_rotate_tree_left(struct inode *inode,
 		 * Caller might still want to make changes to the
 		 * tree root, so re-add it to the journal here.
 		 */
-		ret = ocfs2_journal_access(handle, inode,
-					   path_root_bh(left_path),
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode,
+						   left_path, 0);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2785,8 +2881,7 @@ static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
 		 * We have a path to the left of this one - it needs
 		 * an update too.
 		 */
-		left_path = ocfs2_new_path(path_root_bh(path),
-					   path_root_el(path));
+		left_path = ocfs2_new_path_from_path(path);
 		if (!left_path) {
 			ret = -ENOMEM;
 			mlog_errno(ret);
@@ -2875,8 +2970,7 @@ rightmost_no_delete:
 		 * it up front.
 		 */
 		ret = ocfs2_rotate_rightmost_leaf_left(inode, handle,
-						       path_leaf_bh(path),
-						       path_leaf_el(path));
+						       path);
 		if (ret)
 			mlog_errno(ret);
 		goto out;
@@ -3027,8 +3121,7 @@ static int ocfs2_get_right_path(struct inode *inode,
 	/* This function shouldn't be called for the rightmost leaf. */
 	BUG_ON(right_cpos == 0);
 
-	right_path = ocfs2_new_path(path_root_bh(left_path),
-				    path_root_el(left_path));
+	right_path = ocfs2_new_path_from_path(left_path);
 	if (!right_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -3111,8 +3204,8 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 		root_bh = left_path->p_node[subtree_index].bh;
 		BUG_ON(root_bh != right_path->p_node[subtree_index].bh);
 
-		ret = ocfs2_journal_access(handle, inode, root_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+						   subtree_index);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3120,17 +3213,15 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 
 		for (i = subtree_index + 1;
 		     i < path_num_items(right_path); i++) {
-			ret = ocfs2_journal_access(handle, inode,
-						   right_path->p_node[i].bh,
-						   OCFS2_JOURNAL_ACCESS_WRITE);
+			ret = ocfs2_path_bh_journal_access(handle, inode,
+							   right_path, i);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
 			}
 
-			ret = ocfs2_journal_access(handle, inode,
-						   left_path->p_node[i].bh,
-						   OCFS2_JOURNAL_ACCESS_WRITE);
+			ret = ocfs2_path_bh_journal_access(handle, inode,
+							   left_path, i);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -3142,8 +3233,8 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 		right_rec = &el->l_recs[index + 1];
 	}
 
-	ret = ocfs2_journal_access(handle, inode, bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_path_bh_journal_access(handle, inode, left_path,
+					   path_num_items(left_path) - 1);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3199,8 +3290,7 @@ static int ocfs2_get_left_path(struct inode *inode,
 	/* This function shouldn't be called for the leftmost leaf. */
 	BUG_ON(left_cpos == 0);
 
-	left_path = ocfs2_new_path(path_root_bh(right_path),
-				   path_root_el(right_path));
+	left_path = ocfs2_new_path_from_path(right_path);
 	if (!left_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -3283,8 +3373,8 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 		root_bh = left_path->p_node[subtree_index].bh;
 		BUG_ON(root_bh != right_path->p_node[subtree_index].bh);
 
-		ret = ocfs2_journal_access(handle, inode, root_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+						   subtree_index);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3292,17 +3382,15 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 
 		for (i = subtree_index + 1;
 		     i < path_num_items(right_path); i++) {
-			ret = ocfs2_journal_access(handle, inode,
-						   right_path->p_node[i].bh,
-						   OCFS2_JOURNAL_ACCESS_WRITE);
+			ret = ocfs2_path_bh_journal_access(handle, inode,
+							   right_path, i);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
 			}
 
-			ret = ocfs2_journal_access(handle, inode,
-						   left_path->p_node[i].bh,
-						   OCFS2_JOURNAL_ACCESS_WRITE);
+			ret = ocfs2_path_bh_journal_access(handle, inode,
+							   left_path, i);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -3314,8 +3402,8 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 			has_empty_extent = 1;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+					   path_num_items(right_path) - 1);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3732,8 +3820,7 @@ static int ocfs2_append_rec_to_path(struct inode *inode, handle_t *handle,
 		 * leftmost leaf.
 		 */
 		if (left_cpos) {
-			left_path = ocfs2_new_path(path_root_bh(right_path),
-						   path_root_el(right_path));
+			left_path = ocfs2_new_path_from_path(right_path);
 			if (!left_path) {
 				ret = -ENOMEM;
 				mlog_errno(ret);
@@ -3781,7 +3868,7 @@ static void ocfs2_split_record(struct inode *inode,
 	struct ocfs2_extent_list *left_el = NULL, *right_el, *insert_el, *el;
 	struct ocfs2_extent_rec *rec, *tmprec;
 
-	right_el = path_leaf_el(right_path);;
+	right_el = path_leaf_el(right_path);
 	if (left_path)
 		left_el = path_leaf_el(left_path);
 
@@ -3958,8 +4045,8 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 
 	el = et->et_root_el;
 
-	ret = ocfs2_journal_access(handle, inode, et->et_root_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_et_root_journal_access(handle, inode, et,
+					   OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3970,7 +4057,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 		goto out_update_clusters;
 	}
 
-	right_path = ocfs2_new_path(et->et_root_bh, et->et_root_el);
+	right_path = ocfs2_new_path_from_et(et);
 	if (!right_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -4020,8 +4107,8 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 		 * ocfs2_rotate_tree_right() might have extended the
 		 * transaction without re-journaling our tree root.
 		 */
-		ret = ocfs2_journal_access(handle, inode, et->et_root_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = ocfs2_et_root_journal_access(handle, inode, et,
+						   OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -4082,8 +4169,7 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 			goto out;
 
 		if (left_cpos != 0) {
-			left_path = ocfs2_new_path(path_root_bh(path),
-						   path_root_el(path));
+			left_path = ocfs2_new_path_from_path(path);
 			if (!left_path)
 				goto out;
 
@@ -4097,8 +4183,15 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 			    le16_to_cpu(new_el->l_count)) {
 				bh = path_leaf_bh(left_path);
 				eb = (struct ocfs2_extent_block *)bh->b_data;
-				OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb,
-								 eb);
+				ocfs2_error(inode->i_sb,
+					    "Extent block #%llu has an "
+					    "invalid l_next_free_rec of "
+					    "%d.  It should have "
+					    "matched the l_count of %d",
+					    (unsigned long long)le64_to_cpu(eb->h_blkno),
+					    le16_to_cpu(new_el->l_next_free_rec),
+					    le16_to_cpu(new_el->l_count));
+				status = -EINVAL;
 				goto out;
 			}
 			rec = &new_el->l_recs[
@@ -4132,8 +4225,7 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 		if (right_cpos == 0)
 			goto out;
 
-		right_path = ocfs2_new_path(path_root_bh(path),
-					    path_root_el(path));
+		right_path = ocfs2_new_path_from_path(path);
 		if (!right_path)
 			goto out;
 
@@ -4147,8 +4239,12 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 			if (le16_to_cpu(new_el->l_next_free_rec) <= 1) {
 				bh = path_leaf_bh(right_path);
 				eb = (struct ocfs2_extent_block *)bh->b_data;
-				OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb,
-								 eb);
+				ocfs2_error(inode->i_sb,
+					    "Extent block #%llu has an "
+					    "invalid l_next_free_rec of %d",
+					    (unsigned long long)le64_to_cpu(eb->h_blkno),
+					    le16_to_cpu(new_el->l_next_free_rec));
+				status = -EINVAL;
 				goto out;
 			}
 			rec = &new_el->l_recs[1];
@@ -4294,7 +4390,9 @@ static int ocfs2_figure_insert_type(struct inode *inode,
 		 * ocfs2_figure_insert_type() and ocfs2_add_branch()
 		 * may want it later.
 		 */
-		ret = ocfs2_read_block(inode, ocfs2_et_get_last_eb_blk(et), &bh);
+		ret = ocfs2_read_extent_block(inode,
+					      ocfs2_et_get_last_eb_blk(et),
+					      &bh);
 		if (ret) {
 			mlog_exit(ret);
 			goto out;
@@ -4320,7 +4418,7 @@ static int ocfs2_figure_insert_type(struct inode *inode,
 		return 0;
 	}
 
-	path = ocfs2_new_path(et->et_root_bh, et->et_root_el);
+	path = ocfs2_new_path_from_et(et);
 	if (!path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -4531,9 +4629,9 @@ int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
 
 	BUG_ON(num_bits > clusters_to_add);
 
-	/* reserve our write early -- insert_extent may update the inode */
-	status = ocfs2_journal_access(handle, inode, et->et_root_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	/* reserve our write early -- insert_extent may update the tree root */
+	status = ocfs2_et_root_journal_access(handle, inode, et,
+					      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -4760,20 +4858,15 @@ static int __ocfs2_mark_extent_written(struct inode *inode,
 	if (path->p_tree_depth) {
 		struct ocfs2_extent_block *eb;
 
-		ret = ocfs2_read_block(inode, ocfs2_et_get_last_eb_blk(et),
-				       &last_eb_bh);
+		ret = ocfs2_read_extent_block(inode,
+					      ocfs2_et_get_last_eb_blk(et),
+					      &last_eb_bh);
 		if (ret) {
 			mlog_exit(ret);
 			goto out;
 		}
 
 		eb = (struct ocfs2_extent_block *) last_eb_bh->b_data;
-		if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-			OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-			ret = -EROFS;
-			goto out;
-		}
-
 		rightmost_el = &eb->h_list;
 	} else
 		rightmost_el = path_root_el(path);
@@ -4854,7 +4947,7 @@ int ocfs2_mark_extent_written(struct inode *inode,
 	if (et->et_ops == &ocfs2_dinode_et_ops)
 		ocfs2_extent_map_trunc(inode, 0);
 
-	left_path = ocfs2_new_path(et->et_root_bh, et->et_root_el);
+	left_path = ocfs2_new_path_from_et(et);
 	if (!left_path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -4918,8 +5011,9 @@ static int ocfs2_split_tree(struct inode *inode, struct ocfs2_extent_tree *et,
 
 	depth = path->p_tree_depth;
 	if (depth > 0) {
-		ret = ocfs2_read_block(inode, ocfs2_et_get_last_eb_blk(et),
-				       &last_eb_bh);
+		ret = ocfs2_read_extent_block(inode,
+					      ocfs2_et_get_last_eb_blk(et),
+					      &last_eb_bh);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -5025,8 +5119,7 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 		}
 
 		if (left_cpos && le16_to_cpu(el->l_next_free_rec) > 1) {
-			left_path = ocfs2_new_path(path_root_bh(path),
-						   path_root_el(path));
+			left_path = ocfs2_new_path_from_path(path);
 			if (!left_path) {
 				ret = -ENOMEM;
 				mlog_errno(ret);
@@ -5135,7 +5228,7 @@ int ocfs2_remove_extent(struct inode *inode,
 
 	ocfs2_extent_map_trunc(inode, 0);
 
-	path = ocfs2_new_path(et->et_root_bh, et->et_root_el);
+	path = ocfs2_new_path_from_et(et);
 	if (!path) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
@@ -5255,6 +5348,78 @@ out:
 	return ret;
 }
 
+int ocfs2_remove_btree_range(struct inode *inode,
+			     struct ocfs2_extent_tree *et,
+			     u32 cpos, u32 phys_cpos, u32 len,
+			     struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret;
+	u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct inode *tl_inode = osb->osb_tl_inode;
+	handle_t *handle;
+	struct ocfs2_alloc_context *meta_ac = NULL;
+
+	ret = ocfs2_lock_allocators(inode, et, 0, 1, NULL, &meta_ac);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	mutex_lock(&tl_inode->i_mutex);
+
+	if (ocfs2_truncate_log_needs_flush(osb)) {
+		ret = __ocfs2_flush_truncate_log(osb);
+		if (ret < 0) {
+			mlog_errno(ret);
+			goto out;
+		}
+	}
+
+	handle = ocfs2_start_trans(osb, ocfs2_remove_extent_credits(osb->sb));
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_et_root_journal_access(handle, inode, et,
+					   OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_remove_extent(inode, et, cpos, len, handle, meta_ac,
+				  dealloc);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ocfs2_et_update_clusters(inode, et, -len);
+
+	ret = ocfs2_journal_dirty(handle, et->et_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ret = ocfs2_truncate_log_append(osb, handle, phys_blkno, len);
+	if (ret)
+		mlog_errno(ret);
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+out:
+	mutex_unlock(&tl_inode->i_mutex);
+
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+
+	return ret;
+}
+
 int ocfs2_truncate_log_needs_flush(struct ocfs2_super *osb)
 {
 	struct buffer_head *tl_bh = osb->osb_tl_bh;
@@ -5308,13 +5473,13 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
 	start_cluster = ocfs2_blocks_to_clusters(osb->sb, start_blk);
 
 	di = (struct ocfs2_dinode *) tl_bh->b_data;
-	tl = &di->id2.i_dealloc;
-	if (!OCFS2_IS_VALID_DINODE(di)) {
-		OCFS2_RO_ON_INVALID_DINODE(osb->sb, di);
-		status = -EIO;
-		goto bail;
-	}
 
+	/* tl_bh is loaded from ocfs2_truncate_log_init().  It's validated
+	 * by the underlying call to ocfs2_read_inode_block(), so any
+	 * corruption is a code bug */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(di));
+
+	tl = &di->id2.i_dealloc;
 	tl_count = le16_to_cpu(tl->tl_count);
 	mlog_bug_on_msg(tl_count > ocfs2_truncate_recs_per_inode(osb->sb) ||
 			tl_count == 0,
@@ -5332,8 +5497,8 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_journal_access(handle, tl_inode, tl_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, tl_inode, tl_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -5394,8 +5559,8 @@ static int ocfs2_replay_truncate_records(struct ocfs2_super *osb,
 	while (i >= 0) {
 		/* Caller has given us at least enough credits to
 		 * update the truncate log dinode */
-		status = ocfs2_journal_access(handle, tl_inode, tl_bh,
-					      OCFS2_JOURNAL_ACCESS_WRITE);
+		status = ocfs2_journal_access_di(handle, tl_inode, tl_bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -5464,13 +5629,13 @@ int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
 	BUG_ON(mutex_trylock(&tl_inode->i_mutex));
 
 	di = (struct ocfs2_dinode *) tl_bh->b_data;
-	tl = &di->id2.i_dealloc;
-	if (!OCFS2_IS_VALID_DINODE(di)) {
-		OCFS2_RO_ON_INVALID_DINODE(osb->sb, di);
-		status = -EIO;
-		goto out;
-	}
 
+	/* tl_bh is loaded from ocfs2_truncate_log_init().  It's validated
+	 * by the underlying call to ocfs2_read_inode_block(), so any
+	 * corruption is a code bug */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(di));
+
+	tl = &di->id2.i_dealloc;
 	num_to_flush = le16_to_cpu(tl->tl_used);
 	mlog(0, "Flush %u records from truncate log #%llu\n",
 	     num_to_flush, (unsigned long long)OCFS2_I(tl_inode)->ip_blkno);
@@ -5586,7 +5751,7 @@ static int ocfs2_get_truncate_log_info(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, &bh);
+	status = ocfs2_read_inode_block(inode, &bh);
 	if (status < 0) {
 		iput(inode);
 		mlog_errno(status);
@@ -5625,13 +5790,13 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
 	}
 
 	di = (struct ocfs2_dinode *) tl_bh->b_data;
-	tl = &di->id2.i_dealloc;
-	if (!OCFS2_IS_VALID_DINODE(di)) {
-		OCFS2_RO_ON_INVALID_DINODE(tl_inode->i_sb, di);
-		status = -EIO;
-		goto bail;
-	}
 
+	/* tl_bh is loaded from ocfs2_get_truncate_log_info().  It's
+	 * validated by the underlying call to ocfs2_read_inode_block(),
+	 * so any corruption is a code bug */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(di));
+
+	tl = &di->id2.i_dealloc;
 	if (le16_to_cpu(tl->tl_used)) {
 		mlog(0, "We'll have %u logs to recover\n",
 		     le16_to_cpu(tl->tl_used));
@@ -5651,6 +5816,7 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
 		 * tl_used. */
 		tl->tl_used = 0;
 
+		ocfs2_compute_meta_ecc(osb->sb, tl_bh->b_data, &di->i_check);
 		status = ocfs2_write_block(osb, tl_bh, tl_inode);
 		if (status < 0) {
 			mlog_errno(status);
@@ -5800,7 +5966,10 @@ int ocfs2_truncate_log_init(struct ocfs2_super *osb)
  */
 
 /*
- * Describes a single block free from a suballocator
+ * Describe a single bit freed from a suballocator.  For the block
+ * suballocators, it represents one block.  For the global cluster
+ * allocator, it represents some clusters and free_bit indicates
+ * clusters number.
  */
 struct ocfs2_cached_block_free {
 	struct ocfs2_cached_block_free		*free_next;
@@ -5815,10 +5984,10 @@ struct ocfs2_per_slot_free_list {
 	struct ocfs2_cached_block_free		*f_first;
 };
 
-static int ocfs2_free_cached_items(struct ocfs2_super *osb,
-				   int sysfile_type,
-				   int slot,
-				   struct ocfs2_cached_block_free *head)
+static int ocfs2_free_cached_blocks(struct ocfs2_super *osb,
+				    int sysfile_type,
+				    int slot,
+				    struct ocfs2_cached_block_free *head)
 {
 	int ret;
 	u64 bg_blkno;
@@ -5893,6 +6062,82 @@ out:
 	return ret;
 }
 
+int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
+				u64 blkno, unsigned int bit)
+{
+	int ret = 0;
+	struct ocfs2_cached_block_free *item;
+
+	item = kmalloc(sizeof(*item), GFP_NOFS);
+	if (item == NULL) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		return ret;
+	}
+
+	mlog(0, "Insert clusters: (bit %u, blk %llu)\n",
+	     bit, (unsigned long long)blkno);
+
+	item->free_blk = blkno;
+	item->free_bit = bit;
+	item->free_next = ctxt->c_global_allocator;
+
+	ctxt->c_global_allocator = item;
+	return ret;
+}
+
+static int ocfs2_free_cached_clusters(struct ocfs2_super *osb,
+				      struct ocfs2_cached_block_free *head)
+{
+	struct ocfs2_cached_block_free *tmp;
+	struct inode *tl_inode = osb->osb_tl_inode;
+	handle_t *handle;
+	int ret = 0;
+
+	mutex_lock(&tl_inode->i_mutex);
+
+	while (head) {
+		if (ocfs2_truncate_log_needs_flush(osb)) {
+			ret = __ocfs2_flush_truncate_log(osb);
+			if (ret < 0) {
+				mlog_errno(ret);
+				break;
+			}
+		}
+
+		handle = ocfs2_start_trans(osb, OCFS2_TRUNCATE_LOG_UPDATE);
+		if (IS_ERR(handle)) {
+			ret = PTR_ERR(handle);
+			mlog_errno(ret);
+			break;
+		}
+
+		ret = ocfs2_truncate_log_append(osb, handle, head->free_blk,
+						head->free_bit);
+
+		ocfs2_commit_trans(osb, handle);
+		tmp = head;
+		head = head->free_next;
+		kfree(tmp);
+
+		if (ret < 0) {
+			mlog_errno(ret);
+			break;
+		}
+	}
+
+	mutex_unlock(&tl_inode->i_mutex);
+
+	while (head) {
+		/* Premature exit may have left some dangling items. */
+		tmp = head;
+		head = head->free_next;
+		kfree(tmp);
+	}
+
+	return ret;
+}
+
 int ocfs2_run_deallocs(struct ocfs2_super *osb,
 		       struct ocfs2_cached_dealloc_ctxt *ctxt)
 {
@@ -5908,8 +6153,10 @@ int ocfs2_run_deallocs(struct ocfs2_super *osb,
 		if (fl->f_first) {
 			mlog(0, "Free items: (type %u, slot %d)\n",
 			     fl->f_inode_type, fl->f_slot);
-			ret2 = ocfs2_free_cached_items(osb, fl->f_inode_type,
-						       fl->f_slot, fl->f_first);
+			ret2 = ocfs2_free_cached_blocks(osb,
+							fl->f_inode_type,
+							fl->f_slot,
+							fl->f_first);
 			if (ret2)
 				mlog_errno(ret2);
 			if (!ret)
@@ -5920,6 +6167,17 @@ int ocfs2_run_deallocs(struct ocfs2_super *osb,
 		kfree(fl);
 	}
 
+	if (ctxt->c_global_allocator) {
+		ret2 = ocfs2_free_cached_clusters(osb,
+						  ctxt->c_global_allocator);
+		if (ret2)
+			mlog_errno(ret2);
+		if (!ret)
+			ret = ret2;
+
+		ctxt->c_global_allocator = NULL;
+	}
+
 	return ret;
 }
 
@@ -6075,11 +6333,10 @@ static int ocfs2_find_new_last_ext_blk(struct inode *inode,
 
 	eb = (struct ocfs2_extent_block *) bh->b_data;
 	el = &eb->h_list;
-	if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-		OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-		ret = -EROFS;
-		goto out;
-	}
+
+	/* ocfs2_find_leaf() gets the eb from ocfs2_read_extent_block().
+	 * Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_EXTENT_BLOCK(eb));
 
 	*new_last_eb = bh;
 	get_bh(*new_last_eb);
@@ -6326,8 +6583,8 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 	}
 
 	if (last_eb_bh) {
-		status = ocfs2_journal_access(handle, inode, last_eb_bh,
-					      OCFS2_JOURNAL_ACCESS_WRITE);
+		status = ocfs2_journal_access_eb(handle, inode, last_eb_bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -6350,6 +6607,8 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 		goto bail;
 	}
 
+	vfs_dq_free_space_nodirty(inode,
+			ocfs2_clusters_to_bytes(osb->sb, clusters_to_del));
 	spin_lock(&OCFS2_I(inode)->ip_lock);
 	OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters) -
 				      clusters_to_del;
@@ -6436,11 +6695,6 @@ static void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
 		mlog_errno(ret);
 	else if (ocfs2_should_order_data(inode)) {
 		ret = ocfs2_jbd2_file_inode(handle, inode);
-#ifdef CONFIG_OCFS2_COMPAT_JBD
-		ret = walk_page_buffers(handle, page_buffers(page),
-					from, to, &partial,
-					ocfs2_journal_dirty_data);
-#endif
 		if (ret < 0)
 			mlog_errno(ret);
 	}
@@ -6663,6 +6917,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 	struct page **pages = NULL;
 	loff_t end = osb->s_clustersize;
 	struct ocfs2_extent_tree et;
+	int did_quota = 0;
 
 	has_data = i_size_read(inode) ? 1 : 0;
 
@@ -6682,15 +6937,16 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_INLINE_TO_EXTENTS_CREDITS);
+	handle = ocfs2_start_trans(osb,
+				   ocfs2_inline_to_extents_credits(osb->sb));
 	if (IS_ERR(handle)) {
 		ret = PTR_ERR(handle);
 		mlog_errno(ret);
 		goto out_unlock;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -6701,6 +6957,13 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 		unsigned int page_end;
 		u64 phys;
 
+		if (vfs_dq_alloc_space_nodirty(inode,
+				       ocfs2_clusters_to_bytes(osb->sb, 1))) {
+			ret = -EDQUOT;
+			goto out_commit;
+		}
+		did_quota = 1;
+
 		ret = ocfs2_claim_clusters(osb, handle, data_ac, 1, &bit_off,
 					   &num);
 		if (ret) {
@@ -6774,6 +7037,10 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 	}
 
 out_commit:
+	if (ret < 0 && did_quota)
+		vfs_dq_free_space_nodirty(inode,
+					  ocfs2_clusters_to_bytes(osb->sb, 1));
+
 	ocfs2_commit_trans(osb, handle);
 
 out_unlock:
@@ -6813,7 +7080,8 @@ int ocfs2_commit_truncate(struct ocfs2_super *osb,
 	new_highest_cpos = ocfs2_clusters_for_bytes(osb->sb,
 						     i_size_read(inode));
 
-	path = ocfs2_new_path(fe_bh, &di->id2.i_list);
+	path = ocfs2_new_path(fe_bh, &di->id2.i_list,
+			      ocfs2_journal_access_di);
 	if (!path) {
 		status = -ENOMEM;
 		mlog_errno(status);
@@ -6984,20 +7252,14 @@ int ocfs2_prepare_truncate(struct ocfs2_super *osb,
 	ocfs2_init_dealloc_ctxt(&(*tc)->tc_dealloc);
 
 	if (fe->id2.i_list.l_tree_depth) {
-		status = ocfs2_read_block(inode, le64_to_cpu(fe->i_last_eb_blk),
-					  &last_eb_bh);
+		status = ocfs2_read_extent_block(inode,
+						 le64_to_cpu(fe->i_last_eb_blk),
+						 &last_eb_bh);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
 		}
 		eb = (struct ocfs2_extent_block *) last_eb_bh->b_data;
-		if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-			OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-
-			brelse(last_eb_bh);
-			status = -EIO;
-			goto bail;
-		}
 	}
 
 	(*tc)->tc_last_eb_bh = last_eb_bh;
@@ -7052,8 +7314,8 @@ int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h
index 70257c84cfb..cceff5c37f4 100644
--- a/fs/ocfs2/alloc.h
+++ b/fs/ocfs2/alloc.h
@@ -45,7 +45,9 @@
  *
  * ocfs2_extent_tree contains info for the root of the b-tree, it must have a
  * root ocfs2_extent_list and a root_bh so that they can be used in the b-tree
- * functions.
+ * functions.  With metadata ecc, we now call different journal_access
+ * functions for each type of metadata, so it must have the
+ * root_journal_access function.
  * ocfs2_extent_tree_operations abstract the normal operations we do for
  * the root of extent b-tree.
  */
@@ -54,6 +56,7 @@ struct ocfs2_extent_tree {
 	struct ocfs2_extent_tree_operations	*et_ops;
 	struct buffer_head			*et_root_bh;
 	struct ocfs2_extent_list		*et_root_el;
+	ocfs2_journal_access_func		et_root_journal_access;
 	void					*et_object;
 	unsigned int				et_max_leaf_clusters;
 };
@@ -68,10 +71,18 @@ void ocfs2_init_dinode_extent_tree(struct ocfs2_extent_tree *et,
 void ocfs2_init_xattr_tree_extent_tree(struct ocfs2_extent_tree *et,
 				       struct inode *inode,
 				       struct buffer_head *bh);
+struct ocfs2_xattr_value_buf;
 void ocfs2_init_xattr_value_extent_tree(struct ocfs2_extent_tree *et,
 					struct inode *inode,
-					struct buffer_head *bh,
-					struct ocfs2_xattr_value_root *xv);
+					struct ocfs2_xattr_value_buf *vb);
+
+/*
+ * Read an extent block into *bh.  If *bh is NULL, a bh will be
+ * allocated.  This is a cached read.  The extent block will be validated
+ * with ocfs2_validate_extent_block().
+ */
+int ocfs2_read_extent_block(struct inode *inode, u64 eb_blkno,
+			    struct buffer_head **bh);
 
 struct ocfs2_alloc_context;
 int ocfs2_insert_extent(struct ocfs2_super *osb,
@@ -110,6 +121,11 @@ int ocfs2_remove_extent(struct inode *inode,
 			u32 cpos, u32 len, handle_t *handle,
 			struct ocfs2_alloc_context *meta_ac,
 			struct ocfs2_cached_dealloc_ctxt *dealloc);
+int ocfs2_remove_btree_range(struct inode *inode,
+			     struct ocfs2_extent_tree *et,
+			     u32 cpos, u32 phys_cpos, u32 len,
+			     struct ocfs2_cached_dealloc_ctxt *dealloc);
+
 int ocfs2_num_free_extents(struct ocfs2_super *osb,
 			   struct inode *inode,
 			   struct ocfs2_extent_tree *et);
@@ -167,10 +183,18 @@ int __ocfs2_flush_truncate_log(struct ocfs2_super *osb);
  */
 struct ocfs2_cached_dealloc_ctxt {
 	struct ocfs2_per_slot_free_list		*c_first_suballocator;
+	struct ocfs2_cached_block_free 		*c_global_allocator;
 };
 static inline void ocfs2_init_dealloc_ctxt(struct ocfs2_cached_dealloc_ctxt *c)
 {
 	c->c_first_suballocator = NULL;
+	c->c_global_allocator = NULL;
+}
+int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
+				u64 blkno, unsigned int bit);
+static inline int ocfs2_dealloc_has_cluster(struct ocfs2_cached_dealloc_ctxt *c)
+{
+	return c->c_global_allocator != NULL;
 }
 int ocfs2_run_deallocs(struct ocfs2_super *osb,
 		       struct ocfs2_cached_dealloc_ctxt *ctxt);
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index c22543b3342..a067a6cffb0 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -27,6 +27,7 @@
 #include <linux/swap.h>
 #include <linux/pipe_fs_i.h>
 #include <linux/mpage.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_FILE_IO
 #include <cluster/masklog.h>
@@ -68,20 +69,13 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
 		goto bail;
 	}
 
-	status = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, &bh);
+	status = ocfs2_read_inode_block(inode, &bh);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 	fe = (struct ocfs2_dinode *) bh->b_data;
 
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		mlog(ML_ERROR, "Invalid dinode #%llu: signature = %.*s\n",
-		     (unsigned long long)le64_to_cpu(fe->i_blkno), 7,
-		     fe->i_signature);
-		goto bail;
-	}
-
 	if ((u64)iblock >= ocfs2_clusters_to_blocks(inode->i_sb,
 						    le32_to_cpu(fe->i_clusters))) {
 		mlog(ML_ERROR, "block offset is outside the allocated size: "
@@ -262,7 +256,7 @@ static int ocfs2_readpage_inline(struct inode *inode, struct page *page)
 	BUG_ON(!PageLocked(page));
 	BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL));
 
-	ret = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, &di_bh);
+	ret = ocfs2_read_inode_block(inode, &di_bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -481,12 +475,6 @@ handle_t *ocfs2_start_walk_page_trans(struct inode *inode,
 
 	if (ocfs2_should_order_data(inode)) {
 		ret = ocfs2_jbd2_file_inode(handle, inode);
-#ifdef CONFIG_OCFS2_COMPAT_JBD
-		ret = walk_page_buffers(handle,
-					page_buffers(page),
-					from, to, NULL,
-					ocfs2_journal_dirty_data);
-#endif
 		if (ret < 0)
 			mlog_errno(ret);
 	}
@@ -1072,15 +1060,8 @@ static void ocfs2_write_failure(struct inode *inode,
 		tmppage = wc->w_pages[i];
 
 		if (page_has_buffers(tmppage)) {
-			if (ocfs2_should_order_data(inode)) {
+			if (ocfs2_should_order_data(inode))
 				ocfs2_jbd2_file_inode(wc->w_handle, inode);
-#ifdef CONFIG_OCFS2_COMPAT_JBD
-				walk_page_buffers(wc->w_handle,
-						  page_buffers(tmppage),
-						  from, to, NULL,
-						  ocfs2_journal_dirty_data);
-#endif
-			}
 
 			block_commit_write(tmppage, from, to);
 		}
@@ -1531,8 +1512,8 @@ static int ocfs2_write_begin_inline(struct address_space *mapping,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, wc->w_di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, wc->w_di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		ocfs2_commit_trans(osb, handle);
 
@@ -1750,15 +1731,20 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 
 	wc->w_handle = handle;
 
+	if (clusters_to_alloc && vfs_dq_alloc_space_nodirty(inode,
+			ocfs2_clusters_to_bytes(osb->sb, clusters_to_alloc))) {
+		ret = -EDQUOT;
+		goto out_commit;
+	}
 	/*
 	 * We don't want this to fail in ocfs2_write_end(), so do it
 	 * here.
 	 */
-	ret = ocfs2_journal_access(handle, inode, wc->w_di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, wc->w_di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out_quota;
 	}
 
 	/*
@@ -1771,14 +1757,14 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 					 mmap_page);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out_quota;
 	}
 
 	ret = ocfs2_write_cluster_by_desc(mapping, data_ac, meta_ac, wc, pos,
 					  len);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out_quota;
 	}
 
 	if (data_ac)
@@ -1790,6 +1776,10 @@ success:
 	*pagep = wc->w_target_page;
 	*fsdata = wc;
 	return 0;
+out_quota:
+	if (clusters_to_alloc)
+		vfs_dq_free_space(inode,
+			  ocfs2_clusters_to_bytes(osb->sb, clusters_to_alloc));
 out_commit:
 	ocfs2_commit_trans(osb, handle);
 
@@ -1919,15 +1909,8 @@ int ocfs2_write_end_nolock(struct address_space *mapping,
 		}
 
 		if (page_has_buffers(tmppage)) {
-			if (ocfs2_should_order_data(inode)) {
+			if (ocfs2_should_order_data(inode))
 				ocfs2_jbd2_file_inode(wc->w_handle, inode);
-#ifdef CONFIG_OCFS2_COMPAT_JBD
-				walk_page_buffers(wc->w_handle,
-						  page_buffers(tmppage),
-						  from, to, NULL,
-						  ocfs2_journal_dirty_data);
-#endif
-			}
 			block_commit_write(tmppage, from, to);
 		}
 	}
diff --git a/fs/ocfs2/blockcheck.c b/fs/ocfs2/blockcheck.c
new file mode 100644
index 00000000000..2a947c44e59
--- /dev/null
+++ b/fs/ocfs2/blockcheck.c
@@ -0,0 +1,477 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * blockcheck.c
+ *
+ * Checksum and ECC codes for the OCFS2 userspace library.
+ *
+ * Copyright (C) 2006, 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/crc32.h>
+#include <linux/buffer_head.h>
+#include <linux/bitops.h>
+#include <asm/byteorder.h>
+
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+
+#include "blockcheck.h"
+
+
+/*
+ * We use the following conventions:
+ *
+ * d = # data bits
+ * p = # parity bits
+ * c = # total code bits (d + p)
+ */
+
+
+/*
+ * Calculate the bit offset in the hamming code buffer based on the bit's
+ * offset in the data buffer.  Since the hamming code reserves all
+ * power-of-two bits for parity, the data bit number and the code bit
+ * number are offest by all the parity bits beforehand.
+ *
+ * Recall that bit numbers in hamming code are 1-based.  This function
+ * takes the 0-based data bit from the caller.
+ *
+ * An example.  Take bit 1 of the data buffer.  1 is a power of two (2^0),
+ * so it's a parity bit.  2 is a power of two (2^1), so it's a parity bit.
+ * 3 is not a power of two.  So bit 1 of the data buffer ends up as bit 3
+ * in the code buffer.
+ *
+ * The caller can pass in *p if it wants to keep track of the most recent
+ * number of parity bits added.  This allows the function to start the
+ * calculation at the last place.
+ */
+static unsigned int calc_code_bit(unsigned int i, unsigned int *p_cache)
+{
+	unsigned int b, p = 0;
+
+	/*
+	 * Data bits are 0-based, but we're talking code bits, which
+	 * are 1-based.
+	 */
+	b = i + 1;
+
+	/* Use the cache if it is there */
+	if (p_cache)
+		p = *p_cache;
+        b += p;
+
+	/*
+	 * For every power of two below our bit number, bump our bit.
+	 *
+	 * We compare with (b + 1) because we have to compare with what b
+	 * would be _if_ it were bumped up by the parity bit.  Capice?
+	 *
+	 * p is set above.
+	 */
+	for (; (1 << p) < (b + 1); p++)
+		b++;
+
+	if (p_cache)
+		*p_cache = p;
+
+	return b;
+}
+
+/*
+ * This is the low level encoder function.  It can be called across
+ * multiple hunks just like the crc32 code.  'd' is the number of bits
+ * _in_this_hunk_.  nr is the bit offset of this hunk.  So, if you had
+ * two 512B buffers, you would do it like so:
+ *
+ * parity = ocfs2_hamming_encode(0, buf1, 512 * 8, 0);
+ * parity = ocfs2_hamming_encode(parity, buf2, 512 * 8, 512 * 8);
+ *
+ * If you just have one buffer, use ocfs2_hamming_encode_block().
+ */
+u32 ocfs2_hamming_encode(u32 parity, void *data, unsigned int d, unsigned int nr)
+{
+	unsigned int i, b, p = 0;
+
+	BUG_ON(!d);
+
+	/*
+	 * b is the hamming code bit number.  Hamming code specifies a
+	 * 1-based array, but C uses 0-based.  So 'i' is for C, and 'b' is
+	 * for the algorithm.
+	 *
+	 * The i++ in the for loop is so that the start offset passed
+	 * to ocfs2_find_next_bit_set() is one greater than the previously
+	 * found bit.
+	 */
+	for (i = 0; (i = ocfs2_find_next_bit(data, d, i)) < d; i++)
+	{
+		/*
+		 * i is the offset in this hunk, nr + i is the total bit
+		 * offset.
+		 */
+		b = calc_code_bit(nr + i, &p);
+
+		/*
+		 * Data bits in the resultant code are checked by
+		 * parity bits that are part of the bit number
+		 * representation.  Huh?
+		 *
+		 * <wikipedia href="http://en.wikipedia.org/wiki/Hamming_code">
+		 * In other words, the parity bit at position 2^k
+		 * checks bits in positions having bit k set in
+		 * their binary representation.  Conversely, for
+		 * instance, bit 13, i.e. 1101(2), is checked by
+		 * bits 1000(2) = 8, 0100(2)=4 and 0001(2) = 1.
+		 * </wikipedia>
+		 *
+		 * Note that 'k' is the _code_ bit number.  'b' in
+		 * our loop.
+		 */
+		parity ^= b;
+	}
+
+	/* While the data buffer was treated as little endian, the
+	 * return value is in host endian. */
+	return parity;
+}
+
+u32 ocfs2_hamming_encode_block(void *data, unsigned int blocksize)
+{
+	return ocfs2_hamming_encode(0, data, blocksize * 8, 0);
+}
+
+/*
+ * Like ocfs2_hamming_encode(), this can handle hunks.  nr is the bit
+ * offset of the current hunk.  If bit to be fixed is not part of the
+ * current hunk, this does nothing.
+ *
+ * If you only have one hunk, use ocfs2_hamming_fix_block().
+ */
+void ocfs2_hamming_fix(void *data, unsigned int d, unsigned int nr,
+		       unsigned int fix)
+{
+	unsigned int i, b;
+
+	BUG_ON(!d);
+
+	/*
+	 * If the bit to fix has an hweight of 1, it's a parity bit.  One
+	 * busted parity bit is its own error.  Nothing to do here.
+	 */
+	if (hweight32(fix) == 1)
+		return;
+
+	/*
+	 * nr + d is the bit right past the data hunk we're looking at.
+	 * If fix after that, nothing to do
+	 */
+	if (fix >= calc_code_bit(nr + d, NULL))
+		return;
+
+	/*
+	 * nr is the offset in the data hunk we're starting at.  Let's
+	 * start b at the offset in the code buffer.  See hamming_encode()
+	 * for a more detailed description of 'b'.
+	 */
+	b = calc_code_bit(nr, NULL);
+	/* If the fix is before this hunk, nothing to do */
+	if (fix < b)
+		return;
+
+	for (i = 0; i < d; i++, b++)
+	{
+		/* Skip past parity bits */
+		while (hweight32(b) == 1)
+			b++;
+
+		/*
+		 * i is the offset in this data hunk.
+		 * nr + i is the offset in the total data buffer.
+		 * b is the offset in the total code buffer.
+		 *
+		 * Thus, when b == fix, bit i in the current hunk needs
+		 * fixing.
+		 */
+		if (b == fix)
+		{
+			if (ocfs2_test_bit(i, data))
+				ocfs2_clear_bit(i, data);
+			else
+				ocfs2_set_bit(i, data);
+			break;
+		}
+	}
+}
+
+void ocfs2_hamming_fix_block(void *data, unsigned int blocksize,
+			     unsigned int fix)
+{
+	ocfs2_hamming_fix(data, blocksize * 8, 0, fix);
+}
+
+/*
+ * This function generates check information for a block.
+ * data is the block to be checked.  bc is a pointer to the
+ * ocfs2_block_check structure describing the crc32 and the ecc.
+ *
+ * bc should be a pointer inside data, as the function will
+ * take care of zeroing it before calculating the check information.  If
+ * bc does not point inside data, the caller must make sure any inline
+ * ocfs2_block_check structures are zeroed.
+ *
+ * The data buffer must be in on-disk endian (little endian for ocfs2).
+ * bc will be filled with little-endian values and will be ready to go to
+ * disk.
+ */
+void ocfs2_block_check_compute(void *data, size_t blocksize,
+			       struct ocfs2_block_check *bc)
+{
+	u32 crc;
+	u32 ecc;
+
+	memset(bc, 0, sizeof(struct ocfs2_block_check));
+
+	crc = crc32_le(~0, data, blocksize);
+	ecc = ocfs2_hamming_encode_block(data, blocksize);
+
+	/*
+	 * No ecc'd ocfs2 structure is larger than 4K, so ecc will be no
+	 * larger than 16 bits.
+	 */
+	BUG_ON(ecc > USHORT_MAX);
+
+	bc->bc_crc32e = cpu_to_le32(crc);
+	bc->bc_ecc = cpu_to_le16((u16)ecc);
+}
+
+/*
+ * This function validates existing check information.  Like _compute,
+ * the function will take care of zeroing bc before calculating check codes.
+ * If bc is not a pointer inside data, the caller must have zeroed any
+ * inline ocfs2_block_check structures.
+ *
+ * Again, the data passed in should be the on-disk endian.
+ */
+int ocfs2_block_check_validate(void *data, size_t blocksize,
+			       struct ocfs2_block_check *bc)
+{
+	int rc = 0;
+	struct ocfs2_block_check check;
+	u32 crc, ecc;
+
+	check.bc_crc32e = le32_to_cpu(bc->bc_crc32e);
+	check.bc_ecc = le16_to_cpu(bc->bc_ecc);
+
+	memset(bc, 0, sizeof(struct ocfs2_block_check));
+
+	/* Fast path - if the crc32 validates, we're good to go */
+	crc = crc32_le(~0, data, blocksize);
+	if (crc == check.bc_crc32e)
+		goto out;
+
+	mlog(ML_ERROR,
+	     "CRC32 failed: stored: %u, computed %u.  Applying ECC.\n",
+	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
+
+	/* Ok, try ECC fixups */
+	ecc = ocfs2_hamming_encode_block(data, blocksize);
+	ocfs2_hamming_fix_block(data, blocksize, ecc ^ check.bc_ecc);
+
+	/* And check the crc32 again */
+	crc = crc32_le(~0, data, blocksize);
+	if (crc == check.bc_crc32e)
+		goto out;
+
+	mlog(ML_ERROR, "Fixed CRC32 failed: stored: %u, computed %u\n",
+	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
+
+	rc = -EIO;
+
+out:
+	bc->bc_crc32e = cpu_to_le32(check.bc_crc32e);
+	bc->bc_ecc = cpu_to_le16(check.bc_ecc);
+
+	return rc;
+}
+
+/*
+ * This function generates check information for a list of buffer_heads.
+ * bhs is the blocks to be checked.  bc is a pointer to the
+ * ocfs2_block_check structure describing the crc32 and the ecc.
+ *
+ * bc should be a pointer inside data, as the function will
+ * take care of zeroing it before calculating the check information.  If
+ * bc does not point inside data, the caller must make sure any inline
+ * ocfs2_block_check structures are zeroed.
+ *
+ * The data buffer must be in on-disk endian (little endian for ocfs2).
+ * bc will be filled with little-endian values and will be ready to go to
+ * disk.
+ */
+void ocfs2_block_check_compute_bhs(struct buffer_head **bhs, int nr,
+				   struct ocfs2_block_check *bc)
+{
+	int i;
+	u32 crc, ecc;
+
+	BUG_ON(nr < 0);
+
+	if (!nr)
+		return;
+
+	memset(bc, 0, sizeof(struct ocfs2_block_check));
+
+	for (i = 0, crc = ~0, ecc = 0; i < nr; i++) {
+		crc = crc32_le(crc, bhs[i]->b_data, bhs[i]->b_size);
+		/*
+		 * The number of bits in a buffer is obviously b_size*8.
+		 * The offset of this buffer is b_size*i, so the bit offset
+		 * of this buffer is b_size*8*i.
+		 */
+		ecc = (u16)ocfs2_hamming_encode(ecc, bhs[i]->b_data,
+						bhs[i]->b_size * 8,
+						bhs[i]->b_size * 8 * i);
+	}
+
+	/*
+	 * No ecc'd ocfs2 structure is larger than 4K, so ecc will be no
+	 * larger than 16 bits.
+	 */
+	BUG_ON(ecc > USHORT_MAX);
+
+	bc->bc_crc32e = cpu_to_le32(crc);
+	bc->bc_ecc = cpu_to_le16((u16)ecc);
+}
+
+/*
+ * This function validates existing check information on a list of
+ * buffer_heads.  Like _compute_bhs, the function will take care of
+ * zeroing bc before calculating check codes.  If bc is not a pointer
+ * inside data, the caller must have zeroed any inline
+ * ocfs2_block_check structures.
+ *
+ * Again, the data passed in should be the on-disk endian.
+ */
+int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
+				   struct ocfs2_block_check *bc)
+{
+	int i, rc = 0;
+	struct ocfs2_block_check check;
+	u32 crc, ecc, fix;
+
+	BUG_ON(nr < 0);
+
+	if (!nr)
+		return 0;
+
+	check.bc_crc32e = le32_to_cpu(bc->bc_crc32e);
+	check.bc_ecc = le16_to_cpu(bc->bc_ecc);
+
+	memset(bc, 0, sizeof(struct ocfs2_block_check));
+
+	/* Fast path - if the crc32 validates, we're good to go */
+	for (i = 0, crc = ~0; i < nr; i++)
+		crc = crc32_le(crc, bhs[i]->b_data, bhs[i]->b_size);
+	if (crc == check.bc_crc32e)
+		goto out;
+
+	mlog(ML_ERROR,
+	     "CRC32 failed: stored: %u, computed %u.  Applying ECC.\n",
+	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
+
+	/* Ok, try ECC fixups */
+	for (i = 0, ecc = 0; i < nr; i++) {
+		/*
+		 * The number of bits in a buffer is obviously b_size*8.
+		 * The offset of this buffer is b_size*i, so the bit offset
+		 * of this buffer is b_size*8*i.
+		 */
+		ecc = (u16)ocfs2_hamming_encode(ecc, bhs[i]->b_data,
+						bhs[i]->b_size * 8,
+						bhs[i]->b_size * 8 * i);
+	}
+	fix = ecc ^ check.bc_ecc;
+	for (i = 0; i < nr; i++) {
+		/*
+		 * Try the fix against each buffer.  It will only affect
+		 * one of them.
+		 */
+		ocfs2_hamming_fix(bhs[i]->b_data, bhs[i]->b_size * 8,
+				  bhs[i]->b_size * 8 * i, fix);
+	}
+
+	/* And check the crc32 again */
+	for (i = 0, crc = ~0; i < nr; i++)
+		crc = crc32_le(crc, bhs[i]->b_data, bhs[i]->b_size);
+	if (crc == check.bc_crc32e)
+		goto out;
+
+	mlog(ML_ERROR, "Fixed CRC32 failed: stored: %u, computed %u\n",
+	     (unsigned int)check.bc_crc32e, (unsigned int)crc);
+
+	rc = -EIO;
+
+out:
+	bc->bc_crc32e = cpu_to_le32(check.bc_crc32e);
+	bc->bc_ecc = cpu_to_le16(check.bc_ecc);
+
+	return rc;
+}
+
+/*
+ * These are the main API.  They check the superblock flag before
+ * calling the underlying operations.
+ *
+ * They expect the buffer(s) to be in disk format.
+ */
+void ocfs2_compute_meta_ecc(struct super_block *sb, void *data,
+			    struct ocfs2_block_check *bc)
+{
+	if (ocfs2_meta_ecc(OCFS2_SB(sb)))
+		ocfs2_block_check_compute(data, sb->s_blocksize, bc);
+}
+
+int ocfs2_validate_meta_ecc(struct super_block *sb, void *data,
+			    struct ocfs2_block_check *bc)
+{
+	int rc = 0;
+
+	if (ocfs2_meta_ecc(OCFS2_SB(sb)))
+		rc = ocfs2_block_check_validate(data, sb->s_blocksize, bc);
+
+	return rc;
+}
+
+void ocfs2_compute_meta_ecc_bhs(struct super_block *sb,
+				struct buffer_head **bhs, int nr,
+				struct ocfs2_block_check *bc)
+{
+	if (ocfs2_meta_ecc(OCFS2_SB(sb)))
+		ocfs2_block_check_compute_bhs(bhs, nr, bc);
+}
+
+int ocfs2_validate_meta_ecc_bhs(struct super_block *sb,
+				struct buffer_head **bhs, int nr,
+				struct ocfs2_block_check *bc)
+{
+	int rc = 0;
+
+	if (ocfs2_meta_ecc(OCFS2_SB(sb)))
+		rc = ocfs2_block_check_validate_bhs(bhs, nr, bc);
+
+	return rc;
+}
+
diff --git a/fs/ocfs2/blockcheck.h b/fs/ocfs2/blockcheck.h
new file mode 100644
index 00000000000..70ec3feda32
--- /dev/null
+++ b/fs/ocfs2/blockcheck.h
@@ -0,0 +1,82 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * blockcheck.h
+ *
+ * Checksum and ECC codes for the OCFS2 userspace library.
+ *
+ * Copyright (C) 2004, 2008 Oracle.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#ifndef OCFS2_BLOCKCHECK_H
+#define OCFS2_BLOCKCHECK_H
+
+
+/* High level block API */
+void ocfs2_compute_meta_ecc(struct super_block *sb, void *data,
+			    struct ocfs2_block_check *bc);
+int ocfs2_validate_meta_ecc(struct super_block *sb, void *data,
+			    struct ocfs2_block_check *bc);
+void ocfs2_compute_meta_ecc_bhs(struct super_block *sb,
+				struct buffer_head **bhs, int nr,
+				struct ocfs2_block_check *bc);
+int ocfs2_validate_meta_ecc_bhs(struct super_block *sb,
+				struct buffer_head **bhs, int nr,
+				struct ocfs2_block_check *bc);
+
+/* Lower level API */
+void ocfs2_block_check_compute(void *data, size_t blocksize,
+			       struct ocfs2_block_check *bc);
+int ocfs2_block_check_validate(void *data, size_t blocksize,
+			       struct ocfs2_block_check *bc);
+void ocfs2_block_check_compute_bhs(struct buffer_head **bhs, int nr,
+				   struct ocfs2_block_check *bc);
+int ocfs2_block_check_validate_bhs(struct buffer_head **bhs, int nr,
+				   struct ocfs2_block_check *bc);
+
+/*
+ * Hamming code functions
+ */
+
+/*
+ * Encoding hamming code parity bits for a buffer.
+ *
+ * This is the low level encoder function.  It can be called across
+ * multiple hunks just like the crc32 code.  'd' is the number of bits
+ * _in_this_hunk_.  nr is the bit offset of this hunk.  So, if you had
+ * two 512B buffers, you would do it like so:
+ *
+ * parity = ocfs2_hamming_encode(0, buf1, 512 * 8, 0);
+ * parity = ocfs2_hamming_encode(parity, buf2, 512 * 8, 512 * 8);
+ *
+ * If you just have one buffer, use ocfs2_hamming_encode_block().
+ */
+u32 ocfs2_hamming_encode(u32 parity, void *data, unsigned int d,
+			 unsigned int nr);
+/*
+ * Fix a buffer with a bit error.  The 'fix' is the original parity
+ * xor'd with the parity calculated now.
+ *
+ * Like ocfs2_hamming_encode(), this can handle hunks.  nr is the bit
+ * offset of the current hunk.  If bit to be fixed is not part of the
+ * current hunk, this does nothing.
+ *
+ * If you only have one buffer, use ocfs2_hamming_fix_block().
+ */
+void ocfs2_hamming_fix(void *data, unsigned int d, unsigned int nr,
+		       unsigned int fix);
+
+/* Convenience wrappers for a single buffer of data */
+extern u32 ocfs2_hamming_encode_block(void *data, unsigned int blocksize);
+extern void ocfs2_hamming_fix_block(void *data, unsigned int blocksize,
+				    unsigned int fix);
+#endif
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index 3a178ec48d7..15c8e6deee2 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -39,6 +39,18 @@
 
 #include "buffer_head_io.h"
 
+/*
+ * Bits on bh->b_state used by ocfs2.
+ *
+ * These MUST be after the JBD2 bits.  Hence, we use BH_JBDPrivateStart.
+ */
+enum ocfs2_state_bits {
+	BH_NeedsValidate = BH_JBDPrivateStart,
+};
+
+/* Expand the magic b_state functions */
+BUFFER_FNS(NeedsValidate, needs_validate);
+
 int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
 		      struct inode *inode)
 {
@@ -166,7 +178,9 @@ bail:
 }
 
 int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
-		      struct buffer_head *bhs[], int flags)
+		      struct buffer_head *bhs[], int flags,
+		      int (*validate)(struct super_block *sb,
+				      struct buffer_head *bh))
 {
 	int status = 0;
 	int i, ignore_cache = 0;
@@ -298,6 +312,8 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 
 			clear_buffer_uptodate(bh);
 			get_bh(bh); /* for end_buffer_read_sync() */
+			if (validate)
+				set_buffer_needs_validate(bh);
 			bh->b_end_io = end_buffer_read_sync;
 			submit_bh(READ, bh);
 			continue;
@@ -328,6 +344,20 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 				bhs[i] = NULL;
 				continue;
 			}
+
+			if (buffer_needs_validate(bh)) {
+				/* We never set NeedsValidate if the
+				 * buffer was held by the journal, so
+				 * that better not have changed */
+				BUG_ON(buffer_jbd(bh));
+				clear_buffer_needs_validate(bh);
+				status = validate(inode->i_sb, bh);
+				if (status) {
+					put_bh(bh);
+					bhs[i] = NULL;
+					continue;
+				}
+			}
 		}
 
 		/* Always set the buffer in the cache, even if it was
diff --git a/fs/ocfs2/buffer_head_io.h b/fs/ocfs2/buffer_head_io.h
index 75e1dcb1ade..c75d682dadd 100644
--- a/fs/ocfs2/buffer_head_io.h
+++ b/fs/ocfs2/buffer_head_io.h
@@ -31,21 +31,24 @@
 void ocfs2_end_buffer_io_sync(struct buffer_head *bh,
 			     int uptodate);
 
-static inline int ocfs2_read_block(struct inode	       *inode,
-				   u64                  off,
-				   struct buffer_head **bh);
-
 int ocfs2_write_block(struct ocfs2_super          *osb,
 		      struct buffer_head  *bh,
 		      struct inode        *inode);
-int ocfs2_read_blocks(struct inode	  *inode,
-		      u64                  block,
-		      int                  nr,
-		      struct buffer_head  *bhs[],
-		      int                  flags);
 int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block,
 			   unsigned int nr, struct buffer_head *bhs[]);
 
+/*
+ * If not NULL, validate() will be called on a buffer that is freshly
+ * read from disk.  It will not be called if the buffer was in cache.
+ * Note that if validate() is being used for this buffer, it needs to
+ * be set even for a READAHEAD call, as it marks the buffer for later
+ * validation.
+ */
+int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
+		      struct buffer_head *bhs[], int flags,
+		      int (*validate)(struct super_block *sb,
+				      struct buffer_head *bh));
+
 int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
 				struct buffer_head *bh);
 
@@ -53,7 +56,9 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
 #define OCFS2_BH_READAHEAD         8
 
 static inline int ocfs2_read_block(struct inode *inode, u64 off,
-				   struct buffer_head **bh)
+				   struct buffer_head **bh,
+				   int (*validate)(struct super_block *sb,
+						   struct buffer_head *bh))
 {
 	int status = 0;
 
@@ -63,7 +68,7 @@ static inline int ocfs2_read_block(struct inode *inode, u64 off,
 		goto bail;
 	}
 
-	status = ocfs2_read_blocks(inode, off, 1, bh, 0);
+	status = ocfs2_read_blocks(inode, off, 1, bh, 0, validate);
 
 bail:
 	return status;
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index 6ebaa58e2c0..04697ba7f73 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -854,7 +854,7 @@ static int o2hb_thread(void *data)
 
 	while (!kthread_should_stop() && !reg->hr_unclean_stop) {
 		/* We track the time spent inside
-		 * o2hb_do_disk_heartbeat so that we avoid more then
+		 * o2hb_do_disk_heartbeat so that we avoid more than
 		 * hr_timeout_ms between disk writes. On busy systems
 		 * this should result in a heartbeat which is less
 		 * likely to time itself out. */
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index d8a0cb92cef..96df5416993 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -110,6 +110,7 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
 	define_mask(QUORUM),
 	define_mask(EXPORT),
 	define_mask(XATTR),
+	define_mask(QUOTA),
 	define_mask(ERROR),
 	define_mask(NOTICE),
 	define_mask(KTHREAD),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 57670c68047..7e72a81bc2d 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -113,6 +113,7 @@
 #define ML_QUORUM	0x0000000008000000ULL /* net connection quorum */
 #define ML_EXPORT	0x0000000010000000ULL /* ocfs2 export operations */
 #define ML_XATTR	0x0000000020000000ULL /* ocfs2 extended attributes */
+#define ML_QUOTA	0x0000000040000000ULL /* ocfs2 quota operations */
 /* bits that are infrequently given and frequently matched in the high word */
 #define ML_ERROR	0x0000000100000000ULL /* sent to KERN_ERR */
 #define ML_NOTICE	0x0000000200000000ULL /* setn to KERN_NOTICE */
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 026e6eb8518..f2c4098cf33 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -40,6 +40,7 @@
 #include <linux/types.h>
 #include <linux/slab.h>
 #include <linux/highmem.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_NAMEI
 #include <cluster/masklog.h>
@@ -47,6 +48,7 @@
 #include "ocfs2.h"
 
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dir.h"
 #include "dlmglue.h"
 #include "extent_map.h"
@@ -82,47 +84,72 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
 			       struct ocfs2_alloc_context *meta_ac,
 			       struct buffer_head **new_bh);
 
-static struct buffer_head *ocfs2_bread(struct inode *inode,
-				       int block, int *err, int reada)
+/*
+ * These are distinct checks because future versions of the file system will
+ * want to have a trailing dirent structure independent of indexing.
+ */
+static int ocfs2_dir_has_trailer(struct inode *dir)
 {
-	struct buffer_head *bh = NULL;
-	int tmperr;
-	u64 p_blkno;
-	int readflags = 0;
+	if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
+		return 0;
 
-	if (reada)
-		readflags |= OCFS2_BH_READAHEAD;
+	return ocfs2_meta_ecc(OCFS2_SB(dir->i_sb));
+}
 
-	if (((u64)block << inode->i_sb->s_blocksize_bits) >=
-	    i_size_read(inode)) {
-		BUG_ON(!reada);
-		return NULL;
-	}
+static int ocfs2_supports_dir_trailer(struct ocfs2_super *osb)
+{
+	return ocfs2_meta_ecc(osb);
+}
 
-	down_read(&OCFS2_I(inode)->ip_alloc_sem);
-	tmperr = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
-					     NULL);
-	up_read(&OCFS2_I(inode)->ip_alloc_sem);
-	if (tmperr < 0) {
-		mlog_errno(tmperr);
-		goto fail;
-	}
+static inline unsigned int ocfs2_dir_trailer_blk_off(struct super_block *sb)
+{
+	return sb->s_blocksize - sizeof(struct ocfs2_dir_block_trailer);
+}
 
-	tmperr = ocfs2_read_blocks(inode, p_blkno, 1, &bh, readflags);
-	if (tmperr < 0)
-		goto fail;
+#define ocfs2_trailer_from_bh(_bh, _sb) ((struct ocfs2_dir_block_trailer *) ((_bh)->b_data + ocfs2_dir_trailer_blk_off((_sb))))
 
-	tmperr = 0;
+/* XXX ocfs2_block_dqtrailer() is similar but not quite - can we make
+ * them more consistent? */
+struct ocfs2_dir_block_trailer *ocfs2_dir_trailer_from_size(int blocksize,
+							    void *data)
+{
+	char *p = data;
 
-	*err = 0;
-	return bh;
+	p += blocksize - sizeof(struct ocfs2_dir_block_trailer);
+	return (struct ocfs2_dir_block_trailer *)p;
+}
 
-fail:
-	brelse(bh);
-	bh = NULL;
+/*
+ * XXX: This is executed once on every dirent. We should consider optimizing
+ * it.
+ */
+static int ocfs2_skip_dir_trailer(struct inode *dir,
+				  struct ocfs2_dir_entry *de,
+				  unsigned long offset,
+				  unsigned long blklen)
+{
+	unsigned long toff = blklen - sizeof(struct ocfs2_dir_block_trailer);
 
-	*err = -EIO;
-	return NULL;
+	if (!ocfs2_dir_has_trailer(dir))
+		return 0;
+
+	if (offset != toff)
+		return 0;
+
+	return 1;
+}
+
+static void ocfs2_init_dir_trailer(struct inode *inode,
+				   struct buffer_head *bh)
+{
+	struct ocfs2_dir_block_trailer *trailer;
+
+	trailer = ocfs2_trailer_from_bh(bh, inode->i_sb);
+	strcpy(trailer->db_signature, OCFS2_DIR_TRAILER_SIGNATURE);
+	trailer->db_compat_rec_len =
+			cpu_to_le16(sizeof(struct ocfs2_dir_block_trailer));
+	trailer->db_parent_dinode = cpu_to_le64(OCFS2_I(inode)->ip_blkno);
+	trailer->db_blkno = cpu_to_le64(bh->b_blocknr);
 }
 
 /*
@@ -231,7 +258,7 @@ static struct buffer_head *ocfs2_find_entry_id(const char *name,
 	struct ocfs2_dinode *di;
 	struct ocfs2_inline_data *data;
 
-	ret = ocfs2_read_block(dir, OCFS2_I(dir)->ip_blkno, &di_bh);
+	ret = ocfs2_read_inode_block(dir, &di_bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -250,6 +277,108 @@ out:
 	return NULL;
 }
 
+static int ocfs2_validate_dir_block(struct super_block *sb,
+				    struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_dir_block_trailer *trailer =
+		ocfs2_trailer_from_bh(bh, sb);
+
+
+	/*
+	 * We don't validate dirents here, that's handled
+	 * in-place when the code walks them.
+	 */
+	mlog(0, "Validating dirblock %llu\n",
+	     (unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 *
+	 * Note that we are safe to call this even if the directory
+	 * doesn't have a trailer.  Filesystems without metaecc will do
+	 * nothing, and filesystems with it will have one.
+	 */
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &trailer->db_check);
+	if (rc)
+		mlog(ML_ERROR, "Checksum failed for dinode %llu\n",
+		     (unsigned long long)bh->b_blocknr);
+
+	return rc;
+}
+
+/*
+ * This function forces all errors to -EIO for consistency with its
+ * predecessor, ocfs2_bread().  We haven't audited what returning the
+ * real error codes would do to callers.  We log the real codes with
+ * mlog_errno() before we squash them.
+ */
+static int ocfs2_read_dir_block(struct inode *inode, u64 v_block,
+				struct buffer_head **bh, int flags)
+{
+	int rc = 0;
+	struct buffer_head *tmp = *bh;
+	struct ocfs2_dir_block_trailer *trailer;
+
+	rc = ocfs2_read_virt_blocks(inode, v_block, 1, &tmp, flags,
+				    ocfs2_validate_dir_block);
+	if (rc) {
+		mlog_errno(rc);
+		goto out;
+	}
+
+	/*
+	 * We check the trailer here rather than in
+	 * ocfs2_validate_dir_block() because that function doesn't have
+	 * the inode to test.
+	 */
+	if (!(flags & OCFS2_BH_READAHEAD) &&
+	    ocfs2_dir_has_trailer(inode)) {
+		trailer = ocfs2_trailer_from_bh(tmp, inode->i_sb);
+		if (!OCFS2_IS_VALID_DIR_TRAILER(trailer)) {
+			rc = -EINVAL;
+			ocfs2_error(inode->i_sb,
+				    "Invalid dirblock #%llu: "
+				    "signature = %.*s\n",
+				    (unsigned long long)tmp->b_blocknr, 7,
+				    trailer->db_signature);
+			goto out;
+		}
+		if (le64_to_cpu(trailer->db_blkno) != tmp->b_blocknr) {
+			rc = -EINVAL;
+			ocfs2_error(inode->i_sb,
+				    "Directory block #%llu has an invalid "
+				    "db_blkno of %llu",
+				    (unsigned long long)tmp->b_blocknr,
+				    (unsigned long long)le64_to_cpu(trailer->db_blkno));
+			goto out;
+		}
+		if (le64_to_cpu(trailer->db_parent_dinode) !=
+		    OCFS2_I(inode)->ip_blkno) {
+			rc = -EINVAL;
+			ocfs2_error(inode->i_sb,
+				    "Directory block #%llu on dinode "
+				    "#%llu has an invalid parent_dinode "
+				    "of %llu",
+				    (unsigned long long)tmp->b_blocknr,
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+				    (unsigned long long)le64_to_cpu(trailer->db_blkno));
+			goto out;
+		}
+	}
+
+	/* If ocfs2_read_virt_blocks() got us a new bh, pass it up. */
+	if (!*bh)
+		*bh = tmp;
+
+out:
+	return rc ? -EIO : 0;
+}
+
 static struct buffer_head *ocfs2_find_entry_el(const char *name, int namelen,
 					       struct inode *dir,
 					       struct ocfs2_dir_entry **res_dir)
@@ -296,15 +425,17 @@ restart:
 				}
 				num++;
 
-				bh = ocfs2_bread(dir, b++, &err, 1);
+				bh = NULL;
+				err = ocfs2_read_dir_block(dir, b++, &bh,
+							   OCFS2_BH_READAHEAD);
 				bh_use[ra_max] = bh;
 			}
 		}
 		if ((bh = bh_use[ra_ptr++]) == NULL)
 			goto next;
-		if (ocfs2_read_block(dir, block, &bh)) {
+		if (ocfs2_read_dir_block(dir, block, &bh, 0)) {
 			/* read error, skip block & hope for the best.
-			 * ocfs2_read_block() has released the bh. */
+			 * ocfs2_read_dir_block() has released the bh. */
 			ocfs2_error(dir->i_sb, "reading directory %llu, "
 				    "offset %lu\n",
 				    (unsigned long long)OCFS2_I(dir)->ip_blkno,
@@ -381,14 +512,18 @@ int ocfs2_update_entry(struct inode *dir, handle_t *handle,
 		       struct inode *new_entry_inode)
 {
 	int ret;
+	ocfs2_journal_access_func access = ocfs2_journal_access_db;
 
 	/*
 	 * The same code works fine for both inline-data and extent
-	 * based directories, so no need to split this up.
+	 * based directories, so no need to split this up.  The only
+	 * difference is the journal_access function.
 	 */
 
-	ret = ocfs2_journal_access(handle, dir, de_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
+		access = ocfs2_journal_access_di;
+
+	ret = access(handle, dir, de_bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -410,9 +545,13 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
 {
 	struct ocfs2_dir_entry *de, *pde;
 	int i, status = -ENOENT;
+	ocfs2_journal_access_func access = ocfs2_journal_access_db;
 
 	mlog_entry("(0x%p, 0x%p, 0x%p, 0x%p)\n", handle, dir, de_del, bh);
 
+	if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
+		access = ocfs2_journal_access_di;
+
 	i = 0;
 	pde = NULL;
 	de = (struct ocfs2_dir_entry *) first_de;
@@ -423,8 +562,8 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
 			goto bail;
 		}
 		if (de == de_del)  {
-			status = ocfs2_journal_access(handle, dir, bh,
-						      OCFS2_JOURNAL_ACCESS_WRITE);
+			status = access(handle, dir, bh,
+					OCFS2_JOURNAL_ACCESS_WRITE);
 			if (status < 0) {
 				status = -EIO;
 				mlog_errno(status);
@@ -458,7 +597,7 @@ static inline int ocfs2_delete_entry_id(handle_t *handle,
 	struct ocfs2_dinode *di;
 	struct ocfs2_inline_data *data;
 
-	ret = ocfs2_read_block(dir, OCFS2_I(dir)->ip_blkno, &di_bh);
+	ret = ocfs2_read_inode_block(dir, &di_bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -576,6 +715,16 @@ int __ocfs2_add_entry(handle_t *handle,
 			goto bail;
 		}
 
+		/* We're guaranteed that we should have space, so we
+		 * can't possibly have hit the trailer...right? */
+		mlog_bug_on_msg(ocfs2_skip_dir_trailer(dir, de, offset, size),
+				"Hit dir trailer trying to insert %.*s "
+			        "(namelen %d) into directory %llu.  "
+				"offset is %lu, trailer offset is %d\n",
+				namelen, name, namelen,
+				(unsigned long long)parent_fe_bh->b_blocknr,
+				offset, ocfs2_dir_trailer_blk_off(dir->i_sb));
+
 		if (ocfs2_dirent_would_fit(de, rec_len)) {
 			dir->i_mtime = dir->i_ctime = CURRENT_TIME;
 			retval = ocfs2_mark_inode_dirty(handle, dir, parent_fe_bh);
@@ -584,8 +733,14 @@ int __ocfs2_add_entry(handle_t *handle,
 				goto bail;
 			}
 
-			status = ocfs2_journal_access(handle, dir, insert_bh,
-						      OCFS2_JOURNAL_ACCESS_WRITE);
+			if (insert_bh == parent_fe_bh)
+				status = ocfs2_journal_access_di(handle, dir,
+								 insert_bh,
+								 OCFS2_JOURNAL_ACCESS_WRITE);
+			else
+				status = ocfs2_journal_access_db(handle, dir,
+								 insert_bh,
+								 OCFS2_JOURNAL_ACCESS_WRITE);
 			/* By now the buffer is marked for journaling */
 			offset += le16_to_cpu(de->rec_len);
 			if (le64_to_cpu(de->inode)) {
@@ -611,6 +766,7 @@ int __ocfs2_add_entry(handle_t *handle,
 			retval = 0;
 			goto bail;
 		}
+
 		offset += le16_to_cpu(de->rec_len);
 		de = (struct ocfs2_dir_entry *) ((char *) de + le16_to_cpu(de->rec_len));
 	}
@@ -636,7 +792,7 @@ static int ocfs2_dir_foreach_blk_id(struct inode *inode,
 	struct ocfs2_inline_data *data;
 	struct ocfs2_dir_entry *de;
 
-	ret = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, &di_bh);
+	ret = ocfs2_read_inode_block(inode, &di_bh);
 	if (ret) {
 		mlog(ML_ERROR, "Unable to read inode block for dir %llu\n",
 		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
@@ -724,7 +880,6 @@ static int ocfs2_dir_foreach_blk_el(struct inode *inode,
 	int i, stored;
 	struct buffer_head * bh, * tmp;
 	struct ocfs2_dir_entry * de;
-	int err;
 	struct super_block * sb = inode->i_sb;
 	unsigned int ra_sectors = 16;
 
@@ -735,12 +890,8 @@ static int ocfs2_dir_foreach_blk_el(struct inode *inode,
 
 	while (!error && !stored && *f_pos < i_size_read(inode)) {
 		blk = (*f_pos) >> sb->s_blocksize_bits;
-		bh = ocfs2_bread(inode, blk, &err, 0);
-		if (!bh) {
-			mlog(ML_ERROR,
-			     "directory #%llu contains a hole at offset %lld\n",
-			     (unsigned long long)OCFS2_I(inode)->ip_blkno,
-			     *f_pos);
+		if (ocfs2_read_dir_block(inode, blk, &bh, 0)) {
+			/* Skip the corrupt dirblock and keep trying */
 			*f_pos += sb->s_blocksize - offset;
 			continue;
 		}
@@ -754,8 +905,10 @@ static int ocfs2_dir_foreach_blk_el(struct inode *inode,
 		    || (((last_ra_blk - blk) << 9) <= (ra_sectors / 2))) {
 			for (i = ra_sectors >> (sb->s_blocksize_bits - 9);
 			     i > 0; i--) {
-				tmp = ocfs2_bread(inode, ++blk, &err, 1);
-				brelse(tmp);
+				tmp = NULL;
+				if (!ocfs2_read_dir_block(inode, ++blk, &tmp,
+							  OCFS2_BH_READAHEAD))
+					brelse(tmp);
 			}
 			last_ra_blk = blk;
 			ra_sectors = 8;
@@ -828,6 +981,7 @@ revalidate:
 		}
 		offset = 0;
 		brelse(bh);
+		bh = NULL;
 	}
 
 	stored = 0;
@@ -1050,9 +1204,15 @@ int ocfs2_empty_dir(struct inode *inode)
 	return !priv.seen_other;
 }
 
-static void ocfs2_fill_initial_dirents(struct inode *inode,
-				       struct inode *parent,
-				       char *start, unsigned int size)
+/*
+ * Fills "." and ".." dirents in a new directory block. Returns dirent for
+ * "..", which might be used during creation of a directory with a trailing
+ * header. It is otherwise safe to ignore the return code.
+ */
+static struct ocfs2_dir_entry *ocfs2_fill_initial_dirents(struct inode *inode,
+							  struct inode *parent,
+							  char *start,
+							  unsigned int size)
 {
 	struct ocfs2_dir_entry *de = (struct ocfs2_dir_entry *)start;
 
@@ -1069,6 +1229,8 @@ static void ocfs2_fill_initial_dirents(struct inode *inode,
 	de->name_len = 2;
 	strcpy(de->name, "..");
 	ocfs2_set_de_type(de, S_IFDIR);
+
+	return de;
 }
 
 /*
@@ -1086,8 +1248,8 @@ static int ocfs2_fill_new_dir_id(struct ocfs2_super *osb,
 	struct ocfs2_inline_data *data = &di->id2.i_data;
 	unsigned int size = le16_to_cpu(data->id_count);
 
-	ret = ocfs2_journal_access(handle, inode, di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -1121,10 +1283,15 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb,
 				 struct ocfs2_alloc_context *data_ac)
 {
 	int status;
+	unsigned int size = osb->sb->s_blocksize;
 	struct buffer_head *new_bh = NULL;
+	struct ocfs2_dir_entry *de;
 
 	mlog_entry_void();
 
+	if (ocfs2_supports_dir_trailer(osb))
+		size = ocfs2_dir_trailer_blk_off(parent->i_sb);
+
 	status = ocfs2_do_extend_dir(osb->sb, handle, inode, fe_bh,
 				     data_ac, NULL, &new_bh);
 	if (status < 0) {
@@ -1134,16 +1301,17 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb,
 
 	ocfs2_set_new_buffer_uptodate(inode, new_bh);
 
-	status = ocfs2_journal_access(handle, inode, new_bh,
-				      OCFS2_JOURNAL_ACCESS_CREATE);
+	status = ocfs2_journal_access_db(handle, inode, new_bh,
+					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 	memset(new_bh->b_data, 0, osb->sb->s_blocksize);
 
-	ocfs2_fill_initial_dirents(inode, parent, new_bh->b_data,
-				   osb->sb->s_blocksize);
+	de = ocfs2_fill_initial_dirents(inode, parent, new_bh->b_data, size);
+	if (ocfs2_supports_dir_trailer(osb))
+		ocfs2_init_dir_trailer(inode, new_bh);
 
 	status = ocfs2_journal_dirty(handle, new_bh);
 	if (status < 0) {
@@ -1184,13 +1352,27 @@ int ocfs2_fill_new_dir(struct ocfs2_super *osb,
 				     data_ac);
 }
 
+/*
+ * Expand rec_len of the rightmost dirent in a directory block so that it
+ * contains the end of our valid space for dirents. We do this during
+ * expansion from an inline directory to one with extents. The first dir block
+ * in that case is taken from the inline data portion of the inode block.
+ *
+ * We add the dir trailer if this filesystem wants it.
+ */
 static void ocfs2_expand_last_dirent(char *start, unsigned int old_size,
-				     unsigned int new_size)
+				     struct super_block *sb)
 {
 	struct ocfs2_dir_entry *de;
 	struct ocfs2_dir_entry *prev_de;
 	char *de_buf, *limit;
-	unsigned int bytes = new_size - old_size;
+	unsigned int new_size = sb->s_blocksize;
+	unsigned int bytes;
+
+	if (ocfs2_supports_dir_trailer(OCFS2_SB(sb)))
+		new_size = ocfs2_dir_trailer_blk_off(sb);
+
+	bytes = new_size - old_size;
 
 	limit = start + old_size;
 	de_buf = start;
@@ -1216,9 +1398,9 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 				   unsigned int blocks_wanted,
 				   struct buffer_head **first_block_bh)
 {
-	int ret, credits = OCFS2_INLINE_TO_EXTENTS_CREDITS;
 	u32 alloc, bit_off, len;
 	struct super_block *sb = dir->i_sb;
+	int ret, credits = ocfs2_inline_to_extents_credits(sb);
 	u64 blkno, bytes = blocks_wanted << sb->s_blocksize_bits;
 	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
 	struct ocfs2_inode_info *oi = OCFS2_I(dir);
@@ -1227,6 +1409,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
 	handle_t *handle;
 	struct ocfs2_extent_tree et;
+	int did_quota = 0;
 
 	ocfs2_init_dinode_extent_tree(&et, dir, di_bh);
 
@@ -1264,6 +1447,12 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 		goto out_sem;
 	}
 
+	if (vfs_dq_alloc_space_nodirty(dir,
+				ocfs2_clusters_to_bytes(osb->sb, alloc))) {
+		ret = -EDQUOT;
+		goto out_commit;
+	}
+	did_quota = 1;
 	/*
 	 * Try to claim as many clusters as the bitmap can give though
 	 * if we only get one now, that's enough to continue. The rest
@@ -1290,8 +1479,8 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 
 	ocfs2_set_new_buffer_uptodate(dir, dirdata_bh);
 
-	ret = ocfs2_journal_access(handle, dir, dirdata_bh,
-				   OCFS2_JOURNAL_ACCESS_CREATE);
+	ret = ocfs2_journal_access_db(handle, dir, dirdata_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -1300,8 +1489,9 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	memcpy(dirdata_bh->b_data, di->id2.i_data.id_data, i_size_read(dir));
 	memset(dirdata_bh->b_data + i_size_read(dir), 0,
 	       sb->s_blocksize - i_size_read(dir));
-	ocfs2_expand_last_dirent(dirdata_bh->b_data, i_size_read(dir),
-				 sb->s_blocksize);
+	ocfs2_expand_last_dirent(dirdata_bh->b_data, i_size_read(dir), sb);
+	if (ocfs2_supports_dir_trailer(osb))
+		ocfs2_init_dir_trailer(dir, dirdata_bh);
 
 	ret = ocfs2_journal_dirty(handle, dirdata_bh);
 	if (ret) {
@@ -1317,8 +1507,8 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	 * We let the later dirent insert modify c/mtime - to the user
 	 * the data hasn't changed.
 	 */
-	ret = ocfs2_journal_access(handle, dir, di_bh,
-				   OCFS2_JOURNAL_ACCESS_CREATE);
+	ret = ocfs2_journal_access_di(handle, dir, di_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -1386,6 +1576,9 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	dirdata_bh = NULL;
 
 out_commit:
+	if (ret < 0 && did_quota)
+		vfs_dq_free_space_nodirty(dir,
+			ocfs2_clusters_to_bytes(osb->sb, 2));
 	ocfs2_commit_trans(osb, handle);
 
 out_sem:
@@ -1410,7 +1603,7 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
 			       struct buffer_head **new_bh)
 {
 	int status;
-	int extend;
+	int extend, did_quota = 0;
 	u64 p_blkno, v_blkno;
 
 	spin_lock(&OCFS2_I(dir)->ip_lock);
@@ -1420,6 +1613,13 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
 	if (extend) {
 		u32 offset = OCFS2_I(dir)->ip_clusters;
 
+		if (vfs_dq_alloc_space_nodirty(dir,
+					ocfs2_clusters_to_bytes(sb, 1))) {
+			status = -EDQUOT;
+			goto bail;
+		}
+		did_quota = 1;
+
 		status = ocfs2_add_inode_data(OCFS2_SB(sb), dir, &offset,
 					      1, 0, parent_fe_bh, handle,
 					      data_ac, meta_ac, NULL);
@@ -1445,6 +1645,8 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
 	}
 	status = 0;
 bail:
+	if (did_quota && status < 0)
+		vfs_dq_free_space_nodirty(dir, ocfs2_clusters_to_bytes(sb, 1));
 	mlog_exit(status);
 	return status;
 }
@@ -1569,16 +1771,22 @@ do_extend:
 
 	ocfs2_set_new_buffer_uptodate(dir, new_bh);
 
-	status = ocfs2_journal_access(handle, dir, new_bh,
-				      OCFS2_JOURNAL_ACCESS_CREATE);
+	status = ocfs2_journal_access_db(handle, dir, new_bh,
+					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 	memset(new_bh->b_data, 0, sb->s_blocksize);
+
 	de = (struct ocfs2_dir_entry *) new_bh->b_data;
 	de->inode = 0;
-	de->rec_len = cpu_to_le16(sb->s_blocksize);
+	if (ocfs2_dir_has_trailer(dir)) {
+		de->rec_len = cpu_to_le16(ocfs2_dir_trailer_blk_off(sb));
+		ocfs2_init_dir_trailer(dir, new_bh);
+	} else {
+		de->rec_len = cpu_to_le16(sb->s_blocksize);
+	}
 	status = ocfs2_journal_dirty(handle, new_bh);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1620,11 +1828,21 @@ static int ocfs2_find_dir_space_id(struct inode *dir, struct buffer_head *di_bh,
 				   unsigned int *blocks_wanted)
 {
 	int ret;
+	struct super_block *sb = dir->i_sb;
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
 	struct ocfs2_dir_entry *de, *last_de = NULL;
 	char *de_buf, *limit;
 	unsigned long offset = 0;
-	unsigned int rec_len, new_rec_len;
+	unsigned int rec_len, new_rec_len, free_space = dir->i_sb->s_blocksize;
+
+	/*
+	 * This calculates how many free bytes we'd have in block zero, should
+	 * this function force expansion to an extent tree.
+	 */
+	if (ocfs2_supports_dir_trailer(OCFS2_SB(sb)))
+		free_space = ocfs2_dir_trailer_blk_off(sb) - i_size_read(dir);
+	else
+		free_space = dir->i_sb->s_blocksize - i_size_read(dir);
 
 	de_buf = di->id2.i_data.id_data;
 	limit = de_buf + i_size_read(dir);
@@ -1641,6 +1859,11 @@ static int ocfs2_find_dir_space_id(struct inode *dir, struct buffer_head *di_bh,
 			ret = -EEXIST;
 			goto out;
 		}
+		/*
+		 * No need to check for a trailing dirent record here as
+		 * they're not used for inline dirs.
+		 */
+
 		if (ocfs2_dirent_would_fit(de, rec_len)) {
 			/* Ok, we found a spot. Return this bh and let
 			 * the caller actually fill it in. */
@@ -1661,7 +1884,7 @@ static int ocfs2_find_dir_space_id(struct inode *dir, struct buffer_head *di_bh,
 	 * dirent can be found.
 	 */
 	*blocks_wanted = 1;
-	new_rec_len = le16_to_cpu(last_de->rec_len) + (dir->i_sb->s_blocksize - i_size_read(dir));
+	new_rec_len = le16_to_cpu(last_de->rec_len) + free_space;
 	if (new_rec_len < (rec_len + OCFS2_DIR_REC_LEN(last_de->name_len)))
 		*blocks_wanted = 2;
 
@@ -1679,9 +1902,10 @@ static int ocfs2_find_dir_space_el(struct inode *dir, const char *name,
 	struct ocfs2_dir_entry *de;
 	struct super_block *sb = dir->i_sb;
 	int status;
+	int blocksize = dir->i_sb->s_blocksize;
 
-	bh = ocfs2_bread(dir, 0, &status, 0);
-	if (!bh) {
+	status = ocfs2_read_dir_block(dir, 0, &bh, 0);
+	if (status) {
 		mlog_errno(status);
 		goto bail;
 	}
@@ -1702,11 +1926,10 @@ static int ocfs2_find_dir_space_el(struct inode *dir, const char *name,
 				status = -ENOSPC;
 				goto bail;
 			}
-			bh = ocfs2_bread(dir,
-					 offset >> sb->s_blocksize_bits,
-					 &status,
-					 0);
-			if (!bh) {
+			status = ocfs2_read_dir_block(dir,
+					     offset >> sb->s_blocksize_bits,
+					     &bh, 0);
+			if (status) {
 				mlog_errno(status);
 				goto bail;
 			}
@@ -1721,6 +1944,11 @@ static int ocfs2_find_dir_space_el(struct inode *dir, const char *name,
 			status = -EEXIST;
 			goto bail;
 		}
+
+		if (ocfs2_skip_dir_trailer(dir, de, offset % blocksize,
+					   blocksize))
+			goto next;
+
 		if (ocfs2_dirent_would_fit(de, rec_len)) {
 			/* Ok, we found a spot. Return this bh and let
 			 * the caller actually fill it in. */
@@ -1729,6 +1957,7 @@ static int ocfs2_find_dir_space_el(struct inode *dir, const char *name,
 			status = 0;
 			goto bail;
 		}
+next:
 		offset += le16_to_cpu(de->rec_len);
 		de = (struct ocfs2_dir_entry *)((char *) de + le16_to_cpu(de->rec_len));
 	}
diff --git a/fs/ocfs2/dir.h b/fs/ocfs2/dir.h
index ce48b9080d8..c511e2e18e9 100644
--- a/fs/ocfs2/dir.h
+++ b/fs/ocfs2/dir.h
@@ -83,4 +83,6 @@ int ocfs2_fill_new_dir(struct ocfs2_super *osb,
 		       struct buffer_head *fe_bh,
 		       struct ocfs2_alloc_context *data_ac);
 
+struct ocfs2_dir_block_trailer *ocfs2_dir_trailer_from_size(int blocksize,
+							    void *data);
 #endif /* OCFS2_DIR_H */
diff --git a/fs/ocfs2/dlm/dlmast.c b/fs/ocfs2/dlm/dlmast.c
index 644bee55d8b..d07ddbe4b28 100644
--- a/fs/ocfs2/dlm/dlmast.c
+++ b/fs/ocfs2/dlm/dlmast.c
@@ -275,6 +275,7 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data,
 	struct list_head *iter, *head=NULL;
 	u64 cookie;
 	u32 flags;
+	u8 node;
 
 	if (!dlm_grab(dlm)) {
 		dlm_error(DLM_REJECTED);
@@ -286,18 +287,21 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data,
 
 	name = past->name;
 	locklen = past->namelen;
-	cookie = be64_to_cpu(past->cookie);
+	cookie = past->cookie;
 	flags = be32_to_cpu(past->flags);
+	node = past->node_idx;
 
 	if (locklen > DLM_LOCKID_NAME_MAX) {
 		ret = DLM_IVBUFLEN;
-		mlog(ML_ERROR, "Invalid name length in proxy ast handler!\n");
+		mlog(ML_ERROR, "Invalid name length (%d) in proxy ast "
+		     "handler!\n", locklen);
 		goto leave;
 	}
 
 	if ((flags & (LKM_PUT_LVB|LKM_GET_LVB)) ==
 	     (LKM_PUT_LVB|LKM_GET_LVB)) {
-		mlog(ML_ERROR, "both PUT and GET lvb specified\n");
+		mlog(ML_ERROR, "Both PUT and GET lvb specified, (0x%x)\n",
+		     flags);
 		ret = DLM_BADARGS;
 		goto leave;
 	}
@@ -310,22 +314,21 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data,
 	if (past->type != DLM_AST &&
 	    past->type != DLM_BAST) {
 		mlog(ML_ERROR, "Unknown ast type! %d, cookie=%u:%llu"
-		     "name=%.*s\n", past->type, 
-		     dlm_get_lock_cookie_node(cookie),
-		     dlm_get_lock_cookie_seq(cookie),
-		     locklen, name);
+		     "name=%.*s, node=%u\n", past->type,
+		     dlm_get_lock_cookie_node(be64_to_cpu(cookie)),
+		     dlm_get_lock_cookie_seq(be64_to_cpu(cookie)),
+		     locklen, name, node);
 		ret = DLM_IVLOCKID;
 		goto leave;
 	}
 
 	res = dlm_lookup_lockres(dlm, name, locklen);
 	if (!res) {
-		mlog(0, "got %sast for unknown lockres! "
-		     "cookie=%u:%llu, name=%.*s, namelen=%u\n",
-		     past->type == DLM_AST ? "" : "b",
-		     dlm_get_lock_cookie_node(cookie),
-		     dlm_get_lock_cookie_seq(cookie),
-		     locklen, name, locklen);
+		mlog(0, "Got %sast for unknown lockres! cookie=%u:%llu, "
+		     "name=%.*s, node=%u\n", (past->type == DLM_AST ? "" : "b"),
+		     dlm_get_lock_cookie_node(be64_to_cpu(cookie)),
+		     dlm_get_lock_cookie_seq(be64_to_cpu(cookie)),
+		     locklen, name, node);
 		ret = DLM_IVLOCKID;
 		goto leave;
 	}
@@ -337,12 +340,12 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data,
 
 	spin_lock(&res->spinlock);
 	if (res->state & DLM_LOCK_RES_RECOVERING) {
-		mlog(0, "responding with DLM_RECOVERING!\n");
+		mlog(0, "Responding with DLM_RECOVERING!\n");
 		ret = DLM_RECOVERING;
 		goto unlock_out;
 	}
 	if (res->state & DLM_LOCK_RES_MIGRATING) {
-		mlog(0, "responding with DLM_MIGRATING!\n");
+		mlog(0, "Responding with DLM_MIGRATING!\n");
 		ret = DLM_MIGRATING;
 		goto unlock_out;
 	}
@@ -351,7 +354,7 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data,
 	lock = NULL;
 	list_for_each(iter, head) {
 		lock = list_entry (iter, struct dlm_lock, list);
-		if (be64_to_cpu(lock->ml.cookie) == cookie)
+		if (lock->ml.cookie == cookie)
 			goto do_ast;
 	}
 
@@ -363,15 +366,15 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data,
 
 	list_for_each(iter, head) {
 		lock = list_entry (iter, struct dlm_lock, list);
-		if (be64_to_cpu(lock->ml.cookie) == cookie)
+		if (lock->ml.cookie == cookie)
 			goto do_ast;
 	}
 
-	mlog(0, "got %sast for unknown lock!  cookie=%u:%llu, "
-	     "name=%.*s, namelen=%u\n", past->type == DLM_AST ? "" : "b", 
-	     dlm_get_lock_cookie_node(cookie),
-	     dlm_get_lock_cookie_seq(cookie),
-	     locklen, name, locklen);
+	mlog(0, "Got %sast for unknown lock! cookie=%u:%llu, name=%.*s, "
+	     "node=%u\n", past->type == DLM_AST ? "" : "b",
+	     dlm_get_lock_cookie_node(be64_to_cpu(cookie)),
+	     dlm_get_lock_cookie_seq(be64_to_cpu(cookie)),
+	     locklen, name, node);
 
 	ret = DLM_NORMAL;
 unlock_out:
@@ -383,8 +386,8 @@ do_ast:
 	if (past->type == DLM_AST) {
 		/* do not alter lock refcount.  switching lists. */
 		list_move_tail(&lock->list, &res->granted);
-		mlog(0, "ast: adding to granted list... type=%d, "
-			  "convert_type=%d\n", lock->ml.type, lock->ml.convert_type);
+		mlog(0, "ast: Adding to granted list... type=%d, "
+		     "convert_type=%d\n", lock->ml.type, lock->ml.convert_type);
 		if (lock->ml.convert_type != LKM_IVMODE) {
 			lock->ml.type = lock->ml.convert_type;
 			lock->ml.convert_type = LKM_IVMODE;
@@ -408,7 +411,6 @@ do_ast:
 		dlm_do_local_bast(dlm, res, lock, past->blocked_type);
 
 leave:
-
 	if (res)
 		dlm_lockres_put(res);
 
diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h
index d5a86fb81a4..bb53714813a 100644
--- a/fs/ocfs2/dlm/dlmcommon.h
+++ b/fs/ocfs2/dlm/dlmcommon.h
@@ -140,6 +140,7 @@ struct dlm_ctxt
 	unsigned int purge_count;
 	spinlock_t spinlock;
 	spinlock_t ast_lock;
+	spinlock_t track_lock;
 	char *name;
 	u8 node_num;
 	u32 key;
@@ -316,6 +317,8 @@ struct dlm_lock_resource
 	 * put on a list for the dlm thread to run. */
 	unsigned long    last_used;
 
+	struct dlm_ctxt *dlm;
+
 	unsigned migration_pending:1;
 	atomic_t asts_reserved;
 	spinlock_t spinlock;
diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c
index 1b81dcba175..b32f60a5acf 100644
--- a/fs/ocfs2/dlm/dlmdebug.c
+++ b/fs/ocfs2/dlm/dlmdebug.c
@@ -630,43 +630,38 @@ static void *lockres_seq_start(struct seq_file *m, loff_t *pos)
 {
 	struct debug_lockres *dl = m->private;
 	struct dlm_ctxt *dlm = dl->dl_ctxt;
+	struct dlm_lock_resource *oldres = dl->dl_res;
 	struct dlm_lock_resource *res = NULL;
+	struct list_head *track_list;
 
-	spin_lock(&dlm->spinlock);
+	spin_lock(&dlm->track_lock);
+	if (oldres)
+		track_list = &oldres->tracking;
+	else
+		track_list = &dlm->tracking_list;
 
-	if (dl->dl_res) {
-		list_for_each_entry(res, &dl->dl_res->tracking, tracking) {
-			if (dl->dl_res) {
-				dlm_lockres_put(dl->dl_res);
-				dl->dl_res = NULL;
-			}
-			if (&res->tracking == &dlm->tracking_list) {
-				mlog(0, "End of list found, %p\n", res);
-				dl = NULL;
-				break;
-			}
+	list_for_each_entry(res, track_list, tracking) {
+		if (&res->tracking == &dlm->tracking_list)
+			res = NULL;
+		else
 			dlm_lockres_get(res);
-			dl->dl_res = res;
-			break;
-		}
-	} else {
-		if (!list_empty(&dlm->tracking_list)) {
-			list_for_each_entry(res, &dlm->tracking_list, tracking)
-				break;
-			dlm_lockres_get(res);
-			dl->dl_res = res;
-		} else
-			dl = NULL;
+		break;
 	}
+	spin_unlock(&dlm->track_lock);
 
-	if (dl) {
-		spin_lock(&dl->dl_res->spinlock);
-		dump_lockres(dl->dl_res, dl->dl_buf, dl->dl_len - 1);
-		spin_unlock(&dl->dl_res->spinlock);
-	}
+	if (oldres)
+		dlm_lockres_put(oldres);
 
-	spin_unlock(&dlm->spinlock);
+	dl->dl_res = res;
+
+	if (res) {
+		spin_lock(&res->spinlock);
+		dump_lockres(res, dl->dl_buf, dl->dl_len - 1);
+		spin_unlock(&res->spinlock);
+	} else
+		dl = NULL;
 
+	/* passed to seq_show */
 	return dl;
 }
 
diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c
index 63f8125824e..d8d578f4561 100644
--- a/fs/ocfs2/dlm/dlmdomain.c
+++ b/fs/ocfs2/dlm/dlmdomain.c
@@ -1550,6 +1550,7 @@ static struct dlm_ctxt *dlm_alloc_ctxt(const char *domain,
 	spin_lock_init(&dlm->spinlock);
 	spin_lock_init(&dlm->master_lock);
 	spin_lock_init(&dlm->ast_lock);
+	spin_lock_init(&dlm->track_lock);
 	INIT_LIST_HEAD(&dlm->list);
 	INIT_LIST_HEAD(&dlm->dirty_list);
 	INIT_LIST_HEAD(&dlm->reco.resources);
diff --git a/fs/ocfs2/dlm/dlmfs.c b/fs/ocfs2/dlm/dlmfs.c
index 6f7a77d5402..1c9efb406a9 100644
--- a/fs/ocfs2/dlm/dlmfs.c
+++ b/fs/ocfs2/dlm/dlmfs.c
@@ -341,7 +341,6 @@ static struct inode *dlmfs_get_root_inode(struct super_block *sb)
 		inode->i_mode = mode;
 		inode->i_uid = current_fsuid();
 		inode->i_gid = current_fsgid();
-		inode->i_blocks = 0;
 		inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info;
 		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 		inc_nlink(inode);
@@ -367,7 +366,6 @@ static struct inode *dlmfs_get_inode(struct inode *parent,
 	inode->i_mode = mode;
 	inode->i_uid = current_fsuid();
 	inode->i_gid = current_fsgid();
-	inode->i_blocks = 0;
 	inode->i_mapping->backing_dev_info = &dlmfs_backing_dev_info;
 	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 44f87caf368..54e182a27ca 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -505,8 +505,10 @@ void dlm_change_lockres_owner(struct dlm_ctxt *dlm,
 static void dlm_lockres_release(struct kref *kref)
 {
 	struct dlm_lock_resource *res;
+	struct dlm_ctxt *dlm;
 
 	res = container_of(kref, struct dlm_lock_resource, refs);
+	dlm = res->dlm;
 
 	/* This should not happen -- all lockres' have a name
 	 * associated with them at init time. */
@@ -515,6 +517,7 @@ static void dlm_lockres_release(struct kref *kref)
 	mlog(0, "destroying lockres %.*s\n", res->lockname.len,
 	     res->lockname.name);
 
+	spin_lock(&dlm->track_lock);
 	if (!list_empty(&res->tracking))
 		list_del_init(&res->tracking);
 	else {
@@ -522,6 +525,9 @@ static void dlm_lockres_release(struct kref *kref)
 		     res->lockname.len, res->lockname.name);
 		dlm_print_one_lock_resource(res);
 	}
+	spin_unlock(&dlm->track_lock);
+
+	dlm_put(dlm);
 
 	if (!hlist_unhashed(&res->hash_node) ||
 	    !list_empty(&res->granted) ||
@@ -595,6 +601,10 @@ static void dlm_init_lockres(struct dlm_ctxt *dlm,
 	res->migration_pending = 0;
 	res->inflight_locks = 0;
 
+	/* put in dlm_lockres_release */
+	dlm_grab(dlm);
+	res->dlm = dlm;
+
 	kref_init(&res->refs);
 
 	/* just for consistency */
@@ -722,14 +732,21 @@ lookup:
 	if (tmpres) {
 		int dropping_ref = 0;
 
+		spin_unlock(&dlm->spinlock);
+
 		spin_lock(&tmpres->spinlock);
+		/* We wait for the other thread that is mastering the resource */
+		if (tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN) {
+			__dlm_wait_on_lockres(tmpres);
+			BUG_ON(tmpres->owner == DLM_LOCK_RES_OWNER_UNKNOWN);
+		}
+
 		if (tmpres->owner == dlm->node_num) {
 			BUG_ON(tmpres->state & DLM_LOCK_RES_DROPPING_REF);
 			dlm_lockres_grab_inflight_ref(dlm, tmpres);
 		} else if (tmpres->state & DLM_LOCK_RES_DROPPING_REF)
 			dropping_ref = 1;
 		spin_unlock(&tmpres->spinlock);
-		spin_unlock(&dlm->spinlock);
 
 		/* wait until done messaging the master, drop our ref to allow
 		 * the lockres to be purged, start over. */
@@ -2949,7 +2966,7 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm,
 				  struct dlm_node_iter *iter)
 {
 	struct dlm_migrate_request migrate;
-	int ret, status = 0;
+	int ret, skip, status = 0;
 	int nodenum;
 
 	memset(&migrate, 0, sizeof(migrate));
@@ -2966,12 +2983,27 @@ static int dlm_do_migrate_request(struct dlm_ctxt *dlm,
 		    nodenum == new_master)
 			continue;
 
+		/* We could race exit domain. If exited, skip. */
+		spin_lock(&dlm->spinlock);
+		skip = (!test_bit(nodenum, dlm->domain_map));
+		spin_unlock(&dlm->spinlock);
+		if (skip) {
+			clear_bit(nodenum, iter->node_map);
+			continue;
+		}
+
 		ret = o2net_send_message(DLM_MIGRATE_REQUEST_MSG, dlm->key,
 					 &migrate, sizeof(migrate), nodenum,
 					 &status);
-		if (ret < 0)
-			mlog_errno(ret);
-		else if (status < 0) {
+		if (ret < 0) {
+			mlog(0, "migrate_request returned %d!\n", ret);
+			if (!dlm_is_host_down(ret)) {
+				mlog(ML_ERROR, "unhandled error=%d!\n", ret);
+				BUG();
+			}
+			clear_bit(nodenum, iter->node_map);
+			ret = 0;
+		} else if (status < 0) {
 			mlog(0, "migrate request (node %u) returned %d!\n",
 			     nodenum, status);
 			ret = status;
diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c
index 4060bb328bc..d1295203029 100644
--- a/fs/ocfs2/dlm/dlmthread.c
+++ b/fs/ocfs2/dlm/dlmthread.c
@@ -181,7 +181,8 @@ static int dlm_purge_lockres(struct dlm_ctxt *dlm,
 
 		spin_lock(&res->spinlock);
 		/* This ensures that clear refmap is sent after the set */
-		__dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG);
+		__dlm_wait_on_lockres_flags(res, (DLM_LOCK_RES_SETREF_INPROG |
+						  DLM_LOCK_RES_MIGRATING));
 		spin_unlock(&res->spinlock);
 
 		/* clear our bit from the master's refmap, ignore errors */
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 6e6cc0a2e5f..b0c4cadd4c4 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -32,6 +32,7 @@
 #include <linux/debugfs.h>
 #include <linux/seq_file.h>
 #include <linux/time.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_DLM_GLUE
 #include <cluster/masklog.h>
@@ -51,6 +52,7 @@
 #include "slot_map.h"
 #include "super.h"
 #include "uptodate.h"
+#include "quota.h"
 
 #include "buffer_head_io.h"
 
@@ -68,6 +70,7 @@ struct ocfs2_mask_waiter {
 static struct ocfs2_super *ocfs2_get_dentry_osb(struct ocfs2_lock_res *lockres);
 static struct ocfs2_super *ocfs2_get_inode_osb(struct ocfs2_lock_res *lockres);
 static struct ocfs2_super *ocfs2_get_file_osb(struct ocfs2_lock_res *lockres);
+static struct ocfs2_super *ocfs2_get_qinfo_osb(struct ocfs2_lock_res *lockres);
 
 /*
  * Return value from ->downconvert_worker functions.
@@ -102,6 +105,7 @@ static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
 static void ocfs2_dentry_post_unlock(struct ocfs2_super *osb,
 				     struct ocfs2_lock_res *lockres);
 
+static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres);
 
 #define mlog_meta_lvb(__level, __lockres) ocfs2_dump_meta_lvb_info(__level, __PRETTY_FUNCTION__, __LINE__, __lockres)
 
@@ -111,8 +115,7 @@ static void ocfs2_dump_meta_lvb_info(u64 level,
 				     unsigned int line,
 				     struct ocfs2_lock_res *lockres)
 {
-	struct ocfs2_meta_lvb *lvb =
-		(struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb);
+	struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
 
 	mlog(level, "LVB information for %s (called from %s:%u):\n",
 	     lockres->l_name, function, line);
@@ -258,6 +261,12 @@ static struct ocfs2_lock_res_ops ocfs2_flock_lops = {
 	.flags		= 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_qinfo_lops = {
+	.set_lvb	= ocfs2_set_qinfo_lvb,
+	.get_osb	= ocfs2_get_qinfo_osb,
+	.flags		= LOCK_TYPE_REQUIRES_REFRESH | LOCK_TYPE_USES_LVB,
+};
+
 static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres)
 {
 	return lockres->l_type == OCFS2_LOCK_TYPE_META ||
@@ -279,6 +288,13 @@ static inline struct ocfs2_dentry_lock *ocfs2_lock_res_dl(struct ocfs2_lock_res
 	return (struct ocfs2_dentry_lock *)lockres->l_priv;
 }
 
+static inline struct ocfs2_mem_dqinfo *ocfs2_lock_res_qinfo(struct ocfs2_lock_res *lockres)
+{
+	BUG_ON(lockres->l_type != OCFS2_LOCK_TYPE_QINFO);
+
+	return (struct ocfs2_mem_dqinfo *)lockres->l_priv;
+}
+
 static inline struct ocfs2_super *ocfs2_get_lockres_osb(struct ocfs2_lock_res *lockres)
 {
 	if (lockres->l_ops->get_osb)
@@ -507,6 +523,13 @@ static struct ocfs2_super *ocfs2_get_inode_osb(struct ocfs2_lock_res *lockres)
 	return OCFS2_SB(inode->i_sb);
 }
 
+static struct ocfs2_super *ocfs2_get_qinfo_osb(struct ocfs2_lock_res *lockres)
+{
+	struct ocfs2_mem_dqinfo *info = lockres->l_priv;
+
+	return OCFS2_SB(info->dqi_gi.dqi_sb);
+}
+
 static struct ocfs2_super *ocfs2_get_file_osb(struct ocfs2_lock_res *lockres)
 {
 	struct ocfs2_file_private *fp = lockres->l_priv;
@@ -609,6 +632,17 @@ void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 	lockres->l_flags |= OCFS2_LOCK_NOCACHE;
 }
 
+void ocfs2_qinfo_lock_res_init(struct ocfs2_lock_res *lockres,
+			       struct ocfs2_mem_dqinfo *info)
+{
+	ocfs2_lock_res_init_once(lockres);
+	ocfs2_build_lock_name(OCFS2_LOCK_TYPE_QINFO, info->dqi_gi.dqi_type,
+			      0, lockres->l_name);
+	ocfs2_lock_res_init_common(OCFS2_SB(info->dqi_gi.dqi_sb), lockres,
+				   OCFS2_LOCK_TYPE_QINFO, &ocfs2_qinfo_lops,
+				   info);
+}
+
 void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
 {
 	mlog_entry_void();
@@ -1290,7 +1324,7 @@ again:
 			goto out;
 		}
 
-		mlog(0, "lock %s, successfull return from ocfs2_dlm_lock\n",
+		mlog(0, "lock %s, successful return from ocfs2_dlm_lock\n",
 		     lockres->l_name);
 
 		/* At this point we've gone inside the dlm and need to
@@ -1829,7 +1863,7 @@ static void __ocfs2_stuff_meta_lvb(struct inode *inode)
 
 	mlog_entry_void();
 
-	lvb = (struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
 
 	/*
 	 * Invalidate the LVB of a deleted inode - this way other
@@ -1881,7 +1915,7 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode)
 
 	mlog_meta_lvb(0, lockres);
 
-	lvb = (struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
 
 	/* We're safe here without the lockres lock... */
 	spin_lock(&oi->ip_lock);
@@ -1916,8 +1950,7 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode)
 static inline int ocfs2_meta_lvb_is_trustable(struct inode *inode,
 					      struct ocfs2_lock_res *lockres)
 {
-	struct ocfs2_meta_lvb *lvb =
-		(struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb);
+	struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
 
 	if (lvb->lvb_version == OCFS2_LVB_VERSION
 	    && be32_to_cpu(lvb->lvb_igeneration) == inode->i_generation)
@@ -2024,7 +2057,7 @@ static int ocfs2_inode_lock_update(struct inode *inode,
 	} else {
 		/* Boo, we have to go to disk. */
 		/* read bh, cast, ocfs2_refresh_inode */
-		status = ocfs2_read_block(inode, oi->ip_blkno, bh);
+		status = ocfs2_read_inode_block(inode, bh);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail_refresh;
@@ -2032,18 +2065,14 @@ static int ocfs2_inode_lock_update(struct inode *inode,
 		fe = (struct ocfs2_dinode *) (*bh)->b_data;
 
 		/* This is a good chance to make sure we're not
-		 * locking an invalid object.
+		 * locking an invalid object.  ocfs2_read_inode_block()
+		 * already checked that the inode block is sane.
 		 *
 		 * We bug on a stale inode here because we checked
 		 * above whether it was wiped from disk. The wiping
 		 * node provides a guarantee that we receive that
 		 * message and can mark the inode before dropping any
 		 * locks associated with it. */
-		if (!OCFS2_IS_VALID_DINODE(fe)) {
-			OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, fe);
-			status = -EIO;
-			goto bail_refresh;
-		}
 		mlog_bug_on_msg(inode->i_generation !=
 				le32_to_cpu(fe->i_generation),
 				"Invalid dinode %llu disk generation: %u "
@@ -2085,7 +2114,7 @@ static int ocfs2_assign_bh(struct inode *inode,
 		return 0;
 	}
 
-	status = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, ret_bh);
+	status = ocfs2_read_inode_block(inode, ret_bh);
 	if (status < 0)
 		mlog_errno(status);
 
@@ -2922,7 +2951,7 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb,
 		ocfs2_dlm_dump_lksb(&lockres->l_lksb);
 		BUG();
 	}
-	mlog(0, "lock %s, successfull return from ocfs2_dlm_unlock\n",
+	mlog(0, "lock %s, successful return from ocfs2_dlm_unlock\n",
 	     lockres->l_name);
 
 	ocfs2_wait_on_busy_lock(lockres);
@@ -3449,6 +3478,117 @@ static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
 	return UNBLOCK_CONTINUE_POST;
 }
 
+static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres)
+{
+	struct ocfs2_qinfo_lvb *lvb;
+	struct ocfs2_mem_dqinfo *oinfo = ocfs2_lock_res_qinfo(lockres);
+	struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
+					    oinfo->dqi_gi.dqi_type);
+
+	mlog_entry_void();
+
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb->lvb_version = OCFS2_QINFO_LVB_VERSION;
+	lvb->lvb_bgrace = cpu_to_be32(info->dqi_bgrace);
+	lvb->lvb_igrace = cpu_to_be32(info->dqi_igrace);
+	lvb->lvb_syncms = cpu_to_be32(oinfo->dqi_syncms);
+	lvb->lvb_blocks = cpu_to_be32(oinfo->dqi_gi.dqi_blocks);
+	lvb->lvb_free_blk = cpu_to_be32(oinfo->dqi_gi.dqi_free_blk);
+	lvb->lvb_free_entry = cpu_to_be32(oinfo->dqi_gi.dqi_free_entry);
+
+	mlog_exit_void();
+}
+
+void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex)
+{
+	struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
+	struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+	mlog_entry_void();
+	if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
+		ocfs2_cluster_unlock(osb, lockres, level);
+	mlog_exit_void();
+}
+
+static int ocfs2_refresh_qinfo(struct ocfs2_mem_dqinfo *oinfo)
+{
+	struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
+					    oinfo->dqi_gi.dqi_type);
+	struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
+	struct ocfs2_qinfo_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	struct buffer_head *bh = NULL;
+	struct ocfs2_global_disk_dqinfo *gdinfo;
+	int status = 0;
+
+	if (lvb->lvb_version == OCFS2_QINFO_LVB_VERSION) {
+		info->dqi_bgrace = be32_to_cpu(lvb->lvb_bgrace);
+		info->dqi_igrace = be32_to_cpu(lvb->lvb_igrace);
+		oinfo->dqi_syncms = be32_to_cpu(lvb->lvb_syncms);
+		oinfo->dqi_gi.dqi_blocks = be32_to_cpu(lvb->lvb_blocks);
+		oinfo->dqi_gi.dqi_free_blk = be32_to_cpu(lvb->lvb_free_blk);
+		oinfo->dqi_gi.dqi_free_entry =
+					be32_to_cpu(lvb->lvb_free_entry);
+	} else {
+		status = ocfs2_read_quota_block(oinfo->dqi_gqinode, 0, &bh);
+		if (status) {
+			mlog_errno(status);
+			goto bail;
+		}
+		gdinfo = (struct ocfs2_global_disk_dqinfo *)
+					(bh->b_data + OCFS2_GLOBAL_INFO_OFF);
+		info->dqi_bgrace = le32_to_cpu(gdinfo->dqi_bgrace);
+		info->dqi_igrace = le32_to_cpu(gdinfo->dqi_igrace);
+		oinfo->dqi_syncms = le32_to_cpu(gdinfo->dqi_syncms);
+		oinfo->dqi_gi.dqi_blocks = le32_to_cpu(gdinfo->dqi_blocks);
+		oinfo->dqi_gi.dqi_free_blk = le32_to_cpu(gdinfo->dqi_free_blk);
+		oinfo->dqi_gi.dqi_free_entry =
+					le32_to_cpu(gdinfo->dqi_free_entry);
+		brelse(bh);
+		ocfs2_track_lock_refresh(lockres);
+	}
+
+bail:
+	return status;
+}
+
+/* Lock quota info, this function expects at least shared lock on the quota file
+ * so that we can safely refresh quota info from disk. */
+int ocfs2_qinfo_lock(struct ocfs2_mem_dqinfo *oinfo, int ex)
+{
+	struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
+	struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	int status = 0;
+
+	mlog_entry_void();
+
+	/* On RO devices, locking really isn't needed... */
+	if (ocfs2_is_hard_readonly(osb)) {
+		if (ex)
+			status = -EROFS;
+		goto bail;
+	}
+	if (ocfs2_mount_local(osb))
+		goto bail;
+
+	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
+	if (status < 0) {
+		mlog_errno(status);
+		goto bail;
+	}
+	if (!ocfs2_should_refresh_lock_res(lockres))
+		goto bail;
+	/* OK, we have the lock but we need to refresh the quota info */
+	status = ocfs2_refresh_qinfo(oinfo);
+	if (status)
+		ocfs2_qinfo_unlock(oinfo, ex);
+	ocfs2_complete_lock_res_refresh(lockres, status);
+bail:
+	mlog_exit(status);
+	return status;
+}
+
 /*
  * This is the filesystem locking protocol.  It provides the lock handling
  * hooks for the underlying DLM.  It has a maximum version number.
diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h
index 2bb01f09c1b..3f8d9986b8e 100644
--- a/fs/ocfs2/dlmglue.h
+++ b/fs/ocfs2/dlmglue.h
@@ -49,6 +49,19 @@ struct ocfs2_meta_lvb {
 	__be32       lvb_reserved2;
 };
 
+#define OCFS2_QINFO_LVB_VERSION 1
+
+struct ocfs2_qinfo_lvb {
+	__u8	lvb_version;
+	__u8	lvb_reserved[3];
+	__be32	lvb_bgrace;
+	__be32	lvb_igrace;
+	__be32	lvb_syncms;
+	__be32	lvb_blocks;
+	__be32	lvb_free_blk;
+	__be32	lvb_free_entry;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY	(0x01)
@@ -69,6 +82,9 @@ void ocfs2_dentry_lock_res_init(struct ocfs2_dentry_lock *dl,
 struct ocfs2_file_private;
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 			      struct ocfs2_file_private *fp);
+struct ocfs2_mem_dqinfo;
+void ocfs2_qinfo_lock_res_init(struct ocfs2_lock_res *lockres,
+                               struct ocfs2_mem_dqinfo *info);
 void ocfs2_lock_res_free(struct ocfs2_lock_res *res);
 int ocfs2_create_new_inode_locks(struct inode *inode);
 int ocfs2_drop_inode_locks(struct inode *inode);
@@ -103,6 +119,9 @@ int ocfs2_dentry_lock(struct dentry *dentry, int ex);
 void ocfs2_dentry_unlock(struct dentry *dentry, int ex);
 int ocfs2_file_lock(struct file *file, int ex, int trylock);
 void ocfs2_file_unlock(struct file *file);
+int ocfs2_qinfo_lock(struct ocfs2_mem_dqinfo *oinfo, int ex);
+void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex);
+
 
 void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres);
 void ocfs2_simple_drop_lockres(struct ocfs2_super *osb,
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index 2baedac5823..f2bb1a04d25 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -293,7 +293,7 @@ static int ocfs2_last_eb_is_empty(struct inode *inode,
 	struct ocfs2_extent_block *eb;
 	struct ocfs2_extent_list *el;
 
-	ret = ocfs2_read_block(inode, last_eb_blk, &eb_bh);
+	ret = ocfs2_read_extent_block(inode, last_eb_blk, &eb_bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -302,12 +302,6 @@ static int ocfs2_last_eb_is_empty(struct inode *inode,
 	eb = (struct ocfs2_extent_block *) eb_bh->b_data;
 	el = &eb->h_list;
 
-	if (!OCFS2_IS_VALID_EXTENT_BLOCK(eb)) {
-		ret = -EROFS;
-		OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, eb);
-		goto out;
-	}
-
 	if (el->l_tree_depth) {
 		ocfs2_error(inode->i_sb,
 			    "Inode %lu has non zero tree depth in "
@@ -381,23 +375,16 @@ static int ocfs2_figure_hole_clusters(struct inode *inode,
 		if (le64_to_cpu(eb->h_next_leaf_blk) == 0ULL)
 			goto no_more_extents;
 
-		ret = ocfs2_read_block(inode,
-				       le64_to_cpu(eb->h_next_leaf_blk),
-				       &next_eb_bh);
+		ret = ocfs2_read_extent_block(inode,
+					      le64_to_cpu(eb->h_next_leaf_blk),
+					      &next_eb_bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
-		next_eb = (struct ocfs2_extent_block *)next_eb_bh->b_data;
-
-		if (!OCFS2_IS_VALID_EXTENT_BLOCK(next_eb)) {
-			ret = -EROFS;
-			OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, next_eb);
-			goto out;
-		}
 
+		next_eb = (struct ocfs2_extent_block *)next_eb_bh->b_data;
 		el = &next_eb->h_list;
-
 		i = ocfs2_search_for_hole_index(el, v_cluster);
 	}
 
@@ -630,7 +617,7 @@ int ocfs2_get_clusters(struct inode *inode, u32 v_cluster,
 	if (ret == 0)
 		goto out;
 
-	ret = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, &di_bh);
+	ret = ocfs2_read_inode_block(inode, &di_bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -819,3 +806,74 @@ out:
 
 	return ret;
 }
+
+int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr,
+			   struct buffer_head *bhs[], int flags,
+			   int (*validate)(struct super_block *sb,
+					   struct buffer_head *bh))
+{
+	int rc = 0;
+	u64 p_block, p_count;
+	int i, count, done = 0;
+
+	mlog_entry("(inode = %p, v_block = %llu, nr = %d, bhs = %p, "
+		   "flags = %x, validate = %p)\n",
+		   inode, (unsigned long long)v_block, nr, bhs, flags,
+		   validate);
+
+	if (((v_block + nr - 1) << inode->i_sb->s_blocksize_bits) >=
+	    i_size_read(inode)) {
+		BUG_ON(!(flags & OCFS2_BH_READAHEAD));
+		goto out;
+	}
+
+	while (done < nr) {
+		down_read(&OCFS2_I(inode)->ip_alloc_sem);
+		rc = ocfs2_extent_map_get_blocks(inode, v_block + done,
+						 &p_block, &p_count, NULL);
+		up_read(&OCFS2_I(inode)->ip_alloc_sem);
+		if (rc) {
+			mlog_errno(rc);
+			break;
+		}
+
+		if (!p_block) {
+			rc = -EIO;
+			mlog(ML_ERROR,
+			     "Inode #%llu contains a hole at offset %llu\n",
+			     (unsigned long long)OCFS2_I(inode)->ip_blkno,
+			     (unsigned long long)(v_block + done) <<
+			     inode->i_sb->s_blocksize_bits);
+			break;
+		}
+
+		count = nr - done;
+		if (p_count < count)
+			count = p_count;
+
+		/*
+		 * If the caller passed us bhs, they should have come
+		 * from a previous readahead call to this function.  Thus,
+		 * they should have the right b_blocknr.
+		 */
+		for (i = 0; i < count; i++) {
+			if (!bhs[done + i])
+				continue;
+			BUG_ON(bhs[done + i]->b_blocknr != (p_block + i));
+		}
+
+		rc = ocfs2_read_blocks(inode, p_block, count, bhs + done,
+				       flags, validate);
+		if (rc) {
+			mlog_errno(rc);
+			break;
+		}
+		done += count;
+	}
+
+out:
+	mlog_exit(rc);
+	return rc;
+}
+
+
diff --git a/fs/ocfs2/extent_map.h b/fs/ocfs2/extent_map.h
index 1c4aa8b06f3..b7dd9731b46 100644
--- a/fs/ocfs2/extent_map.h
+++ b/fs/ocfs2/extent_map.h
@@ -57,4 +57,28 @@ int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster,
 			     u32 *p_cluster, u32 *num_clusters,
 			     struct ocfs2_extent_list *el);
 
+int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr,
+			   struct buffer_head *bhs[], int flags,
+			   int (*validate)(struct super_block *sb,
+					   struct buffer_head *bh));
+static inline int ocfs2_read_virt_block(struct inode *inode, u64 v_block,
+					struct buffer_head **bh,
+					int (*validate)(struct super_block *sb,
+							struct buffer_head *bh))
+{
+	int status = 0;
+
+	if (bh == NULL) {
+		printk("ocfs2: bh == NULL\n");
+		status = -EINVAL;
+		goto bail;
+	}
+
+	status = ocfs2_read_virt_blocks(inode, v_block, 1, bh, 0, validate);
+
+bail:
+	return status;
+}
+
+
 #endif  /* _EXTENT_MAP_H */
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index e2570a3bc2b..a5887df2cd8 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -35,6 +35,7 @@
 #include <linux/mount.h>
 #include <linux/writeback.h>
 #include <linux/falloc.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_INODE
 #include <cluster/masklog.h>
@@ -56,6 +57,8 @@
 #include "suballoc.h"
 #include "super.h"
 #include "xattr.h"
+#include "acl.h"
+#include "quota.h"
 
 #include "buffer_head_io.h"
 
@@ -253,8 +256,8 @@ int ocfs2_update_inode_atime(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -303,9 +306,9 @@ bail:
 	return status;
 }
 
-static int ocfs2_simple_size_update(struct inode *inode,
-				    struct buffer_head *di_bh,
-				    u64 new_i_size)
+int ocfs2_simple_size_update(struct inode *inode,
+			     struct buffer_head *di_bh,
+			     u64 new_i_size)
 {
 	int ret;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
@@ -350,8 +353,8 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb,
 		goto out;
 	}
 
-	status = ocfs2_journal_access(handle, inode, fe_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, inode, fe_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_commit;
@@ -401,12 +404,9 @@ static int ocfs2_truncate_file(struct inode *inode,
 		   (unsigned long long)OCFS2_I(inode)->ip_blkno,
 		   (unsigned long long)new_i_size);
 
+	/* We trust di_bh because it comes from ocfs2_inode_lock(), which
+	 * already validated it */
 	fe = (struct ocfs2_dinode *) di_bh->b_data;
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, fe);
-		status = -EIO;
-		goto bail;
-	}
 
 	mlog_bug_on_msg(le64_to_cpu(fe->i_size) != i_size_read(inode),
 			"Inode %llu, inode i_size = %lld != di "
@@ -536,6 +536,7 @@ static int __ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
 	enum ocfs2_alloc_restarted why;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_extent_tree et;
+	int did_quota = 0;
 
 	mlog_entry("(clusters_to_add = %u)\n", clusters_to_add);
 
@@ -545,18 +546,12 @@ static int __ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
 	 */
 	BUG_ON(mark_unwritten && !ocfs2_sparse_alloc(osb));
 
-	status = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, &bh);
+	status = ocfs2_read_inode_block(inode, &bh);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
 	}
-
 	fe = (struct ocfs2_dinode *) bh->b_data;
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, fe);
-		status = -EIO;
-		goto leave;
-	}
 
 restart_all:
 	BUG_ON(le32_to_cpu(fe->i_clusters) != OCFS2_I(inode)->ip_clusters);
@@ -585,11 +580,18 @@ restart_all:
 	}
 
 restarted_transaction:
+	if (vfs_dq_alloc_space_nodirty(inode, ocfs2_clusters_to_bytes(osb->sb,
+	    clusters_to_add))) {
+		status = -EDQUOT;
+		goto leave;
+	}
+	did_quota = 1;
+
 	/* reserve a write to the file entry early on - that we if we
 	 * run out of credits in the allocation path, we can still
 	 * update i_size. */
-	status = ocfs2_journal_access(handle, inode, bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, inode, bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -622,6 +624,10 @@ restarted_transaction:
 	spin_lock(&OCFS2_I(inode)->ip_lock);
 	clusters_to_add -= (OCFS2_I(inode)->ip_clusters - prev_clusters);
 	spin_unlock(&OCFS2_I(inode)->ip_lock);
+	/* Release unused quota reservation */
+	vfs_dq_free_space(inode,
+			ocfs2_clusters_to_bytes(osb->sb, clusters_to_add));
+	did_quota = 0;
 
 	if (why != RESTART_NONE && clusters_to_add) {
 		if (why == RESTART_META) {
@@ -654,6 +660,9 @@ restarted_transaction:
 	     OCFS2_I(inode)->ip_clusters, (long long)i_size_read(inode));
 
 leave:
+	if (status < 0 && did_quota)
+		vfs_dq_free_space(inode,
+			ocfs2_clusters_to_bytes(osb->sb, clusters_to_add));
 	if (handle) {
 		ocfs2_commit_trans(osb, handle);
 		handle = NULL;
@@ -885,6 +894,9 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 	struct ocfs2_super *osb = OCFS2_SB(sb);
 	struct buffer_head *bh = NULL;
 	handle_t *handle = NULL;
+	int locked[MAXQUOTAS] = {0, 0};
+	int credits, qtype;
+	struct ocfs2_mem_dqinfo *oinfo;
 
 	mlog_entry("(0x%p, '%.*s')\n", dentry,
 	           dentry->d_name.len, dentry->d_name.name);
@@ -955,11 +967,47 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
-	if (IS_ERR(handle)) {
-		status = PTR_ERR(handle);
-		mlog_errno(status);
-		goto bail_unlock;
+	if ((attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
+	    (attr->ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
+		credits = OCFS2_INODE_UPDATE_CREDITS;
+		if (attr->ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid
+		    && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+		    OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
+			oinfo = sb_dqinfo(sb, USRQUOTA)->dqi_priv;
+			status = ocfs2_lock_global_qf(oinfo, 1);
+			if (status < 0)
+				goto bail_unlock;
+			credits += ocfs2_calc_qinit_credits(sb, USRQUOTA) +
+				ocfs2_calc_qdel_credits(sb, USRQUOTA);
+			locked[USRQUOTA] = 1;
+		}
+		if (attr->ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid
+		    && OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+		    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
+			oinfo = sb_dqinfo(sb, GRPQUOTA)->dqi_priv;
+			status = ocfs2_lock_global_qf(oinfo, 1);
+			if (status < 0)
+				goto bail_unlock;
+			credits += ocfs2_calc_qinit_credits(sb, GRPQUOTA) +
+				   ocfs2_calc_qdel_credits(sb, GRPQUOTA);
+			locked[GRPQUOTA] = 1;
+		}
+		handle = ocfs2_start_trans(osb, credits);
+		if (IS_ERR(handle)) {
+			status = PTR_ERR(handle);
+			mlog_errno(status);
+			goto bail_unlock;
+		}
+		status = vfs_dq_transfer(inode, attr) ? -EDQUOT : 0;
+		if (status < 0)
+			goto bail_commit;
+	} else {
+		handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
+		if (IS_ERR(handle)) {
+			status = PTR_ERR(handle);
+			mlog_errno(status);
+			goto bail_unlock;
+		}
 	}
 
 	/*
@@ -982,6 +1030,12 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
 bail_commit:
 	ocfs2_commit_trans(osb, handle);
 bail_unlock:
+	for (qtype = 0; qtype < MAXQUOTAS; qtype++) {
+		if (!locked[qtype])
+			continue;
+		oinfo = sb_dqinfo(sb, qtype)->dqi_priv;
+		ocfs2_unlock_global_qf(oinfo, 1);
+	}
 	ocfs2_inode_unlock(inode, 1);
 bail_unlock_rw:
 	if (size_change)
@@ -989,6 +1043,12 @@ bail_unlock_rw:
 bail:
 	brelse(bh);
 
+	if (!status && attr->ia_valid & ATTR_MODE) {
+		status = ocfs2_acl_chmod(inode);
+		if (status < 0)
+			mlog_errno(status);
+	}
+
 	mlog_exit(status);
 	return status;
 }
@@ -1035,7 +1095,7 @@ int ocfs2_permission(struct inode *inode, int mask)
 		goto out;
 	}
 
-	ret = generic_permission(inode, mask, NULL);
+	ret = generic_permission(inode, mask, ocfs2_check_acl);
 
 	ocfs2_inode_unlock(inode, 0);
 out:
@@ -1061,8 +1121,8 @@ static int __ocfs2_write_remove_suid(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out_trans;
@@ -1128,9 +1188,8 @@ static int ocfs2_write_remove_suid(struct inode *inode)
 {
 	int ret;
 	struct buffer_head *bh = NULL;
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 
-	ret = ocfs2_read_block(inode, oi->ip_blkno, &bh);
+	ret = ocfs2_read_inode_block(inode, &bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -1156,8 +1215,7 @@ static int ocfs2_allocate_unwritten_extents(struct inode *inode,
 	struct buffer_head *di_bh = NULL;
 
 	if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
-		ret = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno,
-				       &di_bh);
+		ret = ocfs2_read_inode_block(inode, &di_bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -1226,83 +1284,6 @@ out:
 	return ret;
 }
 
-static int __ocfs2_remove_inode_range(struct inode *inode,
-				      struct buffer_head *di_bh,
-				      u32 cpos, u32 phys_cpos, u32 len,
-				      struct ocfs2_cached_dealloc_ctxt *dealloc)
-{
-	int ret;
-	u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	struct inode *tl_inode = osb->osb_tl_inode;
-	handle_t *handle;
-	struct ocfs2_alloc_context *meta_ac = NULL;
-	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
-	struct ocfs2_extent_tree et;
-
-	ocfs2_init_dinode_extent_tree(&et, inode, di_bh);
-
-	ret = ocfs2_lock_allocators(inode, &et, 0, 1, NULL, &meta_ac);
-	if (ret) {
-		mlog_errno(ret);
-		return ret;
-	}
-
-	mutex_lock(&tl_inode->i_mutex);
-
-	if (ocfs2_truncate_log_needs_flush(osb)) {
-		ret = __ocfs2_flush_truncate_log(osb);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto out;
-		}
-	}
-
-	handle = ocfs2_start_trans(osb, OCFS2_REMOVE_EXTENT_CREDITS);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		goto out;
-	}
-
-	ret = ocfs2_journal_access(handle, inode, di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
-	if (ret) {
-		mlog_errno(ret);
-		goto out;
-	}
-
-	ret = ocfs2_remove_extent(inode, &et, cpos, len, handle, meta_ac,
-				  dealloc);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_commit;
-	}
-
-	OCFS2_I(inode)->ip_clusters -= len;
-	di->i_clusters = cpu_to_le32(OCFS2_I(inode)->ip_clusters);
-
-	ret = ocfs2_journal_dirty(handle, di_bh);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_commit;
-	}
-
-	ret = ocfs2_truncate_log_append(osb, handle, phys_blkno, len);
-	if (ret)
-		mlog_errno(ret);
-
-out_commit:
-	ocfs2_commit_trans(osb, handle);
-out:
-	mutex_unlock(&tl_inode->i_mutex);
-
-	if (meta_ac)
-		ocfs2_free_alloc_context(meta_ac);
-
-	return ret;
-}
-
 /*
  * Truncate a byte range, avoiding pages within partial clusters. This
  * preserves those pages for the zeroing code to write to.
@@ -1402,7 +1383,9 @@ static int ocfs2_remove_inode_range(struct inode *inode,
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_cached_dealloc_ctxt dealloc;
 	struct address_space *mapping = inode->i_mapping;
+	struct ocfs2_extent_tree et;
 
+	ocfs2_init_dinode_extent_tree(&et, inode, di_bh);
 	ocfs2_init_dealloc_ctxt(&dealloc);
 
 	if (byte_len == 0)
@@ -1458,9 +1441,9 @@ static int ocfs2_remove_inode_range(struct inode *inode,
 
 		/* Only do work for non-holes */
 		if (phys_cpos != 0) {
-			ret = __ocfs2_remove_inode_range(inode, di_bh, cpos,
-							 phys_cpos, alloc_size,
-							 &dealloc);
+			ret = ocfs2_remove_btree_range(inode, &et, cpos,
+						       phys_cpos, alloc_size,
+						       &dealloc);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -1622,7 +1605,7 @@ int ocfs2_change_file_space(struct file *file, unsigned int cmd,
 			    struct ocfs2_space_resv *sr)
 {
 	struct inode *inode = file->f_path.dentry->d_inode;
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
 	if ((cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) &&
 	    !ocfs2_writes_unwritten_extents(osb))
diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h
index e92382cbca5..172f9fbc9fc 100644
--- a/fs/ocfs2/file.h
+++ b/fs/ocfs2/file.h
@@ -51,6 +51,9 @@ int ocfs2_add_inode_data(struct ocfs2_super *osb,
 			 struct ocfs2_alloc_context *data_ac,
 			 struct ocfs2_alloc_context *meta_ac,
 			 enum ocfs2_alloc_restarted *reason_ret);
+int ocfs2_simple_size_update(struct inode *inode,
+			     struct buffer_head *di_bh,
+			     u64 new_i_size);
 int ocfs2_extend_no_holes(struct inode *inode, u64 new_i_size,
 			  u64 zero_to);
 int ocfs2_setattr(struct dentry *dentry, struct iattr *attr);
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 7aa00d51187..229e707bc05 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -28,6 +28,7 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
+#include <linux/quotaops.h>
 
 #include <asm/byteorder.h>
 
@@ -37,6 +38,7 @@
 #include "ocfs2.h"
 
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dlmglue.h"
 #include "extent_map.h"
 #include "file.h"
@@ -214,12 +216,11 @@ static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
 	return 0;
 }
 
-int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
-		     	 int create_ino)
+void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
+			  int create_ino)
 {
 	struct super_block *sb;
 	struct ocfs2_super *osb;
-	int status = -EINVAL;
 	int use_plocks = 1;
 
 	mlog_entry("(0x%p, size:%llu)\n", inode,
@@ -232,25 +233,17 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 	    ocfs2_mount_local(osb) || !ocfs2_stack_supports_plocks())
 		use_plocks = 0;
 
-	/* this means that read_inode cannot create a superblock inode
-	 * today.  change if needed. */
-	if (!OCFS2_IS_VALID_DINODE(fe) ||
-	    !(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
-		mlog(0, "Invalid dinode: i_ino=%lu, i_blkno=%llu, "
-		     "signature = %.*s, flags = 0x%x\n",
-		     inode->i_ino,
-		     (unsigned long long)le64_to_cpu(fe->i_blkno), 7,
-		     fe->i_signature, le32_to_cpu(fe->i_flags));
-		goto bail;
-	}
+	/*
+	 * These have all been checked by ocfs2_read_inode_block() or set
+	 * by ocfs2_mknod_locked(), so a failure is a code bug.
+	 */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));  /* This means that read_inode
+						cannot create a superblock
+						inode today.  change if
+						that is needed. */
+	BUG_ON(!(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL)));
+	BUG_ON(le32_to_cpu(fe->i_fs_generation) != osb->fs_generation);
 
-	if (le32_to_cpu(fe->i_fs_generation) != osb->fs_generation) {
-		mlog(ML_ERROR, "file entry generation does not match "
-		     "superblock! osb->fs_generation=%x, "
-		     "fe->i_fs_generation=%x\n",
-		     osb->fs_generation, le32_to_cpu(fe->i_fs_generation));
-		goto bail;
-	}
 
 	OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
 	OCFS2_I(inode)->ip_attr = le32_to_cpu(fe->i_attr);
@@ -284,14 +277,18 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 
 	inode->i_nlink = le16_to_cpu(fe->i_links_count);
 
-	if (fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL))
+	if (fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) {
 		OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SYSTEM_FILE;
+		inode->i_flags |= S_NOQUOTA;
+	}
 
 	if (fe->i_flags & cpu_to_le32(OCFS2_LOCAL_ALLOC_FL)) {
 		OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
 		mlog(0, "local alloc inode: i_ino=%lu\n", inode->i_ino);
 	} else if (fe->i_flags & cpu_to_le32(OCFS2_BITMAP_FL)) {
 		OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
+	} else if (fe->i_flags & cpu_to_le32(OCFS2_QUOTA_FL)) {
+		inode->i_flags |= S_NOQUOTA;
 	} else if (fe->i_flags & cpu_to_le32(OCFS2_SUPER_BLOCK_FL)) {
 		mlog(0, "superblock inode: i_ino=%lu\n", inode->i_ino);
 		/* we can't actually hit this as read_inode can't
@@ -354,10 +351,7 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 
 	ocfs2_set_inode_flags(inode);
 
-	status = 0;
-bail:
-	mlog_exit(status);
-	return status;
+	mlog_exit_void();
 }
 
 static int ocfs2_read_locked_inode(struct inode *inode,
@@ -460,11 +454,14 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 		}
 	}
 
-	if (can_lock)
-		status = ocfs2_read_blocks(inode, args->fi_blkno, 1, &bh,
-					   OCFS2_BH_IGNORE_CACHE);
-	else
+	if (can_lock) {
+		status = ocfs2_read_inode_block_full(inode, &bh,
+						     OCFS2_BH_IGNORE_CACHE);
+	} else {
 		status = ocfs2_read_blocks_sync(osb, args->fi_blkno, 1, &bh);
+		if (!status)
+			status = ocfs2_validate_inode_block(osb->sb, bh);
+	}
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -472,12 +469,6 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 
 	status = -EINVAL;
 	fe = (struct ocfs2_dinode *) bh->b_data;
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		mlog(0, "Invalid dinode #%llu: signature = %.*s\n",
-		     (unsigned long long)args->fi_blkno, 7,
-		     fe->i_signature);
-		goto bail;
-	}
 
 	/*
 	 * This is a code bug. Right now the caller needs to
@@ -491,10 +482,9 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 
 	if (S_ISCHR(le16_to_cpu(fe->i_mode)) ||
 	    S_ISBLK(le16_to_cpu(fe->i_mode)))
-    		inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
+		inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
 
-	if (ocfs2_populate_inode(inode, fe, 0) < 0)
-		goto bail;
+	ocfs2_populate_inode(inode, fe, 0);
 
 	BUG_ON(args->fi_blkno != le64_to_cpu(fe->i_blkno));
 
@@ -547,8 +537,8 @@ static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
 			goto out;
 		}
 
-		status = ocfs2_journal_access(handle, inode, fe_bh,
-					      OCFS2_JOURNAL_ACCESS_WRITE);
+		status = ocfs2_journal_access_di(handle, inode, fe_bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto out;
@@ -615,7 +605,8 @@ static int ocfs2_remove_inode(struct inode *inode,
 		goto bail;
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_DELETE_INODE_CREDITS);
+	handle = ocfs2_start_trans(osb, OCFS2_DELETE_INODE_CREDITS +
+					ocfs2_quota_trans_credits(inode->i_sb));
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		mlog_errno(status);
@@ -630,8 +621,8 @@ static int ocfs2_remove_inode(struct inode *inode,
 	}
 
 	/* set the inodes dtime */
-	status = ocfs2_journal_access(handle, inode, di_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, inode, di_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail_commit;
@@ -647,6 +638,7 @@ static int ocfs2_remove_inode(struct inode *inode,
 	}
 
 	ocfs2_remove_from_cache(inode, di_bh);
+	vfs_dq_free_inode(inode);
 
 	status = ocfs2_free_dinode(handle, inode_alloc_inode,
 				   inode_alloc_bh, di);
@@ -929,7 +921,10 @@ void ocfs2_delete_inode(struct inode *inode)
 
 	mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);
 
-	if (is_bad_inode(inode)) {
+	/* When we fail in read_inode() we mark inode as bad. The second test
+	 * catches the case when inode allocation fails before allocating
+	 * a block for inode. */
+	if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno) {
 		mlog(0, "Skipping delete of bad inode\n");
 		goto bail;
 	}
@@ -1195,8 +1190,8 @@ int ocfs2_mark_inode_dirty(handle_t *handle,
 	mlog_entry("(inode %llu)\n",
 		   (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
-	status = ocfs2_journal_access(handle, inode, bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, inode, bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -1264,3 +1259,89 @@ void ocfs2_refresh_inode(struct inode *inode,
 
 	spin_unlock(&OCFS2_I(inode)->ip_lock);
 }
+
+int ocfs2_validate_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	mlog(0, "Validating dinode %llu\n",
+	     (unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 */
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check);
+	if (rc) {
+		mlog(ML_ERROR, "Checksum failed for dinode %llu\n",
+		     (unsigned long long)bh->b_blocknr);
+		goto bail;
+	}
+
+	/*
+	 * Errors after here are fatal.
+	 */
+
+	rc = -EINVAL;
+
+	if (!OCFS2_IS_VALID_DINODE(di)) {
+		ocfs2_error(sb, "Invalid dinode #%llu: signature = %.*s\n",
+			    (unsigned long long)bh->b_blocknr, 7,
+			    di->i_signature);
+		goto bail;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		ocfs2_error(sb, "Invalid dinode #%llu: i_blkno is %llu\n",
+			    (unsigned long long)bh->b_blocknr,
+			    (unsigned long long)le64_to_cpu(di->i_blkno));
+		goto bail;
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		ocfs2_error(sb,
+			    "Invalid dinode #%llu: OCFS2_VALID_FL not set\n",
+			    (unsigned long long)bh->b_blocknr);
+		goto bail;
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		ocfs2_error(sb,
+			    "Invalid dinode #%llu: fs_generation is %u\n",
+			    (unsigned long long)bh->b_blocknr,
+			    le32_to_cpu(di->i_fs_generation));
+		goto bail;
+	}
+
+	rc = 0;
+
+bail:
+	return rc;
+}
+
+int ocfs2_read_inode_block_full(struct inode *inode, struct buffer_head **bh,
+				int flags)
+{
+	int rc;
+	struct buffer_head *tmp = *bh;
+
+	rc = ocfs2_read_blocks(inode, OCFS2_I(inode)->ip_blkno, 1, &tmp,
+			       flags, ocfs2_validate_inode_block);
+
+	/* If ocfs2_read_blocks() got us a new bh, pass it up. */
+	if (!rc && !*bh)
+		*bh = tmp;
+
+	return rc;
+}
+
+int ocfs2_read_inode_block(struct inode *inode, struct buffer_head **bh)
+{
+	return ocfs2_read_inode_block_full(inode, bh, 0);
+}
diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h
index 2f37af9bcc4..eb3c302b38d 100644
--- a/fs/ocfs2/inode.h
+++ b/fs/ocfs2/inode.h
@@ -128,8 +128,8 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned flags,
 			 int sysfile_type);
 int ocfs2_inode_init_private(struct inode *inode);
 int ocfs2_inode_revalidate(struct dentry *dentry);
-int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
-			 int create_ino);
+void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
+			  int create_ino);
 void ocfs2_read_inode(struct inode *inode);
 void ocfs2_read_inode2(struct inode *inode, void *opaque);
 ssize_t ocfs2_rw_direct(int rw, struct file *filp, char *buf,
@@ -142,6 +142,8 @@ int ocfs2_mark_inode_dirty(handle_t *handle,
 			   struct buffer_head *bh);
 int ocfs2_aio_read(struct file *file, struct kiocb *req, struct iocb *iocb);
 int ocfs2_aio_write(struct file *file, struct kiocb *req, struct iocb *iocb);
+struct buffer_head *ocfs2_bread(struct inode *inode,
+				int block, int *err, int reada);
 
 void ocfs2_set_inode_flags(struct inode *inode);
 void ocfs2_get_inode_flags(struct ocfs2_inode_info *oi);
@@ -153,4 +155,16 @@ static inline blkcnt_t ocfs2_inode_sector_count(struct inode *inode)
 	return (blkcnt_t)(OCFS2_I(inode)->ip_clusters << c_to_s_bits);
 }
 
+/* Validate that a bh contains a valid inode */
+int ocfs2_validate_inode_block(struct super_block *sb,
+			       struct buffer_head *bh);
+/*
+ * Read an inode block into *bh.  If *bh is NULL, a bh will be allocated.
+ * This is a cached read.  The inode will be validated with
+ * ocfs2_validate_inode_block().
+ */
+int ocfs2_read_inode_block(struct inode *inode, struct buffer_head **bh);
+/* The same, but can be passed OCFS2_BH_* flags */
+int ocfs2_read_inode_block_full(struct inode *inode, struct buffer_head **bh,
+				int flags);
 #endif /* OCFS2_INODE_H */
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 99fe9d584f3..57d7d25a2b9 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -35,6 +35,7 @@
 #include "ocfs2.h"
 
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dir.h"
 #include "dlmglue.h"
 #include "extent_map.h"
@@ -45,6 +46,7 @@
 #include "slot_map.h"
 #include "super.h"
 #include "sysfile.h"
+#include "quota.h"
 
 #include "buffer_head_io.h"
 
@@ -52,10 +54,10 @@ DEFINE_SPINLOCK(trans_inc_lock);
 
 static int ocfs2_force_read_journal(struct inode *inode);
 static int ocfs2_recover_node(struct ocfs2_super *osb,
-			      int node_num);
+			      int node_num, int slot_num);
 static int __ocfs2_recovery_thread(void *arg);
 static int ocfs2_commit_cache(struct ocfs2_super *osb);
-static int ocfs2_wait_on_mount(struct ocfs2_super *osb);
+static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota);
 static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
 				      int dirty, int replayed);
 static int ocfs2_trylock_journal(struct ocfs2_super *osb,
@@ -64,6 +66,17 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb,
 				 int slot);
 static int ocfs2_commit_thread(void *arg);
 
+static inline int ocfs2_wait_on_mount(struct ocfs2_super *osb)
+{
+	return __ocfs2_wait_on_mount(osb, 0);
+}
+
+static inline int ocfs2_wait_on_quotas(struct ocfs2_super *osb)
+{
+	return __ocfs2_wait_on_mount(osb, 1);
+}
+
+
 
 /*
  * The recovery_list is a simple linked list of node numbers to recover.
@@ -256,11 +269,9 @@ handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs)
 	BUG_ON(osb->journal->j_state == OCFS2_JOURNAL_FREE);
 	BUG_ON(max_buffs <= 0);
 
-	/* JBD might support this, but our journalling code doesn't yet. */
-	if (journal_current_handle()) {
-		mlog(ML_ERROR, "Recursive transaction attempted!\n");
-		BUG();
-	}
+	/* Nested transaction? Just return the handle... */
+	if (journal_current_handle())
+		return jbd2_journal_start(journal, max_buffs);
 
 	down_read(&osb->journal->j_trans_barrier);
 
@@ -285,16 +296,18 @@ handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs)
 int ocfs2_commit_trans(struct ocfs2_super *osb,
 		       handle_t *handle)
 {
-	int ret;
+	int ret, nested;
 	struct ocfs2_journal *journal = osb->journal;
 
 	BUG_ON(!handle);
 
+	nested = handle->h_ref > 1;
 	ret = jbd2_journal_stop(handle);
 	if (ret < 0)
 		mlog_errno(ret);
 
-	up_read(&journal->j_trans_barrier);
+	if (!nested)
+		up_read(&journal->j_trans_barrier);
 
 	return ret;
 }
@@ -357,10 +370,137 @@ bail:
 	return status;
 }
 
-int ocfs2_journal_access(handle_t *handle,
-			 struct inode *inode,
-			 struct buffer_head *bh,
-			 int type)
+struct ocfs2_triggers {
+	struct jbd2_buffer_trigger_type	ot_triggers;
+	int				ot_offset;
+};
+
+static inline struct ocfs2_triggers *to_ocfs2_trigger(struct jbd2_buffer_trigger_type *triggers)
+{
+	return container_of(triggers, struct ocfs2_triggers, ot_triggers);
+}
+
+static void ocfs2_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
+				 struct buffer_head *bh,
+				 void *data, size_t size)
+{
+	struct ocfs2_triggers *ot = to_ocfs2_trigger(triggers);
+
+	/*
+	 * We aren't guaranteed to have the superblock here, so we
+	 * must unconditionally compute the ecc data.
+	 * __ocfs2_journal_access() will only set the triggers if
+	 * metaecc is enabled.
+	 */
+	ocfs2_block_check_compute(data, size, data + ot->ot_offset);
+}
+
+/*
+ * Quota blocks have their own trigger because the struct ocfs2_block_check
+ * offset depends on the blocksize.
+ */
+static void ocfs2_dq_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
+				 struct buffer_head *bh,
+				 void *data, size_t size)
+{
+	struct ocfs2_disk_dqtrailer *dqt =
+		ocfs2_block_dqtrailer(size, data);
+
+	/*
+	 * We aren't guaranteed to have the superblock here, so we
+	 * must unconditionally compute the ecc data.
+	 * __ocfs2_journal_access() will only set the triggers if
+	 * metaecc is enabled.
+	 */
+	ocfs2_block_check_compute(data, size, &dqt->dq_check);
+}
+
+/*
+ * Directory blocks also have their own trigger because the
+ * struct ocfs2_block_check offset depends on the blocksize.
+ */
+static void ocfs2_db_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
+				 struct buffer_head *bh,
+				 void *data, size_t size)
+{
+	struct ocfs2_dir_block_trailer *trailer =
+		ocfs2_dir_trailer_from_size(size, data);
+
+	/*
+	 * We aren't guaranteed to have the superblock here, so we
+	 * must unconditionally compute the ecc data.
+	 * __ocfs2_journal_access() will only set the triggers if
+	 * metaecc is enabled.
+	 */
+	ocfs2_block_check_compute(data, size, &trailer->db_check);
+}
+
+static void ocfs2_abort_trigger(struct jbd2_buffer_trigger_type *triggers,
+				struct buffer_head *bh)
+{
+	mlog(ML_ERROR,
+	     "ocfs2_abort_trigger called by JBD2.  bh = 0x%lx, "
+	     "bh->b_blocknr = %llu\n",
+	     (unsigned long)bh,
+	     (unsigned long long)bh->b_blocknr);
+
+	/* We aren't guaranteed to have the superblock here - but if we
+	 * don't, it'll just crash. */
+	ocfs2_error(bh->b_assoc_map->host->i_sb,
+		    "JBD2 has aborted our journal, ocfs2 cannot continue\n");
+}
+
+static struct ocfs2_triggers di_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+	.ot_offset	= offsetof(struct ocfs2_dinode, i_check),
+};
+
+static struct ocfs2_triggers eb_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+	.ot_offset	= offsetof(struct ocfs2_extent_block, h_check),
+};
+
+static struct ocfs2_triggers gd_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+	.ot_offset	= offsetof(struct ocfs2_group_desc, bg_check),
+};
+
+static struct ocfs2_triggers db_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_db_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+};
+
+static struct ocfs2_triggers xb_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+	.ot_offset	= offsetof(struct ocfs2_xattr_block, xb_check),
+};
+
+static struct ocfs2_triggers dq_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_dq_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+};
+
+static int __ocfs2_journal_access(handle_t *handle,
+				  struct inode *inode,
+				  struct buffer_head *bh,
+				  struct ocfs2_triggers *triggers,
+				  int type)
 {
 	int status;
 
@@ -406,6 +546,8 @@ int ocfs2_journal_access(handle_t *handle,
 		status = -EINVAL;
 		mlog(ML_ERROR, "Uknown access type!\n");
 	}
+	if (!status && ocfs2_meta_ecc(OCFS2_SB(inode->i_sb)) && triggers)
+		jbd2_journal_set_triggers(bh, &triggers->ot_triggers);
 	mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
 
 	if (status < 0)
@@ -416,6 +558,54 @@ int ocfs2_journal_access(handle_t *handle,
 	return status;
 }
 
+int ocfs2_journal_access_di(handle_t *handle, struct inode *inode,
+			       struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, &di_triggers,
+				      type);
+}
+
+int ocfs2_journal_access_eb(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, &eb_triggers,
+				      type);
+}
+
+int ocfs2_journal_access_gd(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, &gd_triggers,
+				      type);
+}
+
+int ocfs2_journal_access_db(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, &db_triggers,
+				      type);
+}
+
+int ocfs2_journal_access_xb(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, &xb_triggers,
+				      type);
+}
+
+int ocfs2_journal_access_dq(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, &dq_triggers,
+				      type);
+}
+
+int ocfs2_journal_access(handle_t *handle, struct inode *inode,
+			 struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, inode, bh, NULL, type);
+}
+
 int ocfs2_journal_dirty(handle_t *handle,
 			struct buffer_head *bh)
 {
@@ -434,20 +624,6 @@ int ocfs2_journal_dirty(handle_t *handle,
 	return status;
 }
 
-#ifdef CONFIG_OCFS2_COMPAT_JBD
-int ocfs2_journal_dirty_data(handle_t *handle,
-			     struct buffer_head *bh)
-{
-	int err = journal_dirty_data(handle, bh);
-	if (err)
-		mlog_errno(err);
-	/* TODO: When we can handle it, abort the handle and go RO on
-	 * error here. */
-
-	return err;
-}
-#endif
-
 #define OCFS2_DEFAULT_COMMIT_INTERVAL	(HZ * JBD2_DEFAULT_MAX_COMMIT_AGE)
 
 void ocfs2_set_journal_params(struct ocfs2_super *osb)
@@ -587,17 +763,11 @@ static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
 	mlog_entry_void();
 
 	fe = (struct ocfs2_dinode *)bh->b_data;
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		/* This is called from startup/shutdown which will
-		 * handle the errors in a specific manner, so no need
-		 * to call ocfs2_error() here. */
-		mlog(ML_ERROR, "Journal dinode %llu  has invalid "
-		     "signature: %.*s",
-		     (unsigned long long)le64_to_cpu(fe->i_blkno), 7,
-		     fe->i_signature);
-		status = -EIO;
-		goto out;
-	}
+
+	/* The journal bh on the osb always comes from ocfs2_journal_init()
+	 * and was validated there inside ocfs2_inode_lock_full().  It's a
+	 * code bug if we mess it up. */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
 
 	flags = le32_to_cpu(fe->id1.journal1.ij_flags);
 	if (dirty)
@@ -609,11 +779,11 @@ static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
 	if (replayed)
 		ocfs2_bump_recovery_generation(fe);
 
+	ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
 	status = ocfs2_write_block(osb, bh, journal->j_inode);
 	if (status < 0)
 		mlog_errno(status);
 
-out:
 	mlog_exit(status);
 	return status;
 }
@@ -878,6 +1048,7 @@ struct ocfs2_la_recovery_item {
 	int			lri_slot;
 	struct ocfs2_dinode	*lri_la_dinode;
 	struct ocfs2_dinode	*lri_tl_dinode;
+	struct ocfs2_quota_recovery *lri_qrec;
 };
 
 /* Does the second half of the recovery process. By this point, the
@@ -898,6 +1069,7 @@ void ocfs2_complete_recovery(struct work_struct *work)
 	struct ocfs2_super *osb = journal->j_osb;
 	struct ocfs2_dinode *la_dinode, *tl_dinode;
 	struct ocfs2_la_recovery_item *item, *n;
+	struct ocfs2_quota_recovery *qrec;
 	LIST_HEAD(tmp_la_list);
 
 	mlog_entry_void();
@@ -913,6 +1085,8 @@ void ocfs2_complete_recovery(struct work_struct *work)
 
 		mlog(0, "Complete recovery for slot %d\n", item->lri_slot);
 
+		ocfs2_wait_on_quotas(osb);
+
 		la_dinode = item->lri_la_dinode;
 		if (la_dinode) {
 			mlog(0, "Clean up local alloc %llu\n",
@@ -943,6 +1117,16 @@ void ocfs2_complete_recovery(struct work_struct *work)
 		if (ret < 0)
 			mlog_errno(ret);
 
+		qrec = item->lri_qrec;
+		if (qrec) {
+			mlog(0, "Recovering quota files");
+			ret = ocfs2_finish_quota_recovery(osb, qrec,
+							  item->lri_slot);
+			if (ret < 0)
+				mlog_errno(ret);
+			/* Recovery info is already freed now */
+		}
+
 		kfree(item);
 	}
 
@@ -956,7 +1140,8 @@ void ocfs2_complete_recovery(struct work_struct *work)
 static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
 					    int slot_num,
 					    struct ocfs2_dinode *la_dinode,
-					    struct ocfs2_dinode *tl_dinode)
+					    struct ocfs2_dinode *tl_dinode,
+					    struct ocfs2_quota_recovery *qrec)
 {
 	struct ocfs2_la_recovery_item *item;
 
@@ -971,6 +1156,9 @@ static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
 		if (tl_dinode)
 			kfree(tl_dinode);
 
+		if (qrec)
+			ocfs2_free_quota_recovery(qrec);
+
 		mlog_errno(-ENOMEM);
 		return;
 	}
@@ -979,6 +1167,7 @@ static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
 	item->lri_la_dinode = la_dinode;
 	item->lri_slot = slot_num;
 	item->lri_tl_dinode = tl_dinode;
+	item->lri_qrec = qrec;
 
 	spin_lock(&journal->j_lock);
 	list_add_tail(&item->lri_list, &journal->j_la_cleanups);
@@ -998,6 +1187,7 @@ void ocfs2_complete_mount_recovery(struct ocfs2_super *osb)
 		ocfs2_queue_recovery_completion(journal,
 						osb->slot_num,
 						osb->local_alloc_copy,
+						NULL,
 						NULL);
 		ocfs2_schedule_truncate_log_flush(osb, 0);
 
@@ -1006,11 +1196,26 @@ void ocfs2_complete_mount_recovery(struct ocfs2_super *osb)
 	}
 }
 
+void ocfs2_complete_quota_recovery(struct ocfs2_super *osb)
+{
+	if (osb->quota_rec) {
+		ocfs2_queue_recovery_completion(osb->journal,
+						osb->slot_num,
+						NULL,
+						NULL,
+						osb->quota_rec);
+		osb->quota_rec = NULL;
+	}
+}
+
 static int __ocfs2_recovery_thread(void *arg)
 {
-	int status, node_num;
+	int status, node_num, slot_num;
 	struct ocfs2_super *osb = arg;
 	struct ocfs2_recovery_map *rm = osb->recovery_map;
+	int *rm_quota = NULL;
+	int rm_quota_used = 0, i;
+	struct ocfs2_quota_recovery *qrec;
 
 	mlog_entry_void();
 
@@ -1019,6 +1224,11 @@ static int __ocfs2_recovery_thread(void *arg)
 		goto bail;
 	}
 
+	rm_quota = kzalloc(osb->max_slots * sizeof(int), GFP_NOFS);
+	if (!rm_quota) {
+		status = -ENOMEM;
+		goto bail;
+	}
 restart:
 	status = ocfs2_super_lock(osb, 1);
 	if (status < 0) {
@@ -1032,8 +1242,28 @@ restart:
 		 * clear it until ocfs2_recover_node() has succeeded. */
 		node_num = rm->rm_entries[0];
 		spin_unlock(&osb->osb_lock);
-
-		status = ocfs2_recover_node(osb, node_num);
+		mlog(0, "checking node %d\n", node_num);
+		slot_num = ocfs2_node_num_to_slot(osb, node_num);
+		if (slot_num == -ENOENT) {
+			status = 0;
+			mlog(0, "no slot for this node, so no recovery"
+			     "required.\n");
+			goto skip_recovery;
+		}
+		mlog(0, "node %d was using slot %d\n", node_num, slot_num);
+
+		/* It is a bit subtle with quota recovery. We cannot do it
+		 * immediately because we have to obtain cluster locks from
+		 * quota files and we also don't want to just skip it because
+		 * then quota usage would be out of sync until some node takes
+		 * the slot. So we remember which nodes need quota recovery
+		 * and when everything else is done, we recover quotas. */
+		for (i = 0; i < rm_quota_used && rm_quota[i] != slot_num; i++);
+		if (i == rm_quota_used)
+			rm_quota[rm_quota_used++] = slot_num;
+
+		status = ocfs2_recover_node(osb, node_num, slot_num);
+skip_recovery:
 		if (!status) {
 			ocfs2_recovery_map_clear(osb, node_num);
 		} else {
@@ -1055,13 +1285,27 @@ restart:
 	if (status < 0)
 		mlog_errno(status);
 
+	/* Now it is right time to recover quotas... We have to do this under
+	 * superblock lock so that noone can start using the slot (and crash)
+	 * before we recover it */
+	for (i = 0; i < rm_quota_used; i++) {
+		qrec = ocfs2_begin_quota_recovery(osb, rm_quota[i]);
+		if (IS_ERR(qrec)) {
+			status = PTR_ERR(qrec);
+			mlog_errno(status);
+			continue;
+		}
+		ocfs2_queue_recovery_completion(osb->journal, rm_quota[i],
+						NULL, NULL, qrec);
+	}
+
 	ocfs2_super_unlock(osb, 1);
 
 	/* We always run recovery on our own orphan dir - the dead
 	 * node(s) may have disallowd a previos inode delete. Re-processing
 	 * is therefore required. */
 	ocfs2_queue_recovery_completion(osb->journal, osb->slot_num, NULL,
-					NULL);
+					NULL, NULL);
 
 bail:
 	mutex_lock(&osb->recovery_lock);
@@ -1076,6 +1320,9 @@ bail:
 
 	mutex_unlock(&osb->recovery_lock);
 
+	if (rm_quota)
+		kfree(rm_quota);
+
 	mlog_exit(status);
 	/* no one is callint kthread_stop() for us so the kthread() api
 	 * requires that we call do_exit().  And it isn't exported, but
@@ -1135,8 +1382,7 @@ static int ocfs2_read_journal_inode(struct ocfs2_super *osb,
 	}
 	SET_INODE_JOURNAL(inode);
 
-	status = ocfs2_read_blocks(inode, OCFS2_I(inode)->ip_blkno, 1, bh,
-				   OCFS2_BH_IGNORE_CACHE);
+	status = ocfs2_read_inode_block_full(inode, bh, OCFS2_BH_IGNORE_CACHE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1268,6 +1514,7 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
 	osb->slot_recovery_generations[slot_num] =
 					ocfs2_get_recovery_generation(fe);
 
+	ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
 	status = ocfs2_write_block(osb, bh, inode);
 	if (status < 0)
 		mlog_errno(status);
@@ -1304,31 +1551,19 @@ done:
  * far less concerning.
  */
 static int ocfs2_recover_node(struct ocfs2_super *osb,
-			      int node_num)
+			      int node_num, int slot_num)
 {
 	int status = 0;
-	int slot_num;
 	struct ocfs2_dinode *la_copy = NULL;
 	struct ocfs2_dinode *tl_copy = NULL;
 
-	mlog_entry("(node_num=%d, osb->node_num = %d)\n",
-		   node_num, osb->node_num);
-
-	mlog(0, "checking node %d\n", node_num);
+	mlog_entry("(node_num=%d, slot_num=%d, osb->node_num = %d)\n",
+		   node_num, slot_num, osb->node_num);
 
 	/* Should not ever be called to recover ourselves -- in that
 	 * case we should've called ocfs2_journal_load instead. */
 	BUG_ON(osb->node_num == node_num);
 
-	slot_num = ocfs2_node_num_to_slot(osb, node_num);
-	if (slot_num == -ENOENT) {
-		status = 0;
-		mlog(0, "no slot for this node, so no recovery required.\n");
-		goto done;
-	}
-
-	mlog(0, "node %d was using slot %d\n", node_num, slot_num);
-
 	status = ocfs2_replay_journal(osb, node_num, slot_num);
 	if (status < 0) {
 		if (status == -EBUSY) {
@@ -1364,7 +1599,7 @@ static int ocfs2_recover_node(struct ocfs2_super *osb,
 
 	/* This will kfree the memory pointed to by la_copy and tl_copy */
 	ocfs2_queue_recovery_completion(osb->journal, slot_num, la_copy,
-					tl_copy);
+					tl_copy, NULL);
 
 	status = 0;
 done:
@@ -1659,13 +1894,14 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb,
 	return ret;
 }
 
-static int ocfs2_wait_on_mount(struct ocfs2_super *osb)
+static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota)
 {
 	/* This check is good because ocfs2 will wait on our recovery
 	 * thread before changing it to something other than MOUNTED
 	 * or DISABLED. */
 	wait_event(osb->osb_mount_event,
-		   atomic_read(&osb->vol_state) == VOLUME_MOUNTED ||
+		  (!quota && atomic_read(&osb->vol_state) == VOLUME_MOUNTED) ||
+		   atomic_read(&osb->vol_state) == VOLUME_MOUNTED_QUOTAS ||
 		   atomic_read(&osb->vol_state) == VOLUME_DISABLED);
 
 	/* If there's an error on mount, then we may never get to the
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index d4d14e9a3ce..3c3532e1307 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -27,12 +27,7 @@
 #define OCFS2_JOURNAL_H
 
 #include <linux/fs.h>
-#ifndef CONFIG_OCFS2_COMPAT_JBD
-# include <linux/jbd2.h>
-#else
-# include <linux/jbd.h>
-# include "ocfs2_jbd_compat.h"
-#endif
+#include <linux/jbd2.h>
 
 enum ocfs2_journal_state {
 	OCFS2_JOURNAL_FREE = 0,
@@ -173,6 +168,7 @@ void   ocfs2_recovery_thread(struct ocfs2_super *osb,
 			     int node_num);
 int    ocfs2_mark_dead_nodes(struct ocfs2_super *osb);
 void   ocfs2_complete_mount_recovery(struct ocfs2_super *osb);
+void ocfs2_complete_quota_recovery(struct ocfs2_super *osb);
 
 static inline void ocfs2_start_checkpoint(struct ocfs2_super *osb)
 {
@@ -216,9 +212,12 @@ static inline void ocfs2_checkpoint_inode(struct inode *inode)
  *  ocfs2_extend_trans     - Extend a handle by nblocks credits. This may
  *                          commit the handle to disk in the process, but will
  *                          not release any locks taken during the transaction.
- *  ocfs2_journal_access   - Notify the handle that we want to journal this
+ *  ocfs2_journal_access* - Notify the handle that we want to journal this
  *                          buffer. Will have to call ocfs2_journal_dirty once
  *                          we've actually dirtied it. Type is one of . or .
+ *                          Always call the specific flavor of
+ *                          ocfs2_journal_access_*() unless you intend to
+ *                          manage the checksum by hand.
  *  ocfs2_journal_dirty    - Mark a journalled buffer as having dirty data.
  *  ocfs2_jbd2_file_inode  - Mark an inode so that its data goes out before
  *                           the current handle commits.
@@ -248,10 +247,29 @@ int			     ocfs2_extend_trans(handle_t *handle, int nblocks);
 #define OCFS2_JOURNAL_ACCESS_WRITE  1
 #define OCFS2_JOURNAL_ACCESS_UNDO   2
 
-int                  ocfs2_journal_access(handle_t *handle,
-					  struct inode *inode,
-					  struct buffer_head *bh,
-					  int type);
+
+/* ocfs2_inode */
+int ocfs2_journal_access_di(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type);
+/* ocfs2_extent_block */
+int ocfs2_journal_access_eb(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type);
+/* ocfs2_group_desc */
+int ocfs2_journal_access_gd(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type);
+/* ocfs2_xattr_block */
+int ocfs2_journal_access_xb(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type);
+/* quota blocks */
+int ocfs2_journal_access_dq(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type);
+/* dirblock */
+int ocfs2_journal_access_db(handle_t *handle, struct inode *inode,
+			    struct buffer_head *bh, int type);
+/* Anything that has no ecc */
+int ocfs2_journal_access(handle_t *handle, struct inode *inode,
+			 struct buffer_head *bh, int type);
+
 /*
  * A word about the journal_access/journal_dirty "dance". It is
  * entirely legal to journal_access a buffer more than once (as long
@@ -273,10 +291,6 @@ int                  ocfs2_journal_access(handle_t *handle,
  */
 int                  ocfs2_journal_dirty(handle_t *handle,
 					 struct buffer_head *bh);
-#ifdef CONFIG_OCFS2_COMPAT_JBD
-int                  ocfs2_journal_dirty_data(handle_t *handle,
-					      struct buffer_head *bh);
-#endif
 
 /*
  *  Credit Macros:
@@ -293,6 +307,37 @@ int                  ocfs2_journal_dirty_data(handle_t *handle,
 /* extended attribute block update */
 #define OCFS2_XATTR_BLOCK_UPDATE_CREDITS 1
 
+/* global quotafile inode update, data block */
+#define OCFS2_QINFO_WRITE_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1)
+
+/*
+ * The two writes below can accidentally see global info dirty due
+ * to set_info() quotactl so make them prepared for the writes.
+ */
+/* quota data block, global info */
+/* Write to local quota file */
+#define OCFS2_QWRITE_CREDITS (OCFS2_QINFO_WRITE_CREDITS + 1)
+
+/* global quota data block, local quota data block, global quota inode,
+ * global quota info */
+#define OCFS2_QSYNC_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 3)
+
+static inline int ocfs2_quota_trans_credits(struct super_block *sb)
+{
+	int credits = 0;
+
+	if (OCFS2_HAS_RO_COMPAT_FEATURE(sb, OCFS2_FEATURE_RO_COMPAT_USRQUOTA))
+		credits += OCFS2_QWRITE_CREDITS;
+	if (OCFS2_HAS_RO_COMPAT_FEATURE(sb, OCFS2_FEATURE_RO_COMPAT_GRPQUOTA))
+		credits += OCFS2_QWRITE_CREDITS;
+	return credits;
+}
+
+/* Number of credits needed for removing quota structure from file */
+int ocfs2_calc_qdel_credits(struct super_block *sb, int type);
+/* Number of credits needed for initialization of new quota structure */
+int ocfs2_calc_qinit_credits(struct super_block *sb, int type);
+
 /* group extend. inode update and last group update. */
 #define OCFS2_GROUP_EXTEND_CREDITS	(OCFS2_INODE_UPDATE_CREDITS + 1)
 
@@ -303,8 +348,11 @@ int                  ocfs2_journal_dirty_data(handle_t *handle,
  * prev. group desc. if we relink. */
 #define OCFS2_SUBALLOC_ALLOC (3)
 
-#define OCFS2_INLINE_TO_EXTENTS_CREDITS (OCFS2_SUBALLOC_ALLOC		\
-					 + OCFS2_INODE_UPDATE_CREDITS)
+static inline int ocfs2_inline_to_extents_credits(struct super_block *sb)
+{
+	return OCFS2_SUBALLOC_ALLOC + OCFS2_INODE_UPDATE_CREDITS +
+	       ocfs2_quota_trans_credits(sb);
+}
 
 /* dinode + group descriptor update. We don't relink on free yet. */
 #define OCFS2_SUBALLOC_FREE  (2)
@@ -313,16 +361,23 @@ int                  ocfs2_journal_dirty_data(handle_t *handle,
 #define OCFS2_TRUNCATE_LOG_FLUSH_ONE_REC (OCFS2_SUBALLOC_FREE 		      \
 					 + OCFS2_TRUNCATE_LOG_UPDATE)
 
-#define OCFS2_REMOVE_EXTENT_CREDITS (OCFS2_TRUNCATE_LOG_UPDATE + OCFS2_INODE_UPDATE_CREDITS)
+static inline int ocfs2_remove_extent_credits(struct super_block *sb)
+{
+	return OCFS2_TRUNCATE_LOG_UPDATE + OCFS2_INODE_UPDATE_CREDITS +
+	       ocfs2_quota_trans_credits(sb);
+}
 
 /* data block for new dir/symlink, 2 for bitmap updates (bitmap fe +
  * bitmap block for the new bit) */
 #define OCFS2_DIR_LINK_ADDITIONAL_CREDITS (1 + 2)
 
 /* parent fe, parent block, new file entry, inode alloc fe, inode alloc
- * group descriptor + mkdir/symlink blocks */
-#define OCFS2_MKNOD_CREDITS (3 + OCFS2_SUBALLOC_ALLOC                         \
-			    + OCFS2_DIR_LINK_ADDITIONAL_CREDITS)
+ * group descriptor + mkdir/symlink blocks + quota update */
+static inline int ocfs2_mknod_credits(struct super_block *sb)
+{
+	return 3 + OCFS2_SUBALLOC_ALLOC + OCFS2_DIR_LINK_ADDITIONAL_CREDITS +
+	       ocfs2_quota_trans_credits(sb);
+}
 
 /* local alloc metadata change + main bitmap updates */
 #define OCFS2_WINDOW_MOVE_CREDITS (OCFS2_INODE_UPDATE_CREDITS                 \
@@ -332,13 +387,21 @@ int                  ocfs2_journal_dirty_data(handle_t *handle,
  * for the dinode, one for the new block. */
 #define OCFS2_SIMPLE_DIR_EXTEND_CREDITS (2)
 
-/* file update (nlink, etc) + directory mtime/ctime + dir entry block */
-#define OCFS2_LINK_CREDITS  (2*OCFS2_INODE_UPDATE_CREDITS + 1)
+/* file update (nlink, etc) + directory mtime/ctime + dir entry block + quota
+ * update on dir */
+static inline int ocfs2_link_credits(struct super_block *sb)
+{
+	return 2*OCFS2_INODE_UPDATE_CREDITS + 1 +
+	       ocfs2_quota_trans_credits(sb);
+}
 
 /* inode + dir inode (if we unlink a dir), + dir entry block + orphan
  * dir inode link */
-#define OCFS2_UNLINK_CREDITS  (2 * OCFS2_INODE_UPDATE_CREDITS + 1             \
-			      + OCFS2_LINK_CREDITS)
+static inline int ocfs2_unlink_credits(struct super_block *sb)
+{
+	/* The quota update from ocfs2_link_credits is unused here... */
+	return 2 * OCFS2_INODE_UPDATE_CREDITS + 1 + ocfs2_link_credits(sb);
+}
 
 /* dinode + orphan dir dinode + inode alloc dinode + orphan dir entry +
  * inode alloc group descriptor */
@@ -347,8 +410,10 @@ int                  ocfs2_journal_dirty_data(handle_t *handle,
 /* dinode update, old dir dinode update, new dir dinode update, old
  * dir dir entry, new dir dir entry, dir entry update for renaming
  * directory + target unlink */
-#define OCFS2_RENAME_CREDITS (3 * OCFS2_INODE_UPDATE_CREDITS + 3              \
-			     + OCFS2_UNLINK_CREDITS)
+static inline int ocfs2_rename_credits(struct super_block *sb)
+{
+	return 3 * OCFS2_INODE_UPDATE_CREDITS + 3 + ocfs2_unlink_credits(sb);
+}
 
 /* global bitmap dinode, group desc., relinked group,
  * suballocator dinode, group desc., relinked group,
@@ -386,18 +451,19 @@ static inline int ocfs2_calc_extend_credits(struct super_block *sb,
 	 * credit for the dinode there. */
 	extent_blocks = 1 + 1 + le16_to_cpu(root_el->l_tree_depth);
 
-	return bitmap_blocks + sysfile_bitmap_blocks + extent_blocks;
+	return bitmap_blocks + sysfile_bitmap_blocks + extent_blocks +
+	       ocfs2_quota_trans_credits(sb);
 }
 
 static inline int ocfs2_calc_symlink_credits(struct super_block *sb)
 {
-	int blocks = OCFS2_MKNOD_CREDITS;
+	int blocks = ocfs2_mknod_credits(sb);
 
 	/* links can be longer than one block so we may update many
 	 * within our single allocated extent. */
 	blocks += ocfs2_clusters_to_blocks(sb, 1);
 
-	return blocks;
+	return blocks + ocfs2_quota_trans_credits(sb);
 }
 
 static inline int ocfs2_calc_group_alloc_credits(struct super_block *sb,
@@ -434,6 +500,8 @@ static inline int ocfs2_calc_tree_trunc_credits(struct super_block *sb,
 	/* update to the truncate log. */
 	credits += OCFS2_TRUNCATE_LOG_UPDATE;
 
+	credits += ocfs2_quota_trans_credits(sb);
+
 	return credits;
 }
 
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index 687b28713c3..ec70cdbe77f 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -36,6 +36,7 @@
 #include "ocfs2.h"
 
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dlmglue.h"
 #include "inode.h"
 #include "journal.h"
@@ -248,8 +249,8 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb)
 		goto bail;
 	}
 
-	status = ocfs2_read_blocks(inode, OCFS2_I(inode)->ip_blkno, 1,
-				   &alloc_bh, OCFS2_BH_IGNORE_CACHE);
+	status = ocfs2_read_inode_block_full(inode, &alloc_bh,
+					     OCFS2_BH_IGNORE_CACHE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -382,8 +383,8 @@ void ocfs2_shutdown_local_alloc(struct ocfs2_super *osb)
 	}
 	memcpy(alloc_copy, alloc, bh->b_size);
 
-	status = ocfs2_journal_access(handle, local_alloc_inode, bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, local_alloc_inode, bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_commit;
@@ -459,8 +460,8 @@ int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
 
 	mutex_lock(&inode->i_mutex);
 
-	status = ocfs2_read_blocks(inode, OCFS2_I(inode)->ip_blkno, 1,
-				   &alloc_bh, OCFS2_BH_IGNORE_CACHE);
+	status = ocfs2_read_inode_block_full(inode, &alloc_bh,
+					     OCFS2_BH_IGNORE_CACHE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -476,6 +477,7 @@ int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
 	alloc = (struct ocfs2_dinode *) alloc_bh->b_data;
 	ocfs2_clear_local_alloc(alloc);
 
+	ocfs2_compute_meta_ecc(osb->sb, alloc_bh->b_data, &alloc->i_check);
 	status = ocfs2_write_block(osb, alloc_bh, inode);
 	if (status < 0)
 		mlog_errno(status);
@@ -762,9 +764,9 @@ int ocfs2_claim_local_alloc_bits(struct ocfs2_super *osb,
 	 * delete bits from it! */
 	*num_bits = bits_wanted;
 
-	status = ocfs2_journal_access(handle, local_alloc_inode,
-				      osb->local_alloc_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, local_alloc_inode,
+					 osb->local_alloc_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1240,9 +1242,9 @@ static int ocfs2_local_alloc_slide_window(struct ocfs2_super *osb,
 	}
 	memcpy(alloc_copy, alloc, osb->local_alloc_bh->b_size);
 
-	status = ocfs2_journal_access(handle, local_alloc_inode,
-				      osb->local_alloc_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, local_alloc_inode,
+					 osb->local_alloc_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index 2545e7402ef..084aba86c3b 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -40,6 +40,7 @@
 #include <linux/types.h>
 #include <linux/slab.h>
 #include <linux/highmem.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_NAMEI
 #include <cluster/masklog.h>
@@ -61,17 +62,18 @@
 #include "sysfile.h"
 #include "uptodate.h"
 #include "xattr.h"
+#include "acl.h"
 
 #include "buffer_head_io.h"
 
 static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 			      struct inode *dir,
-			      struct dentry *dentry, int mode,
+			      struct inode *inode,
+			      struct dentry *dentry,
 			      dev_t dev,
 			      struct buffer_head **new_fe_bh,
 			      struct buffer_head *parent_fe_bh,
 			      handle_t *handle,
-			      struct inode **ret_inode,
 			      struct ocfs2_alloc_context *inode_ac);
 
 static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
@@ -186,6 +188,35 @@ bail:
 	return ret;
 }
 
+static struct inode *ocfs2_get_init_inode(struct inode *dir, int mode)
+{
+	struct inode *inode;
+
+	inode = new_inode(dir->i_sb);
+	if (!inode) {
+		mlog(ML_ERROR, "new_inode failed!\n");
+		return NULL;
+	}
+
+	/* populate as many fields early on as possible - many of
+	 * these are used by the support functions here and in
+	 * callers. */
+	if (S_ISDIR(mode))
+		inode->i_nlink = 2;
+	else
+		inode->i_nlink = 1;
+	inode->i_uid = current_fsuid();
+	if (dir->i_mode & S_ISGID) {
+		inode->i_gid = dir->i_gid;
+		if (S_ISDIR(mode))
+			mode |= S_ISGID;
+	} else
+		inode->i_gid = current_fsgid();
+	inode->i_mode = mode;
+	vfs_dq_init(inode);
+	return inode;
+}
+
 static int ocfs2_mknod(struct inode *dir,
 		       struct dentry *dentry,
 		       int mode,
@@ -201,6 +232,13 @@ static int ocfs2_mknod(struct inode *dir,
 	struct inode *inode = NULL;
 	struct ocfs2_alloc_context *inode_ac = NULL;
 	struct ocfs2_alloc_context *data_ac = NULL;
+	struct ocfs2_alloc_context *xattr_ac = NULL;
+	int want_clusters = 0;
+	int xattr_credits = 0;
+	struct ocfs2_security_xattr_info si = {
+		.enable = 1,
+	};
+	int did_quota_inode = 0;
 
 	mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
 		   (unsigned long)dev, dentry->d_name.len,
@@ -250,17 +288,46 @@ static int ocfs2_mknod(struct inode *dir,
 		goto leave;
 	}
 
-	/* Reserve a cluster if creating an extent based directory. */
-	if (S_ISDIR(mode) && !ocfs2_supports_inline_data(osb)) {
-		status = ocfs2_reserve_clusters(osb, 1, &data_ac);
-		if (status < 0) {
-			if (status != -ENOSPC)
-				mlog_errno(status);
+	inode = ocfs2_get_init_inode(dir, mode);
+	if (!inode) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto leave;
+	}
+
+	/* get security xattr */
+	status = ocfs2_init_security_get(inode, dir, &si);
+	if (status) {
+		if (status == -EOPNOTSUPP)
+			si.enable = 0;
+		else {
+			mlog_errno(status);
 			goto leave;
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_MKNOD_CREDITS);
+	/* calculate meta data/clusters for setting security and acl xattr */
+	status = ocfs2_calc_xattr_init(dir, parent_fe_bh, mode,
+					&si, &want_clusters,
+					&xattr_credits, &xattr_ac);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	/* Reserve a cluster if creating an extent based directory. */
+	if (S_ISDIR(mode) && !ocfs2_supports_inline_data(osb))
+		want_clusters += 1;
+
+	status = ocfs2_reserve_clusters(osb, want_clusters, &data_ac);
+	if (status < 0) {
+		if (status != -ENOSPC)
+			mlog_errno(status);
+		goto leave;
+	}
+
+	handle = ocfs2_start_trans(osb, ocfs2_mknod_credits(osb->sb) +
+				   xattr_credits);
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		handle = NULL;
@@ -268,10 +335,19 @@ static int ocfs2_mknod(struct inode *dir,
 		goto leave;
 	}
 
+	/* We don't use standard VFS wrapper because we don't want vfs_dq_init
+	 * to be called. */
+	if (sb_any_quota_active(osb->sb) &&
+	    osb->sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) {
+		status = -EDQUOT;
+		goto leave;
+	}
+	did_quota_inode = 1;
+
 	/* do the real work now. */
-	status = ocfs2_mknod_locked(osb, dir, dentry, mode, dev,
+	status = ocfs2_mknod_locked(osb, dir, inode, dentry, dev,
 				    &new_fe_bh, parent_fe_bh, handle,
-				    &inode, inode_ac);
+				    inode_ac);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -285,8 +361,8 @@ static int ocfs2_mknod(struct inode *dir,
 			goto leave;
 		}
 
-		status = ocfs2_journal_access(handle, dir, parent_fe_bh,
-					      OCFS2_JOURNAL_ACCESS_WRITE);
+		status = ocfs2_journal_access_di(handle, dir, parent_fe_bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto leave;
@@ -300,6 +376,22 @@ static int ocfs2_mknod(struct inode *dir,
 		inc_nlink(dir);
 	}
 
+	status = ocfs2_init_acl(handle, inode, dir, new_fe_bh, parent_fe_bh,
+				xattr_ac, data_ac);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	if (si.enable) {
+		status = ocfs2_init_security_set(handle, inode, new_fe_bh, &si,
+						 xattr_ac, data_ac);
+		if (status < 0) {
+			mlog_errno(status);
+			goto leave;
+		}
+	}
+
 	status = ocfs2_add_entry(handle, dentry, inode,
 				 OCFS2_I(inode)->ip_blkno, parent_fe_bh,
 				 de_bh);
@@ -320,6 +412,8 @@ static int ocfs2_mknod(struct inode *dir,
 	d_instantiate(dentry, inode);
 	status = 0;
 leave:
+	if (status < 0 && did_quota_inode)
+		vfs_dq_free_inode(inode);
 	if (handle)
 		ocfs2_commit_trans(osb, handle);
 
@@ -331,9 +425,13 @@ leave:
 	brelse(new_fe_bh);
 	brelse(de_bh);
 	brelse(parent_fe_bh);
+	kfree(si.name);
+	kfree(si.value);
 
-	if ((status < 0) && inode)
+	if ((status < 0) && inode) {
+		clear_nlink(inode);
 		iput(inode);
+	}
 
 	if (inode_ac)
 		ocfs2_free_alloc_context(inode_ac);
@@ -341,6 +439,9 @@ leave:
 	if (data_ac)
 		ocfs2_free_alloc_context(data_ac);
 
+	if (xattr_ac)
+		ocfs2_free_alloc_context(xattr_ac);
+
 	mlog_exit(status);
 
 	return status;
@@ -348,12 +449,12 @@ leave:
 
 static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 			      struct inode *dir,
-			      struct dentry *dentry, int mode,
+			      struct inode *inode,
+			      struct dentry *dentry,
 			      dev_t dev,
 			      struct buffer_head **new_fe_bh,
 			      struct buffer_head *parent_fe_bh,
 			      handle_t *handle,
-			      struct inode **ret_inode,
 			      struct ocfs2_alloc_context *inode_ac)
 {
 	int status = 0;
@@ -361,14 +462,12 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	struct ocfs2_extent_list *fel;
 	u64 fe_blkno = 0;
 	u16 suballoc_bit;
-	struct inode *inode = NULL;
 
-	mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
-		   (unsigned long)dev, dentry->d_name.len,
+	mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry,
+		   inode->i_mode, (unsigned long)dev, dentry->d_name.len,
 		   dentry->d_name.name);
 
 	*new_fe_bh = NULL;
-	*ret_inode = NULL;
 
 	status = ocfs2_claim_new_inode(osb, handle, inode_ac, &suballoc_bit,
 				       &fe_blkno);
@@ -377,23 +476,11 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 		goto leave;
 	}
 
-	inode = new_inode(dir->i_sb);
-	if (!inode) {
-		status = -ENOMEM;
-		mlog(ML_ERROR, "new_inode failed!\n");
-		goto leave;
-	}
-
 	/* populate as many fields early on as possible - many of
 	 * these are used by the support functions here and in
 	 * callers. */
 	inode->i_ino = ino_from_blkno(osb->sb, fe_blkno);
 	OCFS2_I(inode)->ip_blkno = fe_blkno;
-	if (S_ISDIR(mode))
-		inode->i_nlink = 2;
-	else
-		inode->i_nlink = 1;
-	inode->i_mode = mode;
 	spin_lock(&osb->osb_lock);
 	inode->i_generation = osb->s_next_generation++;
 	spin_unlock(&osb->osb_lock);
@@ -406,8 +493,8 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	}
 	ocfs2_set_new_buffer_uptodate(inode, *new_fe_bh);
 
-	status = ocfs2_journal_access(handle, inode, *new_fe_bh,
-				      OCFS2_JOURNAL_ACCESS_CREATE);
+	status = ocfs2_journal_access_di(handle, inode, *new_fe_bh,
+					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -421,17 +508,11 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	fe->i_blkno = cpu_to_le64(fe_blkno);
 	fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
 	fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot);
-	fe->i_uid = cpu_to_le32(current_fsuid());
-	if (dir->i_mode & S_ISGID) {
-		fe->i_gid = cpu_to_le32(dir->i_gid);
-		if (S_ISDIR(mode))
-			mode |= S_ISGID;
-	} else
-		fe->i_gid = cpu_to_le32(current_fsgid());
-	fe->i_mode = cpu_to_le16(mode);
-	if (S_ISCHR(mode) || S_ISBLK(mode))
+	fe->i_uid = cpu_to_le32(inode->i_uid);
+	fe->i_gid = cpu_to_le32(inode->i_gid);
+	fe->i_mode = cpu_to_le16(inode->i_mode);
+	if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode))
 		fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
-
 	fe->i_links_count = cpu_to_le16(inode->i_nlink);
 
 	fe->i_last_eb_blk = 0;
@@ -446,7 +527,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	/*
 	 * If supported, directories start with inline data.
 	 */
-	if (S_ISDIR(mode) && ocfs2_supports_inline_data(osb)) {
+	if (S_ISDIR(inode->i_mode) && ocfs2_supports_inline_data(osb)) {
 		u16 feat = le16_to_cpu(fe->i_dyn_features);
 
 		fe->i_dyn_features = cpu_to_le16(feat | OCFS2_INLINE_DATA_FL);
@@ -465,15 +546,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 		goto leave;
 	}
 
-	if (ocfs2_populate_inode(inode, fe, 1) < 0) {
-		mlog(ML_ERROR, "populate inode failed! bh->b_blocknr=%llu, "
-		     "i_blkno=%llu, i_ino=%lu\n",
-		     (unsigned long long)(*new_fe_bh)->b_blocknr,
-		     (unsigned long long)le64_to_cpu(fe->i_blkno),
-		     inode->i_ino);
-		BUG();
-	}
-
+	ocfs2_populate_inode(inode, fe, 1);
 	ocfs2_inode_set_new(osb, inode);
 	if (!ocfs2_mount_local(osb)) {
 		status = ocfs2_create_new_inode_locks(inode);
@@ -484,17 +557,12 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	status = 0; /* error in ocfs2_create_new_inode_locks is not
 		     * critical */
 
-	*ret_inode = inode;
 leave:
 	if (status < 0) {
 		if (*new_fe_bh) {
 			brelse(*new_fe_bh);
 			*new_fe_bh = NULL;
 		}
-		if (inode) {
-			clear_nlink(inode);
-			iput(inode);
-		}
 	}
 
 	mlog_exit(status);
@@ -588,7 +656,7 @@ static int ocfs2_link(struct dentry *old_dentry,
 		goto out_unlock_inode;
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_LINK_CREDITS);
+	handle = ocfs2_start_trans(osb, ocfs2_link_credits(osb->sb));
 	if (IS_ERR(handle)) {
 		err = PTR_ERR(handle);
 		handle = NULL;
@@ -596,8 +664,8 @@ static int ocfs2_link(struct dentry *old_dentry,
 		goto out_unlock_inode;
 	}
 
-	err = ocfs2_journal_access(handle, inode, fe_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	err = ocfs2_journal_access_di(handle, inode, fe_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (err < 0) {
 		mlog_errno(err);
 		goto out_commit;
@@ -775,7 +843,7 @@ static int ocfs2_unlink(struct inode *dir,
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_UNLINK_CREDITS);
+	handle = ocfs2_start_trans(osb, ocfs2_unlink_credits(osb->sb));
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		handle = NULL;
@@ -783,8 +851,8 @@ static int ocfs2_unlink(struct inode *dir,
 		goto leave;
 	}
 
-	status = ocfs2_journal_access(handle, inode, fe_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, inode, fe_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -1181,7 +1249,7 @@ static int ocfs2_rename(struct inode *old_dir,
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_RENAME_CREDITS);
+	handle = ocfs2_start_trans(osb, ocfs2_rename_credits(osb->sb));
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		handle = NULL;
@@ -1197,8 +1265,8 @@ static int ocfs2_rename(struct inode *old_dir,
 				goto bail;
 			}
 		}
-		status = ocfs2_journal_access(handle, new_inode, newfe_bh,
-					      OCFS2_JOURNAL_ACCESS_WRITE);
+		status = ocfs2_journal_access_di(handle, new_inode, newfe_bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -1244,8 +1312,8 @@ static int ocfs2_rename(struct inode *old_dir,
 	old_inode->i_ctime = CURRENT_TIME;
 	mark_inode_dirty(old_inode);
 
-	status = ocfs2_journal_access(handle, old_inode, old_inode_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, old_inode, old_inode_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status >= 0) {
 		old_di = (struct ocfs2_dinode *) old_inode_bh->b_data;
 
@@ -1321,9 +1389,9 @@ static int ocfs2_rename(struct inode *old_dir,
 			     (int)old_dir_nlink, old_dir->i_nlink);
 		} else {
 			struct ocfs2_dinode *fe;
-			status = ocfs2_journal_access(handle, old_dir,
-						      old_dir_bh,
-						      OCFS2_JOURNAL_ACCESS_WRITE);
+			status = ocfs2_journal_access_di(handle, old_dir,
+							 old_dir_bh,
+							 OCFS2_JOURNAL_ACCESS_WRITE);
 			fe = (struct ocfs2_dinode *) old_dir_bh->b_data;
 			fe->i_links_count = cpu_to_le16(old_dir->i_nlink);
 			status = ocfs2_journal_dirty(handle, old_dir_bh);
@@ -1496,6 +1564,13 @@ static int ocfs2_symlink(struct inode *dir,
 	handle_t *handle = NULL;
 	struct ocfs2_alloc_context *inode_ac = NULL;
 	struct ocfs2_alloc_context *data_ac = NULL;
+	struct ocfs2_alloc_context *xattr_ac = NULL;
+	int want_clusters = 0;
+	int xattr_credits = 0;
+	struct ocfs2_security_xattr_info si = {
+		.enable = 1,
+	};
+	int did_quota = 0, did_quota_inode = 0;
 
 	mlog_entry("(0x%p, 0x%p, symname='%s' actual='%.*s')\n", dir,
 		   dentry, symname, dentry->d_name.len, dentry->d_name.name);
@@ -1542,17 +1617,46 @@ static int ocfs2_symlink(struct inode *dir,
 		goto bail;
 	}
 
-	/* don't reserve bitmap space for fast symlinks. */
-	if (l > ocfs2_fast_symlink_chars(sb)) {
-		status = ocfs2_reserve_clusters(osb, 1, &data_ac);
+	inode = ocfs2_get_init_inode(dir, S_IFLNK | S_IRWXUGO);
+	if (!inode) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto bail;
+	}
+
+	/* get security xattr */
+	status = ocfs2_init_security_get(inode, dir, &si);
+	if (status) {
+		if (status == -EOPNOTSUPP)
+			si.enable = 0;
+		else {
+			mlog_errno(status);
+			goto bail;
+		}
+	}
+
+	/* calculate meta data/clusters for setting security xattr */
+	if (si.enable) {
+		status = ocfs2_calc_security_init(dir, &si, &want_clusters,
+						  &xattr_credits, &xattr_ac);
 		if (status < 0) {
-			if (status != -ENOSPC)
-				mlog_errno(status);
+			mlog_errno(status);
 			goto bail;
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, credits);
+	/* don't reserve bitmap space for fast symlinks. */
+	if (l > ocfs2_fast_symlink_chars(sb))
+		want_clusters += 1;
+
+	status = ocfs2_reserve_clusters(osb, want_clusters, &data_ac);
+	if (status < 0) {
+		if (status != -ENOSPC)
+			mlog_errno(status);
+		goto bail;
+	}
+
+	handle = ocfs2_start_trans(osb, credits + xattr_credits);
 	if (IS_ERR(handle)) {
 		status = PTR_ERR(handle);
 		handle = NULL;
@@ -1560,10 +1664,18 @@ static int ocfs2_symlink(struct inode *dir,
 		goto bail;
 	}
 
-	status = ocfs2_mknod_locked(osb, dir, dentry,
-				    S_IFLNK | S_IRWXUGO, 0,
-				    &new_fe_bh, parent_fe_bh, handle,
-				    &inode, inode_ac);
+	/* We don't use standard VFS wrapper because we don't want vfs_dq_init
+	 * to be called. */
+	if (sb_any_quota_active(osb->sb) &&
+	    osb->sb->dq_op->alloc_inode(inode, 1) == NO_QUOTA) {
+		status = -EDQUOT;
+		goto bail;
+	}
+	did_quota_inode = 1;
+
+	status = ocfs2_mknod_locked(osb, dir, inode, dentry,
+				    0, &new_fe_bh, parent_fe_bh, handle,
+				    inode_ac);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1576,6 +1688,12 @@ static int ocfs2_symlink(struct inode *dir,
 		u32 offset = 0;
 
 		inode->i_op = &ocfs2_symlink_inode_operations;
+		if (vfs_dq_alloc_space_nodirty(inode,
+		    ocfs2_clusters_to_bytes(osb->sb, 1))) {
+			status = -EDQUOT;
+			goto bail;
+		}
+		did_quota = 1;
 		status = ocfs2_add_inode_data(osb, inode, &offset, 1, 0,
 					      new_fe_bh,
 					      handle, data_ac, NULL,
@@ -1614,6 +1732,15 @@ static int ocfs2_symlink(struct inode *dir,
 		}
 	}
 
+	if (si.enable) {
+		status = ocfs2_init_security_set(handle, inode, new_fe_bh, &si,
+						 xattr_ac, data_ac);
+		if (status < 0) {
+			mlog_errno(status);
+			goto bail;
+		}
+	}
+
 	status = ocfs2_add_entry(handle, dentry, inode,
 				 le64_to_cpu(fe->i_blkno), parent_fe_bh,
 				 de_bh);
@@ -1632,6 +1759,11 @@ static int ocfs2_symlink(struct inode *dir,
 	dentry->d_op = &ocfs2_dentry_ops;
 	d_instantiate(dentry, inode);
 bail:
+	if (status < 0 && did_quota)
+		vfs_dq_free_space_nodirty(inode,
+					ocfs2_clusters_to_bytes(osb->sb, 1));
+	if (status < 0 && did_quota_inode)
+		vfs_dq_free_inode(inode);
 	if (handle)
 		ocfs2_commit_trans(osb, handle);
 
@@ -1640,12 +1772,18 @@ bail:
 	brelse(new_fe_bh);
 	brelse(parent_fe_bh);
 	brelse(de_bh);
+	kfree(si.name);
+	kfree(si.value);
 	if (inode_ac)
 		ocfs2_free_alloc_context(inode_ac);
 	if (data_ac)
 		ocfs2_free_alloc_context(data_ac);
-	if ((status < 0) && inode)
+	if (xattr_ac)
+		ocfs2_free_alloc_context(xattr_ac);
+	if ((status < 0) && inode) {
+		clear_nlink(inode);
 		iput(inode);
+	}
 
 	mlog_exit(status);
 
@@ -1754,16 +1892,14 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
 
 	mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);
 
-	status = ocfs2_read_block(orphan_dir_inode,
-				  OCFS2_I(orphan_dir_inode)->ip_blkno,
-				  &orphan_dir_bh);
+	status = ocfs2_read_inode_block(orphan_dir_inode, &orphan_dir_bh);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
 	}
 
-	status = ocfs2_journal_access(handle, orphan_dir_inode, orphan_dir_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, orphan_dir_inode, orphan_dir_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
@@ -1850,8 +1986,8 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
 		goto leave;
 	}
 
-	status = ocfs2_journal_access(handle,orphan_dir_inode,  orphan_dir_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle,orphan_dir_inode,  orphan_dir_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index 3fed9e3d899..ad5c24a29ed 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -161,6 +161,7 @@ enum ocfs2_vol_state
 {
 	VOLUME_INIT = 0,
 	VOLUME_MOUNTED,
+	VOLUME_MOUNTED_QUOTAS,
 	VOLUME_DISMOUNTED,
 	VOLUME_DISABLED
 };
@@ -195,6 +196,9 @@ enum ocfs2_mount_options
 	OCFS2_MOUNT_LOCALFLOCKS = 1 << 5, /* No cluster aware user file locks */
 	OCFS2_MOUNT_NOUSERXATTR = 1 << 6, /* No user xattr */
 	OCFS2_MOUNT_INODE64 = 1 << 7,	/* Allow inode numbers > 2^32 */
+	OCFS2_MOUNT_POSIX_ACL = 1 << 8,	/* POSIX access control lists */
+	OCFS2_MOUNT_USRQUOTA = 1 << 9, /* We support user quotas */
+	OCFS2_MOUNT_GRPQUOTA = 1 << 10, /* We support group quotas */
 };
 
 #define OCFS2_OSB_SOFT_RO	0x0001
@@ -205,6 +209,7 @@ enum ocfs2_mount_options
 struct ocfs2_journal;
 struct ocfs2_slot_info;
 struct ocfs2_recovery_map;
+struct ocfs2_quota_recovery;
 struct ocfs2_super
 {
 	struct task_struct *commit_task;
@@ -286,10 +291,11 @@ struct ocfs2_super
 	char *local_alloc_debug_buf;
 #endif
 
-	/* Next two fields are for local node slot recovery during
+	/* Next three fields are for local node slot recovery during
 	 * mount. */
 	int dirty;
 	struct ocfs2_dinode *local_alloc_copy;
+	struct ocfs2_quota_recovery *quota_rec;
 
 	struct ocfs2_alloc_stats alloc_stats;
 	char dev_str[20];		/* "major,minor" of the device */
@@ -333,6 +339,10 @@ struct ocfs2_super
 
 #define OCFS2_SB(sb)	    ((struct ocfs2_super *)(sb)->s_fs_info)
 
+/* Useful typedef for passing around journal access functions */
+typedef int (*ocfs2_journal_access_func)(handle_t *handle, struct inode *inode,
+					 struct buffer_head *bh, int type);
+
 static inline int ocfs2_should_order_data(struct inode *inode)
 {
 	if (!S_ISREG(inode->i_mode))
@@ -376,6 +386,13 @@ static inline int ocfs2_supports_xattr(struct ocfs2_super *osb)
 	return 0;
 }
 
+static inline int ocfs2_meta_ecc(struct ocfs2_super *osb)
+{
+	if (osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_META_ECC)
+		return 1;
+	return 0;
+}
+
 /* set / clear functions because cluster events can make these happen
  * in parallel so we want the transitions to be atomic. this also
  * means that any future flags osb_flags must be protected by spinlock
@@ -443,39 +460,19 @@ static inline int ocfs2_uses_extended_slot_map(struct ocfs2_super *osb)
 #define OCFS2_IS_VALID_DINODE(ptr)					\
 	(!strcmp((ptr)->i_signature, OCFS2_INODE_SIGNATURE))
 
-#define OCFS2_RO_ON_INVALID_DINODE(__sb, __di)	do {			\
-	typeof(__di) ____di = (__di);					\
-	ocfs2_error((__sb), 						\
-		"Dinode # %llu has bad signature %.*s",			\
-		(unsigned long long)le64_to_cpu((____di)->i_blkno), 7, 	\
-		(____di)->i_signature);					\
-} while (0)
-
 #define OCFS2_IS_VALID_EXTENT_BLOCK(ptr)				\
 	(!strcmp((ptr)->h_signature, OCFS2_EXTENT_BLOCK_SIGNATURE))
 
-#define OCFS2_RO_ON_INVALID_EXTENT_BLOCK(__sb, __eb)	do {		\
-	typeof(__eb) ____eb = (__eb);					\
-	ocfs2_error((__sb), 						\
-		"Extent Block # %llu has bad signature %.*s",		\
-		(unsigned long long)le64_to_cpu((____eb)->h_blkno), 7,	\
-		(____eb)->h_signature);					\
-} while (0)
-
 #define OCFS2_IS_VALID_GROUP_DESC(ptr)					\
 	(!strcmp((ptr)->bg_signature, OCFS2_GROUP_DESC_SIGNATURE))
 
-#define OCFS2_RO_ON_INVALID_GROUP_DESC(__sb, __gd)	do {		\
-	typeof(__gd) ____gd = (__gd);					\
-		ocfs2_error((__sb),					\
-		"Group Descriptor # %llu has bad signature %.*s",	\
-		(unsigned long long)le64_to_cpu((____gd)->bg_blkno), 7, \
-		(____gd)->bg_signature);				\
-} while (0)
 
 #define OCFS2_IS_VALID_XATTR_BLOCK(ptr)					\
 	(!strcmp((ptr)->xb_signature, OCFS2_XATTR_BLOCK_SIGNATURE))
 
+#define OCFS2_IS_VALID_DIR_TRAILER(ptr)					\
+	(!strcmp((ptr)->db_signature, OCFS2_DIR_TRAILER_SIGNATURE))
+
 static inline unsigned long ino_from_blkno(struct super_block *sb,
 					   u64 blkno)
 {
@@ -632,5 +629,6 @@ static inline s16 ocfs2_get_inode_steal_slot(struct ocfs2_super *osb)
 #define ocfs2_clear_bit ext2_clear_bit
 #define ocfs2_test_bit ext2_test_bit
 #define ocfs2_find_next_zero_bit ext2_find_next_zero_bit
+#define ocfs2_find_next_bit ext2_find_next_bit
 #endif  /* OCFS2_H */
 
diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h
index 5e0c0d0aef7..c7ae45aaa36 100644
--- a/fs/ocfs2/ocfs2_fs.h
+++ b/fs/ocfs2/ocfs2_fs.h
@@ -65,6 +65,7 @@
 #define OCFS2_EXTENT_BLOCK_SIGNATURE	"EXBLK01"
 #define OCFS2_GROUP_DESC_SIGNATURE      "GROUP01"
 #define OCFS2_XATTR_BLOCK_SIGNATURE	"XATTR01"
+#define OCFS2_DIR_TRAILER_SIGNATURE	"DIRTRL1"
 
 /* Compatibility flags */
 #define OCFS2_HAS_COMPAT_FEATURE(sb,mask)			\
@@ -93,8 +94,11 @@
 					 | OCFS2_FEATURE_INCOMPAT_INLINE_DATA \
 					 | OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP \
 					 | OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK \
-					 | OCFS2_FEATURE_INCOMPAT_XATTR)
-#define OCFS2_FEATURE_RO_COMPAT_SUPP	OCFS2_FEATURE_RO_COMPAT_UNWRITTEN
+					 | OCFS2_FEATURE_INCOMPAT_XATTR \
+					 | OCFS2_FEATURE_INCOMPAT_META_ECC)
+#define OCFS2_FEATURE_RO_COMPAT_SUPP	(OCFS2_FEATURE_RO_COMPAT_UNWRITTEN \
+					 | OCFS2_FEATURE_RO_COMPAT_USRQUOTA \
+					 | OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)
 
 /*
  * Heartbeat-only devices are missing journals and other files.  The
@@ -147,6 +151,9 @@
 /* Support for extended attributes */
 #define OCFS2_FEATURE_INCOMPAT_XATTR		0x0200
 
+/* Metadata checksum and error correction */
+#define OCFS2_FEATURE_INCOMPAT_META_ECC		0x0800
+
 /*
  * backup superblock flag is used to indicate that this volume
  * has backup superblocks.
@@ -163,6 +170,12 @@
  */
 #define OCFS2_FEATURE_RO_COMPAT_UNWRITTEN	0x0001
 
+/*
+ * Maintain quota information for this filesystem
+ */
+#define OCFS2_FEATURE_RO_COMPAT_USRQUOTA	0x0002
+#define OCFS2_FEATURE_RO_COMPAT_GRPQUOTA	0x0004
+
 /* The byte offset of the first backup block will be 1G.
  * The following will be 4G, 16G, 64G, 256G and 1T.
  */
@@ -192,6 +205,7 @@
 #define OCFS2_HEARTBEAT_FL	(0x00000200)	/* Heartbeat area */
 #define OCFS2_CHAIN_FL		(0x00000400)	/* Chain allocator */
 #define OCFS2_DEALLOC_FL	(0x00000800)	/* Truncate log */
+#define OCFS2_QUOTA_FL		(0x00001000)	/* Quota file */
 
 /*
  * Flags on ocfs2_dinode.i_dyn_features
@@ -329,13 +343,17 @@ enum {
 #define OCFS2_FIRST_ONLINE_SYSTEM_INODE SLOT_MAP_SYSTEM_INODE
 	HEARTBEAT_SYSTEM_INODE,
 	GLOBAL_BITMAP_SYSTEM_INODE,
-#define OCFS2_LAST_GLOBAL_SYSTEM_INODE GLOBAL_BITMAP_SYSTEM_INODE
+	USER_QUOTA_SYSTEM_INODE,
+	GROUP_QUOTA_SYSTEM_INODE,
+#define OCFS2_LAST_GLOBAL_SYSTEM_INODE GROUP_QUOTA_SYSTEM_INODE
 	ORPHAN_DIR_SYSTEM_INODE,
 	EXTENT_ALLOC_SYSTEM_INODE,
 	INODE_ALLOC_SYSTEM_INODE,
 	JOURNAL_SYSTEM_INODE,
 	LOCAL_ALLOC_SYSTEM_INODE,
 	TRUNCATE_LOG_SYSTEM_INODE,
+	LOCAL_USER_QUOTA_SYSTEM_INODE,
+	LOCAL_GROUP_QUOTA_SYSTEM_INODE,
 	NUM_SYSTEM_INODES
 };
 
@@ -349,6 +367,8 @@ static struct ocfs2_system_inode_info ocfs2_system_inodes[NUM_SYSTEM_INODES] = {
 	[SLOT_MAP_SYSTEM_INODE]			= { "slot_map", 0, S_IFREG | 0644 },
 	[HEARTBEAT_SYSTEM_INODE]		= { "heartbeat", OCFS2_HEARTBEAT_FL, S_IFREG | 0644 },
 	[GLOBAL_BITMAP_SYSTEM_INODE]		= { "global_bitmap", 0, S_IFREG | 0644 },
+	[USER_QUOTA_SYSTEM_INODE]		= { "aquota.user", OCFS2_QUOTA_FL, S_IFREG | 0644 },
+	[GROUP_QUOTA_SYSTEM_INODE]		= { "aquota.group", OCFS2_QUOTA_FL, S_IFREG | 0644 },
 
 	/* Slot-specific system inodes (one copy per slot) */
 	[ORPHAN_DIR_SYSTEM_INODE]		= { "orphan_dir:%04d", 0, S_IFDIR | 0755 },
@@ -356,7 +376,9 @@ static struct ocfs2_system_inode_info ocfs2_system_inodes[NUM_SYSTEM_INODES] = {
 	[INODE_ALLOC_SYSTEM_INODE]		= { "inode_alloc:%04d", OCFS2_BITMAP_FL | OCFS2_CHAIN_FL, S_IFREG | 0644 },
 	[JOURNAL_SYSTEM_INODE]			= { "journal:%04d", OCFS2_JOURNAL_FL, S_IFREG | 0644 },
 	[LOCAL_ALLOC_SYSTEM_INODE]		= { "local_alloc:%04d", OCFS2_BITMAP_FL | OCFS2_LOCAL_ALLOC_FL, S_IFREG | 0644 },
-	[TRUNCATE_LOG_SYSTEM_INODE]		= { "truncate_log:%04d", OCFS2_DEALLOC_FL, S_IFREG | 0644 }
+	[TRUNCATE_LOG_SYSTEM_INODE]		= { "truncate_log:%04d", OCFS2_DEALLOC_FL, S_IFREG | 0644 },
+	[LOCAL_USER_QUOTA_SYSTEM_INODE]		= { "aquota.user:%04d", OCFS2_QUOTA_FL, S_IFREG | 0644 },
+	[LOCAL_GROUP_QUOTA_SYSTEM_INODE]	= { "aquota.group:%04d", OCFS2_QUOTA_FL, S_IFREG | 0644 },
 };
 
 /* Parameter passed from mount.ocfs2 to module */
@@ -410,6 +432,22 @@ static unsigned char ocfs2_type_by_mode[S_IFMT >> S_SHIFT] = {
 #define OCFS2_RAW_SB(dinode)		(&((dinode)->id2.i_super))
 
 /*
+ * Block checking structure.  This is used in metadata to validate the
+ * contents.  If OCFS2_FEATURE_INCOMPAT_META_ECC is not set, it is all
+ * zeros.
+ */
+struct ocfs2_block_check {
+/*00*/	__le32 bc_crc32e;	/* 802.3 Ethernet II CRC32 */
+	__le16 bc_ecc;		/* Single-error-correction parity vector.
+				   This is a simple Hamming code dependant
+				   on the blocksize.  OCFS2's maximum
+				   blocksize, 4K, requires 16 parity bits,
+				   so we fit in __le16. */
+	__le16 bc_reserved1;
+/*08*/
+};
+
+/*
  * On disk extent record for OCFS2
  * It describes a range of clusters on disk.
  *
@@ -496,7 +534,7 @@ struct ocfs2_truncate_log {
 struct ocfs2_extent_block
 {
 /*00*/	__u8 h_signature[8];		/* Signature for verification */
-	__le64 h_reserved1;
+	struct ocfs2_block_check h_check;	/* Error checking */
 /*10*/	__le16 h_suballoc_slot;		/* Slot suballocator this
 					   extent_header belongs to */
 	__le16 h_suballoc_bit;		/* Bit offset in suballocator
@@ -666,7 +704,8 @@ struct ocfs2_dinode {
 					   was set in i_flags */
 	__le16 i_dyn_features;
 	__le64 i_xattr_loc;
-/*80*/	__le64 i_reserved2[7];
+/*80*/	struct ocfs2_block_check i_check;	/* Error checking */
+/*88*/	__le64 i_reserved2[6];
 /*B8*/	union {
 		__le64 i_pad1;		/* Generic way to refer to this
 					   64bit union */
@@ -715,6 +754,34 @@ struct ocfs2_dir_entry {
 } __attribute__ ((packed));
 
 /*
+ * Per-block record for the unindexed directory btree. This is carefully
+ * crafted so that the rec_len and name_len records of an ocfs2_dir_entry are
+ * mirrored. That way, the directory manipulation code needs a minimal amount
+ * of update.
+ *
+ * NOTE: Keep this structure aligned to a multiple of 4 bytes.
+ */
+struct ocfs2_dir_block_trailer {
+/*00*/	__le64		db_compat_inode;	/* Always zero. Was inode */
+
+	__le16		db_compat_rec_len;	/* Backwards compatible with
+						 * ocfs2_dir_entry. */
+	__u8		db_compat_name_len;	/* Always zero. Was name_len */
+	__u8		db_reserved0;
+	__le16		db_reserved1;
+	__le16		db_free_rec_len;	/* Size of largest empty hole
+						 * in this block. (unused) */
+/*10*/	__u8		db_signature[8];	/* Signature for verification */
+	__le64		db_reserved2;
+	__le64		db_free_next;		/* Next block in list (unused) */
+/*20*/	__le64		db_blkno;		/* Offset on disk, in blocks */
+	__le64		db_parent_dinode;	/* dinode which owns me, in
+						   blocks */
+/*30*/	struct ocfs2_block_check db_check;	/* Error checking */
+/*40*/
+};
+
+/*
  * On disk allocator group structure for OCFS2
  */
 struct ocfs2_group_desc
@@ -733,7 +800,8 @@ struct ocfs2_group_desc
 /*20*/	__le64   bg_parent_dinode;       /* dinode which owns me, in
 					   blocks */
 	__le64   bg_blkno;               /* Offset on disk, in blocks */
-/*30*/	__le64   bg_reserved2[2];
+/*30*/	struct ocfs2_block_check bg_check;	/* Error checking */
+	__le64   bg_reserved2;
 /*40*/	__u8    bg_bitmap[0];
 };
 
@@ -776,7 +844,12 @@ struct ocfs2_xattr_header {
 						   in this extent record,
 						   only valid in the first
 						   bucket. */
-	__le64  xh_csum;
+	struct ocfs2_block_check xh_check;	/* Error checking
+						   (Note, this is only
+						    used for xattr
+						    buckets.  A block uses
+						    xb_check and sets
+						    this field to zero.) */
 	struct ocfs2_xattr_entry xh_entries[0]; /* xattr entry list. */
 };
 
@@ -827,7 +900,7 @@ struct ocfs2_xattr_block {
 					block group */
 	__le32	xb_fs_generation;    /* Must match super block */
 /*10*/	__le64	xb_blkno;            /* Offset on disk, in blocks */
-	__le64	xb_csum;
+	struct ocfs2_block_check xb_check;	/* Error checking */
 /*20*/	__le16	xb_flags;            /* Indicates whether this block contains
 					real xattr or a xattr tree. */
 	__le16	xb_reserved0;
@@ -868,6 +941,128 @@ static inline int ocfs2_xattr_get_type(struct ocfs2_xattr_entry *xe)
 	return xe->xe_type & OCFS2_XATTR_TYPE_MASK;
 }
 
+/*
+ *  On disk structures for global quota file
+ */
+
+/* Magic numbers and known versions for global quota files */
+#define OCFS2_GLOBAL_QMAGICS {\
+	0x0cf52470, /* USRQUOTA */ \
+	0x0cf52471  /* GRPQUOTA */ \
+}
+
+#define OCFS2_GLOBAL_QVERSIONS {\
+	0, \
+	0, \
+}
+
+
+/* Each block of each quota file has a certain fixed number of bytes reserved
+ * for OCFS2 internal use at its end. OCFS2 can use it for things like
+ * checksums, etc. */
+#define OCFS2_QBLK_RESERVED_SPACE 8
+
+/* Generic header of all quota files */
+struct ocfs2_disk_dqheader {
+	__le32 dqh_magic;	/* Magic number identifying file */
+	__le32 dqh_version;	/* Quota format version */
+};
+
+#define OCFS2_GLOBAL_INFO_OFF (sizeof(struct ocfs2_disk_dqheader))
+
+/* Information header of global quota file (immediately follows the generic
+ * header) */
+struct ocfs2_global_disk_dqinfo {
+/*00*/	__le32 dqi_bgrace;	/* Grace time for space softlimit excess */
+	__le32 dqi_igrace;	/* Grace time for inode softlimit excess */
+	__le32 dqi_syncms;	/* Time after which we sync local changes to
+				 * global quota file */
+	__le32 dqi_blocks;	/* Number of blocks in quota file */
+/*10*/	__le32 dqi_free_blk;	/* First free block in quota file */
+	__le32 dqi_free_entry;	/* First block with free dquot entry in quota
+				 * file */
+};
+
+/* Structure with global user / group information. We reserve some space
+ * for future use. */
+struct ocfs2_global_disk_dqblk {
+/*00*/	__le32 dqb_id;          /* ID the structure belongs to */
+	__le32 dqb_use_count;   /* Number of nodes having reference to this structure */
+	__le64 dqb_ihardlimit;  /* absolute limit on allocated inodes */
+/*10*/	__le64 dqb_isoftlimit;  /* preferred inode limit */
+	__le64 dqb_curinodes;   /* current # allocated inodes */
+/*20*/	__le64 dqb_bhardlimit;  /* absolute limit on disk space */
+	__le64 dqb_bsoftlimit;  /* preferred limit on disk space */
+/*30*/	__le64 dqb_curspace;    /* current space occupied */
+	__le64 dqb_btime;       /* time limit for excessive disk use */
+/*40*/	__le64 dqb_itime;       /* time limit for excessive inode use */
+	__le64 dqb_pad1;
+/*50*/	__le64 dqb_pad2;
+};
+
+/*
+ *  On-disk structures for local quota file
+ */
+
+/* Magic numbers and known versions for local quota files */
+#define OCFS2_LOCAL_QMAGICS {\
+	0x0cf524c0, /* USRQUOTA */ \
+	0x0cf524c1  /* GRPQUOTA */ \
+}
+
+#define OCFS2_LOCAL_QVERSIONS {\
+	0, \
+	0, \
+}
+
+/* Quota flags in dqinfo header */
+#define OLQF_CLEAN	0x0001	/* Quota file is empty (this should be after\
+				 * quota has been cleanly turned off) */
+
+#define OCFS2_LOCAL_INFO_OFF (sizeof(struct ocfs2_disk_dqheader))
+
+/* Information header of local quota file (immediately follows the generic
+ * header) */
+struct ocfs2_local_disk_dqinfo {
+	__le32 dqi_flags;	/* Flags for quota file */
+	__le32 dqi_chunks;	/* Number of chunks of quota structures
+				 * with a bitmap */
+	__le32 dqi_blocks;	/* Number of blocks allocated for quota file */
+};
+
+/* Header of one chunk of a quota file */
+struct ocfs2_local_disk_chunk {
+	__le32 dqc_free;	/* Number of free entries in the bitmap */
+	u8 dqc_bitmap[0];	/* Bitmap of entries in the corresponding
+				 * chunk of quota file */
+};
+
+/* One entry in local quota file */
+struct ocfs2_local_disk_dqblk {
+/*00*/	__le64 dqb_id;		/* id this quota applies to */
+	__le64 dqb_spacemod;	/* Change in the amount of used space */
+/*10*/	__le64 dqb_inodemod;	/* Change in the amount of used inodes */
+};
+
+
+/*
+ * The quota trailer lives at the end of each quota block.
+ */
+
+struct ocfs2_disk_dqtrailer {
+/*00*/	struct ocfs2_block_check dq_check;	/* Error checking */
+/*08*/	/* Cannot be larger than OCFS2_QBLK_RESERVED_SPACE */
+};
+
+static inline struct ocfs2_disk_dqtrailer *ocfs2_block_dqtrailer(int blocksize,
+								 void *buf)
+{
+	char *ptr = buf;
+	ptr += blocksize - OCFS2_QBLK_RESERVED_SPACE;
+
+	return (struct ocfs2_disk_dqtrailer *)ptr;
+}
+
 #ifdef __KERNEL__
 static inline int ocfs2_fast_symlink_chars(struct super_block *sb)
 {
diff --git a/fs/ocfs2/ocfs2_jbd_compat.h b/fs/ocfs2/ocfs2_jbd_compat.h
deleted file mode 100644
index b91c78f8f55..00000000000
--- a/fs/ocfs2/ocfs2_jbd_compat.h
+++ /dev/null
@@ -1,82 +0,0 @@
-/* -*- mode: c; c-basic-offset: 8; -*-
- * vim: noexpandtab sw=8 ts=8 sts=0:
- *
- * ocfs2_jbd_compat.h
- *
- * Compatibility defines for JBD.
- *
- * Copyright (C) 2008 Oracle.  All rights reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public
- * License version 2 as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- */
-
-#ifndef OCFS2_JBD_COMPAT_H
-#define OCFS2_JBD_COMPAT_H
-
-#ifndef CONFIG_OCFS2_COMPAT_JBD
-# error Should not have been included
-#endif
-
-struct jbd2_inode {
-	unsigned int dummy;
-};
-
-#define JBD2_BARRIER			JFS_BARRIER
-#define JBD2_DEFAULT_MAX_COMMIT_AGE	JBD_DEFAULT_MAX_COMMIT_AGE
-
-#define jbd2_journal_ack_err			journal_ack_err
-#define jbd2_journal_clear_err			journal_clear_err
-#define jbd2_journal_destroy			journal_destroy
-#define jbd2_journal_dirty_metadata		journal_dirty_metadata
-#define jbd2_journal_errno			journal_errno
-#define jbd2_journal_extend			journal_extend
-#define jbd2_journal_flush			journal_flush
-#define jbd2_journal_force_commit		journal_force_commit
-#define jbd2_journal_get_write_access		journal_get_write_access
-#define jbd2_journal_get_undo_access		journal_get_undo_access
-#define jbd2_journal_init_inode			journal_init_inode
-#define jbd2_journal_invalidatepage		journal_invalidatepage
-#define jbd2_journal_load			journal_load
-#define jbd2_journal_lock_updates		journal_lock_updates
-#define jbd2_journal_restart			journal_restart
-#define jbd2_journal_start			journal_start
-#define jbd2_journal_start_commit		journal_start_commit
-#define jbd2_journal_stop			journal_stop
-#define jbd2_journal_try_to_free_buffers	journal_try_to_free_buffers
-#define jbd2_journal_unlock_updates		journal_unlock_updates
-#define jbd2_journal_wipe			journal_wipe
-#define jbd2_log_wait_commit			log_wait_commit
-
-static inline int jbd2_journal_file_inode(handle_t *handle,
-					  struct jbd2_inode *inode)
-{
-	return 0;
-}
-
-static inline int jbd2_journal_begin_ordered_truncate(struct jbd2_inode *inode,
-						      loff_t new_size)
-{
-	return 0;
-}
-
-static inline void jbd2_journal_init_jbd_inode(struct jbd2_inode *jinode,
-					       struct inode *inode)
-{
-	return;
-}
-
-static inline void jbd2_journal_release_jbd_inode(journal_t *journal,
-						  struct jbd2_inode *jinode)
-{
-	return;
-}
-
-
-#endif  /* OCFS2_JBD_COMPAT_H */
diff --git a/fs/ocfs2/ocfs2_lockid.h b/fs/ocfs2/ocfs2_lockid.h
index 82c200f7a8f..eb6f50c9cec 100644
--- a/fs/ocfs2/ocfs2_lockid.h
+++ b/fs/ocfs2/ocfs2_lockid.h
@@ -46,6 +46,7 @@ enum ocfs2_lock_type {
 	OCFS2_LOCK_TYPE_DENTRY,
 	OCFS2_LOCK_TYPE_OPEN,
 	OCFS2_LOCK_TYPE_FLOCK,
+	OCFS2_LOCK_TYPE_QINFO,
 	OCFS2_NUM_LOCK_TYPES
 };
 
@@ -77,6 +78,9 @@ static inline char ocfs2_lock_type_char(enum ocfs2_lock_type type)
 		case OCFS2_LOCK_TYPE_FLOCK:
 			c = 'F';
 			break;
+		case OCFS2_LOCK_TYPE_QINFO:
+			c = 'Q';
+			break;
 		default:
 			c = '\0';
 	}
@@ -95,6 +99,7 @@ static char *ocfs2_lock_type_strings[] = {
 	[OCFS2_LOCK_TYPE_DENTRY] = "Dentry",
 	[OCFS2_LOCK_TYPE_OPEN] = "Open",
 	[OCFS2_LOCK_TYPE_FLOCK] = "Flock",
+	[OCFS2_LOCK_TYPE_QINFO] = "Quota",
 };
 
 static inline const char *ocfs2_lock_type_string(enum ocfs2_lock_type type)
diff --git a/fs/ocfs2/quota.h b/fs/ocfs2/quota.h
new file mode 100644
index 00000000000..7365e2e0870
--- /dev/null
+++ b/fs/ocfs2/quota.h
@@ -0,0 +1,119 @@
+/*
+ * quota.h for OCFS2
+ *
+ * On disk quota structures for local and global quota file, in-memory
+ * structures.
+ *
+ */
+
+#ifndef _OCFS2_QUOTA_H
+#define _OCFS2_QUOTA_H
+
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <linux/quota.h>
+#include <linux/list.h>
+#include <linux/dqblk_qtree.h>
+
+#include "ocfs2.h"
+
+/* Common stuff */
+/* id number of quota format */
+#define QFMT_OCFS2 3
+
+/*
+ * In-memory structures
+ */
+struct ocfs2_dquot {
+	struct dquot dq_dquot;	/* Generic VFS dquot */
+	loff_t dq_local_off;	/* Offset in the local quota file */
+	struct ocfs2_quota_chunk *dq_chunk;	/* Chunk dquot is in */
+	unsigned int dq_use_count;	/* Number of nodes having reference to this entry in global quota file */
+	s64 dq_origspace;	/* Last globally synced space usage */
+	s64 dq_originodes;	/* Last globally synced inode usage */
+};
+
+/* Description of one chunk to recover in memory */
+struct ocfs2_recovery_chunk {
+	struct list_head rc_list;	/* List of chunks */
+	int rc_chunk;			/* Chunk number */
+	unsigned long *rc_bitmap;	/* Bitmap of entries to recover */
+};
+
+struct ocfs2_quota_recovery {
+	struct list_head r_list[MAXQUOTAS];	/* List of chunks to recover */
+};
+
+/* In-memory structure with quota header information */
+struct ocfs2_mem_dqinfo {
+	unsigned int dqi_type;		/* Quota type this structure describes */
+	unsigned int dqi_chunks;	/* Number of chunks in local quota file */
+	unsigned int dqi_blocks;	/* Number of blocks allocated for local quota file */
+	unsigned int dqi_syncms;	/* How often should we sync with other nodes */
+	unsigned int dqi_syncjiff;	/* Precomputed dqi_syncms in jiffies */
+	struct list_head dqi_chunk;	/* List of chunks */
+	struct inode *dqi_gqinode;	/* Global quota file inode */
+	struct ocfs2_lock_res dqi_gqlock;	/* Lock protecting quota information structure */
+	struct buffer_head *dqi_gqi_bh;	/* Buffer head with global quota file inode - set only if inode lock is obtained */
+	int dqi_gqi_count;		/* Number of holders of dqi_gqi_bh */
+	struct buffer_head *dqi_lqi_bh;	/* Buffer head with local quota file inode */
+	struct buffer_head *dqi_ibh;	/* Buffer with information header */
+	struct qtree_mem_dqinfo dqi_gi;	/* Info about global file */
+	struct delayed_work dqi_sync_work;	/* Work for syncing dquots */
+	struct ocfs2_quota_recovery *dqi_rec;	/* Pointer to recovery
+						 * information, in case we
+						 * enable quotas on file
+						 * needing it */
+};
+
+static inline struct ocfs2_dquot *OCFS2_DQUOT(struct dquot *dquot)
+{
+	return container_of(dquot, struct ocfs2_dquot, dq_dquot);
+}
+
+struct ocfs2_quota_chunk {
+	struct list_head qc_chunk;	/* List of quotafile chunks */
+	int qc_num;			/* Number of quota chunk */
+	struct buffer_head *qc_headerbh;	/* Buffer head with chunk header */
+};
+
+extern struct kmem_cache *ocfs2_dquot_cachep;
+extern struct kmem_cache *ocfs2_qf_chunk_cachep;
+
+extern struct qtree_fmt_operations ocfs2_global_ops;
+
+struct ocfs2_quota_recovery *ocfs2_begin_quota_recovery(
+				struct ocfs2_super *osb, int slot_num);
+int ocfs2_finish_quota_recovery(struct ocfs2_super *osb,
+				struct ocfs2_quota_recovery *rec,
+				int slot_num);
+void ocfs2_free_quota_recovery(struct ocfs2_quota_recovery *rec);
+ssize_t ocfs2_quota_read(struct super_block *sb, int type, char *data,
+			 size_t len, loff_t off);
+ssize_t ocfs2_quota_write(struct super_block *sb, int type,
+			  const char *data, size_t len, loff_t off);
+int ocfs2_global_read_info(struct super_block *sb, int type);
+int ocfs2_global_write_info(struct super_block *sb, int type);
+int ocfs2_global_read_dquot(struct dquot *dquot);
+int __ocfs2_sync_dquot(struct dquot *dquot, int freeing);
+static inline int ocfs2_sync_dquot(struct dquot *dquot)
+{
+	return __ocfs2_sync_dquot(dquot, 0);
+}
+static inline int ocfs2_global_release_dquot(struct dquot *dquot)
+{
+	return __ocfs2_sync_dquot(dquot, 1);
+}
+
+int ocfs2_lock_global_qf(struct ocfs2_mem_dqinfo *oinfo, int ex);
+void ocfs2_unlock_global_qf(struct ocfs2_mem_dqinfo *oinfo, int ex);
+int ocfs2_read_quota_block(struct inode *inode, u64 v_block,
+			   struct buffer_head **bh);
+
+extern struct dquot_operations ocfs2_quota_operations;
+extern struct quota_format_type ocfs2_quota_format;
+
+int ocfs2_quota_setup(void);
+void ocfs2_quota_shutdown(void);
+
+#endif /* _OCFS2_QUOTA_H */
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
new file mode 100644
index 00000000000..6aff8f2d3e4
--- /dev/null
+++ b/fs/ocfs2/quota_global.c
@@ -0,0 +1,1025 @@
+/*
+ *  Implementation of operations over global quota file
+ */
+#include <linux/spinlock.h>
+#include <linux/fs.h>
+#include <linux/quota.h>
+#include <linux/quotaops.h>
+#include <linux/dqblk_qtree.h>
+#include <linux/jiffies.h>
+#include <linux/writeback.h>
+#include <linux/workqueue.h>
+
+#define MLOG_MASK_PREFIX ML_QUOTA
+#include <cluster/masklog.h>
+
+#include "ocfs2_fs.h"
+#include "ocfs2.h"
+#include "alloc.h"
+#include "blockcheck.h"
+#include "inode.h"
+#include "journal.h"
+#include "file.h"
+#include "sysfile.h"
+#include "dlmglue.h"
+#include "uptodate.h"
+#include "quota.h"
+
+static struct workqueue_struct *ocfs2_quota_wq = NULL;
+
+static void qsync_work_fn(struct work_struct *work);
+
+static void ocfs2_global_disk2memdqb(struct dquot *dquot, void *dp)
+{
+	struct ocfs2_global_disk_dqblk *d = dp;
+	struct mem_dqblk *m = &dquot->dq_dqb;
+
+	/* Update from disk only entries not set by the admin */
+	if (!test_bit(DQ_LASTSET_B + QIF_ILIMITS_B, &dquot->dq_flags)) {
+		m->dqb_ihardlimit = le64_to_cpu(d->dqb_ihardlimit);
+		m->dqb_isoftlimit = le64_to_cpu(d->dqb_isoftlimit);
+	}
+	if (!test_bit(DQ_LASTSET_B + QIF_INODES_B, &dquot->dq_flags))
+		m->dqb_curinodes = le64_to_cpu(d->dqb_curinodes);
+	if (!test_bit(DQ_LASTSET_B + QIF_BLIMITS_B, &dquot->dq_flags)) {
+		m->dqb_bhardlimit = le64_to_cpu(d->dqb_bhardlimit);
+		m->dqb_bsoftlimit = le64_to_cpu(d->dqb_bsoftlimit);
+	}
+	if (!test_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags))
+		m->dqb_curspace = le64_to_cpu(d->dqb_curspace);
+	if (!test_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags))
+		m->dqb_btime = le64_to_cpu(d->dqb_btime);
+	if (!test_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags))
+		m->dqb_itime = le64_to_cpu(d->dqb_itime);
+	OCFS2_DQUOT(dquot)->dq_use_count = le32_to_cpu(d->dqb_use_count);
+}
+
+static void ocfs2_global_mem2diskdqb(void *dp, struct dquot *dquot)
+{
+	struct ocfs2_global_disk_dqblk *d = dp;
+	struct mem_dqblk *m = &dquot->dq_dqb;
+
+	d->dqb_id = cpu_to_le32(dquot->dq_id);
+	d->dqb_use_count = cpu_to_le32(OCFS2_DQUOT(dquot)->dq_use_count);
+	d->dqb_ihardlimit = cpu_to_le64(m->dqb_ihardlimit);
+	d->dqb_isoftlimit = cpu_to_le64(m->dqb_isoftlimit);
+	d->dqb_curinodes = cpu_to_le64(m->dqb_curinodes);
+	d->dqb_bhardlimit = cpu_to_le64(m->dqb_bhardlimit);
+	d->dqb_bsoftlimit = cpu_to_le64(m->dqb_bsoftlimit);
+	d->dqb_curspace = cpu_to_le64(m->dqb_curspace);
+	d->dqb_btime = cpu_to_le64(m->dqb_btime);
+	d->dqb_itime = cpu_to_le64(m->dqb_itime);
+}
+
+static int ocfs2_global_is_id(void *dp, struct dquot *dquot)
+{
+	struct ocfs2_global_disk_dqblk *d = dp;
+	struct ocfs2_mem_dqinfo *oinfo =
+			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
+
+	if (qtree_entry_unused(&oinfo->dqi_gi, dp))
+		return 0;
+	return le32_to_cpu(d->dqb_id) == dquot->dq_id;
+}
+
+struct qtree_fmt_operations ocfs2_global_ops = {
+	.mem2disk_dqblk = ocfs2_global_mem2diskdqb,
+	.disk2mem_dqblk = ocfs2_global_disk2memdqb,
+	.is_id = ocfs2_global_is_id,
+};
+
+static int ocfs2_validate_quota_block(struct super_block *sb,
+				      struct buffer_head *bh)
+{
+	struct ocfs2_disk_dqtrailer *dqt =
+		ocfs2_block_dqtrailer(sb->s_blocksize, bh->b_data);
+
+	mlog(0, "Validating quota block %llu\n",
+	     (unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 */
+	return ocfs2_validate_meta_ecc(sb, bh->b_data, &dqt->dq_check);
+}
+
+int ocfs2_read_quota_block(struct inode *inode, u64 v_block,
+			   struct buffer_head **bh)
+{
+	int rc = 0;
+	struct buffer_head *tmp = *bh;
+
+	rc = ocfs2_read_virt_blocks(inode, v_block, 1, &tmp, 0,
+				    ocfs2_validate_quota_block);
+	if (rc)
+		mlog_errno(rc);
+
+	/* If ocfs2_read_virt_blocks() got us a new bh, pass it up. */
+	if (!rc && !*bh)
+		*bh = tmp;
+
+	return rc;
+}
+
+static int ocfs2_get_quota_block(struct inode *inode, int block,
+				 struct buffer_head **bh)
+{
+	u64 pblock, pcount;
+	int err;
+
+	down_read(&OCFS2_I(inode)->ip_alloc_sem);
+	err = ocfs2_extent_map_get_blocks(inode, block, &pblock, &pcount, NULL);
+	up_read(&OCFS2_I(inode)->ip_alloc_sem);
+	if (err) {
+		mlog_errno(err);
+		return err;
+	}
+	*bh = sb_getblk(inode->i_sb, pblock);
+	if (!*bh) {
+		err = -EIO;
+		mlog_errno(err);
+	}
+	return err;;
+}
+
+/* Read data from global quotafile - avoid pagecache and such because we cannot
+ * afford acquiring the locks... We use quota cluster lock to serialize
+ * operations. Caller is responsible for acquiring it. */
+ssize_t ocfs2_quota_read(struct super_block *sb, int type, char *data,
+			 size_t len, loff_t off)
+{
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	struct inode *gqinode = oinfo->dqi_gqinode;
+	loff_t i_size = i_size_read(gqinode);
+	int offset = off & (sb->s_blocksize - 1);
+	sector_t blk = off >> sb->s_blocksize_bits;
+	int err = 0;
+	struct buffer_head *bh;
+	size_t toread, tocopy;
+
+	if (off > i_size)
+		return 0;
+	if (off + len > i_size)
+		len = i_size - off;
+	toread = len;
+	while (toread > 0) {
+		tocopy = min_t(size_t, (sb->s_blocksize - offset), toread);
+		bh = NULL;
+		err = ocfs2_read_quota_block(gqinode, blk, &bh);
+		if (err) {
+			mlog_errno(err);
+			return err;
+		}
+		memcpy(data, bh->b_data + offset, tocopy);
+		brelse(bh);
+		offset = 0;
+		toread -= tocopy;
+		data += tocopy;
+		blk++;
+	}
+	return len;
+}
+
+/* Write to quotafile (we know the transaction is already started and has
+ * enough credits) */
+ssize_t ocfs2_quota_write(struct super_block *sb, int type,
+			  const char *data, size_t len, loff_t off)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct inode *gqinode = oinfo->dqi_gqinode;
+	int offset = off & (sb->s_blocksize - 1);
+	sector_t blk = off >> sb->s_blocksize_bits;
+	int err = 0, new = 0, ja_type;
+	struct buffer_head *bh = NULL;
+	handle_t *handle = journal_current_handle();
+
+	if (!handle) {
+		mlog(ML_ERROR, "Quota write (off=%llu, len=%llu) cancelled "
+		     "because transaction was not started.\n",
+		     (unsigned long long)off, (unsigned long long)len);
+		return -EIO;
+	}
+	if (len > sb->s_blocksize - OCFS2_QBLK_RESERVED_SPACE - offset) {
+		WARN_ON(1);
+		len = sb->s_blocksize - OCFS2_QBLK_RESERVED_SPACE - offset;
+	}
+
+	mutex_lock_nested(&gqinode->i_mutex, I_MUTEX_QUOTA);
+	if (gqinode->i_size < off + len) {
+		down_write(&OCFS2_I(gqinode)->ip_alloc_sem);
+		err = ocfs2_extend_no_holes(gqinode, off + len, off);
+		up_write(&OCFS2_I(gqinode)->ip_alloc_sem);
+		if (err < 0)
+			goto out;
+		err = ocfs2_simple_size_update(gqinode,
+					       oinfo->dqi_gqi_bh,
+					       off + len);
+		if (err < 0)
+			goto out;
+		new = 1;
+	}
+	/* Not rewriting whole block? */
+	if ((offset || len < sb->s_blocksize - OCFS2_QBLK_RESERVED_SPACE) &&
+	    !new) {
+		err = ocfs2_read_quota_block(gqinode, blk, &bh);
+		ja_type = OCFS2_JOURNAL_ACCESS_WRITE;
+	} else {
+		err = ocfs2_get_quota_block(gqinode, blk, &bh);
+		ja_type = OCFS2_JOURNAL_ACCESS_CREATE;
+	}
+	if (err) {
+		mlog_errno(err);
+		return err;
+	}
+	lock_buffer(bh);
+	if (new)
+		memset(bh->b_data, 0, sb->s_blocksize);
+	memcpy(bh->b_data + offset, data, len);
+	flush_dcache_page(bh->b_page);
+	set_buffer_uptodate(bh);
+	unlock_buffer(bh);
+	ocfs2_set_buffer_uptodate(gqinode, bh);
+	err = ocfs2_journal_access_dq(handle, gqinode, bh, ja_type);
+	if (err < 0) {
+		brelse(bh);
+		goto out;
+	}
+	err = ocfs2_journal_dirty(handle, bh);
+	brelse(bh);
+	if (err < 0)
+		goto out;
+out:
+	if (err) {
+		mutex_unlock(&gqinode->i_mutex);
+		mlog_errno(err);
+		return err;
+	}
+	gqinode->i_version++;
+	ocfs2_mark_inode_dirty(handle, gqinode, oinfo->dqi_gqi_bh);
+	mutex_unlock(&gqinode->i_mutex);
+	return len;
+}
+
+int ocfs2_lock_global_qf(struct ocfs2_mem_dqinfo *oinfo, int ex)
+{
+	int status;
+	struct buffer_head *bh = NULL;
+
+	status = ocfs2_inode_lock(oinfo->dqi_gqinode, &bh, ex);
+	if (status < 0)
+		return status;
+	spin_lock(&dq_data_lock);
+	if (!oinfo->dqi_gqi_count++)
+		oinfo->dqi_gqi_bh = bh;
+	else
+		WARN_ON(bh != oinfo->dqi_gqi_bh);
+	spin_unlock(&dq_data_lock);
+	return 0;
+}
+
+void ocfs2_unlock_global_qf(struct ocfs2_mem_dqinfo *oinfo, int ex)
+{
+	ocfs2_inode_unlock(oinfo->dqi_gqinode, ex);
+	brelse(oinfo->dqi_gqi_bh);
+	spin_lock(&dq_data_lock);
+	if (!--oinfo->dqi_gqi_count)
+		oinfo->dqi_gqi_bh = NULL;
+	spin_unlock(&dq_data_lock);
+}
+
+/* Read information header from global quota file */
+int ocfs2_global_read_info(struct super_block *sb, int type)
+{
+	struct inode *gqinode = NULL;
+	unsigned int ino[MAXQUOTAS] = { USER_QUOTA_SYSTEM_INODE,
+					GROUP_QUOTA_SYSTEM_INODE };
+	struct ocfs2_global_disk_dqinfo dinfo;
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	int status;
+
+	mlog_entry_void();
+
+	/* Read global header */
+	gqinode = ocfs2_get_system_file_inode(OCFS2_SB(sb), ino[type],
+			OCFS2_INVALID_SLOT);
+	if (!gqinode) {
+		mlog(ML_ERROR, "failed to get global quota inode (type=%d)\n",
+			type);
+		status = -EINVAL;
+		goto out_err;
+	}
+	oinfo->dqi_gi.dqi_sb = sb;
+	oinfo->dqi_gi.dqi_type = type;
+	ocfs2_qinfo_lock_res_init(&oinfo->dqi_gqlock, oinfo);
+	oinfo->dqi_gi.dqi_entry_size = sizeof(struct ocfs2_global_disk_dqblk);
+	oinfo->dqi_gi.dqi_ops = &ocfs2_global_ops;
+	oinfo->dqi_gqi_bh = NULL;
+	oinfo->dqi_gqi_count = 0;
+	oinfo->dqi_gqinode = gqinode;
+	status = ocfs2_lock_global_qf(oinfo, 0);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_err;
+	}
+	status = sb->s_op->quota_read(sb, type, (char *)&dinfo,
+				      sizeof(struct ocfs2_global_disk_dqinfo),
+				      OCFS2_GLOBAL_INFO_OFF);
+	ocfs2_unlock_global_qf(oinfo, 0);
+	if (status != sizeof(struct ocfs2_global_disk_dqinfo)) {
+		mlog(ML_ERROR, "Cannot read global quota info (%d).\n",
+		     status);
+		if (status >= 0)
+			status = -EIO;
+		mlog_errno(status);
+		goto out_err;
+	}
+	info->dqi_bgrace = le32_to_cpu(dinfo.dqi_bgrace);
+	info->dqi_igrace = le32_to_cpu(dinfo.dqi_igrace);
+	oinfo->dqi_syncms = le32_to_cpu(dinfo.dqi_syncms);
+	oinfo->dqi_syncjiff = msecs_to_jiffies(oinfo->dqi_syncms);
+	oinfo->dqi_gi.dqi_blocks = le32_to_cpu(dinfo.dqi_blocks);
+	oinfo->dqi_gi.dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk);
+	oinfo->dqi_gi.dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry);
+	oinfo->dqi_gi.dqi_blocksize_bits = sb->s_blocksize_bits;
+	oinfo->dqi_gi.dqi_usable_bs = sb->s_blocksize -
+						OCFS2_QBLK_RESERVED_SPACE;
+	oinfo->dqi_gi.dqi_qtree_depth = qtree_depth(&oinfo->dqi_gi);
+	INIT_DELAYED_WORK(&oinfo->dqi_sync_work, qsync_work_fn);
+	queue_delayed_work(ocfs2_quota_wq, &oinfo->dqi_sync_work,
+			   oinfo->dqi_syncjiff);
+
+out_err:
+	mlog_exit(status);
+	return status;
+}
+
+/* Write information to global quota file. Expects exlusive lock on quota
+ * file inode and quota info */
+static int __ocfs2_global_write_info(struct super_block *sb, int type)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct ocfs2_global_disk_dqinfo dinfo;
+	ssize_t size;
+
+	spin_lock(&dq_data_lock);
+	info->dqi_flags &= ~DQF_INFO_DIRTY;
+	dinfo.dqi_bgrace = cpu_to_le32(info->dqi_bgrace);
+	dinfo.dqi_igrace = cpu_to_le32(info->dqi_igrace);
+	spin_unlock(&dq_data_lock);
+	dinfo.dqi_syncms = cpu_to_le32(oinfo->dqi_syncms);
+	dinfo.dqi_blocks = cpu_to_le32(oinfo->dqi_gi.dqi_blocks);
+	dinfo.dqi_free_blk = cpu_to_le32(oinfo->dqi_gi.dqi_free_blk);
+	dinfo.dqi_free_entry = cpu_to_le32(oinfo->dqi_gi.dqi_free_entry);
+	size = sb->s_op->quota_write(sb, type, (char *)&dinfo,
+				     sizeof(struct ocfs2_global_disk_dqinfo),
+				     OCFS2_GLOBAL_INFO_OFF);
+	if (size != sizeof(struct ocfs2_global_disk_dqinfo)) {
+		mlog(ML_ERROR, "Cannot write global quota info structure\n");
+		if (size >= 0)
+			size = -EIO;
+		return size;
+	}
+	return 0;
+}
+
+int ocfs2_global_write_info(struct super_block *sb, int type)
+{
+	int err;
+	struct ocfs2_mem_dqinfo *info = sb_dqinfo(sb, type)->dqi_priv;
+
+	err = ocfs2_qinfo_lock(info, 1);
+	if (err < 0)
+		return err;
+	err = __ocfs2_global_write_info(sb, type);
+	ocfs2_qinfo_unlock(info, 1);
+	return err;
+}
+
+/* Read in information from global quota file and acquire a reference to it.
+ * dquot_acquire() has already started the transaction and locked quota file */
+int ocfs2_global_read_dquot(struct dquot *dquot)
+{
+	int err, err2, ex = 0;
+	struct ocfs2_mem_dqinfo *info =
+			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
+
+	err = ocfs2_qinfo_lock(info, 0);
+	if (err < 0)
+		goto out;
+	err = qtree_read_dquot(&info->dqi_gi, dquot);
+	if (err < 0)
+		goto out_qlock;
+	OCFS2_DQUOT(dquot)->dq_use_count++;
+	OCFS2_DQUOT(dquot)->dq_origspace = dquot->dq_dqb.dqb_curspace;
+	OCFS2_DQUOT(dquot)->dq_originodes = dquot->dq_dqb.dqb_curinodes;
+	if (!dquot->dq_off) {	/* No real quota entry? */
+		/* Upgrade to exclusive lock for allocation */
+		err = ocfs2_qinfo_lock(info, 1);
+		if (err < 0)
+			goto out_qlock;
+		ex = 1;
+	}
+	err = qtree_write_dquot(&info->dqi_gi, dquot);
+	if (ex && info_dirty(sb_dqinfo(dquot->dq_sb, dquot->dq_type))) {
+		err2 = __ocfs2_global_write_info(dquot->dq_sb, dquot->dq_type);
+		if (!err)
+			err = err2;
+	}
+out_qlock:
+	if (ex)
+		ocfs2_qinfo_unlock(info, 1);
+	ocfs2_qinfo_unlock(info, 0);
+out:
+	if (err < 0)
+		mlog_errno(err);
+	return err;
+}
+
+/* Sync local information about quota modifications with global quota file.
+ * Caller must have started the transaction and obtained exclusive lock for
+ * global quota file inode */
+int __ocfs2_sync_dquot(struct dquot *dquot, int freeing)
+{
+	int err, err2;
+	struct super_block *sb = dquot->dq_sb;
+	int type = dquot->dq_type;
+	struct ocfs2_mem_dqinfo *info = sb_dqinfo(sb, type)->dqi_priv;
+	struct ocfs2_global_disk_dqblk dqblk;
+	s64 spacechange, inodechange;
+	time_t olditime, oldbtime;
+
+	err = sb->s_op->quota_read(sb, type, (char *)&dqblk,
+				   sizeof(struct ocfs2_global_disk_dqblk),
+				   dquot->dq_off);
+	if (err != sizeof(struct ocfs2_global_disk_dqblk)) {
+		if (err >= 0) {
+			mlog(ML_ERROR, "Short read from global quota file "
+				       "(%u read)\n", err);
+			err = -EIO;
+		}
+		goto out;
+	}
+
+	/* Update space and inode usage. Get also other information from
+	 * global quota file so that we don't overwrite any changes there.
+	 * We are */
+	spin_lock(&dq_data_lock);
+	spacechange = dquot->dq_dqb.dqb_curspace -
+					OCFS2_DQUOT(dquot)->dq_origspace;
+	inodechange = dquot->dq_dqb.dqb_curinodes -
+					OCFS2_DQUOT(dquot)->dq_originodes;
+	olditime = dquot->dq_dqb.dqb_itime;
+	oldbtime = dquot->dq_dqb.dqb_btime;
+	ocfs2_global_disk2memdqb(dquot, &dqblk);
+	mlog(0, "Syncing global dquot %u space %lld+%lld, inodes %lld+%lld\n",
+	     dquot->dq_id, dquot->dq_dqb.dqb_curspace, (long long)spacechange,
+	     dquot->dq_dqb.dqb_curinodes, (long long)inodechange);
+	if (!test_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags))
+		dquot->dq_dqb.dqb_curspace += spacechange;
+	if (!test_bit(DQ_LASTSET_B + QIF_INODES_B, &dquot->dq_flags))
+		dquot->dq_dqb.dqb_curinodes += inodechange;
+	/* Set properly space grace time... */
+	if (dquot->dq_dqb.dqb_bsoftlimit &&
+	    dquot->dq_dqb.dqb_curspace > dquot->dq_dqb.dqb_bsoftlimit) {
+		if (!test_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags) &&
+		    oldbtime > 0) {
+			if (dquot->dq_dqb.dqb_btime > 0)
+				dquot->dq_dqb.dqb_btime =
+					min(dquot->dq_dqb.dqb_btime, oldbtime);
+			else
+				dquot->dq_dqb.dqb_btime = oldbtime;
+		}
+	} else {
+		dquot->dq_dqb.dqb_btime = 0;
+		clear_bit(DQ_BLKS_B, &dquot->dq_flags);
+	}
+	/* Set properly inode grace time... */
+	if (dquot->dq_dqb.dqb_isoftlimit &&
+	    dquot->dq_dqb.dqb_curinodes > dquot->dq_dqb.dqb_isoftlimit) {
+		if (!test_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags) &&
+		    olditime > 0) {
+			if (dquot->dq_dqb.dqb_itime > 0)
+				dquot->dq_dqb.dqb_itime =
+					min(dquot->dq_dqb.dqb_itime, olditime);
+			else
+				dquot->dq_dqb.dqb_itime = olditime;
+		}
+	} else {
+		dquot->dq_dqb.dqb_itime = 0;
+		clear_bit(DQ_INODES_B, &dquot->dq_flags);
+	}
+	/* All information is properly updated, clear the flags */
+	__clear_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags);
+	__clear_bit(DQ_LASTSET_B + QIF_INODES_B, &dquot->dq_flags);
+	__clear_bit(DQ_LASTSET_B + QIF_BLIMITS_B, &dquot->dq_flags);
+	__clear_bit(DQ_LASTSET_B + QIF_ILIMITS_B, &dquot->dq_flags);
+	__clear_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags);
+	__clear_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags);
+	OCFS2_DQUOT(dquot)->dq_origspace = dquot->dq_dqb.dqb_curspace;
+	OCFS2_DQUOT(dquot)->dq_originodes = dquot->dq_dqb.dqb_curinodes;
+	spin_unlock(&dq_data_lock);
+	err = ocfs2_qinfo_lock(info, freeing);
+	if (err < 0) {
+		mlog(ML_ERROR, "Failed to lock quota info, loosing quota write"
+			       " (type=%d, id=%u)\n", dquot->dq_type,
+			       (unsigned)dquot->dq_id);
+		goto out;
+	}
+	if (freeing)
+		OCFS2_DQUOT(dquot)->dq_use_count--;
+	err = qtree_write_dquot(&info->dqi_gi, dquot);
+	if (err < 0)
+		goto out_qlock;
+	if (freeing && !OCFS2_DQUOT(dquot)->dq_use_count) {
+		err = qtree_release_dquot(&info->dqi_gi, dquot);
+		if (info_dirty(sb_dqinfo(sb, type))) {
+			err2 = __ocfs2_global_write_info(sb, type);
+			if (!err)
+				err = err2;
+		}
+	}
+out_qlock:
+	ocfs2_qinfo_unlock(info, freeing);
+out:
+	if (err < 0)
+		mlog_errno(err);
+	return err;
+}
+
+/*
+ *  Functions for periodic syncing of dquots with global file
+ */
+static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
+{
+	handle_t *handle;
+	struct super_block *sb = dquot->dq_sb;
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	struct ocfs2_super *osb = OCFS2_SB(sb);
+	int status = 0;
+
+	mlog_entry("id=%u qtype=%u type=%lu device=%s\n", dquot->dq_id,
+		   dquot->dq_type, type, sb->s_id);
+	if (type != dquot->dq_type)
+		goto out;
+	status = ocfs2_lock_global_qf(oinfo, 1);
+	if (status < 0)
+		goto out;
+
+	handle = ocfs2_start_trans(osb, OCFS2_QSYNC_CREDITS);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out_ilock;
+	}
+	mutex_lock(&sb_dqopt(sb)->dqio_mutex);
+	status = ocfs2_sync_dquot(dquot);
+	mutex_unlock(&sb_dqopt(sb)->dqio_mutex);
+	if (status < 0)
+		mlog_errno(status);
+	/* We have to write local structure as well... */
+	dquot_mark_dquot_dirty(dquot);
+	status = dquot_commit(dquot);
+	if (status < 0)
+		mlog_errno(status);
+	ocfs2_commit_trans(osb, handle);
+out_ilock:
+	ocfs2_unlock_global_qf(oinfo, 1);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+static void qsync_work_fn(struct work_struct *work)
+{
+	struct ocfs2_mem_dqinfo *oinfo = container_of(work,
+						      struct ocfs2_mem_dqinfo,
+						      dqi_sync_work.work);
+	struct super_block *sb = oinfo->dqi_gqinode->i_sb;
+
+	dquot_scan_active(sb, ocfs2_sync_dquot_helper, oinfo->dqi_type);
+	queue_delayed_work(ocfs2_quota_wq, &oinfo->dqi_sync_work,
+			   oinfo->dqi_syncjiff);
+}
+
+/*
+ *  Wrappers for generic quota functions
+ */
+
+static int ocfs2_write_dquot(struct dquot *dquot)
+{
+	handle_t *handle;
+	struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
+	int status = 0;
+
+	mlog_entry("id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+
+	handle = ocfs2_start_trans(osb, OCFS2_QWRITE_CREDITS);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out;
+	}
+	status = dquot_commit(dquot);
+	ocfs2_commit_trans(osb, handle);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+int ocfs2_calc_qdel_credits(struct super_block *sb, int type)
+{
+	struct ocfs2_mem_dqinfo *oinfo;
+	int features[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
+				    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA };
+
+	if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, features[type]))
+		return 0;
+
+	oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	/* We modify tree, leaf block, global info, local chunk header,
+	 * global and local inode */
+	return oinfo->dqi_gi.dqi_qtree_depth + 2 + 1 +
+	       2 * OCFS2_INODE_UPDATE_CREDITS;
+}
+
+static int ocfs2_release_dquot(struct dquot *dquot)
+{
+	handle_t *handle;
+	struct ocfs2_mem_dqinfo *oinfo =
+			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
+	struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
+	int status = 0;
+
+	mlog_entry("id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+
+	status = ocfs2_lock_global_qf(oinfo, 1);
+	if (status < 0)
+		goto out;
+	handle = ocfs2_start_trans(osb,
+		ocfs2_calc_qdel_credits(dquot->dq_sb, dquot->dq_type));
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out_ilock;
+	}
+	status = dquot_release(dquot);
+	ocfs2_commit_trans(osb, handle);
+out_ilock:
+	ocfs2_unlock_global_qf(oinfo, 1);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+int ocfs2_calc_qinit_credits(struct super_block *sb, int type)
+{
+	struct ocfs2_mem_dqinfo *oinfo;
+	int features[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
+				    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA };
+	struct ocfs2_dinode *lfe, *gfe;
+
+	if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, features[type]))
+		return 0;
+
+	oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	gfe = (struct ocfs2_dinode *)oinfo->dqi_gqi_bh->b_data;
+	lfe = (struct ocfs2_dinode *)oinfo->dqi_lqi_bh->b_data;
+	/* We can extend local file + global file. In local file we
+	 * can modify info, chunk header block and dquot block. In
+	 * global file we can modify info, tree and leaf block */
+	return ocfs2_calc_extend_credits(sb, &lfe->id2.i_list, 0) +
+	       ocfs2_calc_extend_credits(sb, &gfe->id2.i_list, 0) +
+	       3 + oinfo->dqi_gi.dqi_qtree_depth + 2;
+}
+
+static int ocfs2_acquire_dquot(struct dquot *dquot)
+{
+	handle_t *handle;
+	struct ocfs2_mem_dqinfo *oinfo =
+			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
+	struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
+	int status = 0;
+
+	mlog_entry("id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+	/* We need an exclusive lock, because we're going to update use count
+	 * and instantiate possibly new dquot structure */
+	status = ocfs2_lock_global_qf(oinfo, 1);
+	if (status < 0)
+		goto out;
+	handle = ocfs2_start_trans(osb,
+		ocfs2_calc_qinit_credits(dquot->dq_sb, dquot->dq_type));
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out_ilock;
+	}
+	status = dquot_acquire(dquot);
+	ocfs2_commit_trans(osb, handle);
+out_ilock:
+	ocfs2_unlock_global_qf(oinfo, 1);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+static int ocfs2_mark_dquot_dirty(struct dquot *dquot)
+{
+	unsigned long mask = (1 << (DQ_LASTSET_B + QIF_ILIMITS_B)) |
+			     (1 << (DQ_LASTSET_B + QIF_BLIMITS_B)) |
+			     (1 << (DQ_LASTSET_B + QIF_INODES_B)) |
+			     (1 << (DQ_LASTSET_B + QIF_SPACE_B)) |
+			     (1 << (DQ_LASTSET_B + QIF_BTIME_B)) |
+			     (1 << (DQ_LASTSET_B + QIF_ITIME_B));
+	int sync = 0;
+	int status;
+	struct super_block *sb = dquot->dq_sb;
+	int type = dquot->dq_type;
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	handle_t *handle;
+	struct ocfs2_super *osb = OCFS2_SB(sb);
+
+	mlog_entry("id=%u, type=%d", dquot->dq_id, type);
+	dquot_mark_dquot_dirty(dquot);
+
+	/* In case user set some limits, sync dquot immediately to global
+	 * quota file so that information propagates quicker */
+	spin_lock(&dq_data_lock);
+	if (dquot->dq_flags & mask)
+		sync = 1;
+	spin_unlock(&dq_data_lock);
+	if (!sync) {
+		status = ocfs2_write_dquot(dquot);
+		goto out;
+	}
+	status = ocfs2_lock_global_qf(oinfo, 1);
+	if (status < 0)
+		goto out;
+	handle = ocfs2_start_trans(osb, OCFS2_QSYNC_CREDITS);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out_ilock;
+	}
+	status = ocfs2_sync_dquot(dquot);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	/* Now write updated local dquot structure */
+	status = dquot_commit(dquot);
+out_trans:
+	ocfs2_commit_trans(osb, handle);
+out_ilock:
+	ocfs2_unlock_global_qf(oinfo, 1);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+/* This should happen only after set_dqinfo(). */
+static int ocfs2_write_info(struct super_block *sb, int type)
+{
+	handle_t *handle;
+	int status = 0;
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+
+	mlog_entry_void();
+
+	status = ocfs2_lock_global_qf(oinfo, 1);
+	if (status < 0)
+		goto out;
+	handle = ocfs2_start_trans(OCFS2_SB(sb), OCFS2_QINFO_WRITE_CREDITS);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out_ilock;
+	}
+	status = dquot_commit_info(sb, type);
+	ocfs2_commit_trans(OCFS2_SB(sb), handle);
+out_ilock:
+	ocfs2_unlock_global_qf(oinfo, 1);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+/* This is difficult. We have to lock quota inode and start transaction
+ * in this function but we don't want to take the penalty of exlusive
+ * quota file lock when we are just going to use cached structures. So
+ * we just take read lock check whether we have dquot cached and if so,
+ * we don't have to take the write lock... */
+static int ocfs2_dquot_initialize(struct inode *inode, int type)
+{
+	handle_t *handle = NULL;
+	int status = 0;
+	struct super_block *sb = inode->i_sb;
+	struct ocfs2_mem_dqinfo *oinfo;
+	int exclusive = 0;
+	int cnt;
+	qid_t id;
+
+	mlog_entry_void();
+
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+		if (type != -1 && cnt != type)
+			continue;
+		if (!sb_has_quota_active(sb, cnt))
+			continue;
+		oinfo = sb_dqinfo(sb, cnt)->dqi_priv;
+		status = ocfs2_lock_global_qf(oinfo, 0);
+		if (status < 0)
+			goto out;
+		/* This is just a performance optimization not a reliable test.
+		 * Since we hold an inode lock, noone can actually release
+		 * the structure until we are finished with initialization. */
+		if (inode->i_dquot[cnt] != NODQUOT) {
+			ocfs2_unlock_global_qf(oinfo, 0);
+			continue;
+		}
+		/* When we have inode lock, we know that no dquot_release() can
+		 * run and thus we can safely check whether we need to
+		 * read+modify global file to get quota information or whether
+		 * our node already has it. */
+		if (cnt == USRQUOTA)
+			id = inode->i_uid;
+		else if (cnt == GRPQUOTA)
+			id = inode->i_gid;
+		else
+			BUG();
+		/* Obtain exclusion from quota off... */
+		down_write(&sb_dqopt(sb)->dqptr_sem);
+		exclusive = !dquot_is_cached(sb, id, cnt);
+		up_write(&sb_dqopt(sb)->dqptr_sem);
+		if (exclusive) {
+			status = ocfs2_lock_global_qf(oinfo, 1);
+			if (status < 0) {
+				exclusive = 0;
+				mlog_errno(status);
+				goto out_ilock;
+			}
+			handle = ocfs2_start_trans(OCFS2_SB(sb),
+					ocfs2_calc_qinit_credits(sb, cnt));
+			if (IS_ERR(handle)) {
+				status = PTR_ERR(handle);
+				mlog_errno(status);
+				goto out_ilock;
+			}
+		}
+		dquot_initialize(inode, cnt);
+		if (exclusive) {
+			ocfs2_commit_trans(OCFS2_SB(sb), handle);
+			ocfs2_unlock_global_qf(oinfo, 1);
+		}
+		ocfs2_unlock_global_qf(oinfo, 0);
+	}
+	mlog_exit(0);
+	return 0;
+out_ilock:
+	if (exclusive)
+		ocfs2_unlock_global_qf(oinfo, 1);
+	ocfs2_unlock_global_qf(oinfo, 0);
+out:
+	mlog_exit(status);
+	return status;
+}
+
+static int ocfs2_dquot_drop_slow(struct inode *inode)
+{
+	int status = 0;
+	int cnt;
+	int got_lock[MAXQUOTAS] = {0, 0};
+	handle_t *handle;
+	struct super_block *sb = inode->i_sb;
+	struct ocfs2_mem_dqinfo *oinfo;
+
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+		if (!sb_has_quota_active(sb, cnt))
+			continue;
+		oinfo = sb_dqinfo(sb, cnt)->dqi_priv;
+		status = ocfs2_lock_global_qf(oinfo, 1);
+		if (status < 0)
+			goto out;
+		got_lock[cnt] = 1;
+	}
+	handle = ocfs2_start_trans(OCFS2_SB(sb),
+			ocfs2_calc_qinit_credits(sb, USRQUOTA) +
+			ocfs2_calc_qinit_credits(sb, GRPQUOTA));
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out;
+	}
+	dquot_drop(inode);
+	ocfs2_commit_trans(OCFS2_SB(sb), handle);
+out:
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
+		if (got_lock[cnt]) {
+			oinfo = sb_dqinfo(sb, cnt)->dqi_priv;
+			ocfs2_unlock_global_qf(oinfo, 1);
+		}
+	return status;
+}
+
+/* See the comment before ocfs2_dquot_initialize. */
+static int ocfs2_dquot_drop(struct inode *inode)
+{
+	int status = 0;
+	struct super_block *sb = inode->i_sb;
+	struct ocfs2_mem_dqinfo *oinfo;
+	int exclusive = 0;
+	int cnt;
+	int got_lock[MAXQUOTAS] = {0, 0};
+
+	mlog_entry_void();
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+		if (!sb_has_quota_active(sb, cnt))
+			continue;
+		oinfo = sb_dqinfo(sb, cnt)->dqi_priv;
+		status = ocfs2_lock_global_qf(oinfo, 0);
+		if (status < 0)
+			goto out;
+		got_lock[cnt] = 1;
+	}
+	/* Lock against anyone releasing references so that when when we check
+	 * we know we are not going to be last ones to release dquot */
+	down_write(&sb_dqopt(sb)->dqptr_sem);
+	/* Urgh, this is a terrible hack :( */
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
+		if (inode->i_dquot[cnt] != NODQUOT &&
+		    atomic_read(&inode->i_dquot[cnt]->dq_count) > 1) {
+			exclusive = 1;
+			break;
+		}
+	}
+	if (!exclusive)
+		dquot_drop_locked(inode);
+	up_write(&sb_dqopt(sb)->dqptr_sem);
+out:
+	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
+		if (got_lock[cnt]) {
+			oinfo = sb_dqinfo(sb, cnt)->dqi_priv;
+			ocfs2_unlock_global_qf(oinfo, 0);
+		}
+	/* In case we bailed out because we had to do expensive locking
+	 * do it now... */
+	if (exclusive)
+		status = ocfs2_dquot_drop_slow(inode);
+	mlog_exit(status);
+	return status;
+}
+
+static struct dquot *ocfs2_alloc_dquot(struct super_block *sb, int type)
+{
+	struct ocfs2_dquot *dquot =
+				kmem_cache_zalloc(ocfs2_dquot_cachep, GFP_NOFS);
+
+	if (!dquot)
+		return NULL;
+	return &dquot->dq_dquot;
+}
+
+static void ocfs2_destroy_dquot(struct dquot *dquot)
+{
+	kmem_cache_free(ocfs2_dquot_cachep, dquot);
+}
+
+struct dquot_operations ocfs2_quota_operations = {
+	.initialize	= ocfs2_dquot_initialize,
+	.drop		= ocfs2_dquot_drop,
+	.alloc_space	= dquot_alloc_space,
+	.alloc_inode	= dquot_alloc_inode,
+	.free_space	= dquot_free_space,
+	.free_inode	= dquot_free_inode,
+	.transfer	= dquot_transfer,
+	.write_dquot	= ocfs2_write_dquot,
+	.acquire_dquot	= ocfs2_acquire_dquot,
+	.release_dquot	= ocfs2_release_dquot,
+	.mark_dirty	= ocfs2_mark_dquot_dirty,
+	.write_info	= ocfs2_write_info,
+	.alloc_dquot	= ocfs2_alloc_dquot,
+	.destroy_dquot	= ocfs2_destroy_dquot,
+};
+
+int ocfs2_quota_setup(void)
+{
+	ocfs2_quota_wq = create_workqueue("o2quot");
+	if (!ocfs2_quota_wq)
+		return -ENOMEM;
+	return 0;
+}
+
+void ocfs2_quota_shutdown(void)
+{
+	if (ocfs2_quota_wq) {
+		flush_workqueue(ocfs2_quota_wq);
+		destroy_workqueue(ocfs2_quota_wq);
+		ocfs2_quota_wq = NULL;
+	}
+}
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
new file mode 100644
index 00000000000..07deec5e972
--- /dev/null
+++ b/fs/ocfs2/quota_local.c
@@ -0,0 +1,1253 @@
+/*
+ *  Implementation of operations over local quota file
+ */
+
+#include <linux/fs.h>
+#include <linux/quota.h>
+#include <linux/quotaops.h>
+#include <linux/module.h>
+
+#define MLOG_MASK_PREFIX ML_QUOTA
+#include <cluster/masklog.h>
+
+#include "ocfs2_fs.h"
+#include "ocfs2.h"
+#include "inode.h"
+#include "alloc.h"
+#include "file.h"
+#include "buffer_head_io.h"
+#include "journal.h"
+#include "sysfile.h"
+#include "dlmglue.h"
+#include "quota.h"
+
+/* Number of local quota structures per block */
+static inline unsigned int ol_quota_entries_per_block(struct super_block *sb)
+{
+	return ((sb->s_blocksize - OCFS2_QBLK_RESERVED_SPACE) /
+		sizeof(struct ocfs2_local_disk_dqblk));
+}
+
+/* Number of blocks with entries in one chunk */
+static inline unsigned int ol_chunk_blocks(struct super_block *sb)
+{
+	return ((sb->s_blocksize - sizeof(struct ocfs2_local_disk_chunk) -
+		 OCFS2_QBLK_RESERVED_SPACE) << 3) /
+	       ol_quota_entries_per_block(sb);
+}
+
+/* Number of entries in a chunk bitmap */
+static unsigned int ol_chunk_entries(struct super_block *sb)
+{
+	return ol_chunk_blocks(sb) * ol_quota_entries_per_block(sb);
+}
+
+/* Offset of the chunk in quota file */
+static unsigned int ol_quota_chunk_block(struct super_block *sb, int c)
+{
+	/* 1 block for local quota file info, 1 block per chunk for chunk info */
+	return 1 + (ol_chunk_blocks(sb) + 1) * c;
+}
+
+static unsigned int ol_dqblk_block(struct super_block *sb, int c, int off)
+{
+	int epb = ol_quota_entries_per_block(sb);
+
+	return ol_quota_chunk_block(sb, c) + 1 + off / epb;
+}
+
+static unsigned int ol_dqblk_block_off(struct super_block *sb, int c, int off)
+{
+	int epb = ol_quota_entries_per_block(sb);
+
+	return (off % epb) * sizeof(struct ocfs2_local_disk_dqblk);
+}
+
+/* Offset of the dquot structure in the quota file */
+static loff_t ol_dqblk_off(struct super_block *sb, int c, int off)
+{
+	return (ol_dqblk_block(sb, c, off) << sb->s_blocksize_bits) +
+	       ol_dqblk_block_off(sb, c, off);
+}
+
+/* Compute block number from given offset */
+static inline unsigned int ol_dqblk_file_block(struct super_block *sb, loff_t off)
+{
+	return off >> sb->s_blocksize_bits;
+}
+
+static inline unsigned int ol_dqblk_block_offset(struct super_block *sb, loff_t off)
+{
+	return off & ((1 << sb->s_blocksize_bits) - 1);
+}
+
+/* Compute offset in the chunk of a structure with the given offset */
+static int ol_dqblk_chunk_off(struct super_block *sb, int c, loff_t off)
+{
+	int epb = ol_quota_entries_per_block(sb);
+
+	return ((off >> sb->s_blocksize_bits) -
+			ol_quota_chunk_block(sb, c) - 1) * epb
+	       + ((unsigned int)(off & ((1 << sb->s_blocksize_bits) - 1))) /
+		 sizeof(struct ocfs2_local_disk_dqblk);
+}
+
+/* Write bufferhead into the fs */
+static int ocfs2_modify_bh(struct inode *inode, struct buffer_head *bh,
+		void (*modify)(struct buffer_head *, void *), void *private)
+{
+	struct super_block *sb = inode->i_sb;
+	handle_t *handle;
+	int status;
+
+	handle = ocfs2_start_trans(OCFS2_SB(sb), 1);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		return status;
+	}
+	status = ocfs2_journal_access_dq(handle, inode, bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
+	if (status < 0) {
+		mlog_errno(status);
+		ocfs2_commit_trans(OCFS2_SB(sb), handle);
+		return status;
+	}
+	lock_buffer(bh);
+	modify(bh, private);
+	unlock_buffer(bh);
+	status = ocfs2_journal_dirty(handle, bh);
+	if (status < 0) {
+		mlog_errno(status);
+		ocfs2_commit_trans(OCFS2_SB(sb), handle);
+		return status;
+	}
+	status = ocfs2_commit_trans(OCFS2_SB(sb), handle);
+	if (status < 0) {
+		mlog_errno(status);
+		return status;
+	}
+	return 0;
+}
+
+/* Check whether we understand format of quota files */
+static int ocfs2_local_check_quota_file(struct super_block *sb, int type)
+{
+	unsigned int lmagics[MAXQUOTAS] = OCFS2_LOCAL_QMAGICS;
+	unsigned int lversions[MAXQUOTAS] = OCFS2_LOCAL_QVERSIONS;
+	unsigned int gmagics[MAXQUOTAS] = OCFS2_GLOBAL_QMAGICS;
+	unsigned int gversions[MAXQUOTAS] = OCFS2_GLOBAL_QVERSIONS;
+	unsigned int ino[MAXQUOTAS] = { USER_QUOTA_SYSTEM_INODE,
+					GROUP_QUOTA_SYSTEM_INODE };
+	struct buffer_head *bh = NULL;
+	struct inode *linode = sb_dqopt(sb)->files[type];
+	struct inode *ginode = NULL;
+	struct ocfs2_disk_dqheader *dqhead;
+	int status, ret = 0;
+
+	/* First check whether we understand local quota file */
+	status = ocfs2_read_quota_block(linode, 0, &bh);
+	if (status) {
+		mlog_errno(status);
+		mlog(ML_ERROR, "failed to read quota file header (type=%d)\n",
+			type);
+		goto out_err;
+	}
+	dqhead = (struct ocfs2_disk_dqheader *)(bh->b_data);
+	if (le32_to_cpu(dqhead->dqh_magic) != lmagics[type]) {
+		mlog(ML_ERROR, "quota file magic does not match (%u != %u),"
+			" type=%d\n", le32_to_cpu(dqhead->dqh_magic),
+			lmagics[type], type);
+		goto out_err;
+	}
+	if (le32_to_cpu(dqhead->dqh_version) != lversions[type]) {
+		mlog(ML_ERROR, "quota file version does not match (%u != %u),"
+			" type=%d\n", le32_to_cpu(dqhead->dqh_version),
+			lversions[type], type);
+		goto out_err;
+	}
+	brelse(bh);
+	bh = NULL;
+
+	/* Next check whether we understand global quota file */
+	ginode = ocfs2_get_system_file_inode(OCFS2_SB(sb), ino[type],
+						OCFS2_INVALID_SLOT);
+	if (!ginode) {
+		mlog(ML_ERROR, "cannot get global quota file inode "
+				"(type=%d)\n", type);
+		goto out_err;
+	}
+	/* Since the header is read only, we don't care about locking */
+	status = ocfs2_read_quota_block(ginode, 0, &bh);
+	if (status) {
+		mlog_errno(status);
+		mlog(ML_ERROR, "failed to read global quota file header "
+				"(type=%d)\n", type);
+		goto out_err;
+	}
+	dqhead = (struct ocfs2_disk_dqheader *)(bh->b_data);
+	if (le32_to_cpu(dqhead->dqh_magic) != gmagics[type]) {
+		mlog(ML_ERROR, "global quota file magic does not match "
+			"(%u != %u), type=%d\n",
+			le32_to_cpu(dqhead->dqh_magic), gmagics[type], type);
+		goto out_err;
+	}
+	if (le32_to_cpu(dqhead->dqh_version) != gversions[type]) {
+		mlog(ML_ERROR, "global quota file version does not match "
+			"(%u != %u), type=%d\n",
+			le32_to_cpu(dqhead->dqh_version), gversions[type],
+			type);
+		goto out_err;
+	}
+
+	ret = 1;
+out_err:
+	brelse(bh);
+	iput(ginode);
+	return ret;
+}
+
+/* Release given list of quota file chunks */
+static void ocfs2_release_local_quota_bitmaps(struct list_head *head)
+{
+	struct ocfs2_quota_chunk *pos, *next;
+
+	list_for_each_entry_safe(pos, next, head, qc_chunk) {
+		list_del(&pos->qc_chunk);
+		brelse(pos->qc_headerbh);
+		kmem_cache_free(ocfs2_qf_chunk_cachep, pos);
+	}
+}
+
+/* Load quota bitmaps into memory */
+static int ocfs2_load_local_quota_bitmaps(struct inode *inode,
+			struct ocfs2_local_disk_dqinfo *ldinfo,
+			struct list_head *head)
+{
+	struct ocfs2_quota_chunk *newchunk;
+	int i, status;
+
+	INIT_LIST_HEAD(head);
+	for (i = 0; i < le32_to_cpu(ldinfo->dqi_chunks); i++) {
+		newchunk = kmem_cache_alloc(ocfs2_qf_chunk_cachep, GFP_NOFS);
+		if (!newchunk) {
+			ocfs2_release_local_quota_bitmaps(head);
+			return -ENOMEM;
+		}
+		newchunk->qc_num = i;
+		newchunk->qc_headerbh = NULL;
+		status = ocfs2_read_quota_block(inode,
+				ol_quota_chunk_block(inode->i_sb, i),
+				&newchunk->qc_headerbh);
+		if (status) {
+			mlog_errno(status);
+			kmem_cache_free(ocfs2_qf_chunk_cachep, newchunk);
+			ocfs2_release_local_quota_bitmaps(head);
+			return status;
+		}
+		list_add_tail(&newchunk->qc_chunk, head);
+	}
+	return 0;
+}
+
+static void olq_update_info(struct buffer_head *bh, void *private)
+{
+	struct mem_dqinfo *info = private;
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct ocfs2_local_disk_dqinfo *ldinfo;
+
+	ldinfo = (struct ocfs2_local_disk_dqinfo *)(bh->b_data +
+						OCFS2_LOCAL_INFO_OFF);
+	spin_lock(&dq_data_lock);
+	ldinfo->dqi_flags = cpu_to_le32(info->dqi_flags & DQF_MASK);
+	ldinfo->dqi_chunks = cpu_to_le32(oinfo->dqi_chunks);
+	ldinfo->dqi_blocks = cpu_to_le32(oinfo->dqi_blocks);
+	spin_unlock(&dq_data_lock);
+}
+
+static int ocfs2_add_recovery_chunk(struct super_block *sb,
+				    struct ocfs2_local_disk_chunk *dchunk,
+				    int chunk,
+				    struct list_head *head)
+{
+	struct ocfs2_recovery_chunk *rc;
+
+	rc = kmalloc(sizeof(struct ocfs2_recovery_chunk), GFP_NOFS);
+	if (!rc)
+		return -ENOMEM;
+	rc->rc_chunk = chunk;
+	rc->rc_bitmap = kmalloc(sb->s_blocksize, GFP_NOFS);
+	if (!rc->rc_bitmap) {
+		kfree(rc);
+		return -ENOMEM;
+	}
+	memcpy(rc->rc_bitmap, dchunk->dqc_bitmap,
+	       (ol_chunk_entries(sb) + 7) >> 3);
+	list_add_tail(&rc->rc_list, head);
+	return 0;
+}
+
+static void free_recovery_list(struct list_head *head)
+{
+	struct ocfs2_recovery_chunk *next;
+	struct ocfs2_recovery_chunk *rchunk;
+
+	list_for_each_entry_safe(rchunk, next, head, rc_list) {
+		list_del(&rchunk->rc_list);
+		kfree(rchunk->rc_bitmap);
+		kfree(rchunk);
+	}
+}
+
+void ocfs2_free_quota_recovery(struct ocfs2_quota_recovery *rec)
+{
+	int type;
+
+	for (type = 0; type < MAXQUOTAS; type++)
+		free_recovery_list(&(rec->r_list[type]));
+	kfree(rec);
+}
+
+/* Load entries in our quota file we have to recover*/
+static int ocfs2_recovery_load_quota(struct inode *lqinode,
+				     struct ocfs2_local_disk_dqinfo *ldinfo,
+				     int type,
+				     struct list_head *head)
+{
+	struct super_block *sb = lqinode->i_sb;
+	struct buffer_head *hbh;
+	struct ocfs2_local_disk_chunk *dchunk;
+	int i, chunks = le32_to_cpu(ldinfo->dqi_chunks);
+	int status = 0;
+
+	for (i = 0; i < chunks; i++) {
+		hbh = NULL;
+		status = ocfs2_read_quota_block(lqinode,
+						ol_quota_chunk_block(sb, i),
+						&hbh);
+		if (status) {
+			mlog_errno(status);
+			break;
+		}
+		dchunk = (struct ocfs2_local_disk_chunk *)hbh->b_data;
+		if (le32_to_cpu(dchunk->dqc_free) < ol_chunk_entries(sb))
+			status = ocfs2_add_recovery_chunk(sb, dchunk, i, head);
+		brelse(hbh);
+		if (status < 0)
+			break;
+	}
+	if (status < 0)
+		free_recovery_list(head);
+	return status;
+}
+
+static struct ocfs2_quota_recovery *ocfs2_alloc_quota_recovery(void)
+{
+	int type;
+	struct ocfs2_quota_recovery *rec;
+
+	rec = kmalloc(sizeof(struct ocfs2_quota_recovery), GFP_NOFS);
+	if (!rec)
+		return NULL;
+	for (type = 0; type < MAXQUOTAS; type++)
+		INIT_LIST_HEAD(&(rec->r_list[type]));
+	return rec;
+}
+
+/* Load information we need for quota recovery into memory */
+struct ocfs2_quota_recovery *ocfs2_begin_quota_recovery(
+						struct ocfs2_super *osb,
+						int slot_num)
+{
+	unsigned int feature[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
+					    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA};
+	unsigned int ino[MAXQUOTAS] = { LOCAL_USER_QUOTA_SYSTEM_INODE,
+					LOCAL_GROUP_QUOTA_SYSTEM_INODE };
+	struct super_block *sb = osb->sb;
+	struct ocfs2_local_disk_dqinfo *ldinfo;
+	struct inode *lqinode;
+	struct buffer_head *bh;
+	int type;
+	int status = 0;
+	struct ocfs2_quota_recovery *rec;
+
+	mlog(ML_NOTICE, "Beginning quota recovery in slot %u\n", slot_num);
+	rec = ocfs2_alloc_quota_recovery();
+	if (!rec)
+		return ERR_PTR(-ENOMEM);
+	/* First init... */
+
+	for (type = 0; type < MAXQUOTAS; type++) {
+		if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, feature[type]))
+			continue;
+		/* At this point, journal of the slot is already replayed so
+		 * we can trust metadata and data of the quota file */
+		lqinode = ocfs2_get_system_file_inode(osb, ino[type], slot_num);
+		if (!lqinode) {
+			status = -ENOENT;
+			goto out;
+		}
+		status = ocfs2_inode_lock_full(lqinode, NULL, 1,
+					       OCFS2_META_LOCK_RECOVERY);
+		if (status < 0) {
+			mlog_errno(status);
+			goto out_put;
+		}
+		/* Now read local header */
+		bh = NULL;
+		status = ocfs2_read_quota_block(lqinode, 0, &bh);
+		if (status) {
+			mlog_errno(status);
+			mlog(ML_ERROR, "failed to read quota file info header "
+				"(slot=%d type=%d)\n", slot_num, type);
+			goto out_lock;
+		}
+		ldinfo = (struct ocfs2_local_disk_dqinfo *)(bh->b_data +
+							OCFS2_LOCAL_INFO_OFF);
+		status = ocfs2_recovery_load_quota(lqinode, ldinfo, type,
+						   &rec->r_list[type]);
+		brelse(bh);
+out_lock:
+		ocfs2_inode_unlock(lqinode, 1);
+out_put:
+		iput(lqinode);
+		if (status < 0)
+			break;
+	}
+out:
+	if (status < 0) {
+		ocfs2_free_quota_recovery(rec);
+		rec = ERR_PTR(status);
+	}
+	return rec;
+}
+
+/* Sync changes in local quota file into global quota file and
+ * reinitialize local quota file.
+ * The function expects local quota file to be already locked and
+ * dqonoff_mutex locked. */
+static int ocfs2_recover_local_quota_file(struct inode *lqinode,
+					  int type,
+					  struct ocfs2_quota_recovery *rec)
+{
+	struct super_block *sb = lqinode->i_sb;
+	struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
+	struct ocfs2_local_disk_chunk *dchunk;
+	struct ocfs2_local_disk_dqblk *dqblk;
+	struct dquot *dquot;
+	handle_t *handle;
+	struct buffer_head *hbh = NULL, *qbh = NULL;
+	int status = 0;
+	int bit, chunk;
+	struct ocfs2_recovery_chunk *rchunk, *next;
+	qsize_t spacechange, inodechange;
+
+	mlog_entry("ino=%lu type=%u", (unsigned long)lqinode->i_ino, type);
+
+	status = ocfs2_lock_global_qf(oinfo, 1);
+	if (status < 0)
+		goto out;
+
+	list_for_each_entry_safe(rchunk, next, &(rec->r_list[type]), rc_list) {
+		chunk = rchunk->rc_chunk;
+		hbh = NULL;
+		status = ocfs2_read_quota_block(lqinode,
+						ol_quota_chunk_block(sb, chunk),
+						&hbh);
+		if (status) {
+			mlog_errno(status);
+			break;
+		}
+		dchunk = (struct ocfs2_local_disk_chunk *)hbh->b_data;
+		for_each_bit(bit, rchunk->rc_bitmap, ol_chunk_entries(sb)) {
+			qbh = NULL;
+			status = ocfs2_read_quota_block(lqinode,
+						ol_dqblk_block(sb, chunk, bit),
+						&qbh);
+			if (status) {
+				mlog_errno(status);
+				break;
+			}
+			dqblk = (struct ocfs2_local_disk_dqblk *)(qbh->b_data +
+				ol_dqblk_block_off(sb, chunk, bit));
+			dquot = dqget(sb, le64_to_cpu(dqblk->dqb_id), type);
+			if (!dquot) {
+				status = -EIO;
+				mlog(ML_ERROR, "Failed to get quota structure "
+				     "for id %u, type %d. Cannot finish quota "
+				     "file recovery.\n",
+				     (unsigned)le64_to_cpu(dqblk->dqb_id),
+				     type);
+				goto out_put_bh;
+			}
+			handle = ocfs2_start_trans(OCFS2_SB(sb),
+						   OCFS2_QSYNC_CREDITS);
+			if (IS_ERR(handle)) {
+				status = PTR_ERR(handle);
+				mlog_errno(status);
+				goto out_put_dquot;
+			}
+			mutex_lock(&sb_dqopt(sb)->dqio_mutex);
+			spin_lock(&dq_data_lock);
+			/* Add usage from quota entry into quota changes
+			 * of our node. Auxiliary variables are important
+			 * due to signedness */
+			spacechange = le64_to_cpu(dqblk->dqb_spacemod);
+			inodechange = le64_to_cpu(dqblk->dqb_inodemod);
+			dquot->dq_dqb.dqb_curspace += spacechange;
+			dquot->dq_dqb.dqb_curinodes += inodechange;
+			spin_unlock(&dq_data_lock);
+			/* We want to drop reference held by the crashed
+			 * node. Since we have our own reference we know
+			 * global structure actually won't be freed. */
+			status = ocfs2_global_release_dquot(dquot);
+			if (status < 0) {
+				mlog_errno(status);
+				goto out_commit;
+			}
+			/* Release local quota file entry */
+			status = ocfs2_journal_access_dq(handle, lqinode,
+					qbh, OCFS2_JOURNAL_ACCESS_WRITE);
+			if (status < 0) {
+				mlog_errno(status);
+				goto out_commit;
+			}
+			lock_buffer(qbh);
+			WARN_ON(!ocfs2_test_bit(bit, dchunk->dqc_bitmap));
+			ocfs2_clear_bit(bit, dchunk->dqc_bitmap);
+			le32_add_cpu(&dchunk->dqc_free, 1);
+			unlock_buffer(qbh);
+			status = ocfs2_journal_dirty(handle, qbh);
+			if (status < 0)
+				mlog_errno(status);
+out_commit:
+			mutex_unlock(&sb_dqopt(sb)->dqio_mutex);
+			ocfs2_commit_trans(OCFS2_SB(sb), handle);
+out_put_dquot:
+			dqput(dquot);
+out_put_bh:
+			brelse(qbh);
+			if (status < 0)
+				break;
+		}
+		brelse(hbh);
+		list_del(&rchunk->rc_list);
+		kfree(rchunk->rc_bitmap);
+		kfree(rchunk);
+		if (status < 0)
+			break;
+	}
+	ocfs2_unlock_global_qf(oinfo, 1);
+out:
+	if (status < 0)
+		free_recovery_list(&(rec->r_list[type]));
+	mlog_exit(status);
+	return status;
+}
+
+/* Recover local quota files for given node different from us */
+int ocfs2_finish_quota_recovery(struct ocfs2_super *osb,
+				struct ocfs2_quota_recovery *rec,
+				int slot_num)
+{
+	unsigned int ino[MAXQUOTAS] = { LOCAL_USER_QUOTA_SYSTEM_INODE,
+					LOCAL_GROUP_QUOTA_SYSTEM_INODE };
+	struct super_block *sb = osb->sb;
+	struct ocfs2_local_disk_dqinfo *ldinfo;
+	struct buffer_head *bh;
+	handle_t *handle;
+	int type;
+	int status = 0;
+	struct inode *lqinode;
+	unsigned int flags;
+
+	mlog(ML_NOTICE, "Finishing quota recovery in slot %u\n", slot_num);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
+	for (type = 0; type < MAXQUOTAS; type++) {
+		if (list_empty(&(rec->r_list[type])))
+			continue;
+		mlog(0, "Recovering quota in slot %d\n", slot_num);
+		lqinode = ocfs2_get_system_file_inode(osb, ino[type], slot_num);
+		if (!lqinode) {
+			status = -ENOENT;
+			goto out;
+		}
+		status = ocfs2_inode_lock_full(lqinode, NULL, 1,
+						       OCFS2_META_LOCK_NOQUEUE);
+		/* Someone else is holding the lock? Then he must be
+		 * doing the recovery. Just skip the file... */
+		if (status == -EAGAIN) {
+			mlog(ML_NOTICE, "skipping quota recovery for slot %d "
+			     "because quota file is locked.\n", slot_num);
+			status = 0;
+			goto out_put;
+		} else if (status < 0) {
+			mlog_errno(status);
+			goto out_put;
+		}
+		/* Now read local header */
+		bh = NULL;
+		status = ocfs2_read_quota_block(lqinode, 0, &bh);
+		if (status) {
+			mlog_errno(status);
+			mlog(ML_ERROR, "failed to read quota file info header "
+				"(slot=%d type=%d)\n", slot_num, type);
+			goto out_lock;
+		}
+		ldinfo = (struct ocfs2_local_disk_dqinfo *)(bh->b_data +
+							OCFS2_LOCAL_INFO_OFF);
+		/* Is recovery still needed? */
+		flags = le32_to_cpu(ldinfo->dqi_flags);
+		if (!(flags & OLQF_CLEAN))
+			status = ocfs2_recover_local_quota_file(lqinode,
+								type,
+								rec);
+		/* We don't want to mark file as clean when it is actually
+		 * active */
+		if (slot_num == osb->slot_num)
+			goto out_bh;
+		/* Mark quota file as clean if we are recovering quota file of
+		 * some other node. */
+		handle = ocfs2_start_trans(osb, 1);
+		if (IS_ERR(handle)) {
+			status = PTR_ERR(handle);
+			mlog_errno(status);
+			goto out_bh;
+		}
+		status = ocfs2_journal_access_dq(handle, lqinode, bh,
+						 OCFS2_JOURNAL_ACCESS_WRITE);
+		if (status < 0) {
+			mlog_errno(status);
+			goto out_trans;
+		}
+		lock_buffer(bh);
+		ldinfo->dqi_flags = cpu_to_le32(flags | OLQF_CLEAN);
+		unlock_buffer(bh);
+		status = ocfs2_journal_dirty(handle, bh);
+		if (status < 0)
+			mlog_errno(status);
+out_trans:
+		ocfs2_commit_trans(osb, handle);
+out_bh:
+		brelse(bh);
+out_lock:
+		ocfs2_inode_unlock(lqinode, 1);
+out_put:
+		iput(lqinode);
+		if (status < 0)
+			break;
+	}
+out:
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
+	kfree(rec);
+	return status;
+}
+
+/* Read information header from quota file */
+static int ocfs2_local_read_info(struct super_block *sb, int type)
+{
+	struct ocfs2_local_disk_dqinfo *ldinfo;
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo;
+	struct inode *lqinode = sb_dqopt(sb)->files[type];
+	int status;
+	struct buffer_head *bh = NULL;
+	struct ocfs2_quota_recovery *rec;
+	int locked = 0;
+
+	info->dqi_maxblimit = 0x7fffffffffffffffLL;
+	info->dqi_maxilimit = 0x7fffffffffffffffLL;
+	oinfo = kmalloc(sizeof(struct ocfs2_mem_dqinfo), GFP_NOFS);
+	if (!oinfo) {
+		mlog(ML_ERROR, "failed to allocate memory for ocfs2 quota"
+			       " info.");
+		goto out_err;
+	}
+	info->dqi_priv = oinfo;
+	oinfo->dqi_type = type;
+	INIT_LIST_HEAD(&oinfo->dqi_chunk);
+	oinfo->dqi_rec = NULL;
+	oinfo->dqi_lqi_bh = NULL;
+	oinfo->dqi_ibh = NULL;
+
+	status = ocfs2_global_read_info(sb, type);
+	if (status < 0)
+		goto out_err;
+
+	status = ocfs2_inode_lock(lqinode, &oinfo->dqi_lqi_bh, 1);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_err;
+	}
+	locked = 1;
+
+	/* Now read local header */
+	status = ocfs2_read_quota_block(lqinode, 0, &bh);
+	if (status) {
+		mlog_errno(status);
+		mlog(ML_ERROR, "failed to read quota file info header "
+			"(type=%d)\n", type);
+		goto out_err;
+	}
+	ldinfo = (struct ocfs2_local_disk_dqinfo *)(bh->b_data +
+						OCFS2_LOCAL_INFO_OFF);
+	info->dqi_flags = le32_to_cpu(ldinfo->dqi_flags);
+	oinfo->dqi_chunks = le32_to_cpu(ldinfo->dqi_chunks);
+	oinfo->dqi_blocks = le32_to_cpu(ldinfo->dqi_blocks);
+	oinfo->dqi_ibh = bh;
+
+	/* We crashed when using local quota file? */
+	if (!(info->dqi_flags & OLQF_CLEAN)) {
+		rec = OCFS2_SB(sb)->quota_rec;
+		if (!rec) {
+			rec = ocfs2_alloc_quota_recovery();
+			if (!rec) {
+				status = -ENOMEM;
+				mlog_errno(status);
+				goto out_err;
+			}
+			OCFS2_SB(sb)->quota_rec = rec;
+		}
+
+		status = ocfs2_recovery_load_quota(lqinode, ldinfo, type,
+                                                   &rec->r_list[type]);
+		if (status < 0) {
+			mlog_errno(status);
+			goto out_err;
+		}
+	}
+
+	status = ocfs2_load_local_quota_bitmaps(lqinode,
+						ldinfo,
+						&oinfo->dqi_chunk);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_err;
+	}
+
+	/* Now mark quota file as used */
+	info->dqi_flags &= ~OLQF_CLEAN;
+	status = ocfs2_modify_bh(lqinode, bh, olq_update_info, info);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_err;
+	}
+
+	return 0;
+out_err:
+	if (oinfo) {
+		iput(oinfo->dqi_gqinode);
+		ocfs2_simple_drop_lockres(OCFS2_SB(sb), &oinfo->dqi_gqlock);
+		ocfs2_lock_res_free(&oinfo->dqi_gqlock);
+		brelse(oinfo->dqi_lqi_bh);
+		if (locked)
+			ocfs2_inode_unlock(lqinode, 1);
+		ocfs2_release_local_quota_bitmaps(&oinfo->dqi_chunk);
+		kfree(oinfo);
+	}
+	brelse(bh);
+	return -1;
+}
+
+/* Write local info to quota file */
+static int ocfs2_local_write_info(struct super_block *sb, int type)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct buffer_head *bh = ((struct ocfs2_mem_dqinfo *)info->dqi_priv)
+						->dqi_ibh;
+	int status;
+
+	status = ocfs2_modify_bh(sb_dqopt(sb)->files[type], bh, olq_update_info,
+				 info);
+	if (status < 0) {
+		mlog_errno(status);
+		return -1;
+	}
+
+	return 0;
+}
+
+/* Release info from memory */
+static int ocfs2_local_free_info(struct super_block *sb, int type)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct ocfs2_quota_chunk *chunk;
+	struct ocfs2_local_disk_chunk *dchunk;
+	int mark_clean = 1, len;
+	int status;
+
+	/* At this point we know there are no more dquots and thus
+	 * even if there's some sync in the pdflush queue, it won't
+	 * find any dquots and return without doing anything */
+	cancel_delayed_work_sync(&oinfo->dqi_sync_work);
+	iput(oinfo->dqi_gqinode);
+	ocfs2_simple_drop_lockres(OCFS2_SB(sb), &oinfo->dqi_gqlock);
+	ocfs2_lock_res_free(&oinfo->dqi_gqlock);
+	list_for_each_entry(chunk, &oinfo->dqi_chunk, qc_chunk) {
+		dchunk = (struct ocfs2_local_disk_chunk *)
+					(chunk->qc_headerbh->b_data);
+		if (chunk->qc_num < oinfo->dqi_chunks - 1) {
+			len = ol_chunk_entries(sb);
+		} else {
+			len = (oinfo->dqi_blocks -
+			       ol_quota_chunk_block(sb, chunk->qc_num) - 1)
+			      * ol_quota_entries_per_block(sb);
+		}
+		/* Not all entries free? Bug! */
+		if (le32_to_cpu(dchunk->dqc_free) != len) {
+			mlog(ML_ERROR, "releasing quota file with used "
+					"entries (type=%d)\n", type);
+			mark_clean = 0;
+		}
+	}
+	ocfs2_release_local_quota_bitmaps(&oinfo->dqi_chunk);
+
+	/* dqonoff_mutex protects us against racing with recovery thread... */
+	if (oinfo->dqi_rec) {
+		ocfs2_free_quota_recovery(oinfo->dqi_rec);
+		mark_clean = 0;
+	}
+
+	if (!mark_clean)
+		goto out;
+
+	/* Mark local file as clean */
+	info->dqi_flags |= OLQF_CLEAN;
+	status = ocfs2_modify_bh(sb_dqopt(sb)->files[type],
+				 oinfo->dqi_ibh,
+				 olq_update_info,
+				 info);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+out:
+	ocfs2_inode_unlock(sb_dqopt(sb)->files[type], 1);
+	brelse(oinfo->dqi_ibh);
+	brelse(oinfo->dqi_lqi_bh);
+	kfree(oinfo);
+	return 0;
+}
+
+static void olq_set_dquot(struct buffer_head *bh, void *private)
+{
+	struct ocfs2_dquot *od = private;
+	struct ocfs2_local_disk_dqblk *dqblk;
+	struct super_block *sb = od->dq_dquot.dq_sb;
+
+	dqblk = (struct ocfs2_local_disk_dqblk *)(bh->b_data
+		+ ol_dqblk_block_offset(sb, od->dq_local_off));
+
+	dqblk->dqb_id = cpu_to_le64(od->dq_dquot.dq_id);
+	spin_lock(&dq_data_lock);
+	dqblk->dqb_spacemod = cpu_to_le64(od->dq_dquot.dq_dqb.dqb_curspace -
+					  od->dq_origspace);
+	dqblk->dqb_inodemod = cpu_to_le64(od->dq_dquot.dq_dqb.dqb_curinodes -
+					  od->dq_originodes);
+	spin_unlock(&dq_data_lock);
+	mlog(0, "Writing local dquot %u space %lld inodes %lld\n",
+	     od->dq_dquot.dq_id, (long long)le64_to_cpu(dqblk->dqb_spacemod),
+	     (long long)le64_to_cpu(dqblk->dqb_inodemod));
+}
+
+/* Write dquot to local quota file */
+static int ocfs2_local_write_dquot(struct dquot *dquot)
+{
+	struct super_block *sb = dquot->dq_sb;
+	struct ocfs2_dquot *od = OCFS2_DQUOT(dquot);
+	struct buffer_head *bh = NULL;
+	int status;
+
+	status = ocfs2_read_quota_block(sb_dqopt(sb)->files[dquot->dq_type],
+				    ol_dqblk_file_block(sb, od->dq_local_off),
+				    &bh);
+	if (status) {
+		mlog_errno(status);
+		goto out;
+	}
+	status = ocfs2_modify_bh(sb_dqopt(sb)->files[dquot->dq_type], bh,
+				 olq_set_dquot, od);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+out:
+	brelse(bh);
+	return status;
+}
+
+/* Find free entry in local quota file */
+static struct ocfs2_quota_chunk *ocfs2_find_free_entry(struct super_block *sb,
+						       int type,
+						       int *offset)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct ocfs2_quota_chunk *chunk;
+	struct ocfs2_local_disk_chunk *dchunk;
+	int found = 0, len;
+
+	list_for_each_entry(chunk, &oinfo->dqi_chunk, qc_chunk) {
+		dchunk = (struct ocfs2_local_disk_chunk *)
+						chunk->qc_headerbh->b_data;
+		if (le32_to_cpu(dchunk->dqc_free) > 0) {
+			found = 1;
+			break;
+		}
+	}
+	if (!found)
+		return NULL;
+
+	if (chunk->qc_num < oinfo->dqi_chunks - 1) {
+		len = ol_chunk_entries(sb);
+	} else {
+		len = (oinfo->dqi_blocks -
+		       ol_quota_chunk_block(sb, chunk->qc_num) - 1)
+		      * ol_quota_entries_per_block(sb);
+	}
+
+	found = ocfs2_find_next_zero_bit(dchunk->dqc_bitmap, len, 0);
+	/* We failed? */
+	if (found == len) {
+		mlog(ML_ERROR, "Did not find empty entry in chunk %d with %u"
+		     " entries free (type=%d)\n", chunk->qc_num,
+		     le32_to_cpu(dchunk->dqc_free), type);
+		return ERR_PTR(-EIO);
+	}
+	*offset = found;
+	return chunk;
+}
+
+/* Add new chunk to the local quota file */
+static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
+							struct super_block *sb,
+							int type,
+							int *offset)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct inode *lqinode = sb_dqopt(sb)->files[type];
+	struct ocfs2_quota_chunk *chunk = NULL;
+	struct ocfs2_local_disk_chunk *dchunk;
+	int status;
+	handle_t *handle;
+	struct buffer_head *bh = NULL;
+	u64 p_blkno;
+
+	/* We are protected by dqio_sem so no locking needed */
+	status = ocfs2_extend_no_holes(lqinode,
+				       lqinode->i_size + 2 * sb->s_blocksize,
+				       lqinode->i_size);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	status = ocfs2_simple_size_update(lqinode, oinfo->dqi_lqi_bh,
+					  lqinode->i_size + 2 * sb->s_blocksize);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	chunk = kmem_cache_alloc(ocfs2_qf_chunk_cachep, GFP_NOFS);
+	if (!chunk) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto out;
+	}
+
+	down_read(&OCFS2_I(lqinode)->ip_alloc_sem);
+	status = ocfs2_extent_map_get_blocks(lqinode, oinfo->dqi_blocks,
+					     &p_blkno, NULL, NULL);
+	up_read(&OCFS2_I(lqinode)->ip_alloc_sem);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	bh = sb_getblk(sb, p_blkno);
+	if (!bh) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto out;
+	}
+	dchunk = (struct ocfs2_local_disk_chunk *)bh->b_data;
+
+	handle = ocfs2_start_trans(OCFS2_SB(sb), 2);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out;
+	}
+
+	status = ocfs2_journal_access_dq(handle, lqinode, bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	lock_buffer(bh);
+	dchunk->dqc_free = cpu_to_le32(ol_quota_entries_per_block(sb));
+	memset(dchunk->dqc_bitmap, 0,
+	       sb->s_blocksize - sizeof(struct ocfs2_local_disk_chunk) -
+	       OCFS2_QBLK_RESERVED_SPACE);
+	set_buffer_uptodate(bh);
+	unlock_buffer(bh);
+	status = ocfs2_journal_dirty(handle, bh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+
+	oinfo->dqi_blocks += 2;
+	oinfo->dqi_chunks++;
+	status = ocfs2_local_write_info(sb, type);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	status = ocfs2_commit_trans(OCFS2_SB(sb), handle);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	list_add_tail(&chunk->qc_chunk, &oinfo->dqi_chunk);
+	chunk->qc_num = list_entry(chunk->qc_chunk.prev,
+				   struct ocfs2_quota_chunk,
+				   qc_chunk)->qc_num + 1;
+	chunk->qc_headerbh = bh;
+	*offset = 0;
+	return chunk;
+out_trans:
+	ocfs2_commit_trans(OCFS2_SB(sb), handle);
+out:
+	brelse(bh);
+	kmem_cache_free(ocfs2_qf_chunk_cachep, chunk);
+	return ERR_PTR(status);
+}
+
+/* Find free entry in local quota file */
+static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
+						       struct super_block *sb,
+						       int type,
+						       int *offset)
+{
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct ocfs2_mem_dqinfo *oinfo = info->dqi_priv;
+	struct ocfs2_quota_chunk *chunk;
+	struct inode *lqinode = sb_dqopt(sb)->files[type];
+	struct ocfs2_local_disk_chunk *dchunk;
+	int epb = ol_quota_entries_per_block(sb);
+	unsigned int chunk_blocks;
+	int status;
+	handle_t *handle;
+
+	if (list_empty(&oinfo->dqi_chunk))
+		return ocfs2_local_quota_add_chunk(sb, type, offset);
+	/* Is the last chunk full? */
+	chunk = list_entry(oinfo->dqi_chunk.prev,
+			struct ocfs2_quota_chunk, qc_chunk);
+	chunk_blocks = oinfo->dqi_blocks -
+			ol_quota_chunk_block(sb, chunk->qc_num) - 1;
+	if (ol_chunk_blocks(sb) == chunk_blocks)
+		return ocfs2_local_quota_add_chunk(sb, type, offset);
+
+	/* We are protected by dqio_sem so no locking needed */
+	status = ocfs2_extend_no_holes(lqinode,
+				       lqinode->i_size + sb->s_blocksize,
+				       lqinode->i_size);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	status = ocfs2_simple_size_update(lqinode, oinfo->dqi_lqi_bh,
+					  lqinode->i_size + sb->s_blocksize);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	handle = ocfs2_start_trans(OCFS2_SB(sb), 2);
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		mlog_errno(status);
+		goto out;
+	}
+	status = ocfs2_journal_access_dq(handle, lqinode, chunk->qc_headerbh,
+				 OCFS2_JOURNAL_ACCESS_WRITE);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+
+	dchunk = (struct ocfs2_local_disk_chunk *)chunk->qc_headerbh->b_data;
+	lock_buffer(chunk->qc_headerbh);
+	le32_add_cpu(&dchunk->dqc_free, ol_quota_entries_per_block(sb));
+	unlock_buffer(chunk->qc_headerbh);
+	status = ocfs2_journal_dirty(handle, chunk->qc_headerbh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+	oinfo->dqi_blocks++;
+	status = ocfs2_local_write_info(sb, type);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_trans;
+	}
+
+	status = ocfs2_commit_trans(OCFS2_SB(sb), handle);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	*offset = chunk_blocks * epb;
+	return chunk;
+out_trans:
+	ocfs2_commit_trans(OCFS2_SB(sb), handle);
+out:
+	return ERR_PTR(status);
+}
+
+static void olq_alloc_dquot(struct buffer_head *bh, void *private)
+{
+	int *offset = private;
+	struct ocfs2_local_disk_chunk *dchunk;
+
+	dchunk = (struct ocfs2_local_disk_chunk *)bh->b_data;
+	ocfs2_set_bit(*offset, dchunk->dqc_bitmap);
+	le32_add_cpu(&dchunk->dqc_free, -1);
+}
+
+/* Create dquot in the local file for given id */
+static int ocfs2_create_local_dquot(struct dquot *dquot)
+{
+	struct super_block *sb = dquot->dq_sb;
+	int type = dquot->dq_type;
+	struct inode *lqinode = sb_dqopt(sb)->files[type];
+	struct ocfs2_quota_chunk *chunk;
+	struct ocfs2_dquot *od = OCFS2_DQUOT(dquot);
+	int offset;
+	int status;
+
+	chunk = ocfs2_find_free_entry(sb, type, &offset);
+	if (!chunk) {
+		chunk = ocfs2_extend_local_quota_file(sb, type, &offset);
+		if (IS_ERR(chunk))
+			return PTR_ERR(chunk);
+	} else if (IS_ERR(chunk)) {
+		return PTR_ERR(chunk);
+	}
+	od->dq_local_off = ol_dqblk_off(sb, chunk->qc_num, offset);
+	od->dq_chunk = chunk;
+
+	/* Initialize dquot structure on disk */
+	status = ocfs2_local_write_dquot(dquot);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	/* Mark structure as allocated */
+	status = ocfs2_modify_bh(lqinode, chunk->qc_headerbh, olq_alloc_dquot,
+				 &offset);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+out:
+	return status;
+}
+
+/* Create entry in local file for dquot, load data from the global file */
+static int ocfs2_local_read_dquot(struct dquot *dquot)
+{
+	int status;
+
+	mlog_entry("id=%u, type=%d\n", dquot->dq_id, dquot->dq_type);
+
+	status = ocfs2_global_read_dquot(dquot);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_err;
+	}
+
+	/* Now create entry in the local quota file */
+	status = ocfs2_create_local_dquot(dquot);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_err;
+	}
+	mlog_exit(0);
+	return 0;
+out_err:
+	mlog_exit(status);
+	return status;
+}
+
+/* Release dquot structure from local quota file. ocfs2_release_dquot() has
+ * already started a transaction and obtained exclusive lock for global
+ * quota file. */
+static int ocfs2_local_release_dquot(struct dquot *dquot)
+{
+	int status;
+	int type = dquot->dq_type;
+	struct ocfs2_dquot *od = OCFS2_DQUOT(dquot);
+	struct super_block *sb = dquot->dq_sb;
+	struct ocfs2_local_disk_chunk *dchunk;
+	int offset;
+	handle_t *handle = journal_current_handle();
+
+	BUG_ON(!handle);
+	/* First write all local changes to global file */
+	status = ocfs2_global_release_dquot(dquot);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	status = ocfs2_journal_access_dq(handle, sb_dqopt(sb)->files[type],
+			od->dq_chunk->qc_headerbh, OCFS2_JOURNAL_ACCESS_WRITE);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	offset = ol_dqblk_chunk_off(sb, od->dq_chunk->qc_num,
+					     od->dq_local_off);
+	dchunk = (struct ocfs2_local_disk_chunk *)
+			(od->dq_chunk->qc_headerbh->b_data);
+	/* Mark structure as freed */
+	lock_buffer(od->dq_chunk->qc_headerbh);
+	ocfs2_clear_bit(offset, dchunk->dqc_bitmap);
+	le32_add_cpu(&dchunk->dqc_free, 1);
+	unlock_buffer(od->dq_chunk->qc_headerbh);
+	status = ocfs2_journal_dirty(handle, od->dq_chunk->qc_headerbh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out;
+	}
+	status = 0;
+out:
+	/* Clear the read bit so that next time someone uses this
+	 * dquot he reads fresh info from disk and allocates local
+	 * dquot structure */
+	clear_bit(DQ_READ_B, &dquot->dq_flags);
+	return status;
+}
+
+static struct quota_format_ops ocfs2_format_ops = {
+	.check_quota_file	= ocfs2_local_check_quota_file,
+	.read_file_info		= ocfs2_local_read_info,
+	.write_file_info	= ocfs2_global_write_info,
+	.free_file_info		= ocfs2_local_free_info,
+	.read_dqblk		= ocfs2_local_read_dquot,
+	.commit_dqblk		= ocfs2_local_write_dquot,
+	.release_dqblk		= ocfs2_local_release_dquot,
+};
+
+struct quota_format_type ocfs2_quota_format = {
+	.qf_fmt_id = QFMT_OCFS2,
+	.qf_ops = &ocfs2_format_ops,
+	.qf_owner = THIS_MODULE
+};
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index ffd48db229a..424adaa5f90 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -106,8 +106,8 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
 	mlog_entry("(new_clusters=%d, first_new_cluster = %u)\n",
 		   new_clusters, first_new_cluster);
 
-	ret = ocfs2_journal_access(handle, bm_inode, group_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_gd(handle, bm_inode, group_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -141,8 +141,8 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
 	}
 
 	/* update the inode accordingly. */
-	ret = ocfs2_journal_access(handle, bm_inode, bm_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, bm_inode, bm_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out_rollback;
@@ -314,6 +314,10 @@ int ocfs2_group_extend(struct inode * inode, int new_clusters)
 
 	fe = (struct ocfs2_dinode *)main_bm_bh->b_data;
 
+	/* main_bm_bh is validated by inode read inside ocfs2_inode_lock(),
+	 * so any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
+
 	if (le16_to_cpu(fe->id2.i_chain.cl_cpg) !=
 				 ocfs2_group_bitmap_size(osb->sb) * 8) {
 		mlog(ML_ERROR, "The disk is too old and small. "
@@ -322,30 +326,18 @@ int ocfs2_group_extend(struct inode * inode, int new_clusters)
 		goto out_unlock;
 	}
 
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(main_bm_inode->i_sb, fe);
-		ret = -EIO;
-		goto out_unlock;
-	}
-
 	first_new_cluster = le32_to_cpu(fe->i_clusters);
 	lgd_blkno = ocfs2_which_cluster_group(main_bm_inode,
 					      first_new_cluster - 1);
 
-	ret = ocfs2_read_block(main_bm_inode, lgd_blkno, &group_bh);
+	ret = ocfs2_read_group_descriptor(main_bm_inode, fe, lgd_blkno,
+					  &group_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out_unlock;
 	}
-
 	group = (struct ocfs2_group_desc *)group_bh->b_data;
 
-	ret = ocfs2_check_group_descriptor(inode->i_sb, fe, group);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_unlock;
-	}
-
 	cl_bpc = le16_to_cpu(fe->id2.i_chain.cl_bpc);
 	if (le16_to_cpu(group->bg_bits) / cl_bpc + new_clusters >
 		le16_to_cpu(fe->id2.i_chain.cl_cpg)) {
@@ -398,41 +390,16 @@ static int ocfs2_check_new_group(struct inode *inode,
 				 struct buffer_head *group_bh)
 {
 	int ret;
-	struct ocfs2_group_desc *gd;
+	struct ocfs2_group_desc *gd =
+		(struct ocfs2_group_desc *)group_bh->b_data;
 	u16 cl_bpc = le16_to_cpu(di->id2.i_chain.cl_bpc);
-	unsigned int max_bits = le16_to_cpu(di->id2.i_chain.cl_cpg) *
-				le16_to_cpu(di->id2.i_chain.cl_bpc);
-
 
-	gd = (struct ocfs2_group_desc *)group_bh->b_data;
+	ret = ocfs2_check_group_descriptor(inode->i_sb, di, group_bh);
+	if (ret)
+		goto out;
 
-	ret = -EIO;
-	if (!OCFS2_IS_VALID_GROUP_DESC(gd))
-		mlog(ML_ERROR, "Group descriptor # %llu isn't valid.\n",
-		     (unsigned long long)le64_to_cpu(gd->bg_blkno));
-	else if (di->i_blkno != gd->bg_parent_dinode)
-		mlog(ML_ERROR, "Group descriptor # %llu has bad parent "
-		     "pointer (%llu, expected %llu)\n",
-		     (unsigned long long)le64_to_cpu(gd->bg_blkno),
-		     (unsigned long long)le64_to_cpu(gd->bg_parent_dinode),
-		     (unsigned long long)le64_to_cpu(di->i_blkno));
-	else if (le16_to_cpu(gd->bg_bits) > max_bits)
-		mlog(ML_ERROR, "Group descriptor # %llu has bit count of %u\n",
-		     (unsigned long long)le64_to_cpu(gd->bg_blkno),
-		     le16_to_cpu(gd->bg_bits));
-	else if (le16_to_cpu(gd->bg_free_bits_count) > le16_to_cpu(gd->bg_bits))
-		mlog(ML_ERROR, "Group descriptor # %llu has bit count %u but "
-		     "claims that %u are free\n",
-		     (unsigned long long)le64_to_cpu(gd->bg_blkno),
-		     le16_to_cpu(gd->bg_bits),
-		     le16_to_cpu(gd->bg_free_bits_count));
-	else if (le16_to_cpu(gd->bg_bits) > (8 * le16_to_cpu(gd->bg_size)))
-		mlog(ML_ERROR, "Group descriptor # %llu has bit count %u but "
-		     "max bitmap bits of %u\n",
-		     (unsigned long long)le64_to_cpu(gd->bg_blkno),
-		     le16_to_cpu(gd->bg_bits),
-		     8 * le16_to_cpu(gd->bg_size));
-	else if (le16_to_cpu(gd->bg_chain) != input->chain)
+	ret = -EINVAL;
+	if (le16_to_cpu(gd->bg_chain) != input->chain)
 		mlog(ML_ERROR, "Group descriptor # %llu has bad chain %u "
 		     "while input has %u set.\n",
 		     (unsigned long long)le64_to_cpu(gd->bg_blkno),
@@ -451,6 +418,7 @@ static int ocfs2_check_new_group(struct inode *inode,
 	else
 		ret = 0;
 
+out:
 	return ret;
 }
 
@@ -568,8 +536,8 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
 	cl = &fe->id2.i_chain;
 	cr = &cl->cl_recs[input->chain];
 
-	ret = ocfs2_journal_access(handle, main_bm_inode, group_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_gd(handle, main_bm_inode, group_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -584,8 +552,8 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
 		goto out_commit;
 	}
 
-	ret = ocfs2_journal_access(handle, main_bm_inode, main_bm_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, main_bm_inode, main_bm_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out_commit;
diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c
index bdda2d8f850..40661e7824e 100644
--- a/fs/ocfs2/slot_map.c
+++ b/fs/ocfs2/slot_map.c
@@ -151,7 +151,7 @@ int ocfs2_refresh_slot_info(struct ocfs2_super *osb)
 	 * this is not true, the read of -1 (UINT64_MAX) will fail.
 	 */
 	ret = ocfs2_read_blocks(si->si_inode, -1, si->si_blocks, si->si_bh,
-				OCFS2_BH_IGNORE_CACHE);
+				OCFS2_BH_IGNORE_CACHE, NULL);
 	if (ret == 0) {
 		spin_lock(&osb->osb_lock);
 		ocfs2_update_slot_info(si);
@@ -405,7 +405,7 @@ static int ocfs2_map_slot_buffers(struct ocfs2_super *osb,
 
 		bh = NULL;  /* Acquire a fresh bh */
 		status = ocfs2_read_blocks(si->si_inode, blkno, 1, &bh,
-					   OCFS2_BH_IGNORE_CACHE);
+					   OCFS2_BH_IGNORE_CACHE, NULL);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index c5ff18b46b5..a69628603e1 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -35,6 +35,7 @@
 #include "ocfs2.h"
 
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dlmglue.h"
 #include "inode.h"
 #include "journal.h"
@@ -145,62 +146,183 @@ static u32 ocfs2_bits_per_group(struct ocfs2_chain_list *cl)
 	return (u32)le16_to_cpu(cl->cl_cpg) * (u32)le16_to_cpu(cl->cl_bpc);
 }
 
-/* somewhat more expensive than our other checks, so use sparingly. */
-int ocfs2_check_group_descriptor(struct super_block *sb,
-				 struct ocfs2_dinode *di,
-				 struct ocfs2_group_desc *gd)
+#define do_error(fmt, ...)						\
+	do{								\
+		if (clean_error)					\
+			mlog(ML_ERROR, fmt "\n", ##__VA_ARGS__);	\
+		else							\
+			ocfs2_error(sb, fmt, ##__VA_ARGS__);		\
+	} while (0)
+
+static int ocfs2_validate_gd_self(struct super_block *sb,
+				  struct buffer_head *bh,
+				  int clean_error)
 {
-	unsigned int max_bits;
+	struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
 
 	if (!OCFS2_IS_VALID_GROUP_DESC(gd)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(sb, gd);
-		return -EIO;
+		do_error("Group descriptor #%llu has bad signature %.*s",
+			 (unsigned long long)bh->b_blocknr, 7,
+			 gd->bg_signature);
+		return -EINVAL;
 	}
 
+	if (le64_to_cpu(gd->bg_blkno) != bh->b_blocknr) {
+		do_error("Group descriptor #%llu has an invalid bg_blkno "
+			 "of %llu",
+			 (unsigned long long)bh->b_blocknr,
+			 (unsigned long long)le64_to_cpu(gd->bg_blkno));
+		return -EINVAL;
+	}
+
+	if (le32_to_cpu(gd->bg_generation) != OCFS2_SB(sb)->fs_generation) {
+		do_error("Group descriptor #%llu has an invalid "
+			 "fs_generation of #%u",
+			 (unsigned long long)bh->b_blocknr,
+			 le32_to_cpu(gd->bg_generation));
+		return -EINVAL;
+	}
+
+	if (le16_to_cpu(gd->bg_free_bits_count) > le16_to_cpu(gd->bg_bits)) {
+		do_error("Group descriptor #%llu has bit count %u but "
+			 "claims that %u are free",
+			 (unsigned long long)bh->b_blocknr,
+			 le16_to_cpu(gd->bg_bits),
+			 le16_to_cpu(gd->bg_free_bits_count));
+		return -EINVAL;
+	}
+
+	if (le16_to_cpu(gd->bg_bits) > (8 * le16_to_cpu(gd->bg_size))) {
+		do_error("Group descriptor #%llu has bit count %u but "
+			 "max bitmap bits of %u",
+			 (unsigned long long)bh->b_blocknr,
+			 le16_to_cpu(gd->bg_bits),
+			 8 * le16_to_cpu(gd->bg_size));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int ocfs2_validate_gd_parent(struct super_block *sb,
+				    struct ocfs2_dinode *di,
+				    struct buffer_head *bh,
+				    int clean_error)
+{
+	unsigned int max_bits;
+	struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
+
 	if (di->i_blkno != gd->bg_parent_dinode) {
-		ocfs2_error(sb, "Group descriptor # %llu has bad parent "
-			    "pointer (%llu, expected %llu)",
-			    (unsigned long long)le64_to_cpu(gd->bg_blkno),
-			    (unsigned long long)le64_to_cpu(gd->bg_parent_dinode),
-			    (unsigned long long)le64_to_cpu(di->i_blkno));
-		return -EIO;
+		do_error("Group descriptor #%llu has bad parent "
+			 "pointer (%llu, expected %llu)",
+			 (unsigned long long)bh->b_blocknr,
+			 (unsigned long long)le64_to_cpu(gd->bg_parent_dinode),
+			 (unsigned long long)le64_to_cpu(di->i_blkno));
+		return -EINVAL;
 	}
 
 	max_bits = le16_to_cpu(di->id2.i_chain.cl_cpg) * le16_to_cpu(di->id2.i_chain.cl_bpc);
 	if (le16_to_cpu(gd->bg_bits) > max_bits) {
-		ocfs2_error(sb, "Group descriptor # %llu has bit count of %u",
-			    (unsigned long long)le64_to_cpu(gd->bg_blkno),
-			    le16_to_cpu(gd->bg_bits));
-		return -EIO;
+		do_error("Group descriptor #%llu has bit count of %u",
+			 (unsigned long long)bh->b_blocknr,
+			 le16_to_cpu(gd->bg_bits));
+		return -EINVAL;
 	}
 
 	if (le16_to_cpu(gd->bg_chain) >=
 	    le16_to_cpu(di->id2.i_chain.cl_next_free_rec)) {
-		ocfs2_error(sb, "Group descriptor # %llu has bad chain %u",
-			    (unsigned long long)le64_to_cpu(gd->bg_blkno),
-			    le16_to_cpu(gd->bg_chain));
-		return -EIO;
+		do_error("Group descriptor #%llu has bad chain %u",
+			 (unsigned long long)bh->b_blocknr,
+			 le16_to_cpu(gd->bg_chain));
+		return -EINVAL;
 	}
 
-	if (le16_to_cpu(gd->bg_free_bits_count) > le16_to_cpu(gd->bg_bits)) {
-		ocfs2_error(sb, "Group descriptor # %llu has bit count %u but "
-			    "claims that %u are free",
-			    (unsigned long long)le64_to_cpu(gd->bg_blkno),
-			    le16_to_cpu(gd->bg_bits),
-			    le16_to_cpu(gd->bg_free_bits_count));
-		return -EIO;
-	}
+	return 0;
+}
 
-	if (le16_to_cpu(gd->bg_bits) > (8 * le16_to_cpu(gd->bg_size))) {
-		ocfs2_error(sb, "Group descriptor # %llu has bit count %u but "
-			    "max bitmap bits of %u",
-			    (unsigned long long)le64_to_cpu(gd->bg_blkno),
-			    le16_to_cpu(gd->bg_bits),
-			    8 * le16_to_cpu(gd->bg_size));
-		return -EIO;
+#undef do_error
+
+/*
+ * This version only prints errors.  It does not fail the filesystem, and
+ * exists only for resize.
+ */
+int ocfs2_check_group_descriptor(struct super_block *sb,
+				 struct ocfs2_dinode *di,
+				 struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 */
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &gd->bg_check);
+	if (rc) {
+		mlog(ML_ERROR,
+		     "Checksum failed for group descriptor %llu\n",
+		     (unsigned long long)bh->b_blocknr);
+	} else
+		rc = ocfs2_validate_gd_self(sb, bh, 1);
+	if (!rc)
+		rc = ocfs2_validate_gd_parent(sb, di, bh, 1);
+
+	return rc;
+}
+
+static int ocfs2_validate_group_descriptor(struct super_block *sb,
+					   struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
+
+	mlog(0, "Validating group descriptor %llu\n",
+	     (unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 */
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &gd->bg_check);
+	if (rc)
+		return rc;
+
+	/*
+	 * Errors after here are fatal.
+	 */
+
+	return ocfs2_validate_gd_self(sb, bh, 0);
+}
+
+int ocfs2_read_group_descriptor(struct inode *inode, struct ocfs2_dinode *di,
+				u64 gd_blkno, struct buffer_head **bh)
+{
+	int rc;
+	struct buffer_head *tmp = *bh;
+
+	rc = ocfs2_read_block(inode, gd_blkno, &tmp,
+			      ocfs2_validate_group_descriptor);
+	if (rc)
+		goto out;
+
+	rc = ocfs2_validate_gd_parent(inode->i_sb, di, tmp, 0);
+	if (rc) {
+		brelse(tmp);
+		goto out;
 	}
 
-	return 0;
+	/* If ocfs2_read_block() got us a new bh, pass it up. */
+	if (!*bh)
+		*bh = tmp;
+
+out:
+	return rc;
 }
 
 static int ocfs2_block_group_fill(handle_t *handle,
@@ -225,10 +347,10 @@ static int ocfs2_block_group_fill(handle_t *handle,
 		goto bail;
 	}
 
-	status = ocfs2_journal_access(handle,
-				      alloc_inode,
-				      bg_bh,
-				      OCFS2_JOURNAL_ACCESS_CREATE);
+	status = ocfs2_journal_access_gd(handle,
+					 alloc_inode,
+					 bg_bh,
+					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -358,8 +480,8 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
 
 	bg = (struct ocfs2_group_desc *) bg_bh->b_data;
 
-	status = ocfs2_journal_access(handle, alloc_inode,
-				      bh, OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, alloc_inode,
+					 bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -441,11 +563,11 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
 	ac->ac_alloc_slot = slot;
 
 	fe = (struct ocfs2_dinode *) bh->b_data;
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
-		status = -EIO;
-		goto bail;
-	}
+
+	/* The bh was validated by the inode read inside
+	 * ocfs2_inode_lock().  Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
+
 	if (!(fe->i_flags & cpu_to_le32(OCFS2_CHAIN_FL))) {
 		ocfs2_error(alloc_inode->i_sb, "Invalid chain allocator %llu",
 			    (unsigned long long)le64_to_cpu(fe->i_blkno));
@@ -790,10 +912,9 @@ static int ocfs2_block_group_find_clear_bits(struct ocfs2_super *osb,
 	int offset, start, found, status = 0;
 	struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
 
-	if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(osb->sb, bg);
-		return -EIO;
-	}
+	/* Callers got this descriptor from
+	 * ocfs2_read_group_descriptor().  Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
 
 	found = start = best_offset = best_size = 0;
 	bitmap = bg->bg_bitmap;
@@ -858,11 +979,9 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
 
 	mlog_entry_void();
 
-	if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
-		status = -EIO;
-		goto bail;
-	}
+	/* All callers get the descriptor via
+	 * ocfs2_read_group_descriptor().  Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
 	BUG_ON(le16_to_cpu(bg->bg_free_bits_count) < num_bits);
 
 	mlog(0, "block_group_set_bits: off = %u, num = %u\n", bit_off,
@@ -871,10 +990,10 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
 	if (ocfs2_is_cluster_bitmap(alloc_inode))
 		journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
 
-	status = ocfs2_journal_access(handle,
-				      alloc_inode,
-				      group_bh,
-				      journal_type);
+	status = ocfs2_journal_access_gd(handle,
+					 alloc_inode,
+					 group_bh,
+					 journal_type);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -931,21 +1050,10 @@ static int ocfs2_relink_block_group(handle_t *handle,
 	struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
 	struct ocfs2_group_desc *prev_bg = (struct ocfs2_group_desc *) prev_bg_bh->b_data;
 
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
-		status = -EIO;
-		goto out;
-	}
-	if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
-		status = -EIO;
-		goto out;
-	}
-	if (!OCFS2_IS_VALID_GROUP_DESC(prev_bg)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, prev_bg);
-		status = -EIO;
-		goto out;
-	}
+	/* The caller got these descriptors from
+	 * ocfs2_read_group_descriptor().  Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
+	BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(prev_bg));
 
 	mlog(0, "Suballoc %llu, chain %u, move group %llu to top, prev = %llu\n",
 	     (unsigned long long)le64_to_cpu(fe->i_blkno), chain,
@@ -956,8 +1064,8 @@ static int ocfs2_relink_block_group(handle_t *handle,
 	bg_ptr = le64_to_cpu(bg->bg_next_group);
 	prev_bg_ptr = le64_to_cpu(prev_bg->bg_next_group);
 
-	status = ocfs2_journal_access(handle, alloc_inode, prev_bg_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_gd(handle, alloc_inode, prev_bg_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_rollback;
@@ -971,8 +1079,8 @@ static int ocfs2_relink_block_group(handle_t *handle,
 		goto out_rollback;
 	}
 
-	status = ocfs2_journal_access(handle, alloc_inode, bg_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_gd(handle, alloc_inode, bg_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_rollback;
@@ -986,8 +1094,8 @@ static int ocfs2_relink_block_group(handle_t *handle,
 		goto out_rollback;
 	}
 
-	status = ocfs2_journal_access(handle, alloc_inode, fe_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, alloc_inode, fe_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_rollback;
@@ -1008,7 +1116,7 @@ out_rollback:
 		bg->bg_next_group = cpu_to_le64(bg_ptr);
 		prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr);
 	}
-out:
+
 	mlog_exit(status);
 	return status;
 }
@@ -1138,8 +1246,8 @@ static int ocfs2_alloc_dinode_update_counts(struct inode *inode,
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data;
 	struct ocfs2_chain_list *cl = (struct ocfs2_chain_list *) &di->id2.i_chain;
 
-	ret = ocfs2_journal_access(handle, inode, di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -1170,21 +1278,17 @@ static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac,
 	u16 found;
 	struct buffer_head *group_bh = NULL;
 	struct ocfs2_group_desc *gd;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)ac->ac_bh->b_data;
 	struct inode *alloc_inode = ac->ac_inode;
 
-	ret = ocfs2_read_block(alloc_inode, gd_blkno, &group_bh);
+	ret = ocfs2_read_group_descriptor(alloc_inode, di, gd_blkno,
+					  &group_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		return ret;
 	}
 
 	gd = (struct ocfs2_group_desc *) group_bh->b_data;
-	if (!OCFS2_IS_VALID_GROUP_DESC(gd)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, gd);
-		ret = -EIO;
-		goto out;
-	}
-
 	ret = ac->ac_group_search(alloc_inode, group_bh, bits_wanted, min_bits,
 				  ac->ac_max_block, bit_off, &found);
 	if (ret < 0) {
@@ -1241,19 +1345,14 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 	     bits_wanted, chain,
 	     (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno);
 
-	status = ocfs2_read_block(alloc_inode,
-				  le64_to_cpu(cl->cl_recs[chain].c_blkno),
-				  &group_bh);
+	status = ocfs2_read_group_descriptor(alloc_inode, fe,
+					     le64_to_cpu(cl->cl_recs[chain].c_blkno),
+					     &group_bh);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 	bg = (struct ocfs2_group_desc *) group_bh->b_data;
-	status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, bg);
-	if (status) {
-		mlog_errno(status);
-		goto bail;
-	}
 
 	status = -ENOSPC;
 	/* for now, the chain search is a bit simplistic. We just use
@@ -1271,18 +1370,13 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 		next_group = le64_to_cpu(bg->bg_next_group);
 		prev_group_bh = group_bh;
 		group_bh = NULL;
-		status = ocfs2_read_block(alloc_inode,
-					  next_group, &group_bh);
+		status = ocfs2_read_group_descriptor(alloc_inode, fe,
+						     next_group, &group_bh);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
 		}
 		bg = (struct ocfs2_group_desc *) group_bh->b_data;
-		status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, bg);
-		if (status) {
-			mlog_errno(status);
-			goto bail;
-		}
 	}
 	if (status < 0) {
 		if (status != -ENOSPC)
@@ -1324,10 +1418,10 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 
 	/* Ok, claim our bits now: set the info on dinode, chainlist
 	 * and then the group */
-	status = ocfs2_journal_access(handle,
-				      alloc_inode,
-				      ac->ac_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle,
+					 alloc_inode,
+					 ac->ac_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1392,11 +1486,11 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
 	BUG_ON(!ac->ac_bh);
 
 	fe = (struct ocfs2_dinode *) ac->ac_bh->b_data;
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(osb->sb, fe);
-		status = -EIO;
-		goto bail;
-	}
+
+	/* The bh was validated by the inode read during
+	 * ocfs2_reserve_suballoc_bits().  Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
+
 	if (le32_to_cpu(fe->id1.bitmap1.i_used) >=
 	    le32_to_cpu(fe->id1.bitmap1.i_total)) {
 		ocfs2_error(osb->sb, "Chain allocator dinode %llu has %u used "
@@ -1725,19 +1819,17 @@ static inline int ocfs2_block_group_clear_bits(handle_t *handle,
 
 	mlog_entry_void();
 
-	if (!OCFS2_IS_VALID_GROUP_DESC(bg)) {
-		OCFS2_RO_ON_INVALID_GROUP_DESC(alloc_inode->i_sb, bg);
-		status = -EIO;
-		goto bail;
-	}
+	/* The caller got this descriptor from
+	 * ocfs2_read_group_descriptor().  Any corruption is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
 
 	mlog(0, "off = %u, num = %u\n", bit_off, num_bits);
 
 	if (ocfs2_is_cluster_bitmap(alloc_inode))
 		journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
 
-	status = ocfs2_journal_access(handle, alloc_inode, group_bh,
-				      journal_type);
+	status = ocfs2_journal_access_gd(handle, alloc_inode, group_bh,
+					 journal_type);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1782,29 +1874,26 @@ int ocfs2_free_suballoc_bits(handle_t *handle,
 
 	mlog_entry_void();
 
-	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		OCFS2_RO_ON_INVALID_DINODE(alloc_inode->i_sb, fe);
-		status = -EIO;
-		goto bail;
-	}
+	/* The alloc_bh comes from ocfs2_free_dinode() or
+	 * ocfs2_free_clusters().  The callers have all locked the
+	 * allocator and gotten alloc_bh from the lock call.  This
+	 * validates the dinode buffer.  Any corruption that has happended
+	 * is a code bug. */
+	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
 	BUG_ON((count + start_bit) > ocfs2_bits_per_group(cl));
 
 	mlog(0, "%llu: freeing %u bits from group %llu, starting at %u\n",
 	     (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno, count,
 	     (unsigned long long)bg_blkno, start_bit);
 
-	status = ocfs2_read_block(alloc_inode, bg_blkno, &group_bh);
+	status = ocfs2_read_group_descriptor(alloc_inode, fe, bg_blkno,
+					     &group_bh);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
-
 	group = (struct ocfs2_group_desc *) group_bh->b_data;
-	status = ocfs2_check_group_descriptor(alloc_inode->i_sb, fe, group);
-	if (status) {
-		mlog_errno(status);
-		goto bail;
-	}
+
 	BUG_ON((count + start_bit) > le16_to_cpu(group->bg_bits));
 
 	status = ocfs2_block_group_clear_bits(handle, alloc_inode,
@@ -1815,8 +1904,8 @@ int ocfs2_free_suballoc_bits(handle_t *handle,
 		goto bail;
 	}
 
-	status = ocfs2_journal_access(handle, alloc_inode, alloc_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, alloc_inode, alloc_bh,
+					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h
index 4df159d8f45..e3c13c77f9e 100644
--- a/fs/ocfs2/suballoc.h
+++ b/fs/ocfs2/suballoc.h
@@ -164,10 +164,24 @@ void ocfs2_free_ac_resource(struct ocfs2_alloc_context *ac);
  * and return that block offset. */
 u64 ocfs2_which_cluster_group(struct inode *inode, u32 cluster);
 
-/* somewhat more expensive than our other checks, so use sparingly. */
+/*
+ * By default, ocfs2_read_group_descriptor() calls ocfs2_error() when it
+ * finds a problem.  A caller that wants to check a group descriptor
+ * without going readonly should read the block with ocfs2_read_block[s]()
+ * and then checking it with this function.  This is only resize, really.
+ * Everyone else should be using ocfs2_read_group_descriptor().
+ */
 int ocfs2_check_group_descriptor(struct super_block *sb,
 				 struct ocfs2_dinode *di,
-				 struct ocfs2_group_desc *gd);
+				 struct buffer_head *bh);
+/*
+ * Read a group descriptor block into *bh.  If *bh is NULL, a bh will be
+ * allocated.  This is a cached read.  The descriptor will be validated with
+ * ocfs2_validate_group_descriptor().
+ */
+int ocfs2_read_group_descriptor(struct inode *inode, struct ocfs2_dinode *di,
+				u64 gd_blkno, struct buffer_head **bh);
+
 int ocfs2_lock_allocators(struct inode *inode, struct ocfs2_extent_tree *et,
 			  u32 clusters_to_add, u32 extents_to_split,
 			  struct ocfs2_alloc_context **data_ac,
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 304b63ac78c..43ed11345b5 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -41,6 +41,7 @@
 #include <linux/debugfs.h>
 #include <linux/mount.h>
 #include <linux/seq_file.h>
+#include <linux/quotaops.h>
 
 #define MLOG_MASK_PREFIX ML_SUPER
 #include <cluster/masklog.h>
@@ -51,6 +52,7 @@
 #include "ocfs1_fs_compat.h"
 
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dlmglue.h"
 #include "export.h"
 #include "extent_map.h"
@@ -65,10 +67,13 @@
 #include "uptodate.h"
 #include "ver.h"
 #include "xattr.h"
+#include "quota.h"
 
 #include "buffer_head_io.h"
 
 static struct kmem_cache *ocfs2_inode_cachep = NULL;
+struct kmem_cache *ocfs2_dquot_cachep;
+struct kmem_cache *ocfs2_qf_chunk_cachep;
 
 /* OCFS2 needs to schedule several differnt types of work which
  * require cluster locking, disk I/O, recovery waits, etc. Since these
@@ -124,6 +129,9 @@ static int ocfs2_get_sector(struct super_block *sb,
 static void ocfs2_write_super(struct super_block *sb);
 static struct inode *ocfs2_alloc_inode(struct super_block *sb);
 static void ocfs2_destroy_inode(struct inode *inode);
+static int ocfs2_susp_quotas(struct ocfs2_super *osb, int unsuspend);
+static int ocfs2_enable_quotas(struct ocfs2_super *osb);
+static void ocfs2_disable_quotas(struct ocfs2_super *osb);
 
 static const struct super_operations ocfs2_sops = {
 	.statfs		= ocfs2_statfs,
@@ -137,6 +145,8 @@ static const struct super_operations ocfs2_sops = {
 	.put_super	= ocfs2_put_super,
 	.remount_fs	= ocfs2_remount,
 	.show_options   = ocfs2_show_options,
+	.quota_read	= ocfs2_quota_read,
+	.quota_write	= ocfs2_quota_write,
 };
 
 enum {
@@ -158,6 +168,10 @@ enum {
 	Opt_user_xattr,
 	Opt_nouser_xattr,
 	Opt_inode64,
+	Opt_acl,
+	Opt_noacl,
+	Opt_usrquota,
+	Opt_grpquota,
 	Opt_err,
 };
 
@@ -180,6 +194,10 @@ static const match_table_t tokens = {
 	{Opt_user_xattr, "user_xattr"},
 	{Opt_nouser_xattr, "nouser_xattr"},
 	{Opt_inode64, "inode64"},
+	{Opt_acl, "acl"},
+	{Opt_noacl, "noacl"},
+	{Opt_usrquota, "usrquota"},
+	{Opt_grpquota, "grpquota"},
 	{Opt_err, NULL}
 };
 
@@ -221,6 +239,19 @@ static int ocfs2_sync_fs(struct super_block *sb, int wait)
 	return 0;
 }
 
+static int ocfs2_need_system_inode(struct ocfs2_super *osb, int ino)
+{
+	if (!OCFS2_HAS_RO_COMPAT_FEATURE(osb->sb, OCFS2_FEATURE_RO_COMPAT_USRQUOTA)
+	    && (ino == USER_QUOTA_SYSTEM_INODE
+		|| ino == LOCAL_USER_QUOTA_SYSTEM_INODE))
+		return 0;
+	if (!OCFS2_HAS_RO_COMPAT_FEATURE(osb->sb, OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)
+	    && (ino == GROUP_QUOTA_SYSTEM_INODE
+		|| ino == LOCAL_GROUP_QUOTA_SYSTEM_INODE))
+		return 0;
+	return 1;
+}
+
 static int ocfs2_init_global_system_inodes(struct ocfs2_super *osb)
 {
 	struct inode *new = NULL;
@@ -247,6 +278,8 @@ static int ocfs2_init_global_system_inodes(struct ocfs2_super *osb)
 
 	for (i = OCFS2_FIRST_ONLINE_SYSTEM_INODE;
 	     i <= OCFS2_LAST_GLOBAL_SYSTEM_INODE; i++) {
+		if (!ocfs2_need_system_inode(osb, i))
+			continue;
 		new = ocfs2_get_system_file_inode(osb, i, osb->slot_num);
 		if (!new) {
 			ocfs2_release_system_inodes(osb);
@@ -277,6 +310,8 @@ static int ocfs2_init_local_system_inodes(struct ocfs2_super *osb)
 	for (i = OCFS2_LAST_GLOBAL_SYSTEM_INODE + 1;
 	     i < NUM_SYSTEM_INODES;
 	     i++) {
+		if (!ocfs2_need_system_inode(osb, i))
+			continue;
 		new = ocfs2_get_system_file_inode(osb, i, osb->slot_num);
 		if (!new) {
 			ocfs2_release_system_inodes(osb);
@@ -426,6 +461,12 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data)
 
 	/* We're going to/from readonly mode. */
 	if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY)) {
+		/* Disable quota accounting before remounting RO */
+		if (*flags & MS_RDONLY) {
+			ret = ocfs2_susp_quotas(osb, 0);
+			if (ret < 0)
+				goto out;
+		}
 		/* Lock here so the check of HARD_RO and the potential
 		 * setting of SOFT_RO is atomic. */
 		spin_lock(&osb->osb_lock);
@@ -461,11 +502,28 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data)
 		}
 unlock_osb:
 		spin_unlock(&osb->osb_lock);
+		/* Enable quota accounting after remounting RW */
+		if (!ret && !(*flags & MS_RDONLY)) {
+			if (sb_any_quota_suspended(sb))
+				ret = ocfs2_susp_quotas(osb, 1);
+			else
+				ret = ocfs2_enable_quotas(osb);
+			if (ret < 0) {
+				/* Return back changes... */
+				spin_lock(&osb->osb_lock);
+				sb->s_flags |= MS_RDONLY;
+				osb->osb_flags |= OCFS2_OSB_SOFT_RO;
+				spin_unlock(&osb->osb_lock);
+				goto out;
+			}
+		}
 	}
 
 	if (!ret) {
 		/* Only save off the new mount options in case of a successful
 		 * remount. */
+		if (!(osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_XATTR))
+			parsed_options.mount_opt &= ~OCFS2_MOUNT_POSIX_ACL;
 		osb->s_mount_opt = parsed_options.mount_opt;
 		osb->s_atime_quantum = parsed_options.atime_quantum;
 		osb->preferred_slot = parsed_options.slot;
@@ -619,6 +677,131 @@ static int ocfs2_verify_userspace_stack(struct ocfs2_super *osb,
 	return 0;
 }
 
+static int ocfs2_susp_quotas(struct ocfs2_super *osb, int unsuspend)
+{
+	int type;
+	struct super_block *sb = osb->sb;
+	unsigned int feature[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
+					     OCFS2_FEATURE_RO_COMPAT_GRPQUOTA};
+	int status = 0;
+
+	for (type = 0; type < MAXQUOTAS; type++) {
+		if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, feature[type]))
+			continue;
+		if (unsuspend)
+			status = vfs_quota_enable(
+					sb_dqopt(sb)->files[type],
+					type, QFMT_OCFS2,
+					DQUOT_SUSPENDED);
+		else
+			status = vfs_quota_disable(sb, type,
+						   DQUOT_SUSPENDED);
+		if (status < 0)
+			break;
+	}
+	if (status < 0)
+		mlog(ML_ERROR, "Failed to suspend/unsuspend quotas on "
+		     "remount (error = %d).\n", status);
+	return status;
+}
+
+static int ocfs2_enable_quotas(struct ocfs2_super *osb)
+{
+	struct inode *inode[MAXQUOTAS] = { NULL, NULL };
+	struct super_block *sb = osb->sb;
+	unsigned int feature[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
+					     OCFS2_FEATURE_RO_COMPAT_GRPQUOTA};
+	unsigned int ino[MAXQUOTAS] = { LOCAL_USER_QUOTA_SYSTEM_INODE,
+					LOCAL_GROUP_QUOTA_SYSTEM_INODE };
+	int status;
+	int type;
+
+	sb_dqopt(sb)->flags |= DQUOT_QUOTA_SYS_FILE | DQUOT_NEGATIVE_USAGE;
+	for (type = 0; type < MAXQUOTAS; type++) {
+		if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, feature[type]))
+			continue;
+		inode[type] = ocfs2_get_system_file_inode(osb, ino[type],
+							osb->slot_num);
+		if (!inode[type]) {
+			status = -ENOENT;
+			goto out_quota_off;
+		}
+		status = vfs_quota_enable(inode[type], type, QFMT_OCFS2,
+						DQUOT_USAGE_ENABLED);
+		if (status < 0)
+			goto out_quota_off;
+	}
+
+	for (type = 0; type < MAXQUOTAS; type++)
+		iput(inode[type]);
+	return 0;
+out_quota_off:
+	ocfs2_disable_quotas(osb);
+	for (type = 0; type < MAXQUOTAS; type++)
+		iput(inode[type]);
+	mlog_errno(status);
+	return status;
+}
+
+static void ocfs2_disable_quotas(struct ocfs2_super *osb)
+{
+	int type;
+	struct inode *inode;
+	struct super_block *sb = osb->sb;
+
+	/* We mostly ignore errors in this function because there's not much
+	 * we can do when we see them */
+	for (type = 0; type < MAXQUOTAS; type++) {
+		if (!sb_has_quota_loaded(sb, type))
+			continue;
+		inode = igrab(sb->s_dquot.files[type]);
+		/* Turn off quotas. This will remove all dquot structures from
+		 * memory and so they will be automatically synced to global
+		 * quota files */
+		vfs_quota_disable(sb, type, DQUOT_USAGE_ENABLED |
+					    DQUOT_LIMITS_ENABLED);
+		if (!inode)
+			continue;
+		iput(inode);
+	}
+}
+
+/* Handle quota on quotactl */
+static int ocfs2_quota_on(struct super_block *sb, int type, int format_id,
+			  char *path, int remount)
+{
+	unsigned int feature[MAXQUOTAS] = { OCFS2_FEATURE_RO_COMPAT_USRQUOTA,
+					     OCFS2_FEATURE_RO_COMPAT_GRPQUOTA};
+
+	if (!OCFS2_HAS_RO_COMPAT_FEATURE(sb, feature[type]))
+		return -EINVAL;
+
+	if (remount)
+		return 0;	/* Just ignore it has been handled in
+				 * ocfs2_remount() */
+	return vfs_quota_enable(sb_dqopt(sb)->files[type], type,
+				    format_id, DQUOT_LIMITS_ENABLED);
+}
+
+/* Handle quota off quotactl */
+static int ocfs2_quota_off(struct super_block *sb, int type, int remount)
+{
+	if (remount)
+		return 0;	/* Ignore now and handle later in
+				 * ocfs2_remount() */
+	return vfs_quota_disable(sb, type, DQUOT_LIMITS_ENABLED);
+}
+
+static struct quotactl_ops ocfs2_quotactl_ops = {
+	.quota_on	= ocfs2_quota_on,
+	.quota_off	= ocfs2_quota_off,
+	.quota_sync	= vfs_quota_sync,
+	.get_info	= vfs_get_dqinfo,
+	.set_info	= vfs_set_dqinfo,
+	.get_dqblk	= vfs_get_dqblk,
+	.set_dqblk	= vfs_set_dqblk,
+};
+
 static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 {
 	struct dentry *root;
@@ -651,12 +834,32 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	}
 	brelse(bh);
 	bh = NULL;
+
+	if (!(osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_XATTR))
+		parsed_options.mount_opt &= ~OCFS2_MOUNT_POSIX_ACL;
+
 	osb->s_mount_opt = parsed_options.mount_opt;
 	osb->s_atime_quantum = parsed_options.atime_quantum;
 	osb->preferred_slot = parsed_options.slot;
 	osb->osb_commit_interval = parsed_options.commit_interval;
 	osb->local_alloc_default_bits = ocfs2_megabytes_to_clusters(sb, parsed_options.localalloc_opt);
 	osb->local_alloc_bits = osb->local_alloc_default_bits;
+	if (osb->s_mount_opt & OCFS2_MOUNT_USRQUOTA &&
+	    !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+					 OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
+		status = -EINVAL;
+		mlog(ML_ERROR, "User quotas were requested, but this "
+		     "filesystem does not have the feature enabled.\n");
+		goto read_super_error;
+	}
+	if (osb->s_mount_opt & OCFS2_MOUNT_GRPQUOTA &&
+	    !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+					 OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
+		status = -EINVAL;
+		mlog(ML_ERROR, "Group quotas were requested, but this "
+		     "filesystem does not have the feature enabled.\n");
+		goto read_super_error;
+	}
 
 	status = ocfs2_verify_userspace_stack(osb, &parsed_options);
 	if (status)
@@ -664,6 +867,9 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 
 	sb->s_magic = OCFS2_SUPER_MAGIC;
 
+	sb->s_flags = (sb->s_flags & ~MS_POSIXACL) |
+		((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0);
+
 	/* Hard readonly mode only if: bdev_read_only, MS_RDONLY,
 	 * heartbeat=none */
 	if (bdev_read_only(sb->s_bdev)) {
@@ -758,6 +964,28 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	atomic_set(&osb->vol_state, VOLUME_MOUNTED);
 	wake_up(&osb->osb_mount_event);
 
+	/* Now we can initialize quotas because we can afford to wait
+	 * for cluster locks recovery now. That also means that truncation
+	 * log recovery can happen but that waits for proper quota setup */
+	if (!(sb->s_flags & MS_RDONLY)) {
+		status = ocfs2_enable_quotas(osb);
+		if (status < 0) {
+			/* We have to err-out specially here because
+			 * s_root is already set */
+			mlog_errno(status);
+			atomic_set(&osb->vol_state, VOLUME_DISABLED);
+			wake_up(&osb->osb_mount_event);
+			mlog_exit(status);
+			return status;
+		}
+	}
+
+	ocfs2_complete_quota_recovery(osb);
+
+	/* Now we wake up again for processes waiting for quotas */
+	atomic_set(&osb->vol_state, VOLUME_MOUNTED_QUOTAS);
+	wake_up(&osb->osb_mount_event);
+
 	mlog_exit(status);
 	return status;
 
@@ -945,6 +1173,41 @@ static int ocfs2_parse_options(struct super_block *sb,
 		case Opt_inode64:
 			mopt->mount_opt |= OCFS2_MOUNT_INODE64;
 			break;
+		case Opt_usrquota:
+			/* We check only on remount, otherwise features
+			 * aren't yet initialized. */
+			if (is_remount && !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+			    OCFS2_FEATURE_RO_COMPAT_USRQUOTA)) {
+				mlog(ML_ERROR, "User quota requested but "
+				     "filesystem feature is not set\n");
+				status = 0;
+				goto bail;
+			}
+			mopt->mount_opt |= OCFS2_MOUNT_USRQUOTA;
+			break;
+		case Opt_grpquota:
+			if (is_remount && !OCFS2_HAS_RO_COMPAT_FEATURE(sb,
+			    OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)) {
+				mlog(ML_ERROR, "Group quota requested but "
+				     "filesystem feature is not set\n");
+				status = 0;
+				goto bail;
+			}
+			mopt->mount_opt |= OCFS2_MOUNT_GRPQUOTA;
+			break;
+#ifdef CONFIG_OCFS2_FS_POSIX_ACL
+		case Opt_acl:
+			mopt->mount_opt |= OCFS2_MOUNT_POSIX_ACL;
+			break;
+		case Opt_noacl:
+			mopt->mount_opt &= ~OCFS2_MOUNT_POSIX_ACL;
+			break;
+#else
+		case Opt_acl:
+		case Opt_noacl:
+			printk(KERN_INFO "ocfs2 (no)acl options not supported\n");
+			break;
+#endif
 		default:
 			mlog(ML_ERROR,
 			     "Unrecognized mount option \"%s\" "
@@ -1008,6 +1271,10 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 	if (osb->osb_cluster_stack[0])
 		seq_printf(s, ",cluster_stack=%.*s", OCFS2_STACK_LABEL_LEN,
 			   osb->osb_cluster_stack);
+	if (opts & OCFS2_MOUNT_USRQUOTA)
+		seq_printf(s, ",usrquota");
+	if (opts & OCFS2_MOUNT_GRPQUOTA)
+		seq_printf(s, ",grpquota");
 
 	if (opts & OCFS2_MOUNT_NOUSERXATTR)
 		seq_printf(s, ",nouser_xattr");
@@ -1017,6 +1284,13 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 	if (opts & OCFS2_MOUNT_INODE64)
 		seq_printf(s, ",inode64");
 
+#ifdef CONFIG_OCFS2_FS_POSIX_ACL
+	if (opts & OCFS2_MOUNT_POSIX_ACL)
+		seq_printf(s, ",acl");
+	else
+		seq_printf(s, ",noacl");
+#endif
+
 	return 0;
 }
 
@@ -1052,10 +1326,16 @@ static int __init ocfs2_init(void)
 		mlog(ML_ERROR, "Unable to create ocfs2 debugfs root.\n");
 	}
 
+	status = ocfs2_quota_setup();
+	if (status)
+		goto leave;
+
 	ocfs2_set_locking_protocol();
 
+	status = register_quota_format(&ocfs2_quota_format);
 leave:
 	if (status < 0) {
+		ocfs2_quota_shutdown();
 		ocfs2_free_mem_caches();
 		exit_ocfs2_uptodate_cache();
 	}
@@ -1072,11 +1352,15 @@ static void __exit ocfs2_exit(void)
 {
 	mlog_entry_void();
 
+	ocfs2_quota_shutdown();
+
 	if (ocfs2_wq) {
 		flush_workqueue(ocfs2_wq);
 		destroy_workqueue(ocfs2_wq);
 	}
 
+	unregister_quota_format(&ocfs2_quota_format);
+
 	debugfs_remove(ocfs2_debugfs_root);
 
 	ocfs2_free_mem_caches();
@@ -1192,8 +1476,27 @@ static int ocfs2_initialize_mem_caches(void)
 				       (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
 						SLAB_MEM_SPREAD),
 				       ocfs2_inode_init_once);
-	if (!ocfs2_inode_cachep)
+	ocfs2_dquot_cachep = kmem_cache_create("ocfs2_dquot_cache",
+					sizeof(struct ocfs2_dquot),
+					0,
+					(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
+					NULL);
+	ocfs2_qf_chunk_cachep = kmem_cache_create("ocfs2_qf_chunk_cache",
+					sizeof(struct ocfs2_quota_chunk),
+					0,
+					(SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD),
+					NULL);
+	if (!ocfs2_inode_cachep || !ocfs2_dquot_cachep ||
+	    !ocfs2_qf_chunk_cachep) {
+		if (ocfs2_inode_cachep)
+			kmem_cache_destroy(ocfs2_inode_cachep);
+		if (ocfs2_dquot_cachep)
+			kmem_cache_destroy(ocfs2_dquot_cachep);
+		if (ocfs2_qf_chunk_cachep)
+			kmem_cache_destroy(ocfs2_qf_chunk_cachep);
 		return -ENOMEM;
+	}
 
 	return 0;
 }
@@ -1202,8 +1505,15 @@ static void ocfs2_free_mem_caches(void)
 {
 	if (ocfs2_inode_cachep)
 		kmem_cache_destroy(ocfs2_inode_cachep);
-
 	ocfs2_inode_cachep = NULL;
+
+	if (ocfs2_dquot_cachep)
+		kmem_cache_destroy(ocfs2_dquot_cachep);
+	ocfs2_dquot_cachep = NULL;
+
+	if (ocfs2_qf_chunk_cachep)
+		kmem_cache_destroy(ocfs2_qf_chunk_cachep);
+	ocfs2_qf_chunk_cachep = NULL;
 }
 
 static int ocfs2_get_sector(struct super_block *sb,
@@ -1303,6 +1613,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 	osb = OCFS2_SB(sb);
 	BUG_ON(!osb);
 
+	ocfs2_disable_quotas(osb);
+
 	ocfs2_shutdown_local_alloc(osb);
 
 	ocfs2_truncate_log_shutdown(osb);
@@ -1413,6 +1725,8 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	sb->s_fs_info = osb;
 	sb->s_op = &ocfs2_sops;
 	sb->s_export_op = &ocfs2_export_ops;
+	sb->s_qcop = &ocfs2_quotactl_ops;
+	sb->dq_op = &ocfs2_quota_operations;
 	sb->s_xattr = ocfs2_xattr_handlers;
 	sb->s_time_gran = 1;
 	sb->s_flags |= MS_NOATIME;
@@ -1676,6 +1990,15 @@ static int ocfs2_verify_volume(struct ocfs2_dinode *di,
 
 	if (memcmp(di->i_signature, OCFS2_SUPER_BLOCK_SIGNATURE,
 		   strlen(OCFS2_SUPER_BLOCK_SIGNATURE)) == 0) {
+		/* We have to do a raw check of the feature here */
+		if (le32_to_cpu(di->id2.i_super.s_feature_incompat) &
+		    OCFS2_FEATURE_INCOMPAT_META_ECC) {
+			status = ocfs2_block_check_validate(bh->b_data,
+							    bh->b_size,
+							    &di->i_check);
+			if (status)
+				goto out;
+		}
 		status = -EINVAL;
 		if ((1 << le32_to_cpu(di->id2.i_super.s_blocksize_bits)) != blksz) {
 			mlog(ML_ERROR, "found superblock with incorrect block "
@@ -1717,6 +2040,7 @@ static int ocfs2_verify_volume(struct ocfs2_dinode *di,
 		}
 	}
 
+out:
 	mlog_exit(status);
 	return status;
 }
diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c
index cbd03dfdc7b..ed0a0cfd68d 100644
--- a/fs/ocfs2/symlink.c
+++ b/fs/ocfs2/symlink.c
@@ -84,7 +84,7 @@ static char *ocfs2_fast_symlink_getlink(struct inode *inode,
 
 	mlog_entry_void();
 
-	status = ocfs2_read_block(inode, OCFS2_I(inode)->ip_blkno, bh);
+	status = ocfs2_read_inode_block(inode, bh);
 	if (status < 0) {
 		mlog_errno(status);
 		link = ERR_PTR(status);
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 74d7367ade1..e1d638af6ac 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -35,12 +35,14 @@
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/string.h>
+#include <linux/security.h>
 
 #define MLOG_MASK_PREFIX ML_XATTR
 #include <cluster/masklog.h>
 
 #include "ocfs2.h"
 #include "alloc.h"
+#include "blockcheck.h"
 #include "dlmglue.h"
 #include "file.h"
 #include "symlink.h"
@@ -61,12 +63,32 @@ struct ocfs2_xattr_def_value_root {
 };
 
 struct ocfs2_xattr_bucket {
-	struct buffer_head *bhs[OCFS2_XATTR_MAX_BLOCKS_PER_BUCKET];
-	struct ocfs2_xattr_header *xh;
+	/* The inode these xattrs are associated with */
+	struct inode *bu_inode;
+
+	/* The actual buffers that make up the bucket */
+	struct buffer_head *bu_bhs[OCFS2_XATTR_MAX_BLOCKS_PER_BUCKET];
+
+	/* How many blocks make up one bucket for this filesystem */
+	int bu_blocks;
+};
+
+struct ocfs2_xattr_set_ctxt {
+	handle_t *handle;
+	struct ocfs2_alloc_context *meta_ac;
+	struct ocfs2_alloc_context *data_ac;
+	struct ocfs2_cached_dealloc_ctxt dealloc;
 };
 
 #define OCFS2_XATTR_ROOT_SIZE	(sizeof(struct ocfs2_xattr_def_value_root))
 #define OCFS2_XATTR_INLINE_SIZE	80
+#define OCFS2_XATTR_FREE_IN_IBODY	(OCFS2_MIN_XATTR_INLINE_SIZE \
+					 - sizeof(struct ocfs2_xattr_header) \
+					 - sizeof(__u32))
+#define OCFS2_XATTR_FREE_IN_BLOCK(ptr)	((ptr)->i_sb->s_blocksize \
+					 - sizeof(struct ocfs2_xattr_block) \
+					 - sizeof(struct ocfs2_xattr_header) \
+					 - sizeof(__u32))
 
 static struct ocfs2_xattr_def_value_root def_xv = {
 	.xv.xr_list.l_count = cpu_to_le16(1),
@@ -74,13 +96,25 @@ static struct ocfs2_xattr_def_value_root def_xv = {
 
 struct xattr_handler *ocfs2_xattr_handlers[] = {
 	&ocfs2_xattr_user_handler,
+#ifdef CONFIG_OCFS2_FS_POSIX_ACL
+	&ocfs2_xattr_acl_access_handler,
+	&ocfs2_xattr_acl_default_handler,
+#endif
 	&ocfs2_xattr_trusted_handler,
+	&ocfs2_xattr_security_handler,
 	NULL
 };
 
 static struct xattr_handler *ocfs2_xattr_handler_map[OCFS2_XATTR_MAX] = {
 	[OCFS2_XATTR_INDEX_USER]	= &ocfs2_xattr_user_handler,
+#ifdef CONFIG_OCFS2_FS_POSIX_ACL
+	[OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS]
+					= &ocfs2_xattr_acl_access_handler,
+	[OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT]
+					= &ocfs2_xattr_acl_default_handler,
+#endif
 	[OCFS2_XATTR_INDEX_TRUSTED]	= &ocfs2_xattr_trusted_handler,
+	[OCFS2_XATTR_INDEX_SECURITY]	= &ocfs2_xattr_security_handler,
 };
 
 struct ocfs2_xattr_info {
@@ -98,7 +132,7 @@ struct ocfs2_xattr_search {
 	 */
 	struct buffer_head *xattr_bh;
 	struct ocfs2_xattr_header *header;
-	struct ocfs2_xattr_bucket bucket;
+	struct ocfs2_xattr_bucket *bucket;
 	void *base;
 	void *end;
 	struct ocfs2_xattr_entry *here;
@@ -127,14 +161,20 @@ static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
 					size_t buffer_size);
 
 static int ocfs2_xattr_create_index_block(struct inode *inode,
-					  struct ocfs2_xattr_search *xs);
+					  struct ocfs2_xattr_search *xs,
+					  struct ocfs2_xattr_set_ctxt *ctxt);
 
 static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
 					     struct ocfs2_xattr_info *xi,
-					     struct ocfs2_xattr_search *xs);
+					     struct ocfs2_xattr_search *xs,
+					     struct ocfs2_xattr_set_ctxt *ctxt);
 
 static int ocfs2_delete_xattr_index_block(struct inode *inode,
 					  struct buffer_head *xb_bh);
+static int ocfs2_mv_xattr_buckets(struct inode *inode, handle_t *handle,
+				  u64 src_blk, u64 last_blk, u64 to_blk,
+				  unsigned int start_bucket,
+				  u32 *first_hash);
 
 static inline u16 ocfs2_xattr_buckets_per_cluster(struct ocfs2_super *osb)
 {
@@ -154,6 +194,216 @@ static inline u16 ocfs2_xattr_max_xe_in_bucket(struct super_block *sb)
 	return len / sizeof(struct ocfs2_xattr_entry);
 }
 
+#define bucket_blkno(_b) ((_b)->bu_bhs[0]->b_blocknr)
+#define bucket_block(_b, _n) ((_b)->bu_bhs[(_n)]->b_data)
+#define bucket_xh(_b) ((struct ocfs2_xattr_header *)bucket_block((_b), 0))
+
+static struct ocfs2_xattr_bucket *ocfs2_xattr_bucket_new(struct inode *inode)
+{
+	struct ocfs2_xattr_bucket *bucket;
+	int blks = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+
+	BUG_ON(blks > OCFS2_XATTR_MAX_BLOCKS_PER_BUCKET);
+
+	bucket = kzalloc(sizeof(struct ocfs2_xattr_bucket), GFP_NOFS);
+	if (bucket) {
+		bucket->bu_inode = inode;
+		bucket->bu_blocks = blks;
+	}
+
+	return bucket;
+}
+
+static void ocfs2_xattr_bucket_relse(struct ocfs2_xattr_bucket *bucket)
+{
+	int i;
+
+	for (i = 0; i < bucket->bu_blocks; i++) {
+		brelse(bucket->bu_bhs[i]);
+		bucket->bu_bhs[i] = NULL;
+	}
+}
+
+static void ocfs2_xattr_bucket_free(struct ocfs2_xattr_bucket *bucket)
+{
+	if (bucket) {
+		ocfs2_xattr_bucket_relse(bucket);
+		bucket->bu_inode = NULL;
+		kfree(bucket);
+	}
+}
+
+/*
+ * A bucket that has never been written to disk doesn't need to be
+ * read.  We just need the buffer_heads.  Don't call this for
+ * buckets that are already on disk.  ocfs2_read_xattr_bucket() initializes
+ * them fully.
+ */
+static int ocfs2_init_xattr_bucket(struct ocfs2_xattr_bucket *bucket,
+				   u64 xb_blkno)
+{
+	int i, rc = 0;
+
+	for (i = 0; i < bucket->bu_blocks; i++) {
+		bucket->bu_bhs[i] = sb_getblk(bucket->bu_inode->i_sb,
+					      xb_blkno + i);
+		if (!bucket->bu_bhs[i]) {
+			rc = -EIO;
+			mlog_errno(rc);
+			break;
+		}
+
+		if (!ocfs2_buffer_uptodate(bucket->bu_inode,
+					   bucket->bu_bhs[i]))
+			ocfs2_set_new_buffer_uptodate(bucket->bu_inode,
+						      bucket->bu_bhs[i]);
+	}
+
+	if (rc)
+		ocfs2_xattr_bucket_relse(bucket);
+	return rc;
+}
+
+/* Read the xattr bucket at xb_blkno */
+static int ocfs2_read_xattr_bucket(struct ocfs2_xattr_bucket *bucket,
+				   u64 xb_blkno)
+{
+	int rc;
+
+	rc = ocfs2_read_blocks(bucket->bu_inode, xb_blkno,
+			       bucket->bu_blocks, bucket->bu_bhs, 0,
+			       NULL);
+	if (!rc) {
+		rc = ocfs2_validate_meta_ecc_bhs(bucket->bu_inode->i_sb,
+						 bucket->bu_bhs,
+						 bucket->bu_blocks,
+						 &bucket_xh(bucket)->xh_check);
+		if (rc)
+			mlog_errno(rc);
+	}
+
+	if (rc)
+		ocfs2_xattr_bucket_relse(bucket);
+	return rc;
+}
+
+static int ocfs2_xattr_bucket_journal_access(handle_t *handle,
+					     struct ocfs2_xattr_bucket *bucket,
+					     int type)
+{
+	int i, rc = 0;
+
+	for (i = 0; i < bucket->bu_blocks; i++) {
+		rc = ocfs2_journal_access(handle, bucket->bu_inode,
+					  bucket->bu_bhs[i], type);
+		if (rc) {
+			mlog_errno(rc);
+			break;
+		}
+	}
+
+	return rc;
+}
+
+static void ocfs2_xattr_bucket_journal_dirty(handle_t *handle,
+					     struct ocfs2_xattr_bucket *bucket)
+{
+	int i;
+
+	ocfs2_compute_meta_ecc_bhs(bucket->bu_inode->i_sb,
+				   bucket->bu_bhs, bucket->bu_blocks,
+				   &bucket_xh(bucket)->xh_check);
+
+	for (i = 0; i < bucket->bu_blocks; i++)
+		ocfs2_journal_dirty(handle, bucket->bu_bhs[i]);
+}
+
+static void ocfs2_xattr_bucket_copy_data(struct ocfs2_xattr_bucket *dest,
+					 struct ocfs2_xattr_bucket *src)
+{
+	int i;
+	int blocksize = src->bu_inode->i_sb->s_blocksize;
+
+	BUG_ON(dest->bu_blocks != src->bu_blocks);
+	BUG_ON(dest->bu_inode != src->bu_inode);
+
+	for (i = 0; i < src->bu_blocks; i++) {
+		memcpy(bucket_block(dest, i), bucket_block(src, i),
+		       blocksize);
+	}
+}
+
+static int ocfs2_validate_xattr_block(struct super_block *sb,
+				      struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_xattr_block *xb =
+		(struct ocfs2_xattr_block *)bh->b_data;
+
+	mlog(0, "Validating xattr block %llu\n",
+	     (unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	/*
+	 * If the ecc fails, we return the error but otherwise
+	 * leave the filesystem running.  We know any error is
+	 * local to this block.
+	 */
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &xb->xb_check);
+	if (rc)
+		return rc;
+
+	/*
+	 * Errors after here are fatal
+	 */
+
+	if (!OCFS2_IS_VALID_XATTR_BLOCK(xb)) {
+		ocfs2_error(sb,
+			    "Extended attribute block #%llu has bad "
+			    "signature %.*s",
+			    (unsigned long long)bh->b_blocknr, 7,
+			    xb->xb_signature);
+		return -EINVAL;
+	}
+
+	if (le64_to_cpu(xb->xb_blkno) != bh->b_blocknr) {
+		ocfs2_error(sb,
+			    "Extended attribute block #%llu has an "
+			    "invalid xb_blkno of %llu",
+			    (unsigned long long)bh->b_blocknr,
+			    (unsigned long long)le64_to_cpu(xb->xb_blkno));
+		return -EINVAL;
+	}
+
+	if (le32_to_cpu(xb->xb_fs_generation) != OCFS2_SB(sb)->fs_generation) {
+		ocfs2_error(sb,
+			    "Extended attribute block #%llu has an invalid "
+			    "xb_fs_generation of #%u",
+			    (unsigned long long)bh->b_blocknr,
+			    le32_to_cpu(xb->xb_fs_generation));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int ocfs2_read_xattr_block(struct inode *inode, u64 xb_blkno,
+				  struct buffer_head **bh)
+{
+	int rc;
+	struct buffer_head *tmp = *bh;
+
+	rc = ocfs2_read_block(inode, xb_blkno, &tmp,
+			      ocfs2_validate_xattr_block);
+
+	/* If ocfs2_read_block() got us a new bh, pass it up. */
+	if (!rc && !*bh)
+		*bh = tmp;
+
+	return rc;
+}
+
 static inline const char *ocfs2_xattr_prefix(int name_index)
 {
 	struct xattr_handler *handler = NULL;
@@ -200,54 +450,163 @@ static void ocfs2_xattr_hash_entry(struct inode *inode,
 	return;
 }
 
+static int ocfs2_xattr_entry_real_size(int name_len, size_t value_len)
+{
+	int size = 0;
+
+	if (value_len <= OCFS2_XATTR_INLINE_SIZE)
+		size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_SIZE(value_len);
+	else
+		size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
+	size += sizeof(struct ocfs2_xattr_entry);
+
+	return size;
+}
+
+int ocfs2_calc_security_init(struct inode *dir,
+			     struct ocfs2_security_xattr_info *si,
+			     int *want_clusters,
+			     int *xattr_credits,
+			     struct ocfs2_alloc_context **xattr_ac)
+{
+	int ret = 0;
+	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
+	int s_size = ocfs2_xattr_entry_real_size(strlen(si->name),
+						 si->value_len);
+
+	/*
+	 * The max space of security xattr taken inline is
+	 * 256(name) + 80(value) + 16(entry) = 352 bytes,
+	 * So reserve one metadata block for it is ok.
+	 */
+	if (dir->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE ||
+	    s_size > OCFS2_XATTR_FREE_IN_IBODY) {
+		ret = ocfs2_reserve_new_metadata_blocks(osb, 1, xattr_ac);
+		if (ret) {
+			mlog_errno(ret);
+			return ret;
+		}
+		*xattr_credits += OCFS2_XATTR_BLOCK_CREATE_CREDITS;
+	}
+
+	/* reserve clusters for xattr value which will be set in B tree*/
+	if (si->value_len > OCFS2_XATTR_INLINE_SIZE) {
+		int new_clusters = ocfs2_clusters_for_bytes(dir->i_sb,
+							    si->value_len);
+
+		*xattr_credits += ocfs2_clusters_to_blocks(dir->i_sb,
+							   new_clusters);
+		*want_clusters += new_clusters;
+	}
+	return ret;
+}
+
+int ocfs2_calc_xattr_init(struct inode *dir,
+			  struct buffer_head *dir_bh,
+			  int mode,
+			  struct ocfs2_security_xattr_info *si,
+			  int *want_clusters,
+			  int *xattr_credits,
+			  struct ocfs2_alloc_context **xattr_ac)
+{
+	int ret = 0;
+	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
+	int s_size = 0, a_size = 0, acl_len = 0, new_clusters;
+
+	if (si->enable)
+		s_size = ocfs2_xattr_entry_real_size(strlen(si->name),
+						     si->value_len);
+
+	if (osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) {
+		acl_len = ocfs2_xattr_get_nolock(dir, dir_bh,
+					OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT,
+					"", NULL, 0);
+		if (acl_len > 0) {
+			a_size = ocfs2_xattr_entry_real_size(0, acl_len);
+			if (S_ISDIR(mode))
+				a_size <<= 1;
+		} else if (acl_len != 0 && acl_len != -ENODATA) {
+			mlog_errno(ret);
+			return ret;
+		}
+	}
+
+	if (!(s_size + a_size))
+		return ret;
+
+	/*
+	 * The max space of security xattr taken inline is
+	 * 256(name) + 80(value) + 16(entry) = 352 bytes,
+	 * The max space of acl xattr taken inline is
+	 * 80(value) + 16(entry) * 2(if directory) = 192 bytes,
+	 * when blocksize = 512, may reserve one more cluser for
+	 * xattr bucket, otherwise reserve one metadata block
+	 * for them is ok.
+	 */
+	if (dir->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE ||
+	    (s_size + a_size) > OCFS2_XATTR_FREE_IN_IBODY) {
+		ret = ocfs2_reserve_new_metadata_blocks(osb, 1, xattr_ac);
+		if (ret) {
+			mlog_errno(ret);
+			return ret;
+		}
+		*xattr_credits += OCFS2_XATTR_BLOCK_CREATE_CREDITS;
+	}
+
+	if (dir->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE &&
+	    (s_size + a_size) > OCFS2_XATTR_FREE_IN_BLOCK(dir)) {
+		*want_clusters += 1;
+		*xattr_credits += ocfs2_blocks_per_xattr_bucket(dir->i_sb);
+	}
+
+	/*
+	 * reserve credits and clusters for xattrs which has large value
+	 * and have to be set outside
+	 */
+	if (si->enable && si->value_len > OCFS2_XATTR_INLINE_SIZE) {
+		new_clusters = ocfs2_clusters_for_bytes(dir->i_sb,
+							si->value_len);
+		*xattr_credits += ocfs2_clusters_to_blocks(dir->i_sb,
+							   new_clusters);
+		*want_clusters += new_clusters;
+	}
+	if (osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL &&
+	    acl_len > OCFS2_XATTR_INLINE_SIZE) {
+		/* for directory, it has DEFAULT and ACCESS two types of acls */
+		new_clusters = (S_ISDIR(mode) ? 2 : 1) *
+				ocfs2_clusters_for_bytes(dir->i_sb, acl_len);
+		*xattr_credits += ocfs2_clusters_to_blocks(dir->i_sb,
+							   new_clusters);
+		*want_clusters += new_clusters;
+	}
+
+	return ret;
+}
+
 static int ocfs2_xattr_extend_allocation(struct inode *inode,
 					 u32 clusters_to_add,
-					 struct buffer_head *xattr_bh,
-					 struct ocfs2_xattr_value_root *xv)
+					 struct ocfs2_xattr_value_buf *vb,
+					 struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int status = 0;
-	int restart_func = 0;
-	int credits = 0;
-	handle_t *handle = NULL;
-	struct ocfs2_alloc_context *data_ac = NULL;
-	struct ocfs2_alloc_context *meta_ac = NULL;
+	handle_t *handle = ctxt->handle;
 	enum ocfs2_alloc_restarted why;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	u32 prev_clusters, logical_start = le32_to_cpu(xv->xr_clusters);
+	u32 prev_clusters, logical_start = le32_to_cpu(vb->vb_xv->xr_clusters);
 	struct ocfs2_extent_tree et;
 
 	mlog(0, "(clusters_to_add for xattr= %u)\n", clusters_to_add);
 
-	ocfs2_init_xattr_value_extent_tree(&et, inode, xattr_bh, xv);
-
-restart_all:
-
-	status = ocfs2_lock_allocators(inode, &et, clusters_to_add, 0,
-				       &data_ac, &meta_ac);
-	if (status) {
-		mlog_errno(status);
-		goto leave;
-	}
-
-	credits = ocfs2_calc_extend_credits(osb->sb, et.et_root_el,
-					    clusters_to_add);
-	handle = ocfs2_start_trans(osb, credits);
-	if (IS_ERR(handle)) {
-		status = PTR_ERR(handle);
-		handle = NULL;
-		mlog_errno(status);
-		goto leave;
-	}
+	ocfs2_init_xattr_value_extent_tree(&et, inode, vb);
 
-restarted_transaction:
-	status = ocfs2_journal_access(handle, inode, xattr_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	status = vb->vb_access(handle, inode, vb->vb_bh,
+			      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
 	}
 
-	prev_clusters = le32_to_cpu(xv->xr_clusters);
+	prev_clusters = le32_to_cpu(vb->vb_xv->xr_clusters);
 	status = ocfs2_add_clusters_in_btree(osb,
 					     inode,
 					     &logical_start,
@@ -255,157 +614,84 @@ restarted_transaction:
 					     0,
 					     &et,
 					     handle,
-					     data_ac,
-					     meta_ac,
+					     ctxt->data_ac,
+					     ctxt->meta_ac,
 					     &why);
-	if ((status < 0) && (status != -EAGAIN)) {
-		if (status != -ENOSPC)
-			mlog_errno(status);
+	if (status < 0) {
+		mlog_errno(status);
 		goto leave;
 	}
 
-	status = ocfs2_journal_dirty(handle, xattr_bh);
+	status = ocfs2_journal_dirty(handle, vb->vb_bh);
 	if (status < 0) {
 		mlog_errno(status);
 		goto leave;
 	}
 
-	clusters_to_add -= le32_to_cpu(xv->xr_clusters) - prev_clusters;
+	clusters_to_add -= le32_to_cpu(vb->vb_xv->xr_clusters) - prev_clusters;
 
-	if (why != RESTART_NONE && clusters_to_add) {
-		if (why == RESTART_META) {
-			mlog(0, "restarting function.\n");
-			restart_func = 1;
-		} else {
-			BUG_ON(why != RESTART_TRANS);
-
-			mlog(0, "restarting transaction.\n");
-			/* TODO: This can be more intelligent. */
-			credits = ocfs2_calc_extend_credits(osb->sb,
-							    et.et_root_el,
-							    clusters_to_add);
-			status = ocfs2_extend_trans(handle, credits);
-			if (status < 0) {
-				/* handle still has to be committed at
-				 * this point. */
-				status = -ENOMEM;
-				mlog_errno(status);
-				goto leave;
-			}
-			goto restarted_transaction;
-		}
-	}
+	/*
+	 * We should have already allocated enough space before the transaction,
+	 * so no need to restart.
+	 */
+	BUG_ON(why != RESTART_NONE || clusters_to_add);
 
 leave:
-	if (handle) {
-		ocfs2_commit_trans(osb, handle);
-		handle = NULL;
-	}
-	if (data_ac) {
-		ocfs2_free_alloc_context(data_ac);
-		data_ac = NULL;
-	}
-	if (meta_ac) {
-		ocfs2_free_alloc_context(meta_ac);
-		meta_ac = NULL;
-	}
-	if ((!status) && restart_func) {
-		restart_func = 0;
-		goto restart_all;
-	}
 
 	return status;
 }
 
 static int __ocfs2_remove_xattr_range(struct inode *inode,
-				      struct buffer_head *root_bh,
-				      struct ocfs2_xattr_value_root *xv,
+				      struct ocfs2_xattr_value_buf *vb,
 				      u32 cpos, u32 phys_cpos, u32 len,
-				      struct ocfs2_cached_dealloc_ctxt *dealloc)
+				      struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret;
 	u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	struct inode *tl_inode = osb->osb_tl_inode;
-	handle_t *handle;
-	struct ocfs2_alloc_context *meta_ac = NULL;
+	handle_t *handle = ctxt->handle;
 	struct ocfs2_extent_tree et;
 
-	ocfs2_init_xattr_value_extent_tree(&et, inode, root_bh, xv);
+	ocfs2_init_xattr_value_extent_tree(&et, inode, vb);
 
-	ret = ocfs2_lock_allocators(inode, &et, 0, 1, NULL, &meta_ac);
+	ret = vb->vb_access(handle, inode, vb->vb_bh,
+			    OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		return ret;
-	}
-
-	mutex_lock(&tl_inode->i_mutex);
-
-	if (ocfs2_truncate_log_needs_flush(osb)) {
-		ret = __ocfs2_flush_truncate_log(osb);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto out;
-		}
-	}
-
-	handle = ocfs2_start_trans(osb, OCFS2_REMOVE_EXTENT_CREDITS);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, root_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_commit;
-	}
-
-	ret = ocfs2_remove_extent(inode, &et, cpos, len, handle, meta_ac,
-				  dealloc);
+	ret = ocfs2_remove_extent(inode, &et, cpos, len, handle, ctxt->meta_ac,
+				  &ctxt->dealloc);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
-	le32_add_cpu(&xv->xr_clusters, -len);
+	le32_add_cpu(&vb->vb_xv->xr_clusters, -len);
 
-	ret = ocfs2_journal_dirty(handle, root_bh);
+	ret = ocfs2_journal_dirty(handle, vb->vb_bh);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
-	ret = ocfs2_truncate_log_append(osb, handle, phys_blkno, len);
+	ret = ocfs2_cache_cluster_dealloc(&ctxt->dealloc, phys_blkno, len);
 	if (ret)
 		mlog_errno(ret);
 
-out_commit:
-	ocfs2_commit_trans(osb, handle);
 out:
-	mutex_unlock(&tl_inode->i_mutex);
-
-	if (meta_ac)
-		ocfs2_free_alloc_context(meta_ac);
-
 	return ret;
 }
 
 static int ocfs2_xattr_shrink_size(struct inode *inode,
 				   u32 old_clusters,
 				   u32 new_clusters,
-				   struct buffer_head *root_bh,
-				   struct ocfs2_xattr_value_root *xv)
+				   struct ocfs2_xattr_value_buf *vb,
+				   struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret = 0;
 	u32 trunc_len, cpos, phys_cpos, alloc_size;
 	u64 block;
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	struct ocfs2_cached_dealloc_ctxt dealloc;
-
-	ocfs2_init_dealloc_ctxt(&dealloc);
 
 	if (old_clusters <= new_clusters)
 		return 0;
@@ -414,7 +700,8 @@ static int ocfs2_xattr_shrink_size(struct inode *inode,
 	trunc_len = old_clusters - new_clusters;
 	while (trunc_len) {
 		ret = ocfs2_xattr_get_clusters(inode, cpos, &phys_cpos,
-					       &alloc_size, &xv->xr_list);
+					       &alloc_size,
+					       &vb->vb_xv->xr_list);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -423,9 +710,9 @@ static int ocfs2_xattr_shrink_size(struct inode *inode,
 		if (alloc_size > trunc_len)
 			alloc_size = trunc_len;
 
-		ret = __ocfs2_remove_xattr_range(inode, root_bh, xv, cpos,
+		ret = __ocfs2_remove_xattr_range(inode, vb, cpos,
 						 phys_cpos, alloc_size,
-						 &dealloc);
+						 ctxt);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -439,20 +726,17 @@ static int ocfs2_xattr_shrink_size(struct inode *inode,
 	}
 
 out:
-	ocfs2_schedule_truncate_log_flush(osb, 1);
-	ocfs2_run_deallocs(osb, &dealloc);
-
 	return ret;
 }
 
 static int ocfs2_xattr_value_truncate(struct inode *inode,
-				      struct buffer_head *root_bh,
-				      struct ocfs2_xattr_value_root *xv,
-				      int len)
+				      struct ocfs2_xattr_value_buf *vb,
+				      int len,
+				      struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret;
 	u32 new_clusters = ocfs2_clusters_for_bytes(inode->i_sb, len);
-	u32 old_clusters = le32_to_cpu(xv->xr_clusters);
+	u32 old_clusters = le32_to_cpu(vb->vb_xv->xr_clusters);
 
 	if (new_clusters == old_clusters)
 		return 0;
@@ -460,11 +744,11 @@ static int ocfs2_xattr_value_truncate(struct inode *inode,
 	if (new_clusters > old_clusters)
 		ret = ocfs2_xattr_extend_allocation(inode,
 						    new_clusters - old_clusters,
-						    root_bh, xv);
+						    vb, ctxt);
 	else
 		ret = ocfs2_xattr_shrink_size(inode,
 					      old_clusters, new_clusters,
-					      root_bh, xv);
+					      vb, ctxt);
 
 	return ret;
 }
@@ -554,18 +838,14 @@ static int ocfs2_xattr_block_list(struct inode *inode,
 	if (!di->i_xattr_loc)
 		return ret;
 
-	ret = ocfs2_read_block(inode, le64_to_cpu(di->i_xattr_loc), &blk_bh);
+	ret = ocfs2_read_xattr_block(inode, le64_to_cpu(di->i_xattr_loc),
+				     &blk_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		return ret;
 	}
 
 	xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
-	if (!OCFS2_IS_VALID_XATTR_BLOCK(xb)) {
-		ret = -EIO;
-		goto cleanup;
-	}
-
 	if (!(le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED)) {
 		struct ocfs2_xattr_header *header = &xb->xb_attrs.xb_header;
 		ret = ocfs2_xattr_list_entries(inode, header,
@@ -575,7 +855,7 @@ static int ocfs2_xattr_block_list(struct inode *inode,
 		ret = ocfs2_xattr_tree_list_index_block(inode, xt,
 						   buffer, buffer_size);
 	}
-cleanup:
+
 	brelse(blk_bh);
 
 	return ret;
@@ -685,7 +965,7 @@ static int ocfs2_xattr_get_value_outside(struct inode *inode,
 		blkno = ocfs2_clusters_to_blocks(inode->i_sb, p_cluster);
 		/* Copy ocfs2_xattr_value */
 		for (i = 0; i < num_clusters * bpc; i++, blkno++) {
-			ret = ocfs2_read_block(inode, blkno, &bh);
+			ret = ocfs2_read_block(inode, blkno, &bh, NULL);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -769,7 +1049,12 @@ static int ocfs2_xattr_block_get(struct inode *inode,
 	size_t size;
 	int ret = -ENODATA, name_offset, name_len, block_off, i;
 
-	memset(&xs->bucket, 0, sizeof(xs->bucket));
+	xs->bucket = ocfs2_xattr_bucket_new(inode);
+	if (!xs->bucket) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto cleanup;
+	}
 
 	ret = ocfs2_xattr_block_find(inode, name_index, name, xs);
 	if (ret) {
@@ -795,11 +1080,11 @@ static int ocfs2_xattr_block_get(struct inode *inode,
 
 		if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED) {
 			ret = ocfs2_xattr_bucket_get_name_value(inode,
-								xs->bucket.xh,
+								bucket_xh(xs->bucket),
 								i,
 								&block_off,
 								&name_offset);
-			xs->base = xs->bucket.bhs[block_off]->b_data;
+			xs->base = bucket_block(xs->bucket, block_off);
 		}
 		if (ocfs2_xattr_is_local(xs->here)) {
 			memcpy(buffer, (void *)xs->base +
@@ -817,21 +1102,15 @@ static int ocfs2_xattr_block_get(struct inode *inode,
 	}
 	ret = size;
 cleanup:
-	for (i = 0; i < OCFS2_XATTR_MAX_BLOCKS_PER_BUCKET; i++)
-		brelse(xs->bucket.bhs[i]);
-	memset(&xs->bucket, 0, sizeof(xs->bucket));
+	ocfs2_xattr_bucket_free(xs->bucket);
 
 	brelse(xs->xattr_bh);
 	xs->xattr_bh = NULL;
 	return ret;
 }
 
-/* ocfs2_xattr_get()
- *
- * Copy an extended attribute into the buffer provided.
- * Buffer is NULL to compute the size of buffer required.
- */
-static int ocfs2_xattr_get(struct inode *inode,
+int ocfs2_xattr_get_nolock(struct inode *inode,
+			   struct buffer_head *di_bh,
 			   int name_index,
 			   const char *name,
 			   void *buffer,
@@ -839,7 +1118,6 @@ static int ocfs2_xattr_get(struct inode *inode,
 {
 	int ret;
 	struct ocfs2_dinode *di = NULL;
-	struct buffer_head *di_bh = NULL;
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 	struct ocfs2_xattr_search xis = {
 		.not_found = -ENODATA,
@@ -854,11 +1132,6 @@ static int ocfs2_xattr_get(struct inode *inode,
 	if (!(oi->ip_dyn_features & OCFS2_HAS_XATTR_FL))
 		ret = -ENODATA;
 
-	ret = ocfs2_inode_lock(inode, &di_bh, 0);
-	if (ret < 0) {
-		mlog_errno(ret);
-		return ret;
-	}
 	xis.inode_bh = xbs.inode_bh = di_bh;
 	di = (struct ocfs2_dinode *)di_bh->b_data;
 
@@ -869,6 +1142,32 @@ static int ocfs2_xattr_get(struct inode *inode,
 		ret = ocfs2_xattr_block_get(inode, name_index, name, buffer,
 					    buffer_size, &xbs);
 	up_read(&oi->ip_xattr_sem);
+
+	return ret;
+}
+
+/* ocfs2_xattr_get()
+ *
+ * Copy an extended attribute into the buffer provided.
+ * Buffer is NULL to compute the size of buffer required.
+ */
+static int ocfs2_xattr_get(struct inode *inode,
+			   int name_index,
+			   const char *name,
+			   void *buffer,
+			   size_t buffer_size)
+{
+	int ret;
+	struct buffer_head *di_bh = NULL;
+
+	ret = ocfs2_inode_lock(inode, &di_bh, 0);
+	if (ret < 0) {
+		mlog_errno(ret);
+		return ret;
+	}
+	ret = ocfs2_xattr_get_nolock(inode, di_bh, name_index,
+				     name, buffer, buffer_size);
+
 	ocfs2_inode_unlock(inode, 0);
 
 	brelse(di_bh);
@@ -877,44 +1176,36 @@ static int ocfs2_xattr_get(struct inode *inode,
 }
 
 static int __ocfs2_xattr_set_value_outside(struct inode *inode,
+					   handle_t *handle,
 					   struct ocfs2_xattr_value_root *xv,
 					   const void *value,
 					   int value_len)
 {
-	int ret = 0, i, cp_len, credits;
+	int ret = 0, i, cp_len;
 	u16 blocksize = inode->i_sb->s_blocksize;
 	u32 p_cluster, num_clusters;
 	u32 cpos = 0, bpc = ocfs2_clusters_to_blocks(inode->i_sb, 1);
 	u32 clusters = ocfs2_clusters_for_bytes(inode->i_sb, value_len);
 	u64 blkno;
 	struct buffer_head *bh = NULL;
-	handle_t *handle;
 
 	BUG_ON(clusters > le32_to_cpu(xv->xr_clusters));
 
-	credits = clusters * bpc;
-	handle = ocfs2_start_trans(OCFS2_SB(inode->i_sb), credits);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		goto out;
-	}
-
 	while (cpos < clusters) {
 		ret = ocfs2_xattr_get_clusters(inode, cpos, &p_cluster,
 					       &num_clusters, &xv->xr_list);
 		if (ret) {
 			mlog_errno(ret);
-			goto out_commit;
+			goto out;
 		}
 
 		blkno = ocfs2_clusters_to_blocks(inode->i_sb, p_cluster);
 
 		for (i = 0; i < num_clusters * bpc; i++, blkno++) {
-			ret = ocfs2_read_block(inode, blkno, &bh);
+			ret = ocfs2_read_block(inode, blkno, &bh, NULL);
 			if (ret) {
 				mlog_errno(ret);
-				goto out_commit;
+				goto out;
 			}
 
 			ret = ocfs2_journal_access(handle,
@@ -923,7 +1214,7 @@ static int __ocfs2_xattr_set_value_outside(struct inode *inode,
 						   OCFS2_JOURNAL_ACCESS_WRITE);
 			if (ret < 0) {
 				mlog_errno(ret);
-				goto out_commit;
+				goto out;
 			}
 
 			cp_len = value_len > blocksize ? blocksize : value_len;
@@ -937,7 +1228,7 @@ static int __ocfs2_xattr_set_value_outside(struct inode *inode,
 			ret = ocfs2_journal_dirty(handle, bh);
 			if (ret < 0) {
 				mlog_errno(ret);
-				goto out_commit;
+				goto out;
 			}
 			brelse(bh);
 			bh = NULL;
@@ -951,8 +1242,6 @@ static int __ocfs2_xattr_set_value_outside(struct inode *inode,
 		}
 		cpos += num_clusters;
 	}
-out_commit:
-	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
 out:
 	brelse(bh);
 
@@ -960,28 +1249,22 @@ out:
 }
 
 static int ocfs2_xattr_cleanup(struct inode *inode,
+			       handle_t *handle,
 			       struct ocfs2_xattr_info *xi,
 			       struct ocfs2_xattr_search *xs,
+			       struct ocfs2_xattr_value_buf *vb,
 			       size_t offs)
 {
-	handle_t *handle = NULL;
 	int ret = 0;
 	size_t name_len = strlen(xi->name);
 	void *val = xs->base + offs;
 	size_t size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
 
-	handle = ocfs2_start_trans((OCFS2_SB(inode->i_sb)),
-				   OCFS2_XATTR_BLOCK_UPDATE_CREDITS);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		goto out;
-	}
-	ret = ocfs2_journal_access(handle, inode, xs->xattr_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = vb->vb_access(handle, inode, vb->vb_bh,
+			    OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 	/* Decrease xattr count */
 	le16_add_cpu(&xs->header->xh_count, -1);
@@ -989,35 +1272,27 @@ static int ocfs2_xattr_cleanup(struct inode *inode,
 	memset((void *)xs->here, 0, sizeof(struct ocfs2_xattr_entry));
 	memset(val, 0, size);
 
-	ret = ocfs2_journal_dirty(handle, xs->xattr_bh);
+	ret = ocfs2_journal_dirty(handle, vb->vb_bh);
 	if (ret < 0)
 		mlog_errno(ret);
-out_commit:
-	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
 out:
 	return ret;
 }
 
 static int ocfs2_xattr_update_entry(struct inode *inode,
+				    handle_t *handle,
 				    struct ocfs2_xattr_info *xi,
 				    struct ocfs2_xattr_search *xs,
+				    struct ocfs2_xattr_value_buf *vb,
 				    size_t offs)
 {
-	handle_t *handle = NULL;
-	int ret = 0;
+	int ret;
 
-	handle = ocfs2_start_trans((OCFS2_SB(inode->i_sb)),
-				   OCFS2_XATTR_BLOCK_UPDATE_CREDITS);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		goto out;
-	}
-	ret = ocfs2_journal_access(handle, inode, xs->xattr_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = vb->vb_access(handle, inode, vb->vb_bh,
+			    OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
 	xs->here->xe_name_offset = cpu_to_le16(offs);
@@ -1028,11 +1303,9 @@ static int ocfs2_xattr_update_entry(struct inode *inode,
 		ocfs2_xattr_set_local(xs->here, 0);
 	ocfs2_xattr_hash_entry(inode, xs->header, xs->here);
 
-	ret = ocfs2_journal_dirty(handle, xs->xattr_bh);
+	ret = ocfs2_journal_dirty(handle, vb->vb_bh);
 	if (ret < 0)
 		mlog_errno(ret);
-out_commit:
-	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
 out:
 	return ret;
 }
@@ -1045,6 +1318,8 @@ out:
 static int ocfs2_xattr_set_value_outside(struct inode *inode,
 					 struct ocfs2_xattr_info *xi,
 					 struct ocfs2_xattr_search *xs,
+					 struct ocfs2_xattr_set_ctxt *ctxt,
+					 struct ocfs2_xattr_value_buf *vb,
 					 size_t offs)
 {
 	size_t name_len = strlen(xi->name);
@@ -1062,20 +1337,20 @@ static int ocfs2_xattr_set_value_outside(struct inode *inode,
 	xv->xr_list.l_tree_depth = 0;
 	xv->xr_list.l_count = cpu_to_le16(1);
 	xv->xr_list.l_next_free_rec = 0;
+	vb->vb_xv = xv;
 
-	ret = ocfs2_xattr_value_truncate(inode, xs->xattr_bh, xv,
-					 xi->value_len);
+	ret = ocfs2_xattr_value_truncate(inode, vb, xi->value_len, ctxt);
 	if (ret < 0) {
 		mlog_errno(ret);
 		return ret;
 	}
-	ret = __ocfs2_xattr_set_value_outside(inode, xv, xi->value,
-					      xi->value_len);
+	ret = ocfs2_xattr_update_entry(inode, ctxt->handle, xi, xs, vb, offs);
 	if (ret < 0) {
 		mlog_errno(ret);
 		return ret;
 	}
-	ret = ocfs2_xattr_update_entry(inode, xi, xs, offs);
+	ret = __ocfs2_xattr_set_value_outside(inode, ctxt->handle, vb->vb_xv,
+					      xi->value, xi->value_len);
 	if (ret < 0)
 		mlog_errno(ret);
 
@@ -1195,6 +1470,7 @@ static void ocfs2_xattr_set_entry_local(struct inode *inode,
 static int ocfs2_xattr_set_entry(struct inode *inode,
 				 struct ocfs2_xattr_info *xi,
 				 struct ocfs2_xattr_search *xs,
+				 struct ocfs2_xattr_set_ctxt *ctxt,
 				 int flag)
 {
 	struct ocfs2_xattr_entry *last;
@@ -1202,7 +1478,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)xs->inode_bh->b_data;
 	size_t min_offs = xs->end - xs->base, name_len = strlen(xi->name);
 	size_t size_l = 0;
-	handle_t *handle = NULL;
+	handle_t *handle = ctxt->handle;
 	int free, i, ret;
 	struct ocfs2_xattr_info xi_l = {
 		.name_index = xi->name_index,
@@ -1210,6 +1486,16 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 		.value = xi->value,
 		.value_len = xi->value_len,
 	};
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_bh = xs->xattr_bh,
+		.vb_access = ocfs2_journal_access_di,
+	};
+
+	if (!(flag & OCFS2_INLINE_XATTR_FL)) {
+		BUG_ON(xs->xattr_bh == xs->inode_bh);
+		vb.vb_access = ocfs2_journal_access_xb;
+	} else
+		BUG_ON(xs->xattr_bh != xs->inode_bh);
 
 	/* Compute min_offs, last and free space. */
 	last = xs->header->xh_entries;
@@ -1265,15 +1551,14 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 		if (ocfs2_xattr_is_local(xs->here) && size == size_l) {
 			/* Replace existing local xattr with tree root */
 			ret = ocfs2_xattr_set_value_outside(inode, xi, xs,
-							    offs);
+							    ctxt, &vb, offs);
 			if (ret < 0)
 				mlog_errno(ret);
 			goto out;
 		} else if (!ocfs2_xattr_is_local(xs->here)) {
 			/* For existing xattr which has value outside */
-			struct ocfs2_xattr_value_root *xv = NULL;
-			xv = (struct ocfs2_xattr_value_root *)(val +
-				OCFS2_XATTR_SIZE(name_len));
+			vb.vb_xv = (struct ocfs2_xattr_value_root *)
+				(val + OCFS2_XATTR_SIZE(name_len));
 
 			if (xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
 				/*
@@ -1282,27 +1567,30 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 				 * then set new value with set_value_outside().
 				 */
 				ret = ocfs2_xattr_value_truncate(inode,
-								 xs->xattr_bh,
-								 xv,
-								 xi->value_len);
+								 &vb,
+								 xi->value_len,
+								 ctxt);
 				if (ret < 0) {
 					mlog_errno(ret);
 					goto out;
 				}
 
-				ret = __ocfs2_xattr_set_value_outside(inode,
-								xv,
-								xi->value,
-								xi->value_len);
+				ret = ocfs2_xattr_update_entry(inode,
+							       handle,
+							       xi,
+							       xs,
+							       &vb,
+							       offs);
 				if (ret < 0) {
 					mlog_errno(ret);
 					goto out;
 				}
 
-				ret = ocfs2_xattr_update_entry(inode,
-							       xi,
-							       xs,
-							       offs);
+				ret = __ocfs2_xattr_set_value_outside(inode,
+								handle,
+								vb.vb_xv,
+								xi->value,
+								xi->value_len);
 				if (ret < 0)
 					mlog_errno(ret);
 				goto out;
@@ -1312,44 +1600,28 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 				 * just trucate old value to zero.
 				 */
 				 ret = ocfs2_xattr_value_truncate(inode,
-								 xs->xattr_bh,
-								 xv,
-								 0);
+								  &vb,
+								  0,
+								  ctxt);
 				if (ret < 0)
 					mlog_errno(ret);
 			}
 		}
 	}
 
-	handle = ocfs2_start_trans((OCFS2_SB(inode->i_sb)),
-				   OCFS2_INODE_UPDATE_CREDITS);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		goto out;
-	}
-
-	ret = ocfs2_journal_access(handle, inode, xs->inode_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, xs->inode_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
 	if (!(flag & OCFS2_INLINE_XATTR_FL)) {
-		/* set extended attribute in external block. */
-		ret = ocfs2_extend_trans(handle,
-					 OCFS2_INODE_UPDATE_CREDITS +
-					 OCFS2_XATTR_BLOCK_UPDATE_CREDITS);
-		if (ret) {
-			mlog_errno(ret);
-			goto out_commit;
-		}
-		ret = ocfs2_journal_access(handle, inode, xs->xattr_bh,
-					   OCFS2_JOURNAL_ACCESS_WRITE);
+		ret = vb.vb_access(handle, inode, vb.vb_bh,
+				   OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
-			goto out_commit;
+			goto out;
 		}
 	}
 
@@ -1363,7 +1635,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 		ret = ocfs2_journal_dirty(handle, xs->xattr_bh);
 		if (ret < 0) {
 			mlog_errno(ret);
-			goto out_commit;
+			goto out;
 		}
 	}
 
@@ -1391,25 +1663,19 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 	oi->ip_dyn_features |= flag;
 	di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
 	spin_unlock(&oi->ip_lock);
-	/* Update inode ctime */
-	inode->i_ctime = CURRENT_TIME;
-	di->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec);
-	di->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec);
 
 	ret = ocfs2_journal_dirty(handle, xs->inode_bh);
 	if (ret < 0)
 		mlog_errno(ret);
 
-out_commit:
-	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
-
 	if (!ret && xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
 		/*
 		 * Set value outside in B tree.
 		 * This is the second step for value size > INLINE_SIZE.
 		 */
 		size_t offs = le16_to_cpu(xs->here->xe_name_offset);
-		ret = ocfs2_xattr_set_value_outside(inode, xi, xs, offs);
+		ret = ocfs2_xattr_set_value_outside(inode, xi, xs, ctxt,
+						    &vb, offs);
 		if (ret < 0) {
 			int ret2;
 
@@ -1418,41 +1684,56 @@ out_commit:
 			 * If set value outside failed, we have to clean
 			 * the junk tree root we have already set in local.
 			 */
-			ret2 = ocfs2_xattr_cleanup(inode, xi, xs, offs);
+			ret2 = ocfs2_xattr_cleanup(inode, ctxt->handle,
+						   xi, xs, &vb, offs);
 			if (ret2 < 0)
 				mlog_errno(ret2);
 		}
 	}
 out:
 	return ret;
-
 }
 
 static int ocfs2_remove_value_outside(struct inode*inode,
-				      struct buffer_head *bh,
+				      struct ocfs2_xattr_value_buf *vb,
 				      struct ocfs2_xattr_header *header)
 {
 	int ret = 0, i;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_xattr_set_ctxt ctxt = { NULL, NULL, };
+
+	ocfs2_init_dealloc_ctxt(&ctxt.dealloc);
+
+	ctxt.handle = ocfs2_start_trans(osb,
+					ocfs2_remove_extent_credits(osb->sb));
+	if (IS_ERR(ctxt.handle)) {
+		ret = PTR_ERR(ctxt.handle);
+		mlog_errno(ret);
+		goto out;
+	}
 
 	for (i = 0; i < le16_to_cpu(header->xh_count); i++) {
 		struct ocfs2_xattr_entry *entry = &header->xh_entries[i];
 
 		if (!ocfs2_xattr_is_local(entry)) {
-			struct ocfs2_xattr_value_root *xv;
 			void *val;
 
 			val = (void *)header +
 				le16_to_cpu(entry->xe_name_offset);
-			xv = (struct ocfs2_xattr_value_root *)
+			vb->vb_xv = (struct ocfs2_xattr_value_root *)
 				(val + OCFS2_XATTR_SIZE(entry->xe_name_len));
-			ret = ocfs2_xattr_value_truncate(inode, bh, xv, 0);
+			ret = ocfs2_xattr_value_truncate(inode, vb, 0, &ctxt);
 			if (ret < 0) {
 				mlog_errno(ret);
-				return ret;
+				break;
 			}
 		}
 	}
 
+	ocfs2_commit_trans(osb, ctxt.handle);
+	ocfs2_schedule_truncate_log_flush(osb, 1);
+	ocfs2_run_deallocs(osb, &ctxt.dealloc);
+out:
 	return ret;
 }
 
@@ -1463,12 +1744,16 @@ static int ocfs2_xattr_ibody_remove(struct inode *inode,
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
 	struct ocfs2_xattr_header *header;
 	int ret;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_bh = di_bh,
+		.vb_access = ocfs2_journal_access_di,
+	};
 
 	header = (struct ocfs2_xattr_header *)
 		 ((void *)di + inode->i_sb->s_blocksize -
 		 le16_to_cpu(di->i_xattr_inline_size));
 
-	ret = ocfs2_remove_value_outside(inode, di_bh, header);
+	ret = ocfs2_remove_value_outside(inode, &vb, header);
 
 	return ret;
 }
@@ -1478,11 +1763,15 @@ static int ocfs2_xattr_block_remove(struct inode *inode,
 {
 	struct ocfs2_xattr_block *xb;
 	int ret = 0;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_bh = blk_bh,
+		.vb_access = ocfs2_journal_access_xb,
+	};
 
 	xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
 	if (!(le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED)) {
 		struct ocfs2_xattr_header *header = &(xb->xb_attrs.xb_header);
-		ret = ocfs2_remove_value_outside(inode, blk_bh, header);
+		ret = ocfs2_remove_value_outside(inode, &vb, header);
 	} else
 		ret = ocfs2_delete_xattr_index_block(inode, blk_bh);
 
@@ -1502,24 +1791,19 @@ static int ocfs2_xattr_free_block(struct inode *inode,
 	u64 blk, bg_blkno;
 	u16 bit;
 
-	ret = ocfs2_read_block(inode, block, &blk_bh);
+	ret = ocfs2_read_xattr_block(inode, block, &blk_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
-	if (!OCFS2_IS_VALID_XATTR_BLOCK(xb)) {
-		ret = -EIO;
-		goto out;
-	}
-
 	ret = ocfs2_xattr_block_remove(inode, blk_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
 	}
 
+	xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
 	blk = le64_to_cpu(xb->xb_blkno);
 	bit = le16_to_cpu(xb->xb_suballoc_bit);
 	bg_blkno = ocfs2_which_suballoc_group(blk, bit);
@@ -1606,8 +1890,8 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh)
 		mlog_errno(ret);
 		goto out;
 	}
-	ret = ocfs2_journal_access(handle, inode, di_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -1714,7 +1998,8 @@ static int ocfs2_xattr_ibody_find(struct inode *inode,
  */
 static int ocfs2_xattr_ibody_set(struct inode *inode,
 				 struct ocfs2_xattr_info *xi,
-				 struct ocfs2_xattr_search *xs)
+				 struct ocfs2_xattr_search *xs,
+				 struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)xs->inode_bh->b_data;
@@ -1731,7 +2016,7 @@ static int ocfs2_xattr_ibody_set(struct inode *inode,
 		}
 	}
 
-	ret = ocfs2_xattr_set_entry(inode, xi, xs,
+	ret = ocfs2_xattr_set_entry(inode, xi, xs, ctxt,
 				(OCFS2_INLINE_XATTR_FL | OCFS2_HAS_XATTR_FL));
 out:
 	up_write(&oi->ip_alloc_sem);
@@ -1758,19 +2043,15 @@ static int ocfs2_xattr_block_find(struct inode *inode,
 	if (!di->i_xattr_loc)
 		return ret;
 
-	ret = ocfs2_read_block(inode, le64_to_cpu(di->i_xattr_loc), &blk_bh);
+	ret = ocfs2_read_xattr_block(inode, le64_to_cpu(di->i_xattr_loc),
+				     &blk_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		return ret;
 	}
 
-	xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
-	if (!OCFS2_IS_VALID_XATTR_BLOCK(xb)) {
-		ret = -EIO;
-		goto cleanup;
-	}
-
 	xs->xattr_bh = blk_bh;
+	xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
 
 	if (!(le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED)) {
 		xs->header = &xb->xb_attrs.xb_header;
@@ -1804,13 +2085,13 @@ cleanup:
  */
 static int ocfs2_xattr_block_set(struct inode *inode,
 				 struct ocfs2_xattr_info *xi,
-				 struct ocfs2_xattr_search *xs)
+				 struct ocfs2_xattr_search *xs,
+				 struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	struct buffer_head *new_bh = NULL;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_dinode *di =  (struct ocfs2_dinode *)xs->inode_bh->b_data;
-	struct ocfs2_alloc_context *meta_ac = NULL;
-	handle_t *handle = NULL;
+	handle_t *handle = ctxt->handle;
 	struct ocfs2_xattr_block *xblk = NULL;
 	u16 suballoc_bit_start;
 	u32 num_got;
@@ -1818,45 +2099,29 @@ static int ocfs2_xattr_block_set(struct inode *inode,
 	int ret;
 
 	if (!xs->xattr_bh) {
-		/*
-		 * Alloc one external block for extended attribute
-		 * outside of inode.
-		 */
-		ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
+		ret = ocfs2_journal_access_di(handle, inode, xs->inode_bh,
+					      OCFS2_JOURNAL_ACCESS_CREATE);
 		if (ret < 0) {
 			mlog_errno(ret);
-			goto out;
-		}
-		handle = ocfs2_start_trans(osb,
-					   OCFS2_XATTR_BLOCK_CREATE_CREDITS);
-		if (IS_ERR(handle)) {
-			ret = PTR_ERR(handle);
-			mlog_errno(ret);
-			goto out;
-		}
-		ret = ocfs2_journal_access(handle, inode, xs->inode_bh,
-					   OCFS2_JOURNAL_ACCESS_CREATE);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto out_commit;
+			goto end;
 		}
 
-		ret = ocfs2_claim_metadata(osb, handle, meta_ac, 1,
+		ret = ocfs2_claim_metadata(osb, handle, ctxt->meta_ac, 1,
 					   &suballoc_bit_start, &num_got,
 					   &first_blkno);
 		if (ret < 0) {
 			mlog_errno(ret);
-			goto out_commit;
+			goto end;
 		}
 
 		new_bh = sb_getblk(inode->i_sb, first_blkno);
 		ocfs2_set_new_buffer_uptodate(inode, new_bh);
 
-		ret = ocfs2_journal_access(handle, inode, new_bh,
-					   OCFS2_JOURNAL_ACCESS_CREATE);
+		ret = ocfs2_journal_access_xb(handle, inode, new_bh,
+					      OCFS2_JOURNAL_ACCESS_CREATE);
 		if (ret < 0) {
 			mlog_errno(ret);
-			goto out_commit;
+			goto end;
 		}
 
 		/* Initialize ocfs2_xattr_block */
@@ -1874,44 +2139,555 @@ static int ocfs2_xattr_block_set(struct inode *inode,
 		xs->end = (void *)xblk + inode->i_sb->s_blocksize;
 		xs->here = xs->header->xh_entries;
 
-
 		ret = ocfs2_journal_dirty(handle, new_bh);
 		if (ret < 0) {
 			mlog_errno(ret);
-			goto out_commit;
+			goto end;
 		}
 		di->i_xattr_loc = cpu_to_le64(first_blkno);
-		ret = ocfs2_journal_dirty(handle, xs->inode_bh);
-		if (ret < 0)
-			mlog_errno(ret);
-out_commit:
-		ocfs2_commit_trans(osb, handle);
-out:
-		if (meta_ac)
-			ocfs2_free_alloc_context(meta_ac);
-		if (ret < 0)
-			return ret;
+		ocfs2_journal_dirty(handle, xs->inode_bh);
 	} else
 		xblk = (struct ocfs2_xattr_block *)xs->xattr_bh->b_data;
 
 	if (!(le16_to_cpu(xblk->xb_flags) & OCFS2_XATTR_INDEXED)) {
 		/* Set extended attribute into external block */
-		ret = ocfs2_xattr_set_entry(inode, xi, xs, OCFS2_HAS_XATTR_FL);
+		ret = ocfs2_xattr_set_entry(inode, xi, xs, ctxt,
+					    OCFS2_HAS_XATTR_FL);
 		if (!ret || ret != -ENOSPC)
 			goto end;
 
-		ret = ocfs2_xattr_create_index_block(inode, xs);
+		ret = ocfs2_xattr_create_index_block(inode, xs, ctxt);
 		if (ret)
 			goto end;
 	}
 
-	ret = ocfs2_xattr_set_entry_index_block(inode, xi, xs);
+	ret = ocfs2_xattr_set_entry_index_block(inode, xi, xs, ctxt);
 
 end:
 
 	return ret;
 }
 
+/* Check whether the new xattr can be inserted into the inode. */
+static int ocfs2_xattr_can_be_in_inode(struct inode *inode,
+				       struct ocfs2_xattr_info *xi,
+				       struct ocfs2_xattr_search *xs)
+{
+	u64 value_size;
+	struct ocfs2_xattr_entry *last;
+	int free, i;
+	size_t min_offs = xs->end - xs->base;
+
+	if (!xs->header)
+		return 0;
+
+	last = xs->header->xh_entries;
+
+	for (i = 0; i < le16_to_cpu(xs->header->xh_count); i++) {
+		size_t offs = le16_to_cpu(last->xe_name_offset);
+		if (offs < min_offs)
+			min_offs = offs;
+		last += 1;
+	}
+
+	free = min_offs - ((void *)last - xs->base) - sizeof(__u32);
+	if (free < 0)
+		return 0;
+
+	BUG_ON(!xs->not_found);
+
+	if (xi->value_len > OCFS2_XATTR_INLINE_SIZE)
+		value_size = OCFS2_XATTR_ROOT_SIZE;
+	else
+		value_size = OCFS2_XATTR_SIZE(xi->value_len);
+
+	if (free >= sizeof(struct ocfs2_xattr_entry) +
+		   OCFS2_XATTR_SIZE(strlen(xi->name)) + value_size)
+		return 1;
+
+	return 0;
+}
+
+static int ocfs2_calc_xattr_set_need(struct inode *inode,
+				     struct ocfs2_dinode *di,
+				     struct ocfs2_xattr_info *xi,
+				     struct ocfs2_xattr_search *xis,
+				     struct ocfs2_xattr_search *xbs,
+				     int *clusters_need,
+				     int *meta_need,
+				     int *credits_need)
+{
+	int ret = 0, old_in_xb = 0;
+	int clusters_add = 0, meta_add = 0, credits = 0;
+	struct buffer_head *bh = NULL;
+	struct ocfs2_xattr_block *xb = NULL;
+	struct ocfs2_xattr_entry *xe = NULL;
+	struct ocfs2_xattr_value_root *xv = NULL;
+	char *base = NULL;
+	int name_offset, name_len = 0;
+	u32 new_clusters = ocfs2_clusters_for_bytes(inode->i_sb,
+						    xi->value_len);
+	u64 value_size;
+
+	/*
+	 * Calculate the clusters we need to write.
+	 * No matter whether we replace an old one or add a new one,
+	 * we need this for writing.
+	 */
+	if (xi->value_len > OCFS2_XATTR_INLINE_SIZE)
+		credits += new_clusters *
+			   ocfs2_clusters_to_blocks(inode->i_sb, 1);
+
+	if (xis->not_found && xbs->not_found) {
+		credits += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+
+		if (xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
+			clusters_add += new_clusters;
+			credits += ocfs2_calc_extend_credits(inode->i_sb,
+							&def_xv.xv.xr_list,
+							new_clusters);
+		}
+
+		goto meta_guess;
+	}
+
+	if (!xis->not_found) {
+		xe = xis->here;
+		name_offset = le16_to_cpu(xe->xe_name_offset);
+		name_len = OCFS2_XATTR_SIZE(xe->xe_name_len);
+		base = xis->base;
+		credits += OCFS2_INODE_UPDATE_CREDITS;
+	} else {
+		int i, block_off = 0;
+		xb = (struct ocfs2_xattr_block *)xbs->xattr_bh->b_data;
+		xe = xbs->here;
+		name_offset = le16_to_cpu(xe->xe_name_offset);
+		name_len = OCFS2_XATTR_SIZE(xe->xe_name_len);
+		i = xbs->here - xbs->header->xh_entries;
+		old_in_xb = 1;
+
+		if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED) {
+			ret = ocfs2_xattr_bucket_get_name_value(inode,
+							bucket_xh(xbs->bucket),
+							i, &block_off,
+							&name_offset);
+			base = bucket_block(xbs->bucket, block_off);
+			credits += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+		} else {
+			base = xbs->base;
+			credits += OCFS2_XATTR_BLOCK_UPDATE_CREDITS;
+		}
+	}
+
+	/*
+	 * delete a xattr doesn't need metadata and cluster allocation.
+	 * so just calculate the credits and return.
+	 *
+	 * The credits for removing the value tree will be extended
+	 * by ocfs2_remove_extent itself.
+	 */
+	if (!xi->value) {
+		if (!ocfs2_xattr_is_local(xe))
+			credits += ocfs2_remove_extent_credits(inode->i_sb);
+
+		goto out;
+	}
+
+	/* do cluster allocation guess first. */
+	value_size = le64_to_cpu(xe->xe_value_size);
+
+	if (old_in_xb) {
+		/*
+		 * In xattr set, we always try to set the xe in inode first,
+		 * so if it can be inserted into inode successfully, the old
+		 * one will be removed from the xattr block, and this xattr
+		 * will be inserted into inode as a new xattr in inode.
+		 */
+		if (ocfs2_xattr_can_be_in_inode(inode, xi, xis)) {
+			clusters_add += new_clusters;
+			credits += ocfs2_remove_extent_credits(inode->i_sb) +
+				    OCFS2_INODE_UPDATE_CREDITS;
+			if (!ocfs2_xattr_is_local(xe))
+				credits += ocfs2_calc_extend_credits(
+							inode->i_sb,
+							&def_xv.xv.xr_list,
+							new_clusters);
+			goto out;
+		}
+	}
+
+	if (xi->value_len > OCFS2_XATTR_INLINE_SIZE) {
+		/* the new values will be stored outside. */
+		u32 old_clusters = 0;
+
+		if (!ocfs2_xattr_is_local(xe)) {
+			old_clusters =	ocfs2_clusters_for_bytes(inode->i_sb,
+								 value_size);
+			xv = (struct ocfs2_xattr_value_root *)
+			     (base + name_offset + name_len);
+			value_size = OCFS2_XATTR_ROOT_SIZE;
+		} else
+			xv = &def_xv.xv;
+
+		if (old_clusters >= new_clusters) {
+			credits += ocfs2_remove_extent_credits(inode->i_sb);
+			goto out;
+		} else {
+			meta_add += ocfs2_extend_meta_needed(&xv->xr_list);
+			clusters_add += new_clusters - old_clusters;
+			credits += ocfs2_calc_extend_credits(inode->i_sb,
+							     &xv->xr_list,
+							     new_clusters -
+							     old_clusters);
+			if (value_size >= OCFS2_XATTR_ROOT_SIZE)
+				goto out;
+		}
+	} else {
+		/*
+		 * Now the new value will be stored inside. So if the new
+		 * value is smaller than the size of value root or the old
+		 * value, we don't need any allocation, otherwise we have
+		 * to guess metadata allocation.
+		 */
+		if ((ocfs2_xattr_is_local(xe) && value_size >= xi->value_len) ||
+		    (!ocfs2_xattr_is_local(xe) &&
+		     OCFS2_XATTR_ROOT_SIZE >= xi->value_len))
+			goto out;
+	}
+
+meta_guess:
+	/* calculate metadata allocation. */
+	if (di->i_xattr_loc) {
+		if (!xbs->xattr_bh) {
+			ret = ocfs2_read_xattr_block(inode,
+						     le64_to_cpu(di->i_xattr_loc),
+						     &bh);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+
+			xb = (struct ocfs2_xattr_block *)bh->b_data;
+		} else
+			xb = (struct ocfs2_xattr_block *)xbs->xattr_bh->b_data;
+
+		/*
+		 * If there is already an xattr tree, good, we can calculate
+		 * like other b-trees. Otherwise we may have the chance of
+		 * create a tree, the credit calculation is borrowed from
+		 * ocfs2_calc_extend_credits with root_el = NULL. And the
+		 * new tree will be cluster based, so no meta is needed.
+		 */
+		if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED) {
+			struct ocfs2_extent_list *el =
+				 &xb->xb_attrs.xb_root.xt_list;
+			meta_add += ocfs2_extend_meta_needed(el);
+			credits += ocfs2_calc_extend_credits(inode->i_sb,
+							     el, 1);
+		} else
+			credits += OCFS2_SUBALLOC_ALLOC + 1;
+
+		/*
+		 * This cluster will be used either for new bucket or for
+		 * new xattr block.
+		 * If the cluster size is the same as the bucket size, one
+		 * more is needed since we may need to extend the bucket
+		 * also.
+		 */
+		clusters_add += 1;
+		credits += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+		if (OCFS2_XATTR_BUCKET_SIZE ==
+			OCFS2_SB(inode->i_sb)->s_clustersize) {
+			credits += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+			clusters_add += 1;
+		}
+	} else {
+		meta_add += 1;
+		credits += OCFS2_XATTR_BLOCK_CREATE_CREDITS;
+	}
+out:
+	if (clusters_need)
+		*clusters_need = clusters_add;
+	if (meta_need)
+		*meta_need = meta_add;
+	if (credits_need)
+		*credits_need = credits;
+	brelse(bh);
+	return ret;
+}
+
+static int ocfs2_init_xattr_set_ctxt(struct inode *inode,
+				     struct ocfs2_dinode *di,
+				     struct ocfs2_xattr_info *xi,
+				     struct ocfs2_xattr_search *xis,
+				     struct ocfs2_xattr_search *xbs,
+				     struct ocfs2_xattr_set_ctxt *ctxt,
+				     int *credits)
+{
+	int clusters_add, meta_add, ret;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+
+	memset(ctxt, 0, sizeof(struct ocfs2_xattr_set_ctxt));
+
+	ocfs2_init_dealloc_ctxt(&ctxt->dealloc);
+
+	ret = ocfs2_calc_xattr_set_need(inode, di, xi, xis, xbs,
+					&clusters_add, &meta_add, credits);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	mlog(0, "Set xattr %s, reserve meta blocks = %d, clusters = %d, "
+	     "credits = %d\n", xi->name, meta_add, clusters_add, *credits);
+
+	if (meta_add) {
+		ret = ocfs2_reserve_new_metadata_blocks(osb, meta_add,
+							&ctxt->meta_ac);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+	}
+
+	if (clusters_add) {
+		ret = ocfs2_reserve_clusters(osb, clusters_add, &ctxt->data_ac);
+		if (ret)
+			mlog_errno(ret);
+	}
+out:
+	if (ret) {
+		if (ctxt->meta_ac) {
+			ocfs2_free_alloc_context(ctxt->meta_ac);
+			ctxt->meta_ac = NULL;
+		}
+
+		/*
+		 * We cannot have an error and a non null ctxt->data_ac.
+		 */
+	}
+
+	return ret;
+}
+
+static int __ocfs2_xattr_set_handle(struct inode *inode,
+				    struct ocfs2_dinode *di,
+				    struct ocfs2_xattr_info *xi,
+				    struct ocfs2_xattr_search *xis,
+				    struct ocfs2_xattr_search *xbs,
+				    struct ocfs2_xattr_set_ctxt *ctxt)
+{
+	int ret = 0, credits, old_found;
+
+	if (!xi->value) {
+		/* Remove existing extended attribute */
+		if (!xis->not_found)
+			ret = ocfs2_xattr_ibody_set(inode, xi, xis, ctxt);
+		else if (!xbs->not_found)
+			ret = ocfs2_xattr_block_set(inode, xi, xbs, ctxt);
+	} else {
+		/* We always try to set extended attribute into inode first*/
+		ret = ocfs2_xattr_ibody_set(inode, xi, xis, ctxt);
+		if (!ret && !xbs->not_found) {
+			/*
+			 * If succeed and that extended attribute existing in
+			 * external block, then we will remove it.
+			 */
+			xi->value = NULL;
+			xi->value_len = 0;
+
+			old_found = xis->not_found;
+			xis->not_found = -ENODATA;
+			ret = ocfs2_calc_xattr_set_need(inode,
+							di,
+							xi,
+							xis,
+							xbs,
+							NULL,
+							NULL,
+							&credits);
+			xis->not_found = old_found;
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+
+			ret = ocfs2_extend_trans(ctxt->handle, credits +
+					ctxt->handle->h_buffer_credits);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+			ret = ocfs2_xattr_block_set(inode, xi, xbs, ctxt);
+		} else if (ret == -ENOSPC) {
+			if (di->i_xattr_loc && !xbs->xattr_bh) {
+				ret = ocfs2_xattr_block_find(inode,
+							     xi->name_index,
+							     xi->name, xbs);
+				if (ret)
+					goto out;
+
+				old_found = xis->not_found;
+				xis->not_found = -ENODATA;
+				ret = ocfs2_calc_xattr_set_need(inode,
+								di,
+								xi,
+								xis,
+								xbs,
+								NULL,
+								NULL,
+								&credits);
+				xis->not_found = old_found;
+				if (ret) {
+					mlog_errno(ret);
+					goto out;
+				}
+
+				ret = ocfs2_extend_trans(ctxt->handle, credits +
+					ctxt->handle->h_buffer_credits);
+				if (ret) {
+					mlog_errno(ret);
+					goto out;
+				}
+			}
+			/*
+			 * If no space in inode, we will set extended attribute
+			 * into external block.
+			 */
+			ret = ocfs2_xattr_block_set(inode, xi, xbs, ctxt);
+			if (ret)
+				goto out;
+			if (!xis->not_found) {
+				/*
+				 * If succeed and that extended attribute
+				 * existing in inode, we will remove it.
+				 */
+				xi->value = NULL;
+				xi->value_len = 0;
+				xbs->not_found = -ENODATA;
+				ret = ocfs2_calc_xattr_set_need(inode,
+								di,
+								xi,
+								xis,
+								xbs,
+								NULL,
+								NULL,
+								&credits);
+				if (ret) {
+					mlog_errno(ret);
+					goto out;
+				}
+
+				ret = ocfs2_extend_trans(ctxt->handle, credits +
+						ctxt->handle->h_buffer_credits);
+				if (ret) {
+					mlog_errno(ret);
+					goto out;
+				}
+				ret = ocfs2_xattr_ibody_set(inode, xi,
+							    xis, ctxt);
+			}
+		}
+	}
+
+	if (!ret) {
+		/* Update inode ctime. */
+		ret = ocfs2_journal_access(ctxt->handle, inode, xis->inode_bh,
+					   OCFS2_JOURNAL_ACCESS_WRITE);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		inode->i_ctime = CURRENT_TIME;
+		di->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec);
+		di->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec);
+		ocfs2_journal_dirty(ctxt->handle, xis->inode_bh);
+	}
+out:
+	return ret;
+}
+
+/*
+ * This function only called duing creating inode
+ * for init security/acl xattrs of the new inode.
+ * All transanction credits have been reserved in mknod.
+ */
+int ocfs2_xattr_set_handle(handle_t *handle,
+			   struct inode *inode,
+			   struct buffer_head *di_bh,
+			   int name_index,
+			   const char *name,
+			   const void *value,
+			   size_t value_len,
+			   int flags,
+			   struct ocfs2_alloc_context *meta_ac,
+			   struct ocfs2_alloc_context *data_ac)
+{
+	struct ocfs2_dinode *di;
+	int ret;
+
+	struct ocfs2_xattr_info xi = {
+		.name_index = name_index,
+		.name = name,
+		.value = value,
+		.value_len = value_len,
+	};
+
+	struct ocfs2_xattr_search xis = {
+		.not_found = -ENODATA,
+	};
+
+	struct ocfs2_xattr_search xbs = {
+		.not_found = -ENODATA,
+	};
+
+	struct ocfs2_xattr_set_ctxt ctxt = {
+		.handle = handle,
+		.meta_ac = meta_ac,
+		.data_ac = data_ac,
+	};
+
+	if (!ocfs2_supports_xattr(OCFS2_SB(inode->i_sb)))
+		return -EOPNOTSUPP;
+
+	/*
+	 * In extreme situation, may need xattr bucket when
+	 * block size is too small. And we have already reserved
+	 * the credits for bucket in mknod.
+	 */
+	if (inode->i_sb->s_blocksize == OCFS2_MIN_BLOCKSIZE) {
+		xbs.bucket = ocfs2_xattr_bucket_new(inode);
+		if (!xbs.bucket) {
+			mlog_errno(-ENOMEM);
+			return -ENOMEM;
+		}
+	}
+
+	xis.inode_bh = xbs.inode_bh = di_bh;
+	di = (struct ocfs2_dinode *)di_bh->b_data;
+
+	down_write(&OCFS2_I(inode)->ip_xattr_sem);
+
+	ret = ocfs2_xattr_ibody_find(inode, name_index, name, &xis);
+	if (ret)
+		goto cleanup;
+	if (xis.not_found) {
+		ret = ocfs2_xattr_block_find(inode, name_index, name, &xbs);
+		if (ret)
+			goto cleanup;
+	}
+
+	ret = __ocfs2_xattr_set_handle(inode, di, &xi, &xis, &xbs, &ctxt);
+
+cleanup:
+	up_write(&OCFS2_I(inode)->ip_xattr_sem);
+	brelse(xbs.xattr_bh);
+	ocfs2_xattr_bucket_free(xbs.bucket);
+
+	return ret;
+}
+
 /*
  * ocfs2_xattr_set()
  *
@@ -1928,8 +2704,10 @@ int ocfs2_xattr_set(struct inode *inode,
 {
 	struct buffer_head *di_bh = NULL;
 	struct ocfs2_dinode *di;
-	int ret;
-	u16 i, blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+	int ret, credits;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct inode *tl_inode = osb->osb_tl_inode;
+	struct ocfs2_xattr_set_ctxt ctxt = { NULL, NULL, };
 
 	struct ocfs2_xattr_info xi = {
 		.name_index = name_index,
@@ -1949,10 +2727,20 @@ int ocfs2_xattr_set(struct inode *inode,
 	if (!ocfs2_supports_xattr(OCFS2_SB(inode->i_sb)))
 		return -EOPNOTSUPP;
 
+	/*
+	 * Only xbs will be used on indexed trees.  xis doesn't need a
+	 * bucket.
+	 */
+	xbs.bucket = ocfs2_xattr_bucket_new(inode);
+	if (!xbs.bucket) {
+		mlog_errno(-ENOMEM);
+		return -ENOMEM;
+	}
+
 	ret = ocfs2_inode_lock(inode, &di_bh, 1);
 	if (ret < 0) {
 		mlog_errno(ret);
-		return ret;
+		goto cleanup_nolock;
 	}
 	xis.inode_bh = xbs.inode_bh = di_bh;
 	di = (struct ocfs2_dinode *)di_bh->b_data;
@@ -1984,55 +2772,53 @@ int ocfs2_xattr_set(struct inode *inode,
 			goto cleanup;
 	}
 
-	if (!value) {
-		/* Remove existing extended attribute */
-		if (!xis.not_found)
-			ret = ocfs2_xattr_ibody_set(inode, &xi, &xis);
-		else if (!xbs.not_found)
-			ret = ocfs2_xattr_block_set(inode, &xi, &xbs);
-	} else {
-		/* We always try to set extended attribute into inode first*/
-		ret = ocfs2_xattr_ibody_set(inode, &xi, &xis);
-		if (!ret && !xbs.not_found) {
-			/*
-			 * If succeed and that extended attribute existing in
-			 * external block, then we will remove it.
-			 */
-			xi.value = NULL;
-			xi.value_len = 0;
-			ret = ocfs2_xattr_block_set(inode, &xi, &xbs);
-		} else if (ret == -ENOSPC) {
-			if (di->i_xattr_loc && !xbs.xattr_bh) {
-				ret = ocfs2_xattr_block_find(inode, name_index,
-							     name, &xbs);
-				if (ret)
-					goto cleanup;
-			}
-			/*
-			 * If no space in inode, we will set extended attribute
-			 * into external block.
-			 */
-			ret = ocfs2_xattr_block_set(inode, &xi, &xbs);
-			if (ret)
-				goto cleanup;
-			if (!xis.not_found) {
-				/*
-				 * If succeed and that extended attribute
-				 * existing in inode, we will remove it.
-				 */
-				xi.value = NULL;
-				xi.value_len = 0;
-				ret = ocfs2_xattr_ibody_set(inode, &xi, &xis);
-			}
+
+	mutex_lock(&tl_inode->i_mutex);
+
+	if (ocfs2_truncate_log_needs_flush(osb)) {
+		ret = __ocfs2_flush_truncate_log(osb);
+		if (ret < 0) {
+			mutex_unlock(&tl_inode->i_mutex);
+			mlog_errno(ret);
+			goto cleanup;
 		}
 	}
+	mutex_unlock(&tl_inode->i_mutex);
+
+	ret = ocfs2_init_xattr_set_ctxt(inode, di, &xi, &xis,
+					&xbs, &ctxt, &credits);
+	if (ret) {
+		mlog_errno(ret);
+		goto cleanup;
+	}
+
+	/* we need to update inode's ctime field, so add credit for it. */
+	credits += OCFS2_INODE_UPDATE_CREDITS;
+	ctxt.handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(ctxt.handle)) {
+		ret = PTR_ERR(ctxt.handle);
+		mlog_errno(ret);
+		goto cleanup;
+	}
+
+	ret = __ocfs2_xattr_set_handle(inode, di, &xi, &xis, &xbs, &ctxt);
+
+	ocfs2_commit_trans(osb, ctxt.handle);
+
+	if (ctxt.data_ac)
+		ocfs2_free_alloc_context(ctxt.data_ac);
+	if (ctxt.meta_ac)
+		ocfs2_free_alloc_context(ctxt.meta_ac);
+	if (ocfs2_dealloc_has_cluster(&ctxt.dealloc))
+		ocfs2_schedule_truncate_log_flush(osb, 1);
+	ocfs2_run_deallocs(osb, &ctxt.dealloc);
 cleanup:
 	up_write(&OCFS2_I(inode)->ip_xattr_sem);
 	ocfs2_inode_unlock(inode, 1);
+cleanup_nolock:
 	brelse(di_bh);
 	brelse(xbs.xattr_bh);
-	for (i = 0; i < blk_per_bucket; i++)
-		brelse(xbs.bucket.bhs[i]);
+	ocfs2_xattr_bucket_free(xbs.bucket);
 
 	return ret;
 }
@@ -2107,7 +2893,7 @@ typedef int (xattr_bucket_func)(struct inode *inode,
 				void *para);
 
 static int ocfs2_find_xe_in_bucket(struct inode *inode,
-				   struct buffer_head *header_bh,
+				   struct ocfs2_xattr_bucket *bucket,
 				   int name_index,
 				   const char *name,
 				   u32 name_hash,
@@ -2115,11 +2901,9 @@ static int ocfs2_find_xe_in_bucket(struct inode *inode,
 				   int *found)
 {
 	int i, ret = 0, cmp = 1, block_off, new_offset;
-	struct ocfs2_xattr_header *xh =
-			(struct ocfs2_xattr_header *)header_bh->b_data;
+	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
 	size_t name_len = strlen(name);
 	struct ocfs2_xattr_entry *xe = NULL;
-	struct buffer_head *name_bh = NULL;
 	char *xe_name;
 
 	/*
@@ -2150,19 +2934,9 @@ static int ocfs2_find_xe_in_bucket(struct inode *inode,
 			break;
 		}
 
-		ret = ocfs2_read_block(inode, header_bh->b_blocknr + block_off,
-				       &name_bh);
-		if (ret) {
-			mlog_errno(ret);
-			break;
-		}
-		xe_name = name_bh->b_data + new_offset;
 
-		cmp = memcmp(name, xe_name, name_len);
-		brelse(name_bh);
-		name_bh = NULL;
-
-		if (cmp == 0) {
+		xe_name = bucket_block(bucket, block_off) + new_offset;
+		if (!memcmp(name, xe_name, name_len)) {
 			*xe_index = i;
 			*found = 1;
 			ret = 0;
@@ -2192,39 +2966,42 @@ static int ocfs2_xattr_bucket_find(struct inode *inode,
 				   struct ocfs2_xattr_search *xs)
 {
 	int ret, found = 0;
-	struct buffer_head *bh = NULL;
-	struct buffer_head *lower_bh = NULL;
 	struct ocfs2_xattr_header *xh = NULL;
 	struct ocfs2_xattr_entry *xe = NULL;
 	u16 index = 0;
 	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 	int low_bucket = 0, bucket, high_bucket;
+	struct ocfs2_xattr_bucket *search;
 	u32 last_hash;
-	u64 blkno;
+	u64 blkno, lower_blkno = 0;
 
-	ret = ocfs2_read_block(inode, p_blkno, &bh);
+	search = ocfs2_xattr_bucket_new(inode);
+	if (!search) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_read_xattr_bucket(search, p_blkno);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	xh = (struct ocfs2_xattr_header *)bh->b_data;
+	xh = bucket_xh(search);
 	high_bucket = le16_to_cpu(xh->xh_num_buckets) - 1;
-
 	while (low_bucket <= high_bucket) {
-		brelse(bh);
-		bh = NULL;
-		bucket = (low_bucket + high_bucket) / 2;
+		ocfs2_xattr_bucket_relse(search);
 
+		bucket = (low_bucket + high_bucket) / 2;
 		blkno = p_blkno + bucket * blk_per_bucket;
-
-		ret = ocfs2_read_block(inode, blkno, &bh);
+		ret = ocfs2_read_xattr_bucket(search, blkno);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		xh = (struct ocfs2_xattr_header *)bh->b_data;
+		xh = bucket_xh(search);
 		xe = &xh->xh_entries[0];
 		if (name_hash < le32_to_cpu(xe->xe_name_hash)) {
 			high_bucket = bucket - 1;
@@ -2241,10 +3018,8 @@ static int ocfs2_xattr_bucket_find(struct inode *inode,
 
 		last_hash = le32_to_cpu(xe->xe_name_hash);
 
-		/* record lower_bh which may be the insert place. */
-		brelse(lower_bh);
-		lower_bh = bh;
-		bh = NULL;
+		/* record lower_blkno which may be the insert place. */
+		lower_blkno = blkno;
 
 		if (name_hash > le32_to_cpu(xe->xe_name_hash)) {
 			low_bucket = bucket + 1;
@@ -2252,7 +3027,7 @@ static int ocfs2_xattr_bucket_find(struct inode *inode,
 		}
 
 		/* the searched xattr should reside in this bucket if exists. */
-		ret = ocfs2_find_xe_in_bucket(inode, lower_bh,
+		ret = ocfs2_find_xe_in_bucket(inode, search,
 					      name_index, name, name_hash,
 					      &index, &found);
 		if (ret) {
@@ -2267,46 +3042,29 @@ static int ocfs2_xattr_bucket_find(struct inode *inode,
 	 * When the xattr's hash value is in the gap of 2 buckets, we will
 	 * always set it to the previous bucket.
 	 */
-	if (!lower_bh) {
-		/*
-		 * We can't find any bucket whose first name_hash is less
-		 * than the find name_hash.
-		 */
-		BUG_ON(bh->b_blocknr != p_blkno);
-		lower_bh = bh;
-		bh = NULL;
+	if (!lower_blkno)
+		lower_blkno = p_blkno;
+
+	/* This should be in cache - we just read it during the search */
+	ret = ocfs2_read_xattr_bucket(xs->bucket, lower_blkno);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
 	}
-	xs->bucket.bhs[0] = lower_bh;
-	xs->bucket.xh = (struct ocfs2_xattr_header *)
-					xs->bucket.bhs[0]->b_data;
-	lower_bh = NULL;
 
-	xs->header = xs->bucket.xh;
-	xs->base = xs->bucket.bhs[0]->b_data;
+	xs->header = bucket_xh(xs->bucket);
+	xs->base = bucket_block(xs->bucket, 0);
 	xs->end = xs->base + inode->i_sb->s_blocksize;
 
 	if (found) {
-		/*
-		 * If we have found the xattr enty, read all the blocks in
-		 * this bucket.
-		 */
-		ret = ocfs2_read_blocks(inode, xs->bucket.bhs[0]->b_blocknr + 1,
-					blk_per_bucket - 1, &xs->bucket.bhs[1],
-					0);
-		if (ret) {
-			mlog_errno(ret);
-			goto out;
-		}
-
 		xs->here = &xs->header->xh_entries[index];
 		mlog(0, "find xattr %s in bucket %llu, entry = %u\n", name,
-		     (unsigned long long)xs->bucket.bhs[0]->b_blocknr, index);
+		     (unsigned long long)bucket_blkno(xs->bucket), index);
 	} else
 		ret = -ENODATA;
 
 out:
-	brelse(bh);
-	brelse(lower_bh);
+	ocfs2_xattr_bucket_free(search);
 	return ret;
 }
 
@@ -2357,53 +3115,50 @@ static int ocfs2_iterate_xattr_buckets(struct inode *inode,
 				       xattr_bucket_func *func,
 				       void *para)
 {
-	int i, j, ret = 0;
-	int blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+	int i, ret = 0;
 	u32 bpc = ocfs2_xattr_buckets_per_cluster(OCFS2_SB(inode->i_sb));
 	u32 num_buckets = clusters * bpc;
-	struct ocfs2_xattr_bucket bucket;
+	struct ocfs2_xattr_bucket *bucket;
 
-	memset(&bucket, 0, sizeof(bucket));
+	bucket = ocfs2_xattr_bucket_new(inode);
+	if (!bucket) {
+		mlog_errno(-ENOMEM);
+		return -ENOMEM;
+	}
 
 	mlog(0, "iterating xattr buckets in %u clusters starting from %llu\n",
 	     clusters, (unsigned long long)blkno);
 
-	for (i = 0; i < num_buckets; i++, blkno += blk_per_bucket) {
-		ret = ocfs2_read_blocks(inode, blkno, blk_per_bucket,
-					bucket.bhs, 0);
+	for (i = 0; i < num_buckets; i++, blkno += bucket->bu_blocks) {
+		ret = ocfs2_read_xattr_bucket(bucket, blkno);
 		if (ret) {
 			mlog_errno(ret);
-			goto out;
+			break;
 		}
 
-		bucket.xh = (struct ocfs2_xattr_header *)bucket.bhs[0]->b_data;
 		/*
 		 * The real bucket num in this series of blocks is stored
 		 * in the 1st bucket.
 		 */
 		if (i == 0)
-			num_buckets = le16_to_cpu(bucket.xh->xh_num_buckets);
+			num_buckets = le16_to_cpu(bucket_xh(bucket)->xh_num_buckets);
 
 		mlog(0, "iterating xattr bucket %llu, first hash %u\n",
 		     (unsigned long long)blkno,
-		     le32_to_cpu(bucket.xh->xh_entries[0].xe_name_hash));
+		     le32_to_cpu(bucket_xh(bucket)->xh_entries[0].xe_name_hash));
 		if (func) {
-			ret = func(inode, &bucket, para);
-			if (ret) {
+			ret = func(inode, bucket, para);
+			if (ret)
 				mlog_errno(ret);
-				break;
-			}
+			/* Fall through to bucket_relse() */
 		}
 
-		for (j = 0; j < blk_per_bucket; j++)
-			brelse(bucket.bhs[j]);
-		memset(&bucket, 0, sizeof(bucket));
+		ocfs2_xattr_bucket_relse(bucket);
+		if (ret)
+			break;
 	}
 
-out:
-	for (j = 0; j < blk_per_bucket; j++)
-		brelse(bucket.bhs[j]);
-
+	ocfs2_xattr_bucket_free(bucket);
 	return ret;
 }
 
@@ -2441,21 +3196,21 @@ static int ocfs2_list_xattr_bucket(struct inode *inode,
 	int i, block_off, new_offset;
 	const char *prefix, *name;
 
-	for (i = 0 ; i < le16_to_cpu(bucket->xh->xh_count); i++) {
-		struct ocfs2_xattr_entry *entry = &bucket->xh->xh_entries[i];
+	for (i = 0 ; i < le16_to_cpu(bucket_xh(bucket)->xh_count); i++) {
+		struct ocfs2_xattr_entry *entry = &bucket_xh(bucket)->xh_entries[i];
 		type = ocfs2_xattr_get_type(entry);
 		prefix = ocfs2_xattr_prefix(type);
 
 		if (prefix) {
 			ret = ocfs2_xattr_bucket_get_name_value(inode,
-								bucket->xh,
+								bucket_xh(bucket),
 								i,
 								&block_off,
 								&new_offset);
 			if (ret)
 				break;
 
-			name = (const char *)bucket->bhs[block_off]->b_data +
+			name = (const char *)bucket_block(bucket, block_off) +
 				new_offset;
 			ret = ocfs2_xattr_list_entry(xl->buffer,
 						     xl->buffer_size,
@@ -2540,32 +3295,34 @@ static void swap_xe(void *a, void *b, int size)
 /*
  * When the ocfs2_xattr_block is filled up, new bucket will be created
  * and all the xattr entries will be moved to the new bucket.
+ * The header goes at the start of the bucket, and the names+values are
+ * filled from the end.  This is why *target starts as the last buffer.
  * Note: we need to sort the entries since they are not saved in order
  * in the ocfs2_xattr_block.
  */
 static void ocfs2_cp_xattr_block_to_bucket(struct inode *inode,
 					   struct buffer_head *xb_bh,
-					   struct buffer_head *xh_bh,
-					   struct buffer_head *data_bh)
+					   struct ocfs2_xattr_bucket *bucket)
 {
 	int i, blocksize = inode->i_sb->s_blocksize;
+	int blks = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 	u16 offset, size, off_change;
 	struct ocfs2_xattr_entry *xe;
 	struct ocfs2_xattr_block *xb =
 				(struct ocfs2_xattr_block *)xb_bh->b_data;
 	struct ocfs2_xattr_header *xb_xh = &xb->xb_attrs.xb_header;
-	struct ocfs2_xattr_header *xh =
-				(struct ocfs2_xattr_header *)xh_bh->b_data;
+	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
 	u16 count = le16_to_cpu(xb_xh->xh_count);
-	char *target = xh_bh->b_data, *src = xb_bh->b_data;
+	char *src = xb_bh->b_data;
+	char *target = bucket_block(bucket, blks - 1);
 
 	mlog(0, "cp xattr from block %llu to bucket %llu\n",
 	     (unsigned long long)xb_bh->b_blocknr,
-	     (unsigned long long)xh_bh->b_blocknr);
+	     (unsigned long long)bucket_blkno(bucket));
+
+	for (i = 0; i < blks; i++)
+		memset(bucket_block(bucket, i), 0, blocksize);
 
-	memset(xh_bh->b_data, 0, blocksize);
-	if (data_bh)
-		memset(data_bh->b_data, 0, blocksize);
 	/*
 	 * Since the xe_name_offset is based on ocfs2_xattr_header,
 	 * there is a offset change corresponding to the change of
@@ -2577,8 +3334,6 @@ static void ocfs2_cp_xattr_block_to_bucket(struct inode *inode,
 	size = blocksize - offset;
 
 	/* copy all the names and values. */
-	if (data_bh)
-		target = data_bh->b_data;
 	memcpy(target + offset, src + offset, size);
 
 	/* Init new header now. */
@@ -2588,7 +3343,7 @@ static void ocfs2_cp_xattr_block_to_bucket(struct inode *inode,
 	xh->xh_free_start = cpu_to_le16(OCFS2_XATTR_BUCKET_SIZE - size);
 
 	/* copy all the entries. */
-	target = xh_bh->b_data;
+	target = bucket_block(bucket, 0);
 	offset = offsetof(struct ocfs2_xattr_header, xh_entries);
 	size = count * sizeof(struct ocfs2_xattr_entry);
 	memcpy(target + offset, (char *)xb_xh + offset, size);
@@ -2614,73 +3369,47 @@ static void ocfs2_cp_xattr_block_to_bucket(struct inode *inode,
  * While if the entry is in index b-tree, "bucket" indicates the
  * real place of the xattr.
  */
-static int ocfs2_xattr_update_xattr_search(struct inode *inode,
-					   struct ocfs2_xattr_search *xs,
-					   struct buffer_head *old_bh,
-					   struct buffer_head *new_bh)
+static void ocfs2_xattr_update_xattr_search(struct inode *inode,
+					    struct ocfs2_xattr_search *xs,
+					    struct buffer_head *old_bh)
 {
-	int ret = 0;
 	char *buf = old_bh->b_data;
 	struct ocfs2_xattr_block *old_xb = (struct ocfs2_xattr_block *)buf;
 	struct ocfs2_xattr_header *old_xh = &old_xb->xb_attrs.xb_header;
-	int i, blocksize = inode->i_sb->s_blocksize;
-	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-
-	xs->bucket.bhs[0] = new_bh;
-	get_bh(new_bh);
-	xs->bucket.xh = (struct ocfs2_xattr_header *)xs->bucket.bhs[0]->b_data;
-	xs->header = xs->bucket.xh;
+	int i;
 
-	xs->base = new_bh->b_data;
+	xs->header = bucket_xh(xs->bucket);
+	xs->base = bucket_block(xs->bucket, 0);
 	xs->end = xs->base + inode->i_sb->s_blocksize;
 
-	if (!xs->not_found) {
-		if (OCFS2_XATTR_BUCKET_SIZE != blocksize) {
-			ret = ocfs2_read_blocks(inode,
-					xs->bucket.bhs[0]->b_blocknr + 1,
-					blk_per_bucket - 1, &xs->bucket.bhs[1],
-					0);
-			if (ret) {
-				mlog_errno(ret);
-				return ret;
-			}
-
-		}
-		i = xs->here - old_xh->xh_entries;
-		xs->here = &xs->header->xh_entries[i];
-	}
+	if (xs->not_found)
+		return;
 
-	return ret;
+	i = xs->here - old_xh->xh_entries;
+	xs->here = &xs->header->xh_entries[i];
 }
 
 static int ocfs2_xattr_create_index_block(struct inode *inode,
-					  struct ocfs2_xattr_search *xs)
+					  struct ocfs2_xattr_search *xs,
+					  struct ocfs2_xattr_set_ctxt *ctxt)
 {
-	int ret, credits = OCFS2_SUBALLOC_ALLOC;
+	int ret;
 	u32 bit_off, len;
 	u64 blkno;
-	handle_t *handle;
+	handle_t *handle = ctxt->handle;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
-	struct ocfs2_alloc_context *data_ac;
-	struct buffer_head *xh_bh = NULL, *data_bh = NULL;
 	struct buffer_head *xb_bh = xs->xattr_bh;
 	struct ocfs2_xattr_block *xb =
 			(struct ocfs2_xattr_block *)xb_bh->b_data;
 	struct ocfs2_xattr_tree_root *xr;
 	u16 xb_flags = le16_to_cpu(xb->xb_flags);
-	u16 bpb = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 
 	mlog(0, "create xattr index block for %llu\n",
 	     (unsigned long long)xb_bh->b_blocknr);
 
 	BUG_ON(xb_flags & OCFS2_XATTR_INDEXED);
-
-	ret = ocfs2_reserve_clusters(osb, 1, &data_ac);
-	if (ret) {
-		mlog_errno(ret);
-		goto out;
-	}
+	BUG_ON(!xs->bucket);
 
 	/*
 	 * XXX:
@@ -2689,29 +3418,18 @@ static int ocfs2_xattr_create_index_block(struct inode *inode,
 	 */
 	down_write(&oi->ip_alloc_sem);
 
-	/*
-	 * 3 more credits, one for xattr block update, one for the 1st block
-	 * of the new xattr bucket and one for the value/data.
-	 */
-	credits += 3;
-	handle = ocfs2_start_trans(osb, credits);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		goto out_sem;
-	}
-
-	ret = ocfs2_journal_access(handle, inode, xb_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_xb(handle, inode, xb_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
-	ret = ocfs2_claim_clusters(osb, handle, data_ac, 1, &bit_off, &len);
+	ret = __ocfs2_claim_clusters(osb, handle, ctxt->data_ac,
+				     1, 1, &bit_off, &len);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
 	/*
@@ -2724,51 +3442,23 @@ static int ocfs2_xattr_create_index_block(struct inode *inode,
 	mlog(0, "allocate 1 cluster from %llu to xattr block\n",
 	     (unsigned long long)blkno);
 
-	xh_bh = sb_getblk(inode->i_sb, blkno);
-	if (!xh_bh) {
-		ret = -EIO;
+	ret = ocfs2_init_xattr_bucket(xs->bucket, blkno);
+	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		goto out;
 	}
 
-	ocfs2_set_new_buffer_uptodate(inode, xh_bh);
-
-	ret = ocfs2_journal_access(handle, inode, xh_bh,
-				   OCFS2_JOURNAL_ACCESS_CREATE);
+	ret = ocfs2_xattr_bucket_journal_access(handle, xs->bucket,
+						OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
-	}
-
-	if (bpb > 1) {
-		data_bh = sb_getblk(inode->i_sb, blkno + bpb - 1);
-		if (!data_bh) {
-			ret = -EIO;
-			mlog_errno(ret);
-			goto out_commit;
-		}
-
-		ocfs2_set_new_buffer_uptodate(inode, data_bh);
-
-		ret = ocfs2_journal_access(handle, inode, data_bh,
-					   OCFS2_JOURNAL_ACCESS_CREATE);
-		if (ret) {
-			mlog_errno(ret);
-			goto out_commit;
-		}
+		goto out;
 	}
 
-	ocfs2_cp_xattr_block_to_bucket(inode, xb_bh, xh_bh, data_bh);
+	ocfs2_cp_xattr_block_to_bucket(inode, xb_bh, xs->bucket);
+	ocfs2_xattr_bucket_journal_dirty(handle, xs->bucket);
 
-	ocfs2_journal_dirty(handle, xh_bh);
-	if (data_bh)
-		ocfs2_journal_dirty(handle, data_bh);
-
-	ret = ocfs2_xattr_update_xattr_search(inode, xs, xb_bh, xh_bh);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_commit;
-	}
+	ocfs2_xattr_update_xattr_search(inode, xs, xb_bh);
 
 	/* Change from ocfs2_xattr_header to ocfs2_xattr_tree_root */
 	memset(&xb->xb_attrs, 0, inode->i_sb->s_blocksize -
@@ -2787,24 +3477,10 @@ static int ocfs2_xattr_create_index_block(struct inode *inode,
 
 	xb->xb_flags = cpu_to_le16(xb_flags | OCFS2_XATTR_INDEXED);
 
-	ret = ocfs2_journal_dirty(handle, xb_bh);
-	if (ret) {
-		mlog_errno(ret);
-		goto out_commit;
-	}
-
-out_commit:
-	ocfs2_commit_trans(osb, handle);
-
-out_sem:
-	up_write(&oi->ip_alloc_sem);
+	ocfs2_journal_dirty(handle, xb_bh);
 
 out:
-	if (data_ac)
-		ocfs2_free_alloc_context(data_ac);
-
-	brelse(xh_bh);
-	brelse(data_bh);
+	up_write(&oi->ip_alloc_sem);
 
 	return ret;
 }
@@ -2829,29 +3505,18 @@ static int cmp_xe_offset(const void *a, const void *b)
  * so that we can spare some space for insertion.
  */
 static int ocfs2_defrag_xattr_bucket(struct inode *inode,
+				     handle_t *handle,
 				     struct ocfs2_xattr_bucket *bucket)
 {
 	int ret, i;
 	size_t end, offset, len, value_len;
 	struct ocfs2_xattr_header *xh;
 	char *entries, *buf, *bucket_buf = NULL;
-	u64 blkno = bucket->bhs[0]->b_blocknr;
-	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+	u64 blkno = bucket_blkno(bucket);
 	u16 xh_free_start;
 	size_t blocksize = inode->i_sb->s_blocksize;
-	handle_t *handle;
-	struct buffer_head **bhs;
 	struct ocfs2_xattr_entry *xe;
 
-	bhs = kzalloc(sizeof(struct buffer_head *) * blk_per_bucket,
-			GFP_NOFS);
-	if (!bhs)
-		return -ENOMEM;
-
-	ret = ocfs2_read_blocks(inode, blkno, blk_per_bucket, bhs, 0);
-	if (ret)
-		goto out;
-
 	/*
 	 * In order to make the operation more efficient and generic,
 	 * we copy all the blocks into a contiguous memory and do the
@@ -2865,26 +3530,16 @@ static int ocfs2_defrag_xattr_bucket(struct inode *inode,
 	}
 
 	buf = bucket_buf;
-	for (i = 0; i < blk_per_bucket; i++, buf += blocksize)
-		memcpy(buf, bhs[i]->b_data, blocksize);
+	for (i = 0; i < bucket->bu_blocks; i++, buf += blocksize)
+		memcpy(buf, bucket_block(bucket, i), blocksize);
 
-	handle = ocfs2_start_trans((OCFS2_SB(inode->i_sb)), blk_per_bucket);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		handle = NULL;
+	ret = ocfs2_xattr_bucket_journal_access(handle, bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	for (i = 0; i < blk_per_bucket; i++) {
-		ret = ocfs2_journal_access(handle, inode, bhs[i],
-					   OCFS2_JOURNAL_ACCESS_WRITE);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto commit;
-		}
-	}
-
 	xh = (struct ocfs2_xattr_header *)bucket_buf;
 	entries = (char *)xh->xh_entries;
 	xh_free_start = le16_to_cpu(xh->xh_free_start);
@@ -2940,7 +3595,7 @@ static int ocfs2_defrag_xattr_bucket(struct inode *inode,
 			"bucket %llu\n", (unsigned long long)blkno);
 
 	if (xh_free_start == end)
-		goto commit;
+		goto out;
 
 	memset(bucket_buf + xh_free_start, 0, end - xh_free_start);
 	xh->xh_free_start = cpu_to_le16(end);
@@ -2951,169 +3606,94 @@ static int ocfs2_defrag_xattr_bucket(struct inode *inode,
 	     cmp_xe, swap_xe);
 
 	buf = bucket_buf;
-	for (i = 0; i < blk_per_bucket; i++, buf += blocksize) {
-		memcpy(bhs[i]->b_data, buf, blocksize);
-		ocfs2_journal_dirty(handle, bhs[i]);
-	}
+	for (i = 0; i < bucket->bu_blocks; i++, buf += blocksize)
+		memcpy(bucket_block(bucket, i), buf, blocksize);
+	ocfs2_xattr_bucket_journal_dirty(handle, bucket);
 
-commit:
-	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
 out:
-
-	if (bhs) {
-		for (i = 0; i < blk_per_bucket; i++)
-			brelse(bhs[i]);
-	}
-	kfree(bhs);
-
 	kfree(bucket_buf);
 	return ret;
 }
 
 /*
- * Move half nums of the xattr bucket in the previous cluster to this new
- * cluster. We only touch the last cluster of the previous extend record.
+ * prev_blkno points to the start of an existing extent.  new_blkno
+ * points to a newly allocated extent.  Because we know each of our
+ * clusters contains more than bucket, we can easily split one cluster
+ * at a bucket boundary.  So we take the last cluster of the existing
+ * extent and split it down the middle.  We move the last half of the
+ * buckets in the last cluster of the existing extent over to the new
+ * extent.
+ *
+ * first_bh is the buffer at prev_blkno so we can update the existing
+ * extent's bucket count.  header_bh is the bucket were we were hoping
+ * to insert our xattr.  If the bucket move places the target in the new
+ * extent, we'll update first_bh and header_bh after modifying the old
+ * extent.
  *
- * first_bh is the first buffer_head of a series of bucket in the same
- * extent rec and header_bh is the header of one bucket in this cluster.
- * They will be updated if we move the data header_bh contains to the new
- * cluster. first_hash will be set as the 1st xe's name_hash of the new cluster.
+ * first_hash will be set as the 1st xe's name_hash in the new extent.
  */
 static int ocfs2_mv_xattr_bucket_cross_cluster(struct inode *inode,
 					       handle_t *handle,
-					       struct buffer_head **first_bh,
-					       struct buffer_head **header_bh,
+					       struct ocfs2_xattr_bucket *first,
+					       struct ocfs2_xattr_bucket *target,
 					       u64 new_blkno,
-					       u64 prev_blkno,
 					       u32 num_clusters,
 					       u32 *first_hash)
 {
-	int i, ret, credits;
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	int bpc = ocfs2_clusters_to_blocks(inode->i_sb, 1);
-	int num_buckets = ocfs2_xattr_buckets_per_cluster(osb);
-	int blocksize = inode->i_sb->s_blocksize;
-	struct buffer_head *old_bh, *new_bh, *prev_bh, *new_first_bh = NULL;
-	struct ocfs2_xattr_header *new_xh;
-	struct ocfs2_xattr_header *xh =
-			(struct ocfs2_xattr_header *)((*first_bh)->b_data);
-
-	BUG_ON(le16_to_cpu(xh->xh_num_buckets) < num_buckets);
-	BUG_ON(OCFS2_XATTR_BUCKET_SIZE == osb->s_clustersize);
-
-	prev_bh = *first_bh;
-	get_bh(prev_bh);
-	xh = (struct ocfs2_xattr_header *)prev_bh->b_data;
+	int ret;
+	struct super_block *sb = inode->i_sb;
+	int blks_per_bucket = ocfs2_blocks_per_xattr_bucket(sb);
+	int num_buckets = ocfs2_xattr_buckets_per_cluster(OCFS2_SB(sb));
+	int to_move = num_buckets / 2;
+	u64 src_blkno;
+	u64 last_cluster_blkno = bucket_blkno(first) +
+		((num_clusters - 1) * ocfs2_clusters_to_blocks(sb, 1));
 
-	prev_blkno += (num_clusters - 1) * bpc + bpc / 2;
+	BUG_ON(le16_to_cpu(bucket_xh(first)->xh_num_buckets) < num_buckets);
+	BUG_ON(OCFS2_XATTR_BUCKET_SIZE == OCFS2_SB(sb)->s_clustersize);
 
 	mlog(0, "move half of xattrs in cluster %llu to %llu\n",
-	     (unsigned long long)prev_blkno, (unsigned long long)new_blkno);
+	     (unsigned long long)last_cluster_blkno, (unsigned long long)new_blkno);
 
-	/*
-	 * We need to update the 1st half of the new cluster and
-	 * 1 more for the update of the 1st bucket of the previous
-	 * extent record.
-	 */
-	credits = bpc / 2 + 1;
-	ret = ocfs2_extend_trans(handle, credits);
+	ret = ocfs2_mv_xattr_buckets(inode, handle, bucket_blkno(first),
+				     last_cluster_blkno, new_blkno,
+				     to_move, first_hash);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, prev_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
-	if (ret) {
-		mlog_errno(ret);
-		goto out;
-	}
+	/* This is the first bucket that got moved */
+	src_blkno = last_cluster_blkno + (to_move * blks_per_bucket);
 
-	for (i = 0; i < bpc / 2; i++, prev_blkno++, new_blkno++) {
-		old_bh = new_bh = NULL;
-		new_bh = sb_getblk(inode->i_sb, new_blkno);
-		if (!new_bh) {
-			ret = -EIO;
-			mlog_errno(ret);
-			goto out;
-		}
+	/*
+	 * If the target bucket was part of the moved buckets, we need to
+	 * update first and target.
+	 */
+	if (bucket_blkno(target) >= src_blkno) {
+		/* Find the block for the new target bucket */
+		src_blkno = new_blkno +
+			(bucket_blkno(target) - src_blkno);
 
-		ocfs2_set_new_buffer_uptodate(inode, new_bh);
+		ocfs2_xattr_bucket_relse(first);
+		ocfs2_xattr_bucket_relse(target);
 
-		ret = ocfs2_journal_access(handle, inode, new_bh,
-					   OCFS2_JOURNAL_ACCESS_CREATE);
-		if (ret < 0) {
+		/*
+		 * These shouldn't fail - the buffers are in the
+		 * journal from ocfs2_cp_xattr_bucket().
+		 */
+		ret = ocfs2_read_xattr_bucket(first, new_blkno);
+		if (ret) {
 			mlog_errno(ret);
-			brelse(new_bh);
 			goto out;
 		}
-
-		ret = ocfs2_read_block(inode, prev_blkno, &old_bh);
-		if (ret < 0) {
+		ret = ocfs2_read_xattr_bucket(target, src_blkno);
+		if (ret)
 			mlog_errno(ret);
-			brelse(new_bh);
-			goto out;
-		}
 
-		memcpy(new_bh->b_data, old_bh->b_data, blocksize);
-
-		if (i == 0) {
-			new_xh = (struct ocfs2_xattr_header *)new_bh->b_data;
-			new_xh->xh_num_buckets = cpu_to_le16(num_buckets / 2);
-
-			if (first_hash)
-				*first_hash = le32_to_cpu(
-					new_xh->xh_entries[0].xe_name_hash);
-			new_first_bh = new_bh;
-			get_bh(new_first_bh);
-		}
-
-		ocfs2_journal_dirty(handle, new_bh);
-
-		if (*header_bh == old_bh) {
-			brelse(*header_bh);
-			*header_bh = new_bh;
-			get_bh(*header_bh);
-
-			brelse(*first_bh);
-			*first_bh = new_first_bh;
-			get_bh(*first_bh);
-		}
-		brelse(new_bh);
-		brelse(old_bh);
 	}
 
-	le16_add_cpu(&xh->xh_num_buckets, -(num_buckets / 2));
-
-	ocfs2_journal_dirty(handle, prev_bh);
 out:
-	brelse(prev_bh);
-	brelse(new_first_bh);
-	return ret;
-}
-
-static int ocfs2_read_xattr_bucket(struct inode *inode,
-				   u64 blkno,
-				   struct buffer_head **bhs,
-				   int new)
-{
-	int ret = 0;
-	u16 i, blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-
-	if (!new)
-		return ocfs2_read_blocks(inode, blkno,
-					 blk_per_bucket, bhs, 0);
-
-	for (i = 0; i < blk_per_bucket; i++) {
-		bhs[i] = sb_getblk(inode->i_sb, blkno + i);
-		if (bhs[i] == NULL) {
-			ret = -EIO;
-			mlog_errno(ret);
-			break;
-		}
-		ocfs2_set_new_buffer_uptodate(inode, bhs[i]);
-	}
-
 	return ret;
 }
 
@@ -3178,8 +3758,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
 {
 	int ret, i;
 	int count, start, len, name_value_len = 0, xe_len, name_offset = 0;
-	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-	struct buffer_head **s_bhs, **t_bhs = NULL;
+	struct ocfs2_xattr_bucket *s_bucket = NULL, *t_bucket = NULL;
 	struct ocfs2_xattr_header *xh;
 	struct ocfs2_xattr_entry *xe;
 	int blocksize = inode->i_sb->s_blocksize;
@@ -3187,47 +3766,52 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
 	mlog(0, "move some of xattrs from bucket %llu to %llu\n",
 	     (unsigned long long)blk, (unsigned long long)new_blk);
 
-	s_bhs = kcalloc(blk_per_bucket, sizeof(struct buffer_head *), GFP_NOFS);
-	if (!s_bhs)
-		return -ENOMEM;
-
-	ret = ocfs2_read_xattr_bucket(inode, blk, s_bhs, 0);
-	if (ret) {
+	s_bucket = ocfs2_xattr_bucket_new(inode);
+	t_bucket = ocfs2_xattr_bucket_new(inode);
+	if (!s_bucket || !t_bucket) {
+		ret = -ENOMEM;
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, s_bhs[0],
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_read_xattr_bucket(s_bucket, blk);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	t_bhs = kcalloc(blk_per_bucket, sizeof(struct buffer_head *), GFP_NOFS);
-	if (!t_bhs) {
-		ret = -ENOMEM;
+	ret = ocfs2_xattr_bucket_journal_access(handle, s_bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_read_xattr_bucket(inode, new_blk, t_bhs, new_bucket_head);
+	/*
+	 * Even if !new_bucket_head, we're overwriting t_bucket.  Thus,
+	 * there's no need to read it.
+	 */
+	ret = ocfs2_init_xattr_bucket(t_bucket, new_blk);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	for (i = 0; i < blk_per_bucket; i++) {
-		ret = ocfs2_journal_access(handle, inode, t_bhs[i],
-					   new_bucket_head ?
-					   OCFS2_JOURNAL_ACCESS_CREATE :
-					   OCFS2_JOURNAL_ACCESS_WRITE);
-		if (ret) {
-			mlog_errno(ret);
-			goto out;
-		}
+	/*
+	 * Hey, if we're overwriting t_bucket, what difference does
+	 * ACCESS_CREATE vs ACCESS_WRITE make?  See the comment in the
+	 * same part of ocfs2_cp_xattr_bucket().
+	 */
+	ret = ocfs2_xattr_bucket_journal_access(handle, t_bucket,
+						new_bucket_head ?
+						OCFS2_JOURNAL_ACCESS_CREATE :
+						OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
 	}
 
-	xh = (struct ocfs2_xattr_header *)s_bhs[0]->b_data;
+	xh = bucket_xh(s_bucket);
 	count = le16_to_cpu(xh->xh_count);
 	start = ocfs2_xattr_find_divide_pos(xh);
 
@@ -3239,10 +3823,10 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
 		 * The hash value is set as one larger than
 		 * that of the last entry in the previous bucket.
 		 */
-		for (i = 0; i < blk_per_bucket; i++)
-			memset(t_bhs[i]->b_data, 0, blocksize);
+		for (i = 0; i < t_bucket->bu_blocks; i++)
+			memset(bucket_block(t_bucket, i), 0, blocksize);
 
-		xh = (struct ocfs2_xattr_header *)t_bhs[0]->b_data;
+		xh = bucket_xh(t_bucket);
 		xh->xh_free_start = cpu_to_le16(blocksize);
 		xh->xh_entries[0].xe_name_hash = xe->xe_name_hash;
 		le32_add_cpu(&xh->xh_entries[0].xe_name_hash, 1);
@@ -3251,11 +3835,10 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
 	}
 
 	/* copy the whole bucket to the new first. */
-	for (i = 0; i < blk_per_bucket; i++)
-		memcpy(t_bhs[i]->b_data, s_bhs[i]->b_data, blocksize);
+	ocfs2_xattr_bucket_copy_data(t_bucket, s_bucket);
 
 	/* update the new bucket. */
-	xh = (struct ocfs2_xattr_header *)t_bhs[0]->b_data;
+	xh = bucket_xh(t_bucket);
 
 	/*
 	 * Calculate the total name/value len and xh_free_start for
@@ -3319,11 +3902,7 @@ set_num_buckets:
 	else
 		xh->xh_num_buckets = 0;
 
-	for (i = 0; i < blk_per_bucket; i++) {
-		ocfs2_journal_dirty(handle, t_bhs[i]);
-		if (ret)
-			mlog_errno(ret);
-	}
+	ocfs2_xattr_bucket_journal_dirty(handle, t_bucket);
 
 	/* store the first_hash of the new bucket. */
 	if (first_hash)
@@ -3337,29 +3916,18 @@ set_num_buckets:
 	if (start == count)
 		goto out;
 
-	xh = (struct ocfs2_xattr_header *)s_bhs[0]->b_data;
+	xh = bucket_xh(s_bucket);
 	memset(&xh->xh_entries[start], 0,
 	       sizeof(struct ocfs2_xattr_entry) * (count - start));
 	xh->xh_count = cpu_to_le16(start);
 	xh->xh_free_start = cpu_to_le16(name_offset);
 	xh->xh_name_value_len = cpu_to_le16(name_value_len);
 
-	ocfs2_journal_dirty(handle, s_bhs[0]);
-	if (ret)
-		mlog_errno(ret);
+	ocfs2_xattr_bucket_journal_dirty(handle, s_bucket);
 
 out:
-	if (s_bhs) {
-		for (i = 0; i < blk_per_bucket; i++)
-			brelse(s_bhs[i]);
-	}
-	kfree(s_bhs);
-
-	if (t_bhs) {
-		for (i = 0; i < blk_per_bucket; i++)
-			brelse(t_bhs[i]);
-	}
-	kfree(t_bhs);
+	ocfs2_xattr_bucket_free(s_bucket);
+	ocfs2_xattr_bucket_free(t_bucket);
 
 	return ret;
 }
@@ -3376,10 +3944,8 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode,
 				 u64 t_blkno,
 				 int t_is_new)
 {
-	int ret, i;
-	int blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-	int blocksize = inode->i_sb->s_blocksize;
-	struct buffer_head **s_bhs, **t_bhs = NULL;
+	int ret;
+	struct ocfs2_xattr_bucket *s_bucket = NULL, *t_bucket = NULL;
 
 	BUG_ON(s_blkno == t_blkno);
 
@@ -3387,92 +3953,115 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode,
 	     (unsigned long long)s_blkno, (unsigned long long)t_blkno,
 	     t_is_new);
 
-	s_bhs = kzalloc(sizeof(struct buffer_head *) * blk_per_bucket,
-			GFP_NOFS);
-	if (!s_bhs)
-		return -ENOMEM;
+	s_bucket = ocfs2_xattr_bucket_new(inode);
+	t_bucket = ocfs2_xattr_bucket_new(inode);
+	if (!s_bucket || !t_bucket) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
 
-	ret = ocfs2_read_xattr_bucket(inode, s_blkno, s_bhs, 0);
+	ret = ocfs2_read_xattr_bucket(s_bucket, s_blkno);
 	if (ret)
 		goto out;
 
-	t_bhs = kzalloc(sizeof(struct buffer_head *) * blk_per_bucket,
-			GFP_NOFS);
-	if (!t_bhs) {
-		ret = -ENOMEM;
+	/*
+	 * Even if !t_is_new, we're overwriting t_bucket.  Thus,
+	 * there's no need to read it.
+	 */
+	ret = ocfs2_init_xattr_bucket(t_bucket, t_blkno);
+	if (ret)
 		goto out;
-	}
 
-	ret = ocfs2_read_xattr_bucket(inode, t_blkno, t_bhs, t_is_new);
+	/*
+	 * Hey, if we're overwriting t_bucket, what difference does
+	 * ACCESS_CREATE vs ACCESS_WRITE make?  Well, if we allocated a new
+	 * cluster to fill, we came here from
+	 * ocfs2_mv_xattr_buckets(), and it is really new -
+	 * ACCESS_CREATE is required.  But we also might have moved data
+	 * out of t_bucket before extending back into it.
+	 * ocfs2_add_new_xattr_bucket() can do this - its call to
+	 * ocfs2_add_new_xattr_cluster() may have created a new extent
+	 * and copied out the end of the old extent.  Then it re-extends
+	 * the old extent back to create space for new xattrs.  That's
+	 * how we get here, and the bucket isn't really new.
+	 */
+	ret = ocfs2_xattr_bucket_journal_access(handle, t_bucket,
+						t_is_new ?
+						OCFS2_JOURNAL_ACCESS_CREATE :
+						OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret)
 		goto out;
 
-	for (i = 0; i < blk_per_bucket; i++) {
-		ret = ocfs2_journal_access(handle, inode, t_bhs[i],
-					   t_is_new ?
-					   OCFS2_JOURNAL_ACCESS_CREATE :
-					   OCFS2_JOURNAL_ACCESS_WRITE);
-		if (ret)
-			goto out;
-	}
-
-	for (i = 0; i < blk_per_bucket; i++) {
-		memcpy(t_bhs[i]->b_data, s_bhs[i]->b_data, blocksize);
-		ocfs2_journal_dirty(handle, t_bhs[i]);
-	}
+	ocfs2_xattr_bucket_copy_data(t_bucket, s_bucket);
+	ocfs2_xattr_bucket_journal_dirty(handle, t_bucket);
 
 out:
-	if (s_bhs) {
-		for (i = 0; i < blk_per_bucket; i++)
-			brelse(s_bhs[i]);
-	}
-	kfree(s_bhs);
-
-	if (t_bhs) {
-		for (i = 0; i < blk_per_bucket; i++)
-			brelse(t_bhs[i]);
-	}
-	kfree(t_bhs);
+	ocfs2_xattr_bucket_free(t_bucket);
+	ocfs2_xattr_bucket_free(s_bucket);
 
 	return ret;
 }
 
 /*
- * Copy one xattr cluster from src_blk to to_blk.
- * The to_blk will become the first bucket header of the cluster, so its
- * xh_num_buckets will be initialized as the bucket num in the cluster.
+ * src_blk points to the start of an existing extent.  last_blk points to
+ * last cluster in that extent.  to_blk points to a newly allocated
+ * extent.  We copy the buckets from the cluster at last_blk to the new
+ * extent.  If start_bucket is non-zero, we skip that many buckets before
+ * we start copying.  The new extent's xh_num_buckets gets set to the
+ * number of buckets we copied.  The old extent's xh_num_buckets shrinks
+ * by the same amount.
  */
-static int ocfs2_cp_xattr_cluster(struct inode *inode,
-				  handle_t *handle,
-				  struct buffer_head *first_bh,
-				  u64 src_blk,
-				  u64 to_blk,
+static int ocfs2_mv_xattr_buckets(struct inode *inode, handle_t *handle,
+				  u64 src_blk, u64 last_blk, u64 to_blk,
+				  unsigned int start_bucket,
 				  u32 *first_hash)
 {
 	int i, ret, credits;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	int bpc = ocfs2_clusters_to_blocks(inode->i_sb, 1);
+	int blks_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 	int num_buckets = ocfs2_xattr_buckets_per_cluster(osb);
-	struct buffer_head *bh = NULL;
-	struct ocfs2_xattr_header *xh;
-	u64 to_blk_start = to_blk;
+	struct ocfs2_xattr_bucket *old_first, *new_first;
+
+	mlog(0, "mv xattrs from cluster %llu to %llu\n",
+	     (unsigned long long)last_blk, (unsigned long long)to_blk);
+
+	BUG_ON(start_bucket >= num_buckets);
+	if (start_bucket) {
+		num_buckets -= start_bucket;
+		last_blk += (start_bucket * blks_per_bucket);
+	}
+
+	/* The first bucket of the original extent */
+	old_first = ocfs2_xattr_bucket_new(inode);
+	/* The first bucket of the new extent */
+	new_first = ocfs2_xattr_bucket_new(inode);
+	if (!old_first || !new_first) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
 
-	mlog(0, "cp xattrs from cluster %llu to %llu\n",
-	     (unsigned long long)src_blk, (unsigned long long)to_blk);
+	ret = ocfs2_read_xattr_bucket(old_first, src_blk);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
 
 	/*
-	 * We need to update the new cluster and 1 more for the update of
-	 * the 1st bucket of the previous extent rec.
+	 * We need to update the first bucket of the old extent and all
+	 * the buckets going to the new extent.
 	 */
-	credits = bpc + 1;
+	credits = ((num_buckets + 1) * blks_per_bucket) +
+		handle->h_buffer_credits;
 	ret = ocfs2_extend_trans(handle, credits);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, first_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_xattr_bucket_journal_access(handle, old_first,
+						OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3480,45 +4069,45 @@ static int ocfs2_cp_xattr_cluster(struct inode *inode,
 
 	for (i = 0; i < num_buckets; i++) {
 		ret = ocfs2_cp_xattr_bucket(inode, handle,
-					    src_blk, to_blk, 1);
+					    last_blk + (i * blks_per_bucket),
+					    to_blk + (i * blks_per_bucket),
+					    1);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
-
-		src_blk += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-		to_blk += ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 	}
 
-	/* update the old bucket header. */
-	xh = (struct ocfs2_xattr_header *)first_bh->b_data;
-	le16_add_cpu(&xh->xh_num_buckets, -num_buckets);
-
-	ocfs2_journal_dirty(handle, first_bh);
-
-	/* update the new bucket header. */
-	ret = ocfs2_read_block(inode, to_blk_start, &bh);
-	if (ret < 0) {
+	/*
+	 * Get the new bucket ready before we dirty anything
+	 * (This actually shouldn't fail, because we already dirtied
+	 * it once in ocfs2_cp_xattr_bucket()).
+	 */
+	ret = ocfs2_read_xattr_bucket(new_first, to_blk);
+	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
-
-	ret = ocfs2_journal_access(handle, inode, bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_xattr_bucket_journal_access(handle, new_first,
+						OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	xh = (struct ocfs2_xattr_header *)bh->b_data;
-	xh->xh_num_buckets = cpu_to_le16(num_buckets);
+	/* Now update the headers */
+	le16_add_cpu(&bucket_xh(old_first)->xh_num_buckets, -num_buckets);
+	ocfs2_xattr_bucket_journal_dirty(handle, old_first);
 
-	ocfs2_journal_dirty(handle, bh);
+	bucket_xh(new_first)->xh_num_buckets = cpu_to_le16(num_buckets);
+	ocfs2_xattr_bucket_journal_dirty(handle, new_first);
 
 	if (first_hash)
-		*first_hash = le32_to_cpu(xh->xh_entries[0].xe_name_hash);
+		*first_hash = le32_to_cpu(bucket_xh(new_first)->xh_entries[0].xe_name_hash);
+
 out:
-	brelse(bh);
+	ocfs2_xattr_bucket_free(new_first);
+	ocfs2_xattr_bucket_free(old_first);
 	return ret;
 }
 
@@ -3534,7 +4123,7 @@ static int ocfs2_divide_xattr_cluster(struct inode *inode,
 				      u32 *first_hash)
 {
 	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-	int ret, credits = 2 * blk_per_bucket;
+	int ret, credits = 2 * blk_per_bucket + handle->h_buffer_credits;
 
 	BUG_ON(OCFS2_XATTR_BUCKET_SIZE < OCFS2_SB(inode->i_sb)->s_clustersize);
 
@@ -3577,43 +4166,49 @@ static int ocfs2_divide_xattr_cluster(struct inode *inode,
  */
 static int ocfs2_adjust_xattr_cross_cluster(struct inode *inode,
 					    handle_t *handle,
-					    struct buffer_head **first_bh,
-					    struct buffer_head **header_bh,
+					    struct ocfs2_xattr_bucket *first,
+					    struct ocfs2_xattr_bucket *target,
 					    u64 new_blk,
-					    u64 prev_blk,
 					    u32 prev_clusters,
 					    u32 *v_start,
 					    int *extend)
 {
-	int ret = 0;
-	int bpc = ocfs2_clusters_to_blocks(inode->i_sb, 1);
+	int ret;
 
 	mlog(0, "adjust xattrs from cluster %llu len %u to %llu\n",
-	     (unsigned long long)prev_blk, prev_clusters,
+	     (unsigned long long)bucket_blkno(first), prev_clusters,
 	     (unsigned long long)new_blk);
 
-	if (ocfs2_xattr_buckets_per_cluster(OCFS2_SB(inode->i_sb)) > 1)
+	if (ocfs2_xattr_buckets_per_cluster(OCFS2_SB(inode->i_sb)) > 1) {
 		ret = ocfs2_mv_xattr_bucket_cross_cluster(inode,
 							  handle,
-							  first_bh,
-							  header_bh,
+							  first, target,
 							  new_blk,
-							  prev_blk,
 							  prev_clusters,
 							  v_start);
-	else {
-		u64 last_blk = prev_blk + bpc * (prev_clusters - 1);
-
-		if (prev_clusters > 1 && (*header_bh)->b_blocknr != last_blk)
-			ret = ocfs2_cp_xattr_cluster(inode, handle, *first_bh,
-						     last_blk, new_blk,
+		if (ret)
+			mlog_errno(ret);
+	} else {
+		/* The start of the last cluster in the first extent */
+		u64 last_blk = bucket_blkno(first) +
+			((prev_clusters - 1) *
+			 ocfs2_clusters_to_blocks(inode->i_sb, 1));
+
+		if (prev_clusters > 1 && bucket_blkno(target) != last_blk) {
+			ret = ocfs2_mv_xattr_buckets(inode, handle,
+						     bucket_blkno(first),
+						     last_blk, new_blk, 0,
 						     v_start);
-		else {
+			if (ret)
+				mlog_errno(ret);
+		} else {
 			ret = ocfs2_divide_xattr_cluster(inode, handle,
 							 last_blk, new_blk,
 							 v_start);
+			if (ret)
+				mlog_errno(ret);
 
-			if ((*header_bh)->b_blocknr == last_blk && extend)
+			if ((bucket_blkno(target) == last_blk) && extend)
 				*extend = 0;
 		}
 	}
@@ -3639,56 +4234,37 @@ static int ocfs2_adjust_xattr_cross_cluster(struct inode *inode,
  */
 static int ocfs2_add_new_xattr_cluster(struct inode *inode,
 				       struct buffer_head *root_bh,
-				       struct buffer_head **first_bh,
-				       struct buffer_head **header_bh,
+				       struct ocfs2_xattr_bucket *first,
+				       struct ocfs2_xattr_bucket *target,
 				       u32 *num_clusters,
 				       u32 prev_cpos,
-				       u64 prev_blkno,
-				       int *extend)
+				       int *extend,
+				       struct ocfs2_xattr_set_ctxt *ctxt)
 {
-	int ret, credits;
+	int ret;
 	u16 bpc = ocfs2_clusters_to_blocks(inode->i_sb, 1);
 	u32 prev_clusters = *num_clusters;
 	u32 clusters_to_add = 1, bit_off, num_bits, v_start = 0;
 	u64 block;
-	handle_t *handle = NULL;
-	struct ocfs2_alloc_context *data_ac = NULL;
-	struct ocfs2_alloc_context *meta_ac = NULL;
+	handle_t *handle = ctxt->handle;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_extent_tree et;
 
 	mlog(0, "Add new xattr cluster for %llu, previous xattr hash = %u, "
 	     "previous xattr blkno = %llu\n",
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
-	     prev_cpos, (unsigned long long)prev_blkno);
+	     prev_cpos, (unsigned long long)bucket_blkno(first));
 
 	ocfs2_init_xattr_tree_extent_tree(&et, inode, root_bh);
 
-	ret = ocfs2_lock_allocators(inode, &et, clusters_to_add, 0,
-				    &data_ac, &meta_ac);
-	if (ret) {
-		mlog_errno(ret);
-		goto leave;
-	}
-
-	credits = ocfs2_calc_extend_credits(osb->sb, et.et_root_el,
-					    clusters_to_add);
-	handle = ocfs2_start_trans(osb, credits);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		handle = NULL;
-		mlog_errno(ret);
-		goto leave;
-	}
-
-	ret = ocfs2_journal_access(handle, inode, root_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_xb(handle, inode, root_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto leave;
 	}
 
-	ret = __ocfs2_claim_clusters(osb, handle, data_ac, 1,
+	ret = __ocfs2_claim_clusters(osb, handle, ctxt->data_ac, 1,
 				     clusters_to_add, &bit_off, &num_bits);
 	if (ret < 0) {
 		if (ret != -ENOSPC)
@@ -3702,7 +4278,7 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
 	mlog(0, "Allocating %u clusters at block %u for xattr in inode %llu\n",
 	     num_bits, bit_off, (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
-	if (prev_blkno + prev_clusters * bpc == block &&
+	if (bucket_blkno(first) + (prev_clusters * bpc) == block &&
 	    (prev_clusters + num_bits) << osb->s_clustersize_bits <=
 	     OCFS2_MAX_XATTR_TREE_LEAF_SIZE) {
 		/*
@@ -3721,10 +4297,9 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
 	} else {
 		ret = ocfs2_adjust_xattr_cross_cluster(inode,
 						       handle,
-						       first_bh,
-						       header_bh,
+						       first,
+						       target,
 						       block,
-						       prev_blkno,
 						       prev_clusters,
 						       &v_start,
 						       extend);
@@ -3734,149 +4309,137 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
 		}
 	}
 
-	if (handle->h_buffer_credits < credits) {
-		/*
-		 * The journal has been restarted before, and don't
-		 * have enough space for the insertion, so extend it
-		 * here.
-		 */
-		ret = ocfs2_extend_trans(handle, credits);
-		if (ret) {
-			mlog_errno(ret);
-			goto leave;
-		}
-	}
 	mlog(0, "Insert %u clusters at block %llu for xattr at %u\n",
 	     num_bits, (unsigned long long)block, v_start);
 	ret = ocfs2_insert_extent(osb, handle, inode, &et, v_start, block,
-				  num_bits, 0, meta_ac);
+				  num_bits, 0, ctxt->meta_ac);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto leave;
 	}
 
 	ret = ocfs2_journal_dirty(handle, root_bh);
-	if (ret < 0) {
+	if (ret < 0)
 		mlog_errno(ret);
-		goto leave;
-	}
 
 leave:
-	if (handle)
-		ocfs2_commit_trans(osb, handle);
-	if (data_ac)
-		ocfs2_free_alloc_context(data_ac);
-	if (meta_ac)
-		ocfs2_free_alloc_context(meta_ac);
-
 	return ret;
 }
 
 /*
- * Extend a new xattr bucket and move xattrs to the end one by one until
- * We meet with start_bh. Only move half of the xattrs to the bucket after it.
+ * We are given an extent.  'first' is the bucket at the very front of
+ * the extent.  The extent has space for an additional bucket past
+ * bucket_xh(first)->xh_num_buckets.  'target_blkno' is the block number
+ * of the target bucket.  We wish to shift every bucket past the target
+ * down one, filling in that additional space.  When we get back to the
+ * target, we split the target between itself and the now-empty bucket
+ * at target+1 (aka, target_blkno + blks_per_bucket).
  */
 static int ocfs2_extend_xattr_bucket(struct inode *inode,
-				     struct buffer_head *first_bh,
-				     struct buffer_head *start_bh,
+				     handle_t *handle,
+				     struct ocfs2_xattr_bucket *first,
+				     u64 target_blk,
 				     u32 num_clusters)
 {
 	int ret, credits;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-	u64 start_blk = start_bh->b_blocknr, end_blk;
-	u32 num_buckets = num_clusters * ocfs2_xattr_buckets_per_cluster(osb);
-	handle_t *handle;
-	struct ocfs2_xattr_header *first_xh =
-				(struct ocfs2_xattr_header *)first_bh->b_data;
-	u16 bucket = le16_to_cpu(first_xh->xh_num_buckets);
+	u64 end_blk;
+	u16 new_bucket = le16_to_cpu(bucket_xh(first)->xh_num_buckets);
 
 	mlog(0, "extend xattr bucket in %llu, xattr extend rec starting "
-	     "from %llu, len = %u\n", (unsigned long long)start_blk,
-	     (unsigned long long)first_bh->b_blocknr, num_clusters);
+	     "from %llu, len = %u\n", (unsigned long long)target_blk,
+	     (unsigned long long)bucket_blkno(first), num_clusters);
 
-	BUG_ON(bucket >= num_buckets);
+	/* The extent must have room for an additional bucket */
+	BUG_ON(new_bucket >=
+	       (num_clusters * ocfs2_xattr_buckets_per_cluster(osb)));
 
-	end_blk = first_bh->b_blocknr + (bucket - 1) * blk_per_bucket;
+	/* end_blk points to the last existing bucket */
+	end_blk = bucket_blkno(first) + ((new_bucket - 1) * blk_per_bucket);
 
 	/*
-	 * We will touch all the buckets after the start_bh(include it).
-	 * Add one more bucket and modify the first_bh.
+	 * end_blk is the start of the last existing bucket.
+	 * Thus, (end_blk - target_blk) covers the target bucket and
+	 * every bucket after it up to, but not including, the last
+	 * existing bucket.  Then we add the last existing bucket, the
+	 * new bucket, and the first bucket (3 * blk_per_bucket).
 	 */
-	credits = end_blk - start_blk + 2 * blk_per_bucket + 1;
-	handle = ocfs2_start_trans(osb, credits);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		handle = NULL;
+	credits = (end_blk - target_blk) + (3 * blk_per_bucket) +
+		  handle->h_buffer_credits;
+	ret = ocfs2_extend_trans(handle, credits);
+	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, first_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_xattr_bucket_journal_access(handle, first,
+						OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto commit;
+		goto out;
 	}
 
-	while (end_blk != start_blk) {
+	while (end_blk != target_blk) {
 		ret = ocfs2_cp_xattr_bucket(inode, handle, end_blk,
 					    end_blk + blk_per_bucket, 0);
 		if (ret)
-			goto commit;
+			goto out;
 		end_blk -= blk_per_bucket;
 	}
 
-	/* Move half of the xattr in start_blk to the next bucket. */
-	ret = ocfs2_divide_xattr_bucket(inode, handle, start_blk,
-					start_blk + blk_per_bucket, NULL, 0);
+	/* Move half of the xattr in target_blkno to the next bucket. */
+	ret = ocfs2_divide_xattr_bucket(inode, handle, target_blk,
+					target_blk + blk_per_bucket, NULL, 0);
 
-	le16_add_cpu(&first_xh->xh_num_buckets, 1);
-	ocfs2_journal_dirty(handle, first_bh);
+	le16_add_cpu(&bucket_xh(first)->xh_num_buckets, 1);
+	ocfs2_xattr_bucket_journal_dirty(handle, first);
 
-commit:
-	ocfs2_commit_trans(osb, handle);
 out:
 	return ret;
 }
 
 /*
- * Add new xattr bucket in an extent record and adjust the buckets accordingly.
- * xb_bh is the ocfs2_xattr_block.
- * We will move all the buckets starting from header_bh to the next place. As
- * for this one, half num of its xattrs will be moved to the next one.
+ * Add new xattr bucket in an extent record and adjust the buckets
+ * accordingly.  xb_bh is the ocfs2_xattr_block, and target is the
+ * bucket we want to insert into.
+ *
+ * In the easy case, we will move all the buckets after target down by
+ * one. Half of target's xattrs will be moved to the next bucket.
  *
- * We will allocate a new cluster if current cluster is full and adjust
- * header_bh and first_bh if the insert place is moved to the new cluster.
+ * If current cluster is full, we'll allocate a new one.  This may not
+ * be contiguous.  The underlying calls will make sure that there is
+ * space for the insert, shifting buckets around if necessary.
+ * 'target' may be moved by those calls.
  */
 static int ocfs2_add_new_xattr_bucket(struct inode *inode,
 				      struct buffer_head *xb_bh,
-				      struct buffer_head *header_bh)
+				      struct ocfs2_xattr_bucket *target,
+				      struct ocfs2_xattr_set_ctxt *ctxt)
 {
-	struct ocfs2_xattr_header *first_xh = NULL;
-	struct buffer_head *first_bh = NULL;
 	struct ocfs2_xattr_block *xb =
 			(struct ocfs2_xattr_block *)xb_bh->b_data;
 	struct ocfs2_xattr_tree_root *xb_root = &xb->xb_attrs.xb_root;
 	struct ocfs2_extent_list *el = &xb_root->xt_list;
-	struct ocfs2_xattr_header *xh =
-			(struct ocfs2_xattr_header *)header_bh->b_data;
-	u32 name_hash = le32_to_cpu(xh->xh_entries[0].xe_name_hash);
-	struct super_block *sb = inode->i_sb;
-	struct ocfs2_super *osb = OCFS2_SB(sb);
+	u32 name_hash =
+		le32_to_cpu(bucket_xh(target)->xh_entries[0].xe_name_hash);
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	int ret, num_buckets, extend = 1;
 	u64 p_blkno;
 	u32 e_cpos, num_clusters;
+	/* The bucket at the front of the extent */
+	struct ocfs2_xattr_bucket *first;
 
-	mlog(0, "Add new xattr bucket starting form %llu\n",
-	     (unsigned long long)header_bh->b_blocknr);
+	mlog(0, "Add new xattr bucket starting from %llu\n",
+	     (unsigned long long)bucket_blkno(target));
 
-	/*
-	 * Add refrence for header_bh here because it may be
-	 * changed in ocfs2_add_new_xattr_cluster and we need
-	 * to free it in the end.
-	 */
-	get_bh(header_bh);
+	/* The first bucket of the original extent */
+	first = ocfs2_xattr_bucket_new(inode);
+	if (!first) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
 
 	ret = ocfs2_xattr_get_rec(inode, name_hash, &p_blkno, &e_cpos,
 				  &num_clusters, el);
@@ -3885,40 +4448,45 @@ static int ocfs2_add_new_xattr_bucket(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_read_block(inode, p_blkno, &first_bh);
+	ret = ocfs2_read_xattr_bucket(first, p_blkno);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
 	num_buckets = ocfs2_xattr_buckets_per_cluster(osb) * num_clusters;
-	first_xh = (struct ocfs2_xattr_header *)first_bh->b_data;
-
-	if (num_buckets == le16_to_cpu(first_xh->xh_num_buckets)) {
+	if (num_buckets == le16_to_cpu(bucket_xh(first)->xh_num_buckets)) {
+		/*
+		 * This can move first+target if the target bucket moves
+		 * to the new extent.
+		 */
 		ret = ocfs2_add_new_xattr_cluster(inode,
 						  xb_bh,
-						  &first_bh,
-						  &header_bh,
+						  first,
+						  target,
 						  &num_clusters,
 						  e_cpos,
-						  p_blkno,
-						  &extend);
+						  &extend,
+						  ctxt);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 	}
 
-	if (extend)
+	if (extend) {
 		ret = ocfs2_extend_xattr_bucket(inode,
-						first_bh,
-						header_bh,
+						ctxt->handle,
+						first,
+						bucket_blkno(target),
 						num_clusters);
-	if (ret)
-		mlog_errno(ret);
+		if (ret)
+			mlog_errno(ret);
+	}
+
 out:
-	brelse(first_bh);
-	brelse(header_bh);
+	ocfs2_xattr_bucket_free(first);
+
 	return ret;
 }
 
@@ -3929,7 +4497,7 @@ static inline char *ocfs2_xattr_bucket_get_val(struct inode *inode,
 	int block_off = offs >> inode->i_sb->s_blocksize_bits;
 
 	offs = offs % inode->i_sb->s_blocksize;
-	return bucket->bhs[block_off]->b_data + offs;
+	return bucket_block(bucket, block_off) + offs;
 }
 
 /*
@@ -3984,7 +4552,7 @@ static void ocfs2_xattr_set_entry_normal(struct inode *inode,
 				xe->xe_value_size = 0;
 
 			val = ocfs2_xattr_bucket_get_val(inode,
-							 &xs->bucket, offs);
+							 xs->bucket, offs);
 			memset(val + OCFS2_XATTR_SIZE(name_len), 0,
 			       size - OCFS2_XATTR_SIZE(name_len));
 			if (OCFS2_XATTR_SIZE(xi->value_len) > 0)
@@ -4062,8 +4630,7 @@ set_new_name_value:
 		xh->xh_free_start = cpu_to_le16(offs);
 	}
 
-	val = ocfs2_xattr_bucket_get_val(inode,
-					 &xs->bucket, offs - size);
+	val = ocfs2_xattr_bucket_get_val(inode, xs->bucket, offs - size);
 	xe->xe_name_offset = cpu_to_le16(offs - size);
 
 	memset(val, 0, size);
@@ -4079,125 +4646,45 @@ set_new_name_value:
 	return;
 }
 
-static int ocfs2_xattr_bucket_handle_journal(struct inode *inode,
-					     handle_t *handle,
-					     struct ocfs2_xattr_search *xs,
-					     struct buffer_head **bhs,
-					     u16 bh_num)
-{
-	int ret = 0, off, block_off;
-	struct ocfs2_xattr_entry *xe = xs->here;
-
-	/*
-	 * First calculate all the blocks we should journal_access
-	 * and journal_dirty. The first block should always be touched.
-	 */
-	ret = ocfs2_journal_dirty(handle, bhs[0]);
-	if (ret)
-		mlog_errno(ret);
-
-	/* calc the data. */
-	off = le16_to_cpu(xe->xe_name_offset);
-	block_off = off >> inode->i_sb->s_blocksize_bits;
-	ret = ocfs2_journal_dirty(handle, bhs[block_off]);
-	if (ret)
-		mlog_errno(ret);
-
-	return ret;
-}
-
 /*
  * Set the xattr entry in the specified bucket.
  * The bucket is indicated by xs->bucket and it should have the enough
  * space for the xattr insertion.
  */
 static int ocfs2_xattr_set_entry_in_bucket(struct inode *inode,
+					   handle_t *handle,
 					   struct ocfs2_xattr_info *xi,
 					   struct ocfs2_xattr_search *xs,
 					   u32 name_hash,
 					   int local)
 {
-	int i, ret;
-	handle_t *handle = NULL;
-	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	int ret;
+	u64 blkno;
 
 	mlog(0, "Set xattr entry len = %lu index = %d in bucket %llu\n",
 	     (unsigned long)xi->value_len, xi->name_index,
-	     (unsigned long long)xs->bucket.bhs[0]->b_blocknr);
+	     (unsigned long long)bucket_blkno(xs->bucket));
 
-	if (!xs->bucket.bhs[1]) {
-		ret = ocfs2_read_blocks(inode,
-					xs->bucket.bhs[0]->b_blocknr + 1,
-					blk_per_bucket - 1, &xs->bucket.bhs[1],
-					0);
+	if (!xs->bucket->bu_bhs[1]) {
+		blkno = bucket_blkno(xs->bucket);
+		ocfs2_xattr_bucket_relse(xs->bucket);
+		ret = ocfs2_read_xattr_bucket(xs->bucket, blkno);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, blk_per_bucket);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		handle = NULL;
+	ret = ocfs2_xattr_bucket_journal_access(handle, xs->bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	for (i = 0; i < blk_per_bucket; i++) {
-		ret = ocfs2_journal_access(handle, inode, xs->bucket.bhs[i],
-					   OCFS2_JOURNAL_ACCESS_WRITE);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto out;
-		}
-	}
-
 	ocfs2_xattr_set_entry_normal(inode, xi, xs, name_hash, local);
+	ocfs2_xattr_bucket_journal_dirty(handle, xs->bucket);
 
-	/*Only dirty the blocks we have touched in set xattr. */
-	ret = ocfs2_xattr_bucket_handle_journal(inode, handle, xs,
-						xs->bucket.bhs, blk_per_bucket);
-	if (ret)
-		mlog_errno(ret);
-out:
-	ocfs2_commit_trans(osb, handle);
-
-	return ret;
-}
-
-static int ocfs2_xattr_value_update_size(struct inode *inode,
-					 struct buffer_head *xe_bh,
-					 struct ocfs2_xattr_entry *xe,
-					 u64 new_size)
-{
-	int ret;
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
-	handle_t *handle = NULL;
-
-	handle = ocfs2_start_trans(osb, 1);
-	if (IS_ERR(handle)) {
-		ret = -ENOMEM;
-		mlog_errno(ret);
-		goto out;
-	}
-
-	ret = ocfs2_journal_access(handle, inode, xe_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
-	if (ret < 0) {
-		mlog_errno(ret);
-		goto out_commit;
-	}
-
-	xe->xe_value_size = cpu_to_le64(new_size);
-
-	ret = ocfs2_journal_dirty(handle, xe_bh);
-	if (ret < 0)
-		mlog_errno(ret);
-
-out_commit:
-	ocfs2_commit_trans(osb, handle);
 out:
 	return ret;
 }
@@ -4210,18 +4697,19 @@ out:
  * Copy the new updated xe and xe_value_root to new_xe and new_xv if needed.
  */
 static int ocfs2_xattr_bucket_value_truncate(struct inode *inode,
-					     struct buffer_head *header_bh,
+					     struct ocfs2_xattr_bucket *bucket,
 					     int xe_off,
-					     int len)
+					     int len,
+					     struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret, offset;
 	u64 value_blk;
-	struct buffer_head *value_bh = NULL;
-	struct ocfs2_xattr_value_root *xv;
 	struct ocfs2_xattr_entry *xe;
-	struct ocfs2_xattr_header *xh =
-			(struct ocfs2_xattr_header *)header_bh->b_data;
+	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
 	size_t blocksize = inode->i_sb->s_blocksize;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_access = ocfs2_journal_access,
+	};
 
 	xe = &xh->xh_entries[xe_off];
 
@@ -4234,49 +4722,58 @@ static int ocfs2_xattr_bucket_value_truncate(struct inode *inode,
 
 	/* We don't allow ocfs2_xattr_value to be stored in different block. */
 	BUG_ON(value_blk != (offset + OCFS2_XATTR_ROOT_SIZE - 1) / blocksize);
-	value_blk += header_bh->b_blocknr;
 
-	ret = ocfs2_read_block(inode, value_blk, &value_bh);
-	if (ret) {
-		mlog_errno(ret);
-		goto out;
-	}
+	vb.vb_bh = bucket->bu_bhs[value_blk];
+	BUG_ON(!vb.vb_bh);
 
-	xv = (struct ocfs2_xattr_value_root *)
-		(value_bh->b_data + offset % blocksize);
+	vb.vb_xv = (struct ocfs2_xattr_value_root *)
+		(vb.vb_bh->b_data + offset % blocksize);
 
-	mlog(0, "truncate %u in xattr bucket %llu to %d bytes.\n",
-	     xe_off, (unsigned long long)header_bh->b_blocknr, len);
-	ret = ocfs2_xattr_value_truncate(inode, value_bh, xv, len);
+	ret = ocfs2_xattr_bucket_journal_access(ctxt->handle, bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_xattr_value_update_size(inode, header_bh, xe, len);
+	/*
+	 * From here on out we have to dirty the bucket.  The generic
+	 * value calls only modify one of the bucket's bhs, but we need
+	 * to send the bucket at once.  So if they error, they *could* have
+	 * modified something.  We have to assume they did, and dirty
+	 * the whole bucket.  This leaves us in a consistent state.
+	 */
+	mlog(0, "truncate %u in xattr bucket %llu to %d bytes.\n",
+	     xe_off, (unsigned long long)bucket_blkno(bucket), len);
+	ret = ocfs2_xattr_value_truncate(inode, &vb, len, ctxt);
 	if (ret) {
 		mlog_errno(ret);
-		goto out;
+		goto out_dirty;
 	}
 
+	xe->xe_value_size = cpu_to_le64(len);
+
+out_dirty:
+	ocfs2_xattr_bucket_journal_dirty(ctxt->handle, bucket);
+
 out:
-	brelse(value_bh);
 	return ret;
 }
 
 static int ocfs2_xattr_bucket_value_truncate_xs(struct inode *inode,
-						struct ocfs2_xattr_search *xs,
-						int len)
+					struct ocfs2_xattr_search *xs,
+					int len,
+					struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret, offset;
 	struct ocfs2_xattr_entry *xe = xs->here;
 	struct ocfs2_xattr_header *xh = (struct ocfs2_xattr_header *)xs->base;
 
-	BUG_ON(!xs->bucket.bhs[0] || !xe || ocfs2_xattr_is_local(xe));
+	BUG_ON(!xs->bucket->bu_bhs[0] || !xe || ocfs2_xattr_is_local(xe));
 
 	offset = xe - xh->xh_entries;
-	ret = ocfs2_xattr_bucket_value_truncate(inode, xs->bucket.bhs[0],
-						offset, len);
+	ret = ocfs2_xattr_bucket_value_truncate(inode, xs->bucket,
+						offset, len, ctxt);
 	if (ret)
 		mlog_errno(ret);
 
@@ -4284,6 +4781,7 @@ static int ocfs2_xattr_bucket_value_truncate_xs(struct inode *inode,
 }
 
 static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode,
+						handle_t *handle,
 						struct ocfs2_xattr_search *xs,
 						char *val,
 						int value_len)
@@ -4299,7 +4797,8 @@ static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode,
 
 	xv = (struct ocfs2_xattr_value_root *)(xs->base + offset);
 
-	return __ocfs2_xattr_set_value_outside(inode, xv, val, value_len);
+	return __ocfs2_xattr_set_value_outside(inode, handle,
+					       xv, val, value_len);
 }
 
 static int ocfs2_rm_xattr_cluster(struct inode *inode,
@@ -4343,15 +4842,15 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode,
 		}
 	}
 
-	handle = ocfs2_start_trans(osb, OCFS2_REMOVE_EXTENT_CREDITS);
+	handle = ocfs2_start_trans(osb, ocfs2_remove_extent_credits(osb->sb));
 	if (IS_ERR(handle)) {
 		ret = -ENOMEM;
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access(handle, inode, root_bh,
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_xb(handle, inode, root_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -4392,26 +4891,19 @@ out:
 }
 
 static void ocfs2_xattr_bucket_remove_xs(struct inode *inode,
+					 handle_t *handle,
 					 struct ocfs2_xattr_search *xs)
 {
-	handle_t *handle = NULL;
-	struct ocfs2_xattr_header *xh = xs->bucket.xh;
+	struct ocfs2_xattr_header *xh = bucket_xh(xs->bucket);
 	struct ocfs2_xattr_entry *last = &xh->xh_entries[
 						le16_to_cpu(xh->xh_count) - 1];
 	int ret = 0;
 
-	handle = ocfs2_start_trans((OCFS2_SB(inode->i_sb)), 1);
-	if (IS_ERR(handle)) {
-		ret = PTR_ERR(handle);
-		mlog_errno(ret);
-		return;
-	}
-
-	ret = ocfs2_journal_access(handle, inode, xs->bucket.bhs[0],
-				   OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_xattr_bucket_journal_access(handle, xs->bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
-		goto out_commit;
+		return;
 	}
 
 	/* Remove the old entry. */
@@ -4420,11 +4912,7 @@ static void ocfs2_xattr_bucket_remove_xs(struct inode *inode,
 	memset(last, 0, sizeof(struct ocfs2_xattr_entry));
 	le16_add_cpu(&xh->xh_count, -1);
 
-	ret = ocfs2_journal_dirty(handle, xs->bucket.bhs[0]);
-	if (ret < 0)
-		mlog_errno(ret);
-out_commit:
-	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
+	ocfs2_xattr_bucket_journal_dirty(handle, xs->bucket);
 }
 
 /*
@@ -4440,7 +4928,8 @@ out_commit:
  */
 static int ocfs2_xattr_set_in_bucket(struct inode *inode,
 				     struct ocfs2_xattr_info *xi,
-				     struct ocfs2_xattr_search *xs)
+				     struct ocfs2_xattr_search *xs,
+				     struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret, local = 1;
 	size_t value_len;
@@ -4468,7 +4957,8 @@ static int ocfs2_xattr_set_in_bucket(struct inode *inode,
 			value_len = 0;
 
 		ret = ocfs2_xattr_bucket_value_truncate_xs(inode, xs,
-							   value_len);
+							   value_len,
+							   ctxt);
 		if (ret)
 			goto out;
 
@@ -4488,7 +4978,8 @@ static int ocfs2_xattr_set_in_bucket(struct inode *inode,
 		xi->value_len = OCFS2_XATTR_ROOT_SIZE;
 	}
 
-	ret = ocfs2_xattr_set_entry_in_bucket(inode, xi, xs, name_hash, local);
+	ret = ocfs2_xattr_set_entry_in_bucket(inode, ctxt->handle, xi, xs,
+					      name_hash, local);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -4499,7 +4990,7 @@ static int ocfs2_xattr_set_in_bucket(struct inode *inode,
 
 	/* allocate the space now for the outside block storage. */
 	ret = ocfs2_xattr_bucket_value_truncate_xs(inode, xs,
-						   value_len);
+						   value_len, ctxt);
 	if (ret) {
 		mlog_errno(ret);
 
@@ -4509,13 +5000,14 @@ static int ocfs2_xattr_set_in_bucket(struct inode *inode,
 			 * storage and we have allocated xattr already,
 			 * so need to remove it.
 			 */
-			ocfs2_xattr_bucket_remove_xs(inode, xs);
+			ocfs2_xattr_bucket_remove_xs(inode, ctxt->handle, xs);
 		}
 		goto out;
 	}
 
 set_value_outside:
-	ret = ocfs2_xattr_bucket_set_value_outside(inode, xs, val, value_len);
+	ret = ocfs2_xattr_bucket_set_value_outside(inode, ctxt->handle,
+						   xs, val, value_len);
 out:
 	return ret;
 }
@@ -4530,7 +5022,7 @@ static int ocfs2_check_xattr_bucket_collision(struct inode *inode,
 					      struct ocfs2_xattr_bucket *bucket,
 					      const char *name)
 {
-	struct ocfs2_xattr_header *xh = bucket->xh;
+	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
 	u32 name_hash = ocfs2_xattr_name_hash(inode, name, strlen(name));
 
 	if (name_hash != le32_to_cpu(xh->xh_entries[0].xe_name_hash))
@@ -4540,7 +5032,7 @@ static int ocfs2_check_xattr_bucket_collision(struct inode *inode,
 	    xh->xh_entries[0].xe_name_hash) {
 		mlog(ML_ERROR, "Too much hash collision in xattr bucket %llu, "
 		     "hash = %u\n",
-		     (unsigned long long)bucket->bhs[0]->b_blocknr,
+		     (unsigned long long)bucket_blkno(bucket),
 		     le32_to_cpu(xh->xh_entries[0].xe_name_hash));
 		return -ENOSPC;
 	}
@@ -4550,16 +5042,16 @@ static int ocfs2_check_xattr_bucket_collision(struct inode *inode,
 
 static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
 					     struct ocfs2_xattr_info *xi,
-					     struct ocfs2_xattr_search *xs)
+					     struct ocfs2_xattr_search *xs,
+					     struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	struct ocfs2_xattr_header *xh;
 	struct ocfs2_xattr_entry *xe;
 	u16 count, header_size, xh_free_start;
-	int i, free, max_free, need, old;
+	int free, max_free, need, old;
 	size_t value_size = 0, name_len = strlen(xi->name);
 	size_t blocksize = inode->i_sb->s_blocksize;
 	int ret, allocation = 0;
-	u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
 
 	mlog_entry("Set xattr %s in xattr index block\n", xi->name);
 
@@ -4574,7 +5066,7 @@ try_again:
 
 	mlog_bug_on_msg(header_size > blocksize, "bucket %llu has header size "
 			"of %u which exceed block size\n",
-			(unsigned long long)xs->bucket.bhs[0]->b_blocknr,
+			(unsigned long long)bucket_blkno(xs->bucket),
 			header_size);
 
 	if (xi->value && xi->value_len > OCFS2_XATTR_INLINE_SIZE)
@@ -4614,11 +5106,13 @@ try_again:
 	mlog(0, "xs->not_found = %d, in xattr bucket %llu: free = %d, "
 	     "need = %d, max_free = %d, xh_free_start = %u, xh_name_value_len ="
 	     " %u\n", xs->not_found,
-	     (unsigned long long)xs->bucket.bhs[0]->b_blocknr,
+	     (unsigned long long)bucket_blkno(xs->bucket),
 	     free, need, max_free, le16_to_cpu(xh->xh_free_start),
 	     le16_to_cpu(xh->xh_name_value_len));
 
-	if (free < need || count == ocfs2_xattr_max_xe_in_bucket(inode->i_sb)) {
+	if (free < need ||
+	    (xs->not_found &&
+	     count == ocfs2_xattr_max_xe_in_bucket(inode->i_sb))) {
 		if (need <= max_free &&
 		    count < ocfs2_xattr_max_xe_in_bucket(inode->i_sb)) {
 			/*
@@ -4626,7 +5120,8 @@ try_again:
 			 * name/value will be moved, the xe shouldn't be changed
 			 * in xs.
 			 */
-			ret = ocfs2_defrag_xattr_bucket(inode, &xs->bucket);
+			ret = ocfs2_defrag_xattr_bucket(inode, ctxt->handle,
+							xs->bucket);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -4658,7 +5153,7 @@ try_again:
 		 * add a new bucket for the insert.
 		 */
 		ret = ocfs2_check_xattr_bucket_collision(inode,
-							 &xs->bucket,
+							 xs->bucket,
 							 xi->name);
 		if (ret) {
 			mlog_errno(ret);
@@ -4667,17 +5162,21 @@ try_again:
 
 		ret = ocfs2_add_new_xattr_bucket(inode,
 						 xs->xattr_bh,
-						 xs->bucket.bhs[0]);
+						 xs->bucket,
+						 ctxt);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		for (i = 0; i < blk_per_bucket; i++)
-			brelse(xs->bucket.bhs[i]);
-
-		memset(&xs->bucket, 0, sizeof(xs->bucket));
-
+		/*
+		 * ocfs2_add_new_xattr_bucket() will have updated
+		 * xs->bucket if it moved, but it will not have updated
+		 * any of the other search fields.  Thus, we drop it and
+		 * re-search.  Everything should be cached, so it'll be
+		 * quick.
+		 */
+		ocfs2_xattr_bucket_relse(xs->bucket);
 		ret = ocfs2_xattr_index_block_find(inode, xs->xattr_bh,
 						   xi->name_index,
 						   xi->name, xs);
@@ -4689,7 +5188,7 @@ try_again:
 	}
 
 xattr_set:
-	ret = ocfs2_xattr_set_in_bucket(inode, xi, xs);
+	ret = ocfs2_xattr_set_in_bucket(inode, xi, xs, ctxt);
 out:
 	mlog_exit(ret);
 	return ret;
@@ -4700,24 +5199,41 @@ static int ocfs2_delete_xattr_in_bucket(struct inode *inode,
 					void *para)
 {
 	int ret = 0;
-	struct ocfs2_xattr_header *xh = bucket->xh;
+	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
 	u16 i;
 	struct ocfs2_xattr_entry *xe;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_xattr_set_ctxt ctxt = {NULL, NULL,};
+	int credits = ocfs2_remove_extent_credits(osb->sb) +
+		ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+
+
+	ocfs2_init_dealloc_ctxt(&ctxt.dealloc);
 
 	for (i = 0; i < le16_to_cpu(xh->xh_count); i++) {
 		xe = &xh->xh_entries[i];
 		if (ocfs2_xattr_is_local(xe))
 			continue;
 
-		ret = ocfs2_xattr_bucket_value_truncate(inode,
-							bucket->bhs[0],
-							i, 0);
+		ctxt.handle = ocfs2_start_trans(osb, credits);
+		if (IS_ERR(ctxt.handle)) {
+			ret = PTR_ERR(ctxt.handle);
+			mlog_errno(ret);
+			break;
+		}
+
+		ret = ocfs2_xattr_bucket_value_truncate(inode, bucket,
+							i, 0, &ctxt);
+
+		ocfs2_commit_trans(osb, ctxt.handle);
 		if (ret) {
 			mlog_errno(ret);
 			break;
 		}
 	}
 
+	ocfs2_schedule_truncate_log_flush(osb, 1);
+	ocfs2_run_deallocs(osb, &ctxt.dealloc);
 	return ret;
 }
 
@@ -4768,6 +5284,74 @@ out:
 }
 
 /*
+ * 'security' attributes support
+ */
+static size_t ocfs2_xattr_security_list(struct inode *inode, char *list,
+					size_t list_size, const char *name,
+					size_t name_len)
+{
+	const size_t prefix_len = XATTR_SECURITY_PREFIX_LEN;
+	const size_t total_len = prefix_len + name_len + 1;
+
+	if (list && total_len <= list_size) {
+		memcpy(list, XATTR_SECURITY_PREFIX, prefix_len);
+		memcpy(list + prefix_len, name, name_len);
+		list[prefix_len + name_len] = '\0';
+	}
+	return total_len;
+}
+
+static int ocfs2_xattr_security_get(struct inode *inode, const char *name,
+				    void *buffer, size_t size)
+{
+	if (strcmp(name, "") == 0)
+		return -EINVAL;
+	return ocfs2_xattr_get(inode, OCFS2_XATTR_INDEX_SECURITY, name,
+			       buffer, size);
+}
+
+static int ocfs2_xattr_security_set(struct inode *inode, const char *name,
+				    const void *value, size_t size, int flags)
+{
+	if (strcmp(name, "") == 0)
+		return -EINVAL;
+
+	return ocfs2_xattr_set(inode, OCFS2_XATTR_INDEX_SECURITY, name, value,
+			       size, flags);
+}
+
+int ocfs2_init_security_get(struct inode *inode,
+			    struct inode *dir,
+			    struct ocfs2_security_xattr_info *si)
+{
+	/* check whether ocfs2 support feature xattr */
+	if (!ocfs2_supports_xattr(OCFS2_SB(dir->i_sb)))
+		return -EOPNOTSUPP;
+	return security_inode_init_security(inode, dir, &si->name, &si->value,
+					    &si->value_len);
+}
+
+int ocfs2_init_security_set(handle_t *handle,
+			    struct inode *inode,
+			    struct buffer_head *di_bh,
+			    struct ocfs2_security_xattr_info *si,
+			    struct ocfs2_alloc_context *xattr_ac,
+			    struct ocfs2_alloc_context *data_ac)
+{
+	return ocfs2_xattr_set_handle(handle, inode, di_bh,
+				     OCFS2_XATTR_INDEX_SECURITY,
+				     si->name, si->value, si->value_len, 0,
+				     xattr_ac, data_ac);
+}
+
+struct xattr_handler ocfs2_xattr_security_handler = {
+	.prefix	= XATTR_SECURITY_PREFIX,
+	.list	= ocfs2_xattr_security_list,
+	.get	= ocfs2_xattr_security_get,
+	.set	= ocfs2_xattr_security_set,
+};
+
+/*
  * 'trusted' attributes support
  */
 static size_t ocfs2_xattr_trusted_list(struct inode *inode, char *list,
diff --git a/fs/ocfs2/xattr.h b/fs/ocfs2/xattr.h
index 1d8314c7656..5a1ebc789f7 100644
--- a/fs/ocfs2/xattr.h
+++ b/fs/ocfs2/xattr.h
@@ -30,13 +30,58 @@ enum ocfs2_xattr_type {
 	OCFS2_XATTR_MAX
 };
 
+struct ocfs2_security_xattr_info {
+	int enable;
+	char *name;
+	void *value;
+	size_t value_len;
+};
+
 extern struct xattr_handler ocfs2_xattr_user_handler;
 extern struct xattr_handler ocfs2_xattr_trusted_handler;
+extern struct xattr_handler ocfs2_xattr_security_handler;
+#ifdef CONFIG_OCFS2_FS_POSIX_ACL
+extern struct xattr_handler ocfs2_xattr_acl_access_handler;
+extern struct xattr_handler ocfs2_xattr_acl_default_handler;
+#endif
 extern struct xattr_handler *ocfs2_xattr_handlers[];
 
 ssize_t ocfs2_listxattr(struct dentry *, char *, size_t);
+int ocfs2_xattr_get_nolock(struct inode *, struct buffer_head *, int,
+			   const char *, void *, size_t);
 int ocfs2_xattr_set(struct inode *, int, const char *, const void *,
 		    size_t, int);
+int ocfs2_xattr_set_handle(handle_t *, struct inode *, struct buffer_head *,
+			   int, const char *, const void *, size_t, int,
+			   struct ocfs2_alloc_context *,
+			   struct ocfs2_alloc_context *);
 int ocfs2_xattr_remove(struct inode *, struct buffer_head *);
+int ocfs2_init_security_get(struct inode *, struct inode *,
+			    struct ocfs2_security_xattr_info *);
+int ocfs2_init_security_set(handle_t *, struct inode *,
+			    struct buffer_head *,
+			    struct ocfs2_security_xattr_info *,
+			    struct ocfs2_alloc_context *,
+			    struct ocfs2_alloc_context *);
+int ocfs2_calc_security_init(struct inode *,
+			     struct ocfs2_security_xattr_info *,
+			     int *, int *, struct ocfs2_alloc_context **);
+int ocfs2_calc_xattr_init(struct inode *, struct buffer_head *,
+			  int, struct ocfs2_security_xattr_info *,
+			  int *, int *, struct ocfs2_alloc_context **);
+
+/*
+ * xattrs can live inside an inode, as part of an external xattr block,
+ * or inside an xattr bucket, which is the leaf of a tree rooted in an
+ * xattr block.  Some of the xattr calls, especially the value setting
+ * functions, want to treat each of these locations as equal.  Let's wrap
+ * them in a structure that we can pass around instead of raw buffer_heads.
+ */
+struct ocfs2_xattr_value_buf {
+	struct buffer_head		*vb_bh;
+	ocfs2_journal_access_func	vb_access;
+	struct ocfs2_xattr_value_root	*vb_xv;
+};
+
 
 #endif /* OCFS2_XATTR_H */
diff --git a/fs/omfs/inode.c b/fs/omfs/inode.c
index 6afe57c84f8..633e9dc972b 100644
--- a/fs/omfs/inode.c
+++ b/fs/omfs/inode.c
@@ -39,7 +39,6 @@ struct inode *omfs_new_inode(struct inode *dir, int mode)
 	inode->i_mode = mode;
 	inode->i_uid = current_fsuid();
 	inode->i_gid = current_fsgid();
-	inode->i_blocks = 0;
 	inode->i_mapping->a_ops = &omfs_aops;
 
 	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
diff --git a/fs/open.c b/fs/open.c
index c0a426d5766..d882fd2351d 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -272,6 +272,8 @@ static long do_sys_truncate(const char __user *pathname, loff_t length)
 		goto put_write_and_out;
 
 	error = locks_verify_truncate(inode, NULL, length);
+	if (!error)
+		error = security_path_truncate(&path, length, 0);
 	if (!error) {
 		DQUOT_INIT(inode);
 		error = do_truncate(path.dentry, length, 0, NULL);
@@ -329,6 +331,9 @@ static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
 
 	error = locks_verify_truncate(inode, file, length);
 	if (!error)
+		error = security_path_truncate(&file->f_path, length,
+					       ATTR_MTIME|ATTR_CTIME);
+	if (!error)
 		error = do_truncate(dentry, length, ATTR_MTIME|ATTR_CTIME, file);
 out_putf:
 	fput(file);
@@ -407,7 +412,7 @@ asmlinkage long sys_fallocate(int fd, int mode, loff_t offset, loff_t len)
 	if (((offset + len) > inode->i_sb->s_maxbytes) || ((offset + len) < 0))
 		goto out_fput;
 
-	if (inode->i_op && inode->i_op->fallocate)
+	if (inode->i_op->fallocate)
 		ret = inode->i_op->fallocate(inode, mode, offset, len);
 	else
 		ret = -EOPNOTSUPP;
diff --git a/fs/openpromfs/inode.c b/fs/openpromfs/inode.c
index d41bdc784de..ffcd04f0012 100644
--- a/fs/openpromfs/inode.c
+++ b/fs/openpromfs/inode.c
@@ -256,9 +256,6 @@ found:
 		break;
 	}
 
-	inode->i_gid = 0;
-	inode->i_uid = 0;
-
 	d_add(dentry, inode);
 	return NULL;
 }
diff --git a/fs/partitions/check.c b/fs/partitions/check.c
index 6d5b213b8a9..6d720243f5f 100644
--- a/fs/partitions/check.c
+++ b/fs/partitions/check.c
@@ -334,6 +334,7 @@ void delete_partition(struct gendisk *disk, int partno)
 
 	blk_free_devt(part_devt(part));
 	rcu_assign_pointer(ptbl->part[partno], NULL);
+	rcu_assign_pointer(ptbl->last_lookup, NULL);
 	kobject_put(part->holder_dir);
 	device_del(part_to_dev(part));
 
@@ -384,9 +385,9 @@ struct hd_struct *add_partition(struct gendisk *disk, int partno,
 
 	dname = dev_name(ddev);
 	if (isdigit(dname[strlen(dname) - 1]))
-		snprintf(pdev->bus_id, BUS_ID_SIZE, "%sp%d", dname, partno);
+		dev_set_name(pdev, "%sp%d", dname, partno);
 	else
-		snprintf(pdev->bus_id, BUS_ID_SIZE, "%s%d", dname, partno);
+		dev_set_name(pdev, "%s%d", dname, partno);
 
 	device_initialize(pdev);
 	pdev->class = &block_class;
@@ -447,16 +448,11 @@ void register_disk(struct gendisk *disk)
 	struct block_device *bdev;
 	struct disk_part_iter piter;
 	struct hd_struct *part;
-	char *s;
 	int err;
 
 	ddev->parent = disk->driverfs_dev;
 
-	strlcpy(ddev->bus_id, disk->disk_name, BUS_ID_SIZE);
-	/* ewww... some of these buggers have / in the name... */
-	s = strchr(ddev->bus_id, '/');
-	if (s)
-		*s = '!';
+	dev_set_name(ddev, disk->disk_name);
 
 	/* delay uevents, until we scanned partition table */
 	ddev->uevent_suppress = 1;
diff --git a/fs/pipe.c b/fs/pipe.c
index aaf797bd57b..891697112f6 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -1016,10 +1016,7 @@ int do_pipe_flags(int *fd, int flags)
 		goto err_fdr;
 	fdw = error;
 
-	error = audit_fd_pair(fdr, fdw);
-	if (error < 0)
-		goto err_fdw;
-
+	audit_fd_pair(fdr, fdw);
 	fd_install(fdr, fr);
 	fd_install(fdw, fw);
 	fd[0] = fdr;
@@ -1027,8 +1024,6 @@ int do_pipe_flags(int *fd, int flags)
 
 	return 0;
 
- err_fdw:
-	put_unused_fd(fdw);
  err_fdr:
 	put_unused_fd(fdr);
  err_read_pipe:
diff --git a/fs/proc/base.c b/fs/proc/base.c
index cad92c1ac2b..0c9de19a163 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -65,6 +65,7 @@
 #include <linux/mm.h>
 #include <linux/rcupdate.h>
 #include <linux/kallsyms.h>
+#include <linux/stacktrace.h>
 #include <linux/resource.h>
 #include <linux/module.h>
 #include <linux/mount.h>
@@ -109,25 +110,22 @@ struct pid_entry {
 	.op   = OP,					\
 }
 
-#define DIR(NAME, MODE, OTYPE)							\
-	NOD(NAME, (S_IFDIR|(MODE)),						\
-		&proc_##OTYPE##_inode_operations, &proc_##OTYPE##_operations,	\
-		{} )
-#define LNK(NAME, OTYPE)					\
+#define DIR(NAME, MODE, iops, fops)	\
+	NOD(NAME, (S_IFDIR|(MODE)), &iops, &fops, {} )
+#define LNK(NAME, get_link)					\
 	NOD(NAME, (S_IFLNK|S_IRWXUGO),				\
 		&proc_pid_link_inode_operations, NULL,		\
-		{ .proc_get_link = &proc_##OTYPE##_link } )
-#define REG(NAME, MODE, OTYPE)				\
-	NOD(NAME, (S_IFREG|(MODE)), NULL,		\
-		&proc_##OTYPE##_operations, {})
-#define INF(NAME, MODE, OTYPE)				\
+		{ .proc_get_link = get_link } )
+#define REG(NAME, MODE, fops)				\
+	NOD(NAME, (S_IFREG|(MODE)), NULL, &fops, {})
+#define INF(NAME, MODE, read)				\
 	NOD(NAME, (S_IFREG|(MODE)), 			\
 		NULL, &proc_info_file_operations,	\
-		{ .proc_read = &proc_##OTYPE } )
-#define ONE(NAME, MODE, OTYPE)				\
+		{ .proc_read = read } )
+#define ONE(NAME, MODE, show)				\
 	NOD(NAME, (S_IFREG|(MODE)), 			\
 		NULL, &proc_single_file_operations,	\
-		{ .proc_show = &proc_##OTYPE } )
+		{ .proc_show = show } )
 
 /*
  * Count the number of hardlinks for the pid_entry table, excluding the .
@@ -308,9 +306,9 @@ static int proc_pid_auxv(struct task_struct *task, char *buffer)
 	struct mm_struct *mm = get_task_mm(task);
 	if (mm) {
 		unsigned int nwords = 0;
-		do
+		do {
 			nwords += 2;
-		while (mm->saved_auxv[nwords - 2] != 0); /* AT_NULL */
+		} while (mm->saved_auxv[nwords - 2] != 0); /* AT_NULL */
 		res = nwords * sizeof(mm->saved_auxv[0]);
 		if (res > PAGE_SIZE)
 			res = PAGE_SIZE;
@@ -340,6 +338,37 @@ static int proc_pid_wchan(struct task_struct *task, char *buffer)
 }
 #endif /* CONFIG_KALLSYMS */
 
+#ifdef CONFIG_STACKTRACE
+
+#define MAX_STACK_TRACE_DEPTH	64
+
+static int proc_pid_stack(struct seq_file *m, struct pid_namespace *ns,
+			  struct pid *pid, struct task_struct *task)
+{
+	struct stack_trace trace;
+	unsigned long *entries;
+	int i;
+
+	entries = kmalloc(MAX_STACK_TRACE_DEPTH * sizeof(*entries), GFP_KERNEL);
+	if (!entries)
+		return -ENOMEM;
+
+	trace.nr_entries	= 0;
+	trace.max_entries	= MAX_STACK_TRACE_DEPTH;
+	trace.entries		= entries;
+	trace.skip		= 0;
+	save_stack_trace_tsk(task, &trace);
+
+	for (i = 0; i < trace.nr_entries; i++) {
+		seq_printf(m, "[<%p>] %pS\n",
+			   (void *)entries[i], (void *)entries[i]);
+	}
+	kfree(entries);
+
+	return 0;
+}
+#endif
+
 #ifdef CONFIG_SCHEDSTATS
 /*
  * Provides /proc/PID/schedstat
@@ -1186,8 +1215,6 @@ static int sched_show(struct seq_file *m, void *v)
 	struct inode *inode = m->private;
 	struct task_struct *p;
 
-	WARN_ON(!inode);
-
 	p = get_proc_task(inode);
 	if (!p)
 		return -ESRCH;
@@ -1205,8 +1232,6 @@ sched_write(struct file *file, const char __user *buf,
 	struct inode *inode = file->f_path.dentry->d_inode;
 	struct task_struct *p;
 
-	WARN_ON(!inode);
-
 	p = get_proc_task(inode);
 	if (!p)
 		return -ESRCH;
@@ -1426,8 +1451,6 @@ static struct inode *proc_pid_make_inode(struct super_block * sb, struct task_st
 	if (!ei->pid)
 		goto out_unlock;
 
-	inode->i_uid = 0;
-	inode->i_gid = 0;
 	if (task_dumpable(task)) {
 		rcu_read_lock();
 		cred = __task_cred(task);
@@ -1976,13 +1999,11 @@ static struct dentry *proc_pident_lookup(struct inode *dir,
 					 const struct pid_entry *ents,
 					 unsigned int nents)
 {
-	struct inode *inode;
 	struct dentry *error;
 	struct task_struct *task = get_proc_task(dir);
 	const struct pid_entry *p, *last;
 
 	error = ERR_PTR(-ENOENT);
-	inode = NULL;
 
 	if (!task)
 		goto out_no_task;
@@ -2138,12 +2159,12 @@ static const struct file_operations proc_pid_attr_operations = {
 };
 
 static const struct pid_entry attr_dir_stuff[] = {
-	REG("current",    S_IRUGO|S_IWUGO, pid_attr),
-	REG("prev",       S_IRUGO,	   pid_attr),
-	REG("exec",       S_IRUGO|S_IWUGO, pid_attr),
-	REG("fscreate",   S_IRUGO|S_IWUGO, pid_attr),
-	REG("keycreate",  S_IRUGO|S_IWUGO, pid_attr),
-	REG("sockcreate", S_IRUGO|S_IWUGO, pid_attr),
+	REG("current",    S_IRUGO|S_IWUGO, proc_pid_attr_operations),
+	REG("prev",       S_IRUGO,	   proc_pid_attr_operations),
+	REG("exec",       S_IRUGO|S_IWUGO, proc_pid_attr_operations),
+	REG("fscreate",   S_IRUGO|S_IWUGO, proc_pid_attr_operations),
+	REG("keycreate",  S_IRUGO|S_IWUGO, proc_pid_attr_operations),
+	REG("sockcreate", S_IRUGO|S_IWUGO, proc_pid_attr_operations),
 };
 
 static int proc_attr_dir_readdir(struct file * filp,
@@ -2349,8 +2370,6 @@ static struct dentry *proc_base_instantiate(struct inode *dir,
 	if (!ei->pid)
 		goto out_iput;
 
-	inode->i_uid = 0;
-	inode->i_gid = 0;
 	inode->i_mode = p->mode;
 	if (S_ISDIR(inode->i_mode))
 		inode->i_nlink = 2;
@@ -2465,74 +2484,77 @@ static const struct file_operations proc_task_operations;
 static const struct inode_operations proc_task_inode_operations;
 
 static const struct pid_entry tgid_base_stuff[] = {
-	DIR("task",       S_IRUGO|S_IXUGO, task),
-	DIR("fd",         S_IRUSR|S_IXUSR, fd),
-	DIR("fdinfo",     S_IRUSR|S_IXUSR, fdinfo),
+	DIR("task",       S_IRUGO|S_IXUGO, proc_task_inode_operations, proc_task_operations),
+	DIR("fd",         S_IRUSR|S_IXUSR, proc_fd_inode_operations, proc_fd_operations),
+	DIR("fdinfo",     S_IRUSR|S_IXUSR, proc_fdinfo_inode_operations, proc_fdinfo_operations),
 #ifdef CONFIG_NET
-	DIR("net",        S_IRUGO|S_IXUGO, net),
+	DIR("net",        S_IRUGO|S_IXUGO, proc_net_inode_operations, proc_net_operations),
 #endif
-	REG("environ",    S_IRUSR, environ),
-	INF("auxv",       S_IRUSR, pid_auxv),
-	ONE("status",     S_IRUGO, pid_status),
-	ONE("personality", S_IRUSR, pid_personality),
-	INF("limits",	  S_IRUSR, pid_limits),
+	REG("environ",    S_IRUSR, proc_environ_operations),
+	INF("auxv",       S_IRUSR, proc_pid_auxv),
+	ONE("status",     S_IRUGO, proc_pid_status),
+	ONE("personality", S_IRUSR, proc_pid_personality),
+	INF("limits",	  S_IRUSR, proc_pid_limits),
 #ifdef CONFIG_SCHED_DEBUG
-	REG("sched",      S_IRUGO|S_IWUSR, pid_sched),
+	REG("sched",      S_IRUGO|S_IWUSR, proc_pid_sched_operations),
 #endif
 #ifdef CONFIG_HAVE_ARCH_TRACEHOOK
-	INF("syscall",    S_IRUSR, pid_syscall),
+	INF("syscall",    S_IRUSR, proc_pid_syscall),
 #endif
-	INF("cmdline",    S_IRUGO, pid_cmdline),
-	ONE("stat",       S_IRUGO, tgid_stat),
-	ONE("statm",      S_IRUGO, pid_statm),
-	REG("maps",       S_IRUGO, maps),
+	INF("cmdline",    S_IRUGO, proc_pid_cmdline),
+	ONE("stat",       S_IRUGO, proc_tgid_stat),
+	ONE("statm",      S_IRUGO, proc_pid_statm),
+	REG("maps",       S_IRUGO, proc_maps_operations),
 #ifdef CONFIG_NUMA
-	REG("numa_maps",  S_IRUGO, numa_maps),
+	REG("numa_maps",  S_IRUGO, proc_numa_maps_operations),
 #endif
-	REG("mem",        S_IRUSR|S_IWUSR, mem),
-	LNK("cwd",        cwd),
-	LNK("root",       root),
-	LNK("exe",        exe),
-	REG("mounts",     S_IRUGO, mounts),
-	REG("mountinfo",  S_IRUGO, mountinfo),
-	REG("mountstats", S_IRUSR, mountstats),
+	REG("mem",        S_IRUSR|S_IWUSR, proc_mem_operations),
+	LNK("cwd",        proc_cwd_link),
+	LNK("root",       proc_root_link),
+	LNK("exe",        proc_exe_link),
+	REG("mounts",     S_IRUGO, proc_mounts_operations),
+	REG("mountinfo",  S_IRUGO, proc_mountinfo_operations),
+	REG("mountstats", S_IRUSR, proc_mountstats_operations),
 #ifdef CONFIG_PROC_PAGE_MONITOR
-	REG("clear_refs", S_IWUSR, clear_refs),
-	REG("smaps",      S_IRUGO, smaps),
-	REG("pagemap",    S_IRUSR, pagemap),
+	REG("clear_refs", S_IWUSR, proc_clear_refs_operations),
+	REG("smaps",      S_IRUGO, proc_smaps_operations),
+	REG("pagemap",    S_IRUSR, proc_pagemap_operations),
 #endif
 #ifdef CONFIG_SECURITY
-	DIR("attr",       S_IRUGO|S_IXUGO, attr_dir),
+	DIR("attr",       S_IRUGO|S_IXUGO, proc_attr_dir_inode_operations, proc_attr_dir_operations),
 #endif
 #ifdef CONFIG_KALLSYMS
-	INF("wchan",      S_IRUGO, pid_wchan),
+	INF("wchan",      S_IRUGO, proc_pid_wchan),
+#endif
+#ifdef CONFIG_STACKTRACE
+	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
 #ifdef CONFIG_SCHEDSTATS
-	INF("schedstat",  S_IRUGO, pid_schedstat),
+	INF("schedstat",  S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
-	REG("latency",  S_IRUGO, lstats),
+	REG("latency",  S_IRUGO, proc_lstats_operations),
 #endif
 #ifdef CONFIG_PROC_PID_CPUSET
-	REG("cpuset",     S_IRUGO, cpuset),
+	REG("cpuset",     S_IRUGO, proc_cpuset_operations),
 #endif
 #ifdef CONFIG_CGROUPS
-	REG("cgroup",  S_IRUGO, cgroup),
+	REG("cgroup",  S_IRUGO, proc_cgroup_operations),
 #endif
-	INF("oom_score",  S_IRUGO, oom_score),
-	REG("oom_adj",    S_IRUGO|S_IWUSR, oom_adjust),
+	INF("oom_score",  S_IRUGO, proc_oom_score),
+	REG("oom_adj",    S_IRUGO|S_IWUSR, proc_oom_adjust_operations),
 #ifdef CONFIG_AUDITSYSCALL
-	REG("loginuid",   S_IWUSR|S_IRUGO, loginuid),
-	REG("sessionid",  S_IRUGO, sessionid),
+	REG("loginuid",   S_IWUSR|S_IRUGO, proc_loginuid_operations),
+	REG("sessionid",  S_IRUGO, proc_sessionid_operations),
 #endif
 #ifdef CONFIG_FAULT_INJECTION
-	REG("make-it-fail", S_IRUGO|S_IWUSR, fault_inject),
+	REG("make-it-fail", S_IRUGO|S_IWUSR, proc_fault_inject_operations),
 #endif
 #if defined(USE_ELF_CORE_DUMP) && defined(CONFIG_ELF_CORE)
-	REG("coredump_filter", S_IRUGO|S_IWUSR, coredump_filter),
+	REG("coredump_filter", S_IRUGO|S_IWUSR, proc_coredump_filter_operations),
 #endif
 #ifdef CONFIG_TASK_IO_ACCOUNTING
-	INF("io",	S_IRUGO, tgid_io_accounting),
+	INF("io",	S_IRUGO, proc_tgid_io_accounting),
 #endif
 };
 
@@ -2805,66 +2827,69 @@ out_no_task:
  * Tasks
  */
 static const struct pid_entry tid_base_stuff[] = {
-	DIR("fd",        S_IRUSR|S_IXUSR, fd),
-	DIR("fdinfo",    S_IRUSR|S_IXUSR, fdinfo),
-	REG("environ",   S_IRUSR, environ),
-	INF("auxv",      S_IRUSR, pid_auxv),
-	ONE("status",    S_IRUGO, pid_status),
-	ONE("personality", S_IRUSR, pid_personality),
-	INF("limits",	 S_IRUSR, pid_limits),
+	DIR("fd",        S_IRUSR|S_IXUSR, proc_fd_inode_operations, proc_fd_operations),
+	DIR("fdinfo",    S_IRUSR|S_IXUSR, proc_fdinfo_inode_operations, proc_fd_operations),
+	REG("environ",   S_IRUSR, proc_environ_operations),
+	INF("auxv",      S_IRUSR, proc_pid_auxv),
+	ONE("status",    S_IRUGO, proc_pid_status),
+	ONE("personality", S_IRUSR, proc_pid_personality),
+	INF("limits",	 S_IRUSR, proc_pid_limits),
 #ifdef CONFIG_SCHED_DEBUG
-	REG("sched",     S_IRUGO|S_IWUSR, pid_sched),
+	REG("sched",     S_IRUGO|S_IWUSR, proc_pid_sched_operations),
 #endif
 #ifdef CONFIG_HAVE_ARCH_TRACEHOOK
-	INF("syscall",   S_IRUSR, pid_syscall),
+	INF("syscall",   S_IRUSR, proc_pid_syscall),
 #endif
-	INF("cmdline",   S_IRUGO, pid_cmdline),
-	ONE("stat",      S_IRUGO, tid_stat),
-	ONE("statm",     S_IRUGO, pid_statm),
-	REG("maps",      S_IRUGO, maps),
+	INF("cmdline",   S_IRUGO, proc_pid_cmdline),
+	ONE("stat",      S_IRUGO, proc_tid_stat),
+	ONE("statm",     S_IRUGO, proc_pid_statm),
+	REG("maps",      S_IRUGO, proc_maps_operations),
 #ifdef CONFIG_NUMA
-	REG("numa_maps", S_IRUGO, numa_maps),
+	REG("numa_maps", S_IRUGO, proc_numa_maps_operations),
 #endif
-	REG("mem",       S_IRUSR|S_IWUSR, mem),
-	LNK("cwd",       cwd),
-	LNK("root",      root),
-	LNK("exe",       exe),
-	REG("mounts",    S_IRUGO, mounts),
-	REG("mountinfo",  S_IRUGO, mountinfo),
+	REG("mem",       S_IRUSR|S_IWUSR, proc_mem_operations),
+	LNK("cwd",       proc_cwd_link),
+	LNK("root",      proc_root_link),
+	LNK("exe",       proc_exe_link),
+	REG("mounts",    S_IRUGO, proc_mounts_operations),
+	REG("mountinfo",  S_IRUGO, proc_mountinfo_operations),
 #ifdef CONFIG_PROC_PAGE_MONITOR
-	REG("clear_refs", S_IWUSR, clear_refs),
-	REG("smaps",     S_IRUGO, smaps),
-	REG("pagemap",    S_IRUSR, pagemap),
+	REG("clear_refs", S_IWUSR, proc_clear_refs_operations),
+	REG("smaps",     S_IRUGO, proc_smaps_operations),
+	REG("pagemap",    S_IRUSR, proc_pagemap_operations),
 #endif
 #ifdef CONFIG_SECURITY
-	DIR("attr",      S_IRUGO|S_IXUGO, attr_dir),
+	DIR("attr",      S_IRUGO|S_IXUGO, proc_attr_dir_inode_operations, proc_attr_dir_operations),
 #endif
 #ifdef CONFIG_KALLSYMS
-	INF("wchan",     S_IRUGO, pid_wchan),
+	INF("wchan",     S_IRUGO, proc_pid_wchan),
+#endif
+#ifdef CONFIG_STACKTRACE
+	ONE("stack",      S_IRUSR, proc_pid_stack),
 #endif
 #ifdef CONFIG_SCHEDSTATS
-	INF("schedstat", S_IRUGO, pid_schedstat),
+	INF("schedstat", S_IRUGO, proc_pid_schedstat),
 #endif
 #ifdef CONFIG_LATENCYTOP
-	REG("latency",  S_IRUGO, lstats),
+	REG("latency",  S_IRUGO, proc_lstats_operations),
 #endif
 #ifdef CONFIG_PROC_PID_CPUSET
-	REG("cpuset",    S_IRUGO, cpuset),
+	REG("cpuset",    S_IRUGO, proc_cpuset_operations),
 #endif
 #ifdef CONFIG_CGROUPS
-	REG("cgroup",  S_IRUGO, cgroup),
+	REG("cgroup",  S_IRUGO, proc_cgroup_operations),
 #endif
-	INF("oom_score", S_IRUGO, oom_score),
-	REG("oom_adj",   S_IRUGO|S_IWUSR, oom_adjust),
+	INF("oom_score", S_IRUGO, proc_oom_score),
+	REG("oom_adj",   S_IRUGO|S_IWUSR, proc_oom_adjust_operations),
 #ifdef CONFIG_AUDITSYSCALL
-	REG("loginuid",  S_IWUSR|S_IRUGO, loginuid),
-	REG("sessionid",  S_IRUSR, sessionid),
+	REG("loginuid",  S_IWUSR|S_IRUGO, proc_loginuid_operations),
+	REG("sessionid",  S_IRUSR, proc_sessionid_operations),
 #endif
 #ifdef CONFIG_FAULT_INJECTION
-	REG("make-it-fail", S_IRUGO|S_IWUSR, fault_inject),
+	REG("make-it-fail", S_IRUGO|S_IWUSR, proc_fault_inject_operations),
 #endif
 #ifdef CONFIG_TASK_IO_ACCOUNTING
-	INF("io",	S_IRUGO, tid_io_accounting),
+	INF("io",	S_IRUGO, proc_tid_io_accounting),
 #endif
 };
 
diff --git a/fs/proc/generic.c b/fs/proc/generic.c
index 60a359b3558..db7fa5cab98 100644
--- a/fs/proc/generic.c
+++ b/fs/proc/generic.c
@@ -14,7 +14,6 @@
 #include <linux/stat.h>
 #include <linux/module.h>
 #include <linux/mount.h>
-#include <linux/smp_lock.h>
 #include <linux/init.h>
 #include <linux/idr.h>
 #include <linux/namei.h>
@@ -379,7 +378,6 @@ struct dentry *proc_lookup_de(struct proc_dir_entry *de, struct inode *dir,
 	struct inode *inode = NULL;
 	int error = -ENOENT;
 
-	lock_kernel();
 	spin_lock(&proc_subdir_lock);
 	for (de = de->subdir; de ; de = de->next) {
 		if (de->namelen != dentry->d_name.len)
@@ -397,7 +395,6 @@ struct dentry *proc_lookup_de(struct proc_dir_entry *de, struct inode *dir,
 	}
 	spin_unlock(&proc_subdir_lock);
 out_unlock:
-	unlock_kernel();
 
 	if (inode) {
 		dentry->d_op = &proc_dentry_operations;
@@ -432,8 +429,6 @@ int proc_readdir_de(struct proc_dir_entry *de, struct file *filp, void *dirent,
 	struct inode *inode = filp->f_path.dentry->d_inode;
 	int ret = 0;
 
-	lock_kernel();
-
 	ino = inode->i_ino;
 	i = filp->f_pos;
 	switch (i) {
@@ -487,7 +482,7 @@ int proc_readdir_de(struct proc_dir_entry *de, struct file *filp, void *dirent,
 			spin_unlock(&proc_subdir_lock);
 	}
 	ret = 1;
-out:	unlock_kernel();
+out:
 	return ret;	
 }
 
@@ -504,6 +499,7 @@ int proc_readdir(struct file *filp, void *dirent, filldir_t filldir)
  * the /proc directory.
  */
 static const struct file_operations proc_dir_operations = {
+	.llseek			= generic_file_llseek,
 	.read			= generic_read_dir,
 	.readdir		= proc_readdir,
 };
diff --git a/fs/proc/inode.c b/fs/proc/inode.c
index 2543fd00c65..3e76bb9b3ad 100644
--- a/fs/proc/inode.c
+++ b/fs/proc/inode.c
@@ -35,16 +35,13 @@ struct proc_dir_entry *de_get(struct proc_dir_entry *de)
  */
 void de_put(struct proc_dir_entry *de)
 {
-	lock_kernel();
 	if (!atomic_read(&de->count)) {
 		printk("de_put: entry %s already free!\n", de->name);
-		unlock_kernel();
 		return;
 	}
 
 	if (atomic_dec_and_test(&de->count))
 		free_proc_entry(de);
-	unlock_kernel();
 }
 
 /*
diff --git a/fs/proc/internal.h b/fs/proc/internal.h
index 3e8aeb8b61c..cd53ff83849 100644
--- a/fs/proc/internal.h
+++ b/fs/proc/internal.h
@@ -41,8 +41,6 @@ do {						\
 	(vmi)->used = 0;			\
 	(vmi)->largest_chunk = 0;		\
 } while(0)
-
-extern int nommu_vma_show(struct seq_file *, struct vm_area_struct *);
 #endif
 
 extern int proc_tid_stat(struct seq_file *m, struct pid_namespace *ns,
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index b1675c4e66d..43d23948384 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -74,6 +74,9 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		"LowTotal:       %8lu kB\n"
 		"LowFree:        %8lu kB\n"
 #endif
+#ifndef CONFIG_MMU
+		"MmapCopy:       %8lu kB\n"
+#endif
 		"SwapTotal:      %8lu kB\n"
 		"SwapFree:       %8lu kB\n"
 		"Dirty:          %8lu kB\n"
@@ -116,6 +119,9 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		K(i.totalram-i.totalhigh),
 		K(i.freeram-i.freehigh),
 #endif
+#ifndef CONFIG_MMU
+		K((unsigned long) atomic_read(&mmap_pages_allocated)),
+#endif
 		K(i.totalswap),
 		K(i.freeswap),
 		K(global_page_state(NR_FILE_DIRTY)),
diff --git a/fs/proc/nommu.c b/fs/proc/nommu.c
index 3f87d263294..b446d7ad0b0 100644
--- a/fs/proc/nommu.c
+++ b/fs/proc/nommu.c
@@ -33,33 +33,33 @@
 #include "internal.h"
 
 /*
- * display a single VMA to a sequenced file
+ * display a single region to a sequenced file
  */
-int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma)
+static int nommu_region_show(struct seq_file *m, struct vm_region *region)
 {
 	unsigned long ino = 0;
 	struct file *file;
 	dev_t dev = 0;
 	int flags, len;
 
-	flags = vma->vm_flags;
-	file = vma->vm_file;
+	flags = region->vm_flags;
+	file = region->vm_file;
 
 	if (file) {
-		struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
+		struct inode *inode = region->vm_file->f_path.dentry->d_inode;
 		dev = inode->i_sb->s_dev;
 		ino = inode->i_ino;
 	}
 
 	seq_printf(m,
 		   "%08lx-%08lx %c%c%c%c %08llx %02x:%02x %lu %n",
-		   vma->vm_start,
-		   vma->vm_end,
+		   region->vm_start,
+		   region->vm_end,
 		   flags & VM_READ ? 'r' : '-',
 		   flags & VM_WRITE ? 'w' : '-',
 		   flags & VM_EXEC ? 'x' : '-',
 		   flags & VM_MAYSHARE ? flags & VM_SHARED ? 'S' : 's' : 'p',
-		   ((loff_t)vma->vm_pgoff) << PAGE_SHIFT,
+		   ((loff_t)region->vm_pgoff) << PAGE_SHIFT,
 		   MAJOR(dev), MINOR(dev), ino, &len);
 
 	if (file) {
@@ -75,61 +75,54 @@ int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma)
 }
 
 /*
- * display a list of all the VMAs the kernel knows about
+ * display a list of all the REGIONs the kernel knows about
  * - nommu kernals have a single flat list
  */
-static int nommu_vma_list_show(struct seq_file *m, void *v)
+static int nommu_region_list_show(struct seq_file *m, void *_p)
 {
-	struct vm_area_struct *vma;
+	struct rb_node *p = _p;
 
-	vma = rb_entry((struct rb_node *) v, struct vm_area_struct, vm_rb);
-	return nommu_vma_show(m, vma);
+	return nommu_region_show(m, rb_entry(p, struct vm_region, vm_rb));
 }
 
-static void *nommu_vma_list_start(struct seq_file *m, loff_t *_pos)
+static void *nommu_region_list_start(struct seq_file *m, loff_t *_pos)
 {
-	struct rb_node *_rb;
+	struct rb_node *p;
 	loff_t pos = *_pos;
-	void *next = NULL;
 
-	down_read(&nommu_vma_sem);
+	down_read(&nommu_region_sem);
 
-	for (_rb = rb_first(&nommu_vma_tree); _rb; _rb = rb_next(_rb)) {
-		if (pos == 0) {
-			next = _rb;
-			break;
-		}
-		pos--;
-	}
-
-	return next;
+	for (p = rb_first(&nommu_region_tree); p; p = rb_next(p))
+		if (pos-- == 0)
+			return p;
+	return NULL;
 }
 
-static void nommu_vma_list_stop(struct seq_file *m, void *v)
+static void nommu_region_list_stop(struct seq_file *m, void *v)
 {
-	up_read(&nommu_vma_sem);
+	up_read(&nommu_region_sem);
 }
 
-static void *nommu_vma_list_next(struct seq_file *m, void *v, loff_t *pos)
+static void *nommu_region_list_next(struct seq_file *m, void *v, loff_t *pos)
 {
 	(*pos)++;
 	return rb_next((struct rb_node *) v);
 }
 
-static const struct seq_operations proc_nommu_vma_list_seqop = {
-	.start	= nommu_vma_list_start,
-	.next	= nommu_vma_list_next,
-	.stop	= nommu_vma_list_stop,
-	.show	= nommu_vma_list_show
+static struct seq_operations proc_nommu_region_list_seqop = {
+	.start	= nommu_region_list_start,
+	.next	= nommu_region_list_next,
+	.stop	= nommu_region_list_stop,
+	.show	= nommu_region_list_show
 };
 
-static int proc_nommu_vma_list_open(struct inode *inode, struct file *file)
+static int proc_nommu_region_list_open(struct inode *inode, struct file *file)
 {
-	return seq_open(file, &proc_nommu_vma_list_seqop);
+	return seq_open(file, &proc_nommu_region_list_seqop);
 }
 
-static const struct file_operations proc_nommu_vma_list_operations = {
-	.open    = proc_nommu_vma_list_open,
+static const struct file_operations proc_nommu_region_list_operations = {
+	.open    = proc_nommu_region_list_open,
 	.read    = seq_read,
 	.llseek  = seq_lseek,
 	.release = seq_release,
@@ -137,7 +130,7 @@ static const struct file_operations proc_nommu_vma_list_operations = {
 
 static int __init proc_nommu_init(void)
 {
-	proc_create("maps", S_IRUGO, NULL, &proc_nommu_vma_list_operations);
+	proc_create("maps", S_IRUGO, NULL, &proc_nommu_region_list_operations);
 	return 0;
 }
 
diff --git a/fs/proc/proc_net.c b/fs/proc/proc_net.c
index 7bc296f424a..04d1270f1c3 100644
--- a/fs/proc/proc_net.c
+++ b/fs/proc/proc_net.c
@@ -18,7 +18,6 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 #include <linux/bitops.h>
-#include <linux/smp_lock.h>
 #include <linux/mount.h>
 #include <linux/nsproxy.h>
 #include <net/net_namespace.h>
@@ -172,6 +171,7 @@ static int proc_tgid_net_readdir(struct file *filp, void *dirent,
 }
 
 const struct file_operations proc_net_operations = {
+	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= proc_tgid_net_readdir,
 };
diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c
index 06ed10b7da9..94fcfff6863 100644
--- a/fs/proc/proc_sysctl.c
+++ b/fs/proc/proc_sysctl.c
@@ -31,7 +31,6 @@ static struct inode *proc_sys_make_inode(struct super_block *sb,
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
 	inode->i_flags |= S_PRIVATE; /* tell selinux to ignore this inode */
 	inode->i_mode = table->mode;
-	inode->i_uid = inode->i_gid = 0;
 	if (!table->child) {
 		inode->i_mode |= S_IFREG;
 		inode->i_op = &proc_sys_inode_operations;
diff --git a/fs/proc/root.c b/fs/proc/root.c
index 7761602af9d..f6299a25594 100644
--- a/fs/proc/root.c
+++ b/fs/proc/root.c
@@ -16,7 +16,6 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 #include <linux/bitops.h>
-#include <linux/smp_lock.h>
 #include <linux/mount.h>
 #include <linux/pid_namespace.h>
 
@@ -162,17 +161,12 @@ static int proc_root_readdir(struct file * filp,
 	unsigned int nr = filp->f_pos;
 	int ret;
 
-	lock_kernel();
-
 	if (nr < FIRST_PROCESS_ENTRY) {
 		int error = proc_readdir(filp, dirent, filldir);
-		if (error <= 0) {
-			unlock_kernel();
+		if (error <= 0)
 			return error;
-		}
 		filp->f_pos = FIRST_PROCESS_ENTRY;
 	}
-	unlock_kernel();
 
 	ret = proc_pid_readdir(filp, dirent, filldir);
 	return ret;
diff --git a/fs/proc/stat.c b/fs/proc/stat.c
index 81904f07679..f75efa22df5 100644
--- a/fs/proc/stat.c
+++ b/fs/proc/stat.c
@@ -9,6 +9,7 @@
 #include <linux/seq_file.h>
 #include <linux/slab.h>
 #include <linux/time.h>
+#include <linux/irqnr.h>
 #include <asm/cputime.h>
 
 #ifndef arch_irq_stat_cpu
@@ -44,10 +45,9 @@ static int show_stat(struct seq_file *p, void *v)
 		softirq = cputime64_add(softirq, kstat_cpu(i).cpustat.softirq);
 		steal = cputime64_add(steal, kstat_cpu(i).cpustat.steal);
 		guest = cputime64_add(guest, kstat_cpu(i).cpustat.guest);
-
-		for_each_irq_nr(j)
+		for_each_irq_nr(j) {
 			sum += kstat_irqs_cpu(j, i);
-
+		}
 		sum += arch_irq_stat_cpu(i);
 	}
 	sum += arch_irq_stat();
@@ -92,7 +92,6 @@ static int show_stat(struct seq_file *p, void *v)
 	/* sum again ? it could be updated? */
 	for_each_irq_nr(j) {
 		per_irq_sum = 0;
-
 		for_each_possible_cpu(i)
 			per_irq_sum += kstat_irqs_cpu(j, i);
 
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 3a8bdd7f575..94063840832 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -396,7 +396,9 @@ static int show_smap(struct seq_file *m, void *v)
 		   "Private_Clean:  %8lu kB\n"
 		   "Private_Dirty:  %8lu kB\n"
 		   "Referenced:     %8lu kB\n"
-		   "Swap:           %8lu kB\n",
+		   "Swap:           %8lu kB\n"
+		   "KernelPageSize: %8lu kB\n"
+		   "MMUPageSize:    %8lu kB\n",
 		   (vma->vm_end - vma->vm_start) >> 10,
 		   mss.resident >> 10,
 		   (unsigned long)(mss.pss >> (10 + PSS_SHIFT)),
@@ -405,7 +407,9 @@ static int show_smap(struct seq_file *m, void *v)
 		   mss.private_clean >> 10,
 		   mss.private_dirty >> 10,
 		   mss.referenced >> 10,
-		   mss.swap >> 10);
+		   mss.swap >> 10,
+		   vma_kernel_pagesize(vma) >> 10,
+		   vma_mmu_pagesize(vma) >> 10);
 
 	if (m->count < m->size)  /* vma is copied successfully */
 		m->version = (vma != get_gate_vma(task)) ? vma->vm_start : 0;
diff --git a/fs/proc/task_nommu.c b/fs/proc/task_nommu.c
index 219bd79ea89..343ea1216bc 100644
--- a/fs/proc/task_nommu.c
+++ b/fs/proc/task_nommu.c
@@ -9,31 +9,38 @@
 
 /*
  * Logic: we've got two memory sums for each process, "shared", and
- * "non-shared". Shared memory may get counted more then once, for
+ * "non-shared". Shared memory may get counted more than once, for
  * each process that owns it. Non-shared memory is counted
  * accurately.
  */
 void task_mem(struct seq_file *m, struct mm_struct *mm)
 {
-	struct vm_list_struct *vml;
-	unsigned long bytes = 0, sbytes = 0, slack = 0;
+	struct vm_area_struct *vma;
+	struct vm_region *region;
+	struct rb_node *p;
+	unsigned long bytes = 0, sbytes = 0, slack = 0, size;
         
 	down_read(&mm->mmap_sem);
-	for (vml = mm->context.vmlist; vml; vml = vml->next) {
-		if (!vml->vma)
-			continue;
+	for (p = rb_first(&mm->mm_rb); p; p = rb_next(p)) {
+		vma = rb_entry(p, struct vm_area_struct, vm_rb);
+
+		bytes += kobjsize(vma);
+
+		region = vma->vm_region;
+		if (region) {
+			size = kobjsize(region);
+			size += region->vm_end - region->vm_start;
+		} else {
+			size = vma->vm_end - vma->vm_start;
+		}
 
-		bytes += kobjsize(vml);
 		if (atomic_read(&mm->mm_count) > 1 ||
-		    atomic_read(&vml->vma->vm_usage) > 1
-		    ) {
-			sbytes += kobjsize((void *) vml->vma->vm_start);
-			sbytes += kobjsize(vml->vma);
+		    vma->vm_flags & VM_MAYSHARE) {
+			sbytes += size;
 		} else {
-			bytes += kobjsize((void *) vml->vma->vm_start);
-			bytes += kobjsize(vml->vma);
-			slack += kobjsize((void *) vml->vma->vm_start) -
-				(vml->vma->vm_end - vml->vma->vm_start);
+			bytes += size;
+			if (region)
+				slack = region->vm_end - vma->vm_end;
 		}
 	}
 
@@ -70,13 +77,14 @@ void task_mem(struct seq_file *m, struct mm_struct *mm)
 
 unsigned long task_vsize(struct mm_struct *mm)
 {
-	struct vm_list_struct *tbp;
+	struct vm_area_struct *vma;
+	struct rb_node *p;
 	unsigned long vsize = 0;
 
 	down_read(&mm->mmap_sem);
-	for (tbp = mm->context.vmlist; tbp; tbp = tbp->next) {
-		if (tbp->vma)
-			vsize += kobjsize((void *) tbp->vma->vm_start);
+	for (p = rb_first(&mm->mm_rb); p; p = rb_next(p)) {
+		vma = rb_entry(p, struct vm_area_struct, vm_rb);
+		vsize += vma->vm_end - vma->vm_start;
 	}
 	up_read(&mm->mmap_sem);
 	return vsize;
@@ -85,15 +93,19 @@ unsigned long task_vsize(struct mm_struct *mm)
 int task_statm(struct mm_struct *mm, int *shared, int *text,
 	       int *data, int *resident)
 {
-	struct vm_list_struct *tbp;
+	struct vm_area_struct *vma;
+	struct vm_region *region;
+	struct rb_node *p;
 	int size = kobjsize(mm);
 
 	down_read(&mm->mmap_sem);
-	for (tbp = mm->context.vmlist; tbp; tbp = tbp->next) {
-		size += kobjsize(tbp);
-		if (tbp->vma) {
-			size += kobjsize(tbp->vma);
-			size += kobjsize((void *) tbp->vma->vm_start);
+	for (p = rb_first(&mm->mm_rb); p; p = rb_next(p)) {
+		vma = rb_entry(p, struct vm_area_struct, vm_rb);
+		size += kobjsize(vma);
+		region = vma->vm_region;
+		if (region) {
+			size += kobjsize(region);
+			size += region->vm_end - region->vm_start;
 		}
 	}
 
@@ -105,20 +117,62 @@ int task_statm(struct mm_struct *mm, int *shared, int *text,
 }
 
 /*
+ * display a single VMA to a sequenced file
+ */
+static int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma)
+{
+	unsigned long ino = 0;
+	struct file *file;
+	dev_t dev = 0;
+	int flags, len;
+
+	flags = vma->vm_flags;
+	file = vma->vm_file;
+
+	if (file) {
+		struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
+		dev = inode->i_sb->s_dev;
+		ino = inode->i_ino;
+	}
+
+	seq_printf(m,
+		   "%08lx-%08lx %c%c%c%c %08lx %02x:%02x %lu %n",
+		   vma->vm_start,
+		   vma->vm_end,
+		   flags & VM_READ ? 'r' : '-',
+		   flags & VM_WRITE ? 'w' : '-',
+		   flags & VM_EXEC ? 'x' : '-',
+		   flags & VM_MAYSHARE ? flags & VM_SHARED ? 'S' : 's' : 'p',
+		   vma->vm_pgoff << PAGE_SHIFT,
+		   MAJOR(dev), MINOR(dev), ino, &len);
+
+	if (file) {
+		len = 25 + sizeof(void *) * 6 - len;
+		if (len < 1)
+			len = 1;
+		seq_printf(m, "%*c", len, ' ');
+		seq_path(m, &file->f_path, "");
+	}
+
+	seq_putc(m, '\n');
+	return 0;
+}
+
+/*
  * display mapping lines for a particular process's /proc/pid/maps
  */
-static int show_map(struct seq_file *m, void *_vml)
+static int show_map(struct seq_file *m, void *_p)
 {
-	struct vm_list_struct *vml = _vml;
+	struct rb_node *p = _p;
 
-	return nommu_vma_show(m, vml->vma);
+	return nommu_vma_show(m, rb_entry(p, struct vm_area_struct, vm_rb));
 }
 
 static void *m_start(struct seq_file *m, loff_t *pos)
 {
 	struct proc_maps_private *priv = m->private;
-	struct vm_list_struct *vml;
 	struct mm_struct *mm;
+	struct rb_node *p;
 	loff_t n = *pos;
 
 	/* pin the task and mm whilst we play with them */
@@ -134,9 +188,9 @@ static void *m_start(struct seq_file *m, loff_t *pos)
 	}
 
 	/* start from the Nth VMA */
-	for (vml = mm->context.vmlist; vml; vml = vml->next)
+	for (p = rb_first(&mm->mm_rb); p; p = rb_next(p))
 		if (n-- == 0)
-			return vml;
+			return p;
 	return NULL;
 }
 
@@ -152,12 +206,12 @@ static void m_stop(struct seq_file *m, void *_vml)
 	}
 }
 
-static void *m_next(struct seq_file *m, void *_vml, loff_t *pos)
+static void *m_next(struct seq_file *m, void *_p, loff_t *pos)
 {
-	struct vm_list_struct *vml = _vml;
+	struct rb_node *p = _p;
 
 	(*pos)++;
-	return vml ? vml->next : NULL;
+	return p ? rb_next(p) : NULL;
 }
 
 static const struct seq_operations proc_pid_maps_ops = {
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 03ec5950490..5edcc3f92ba 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -47,8 +47,6 @@ static ssize_t read_from_oldmem(char *buf, size_t count,
 
 	offset = (unsigned long)(*ppos % PAGE_SIZE);
 	pfn = (unsigned long)(*ppos / PAGE_SIZE);
-	if (pfn > saved_max_pfn)
-		return -EINVAL;
 
 	do {
 		if (count > (PAGE_SIZE - offset))
diff --git a/fs/quota.c b/fs/quota.c
index b7fe44e0161..4a8c94f05f7 100644
--- a/fs/quota.c
+++ b/fs/quota.c
@@ -73,7 +73,7 @@ static int generic_quotactl_valid(struct super_block *sb, int type, int cmd, qid
 		case Q_SETQUOTA:
 		case Q_GETQUOTA:
 			/* This is just informative test so we are satisfied without a lock */
-			if (!sb_has_quota_enabled(sb, type))
+			if (!sb_has_quota_active(sb, type))
 				return -ESRCH;
 	}
 
@@ -160,6 +160,9 @@ static void quota_sync_sb(struct super_block *sb, int type)
 	int cnt;
 
 	sb->s_qcop->quota_sync(sb, type);
+
+	if (sb_dqopt(sb)->flags & DQUOT_QUOTA_SYS_FILE)
+		return;
 	/* This is not very clever (and fast) but currently I don't know about
 	 * any other simple way of getting quota data to disk and we must get
 	 * them there for userspace to be visible... */
@@ -175,7 +178,7 @@ static void quota_sync_sb(struct super_block *sb, int type)
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (type != -1 && cnt != type)
 			continue;
-		if (!sb_has_quota_enabled(sb, cnt))
+		if (!sb_has_quota_active(sb, cnt))
 			continue;
 		mutex_lock_nested(&sb_dqopt(sb)->files[cnt]->i_mutex, I_MUTEX_QUOTA);
 		truncate_inode_pages(&sb_dqopt(sb)->files[cnt]->i_data, 0);
@@ -201,7 +204,7 @@ restart:
 		for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 			if (type != -1 && type != cnt)
 				continue;
-			if (!sb_has_quota_enabled(sb, cnt))
+			if (!sb_has_quota_active(sb, cnt))
 				continue;
 			if (!info_dirty(&sb_dqopt(sb)->info[cnt]) &&
 			    list_empty(&sb_dqopt(sb)->info[cnt].dqi_dirty_list))
@@ -245,7 +248,7 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id, void
 			__u32 fmt;
 
 			down_read(&sb_dqopt(sb)->dqptr_sem);
-			if (!sb_has_quota_enabled(sb, type)) {
+			if (!sb_has_quota_active(sb, type)) {
 				up_read(&sb_dqopt(sb)->dqptr_sem);
 				return -ESRCH;
 			}
diff --git a/fs/quota_tree.c b/fs/quota_tree.c
new file mode 100644
index 00000000000..953404c95b1
--- /dev/null
+++ b/fs/quota_tree.c
@@ -0,0 +1,645 @@
+/*
+ *	vfsv0 quota IO operations on file
+ */
+
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/mount.h>
+#include <linux/dqblk_v2.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/quotaops.h>
+
+#include <asm/byteorder.h>
+
+#include "quota_tree.h"
+
+MODULE_AUTHOR("Jan Kara");
+MODULE_DESCRIPTION("Quota trie support");
+MODULE_LICENSE("GPL");
+
+#define __QUOTA_QT_PARANOIA
+
+typedef char *dqbuf_t;
+
+static int get_index(struct qtree_mem_dqinfo *info, qid_t id, int depth)
+{
+	unsigned int epb = info->dqi_usable_bs >> 2;
+
+	depth = info->dqi_qtree_depth - depth - 1;
+	while (depth--)
+		id /= epb;
+	return id % epb;
+}
+
+/* Number of entries in one blocks */
+static inline int qtree_dqstr_in_blk(struct qtree_mem_dqinfo *info)
+{
+	return (info->dqi_usable_bs - sizeof(struct qt_disk_dqdbheader))
+	       / info->dqi_entry_size;
+}
+
+static dqbuf_t getdqbuf(size_t size)
+{
+	dqbuf_t buf = kmalloc(size, GFP_NOFS);
+	if (!buf)
+		printk(KERN_WARNING "VFS: Not enough memory for quota buffers.\n");
+	return buf;
+}
+
+static inline void freedqbuf(dqbuf_t buf)
+{
+	kfree(buf);
+}
+
+static inline ssize_t read_blk(struct qtree_mem_dqinfo *info, uint blk, dqbuf_t buf)
+{
+	struct super_block *sb = info->dqi_sb;
+
+	memset(buf, 0, info->dqi_usable_bs);
+	return sb->s_op->quota_read(sb, info->dqi_type, (char *)buf,
+	       info->dqi_usable_bs, blk << info->dqi_blocksize_bits);
+}
+
+static inline ssize_t write_blk(struct qtree_mem_dqinfo *info, uint blk, dqbuf_t buf)
+{
+	struct super_block *sb = info->dqi_sb;
+
+	return sb->s_op->quota_write(sb, info->dqi_type, (char *)buf,
+	       info->dqi_usable_bs, blk << info->dqi_blocksize_bits);
+}
+
+/* Remove empty block from list and return it */
+static int get_free_dqblk(struct qtree_mem_dqinfo *info)
+{
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
+	int ret, blk;
+
+	if (!buf)
+		return -ENOMEM;
+	if (info->dqi_free_blk) {
+		blk = info->dqi_free_blk;
+		ret = read_blk(info, blk, buf);
+		if (ret < 0)
+			goto out_buf;
+		info->dqi_free_blk = le32_to_cpu(dh->dqdh_next_free);
+	}
+	else {
+		memset(buf, 0, info->dqi_usable_bs);
+		/* Assure block allocation... */
+		ret = write_blk(info, info->dqi_blocks, buf);
+		if (ret < 0)
+			goto out_buf;
+		blk = info->dqi_blocks++;
+	}
+	mark_info_dirty(info->dqi_sb, info->dqi_type);
+	ret = blk;
+out_buf:
+	freedqbuf(buf);
+	return ret;
+}
+
+/* Insert empty block to the list */
+static int put_free_dqblk(struct qtree_mem_dqinfo *info, dqbuf_t buf, uint blk)
+{
+	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
+	int err;
+
+	dh->dqdh_next_free = cpu_to_le32(info->dqi_free_blk);
+	dh->dqdh_prev_free = cpu_to_le32(0);
+	dh->dqdh_entries = cpu_to_le16(0);
+	err = write_blk(info, blk, buf);
+	if (err < 0)
+		return err;
+	info->dqi_free_blk = blk;
+	mark_info_dirty(info->dqi_sb, info->dqi_type);
+	return 0;
+}
+
+/* Remove given block from the list of blocks with free entries */
+static int remove_free_dqentry(struct qtree_mem_dqinfo *info, dqbuf_t buf, uint blk)
+{
+	dqbuf_t tmpbuf = getdqbuf(info->dqi_usable_bs);
+	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
+	uint nextblk = le32_to_cpu(dh->dqdh_next_free);
+	uint prevblk = le32_to_cpu(dh->dqdh_prev_free);
+	int err;
+
+	if (!tmpbuf)
+		return -ENOMEM;
+	if (nextblk) {
+		err = read_blk(info, nextblk, tmpbuf);
+		if (err < 0)
+			goto out_buf;
+		((struct qt_disk_dqdbheader *)tmpbuf)->dqdh_prev_free =
+							dh->dqdh_prev_free;
+		err = write_blk(info, nextblk, tmpbuf);
+		if (err < 0)
+			goto out_buf;
+	}
+	if (prevblk) {
+		err = read_blk(info, prevblk, tmpbuf);
+		if (err < 0)
+			goto out_buf;
+		((struct qt_disk_dqdbheader *)tmpbuf)->dqdh_next_free =
+							dh->dqdh_next_free;
+		err = write_blk(info, prevblk, tmpbuf);
+		if (err < 0)
+			goto out_buf;
+	} else {
+		info->dqi_free_entry = nextblk;
+		mark_info_dirty(info->dqi_sb, info->dqi_type);
+	}
+	freedqbuf(tmpbuf);
+	dh->dqdh_next_free = dh->dqdh_prev_free = cpu_to_le32(0);
+	/* No matter whether write succeeds block is out of list */
+	if (write_blk(info, blk, buf) < 0)
+		printk(KERN_ERR "VFS: Can't write block (%u) with free entries.\n", blk);
+	return 0;
+out_buf:
+	freedqbuf(tmpbuf);
+	return err;
+}
+
+/* Insert given block to the beginning of list with free entries */
+static int insert_free_dqentry(struct qtree_mem_dqinfo *info, dqbuf_t buf, uint blk)
+{
+	dqbuf_t tmpbuf = getdqbuf(info->dqi_usable_bs);
+	struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf;
+	int err;
+
+	if (!tmpbuf)
+		return -ENOMEM;
+	dh->dqdh_next_free = cpu_to_le32(info->dqi_free_entry);
+	dh->dqdh_prev_free = cpu_to_le32(0);
+	err = write_blk(info, blk, buf);
+	if (err < 0)
+		goto out_buf;
+	if (info->dqi_free_entry) {
+		err = read_blk(info, info->dqi_free_entry, tmpbuf);
+		if (err < 0)
+			goto out_buf;
+		((struct qt_disk_dqdbheader *)tmpbuf)->dqdh_prev_free =
+							cpu_to_le32(blk);
+		err = write_blk(info, info->dqi_free_entry, tmpbuf);
+		if (err < 0)
+			goto out_buf;
+	}
+	freedqbuf(tmpbuf);
+	info->dqi_free_entry = blk;
+	mark_info_dirty(info->dqi_sb, info->dqi_type);
+	return 0;
+out_buf:
+	freedqbuf(tmpbuf);
+	return err;
+}
+
+/* Is the entry in the block free? */
+int qtree_entry_unused(struct qtree_mem_dqinfo *info, char *disk)
+{
+	int i;
+
+	for (i = 0; i < info->dqi_entry_size; i++)
+		if (disk[i])
+			return 0;
+	return 1;
+}
+EXPORT_SYMBOL(qtree_entry_unused);
+
+/* Find space for dquot */
+static uint find_free_dqentry(struct qtree_mem_dqinfo *info,
+			      struct dquot *dquot, int *err)
+{
+	uint blk, i;
+	struct qt_disk_dqdbheader *dh;
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	char *ddquot;
+
+	*err = 0;
+	if (!buf) {
+		*err = -ENOMEM;
+		return 0;
+	}
+	dh = (struct qt_disk_dqdbheader *)buf;
+	if (info->dqi_free_entry) {
+		blk = info->dqi_free_entry;
+		*err = read_blk(info, blk, buf);
+		if (*err < 0)
+			goto out_buf;
+	} else {
+		blk = get_free_dqblk(info);
+		if ((int)blk < 0) {
+			*err = blk;
+			freedqbuf(buf);
+			return 0;
+		}
+		memset(buf, 0, info->dqi_usable_bs);
+		/* This is enough as block is already zeroed and entry list is empty... */
+		info->dqi_free_entry = blk;
+		mark_info_dirty(dquot->dq_sb, dquot->dq_type);
+	}
+	/* Block will be full? */
+	if (le16_to_cpu(dh->dqdh_entries) + 1 >= qtree_dqstr_in_blk(info)) {
+		*err = remove_free_dqentry(info, buf, blk);
+		if (*err < 0) {
+			printk(KERN_ERR "VFS: find_free_dqentry(): Can't "
+			       "remove block (%u) from entry free list.\n",
+			       blk);
+			goto out_buf;
+		}
+	}
+	le16_add_cpu(&dh->dqdh_entries, 1);
+	/* Find free structure in block */
+	for (i = 0, ddquot = ((char *)buf) + sizeof(struct qt_disk_dqdbheader);
+	     i < qtree_dqstr_in_blk(info) && !qtree_entry_unused(info, ddquot);
+	     i++, ddquot += info->dqi_entry_size);
+#ifdef __QUOTA_QT_PARANOIA
+	if (i == qtree_dqstr_in_blk(info)) {
+		printk(KERN_ERR "VFS: find_free_dqentry(): Data block full "
+				"but it shouldn't.\n");
+		*err = -EIO;
+		goto out_buf;
+	}
+#endif
+	*err = write_blk(info, blk, buf);
+	if (*err < 0) {
+		printk(KERN_ERR "VFS: find_free_dqentry(): Can't write quota "
+				"data block %u.\n", blk);
+		goto out_buf;
+	}
+	dquot->dq_off = (blk << info->dqi_blocksize_bits) +
+			sizeof(struct qt_disk_dqdbheader) +
+			i * info->dqi_entry_size;
+	freedqbuf(buf);
+	return blk;
+out_buf:
+	freedqbuf(buf);
+	return 0;
+}
+
+/* Insert reference to structure into the trie */
+static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
+			  uint *treeblk, int depth)
+{
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	int ret = 0, newson = 0, newact = 0;
+	__le32 *ref;
+	uint newblk;
+
+	if (!buf)
+		return -ENOMEM;
+	if (!*treeblk) {
+		ret = get_free_dqblk(info);
+		if (ret < 0)
+			goto out_buf;
+		*treeblk = ret;
+		memset(buf, 0, info->dqi_usable_bs);
+		newact = 1;
+	} else {
+		ret = read_blk(info, *treeblk, buf);
+		if (ret < 0) {
+			printk(KERN_ERR "VFS: Can't read tree quota block "
+					"%u.\n", *treeblk);
+			goto out_buf;
+		}
+	}
+	ref = (__le32 *)buf;
+	newblk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]);
+	if (!newblk)
+		newson = 1;
+	if (depth == info->dqi_qtree_depth - 1) {
+#ifdef __QUOTA_QT_PARANOIA
+		if (newblk) {
+			printk(KERN_ERR "VFS: Inserting already present quota "
+					"entry (block %u).\n",
+			       le32_to_cpu(ref[get_index(info,
+						dquot->dq_id, depth)]));
+			ret = -EIO;
+			goto out_buf;
+		}
+#endif
+		newblk = find_free_dqentry(info, dquot, &ret);
+	} else {
+		ret = do_insert_tree(info, dquot, &newblk, depth+1);
+	}
+	if (newson && ret >= 0) {
+		ref[get_index(info, dquot->dq_id, depth)] =
+							cpu_to_le32(newblk);
+		ret = write_blk(info, *treeblk, buf);
+	} else if (newact && ret < 0) {
+		put_free_dqblk(info, buf, *treeblk);
+	}
+out_buf:
+	freedqbuf(buf);
+	return ret;
+}
+
+/* Wrapper for inserting quota structure into tree */
+static inline int dq_insert_tree(struct qtree_mem_dqinfo *info,
+				 struct dquot *dquot)
+{
+	int tmp = QT_TREEOFF;
+	return do_insert_tree(info, dquot, &tmp, 0);
+}
+
+/*
+ *	We don't have to be afraid of deadlocks as we never have quotas on quota files...
+ */
+int qtree_write_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
+{
+	int type = dquot->dq_type;
+	struct super_block *sb = dquot->dq_sb;
+	ssize_t ret;
+	dqbuf_t ddquot = getdqbuf(info->dqi_entry_size);
+
+	if (!ddquot)
+		return -ENOMEM;
+
+	/* dq_off is guarded by dqio_mutex */
+	if (!dquot->dq_off) {
+		ret = dq_insert_tree(info, dquot);
+		if (ret < 0) {
+			printk(KERN_ERR "VFS: Error %zd occurred while "
+					"creating quota.\n", ret);
+			freedqbuf(ddquot);
+			return ret;
+		}
+	}
+	spin_lock(&dq_data_lock);
+	info->dqi_ops->mem2disk_dqblk(ddquot, dquot);
+	spin_unlock(&dq_data_lock);
+	ret = sb->s_op->quota_write(sb, type, (char *)ddquot,
+					info->dqi_entry_size, dquot->dq_off);
+	if (ret != info->dqi_entry_size) {
+		printk(KERN_WARNING "VFS: dquota write failed on dev %s\n",
+		       sb->s_id);
+		if (ret >= 0)
+			ret = -ENOSPC;
+	} else {
+		ret = 0;
+	}
+	dqstats.writes++;
+	freedqbuf(ddquot);
+
+	return ret;
+}
+EXPORT_SYMBOL(qtree_write_dquot);
+
+/* Free dquot entry in data block */
+static int free_dqentry(struct qtree_mem_dqinfo *info, struct dquot *dquot,
+			uint blk)
+{
+	struct qt_disk_dqdbheader *dh;
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	int ret = 0;
+
+	if (!buf)
+		return -ENOMEM;
+	if (dquot->dq_off >> info->dqi_blocksize_bits != blk) {
+		printk(KERN_ERR "VFS: Quota structure has offset to other "
+		  "block (%u) than it should (%u).\n", blk,
+		  (uint)(dquot->dq_off >> info->dqi_blocksize_bits));
+		goto out_buf;
+	}
+	ret = read_blk(info, blk, buf);
+	if (ret < 0) {
+		printk(KERN_ERR "VFS: Can't read quota data block %u\n", blk);
+		goto out_buf;
+	}
+	dh = (struct qt_disk_dqdbheader *)buf;
+	le16_add_cpu(&dh->dqdh_entries, -1);
+	if (!le16_to_cpu(dh->dqdh_entries)) {	/* Block got free? */
+		ret = remove_free_dqentry(info, buf, blk);
+		if (ret >= 0)
+			ret = put_free_dqblk(info, buf, blk);
+		if (ret < 0) {
+			printk(KERN_ERR "VFS: Can't move quota data block (%u) "
+			  "to free list.\n", blk);
+			goto out_buf;
+		}
+	} else {
+		memset(buf +
+		       (dquot->dq_off & ((1 << info->dqi_blocksize_bits) - 1)),
+		       0, info->dqi_entry_size);
+		if (le16_to_cpu(dh->dqdh_entries) ==
+		    qtree_dqstr_in_blk(info) - 1) {
+			/* Insert will write block itself */
+			ret = insert_free_dqentry(info, buf, blk);
+			if (ret < 0) {
+				printk(KERN_ERR "VFS: Can't insert quota data "
+				       "block (%u) to free entry list.\n", blk);
+				goto out_buf;
+			}
+		} else {
+			ret = write_blk(info, blk, buf);
+			if (ret < 0) {
+				printk(KERN_ERR "VFS: Can't write quota data "
+				  "block %u\n", blk);
+				goto out_buf;
+			}
+		}
+	}
+	dquot->dq_off = 0;	/* Quota is now unattached */
+out_buf:
+	freedqbuf(buf);
+	return ret;
+}
+
+/* Remove reference to dquot from tree */
+static int remove_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot,
+		       uint *blk, int depth)
+{
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	int ret = 0;
+	uint newblk;
+	__le32 *ref = (__le32 *)buf;
+
+	if (!buf)
+		return -ENOMEM;
+	ret = read_blk(info, *blk, buf);
+	if (ret < 0) {
+		printk(KERN_ERR "VFS: Can't read quota data block %u\n", *blk);
+		goto out_buf;
+	}
+	newblk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]);
+	if (depth == info->dqi_qtree_depth - 1) {
+		ret = free_dqentry(info, dquot, newblk);
+		newblk = 0;
+	} else {
+		ret = remove_tree(info, dquot, &newblk, depth+1);
+	}
+	if (ret >= 0 && !newblk) {
+		int i;
+		ref[get_index(info, dquot->dq_id, depth)] = cpu_to_le32(0);
+		/* Block got empty? */
+		for (i = 0;
+		     i < (info->dqi_usable_bs >> 2) && !ref[i];
+		     i++);
+		/* Don't put the root block into the free block list */
+		if (i == (info->dqi_usable_bs >> 2)
+		    && *blk != QT_TREEOFF) {
+			put_free_dqblk(info, buf, *blk);
+			*blk = 0;
+		} else {
+			ret = write_blk(info, *blk, buf);
+			if (ret < 0)
+				printk(KERN_ERR "VFS: Can't write quota tree "
+				  "block %u.\n", *blk);
+		}
+	}
+out_buf:
+	freedqbuf(buf);
+	return ret;
+}
+
+/* Delete dquot from tree */
+int qtree_delete_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
+{
+	uint tmp = QT_TREEOFF;
+
+	if (!dquot->dq_off)	/* Even not allocated? */
+		return 0;
+	return remove_tree(info, dquot, &tmp, 0);
+}
+EXPORT_SYMBOL(qtree_delete_dquot);
+
+/* Find entry in block */
+static loff_t find_block_dqentry(struct qtree_mem_dqinfo *info,
+				 struct dquot *dquot, uint blk)
+{
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	loff_t ret = 0;
+	int i;
+	char *ddquot;
+
+	if (!buf)
+		return -ENOMEM;
+	ret = read_blk(info, blk, buf);
+	if (ret < 0) {
+		printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk);
+		goto out_buf;
+	}
+	for (i = 0, ddquot = ((char *)buf) + sizeof(struct qt_disk_dqdbheader);
+	     i < qtree_dqstr_in_blk(info) && !info->dqi_ops->is_id(ddquot, dquot);
+	     i++, ddquot += info->dqi_entry_size);
+	if (i == qtree_dqstr_in_blk(info)) {
+		printk(KERN_ERR "VFS: Quota for id %u referenced "
+		  "but not present.\n", dquot->dq_id);
+		ret = -EIO;
+		goto out_buf;
+	} else {
+		ret = (blk << info->dqi_blocksize_bits) + sizeof(struct
+		  qt_disk_dqdbheader) + i * info->dqi_entry_size;
+	}
+out_buf:
+	freedqbuf(buf);
+	return ret;
+}
+
+/* Find entry for given id in the tree */
+static loff_t find_tree_dqentry(struct qtree_mem_dqinfo *info,
+				struct dquot *dquot, uint blk, int depth)
+{
+	dqbuf_t buf = getdqbuf(info->dqi_usable_bs);
+	loff_t ret = 0;
+	__le32 *ref = (__le32 *)buf;
+
+	if (!buf)
+		return -ENOMEM;
+	ret = read_blk(info, blk, buf);
+	if (ret < 0) {
+		printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk);
+		goto out_buf;
+	}
+	ret = 0;
+	blk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]);
+	if (!blk)	/* No reference? */
+		goto out_buf;
+	if (depth < info->dqi_qtree_depth - 1)
+		ret = find_tree_dqentry(info, dquot, blk, depth+1);
+	else
+		ret = find_block_dqentry(info, dquot, blk);
+out_buf:
+	freedqbuf(buf);
+	return ret;
+}
+
+/* Find entry for given id in the tree - wrapper function */
+static inline loff_t find_dqentry(struct qtree_mem_dqinfo *info,
+				  struct dquot *dquot)
+{
+	return find_tree_dqentry(info, dquot, QT_TREEOFF, 0);
+}
+
+int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
+{
+	int type = dquot->dq_type;
+	struct super_block *sb = dquot->dq_sb;
+	loff_t offset;
+	dqbuf_t ddquot;
+	int ret = 0;
+
+#ifdef __QUOTA_QT_PARANOIA
+	/* Invalidated quota? */
+	if (!sb_dqopt(dquot->dq_sb)->files[type]) {
+		printk(KERN_ERR "VFS: Quota invalidated while reading!\n");
+		return -EIO;
+	}
+#endif
+	/* Do we know offset of the dquot entry in the quota file? */
+	if (!dquot->dq_off) {
+		offset = find_dqentry(info, dquot);
+		if (offset <= 0) {	/* Entry not present? */
+			if (offset < 0)
+				printk(KERN_ERR "VFS: Can't read quota "
+				  "structure for id %u.\n", dquot->dq_id);
+			dquot->dq_off = 0;
+			set_bit(DQ_FAKE_B, &dquot->dq_flags);
+			memset(&dquot->dq_dqb, 0, sizeof(struct mem_dqblk));
+			ret = offset;
+			goto out;
+		}
+		dquot->dq_off = offset;
+	}
+	ddquot = getdqbuf(info->dqi_entry_size);
+	if (!ddquot)
+		return -ENOMEM;
+	ret = sb->s_op->quota_read(sb, type, (char *)ddquot,
+				   info->dqi_entry_size, dquot->dq_off);
+	if (ret != info->dqi_entry_size) {
+		if (ret >= 0)
+			ret = -EIO;
+		printk(KERN_ERR "VFS: Error while reading quota "
+				"structure for id %u.\n", dquot->dq_id);
+		set_bit(DQ_FAKE_B, &dquot->dq_flags);
+		memset(&dquot->dq_dqb, 0, sizeof(struct mem_dqblk));
+		freedqbuf(ddquot);
+		goto out;
+	}
+	spin_lock(&dq_data_lock);
+	info->dqi_ops->disk2mem_dqblk(dquot, ddquot);
+	if (!dquot->dq_dqb.dqb_bhardlimit &&
+	    !dquot->dq_dqb.dqb_bsoftlimit &&
+	    !dquot->dq_dqb.dqb_ihardlimit &&
+	    !dquot->dq_dqb.dqb_isoftlimit)
+		set_bit(DQ_FAKE_B, &dquot->dq_flags);
+	spin_unlock(&dq_data_lock);
+	freedqbuf(ddquot);
+out:
+	dqstats.reads++;
+	return ret;
+}
+EXPORT_SYMBOL(qtree_read_dquot);
+
+/* Check whether dquot should not be deleted. We know we are
+ * the only one operating on dquot (thanks to dq_lock) */
+int qtree_release_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot)
+{
+	if (test_bit(DQ_FAKE_B, &dquot->dq_flags) && !(dquot->dq_dqb.dqb_curinodes | dquot->dq_dqb.dqb_curspace))
+		return qtree_delete_dquot(info, dquot);
+	return 0;
+}
+EXPORT_SYMBOL(qtree_release_dquot);
diff --git a/fs/quota_tree.h b/fs/quota_tree.h
new file mode 100644
index 00000000000..a1ab8db81a5
--- /dev/null
+++ b/fs/quota_tree.h
@@ -0,0 +1,25 @@
+/*
+ *	Definitions of structures for vfsv0 quota format
+ */
+
+#ifndef _LINUX_QUOTA_TREE_H
+#define _LINUX_QUOTA_TREE_H
+
+#include <linux/types.h>
+#include <linux/quota.h>
+
+/*
+ *  Structure of header of block with quota structures. It is padded to 16 bytes so
+ *  there will be space for exactly 21 quota-entries in a block
+ */
+struct qt_disk_dqdbheader {
+	__le32 dqdh_next_free;	/* Number of next block with free entry */
+	__le32 dqdh_prev_free;	/* Number of previous block with free entry */
+	__le16 dqdh_entries;	/* Number of valid entries in block */
+	__le16 dqdh_pad1;
+	__le32 dqdh_pad2;
+};
+
+#define QT_TREEOFF	1		/* Offset of tree in file in blocks */
+
+#endif /* _LINUX_QUOTAIO_TREE_H */
diff --git a/fs/quota_v1.c b/fs/quota_v1.c
index 5ae15b13eeb..b4af1c69ad1 100644
--- a/fs/quota_v1.c
+++ b/fs/quota_v1.c
@@ -3,25 +3,39 @@
 #include <linux/quota.h>
 #include <linux/quotaops.h>
 #include <linux/dqblk_v1.h>
-#include <linux/quotaio_v1.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/module.h>
 
 #include <asm/byteorder.h>
 
+#include "quotaio_v1.h"
+
 MODULE_AUTHOR("Jan Kara");
 MODULE_DESCRIPTION("Old quota format support");
 MODULE_LICENSE("GPL");
 
+#define QUOTABLOCK_BITS 10
+#define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS)
+
+static inline qsize_t v1_stoqb(qsize_t space)
+{
+	return (space + QUOTABLOCK_SIZE - 1) >> QUOTABLOCK_BITS;
+}
+
+static inline qsize_t v1_qbtos(qsize_t blocks)
+{
+	return blocks << QUOTABLOCK_BITS;
+}
+
 static void v1_disk2mem_dqblk(struct mem_dqblk *m, struct v1_disk_dqblk *d)
 {
 	m->dqb_ihardlimit = d->dqb_ihardlimit;
 	m->dqb_isoftlimit = d->dqb_isoftlimit;
 	m->dqb_curinodes = d->dqb_curinodes;
-	m->dqb_bhardlimit = d->dqb_bhardlimit;
-	m->dqb_bsoftlimit = d->dqb_bsoftlimit;
-	m->dqb_curspace = ((qsize_t)d->dqb_curblocks) << QUOTABLOCK_BITS;
+	m->dqb_bhardlimit = v1_qbtos(d->dqb_bhardlimit);
+	m->dqb_bsoftlimit = v1_qbtos(d->dqb_bsoftlimit);
+	m->dqb_curspace = v1_qbtos(d->dqb_curblocks);
 	m->dqb_itime = d->dqb_itime;
 	m->dqb_btime = d->dqb_btime;
 }
@@ -31,9 +45,9 @@ static void v1_mem2disk_dqblk(struct v1_disk_dqblk *d, struct mem_dqblk *m)
 	d->dqb_ihardlimit = m->dqb_ihardlimit;
 	d->dqb_isoftlimit = m->dqb_isoftlimit;
 	d->dqb_curinodes = m->dqb_curinodes;
-	d->dqb_bhardlimit = m->dqb_bhardlimit;
-	d->dqb_bsoftlimit = m->dqb_bsoftlimit;
-	d->dqb_curblocks = toqb(m->dqb_curspace);
+	d->dqb_bhardlimit = v1_stoqb(m->dqb_bhardlimit);
+	d->dqb_bsoftlimit = v1_stoqb(m->dqb_bsoftlimit);
+	d->dqb_curblocks = v1_stoqb(m->dqb_curspace);
 	d->dqb_itime = m->dqb_itime;
 	d->dqb_btime = m->dqb_btime;
 }
diff --git a/fs/quota_v2.c b/fs/quota_v2.c
index b53827dc02d..b618b563635 100644
--- a/fs/quota_v2.c
+++ b/fs/quota_v2.c
@@ -6,7 +6,6 @@
 #include <linux/fs.h>
 #include <linux/mount.h>
 #include <linux/dqblk_v2.h>
-#include <linux/quotaio_v2.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/module.h>
@@ -15,16 +14,37 @@
 
 #include <asm/byteorder.h>
 
+#include "quota_tree.h"
+#include "quotaio_v2.h"
+
 MODULE_AUTHOR("Jan Kara");
 MODULE_DESCRIPTION("Quota format v2 support");
 MODULE_LICENSE("GPL");
 
 #define __QUOTA_V2_PARANOIA
 
-typedef char *dqbuf_t;
+static void v2_mem2diskdqb(void *dp, struct dquot *dquot);
+static void v2_disk2memdqb(struct dquot *dquot, void *dp);
+static int v2_is_id(void *dp, struct dquot *dquot);
+
+static struct qtree_fmt_operations v2_qtree_ops = {
+	.mem2disk_dqblk = v2_mem2diskdqb,
+	.disk2mem_dqblk = v2_disk2memdqb,
+	.is_id = v2_is_id,
+};
+
+#define QUOTABLOCK_BITS 10
+#define QUOTABLOCK_SIZE (1 << QUOTABLOCK_BITS)
 
-#define GETIDINDEX(id, depth) (((id) >> ((V2_DQTREEDEPTH-(depth)-1)*8)) & 0xff)
-#define GETENTRIES(buf) ((struct v2_disk_dqblk *)(((char *)buf)+sizeof(struct v2_disk_dqdbheader)))
+static inline qsize_t v2_stoqb(qsize_t space)
+{
+	return (space + QUOTABLOCK_SIZE - 1) >> QUOTABLOCK_BITS;
+}
+
+static inline qsize_t v2_qbtos(qsize_t blocks)
+{
+	return blocks << QUOTABLOCK_BITS;
+}
 
 /* Check whether given file is really vfsv0 quotafile */
 static int v2_check_quota_file(struct super_block *sb, int type)
@@ -50,7 +70,8 @@ static int v2_check_quota_file(struct super_block *sb, int type)
 static int v2_read_file_info(struct super_block *sb, int type)
 {
 	struct v2_disk_dqinfo dinfo;
-	struct mem_dqinfo *info = sb_dqopt(sb)->info+type;
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct qtree_mem_dqinfo *qinfo;
 	ssize_t size;
 
 	size = sb->s_op->quota_read(sb, type, (char *)&dinfo,
@@ -60,15 +81,29 @@ static int v2_read_file_info(struct super_block *sb, int type)
 			sb->s_id);
 		return -1;
 	}
+	info->dqi_priv = kmalloc(sizeof(struct qtree_mem_dqinfo), GFP_NOFS);
+	if (!info->dqi_priv) {
+		printk(KERN_WARNING
+		       "Not enough memory for quota information structure.\n");
+		return -1;
+	}
+	qinfo = info->dqi_priv;
 	/* limits are stored as unsigned 32-bit data */
 	info->dqi_maxblimit = 0xffffffff;
 	info->dqi_maxilimit = 0xffffffff;
 	info->dqi_bgrace = le32_to_cpu(dinfo.dqi_bgrace);
 	info->dqi_igrace = le32_to_cpu(dinfo.dqi_igrace);
 	info->dqi_flags = le32_to_cpu(dinfo.dqi_flags);
-	info->u.v2_i.dqi_blocks = le32_to_cpu(dinfo.dqi_blocks);
-	info->u.v2_i.dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk);
-	info->u.v2_i.dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry);
+	qinfo->dqi_sb = sb;
+	qinfo->dqi_type = type;
+	qinfo->dqi_blocks = le32_to_cpu(dinfo.dqi_blocks);
+	qinfo->dqi_free_blk = le32_to_cpu(dinfo.dqi_free_blk);
+	qinfo->dqi_free_entry = le32_to_cpu(dinfo.dqi_free_entry);
+	qinfo->dqi_blocksize_bits = V2_DQBLKSIZE_BITS;
+	qinfo->dqi_usable_bs = 1 << V2_DQBLKSIZE_BITS;
+	qinfo->dqi_qtree_depth = qtree_depth(qinfo);
+	qinfo->dqi_entry_size = sizeof(struct v2_disk_dqblk);
+	qinfo->dqi_ops = &v2_qtree_ops;
 	return 0;
 }
 
@@ -76,7 +111,8 @@ static int v2_read_file_info(struct super_block *sb, int type)
 static int v2_write_file_info(struct super_block *sb, int type)
 {
 	struct v2_disk_dqinfo dinfo;
-	struct mem_dqinfo *info = sb_dqopt(sb)->info+type;
+	struct mem_dqinfo *info = sb_dqinfo(sb, type);
+	struct qtree_mem_dqinfo *qinfo = info->dqi_priv;
 	ssize_t size;
 
 	spin_lock(&dq_data_lock);
@@ -85,9 +121,9 @@ static int v2_write_file_info(struct super_block *sb, int type)
 	dinfo.dqi_igrace = cpu_to_le32(info->dqi_igrace);
 	dinfo.dqi_flags = cpu_to_le32(info->dqi_flags & DQF_MASK);
 	spin_unlock(&dq_data_lock);
-	dinfo.dqi_blocks = cpu_to_le32(info->u.v2_i.dqi_blocks);
-	dinfo.dqi_free_blk = cpu_to_le32(info->u.v2_i.dqi_free_blk);
-	dinfo.dqi_free_entry = cpu_to_le32(info->u.v2_i.dqi_free_entry);
+	dinfo.dqi_blocks = cpu_to_le32(qinfo->dqi_blocks);
+	dinfo.dqi_free_blk = cpu_to_le32(qinfo->dqi_free_blk);
+	dinfo.dqi_free_entry = cpu_to_le32(qinfo->dqi_free_entry);
 	size = sb->s_op->quota_write(sb, type, (char *)&dinfo,
 	       sizeof(struct v2_disk_dqinfo), V2_DQINFOOFF);
 	if (size != sizeof(struct v2_disk_dqinfo)) {
@@ -98,574 +134,75 @@ static int v2_write_file_info(struct super_block *sb, int type)
 	return 0;
 }
 
-static void disk2memdqb(struct mem_dqblk *m, struct v2_disk_dqblk *d)
+static void v2_disk2memdqb(struct dquot *dquot, void *dp)
 {
+	struct v2_disk_dqblk *d = dp, empty;
+	struct mem_dqblk *m = &dquot->dq_dqb;
+
 	m->dqb_ihardlimit = le32_to_cpu(d->dqb_ihardlimit);
 	m->dqb_isoftlimit = le32_to_cpu(d->dqb_isoftlimit);
 	m->dqb_curinodes = le32_to_cpu(d->dqb_curinodes);
 	m->dqb_itime = le64_to_cpu(d->dqb_itime);
-	m->dqb_bhardlimit = le32_to_cpu(d->dqb_bhardlimit);
-	m->dqb_bsoftlimit = le32_to_cpu(d->dqb_bsoftlimit);
+	m->dqb_bhardlimit = v2_qbtos(le32_to_cpu(d->dqb_bhardlimit));
+	m->dqb_bsoftlimit = v2_qbtos(le32_to_cpu(d->dqb_bsoftlimit));
 	m->dqb_curspace = le64_to_cpu(d->dqb_curspace);
 	m->dqb_btime = le64_to_cpu(d->dqb_btime);
+	/* We need to escape back all-zero structure */
+	memset(&empty, 0, sizeof(struct v2_disk_dqblk));
+	empty.dqb_itime = cpu_to_le64(1);
+	if (!memcmp(&empty, dp, sizeof(struct v2_disk_dqblk)))
+		m->dqb_itime = 0;
 }
 
-static void mem2diskdqb(struct v2_disk_dqblk *d, struct mem_dqblk *m, qid_t id)
+static void v2_mem2diskdqb(void *dp, struct dquot *dquot)
 {
+	struct v2_disk_dqblk *d = dp;
+	struct mem_dqblk *m = &dquot->dq_dqb;
+	struct qtree_mem_dqinfo *info =
+			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
+
 	d->dqb_ihardlimit = cpu_to_le32(m->dqb_ihardlimit);
 	d->dqb_isoftlimit = cpu_to_le32(m->dqb_isoftlimit);
 	d->dqb_curinodes = cpu_to_le32(m->dqb_curinodes);
 	d->dqb_itime = cpu_to_le64(m->dqb_itime);
-	d->dqb_bhardlimit = cpu_to_le32(m->dqb_bhardlimit);
-	d->dqb_bsoftlimit = cpu_to_le32(m->dqb_bsoftlimit);
+	d->dqb_bhardlimit = cpu_to_le32(v2_stoqb(m->dqb_bhardlimit));
+	d->dqb_bsoftlimit = cpu_to_le32(v2_stoqb(m->dqb_bsoftlimit));
 	d->dqb_curspace = cpu_to_le64(m->dqb_curspace);
 	d->dqb_btime = cpu_to_le64(m->dqb_btime);
-	d->dqb_id = cpu_to_le32(id);
-}
-
-static dqbuf_t getdqbuf(void)
-{
-	dqbuf_t buf = kmalloc(V2_DQBLKSIZE, GFP_NOFS);
-	if (!buf)
-		printk(KERN_WARNING "VFS: Not enough memory for quota buffers.\n");
-	return buf;
-}
-
-static inline void freedqbuf(dqbuf_t buf)
-{
-	kfree(buf);
-}
-
-static inline ssize_t read_blk(struct super_block *sb, int type, uint blk, dqbuf_t buf)
-{
-	memset(buf, 0, V2_DQBLKSIZE);
-	return sb->s_op->quota_read(sb, type, (char *)buf,
-	       V2_DQBLKSIZE, blk << V2_DQBLKSIZE_BITS);
-}
-
-static inline ssize_t write_blk(struct super_block *sb, int type, uint blk, dqbuf_t buf)
-{
-	return sb->s_op->quota_write(sb, type, (char *)buf,
-	       V2_DQBLKSIZE, blk << V2_DQBLKSIZE_BITS);
-}
-
-/* Remove empty block from list and return it */
-static int get_free_dqblk(struct super_block *sb, int type)
-{
-	dqbuf_t buf = getdqbuf();
-	struct mem_dqinfo *info = sb_dqinfo(sb, type);
-	struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf;
-	int ret, blk;
-
-	if (!buf)
-		return -ENOMEM;
-	if (info->u.v2_i.dqi_free_blk) {
-		blk = info->u.v2_i.dqi_free_blk;
-		if ((ret = read_blk(sb, type, blk, buf)) < 0)
-			goto out_buf;
-		info->u.v2_i.dqi_free_blk = le32_to_cpu(dh->dqdh_next_free);
-	}
-	else {
-		memset(buf, 0, V2_DQBLKSIZE);
-		/* Assure block allocation... */
-		if ((ret = write_blk(sb, type, info->u.v2_i.dqi_blocks, buf)) < 0)
-			goto out_buf;
-		blk = info->u.v2_i.dqi_blocks++;
-	}
-	mark_info_dirty(sb, type);
-	ret = blk;
-out_buf:
-	freedqbuf(buf);
-	return ret;
-}
-
-/* Insert empty block to the list */
-static int put_free_dqblk(struct super_block *sb, int type, dqbuf_t buf, uint blk)
-{
-	struct mem_dqinfo *info = sb_dqinfo(sb, type);
-	struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf;
-	int err;
-
-	dh->dqdh_next_free = cpu_to_le32(info->u.v2_i.dqi_free_blk);
-	dh->dqdh_prev_free = cpu_to_le32(0);
-	dh->dqdh_entries = cpu_to_le16(0);
-	info->u.v2_i.dqi_free_blk = blk;
-	mark_info_dirty(sb, type);
-	/* Some strange block. We had better leave it... */
-	if ((err = write_blk(sb, type, blk, buf)) < 0)
-		return err;
-	return 0;
+	d->dqb_id = cpu_to_le32(dquot->dq_id);
+	if (qtree_entry_unused(info, dp))
+		d->dqb_itime = cpu_to_le64(1);
 }
 
-/* Remove given block from the list of blocks with free entries */
-static int remove_free_dqentry(struct super_block *sb, int type, dqbuf_t buf, uint blk)
+static int v2_is_id(void *dp, struct dquot *dquot)
 {
-	dqbuf_t tmpbuf = getdqbuf();
-	struct mem_dqinfo *info = sb_dqinfo(sb, type);
-	struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf;
-	uint nextblk = le32_to_cpu(dh->dqdh_next_free), prevblk = le32_to_cpu(dh->dqdh_prev_free);
-	int err;
+	struct v2_disk_dqblk *d = dp;
+	struct qtree_mem_dqinfo *info =
+			sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv;
 
-	if (!tmpbuf)
-		return -ENOMEM;
-	if (nextblk) {
-		if ((err = read_blk(sb, type, nextblk, tmpbuf)) < 0)
-			goto out_buf;
-		((struct v2_disk_dqdbheader *)tmpbuf)->dqdh_prev_free = dh->dqdh_prev_free;
-		if ((err = write_blk(sb, type, nextblk, tmpbuf)) < 0)
-			goto out_buf;
-	}
-	if (prevblk) {
-		if ((err = read_blk(sb, type, prevblk, tmpbuf)) < 0)
-			goto out_buf;
-		((struct v2_disk_dqdbheader *)tmpbuf)->dqdh_next_free = dh->dqdh_next_free;
-		if ((err = write_blk(sb, type, prevblk, tmpbuf)) < 0)
-			goto out_buf;
-	}
-	else {
-		info->u.v2_i.dqi_free_entry = nextblk;
-		mark_info_dirty(sb, type);
-	}
-	freedqbuf(tmpbuf);
-	dh->dqdh_next_free = dh->dqdh_prev_free = cpu_to_le32(0);
-	/* No matter whether write succeeds block is out of list */
-	if (write_blk(sb, type, blk, buf) < 0)
-		printk(KERN_ERR "VFS: Can't write block (%u) with free entries.\n", blk);
-	return 0;
-out_buf:
-	freedqbuf(tmpbuf);
-	return err;
-}
-
-/* Insert given block to the beginning of list with free entries */
-static int insert_free_dqentry(struct super_block *sb, int type, dqbuf_t buf, uint blk)
-{
-	dqbuf_t tmpbuf = getdqbuf();
-	struct mem_dqinfo *info = sb_dqinfo(sb, type);
-	struct v2_disk_dqdbheader *dh = (struct v2_disk_dqdbheader *)buf;
-	int err;
-
-	if (!tmpbuf)
-		return -ENOMEM;
-	dh->dqdh_next_free = cpu_to_le32(info->u.v2_i.dqi_free_entry);
-	dh->dqdh_prev_free = cpu_to_le32(0);
-	if ((err = write_blk(sb, type, blk, buf)) < 0)
-		goto out_buf;
-	if (info->u.v2_i.dqi_free_entry) {
-		if ((err = read_blk(sb, type, info->u.v2_i.dqi_free_entry, tmpbuf)) < 0)
-			goto out_buf;
-		((struct v2_disk_dqdbheader *)tmpbuf)->dqdh_prev_free = cpu_to_le32(blk);
-		if ((err = write_blk(sb, type, info->u.v2_i.dqi_free_entry, tmpbuf)) < 0)
-			goto out_buf;
-	}
-	freedqbuf(tmpbuf);
-	info->u.v2_i.dqi_free_entry = blk;
-	mark_info_dirty(sb, type);
-	return 0;
-out_buf:
-	freedqbuf(tmpbuf);
-	return err;
-}
-
-/* Find space for dquot */
-static uint find_free_dqentry(struct dquot *dquot, int *err)
-{
-	struct super_block *sb = dquot->dq_sb;
-	struct mem_dqinfo *info = sb_dqopt(sb)->info+dquot->dq_type;
-	uint blk, i;
-	struct v2_disk_dqdbheader *dh;
-	struct v2_disk_dqblk *ddquot;
-	struct v2_disk_dqblk fakedquot;
-	dqbuf_t buf;
-
-	*err = 0;
-	if (!(buf = getdqbuf())) {
-		*err = -ENOMEM;
+	if (qtree_entry_unused(info, dp))
 		return 0;
-	}
-	dh = (struct v2_disk_dqdbheader *)buf;
-	ddquot = GETENTRIES(buf);
-	if (info->u.v2_i.dqi_free_entry) {
-		blk = info->u.v2_i.dqi_free_entry;
-		if ((*err = read_blk(sb, dquot->dq_type, blk, buf)) < 0)
-			goto out_buf;
-	}
-	else {
-		blk = get_free_dqblk(sb, dquot->dq_type);
-		if ((int)blk < 0) {
-			*err = blk;
-			freedqbuf(buf);
-			return 0;
-		}
-		memset(buf, 0, V2_DQBLKSIZE);
-		/* This is enough as block is already zeroed and entry list is empty... */
-		info->u.v2_i.dqi_free_entry = blk;
-		mark_info_dirty(sb, dquot->dq_type);
-	}
-	if (le16_to_cpu(dh->dqdh_entries)+1 >= V2_DQSTRINBLK)	/* Block will be full? */
-		if ((*err = remove_free_dqentry(sb, dquot->dq_type, buf, blk)) < 0) {
-			printk(KERN_ERR "VFS: find_free_dqentry(): Can't remove block (%u) from entry free list.\n", blk);
-			goto out_buf;
-		}
-	le16_add_cpu(&dh->dqdh_entries, 1);
-	memset(&fakedquot, 0, sizeof(struct v2_disk_dqblk));
-	/* Find free structure in block */
-	for (i = 0; i < V2_DQSTRINBLK && memcmp(&fakedquot, ddquot+i, sizeof(struct v2_disk_dqblk)); i++);
-#ifdef __QUOTA_V2_PARANOIA
-	if (i == V2_DQSTRINBLK) {
-		printk(KERN_ERR "VFS: find_free_dqentry(): Data block full but it shouldn't.\n");
-		*err = -EIO;
-		goto out_buf;
-	}
-#endif
-	if ((*err = write_blk(sb, dquot->dq_type, blk, buf)) < 0) {
-		printk(KERN_ERR "VFS: find_free_dqentry(): Can't write quota data block %u.\n", blk);
-		goto out_buf;
-	}
-	dquot->dq_off = (blk<<V2_DQBLKSIZE_BITS)+sizeof(struct v2_disk_dqdbheader)+i*sizeof(struct v2_disk_dqblk);
-	freedqbuf(buf);
-	return blk;
-out_buf:
-	freedqbuf(buf);
-	return 0;
-}
-
-/* Insert reference to structure into the trie */
-static int do_insert_tree(struct dquot *dquot, uint *treeblk, int depth)
-{
-	struct super_block *sb = dquot->dq_sb;
-	dqbuf_t buf;
-	int ret = 0, newson = 0, newact = 0;
-	__le32 *ref;
-	uint newblk;
-
-	if (!(buf = getdqbuf()))
-		return -ENOMEM;
-	if (!*treeblk) {
-		ret = get_free_dqblk(sb, dquot->dq_type);
-		if (ret < 0)
-			goto out_buf;
-		*treeblk = ret;
-		memset(buf, 0, V2_DQBLKSIZE);
-		newact = 1;
-	}
-	else {
-		if ((ret = read_blk(sb, dquot->dq_type, *treeblk, buf)) < 0) {
-			printk(KERN_ERR "VFS: Can't read tree quota block %u.\n", *treeblk);
-			goto out_buf;
-		}
-	}
-	ref = (__le32 *)buf;
-	newblk = le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]);
-	if (!newblk)
-		newson = 1;
-	if (depth == V2_DQTREEDEPTH-1) {
-#ifdef __QUOTA_V2_PARANOIA
-		if (newblk) {
-			printk(KERN_ERR "VFS: Inserting already present quota entry (block %u).\n", le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]));
-			ret = -EIO;
-			goto out_buf;
-		}
-#endif
-		newblk = find_free_dqentry(dquot, &ret);
-	}
-	else
-		ret = do_insert_tree(dquot, &newblk, depth+1);
-	if (newson && ret >= 0) {
-		ref[GETIDINDEX(dquot->dq_id, depth)] = cpu_to_le32(newblk);
-		ret = write_blk(sb, dquot->dq_type, *treeblk, buf);
-	}
-	else if (newact && ret < 0)
-		put_free_dqblk(sb, dquot->dq_type, buf, *treeblk);
-out_buf:
-	freedqbuf(buf);
-	return ret;
+	return le32_to_cpu(d->dqb_id) == dquot->dq_id;
 }
 
-/* Wrapper for inserting quota structure into tree */
-static inline int dq_insert_tree(struct dquot *dquot)
+static int v2_read_dquot(struct dquot *dquot)
 {
-	int tmp = V2_DQTREEOFF;
-	return do_insert_tree(dquot, &tmp, 0);
+	return qtree_read_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv, dquot);
 }
 
-/*
- *	We don't have to be afraid of deadlocks as we never have quotas on quota files...
- */
 static int v2_write_dquot(struct dquot *dquot)
 {
-	int type = dquot->dq_type;
-	ssize_t ret;
-	struct v2_disk_dqblk ddquot, empty;
-
-	/* dq_off is guarded by dqio_mutex */
-	if (!dquot->dq_off)
-		if ((ret = dq_insert_tree(dquot)) < 0) {
-			printk(KERN_ERR "VFS: Error %zd occurred while creating quota.\n", ret);
-			return ret;
-		}
-	spin_lock(&dq_data_lock);
-	mem2diskdqb(&ddquot, &dquot->dq_dqb, dquot->dq_id);
-	/* Argh... We may need to write structure full of zeroes but that would be
-	 * treated as an empty place by the rest of the code. Format change would
-	 * be definitely cleaner but the problems probably are not worth it */
-	memset(&empty, 0, sizeof(struct v2_disk_dqblk));
-	if (!memcmp(&empty, &ddquot, sizeof(struct v2_disk_dqblk)))
-		ddquot.dqb_itime = cpu_to_le64(1);
-	spin_unlock(&dq_data_lock);
-	ret = dquot->dq_sb->s_op->quota_write(dquot->dq_sb, type,
-	      (char *)&ddquot, sizeof(struct v2_disk_dqblk), dquot->dq_off);
-	if (ret != sizeof(struct v2_disk_dqblk)) {
-		printk(KERN_WARNING "VFS: dquota write failed on dev %s\n", dquot->dq_sb->s_id);
-		if (ret >= 0)
-			ret = -ENOSPC;
-	}
-	else
-		ret = 0;
-	dqstats.writes++;
-
-	return ret;
+	return qtree_write_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv, dquot);
 }
 
-/* Free dquot entry in data block */
-static int free_dqentry(struct dquot *dquot, uint blk)
-{
-	struct super_block *sb = dquot->dq_sb;
-	int type = dquot->dq_type;
-	struct v2_disk_dqdbheader *dh;
-	dqbuf_t buf = getdqbuf();
-	int ret = 0;
-
-	if (!buf)
-		return -ENOMEM;
-	if (dquot->dq_off >> V2_DQBLKSIZE_BITS != blk) {
-		printk(KERN_ERR "VFS: Quota structure has offset to other "
-		  "block (%u) than it should (%u).\n", blk,
-		  (uint)(dquot->dq_off >> V2_DQBLKSIZE_BITS));
-		goto out_buf;
-	}
-	if ((ret = read_blk(sb, type, blk, buf)) < 0) {
-		printk(KERN_ERR "VFS: Can't read quota data block %u\n", blk);
-		goto out_buf;
-	}
-	dh = (struct v2_disk_dqdbheader *)buf;
-	le16_add_cpu(&dh->dqdh_entries, -1);
-	if (!le16_to_cpu(dh->dqdh_entries)) {	/* Block got free? */
-		if ((ret = remove_free_dqentry(sb, type, buf, blk)) < 0 ||
-		    (ret = put_free_dqblk(sb, type, buf, blk)) < 0) {
-			printk(KERN_ERR "VFS: Can't move quota data block (%u) "
-			  "to free list.\n", blk);
-			goto out_buf;
-		}
-	}
-	else {
-		memset(buf+(dquot->dq_off & ((1 << V2_DQBLKSIZE_BITS)-1)), 0,
-		  sizeof(struct v2_disk_dqblk));
-		if (le16_to_cpu(dh->dqdh_entries) == V2_DQSTRINBLK-1) {
-			/* Insert will write block itself */
-			if ((ret = insert_free_dqentry(sb, type, buf, blk)) < 0) {
-				printk(KERN_ERR "VFS: Can't insert quota data block (%u) to free entry list.\n", blk);
-				goto out_buf;
-			}
-		}
-		else
-			if ((ret = write_blk(sb, type, blk, buf)) < 0) {
-				printk(KERN_ERR "VFS: Can't write quota data "
-				  "block %u\n", blk);
-				goto out_buf;
-			}
-	}
-	dquot->dq_off = 0;	/* Quota is now unattached */
-out_buf:
-	freedqbuf(buf);
-	return ret;
-}
-
-/* Remove reference to dquot from tree */
-static int remove_tree(struct dquot *dquot, uint *blk, int depth)
-{
-	struct super_block *sb = dquot->dq_sb;
-	int type = dquot->dq_type;
-	dqbuf_t buf = getdqbuf();
-	int ret = 0;
-	uint newblk;
-	__le32 *ref = (__le32 *)buf;
-	
-	if (!buf)
-		return -ENOMEM;
-	if ((ret = read_blk(sb, type, *blk, buf)) < 0) {
-		printk(KERN_ERR "VFS: Can't read quota data block %u\n", *blk);
-		goto out_buf;
-	}
-	newblk = le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]);
-	if (depth == V2_DQTREEDEPTH-1) {
-		ret = free_dqentry(dquot, newblk);
-		newblk = 0;
-	}
-	else
-		ret = remove_tree(dquot, &newblk, depth+1);
-	if (ret >= 0 && !newblk) {
-		int i;
-		ref[GETIDINDEX(dquot->dq_id, depth)] = cpu_to_le32(0);
-		for (i = 0; i < V2_DQBLKSIZE && !buf[i]; i++);	/* Block got empty? */
-		/* Don't put the root block into the free block list */
-		if (i == V2_DQBLKSIZE && *blk != V2_DQTREEOFF) {
-			put_free_dqblk(sb, type, buf, *blk);
-			*blk = 0;
-		}
-		else
-			if ((ret = write_blk(sb, type, *blk, buf)) < 0)
-				printk(KERN_ERR "VFS: Can't write quota tree "
-				  "block %u.\n", *blk);
-	}
-out_buf:
-	freedqbuf(buf);
-	return ret;	
-}
-
-/* Delete dquot from tree */
-static int v2_delete_dquot(struct dquot *dquot)
-{
-	uint tmp = V2_DQTREEOFF;
-
-	if (!dquot->dq_off)	/* Even not allocated? */
-		return 0;
-	return remove_tree(dquot, &tmp, 0);
-}
-
-/* Find entry in block */
-static loff_t find_block_dqentry(struct dquot *dquot, uint blk)
-{
-	dqbuf_t buf = getdqbuf();
-	loff_t ret = 0;
-	int i;
-	struct v2_disk_dqblk *ddquot = GETENTRIES(buf);
-
-	if (!buf)
-		return -ENOMEM;
-	if ((ret = read_blk(dquot->dq_sb, dquot->dq_type, blk, buf)) < 0) {
-		printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk);
-		goto out_buf;
-	}
-	if (dquot->dq_id)
-		for (i = 0; i < V2_DQSTRINBLK &&
-		     le32_to_cpu(ddquot[i].dqb_id) != dquot->dq_id; i++);
-	else {	/* ID 0 as a bit more complicated searching... */
-		struct v2_disk_dqblk fakedquot;
-
-		memset(&fakedquot, 0, sizeof(struct v2_disk_dqblk));
-		for (i = 0; i < V2_DQSTRINBLK; i++)
-			if (!le32_to_cpu(ddquot[i].dqb_id) &&
-			    memcmp(&fakedquot, ddquot+i, sizeof(struct v2_disk_dqblk)))
-				break;
-	}
-	if (i == V2_DQSTRINBLK) {
-		printk(KERN_ERR "VFS: Quota for id %u referenced "
-		  "but not present.\n", dquot->dq_id);
-		ret = -EIO;
-		goto out_buf;
-	}
-	else
-		ret = (blk << V2_DQBLKSIZE_BITS) + sizeof(struct
-		  v2_disk_dqdbheader) + i * sizeof(struct v2_disk_dqblk);
-out_buf:
-	freedqbuf(buf);
-	return ret;
-}
-
-/* Find entry for given id in the tree */
-static loff_t find_tree_dqentry(struct dquot *dquot, uint blk, int depth)
-{
-	dqbuf_t buf = getdqbuf();
-	loff_t ret = 0;
-	__le32 *ref = (__le32 *)buf;
-
-	if (!buf)
-		return -ENOMEM;
-	if ((ret = read_blk(dquot->dq_sb, dquot->dq_type, blk, buf)) < 0) {
-		printk(KERN_ERR "VFS: Can't read quota tree block %u.\n", blk);
-		goto out_buf;
-	}
-	ret = 0;
-	blk = le32_to_cpu(ref[GETIDINDEX(dquot->dq_id, depth)]);
-	if (!blk)	/* No reference? */
-		goto out_buf;
-	if (depth < V2_DQTREEDEPTH-1)
-		ret = find_tree_dqentry(dquot, blk, depth+1);
-	else
-		ret = find_block_dqentry(dquot, blk);
-out_buf:
-	freedqbuf(buf);
-	return ret;
-}
-
-/* Find entry for given id in the tree - wrapper function */
-static inline loff_t find_dqentry(struct dquot *dquot)
-{
-	return find_tree_dqentry(dquot, V2_DQTREEOFF, 0);
-}
-
-static int v2_read_dquot(struct dquot *dquot)
+static int v2_release_dquot(struct dquot *dquot)
 {
-	int type = dquot->dq_type;
-	loff_t offset;
-	struct v2_disk_dqblk ddquot, empty;
-	int ret = 0;
-
-#ifdef __QUOTA_V2_PARANOIA
-	/* Invalidated quota? */
-	if (!dquot->dq_sb || !sb_dqopt(dquot->dq_sb)->files[type]) {
-		printk(KERN_ERR "VFS: Quota invalidated while reading!\n");
-		return -EIO;
-	}
-#endif
-	offset = find_dqentry(dquot);
-	if (offset <= 0) {	/* Entry not present? */
-		if (offset < 0)
-			printk(KERN_ERR "VFS: Can't read quota "
-			  "structure for id %u.\n", dquot->dq_id);
-		dquot->dq_off = 0;
-		set_bit(DQ_FAKE_B, &dquot->dq_flags);
-		memset(&dquot->dq_dqb, 0, sizeof(struct mem_dqblk));
-		ret = offset;
-	}
-	else {
-		dquot->dq_off = offset;
-		if ((ret = dquot->dq_sb->s_op->quota_read(dquot->dq_sb, type,
-		    (char *)&ddquot, sizeof(struct v2_disk_dqblk), offset))
-		    != sizeof(struct v2_disk_dqblk)) {
-			if (ret >= 0)
-				ret = -EIO;
-			printk(KERN_ERR "VFS: Error while reading quota "
-			  "structure for id %u.\n", dquot->dq_id);
-			memset(&ddquot, 0, sizeof(struct v2_disk_dqblk));
-		}
-		else {
-			ret = 0;
-			/* We need to escape back all-zero structure */
-			memset(&empty, 0, sizeof(struct v2_disk_dqblk));
-			empty.dqb_itime = cpu_to_le64(1);
-			if (!memcmp(&empty, &ddquot, sizeof(struct v2_disk_dqblk)))
-				ddquot.dqb_itime = 0;
-		}
-		disk2memdqb(&dquot->dq_dqb, &ddquot);
-		if (!dquot->dq_dqb.dqb_bhardlimit &&
-			!dquot->dq_dqb.dqb_bsoftlimit &&
-			!dquot->dq_dqb.dqb_ihardlimit &&
-			!dquot->dq_dqb.dqb_isoftlimit)
-			set_bit(DQ_FAKE_B, &dquot->dq_flags);
-	}
-	dqstats.reads++;
-
-	return ret;
+	return qtree_release_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_type)->dqi_priv, dquot);
 }
 
-/* Check whether dquot should not be deleted. We know we are
- * the only one operating on dquot (thanks to dq_lock) */
-static int v2_release_dquot(struct dquot *dquot)
+static int v2_free_file_info(struct super_block *sb, int type)
 {
-	if (test_bit(DQ_FAKE_B, &dquot->dq_flags) && !(dquot->dq_dqb.dqb_curinodes | dquot->dq_dqb.dqb_curspace))
-		return v2_delete_dquot(dquot);
+	kfree(sb_dqinfo(sb, type)->dqi_priv);
 	return 0;
 }
 
@@ -673,7 +210,7 @@ static struct quota_format_ops v2_format_ops = {
 	.check_quota_file	= v2_check_quota_file,
 	.read_file_info		= v2_read_file_info,
 	.write_file_info	= v2_write_file_info,
-	.free_file_info		= NULL,
+	.free_file_info		= v2_free_file_info,
 	.read_dqblk		= v2_read_dquot,
 	.commit_dqblk		= v2_write_dquot,
 	.release_dqblk		= v2_release_dquot,
diff --git a/fs/quotaio_v1.h b/fs/quotaio_v1.h
new file mode 100644
index 00000000000..746654b5de7
--- /dev/null
+++ b/fs/quotaio_v1.h
@@ -0,0 +1,33 @@
+#ifndef _LINUX_QUOTAIO_V1_H
+#define _LINUX_QUOTAIO_V1_H
+
+#include <linux/types.h>
+
+/*
+ * The following constants define the amount of time given a user
+ * before the soft limits are treated as hard limits (usually resulting
+ * in an allocation failure). The timer is started when the user crosses
+ * their soft limit, it is reset when they go below their soft limit.
+ */
+#define MAX_IQ_TIME  604800	/* (7*24*60*60) 1 week */
+#define MAX_DQ_TIME  604800	/* (7*24*60*60) 1 week */
+
+/*
+ * The following structure defines the format of the disk quota file
+ * (as it appears on disk) - the file is an array of these structures
+ * indexed by user or group number.
+ */
+struct v1_disk_dqblk {
+	__u32 dqb_bhardlimit;	/* absolute limit on disk blks alloc */
+	__u32 dqb_bsoftlimit;	/* preferred limit on disk blks */
+	__u32 dqb_curblocks;	/* current block count */
+	__u32 dqb_ihardlimit;	/* absolute limit on allocated inodes */
+	__u32 dqb_isoftlimit;	/* preferred inode limit */
+	__u32 dqb_curinodes;	/* current # allocated inodes */
+	time_t dqb_btime;	/* time limit for excessive disk use */
+	time_t dqb_itime;	/* time limit for excessive inode use */
+};
+
+#define v1_dqoff(UID)      ((loff_t)((UID) * sizeof (struct v1_disk_dqblk)))
+
+#endif	/* _LINUX_QUOTAIO_V1_H */
diff --git a/fs/quotaio_v2.h b/fs/quotaio_v2.h
new file mode 100644
index 00000000000..530fe580685
--- /dev/null
+++ b/fs/quotaio_v2.h
@@ -0,0 +1,60 @@
+/*
+ *	Definitions of structures for vfsv0 quota format
+ */
+
+#ifndef _LINUX_QUOTAIO_V2_H
+#define _LINUX_QUOTAIO_V2_H
+
+#include <linux/types.h>
+#include <linux/quota.h>
+
+/*
+ * Definitions of magics and versions of current quota files
+ */
+#define V2_INITQMAGICS {\
+	0xd9c01f11,	/* USRQUOTA */\
+	0xd9c01927	/* GRPQUOTA */\
+}
+
+#define V2_INITQVERSIONS {\
+	0,		/* USRQUOTA */\
+	0		/* GRPQUOTA */\
+}
+
+/* First generic header */
+struct v2_disk_dqheader {
+	__le32 dqh_magic;	/* Magic number identifying file */
+	__le32 dqh_version;	/* File version */
+};
+
+/*
+ * The following structure defines the format of the disk quota file
+ * (as it appears on disk) - the file is a radix tree whose leaves point
+ * to blocks of these structures.
+ */
+struct v2_disk_dqblk {
+	__le32 dqb_id;		/* id this quota applies to */
+	__le32 dqb_ihardlimit;	/* absolute limit on allocated inodes */
+	__le32 dqb_isoftlimit;	/* preferred inode limit */
+	__le32 dqb_curinodes;	/* current # allocated inodes */
+	__le32 dqb_bhardlimit;	/* absolute limit on disk space (in QUOTABLOCK_SIZE) */
+	__le32 dqb_bsoftlimit;	/* preferred limit on disk space (in QUOTABLOCK_SIZE) */
+	__le64 dqb_curspace;	/* current space occupied (in bytes) */
+	__le64 dqb_btime;	/* time limit for excessive disk use */
+	__le64 dqb_itime;	/* time limit for excessive inode use */
+};
+
+/* Header with type and version specific information */
+struct v2_disk_dqinfo {
+	__le32 dqi_bgrace;	/* Time before block soft limit becomes hard limit */
+	__le32 dqi_igrace;	/* Time before inode soft limit becomes hard limit */
+	__le32 dqi_flags;	/* Flags for quotafile (DQF_*) */
+	__le32 dqi_blocks;	/* Number of blocks in file */
+	__le32 dqi_free_blk;	/* Number of first free block in the list */
+	__le32 dqi_free_entry;	/* Number of block with at least one free entry */
+};
+
+#define V2_DQINFOOFF	sizeof(struct v2_disk_dqheader)	/* Offset of info header in file */
+#define V2_DQBLKSIZE_BITS 10				/* Size of leaf block in tree */
+
+#endif /* _LINUX_QUOTAIO_V2_H */
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index 76acdbc3461..b9b567a2837 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -262,11 +262,11 @@ unsigned long ramfs_nommu_get_unmapped_area(struct file *file,
 	ret = -ENOMEM;
 	pages = kzalloc(lpages * sizeof(struct page *), GFP_KERNEL);
 	if (!pages)
-		goto out;
+		goto out_free;
 
 	nr = find_get_pages(inode->i_mapping, pgoff, lpages, pages);
 	if (nr != lpages)
-		goto out; /* leave if some pages were missing */
+		goto out_free_pages; /* leave if some pages were missing */
 
 	/* check the pages for physical adjacency */
 	ptr = pages;
@@ -274,19 +274,18 @@ unsigned long ramfs_nommu_get_unmapped_area(struct file *file,
 	page++;
 	for (loop = lpages; loop > 1; loop--)
 		if (*ptr++ != page++)
-			goto out;
+			goto out_free_pages;
 
 	/* okay - all conditions fulfilled */
 	ret = (unsigned long) page_address(pages[0]);
 
- out:
-	if (pages) {
-		ptr = pages;
-		for (loop = lpages; loop > 0; loop--)
-			put_page(*ptr++);
-		kfree(pages);
-	}
-
+out_free_pages:
+	ptr = pages;
+	for (loop = nr; loop > 0; loop--)
+		put_page(*ptr++);
+out_free:
+	kfree(pages);
+out:
 	return ret;
 }
 
diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c
index a83a3518ae3..b7e6ac706b8 100644
--- a/fs/ramfs/inode.c
+++ b/fs/ramfs/inode.c
@@ -57,7 +57,6 @@ struct inode *ramfs_get_inode(struct super_block *sb, int mode, dev_t dev)
 		inode->i_mode = mode;
 		inode->i_uid = current_fsuid();
 		inode->i_gid = current_fsgid();
-		inode->i_blocks = 0;
 		inode->i_mapping->a_ops = &ramfs_aops;
 		inode->i_mapping->backing_dev_info = &ramfs_backing_dev_info;
 		mapping_set_gfp_mask(inode->i_mapping, GFP_HIGHUSER);
diff --git a/fs/read_write.c b/fs/read_write.c
index 969a6d9c020..5cc6924eb15 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -50,6 +50,14 @@ generic_file_llseek_unlocked(struct file *file, loff_t offset, int origin)
 		offset += inode->i_size;
 		break;
 	case SEEK_CUR:
+		/*
+		 * Here we special-case the lseek(fd, 0, SEEK_CUR)
+		 * position-querying operation.  Avoid rewriting the "same"
+		 * f_pos value back to the file because a concurrent read(),
+		 * write() or lseek() might have altered it
+		 */
+		if (offset == 0)
+			return file->f_pos;
 		offset += file->f_pos;
 		break;
 	}
@@ -105,6 +113,10 @@ loff_t default_llseek(struct file *file, loff_t offset, int origin)
 			offset += i_size_read(file->f_path.dentry->d_inode);
 			break;
 		case SEEK_CUR:
+			if (offset == 0) {
+				retval = file->f_pos;
+				goto out;
+			}
 			offset += file->f_pos;
 	}
 	retval = -EINVAL;
@@ -115,6 +127,7 @@ loff_t default_llseek(struct file *file, loff_t offset, int origin)
 		}
 		retval = offset;
 	}
+out:
 	unlock_kernel();
 	return retval;
 }
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index 6c4c2c69449..55fce92cdf1 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -1753,6 +1753,7 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
 		       struct inode *inode)
 {
 	struct super_block *sb;
+	struct reiserfs_iget_args args;
 	INITIALIZE_PATH(path_to_key);
 	struct cpu_key key;
 	struct item_head ih;
@@ -1780,6 +1781,20 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
 		err = -ENOMEM;
 		goto out_bad_inode;
 	}
+	args.objectid = inode->i_ino = le32_to_cpu(ih.ih_key.k_objectid);
+	if (old_format_only(sb))
+		make_le_item_head(&ih, NULL, KEY_FORMAT_3_5, SD_OFFSET,
+				  TYPE_STAT_DATA, SD_V1_SIZE, MAX_US_INT);
+	else
+		make_le_item_head(&ih, NULL, KEY_FORMAT_3_6, SD_OFFSET,
+				  TYPE_STAT_DATA, SD_SIZE, MAX_US_INT);
+	memcpy(INODE_PKEY(inode), &(ih.ih_key), KEY_SIZE);
+	args.dirid = le32_to_cpu(ih.ih_key.k_dir_id);
+	if (insert_inode_locked4(inode, args.objectid,
+			     reiserfs_find_actor, &args) < 0) {
+		err = -EINVAL;
+		goto out_bad_inode;
+	}
 	if (old_format_only(sb))
 		/* not a perfect generation count, as object ids can be reused, but 
 		 ** this is as good as reiserfs can do right now.
@@ -1825,13 +1840,6 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
 	reiserfs_init_acl_default(inode);
 	reiserfs_init_xattr_rwsem(inode);
 
-	if (old_format_only(sb))
-		make_le_item_head(&ih, NULL, KEY_FORMAT_3_5, SD_OFFSET,
-				  TYPE_STAT_DATA, SD_V1_SIZE, MAX_US_INT);
-	else
-		make_le_item_head(&ih, NULL, KEY_FORMAT_3_6, SD_OFFSET,
-				  TYPE_STAT_DATA, SD_SIZE, MAX_US_INT);
-
 	/* key to search for correct place for new stat data */
 	_make_cpu_key(&key, KEY_FORMAT_3_6, le32_to_cpu(ih.ih_key.k_dir_id),
 		      le32_to_cpu(ih.ih_key.k_objectid), SD_OFFSET,
@@ -1859,13 +1867,9 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
 	} else {
 		inode2sd(&sd, inode, inode->i_size);
 	}
-	// these do not go to on-disk stat data
-	inode->i_ino = le32_to_cpu(ih.ih_key.k_objectid);
-
 	// store in in-core inode the key of stat data and version all
 	// object items will have (directory items will have old offset
 	// format, other new objects will consist of new items)
-	memcpy(INODE_PKEY(inode), &(ih.ih_key), KEY_SIZE);
 	if (old_format_only(sb) || S_ISDIR(mode) || S_ISLNK(mode))
 		set_inode_item_key_version(inode, KEY_FORMAT_3_5);
 	else
@@ -1929,7 +1933,6 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
 		reiserfs_mark_inode_private(inode);
 	}
 
-	insert_inode_hash(inode);
 	reiserfs_update_sd(th, inode);
 	reiserfs_check_path(&path_to_key);
 
@@ -1956,6 +1959,7 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
       out_inserted_sd:
 	inode->i_nlink = 0;
 	th->t_trans_id = 0;	/* so the caller can't use this handle later */
+	unlock_new_inode(inode); /* OK to do even if we hadn't locked it */
 
 	/* If we were inheriting an ACL, we need to release the lock so that
 	 * iput doesn't deadlock in reiserfs_delete_xattrs. The locking
@@ -2556,7 +2560,7 @@ static int reiserfs_write_begin(struct file *file,
 	}
 
 	index = pos >> PAGE_CACHE_SHIFT;
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
 		return -ENOMEM;
 	*pagep = page;
diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c
index 4f322e5ed84..738967f6c8e 100644
--- a/fs/reiserfs/namei.c
+++ b/fs/reiserfs/namei.c
@@ -646,6 +646,7 @@ static int reiserfs_create(struct inode *dir, struct dentry *dentry, int mode,
 		err = journal_end(&th, dir->i_sb, jbegin_count);
 		if (err)
 			retval = err;
+		unlock_new_inode(inode);
 		iput(inode);
 		goto out_failed;
 	}
@@ -653,6 +654,7 @@ static int reiserfs_create(struct inode *dir, struct dentry *dentry, int mode,
 	reiserfs_update_inode_transaction(dir);
 
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 	retval = journal_end(&th, dir->i_sb, jbegin_count);
 
       out_failed:
@@ -727,11 +729,13 @@ static int reiserfs_mknod(struct inode *dir, struct dentry *dentry, int mode,
 		err = journal_end(&th, dir->i_sb, jbegin_count);
 		if (err)
 			retval = err;
+		unlock_new_inode(inode);
 		iput(inode);
 		goto out_failed;
 	}
 
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 	retval = journal_end(&th, dir->i_sb, jbegin_count);
 
       out_failed:
@@ -812,6 +816,7 @@ static int reiserfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 		err = journal_end(&th, dir->i_sb, jbegin_count);
 		if (err)
 			retval = err;
+		unlock_new_inode(inode);
 		iput(inode);
 		goto out_failed;
 	}
@@ -819,6 +824,7 @@ static int reiserfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	reiserfs_update_sd(&th, dir);
 
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 	retval = journal_end(&th, dir->i_sb, jbegin_count);
       out_failed:
 	if (locked)
@@ -1096,11 +1102,13 @@ static int reiserfs_symlink(struct inode *parent_dir,
 		err = journal_end(&th, parent_dir->i_sb, jbegin_count);
 		if (err)
 			retval = err;
+		unlock_new_inode(inode);
 		iput(inode);
 		goto out_failed;
 	}
 
 	d_instantiate(dentry, inode);
+	unlock_new_inode(inode);
 	retval = journal_end(&th, parent_dir->i_sb, jbegin_count);
       out_failed:
 	reiserfs_write_unlock(parent_dir->i_sb);
diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c
index 663a91f5dce..f3c820b7582 100644
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -83,7 +83,7 @@ static void reiserfs_write_super(struct super_block *s)
 	reiserfs_sync_fs(s, 1);
 }
 
-static void reiserfs_write_super_lockfs(struct super_block *s)
+static int reiserfs_freeze(struct super_block *s)
 {
 	struct reiserfs_transaction_handle th;
 	reiserfs_write_lock(s);
@@ -101,11 +101,13 @@ static void reiserfs_write_super_lockfs(struct super_block *s)
 	}
 	s->s_dirt = 0;
 	reiserfs_write_unlock(s);
+	return 0;
 }
 
-static void reiserfs_unlockfs(struct super_block *s)
+static int reiserfs_unfreeze(struct super_block *s)
 {
 	reiserfs_allow_writes(s);
+	return 0;
 }
 
 extern const struct in_core_key MAX_IN_CORE_KEY;
@@ -613,8 +615,8 @@ static const struct super_operations reiserfs_sops = {
 	.put_super = reiserfs_put_super,
 	.write_super = reiserfs_write_super,
 	.sync_fs = reiserfs_sync_fs,
-	.write_super_lockfs = reiserfs_write_super_lockfs,
-	.unlockfs = reiserfs_unlockfs,
+	.freeze_fs = reiserfs_freeze,
+	.unfreeze_fs = reiserfs_unfreeze,
 	.statfs = reiserfs_statfs,
 	.remount_fs = reiserfs_remount,
 	.show_options = generic_show_options,
@@ -649,6 +651,8 @@ static struct dquot_operations reiserfs_quota_operations = {
 	.release_dquot = reiserfs_release_dquot,
 	.mark_dirty = reiserfs_mark_dquot_dirty,
 	.write_info = reiserfs_write_info,
+	.alloc_dquot	= dquot_alloc,
+	.destroy_dquot	= dquot_destroy,
 };
 
 static struct quotactl_ops reiserfs_qctl_operations = {
@@ -994,8 +998,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options,	/* strin
 		if (c == 'u' || c == 'g') {
 			int qtype = c == 'u' ? USRQUOTA : GRPQUOTA;
 
-			if ((sb_any_quota_enabled(s) ||
-			     sb_any_quota_suspended(s)) &&
+			if (sb_any_quota_loaded(s) &&
 			    (!*arg != !REISERFS_SB(s)->s_qf_names[qtype])) {
 				reiserfs_warning(s,
 						 "reiserfs_parse_options: cannot change journaled quota options when quota turned on.");
@@ -1041,8 +1044,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options,	/* strin
 						 "reiserfs_parse_options: unknown quota format specified.");
 				return 0;
 			}
-			if ((sb_any_quota_enabled(s) ||
-			     sb_any_quota_suspended(s)) &&
+			if (sb_any_quota_loaded(s) &&
 			    *qfmt != REISERFS_SB(s)->s_jquota_fmt) {
 				reiserfs_warning(s,
 						 "reiserfs_parse_options: cannot change journaled quota options when quota turned on.");
@@ -1067,7 +1069,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options,	/* strin
 	}
 	/* This checking is not precise wrt the quota type but for our purposes it is sufficient */
 	if (!(*mount_options & (1 << REISERFS_QUOTA))
-	    && sb_any_quota_enabled(s)) {
+	    && sb_any_quota_loaded(s)) {
 		reiserfs_warning(s,
 				 "reiserfs_parse_options: quota options must be present when quota is turned on.");
 		return 0;
diff --git a/fs/romfs/inode.c b/fs/romfs/inode.c
index 60d2f822e87..98a232f7196 100644
--- a/fs/romfs/inode.c
+++ b/fs/romfs/inode.c
@@ -490,7 +490,7 @@ static mode_t romfs_modemap[] =
 static struct inode *
 romfs_iget(struct super_block *sb, unsigned long ino)
 {
-	int nextfh;
+	int nextfh, ret;
 	struct romfs_inode ri;
 	struct inode *i;
 
@@ -524,14 +524,13 @@ romfs_iget(struct super_block *sb, unsigned long ino)
 	i->i_size = be32_to_cpu(ri.size);
 	i->i_mtime.tv_sec = i->i_atime.tv_sec = i->i_ctime.tv_sec = 0;
 	i->i_mtime.tv_nsec = i->i_atime.tv_nsec = i->i_ctime.tv_nsec = 0;
-	i->i_uid = i->i_gid = 0;
 
         /* Precalculate the data offset */
-        ino = romfs_strnlen(i, ino+ROMFH_SIZE, ROMFS_MAXFN);
-        if (ino >= 0)
-                ino = ((ROMFH_SIZE+ino+1+ROMFH_PAD)&ROMFH_MASK);
-        else
-                ino = 0;
+	ret = romfs_strnlen(i, ino + ROMFH_SIZE, ROMFS_MAXFN);
+	if (ret >= 0)
+		ino = (ROMFH_SIZE + ret + 1 + ROMFH_PAD) & ROMFH_MASK;
+	else
+		ino = 0;
 
         ROMFS_I(i)->i_metasize = ino;
         ROMFS_I(i)->i_dataoffset = ino+(i->i_ino&ROMFH_MASK);
diff --git a/fs/select.c b/fs/select.c
index 87df51eadcf..08b91beed80 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -109,11 +109,11 @@ static void __pollwait(struct file *filp, wait_queue_head_t *wait_address,
 void poll_initwait(struct poll_wqueues *pwq)
 {
 	init_poll_funcptr(&pwq->pt, __pollwait);
+	pwq->polling_task = current;
 	pwq->error = 0;
 	pwq->table = NULL;
 	pwq->inline_index = 0;
 }
-
 EXPORT_SYMBOL(poll_initwait);
 
 static void free_poll_entry(struct poll_table_entry *entry)
@@ -142,12 +142,10 @@ void poll_freewait(struct poll_wqueues *pwq)
 		free_page((unsigned long) old);
 	}
 }
-
 EXPORT_SYMBOL(poll_freewait);
 
-static struct poll_table_entry *poll_get_entry(poll_table *_p)
+static struct poll_table_entry *poll_get_entry(struct poll_wqueues *p)
 {
-	struct poll_wqueues *p = container_of(_p, struct poll_wqueues, pt);
 	struct poll_table_page *table = p->table;
 
 	if (p->inline_index < N_INLINE_POLL_ENTRIES)
@@ -159,7 +157,6 @@ static struct poll_table_entry *poll_get_entry(poll_table *_p)
 		new_table = (struct poll_table_page *) __get_free_page(GFP_KERNEL);
 		if (!new_table) {
 			p->error = -ENOMEM;
-			__set_current_state(TASK_RUNNING);
 			return NULL;
 		}
 		new_table->entry = new_table->entries;
@@ -171,20 +168,75 @@ static struct poll_table_entry *poll_get_entry(poll_table *_p)
 	return table->entry++;
 }
 
+static int pollwake(wait_queue_t *wait, unsigned mode, int sync, void *key)
+{
+	struct poll_wqueues *pwq = wait->private;
+	DECLARE_WAITQUEUE(dummy_wait, pwq->polling_task);
+
+	/*
+	 * Although this function is called under waitqueue lock, LOCK
+	 * doesn't imply write barrier and the users expect write
+	 * barrier semantics on wakeup functions.  The following
+	 * smp_wmb() is equivalent to smp_wmb() in try_to_wake_up()
+	 * and is paired with set_mb() in poll_schedule_timeout.
+	 */
+	smp_wmb();
+	pwq->triggered = 1;
+
+	/*
+	 * Perform the default wake up operation using a dummy
+	 * waitqueue.
+	 *
+	 * TODO: This is hacky but there currently is no interface to
+	 * pass in @sync.  @sync is scheduled to be removed and once
+	 * that happens, wake_up_process() can be used directly.
+	 */
+	return default_wake_function(&dummy_wait, mode, sync, key);
+}
+
 /* Add a new entry */
 static void __pollwait(struct file *filp, wait_queue_head_t *wait_address,
 				poll_table *p)
 {
-	struct poll_table_entry *entry = poll_get_entry(p);
+	struct poll_wqueues *pwq = container_of(p, struct poll_wqueues, pt);
+	struct poll_table_entry *entry = poll_get_entry(pwq);
 	if (!entry)
 		return;
 	get_file(filp);
 	entry->filp = filp;
 	entry->wait_address = wait_address;
-	init_waitqueue_entry(&entry->wait, current);
+	init_waitqueue_func_entry(&entry->wait, pollwake);
+	entry->wait.private = pwq;
 	add_wait_queue(wait_address, &entry->wait);
 }
 
+int poll_schedule_timeout(struct poll_wqueues *pwq, int state,
+			  ktime_t *expires, unsigned long slack)
+{
+	int rc = -EINTR;
+
+	set_current_state(state);
+	if (!pwq->triggered)
+		rc = schedule_hrtimeout_range(expires, slack, HRTIMER_MODE_ABS);
+	__set_current_state(TASK_RUNNING);
+
+	/*
+	 * Prepare for the next iteration.
+	 *
+	 * The following set_mb() serves two purposes.  First, it's
+	 * the counterpart rmb of the wmb in pollwake() such that data
+	 * written before wake up is always visible after wake up.
+	 * Second, the full barrier guarantees that triggered clearing
+	 * doesn't pass event check of the next iteration.  Note that
+	 * this problem doesn't exist for the first iteration as
+	 * add_wait_queue() has full barrier semantics.
+	 */
+	set_mb(pwq->triggered, 0);
+
+	return rc;
+}
+EXPORT_SYMBOL(poll_schedule_timeout);
+
 /**
  * poll_select_set_timeout - helper function to setup the timeout value
  * @to:		pointer to timespec variable for the final timeout
@@ -340,8 +392,6 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
 	for (;;) {
 		unsigned long *rinp, *routp, *rexp, *inp, *outp, *exp;
 
-		set_current_state(TASK_INTERRUPTIBLE);
-
 		inp = fds->in; outp = fds->out; exp = fds->ex;
 		rinp = fds->res_in; routp = fds->res_out; rexp = fds->res_ex;
 
@@ -411,10 +461,10 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
 			to = &expire;
 		}
 
-		if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS))
+		if (!poll_schedule_timeout(&table, TASK_INTERRUPTIBLE,
+					   to, slack))
 			timed_out = 1;
 	}
-	__set_current_state(TASK_RUNNING);
 
 	poll_freewait(&table);
 
@@ -666,7 +716,6 @@ static int do_poll(unsigned int nfds,  struct poll_list *list,
 	for (;;) {
 		struct poll_list *walk;
 
-		set_current_state(TASK_INTERRUPTIBLE);
 		for (walk = list; walk != NULL; walk = walk->next) {
 			struct pollfd * pfd, * pfd_end;
 
@@ -709,10 +758,9 @@ static int do_poll(unsigned int nfds,  struct poll_list *list,
 			to = &expire;
 		}
 
-		if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS))
+		if (!poll_schedule_timeout(wait, TASK_INTERRUPTIBLE, to, slack))
 			timed_out = 1;
 	}
-	__set_current_state(TASK_RUNNING);
 	return count;
 }
 
diff --git a/fs/seq_file.c b/fs/seq_file.c
index 16c211558c2..b569ff1c4dc 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -389,8 +389,14 @@ char *mangle_path(char *s, char *p, char *esc)
 }
 EXPORT_SYMBOL(mangle_path);
 
-/*
- * return the absolute path of 'dentry' residing in mount 'mnt'.
+/**
+ * seq_path - seq_file interface to print a pathname
+ * @m: the seq_file handle
+ * @path: the struct path to print
+ * @esc: set of characters to escape in the output
+ *
+ * return the absolute path of 'path', as represented by the
+ * dentry / mnt pair in the path parameter.
  */
 int seq_path(struct seq_file *m, struct path *path, char *esc)
 {
@@ -462,7 +468,8 @@ int seq_dentry(struct seq_file *m, struct dentry *dentry, char *esc)
 	return -1;
 }
 
-int seq_bitmap(struct seq_file *m, unsigned long *bits, unsigned int nr_bits)
+int seq_bitmap(struct seq_file *m, const unsigned long *bits,
+				   unsigned int nr_bits)
 {
 	if (m->count < m->size) {
 		int len = bitmap_scnprintf(m->buf + m->count,
diff --git a/fs/smbfs/file.c b/fs/smbfs/file.c
index e4f8d51a555..92d5e8ffb63 100644
--- a/fs/smbfs/file.c
+++ b/fs/smbfs/file.c
@@ -297,7 +297,7 @@ static int smb_write_begin(struct file *file, struct address_space *mapping,
 			struct page **pagep, void **fsdata)
 {
 	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
-	*pagep = __grab_cache_page(mapping, index);
+	*pagep = grab_cache_page_write_begin(mapping, index, flags);
 	if (!*pagep)
 		return -ENOMEM;
 	return 0;
diff --git a/fs/splice.c b/fs/splice.c
index 1abab5cee4b..a54b3e3f10a 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -21,6 +21,7 @@
 #include <linux/file.h>
 #include <linux/pagemap.h>
 #include <linux/splice.h>
+#include <linux/memcontrol.h>
 #include <linux/mm_inline.h>
 #include <linux/swap.h>
 #include <linux/writeback.h>
diff --git a/fs/squashfs/Makefile b/fs/squashfs/Makefile
new file mode 100644
index 00000000000..8258cf9a031
--- /dev/null
+++ b/fs/squashfs/Makefile
@@ -0,0 +1,8 @@
+#
+# Makefile for the linux squashfs routines.
+#
+
+obj-$(CONFIG_SQUASHFS) += squashfs.o
+squashfs-y += block.o cache.o dir.o export.o file.o fragment.o id.o inode.o
+squashfs-y += namei.o super.o symlink.o
+#squashfs-y += squashfs2_0.o
diff --git a/fs/squashfs/block.c b/fs/squashfs/block.c
new file mode 100644
index 00000000000..c837dfc2b3c
--- /dev/null
+++ b/fs/squashfs/block.c
@@ -0,0 +1,274 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * block.c
+ */
+
+/*
+ * This file implements the low-level routines to read and decompress
+ * datablocks and metadata blocks.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/mutex.h>
+#include <linux/string.h>
+#include <linux/buffer_head.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Read the metadata block length, this is stored in the first two
+ * bytes of the metadata block.
+ */
+static struct buffer_head *get_block_length(struct super_block *sb,
+			u64 *cur_index, int *offset, int *length)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	struct buffer_head *bh;
+
+	bh = sb_bread(sb, *cur_index);
+	if (bh == NULL)
+		return NULL;
+
+	if (msblk->devblksize - *offset == 1) {
+		*length = (unsigned char) bh->b_data[*offset];
+		put_bh(bh);
+		bh = sb_bread(sb, ++(*cur_index));
+		if (bh == NULL)
+			return NULL;
+		*length |= (unsigned char) bh->b_data[0] << 8;
+		*offset = 1;
+	} else {
+		*length = (unsigned char) bh->b_data[*offset] |
+			(unsigned char) bh->b_data[*offset + 1] << 8;
+		*offset += 2;
+	}
+
+	return bh;
+}
+
+
+/*
+ * Read and decompress a metadata block or datablock.  Length is non-zero
+ * if a datablock is being read (the size is stored elsewhere in the
+ * filesystem), otherwise the length is obtained from the first two bytes of
+ * the metadata block.  A bit in the length field indicates if the block
+ * is stored uncompressed in the filesystem (usually because compression
+ * generated a larger block - this does occasionally happen with zlib).
+ */
+int squashfs_read_data(struct super_block *sb, void **buffer, u64 index,
+			int length, u64 *next_index, int srclength)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	struct buffer_head **bh;
+	int offset = index & ((1 << msblk->devblksize_log2) - 1);
+	u64 cur_index = index >> msblk->devblksize_log2;
+	int bytes, compressed, b = 0, k = 0, page = 0, avail;
+
+
+	bh = kcalloc((msblk->block_size >> msblk->devblksize_log2) + 1,
+				sizeof(*bh), GFP_KERNEL);
+	if (bh == NULL)
+		return -ENOMEM;
+
+	if (length) {
+		/*
+		 * Datablock.
+		 */
+		bytes = -offset;
+		compressed = SQUASHFS_COMPRESSED_BLOCK(length);
+		length = SQUASHFS_COMPRESSED_SIZE_BLOCK(length);
+		if (next_index)
+			*next_index = index + length;
+
+		TRACE("Block @ 0x%llx, %scompressed size %d, src size %d\n",
+			index, compressed ? "" : "un", length, srclength);
+
+		if (length < 0 || length > srclength ||
+				(index + length) > msblk->bytes_used)
+			goto read_failure;
+
+		for (b = 0; bytes < length; b++, cur_index++) {
+			bh[b] = sb_getblk(sb, cur_index);
+			if (bh[b] == NULL)
+				goto block_release;
+			bytes += msblk->devblksize;
+		}
+		ll_rw_block(READ, b, bh);
+	} else {
+		/*
+		 * Metadata block.
+		 */
+		if ((index + 2) > msblk->bytes_used)
+			goto read_failure;
+
+		bh[0] = get_block_length(sb, &cur_index, &offset, &length);
+		if (bh[0] == NULL)
+			goto read_failure;
+		b = 1;
+
+		bytes = msblk->devblksize - offset;
+		compressed = SQUASHFS_COMPRESSED(length);
+		length = SQUASHFS_COMPRESSED_SIZE(length);
+		if (next_index)
+			*next_index = index + length + 2;
+
+		TRACE("Block @ 0x%llx, %scompressed size %d\n", index,
+				compressed ? "" : "un", length);
+
+		if (length < 0 || length > srclength ||
+					(index + length) > msblk->bytes_used)
+			goto block_release;
+
+		for (; bytes < length; b++) {
+			bh[b] = sb_getblk(sb, ++cur_index);
+			if (bh[b] == NULL)
+				goto block_release;
+			bytes += msblk->devblksize;
+		}
+		ll_rw_block(READ, b - 1, bh + 1);
+	}
+
+	if (compressed) {
+		int zlib_err = 0, zlib_init = 0;
+
+		/*
+		 * Uncompress block.
+		 */
+
+		mutex_lock(&msblk->read_data_mutex);
+
+		msblk->stream.avail_out = 0;
+		msblk->stream.avail_in = 0;
+
+		bytes = length;
+		do {
+			if (msblk->stream.avail_in == 0 && k < b) {
+				avail = min(bytes, msblk->devblksize - offset);
+				bytes -= avail;
+				wait_on_buffer(bh[k]);
+				if (!buffer_uptodate(bh[k]))
+					goto release_mutex;
+
+				if (avail == 0) {
+					offset = 0;
+					put_bh(bh[k++]);
+					continue;
+				}
+
+				msblk->stream.next_in = bh[k]->b_data + offset;
+				msblk->stream.avail_in = avail;
+				offset = 0;
+			}
+
+			if (msblk->stream.avail_out == 0) {
+				msblk->stream.next_out = buffer[page++];
+				msblk->stream.avail_out = PAGE_CACHE_SIZE;
+			}
+
+			if (!zlib_init) {
+				zlib_err = zlib_inflateInit(&msblk->stream);
+				if (zlib_err != Z_OK) {
+					ERROR("zlib_inflateInit returned"
+						" unexpected result 0x%x,"
+						" srclength %d\n", zlib_err,
+						srclength);
+					goto release_mutex;
+				}
+				zlib_init = 1;
+			}
+
+			zlib_err = zlib_inflate(&msblk->stream, Z_NO_FLUSH);
+
+			if (msblk->stream.avail_in == 0 && k < b)
+				put_bh(bh[k++]);
+		} while (zlib_err == Z_OK);
+
+		if (zlib_err != Z_STREAM_END) {
+			ERROR("zlib_inflate returned unexpected result"
+				" 0x%x, srclength %d, avail_in %d,"
+				" avail_out %d\n", zlib_err, srclength,
+				msblk->stream.avail_in,
+				msblk->stream.avail_out);
+			goto release_mutex;
+		}
+
+		zlib_err = zlib_inflateEnd(&msblk->stream);
+		if (zlib_err != Z_OK) {
+			ERROR("zlib_inflateEnd returned unexpected result 0x%x,"
+				" srclength %d\n", zlib_err, srclength);
+			goto release_mutex;
+		}
+		length = msblk->stream.total_out;
+		mutex_unlock(&msblk->read_data_mutex);
+	} else {
+		/*
+		 * Block is uncompressed.
+		 */
+		int i, in, pg_offset = 0;
+
+		for (i = 0; i < b; i++) {
+			wait_on_buffer(bh[i]);
+			if (!buffer_uptodate(bh[i]))
+				goto block_release;
+		}
+
+		for (bytes = length; k < b; k++) {
+			in = min(bytes, msblk->devblksize - offset);
+			bytes -= in;
+			while (in) {
+				if (pg_offset == PAGE_CACHE_SIZE) {
+					page++;
+					pg_offset = 0;
+				}
+				avail = min_t(int, in, PAGE_CACHE_SIZE -
+						pg_offset);
+				memcpy(buffer[page] + pg_offset,
+						bh[k]->b_data + offset, avail);
+				in -= avail;
+				pg_offset += avail;
+				offset += avail;
+			}
+			offset = 0;
+			put_bh(bh[k]);
+		}
+	}
+
+	kfree(bh);
+	return length;
+
+release_mutex:
+	mutex_unlock(&msblk->read_data_mutex);
+
+block_release:
+	for (; k < b; k++)
+		put_bh(bh[k]);
+
+read_failure:
+	ERROR("sb_bread failed reading block 0x%llx\n", cur_index);
+	kfree(bh);
+	return -EIO;
+}
diff --git a/fs/squashfs/cache.c b/fs/squashfs/cache.c
new file mode 100644
index 00000000000..f29eda16d25
--- /dev/null
+++ b/fs/squashfs/cache.c
@@ -0,0 +1,412 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * cache.c
+ */
+
+/*
+ * Blocks in Squashfs are compressed.  To avoid repeatedly decompressing
+ * recently accessed data Squashfs uses two small metadata and fragment caches.
+ *
+ * This file implements a generic cache implementation used for both caches,
+ * plus functions layered ontop of the generic cache implementation to
+ * access the metadata and fragment caches.
+ *
+ * To avoid out of memory and fragmentation isssues with vmalloc the cache
+ * uses sequences of kmalloced PAGE_CACHE_SIZE buffers.
+ *
+ * It should be noted that the cache is not used for file datablocks, these
+ * are decompressed and cached in the page-cache in the normal way.  The
+ * cache is only used to temporarily cache fragment and metadata blocks
+ * which have been read as as a result of a metadata (i.e. inode or
+ * directory) or fragment access.  Because metadata and fragments are packed
+ * together into blocks (to gain greater compression) the read of a particular
+ * piece of metadata or fragment will retrieve other metadata/fragments which
+ * have been packed with it, these because of locality-of-reference may be read
+ * in the near future. Temporarily caching them ensures they are available for
+ * near future access without requiring an additional read and decompress.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/sched.h>
+#include <linux/spinlock.h>
+#include <linux/wait.h>
+#include <linux/zlib.h>
+#include <linux/pagemap.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Look-up block in cache, and increment usage count.  If not in cache, read
+ * and decompress it from disk.
+ */
+struct squashfs_cache_entry *squashfs_cache_get(struct super_block *sb,
+	struct squashfs_cache *cache, u64 block, int length)
+{
+	int i, n;
+	struct squashfs_cache_entry *entry;
+
+	spin_lock(&cache->lock);
+
+	while (1) {
+		for (i = 0; i < cache->entries; i++)
+			if (cache->entry[i].block == block)
+				break;
+
+		if (i == cache->entries) {
+			/*
+			 * Block not in cache, if all cache entries are used
+			 * go to sleep waiting for one to become available.
+			 */
+			if (cache->unused == 0) {
+				cache->num_waiters++;
+				spin_unlock(&cache->lock);
+				wait_event(cache->wait_queue, cache->unused);
+				spin_lock(&cache->lock);
+				cache->num_waiters--;
+				continue;
+			}
+
+			/*
+			 * At least one unused cache entry.  A simple
+			 * round-robin strategy is used to choose the entry to
+			 * be evicted from the cache.
+			 */
+			i = cache->next_blk;
+			for (n = 0; n < cache->entries; n++) {
+				if (cache->entry[i].refcount == 0)
+					break;
+				i = (i + 1) % cache->entries;
+			}
+
+			cache->next_blk = (i + 1) % cache->entries;
+			entry = &cache->entry[i];
+
+			/*
+			 * Initialise choosen cache entry, and fill it in from
+			 * disk.
+			 */
+			cache->unused--;
+			entry->block = block;
+			entry->refcount = 1;
+			entry->pending = 1;
+			entry->num_waiters = 0;
+			entry->error = 0;
+			spin_unlock(&cache->lock);
+
+			entry->length = squashfs_read_data(sb, entry->data,
+				block, length, &entry->next_index,
+				cache->block_size);
+
+			spin_lock(&cache->lock);
+
+			if (entry->length < 0)
+				entry->error = entry->length;
+
+			entry->pending = 0;
+
+			/*
+			 * While filling this entry one or more other processes
+			 * have looked it up in the cache, and have slept
+			 * waiting for it to become available.
+			 */
+			if (entry->num_waiters) {
+				spin_unlock(&cache->lock);
+				wake_up_all(&entry->wait_queue);
+			} else
+				spin_unlock(&cache->lock);
+
+			goto out;
+		}
+
+		/*
+		 * Block already in cache.  Increment refcount so it doesn't
+		 * get reused until we're finished with it, if it was
+		 * previously unused there's one less cache entry available
+		 * for reuse.
+		 */
+		entry = &cache->entry[i];
+		if (entry->refcount == 0)
+			cache->unused--;
+		entry->refcount++;
+
+		/*
+		 * If the entry is currently being filled in by another process
+		 * go to sleep waiting for it to become available.
+		 */
+		if (entry->pending) {
+			entry->num_waiters++;
+			spin_unlock(&cache->lock);
+			wait_event(entry->wait_queue, !entry->pending);
+		} else
+			spin_unlock(&cache->lock);
+
+		goto out;
+	}
+
+out:
+	TRACE("Got %s %d, start block %lld, refcount %d, error %d\n",
+		cache->name, i, entry->block, entry->refcount, entry->error);
+
+	if (entry->error)
+		ERROR("Unable to read %s cache entry [%llx]\n", cache->name,
+							block);
+	return entry;
+}
+
+
+/*
+ * Release cache entry, once usage count is zero it can be reused.
+ */
+void squashfs_cache_put(struct squashfs_cache_entry *entry)
+{
+	struct squashfs_cache *cache = entry->cache;
+
+	spin_lock(&cache->lock);
+	entry->refcount--;
+	if (entry->refcount == 0) {
+		cache->unused++;
+		/*
+		 * If there's any processes waiting for a block to become
+		 * available, wake one up.
+		 */
+		if (cache->num_waiters) {
+			spin_unlock(&cache->lock);
+			wake_up(&cache->wait_queue);
+			return;
+		}
+	}
+	spin_unlock(&cache->lock);
+}
+
+/*
+ * Delete cache reclaiming all kmalloced buffers.
+ */
+void squashfs_cache_delete(struct squashfs_cache *cache)
+{
+	int i, j;
+
+	if (cache == NULL)
+		return;
+
+	for (i = 0; i < cache->entries; i++) {
+		if (cache->entry[i].data) {
+			for (j = 0; j < cache->pages; j++)
+				kfree(cache->entry[i].data[j]);
+			kfree(cache->entry[i].data);
+		}
+	}
+
+	kfree(cache->entry);
+	kfree(cache);
+}
+
+
+/*
+ * Initialise cache allocating the specified number of entries, each of
+ * size block_size.  To avoid vmalloc fragmentation issues each entry
+ * is allocated as a sequence of kmalloced PAGE_CACHE_SIZE buffers.
+ */
+struct squashfs_cache *squashfs_cache_init(char *name, int entries,
+	int block_size)
+{
+	int i, j;
+	struct squashfs_cache *cache = kzalloc(sizeof(*cache), GFP_KERNEL);
+
+	if (cache == NULL) {
+		ERROR("Failed to allocate %s cache\n", name);
+		return NULL;
+	}
+
+	cache->entry = kcalloc(entries, sizeof(*(cache->entry)), GFP_KERNEL);
+	if (cache->entry == NULL) {
+		ERROR("Failed to allocate %s cache\n", name);
+		goto cleanup;
+	}
+
+	cache->next_blk = 0;
+	cache->unused = entries;
+	cache->entries = entries;
+	cache->block_size = block_size;
+	cache->pages = block_size >> PAGE_CACHE_SHIFT;
+	cache->name = name;
+	cache->num_waiters = 0;
+	spin_lock_init(&cache->lock);
+	init_waitqueue_head(&cache->wait_queue);
+
+	for (i = 0; i < entries; i++) {
+		struct squashfs_cache_entry *entry = &cache->entry[i];
+
+		init_waitqueue_head(&cache->entry[i].wait_queue);
+		entry->cache = cache;
+		entry->block = SQUASHFS_INVALID_BLK;
+		entry->data = kcalloc(cache->pages, sizeof(void *), GFP_KERNEL);
+		if (entry->data == NULL) {
+			ERROR("Failed to allocate %s cache entry\n", name);
+			goto cleanup;
+		}
+
+		for (j = 0; j < cache->pages; j++) {
+			entry->data[j] = kmalloc(PAGE_CACHE_SIZE, GFP_KERNEL);
+			if (entry->data[j] == NULL) {
+				ERROR("Failed to allocate %s buffer\n", name);
+				goto cleanup;
+			}
+		}
+	}
+
+	return cache;
+
+cleanup:
+	squashfs_cache_delete(cache);
+	return NULL;
+}
+
+
+/*
+ * Copy upto length bytes from cache entry to buffer starting at offset bytes
+ * into the cache entry.  If there's not length bytes then copy the number of
+ * bytes available.  In all cases return the number of bytes copied.
+ */
+int squashfs_copy_data(void *buffer, struct squashfs_cache_entry *entry,
+		int offset, int length)
+{
+	int remaining = length;
+
+	if (length == 0)
+		return 0;
+	else if (buffer == NULL)
+		return min(length, entry->length - offset);
+
+	while (offset < entry->length) {
+		void *buff = entry->data[offset / PAGE_CACHE_SIZE]
+				+ (offset % PAGE_CACHE_SIZE);
+		int bytes = min_t(int, entry->length - offset,
+				PAGE_CACHE_SIZE - (offset % PAGE_CACHE_SIZE));
+
+		if (bytes >= remaining) {
+			memcpy(buffer, buff, remaining);
+			remaining = 0;
+			break;
+		}
+
+		memcpy(buffer, buff, bytes);
+		buffer += bytes;
+		remaining -= bytes;
+		offset += bytes;
+	}
+
+	return length - remaining;
+}
+
+
+/*
+ * Read length bytes from metadata position <block, offset> (block is the
+ * start of the compressed block on disk, and offset is the offset into
+ * the block once decompressed).  Data is packed into consecutive blocks,
+ * and length bytes may require reading more than one block.
+ */
+int squashfs_read_metadata(struct super_block *sb, void *buffer,
+		u64 *block, int *offset, int length)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int bytes, copied = length;
+	struct squashfs_cache_entry *entry;
+
+	TRACE("Entered squashfs_read_metadata [%llx:%x]\n", *block, *offset);
+
+	while (length) {
+		entry = squashfs_cache_get(sb, msblk->block_cache, *block, 0);
+		if (entry->error)
+			return entry->error;
+		else if (*offset >= entry->length)
+			return -EIO;
+
+		bytes = squashfs_copy_data(buffer, entry, *offset, length);
+		if (buffer)
+			buffer += bytes;
+		length -= bytes;
+		*offset += bytes;
+
+		if (*offset == entry->length) {
+			*block = entry->next_index;
+			*offset = 0;
+		}
+
+		squashfs_cache_put(entry);
+	}
+
+	return copied;
+}
+
+
+/*
+ * Look-up in the fragmment cache the fragment located at <start_block> in the
+ * filesystem.  If necessary read and decompress it from disk.
+ */
+struct squashfs_cache_entry *squashfs_get_fragment(struct super_block *sb,
+				u64 start_block, int length)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+
+	return squashfs_cache_get(sb, msblk->fragment_cache, start_block,
+		length);
+}
+
+
+/*
+ * Read and decompress the datablock located at <start_block> in the
+ * filesystem.  The cache is used here to avoid duplicating locking and
+ * read/decompress code.
+ */
+struct squashfs_cache_entry *squashfs_get_datablock(struct super_block *sb,
+				u64 start_block, int length)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+
+	return squashfs_cache_get(sb, msblk->read_page, start_block, length);
+}
+
+
+/*
+ * Read a filesystem table (uncompressed sequence of bytes) from disk
+ */
+int squashfs_read_table(struct super_block *sb, void *buffer, u64 block,
+	int length)
+{
+	int pages = (length + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+	int i, res;
+	void **data = kcalloc(pages, sizeof(void *), GFP_KERNEL);
+	if (data == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < pages; i++, buffer += PAGE_CACHE_SIZE)
+		data[i] = buffer;
+	res = squashfs_read_data(sb, data, block, length |
+		SQUASHFS_COMPRESSED_BIT_BLOCK, NULL, length);
+	kfree(data);
+	return res;
+}
diff --git a/fs/squashfs/dir.c b/fs/squashfs/dir.c
new file mode 100644
index 00000000000..566b0eaed86
--- /dev/null
+++ b/fs/squashfs/dir.c
@@ -0,0 +1,235 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * dir.c
+ */
+
+/*
+ * This file implements code to read directories from disk.
+ *
+ * See namei.c for a description of directory organisation on disk.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+static const unsigned char squashfs_filetype_table[] = {
+	DT_UNKNOWN, DT_DIR, DT_REG, DT_LNK, DT_BLK, DT_CHR, DT_FIFO, DT_SOCK
+};
+
+/*
+ * Lookup offset (f_pos) in the directory index, returning the
+ * metadata block containing it.
+ *
+ * If we get an error reading the index then return the part of the index
+ * (if any) we have managed to read - the index isn't essential, just
+ * quicker.
+ */
+static int get_dir_index_using_offset(struct super_block *sb,
+	u64 *next_block, int *next_offset, u64 index_start, int index_offset,
+	int i_count, u64 f_pos)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int err, i, index, length = 0;
+	struct squashfs_dir_index dir_index;
+
+	TRACE("Entered get_dir_index_using_offset, i_count %d, f_pos %lld\n",
+					i_count, f_pos);
+
+	/*
+	 * Translate from external f_pos to the internal f_pos.  This
+	 * is offset by 3 because we invent "." and ".." entries which are
+	 * not actually stored in the directory.
+	 */
+	if (f_pos < 3)
+		return f_pos;
+	f_pos -= 3;
+
+	for (i = 0; i < i_count; i++) {
+		err = squashfs_read_metadata(sb, &dir_index, &index_start,
+				&index_offset, sizeof(dir_index));
+		if (err < 0)
+			break;
+
+		index = le32_to_cpu(dir_index.index);
+		if (index > f_pos)
+			/*
+			 * Found the index we're looking for.
+			 */
+			break;
+
+		err = squashfs_read_metadata(sb, NULL, &index_start,
+				&index_offset, le32_to_cpu(dir_index.size) + 1);
+		if (err < 0)
+			break;
+
+		length = index;
+		*next_block = le32_to_cpu(dir_index.start_block) +
+					msblk->directory_table;
+	}
+
+	*next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE;
+
+	/*
+	 * Translate back from internal f_pos to external f_pos.
+	 */
+	return length + 3;
+}
+
+
+static int squashfs_readdir(struct file *file, void *dirent, filldir_t filldir)
+{
+	struct inode *inode = file->f_dentry->d_inode;
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	u64 block = squashfs_i(inode)->start + msblk->directory_table;
+	int offset = squashfs_i(inode)->offset, length = 0, dir_count, size,
+				type, err;
+	unsigned int inode_number;
+	struct squashfs_dir_header dirh;
+	struct squashfs_dir_entry *dire;
+
+	TRACE("Entered squashfs_readdir [%llx:%x]\n", block, offset);
+
+	dire = kmalloc(sizeof(*dire) + SQUASHFS_NAME_LEN + 1, GFP_KERNEL);
+	if (dire == NULL) {
+		ERROR("Failed to allocate squashfs_dir_entry\n");
+		goto finish;
+	}
+
+	/*
+	 * Return "." and  ".." entries as the first two filenames in the
+	 * directory.  To maximise compression these two entries are not
+	 * stored in the directory, and so we invent them here.
+	 *
+	 * It also means that the external f_pos is offset by 3 from the
+	 * on-disk directory f_pos.
+	 */
+	while (file->f_pos < 3) {
+		char *name;
+		int i_ino;
+
+		if (file->f_pos == 0) {
+			name = ".";
+			size = 1;
+			i_ino = inode->i_ino;
+		} else {
+			name = "..";
+			size = 2;
+			i_ino = squashfs_i(inode)->parent;
+		}
+
+		TRACE("Calling filldir(%p, %s, %d, %lld, %d, %d)\n",
+				dirent, name, size, file->f_pos, i_ino,
+				squashfs_filetype_table[1]);
+
+		if (filldir(dirent, name, size, file->f_pos, i_ino,
+				squashfs_filetype_table[1]) < 0) {
+				TRACE("Filldir returned less than 0\n");
+			goto finish;
+		}
+
+		file->f_pos += size;
+	}
+
+	length = get_dir_index_using_offset(inode->i_sb, &block, &offset,
+				squashfs_i(inode)->dir_idx_start,
+				squashfs_i(inode)->dir_idx_offset,
+				squashfs_i(inode)->dir_idx_cnt,
+				file->f_pos);
+
+	while (length < i_size_read(inode)) {
+		/*
+		 * Read directory header
+		 */
+		err = squashfs_read_metadata(inode->i_sb, &dirh, &block,
+					&offset, sizeof(dirh));
+		if (err < 0)
+			goto failed_read;
+
+		length += sizeof(dirh);
+
+		dir_count = le32_to_cpu(dirh.count) + 1;
+		while (dir_count--) {
+			/*
+			 * Read directory entry.
+			 */
+			err = squashfs_read_metadata(inode->i_sb, dire, &block,
+					&offset, sizeof(*dire));
+			if (err < 0)
+				goto failed_read;
+
+			size = le16_to_cpu(dire->size) + 1;
+
+			err = squashfs_read_metadata(inode->i_sb, dire->name,
+					&block, &offset, size);
+			if (err < 0)
+				goto failed_read;
+
+			length += sizeof(*dire) + size;
+
+			if (file->f_pos >= length)
+				continue;
+
+			dire->name[size] = '\0';
+			inode_number = le32_to_cpu(dirh.inode_number) +
+				((short) le16_to_cpu(dire->inode_number));
+			type = le16_to_cpu(dire->type);
+
+			TRACE("Calling filldir(%p, %s, %d, %lld, %x:%x, %d, %d)"
+					"\n", dirent, dire->name, size,
+					file->f_pos,
+					le32_to_cpu(dirh.start_block),
+					le16_to_cpu(dire->offset),
+					inode_number,
+					squashfs_filetype_table[type]);
+
+			if (filldir(dirent, dire->name, size, file->f_pos,
+					inode_number,
+					squashfs_filetype_table[type]) < 0) {
+				TRACE("Filldir returned less than 0\n");
+				goto finish;
+			}
+
+			file->f_pos = length;
+		}
+	}
+
+finish:
+	kfree(dire);
+	return 0;
+
+failed_read:
+	ERROR("Unable to read directory block [%llx:%x]\n", block, offset);
+	kfree(dire);
+	return 0;
+}
+
+
+const struct file_operations squashfs_dir_ops = {
+	.read = generic_read_dir,
+	.readdir = squashfs_readdir
+};
diff --git a/fs/squashfs/export.c b/fs/squashfs/export.c
new file mode 100644
index 00000000000..69e971d5ddc
--- /dev/null
+++ b/fs/squashfs/export.c
@@ -0,0 +1,155 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * export.c
+ */
+
+/*
+ * This file implements code to make Squashfs filesystems exportable (NFS etc.)
+ *
+ * The export code uses an inode lookup table to map inode numbers passed in
+ * filehandles to an inode location on disk.  This table is stored compressed
+ * into metadata blocks.  A second index table is used to locate these.  This
+ * second index table for speed of access (and because it is small) is read at
+ * mount time and cached in memory.
+ *
+ * The inode lookup table is used only by the export code, inode disk
+ * locations are directly encoded in directories, enabling direct access
+ * without an intermediate lookup for all operations except the export ops.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/dcache.h>
+#include <linux/exportfs.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Look-up inode number (ino) in table, returning the inode location.
+ */
+static long long squashfs_inode_lookup(struct super_block *sb, int ino_num)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int blk = SQUASHFS_LOOKUP_BLOCK(ino_num - 1);
+	int offset = SQUASHFS_LOOKUP_BLOCK_OFFSET(ino_num - 1);
+	u64 start = le64_to_cpu(msblk->inode_lookup_table[blk]);
+	__le64 ino;
+	int err;
+
+	TRACE("Entered squashfs_inode_lookup, inode_number = %d\n", ino_num);
+
+	err = squashfs_read_metadata(sb, &ino, &start, &offset, sizeof(ino));
+	if (err < 0)
+		return err;
+
+	TRACE("squashfs_inode_lookup, inode = 0x%llx\n",
+		(u64) le64_to_cpu(ino));
+
+	return le64_to_cpu(ino);
+}
+
+
+static struct dentry *squashfs_export_iget(struct super_block *sb,
+	unsigned int ino_num)
+{
+	long long ino;
+	struct dentry *dentry = ERR_PTR(-ENOENT);
+
+	TRACE("Entered squashfs_export_iget\n");
+
+	ino = squashfs_inode_lookup(sb, ino_num);
+	if (ino >= 0)
+		dentry = d_obtain_alias(squashfs_iget(sb, ino, ino_num));
+
+	return dentry;
+}
+
+
+static struct dentry *squashfs_fh_to_dentry(struct super_block *sb,
+		struct fid *fid, int fh_len, int fh_type)
+{
+	if ((fh_type != FILEID_INO32_GEN && fh_type != FILEID_INO32_GEN_PARENT)
+			|| fh_len < 2)
+		return NULL;
+
+	return squashfs_export_iget(sb, fid->i32.ino);
+}
+
+
+static struct dentry *squashfs_fh_to_parent(struct super_block *sb,
+		struct fid *fid, int fh_len, int fh_type)
+{
+	if (fh_type != FILEID_INO32_GEN_PARENT || fh_len < 4)
+		return NULL;
+
+	return squashfs_export_iget(sb, fid->i32.parent_ino);
+}
+
+
+static struct dentry *squashfs_get_parent(struct dentry *child)
+{
+	struct inode *inode = child->d_inode;
+	unsigned int parent_ino = squashfs_i(inode)->parent;
+
+	return squashfs_export_iget(inode->i_sb, parent_ino);
+}
+
+
+/*
+ * Read uncompressed inode lookup table indexes off disk into memory
+ */
+__le64 *squashfs_read_inode_lookup_table(struct super_block *sb,
+		u64 lookup_table_start, unsigned int inodes)
+{
+	unsigned int length = SQUASHFS_LOOKUP_BLOCK_BYTES(inodes);
+	__le64 *inode_lookup_table;
+	int err;
+
+	TRACE("In read_inode_lookup_table, length %d\n", length);
+
+	/* Allocate inode lookup table indexes */
+	inode_lookup_table = kmalloc(length, GFP_KERNEL);
+	if (inode_lookup_table == NULL) {
+		ERROR("Failed to allocate inode lookup table\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	err = squashfs_read_table(sb, inode_lookup_table, lookup_table_start,
+			length);
+	if (err < 0) {
+		ERROR("unable to read inode lookup table\n");
+		kfree(inode_lookup_table);
+		return ERR_PTR(err);
+	}
+
+	return inode_lookup_table;
+}
+
+
+const struct export_operations squashfs_export_ops = {
+	.fh_to_dentry = squashfs_fh_to_dentry,
+	.fh_to_parent = squashfs_fh_to_parent,
+	.get_parent = squashfs_get_parent
+};
diff --git a/fs/squashfs/file.c b/fs/squashfs/file.c
new file mode 100644
index 00000000000..717767d831d
--- /dev/null
+++ b/fs/squashfs/file.c
@@ -0,0 +1,502 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * file.c
+ */
+
+/*
+ * This file contains code for handling regular files.  A regular file
+ * consists of a sequence of contiguous compressed blocks, and/or a
+ * compressed fragment block (tail-end packed block).   The compressed size
+ * of each datablock is stored in a block list contained within the
+ * file inode (itself stored in one or more compressed metadata blocks).
+ *
+ * To speed up access to datablocks when reading 'large' files (256 Mbytes or
+ * larger), the code implements an index cache that caches the mapping from
+ * block index to datablock location on disk.
+ *
+ * The index cache allows Squashfs to handle large files (up to 1.75 TiB) while
+ * retaining a simple and space-efficient block list on disk.  The cache
+ * is split into slots, caching up to eight 224 GiB files (128 KiB blocks).
+ * Larger files use multiple slots, with 1.75 TiB files using all 8 slots.
+ * The index cache is designed to be memory efficient, and by default uses
+ * 16 KiB.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/pagemap.h>
+#include <linux/mutex.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Locate cache slot in range [offset, index] for specified inode.  If
+ * there's more than one return the slot closest to index.
+ */
+static struct meta_index *locate_meta_index(struct inode *inode, int offset,
+				int index)
+{
+	struct meta_index *meta = NULL;
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	int i;
+
+	mutex_lock(&msblk->meta_index_mutex);
+
+	TRACE("locate_meta_index: index %d, offset %d\n", index, offset);
+
+	if (msblk->meta_index == NULL)
+		goto not_allocated;
+
+	for (i = 0; i < SQUASHFS_META_SLOTS; i++) {
+		if (msblk->meta_index[i].inode_number == inode->i_ino &&
+				msblk->meta_index[i].offset >= offset &&
+				msblk->meta_index[i].offset <= index &&
+				msblk->meta_index[i].locked == 0) {
+			TRACE("locate_meta_index: entry %d, offset %d\n", i,
+					msblk->meta_index[i].offset);
+			meta = &msblk->meta_index[i];
+			offset = meta->offset;
+		}
+	}
+
+	if (meta)
+		meta->locked = 1;
+
+not_allocated:
+	mutex_unlock(&msblk->meta_index_mutex);
+
+	return meta;
+}
+
+
+/*
+ * Find and initialise an empty cache slot for index offset.
+ */
+static struct meta_index *empty_meta_index(struct inode *inode, int offset,
+				int skip)
+{
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	struct meta_index *meta = NULL;
+	int i;
+
+	mutex_lock(&msblk->meta_index_mutex);
+
+	TRACE("empty_meta_index: offset %d, skip %d\n", offset, skip);
+
+	if (msblk->meta_index == NULL) {
+		/*
+		 * First time cache index has been used, allocate and
+		 * initialise.  The cache index could be allocated at
+		 * mount time but doing it here means it is allocated only
+		 * if a 'large' file is read.
+		 */
+		msblk->meta_index = kcalloc(SQUASHFS_META_SLOTS,
+			sizeof(*(msblk->meta_index)), GFP_KERNEL);
+		if (msblk->meta_index == NULL) {
+			ERROR("Failed to allocate meta_index\n");
+			goto failed;
+		}
+		for (i = 0; i < SQUASHFS_META_SLOTS; i++) {
+			msblk->meta_index[i].inode_number = 0;
+			msblk->meta_index[i].locked = 0;
+		}
+		msblk->next_meta_index = 0;
+	}
+
+	for (i = SQUASHFS_META_SLOTS; i &&
+			msblk->meta_index[msblk->next_meta_index].locked; i--)
+		msblk->next_meta_index = (msblk->next_meta_index + 1) %
+			SQUASHFS_META_SLOTS;
+
+	if (i == 0) {
+		TRACE("empty_meta_index: failed!\n");
+		goto failed;
+	}
+
+	TRACE("empty_meta_index: returned meta entry %d, %p\n",
+			msblk->next_meta_index,
+			&msblk->meta_index[msblk->next_meta_index]);
+
+	meta = &msblk->meta_index[msblk->next_meta_index];
+	msblk->next_meta_index = (msblk->next_meta_index + 1) %
+			SQUASHFS_META_SLOTS;
+
+	meta->inode_number = inode->i_ino;
+	meta->offset = offset;
+	meta->skip = skip;
+	meta->entries = 0;
+	meta->locked = 1;
+
+failed:
+	mutex_unlock(&msblk->meta_index_mutex);
+	return meta;
+}
+
+
+static void release_meta_index(struct inode *inode, struct meta_index *meta)
+{
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	mutex_lock(&msblk->meta_index_mutex);
+	meta->locked = 0;
+	mutex_unlock(&msblk->meta_index_mutex);
+}
+
+
+/*
+ * Read the next n blocks from the block list, starting from
+ * metadata block <start_block, offset>.
+ */
+static long long read_indexes(struct super_block *sb, int n,
+				u64 *start_block, int *offset)
+{
+	int err, i;
+	long long block = 0;
+	__le32 *blist = kmalloc(PAGE_CACHE_SIZE, GFP_KERNEL);
+
+	if (blist == NULL) {
+		ERROR("read_indexes: Failed to allocate block_list\n");
+		return -ENOMEM;
+	}
+
+	while (n) {
+		int blocks = min_t(int, n, PAGE_CACHE_SIZE >> 2);
+
+		err = squashfs_read_metadata(sb, blist, start_block,
+				offset, blocks << 2);
+		if (err < 0) {
+			ERROR("read_indexes: reading block [%llx:%x]\n",
+				*start_block, *offset);
+			goto failure;
+		}
+
+		for (i = 0; i < blocks; i++) {
+			int size = le32_to_cpu(blist[i]);
+			block += SQUASHFS_COMPRESSED_SIZE_BLOCK(size);
+		}
+		n -= blocks;
+	}
+
+	kfree(blist);
+	return block;
+
+failure:
+	kfree(blist);
+	return err;
+}
+
+
+/*
+ * Each cache index slot has SQUASHFS_META_ENTRIES, each of which
+ * can cache one index -> datablock/blocklist-block mapping.  We wish
+ * to distribute these over the length of the file, entry[0] maps index x,
+ * entry[1] maps index x + skip, entry[2] maps index x + 2 * skip, and so on.
+ * The larger the file, the greater the skip factor.  The skip factor is
+ * limited to the size of the metadata cache (SQUASHFS_CACHED_BLKS) to ensure
+ * the number of metadata blocks that need to be read fits into the cache.
+ * If the skip factor is limited in this way then the file will use multiple
+ * slots.
+ */
+static inline int calculate_skip(int blocks)
+{
+	int skip = blocks / ((SQUASHFS_META_ENTRIES + 1)
+		 * SQUASHFS_META_INDEXES);
+	return min(SQUASHFS_CACHED_BLKS - 1, skip + 1);
+}
+
+
+/*
+ * Search and grow the index cache for the specified inode, returning the
+ * on-disk locations of the datablock and block list metadata block
+ * <index_block, index_offset> for index (scaled to nearest cache index).
+ */
+static int fill_meta_index(struct inode *inode, int index,
+		u64 *index_block, int *index_offset, u64 *data_block)
+{
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	int skip = calculate_skip(i_size_read(inode) >> msblk->block_log);
+	int offset = 0;
+	struct meta_index *meta;
+	struct meta_entry *meta_entry;
+	u64 cur_index_block = squashfs_i(inode)->block_list_start;
+	int cur_offset = squashfs_i(inode)->offset;
+	u64 cur_data_block = squashfs_i(inode)->start;
+	int err, i;
+
+	/*
+	 * Scale index to cache index (cache slot entry)
+	 */
+	index /= SQUASHFS_META_INDEXES * skip;
+
+	while (offset < index) {
+		meta = locate_meta_index(inode, offset + 1, index);
+
+		if (meta == NULL) {
+			meta = empty_meta_index(inode, offset + 1, skip);
+			if (meta == NULL)
+				goto all_done;
+		} else {
+			offset = index < meta->offset + meta->entries ? index :
+				meta->offset + meta->entries - 1;
+			meta_entry = &meta->meta_entry[offset - meta->offset];
+			cur_index_block = meta_entry->index_block +
+				msblk->inode_table;
+			cur_offset = meta_entry->offset;
+			cur_data_block = meta_entry->data_block;
+			TRACE("get_meta_index: offset %d, meta->offset %d, "
+				"meta->entries %d\n", offset, meta->offset,
+				meta->entries);
+			TRACE("get_meta_index: index_block 0x%llx, offset 0x%x"
+				" data_block 0x%llx\n", cur_index_block,
+				cur_offset, cur_data_block);
+		}
+
+		/*
+		 * If necessary grow cache slot by reading block list.  Cache
+		 * slot is extended up to index or to the end of the slot, in
+		 * which case further slots will be used.
+		 */
+		for (i = meta->offset + meta->entries; i <= index &&
+				i < meta->offset + SQUASHFS_META_ENTRIES; i++) {
+			int blocks = skip * SQUASHFS_META_INDEXES;
+			long long res = read_indexes(inode->i_sb, blocks,
+					&cur_index_block, &cur_offset);
+
+			if (res < 0) {
+				if (meta->entries == 0)
+					/*
+					 * Don't leave an empty slot on read
+					 * error allocated to this inode...
+					 */
+					meta->inode_number = 0;
+				err = res;
+				goto failed;
+			}
+
+			cur_data_block += res;
+			meta_entry = &meta->meta_entry[i - meta->offset];
+			meta_entry->index_block = cur_index_block -
+				msblk->inode_table;
+			meta_entry->offset = cur_offset;
+			meta_entry->data_block = cur_data_block;
+			meta->entries++;
+			offset++;
+		}
+
+		TRACE("get_meta_index: meta->offset %d, meta->entries %d\n",
+				meta->offset, meta->entries);
+
+		release_meta_index(inode, meta);
+	}
+
+all_done:
+	*index_block = cur_index_block;
+	*index_offset = cur_offset;
+	*data_block = cur_data_block;
+
+	/*
+	 * Scale cache index (cache slot entry) to index
+	 */
+	return offset * SQUASHFS_META_INDEXES * skip;
+
+failed:
+	release_meta_index(inode, meta);
+	return err;
+}
+
+
+/*
+ * Get the on-disk location and compressed size of the datablock
+ * specified by index.  Fill_meta_index() does most of the work.
+ */
+static int read_blocklist(struct inode *inode, int index, u64 *block)
+{
+	u64 start;
+	long long blks;
+	int offset;
+	__le32 size;
+	int res = fill_meta_index(inode, index, &start, &offset, block);
+
+	TRACE("read_blocklist: res %d, index %d, start 0x%llx, offset"
+		       " 0x%x, block 0x%llx\n", res, index, start, offset,
+			*block);
+
+	if (res < 0)
+		return res;
+
+	/*
+	 * res contains the index of the mapping returned by fill_meta_index(),
+	 * this will likely be less than the desired index (because the
+	 * meta_index cache works at a higher granularity).  Read any
+	 * extra block indexes needed.
+	 */
+	if (res < index) {
+		blks = read_indexes(inode->i_sb, index - res, &start, &offset);
+		if (blks < 0)
+			return (int) blks;
+		*block += blks;
+	}
+
+	/*
+	 * Read length of block specified by index.
+	 */
+	res = squashfs_read_metadata(inode->i_sb, &size, &start, &offset,
+			sizeof(size));
+	if (res < 0)
+		return res;
+	return le32_to_cpu(size);
+}
+
+
+static int squashfs_readpage(struct file *file, struct page *page)
+{
+	struct inode *inode = page->mapping->host;
+	struct squashfs_sb_info *msblk = inode->i_sb->s_fs_info;
+	int bytes, i, offset = 0, sparse = 0;
+	struct squashfs_cache_entry *buffer = NULL;
+	void *pageaddr;
+
+	int mask = (1 << (msblk->block_log - PAGE_CACHE_SHIFT)) - 1;
+	int index = page->index >> (msblk->block_log - PAGE_CACHE_SHIFT);
+	int start_index = page->index & ~mask;
+	int end_index = start_index | mask;
+	int file_end = i_size_read(inode) >> msblk->block_log;
+
+	TRACE("Entered squashfs_readpage, page index %lx, start block %llx\n",
+				page->index, squashfs_i(inode)->start);
+
+	if (page->index >= ((i_size_read(inode) + PAGE_CACHE_SIZE - 1) >>
+					PAGE_CACHE_SHIFT))
+		goto out;
+
+	if (index < file_end || squashfs_i(inode)->fragment_block ==
+					SQUASHFS_INVALID_BLK) {
+		/*
+		 * Reading a datablock from disk.  Need to read block list
+		 * to get location and block size.
+		 */
+		u64 block = 0;
+		int bsize = read_blocklist(inode, index, &block);
+		if (bsize < 0)
+			goto error_out;
+
+		if (bsize == 0) { /* hole */
+			bytes = index == file_end ?
+				(i_size_read(inode) & (msblk->block_size - 1)) :
+				 msblk->block_size;
+			sparse = 1;
+		} else {
+			/*
+			 * Read and decompress datablock.
+			 */
+			buffer = squashfs_get_datablock(inode->i_sb,
+								block, bsize);
+			if (buffer->error) {
+				ERROR("Unable to read page, block %llx, size %x"
+					"\n", block, bsize);
+				squashfs_cache_put(buffer);
+				goto error_out;
+			}
+			bytes = buffer->length;
+		}
+	} else {
+		/*
+		 * Datablock is stored inside a fragment (tail-end packed
+		 * block).
+		 */
+		buffer = squashfs_get_fragment(inode->i_sb,
+				squashfs_i(inode)->fragment_block,
+				squashfs_i(inode)->fragment_size);
+
+		if (buffer->error) {
+			ERROR("Unable to read page, block %llx, size %x\n",
+				squashfs_i(inode)->fragment_block,
+				squashfs_i(inode)->fragment_size);
+			squashfs_cache_put(buffer);
+			goto error_out;
+		}
+		bytes = i_size_read(inode) & (msblk->block_size - 1);
+		offset = squashfs_i(inode)->fragment_offset;
+	}
+
+	/*
+	 * Loop copying datablock into pages.  As the datablock likely covers
+	 * many PAGE_CACHE_SIZE pages (default block size is 128 KiB) explicitly
+	 * grab the pages from the page cache, except for the page that we've
+	 * been called to fill.
+	 */
+	for (i = start_index; i <= end_index && bytes > 0; i++,
+			bytes -= PAGE_CACHE_SIZE, offset += PAGE_CACHE_SIZE) {
+		struct page *push_page;
+		int avail = sparse ? 0 : min_t(int, bytes, PAGE_CACHE_SIZE);
+
+		TRACE("bytes %d, i %d, available_bytes %d\n", bytes, i, avail);
+
+		push_page = (i == page->index) ? page :
+			grab_cache_page_nowait(page->mapping, i);
+
+		if (!push_page)
+			continue;
+
+		if (PageUptodate(push_page))
+			goto skip_page;
+
+		pageaddr = kmap_atomic(push_page, KM_USER0);
+		squashfs_copy_data(pageaddr, buffer, offset, avail);
+		memset(pageaddr + avail, 0, PAGE_CACHE_SIZE - avail);
+		kunmap_atomic(pageaddr, KM_USER0);
+		flush_dcache_page(push_page);
+		SetPageUptodate(push_page);
+skip_page:
+		unlock_page(push_page);
+		if (i != page->index)
+			page_cache_release(push_page);
+	}
+
+	if (!sparse)
+		squashfs_cache_put(buffer);
+
+	return 0;
+
+error_out:
+	SetPageError(page);
+out:
+	pageaddr = kmap_atomic(page, KM_USER0);
+	memset(pageaddr, 0, PAGE_CACHE_SIZE);
+	kunmap_atomic(pageaddr, KM_USER0);
+	flush_dcache_page(page);
+	if (!PageError(page))
+		SetPageUptodate(page);
+	unlock_page(page);
+
+	return 0;
+}
+
+
+const struct address_space_operations squashfs_aops = {
+	.readpage = squashfs_readpage
+};
diff --git a/fs/squashfs/fragment.c b/fs/squashfs/fragment.c
new file mode 100644
index 00000000000..b5a2c15bbbc
--- /dev/null
+++ b/fs/squashfs/fragment.c
@@ -0,0 +1,98 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * fragment.c
+ */
+
+/*
+ * This file implements code to handle compressed fragments (tail-end packed
+ * datablocks).
+ *
+ * Regular files contain a fragment index which is mapped to a fragment
+ * location on disk and compressed size using a fragment lookup table.
+ * Like everything in Squashfs this fragment lookup table is itself stored
+ * compressed into metadata blocks.  A second index table is used to locate
+ * these.  This second index table for speed of access (and because it
+ * is small) is read at mount time and cached in memory.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Look-up fragment using the fragment index table.  Return the on disk
+ * location of the fragment and its compressed size
+ */
+int squashfs_frag_lookup(struct super_block *sb, unsigned int fragment,
+				u64 *fragment_block)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int block = SQUASHFS_FRAGMENT_INDEX(fragment);
+	int offset = SQUASHFS_FRAGMENT_INDEX_OFFSET(fragment);
+	u64 start_block = le64_to_cpu(msblk->fragment_index[block]);
+	struct squashfs_fragment_entry fragment_entry;
+	int size;
+
+	size = squashfs_read_metadata(sb, &fragment_entry, &start_block,
+					&offset, sizeof(fragment_entry));
+	if (size < 0)
+		return size;
+
+	*fragment_block = le64_to_cpu(fragment_entry.start_block);
+	size = le32_to_cpu(fragment_entry.size);
+
+	return size;
+}
+
+
+/*
+ * Read the uncompressed fragment lookup table indexes off disk into memory
+ */
+__le64 *squashfs_read_fragment_index_table(struct super_block *sb,
+	u64 fragment_table_start, unsigned int fragments)
+{
+	unsigned int length = SQUASHFS_FRAGMENT_INDEX_BYTES(fragments);
+	__le64 *fragment_index;
+	int err;
+
+	/* Allocate fragment lookup table indexes */
+	fragment_index = kmalloc(length, GFP_KERNEL);
+	if (fragment_index == NULL) {
+		ERROR("Failed to allocate fragment index table\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	err = squashfs_read_table(sb, fragment_index, fragment_table_start,
+			length);
+	if (err < 0) {
+		ERROR("unable to read fragment index table\n");
+		kfree(fragment_index);
+		return ERR_PTR(err);
+	}
+
+	return fragment_index;
+}
diff --git a/fs/squashfs/id.c b/fs/squashfs/id.c
new file mode 100644
index 00000000000..3795b837ba2
--- /dev/null
+++ b/fs/squashfs/id.c
@@ -0,0 +1,94 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * id.c
+ */
+
+/*
+ * This file implements code to handle uids and gids.
+ *
+ * For space efficiency regular files store uid and gid indexes, which are
+ * converted to 32-bit uids/gids using an id look up table.  This table is
+ * stored compressed into metadata blocks.  A second index table is used to
+ * locate these.  This second index table for speed of access (and because it
+ * is small) is read at mount time and cached in memory.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Map uid/gid index into real 32-bit uid/gid using the id look up table
+ */
+int squashfs_get_id(struct super_block *sb, unsigned int index,
+					unsigned int *id)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int block = SQUASHFS_ID_BLOCK(index);
+	int offset = SQUASHFS_ID_BLOCK_OFFSET(index);
+	u64 start_block = le64_to_cpu(msblk->id_table[block]);
+	__le32 disk_id;
+	int err;
+
+	err = squashfs_read_metadata(sb, &disk_id, &start_block, &offset,
+							sizeof(disk_id));
+	if (err < 0)
+		return err;
+
+	*id = le32_to_cpu(disk_id);
+	return 0;
+}
+
+
+/*
+ * Read uncompressed id lookup table indexes from disk into memory
+ */
+__le64 *squashfs_read_id_index_table(struct super_block *sb,
+			u64 id_table_start, unsigned short no_ids)
+{
+	unsigned int length = SQUASHFS_ID_BLOCK_BYTES(no_ids);
+	__le64 *id_table;
+	int err;
+
+	TRACE("In read_id_index_table, length %d\n", length);
+
+	/* Allocate id lookup table indexes */
+	id_table = kmalloc(length, GFP_KERNEL);
+	if (id_table == NULL) {
+		ERROR("Failed to allocate id index table\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	err = squashfs_read_table(sb, id_table, id_table_start, length);
+	if (err < 0) {
+		ERROR("unable to read id index table\n");
+		kfree(id_table);
+		return ERR_PTR(err);
+	}
+
+	return id_table;
+}
diff --git a/fs/squashfs/inode.c b/fs/squashfs/inode.c
new file mode 100644
index 00000000000..7a63398bb85
--- /dev/null
+++ b/fs/squashfs/inode.c
@@ -0,0 +1,346 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * inode.c
+ */
+
+/*
+ * This file implements code to create and read inodes from disk.
+ *
+ * Inodes in Squashfs are identified by a 48-bit inode which encodes the
+ * location of the compressed metadata block containing the inode, and the byte
+ * offset into that block where the inode is placed (<block, offset>).
+ *
+ * To maximise compression there are different inodes for each file type
+ * (regular file, directory, device, etc.), the inode contents and length
+ * varying with the type.
+ *
+ * To further maximise compression, two types of regular file inode and
+ * directory inode are defined: inodes optimised for frequently occurring
+ * regular files and directories, and extended types where extra
+ * information has to be stored.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Initialise VFS inode with the base inode information common to all
+ * Squashfs inode types.  Sqsh_ino contains the unswapped base inode
+ * off disk.
+ */
+static int squashfs_new_inode(struct super_block *sb, struct inode *inode,
+				struct squashfs_base_inode *sqsh_ino)
+{
+	int err;
+
+	err = squashfs_get_id(sb, le16_to_cpu(sqsh_ino->uid), &inode->i_uid);
+	if (err)
+		return err;
+
+	err = squashfs_get_id(sb, le16_to_cpu(sqsh_ino->guid), &inode->i_gid);
+	if (err)
+		return err;
+
+	inode->i_ino = le32_to_cpu(sqsh_ino->inode_number);
+	inode->i_mtime.tv_sec = le32_to_cpu(sqsh_ino->mtime);
+	inode->i_atime.tv_sec = inode->i_mtime.tv_sec;
+	inode->i_ctime.tv_sec = inode->i_mtime.tv_sec;
+	inode->i_mode = le16_to_cpu(sqsh_ino->mode);
+	inode->i_size = 0;
+
+	return err;
+}
+
+
+struct inode *squashfs_iget(struct super_block *sb, long long ino,
+				unsigned int ino_number)
+{
+	struct inode *inode = iget_locked(sb, ino_number);
+	int err;
+
+	TRACE("Entered squashfs_iget\n");
+
+	if (!inode)
+		return ERR_PTR(-ENOMEM);
+	if (!(inode->i_state & I_NEW))
+		return inode;
+
+	err = squashfs_read_inode(inode, ino);
+	if (err) {
+		iget_failed(inode);
+		return ERR_PTR(err);
+	}
+
+	unlock_new_inode(inode);
+	return inode;
+}
+
+
+/*
+ * Initialise VFS inode by reading inode from inode table (compressed
+ * metadata).  The format and amount of data read depends on type.
+ */
+int squashfs_read_inode(struct inode *inode, long long ino)
+{
+	struct super_block *sb = inode->i_sb;
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	u64 block = SQUASHFS_INODE_BLK(ino) + msblk->inode_table;
+	int err, type, offset = SQUASHFS_INODE_OFFSET(ino);
+	union squashfs_inode squashfs_ino;
+	struct squashfs_base_inode *sqshb_ino = &squashfs_ino.base;
+
+	TRACE("Entered squashfs_read_inode\n");
+
+	/*
+	 * Read inode base common to all inode types.
+	 */
+	err = squashfs_read_metadata(sb, sqshb_ino, &block,
+				&offset, sizeof(*sqshb_ino));
+	if (err < 0)
+		goto failed_read;
+
+	err = squashfs_new_inode(sb, inode, sqshb_ino);
+	if (err)
+		goto failed_read;
+
+	block = SQUASHFS_INODE_BLK(ino) + msblk->inode_table;
+	offset = SQUASHFS_INODE_OFFSET(ino);
+
+	type = le16_to_cpu(sqshb_ino->inode_type);
+	switch (type) {
+	case SQUASHFS_REG_TYPE: {
+		unsigned int frag_offset, frag_size, frag;
+		u64 frag_blk;
+		struct squashfs_reg_inode *sqsh_ino = &squashfs_ino.reg;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+							sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		frag = le32_to_cpu(sqsh_ino->fragment);
+		if (frag != SQUASHFS_INVALID_FRAG) {
+			frag_offset = le32_to_cpu(sqsh_ino->offset);
+			frag_size = squashfs_frag_lookup(sb, frag, &frag_blk);
+			if (frag_size < 0) {
+				err = frag_size;
+				goto failed_read;
+			}
+		} else {
+			frag_blk = SQUASHFS_INVALID_BLK;
+			frag_size = 0;
+			frag_offset = 0;
+		}
+
+		inode->i_nlink = 1;
+		inode->i_size = le32_to_cpu(sqsh_ino->file_size);
+		inode->i_fop = &generic_ro_fops;
+		inode->i_mode |= S_IFREG;
+		inode->i_blocks = ((inode->i_size - 1) >> 9) + 1;
+		squashfs_i(inode)->fragment_block = frag_blk;
+		squashfs_i(inode)->fragment_size = frag_size;
+		squashfs_i(inode)->fragment_offset = frag_offset;
+		squashfs_i(inode)->start = le32_to_cpu(sqsh_ino->start_block);
+		squashfs_i(inode)->block_list_start = block;
+		squashfs_i(inode)->offset = offset;
+		inode->i_data.a_ops = &squashfs_aops;
+
+		TRACE("File inode %x:%x, start_block %llx, block_list_start "
+			"%llx, offset %x\n", SQUASHFS_INODE_BLK(ino),
+			offset, squashfs_i(inode)->start, block, offset);
+		break;
+	}
+	case SQUASHFS_LREG_TYPE: {
+		unsigned int frag_offset, frag_size, frag;
+		u64 frag_blk;
+		struct squashfs_lreg_inode *sqsh_ino = &squashfs_ino.lreg;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+							sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		frag = le32_to_cpu(sqsh_ino->fragment);
+		if (frag != SQUASHFS_INVALID_FRAG) {
+			frag_offset = le32_to_cpu(sqsh_ino->offset);
+			frag_size = squashfs_frag_lookup(sb, frag, &frag_blk);
+			if (frag_size < 0) {
+				err = frag_size;
+				goto failed_read;
+			}
+		} else {
+			frag_blk = SQUASHFS_INVALID_BLK;
+			frag_size = 0;
+			frag_offset = 0;
+		}
+
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		inode->i_size = le64_to_cpu(sqsh_ino->file_size);
+		inode->i_fop = &generic_ro_fops;
+		inode->i_mode |= S_IFREG;
+		inode->i_blocks = ((inode->i_size -
+				le64_to_cpu(sqsh_ino->sparse) - 1) >> 9) + 1;
+
+		squashfs_i(inode)->fragment_block = frag_blk;
+		squashfs_i(inode)->fragment_size = frag_size;
+		squashfs_i(inode)->fragment_offset = frag_offset;
+		squashfs_i(inode)->start = le64_to_cpu(sqsh_ino->start_block);
+		squashfs_i(inode)->block_list_start = block;
+		squashfs_i(inode)->offset = offset;
+		inode->i_data.a_ops = &squashfs_aops;
+
+		TRACE("File inode %x:%x, start_block %llx, block_list_start "
+			"%llx, offset %x\n", SQUASHFS_INODE_BLK(ino),
+			offset, squashfs_i(inode)->start, block, offset);
+		break;
+	}
+	case SQUASHFS_DIR_TYPE: {
+		struct squashfs_dir_inode *sqsh_ino = &squashfs_ino.dir;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		inode->i_size = le16_to_cpu(sqsh_ino->file_size);
+		inode->i_op = &squashfs_dir_inode_ops;
+		inode->i_fop = &squashfs_dir_ops;
+		inode->i_mode |= S_IFDIR;
+		squashfs_i(inode)->start = le32_to_cpu(sqsh_ino->start_block);
+		squashfs_i(inode)->offset = le16_to_cpu(sqsh_ino->offset);
+		squashfs_i(inode)->dir_idx_cnt = 0;
+		squashfs_i(inode)->parent = le32_to_cpu(sqsh_ino->parent_inode);
+
+		TRACE("Directory inode %x:%x, start_block %llx, offset %x\n",
+				SQUASHFS_INODE_BLK(ino), offset,
+				squashfs_i(inode)->start,
+				le16_to_cpu(sqsh_ino->offset));
+		break;
+	}
+	case SQUASHFS_LDIR_TYPE: {
+		struct squashfs_ldir_inode *sqsh_ino = &squashfs_ino.ldir;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		inode->i_size = le32_to_cpu(sqsh_ino->file_size);
+		inode->i_op = &squashfs_dir_inode_ops;
+		inode->i_fop = &squashfs_dir_ops;
+		inode->i_mode |= S_IFDIR;
+		squashfs_i(inode)->start = le32_to_cpu(sqsh_ino->start_block);
+		squashfs_i(inode)->offset = le16_to_cpu(sqsh_ino->offset);
+		squashfs_i(inode)->dir_idx_start = block;
+		squashfs_i(inode)->dir_idx_offset = offset;
+		squashfs_i(inode)->dir_idx_cnt = le16_to_cpu(sqsh_ino->i_count);
+		squashfs_i(inode)->parent = le32_to_cpu(sqsh_ino->parent_inode);
+
+		TRACE("Long directory inode %x:%x, start_block %llx, offset "
+				"%x\n", SQUASHFS_INODE_BLK(ino), offset,
+				squashfs_i(inode)->start,
+				le16_to_cpu(sqsh_ino->offset));
+		break;
+	}
+	case SQUASHFS_SYMLINK_TYPE:
+	case SQUASHFS_LSYMLINK_TYPE: {
+		struct squashfs_symlink_inode *sqsh_ino = &squashfs_ino.symlink;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		inode->i_size = le32_to_cpu(sqsh_ino->symlink_size);
+		inode->i_op = &page_symlink_inode_operations;
+		inode->i_data.a_ops = &squashfs_symlink_aops;
+		inode->i_mode |= S_IFLNK;
+		squashfs_i(inode)->start = block;
+		squashfs_i(inode)->offset = offset;
+
+		TRACE("Symbolic link inode %x:%x, start_block %llx, offset "
+				"%x\n", SQUASHFS_INODE_BLK(ino), offset,
+				block, offset);
+		break;
+	}
+	case SQUASHFS_BLKDEV_TYPE:
+	case SQUASHFS_CHRDEV_TYPE:
+	case SQUASHFS_LBLKDEV_TYPE:
+	case SQUASHFS_LCHRDEV_TYPE: {
+		struct squashfs_dev_inode *sqsh_ino = &squashfs_ino.dev;
+		unsigned int rdev;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		if (type == SQUASHFS_CHRDEV_TYPE)
+			inode->i_mode |= S_IFCHR;
+		else
+			inode->i_mode |= S_IFBLK;
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		rdev = le32_to_cpu(sqsh_ino->rdev);
+		init_special_inode(inode, inode->i_mode, new_decode_dev(rdev));
+
+		TRACE("Device inode %x:%x, rdev %x\n",
+				SQUASHFS_INODE_BLK(ino), offset, rdev);
+		break;
+	}
+	case SQUASHFS_FIFO_TYPE:
+	case SQUASHFS_SOCKET_TYPE:
+	case SQUASHFS_LFIFO_TYPE:
+	case SQUASHFS_LSOCKET_TYPE: {
+		struct squashfs_ipc_inode *sqsh_ino = &squashfs_ino.ipc;
+
+		err = squashfs_read_metadata(sb, sqsh_ino, &block, &offset,
+				sizeof(*sqsh_ino));
+		if (err < 0)
+			goto failed_read;
+
+		if (type == SQUASHFS_FIFO_TYPE)
+			inode->i_mode |= S_IFIFO;
+		else
+			inode->i_mode |= S_IFSOCK;
+		inode->i_nlink = le32_to_cpu(sqsh_ino->nlink);
+		init_special_inode(inode, inode->i_mode, 0);
+		break;
+	}
+	default:
+		ERROR("Unknown inode type %d in squashfs_iget!\n", type);
+		return -EINVAL;
+	}
+
+	return 0;
+
+failed_read:
+	ERROR("Unable to read inode 0x%llx\n", ino);
+	return err;
+}
diff --git a/fs/squashfs/namei.c b/fs/squashfs/namei.c
new file mode 100644
index 00000000000..9e398653b22
--- /dev/null
+++ b/fs/squashfs/namei.c
@@ -0,0 +1,242 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * namei.c
+ */
+
+/*
+ * This file implements code to do filename lookup in directories.
+ *
+ * Like inodes, directories are packed into compressed metadata blocks, stored
+ * in a directory table.  Directories are accessed using the start address of
+ * the metablock containing the directory and the offset into the
+ * decompressed block (<block, offset>).
+ *
+ * Directories are organised in a slightly complex way, and are not simply
+ * a list of file names.  The organisation takes advantage of the
+ * fact that (in most cases) the inodes of the files will be in the same
+ * compressed metadata block, and therefore, can share the start block.
+ * Directories are therefore organised in a two level list, a directory
+ * header containing the shared start block value, and a sequence of directory
+ * entries, each of which share the shared start block.  A new directory header
+ * is written once/if the inode start block changes.  The directory
+ * header/directory entry list is repeated as many times as necessary.
+ *
+ * Directories are sorted, and can contain a directory index to speed up
+ * file lookup.  Directory indexes store one entry per metablock, each entry
+ * storing the index/filename mapping to the first directory header
+ * in each metadata block.  Directories are sorted in alphabetical order,
+ * and at lookup the index is scanned linearly looking for the first filename
+ * alphabetically larger than the filename being looked up.  At this point the
+ * location of the metadata block the filename is in has been found.
+ * The general idea of the index is ensure only one metadata block needs to be
+ * decompressed to do a lookup irrespective of the length of the directory.
+ * This scheme has the advantage that it doesn't require extra memory overhead
+ * and doesn't require much extra storage on disk.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/dcache.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+/*
+ * Lookup name in the directory index, returning the location of the metadata
+ * block containing it, and the directory index this represents.
+ *
+ * If we get an error reading the index then return the part of the index
+ * (if any) we have managed to read - the index isn't essential, just
+ * quicker.
+ */
+static int get_dir_index_using_name(struct super_block *sb,
+			u64 *next_block, int *next_offset, u64 index_start,
+			int index_offset, int i_count, const char *name,
+			int len)
+{
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int i, size, length = 0, err;
+	struct squashfs_dir_index *index;
+	char *str;
+
+	TRACE("Entered get_dir_index_using_name, i_count %d\n", i_count);
+
+	index = kmalloc(sizeof(*index) + SQUASHFS_NAME_LEN * 2 + 2, GFP_KERNEL);
+	if (index == NULL) {
+		ERROR("Failed to allocate squashfs_dir_index\n");
+		goto out;
+	}
+
+	str = &index->name[SQUASHFS_NAME_LEN + 1];
+	strncpy(str, name, len);
+	str[len] = '\0';
+
+	for (i = 0; i < i_count; i++) {
+		err = squashfs_read_metadata(sb, index, &index_start,
+					&index_offset, sizeof(*index));
+		if (err < 0)
+			break;
+
+
+		size = le32_to_cpu(index->size) + 1;
+
+		err = squashfs_read_metadata(sb, index->name, &index_start,
+					&index_offset, size);
+		if (err < 0)
+			break;
+
+		index->name[size] = '\0';
+
+		if (strcmp(index->name, str) > 0)
+			break;
+
+		length = le32_to_cpu(index->index);
+		*next_block = le32_to_cpu(index->start_block) +
+					msblk->directory_table;
+	}
+
+	*next_offset = (length + *next_offset) % SQUASHFS_METADATA_SIZE;
+	kfree(index);
+
+out:
+	/*
+	 * Return index (f_pos) of the looked up metadata block.  Translate
+	 * from internal f_pos to external f_pos which is offset by 3 because
+	 * we invent "." and ".." entries which are not actually stored in the
+	 * directory.
+	 */
+	return length + 3;
+}
+
+
+static struct dentry *squashfs_lookup(struct inode *dir, struct dentry *dentry,
+				 struct nameidata *nd)
+{
+	const unsigned char *name = dentry->d_name.name;
+	int len = dentry->d_name.len;
+	struct inode *inode = NULL;
+	struct squashfs_sb_info *msblk = dir->i_sb->s_fs_info;
+	struct squashfs_dir_header dirh;
+	struct squashfs_dir_entry *dire;
+	u64 block = squashfs_i(dir)->start + msblk->directory_table;
+	int offset = squashfs_i(dir)->offset;
+	int err, length = 0, dir_count, size;
+
+	TRACE("Entered squashfs_lookup [%llx:%x]\n", block, offset);
+
+	dire = kmalloc(sizeof(*dire) + SQUASHFS_NAME_LEN + 1, GFP_KERNEL);
+	if (dire == NULL) {
+		ERROR("Failed to allocate squashfs_dir_entry\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	if (len > SQUASHFS_NAME_LEN) {
+		err = -ENAMETOOLONG;
+		goto failed;
+	}
+
+	length = get_dir_index_using_name(dir->i_sb, &block, &offset,
+				squashfs_i(dir)->dir_idx_start,
+				squashfs_i(dir)->dir_idx_offset,
+				squashfs_i(dir)->dir_idx_cnt, name, len);
+
+	while (length < i_size_read(dir)) {
+		/*
+		 * Read directory header.
+		 */
+		err = squashfs_read_metadata(dir->i_sb, &dirh, &block,
+				&offset, sizeof(dirh));
+		if (err < 0)
+			goto read_failure;
+
+		length += sizeof(dirh);
+
+		dir_count = le32_to_cpu(dirh.count) + 1;
+		while (dir_count--) {
+			/*
+			 * Read directory entry.
+			 */
+			err = squashfs_read_metadata(dir->i_sb, dire, &block,
+					&offset, sizeof(*dire));
+			if (err < 0)
+				goto read_failure;
+
+			size = le16_to_cpu(dire->size) + 1;
+
+			err = squashfs_read_metadata(dir->i_sb, dire->name,
+					&block, &offset, size);
+			if (err < 0)
+				goto read_failure;
+
+			length += sizeof(*dire) + size;
+
+			if (name[0] < dire->name[0])
+				goto exit_lookup;
+
+			if (len == size && !strncmp(name, dire->name, len)) {
+				unsigned int blk, off, ino_num;
+				long long ino;
+				blk = le32_to_cpu(dirh.start_block);
+				off = le16_to_cpu(dire->offset);
+				ino_num = le32_to_cpu(dirh.inode_number) +
+					(short) le16_to_cpu(dire->inode_number);
+				ino = SQUASHFS_MKINODE(blk, off);
+
+				TRACE("calling squashfs_iget for directory "
+					"entry %s, inode  %x:%x, %d\n", name,
+					blk, off, ino_num);
+
+				inode = squashfs_iget(dir->i_sb, ino, ino_num);
+				if (IS_ERR(inode)) {
+					err = PTR_ERR(inode);
+					goto failed;
+				}
+
+				goto exit_lookup;
+			}
+		}
+	}
+
+exit_lookup:
+	kfree(dire);
+	if (inode)
+		return d_splice_alias(inode, dentry);
+	d_add(dentry, inode);
+	return ERR_PTR(0);
+
+read_failure:
+	ERROR("Unable to read directory block [%llx:%x]\n",
+		squashfs_i(dir)->start + msblk->directory_table,
+		squashfs_i(dir)->offset);
+failed:
+	kfree(dire);
+	return ERR_PTR(err);
+}
+
+
+const struct inode_operations squashfs_dir_inode_ops = {
+	.lookup = squashfs_lookup
+};
diff --git a/fs/squashfs/squashfs.h b/fs/squashfs/squashfs.h
new file mode 100644
index 00000000000..6b2515d027d
--- /dev/null
+++ b/fs/squashfs/squashfs.h
@@ -0,0 +1,90 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * squashfs.h
+ */
+
+#define TRACE(s, args...)	pr_debug("SQUASHFS: "s, ## args)
+
+#define ERROR(s, args...)	pr_err("SQUASHFS error: "s, ## args)
+
+#define WARNING(s, args...)	pr_warning("SQUASHFS: "s, ## args)
+
+static inline struct squashfs_inode_info *squashfs_i(struct inode *inode)
+{
+	return list_entry(inode, struct squashfs_inode_info, vfs_inode);
+}
+
+/* block.c */
+extern int squashfs_read_data(struct super_block *, void **, u64, int, u64 *,
+				int);
+
+/* cache.c */
+extern struct squashfs_cache *squashfs_cache_init(char *, int, int);
+extern void squashfs_cache_delete(struct squashfs_cache *);
+extern struct squashfs_cache_entry *squashfs_cache_get(struct super_block *,
+				struct squashfs_cache *, u64, int);
+extern void squashfs_cache_put(struct squashfs_cache_entry *);
+extern int squashfs_copy_data(void *, struct squashfs_cache_entry *, int, int);
+extern int squashfs_read_metadata(struct super_block *, void *, u64 *,
+				int *, int);
+extern struct squashfs_cache_entry *squashfs_get_fragment(struct super_block *,
+				u64, int);
+extern struct squashfs_cache_entry *squashfs_get_datablock(struct super_block *,
+				u64, int);
+extern int squashfs_read_table(struct super_block *, void *, u64, int);
+
+/* export.c */
+extern __le64 *squashfs_read_inode_lookup_table(struct super_block *, u64,
+				unsigned int);
+
+/* fragment.c */
+extern int squashfs_frag_lookup(struct super_block *, unsigned int, u64 *);
+extern __le64 *squashfs_read_fragment_index_table(struct super_block *,
+				u64, unsigned int);
+
+/* id.c */
+extern int squashfs_get_id(struct super_block *, unsigned int, unsigned int *);
+extern __le64 *squashfs_read_id_index_table(struct super_block *, u64,
+				unsigned short);
+
+/* inode.c */
+extern struct inode *squashfs_iget(struct super_block *, long long,
+				unsigned int);
+extern int squashfs_read_inode(struct inode *, long long);
+
+/*
+ * Inodes and files operations
+ */
+
+/* dir.c */
+extern const struct file_operations squashfs_dir_ops;
+
+/* export.c */
+extern const struct export_operations squashfs_export_ops;
+
+/* file.c */
+extern const struct address_space_operations squashfs_aops;
+
+/* namei.c */
+extern const struct inode_operations squashfs_dir_inode_ops;
+
+/* symlink.c */
+extern const struct address_space_operations squashfs_symlink_aops;
diff --git a/fs/squashfs/squashfs_fs.h b/fs/squashfs/squashfs_fs.h
new file mode 100644
index 00000000000..6840da1bf21
--- /dev/null
+++ b/fs/squashfs/squashfs_fs.h
@@ -0,0 +1,381 @@
+#ifndef SQUASHFS_FS
+#define SQUASHFS_FS
+/*
+ * Squashfs
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * squashfs_fs.h
+ */
+
+#define SQUASHFS_CACHED_FRAGMENTS	CONFIG_SQUASHFS_FRAGMENT_CACHE_SIZE
+#define SQUASHFS_MAJOR			4
+#define SQUASHFS_MINOR			0
+#define SQUASHFS_MAGIC			0x73717368
+#define SQUASHFS_START			0
+
+/* size of metadata (inode and directory) blocks */
+#define SQUASHFS_METADATA_SIZE		8192
+#define SQUASHFS_METADATA_LOG		13
+
+/* default size of data blocks */
+#define SQUASHFS_FILE_SIZE		131072
+#define SQUASHFS_FILE_LOG		17
+
+#define SQUASHFS_FILE_MAX_SIZE		1048576
+#define SQUASHFS_FILE_MAX_LOG		20
+
+/* Max number of uids and gids */
+#define SQUASHFS_IDS			65536
+
+/* Max length of filename (not 255) */
+#define SQUASHFS_NAME_LEN		256
+
+#define SQUASHFS_INVALID_FRAG		(0xffffffffU)
+#define SQUASHFS_INVALID_BLK		(-1LL)
+
+/* Filesystem flags */
+#define SQUASHFS_NOI			0
+#define SQUASHFS_NOD			1
+#define SQUASHFS_NOF			3
+#define SQUASHFS_NO_FRAG		4
+#define SQUASHFS_ALWAYS_FRAG		5
+#define SQUASHFS_DUPLICATE		6
+#define SQUASHFS_EXPORT			7
+
+#define SQUASHFS_BIT(flag, bit)		((flag >> bit) & 1)
+
+#define SQUASHFS_UNCOMPRESSED_INODES(flags)	SQUASHFS_BIT(flags, \
+						SQUASHFS_NOI)
+
+#define SQUASHFS_UNCOMPRESSED_DATA(flags)	SQUASHFS_BIT(flags, \
+						SQUASHFS_NOD)
+
+#define SQUASHFS_UNCOMPRESSED_FRAGMENTS(flags)	SQUASHFS_BIT(flags, \
+						SQUASHFS_NOF)
+
+#define SQUASHFS_NO_FRAGMENTS(flags)		SQUASHFS_BIT(flags, \
+						SQUASHFS_NO_FRAG)
+
+#define SQUASHFS_ALWAYS_FRAGMENTS(flags)	SQUASHFS_BIT(flags, \
+						SQUASHFS_ALWAYS_FRAG)
+
+#define SQUASHFS_DUPLICATES(flags)		SQUASHFS_BIT(flags, \
+						SQUASHFS_DUPLICATE)
+
+#define SQUASHFS_EXPORTABLE(flags)		SQUASHFS_BIT(flags, \
+						SQUASHFS_EXPORT)
+
+/* Max number of types and file types */
+#define SQUASHFS_DIR_TYPE		1
+#define SQUASHFS_REG_TYPE		2
+#define SQUASHFS_SYMLINK_TYPE		3
+#define SQUASHFS_BLKDEV_TYPE		4
+#define SQUASHFS_CHRDEV_TYPE		5
+#define SQUASHFS_FIFO_TYPE		6
+#define SQUASHFS_SOCKET_TYPE		7
+#define SQUASHFS_LDIR_TYPE		8
+#define SQUASHFS_LREG_TYPE		9
+#define SQUASHFS_LSYMLINK_TYPE		10
+#define SQUASHFS_LBLKDEV_TYPE		11
+#define SQUASHFS_LCHRDEV_TYPE		12
+#define SQUASHFS_LFIFO_TYPE		13
+#define SQUASHFS_LSOCKET_TYPE		14
+
+/* Flag whether block is compressed or uncompressed, bit is set if block is
+ * uncompressed */
+#define SQUASHFS_COMPRESSED_BIT		(1 << 15)
+
+#define SQUASHFS_COMPRESSED_SIZE(B)	(((B) & ~SQUASHFS_COMPRESSED_BIT) ? \
+		(B) & ~SQUASHFS_COMPRESSED_BIT :  SQUASHFS_COMPRESSED_BIT)
+
+#define SQUASHFS_COMPRESSED(B)		(!((B) & SQUASHFS_COMPRESSED_BIT))
+
+#define SQUASHFS_COMPRESSED_BIT_BLOCK	(1 << 24)
+
+#define SQUASHFS_COMPRESSED_SIZE_BLOCK(B)	((B) & \
+						~SQUASHFS_COMPRESSED_BIT_BLOCK)
+
+#define SQUASHFS_COMPRESSED_BLOCK(B)	(!((B) & SQUASHFS_COMPRESSED_BIT_BLOCK))
+
+/*
+ * Inode number ops.  Inodes consist of a compressed block number, and an
+ * uncompressed offset within that block
+ */
+#define SQUASHFS_INODE_BLK(A)		((unsigned int) ((A) >> 16))
+
+#define SQUASHFS_INODE_OFFSET(A)	((unsigned int) ((A) & 0xffff))
+
+#define SQUASHFS_MKINODE(A, B)		((long long)(((long long) (A)\
+					<< 16) + (B)))
+
+/* Translate between VFS mode and squashfs mode */
+#define SQUASHFS_MODE(A)		((A) & 0xfff)
+
+/* fragment and fragment table defines */
+#define SQUASHFS_FRAGMENT_BYTES(A)	\
+				((A) * sizeof(struct squashfs_fragment_entry))
+
+#define SQUASHFS_FRAGMENT_INDEX(A)	(SQUASHFS_FRAGMENT_BYTES(A) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_FRAGMENT_INDEX_OFFSET(A)	(SQUASHFS_FRAGMENT_BYTES(A) % \
+						SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_FRAGMENT_INDEXES(A)	((SQUASHFS_FRAGMENT_BYTES(A) + \
+					SQUASHFS_METADATA_SIZE - 1) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_FRAGMENT_INDEX_BYTES(A)	(SQUASHFS_FRAGMENT_INDEXES(A) *\
+						sizeof(u64))
+
+/* inode lookup table defines */
+#define SQUASHFS_LOOKUP_BYTES(A)	((A) * sizeof(u64))
+
+#define SQUASHFS_LOOKUP_BLOCK(A)	(SQUASHFS_LOOKUP_BYTES(A) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_LOOKUP_BLOCK_OFFSET(A)	(SQUASHFS_LOOKUP_BYTES(A) % \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_LOOKUP_BLOCKS(A)	((SQUASHFS_LOOKUP_BYTES(A) + \
+					SQUASHFS_METADATA_SIZE - 1) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_LOOKUP_BLOCK_BYTES(A)	(SQUASHFS_LOOKUP_BLOCKS(A) *\
+					sizeof(u64))
+
+/* uid/gid lookup table defines */
+#define SQUASHFS_ID_BYTES(A)		((A) * sizeof(unsigned int))
+
+#define SQUASHFS_ID_BLOCK(A)		(SQUASHFS_ID_BYTES(A) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_ID_BLOCK_OFFSET(A)	(SQUASHFS_ID_BYTES(A) % \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_ID_BLOCKS(A)		((SQUASHFS_ID_BYTES(A) + \
+					SQUASHFS_METADATA_SIZE - 1) / \
+					SQUASHFS_METADATA_SIZE)
+
+#define SQUASHFS_ID_BLOCK_BYTES(A)	(SQUASHFS_ID_BLOCKS(A) *\
+					sizeof(u64))
+
+/* cached data constants for filesystem */
+#define SQUASHFS_CACHED_BLKS		8
+
+#define SQUASHFS_MAX_FILE_SIZE_LOG	64
+
+#define SQUASHFS_MAX_FILE_SIZE		(1LL << \
+					(SQUASHFS_MAX_FILE_SIZE_LOG - 2))
+
+#define SQUASHFS_MARKER_BYTE		0xff
+
+/* meta index cache */
+#define SQUASHFS_META_INDEXES	(SQUASHFS_METADATA_SIZE / sizeof(unsigned int))
+#define SQUASHFS_META_ENTRIES	127
+#define SQUASHFS_META_SLOTS	8
+
+struct meta_entry {
+	u64			data_block;
+	unsigned int		index_block;
+	unsigned short		offset;
+	unsigned short		pad;
+};
+
+struct meta_index {
+	unsigned int		inode_number;
+	unsigned int		offset;
+	unsigned short		entries;
+	unsigned short		skip;
+	unsigned short		locked;
+	unsigned short		pad;
+	struct meta_entry	meta_entry[SQUASHFS_META_ENTRIES];
+};
+
+
+/*
+ * definitions for structures on disk
+ */
+#define ZLIB_COMPRESSION	 1
+
+struct squashfs_super_block {
+	__le32			s_magic;
+	__le32			inodes;
+	__le32			mkfs_time;
+	__le32			block_size;
+	__le32			fragments;
+	__le16			compression;
+	__le16			block_log;
+	__le16			flags;
+	__le16			no_ids;
+	__le16			s_major;
+	__le16			s_minor;
+	__le64			root_inode;
+	__le64			bytes_used;
+	__le64			id_table_start;
+	__le64			xattr_table_start;
+	__le64			inode_table_start;
+	__le64			directory_table_start;
+	__le64			fragment_table_start;
+	__le64			lookup_table_start;
+};
+
+struct squashfs_dir_index {
+	__le32			index;
+	__le32			start_block;
+	__le32			size;
+	unsigned char		name[0];
+};
+
+struct squashfs_base_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+};
+
+struct squashfs_ipc_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			nlink;
+};
+
+struct squashfs_dev_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			nlink;
+	__le32			rdev;
+};
+
+struct squashfs_symlink_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			nlink;
+	__le32			symlink_size;
+	char			symlink[0];
+};
+
+struct squashfs_reg_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			start_block;
+	__le32			fragment;
+	__le32			offset;
+	__le32			file_size;
+	__le16			block_list[0];
+};
+
+struct squashfs_lreg_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le64			start_block;
+	__le64			file_size;
+	__le64			sparse;
+	__le32			nlink;
+	__le32			fragment;
+	__le32			offset;
+	__le32			xattr;
+	__le16			block_list[0];
+};
+
+struct squashfs_dir_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			start_block;
+	__le32			nlink;
+	__le16			file_size;
+	__le16			offset;
+	__le32			parent_inode;
+};
+
+struct squashfs_ldir_inode {
+	__le16			inode_type;
+	__le16			mode;
+	__le16			uid;
+	__le16			guid;
+	__le32			mtime;
+	__le32	 		inode_number;
+	__le32			nlink;
+	__le32			file_size;
+	__le32			start_block;
+	__le32			parent_inode;
+	__le16			i_count;
+	__le16			offset;
+	__le32			xattr;
+	struct squashfs_dir_index	index[0];
+};
+
+union squashfs_inode {
+	struct squashfs_base_inode		base;
+	struct squashfs_dev_inode		dev;
+	struct squashfs_symlink_inode		symlink;
+	struct squashfs_reg_inode		reg;
+	struct squashfs_lreg_inode		lreg;
+	struct squashfs_dir_inode		dir;
+	struct squashfs_ldir_inode		ldir;
+	struct squashfs_ipc_inode		ipc;
+};
+
+struct squashfs_dir_entry {
+	__le16			offset;
+	__le16			inode_number;
+	__le16			type;
+	__le16			size;
+	char			name[0];
+};
+
+struct squashfs_dir_header {
+	__le32			count;
+	__le32			start_block;
+	__le32			inode_number;
+};
+
+struct squashfs_fragment_entry {
+	__le64			start_block;
+	__le32			size;
+	unsigned int		unused;
+};
+
+#endif
diff --git a/fs/squashfs/squashfs_fs_i.h b/fs/squashfs/squashfs_fs_i.h
new file mode 100644
index 00000000000..fbfca30c0c6
--- /dev/null
+++ b/fs/squashfs/squashfs_fs_i.h
@@ -0,0 +1,45 @@
+#ifndef SQUASHFS_FS_I
+#define SQUASHFS_FS_I
+/*
+ * Squashfs
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * squashfs_fs_i.h
+ */
+
+struct squashfs_inode_info {
+	u64		start;
+	int		offset;
+	union {
+		struct {
+			u64		fragment_block;
+			int		fragment_size;
+			int		fragment_offset;
+			u64		block_list_start;
+		};
+		struct {
+			u64		dir_idx_start;
+			int		dir_idx_offset;
+			int		dir_idx_cnt;
+			int		parent;
+		};
+	};
+	struct inode	vfs_inode;
+};
+#endif
diff --git a/fs/squashfs/squashfs_fs_sb.h b/fs/squashfs/squashfs_fs_sb.h
new file mode 100644
index 00000000000..c8c65614dd1
--- /dev/null
+++ b/fs/squashfs/squashfs_fs_sb.h
@@ -0,0 +1,76 @@
+#ifndef SQUASHFS_FS_SB
+#define SQUASHFS_FS_SB
+/*
+ * Squashfs
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * squashfs_fs_sb.h
+ */
+
+#include "squashfs_fs.h"
+
+struct squashfs_cache {
+	char			*name;
+	int			entries;
+	int			next_blk;
+	int			num_waiters;
+	int			unused;
+	int			block_size;
+	int			pages;
+	spinlock_t		lock;
+	wait_queue_head_t	wait_queue;
+	struct squashfs_cache_entry *entry;
+};
+
+struct squashfs_cache_entry {
+	u64			block;
+	int			length;
+	int			refcount;
+	u64			next_index;
+	int			pending;
+	int			error;
+	int			num_waiters;
+	wait_queue_head_t	wait_queue;
+	struct squashfs_cache	*cache;
+	void			**data;
+};
+
+struct squashfs_sb_info {
+	int			devblksize;
+	int			devblksize_log2;
+	struct squashfs_cache	*block_cache;
+	struct squashfs_cache	*fragment_cache;
+	struct squashfs_cache	*read_page;
+	int			next_meta_index;
+	__le64			*id_table;
+	__le64			*fragment_index;
+	unsigned int		*fragment_index_2;
+	struct mutex		read_data_mutex;
+	struct mutex		meta_index_mutex;
+	struct meta_index	*meta_index;
+	z_stream		stream;
+	__le64			*inode_lookup_table;
+	u64			inode_table;
+	u64			directory_table;
+	unsigned int		block_size;
+	unsigned short		block_log;
+	long long		bytes_used;
+	unsigned int		inodes;
+};
+#endif
diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c
new file mode 100644
index 00000000000..a0466d7467b
--- /dev/null
+++ b/fs/squashfs/super.c
@@ -0,0 +1,440 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * super.c
+ */
+
+/*
+ * This file implements code to read the superblock, read and initialise
+ * in-memory structures at mount time, and all the VFS glue code to register
+ * the filesystem.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/slab.h>
+#include <linux/mutex.h>
+#include <linux/pagemap.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+static struct file_system_type squashfs_fs_type;
+static struct super_operations squashfs_super_ops;
+
+static int supported_squashfs_filesystem(short major, short minor, short comp)
+{
+	if (major < SQUASHFS_MAJOR) {
+		ERROR("Major/Minor mismatch, older Squashfs %d.%d "
+			"filesystems are unsupported\n", major, minor);
+		return -EINVAL;
+	} else if (major > SQUASHFS_MAJOR || minor > SQUASHFS_MINOR) {
+		ERROR("Major/Minor mismatch, trying to mount newer "
+			"%d.%d filesystem\n", major, minor);
+		ERROR("Please update your kernel\n");
+		return -EINVAL;
+	}
+
+	if (comp != ZLIB_COMPRESSION)
+		return -EINVAL;
+
+	return 0;
+}
+
+
+static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
+{
+	struct squashfs_sb_info *msblk;
+	struct squashfs_super_block *sblk = NULL;
+	char b[BDEVNAME_SIZE];
+	struct inode *root;
+	long long root_inode;
+	unsigned short flags;
+	unsigned int fragments;
+	u64 lookup_table_start;
+	int err;
+
+	TRACE("Entered squashfs_fill_superblock\n");
+
+	sb->s_fs_info = kzalloc(sizeof(*msblk), GFP_KERNEL);
+	if (sb->s_fs_info == NULL) {
+		ERROR("Failed to allocate squashfs_sb_info\n");
+		return -ENOMEM;
+	}
+	msblk = sb->s_fs_info;
+
+	msblk->stream.workspace = kmalloc(zlib_inflate_workspacesize(),
+		GFP_KERNEL);
+	if (msblk->stream.workspace == NULL) {
+		ERROR("Failed to allocate zlib workspace\n");
+		goto failure;
+	}
+
+	sblk = kzalloc(sizeof(*sblk), GFP_KERNEL);
+	if (sblk == NULL) {
+		ERROR("Failed to allocate squashfs_super_block\n");
+		goto failure;
+	}
+
+	msblk->devblksize = sb_min_blocksize(sb, BLOCK_SIZE);
+	msblk->devblksize_log2 = ffz(~msblk->devblksize);
+
+	mutex_init(&msblk->read_data_mutex);
+	mutex_init(&msblk->meta_index_mutex);
+
+	/*
+	 * msblk->bytes_used is checked in squashfs_read_table to ensure reads
+	 * are not beyond filesystem end.  But as we're using
+	 * squashfs_read_table here to read the superblock (including the value
+	 * of bytes_used) we need to set it to an initial sensible dummy value
+	 */
+	msblk->bytes_used = sizeof(*sblk);
+	err = squashfs_read_table(sb, sblk, SQUASHFS_START, sizeof(*sblk));
+
+	if (err < 0) {
+		ERROR("unable to read squashfs_super_block\n");
+		goto failed_mount;
+	}
+
+	/* Check it is a SQUASHFS superblock */
+	sb->s_magic = le32_to_cpu(sblk->s_magic);
+	if (sb->s_magic != SQUASHFS_MAGIC) {
+		if (!silent)
+			ERROR("Can't find a SQUASHFS superblock on %s\n",
+						bdevname(sb->s_bdev, b));
+		err = -EINVAL;
+		goto failed_mount;
+	}
+
+	/* Check the MAJOR & MINOR versions and compression type */
+	err = supported_squashfs_filesystem(le16_to_cpu(sblk->s_major),
+			le16_to_cpu(sblk->s_minor),
+			le16_to_cpu(sblk->compression));
+	if (err < 0)
+		goto failed_mount;
+
+	err = -EINVAL;
+
+	/*
+	 * Check if there's xattrs in the filesystem.  These are not
+	 * supported in this version, so warn that they will be ignored.
+	 */
+	if (le64_to_cpu(sblk->xattr_table_start) != SQUASHFS_INVALID_BLK)
+		ERROR("Xattrs in filesystem, these will be ignored\n");
+
+	/* Check the filesystem does not extend beyond the end of the
+	   block device */
+	msblk->bytes_used = le64_to_cpu(sblk->bytes_used);
+	if (msblk->bytes_used < 0 || msblk->bytes_used >
+			i_size_read(sb->s_bdev->bd_inode))
+		goto failed_mount;
+
+	/* Check block size for sanity */
+	msblk->block_size = le32_to_cpu(sblk->block_size);
+	if (msblk->block_size > SQUASHFS_FILE_MAX_SIZE)
+		goto failed_mount;
+
+	msblk->block_log = le16_to_cpu(sblk->block_log);
+	if (msblk->block_log > SQUASHFS_FILE_MAX_LOG)
+		goto failed_mount;
+
+	/* Check the root inode for sanity */
+	root_inode = le64_to_cpu(sblk->root_inode);
+	if (SQUASHFS_INODE_OFFSET(root_inode) > SQUASHFS_METADATA_SIZE)
+		goto failed_mount;
+
+	msblk->inode_table = le64_to_cpu(sblk->inode_table_start);
+	msblk->directory_table = le64_to_cpu(sblk->directory_table_start);
+	msblk->inodes = le32_to_cpu(sblk->inodes);
+	flags = le16_to_cpu(sblk->flags);
+
+	TRACE("Found valid superblock on %s\n", bdevname(sb->s_bdev, b));
+	TRACE("Inodes are %scompressed\n", SQUASHFS_UNCOMPRESSED_INODES(flags)
+				? "un" : "");
+	TRACE("Data is %scompressed\n", SQUASHFS_UNCOMPRESSED_DATA(flags)
+				? "un" : "");
+	TRACE("Filesystem size %lld bytes\n", msblk->bytes_used);
+	TRACE("Block size %d\n", msblk->block_size);
+	TRACE("Number of inodes %d\n", msblk->inodes);
+	TRACE("Number of fragments %d\n", le32_to_cpu(sblk->fragments));
+	TRACE("Number of ids %d\n", le16_to_cpu(sblk->no_ids));
+	TRACE("sblk->inode_table_start %llx\n", msblk->inode_table);
+	TRACE("sblk->directory_table_start %llx\n", msblk->directory_table);
+	TRACE("sblk->fragment_table_start %llx\n",
+		(u64) le64_to_cpu(sblk->fragment_table_start));
+	TRACE("sblk->id_table_start %llx\n",
+		(u64) le64_to_cpu(sblk->id_table_start));
+
+	sb->s_maxbytes = MAX_LFS_FILESIZE;
+	sb->s_flags |= MS_RDONLY;
+	sb->s_op = &squashfs_super_ops;
+
+	err = -ENOMEM;
+
+	msblk->block_cache = squashfs_cache_init("metadata",
+			SQUASHFS_CACHED_BLKS, SQUASHFS_METADATA_SIZE);
+	if (msblk->block_cache == NULL)
+		goto failed_mount;
+
+	/* Allocate read_page block */
+	msblk->read_page = squashfs_cache_init("data", 1, msblk->block_size);
+	if (msblk->read_page == NULL) {
+		ERROR("Failed to allocate read_page block\n");
+		goto failed_mount;
+	}
+
+	/* Allocate and read id index table */
+	msblk->id_table = squashfs_read_id_index_table(sb,
+		le64_to_cpu(sblk->id_table_start), le16_to_cpu(sblk->no_ids));
+	if (IS_ERR(msblk->id_table)) {
+		err = PTR_ERR(msblk->id_table);
+		msblk->id_table = NULL;
+		goto failed_mount;
+	}
+
+	fragments = le32_to_cpu(sblk->fragments);
+	if (fragments == 0)
+		goto allocate_lookup_table;
+
+	msblk->fragment_cache = squashfs_cache_init("fragment",
+		SQUASHFS_CACHED_FRAGMENTS, msblk->block_size);
+	if (msblk->fragment_cache == NULL) {
+		err = -ENOMEM;
+		goto failed_mount;
+	}
+
+	/* Allocate and read fragment index table */
+	msblk->fragment_index = squashfs_read_fragment_index_table(sb,
+		le64_to_cpu(sblk->fragment_table_start), fragments);
+	if (IS_ERR(msblk->fragment_index)) {
+		err = PTR_ERR(msblk->fragment_index);
+		msblk->fragment_index = NULL;
+		goto failed_mount;
+	}
+
+allocate_lookup_table:
+	lookup_table_start = le64_to_cpu(sblk->lookup_table_start);
+	if (lookup_table_start == SQUASHFS_INVALID_BLK)
+		goto allocate_root;
+
+	/* Allocate and read inode lookup table */
+	msblk->inode_lookup_table = squashfs_read_inode_lookup_table(sb,
+		lookup_table_start, msblk->inodes);
+	if (IS_ERR(msblk->inode_lookup_table)) {
+		err = PTR_ERR(msblk->inode_lookup_table);
+		msblk->inode_lookup_table = NULL;
+		goto failed_mount;
+	}
+
+	sb->s_export_op = &squashfs_export_ops;
+
+allocate_root:
+	root = new_inode(sb);
+	if (!root) {
+		err = -ENOMEM;
+		goto failed_mount;
+	}
+
+	err = squashfs_read_inode(root, root_inode);
+	if (err) {
+		iget_failed(root);
+		goto failed_mount;
+	}
+	insert_inode_hash(root);
+
+	sb->s_root = d_alloc_root(root);
+	if (sb->s_root == NULL) {
+		ERROR("Root inode create failed\n");
+		err = -ENOMEM;
+		iput(root);
+		goto failed_mount;
+	}
+
+	TRACE("Leaving squashfs_fill_super\n");
+	kfree(sblk);
+	return 0;
+
+failed_mount:
+	squashfs_cache_delete(msblk->block_cache);
+	squashfs_cache_delete(msblk->fragment_cache);
+	squashfs_cache_delete(msblk->read_page);
+	kfree(msblk->inode_lookup_table);
+	kfree(msblk->fragment_index);
+	kfree(msblk->id_table);
+	kfree(msblk->stream.workspace);
+	kfree(sb->s_fs_info);
+	sb->s_fs_info = NULL;
+	kfree(sblk);
+	return err;
+
+failure:
+	kfree(msblk->stream.workspace);
+	kfree(sb->s_fs_info);
+	sb->s_fs_info = NULL;
+	return -ENOMEM;
+}
+
+
+static int squashfs_statfs(struct dentry *dentry, struct kstatfs *buf)
+{
+	struct squashfs_sb_info *msblk = dentry->d_sb->s_fs_info;
+
+	TRACE("Entered squashfs_statfs\n");
+
+	buf->f_type = SQUASHFS_MAGIC;
+	buf->f_bsize = msblk->block_size;
+	buf->f_blocks = ((msblk->bytes_used - 1) >> msblk->block_log) + 1;
+	buf->f_bfree = buf->f_bavail = 0;
+	buf->f_files = msblk->inodes;
+	buf->f_ffree = 0;
+	buf->f_namelen = SQUASHFS_NAME_LEN;
+
+	return 0;
+}
+
+
+static int squashfs_remount(struct super_block *sb, int *flags, char *data)
+{
+	*flags |= MS_RDONLY;
+	return 0;
+}
+
+
+static void squashfs_put_super(struct super_block *sb)
+{
+	if (sb->s_fs_info) {
+		struct squashfs_sb_info *sbi = sb->s_fs_info;
+		squashfs_cache_delete(sbi->block_cache);
+		squashfs_cache_delete(sbi->fragment_cache);
+		squashfs_cache_delete(sbi->read_page);
+		kfree(sbi->id_table);
+		kfree(sbi->fragment_index);
+		kfree(sbi->meta_index);
+		kfree(sbi->stream.workspace);
+		kfree(sb->s_fs_info);
+		sb->s_fs_info = NULL;
+	}
+}
+
+
+static int squashfs_get_sb(struct file_system_type *fs_type, int flags,
+				const char *dev_name, void *data,
+				struct vfsmount *mnt)
+{
+	return get_sb_bdev(fs_type, flags, dev_name, data, squashfs_fill_super,
+				mnt);
+}
+
+
+static struct kmem_cache *squashfs_inode_cachep;
+
+
+static void init_once(void *foo)
+{
+	struct squashfs_inode_info *ei = foo;
+
+	inode_init_once(&ei->vfs_inode);
+}
+
+
+static int __init init_inodecache(void)
+{
+	squashfs_inode_cachep = kmem_cache_create("squashfs_inode_cache",
+		sizeof(struct squashfs_inode_info), 0,
+		SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT, init_once);
+
+	return squashfs_inode_cachep ? 0 : -ENOMEM;
+}
+
+
+static void destroy_inodecache(void)
+{
+	kmem_cache_destroy(squashfs_inode_cachep);
+}
+
+
+static int __init init_squashfs_fs(void)
+{
+	int err = init_inodecache();
+
+	if (err)
+		return err;
+
+	err = register_filesystem(&squashfs_fs_type);
+	if (err) {
+		destroy_inodecache();
+		return err;
+	}
+
+	printk(KERN_INFO "squashfs: version 4.0 (2009/01/03) "
+		"Phillip Lougher\n");
+
+	return 0;
+}
+
+
+static void __exit exit_squashfs_fs(void)
+{
+	unregister_filesystem(&squashfs_fs_type);
+	destroy_inodecache();
+}
+
+
+static struct inode *squashfs_alloc_inode(struct super_block *sb)
+{
+	struct squashfs_inode_info *ei =
+		kmem_cache_alloc(squashfs_inode_cachep, GFP_KERNEL);
+
+	return ei ? &ei->vfs_inode : NULL;
+}
+
+
+static void squashfs_destroy_inode(struct inode *inode)
+{
+	kmem_cache_free(squashfs_inode_cachep, squashfs_i(inode));
+}
+
+
+static struct file_system_type squashfs_fs_type = {
+	.owner = THIS_MODULE,
+	.name = "squashfs",
+	.get_sb = squashfs_get_sb,
+	.kill_sb = kill_block_super,
+	.fs_flags = FS_REQUIRES_DEV
+};
+
+static struct super_operations squashfs_super_ops = {
+	.alloc_inode = squashfs_alloc_inode,
+	.destroy_inode = squashfs_destroy_inode,
+	.statfs = squashfs_statfs,
+	.put_super = squashfs_put_super,
+	.remount_fs = squashfs_remount
+};
+
+module_init(init_squashfs_fs);
+module_exit(exit_squashfs_fs);
+MODULE_DESCRIPTION("squashfs 4.0, a compressed read-only filesystem");
+MODULE_AUTHOR("Phillip Lougher <phillip@lougher.demon.co.uk>");
+MODULE_LICENSE("GPL");
diff --git a/fs/squashfs/symlink.c b/fs/squashfs/symlink.c
new file mode 100644
index 00000000000..83d87880aac
--- /dev/null
+++ b/fs/squashfs/symlink.c
@@ -0,0 +1,118 @@
+/*
+ * Squashfs - a compressed read only filesystem for Linux
+ *
+ * Copyright (c) 2002, 2003, 2004, 2005, 2006, 2007, 2008
+ * Phillip Lougher <phillip@lougher.demon.co.uk>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2,
+ * or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * symlink.c
+ */
+
+/*
+ * This file implements code to handle symbolic links.
+ *
+ * The data contents of symbolic links are stored inside the symbolic
+ * link inode within the inode table.  This allows the normally small symbolic
+ * link to be compressed as part of the inode table, achieving much greater
+ * compression than if the symbolic link was compressed individually.
+ */
+
+#include <linux/fs.h>
+#include <linux/vfs.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/pagemap.h>
+#include <linux/zlib.h>
+
+#include "squashfs_fs.h"
+#include "squashfs_fs_sb.h"
+#include "squashfs_fs_i.h"
+#include "squashfs.h"
+
+static int squashfs_symlink_readpage(struct file *file, struct page *page)
+{
+	struct inode *inode = page->mapping->host;
+	struct super_block *sb = inode->i_sb;
+	struct squashfs_sb_info *msblk = sb->s_fs_info;
+	int index = page->index << PAGE_CACHE_SHIFT;
+	u64 block = squashfs_i(inode)->start;
+	int offset = squashfs_i(inode)->offset;
+	int length = min_t(int, i_size_read(inode) - index, PAGE_CACHE_SIZE);
+	int bytes, copied;
+	void *pageaddr;
+	struct squashfs_cache_entry *entry;
+
+	TRACE("Entered squashfs_symlink_readpage, page index %ld, start block "
+			"%llx, offset %x\n", page->index, block, offset);
+
+	/*
+	 * Skip index bytes into symlink metadata.
+	 */
+	if (index) {
+		bytes = squashfs_read_metadata(sb, NULL, &block, &offset,
+								index);
+		if (bytes < 0) {
+			ERROR("Unable to read symlink [%llx:%x]\n",
+				squashfs_i(inode)->start,
+				squashfs_i(inode)->offset);
+			goto error_out;
+		}
+	}
+
+	/*
+	 * Read length bytes from symlink metadata.  Squashfs_read_metadata
+	 * is not used here because it can sleep and we want to use
+	 * kmap_atomic to map the page.  Instead call the underlying
+	 * squashfs_cache_get routine.  As length bytes may overlap metadata
+	 * blocks, we may need to call squashfs_cache_get multiple times.
+	 */
+	for (bytes = 0; bytes < length; offset = 0, bytes += copied) {
+		entry = squashfs_cache_get(sb, msblk->block_cache, block, 0);
+		if (entry->error) {
+			ERROR("Unable to read symlink [%llx:%x]\n",
+				squashfs_i(inode)->start,
+				squashfs_i(inode)->offset);
+			squashfs_cache_put(entry);
+			goto error_out;
+		}
+
+		pageaddr = kmap_atomic(page, KM_USER0);
+		copied = squashfs_copy_data(pageaddr + bytes, entry, offset,
+								length - bytes);
+		if (copied == length - bytes)
+			memset(pageaddr + length, 0, PAGE_CACHE_SIZE - length);
+		else
+			block = entry->next_index;
+		kunmap_atomic(pageaddr, KM_USER0);
+		squashfs_cache_put(entry);
+	}
+
+	flush_dcache_page(page);
+	SetPageUptodate(page);
+	unlock_page(page);
+	return 0;
+
+error_out:
+	SetPageError(page);
+	unlock_page(page);
+	return 0;
+}
+
+
+const struct address_space_operations squashfs_symlink_aops = {
+	.readpage = squashfs_symlink_readpage
+};
diff --git a/fs/stat.c b/fs/stat.c
index 7c46fbeb8b7..7e12a6f8279 100644
--- a/fs/stat.c
+++ b/fs/stat.c
@@ -305,7 +305,7 @@ asmlinkage long sys_readlinkat(int dfd, const char __user *pathname,
 		struct inode *inode = path.dentry->d_inode;
 
 		error = -EINVAL;
-		if (inode->i_op && inode->i_op->readlink) {
+		if (inode->i_op->readlink) {
 			error = security_inode_readlink(path.dentry);
 			if (!error) {
 				touch_atime(path.mnt, path.dentry);
diff --git a/fs/super.c b/fs/super.c
index ddba069d7a9..ed080c41716 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -38,6 +38,7 @@
 #include <linux/kobject.h>
 #include <linux/mutex.h>
 #include <linux/file.h>
+#include <linux/async.h>
 #include <asm/uaccess.h>
 #include "internal.h"
 
@@ -71,6 +72,7 @@ static struct super_block *alloc_super(struct file_system_type *type)
 		INIT_HLIST_HEAD(&s->s_anon);
 		INIT_LIST_HEAD(&s->s_inodes);
 		INIT_LIST_HEAD(&s->s_dentry_lru);
+		INIT_LIST_HEAD(&s->s_async_list);
 		init_rwsem(&s->s_umount);
 		mutex_init(&s->s_lock);
 		lockdep_set_class(&s->s_umount, &type->s_umount_key);
@@ -289,11 +291,18 @@ void generic_shutdown_super(struct super_block *sb)
 {
 	const struct super_operations *sop = sb->s_op;
 
+
 	if (sb->s_root) {
 		shrink_dcache_for_umount(sb);
 		fsync_super(sb);
 		lock_super(sb);
 		sb->s_flags &= ~MS_ACTIVE;
+
+		/*
+		 * wait for asynchronous fs operations to finish before going further
+		 */
+		async_synchronize_full_special(&sb->s_async_list);
+
 		/* bad name - it should be evict_inodes() */
 		invalidate_inodes(sb);
 		lock_kernel();
@@ -461,6 +470,7 @@ restart:
 		sb->s_count++;
 		spin_unlock(&sb_lock);
 		down_read(&sb->s_umount);
+		async_synchronize_full_special(&sb->s_async_list);
 		if (sb->s_root && (wait || sb->s_dirt))
 			sb->s_op->sync_fs(sb, wait);
 		up_read(&sb->s_umount);
@@ -800,6 +810,7 @@ int get_sb_bdev(struct file_system_type *fs_type,
 		}
 
 		s->s_flags |= MS_ACTIVE;
+		bdev->bd_super = s;
 	}
 
 	return simple_set_mnt(mnt, s);
@@ -819,6 +830,7 @@ void kill_block_super(struct super_block *sb)
 	struct block_device *bdev = sb->s_bdev;
 	fmode_t mode = sb->s_mode;
 
+	bdev->bd_super = 0;
 	generic_shutdown_super(sb);
 	sync_blockdev(bdev);
 	close_bdev_exclusive(bdev, mode);
diff --git a/fs/sync.c b/fs/sync.c
index 2967562d416..ac02b56548b 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -75,14 +75,39 @@ int file_fsync(struct file *filp, struct dentry *dentry, int datasync)
 	return ret;
 }
 
-long do_fsync(struct file *file, int datasync)
+/**
+ * vfs_fsync - perform a fsync or fdatasync on a file
+ * @file:		file to sync
+ * @dentry:		dentry of @file
+ * @data:		only perform a fdatasync operation
+ *
+ * Write back data and metadata for @file to disk.  If @datasync is
+ * set only metadata needed to access modified file data is written.
+ *
+ * In case this function is called from nfsd @file may be %NULL and
+ * only @dentry is set.  This can only happen when the filesystem
+ * implements the export_operations API.
+ */
+int vfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 {
-	int ret;
-	int err;
-	struct address_space *mapping = file->f_mapping;
+	const struct file_operations *fop;
+	struct address_space *mapping;
+	int err, ret;
+
+	/*
+	 * Get mapping and operations from the file in case we have
+	 * as file, or get the default values for them in case we
+	 * don't have a struct file available.  Damn nfsd..
+	 */
+	if (file) {
+		mapping = file->f_mapping;
+		fop = file->f_op;
+	} else {
+		mapping = dentry->d_inode->i_mapping;
+		fop = dentry->d_inode->i_fop;
+	}
 
-	if (!file->f_op || !file->f_op->fsync) {
-		/* Why?  We can still call filemap_fdatawrite */
+	if (!fop || !fop->fsync) {
 		ret = -EINVAL;
 		goto out;
 	}
@@ -94,7 +119,7 @@ long do_fsync(struct file *file, int datasync)
 	 * livelocks in fsync_buffers_list().
 	 */
 	mutex_lock(&mapping->host->i_mutex);
-	err = file->f_op->fsync(file, file->f_path.dentry, datasync);
+	err = fop->fsync(file, dentry, datasync);
 	if (!ret)
 		ret = err;
 	mutex_unlock(&mapping->host->i_mutex);
@@ -104,15 +129,16 @@ long do_fsync(struct file *file, int datasync)
 out:
 	return ret;
 }
+EXPORT_SYMBOL(vfs_fsync);
 
-static long __do_fsync(unsigned int fd, int datasync)
+static int do_fsync(unsigned int fd, int datasync)
 {
 	struct file *file;
 	int ret = -EBADF;
 
 	file = fget(fd);
 	if (file) {
-		ret = do_fsync(file, datasync);
+		ret = vfs_fsync(file, file->f_path.dentry, datasync);
 		fput(file);
 	}
 	return ret;
@@ -120,12 +146,12 @@ static long __do_fsync(unsigned int fd, int datasync)
 
 asmlinkage long sys_fsync(unsigned int fd)
 {
-	return __do_fsync(fd, 0);
+	return do_fsync(fd, 0);
 }
 
 asmlinkage long sys_fdatasync(unsigned int fd)
 {
-	return __do_fsync(fd, 1);
+	return do_fsync(fd, 1);
 }
 
 /*
@@ -269,7 +295,7 @@ int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
 
 	if (flags & SYNC_FILE_RANGE_WRITE) {
 		ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
-						WB_SYNC_NONE);
+						WB_SYNC_ALL);
 		if (ret < 0)
 			goto out;
 	}
diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c
index eb53c632f85..dfa3d94cfc7 100644
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -107,8 +107,6 @@ int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 static inline void set_default_inode_attr(struct inode * inode, mode_t mode)
 {
 	inode->i_mode = mode;
-	inode->i_uid = 0;
-	inode->i_gid = 0;
 	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 }
 
@@ -149,7 +147,6 @@ static void sysfs_init_inode(struct sysfs_dirent *sd, struct inode *inode)
 {
 	struct bin_attribute *bin_attr;
 
-	inode->i_blocks = 0;
 	inode->i_mapping->a_ops = &sysfs_aops;
 	inode->i_mapping->backing_dev_info = &sysfs_backing_dev_info;
 	inode->i_op = &sysfs_inode_operations;
diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c
index df0d435baa4..3d81bf58dae 100644
--- a/fs/sysv/inode.c
+++ b/fs/sysv/inode.c
@@ -27,6 +27,7 @@
 #include <linux/init.h>
 #include <linux/buffer_head.h>
 #include <linux/vfs.h>
+#include <linux/namei.h>
 #include <asm/byteorder.h>
 #include "sysv.h"
 
@@ -163,8 +164,11 @@ void sysv_set_inode(struct inode *inode, dev_t rdev)
 		if (inode->i_blocks) {
 			inode->i_op = &sysv_symlink_inode_operations;
 			inode->i_mapping->a_ops = &sysv_aops;
-		} else
+		} else {
 			inode->i_op = &sysv_fast_symlink_inode_operations;
+			nd_terminate_link(SYSV_I(inode)->i_data, inode->i_size,
+				sizeof(SYSV_I(inode)->i_data) - 1);
+		}
 	} else
 		init_special_inode(inode, inode->i_mode, rdev);
 }
diff --git a/fs/ubifs/Kconfig b/fs/ubifs/Kconfig
index 91ceeda7e5b..e35b54d5059 100644
--- a/fs/ubifs/Kconfig
+++ b/fs/ubifs/Kconfig
@@ -40,7 +40,7 @@ config UBIFS_FS_ZLIB
 	depends on UBIFS_FS
 	default y
 	help
-	  Zlib copresses better then LZO but it is slower. Say 'Y' if unsure.
+	  Zlib compresses better than LZO but it is slower. Say 'Y' if unsure.
 
 # Debugging-related stuff
 config UBIFS_FS_DEBUG
diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c
index 4a18f084cc4..175f9c590b7 100644
--- a/fs/ubifs/budget.c
+++ b/fs/ubifs/budget.c
@@ -32,18 +32,15 @@
 
 #include "ubifs.h"
 #include <linux/writeback.h>
-#include <asm/div64.h>
+#include <linux/math64.h>
 
 /*
  * When pessimistic budget calculations say that there is no enough space,
  * UBIFS starts writing back dirty inodes and pages, doing garbage collection,
- * or committing. The below constants define maximum number of times UBIFS
+ * or committing. The below constant defines maximum number of times UBIFS
  * repeats the operations.
  */
-#define MAX_SHRINK_RETRIES 8
-#define MAX_GC_RETRIES     4
-#define MAX_CMT_RETRIES    2
-#define MAX_NOSPC_RETRIES  1
+#define MAX_MKSPC_RETRIES 3
 
 /*
  * The below constant defines amount of dirty pages which should be written
@@ -52,30 +49,6 @@
 #define NR_TO_WRITE 16
 
 /**
- * struct retries_info - information about re-tries while making free space.
- * @prev_liability: previous liability
- * @shrink_cnt: how many times the liability was shrinked
- * @shrink_retries: count of liability shrink re-tries (increased when
- *                  liability does not shrink)
- * @try_gc: GC should be tried first
- * @gc_retries: how many times GC was run
- * @cmt_retries: how many times commit has been done
- * @nospc_retries: how many times GC returned %-ENOSPC
- *
- * Since we consider budgeting to be the fast-path, and this structure has to
- * be allocated on stack and zeroed out, we make it smaller using bit-fields.
- */
-struct retries_info {
-	long long prev_liability;
-	unsigned int shrink_cnt;
-	unsigned int shrink_retries:5;
-	unsigned int try_gc:1;
-	unsigned int gc_retries:4;
-	unsigned int cmt_retries:3;
-	unsigned int nospc_retries:1;
-};
-
-/**
  * shrink_liability - write-back some dirty pages/inodes.
  * @c: UBIFS file-system description object
  * @nr_to_write: how many dirty pages to write-back
@@ -147,13 +120,29 @@ static int run_gc(struct ubifs_info *c)
 }
 
 /**
+ * get_liability - calculate current liability.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates and returns current UBIFS liability, i.e. the
+ * amount of bytes UBIFS has "promised" to write to the media.
+ */
+static long long get_liability(struct ubifs_info *c)
+{
+	long long liab;
+
+	spin_lock(&c->space_lock);
+	liab = c->budg_idx_growth + c->budg_data_growth + c->budg_dd_growth;
+	spin_unlock(&c->space_lock);
+	return liab;
+}
+
+/**
  * make_free_space - make more free space on the file-system.
  * @c: UBIFS file-system description object
- * @ri: information about previous invocations of this function
  *
  * This function is called when an operation cannot be budgeted because there
  * is supposedly no free space. But in most cases there is some free space:
- *   o budgeting is pessimistic, so it always budgets more then it is actually
+ *   o budgeting is pessimistic, so it always budgets more than it is actually
  *     needed, so shrinking the liability is one way to make free space - the
  *     cached data will take less space then it was budgeted for;
  *   o GC may turn some dark space into free space (budgeting treats dark space
@@ -165,87 +154,42 @@ static int run_gc(struct ubifs_info *c)
  * Returns %-ENOSPC if it couldn't do more free space, and other negative error
  * codes on failures.
  */
-static int make_free_space(struct ubifs_info *c, struct retries_info *ri)
+static int make_free_space(struct ubifs_info *c)
 {
-	int err;
-
-	/*
-	 * If we have some dirty pages and inodes (liability), try to write
-	 * them back unless this was tried too many times without effect
-	 * already.
-	 */
-	if (ri->shrink_retries < MAX_SHRINK_RETRIES && !ri->try_gc) {
-		long long liability;
-
-		spin_lock(&c->space_lock);
-		liability = c->budg_idx_growth + c->budg_data_growth +
-			    c->budg_dd_growth;
-		spin_unlock(&c->space_lock);
+	int err, retries = 0;
+	long long liab1, liab2;
 
-		if (ri->prev_liability >= liability) {
-			/* Liability does not shrink, next time try GC then */
-			ri->shrink_retries += 1;
-			if (ri->gc_retries < MAX_GC_RETRIES)
-				ri->try_gc = 1;
-			dbg_budg("liability did not shrink: retries %d of %d",
-				 ri->shrink_retries, MAX_SHRINK_RETRIES);
-		}
+	do {
+		liab1 = get_liability(c);
+		/*
+		 * We probably have some dirty pages or inodes (liability), try
+		 * to write them back.
+		 */
+		dbg_budg("liability %lld, run write-back", liab1);
+		shrink_liability(c, NR_TO_WRITE);
 
-		dbg_budg("force write-back (count %d)", ri->shrink_cnt);
-		shrink_liability(c, NR_TO_WRITE + ri->shrink_cnt);
+		liab2 = get_liability(c);
+		if (liab2 < liab1)
+			return -EAGAIN;
 
-		ri->prev_liability = liability;
-		ri->shrink_cnt += 1;
-		return -EAGAIN;
-	}
+		dbg_budg("new liability %lld (not shrinked)", liab2);
 
-	/*
-	 * Try to run garbage collector unless it was already tried too many
-	 * times.
-	 */
-	if (ri->gc_retries < MAX_GC_RETRIES) {
-		ri->gc_retries += 1;
-		dbg_budg("run GC, retries %d of %d",
-			 ri->gc_retries, MAX_GC_RETRIES);
-
-		ri->try_gc = 0;
+		/* Liability did not shrink again, try GC */
+		dbg_budg("Run GC");
 		err = run_gc(c);
 		if (!err)
 			return -EAGAIN;
 
-		if (err == -EAGAIN) {
-			dbg_budg("GC asked to commit");
-			err = ubifs_run_commit(c);
-			if (err)
-				return err;
-			return -EAGAIN;
-		}
-
-		if (err != -ENOSPC)
-			return err;
-
-		/*
-		 * GC could not make any progress. If this is the first time,
-		 * then it makes sense to try to commit, because it might make
-		 * some dirty space.
-		 */
-		dbg_budg("GC returned -ENOSPC, retries %d",
-			 ri->nospc_retries);
-		if (ri->nospc_retries >= MAX_NOSPC_RETRIES)
+		if (err != -EAGAIN && err != -ENOSPC)
+			/* Some real error happened */
 			return err;
-		ri->nospc_retries += 1;
-	}
 
-	/* Neither GC nor write-back helped, try to commit */
-	if (ri->cmt_retries < MAX_CMT_RETRIES) {
-		ri->cmt_retries += 1;
-		dbg_budg("run commit, retries %d of %d",
-			 ri->cmt_retries, MAX_CMT_RETRIES);
+		dbg_budg("Run commit (retries %d)", retries);
 		err = ubifs_run_commit(c);
 		if (err)
 			return err;
-		return -EAGAIN;
-	}
+	} while (retries++ < MAX_MKSPC_RETRIES);
+
 	return -ENOSPC;
 }
 
@@ -258,8 +202,8 @@ static int make_free_space(struct ubifs_info *c, struct retries_info *ri)
  */
 int ubifs_calc_min_idx_lebs(struct ubifs_info *c)
 {
-	int ret;
-	uint64_t idx_size;
+	int idx_lebs, eff_leb_size = c->leb_size - c->max_idx_node_sz;
+	long long idx_size;
 
 	idx_size = c->old_idx_sz + c->budg_idx_growth + c->budg_uncommitted_idx;
 
@@ -271,23 +215,16 @@ int ubifs_calc_min_idx_lebs(struct ubifs_info *c)
 	 * pair, nor similarly the two variables for the new index size, so we
 	 * have to do this costly 64-bit division on fast-path.
 	 */
-	if (do_div(idx_size, c->leb_size - c->max_idx_node_sz))
-		ret = idx_size + 1;
-	else
-		ret = idx_size;
+	idx_size += eff_leb_size - 1;
+	idx_lebs = div_u64(idx_size, eff_leb_size);
 	/*
 	 * The index head is not available for the in-the-gaps method, so add an
 	 * extra LEB to compensate.
 	 */
-	ret += 1;
-	/*
-	 * At present the index needs at least 2 LEBs: one for the index head
-	 * and one for in-the-gaps method (which currently does not cater for
-	 * the index head and so excludes it from consideration).
-	 */
-	if (ret < 2)
-		ret = 2;
-	return ret;
+	idx_lebs += 1;
+	if (idx_lebs < MIN_INDEX_LEBS)
+		idx_lebs = MIN_INDEX_LEBS;
+	return idx_lebs;
 }
 
 /**
@@ -530,8 +467,7 @@ static int calc_dd_growth(const struct ubifs_info *c,
 int ubifs_budget_space(struct ubifs_info *c, struct ubifs_budget_req *req)
 {
 	int uninitialized_var(cmt_retries), uninitialized_var(wb_retries);
-	int err, idx_growth, data_growth, dd_growth;
-	struct retries_info ri;
+	int err, idx_growth, data_growth, dd_growth, retried = 0;
 
 	ubifs_assert(req->new_page <= 1);
 	ubifs_assert(req->dirtied_page <= 1);
@@ -549,7 +485,6 @@ int ubifs_budget_space(struct ubifs_info *c, struct ubifs_budget_req *req)
 	if (!data_growth && !dd_growth)
 		return 0;
 	idx_growth = calc_idx_growth(c, req);
-	memset(&ri, 0, sizeof(struct retries_info));
 
 again:
 	spin_lock(&c->space_lock);
@@ -587,12 +522,17 @@ again:
 		return err;
 	}
 
-	err = make_free_space(c, &ri);
+	err = make_free_space(c);
+	cond_resched();
 	if (err == -EAGAIN) {
 		dbg_budg("try again");
-		cond_resched();
 		goto again;
 	} else if (err == -ENOSPC) {
+		if (!retried) {
+			retried = 1;
+			dbg_budg("-ENOSPC, but anyway try once again");
+			goto again;
+		}
 		dbg_budg("FS is full, -ENOSPC");
 		c->nospace = 1;
 		if (can_use_rp(c) || c->rp_size == 0)
@@ -666,7 +606,7 @@ void ubifs_release_budget(struct ubifs_info *c, struct ubifs_budget_req *req)
  * @c: UBIFS file-system description object
  *
  * This function converts budget which was allocated for a new page of data to
- * the budget of changing an existing page of data. The latter is smaller then
+ * the budget of changing an existing page of data. The latter is smaller than
  * the former, so this function only does simple re-calculation and does not
  * involve any write-back.
  */
@@ -712,9 +652,9 @@ void ubifs_release_dirty_inode_budget(struct ubifs_info *c,
  * user-space. User-space application tend to expect that if the file-system
  * (e.g., via the 'statfs()' call) reports that it has N bytes available, they
  * are able to write a file of size N. UBIFS attaches node headers to each data
- * node and it has to write indexind nodes as well. This introduces additional
- * overhead, and UBIFS it has to report sligtly less free space to meet the
- * above expectetion.
+ * node and it has to write indexing nodes as well. This introduces additional
+ * overhead, and UBIFS has to report slightly less free space to meet the above
+ * expectations.
  *
  * This function assumes free space is made up of uncompressed data nodes and
  * full index nodes (one per data node, tripled because we always allow enough
@@ -723,7 +663,7 @@ void ubifs_release_dirty_inode_budget(struct ubifs_info *c,
  * Note, the calculation is pessimistic, which means that most of the time
  * UBIFS reports less space than it actually has.
  */
-long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free)
+long long ubifs_reported_space(const struct ubifs_info *c, long long free)
 {
 	int divisor, factor, f;
 
@@ -737,7 +677,7 @@ long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free)
 	 * of data nodes, f - fanout. Because effective UBIFS fanout is twice
 	 * as less than maximum fanout, we assume that each data node
 	 * introduces 3 * @c->max_idx_node_sz / (@c->fanout/2 - 1) bytes.
-	 * Note, the multiplier 3 is because UBIFS reseves thrice as more space
+	 * Note, the multiplier 3 is because UBIFS reserves thrice as more space
 	 * for the index.
 	 */
 	f = c->fanout > 3 ? c->fanout >> 1 : 2;
@@ -745,8 +685,7 @@ long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free)
 	divisor = UBIFS_MAX_DATA_NODE_SZ;
 	divisor += (c->max_idx_node_sz * 3) / (f - 1);
 	free *= factor;
-	do_div(free, divisor);
-	return free;
+	return div_u64(free, divisor);
 }
 
 /**
@@ -756,10 +695,10 @@ long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free)
  * This function calculates amount of free space to report to user-space.
  *
  * Because UBIFS may introduce substantial overhead (the index, node headers,
- * alighment, wastage at the end of eraseblocks, etc), it cannot report real
+ * alignment, wastage at the end of eraseblocks, etc), it cannot report real
  * amount of free flash space it has (well, because not all dirty space is
- * reclamable, UBIFS does not actually know the real amount). If UBIFS did so,
- * it would bread user expectetion about what free space is. Users seem to
+ * reclaimable, UBIFS does not actually know the real amount). If UBIFS did so,
+ * it would bread user expectations about what free space is. Users seem to
  * accustomed to assume that if the file-system reports N bytes of free space,
  * they would be able to fit a file of N bytes to the FS. This almost works for
  * traditional file-systems, because they have way less overhead than UBIFS.
@@ -771,18 +710,9 @@ long long ubifs_get_free_space(struct ubifs_info *c)
 	long long available, outstanding, free;
 
 	spin_lock(&c->space_lock);
-	min_idx_lebs = ubifs_calc_min_idx_lebs(c);
+	min_idx_lebs = c->min_idx_lebs;
+	ubifs_assert(min_idx_lebs == ubifs_calc_min_idx_lebs(c));
 	outstanding = c->budg_data_growth + c->budg_dd_growth;
-
-	/*
-	 * Force the amount available to the total size reported if the used
-	 * space is zero.
-	 */
-	if (c->lst.total_used <= UBIFS_INO_NODE_SZ && !outstanding) {
-		spin_unlock(&c->space_lock);
-		return (long long)c->block_cnt << UBIFS_BLOCK_SHIFT;
-	}
-
 	available = ubifs_calc_available(c, min_idx_lebs);
 
 	/*
diff --git a/fs/ubifs/commit.c b/fs/ubifs/commit.c
index b49884c8c10..f3a7945527f 100644
--- a/fs/ubifs/commit.c
+++ b/fs/ubifs/commit.c
@@ -470,12 +470,12 @@ int dbg_old_index_check_init(struct ubifs_info *c, struct ubifs_zbranch *zroot)
 {
 	struct ubifs_idx_node *idx;
 	int lnum, offs, len, err = 0;
+	struct ubifs_debug_info *d = c->dbg;
 
-	c->old_zroot = *zroot;
-
-	lnum = c->old_zroot.lnum;
-	offs = c->old_zroot.offs;
-	len = c->old_zroot.len;
+	d->old_zroot = *zroot;
+	lnum = d->old_zroot.lnum;
+	offs = d->old_zroot.offs;
+	len = d->old_zroot.len;
 
 	idx = kmalloc(c->max_idx_node_sz, GFP_NOFS);
 	if (!idx)
@@ -485,8 +485,8 @@ int dbg_old_index_check_init(struct ubifs_info *c, struct ubifs_zbranch *zroot)
 	if (err)
 		goto out;
 
-	c->old_zroot_level = le16_to_cpu(idx->level);
-	c->old_zroot_sqnum = le64_to_cpu(idx->ch.sqnum);
+	d->old_zroot_level = le16_to_cpu(idx->level);
+	d->old_zroot_sqnum = le64_to_cpu(idx->ch.sqnum);
 out:
 	kfree(idx);
 	return err;
@@ -509,6 +509,7 @@ int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot)
 {
 	int lnum, offs, len, err = 0, uninitialized_var(last_level), child_cnt;
 	int first = 1, iip;
+	struct ubifs_debug_info *d = c->dbg;
 	union ubifs_key lower_key, upper_key, l_key, u_key;
 	unsigned long long uninitialized_var(last_sqnum);
 	struct ubifs_idx_node *idx;
@@ -525,9 +526,9 @@ int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot)
 	     UBIFS_IDX_NODE_SZ;
 
 	/* Start at the old zroot */
-	lnum = c->old_zroot.lnum;
-	offs = c->old_zroot.offs;
-	len = c->old_zroot.len;
+	lnum = d->old_zroot.lnum;
+	offs = d->old_zroot.offs;
+	len = d->old_zroot.len;
 	iip = 0;
 
 	/*
@@ -560,11 +561,11 @@ int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot)
 		if (first) {
 			first = 0;
 			/* Check root level and sqnum */
-			if (le16_to_cpu(idx->level) != c->old_zroot_level) {
+			if (le16_to_cpu(idx->level) != d->old_zroot_level) {
 				err = 2;
 				goto out_dump;
 			}
-			if (le64_to_cpu(idx->ch.sqnum) != c->old_zroot_sqnum) {
+			if (le64_to_cpu(idx->ch.sqnum) != d->old_zroot_sqnum) {
 				err = 3;
 				goto out_dump;
 			}
diff --git a/fs/ubifs/compress.c b/fs/ubifs/compress.c
index a0ada596b17..11e4132f314 100644
--- a/fs/ubifs/compress.c
+++ b/fs/ubifs/compress.c
@@ -33,7 +33,7 @@
 /* Fake description object for the "none" compressor */
 static struct ubifs_compressor none_compr = {
 	.compr_type = UBIFS_COMPR_NONE,
-	.name = "no compression",
+	.name = "none",
 	.capi_name = "",
 };
 
@@ -43,13 +43,13 @@ static DEFINE_MUTEX(lzo_mutex);
 static struct ubifs_compressor lzo_compr = {
 	.compr_type = UBIFS_COMPR_LZO,
 	.comp_mutex = &lzo_mutex,
-	.name = "LZO",
+	.name = "lzo",
 	.capi_name = "lzo",
 };
 #else
 static struct ubifs_compressor lzo_compr = {
 	.compr_type = UBIFS_COMPR_LZO,
-	.name = "LZO",
+	.name = "lzo",
 };
 #endif
 
@@ -108,7 +108,7 @@ void ubifs_compress(const void *in_buf, int in_len, void *out_buf, int *out_len,
 	if (compr->comp_mutex)
 		mutex_lock(compr->comp_mutex);
 	err = crypto_comp_compress(compr->cc, in_buf, in_len, out_buf,
-				   out_len);
+				   (unsigned int *)out_len);
 	if (compr->comp_mutex)
 		mutex_unlock(compr->comp_mutex);
 	if (unlikely(err)) {
@@ -119,10 +119,10 @@ void ubifs_compress(const void *in_buf, int in_len, void *out_buf, int *out_len,
 	}
 
 	/*
-	 * Presently, we just require that compression results in less data,
-	 * rather than any defined minimum compression ratio or amount.
+	 * If the data compressed only slightly, it is better to leave it
+	 * uncompressed to improve read speed.
 	 */
-	if (ALIGN(*out_len, 8) >= ALIGN(in_len, 8))
+	if (in_len - *out_len < UBIFS_MIN_COMPRESS_DIFF)
 		goto no_compr;
 
 	return;
@@ -172,7 +172,7 @@ int ubifs_decompress(const void *in_buf, int in_len, void *out_buf,
 	if (compr->decomp_mutex)
 		mutex_lock(compr->decomp_mutex);
 	err = crypto_comp_decompress(compr->cc, in_buf, in_len, out_buf,
-				     out_len);
+				     (unsigned int *)out_len);
 	if (compr->decomp_mutex)
 		mutex_unlock(compr->decomp_mutex);
 	if (err)
@@ -244,7 +244,7 @@ out_lzo:
 /**
  * ubifs_compressors_exit - de-initialize UBIFS compressors.
  */
-void __exit ubifs_compressors_exit(void)
+void ubifs_compressors_exit(void)
 {
 	compr_exit(&lzo_compr);
 	compr_exit(&zlib_compr);
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index 510ffa0bbda..792c5a16c18 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -32,6 +32,8 @@
 #include "ubifs.h"
 #include <linux/module.h>
 #include <linux/moduleparam.h>
+#include <linux/debugfs.h>
+#include <linux/math64.h>
 
 #ifdef CONFIG_UBIFS_FS_DEBUG
 
@@ -596,7 +598,9 @@ void dbg_dump_budg(struct ubifs_info *c)
 	struct rb_node *rb;
 	struct ubifs_bud *bud;
 	struct ubifs_gced_idx_leb *idx_gc;
+	long long available, outstanding, free;
 
+	ubifs_assert(spin_is_locked(&c->space_lock));
 	spin_lock(&dbg_lock);
 	printk(KERN_DEBUG "(pid %d) Budgeting info: budg_data_growth %lld, "
 	       "budg_dd_growth %lld, budg_idx_growth %lld\n", current->pid,
@@ -629,6 +633,17 @@ void dbg_dump_budg(struct ubifs_info *c)
 		printk(KERN_DEBUG "\tGC'ed idx LEB %d unmap %d\n",
 		       idx_gc->lnum, idx_gc->unmap);
 	printk(KERN_DEBUG "\tcommit state %d\n", c->cmt_state);
+
+	/* Print budgeting predictions */
+	available = ubifs_calc_available(c, c->min_idx_lebs);
+	outstanding = c->budg_data_growth + c->budg_dd_growth;
+	if (available > outstanding)
+		free = ubifs_reported_space(c, available - outstanding);
+	else
+		free = 0;
+	printk(KERN_DEBUG "Budgeting predictions:\n");
+	printk(KERN_DEBUG "\tavailable: %lld, outstanding %lld, free %lld\n",
+	       available, outstanding, free);
 	spin_unlock(&dbg_lock);
 }
 
@@ -645,7 +660,8 @@ void dbg_dump_lprops(struct ubifs_info *c)
 	struct ubifs_lprops lp;
 	struct ubifs_lp_stats lst;
 
-	printk(KERN_DEBUG "(pid %d) Dumping LEB properties\n", current->pid);
+	printk(KERN_DEBUG "(pid %d) start dumping LEB properties\n",
+	       current->pid);
 	ubifs_get_lp_stats(c, &lst);
 	dbg_dump_lstats(&lst);
 
@@ -656,6 +672,8 @@ void dbg_dump_lprops(struct ubifs_info *c)
 
 		dbg_dump_lprop(c, &lp);
 	}
+	printk(KERN_DEBUG "(pid %d) finish dumping LEB properties\n",
+	       current->pid);
 }
 
 void dbg_dump_lpt_info(struct ubifs_info *c)
@@ -663,6 +681,7 @@ void dbg_dump_lpt_info(struct ubifs_info *c)
 	int i;
 
 	spin_lock(&dbg_lock);
+	printk(KERN_DEBUG "(pid %d) dumping LPT information\n", current->pid);
 	printk(KERN_DEBUG "\tlpt_sz:        %lld\n", c->lpt_sz);
 	printk(KERN_DEBUG "\tpnode_sz:      %d\n", c->pnode_sz);
 	printk(KERN_DEBUG "\tnnode_sz:      %d\n", c->nnode_sz);
@@ -684,7 +703,8 @@ void dbg_dump_lpt_info(struct ubifs_info *c)
 	printk(KERN_DEBUG "\tLPT root is at %d:%d\n", c->lpt_lnum, c->lpt_offs);
 	printk(KERN_DEBUG "\tLPT head is at %d:%d\n",
 	       c->nhead_lnum, c->nhead_offs);
-	printk(KERN_DEBUG "\tLPT ltab is at %d:%d\n", c->ltab_lnum, c->ltab_offs);
+	printk(KERN_DEBUG "\tLPT ltab is at %d:%d\n",
+	       c->ltab_lnum, c->ltab_offs);
 	if (c->big_lpt)
 		printk(KERN_DEBUG "\tLPT lsave is at %d:%d\n",
 		       c->lsave_lnum, c->lsave_offs);
@@ -703,9 +723,9 @@ void dbg_dump_leb(const struct ubifs_info *c, int lnum)
 	if (dbg_failure_mode)
 		return;
 
-	printk(KERN_DEBUG "(pid %d) Dumping LEB %d\n", current->pid, lnum);
-
-	sleb = ubifs_scan(c, lnum, 0, c->dbg_buf);
+	printk(KERN_DEBUG "(pid %d) start dumping LEB %d\n",
+	       current->pid, lnum);
+	sleb = ubifs_scan(c, lnum, 0, c->dbg->buf);
 	if (IS_ERR(sleb)) {
 		ubifs_err("scan error %d", (int)PTR_ERR(sleb));
 		return;
@@ -721,6 +741,8 @@ void dbg_dump_leb(const struct ubifs_info *c, int lnum)
 		dbg_dump_node(c, snod->node);
 	}
 
+	printk(KERN_DEBUG "(pid %d) finish dumping LEB %d\n",
+	       current->pid, lnum);
 	ubifs_scan_destroy(sleb);
 	return;
 }
@@ -768,7 +790,7 @@ void dbg_dump_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat)
 {
 	int i;
 
-	printk(KERN_DEBUG "(pid %d) Dumping heap cat %d (%d elements)\n",
+	printk(KERN_DEBUG "(pid %d) start dumping heap cat %d (%d elements)\n",
 	       current->pid, cat, heap->cnt);
 	for (i = 0; i < heap->cnt; i++) {
 		struct ubifs_lprops *lprops = heap->arr[i];
@@ -777,6 +799,7 @@ void dbg_dump_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat)
 		       "flags %d\n", i, lprops->lnum, lprops->hpos,
 		       lprops->free, lprops->dirty, lprops->flags);
 	}
+	printk(KERN_DEBUG "(pid %d) finish dumping heap\n", current->pid);
 }
 
 void dbg_dump_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
@@ -784,7 +807,7 @@ void dbg_dump_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
 {
 	int i;
 
-	printk(KERN_DEBUG "(pid %d) Dumping pnode:\n", current->pid);
+	printk(KERN_DEBUG "(pid %d) dumping pnode:\n", current->pid);
 	printk(KERN_DEBUG "\taddress %zx parent %zx cnext %zx\n",
 	       (size_t)pnode, (size_t)parent, (size_t)pnode->cnext);
 	printk(KERN_DEBUG "\tflags %lu iip %d level %d num %d\n",
@@ -803,7 +826,7 @@ void dbg_dump_tnc(struct ubifs_info *c)
 	int level;
 
 	printk(KERN_DEBUG "\n");
-	printk(KERN_DEBUG "(pid %d) Dumping the TNC tree\n", current->pid);
+	printk(KERN_DEBUG "(pid %d) start dumping TNC tree\n", current->pid);
 	znode = ubifs_tnc_levelorder_next(c->zroot.znode, NULL);
 	level = znode->level;
 	printk(KERN_DEBUG "== Level %d ==\n", level);
@@ -815,8 +838,7 @@ void dbg_dump_tnc(struct ubifs_info *c)
 		dbg_dump_znode(c, znode);
 		znode = ubifs_tnc_levelorder_next(c->zroot.znode, znode);
 	}
-
-	printk(KERN_DEBUG "\n");
+	printk(KERN_DEBUG "(pid %d) finish dumping TNC tree\n", current->pid);
 }
 
 static int dump_znode(struct ubifs_info *c, struct ubifs_znode *znode,
@@ -992,8 +1014,8 @@ static int dbg_check_key_order(struct ubifs_info *c, struct ubifs_zbranch *zbr1,
 			zbr1->offs, DBGKEY(&key));
 		dbg_err("but it should have key %s according to tnc",
 			DBGKEY(&zbr1->key));
-			dbg_dump_node(c, dent1);
-			goto out_free;
+		dbg_dump_node(c, dent1);
+		goto out_free;
 	}
 
 	key_read(c, &dent2->key, &key);
@@ -1002,8 +1024,8 @@ static int dbg_check_key_order(struct ubifs_info *c, struct ubifs_zbranch *zbr1,
 			zbr1->offs, DBGKEY(&key));
 		dbg_err("but it should have key %s according to tnc",
 			DBGKEY(&zbr2->key));
-			dbg_dump_node(c, dent2);
-			goto out_free;
+		dbg_dump_node(c, dent2);
+		goto out_free;
 	}
 
 	nlen1 = le16_to_cpu(dent1->nlen);
@@ -1020,9 +1042,9 @@ static int dbg_check_key_order(struct ubifs_info *c, struct ubifs_zbranch *zbr1,
 		dbg_err("bad order of colliding key %s",
 			DBGKEY(&key));
 
-	dbg_msg("first node at %d:%d\n", zbr1->lnum, zbr1->offs);
+	ubifs_msg("first node at %d:%d\n", zbr1->lnum, zbr1->offs);
 	dbg_dump_node(c, dent1);
-	dbg_msg("second node at %d:%d\n", zbr2->lnum, zbr2->offs);
+	ubifs_msg("second node at %d:%d\n", zbr2->lnum, zbr2->offs);
 	dbg_dump_node(c, dent2);
 
 out_free:
@@ -2097,13 +2119,13 @@ static int simple_rand(void)
 	return (next >> 16) & 32767;
 }
 
-void dbg_failure_mode_registration(struct ubifs_info *c)
+static void failure_mode_init(struct ubifs_info *c)
 {
 	struct failure_mode_info *fmi;
 
 	fmi = kmalloc(sizeof(struct failure_mode_info), GFP_NOFS);
 	if (!fmi) {
-		dbg_err("Failed to register failure mode - no memory");
+		ubifs_err("Failed to register failure mode - no memory");
 		return;
 	}
 	fmi->c = c;
@@ -2112,7 +2134,7 @@ void dbg_failure_mode_registration(struct ubifs_info *c)
 	spin_unlock(&fmi_lock);
 }
 
-void dbg_failure_mode_deregistration(struct ubifs_info *c)
+static void failure_mode_exit(struct ubifs_info *c)
 {
 	struct failure_mode_info *fmi, *tmp;
 
@@ -2146,42 +2168,44 @@ static int in_failure_mode(struct ubi_volume_desc *desc)
 	struct ubifs_info *c = dbg_find_info(desc);
 
 	if (c && dbg_failure_mode)
-		return c->failure_mode;
+		return c->dbg->failure_mode;
 	return 0;
 }
 
 static int do_fail(struct ubi_volume_desc *desc, int lnum, int write)
 {
 	struct ubifs_info *c = dbg_find_info(desc);
+	struct ubifs_debug_info *d;
 
 	if (!c || !dbg_failure_mode)
 		return 0;
-	if (c->failure_mode)
+	d = c->dbg;
+	if (d->failure_mode)
 		return 1;
-	if (!c->fail_cnt) {
+	if (!d->fail_cnt) {
 		/* First call - decide delay to failure */
 		if (chance(1, 2)) {
 			unsigned int delay = 1 << (simple_rand() >> 11);
 
 			if (chance(1, 2)) {
-				c->fail_delay = 1;
-				c->fail_timeout = jiffies +
+				d->fail_delay = 1;
+				d->fail_timeout = jiffies +
 						  msecs_to_jiffies(delay);
 				dbg_rcvry("failing after %ums", delay);
 			} else {
-				c->fail_delay = 2;
-				c->fail_cnt_max = delay;
+				d->fail_delay = 2;
+				d->fail_cnt_max = delay;
 				dbg_rcvry("failing after %u calls", delay);
 			}
 		}
-		c->fail_cnt += 1;
+		d->fail_cnt += 1;
 	}
 	/* Determine if failure delay has expired */
-	if (c->fail_delay == 1) {
-		if (time_before(jiffies, c->fail_timeout))
+	if (d->fail_delay == 1) {
+		if (time_before(jiffies, d->fail_timeout))
 			return 0;
-	} else if (c->fail_delay == 2)
-		if (c->fail_cnt++ < c->fail_cnt_max)
+	} else if (d->fail_delay == 2)
+		if (d->fail_cnt++ < d->fail_cnt_max)
 			return 0;
 	if (lnum == UBIFS_SB_LNUM) {
 		if (write) {
@@ -2239,7 +2263,7 @@ static int do_fail(struct ubi_volume_desc *desc, int lnum, int write)
 		dbg_rcvry("failing in bud LEB %d commit not running", lnum);
 	}
 	ubifs_err("*** SETTING FAILURE MODE ON (LEB %d) ***", lnum);
-	c->failure_mode = 1;
+	d->failure_mode = 1;
 	dump_stack();
 	return 1;
 }
@@ -2344,4 +2368,181 @@ int dbg_leb_map(struct ubi_volume_desc *desc, int lnum, int dtype)
 	return 0;
 }
 
+/**
+ * ubifs_debugging_init - initialize UBIFS debugging.
+ * @c: UBIFS file-system description object
+ *
+ * This function initializes debugging-related data for the file system.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_debugging_init(struct ubifs_info *c)
+{
+	c->dbg = kzalloc(sizeof(struct ubifs_debug_info), GFP_KERNEL);
+	if (!c->dbg)
+		return -ENOMEM;
+
+	c->dbg->buf = vmalloc(c->leb_size);
+	if (!c->dbg->buf)
+		goto out;
+
+	failure_mode_init(c);
+	return 0;
+
+out:
+	kfree(c->dbg);
+	return -ENOMEM;
+}
+
+/**
+ * ubifs_debugging_exit - free debugging data.
+ * @c: UBIFS file-system description object
+ */
+void ubifs_debugging_exit(struct ubifs_info *c)
+{
+	failure_mode_exit(c);
+	vfree(c->dbg->buf);
+	kfree(c->dbg);
+}
+
+/*
+ * Root directory for UBIFS stuff in debugfs. Contains sub-directories which
+ * contain the stuff specific to particular file-system mounts.
+ */
+static struct dentry *debugfs_rootdir;
+
+/**
+ * dbg_debugfs_init - initialize debugfs file-system.
+ *
+ * UBIFS uses debugfs file-system to expose various debugging knobs to
+ * user-space. This function creates "ubifs" directory in the debugfs
+ * file-system. Returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+int dbg_debugfs_init(void)
+{
+	debugfs_rootdir = debugfs_create_dir("ubifs", NULL);
+	if (IS_ERR(debugfs_rootdir)) {
+		int err = PTR_ERR(debugfs_rootdir);
+		ubifs_err("cannot create \"ubifs\" debugfs directory, "
+			  "error %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+/**
+ * dbg_debugfs_exit - remove the "ubifs" directory from debugfs file-system.
+ */
+void dbg_debugfs_exit(void)
+{
+	debugfs_remove(debugfs_rootdir);
+}
+
+static int open_debugfs_file(struct inode *inode, struct file *file)
+{
+	file->private_data = inode->i_private;
+	return 0;
+}
+
+static ssize_t write_debugfs_file(struct file *file, const char __user *buf,
+				  size_t count, loff_t *ppos)
+{
+	struct ubifs_info *c = file->private_data;
+	struct ubifs_debug_info *d = c->dbg;
+
+	if (file->f_path.dentry == d->dump_lprops)
+		dbg_dump_lprops(c);
+	else if (file->f_path.dentry == d->dump_budg) {
+		spin_lock(&c->space_lock);
+		dbg_dump_budg(c);
+		spin_unlock(&c->space_lock);
+	} else if (file->f_path.dentry == d->dump_tnc) {
+		mutex_lock(&c->tnc_mutex);
+		dbg_dump_tnc(c);
+		mutex_unlock(&c->tnc_mutex);
+	} else
+		return -EINVAL;
+
+	*ppos += count;
+	return count;
+}
+
+static const struct file_operations debugfs_fops = {
+	.open = open_debugfs_file,
+	.write = write_debugfs_file,
+	.owner = THIS_MODULE,
+};
+
+/**
+ * dbg_debugfs_init_fs - initialize debugfs for UBIFS instance.
+ * @c: UBIFS file-system description object
+ *
+ * This function creates all debugfs files for this instance of UBIFS. Returns
+ * zero in case of success and a negative error code in case of failure.
+ *
+ * Note, the only reason we have not merged this function with the
+ * 'ubifs_debugging_init()' function is because it is better to initialize
+ * debugfs interfaces at the very end of the mount process, and remove them at
+ * the very beginning of the mount process.
+ */
+int dbg_debugfs_init_fs(struct ubifs_info *c)
+{
+	int err;
+	const char *fname;
+	struct dentry *dent;
+	struct ubifs_debug_info *d = c->dbg;
+
+	sprintf(d->debugfs_dir_name, "ubi%d_%d", c->vi.ubi_num, c->vi.vol_id);
+	d->debugfs_dir = debugfs_create_dir(d->debugfs_dir_name,
+					      debugfs_rootdir);
+	if (IS_ERR(d->debugfs_dir)) {
+		err = PTR_ERR(d->debugfs_dir);
+		ubifs_err("cannot create \"%s\" debugfs directory, error %d\n",
+			  d->debugfs_dir_name, err);
+		goto out;
+	}
+
+	fname = "dump_lprops";
+	dent = debugfs_create_file(fname, S_IWUGO, d->debugfs_dir, c,
+				   &debugfs_fops);
+	if (IS_ERR(dent))
+		goto out_remove;
+	d->dump_lprops = dent;
+
+	fname = "dump_budg";
+	dent = debugfs_create_file(fname, S_IWUGO, d->debugfs_dir, c,
+				   &debugfs_fops);
+	if (IS_ERR(dent))
+		goto out_remove;
+	d->dump_budg = dent;
+
+	fname = "dump_tnc";
+	dent = debugfs_create_file(fname, S_IWUGO, d->debugfs_dir, c,
+				   &debugfs_fops);
+	if (IS_ERR(dent))
+		goto out_remove;
+	d->dump_tnc = dent;
+
+	return 0;
+
+out_remove:
+	err = PTR_ERR(dent);
+	ubifs_err("cannot create \"%s\" debugfs directory, error %d\n",
+		  fname, err);
+	debugfs_remove_recursive(d->debugfs_dir);
+out:
+	return err;
+}
+
+/**
+ * dbg_debugfs_exit_fs - remove all debugfs files.
+ * @c: UBIFS file-system description object
+ */
+void dbg_debugfs_exit_fs(struct ubifs_info *c)
+{
+	debugfs_remove_recursive(c->dbg->debugfs_dir);
+}
+
 #endif /* CONFIG_UBIFS_FS_DEBUG */
diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h
index 33d6b95071e..9820d6999f7 100644
--- a/fs/ubifs/debug.h
+++ b/fs/ubifs/debug.h
@@ -25,7 +25,56 @@
 
 #ifdef CONFIG_UBIFS_FS_DEBUG
 
-#define UBIFS_DBG(op) op
+/**
+ * ubifs_debug_info - per-FS debugging information.
+ * @buf: a buffer of LEB size, used for various purposes
+ * @old_zroot: old index root - used by 'dbg_check_old_index()'
+ * @old_zroot_level: old index root level - used by 'dbg_check_old_index()'
+ * @old_zroot_sqnum: old index root sqnum - used by 'dbg_check_old_index()'
+ * @failure_mode: failure mode for recovery testing
+ * @fail_delay: 0=>don't delay, 1=>delay a time, 2=>delay a number of calls
+ * @fail_timeout: time in jiffies when delay of failure mode expires
+ * @fail_cnt: current number of calls to failure mode I/O functions
+ * @fail_cnt_max: number of calls by which to delay failure mode
+ * @chk_lpt_sz: used by LPT tree size checker
+ * @chk_lpt_sz2: used by LPT tree size checker
+ * @chk_lpt_wastage: used by LPT tree size checker
+ * @chk_lpt_lebs: used by LPT tree size checker
+ * @new_nhead_offs: used by LPT tree size checker
+ * @new_ihead_lnum: used by debugging to check ihead_lnum
+ * @new_ihead_offs: used by debugging to check ihead_offs
+ *
+ * debugfs_dir_name: name of debugfs directory containing this file-system's
+ *                   files
+ * debugfs_dir: direntry object of the file-system debugfs directory
+ * dump_lprops: "dump lprops" debugfs knob
+ * dump_budg: "dump budgeting information" debugfs knob
+ * dump_tnc: "dump TNC" debugfs knob
+ */
+struct ubifs_debug_info {
+	void *buf;
+	struct ubifs_zbranch old_zroot;
+	int old_zroot_level;
+	unsigned long long old_zroot_sqnum;
+	int failure_mode;
+	int fail_delay;
+	unsigned long fail_timeout;
+	unsigned int fail_cnt;
+	unsigned int fail_cnt_max;
+	long long chk_lpt_sz;
+	long long chk_lpt_sz2;
+	long long chk_lpt_wastage;
+	int chk_lpt_lebs;
+	int new_nhead_offs;
+	int new_ihead_lnum;
+	int new_ihead_offs;
+
+	char debugfs_dir_name[100];
+	struct dentry *debugfs_dir;
+	struct dentry *dump_lprops;
+	struct dentry *dump_budg;
+	struct dentry *dump_tnc;
+};
 
 #define ubifs_assert(expr) do {                                                \
 	if (unlikely(!(expr))) {                                               \
@@ -211,14 +260,18 @@ extern unsigned int ubifs_msg_flags;
 extern unsigned int ubifs_chk_flags;
 extern unsigned int ubifs_tst_flags;
 
-/* Dump functions */
+int ubifs_debugging_init(struct ubifs_info *c);
+void ubifs_debugging_exit(struct ubifs_info *c);
 
+/* Dump functions */
 const char *dbg_ntype(int type);
 const char *dbg_cstate(int cmt_state);
 const char *dbg_get_key_dump(const struct ubifs_info *c,
 			     const union ubifs_key *key);
 void dbg_dump_inode(const struct ubifs_info *c, const struct inode *inode);
 void dbg_dump_node(const struct ubifs_info *c, const void *node);
+void dbg_dump_lpt_node(const struct ubifs_info *c, void *node, int lnum,
+		       int offs);
 void dbg_dump_budget_req(const struct ubifs_budget_req *req);
 void dbg_dump_lstats(const struct ubifs_lp_stats *lst);
 void dbg_dump_budg(struct ubifs_info *c);
@@ -233,9 +286,9 @@ void dbg_dump_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
 		    struct ubifs_nnode *parent, int iip);
 void dbg_dump_tnc(struct ubifs_info *c);
 void dbg_dump_index(struct ubifs_info *c);
+void dbg_dump_lpt_lebs(const struct ubifs_info *c);
 
 /* Checking helper functions */
-
 typedef int (*dbg_leaf_callback)(struct ubifs_info *c,
 				 struct ubifs_zbranch *zbr, void *priv);
 typedef int (*dbg_znode_callback)(struct ubifs_info *c,
@@ -274,9 +327,6 @@ int dbg_force_in_the_gaps(void);
 
 #define dbg_failure_mode (ubifs_tst_flags & UBIFS_TST_RCVRY)
 
-void dbg_failure_mode_registration(struct ubifs_info *c);
-void dbg_failure_mode_deregistration(struct ubifs_info *c);
-
 #ifndef UBIFS_DBG_PRESERVE_UBI
 
 #define ubi_leb_read   dbg_leb_read
@@ -318,9 +368,13 @@ static inline int dbg_change(struct ubi_volume_desc *desc, int lnum,
 	return dbg_leb_change(desc, lnum, buf, len, UBI_UNKNOWN);
 }
 
-#else /* !CONFIG_UBIFS_FS_DEBUG */
+/* Debugfs-related stuff */
+int dbg_debugfs_init(void);
+void dbg_debugfs_exit(void);
+int dbg_debugfs_init_fs(struct ubifs_info *c);
+void dbg_debugfs_exit_fs(struct ubifs_info *c);
 
-#define UBIFS_DBG(op)
+#else /* !CONFIG_UBIFS_FS_DEBUG */
 
 /* Use "if (0)" to make compiler check arguments even if debugging is off */
 #define ubifs_assert(expr)  do {                                               \
@@ -360,23 +414,28 @@ static inline int dbg_change(struct ubi_volume_desc *desc, int lnum,
 #define DBGKEY(key)  ((char *)(key))
 #define DBGKEY1(key) ((char *)(key))
 
-#define dbg_ntype(type)                       ""
-#define dbg_cstate(cmt_state)                 ""
-#define dbg_get_key_dump(c, key)              ({})
-#define dbg_dump_inode(c, inode)              ({})
-#define dbg_dump_node(c, node)                ({})
-#define dbg_dump_budget_req(req)              ({})
-#define dbg_dump_lstats(lst)                  ({})
-#define dbg_dump_budg(c)                      ({})
-#define dbg_dump_lprop(c, lp)                 ({})
-#define dbg_dump_lprops(c)                    ({})
-#define dbg_dump_lpt_info(c)                  ({})
-#define dbg_dump_leb(c, lnum)                 ({})
-#define dbg_dump_znode(c, znode)              ({})
-#define dbg_dump_heap(c, heap, cat)           ({})
-#define dbg_dump_pnode(c, pnode, parent, iip) ({})
-#define dbg_dump_tnc(c)                       ({})
-#define dbg_dump_index(c)                     ({})
+#define ubifs_debugging_init(c)                0
+#define ubifs_debugging_exit(c)                ({})
+
+#define dbg_ntype(type)                        ""
+#define dbg_cstate(cmt_state)                  ""
+#define dbg_get_key_dump(c, key)               ({})
+#define dbg_dump_inode(c, inode)               ({})
+#define dbg_dump_node(c, node)                 ({})
+#define dbg_dump_lpt_node(c, node, lnum, offs) ({})
+#define dbg_dump_budget_req(req)               ({})
+#define dbg_dump_lstats(lst)                   ({})
+#define dbg_dump_budg(c)                       ({})
+#define dbg_dump_lprop(c, lp)                  ({})
+#define dbg_dump_lprops(c)                     ({})
+#define dbg_dump_lpt_info(c)                   ({})
+#define dbg_dump_leb(c, lnum)                  ({})
+#define dbg_dump_znode(c, znode)               ({})
+#define dbg_dump_heap(c, heap, cat)            ({})
+#define dbg_dump_pnode(c, pnode, parent, iip)  ({})
+#define dbg_dump_tnc(c)                        ({})
+#define dbg_dump_index(c)                      ({})
+#define dbg_dump_lpt_lebs(c)                   ({})
 
 #define dbg_walk_index(c, leaf_cb, znode_cb, priv) 0
 #define dbg_old_index_check_init(c, zroot)         0
@@ -396,9 +455,11 @@ static inline int dbg_change(struct ubi_volume_desc *desc, int lnum,
 #define dbg_force_in_the_gaps_enabled              0
 #define dbg_force_in_the_gaps()                    0
 #define dbg_failure_mode                           0
-#define dbg_failure_mode_registration(c)           ({})
-#define dbg_failure_mode_deregistration(c)         ({})
 
-#endif /* !CONFIG_UBIFS_FS_DEBUG */
+#define dbg_debugfs_init()                         0
+#define dbg_debugfs_exit()
+#define dbg_debugfs_init_fs(c)                     0
+#define dbg_debugfs_exit_fs(c)                     0
 
+#endif /* !CONFIG_UBIFS_FS_DEBUG */
 #endif /* !__UBIFS_DEBUG_H__ */
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index 2624411d975..bf37374567f 100644
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -72,8 +72,8 @@ static int read_block(struct inode *inode, void *addr, unsigned int block,
 		return err;
 	}
 
-	ubifs_assert(le64_to_cpu(dn->ch.sqnum) > ubifs_inode(inode)->creat_sqnum);
-
+	ubifs_assert(le64_to_cpu(dn->ch.sqnum) >
+		     ubifs_inode(inode)->creat_sqnum);
 	len = le32_to_cpu(dn->size);
 	if (len <= 0 || len > UBIFS_BLOCK_SIZE)
 		goto dump;
@@ -219,7 +219,8 @@ static void release_existing_page_budget(struct ubifs_info *c)
 }
 
 static int write_begin_slow(struct address_space *mapping,
-			    loff_t pos, unsigned len, struct page **pagep)
+			    loff_t pos, unsigned len, struct page **pagep,
+			    unsigned flags)
 {
 	struct inode *inode = mapping->host;
 	struct ubifs_info *c = inode->i_sb->s_fs_info;
@@ -247,14 +248,14 @@ static int write_begin_slow(struct address_space *mapping,
 	if (unlikely(err))
 		return err;
 
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (unlikely(!page)) {
 		ubifs_release_budget(c, &req);
 		return -ENOMEM;
 	}
 
 	if (!PageUptodate(page)) {
-		if (!(pos & PAGE_CACHE_MASK) && len == PAGE_CACHE_SIZE)
+		if (!(pos & ~PAGE_CACHE_MASK) && len == PAGE_CACHE_SIZE)
 			SetPageChecked(page);
 		else {
 			err = do_readpage(page);
@@ -438,13 +439,13 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 		return -EROFS;
 
 	/* Try out the fast-path part first */
-	page = __grab_cache_page(mapping, index);
+	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (unlikely(!page))
 		return -ENOMEM;
 
 	if (!PageUptodate(page)) {
 		/* The page is not loaded from the flash */
-		if (!(pos & PAGE_CACHE_MASK) && len == PAGE_CACHE_SIZE)
+		if (!(pos & ~PAGE_CACHE_MASK) && len == PAGE_CACHE_SIZE)
 			/*
 			 * We change whole page so no need to load it. But we
 			 * have to set the @PG_checked flag to make the further
@@ -483,7 +484,7 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 		unlock_page(page);
 		page_cache_release(page);
 
-		return write_begin_slow(mapping, pos, len, pagep);
+		return write_begin_slow(mapping, pos, len, pagep, flags);
 	}
 
 	/*
diff --git a/fs/ubifs/gc.c b/fs/ubifs/gc.c
index 0bef6501d58..9832f9abe28 100644
--- a/fs/ubifs/gc.c
+++ b/fs/ubifs/gc.c
@@ -45,7 +45,7 @@
 #define SMALL_NODE_WM  UBIFS_MAX_DENT_NODE_SZ
 
 /*
- * GC may need to move more then one LEB to make progress. The below constants
+ * GC may need to move more than one LEB to make progress. The below constants
  * define "soft" and "hard" limits on the number of LEBs the garbage collector
  * may move.
  */
diff --git a/fs/ubifs/ioctl.c b/fs/ubifs/ioctl.c
index 5e82cffe969..6db7a6be6c9 100644
--- a/fs/ubifs/ioctl.c
+++ b/fs/ubifs/ioctl.c
@@ -154,6 +154,7 @@ long ubifs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 	case FS_IOC_GETFLAGS:
 		flags = ubifs2ioctl(ubifs_inode(inode)->flags);
 
+		dbg_gen("get flags: %#x, i_flags %#x", flags, inode->i_flags);
 		return put_user(flags, (int __user *) arg);
 
 	case FS_IOC_SETFLAGS: {
@@ -176,6 +177,7 @@ long ubifs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 		err = mnt_want_write(file->f_path.mnt);
 		if (err)
 			return err;
+		dbg_gen("set flags: %#x, i_flags %#x", flags, inode->i_flags);
 		err = setflags(inode, flags);
 		mnt_drop_write(file->f_path.mnt);
 		return err;
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index f91b745908e..9b7c54e0cd2 100644
--- a/fs/ubifs/journal.c
+++ b/fs/ubifs/journal.c
@@ -191,7 +191,7 @@ again:
 	if (wbuf->lnum != -1 && avail >= len) {
 		/*
 		 * Someone else has switched the journal head and we have
-		 * enough space now. This happens when more then one process is
+		 * enough space now. This happens when more than one process is
 		 * trying to write to the same journal head at the same time.
 		 */
 		dbg_jnl("return LEB %d back, already have LEB %d:%d",
@@ -704,7 +704,7 @@ int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
 	data->size = cpu_to_le32(len);
 	zero_data_node_unused(data);
 
-	if (!(ui->flags && UBIFS_COMPR_FL))
+	if (!(ui->flags & UBIFS_COMPR_FL))
 		/* Compression is disabled for this inode */
 		compr_type = UBIFS_COMPR_NONE;
 	else
@@ -1220,7 +1220,7 @@ int ubifs_jnl_truncate(struct ubifs_info *c, const struct inode *inode,
 	data_key_init(c, &key, inum, blk);
 
 	bit = old_size & (UBIFS_BLOCK_SIZE - 1);
-	blk = (old_size >> UBIFS_BLOCK_SHIFT) - (bit ? 0: 1);
+	blk = (old_size >> UBIFS_BLOCK_SHIFT) - (bit ? 0 : 1);
 	data_key_init(c, &to_key, inum, blk);
 
 	err = ubifs_tnc_remove_range(c, &key, &to_key);
diff --git a/fs/ubifs/key.h b/fs/ubifs/key.h
index 3f1f16bc25c..efb3430a258 100644
--- a/fs/ubifs/key.h
+++ b/fs/ubifs/key.h
@@ -38,6 +38,22 @@
 #define __UBIFS_KEY_H__
 
 /**
+ * key_mask_hash - mask a valid hash value.
+ * @val: value to be masked
+ *
+ * We use hash values as offset in directories, so values %0 and %1 are
+ * reserved for "." and "..". %2 is reserved for "end of readdir" marker. This
+ * function makes sure the reserved values are not used.
+ */
+static inline uint32_t key_mask_hash(uint32_t hash)
+{
+	hash &= UBIFS_S_KEY_HASH_MASK;
+	if (unlikely(hash <= 2))
+		hash += 3;
+	return hash;
+}
+
+/**
  * key_r5_hash - R5 hash function (borrowed from reiserfs).
  * @s: direntry name
  * @len: name length
@@ -54,16 +70,7 @@ static inline uint32_t key_r5_hash(const char *s, int len)
 		str++;
 	}
 
-	a &= UBIFS_S_KEY_HASH_MASK;
-
-	/*
-	 * We use hash values as offset in directories, so values %0 and %1 are
-	 * reserved for "." and "..". %2 is reserved for "end of readdir"
-	 * marker.
-	 */
-	if (unlikely(a >= 0 && a <= 2))
-		a += 3;
-	return a;
+	return key_mask_hash(a);
 }
 
 /**
@@ -77,10 +84,7 @@ static inline uint32_t key_test_hash(const char *str, int len)
 
 	len = min_t(uint32_t, len, 4);
 	memcpy(&a, str, len);
-	a &= UBIFS_S_KEY_HASH_MASK;
-	if (unlikely(a >= 0 && a <= 2))
-		a += 3;
-	return a;
+	return key_mask_hash(a);
 }
 
 /**
diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c
index f27176e9b70..dfd2bcece27 100644
--- a/fs/ubifs/lprops.c
+++ b/fs/ubifs/lprops.c
@@ -520,13 +520,13 @@ static int is_lprops_dirty(struct ubifs_info *c, struct ubifs_lprops *lprops)
  * @flags: new flags
  * @idx_gc_cnt: change to the count of idx_gc list
  *
- * This function changes LEB properties. This function does not change a LEB
- * property (@free, @dirty or @flag) if the value passed is %LPROPS_NC.
+ * This function changes LEB properties (@free, @dirty or @flag). However, the
+ * property which has the %LPROPS_NC value is not changed. Returns a pointer to
+ * the updated LEB properties on success and a negative error code on failure.
  *
- * This function returns a pointer to the updated LEB properties on success
- * and a negative error code on failure. N.B. the LEB properties may have had to
- * be copied (due to COW) and consequently the pointer returned may not be the
- * same as the pointer passed.
+ * Note, the LEB properties may have had to be copied (due to COW) and
+ * consequently the pointer returned may not be the same as the pointer
+ * passed.
  */
 const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
 					   const struct ubifs_lprops *lp,
@@ -1088,7 +1088,7 @@ static int scan_check_cb(struct ubifs_info *c,
 		}
 	}
 
-	sleb = ubifs_scan(c, lnum, 0, c->dbg_buf);
+	sleb = ubifs_scan(c, lnum, 0, c->dbg->buf);
 	if (IS_ERR(sleb)) {
 		/*
 		 * After an unclean unmount, empty and freeable LEBs
diff --git a/fs/ubifs/lpt.c b/fs/ubifs/lpt.c
index db8bd0e518b..b2792e84d24 100644
--- a/fs/ubifs/lpt.c
+++ b/fs/ubifs/lpt.c
@@ -36,15 +36,16 @@
  * can be written into a single eraseblock. In that case, garbage collection
  * consists of just writing the whole table, which therefore makes all other
  * eraseblocks reusable. In the case of the big model, dirty eraseblocks are
- * selected for garbage collection, which consists are marking the nodes in
+ * selected for garbage collection, which consists of marking the clean nodes in
  * that LEB as dirty, and then only the dirty nodes are written out. Also, in
  * the case of the big model, a table of LEB numbers is saved so that the entire
  * LPT does not to be scanned looking for empty eraseblocks when UBIFS is first
  * mounted.
  */
 
-#include <linux/crc16.h>
 #include "ubifs.h"
+#include <linux/crc16.h>
+#include <linux/math64.h>
 
 /**
  * do_calc_lpt_geom - calculate sizes for the LPT area.
@@ -135,15 +136,13 @@ static void do_calc_lpt_geom(struct ubifs_info *c)
 int ubifs_calc_lpt_geom(struct ubifs_info *c)
 {
 	int lebs_needed;
-	uint64_t sz;
+	long long sz;
 
 	do_calc_lpt_geom(c);
 
 	/* Verify that lpt_lebs is big enough */
 	sz = c->lpt_sz * 2; /* Must have at least 2 times the size */
-	sz += c->leb_size - 1;
-	do_div(sz, c->leb_size);
-	lebs_needed = sz;
+	lebs_needed = div_u64(sz + c->leb_size - 1, c->leb_size);
 	if (lebs_needed > c->lpt_lebs) {
 		ubifs_err("too few LPT LEBs");
 		return -EINVAL;
@@ -156,7 +155,6 @@ int ubifs_calc_lpt_geom(struct ubifs_info *c)
 	}
 
 	c->check_lpt_free = c->big_lpt;
-
 	return 0;
 }
 
@@ -176,7 +174,7 @@ static int calc_dflt_lpt_geom(struct ubifs_info *c, int *main_lebs,
 			      int *big_lpt)
 {
 	int i, lebs_needed;
-	uint64_t sz;
+	long long sz;
 
 	/* Start by assuming the minimum number of LPT LEBs */
 	c->lpt_lebs = UBIFS_MIN_LPT_LEBS;
@@ -203,9 +201,7 @@ static int calc_dflt_lpt_geom(struct ubifs_info *c, int *main_lebs,
 	/* Now check there are enough LPT LEBs */
 	for (i = 0; i < 64 ; i++) {
 		sz = c->lpt_sz * 4; /* Allow 4 times the size */
-		sz += c->leb_size - 1;
-		do_div(sz, c->leb_size);
-		lebs_needed = sz;
+		lebs_needed = div_u64(sz + c->leb_size - 1, c->leb_size);
 		if (lebs_needed > c->lpt_lebs) {
 			/* Not enough LPT LEBs so try again with more */
 			c->lpt_lebs = lebs_needed;
@@ -558,7 +554,7 @@ static int calc_nnode_num(int row, int col)
  * This function calculates and returns the nnode number based on the parent's
  * nnode number and the index in parent.
  */
-static int calc_nnode_num_from_parent(struct ubifs_info *c,
+static int calc_nnode_num_from_parent(const struct ubifs_info *c,
 				      struct ubifs_nnode *parent, int iip)
 {
 	int num, shft;
@@ -583,7 +579,7 @@ static int calc_nnode_num_from_parent(struct ubifs_info *c,
  * This function calculates and returns the pnode number based on the parent's
  * nnode number and the index in parent.
  */
-static int calc_pnode_num_from_parent(struct ubifs_info *c,
+static int calc_pnode_num_from_parent(const struct ubifs_info *c,
 				      struct ubifs_nnode *parent, int iip)
 {
 	int i, n = c->lpt_hght - 1, pnum = parent->num, num = 0;
@@ -966,7 +962,7 @@ static int check_lpt_type(uint8_t **addr, int *pos, int type)
  *
  * This function returns %0 on success and a negative error code on failure.
  */
-static int unpack_pnode(struct ubifs_info *c, void *buf,
+static int unpack_pnode(const struct ubifs_info *c, void *buf,
 			struct ubifs_pnode *pnode)
 {
 	uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
@@ -996,15 +992,15 @@ static int unpack_pnode(struct ubifs_info *c, void *buf,
 }
 
 /**
- * unpack_nnode - unpack a nnode.
+ * ubifs_unpack_nnode - unpack a nnode.
  * @c: UBIFS file-system description object
  * @buf: buffer containing packed nnode to unpack
  * @nnode: nnode structure to fill
  *
  * This function returns %0 on success and a negative error code on failure.
  */
-static int unpack_nnode(struct ubifs_info *c, void *buf,
-			struct ubifs_nnode *nnode)
+int ubifs_unpack_nnode(const struct ubifs_info *c, void *buf,
+		       struct ubifs_nnode *nnode)
 {
 	uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
 	int i, pos = 0, err;
@@ -1036,7 +1032,7 @@ static int unpack_nnode(struct ubifs_info *c, void *buf,
  *
  * This function returns %0 on success and a negative error code on failure.
  */
-static int unpack_ltab(struct ubifs_info *c, void *buf)
+static int unpack_ltab(const struct ubifs_info *c, void *buf)
 {
 	uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
 	int i, pos = 0, err;
@@ -1068,7 +1064,7 @@ static int unpack_ltab(struct ubifs_info *c, void *buf)
  *
  * This function returns %0 on success and a negative error code on failure.
  */
-static int unpack_lsave(struct ubifs_info *c, void *buf)
+static int unpack_lsave(const struct ubifs_info *c, void *buf)
 {
 	uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
 	int i, pos = 0, err;
@@ -1096,7 +1092,7 @@ static int unpack_lsave(struct ubifs_info *c, void *buf)
  *
  * This function returns %0 on success and a negative error code on failure.
  */
-static int validate_nnode(struct ubifs_info *c, struct ubifs_nnode *nnode,
+static int validate_nnode(const struct ubifs_info *c, struct ubifs_nnode *nnode,
 			  struct ubifs_nnode *parent, int iip)
 {
 	int i, lvl, max_offs;
@@ -1140,7 +1136,7 @@ static int validate_nnode(struct ubifs_info *c, struct ubifs_nnode *nnode,
  *
  * This function returns %0 on success and a negative error code on failure.
  */
-static int validate_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
+static int validate_pnode(const struct ubifs_info *c, struct ubifs_pnode *pnode,
 			  struct ubifs_nnode *parent, int iip)
 {
 	int i;
@@ -1174,7 +1170,8 @@ static int validate_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
  * This function calculates the LEB numbers for the LEB properties it contains
  * based on the pnode number.
  */
-static void set_pnode_lnum(struct ubifs_info *c, struct ubifs_pnode *pnode)
+static void set_pnode_lnum(const struct ubifs_info *c,
+			   struct ubifs_pnode *pnode)
 {
 	int i, lnum;
 
@@ -1227,7 +1224,7 @@ int ubifs_read_nnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip)
 		err = ubi_read(c->ubi, lnum, buf, offs, c->nnode_sz);
 		if (err)
 			goto out;
-		err = unpack_nnode(c, buf, nnode);
+		err = ubifs_unpack_nnode(c, buf, nnode);
 		if (err)
 			goto out;
 	}
@@ -1816,7 +1813,7 @@ static struct ubifs_nnode *scan_get_nnode(struct ubifs_info *c,
 			       c->nnode_sz);
 		if (err)
 			return ERR_PTR(err);
-		err = unpack_nnode(c, buf, nnode);
+		err = ubifs_unpack_nnode(c, buf, nnode);
 		if (err)
 			return ERR_PTR(err);
 	}
diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c
index a41434b4278..96ca9570717 100644
--- a/fs/ubifs/lpt_commit.c
+++ b/fs/ubifs/lpt_commit.c
@@ -320,6 +320,8 @@ no_space:
 	dbg_err("LPT out of space at LEB %d:%d needing %d, done_ltab %d, "
 		"done_lsave %d", lnum, offs, len, done_ltab, done_lsave);
 	dbg_dump_lpt_info(c);
+	dbg_dump_lpt_lebs(c);
+	dump_stack();
 	return err;
 }
 
@@ -546,8 +548,10 @@ static int write_cnodes(struct ubifs_info *c)
 no_space:
 	ubifs_err("LPT out of space mismatch");
 	dbg_err("LPT out of space mismatch at LEB %d:%d needing %d, done_ltab "
-	        "%d, done_lsave %d", lnum, offs, len, done_ltab, done_lsave);
+		"%d, done_lsave %d", lnum, offs, len, done_ltab, done_lsave);
 	dbg_dump_lpt_info(c);
+	dbg_dump_lpt_lebs(c);
+	dump_stack();
 	return err;
 }
 
@@ -749,7 +753,7 @@ static void lpt_tgc_start(struct ubifs_info *c)
  * LPT trivial garbage collection is where a LPT LEB contains only dirty and
  * free space and so may be reused as soon as the next commit is completed.
  * This function is called after the commit is completed (master node has been
- * written) and unmaps LPT LEBs that were marked for trivial GC.
+ * written) and un-maps LPT LEBs that were marked for trivial GC.
  */
 static int lpt_tgc_end(struct ubifs_info *c)
 {
@@ -1025,7 +1029,7 @@ static int make_node_dirty(struct ubifs_info *c, int node_type, int node_num,
  * @c: UBIFS file-system description object
  * @node_type: LPT node type
  */
-static int get_lpt_node_len(struct ubifs_info *c, int node_type)
+static int get_lpt_node_len(const struct ubifs_info *c, int node_type)
 {
 	switch (node_type) {
 	case UBIFS_LPT_NNODE:
@@ -1046,7 +1050,7 @@ static int get_lpt_node_len(struct ubifs_info *c, int node_type)
  * @buf: buffer
  * @len: length of buffer
  */
-static int get_pad_len(struct ubifs_info *c, uint8_t *buf, int len)
+static int get_pad_len(const struct ubifs_info *c, uint8_t *buf, int len)
 {
 	int offs, pad_len;
 
@@ -1063,7 +1067,8 @@ static int get_pad_len(struct ubifs_info *c, uint8_t *buf, int len)
  * @buf: buffer
  * @node_num: node number is returned here
  */
-static int get_lpt_node_type(struct ubifs_info *c, uint8_t *buf, int *node_num)
+static int get_lpt_node_type(const struct ubifs_info *c, uint8_t *buf,
+			     int *node_num)
 {
 	uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
 	int pos = 0, node_type;
@@ -1081,7 +1086,7 @@ static int get_lpt_node_type(struct ubifs_info *c, uint8_t *buf, int *node_num)
  *
  * This function returns %1 if the buffer contains a node or %0 if it does not.
  */
-static int is_a_node(struct ubifs_info *c, uint8_t *buf, int len)
+static int is_a_node(const struct ubifs_info *c, uint8_t *buf, int len)
 {
 	uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
 	int pos = 0, node_type, node_len;
@@ -1105,7 +1110,6 @@ static int is_a_node(struct ubifs_info *c, uint8_t *buf, int len)
 	return 1;
 }
 
-
 /**
  * lpt_gc_lnum - garbage collect a LPT LEB.
  * @c: UBIFS file-system description object
@@ -1463,7 +1467,7 @@ void ubifs_lpt_free(struct ubifs_info *c, int wr_only)
 #ifdef CONFIG_UBIFS_FS_DEBUG
 
 /**
- * dbg_is_all_ff - determine if a buffer contains only 0xff bytes.
+ * dbg_is_all_ff - determine if a buffer contains only 0xFF bytes.
  * @buf: buffer
  * @len: buffer length
  */
@@ -1488,7 +1492,7 @@ static int dbg_is_nnode_dirty(struct ubifs_info *c, int lnum, int offs)
 	struct ubifs_nnode *nnode;
 	int hght;
 
-	/* Entire tree is in memory so first_nnode / next_nnode are ok */
+	/* Entire tree is in memory so first_nnode / next_nnode are OK */
 	nnode = first_nnode(c, &hght);
 	for (; nnode; nnode = next_nnode(c, nnode, &hght)) {
 		struct ubifs_nbranch *branch;
@@ -1602,7 +1606,10 @@ static int dbg_check_ltab_lnum(struct ubifs_info *c, int lnum)
 {
 	int err, len = c->leb_size, dirty = 0, node_type, node_num, node_len;
 	int ret;
-	void *buf = c->dbg_buf;
+	void *buf = c->dbg->buf;
+
+	if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS))
+		return 0;
 
 	dbg_lp("LEB %d", lnum);
 	err = ubi_read(c->ubi, lnum, buf, 0, c->leb_size);
@@ -1704,6 +1711,9 @@ int dbg_chk_lpt_free_spc(struct ubifs_info *c)
 	long long free = 0;
 	int i;
 
+	if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS))
+		return 0;
+
 	for (i = 0; i < c->lpt_lebs; i++) {
 		if (c->ltab[i].tgc || c->ltab[i].cmt)
 			continue;
@@ -1716,6 +1726,8 @@ int dbg_chk_lpt_free_spc(struct ubifs_info *c)
 		dbg_err("LPT space error: free %lld lpt_sz %lld",
 			free, c->lpt_sz);
 		dbg_dump_lpt_info(c);
+		dbg_dump_lpt_lebs(c);
+		dump_stack();
 		return -EINVAL;
 	}
 	return 0;
@@ -1731,15 +1743,19 @@ int dbg_chk_lpt_free_spc(struct ubifs_info *c)
  */
 int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len)
 {
+	struct ubifs_debug_info *d = c->dbg;
 	long long chk_lpt_sz, lpt_sz;
 	int err = 0;
 
+	if (!(ubifs_chk_flags & UBIFS_CHK_LPROPS))
+		return 0;
+
 	switch (action) {
 	case 0:
-		c->chk_lpt_sz = 0;
-		c->chk_lpt_sz2 = 0;
-		c->chk_lpt_lebs = 0;
-		c->chk_lpt_wastage = 0;
+		d->chk_lpt_sz = 0;
+		d->chk_lpt_sz2 = 0;
+		d->chk_lpt_lebs = 0;
+		d->chk_lpt_wastage = 0;
 		if (c->dirty_pn_cnt > c->pnode_cnt) {
 			dbg_err("dirty pnodes %d exceed max %d",
 				c->dirty_pn_cnt, c->pnode_cnt);
@@ -1752,35 +1768,35 @@ int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len)
 		}
 		return err;
 	case 1:
-		c->chk_lpt_sz += len;
+		d->chk_lpt_sz += len;
 		return 0;
 	case 2:
-		c->chk_lpt_sz += len;
-		c->chk_lpt_wastage += len;
-		c->chk_lpt_lebs += 1;
+		d->chk_lpt_sz += len;
+		d->chk_lpt_wastage += len;
+		d->chk_lpt_lebs += 1;
 		return 0;
 	case 3:
 		chk_lpt_sz = c->leb_size;
-		chk_lpt_sz *= c->chk_lpt_lebs;
+		chk_lpt_sz *= d->chk_lpt_lebs;
 		chk_lpt_sz += len - c->nhead_offs;
-		if (c->chk_lpt_sz != chk_lpt_sz) {
+		if (d->chk_lpt_sz != chk_lpt_sz) {
 			dbg_err("LPT wrote %lld but space used was %lld",
-				c->chk_lpt_sz, chk_lpt_sz);
+				d->chk_lpt_sz, chk_lpt_sz);
 			err = -EINVAL;
 		}
-		if (c->chk_lpt_sz > c->lpt_sz) {
+		if (d->chk_lpt_sz > c->lpt_sz) {
 			dbg_err("LPT wrote %lld but lpt_sz is %lld",
-				c->chk_lpt_sz, c->lpt_sz);
+				d->chk_lpt_sz, c->lpt_sz);
 			err = -EINVAL;
 		}
-		if (c->chk_lpt_sz2 && c->chk_lpt_sz != c->chk_lpt_sz2) {
+		if (d->chk_lpt_sz2 && d->chk_lpt_sz != d->chk_lpt_sz2) {
 			dbg_err("LPT layout size %lld but wrote %lld",
-				c->chk_lpt_sz, c->chk_lpt_sz2);
+				d->chk_lpt_sz, d->chk_lpt_sz2);
 			err = -EINVAL;
 		}
-		if (c->chk_lpt_sz2 && c->new_nhead_offs != len) {
+		if (d->chk_lpt_sz2 && d->new_nhead_offs != len) {
 			dbg_err("LPT new nhead offs: expected %d was %d",
-				c->new_nhead_offs, len);
+				d->new_nhead_offs, len);
 			err = -EINVAL;
 		}
 		lpt_sz = (long long)c->pnode_cnt * c->pnode_sz;
@@ -1788,26 +1804,146 @@ int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len)
 		lpt_sz += c->ltab_sz;
 		if (c->big_lpt)
 			lpt_sz += c->lsave_sz;
-		if (c->chk_lpt_sz - c->chk_lpt_wastage > lpt_sz) {
+		if (d->chk_lpt_sz - d->chk_lpt_wastage > lpt_sz) {
 			dbg_err("LPT chk_lpt_sz %lld + waste %lld exceeds %lld",
-				c->chk_lpt_sz, c->chk_lpt_wastage, lpt_sz);
+				d->chk_lpt_sz, d->chk_lpt_wastage, lpt_sz);
 			err = -EINVAL;
 		}
-		if (err)
+		if (err) {
 			dbg_dump_lpt_info(c);
-		c->chk_lpt_sz2 = c->chk_lpt_sz;
-		c->chk_lpt_sz = 0;
-		c->chk_lpt_wastage = 0;
-		c->chk_lpt_lebs = 0;
-		c->new_nhead_offs = len;
+			dbg_dump_lpt_lebs(c);
+			dump_stack();
+		}
+		d->chk_lpt_sz2 = d->chk_lpt_sz;
+		d->chk_lpt_sz = 0;
+		d->chk_lpt_wastage = 0;
+		d->chk_lpt_lebs = 0;
+		d->new_nhead_offs = len;
 		return err;
 	case 4:
-		c->chk_lpt_sz += len;
-		c->chk_lpt_wastage += len;
+		d->chk_lpt_sz += len;
+		d->chk_lpt_wastage += len;
 		return 0;
 	default:
 		return -EINVAL;
 	}
 }
 
+/**
+ * dbg_dump_lpt_leb - dump an LPT LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to dump
+ *
+ * This function dumps an LEB from LPT area. Nodes in this area are very
+ * different to nodes in the main area (e.g., they do not have common headers,
+ * they do not have 8-byte alignments, etc), so we have a separate function to
+ * dump LPT area LEBs. Note, LPT has to be locked by the caller.
+ */
+static void dump_lpt_leb(const struct ubifs_info *c, int lnum)
+{
+	int err, len = c->leb_size, node_type, node_num, node_len, offs;
+	void *buf = c->dbg->buf;
+
+	printk(KERN_DEBUG "(pid %d) start dumping LEB %d\n",
+	       current->pid, lnum);
+	err = ubi_read(c->ubi, lnum, buf, 0, c->leb_size);
+	if (err) {
+		ubifs_err("cannot read LEB %d, error %d", lnum, err);
+		return;
+	}
+	while (1) {
+		offs = c->leb_size - len;
+		if (!is_a_node(c, buf, len)) {
+			int pad_len;
+
+			pad_len = get_pad_len(c, buf, len);
+			if (pad_len) {
+				printk(KERN_DEBUG "LEB %d:%d, pad %d bytes\n",
+				       lnum, offs, pad_len);
+				buf += pad_len;
+				len -= pad_len;
+				continue;
+			}
+			if (len)
+				printk(KERN_DEBUG "LEB %d:%d, free %d bytes\n",
+				       lnum, offs, len);
+			break;
+		}
+
+		node_type = get_lpt_node_type(c, buf, &node_num);
+		switch (node_type) {
+		case UBIFS_LPT_PNODE:
+		{
+			node_len = c->pnode_sz;
+			if (c->big_lpt)
+				printk(KERN_DEBUG "LEB %d:%d, pnode num %d\n",
+				       lnum, offs, node_num);
+			else
+				printk(KERN_DEBUG "LEB %d:%d, pnode\n",
+				       lnum, offs);
+			break;
+		}
+		case UBIFS_LPT_NNODE:
+		{
+			int i;
+			struct ubifs_nnode nnode;
+
+			node_len = c->nnode_sz;
+			if (c->big_lpt)
+				printk(KERN_DEBUG "LEB %d:%d, nnode num %d, ",
+				       lnum, offs, node_num);
+			else
+				printk(KERN_DEBUG "LEB %d:%d, nnode, ",
+				       lnum, offs);
+			err = ubifs_unpack_nnode(c, buf, &nnode);
+			for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+				printk("%d:%d", nnode.nbranch[i].lnum,
+				       nnode.nbranch[i].offs);
+				if (i != UBIFS_LPT_FANOUT - 1)
+					printk(", ");
+			}
+			printk("\n");
+			break;
+		}
+		case UBIFS_LPT_LTAB:
+			node_len = c->ltab_sz;
+			printk(KERN_DEBUG "LEB %d:%d, ltab\n",
+			       lnum, offs);
+			break;
+		case UBIFS_LPT_LSAVE:
+			node_len = c->lsave_sz;
+			printk(KERN_DEBUG "LEB %d:%d, lsave len\n", lnum, offs);
+			break;
+		default:
+			ubifs_err("LPT node type %d not recognized", node_type);
+			return;
+		}
+
+		buf += node_len;
+		len -= node_len;
+	}
+
+	printk(KERN_DEBUG "(pid %d) finish dumping LEB %d\n",
+	       current->pid, lnum);
+}
+
+/**
+ * dbg_dump_lpt_lebs - dump LPT lebs.
+ * @c: UBIFS file-system description object
+ *
+ * This function dumps all LPT LEBs. The caller has to make sure the LPT is
+ * locked.
+ */
+void dbg_dump_lpt_lebs(const struct ubifs_info *c)
+{
+	int i;
+
+	printk(KERN_DEBUG "(pid %d) start dumping all LPT LEBs\n",
+	       current->pid);
+	for (i = 0; i < c->lpt_lebs; i++)
+		dump_lpt_leb(c, i + c->lpt_first);
+	printk(KERN_DEBUG "(pid %d) finish dumping all LPT LEBs\n",
+	       current->pid);
+}
+
 #endif /* CONFIG_UBIFS_FS_DEBUG */
diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
index 9bd5a43d452..9e6f403f170 100644
--- a/fs/ubifs/orphan.c
+++ b/fs/ubifs/orphan.c
@@ -899,7 +899,7 @@ static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
 		struct ubifs_scan_leb *sleb;
 
-		sleb = ubifs_scan(c, lnum, 0, c->dbg_buf);
+		sleb = ubifs_scan(c, lnum, 0, c->dbg->buf);
 		if (IS_ERR(sleb)) {
 			err = PTR_ERR(sleb);
 			break;
diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
index 21f7d047c30..ce42a7b0ca5 100644
--- a/fs/ubifs/replay.c
+++ b/fs/ubifs/replay.c
@@ -144,7 +144,7 @@ static int set_bud_lprops(struct ubifs_info *c, struct replay_entry *r)
 		/*
 		 * If the replay order was perfect the dirty space would now be
 		 * zero. The order is not perfect because the the journal heads
-		 * race with eachother. This is not a problem but is does mean
+		 * race with each other. This is not a problem but is does mean
 		 * that the dirty space may temporarily exceed c->leb_size
 		 * during the replay.
 		 */
@@ -656,7 +656,7 @@ out_dump:
  * @dirty: amount of dirty space from padding and deletion nodes
  *
  * This function inserts a reference node to the replay tree and returns zero
- * in case of success ort a negative error code in case of failure.
+ * in case of success or a negative error code in case of failure.
  */
 static int insert_ref_node(struct ubifs_info *c, int lnum, int offs,
 			   unsigned long long sqnum, int free, int dirty)
@@ -883,7 +883,7 @@ static int replay_log_leb(struct ubifs_info *c, int lnum, int offs, void *sbuf)
 		 * This means that we reached end of log and now
 		 * look to the older log data, which was already
 		 * committed but the eraseblock was not erased (UBIFS
-		 * only unmaps it). So this basically means we have to
+		 * only un-maps it). So this basically means we have to
 		 * exit with "end of log" code.
 		 */
 		err = 1;
@@ -1062,6 +1062,15 @@ int ubifs_replay_journal(struct ubifs_info *c)
 	if (err)
 		goto out;
 
+	/*
+	 * UBIFS budgeting calculations use @c->budg_uncommitted_idx variable
+	 * to roughly estimate index growth. Things like @c->min_idx_lebs
+	 * depend on it. This means we have to initialize it to make sure
+	 * budgeting works properly.
+	 */
+	c->budg_uncommitted_idx = atomic_long_read(&c->dirty_zn_cnt);
+	c->budg_uncommitted_idx *= c->max_idx_node_sz;
+
 	ubifs_assert(c->bud_bytes <= c->max_bud_bytes || c->need_recovery);
 	dbg_mnt("finished, log head LEB %d:%d, max_sqnum %llu, "
 		"highest_inum %lu", c->lhead_lnum, c->lhead_offs, c->max_sqnum,
diff --git a/fs/ubifs/sb.c b/fs/ubifs/sb.c
index 0f392351dc5..e070c643d1b 100644
--- a/fs/ubifs/sb.c
+++ b/fs/ubifs/sb.c
@@ -28,6 +28,7 @@
 
 #include "ubifs.h"
 #include <linux/random.h>
+#include <linux/math64.h>
 
 /*
  * Default journal size in logical eraseblocks as a percent of total
@@ -80,7 +81,7 @@ static int create_default_filesystem(struct ubifs_info *c)
 	int err, tmp, jnl_lebs, log_lebs, max_buds, main_lebs, main_first;
 	int lpt_lebs, lpt_first, orph_lebs, big_lpt, ino_waste, sup_flags = 0;
 	int min_leb_cnt = UBIFS_MIN_LEB_CNT;
-	uint64_t tmp64, main_bytes;
+	long long tmp64, main_bytes;
 	__le64 tmp_le64;
 
 	/* Some functions called from here depend on the @c->key_len filed */
@@ -160,7 +161,7 @@ static int create_default_filesystem(struct ubifs_info *c)
 	if (!sup)
 		return -ENOMEM;
 
-	tmp64 = (uint64_t)max_buds * c->leb_size;
+	tmp64 = (long long)max_buds * c->leb_size;
 	if (big_lpt)
 		sup_flags |= UBIFS_FLG_BIGLPT;
 
@@ -179,14 +180,16 @@ static int create_default_filesystem(struct ubifs_info *c)
 	sup->fanout        = cpu_to_le32(DEFAULT_FANOUT);
 	sup->lsave_cnt     = cpu_to_le32(c->lsave_cnt);
 	sup->fmt_version   = cpu_to_le32(UBIFS_FORMAT_VERSION);
-	sup->default_compr = cpu_to_le16(UBIFS_COMPR_LZO);
 	sup->time_gran     = cpu_to_le32(DEFAULT_TIME_GRAN);
+	if (c->mount_opts.override_compr)
+		sup->default_compr = cpu_to_le16(c->mount_opts.compr_type);
+	else
+		sup->default_compr = cpu_to_le16(UBIFS_COMPR_LZO);
 
 	generate_random_uuid(sup->uuid);
 
-	main_bytes = (uint64_t)main_lebs * c->leb_size;
-	tmp64 = main_bytes * DEFAULT_RP_PERCENT;
-	do_div(tmp64, 100);
+	main_bytes = (long long)main_lebs * c->leb_size;
+	tmp64 = div_u64(main_bytes * DEFAULT_RP_PERCENT, 100);
 	if (tmp64 > DEFAULT_MAX_RP_SIZE)
 		tmp64 = DEFAULT_MAX_RP_SIZE;
 	sup->rp_size = cpu_to_le64(tmp64);
@@ -582,16 +585,15 @@ int ubifs_read_superblock(struct ubifs_info *c)
 	c->jhead_cnt     = le32_to_cpu(sup->jhead_cnt) + NONDATA_JHEADS_CNT;
 	c->fanout        = le32_to_cpu(sup->fanout);
 	c->lsave_cnt     = le32_to_cpu(sup->lsave_cnt);
-	c->default_compr = le16_to_cpu(sup->default_compr);
 	c->rp_size       = le64_to_cpu(sup->rp_size);
 	c->rp_uid        = le32_to_cpu(sup->rp_uid);
 	c->rp_gid        = le32_to_cpu(sup->rp_gid);
 	sup_flags        = le32_to_cpu(sup->flags);
+	if (!c->mount_opts.override_compr)
+		c->default_compr = le16_to_cpu(sup->default_compr);
 
 	c->vfs_sb->s_time_gran = le32_to_cpu(sup->time_gran);
-
 	memcpy(&c->uuid, &sup->uuid, 16);
-
 	c->big_lpt = !!(sup_flags & UBIFS_FLG_BIGLPT);
 
 	/* Automatically increase file system size to the maximum size */
diff --git a/fs/ubifs/shrinker.c b/fs/ubifs/shrinker.c
index f248533841a..e7bab52a141 100644
--- a/fs/ubifs/shrinker.c
+++ b/fs/ubifs/shrinker.c
@@ -151,7 +151,7 @@ static int shrink_tnc(struct ubifs_info *c, int nr, int age, int *contention)
  * @contention: if any contention, this is set to %1
  *
  * This function walks the list of mounted UBIFS file-systems and frees clean
- * znodes which are older then @age, until at least @nr znodes are freed.
+ * znodes which are older than @age, until at least @nr znodes are freed.
  * Returns the number of freed znodes.
  */
 static int shrink_tnc_trees(int nr, int age, int *contention)
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index d80b2aef42b..89556ee7251 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -34,6 +34,8 @@
 #include <linux/parser.h>
 #include <linux/seq_file.h>
 #include <linux/mount.h>
+#include <linux/math64.h>
+#include <linux/writeback.h>
 #include "ubifs.h"
 
 /*
@@ -417,39 +419,61 @@ static int ubifs_show_options(struct seq_file *s, struct vfsmount *mnt)
 	else if (c->mount_opts.chk_data_crc == 1)
 		seq_printf(s, ",no_chk_data_crc");
 
+	if (c->mount_opts.override_compr) {
+		seq_printf(s, ",compr=");
+		seq_printf(s, ubifs_compr_name(c->mount_opts.compr_type));
+	}
+
 	return 0;
 }
 
 static int ubifs_sync_fs(struct super_block *sb, int wait)
 {
+	int i, err;
 	struct ubifs_info *c = sb->s_fs_info;
-	int i, ret = 0, err;
-	long long bud_bytes;
+	struct writeback_control wbc = {
+		.sync_mode   = wait ? WB_SYNC_ALL : WB_SYNC_NONE,
+		.range_start = 0,
+		.range_end   = LLONG_MAX,
+		.nr_to_write = LONG_MAX,
+	};
 
-	if (c->jheads) {
-		for (i = 0; i < c->jhead_cnt; i++) {
-			err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
-			if (err && !ret)
-				ret = err;
-		}
+	/*
+	 * Note by akpm about WB_SYNC_NONE used above: zero @wait is just an
+	 * advisory thing to help the file system shove lots of data into the
+	 * queues. If some gets missed then it'll be picked up on the second
+	 * '->sync_fs()' call, with non-zero @wait.
+	 */
 
-		/* Commit the journal unless it has too little data */
-		spin_lock(&c->buds_lock);
-		bud_bytes = c->bud_bytes;
-		spin_unlock(&c->buds_lock);
-		if (bud_bytes > c->leb_size) {
-			err = ubifs_run_commit(c);
-			if (err)
-				return err;
-		}
+	if (sb->s_flags & MS_RDONLY)
+		return 0;
+
+	/*
+	 * Synchronize write buffers, because 'ubifs_run_commit()' does not
+	 * do this if it waits for an already running commit.
+	 */
+	for (i = 0; i < c->jhead_cnt; i++) {
+		err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
+		if (err)
+			return err;
 	}
 
 	/*
-	 * We ought to call sync for c->ubi but it does not have one. If it had
-	 * it would in turn call mtd->sync, however mtd operations are
-	 * synchronous anyway, so we don't lose any sleep here.
+	 * VFS calls '->sync_fs()' before synchronizing all dirty inodes and
+	 * pages, so synchronize them first, then commit the journal. Strictly
+	 * speaking, it is not necessary to commit the journal here,
+	 * synchronizing write-buffers would be enough. But committing makes
+	 * UBIFS free space predictions much more accurate, so we want to let
+	 * the user be able to get more accurate results of 'statfs()' after
+	 * they synchronize the file system.
 	 */
-	return ret;
+	generic_sync_sb_inodes(sb, &wbc);
+
+	err = ubifs_run_commit(c);
+	if (err)
+		return err;
+
+	return ubi_sync(c->vi.ubi_num);
 }
 
 /**
@@ -596,7 +620,7 @@ static int bud_wbuf_callback(struct ubifs_info *c, int lnum, int free, int pad)
 }
 
 /*
- * init_constants_late - initialize UBIFS constants.
+ * init_constants_sb - initialize UBIFS constants.
  * @c: UBIFS file-system description object
  *
  * This is a helper function which initializes various UBIFS constants after
@@ -604,10 +628,10 @@ static int bud_wbuf_callback(struct ubifs_info *c, int lnum, int free, int pad)
  * makes sure they are all right. Returns zero in case of success and a
  * negative error code in case of failure.
  */
-static int init_constants_late(struct ubifs_info *c)
+static int init_constants_sb(struct ubifs_info *c)
 {
 	int tmp, err;
-	uint64_t tmp64;
+	long long tmp64;
 
 	c->main_bytes = (long long)c->main_lebs * c->leb_size;
 	c->max_znode_sz = sizeof(struct ubifs_znode) +
@@ -634,9 +658,8 @@ static int init_constants_late(struct ubifs_info *c)
 	 * Make sure that the log is large enough to fit reference nodes for
 	 * all buds plus one reserved LEB.
 	 */
-	tmp64 = c->max_bud_bytes;
-	tmp = do_div(tmp64, c->leb_size);
-	c->max_bud_cnt = tmp64 + !!tmp;
+	tmp64 = c->max_bud_bytes + c->leb_size - 1;
+	c->max_bud_cnt = div_u64(tmp64, c->leb_size);
 	tmp = (c->ref_node_alsz * c->max_bud_cnt + c->leb_size - 1);
 	tmp /= c->leb_size;
 	tmp += 1;
@@ -672,7 +695,7 @@ static int init_constants_late(struct ubifs_info *c)
 	 * Consequently, if the journal is too small, UBIFS will treat it as
 	 * always full.
 	 */
-	tmp64 = (uint64_t)(c->jhead_cnt + 1) * c->leb_size + 1;
+	tmp64 = (long long)(c->jhead_cnt + 1) * c->leb_size + 1;
 	if (c->bg_bud_bytes < tmp64)
 		c->bg_bud_bytes = tmp64;
 	if (c->max_bud_bytes < tmp64 + c->leb_size)
@@ -682,6 +705,21 @@ static int init_constants_late(struct ubifs_info *c)
 	if (err)
 		return err;
 
+	return 0;
+}
+
+/*
+ * init_constants_master - initialize UBIFS constants.
+ * @c: UBIFS file-system description object
+ *
+ * This is a helper function which initializes various UBIFS constants after
+ * the master node has been read. It also checks various UBIFS parameters and
+ * makes sure they are all right.
+ */
+static void init_constants_master(struct ubifs_info *c)
+{
+	long long tmp64;
+
 	c->min_idx_lebs = ubifs_calc_min_idx_lebs(c);
 
 	/*
@@ -690,14 +728,13 @@ static int init_constants_late(struct ubifs_info *c)
 	 * necessary to report something for the 'statfs()' call.
 	 *
 	 * Subtract the LEB reserved for GC, the LEB which is reserved for
-	 * deletions, and assume only one journal head is available.
+	 * deletions, minimum LEBs for the index, and assume only one journal
+	 * head is available.
 	 */
-	tmp64 = c->main_lebs - 2 - c->jhead_cnt + 1;
-	tmp64 *= (uint64_t)c->leb_size - c->leb_overhead;
+	tmp64 = c->main_lebs - 1 - 1 - MIN_INDEX_LEBS - c->jhead_cnt + 1;
+	tmp64 *= (long long)c->leb_size - c->leb_overhead;
 	tmp64 = ubifs_reported_space(c, tmp64);
 	c->block_cnt = tmp64 >> UBIFS_BLOCK_SHIFT;
-
-	return 0;
 }
 
 /**
@@ -878,6 +915,7 @@ static int check_volume_empty(struct ubifs_info *c)
  * Opt_no_bulk_read: disable bulk-reads
  * Opt_chk_data_crc: check CRCs when reading data nodes
  * Opt_no_chk_data_crc: do not check CRCs when reading data nodes
+ * Opt_override_compr: override default compressor
  * Opt_err: just end of array marker
  */
 enum {
@@ -887,6 +925,7 @@ enum {
 	Opt_no_bulk_read,
 	Opt_chk_data_crc,
 	Opt_no_chk_data_crc,
+	Opt_override_compr,
 	Opt_err,
 };
 
@@ -897,6 +936,7 @@ static const match_table_t tokens = {
 	{Opt_no_bulk_read, "no_bulk_read"},
 	{Opt_chk_data_crc, "chk_data_crc"},
 	{Opt_no_chk_data_crc, "no_chk_data_crc"},
+	{Opt_override_compr, "compr=%s"},
 	{Opt_err, NULL},
 };
 
@@ -950,6 +990,28 @@ static int ubifs_parse_options(struct ubifs_info *c, char *options,
 			c->mount_opts.chk_data_crc = 1;
 			c->no_chk_data_crc = 1;
 			break;
+		case Opt_override_compr:
+		{
+			char *name = match_strdup(&args[0]);
+
+			if (!name)
+				return -ENOMEM;
+			if (!strcmp(name, "none"))
+				c->mount_opts.compr_type = UBIFS_COMPR_NONE;
+			else if (!strcmp(name, "lzo"))
+				c->mount_opts.compr_type = UBIFS_COMPR_LZO;
+			else if (!strcmp(name, "zlib"))
+				c->mount_opts.compr_type = UBIFS_COMPR_ZLIB;
+			else {
+				ubifs_err("unknown compressor \"%s\"", name);
+				kfree(name);
+				return -EINVAL;
+			}
+			kfree(name);
+			c->mount_opts.override_compr = 1;
+			c->default_compr = c->mount_opts.compr_type;
+			break;
+		}
 		default:
 			ubifs_err("unrecognized mount option \"%s\" "
 				  "or missing value", p);
@@ -1019,6 +1081,30 @@ again:
 }
 
 /**
+ * check_free_space - check if there is enough free space to mount.
+ * @c: UBIFS file-system description object
+ *
+ * This function makes sure UBIFS has enough free space to be mounted in
+ * read/write mode. UBIFS must always have some free space to allow deletions.
+ */
+static int check_free_space(struct ubifs_info *c)
+{
+	ubifs_assert(c->dark_wm > 0);
+	if (c->lst.total_free + c->lst.total_dirty < c->dark_wm) {
+		ubifs_err("insufficient free space to mount in read/write mode");
+		dbg_dump_budg(c);
+		dbg_dump_lprops(c);
+		/*
+		 * We return %-EINVAL instead of %-ENOSPC because it seems to
+		 * be the closest error code mentioned in the mount function
+		 * documentation.
+		 */
+		return -EINVAL;
+	}
+	return 0;
+}
+
+/**
  * mount_ubifs - mount UBIFS file-system.
  * @c: UBIFS file-system description object
  *
@@ -1039,11 +1125,9 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (err)
 		return err;
 
-#ifdef CONFIG_UBIFS_FS_DEBUG
-	c->dbg_buf = vmalloc(c->leb_size);
-	if (!c->dbg_buf)
-		return -ENOMEM;
-#endif
+	err = ubifs_debugging_init(c);
+	if (err)
+		return err;
 
 	err = check_volume_empty(c);
 	if (err)
@@ -1100,27 +1184,25 @@ static int mount_ubifs(struct ubifs_info *c)
 		goto out_free;
 
 	/*
-	 * Make sure the compressor which is set as the default on in the
-	 * superblock was actually compiled in.
+	 * Make sure the compressor which is set as default in the superblock
+	 * or overridden by mount options is actually compiled in.
 	 */
 	if (!ubifs_compr_present(c->default_compr)) {
-		ubifs_warn("'%s' compressor is set by superblock, but not "
-			   "compiled in", ubifs_compr_name(c->default_compr));
-		c->default_compr = UBIFS_COMPR_NONE;
+		ubifs_err("'compressor \"%s\" is not compiled in",
+			  ubifs_compr_name(c->default_compr));
+		goto out_free;
 	}
 
-	dbg_failure_mode_registration(c);
-
-	err = init_constants_late(c);
+	err = init_constants_sb(c);
 	if (err)
-		goto out_dereg;
+		goto out_free;
 
 	sz = ALIGN(c->max_idx_node_sz, c->min_io_size);
 	sz = ALIGN(sz + c->max_idx_node_sz, c->min_io_size);
 	c->cbuf = kmalloc(sz, GFP_NOFS);
 	if (!c->cbuf) {
 		err = -ENOMEM;
-		goto out_dereg;
+		goto out_free;
 	}
 
 	sprintf(c->bgt_name, BGT_NAME_PATTERN, c->vi.ubi_num, c->vi.vol_id);
@@ -1145,6 +1227,8 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (err)
 		goto out_master;
 
+	init_constants_master(c);
+
 	if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) {
 		ubifs_msg("recovery needed");
 		c->need_recovery = 1;
@@ -1183,12 +1267,9 @@ static int mount_ubifs(struct ubifs_info *c)
 	if (!mounted_read_only) {
 		int lnum;
 
-		/* Check for enough free space */
-		if (ubifs_calc_available(c, c->min_idx_lebs) <= 0) {
-			ubifs_err("insufficient available space");
-			err = -EINVAL;
+		err = check_free_space(c);
+		if (err)
 			goto out_orphans;
-		}
 
 		/* Check for enough log space */
 		lnum = c->lhead_lnum + 1;
@@ -1232,6 +1313,10 @@ static int mount_ubifs(struct ubifs_info *c)
 		}
 	}
 
+	err = dbg_debugfs_init_fs(c);
+	if (err)
+		goto out_infos;
+
 	err = dbg_check_filesystem(c);
 	if (err)
 		goto out_infos;
@@ -1283,8 +1368,20 @@ static int mount_ubifs(struct ubifs_info *c)
 	dbg_msg("tree fanout:         %d", c->fanout);
 	dbg_msg("reserved GC LEB:     %d", c->gc_lnum);
 	dbg_msg("first main LEB:      %d", c->main_first);
+	dbg_msg("max. znode size      %d", c->max_znode_sz);
+	dbg_msg("max. index node size %d", c->max_idx_node_sz);
+	dbg_msg("node sizes:          data %zu, inode %zu, dentry %zu",
+		UBIFS_DATA_NODE_SZ, UBIFS_INO_NODE_SZ, UBIFS_DENT_NODE_SZ);
+	dbg_msg("node sizes:          trun %zu, sb %zu, master %zu",
+		UBIFS_TRUN_NODE_SZ, UBIFS_SB_NODE_SZ, UBIFS_MST_NODE_SZ);
+	dbg_msg("node sizes:          ref %zu, cmt. start %zu, orph %zu",
+		UBIFS_REF_NODE_SZ, UBIFS_CS_NODE_SZ, UBIFS_ORPH_NODE_SZ);
+	dbg_msg("max. node sizes:     data %zu, inode %zu dentry %zu",
+	        UBIFS_MAX_DATA_NODE_SZ, UBIFS_MAX_INO_NODE_SZ,
+		UBIFS_MAX_DENT_NODE_SZ);
 	dbg_msg("dead watermark:      %d", c->dead_wm);
 	dbg_msg("dark watermark:      %d", c->dark_wm);
+	dbg_msg("LEB overhead:        %d", c->leb_overhead);
 	x = (long long)c->main_lebs * c->dark_wm;
 	dbg_msg("max. dark space:     %lld (%lld KiB, %lld MiB)",
 		x, x >> 10, x >> 20);
@@ -1320,14 +1417,12 @@ out_wbufs:
 	free_wbufs(c);
 out_cbuf:
 	kfree(c->cbuf);
-out_dereg:
-	dbg_failure_mode_deregistration(c);
 out_free:
 	kfree(c->bu.buf);
 	vfree(c->ileb_buf);
 	vfree(c->sbuf);
 	kfree(c->bottom_up_buf);
-	UBIFS_DBG(vfree(c->dbg_buf));
+	ubifs_debugging_exit(c);
 	return err;
 }
 
@@ -1345,6 +1440,7 @@ static void ubifs_umount(struct ubifs_info *c)
 	dbg_gen("un-mounting UBI device %d, volume %d", c->vi.ubi_num,
 		c->vi.vol_id);
 
+	dbg_debugfs_exit_fs(c);
 	spin_lock(&ubifs_infos_lock);
 	list_del(&c->infos_list);
 	spin_unlock(&ubifs_infos_lock);
@@ -1364,8 +1460,7 @@ static void ubifs_umount(struct ubifs_info *c)
 	vfree(c->ileb_buf);
 	vfree(c->sbuf);
 	kfree(c->bottom_up_buf);
-	UBIFS_DBG(vfree(c->dbg_buf));
-	dbg_failure_mode_deregistration(c);
+	ubifs_debugging_exit(c);
 }
 
 /**
@@ -1387,12 +1482,9 @@ static int ubifs_remount_rw(struct ubifs_info *c)
 	c->remounting_rw = 1;
 	c->always_chk_crc = 1;
 
-	/* Check for enough free space */
-	if (ubifs_calc_available(c, c->min_idx_lebs) <= 0) {
-		ubifs_err("insufficient available space");
-		err = -EINVAL;
+	err = check_free_space(c);
+	if (err)
 		goto out;
-	}
 
 	if (c->old_leb_cnt != c->leb_cnt) {
 		struct ubifs_sb_node *sup;
@@ -1515,20 +1607,24 @@ out:
  * @c: UBIFS file-system description object
  *
  * This function is called during un-mounting and re-mounting, and it commits
- * the journal unless the "fast unmount" mode is enabled. It also avoids
- * committing the journal if it contains too few data.
+ * the journal unless the "fast unmount" mode is enabled.
  */
 static void commit_on_unmount(struct ubifs_info *c)
 {
-	if (!c->fast_unmount) {
-		long long bud_bytes;
+	struct super_block *sb = c->vfs_sb;
+	long long bud_bytes;
 
-		spin_lock(&c->buds_lock);
-		bud_bytes = c->bud_bytes;
-		spin_unlock(&c->buds_lock);
-		if (bud_bytes > c->leb_size)
-			ubifs_run_commit(c);
-	}
+	/*
+	 * This function is called before the background thread is stopped, so
+	 * we may race with ongoing commit, which means we have to take
+	 * @c->bud_lock to access @c->bud_bytes.
+	 */
+	spin_lock(&c->buds_lock);
+	bud_bytes = c->bud_bytes;
+	spin_unlock(&c->buds_lock);
+
+	if (!c->fast_unmount && !(sb->s_flags & MS_RDONLY) && bud_bytes)
+		ubifs_run_commit(c);
 }
 
 /**
@@ -1849,7 +1945,6 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 		goto out_iput;
 
 	mutex_unlock(&c->umount_mutex);
-
 	return 0;
 
 out_iput:
@@ -1955,7 +2050,7 @@ static void ubifs_kill_sb(struct super_block *sb)
 	 * We do 'commit_on_unmount()' here instead of 'ubifs_put_super()'
 	 * in order to be outside BKL.
 	 */
-	if (sb->s_root && !(sb->s_flags & MS_RDONLY))
+	if (sb->s_root)
 		commit_on_unmount(c);
 	/* The un-mount routine is actually done in put_super() */
 	generic_shutdown_super(sb);
@@ -2021,6 +2116,14 @@ static int __init ubifs_init(void)
 	BUILD_BUG_ON(UBIFS_REF_NODE_SZ != 64);
 
 	/*
+	 * We use 2 bit wide bit-fields to store compression type, which should
+	 * be amended if more compressors are added. The bit-fields are:
+	 * @compr_type in 'struct ubifs_inode', @default_compr in
+	 * 'struct ubifs_info' and @compr_type in 'struct ubifs_mount_opts'.
+	 */
+	BUILD_BUG_ON(UBIFS_COMPR_TYPES_CNT > 4);
+
+	/*
 	 * We require that PAGE_CACHE_SIZE is greater-than-or-equal-to
 	 * UBIFS_BLOCK_SIZE. It is assumed that both are powers of 2.
 	 */
@@ -2049,11 +2152,17 @@ static int __init ubifs_init(void)
 
 	err = ubifs_compressors_init();
 	if (err)
+		goto out_shrinker;
+
+	err = dbg_debugfs_init();
+	if (err)
 		goto out_compr;
 
 	return 0;
 
 out_compr:
+	ubifs_compressors_exit();
+out_shrinker:
 	unregister_shrinker(&ubifs_shrinker_info);
 	kmem_cache_destroy(ubifs_inode_slab);
 out_reg:
@@ -2068,6 +2177,7 @@ static void __exit ubifs_exit(void)
 	ubifs_assert(list_empty(&ubifs_infos));
 	ubifs_assert(atomic_long_read(&ubifs_clean_zn_cnt) == 0);
 
+	dbg_debugfs_exit();
 	ubifs_compressors_exit();
 	unregister_shrinker(&ubifs_shrinker_info);
 	kmem_cache_destroy(ubifs_inode_slab);
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index 6eef5344a14..f7e36f54552 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -2245,12 +2245,11 @@ int ubifs_tnc_replace(struct ubifs_info *c, const union ubifs_key *key,
 			if (found) {
 				/* Ensure the znode is dirtied */
 				if (znode->cnext || !ubifs_zn_dirty(znode)) {
-					    znode = dirty_cow_bottom_up(c,
-									znode);
-					    if (IS_ERR(znode)) {
-						    err = PTR_ERR(znode);
-						    goto out_unlock;
-					    }
+					znode = dirty_cow_bottom_up(c, znode);
+					if (IS_ERR(znode)) {
+						err = PTR_ERR(znode);
+						goto out_unlock;
+					}
 				}
 				zbr = &znode->zbranch[n];
 				lnc_free(zbr);
@@ -2317,11 +2316,11 @@ int ubifs_tnc_add_nm(struct ubifs_info *c, const union ubifs_key *key,
 
 		/* Ensure the znode is dirtied */
 		if (znode->cnext || !ubifs_zn_dirty(znode)) {
-			    znode = dirty_cow_bottom_up(c, znode);
-			    if (IS_ERR(znode)) {
-				    err = PTR_ERR(znode);
-				    goto out_unlock;
-			    }
+			znode = dirty_cow_bottom_up(c, znode);
+			if (IS_ERR(znode)) {
+				err = PTR_ERR(znode);
+				goto out_unlock;
+			}
 		}
 
 		if (found == 1) {
@@ -2627,11 +2626,11 @@ int ubifs_tnc_remove_range(struct ubifs_info *c, union ubifs_key *from_key,
 
 		/* Ensure the znode is dirtied */
 		if (znode->cnext || !ubifs_zn_dirty(znode)) {
-			    znode = dirty_cow_bottom_up(c, znode);
-			    if (IS_ERR(znode)) {
-				    err = PTR_ERR(znode);
-				    goto out_unlock;
-			    }
+			znode = dirty_cow_bottom_up(c, znode);
+			if (IS_ERR(znode)) {
+				err = PTR_ERR(znode);
+				goto out_unlock;
+			}
 		}
 
 		/* Remove all keys in range except the first */
diff --git a/fs/ubifs/tnc_commit.c b/fs/ubifs/tnc_commit.c
index 8ac76b1c2d5..fde8d127c76 100644
--- a/fs/ubifs/tnc_commit.c
+++ b/fs/ubifs/tnc_commit.c
@@ -553,8 +553,8 @@ static int layout_in_empty_space(struct ubifs_info *c)
 	}
 
 #ifdef CONFIG_UBIFS_FS_DEBUG
-	c->new_ihead_lnum = lnum;
-	c->new_ihead_offs = buf_offs;
+	c->dbg->new_ihead_lnum = lnum;
+	c->dbg->new_ihead_offs = buf_offs;
 #endif
 
 	return 0;
@@ -802,8 +802,10 @@ int ubifs_tnc_start_commit(struct ubifs_info *c, struct ubifs_zbranch *zroot)
 	 * budgeting subsystem to assume the index is already committed,
 	 * even though it is not.
 	 */
+	ubifs_assert(c->min_idx_lebs == ubifs_calc_min_idx_lebs(c));
 	c->old_idx_sz = c->calc_idx_sz;
 	c->budg_uncommitted_idx = 0;
+	c->min_idx_lebs = ubifs_calc_min_idx_lebs(c);
 	spin_unlock(&c->space_lock);
 	mutex_unlock(&c->tnc_mutex);
 
@@ -1002,7 +1004,8 @@ static int write_index(struct ubifs_info *c)
 	}
 
 #ifdef CONFIG_UBIFS_FS_DEBUG
-	if (lnum != c->new_ihead_lnum || buf_offs != c->new_ihead_offs) {
+	if (lnum != c->dbg->new_ihead_lnum ||
+	    buf_offs != c->dbg->new_ihead_offs) {
 		ubifs_err("inconsistent ihead");
 		return -EINVAL;
 	}
diff --git a/fs/ubifs/ubifs-media.h b/fs/ubifs/ubifs-media.h
index 0b378042a3a..b25fc36cf72 100644
--- a/fs/ubifs/ubifs-media.h
+++ b/fs/ubifs/ubifs-media.h
@@ -51,6 +51,13 @@
  */
 #define UBIFS_MIN_COMPR_LEN 128
 
+/*
+ * If compressed data length is less than %UBIFS_MIN_COMPRESS_DIFF bytes
+ * shorter than uncompressed data length, UBIFS preferes to leave this data
+ * node uncompress, because it'll be read faster.
+ */
+#define UBIFS_MIN_COMPRESS_DIFF 64
+
 /* Root inode number */
 #define UBIFS_ROOT_INO 1
 
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index 46b172560a0..fc2a4cc66d0 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -63,6 +63,14 @@
 #define SQNUM_WARN_WATERMARK 0xFFFFFFFF00000000ULL
 #define SQNUM_WATERMARK      0xFFFFFFFFFF000000ULL
 
+/*
+ * Minimum amount of LEBs reserved for the index. At present the index needs at
+ * least 2 LEBs: one for the index head and one for in-the-gaps method (which
+ * currently does not cater for the index head and so excludes it from
+ * consideration).
+ */
+#define MIN_INDEX_LEBS 2
+
 /* Minimum amount of data UBIFS writes to the flash */
 #define MIN_WRITE_SZ (UBIFS_DATA_NODE_SZ + 8)
 
@@ -386,12 +394,12 @@ struct ubifs_inode {
 	unsigned int dirty:1;
 	unsigned int xattr:1;
 	unsigned int bulk_read:1;
+	unsigned int compr_type:2;
 	struct mutex ui_mutex;
 	spinlock_t ui_lock;
 	loff_t synced_i_size;
 	loff_t ui_size;
 	int flags;
-	int compr_type;
 	pgoff_t last_page_read;
 	pgoff_t read_in_a_row;
 	int data_len;
@@ -419,7 +427,7 @@ struct ubifs_unclean_leb {
  *
  * LPROPS_UNCAT: not categorized
  * LPROPS_DIRTY: dirty > 0, not index
- * LPROPS_DIRTY_IDX: dirty + free > UBIFS_CH_SZ and index
+ * LPROPS_DIRTY_IDX: dirty + free > @c->min_idx_node_sze and index
  * LPROPS_FREE: free > 0, not empty, not index
  * LPROPS_HEAP_CNT: number of heaps used for storing categorized LEBs
  * LPROPS_EMPTY: LEB is empty, not taken
@@ -473,8 +481,8 @@ struct ubifs_lprops {
 struct ubifs_lpt_lprops {
 	int free;
 	int dirty;
-	unsigned tgc : 1;
-	unsigned cmt : 1;
+	unsigned tgc:1;
+	unsigned cmt:1;
 };
 
 /**
@@ -482,24 +490,26 @@ struct ubifs_lpt_lprops {
  * @empty_lebs: number of empty LEBs
  * @taken_empty_lebs: number of taken LEBs
  * @idx_lebs: number of indexing LEBs
- * @total_free: total free space in bytes
- * @total_dirty: total dirty space in bytes
- * @total_used: total used space in bytes (includes only data LEBs)
- * @total_dead: total dead space in bytes (includes only data LEBs)
- * @total_dark: total dark space in bytes (includes only data LEBs)
+ * @total_free: total free space in bytes (includes all LEBs)
+ * @total_dirty: total dirty space in bytes (includes all LEBs)
+ * @total_used: total used space in bytes (does not include index LEBs)
+ * @total_dead: total dead space in bytes (does not include index LEBs)
+ * @total_dark: total dark space in bytes (does not include index LEBs)
+ *
+ * The @taken_empty_lebs field counts the LEBs that are in the transient state
+ * of having been "taken" for use but not yet written to. @taken_empty_lebs is
+ * needed to account correctly for @gc_lnum, otherwise @empty_lebs could be
+ * used by itself (in which case 'unused_lebs' would be a better name). In the
+ * case of @gc_lnum, it is "taken" at mount time or whenever a LEB is retained
+ * by GC, but unlike other empty LEBs that are "taken", it may not be written
+ * straight away (i.e. before the next commit start or unmount), so either
+ * @gc_lnum must be specially accounted for, or the current approach followed
+ * i.e. count it under @taken_empty_lebs.
  *
- * N.B. total_dirty and total_used are different to other total_* fields,
- * because they account _all_ LEBs, not just data LEBs.
+ * @empty_lebs includes @taken_empty_lebs.
  *
- * 'taken_empty_lebs' counts the LEBs that are in the transient state of having
- * been 'taken' for use but not yet written to. 'taken_empty_lebs' is needed
- * to account correctly for gc_lnum, otherwise 'empty_lebs' could be used
- * by itself (in which case 'unused_lebs' would be a better name). In the case
- * of gc_lnum, it is 'taken' at mount time or whenever a LEB is retained by GC,
- * but unlike other empty LEBs that are 'taken', it may not be written straight
- * away (i.e. before the next commit start or unmount), so either gc_lnum must
- * be specially accounted for, or the current approach followed i.e. count it
- * under 'taken_empty_lebs'.
+ * @total_used, @total_dead and @total_dark fields do not account indexing
+ * LEBs.
  */
 struct ubifs_lp_stats {
 	int empty_lebs;
@@ -893,15 +903,25 @@ struct ubifs_orphan {
 /**
  * struct ubifs_mount_opts - UBIFS-specific mount options information.
  * @unmount_mode: selected unmount mode (%0 default, %1 normal, %2 fast)
- * @bulk_read: enable bulk-reads
- * @chk_data_crc: check CRCs when reading data nodes
+ * @bulk_read: enable/disable bulk-reads (%0 default, %1 disabe, %2 enable)
+ * @chk_data_crc: enable/disable CRC data checking when reading data nodes
+ *                (%0 default, %1 disabe, %2 enable)
+ * @override_compr: override default compressor (%0 - do not override and use
+ *                  superblock compressor, %1 - override and use compressor
+ *                  specified in @compr_type)
+ * @compr_type: compressor type to override the superblock compressor with
+ *              (%UBIFS_COMPR_NONE, etc)
  */
 struct ubifs_mount_opts {
 	unsigned int unmount_mode:2;
 	unsigned int bulk_read:2;
 	unsigned int chk_data_crc:2;
+	unsigned int override_compr:1;
+	unsigned int compr_type:2;
 };
 
+struct ubifs_debug_info;
+
 /**
  * struct ubifs_info - UBIFS file-system description data structure
  * (per-superblock).
@@ -946,6 +966,7 @@ struct ubifs_mount_opts {
  * @no_chk_data_crc: do not check CRCs when reading data nodes (except during
  *                   recovery)
  * @bulk_read: enable bulk-reads
+ * @default_compr: default compression algorithm (%UBIFS_COMPR_LZO, etc)
  *
  * @tnc_mutex: protects the Tree Node Cache (TNC), @zroot, @cnext, @enext, and
  *             @calc_idx_sz
@@ -963,8 +984,6 @@ struct ubifs_mount_opts {
  * @ileb_nxt: next pre-allocated index LEBs
  * @old_idx: tree of index nodes obsoleted since the last commit start
  * @bottom_up_buf: a buffer which is used by 'dirty_cow_bottom_up()' in tnc.c
- * @new_ihead_lnum: used by debugging to check ihead_lnum
- * @new_ihead_offs: used by debugging to check ihead_offs
  *
  * @mst_node: master node
  * @mst_offs: offset of valid master node
@@ -986,7 +1005,6 @@ struct ubifs_mount_opts {
  * @main_lebs: count of LEBs in the main area
  * @main_first: first LEB of the main area
  * @main_bytes: main area size in bytes
- * @default_compr: default compression algorithm (%UBIFS_COMPR_LZO, etc)
  *
  * @key_hash_type: type of the key hash
  * @key_hash: direntry key hash function
@@ -1149,15 +1167,7 @@ struct ubifs_mount_opts {
  * @always_chk_crc: always check CRCs (while mounting and remounting rw)
  * @mount_opts: UBIFS-specific mount options
  *
- * @dbg_buf: a buffer of LEB size used for debugging purposes
- * @old_zroot: old index root - used by 'dbg_check_old_index()'
- * @old_zroot_level: old index root level - used by 'dbg_check_old_index()'
- * @old_zroot_sqnum: old index root sqnum - used by 'dbg_check_old_index()'
- * @failure_mode: failure mode for recovery testing
- * @fail_delay: 0=>don't delay, 1=>delay a time, 2=>delay a number of calls
- * @fail_timeout: time in jiffies when delay of failure mode expires
- * @fail_cnt: current number of calls to failure mode I/O functions
- * @fail_cnt_max: number of calls by which to delay failure mode
+ * @dbg: debugging-related information
  */
 struct ubifs_info {
 	struct super_block *vfs_sb;
@@ -1196,6 +1206,7 @@ struct ubifs_info {
 	unsigned int big_lpt:1;
 	unsigned int no_chk_data_crc:1;
 	unsigned int bulk_read:1;
+	unsigned int default_compr:2;
 
 	struct mutex tnc_mutex;
 	struct ubifs_zbranch zroot;
@@ -1212,10 +1223,6 @@ struct ubifs_info {
 	int ileb_nxt;
 	struct rb_root old_idx;
 	int *bottom_up_buf;
-#ifdef CONFIG_UBIFS_FS_DEBUG
-	int new_ihead_lnum;
-	int new_ihead_offs;
-#endif
 
 	struct ubifs_mst_node *mst_node;
 	int mst_offs;
@@ -1237,7 +1244,6 @@ struct ubifs_info {
 	int main_lebs;
 	int main_first;
 	long long main_bytes;
-	int default_compr;
 
 	uint8_t key_hash_type;
 	uint32_t (*key_hash)(const char *str, int len);
@@ -1315,8 +1321,8 @@ struct ubifs_info {
 	void *sbuf;
 	struct list_head idx_gc;
 	int idx_gc_cnt;
-	volatile int gc_seq;
-	volatile int gced_lnum;
+	int gc_seq;
+	int gced_lnum;
 
 	struct list_head infos_list;
 	struct mutex umount_mutex;
@@ -1391,21 +1397,7 @@ struct ubifs_info {
 	struct ubifs_mount_opts mount_opts;
 
 #ifdef CONFIG_UBIFS_FS_DEBUG
-	void *dbg_buf;
-	struct ubifs_zbranch old_zroot;
-	int old_zroot_level;
-	unsigned long long old_zroot_sqnum;
-	int failure_mode;
-	int fail_delay;
-	unsigned long fail_timeout;
-	unsigned int fail_cnt;
-	unsigned int fail_cnt_max;
-	long long chk_lpt_sz;
-	long long chk_lpt_sz2;
-	long long chk_lpt_wastage;
-	int chk_lpt_lebs;
-	int new_nhead_lnum;
-	int new_nhead_offs;
+	struct ubifs_debug_info *dbg;
 #endif
 };
 
@@ -1505,7 +1497,7 @@ void ubifs_cancel_ino_op(struct ubifs_info *c, struct inode *inode,
 long long ubifs_get_free_space(struct ubifs_info *c);
 int ubifs_calc_min_idx_lebs(struct ubifs_info *c);
 void ubifs_convert_page_budget(struct ubifs_info *c);
-long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free);
+long long ubifs_reported_space(const struct ubifs_info *c, long long free);
 long long ubifs_calc_available(const struct ubifs_info *c, int min_idx_lebs);
 
 /* find.c */
@@ -1639,6 +1631,9 @@ void ubifs_add_lpt_dirt(struct ubifs_info *c, int lnum, int dirty);
 void ubifs_add_nnode_dirt(struct ubifs_info *c, struct ubifs_nnode *nnode);
 uint32_t ubifs_unpack_bits(uint8_t **addr, int *pos, int nrbits);
 struct ubifs_nnode *ubifs_first_nnode(struct ubifs_info *c, int *hght);
+/* Needed only in debugging code in lpt_commit.c */
+int ubifs_unpack_nnode(const struct ubifs_info *c, void *buf,
+		       struct ubifs_nnode *nnode);
 
 /* lpt_commit.c */
 int ubifs_lpt_start_commit(struct ubifs_info *c);
@@ -1714,7 +1709,7 @@ long ubifs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 
 /* compressor.c */
 int __init ubifs_compressors_init(void);
-void __exit ubifs_compressors_exit(void);
+void ubifs_compressors_exit(void);
 void ubifs_compress(const void *in_buf, int in_len, void *out_buf, int *out_len,
 		    int *compr_type);
 int ubifs_decompress(const void *buf, int len, void *out, int *out_len,
diff --git a/fs/xattr.c b/fs/xattr.c
index 468377e6653..237804cd6b5 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
@@ -175,7 +175,7 @@ vfs_listxattr(struct dentry *d, char *list, size_t size)
 	if (error)
 		return error;
 	error = -EOPNOTSUPP;
-	if (d->d_inode->i_op && d->d_inode->i_op->listxattr) {
+	if (d->d_inode->i_op->listxattr) {
 		error = d->d_inode->i_op->listxattr(d, list, size);
 	} else {
 		error = security_inode_listsecurity(d->d_inode, list, size);
diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index 737c9a42536..c3dc491fff8 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -85,13 +85,13 @@ xfs-y				+= xfs_alloc.o \
 				   xfs_trans_inode.o \
 				   xfs_trans_item.o \
 				   xfs_utils.o \
-				   xfs_vfsops.o \
 				   xfs_vnodeops.o \
 				   xfs_rw.o \
 				   xfs_dmops.o \
 				   xfs_qmops.o
 
-xfs-$(CONFIG_XFS_TRACE)		+= xfs_dir2_trace.o
+xfs-$(CONFIG_XFS_TRACE)		+= xfs_btree_trace.o \
+				   xfs_dir2_trace.o
 
 # Objects in linux/
 xfs-y				+= $(addprefix $(XFS_LINUX)/, \
@@ -106,7 +106,7 @@ xfs-y				+= $(addprefix $(XFS_LINUX)/, \
 				   xfs_iops.o \
 				   xfs_lrw.o \
 				   xfs_super.o \
-				   xfs_vnode.o \
+				   xfs_sync.o \
 				   xfs_xattr.o)
 
 # Objects in support/
diff --git a/fs/xfs/linux-2.6/sv.h b/fs/xfs/linux-2.6/sv.h
index 351a8f454bd..4dfc7c37081 100644
--- a/fs/xfs/linux-2.6/sv.h
+++ b/fs/xfs/linux-2.6/sv.h
@@ -32,23 +32,15 @@ typedef struct sv_s {
 	wait_queue_head_t waiters;
 } sv_t;
 
-#define SV_FIFO		0x0		/* sv_t is FIFO type */
-#define SV_LIFO		0x2		/* sv_t is LIFO type */
-#define SV_PRIO		0x4		/* sv_t is PRIO type */
-#define SV_KEYED	0x6		/* sv_t is KEYED type */
-#define SV_DEFAULT      SV_FIFO
-
-
-static inline void _sv_wait(sv_t *sv, spinlock_t *lock, int state,
-			     unsigned long timeout)
+static inline void _sv_wait(sv_t *sv, spinlock_t *lock)
 {
 	DECLARE_WAITQUEUE(wait, current);
 
 	add_wait_queue_exclusive(&sv->waiters, &wait);
-	__set_current_state(state);
+	__set_current_state(TASK_UNINTERRUPTIBLE);
 	spin_unlock(lock);
 
-	schedule_timeout(timeout);
+	schedule();
 
 	remove_wait_queue(&sv->waiters, &wait);
 }
@@ -58,13 +50,7 @@ static inline void _sv_wait(sv_t *sv, spinlock_t *lock, int state,
 #define sv_destroy(sv) \
 	/*NOTHING*/
 #define sv_wait(sv, pri, lock, s) \
-	_sv_wait(sv, lock, TASK_UNINTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT)
-#define sv_wait_sig(sv, pri, lock, s)   \
-	_sv_wait(sv, lock, TASK_INTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT)
-#define sv_timedwait(sv, pri, lock, s, svf, ts, rts) \
-	_sv_wait(sv, lock, TASK_UNINTERRUPTIBLE, timespec_to_jiffies(ts))
-#define sv_timedwait_sig(sv, pri, lock, s, svf, ts, rts) \
-	_sv_wait(sv, lock, TASK_INTERRUPTIBLE, timespec_to_jiffies(ts))
+	_sv_wait(sv, lock)
 #define sv_signal(sv) \
 	wake_up(&(sv)->waiters)
 #define sv_broadcast(sv) \
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index a44d68eb50b..de3a198f771 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -42,6 +42,40 @@
 #include <linux/pagevec.h>
 #include <linux/writeback.h>
 
+
+/*
+ * Prime number of hash buckets since address is used as the key.
+ */
+#define NVSYNC		37
+#define to_ioend_wq(v)	(&xfs_ioend_wq[((unsigned long)v) % NVSYNC])
+static wait_queue_head_t xfs_ioend_wq[NVSYNC];
+
+void __init
+xfs_ioend_init(void)
+{
+	int i;
+
+	for (i = 0; i < NVSYNC; i++)
+		init_waitqueue_head(&xfs_ioend_wq[i]);
+}
+
+void
+xfs_ioend_wait(
+	xfs_inode_t	*ip)
+{
+	wait_queue_head_t *wq = to_ioend_wq(ip);
+
+	wait_event(*wq, (atomic_read(&ip->i_iocount) == 0));
+}
+
+STATIC void
+xfs_ioend_wake(
+	xfs_inode_t	*ip)
+{
+	if (atomic_dec_and_test(&ip->i_iocount))
+		wake_up(to_ioend_wq(ip));
+}
+
 STATIC void
 xfs_count_page_state(
 	struct page		*page,
@@ -146,16 +180,25 @@ xfs_destroy_ioend(
 	xfs_ioend_t		*ioend)
 {
 	struct buffer_head	*bh, *next;
+	struct xfs_inode	*ip = XFS_I(ioend->io_inode);
 
 	for (bh = ioend->io_buffer_head; bh; bh = next) {
 		next = bh->b_private;
 		bh->b_end_io(bh, !ioend->io_error);
 	}
-	if (unlikely(ioend->io_error)) {
-		vn_ioerror(XFS_I(ioend->io_inode), ioend->io_error,
-				__FILE__,__LINE__);
+
+	/*
+	 * Volume managers supporting multiple paths can send back ENODEV
+	 * when the final path disappears.  In this case continuing to fill
+	 * the page cache with dirty data which cannot be written out is
+	 * evil, so prevent that.
+	 */
+	if (unlikely(ioend->io_error == -ENODEV)) {
+		xfs_do_force_shutdown(ip->i_mount, SHUTDOWN_DEVICE_REQ,
+				      __FILE__, __LINE__);
 	}
-	vn_iowake(XFS_I(ioend->io_inode));
+
+	xfs_ioend_wake(ip);
 	mempool_free(ioend, xfs_ioend_pool);
 }
 
@@ -191,7 +234,7 @@ xfs_setfilesize(
 		ip->i_d.di_size = isize;
 		ip->i_update_core = 1;
 		ip->i_update_size = 1;
-		mark_inode_dirty_sync(ioend->io_inode);
+		xfs_mark_inode_dirty_sync(ip);
 	}
 
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
@@ -317,14 +360,9 @@ xfs_map_blocks(
 	xfs_iomap_t		*mapp,
 	int			flags)
 {
-	xfs_inode_t		*ip = XFS_I(inode);
-	int			error, nmaps = 1;
-
-	error = xfs_iomap(ip, offset, count,
-				flags, mapp, &nmaps);
-	if (!error && (flags & (BMAPI_WRITE|BMAPI_ALLOCATE)))
-		xfs_iflags_set(ip, XFS_IMODIFIED);
-	return -error;
+	int			nmaps = 1;
+
+	return -xfs_iomap(XFS_I(inode), offset, count, flags, mapp, &nmaps);
 }
 
 STATIC_INLINE int
@@ -512,7 +550,7 @@ xfs_cancel_ioend(
 			unlock_buffer(bh);
 		} while ((bh = next_bh) != NULL);
 
-		vn_iowake(XFS_I(ioend->io_inode));
+		xfs_ioend_wake(XFS_I(ioend->io_inode));
 		mempool_free(ioend, xfs_ioend_pool);
 	} while ((ioend = next) != NULL);
 }
diff --git a/fs/xfs/linux-2.6/xfs_aops.h b/fs/xfs/linux-2.6/xfs_aops.h
index 3ba0631a381..7b26f5ff969 100644
--- a/fs/xfs/linux-2.6/xfs_aops.h
+++ b/fs/xfs/linux-2.6/xfs_aops.h
@@ -43,4 +43,7 @@ typedef struct xfs_ioend {
 extern const struct address_space_operations xfs_address_space_operations;
 extern int xfs_get_blocks(struct inode *, sector_t, struct buffer_head *, int);
 
+extern void xfs_ioend_init(void);
+extern void xfs_ioend_wait(struct xfs_inode *);
+
 #endif /* __XFS_AOPS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index 36d5fcd3f59..cb329edc925 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -630,6 +630,29 @@ xfs_buf_get_flags(
 	return NULL;
 }
 
+STATIC int
+_xfs_buf_read(
+	xfs_buf_t		*bp,
+	xfs_buf_flags_t		flags)
+{
+	int			status;
+
+	XB_TRACE(bp, "_xfs_buf_read", (unsigned long)flags);
+
+	ASSERT(!(flags & (XBF_DELWRI|XBF_WRITE)));
+	ASSERT(bp->b_bn != XFS_BUF_DADDR_NULL);
+
+	bp->b_flags &= ~(XBF_WRITE | XBF_ASYNC | XBF_DELWRI | \
+			XBF_READ_AHEAD | _XBF_RUN_QUEUES);
+	bp->b_flags |= flags & (XBF_READ | XBF_ASYNC | \
+			XBF_READ_AHEAD | _XBF_RUN_QUEUES);
+
+	status = xfs_buf_iorequest(bp);
+	if (!status && !(flags & XBF_ASYNC))
+		status = xfs_buf_iowait(bp);
+	return status;
+}
+
 xfs_buf_t *
 xfs_buf_read_flags(
 	xfs_buftarg_t		*target,
@@ -646,7 +669,7 @@ xfs_buf_read_flags(
 		if (!XFS_BUF_ISDONE(bp)) {
 			XB_TRACE(bp, "read", (unsigned long)flags);
 			XFS_STATS_INC(xb_get_read);
-			xfs_buf_iostart(bp, flags);
+			_xfs_buf_read(bp, flags);
 		} else if (flags & XBF_ASYNC) {
 			XB_TRACE(bp, "read_async", (unsigned long)flags);
 			/*
@@ -1048,50 +1071,39 @@ xfs_buf_ioerror(
 	XB_TRACE(bp, "ioerror", (unsigned long)error);
 }
 
-/*
- *	Initiate I/O on a buffer, based on the flags supplied.
- *	The b_iodone routine in the buffer supplied will only be called
- *	when all of the subsidiary I/O requests, if any, have been completed.
- */
 int
-xfs_buf_iostart(
-	xfs_buf_t		*bp,
-	xfs_buf_flags_t		flags)
+xfs_bawrite(
+	void			*mp,
+	struct xfs_buf		*bp)
 {
-	int			status = 0;
+	XB_TRACE(bp, "bawrite", 0);
 
-	XB_TRACE(bp, "iostart", (unsigned long)flags);
+	ASSERT(bp->b_bn != XFS_BUF_DADDR_NULL);
 
-	if (flags & XBF_DELWRI) {
-		bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_ASYNC);
-		bp->b_flags |= flags & (XBF_DELWRI | XBF_ASYNC);
-		xfs_buf_delwri_queue(bp, 1);
-		return 0;
-	}
+	xfs_buf_delwri_dequeue(bp);
 
-	bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_ASYNC | XBF_DELWRI | \
-			XBF_READ_AHEAD | _XBF_RUN_QUEUES);
-	bp->b_flags |= flags & (XBF_READ | XBF_WRITE | XBF_ASYNC | \
-			XBF_READ_AHEAD | _XBF_RUN_QUEUES);
+	bp->b_flags &= ~(XBF_READ | XBF_DELWRI | XBF_READ_AHEAD);
+	bp->b_flags |= (XBF_WRITE | XBF_ASYNC | _XBF_RUN_QUEUES);
+
+	bp->b_mount = mp;
+	bp->b_strat = xfs_bdstrat_cb;
+	return xfs_bdstrat_cb(bp);
+}
 
-	BUG_ON(bp->b_bn == XFS_BUF_DADDR_NULL);
+void
+xfs_bdwrite(
+	void			*mp,
+	struct xfs_buf		*bp)
+{
+	XB_TRACE(bp, "bdwrite", 0);
 
-	/* For writes allow an alternate strategy routine to precede
-	 * the actual I/O request (which may not be issued at all in
-	 * a shutdown situation, for example).
-	 */
-	status = (flags & XBF_WRITE) ?
-		xfs_buf_iostrategy(bp) : xfs_buf_iorequest(bp);
+	bp->b_strat = xfs_bdstrat_cb;
+	bp->b_mount = mp;
 
-	/* Wait for I/O if we are not an async request.
-	 * Note: async I/O request completion will release the buffer,
-	 * and that can already be done by this point.  So using the
-	 * buffer pointer from here on, after async I/O, is invalid.
-	 */
-	if (!status && !(flags & XBF_ASYNC))
-		status = xfs_buf_iowait(bp);
+	bp->b_flags &= ~XBF_READ;
+	bp->b_flags |= (XBF_DELWRI | XBF_ASYNC);
 
-	return status;
+	xfs_buf_delwri_queue(bp, 1);
 }
 
 STATIC_INLINE void
@@ -1114,8 +1126,7 @@ xfs_buf_bio_end_io(
 	unsigned int		blocksize = bp->b_target->bt_bsize;
 	struct bio_vec		*bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
 
-	if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
-		bp->b_error = EIO;
+	xfs_buf_ioerror(bp, -error);
 
 	do {
 		struct page	*page = bvec->bv_page;
diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h
index 456519a088c..288ae7c4c80 100644
--- a/fs/xfs/linux-2.6/xfs_buf.h
+++ b/fs/xfs/linux-2.6/xfs_buf.h
@@ -168,7 +168,7 @@ typedef struct xfs_buf {
 	struct completion	b_iowait;	/* queue for I/O waiters */
 	void			*b_fspriv;
 	void			*b_fspriv2;
-	void			*b_fspriv3;
+	struct xfs_mount	*b_mount;
 	unsigned short		b_error;	/* error code on I/O */
 	unsigned int		b_page_count;	/* size of page array */
 	unsigned int		b_offset;	/* page offset in first page */
@@ -214,9 +214,10 @@ extern void xfs_buf_lock(xfs_buf_t *);
 extern void xfs_buf_unlock(xfs_buf_t *);
 
 /* Buffer Read and Write Routines */
+extern int xfs_bawrite(void *mp, xfs_buf_t *bp);
+extern void xfs_bdwrite(void *mp, xfs_buf_t *bp);
 extern void xfs_buf_ioend(xfs_buf_t *,	int);
 extern void xfs_buf_ioerror(xfs_buf_t *, int);
-extern int xfs_buf_iostart(xfs_buf_t *, xfs_buf_flags_t);
 extern int xfs_buf_iorequest(xfs_buf_t *);
 extern int xfs_buf_iowait(xfs_buf_t *);
 extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, xfs_caddr_t,
@@ -311,10 +312,6 @@ extern void xfs_buf_trace(xfs_buf_t *, char *, void *, void *);
 #define XFS_BUF_UNORDERED(bp)	((bp)->b_flags &= ~XBF_ORDERED)
 #define XFS_BUF_ISORDERED(bp)	((bp)->b_flags & XBF_ORDERED)
 
-#define XFS_BUF_SHUT(bp)	do { } while (0)
-#define XFS_BUF_UNSHUT(bp)	do { } while (0)
-#define XFS_BUF_ISSHUT(bp)	(0)
-
 #define XFS_BUF_HOLD(bp)	xfs_buf_hold(bp)
 #define XFS_BUF_READ(bp)	((bp)->b_flags |= XBF_READ)
 #define XFS_BUF_UNREAD(bp)	((bp)->b_flags &= ~XBF_READ)
@@ -334,8 +331,6 @@ extern void xfs_buf_trace(xfs_buf_t *, char *, void *, void *);
 #define XFS_BUF_SET_FSPRIVATE(bp, val)		((bp)->b_fspriv = (void*)(val))
 #define XFS_BUF_FSPRIVATE2(bp, type)		((type)(bp)->b_fspriv2)
 #define XFS_BUF_SET_FSPRIVATE2(bp, val)		((bp)->b_fspriv2 = (void*)(val))
-#define XFS_BUF_FSPRIVATE3(bp, type)		((type)(bp)->b_fspriv3)
-#define XFS_BUF_SET_FSPRIVATE3(bp, val)		((bp)->b_fspriv3 = (void*)(val))
 #define XFS_BUF_SET_START(bp)			do { } while (0)
 #define XFS_BUF_SET_BRELSE_FUNC(bp, func)	((bp)->b_relse = (func))
 
@@ -366,14 +361,6 @@ extern void xfs_buf_trace(xfs_buf_t *, char *, void *, void *);
 #define XFS_BUF_TARGET(bp)		((bp)->b_target)
 #define XFS_BUFTARG_NAME(target)	xfs_buf_target_name(target)
 
-static inline int xfs_bawrite(void *mp, xfs_buf_t *bp)
-{
-	bp->b_fspriv3 = mp;
-	bp->b_strat = xfs_bdstrat_cb;
-	xfs_buf_delwri_dequeue(bp);
-	return xfs_buf_iostart(bp, XBF_WRITE | XBF_ASYNC | _XBF_RUN_QUEUES);
-}
-
 static inline void xfs_buf_relse(xfs_buf_t *bp)
 {
 	if (!bp->b_relse)
@@ -414,17 +401,6 @@ static inline int XFS_bwrite(xfs_buf_t *bp)
 	return error;
 }
 
-/*
- * No error can be returned from xfs_buf_iostart for delwri
- * buffers as they are queued and no I/O is issued.
- */
-static inline void xfs_bdwrite(void *mp, xfs_buf_t *bp)
-{
-	bp->b_strat = xfs_bdstrat_cb;
-	bp->b_fspriv3 = mp;
-	(void)xfs_buf_iostart(bp, XBF_DELWRI | XBF_ASYNC);
-}
-
 #define XFS_bdstrat(bp) xfs_buf_iorequest(bp)
 
 #define xfs_iowait(bp)	xfs_buf_iowait(bp)
diff --git a/fs/xfs/linux-2.6/xfs_cred.h b/fs/xfs/linux-2.6/xfs_cred.h
index 8c022cd0ad6..55bddf3b609 100644
--- a/fs/xfs/linux-2.6/xfs_cred.h
+++ b/fs/xfs/linux-2.6/xfs_cred.h
@@ -25,12 +25,4 @@
  */
 typedef const struct cred cred_t;
 
-extern cred_t *sys_cred;
-
-/* this is a hack.. (assumes sys_cred is the only cred_t in the system) */
-static inline int capable_cred(cred_t *cr, int cid)
-{
-	return (cr == sys_cred) ? 1 : capable(cid);
-}
-
 #endif  /* __XFS_CRED_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_export.c b/fs/xfs/linux-2.6/xfs_export.c
index 7f7abec25e1..595751f7835 100644
--- a/fs/xfs/linux-2.6/xfs_export.c
+++ b/fs/xfs/linux-2.6/xfs_export.c
@@ -29,7 +29,6 @@
 #include "xfs_vnodeops.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_inode.h"
-#include "xfs_vfsops.h"
 
 /*
  * Note that we only accept fileids which are long enough rather than allow
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index 3fee790f138..e14c4e3aea0 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -36,89 +36,54 @@
 #include "xfs_inode.h"
 #include "xfs_error.h"
 #include "xfs_rw.h"
-#include "xfs_ioctl32.h"
 #include "xfs_vnodeops.h"
+#include "xfs_da_btree.h"
+#include "xfs_ioctl.h"
 
 #include <linux/dcache.h>
 #include <linux/smp_lock.h>
 
 static struct vm_operations_struct xfs_file_vm_ops;
 
-STATIC_INLINE ssize_t
-__xfs_file_read(
+STATIC ssize_t
+xfs_file_aio_read(
 	struct kiocb		*iocb,
 	const struct iovec	*iov,
 	unsigned long		nr_segs,
-	int			ioflags,
 	loff_t			pos)
 {
 	struct file		*file = iocb->ki_filp;
+	int			ioflags = IO_ISAIO;
 
 	BUG_ON(iocb->ki_pos != pos);
 	if (unlikely(file->f_flags & O_DIRECT))
 		ioflags |= IO_ISDIRECT;
+	if (file->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
 	return xfs_read(XFS_I(file->f_path.dentry->d_inode), iocb, iov,
 				nr_segs, &iocb->ki_pos, ioflags);
 }
 
 STATIC ssize_t
-xfs_file_aio_read(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_read(iocb, iov, nr_segs, IO_ISAIO, pos);
-}
-
-STATIC ssize_t
-xfs_file_aio_read_invis(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_read(iocb, iov, nr_segs, IO_ISAIO|IO_INVIS, pos);
-}
-
-STATIC_INLINE ssize_t
-__xfs_file_write(
+xfs_file_aio_write(
 	struct kiocb		*iocb,
 	const struct iovec	*iov,
 	unsigned long		nr_segs,
-	int			ioflags,
 	loff_t			pos)
 {
-	struct file	*file = iocb->ki_filp;
+	struct file		*file = iocb->ki_filp;
+	int			ioflags = IO_ISAIO;
 
 	BUG_ON(iocb->ki_pos != pos);
 	if (unlikely(file->f_flags & O_DIRECT))
 		ioflags |= IO_ISDIRECT;
+	if (file->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
 	return xfs_write(XFS_I(file->f_mapping->host), iocb, iov, nr_segs,
 				&iocb->ki_pos, ioflags);
 }
 
 STATIC ssize_t
-xfs_file_aio_write(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_write(iocb, iov, nr_segs, IO_ISAIO, pos);
-}
-
-STATIC ssize_t
-xfs_file_aio_write_invis(
-	struct kiocb		*iocb,
-	const struct iovec	*iov,
-	unsigned long		nr_segs,
-	loff_t			pos)
-{
-	return __xfs_file_write(iocb, iov, nr_segs, IO_ISAIO|IO_INVIS, pos);
-}
-
-STATIC ssize_t
 xfs_file_splice_read(
 	struct file		*infilp,
 	loff_t			*ppos,
@@ -126,20 +91,13 @@ xfs_file_splice_read(
 	size_t			len,
 	unsigned int		flags)
 {
-	return xfs_splice_read(XFS_I(infilp->f_path.dentry->d_inode),
-				   infilp, ppos, pipe, len, flags, 0);
-}
+	int			ioflags = 0;
+
+	if (infilp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
 
-STATIC ssize_t
-xfs_file_splice_read_invis(
-	struct file		*infilp,
-	loff_t			*ppos,
-	struct pipe_inode_info	*pipe,
-	size_t			len,
-	unsigned int		flags)
-{
 	return xfs_splice_read(XFS_I(infilp->f_path.dentry->d_inode),
-				   infilp, ppos, pipe, len, flags, IO_INVIS);
+				   infilp, ppos, pipe, len, flags, ioflags);
 }
 
 STATIC ssize_t
@@ -150,30 +108,49 @@ xfs_file_splice_write(
 	size_t			len,
 	unsigned int		flags)
 {
-	return xfs_splice_write(XFS_I(outfilp->f_path.dentry->d_inode),
-				    pipe, outfilp, ppos, len, flags, 0);
-}
+	int			ioflags = 0;
+
+	if (outfilp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
 
-STATIC ssize_t
-xfs_file_splice_write_invis(
-	struct pipe_inode_info	*pipe,
-	struct file		*outfilp,
-	loff_t			*ppos,
-	size_t			len,
-	unsigned int		flags)
-{
 	return xfs_splice_write(XFS_I(outfilp->f_path.dentry->d_inode),
-				    pipe, outfilp, ppos, len, flags, IO_INVIS);
+				    pipe, outfilp, ppos, len, flags, ioflags);
 }
 
 STATIC int
 xfs_file_open(
 	struct inode	*inode,
-	struct file	*filp)
+	struct file	*file)
 {
-	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
+	if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
 		return -EFBIG;
-	return -xfs_open(XFS_I(inode));
+	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
+		return -EIO;
+	return 0;
+}
+
+STATIC int
+xfs_dir_open(
+	struct inode	*inode,
+	struct file	*file)
+{
+	struct xfs_inode *ip = XFS_I(inode);
+	int		mode;
+	int		error;
+
+	error = xfs_file_open(inode, file);
+	if (error)
+		return error;
+
+	/*
+	 * If there are any blocks, read-ahead block 0 as we're almost
+	 * certain to have the next operation be a read there.
+	 */
+	mode = xfs_ilock_map_shared(ip);
+	if (ip->i_d.di_nextents > 0)
+		xfs_da_reada_buf(NULL, ip, 0, XFS_DATA_FORK);
+	xfs_iunlock(ip, mode);
+	return 0;
 }
 
 STATIC int
@@ -227,7 +204,7 @@ xfs_file_readdir(
 	 * point we can change the ->readdir prototype to include the
 	 * buffer size.
 	 */
-	bufsize = (size_t)min_t(loff_t, PAGE_SIZE, inode->i_size);
+	bufsize = (size_t)min_t(loff_t, PAGE_SIZE, ip->i_d.di_size);
 
 	error = xfs_readdir(ip, dirent, bufsize,
 				(xfs_off_t *)&filp->f_pos, filldir);
@@ -248,48 +225,6 @@ xfs_file_mmap(
 	return 0;
 }
 
-STATIC long
-xfs_file_ioctl(
-	struct file	*filp,
-	unsigned int	cmd,
-	unsigned long	p)
-{
-	int		error;
-	struct inode	*inode = filp->f_path.dentry->d_inode;
-
-	error = xfs_ioctl(XFS_I(inode), filp, 0, cmd, (void __user *)p);
-	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
-
-	/* NOTE:  some of the ioctl's return positive #'s as a
-	 *	  byte count indicating success, such as
-	 *	  readlink_by_handle.  So we don't "sign flip"
-	 *	  like most other routines.  This means true
-	 *	  errors need to be returned as a negative value.
-	 */
-	return error;
-}
-
-STATIC long
-xfs_file_ioctl_invis(
-	struct file	*filp,
-	unsigned int	cmd,
-	unsigned long	p)
-{
-	int		error;
-	struct inode	*inode = filp->f_path.dentry->d_inode;
-
-	error = xfs_ioctl(XFS_I(inode), filp, IO_INVIS, cmd, (void __user *)p);
-	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
-
-	/* NOTE:  some of the ioctl's return positive #'s as a
-	 *	  byte count indicating success, such as
-	 *	  readlink_by_handle.  So we don't "sign flip"
-	 *	  like most other routines.  This means true
-	 *	  errors need to be returned as a negative value.
-	 */
-	return error;
-}
-
 /*
  * mmap()d file has taken write protection fault and is being made
  * writable. We can set the page state up correctly for a writable
@@ -325,26 +260,8 @@ const struct file_operations xfs_file_operations = {
 #endif
 };
 
-const struct file_operations xfs_invis_file_operations = {
-	.llseek		= generic_file_llseek,
-	.read		= do_sync_read,
-	.write		= do_sync_write,
-	.aio_read	= xfs_file_aio_read_invis,
-	.aio_write	= xfs_file_aio_write_invis,
-	.splice_read	= xfs_file_splice_read_invis,
-	.splice_write	= xfs_file_splice_write_invis,
-	.unlocked_ioctl	= xfs_file_ioctl_invis,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl	= xfs_file_compat_invis_ioctl,
-#endif
-	.mmap		= xfs_file_mmap,
-	.open		= xfs_file_open,
-	.release	= xfs_file_release,
-	.fsync		= xfs_file_fsync,
-};
-
-
 const struct file_operations xfs_dir_file_operations = {
+	.open		= xfs_dir_open,
 	.read		= generic_read_dir,
 	.readdir	= xfs_file_readdir,
 	.llseek		= generic_file_llseek,
diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c
index 36caa6d957d..5aeb7777696 100644
--- a/fs/xfs/linux-2.6/xfs_fs_subr.c
+++ b/fs/xfs/linux-2.6/xfs_fs_subr.c
@@ -24,6 +24,10 @@ int  fs_noerr(void) { return 0; }
 int  fs_nosys(void) { return ENOSYS; }
 void fs_noval(void) { return; }
 
+/*
+ * note: all filemap functions return negative error codes. These
+ * need to be inverted before returning to the xfs core functions.
+ */
 void
 xfs_tosspages(
 	xfs_inode_t	*ip,
@@ -53,7 +57,7 @@ xfs_flushinval_pages(
 		if (!ret)
 			truncate_inode_pages(mapping, first);
 	}
-	return ret;
+	return -ret;
 }
 
 int
@@ -72,10 +76,23 @@ xfs_flush_pages(
 		xfs_iflags_clear(ip, XFS_ITRUNCATED);
 		ret = filemap_fdatawrite(mapping);
 		if (flags & XFS_B_ASYNC)
-			return ret;
+			return -ret;
 		ret2 = filemap_fdatawait(mapping);
 		if (!ret)
 			ret = ret2;
 	}
-	return ret;
+	return -ret;
+}
+
+int
+xfs_wait_on_pages(
+	xfs_inode_t	*ip,
+	xfs_off_t	first,
+	xfs_off_t	last)
+{
+	struct address_space *mapping = VFS_I(ip)->i_mapping;
+
+	if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK))
+		return -filemap_fdatawait(mapping);
+	return 0;
 }
diff --git a/fs/xfs/linux-2.6/xfs_globals.c b/fs/xfs/linux-2.6/xfs_globals.c
index ef90e64641e..2ae8b1ccb02 100644
--- a/fs/xfs/linux-2.6/xfs_globals.c
+++ b/fs/xfs/linux-2.6/xfs_globals.c
@@ -26,7 +26,6 @@
  */
 xfs_param_t xfs_params = {
 			  /*	MIN		DFLT		MAX	*/
-	.restrict_chown	= {	0,		1,		1	},
 	.sgid_inherit	= {	0,		0,		1	},
 	.symlink_mode	= {	0,		0,		1	},
 	.panic_mask	= {	0,		0,		255	},
@@ -43,10 +42,3 @@ xfs_param_t xfs_params = {
 	.inherit_nodfrg	= {	0,		1,		1	},
 	.fstrm_timer	= {	1,		30*100,		3600*100},
 };
-
-/*
- * Global system credential structure.
- */
-static cred_t sys_cred_val;
-cred_t *sys_cred = &sys_cred_val;
-
diff --git a/fs/xfs/linux-2.6/xfs_globals.h b/fs/xfs/linux-2.6/xfs_globals.h
index 6eda8a3eb6f..69f71caf061 100644
--- a/fs/xfs/linux-2.6/xfs_globals.h
+++ b/fs/xfs/linux-2.6/xfs_globals.h
@@ -19,6 +19,5 @@
 #define __XFS_GLOBALS_H__
 
 extern uint64_t	xfs_panic_mask;		/* set to cause more panics */
-extern cred_t *sys_cred;
 
 #endif	/* __XFS_GLOBALS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c
index 281cbd5a25c..e5be1e0be80 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl.c
@@ -68,26 +68,22 @@
  * XFS_IOC_PATH_TO_HANDLE
  *    returns full handle for a path
  */
-STATIC int
+int
 xfs_find_handle(
 	unsigned int		cmd,
-	void			__user *arg)
+	xfs_fsop_handlereq_t	*hreq)
 {
 	int			hsize;
 	xfs_handle_t		handle;
-	xfs_fsop_handlereq_t	hreq;
 	struct inode		*inode;
 
-	if (copy_from_user(&hreq, arg, sizeof(hreq)))
-		return -XFS_ERROR(EFAULT);
-
 	memset((char *)&handle, 0, sizeof(handle));
 
 	switch (cmd) {
 	case XFS_IOC_PATH_TO_FSHANDLE:
 	case XFS_IOC_PATH_TO_HANDLE: {
 		struct path path;
-		int error = user_lpath((const char __user *)hreq.path, &path);
+		int error = user_lpath((const char __user *)hreq->path, &path);
 		if (error)
 			return error;
 
@@ -101,7 +97,7 @@ xfs_find_handle(
 	case XFS_IOC_FD_TO_HANDLE: {
 		struct file	*file;
 
-		file = fget(hreq.fd);
+		file = fget(hreq->fd);
 		if (!file)
 		    return -EBADF;
 
@@ -158,8 +154,8 @@ xfs_find_handle(
 	}
 
 	/* now copy our handle into the user buffer & write out the size */
-	if (copy_to_user(hreq.ohandle, &handle, hsize) ||
-	    copy_to_user(hreq.ohandlen, &hsize, sizeof(__s32))) {
+	if (copy_to_user(hreq->ohandle, &handle, hsize) ||
+	    copy_to_user(hreq->ohandlen, &hsize, sizeof(__s32))) {
 		iput(inode);
 		return -XFS_ERROR(EFAULT);
 	}
@@ -249,10 +245,10 @@ xfs_vget_fsop_handlereq(
 	return 0;
 }
 
-STATIC int
+int
 xfs_open_by_handle(
 	xfs_mount_t		*mp,
-	void			__user *arg,
+	xfs_fsop_handlereq_t	*hreq,
 	struct file		*parfilp,
 	struct inode		*parinode)
 {
@@ -263,14 +259,11 @@ xfs_open_by_handle(
 	struct file		*filp;
 	struct inode		*inode;
 	struct dentry		*dentry;
-	xfs_fsop_handlereq_t	hreq;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
 
-	error = xfs_vget_fsop_handlereq(mp, parinode, &hreq, &inode);
+	error = xfs_vget_fsop_handlereq(mp, parinode, hreq, &inode);
 	if (error)
 		return -error;
 
@@ -281,10 +274,10 @@ xfs_open_by_handle(
 	}
 
 #if BITS_PER_LONG != 32
-	hreq.oflags |= O_LARGEFILE;
+	hreq->oflags |= O_LARGEFILE;
 #endif
 	/* Put open permission in namei format. */
-	permflag = hreq.oflags;
+	permflag = hreq->oflags;
 	if ((permflag+1) & O_ACCMODE)
 		permflag++;
 	if (permflag & O_TRUNC)
@@ -322,15 +315,16 @@ xfs_open_by_handle(
 	mntget(parfilp->f_path.mnt);
 
 	/* Create file pointer. */
-	filp = dentry_open(dentry, parfilp->f_path.mnt, hreq.oflags, cred);
+	filp = dentry_open(dentry, parfilp->f_path.mnt, hreq->oflags, cred);
 	if (IS_ERR(filp)) {
 		put_unused_fd(new_fd);
 		return -XFS_ERROR(-PTR_ERR(filp));
 	}
+
 	if (inode->i_mode & S_IFREG) {
 		/* invisible operation should not change atime */
 		filp->f_flags |= O_NOATIME;
-		filp->f_op = &xfs_invis_file_operations;
+		filp->f_mode |= FMODE_NOCMTIME;
 	}
 
 	fd_install(new_fd, filp);
@@ -363,24 +357,21 @@ do_readlink(
 }
 
 
-STATIC int
+int
 xfs_readlink_by_handle(
 	xfs_mount_t		*mp,
-	void			__user *arg,
+	xfs_fsop_handlereq_t	*hreq,
 	struct inode		*parinode)
 {
 	struct inode		*inode;
-	xfs_fsop_handlereq_t	hreq;
 	__u32			olen;
 	void			*link;
 	int			error;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -XFS_ERROR(EPERM);
-	if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
-		return -XFS_ERROR(EFAULT);
 
-	error = xfs_vget_fsop_handlereq(mp, parinode, &hreq, &inode);
+	error = xfs_vget_fsop_handlereq(mp, parinode, hreq, &inode);
 	if (error)
 		return -error;
 
@@ -390,7 +381,7 @@ xfs_readlink_by_handle(
 		goto out_iput;
 	}
 
-	if (copy_from_user(&olen, hreq.ohandlen, sizeof(__u32))) {
+	if (copy_from_user(&olen, hreq->ohandlen, sizeof(__u32))) {
 		error = -XFS_ERROR(EFAULT);
 		goto out_iput;
 	}
@@ -402,7 +393,7 @@ xfs_readlink_by_handle(
 	error = -xfs_readlink(XFS_I(inode), link);
 	if (error)
 		goto out_kfree;
-	error = do_readlink(hreq.ohandle, olen, link);
+	error = do_readlink(hreq->ohandle, olen, link);
 	if (error)
 		goto out_kfree;
 
@@ -501,7 +492,7 @@ xfs_attrlist_by_handle(
 	return -error;
 }
 
-STATIC int
+int
 xfs_attrmulti_attr_get(
 	struct inode		*inode,
 	char			*name,
@@ -530,7 +521,7 @@ xfs_attrmulti_attr_get(
 	return error;
 }
 
-STATIC int
+int
 xfs_attrmulti_attr_set(
 	struct inode		*inode,
 	char			*name,
@@ -560,7 +551,7 @@ xfs_attrmulti_attr_set(
 	return error;
 }
 
-STATIC int
+int
 xfs_attrmulti_attr_remove(
 	struct inode		*inode,
 	char			*name,
@@ -662,19 +653,26 @@ xfs_attrmulti_by_handle(
 	return -error;
 }
 
-STATIC int
+int
 xfs_ioc_space(
 	struct xfs_inode	*ip,
 	struct inode		*inode,
 	struct file		*filp,
 	int			ioflags,
 	unsigned int		cmd,
-	void			__user *arg)
+	xfs_flock64_t		*bf)
 {
-	xfs_flock64_t		bf;
 	int			attr_flags = 0;
 	int			error;
 
+	/*
+	 * Only allow the sys admin to reserve space unless
+	 * unwritten extents are enabled.
+	 */
+	if (!xfs_sb_version_hasextflgbit(&ip->i_mount->m_sb) &&
+	    !capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+
 	if (inode->i_flags & (S_IMMUTABLE|S_APPEND))
 		return -XFS_ERROR(EPERM);
 
@@ -684,16 +682,12 @@ xfs_ioc_space(
 	if (!S_ISREG(inode->i_mode))
 		return -XFS_ERROR(EINVAL);
 
-	if (copy_from_user(&bf, arg, sizeof(bf)))
-		return -XFS_ERROR(EFAULT);
-
 	if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
 		attr_flags |= XFS_ATTR_NONBLOCK;
 	if (ioflags & IO_INVIS)
 		attr_flags |= XFS_ATTR_DMI;
 
-	error = xfs_change_file_space(ip, cmd, &bf, filp->f_pos,
-					      NULL, attr_flags);
+	error = xfs_change_file_space(ip, cmd, bf, filp->f_pos, attr_flags);
 	return -error;
 }
 
@@ -1105,10 +1099,6 @@ xfs_ioctl_setattr(
 
 	/*
 	 * Change file ownership.  Must be the owner or privileged.
-	 * If the system was configured with the "restricted_chown"
-	 * option, the owner is not permitted to give away the file,
-	 * and can change the group id only to a group of which he
-	 * or she is a member.
 	 */
 	if (mask & FSX_PROJID) {
 		/*
@@ -1137,7 +1127,7 @@ xfs_ioctl_setattr(
 			 * the superblock version number since projids didn't
 			 * exist before DINODE_VERSION_2 and SB_VERSION_NLINK.
 			 */
-			if (ip->i_d.di_version == XFS_DINODE_VERSION_1)
+			if (ip->i_d.di_version == 1)
 				xfs_bump_ino_vers2(tp, ip);
 		}
 
@@ -1256,43 +1246,67 @@ xfs_ioc_setxflags(
 }
 
 STATIC int
+xfs_getbmap_format(void **ap, struct getbmapx *bmv, int *full)
+{
+	struct getbmap __user	*base = *ap;
+
+	/* copy only getbmap portion (not getbmapx) */
+	if (copy_to_user(base, bmv, sizeof(struct getbmap)))
+		return XFS_ERROR(EFAULT);
+
+	*ap += sizeof(struct getbmap);
+	return 0;
+}
+
+STATIC int
 xfs_ioc_getbmap(
 	struct xfs_inode	*ip,
 	int			ioflags,
 	unsigned int		cmd,
 	void			__user *arg)
 {
-	struct getbmap		bm;
-	int			iflags;
+	struct getbmapx		bmx;
 	int			error;
 
-	if (copy_from_user(&bm, arg, sizeof(bm)))
+	if (copy_from_user(&bmx, arg, sizeof(struct getbmapx)))
 		return -XFS_ERROR(EFAULT);
 
-	if (bm.bmv_count < 2)
+	if (bmx.bmv_count < 2)
 		return -XFS_ERROR(EINVAL);
 
-	iflags = (cmd == XFS_IOC_GETBMAPA ? BMV_IF_ATTRFORK : 0);
+	bmx.bmv_iflags = (cmd == XFS_IOC_GETBMAPA ? BMV_IF_ATTRFORK : 0);
 	if (ioflags & IO_INVIS)
-		iflags |= BMV_IF_NO_DMAPI_READ;
+		bmx.bmv_iflags |= BMV_IF_NO_DMAPI_READ;
 
-	error = xfs_getbmap(ip, &bm, (struct getbmap __user *)arg+1, iflags);
+	error = xfs_getbmap(ip, &bmx, xfs_getbmap_format,
+			    (struct getbmap *)arg+1);
 	if (error)
 		return -error;
 
-	if (copy_to_user(arg, &bm, sizeof(bm)))
+	/* copy back header - only size of getbmap */
+	if (copy_to_user(arg, &bmx, sizeof(struct getbmap)))
 		return -XFS_ERROR(EFAULT);
 	return 0;
 }
 
 STATIC int
+xfs_getbmapx_format(void **ap, struct getbmapx *bmv, int *full)
+{
+	struct getbmapx __user	*base = *ap;
+
+	if (copy_to_user(base, bmv, sizeof(struct getbmapx)))
+		return XFS_ERROR(EFAULT);
+
+	*ap += sizeof(struct getbmapx);
+	return 0;
+}
+
+STATIC int
 xfs_ioc_getbmapx(
 	struct xfs_inode	*ip,
 	void			__user *arg)
 {
 	struct getbmapx		bmx;
-	struct getbmap		bm;
-	int			iflags;
 	int			error;
 
 	if (copy_from_user(&bmx, arg, sizeof(bmx)))
@@ -1301,46 +1315,46 @@ xfs_ioc_getbmapx(
 	if (bmx.bmv_count < 2)
 		return -XFS_ERROR(EINVAL);
 
-	/*
-	 * Map input getbmapx structure to a getbmap
-	 * structure for xfs_getbmap.
-	 */
-	GETBMAP_CONVERT(bmx, bm);
-
-	iflags = bmx.bmv_iflags;
-
-	if (iflags & (~BMV_IF_VALID))
+	if (bmx.bmv_iflags & (~BMV_IF_VALID))
 		return -XFS_ERROR(EINVAL);
 
-	iflags |= BMV_IF_EXTENDED;
-
-	error = xfs_getbmap(ip, &bm, (struct getbmapx __user *)arg+1, iflags);
+	error = xfs_getbmap(ip, &bmx, xfs_getbmapx_format,
+			    (struct getbmapx *)arg+1);
 	if (error)
 		return -error;
 
-	GETBMAP_CONVERT(bm, bmx);
-
-	if (copy_to_user(arg, &bmx, sizeof(bmx)))
+	/* copy back header */
+	if (copy_to_user(arg, &bmx, sizeof(struct getbmapx)))
 		return -XFS_ERROR(EFAULT);
 
 	return 0;
 }
 
-int
-xfs_ioctl(
-	xfs_inode_t		*ip,
+/*
+ * Note: some of the ioctl's return positive numbers as a
+ * byte count indicating success, such as readlink_by_handle.
+ * So we don't "sign flip" like most other routines.  This means
+ * true errors need to be returned as a negative value.
+ */
+long
+xfs_file_ioctl(
 	struct file		*filp,
-	int			ioflags,
 	unsigned int		cmd,
-	void			__user *arg)
+	unsigned long		p)
 {
 	struct inode		*inode = filp->f_path.dentry->d_inode;
-	xfs_mount_t		*mp = ip->i_mount;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	void			__user *arg = (void __user *)p;
+	int			ioflags = 0;
 	int			error;
 
-	xfs_itrace_entry(XFS_I(inode));
-	switch (cmd) {
+	if (filp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
 
+	xfs_itrace_entry(ip);
+
+	switch (cmd) {
 	case XFS_IOC_ALLOCSP:
 	case XFS_IOC_FREESP:
 	case XFS_IOC_RESVSP:
@@ -1348,17 +1362,13 @@ xfs_ioctl(
 	case XFS_IOC_ALLOCSP64:
 	case XFS_IOC_FREESP64:
 	case XFS_IOC_RESVSP64:
-	case XFS_IOC_UNRESVSP64:
-		/*
-		 * Only allow the sys admin to reserve space unless
-		 * unwritten extents are enabled.
-		 */
-		if (!xfs_sb_version_hasextflgbit(&mp->m_sb) &&
-		    !capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, arg);
+	case XFS_IOC_UNRESVSP64: {
+		xfs_flock64_t		bf;
 
+		if (copy_from_user(&bf, arg, sizeof(bf)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, &bf);
+	}
 	case XFS_IOC_DIOINFO: {
 		struct dioattr	da;
 		xfs_buftarg_t	*target =
@@ -1418,18 +1428,30 @@ xfs_ioctl(
 
 	case XFS_IOC_FD_TO_HANDLE:
 	case XFS_IOC_PATH_TO_HANDLE:
-	case XFS_IOC_PATH_TO_FSHANDLE:
-		return xfs_find_handle(cmd, arg);
+	case XFS_IOC_PATH_TO_FSHANDLE: {
+		xfs_fsop_handlereq_t	hreq;
 
-	case XFS_IOC_OPEN_BY_HANDLE:
-		return xfs_open_by_handle(mp, arg, filp, inode);
+		if (copy_from_user(&hreq, arg, sizeof(hreq)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_find_handle(cmd, &hreq);
+	}
+	case XFS_IOC_OPEN_BY_HANDLE: {
+		xfs_fsop_handlereq_t	hreq;
 
+		if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_open_by_handle(mp, &hreq, filp, inode);
+	}
 	case XFS_IOC_FSSETDM_BY_HANDLE:
 		return xfs_fssetdm_by_handle(mp, arg, inode);
 
-	case XFS_IOC_READLINK_BY_HANDLE:
-		return xfs_readlink_by_handle(mp, arg, inode);
+	case XFS_IOC_READLINK_BY_HANDLE: {
+		xfs_fsop_handlereq_t	hreq;
 
+		if (copy_from_user(&hreq, arg, sizeof(xfs_fsop_handlereq_t)))
+			return -XFS_ERROR(EFAULT);
+		return xfs_readlink_by_handle(mp, &hreq, inode);
+	}
 	case XFS_IOC_ATTRLIST_BY_HANDLE:
 		return xfs_attrlist_by_handle(mp, arg, inode);
 
@@ -1437,7 +1459,11 @@ xfs_ioctl(
 		return xfs_attrmulti_by_handle(mp, arg, filp, inode);
 
 	case XFS_IOC_SWAPEXT: {
-		error = xfs_swapext((struct xfs_swapext __user *)arg);
+		struct xfs_swapext	sxp;
+
+		if (copy_from_user(&sxp, arg, sizeof(xfs_swapext_t)))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_swapext(&sxp);
 		return -error;
 	}
 
@@ -1493,9 +1519,6 @@ xfs_ioctl(
 	case XFS_IOC_FSGROWFSDATA: {
 		xfs_growfs_data_t in;
 
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
 		if (copy_from_user(&in, arg, sizeof(in)))
 			return -XFS_ERROR(EFAULT);
 
@@ -1506,9 +1529,6 @@ xfs_ioctl(
 	case XFS_IOC_FSGROWFSLOG: {
 		xfs_growfs_log_t in;
 
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
 		if (copy_from_user(&in, arg, sizeof(in)))
 			return -XFS_ERROR(EFAULT);
 
@@ -1519,9 +1539,6 @@ xfs_ioctl(
 	case XFS_IOC_FSGROWFSRT: {
 		xfs_growfs_rt_t in;
 
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
 		if (copy_from_user(&in, arg, sizeof(in)))
 			return -XFS_ERROR(EFAULT);
 
@@ -1529,21 +1546,6 @@ xfs_ioctl(
 		return -error;
 	}
 
-	case XFS_IOC_FREEZE:
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-
-		if (inode->i_sb->s_frozen == SB_UNFROZEN)
-			freeze_bdev(inode->i_sb->s_bdev);
-		return 0;
-
-	case XFS_IOC_THAW:
-		if (!capable(CAP_SYS_ADMIN))
-			return -EPERM;
-		if (inode->i_sb->s_frozen != SB_UNFROZEN)
-			thaw_bdev(inode->i_sb->s_bdev, inode->i_sb);
-		return 0;
-
 	case XFS_IOC_GOINGDOWN: {
 		__uint32_t in;
 
diff --git a/fs/xfs/linux-2.6/xfs_ioctl.h b/fs/xfs/linux-2.6/xfs_ioctl.h
new file mode 100644
index 00000000000..8c16bf2d7e0
--- /dev/null
+++ b/fs/xfs/linux-2.6/xfs_ioctl.h
@@ -0,0 +1,82 @@
+/*
+ * Copyright (c) 2008 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_IOCTL_H__
+#define __XFS_IOCTL_H__
+
+extern int
+xfs_ioc_space(
+	struct xfs_inode	*ip,
+	struct inode		*inode,
+	struct file		*filp,
+	int			ioflags,
+	unsigned int		cmd,
+	xfs_flock64_t		*bf);
+
+extern int
+xfs_find_handle(
+	unsigned int		cmd,
+	xfs_fsop_handlereq_t	*hreq);
+
+extern int
+xfs_open_by_handle(
+	xfs_mount_t		*mp,
+	xfs_fsop_handlereq_t	*hreq,
+	struct file		*parfilp,
+	struct inode		*parinode);
+
+extern int
+xfs_readlink_by_handle(
+	xfs_mount_t		*mp,
+	xfs_fsop_handlereq_t	*hreq,
+	struct inode		*parinode);
+
+extern int
+xfs_attrmulti_attr_get(
+	struct inode		*inode,
+	char			*name,
+	char			__user *ubuf,
+	__uint32_t		*len,
+	__uint32_t		flags);
+
+extern int
+	xfs_attrmulti_attr_set(
+	struct inode		*inode,
+	char			*name,
+	const char		__user *ubuf,
+	__uint32_t		len,
+	__uint32_t		flags);
+
+extern int
+xfs_attrmulti_attr_remove(
+	struct inode		*inode,
+	char			*name,
+	__uint32_t		flags);
+
+extern long
+xfs_file_ioctl(
+	struct file		*filp,
+	unsigned int		cmd,
+	unsigned long		p);
+
+extern long
+xfs_file_compat_ioctl(
+	struct file		*file,
+	unsigned int		cmd,
+	unsigned long		arg);
+
+#endif
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c
index a4b254eb43b..50903ad3182 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.c
@@ -16,11 +16,7 @@
  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  */
 #include <linux/compat.h>
-#include <linux/init.h>
 #include <linux/ioctl.h>
-#include <linux/syscalls.h>
-#include <linux/types.h>
-#include <linux/fs.h>
 #include <asm/uaccess.h>
 #include "xfs.h"
 #include "xfs_fs.h"
@@ -36,7 +32,6 @@
 #include "xfs_bmap_btree.h"
 #include "xfs_attr_sf.h"
 #include "xfs_dir2_sf.h"
-#include "xfs_vfs.h"
 #include "xfs_vnode.h"
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
@@ -44,221 +39,219 @@
 #include "xfs_error.h"
 #include "xfs_dfrag.h"
 #include "xfs_vnodeops.h"
+#include "xfs_fsops.h"
+#include "xfs_alloc.h"
+#include "xfs_rtalloc.h"
+#include "xfs_attr.h"
+#include "xfs_ioctl.h"
 #include "xfs_ioctl32.h"
 
 #define  _NATIVE_IOC(cmd, type) \
 	  _IOC(_IOC_DIR(cmd), _IOC_TYPE(cmd), _IOC_NR(cmd), sizeof(type))
 
-#if defined(CONFIG_IA64) || defined(CONFIG_X86_64)
-#define BROKEN_X86_ALIGNMENT
-#define _PACKED __attribute__((packed))
-/* on ia32 l_start is on a 32-bit boundary */
-typedef struct xfs_flock64_32 {
-	__s16		l_type;
-	__s16		l_whence;
-	__s64		l_start	__attribute__((packed));
-			/* len == 0 means until end of file */
-	__s64		l_len __attribute__((packed));
-	__s32		l_sysid;
-	__u32		l_pid;
-	__s32		l_pad[4];	/* reserve area */
-} xfs_flock64_32_t;
-
-#define XFS_IOC_ALLOCSP_32	_IOW ('X', 10, struct xfs_flock64_32)
-#define XFS_IOC_FREESP_32	_IOW ('X', 11, struct xfs_flock64_32)
-#define XFS_IOC_ALLOCSP64_32	_IOW ('X', 36, struct xfs_flock64_32)
-#define XFS_IOC_FREESP64_32	_IOW ('X', 37, struct xfs_flock64_32)
-#define XFS_IOC_RESVSP_32	_IOW ('X', 40, struct xfs_flock64_32)
-#define XFS_IOC_UNRESVSP_32	_IOW ('X', 41, struct xfs_flock64_32)
-#define XFS_IOC_RESVSP64_32	_IOW ('X', 42, struct xfs_flock64_32)
-#define XFS_IOC_UNRESVSP64_32	_IOW ('X', 43, struct xfs_flock64_32)
-
-/* just account for different alignment */
-STATIC unsigned long
-xfs_ioctl32_flock(
-	unsigned long		arg)
+#ifdef BROKEN_X86_ALIGNMENT
+STATIC int
+xfs_compat_flock64_copyin(
+	xfs_flock64_t		*bf,
+	compat_xfs_flock64_t	__user *arg32)
 {
-	xfs_flock64_32_t	__user *p32 = (void __user *)arg;
-	xfs_flock64_t		__user *p = compat_alloc_user_space(sizeof(*p));
-
-	if (copy_in_user(&p->l_type,	&p32->l_type,	sizeof(s16)) ||
-	    copy_in_user(&p->l_whence,	&p32->l_whence, sizeof(s16)) ||
-	    copy_in_user(&p->l_start,	&p32->l_start,	sizeof(s64)) ||
-	    copy_in_user(&p->l_len,	&p32->l_len,	sizeof(s64)) ||
-	    copy_in_user(&p->l_sysid,	&p32->l_sysid,	sizeof(s32)) ||
-	    copy_in_user(&p->l_pid,	&p32->l_pid,	sizeof(u32)) ||
-	    copy_in_user(&p->l_pad,	&p32->l_pad,	4*sizeof(u32)))
-		return -EFAULT;
-
-	return (unsigned long)p;
+	if (get_user(bf->l_type,	&arg32->l_type) ||
+	    get_user(bf->l_whence,	&arg32->l_whence) ||
+	    get_user(bf->l_start,	&arg32->l_start) ||
+	    get_user(bf->l_len,		&arg32->l_len) ||
+	    get_user(bf->l_sysid,	&arg32->l_sysid) ||
+	    get_user(bf->l_pid,		&arg32->l_pid) ||
+	    copy_from_user(bf->l_pad,	&arg32->l_pad,	4*sizeof(u32)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
 }
 
-typedef struct compat_xfs_fsop_geom_v1 {
-	__u32		blocksize;	/* filesystem (data) block size */
-	__u32		rtextsize;	/* realtime extent size		*/
-	__u32		agblocks;	/* fsblocks in an AG		*/
-	__u32		agcount;	/* number of allocation groups	*/
-	__u32		logblocks;	/* fsblocks in the log		*/
-	__u32		sectsize;	/* (data) sector size, bytes	*/
-	__u32		inodesize;	/* inode size in bytes		*/
-	__u32		imaxpct;	/* max allowed inode space(%)	*/
-	__u64		datablocks;	/* fsblocks in data subvolume	*/
-	__u64		rtblocks;	/* fsblocks in realtime subvol	*/
-	__u64		rtextents;	/* rt extents in realtime subvol*/
-	__u64		logstart;	/* starting fsblock of the log	*/
-	unsigned char	uuid[16];	/* unique id of the filesystem	*/
-	__u32		sunit;		/* stripe unit, fsblocks	*/
-	__u32		swidth;		/* stripe width, fsblocks	*/
-	__s32		version;	/* structure version		*/
-	__u32		flags;		/* superblock version flags	*/
-	__u32		logsectsize;	/* log sector size, bytes	*/
-	__u32		rtsectsize;	/* realtime sector size, bytes	*/
-	__u32		dirblocksize;	/* directory block size, bytes	*/
-} __attribute__((packed)) compat_xfs_fsop_geom_v1_t;
-
-#define XFS_IOC_FSGEOMETRY_V1_32  \
-	_IOR ('X', 100, struct compat_xfs_fsop_geom_v1)
-
-STATIC unsigned long xfs_ioctl32_geom_v1(unsigned long arg)
+STATIC int
+xfs_compat_ioc_fsgeometry_v1(
+	struct xfs_mount	  *mp,
+	compat_xfs_fsop_geom_v1_t __user *arg32)
 {
-	compat_xfs_fsop_geom_v1_t __user *p32 = (void __user *)arg;
-	xfs_fsop_geom_v1_t __user *p = compat_alloc_user_space(sizeof(*p));
+	xfs_fsop_geom_t		  fsgeo;
+	int			  error;
 
-	if (copy_in_user(p, p32, sizeof(*p32)))
-		return -EFAULT;
-	return (unsigned long)p;
+	error = xfs_fs_geometry(mp, &fsgeo, 3);
+	if (error)
+		return -error;
+	/* The 32-bit variant simply has some padding at the end */
+	if (copy_to_user(arg32, &fsgeo, sizeof(struct compat_xfs_fsop_geom_v1)))
+		return -XFS_ERROR(EFAULT);
+	return 0;
 }
 
-typedef struct compat_xfs_inogrp {
-	__u64		xi_startino;	/* starting inode number	*/
-	__s32		xi_alloccount;	/* # bits set in allocmask	*/
-	__u64		xi_allocmask;	/* mask of allocated inodes	*/
-} __attribute__((packed)) compat_xfs_inogrp_t;
-
-STATIC int xfs_inumbers_fmt_compat(
-	void __user *ubuffer,
-	const xfs_inogrp_t *buffer,
-	long count,
-	long *written)
+STATIC int
+xfs_compat_growfs_data_copyin(
+	struct xfs_growfs_data	 *in,
+	compat_xfs_growfs_data_t __user *arg32)
 {
-	compat_xfs_inogrp_t __user *p32 = ubuffer;
-	long i;
+	if (get_user(in->newblocks, &arg32->newblocks) ||
+	    get_user(in->imaxpct,   &arg32->imaxpct))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_compat_growfs_rt_copyin(
+	struct xfs_growfs_rt	 *in,
+	compat_xfs_growfs_rt_t	__user *arg32)
+{
+	if (get_user(in->newblocks, &arg32->newblocks) ||
+	    get_user(in->extsize,   &arg32->extsize))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
+
+STATIC int
+xfs_inumbers_fmt_compat(
+	void			__user *ubuffer,
+	const xfs_inogrp_t	*buffer,
+	long			count,
+	long			*written)
+{
+	compat_xfs_inogrp_t	__user *p32 = ubuffer;
+	long			i;
 
 	for (i = 0; i < count; i++) {
 		if (put_user(buffer[i].xi_startino,   &p32[i].xi_startino) ||
 		    put_user(buffer[i].xi_alloccount, &p32[i].xi_alloccount) ||
 		    put_user(buffer[i].xi_allocmask,  &p32[i].xi_allocmask))
-			return -EFAULT;
+			return -XFS_ERROR(EFAULT);
 	}
 	*written = count * sizeof(*p32);
 	return 0;
 }
 
 #else
-
 #define xfs_inumbers_fmt_compat xfs_inumbers_fmt
-#define _PACKED
+#endif	/* BROKEN_X86_ALIGNMENT */
 
-#endif
+STATIC int
+xfs_ioctl32_bstime_copyin(
+	xfs_bstime_t		*bstime,
+	compat_xfs_bstime_t	__user *bstime32)
+{
+	compat_time_t		sec32;	/* tv_sec differs on 64 vs. 32 */
 
-/* XFS_IOC_FSBULKSTAT and friends */
+	if (get_user(sec32,		&bstime32->tv_sec)	||
+	    get_user(bstime->tv_nsec,	&bstime32->tv_nsec))
+		return -XFS_ERROR(EFAULT);
+	bstime->tv_sec = sec32;
+	return 0;
+}
+
+/* xfs_bstat_t has differing alignment on intel, & bstime_t sizes everywhere */
+STATIC int
+xfs_ioctl32_bstat_copyin(
+	xfs_bstat_t		*bstat,
+	compat_xfs_bstat_t	__user *bstat32)
+{
+	if (get_user(bstat->bs_ino,	&bstat32->bs_ino)	||
+	    get_user(bstat->bs_mode,	&bstat32->bs_mode)	||
+	    get_user(bstat->bs_nlink,	&bstat32->bs_nlink)	||
+	    get_user(bstat->bs_uid,	&bstat32->bs_uid)	||
+	    get_user(bstat->bs_gid,	&bstat32->bs_gid)	||
+	    get_user(bstat->bs_rdev,	&bstat32->bs_rdev)	||
+	    get_user(bstat->bs_blksize,	&bstat32->bs_blksize)	||
+	    get_user(bstat->bs_size,	&bstat32->bs_size)	||
+	    xfs_ioctl32_bstime_copyin(&bstat->bs_atime, &bstat32->bs_atime) ||
+	    xfs_ioctl32_bstime_copyin(&bstat->bs_mtime, &bstat32->bs_mtime) ||
+	    xfs_ioctl32_bstime_copyin(&bstat->bs_ctime, &bstat32->bs_ctime) ||
+	    get_user(bstat->bs_blocks,	&bstat32->bs_size)	||
+	    get_user(bstat->bs_xflags,	&bstat32->bs_size)	||
+	    get_user(bstat->bs_extsize,	&bstat32->bs_extsize)	||
+	    get_user(bstat->bs_extents,	&bstat32->bs_extents)	||
+	    get_user(bstat->bs_gen,	&bstat32->bs_gen)	||
+	    get_user(bstat->bs_projid,	&bstat32->bs_projid)	||
+	    get_user(bstat->bs_dmevmask, &bstat32->bs_dmevmask)	||
+	    get_user(bstat->bs_dmstate,	&bstat32->bs_dmstate)	||
+	    get_user(bstat->bs_aextents, &bstat32->bs_aextents))
+		return -XFS_ERROR(EFAULT);
+	return 0;
+}
 
-typedef struct compat_xfs_bstime {
-	__s32		tv_sec;		/* seconds		*/
-	__s32		tv_nsec;	/* and nanoseconds	*/
-} compat_xfs_bstime_t;
+/* XFS_IOC_FSBULKSTAT and friends */
 
-STATIC int xfs_bstime_store_compat(
-	compat_xfs_bstime_t __user *p32,
-	const xfs_bstime_t *p)
+STATIC int
+xfs_bstime_store_compat(
+	compat_xfs_bstime_t	__user *p32,
+	const xfs_bstime_t	*p)
 {
-	__s32 sec32;
+	__s32			sec32;
 
 	sec32 = p->tv_sec;
 	if (put_user(sec32, &p32->tv_sec) ||
 	    put_user(p->tv_nsec, &p32->tv_nsec))
-		return -EFAULT;
+		return -XFS_ERROR(EFAULT);
 	return 0;
 }
 
-typedef struct compat_xfs_bstat {
-	__u64		bs_ino;		/* inode number			*/
-	__u16		bs_mode;	/* type and mode		*/
-	__u16		bs_nlink;	/* number of links		*/
-	__u32		bs_uid;		/* user id			*/
-	__u32		bs_gid;		/* group id			*/
-	__u32		bs_rdev;	/* device value			*/
-	__s32		bs_blksize;	/* block size			*/
-	__s64		bs_size;	/* file size			*/
-	compat_xfs_bstime_t bs_atime;	/* access time			*/
-	compat_xfs_bstime_t bs_mtime;	/* modify time			*/
-	compat_xfs_bstime_t bs_ctime;	/* inode change time		*/
-	int64_t		bs_blocks;	/* number of blocks		*/
-	__u32		bs_xflags;	/* extended flags		*/
-	__s32		bs_extsize;	/* extent size			*/
-	__s32		bs_extents;	/* number of extents		*/
-	__u32		bs_gen;		/* generation count		*/
-	__u16		bs_projid;	/* project id			*/
-	unsigned char	bs_pad[14];	/* pad space, unused		*/
-	__u32		bs_dmevmask;	/* DMIG event mask		*/
-	__u16		bs_dmstate;	/* DMIG state info		*/
-	__u16		bs_aextents;	/* attribute number of extents	*/
-} _PACKED compat_xfs_bstat_t;
-
-STATIC int xfs_bulkstat_one_fmt_compat(
+/* Return 0 on success or positive error (to xfs_bulkstat()) */
+STATIC int
+xfs_bulkstat_one_fmt_compat(
 	void			__user *ubuffer,
+	int			ubsize,
+	int			*ubused,
 	const xfs_bstat_t	*buffer)
 {
-	compat_xfs_bstat_t __user *p32 = ubuffer;
-
-	if (put_user(buffer->bs_ino, &p32->bs_ino) ||
-	    put_user(buffer->bs_mode, &p32->bs_mode) ||
-	    put_user(buffer->bs_nlink, &p32->bs_nlink) ||
-	    put_user(buffer->bs_uid, &p32->bs_uid) ||
-	    put_user(buffer->bs_gid, &p32->bs_gid) ||
-	    put_user(buffer->bs_rdev, &p32->bs_rdev) ||
-	    put_user(buffer->bs_blksize, &p32->bs_blksize) ||
-	    put_user(buffer->bs_size, &p32->bs_size) ||
+	compat_xfs_bstat_t	__user *p32 = ubuffer;
+
+	if (ubsize < sizeof(*p32))
+		return XFS_ERROR(ENOMEM);
+
+	if (put_user(buffer->bs_ino,	  &p32->bs_ino)		||
+	    put_user(buffer->bs_mode,	  &p32->bs_mode)	||
+	    put_user(buffer->bs_nlink,	  &p32->bs_nlink)	||
+	    put_user(buffer->bs_uid,	  &p32->bs_uid)		||
+	    put_user(buffer->bs_gid,	  &p32->bs_gid)		||
+	    put_user(buffer->bs_rdev,	  &p32->bs_rdev)	||
+	    put_user(buffer->bs_blksize,  &p32->bs_blksize)	||
+	    put_user(buffer->bs_size,	  &p32->bs_size)	||
 	    xfs_bstime_store_compat(&p32->bs_atime, &buffer->bs_atime) ||
 	    xfs_bstime_store_compat(&p32->bs_mtime, &buffer->bs_mtime) ||
 	    xfs_bstime_store_compat(&p32->bs_ctime, &buffer->bs_ctime) ||
-	    put_user(buffer->bs_blocks, &p32->bs_blocks) ||
-	    put_user(buffer->bs_xflags, &p32->bs_xflags) ||
-	    put_user(buffer->bs_extsize, &p32->bs_extsize) ||
-	    put_user(buffer->bs_extents, &p32->bs_extents) ||
-	    put_user(buffer->bs_gen, &p32->bs_gen) ||
-	    put_user(buffer->bs_projid, &p32->bs_projid) ||
-	    put_user(buffer->bs_dmevmask, &p32->bs_dmevmask) ||
-	    put_user(buffer->bs_dmstate, &p32->bs_dmstate) ||
+	    put_user(buffer->bs_blocks,	  &p32->bs_blocks)	||
+	    put_user(buffer->bs_xflags,	  &p32->bs_xflags)	||
+	    put_user(buffer->bs_extsize,  &p32->bs_extsize)	||
+	    put_user(buffer->bs_extents,  &p32->bs_extents)	||
+	    put_user(buffer->bs_gen,	  &p32->bs_gen)		||
+	    put_user(buffer->bs_projid,	  &p32->bs_projid)	||
+	    put_user(buffer->bs_dmevmask, &p32->bs_dmevmask)	||
+	    put_user(buffer->bs_dmstate,  &p32->bs_dmstate)	||
 	    put_user(buffer->bs_aextents, &p32->bs_aextents))
-		return -EFAULT;
-	return sizeof(*p32);
+		return XFS_ERROR(EFAULT);
+	if (ubused)
+		*ubused = sizeof(*p32);
+	return 0;
 }
 
-
-
-typedef struct compat_xfs_fsop_bulkreq {
-	compat_uptr_t	lastip;		/* last inode # pointer		*/
-	__s32		icount;		/* count of entries in buffer	*/
-	compat_uptr_t	ubuffer;	/* user buffer for inode desc.	*/
-	compat_uptr_t	ocount;		/* output count pointer		*/
-} compat_xfs_fsop_bulkreq_t;
-
-#define XFS_IOC_FSBULKSTAT_32 \
-	_IOWR('X', 101, struct compat_xfs_fsop_bulkreq)
-#define XFS_IOC_FSBULKSTAT_SINGLE_32 \
-	_IOWR('X', 102, struct compat_xfs_fsop_bulkreq)
-#define XFS_IOC_FSINUMBERS_32 \
-	_IOWR('X', 103, struct compat_xfs_fsop_bulkreq)
+STATIC int
+xfs_bulkstat_one_compat(
+	xfs_mount_t	*mp,		/* mount point for filesystem */
+	xfs_ino_t	ino,		/* inode number to get data for */
+	void		__user *buffer,	/* buffer to place output in */
+	int		ubsize,		/* size of buffer */
+	void		*private_data,	/* my private data */
+	xfs_daddr_t	bno,		/* starting bno of inode cluster */
+	int		*ubused,	/* bytes used by me */
+	void		*dibuff,	/* on-disk inode buffer */
+	int		*stat)		/* BULKSTAT_RV_... */
+{
+	return xfs_bulkstat_one_int(mp, ino, buffer, ubsize,
+				    xfs_bulkstat_one_fmt_compat, bno,
+				    ubused, dibuff, stat);
+}
 
 /* copied from xfs_ioctl.c */
 STATIC int
-xfs_ioc_bulkstat_compat(
-	xfs_mount_t		*mp,
-	unsigned int		cmd,
-	void			__user *arg)
+xfs_compat_ioc_bulkstat(
+	xfs_mount_t		  *mp,
+	unsigned int		  cmd,
+	compat_xfs_fsop_bulkreq_t __user *p32)
 {
-	compat_xfs_fsop_bulkreq_t __user *p32 = (void __user *)arg;
 	u32			addr;
 	xfs_fsop_bulkreq_t	bulkreq;
 	int			count;	/* # of records returned */
@@ -270,20 +263,20 @@ xfs_ioc_bulkstat_compat(
 	/* should be called again (unused here, but used in dmapi) */
 
 	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
+		return -XFS_ERROR(EPERM);
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -XFS_ERROR(EIO);
 
 	if (get_user(addr, &p32->lastip))
-		return -EFAULT;
+		return -XFS_ERROR(EFAULT);
 	bulkreq.lastip = compat_ptr(addr);
 	if (get_user(bulkreq.icount, &p32->icount) ||
 	    get_user(addr, &p32->ubuffer))
-		return -EFAULT;
+		return -XFS_ERROR(EFAULT);
 	bulkreq.ubuffer = compat_ptr(addr);
 	if (get_user(addr, &p32->ocount))
-		return -EFAULT;
+		return -XFS_ERROR(EFAULT);
 	bulkreq.ocount = compat_ptr(addr);
 
 	if (copy_from_user(&inlast, bulkreq.lastip, sizeof(__s64)))
@@ -295,17 +288,22 @@ xfs_ioc_bulkstat_compat(
 	if (bulkreq.ubuffer == NULL)
 		return -XFS_ERROR(EINVAL);
 
-	if (cmd == XFS_IOC_FSINUMBERS)
+	if (cmd == XFS_IOC_FSINUMBERS_32) {
 		error = xfs_inumbers(mp, &inlast, &count,
 				bulkreq.ubuffer, xfs_inumbers_fmt_compat);
-	else {
-		/* declare a var to get a warning in case the type changes */
-		bulkstat_one_fmt_pf formatter = xfs_bulkstat_one_fmt_compat;
+	} else if (cmd == XFS_IOC_FSBULKSTAT_SINGLE_32) {
+		int res;
+
+		error = xfs_bulkstat_one_compat(mp, inlast, bulkreq.ubuffer,
+				sizeof(compat_xfs_bstat_t),
+				NULL, 0, NULL, NULL, &res);
+	} else if (cmd == XFS_IOC_FSBULKSTAT_32) {
 		error = xfs_bulkstat(mp, &inlast, &count,
-			xfs_bulkstat_one, formatter,
+			xfs_bulkstat_one_compat, NULL,
 			sizeof(compat_xfs_bstat_t), bulkreq.ubuffer,
 			BULKSTAT_FG_QUICK, &done);
-	}
+	} else
+		error = XFS_ERROR(EINVAL);
 	if (error)
 		return -error;
 
@@ -321,63 +319,306 @@ xfs_ioc_bulkstat_compat(
 	return 0;
 }
 
+STATIC int
+xfs_compat_handlereq_copyin(
+	xfs_fsop_handlereq_t		*hreq,
+	compat_xfs_fsop_handlereq_t	__user *arg32)
+{
+	compat_xfs_fsop_handlereq_t	hreq32;
+
+	if (copy_from_user(&hreq32, arg32, sizeof(compat_xfs_fsop_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	hreq->fd = hreq32.fd;
+	hreq->path = compat_ptr(hreq32.path);
+	hreq->oflags = hreq32.oflags;
+	hreq->ihandle = compat_ptr(hreq32.ihandle);
+	hreq->ihandlen = hreq32.ihandlen;
+	hreq->ohandle = compat_ptr(hreq32.ohandle);
+	hreq->ohandlen = compat_ptr(hreq32.ohandlen);
 
+	return 0;
+}
 
-typedef struct compat_xfs_fsop_handlereq {
-	__u32		fd;		/* fd for FD_TO_HANDLE		*/
-	compat_uptr_t	path;		/* user pathname		*/
-	__u32		oflags;		/* open flags			*/
-	compat_uptr_t	ihandle;	/* user supplied handle		*/
-	__u32		ihandlen;	/* user supplied length		*/
-	compat_uptr_t	ohandle;	/* user buffer for handle	*/
-	compat_uptr_t	ohandlen;	/* user buffer length		*/
-} compat_xfs_fsop_handlereq_t;
-
-#define XFS_IOC_PATH_TO_FSHANDLE_32 \
-	_IOWR('X', 104, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_PATH_TO_HANDLE_32 \
-	_IOWR('X', 105, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_FD_TO_HANDLE_32 \
-	_IOWR('X', 106, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_OPEN_BY_HANDLE_32 \
-	_IOWR('X', 107, struct compat_xfs_fsop_handlereq)
-#define XFS_IOC_READLINK_BY_HANDLE_32 \
-	_IOWR('X', 108, struct compat_xfs_fsop_handlereq)
-
-STATIC unsigned long xfs_ioctl32_fshandle(unsigned long arg)
+/*
+ * Convert userspace handle data into inode.
+ *
+ * We use the fact that all the fsop_handlereq ioctl calls have a data
+ * structure argument whose first component is always a xfs_fsop_handlereq_t,
+ * so we can pass that sub structure into this handy, shared routine.
+ *
+ * If no error, caller must always iput the returned inode.
+ */
+STATIC int
+xfs_vget_fsop_handlereq_compat(
+	xfs_mount_t		*mp,
+	struct inode		*parinode,	/* parent inode pointer    */
+	compat_xfs_fsop_handlereq_t	*hreq,
+	struct inode		**inode)
 {
-	compat_xfs_fsop_handlereq_t __user *p32 = (void __user *)arg;
-	xfs_fsop_handlereq_t __user *p = compat_alloc_user_space(sizeof(*p));
-	u32 addr;
-
-	if (copy_in_user(&p->fd, &p32->fd, sizeof(__u32)) ||
-	    get_user(addr, &p32->path) ||
-	    put_user(compat_ptr(addr), &p->path) ||
-	    copy_in_user(&p->oflags, &p32->oflags, sizeof(__u32)) ||
-	    get_user(addr, &p32->ihandle) ||
-	    put_user(compat_ptr(addr), &p->ihandle) ||
-	    copy_in_user(&p->ihandlen, &p32->ihandlen, sizeof(__u32)) ||
-	    get_user(addr, &p32->ohandle) ||
-	    put_user(compat_ptr(addr), &p->ohandle) ||
-	    get_user(addr, &p32->ohandlen) ||
-	    put_user(compat_ptr(addr), &p->ohandlen))
-		return -EFAULT;
-
-	return (unsigned long)p;
+	void			__user *hanp;
+	size_t			hlen;
+	xfs_fid_t		*xfid;
+	xfs_handle_t		*handlep;
+	xfs_handle_t		handle;
+	xfs_inode_t		*ip;
+	xfs_ino_t		ino;
+	__u32			igen;
+	int			error;
+
+	/*
+	 * Only allow handle opens under a directory.
+	 */
+	if (!S_ISDIR(parinode->i_mode))
+		return XFS_ERROR(ENOTDIR);
+
+	hanp = compat_ptr(hreq->ihandle);
+	hlen = hreq->ihandlen;
+	handlep = &handle;
+
+	if (hlen < sizeof(handlep->ha_fsid) || hlen > sizeof(*handlep))
+		return XFS_ERROR(EINVAL);
+	if (copy_from_user(handlep, hanp, hlen))
+		return XFS_ERROR(EFAULT);
+	if (hlen < sizeof(*handlep))
+		memset(((char *)handlep) + hlen, 0, sizeof(*handlep) - hlen);
+	if (hlen > sizeof(handlep->ha_fsid)) {
+		if (handlep->ha_fid.fid_len !=
+		    (hlen - sizeof(handlep->ha_fsid) -
+			    sizeof(handlep->ha_fid.fid_len)) ||
+		    handlep->ha_fid.fid_pad)
+			return XFS_ERROR(EINVAL);
+	}
+
+	/*
+	 * Crack the handle, obtain the inode # & generation #
+	 */
+	xfid = (struct xfs_fid *)&handlep->ha_fid;
+	if (xfid->fid_len == sizeof(*xfid) - sizeof(xfid->fid_len)) {
+		ino  = xfid->fid_ino;
+		igen = xfid->fid_gen;
+	} else {
+		return XFS_ERROR(EINVAL);
+	}
+
+	/*
+	 * Get the XFS inode, building a Linux inode to go with it.
+	 */
+	error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_SHARED, &ip, 0);
+	if (error)
+		return error;
+	if (ip == NULL)
+		return XFS_ERROR(EIO);
+	if (ip->i_d.di_gen != igen) {
+		xfs_iput_new(ip, XFS_ILOCK_SHARED);
+		return XFS_ERROR(ENOENT);
+	}
+
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	*inode = VFS_I(ip);
+	return 0;
 }
 
+STATIC int
+xfs_compat_attrlist_by_handle(
+	xfs_mount_t		*mp,
+	void			__user *arg,
+	struct inode		*parinode)
+{
+	int			error;
+	attrlist_cursor_kern_t	*cursor;
+	compat_xfs_fsop_attrlist_handlereq_t al_hreq;
+	struct inode		*inode;
+	char			*kbuf;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&al_hreq, arg,
+			   sizeof(compat_xfs_fsop_attrlist_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+	if (al_hreq.buflen > XATTR_LIST_MAX)
+		return -XFS_ERROR(EINVAL);
+
+	/*
+	 * Reject flags, only allow namespaces.
+	 */
+	if (al_hreq.flags & ~(ATTR_ROOT | ATTR_SECURE))
+		return -XFS_ERROR(EINVAL);
+
+	error = xfs_vget_fsop_handlereq_compat(mp, parinode, &al_hreq.hreq,
+					       &inode);
+	if (error)
+		goto out;
+
+	kbuf = kmalloc(al_hreq.buflen, GFP_KERNEL);
+	if (!kbuf)
+		goto out_vn_rele;
+
+	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
+	error = xfs_attr_list(XFS_I(inode), kbuf, al_hreq.buflen,
+					al_hreq.flags, cursor);
+	if (error)
+		goto out_kfree;
+
+	if (copy_to_user(compat_ptr(al_hreq.buffer), kbuf, al_hreq.buflen))
+		error = -EFAULT;
+
+ out_kfree:
+	kfree(kbuf);
+ out_vn_rele:
+	iput(inode);
+ out:
+	return -error;
+}
 
-STATIC long
-xfs_compat_ioctl(
-	int		mode,
-	struct file	*file,
-	unsigned	cmd,
-	unsigned long	arg)
+STATIC int
+xfs_compat_attrmulti_by_handle(
+	xfs_mount_t				*mp,
+	void					__user *arg,
+	struct inode				*parinode)
+{
+	int					error;
+	compat_xfs_attr_multiop_t		*ops;
+	compat_xfs_fsop_attrmulti_handlereq_t	am_hreq;
+	struct inode				*inode;
+	unsigned int				i, size;
+	char					*attr_name;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&am_hreq, arg,
+			   sizeof(compat_xfs_fsop_attrmulti_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	error = xfs_vget_fsop_handlereq_compat(mp, parinode, &am_hreq.hreq,
+					       &inode);
+	if (error)
+		goto out;
+
+	error = E2BIG;
+	size = am_hreq.opcount * sizeof(compat_xfs_attr_multiop_t);
+	if (!size || size > 16 * PAGE_SIZE)
+		goto out_vn_rele;
+
+	error = ENOMEM;
+	ops = kmalloc(size, GFP_KERNEL);
+	if (!ops)
+		goto out_vn_rele;
+
+	error = EFAULT;
+	if (copy_from_user(ops, compat_ptr(am_hreq.ops), size))
+		goto out_kfree_ops;
+
+	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
+	if (!attr_name)
+		goto out_kfree_ops;
+
+
+	error = 0;
+	for (i = 0; i < am_hreq.opcount; i++) {
+		ops[i].am_error = strncpy_from_user(attr_name,
+				compat_ptr(ops[i].am_attrname),
+				MAXNAMELEN);
+		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
+			error = -ERANGE;
+		if (ops[i].am_error < 0)
+			break;
+
+		switch (ops[i].am_opcode) {
+		case ATTR_OP_GET:
+			ops[i].am_error = xfs_attrmulti_attr_get(inode,
+					attr_name,
+					compat_ptr(ops[i].am_attrvalue),
+					&ops[i].am_length, ops[i].am_flags);
+			break;
+		case ATTR_OP_SET:
+			ops[i].am_error = xfs_attrmulti_attr_set(inode,
+					attr_name,
+					compat_ptr(ops[i].am_attrvalue),
+					ops[i].am_length, ops[i].am_flags);
+			break;
+		case ATTR_OP_REMOVE:
+			ops[i].am_error = xfs_attrmulti_attr_remove(inode,
+					attr_name, ops[i].am_flags);
+			break;
+		default:
+			ops[i].am_error = EINVAL;
+		}
+	}
+
+	if (copy_to_user(compat_ptr(am_hreq.ops), ops, size))
+		error = XFS_ERROR(EFAULT);
+
+	kfree(attr_name);
+ out_kfree_ops:
+	kfree(ops);
+ out_vn_rele:
+	iput(inode);
+ out:
+	return -error;
+}
+
+STATIC int
+xfs_compat_fssetdm_by_handle(
+	xfs_mount_t		*mp,
+	void			__user *arg,
+	struct inode		*parinode)
+{
+	int			error;
+	struct fsdmidata	fsd;
+	compat_xfs_fsop_setdm_handlereq_t dmhreq;
+	struct inode		*inode;
+
+	if (!capable(CAP_MKNOD))
+		return -XFS_ERROR(EPERM);
+	if (copy_from_user(&dmhreq, arg,
+			   sizeof(compat_xfs_fsop_setdm_handlereq_t)))
+		return -XFS_ERROR(EFAULT);
+
+	error = xfs_vget_fsop_handlereq_compat(mp, parinode, &dmhreq.hreq,
+					       &inode);
+	if (error)
+		return -error;
+
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) {
+		error = -XFS_ERROR(EPERM);
+		goto out;
+	}
+
+	if (copy_from_user(&fsd, compat_ptr(dmhreq.data), sizeof(fsd))) {
+		error = -XFS_ERROR(EFAULT);
+		goto out;
+	}
+
+	error = -xfs_set_dmattrs(XFS_I(inode), fsd.fsd_dmevmask,
+				 fsd.fsd_dmstate);
+
+out:
+	iput(inode);
+	return error;
+}
+
+long
+xfs_file_compat_ioctl(
+	struct file		*filp,
+	unsigned		cmd,
+	unsigned long		p)
 {
-	struct inode	*inode = file->f_path.dentry->d_inode;
-	int		error;
+	struct inode		*inode = filp->f_path.dentry->d_inode;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	void			__user *arg = (void __user *)p;
+	int			ioflags = 0;
+	int			error;
+
+	if (filp->f_mode & FMODE_NOCMTIME)
+		ioflags |= IO_INVIS;
+
+	xfs_itrace_entry(ip);
 
 	switch (cmd) {
+	/* No size or alignment issues on any arch */
 	case XFS_IOC_DIOINFO:
 	case XFS_IOC_FSGEOMETRY:
 	case XFS_IOC_FSGETXATTR:
@@ -387,48 +628,16 @@ xfs_compat_ioctl(
 	case XFS_IOC_GETBMAP:
 	case XFS_IOC_GETBMAPA:
 	case XFS_IOC_GETBMAPX:
-/* not handled
-	case XFS_IOC_FSSETDM_BY_HANDLE:
-	case XFS_IOC_ATTRLIST_BY_HANDLE:
-	case XFS_IOC_ATTRMULTI_BY_HANDLE:
-*/
 	case XFS_IOC_FSCOUNTS:
 	case XFS_IOC_SET_RESBLKS:
 	case XFS_IOC_GET_RESBLKS:
-	case XFS_IOC_FSGROWFSDATA:
 	case XFS_IOC_FSGROWFSLOG:
-	case XFS_IOC_FSGROWFSRT:
-	case XFS_IOC_FREEZE:
-	case XFS_IOC_THAW:
 	case XFS_IOC_GOINGDOWN:
 	case XFS_IOC_ERROR_INJECTION:
 	case XFS_IOC_ERROR_CLEARALL:
-		break;
-
-	case XFS_IOC32_GETXFLAGS:
-	case XFS_IOC32_SETXFLAGS:
-	case XFS_IOC32_GETVERSION:
-		cmd = _NATIVE_IOC(cmd, long);
-		break;
-#ifdef BROKEN_X86_ALIGNMENT
-	/* xfs_flock_t has wrong u32 vs u64 alignment */
-	case XFS_IOC_ALLOCSP_32:
-	case XFS_IOC_FREESP_32:
-	case XFS_IOC_ALLOCSP64_32:
-	case XFS_IOC_FREESP64_32:
-	case XFS_IOC_RESVSP_32:
-	case XFS_IOC_UNRESVSP_32:
-	case XFS_IOC_RESVSP64_32:
-	case XFS_IOC_UNRESVSP64_32:
-		arg = xfs_ioctl32_flock(arg);
-		cmd = _NATIVE_IOC(cmd, struct xfs_flock64);
-		break;
-	case XFS_IOC_FSGEOMETRY_V1_32:
-		arg = xfs_ioctl32_geom_v1(arg);
-		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_geom_v1);
-		break;
-
-#else /* These are handled fine if no alignment issues */
+		return xfs_file_ioctl(filp, cmd, p);
+#ifndef BROKEN_X86_ALIGNMENT
+	/* These are handled fine if no alignment issues */
 	case XFS_IOC_ALLOCSP:
 	case XFS_IOC_FREESP:
 	case XFS_IOC_RESVSP:
@@ -438,51 +647,97 @@ xfs_compat_ioctl(
 	case XFS_IOC_RESVSP64:
 	case XFS_IOC_UNRESVSP64:
 	case XFS_IOC_FSGEOMETRY_V1:
-		break;
+	case XFS_IOC_FSGROWFSDATA:
+	case XFS_IOC_FSGROWFSRT:
+		return xfs_file_ioctl(filp, cmd, p);
+#else
+	case XFS_IOC_ALLOCSP_32:
+	case XFS_IOC_FREESP_32:
+	case XFS_IOC_ALLOCSP64_32:
+	case XFS_IOC_FREESP64_32:
+	case XFS_IOC_RESVSP_32:
+	case XFS_IOC_UNRESVSP_32:
+	case XFS_IOC_RESVSP64_32:
+	case XFS_IOC_UNRESVSP64_32: {
+		struct xfs_flock64	bf;
 
-	/* xfs_bstat_t still has wrong u32 vs u64 alignment */
-	case XFS_IOC_SWAPEXT:
-		break;
+		if (xfs_compat_flock64_copyin(&bf, arg))
+			return -XFS_ERROR(EFAULT);
+		cmd = _NATIVE_IOC(cmd, struct xfs_flock64);
+		return xfs_ioc_space(ip, inode, filp, ioflags, cmd, &bf);
+	}
+	case XFS_IOC_FSGEOMETRY_V1_32:
+		return xfs_compat_ioc_fsgeometry_v1(mp, arg);
+	case XFS_IOC_FSGROWFSDATA_32: {
+		struct xfs_growfs_data	in;
+
+		if (xfs_compat_growfs_data_copyin(&in, arg))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_growfs_data(mp, &in);
+		return -error;
+	}
+	case XFS_IOC_FSGROWFSRT_32: {
+		struct xfs_growfs_rt	in;
 
+		if (xfs_compat_growfs_rt_copyin(&in, arg))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_growfs_rt(mp, &in);
+		return -error;
+	}
 #endif
+	/* long changes size, but xfs only copiese out 32 bits */
+	case XFS_IOC_GETXFLAGS_32:
+	case XFS_IOC_SETXFLAGS_32:
+	case XFS_IOC_GETVERSION_32:
+		cmd = _NATIVE_IOC(cmd, long);
+		return xfs_file_ioctl(filp, cmd, p);
+	case XFS_IOC_SWAPEXT: {
+		struct xfs_swapext	  sxp;
+		struct compat_xfs_swapext __user *sxu = arg;
+
+		/* Bulk copy in up to the sx_stat field, then copy bstat */
+		if (copy_from_user(&sxp, sxu,
+				   offsetof(struct xfs_swapext, sx_stat)) ||
+		    xfs_ioctl32_bstat_copyin(&sxp.sx_stat, &sxu->sx_stat))
+			return -XFS_ERROR(EFAULT);
+		error = xfs_swapext(&sxp);
+		return -error;
+	}
 	case XFS_IOC_FSBULKSTAT_32:
 	case XFS_IOC_FSBULKSTAT_SINGLE_32:
 	case XFS_IOC_FSINUMBERS_32:
-		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_bulkreq);
-		return xfs_ioc_bulkstat_compat(XFS_I(inode)->i_mount,
-				cmd, (void __user*)arg);
+		return xfs_compat_ioc_bulkstat(mp, cmd, arg);
 	case XFS_IOC_FD_TO_HANDLE_32:
 	case XFS_IOC_PATH_TO_HANDLE_32:
-	case XFS_IOC_PATH_TO_FSHANDLE_32:
-	case XFS_IOC_OPEN_BY_HANDLE_32:
-	case XFS_IOC_READLINK_BY_HANDLE_32:
-		arg = xfs_ioctl32_fshandle(arg);
+	case XFS_IOC_PATH_TO_FSHANDLE_32: {
+		struct xfs_fsop_handlereq	hreq;
+
+		if (xfs_compat_handlereq_copyin(&hreq, arg))
+			return -XFS_ERROR(EFAULT);
 		cmd = _NATIVE_IOC(cmd, struct xfs_fsop_handlereq);
-		break;
-	default:
-		return -ENOIOCTLCMD;
+		return xfs_find_handle(cmd, &hreq);
 	}
+	case XFS_IOC_OPEN_BY_HANDLE_32: {
+		struct xfs_fsop_handlereq	hreq;
 
-	error = xfs_ioctl(XFS_I(inode), file, mode, cmd, (void __user *)arg);
-	xfs_iflags_set(XFS_I(inode), XFS_IMODIFIED);
-
-	return error;
-}
-
-long
-xfs_file_compat_ioctl(
-	struct file		*file,
-	unsigned		cmd,
-	unsigned long		arg)
-{
-	return xfs_compat_ioctl(0, file, cmd, arg);
-}
+		if (xfs_compat_handlereq_copyin(&hreq, arg))
+			return -XFS_ERROR(EFAULT);
+		return xfs_open_by_handle(mp, &hreq, filp, inode);
+	}
+	case XFS_IOC_READLINK_BY_HANDLE_32: {
+		struct xfs_fsop_handlereq	hreq;
 
-long
-xfs_file_compat_invis_ioctl(
-	struct file		*file,
-	unsigned		cmd,
-	unsigned long		arg)
-{
-	return xfs_compat_ioctl(IO_INVIS, file, cmd, arg);
+		if (xfs_compat_handlereq_copyin(&hreq, arg))
+			return -XFS_ERROR(EFAULT);
+		return xfs_readlink_by_handle(mp, &hreq, inode);
+	}
+	case XFS_IOC_ATTRLIST_BY_HANDLE_32:
+		return xfs_compat_attrlist_by_handle(mp, arg, inode);
+	case XFS_IOC_ATTRMULTI_BY_HANDLE_32:
+		return xfs_compat_attrmulti_by_handle(mp, arg, inode);
+	case XFS_IOC_FSSETDM_BY_HANDLE_32:
+		return xfs_compat_fssetdm_by_handle(mp, arg, inode);
+	default:
+		return -XFS_ERROR(ENOIOCTLCMD);
+	}
 }
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.h b/fs/xfs/linux-2.6/xfs_ioctl32.h
index 02de6e62ee3..1024c4f8ba0 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.h
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.h
@@ -18,7 +18,217 @@
 #ifndef __XFS_IOCTL32_H__
 #define __XFS_IOCTL32_H__
 
-extern long xfs_file_compat_ioctl(struct file *, unsigned, unsigned long);
-extern long xfs_file_compat_invis_ioctl(struct file *, unsigned, unsigned long);
+#include <linux/compat.h>
+
+/*
+ * on 32-bit arches, ioctl argument structures may have different sizes
+ * and/or alignment.  We define compat structures which match the
+ * 32-bit sizes/alignments here, and their associated ioctl numbers.
+ *
+ * xfs_ioctl32.c contains routines to copy these structures in and out.
+ */
+
+/* stock kernel-level ioctls we support */
+#define XFS_IOC_GETXFLAGS_32	FS_IOC32_GETFLAGS
+#define XFS_IOC_SETXFLAGS_32	FS_IOC32_SETFLAGS
+#define XFS_IOC_GETVERSION_32	FS_IOC32_GETVERSION
+
+/*
+ * On intel, even if sizes match, alignment and/or padding may differ.
+ */
+#if defined(CONFIG_IA64) || defined(CONFIG_X86_64)
+#define BROKEN_X86_ALIGNMENT
+#define __compat_packed __attribute__((packed))
+#else
+#define __compat_packed
+#endif
+
+typedef struct compat_xfs_bstime {
+	compat_time_t	tv_sec;		/* seconds		*/
+	__s32		tv_nsec;	/* and nanoseconds	*/
+} compat_xfs_bstime_t;
+
+typedef struct compat_xfs_bstat {
+	__u64		bs_ino;		/* inode number			*/
+	__u16		bs_mode;	/* type and mode		*/
+	__u16		bs_nlink;	/* number of links		*/
+	__u32		bs_uid;		/* user id			*/
+	__u32		bs_gid;		/* group id			*/
+	__u32		bs_rdev;	/* device value			*/
+	__s32		bs_blksize;	/* block size			*/
+	__s64		bs_size;	/* file size			*/
+	compat_xfs_bstime_t bs_atime;	/* access time			*/
+	compat_xfs_bstime_t bs_mtime;	/* modify time			*/
+	compat_xfs_bstime_t bs_ctime;	/* inode change time		*/
+	int64_t		bs_blocks;	/* number of blocks		*/
+	__u32		bs_xflags;	/* extended flags		*/
+	__s32		bs_extsize;	/* extent size			*/
+	__s32		bs_extents;	/* number of extents		*/
+	__u32		bs_gen;		/* generation count		*/
+	__u16		bs_projid;	/* project id			*/
+	unsigned char	bs_pad[14];	/* pad space, unused		*/
+	__u32		bs_dmevmask;	/* DMIG event mask		*/
+	__u16		bs_dmstate;	/* DMIG state info		*/
+	__u16		bs_aextents;	/* attribute number of extents	*/
+} __compat_packed compat_xfs_bstat_t;
+
+typedef struct compat_xfs_fsop_bulkreq {
+	compat_uptr_t	lastip;		/* last inode # pointer		*/
+	__s32		icount;		/* count of entries in buffer	*/
+	compat_uptr_t	ubuffer;	/* user buffer for inode desc.	*/
+	compat_uptr_t	ocount;		/* output count pointer		*/
+} compat_xfs_fsop_bulkreq_t;
+
+#define XFS_IOC_FSBULKSTAT_32 \
+	_IOWR('X', 101, struct compat_xfs_fsop_bulkreq)
+#define XFS_IOC_FSBULKSTAT_SINGLE_32 \
+	_IOWR('X', 102, struct compat_xfs_fsop_bulkreq)
+#define XFS_IOC_FSINUMBERS_32 \
+	_IOWR('X', 103, struct compat_xfs_fsop_bulkreq)
+
+typedef struct compat_xfs_fsop_handlereq {
+	__u32		fd;		/* fd for FD_TO_HANDLE		*/
+	compat_uptr_t	path;		/* user pathname		*/
+	__u32		oflags;		/* open flags			*/
+	compat_uptr_t	ihandle;	/* user supplied handle		*/
+	__u32		ihandlen;	/* user supplied length		*/
+	compat_uptr_t	ohandle;	/* user buffer for handle	*/
+	compat_uptr_t	ohandlen;	/* user buffer length		*/
+} compat_xfs_fsop_handlereq_t;
+
+#define XFS_IOC_PATH_TO_FSHANDLE_32 \
+	_IOWR('X', 104, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_PATH_TO_HANDLE_32 \
+	_IOWR('X', 105, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_FD_TO_HANDLE_32 \
+	_IOWR('X', 106, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_OPEN_BY_HANDLE_32 \
+	_IOWR('X', 107, struct compat_xfs_fsop_handlereq)
+#define XFS_IOC_READLINK_BY_HANDLE_32 \
+	_IOWR('X', 108, struct compat_xfs_fsop_handlereq)
+
+/* The bstat field in the swapext struct needs translation */
+typedef struct compat_xfs_swapext {
+	__int64_t		sx_version;	/* version */
+	__int64_t		sx_fdtarget;	/* fd of target file */
+	__int64_t		sx_fdtmp;	/* fd of tmp file */
+	xfs_off_t		sx_offset;	/* offset into file */
+	xfs_off_t		sx_length;	/* leng from offset */
+	char			sx_pad[16];	/* pad space, unused */
+	compat_xfs_bstat_t	sx_stat;	/* stat of target b4 copy */
+} __compat_packed compat_xfs_swapext_t;
+
+#define XFS_IOC_SWAPEXT_32	_IOWR('X', 109, struct compat_xfs_swapext)
+
+typedef struct compat_xfs_fsop_attrlist_handlereq {
+	struct compat_xfs_fsop_handlereq hreq; /* handle interface structure */
+	struct xfs_attrlist_cursor	pos; /* opaque cookie, list offset */
+	__u32				flags;	/* which namespace to use */
+	__u32				buflen;	/* length of buffer supplied */
+	compat_uptr_t			buffer;	/* returned names */
+} __compat_packed compat_xfs_fsop_attrlist_handlereq_t;
+
+/* Note: actually this is read/write */
+#define XFS_IOC_ATTRLIST_BY_HANDLE_32 \
+	_IOW('X', 122, struct compat_xfs_fsop_attrlist_handlereq)
+
+/* am_opcodes defined in xfs_fs.h */
+typedef struct compat_xfs_attr_multiop {
+	__u32		am_opcode;
+	__s32		am_error;
+	compat_uptr_t	am_attrname;
+	compat_uptr_t	am_attrvalue;
+	__u32		am_length;
+	__u32		am_flags;
+} compat_xfs_attr_multiop_t;
+
+typedef struct compat_xfs_fsop_attrmulti_handlereq {
+	struct compat_xfs_fsop_handlereq hreq; /* handle interface structure */
+	__u32				opcount;/* count of following multiop */
+	/* ptr to compat_xfs_attr_multiop */
+	compat_uptr_t			ops; /* attr_multi data */
+} compat_xfs_fsop_attrmulti_handlereq_t;
+
+#define XFS_IOC_ATTRMULTI_BY_HANDLE_32 \
+	_IOW('X', 123, struct compat_xfs_fsop_attrmulti_handlereq)
+
+typedef struct compat_xfs_fsop_setdm_handlereq {
+	struct compat_xfs_fsop_handlereq hreq;	/* handle information   */
+	/* ptr to struct fsdmidata */
+	compat_uptr_t			data;	/* DMAPI data   */
+} compat_xfs_fsop_setdm_handlereq_t;
+
+#define XFS_IOC_FSSETDM_BY_HANDLE_32 \
+	_IOW('X', 121, struct compat_xfs_fsop_setdm_handlereq)
+
+#ifdef BROKEN_X86_ALIGNMENT
+/* on ia32 l_start is on a 32-bit boundary */
+typedef struct compat_xfs_flock64 {
+	__s16		l_type;
+	__s16		l_whence;
+	__s64		l_start	__attribute__((packed));
+			/* len == 0 means until end of file */
+	__s64		l_len __attribute__((packed));
+	__s32		l_sysid;
+	__u32		l_pid;
+	__s32		l_pad[4];	/* reserve area */
+} compat_xfs_flock64_t;
+
+#define XFS_IOC_ALLOCSP_32	_IOW('X', 10, struct compat_xfs_flock64)
+#define XFS_IOC_FREESP_32	_IOW('X', 11, struct compat_xfs_flock64)
+#define XFS_IOC_ALLOCSP64_32	_IOW('X', 36, struct compat_xfs_flock64)
+#define XFS_IOC_FREESP64_32	_IOW('X', 37, struct compat_xfs_flock64)
+#define XFS_IOC_RESVSP_32	_IOW('X', 40, struct compat_xfs_flock64)
+#define XFS_IOC_UNRESVSP_32	_IOW('X', 41, struct compat_xfs_flock64)
+#define XFS_IOC_RESVSP64_32	_IOW('X', 42, struct compat_xfs_flock64)
+#define XFS_IOC_UNRESVSP64_32	_IOW('X', 43, struct compat_xfs_flock64)
+
+typedef struct compat_xfs_fsop_geom_v1 {
+	__u32		blocksize;	/* filesystem (data) block size */
+	__u32		rtextsize;	/* realtime extent size		*/
+	__u32		agblocks;	/* fsblocks in an AG		*/
+	__u32		agcount;	/* number of allocation groups	*/
+	__u32		logblocks;	/* fsblocks in the log		*/
+	__u32		sectsize;	/* (data) sector size, bytes	*/
+	__u32		inodesize;	/* inode size in bytes		*/
+	__u32		imaxpct;	/* max allowed inode space(%)	*/
+	__u64		datablocks;	/* fsblocks in data subvolume	*/
+	__u64		rtblocks;	/* fsblocks in realtime subvol	*/
+	__u64		rtextents;	/* rt extents in realtime subvol*/
+	__u64		logstart;	/* starting fsblock of the log	*/
+	unsigned char	uuid[16];	/* unique id of the filesystem	*/
+	__u32		sunit;		/* stripe unit, fsblocks	*/
+	__u32		swidth;		/* stripe width, fsblocks	*/
+	__s32		version;	/* structure version		*/
+	__u32		flags;		/* superblock version flags	*/
+	__u32		logsectsize;	/* log sector size, bytes	*/
+	__u32		rtsectsize;	/* realtime sector size, bytes	*/
+	__u32		dirblocksize;	/* directory block size, bytes	*/
+} __attribute__((packed)) compat_xfs_fsop_geom_v1_t;
+
+#define XFS_IOC_FSGEOMETRY_V1_32  \
+	_IOR('X', 100, struct compat_xfs_fsop_geom_v1)
+
+typedef struct compat_xfs_inogrp {
+	__u64		xi_startino;	/* starting inode number	*/
+	__s32		xi_alloccount;	/* # bits set in allocmask	*/
+	__u64		xi_allocmask;	/* mask of allocated inodes	*/
+} __attribute__((packed)) compat_xfs_inogrp_t;
+
+/* These growfs input structures have padding on the end, so must translate */
+typedef struct compat_xfs_growfs_data {
+	__u64		newblocks;	/* new data subvol size, fsblocks */
+	__u32		imaxpct;	/* new inode space percentage limit */
+} __attribute__((packed)) compat_xfs_growfs_data_t;
+
+typedef struct compat_xfs_growfs_rt {
+	__u64		newblocks;	/* new realtime size, fsblocks */
+	__u32		extsize;	/* new realtime extent size, fsblocks */
+} __attribute__((packed)) compat_xfs_growfs_rt_t;
+
+#define XFS_IOC_FSGROWFSDATA_32 _IOW('X', 110, struct compat_xfs_growfs_data)
+#define XFS_IOC_FSGROWFSRT_32   _IOW('X', 112, struct compat_xfs_growfs_rt)
+
+#endif /* BROKEN_X86_ALIGNMENT */
 
 #endif /* __XFS_IOCTL32_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index 095d271f343..7aa53fefc67 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -53,6 +53,7 @@
 #include <linux/namei.h>
 #include <linux/security.h>
 #include <linux/falloc.h>
+#include <linux/fiemap.h>
 
 /*
  * Bring the atime in the XFS inode uptodate.
@@ -64,14 +65,14 @@ xfs_synchronize_atime(
 {
 	struct inode	*inode = VFS_I(ip);
 
-	if (inode) {
+	if (!(inode->i_state & I_CLEAR)) {
 		ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec;
 		ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec;
 	}
 }
 
 /*
- * If the linux inode exists, mark it dirty.
+ * If the linux inode is valid, mark it dirty.
  * Used when commiting a dirty inode into a transaction so that
  * the inode will get written back by the linux code
  */
@@ -81,7 +82,7 @@ xfs_mark_inode_dirty_sync(
 {
 	struct inode	*inode = VFS_I(ip);
 
-	if (inode)
+	if (!(inode->i_state & (I_WILL_FREE|I_FREEING|I_CLEAR)))
 		mark_inode_dirty_sync(inode);
 }
 
@@ -128,7 +129,7 @@ xfs_ichgtime(
 	if (sync_it) {
 		SYNCHRONIZE();
 		ip->i_update_core = 1;
-		mark_inode_dirty_sync(inode);
+		xfs_mark_inode_dirty_sync(ip);
 	}
 }
 
@@ -158,8 +159,6 @@ xfs_init_security(
 	}
 
 	error = xfs_attr_set(ip, name, value, length, ATTR_SECURE);
-	if (!error)
-		xfs_iflags_set(ip, XFS_IMODIFIED);
 
 	kfree(name);
 	kfree(value);
@@ -260,7 +259,6 @@ xfs_vn_mknod(
 		error = _ACL_INHERIT(inode, mode, default_acl);
 		if (unlikely(error))
 			goto out_cleanup_inode;
-		xfs_iflags_set(ip, XFS_IMODIFIED);
 		_ACL_FREE(default_acl);
 	}
 
@@ -366,21 +364,17 @@ xfs_vn_link(
 	struct inode	*dir,
 	struct dentry	*dentry)
 {
-	struct inode	*inode;	/* inode of guy being linked to */
+	struct inode	*inode = old_dentry->d_inode;
 	struct xfs_name	name;
 	int		error;
 
-	inode = old_dentry->d_inode;
 	xfs_dentry_to_name(&name, dentry);
 
-	igrab(inode);
 	error = xfs_link(XFS_I(dir), XFS_I(inode), &name);
-	if (unlikely(error)) {
-		iput(inode);
+	if (unlikely(error))
 		return -error;
-	}
 
-	xfs_iflags_set(XFS_I(dir), XFS_IMODIFIED);
+	atomic_inc(&inode->i_count);
 	d_instantiate(dentry, inode);
 	return 0;
 }
@@ -601,7 +595,7 @@ xfs_vn_setattr(
 	struct dentry	*dentry,
 	struct iattr	*iattr)
 {
-	return -xfs_setattr(XFS_I(dentry->d_inode), iattr, 0, NULL);
+	return -xfs_setattr(XFS_I(dentry->d_inode), iattr, 0);
 }
 
 /*
@@ -642,7 +636,7 @@ xfs_vn_fallocate(
 
 	xfs_ilock(ip, XFS_IOLOCK_EXCL);
 	error = xfs_change_file_space(ip, XFS_IOC_RESVSP, &bf,
-				      0, NULL, XFS_ATTR_NOLOCK);
+				      0, XFS_ATTR_NOLOCK);
 	if (!error && !(mode & FALLOC_FL_KEEP_SIZE) &&
 	    offset + len > i_size_read(inode))
 		new_size = offset + len;
@@ -653,7 +647,7 @@ xfs_vn_fallocate(
 
 		iattr.ia_valid = ATTR_SIZE;
 		iattr.ia_size = new_size;
-		error = xfs_setattr(ip, &iattr, XFS_ATTR_NOLOCK, NULL);
+		error = xfs_setattr(ip, &iattr, XFS_ATTR_NOLOCK);
 	}
 
 	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
@@ -661,6 +655,88 @@ out_error:
 	return error;
 }
 
+#define XFS_FIEMAP_FLAGS	(FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
+
+/*
+ * Call fiemap helper to fill in user data.
+ * Returns positive errors to xfs_getbmap.
+ */
+STATIC int
+xfs_fiemap_format(
+	void			**arg,
+	struct getbmapx		*bmv,
+	int			*full)
+{
+	int			error;
+	struct fiemap_extent_info *fieinfo = *arg;
+	u32			fiemap_flags = 0;
+	u64			logical, physical, length;
+
+	/* Do nothing for a hole */
+	if (bmv->bmv_block == -1LL)
+		return 0;
+
+	logical = BBTOB(bmv->bmv_offset);
+	physical = BBTOB(bmv->bmv_block);
+	length = BBTOB(bmv->bmv_length);
+
+	if (bmv->bmv_oflags & BMV_OF_PREALLOC)
+		fiemap_flags |= FIEMAP_EXTENT_UNWRITTEN;
+	else if (bmv->bmv_oflags & BMV_OF_DELALLOC) {
+		fiemap_flags |= FIEMAP_EXTENT_DELALLOC;
+		physical = 0;   /* no block yet */
+	}
+	if (bmv->bmv_oflags & BMV_OF_LAST)
+		fiemap_flags |= FIEMAP_EXTENT_LAST;
+
+	error = fiemap_fill_next_extent(fieinfo, logical, physical,
+					length, fiemap_flags);
+	if (error > 0) {
+		error = 0;
+		*full = 1;	/* user array now full */
+	}
+
+	return -error;
+}
+
+STATIC int
+xfs_vn_fiemap(
+	struct inode		*inode,
+	struct fiemap_extent_info *fieinfo,
+	u64			start,
+	u64			length)
+{
+	xfs_inode_t		*ip = XFS_I(inode);
+	struct getbmapx		bm;
+	int			error;
+
+	error = fiemap_check_flags(fieinfo, XFS_FIEMAP_FLAGS);
+	if (error)
+		return error;
+
+	/* Set up bmap header for xfs internal routine */
+	bm.bmv_offset = BTOBB(start);
+	/* Special case for whole file */
+	if (length == FIEMAP_MAX_OFFSET)
+		bm.bmv_length = -1LL;
+	else
+		bm.bmv_length = BTOBB(length);
+
+	/* our formatter will tell xfs_getbmap when to stop. */
+	bm.bmv_count = MAXEXTNUM;
+	bm.bmv_iflags = BMV_IF_PREALLOC;
+	if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR)
+		bm.bmv_iflags |= BMV_IF_ATTRFORK;
+	if (!(fieinfo->fi_flags & FIEMAP_FLAG_SYNC))
+		bm.bmv_iflags |= BMV_IF_DELALLOC;
+
+	error = xfs_getbmap(ip, &bm, xfs_fiemap_format, fieinfo);
+	if (error)
+		return -error;
+
+	return 0;
+}
+
 static const struct inode_operations xfs_inode_operations = {
 	.permission		= xfs_vn_permission,
 	.truncate		= xfs_vn_truncate,
@@ -671,6 +747,7 @@ static const struct inode_operations xfs_inode_operations = {
 	.removexattr		= generic_removexattr,
 	.listxattr		= xfs_vn_listxattr,
 	.fallocate		= xfs_vn_fallocate,
+	.fiemap			= xfs_vn_fiemap,
 };
 
 static const struct inode_operations xfs_dir_inode_operations = {
@@ -766,12 +843,20 @@ xfs_diflags_to_iflags(
  * When reading existing inodes from disk this is called directly
  * from xfs_iget, when creating a new inode it is called from
  * xfs_ialloc after setting up the inode.
+ *
+ * We are always called with an uninitialised linux inode here.
+ * We need to initialise the necessary fields and take a reference
+ * on it.
  */
 void
 xfs_setup_inode(
 	struct xfs_inode	*ip)
 {
-	struct inode		*inode = ip->i_vnode;
+	struct inode		*inode = &ip->i_vnode;
+
+	inode->i_ino = ip->i_ino;
+	inode->i_state = I_NEW|I_LOCK;
+	inode_add_to_lists(ip->i_mount->m_super, inode);
 
 	inode->i_mode	= ip->i_d.di_mode;
 	inode->i_nlink	= ip->i_d.di_nlink;
@@ -799,7 +884,6 @@ xfs_setup_inode(
 	inode->i_ctime.tv_sec	= ip->i_d.di_ctime.t_sec;
 	inode->i_ctime.tv_nsec	= ip->i_d.di_ctime.t_nsec;
 	xfs_diflags_to_iflags(inode, ip);
-	xfs_iflags_clear(ip, XFS_IMODIFIED);
 
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFREG:
diff --git a/fs/xfs/linux-2.6/xfs_iops.h b/fs/xfs/linux-2.6/xfs_iops.h
index 8b1a1e31dc2..ef41c92ce66 100644
--- a/fs/xfs/linux-2.6/xfs_iops.h
+++ b/fs/xfs/linux-2.6/xfs_iops.h
@@ -22,7 +22,6 @@ struct xfs_inode;
 
 extern const struct file_operations xfs_file_operations;
 extern const struct file_operations xfs_dir_file_operations;
-extern const struct file_operations xfs_invis_file_operations;
 
 extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size);
 
diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h
index cc0f7b3a979..507492d6dcc 100644
--- a/fs/xfs/linux-2.6/xfs_linux.h
+++ b/fs/xfs/linux-2.6/xfs_linux.h
@@ -21,18 +21,12 @@
 #include <linux/types.h>
 
 /*
- * Some types are conditional depending on the target system.
  * XFS_BIG_BLKNOS needs block layer disk addresses to be 64 bits.
- * XFS_BIG_INUMS needs the VFS inode number to be 64 bits, as well
- * as requiring XFS_BIG_BLKNOS to be set.
+ * XFS_BIG_INUMS requires XFS_BIG_BLKNOS to be set.
  */
 #if defined(CONFIG_LBD) || (BITS_PER_LONG == 64)
 # define XFS_BIG_BLKNOS	1
-# if BITS_PER_LONG == 64
-#  define XFS_BIG_INUMS	1
-# else
-#  define XFS_BIG_INUMS	0
-# endif
+# define XFS_BIG_INUMS	1
 #else
 # define XFS_BIG_BLKNOS	0
 # define XFS_BIG_INUMS	0
@@ -77,6 +71,7 @@
 #include <linux/spinlock.h>
 #include <linux/random.h>
 #include <linux/ctype.h>
+#include <linux/writeback.h>
 
 #include <asm/page.h>
 #include <asm/div64.h>
@@ -85,7 +80,6 @@
 #include <asm/byteorder.h>
 #include <asm/unaligned.h>
 
-#include <xfs_vfs.h>
 #include <xfs_cred.h>
 #include <xfs_vnode.h>
 #include <xfs_stats.h>
@@ -107,7 +101,6 @@
 #undef  HAVE_PERCPU_SB	/* per cpu superblock counters are a 2.6 feature */
 #endif
 
-#define restricted_chown	xfs_params.restrict_chown.val
 #define irix_sgid_inherit	xfs_params.sgid_inherit.val
 #define irix_symlink_mode	xfs_params.symlink_mode.val
 #define xfs_panic_mask		xfs_params.panic_mask.val
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
index 1957e5357d0..7e90daa0d1d 100644
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ b/fs/xfs/linux-2.6/xfs_lrw.c
@@ -51,7 +51,6 @@
 #include "xfs_vnodeops.h"
 
 #include <linux/capability.h>
-#include <linux/mount.h>
 #include <linux/writeback.h>
 
 
@@ -243,7 +242,7 @@ xfs_read(
 
 	if (unlikely(ioflags & IO_ISDIRECT)) {
 		if (inode->i_mapping->nrpages)
-			ret = xfs_flushinval_pages(ip, (*offset & PAGE_CACHE_MASK),
+			ret = -xfs_flushinval_pages(ip, (*offset & PAGE_CACHE_MASK),
 						    -1, FI_REMAPF_LOCKED);
 		mutex_unlock(&inode->i_mutex);
 		if (ret) {
@@ -668,15 +667,8 @@ start:
 	if (new_size > xip->i_size)
 		xip->i_new_size = new_size;
 
-	/*
-	 * We're not supposed to change timestamps in readonly-mounted
-	 * filesystems.  Throw it away if anyone asks us.
-	 */
-	if (likely(!(ioflags & IO_INVIS) &&
-		   !mnt_want_write(file->f_path.mnt))) {
+	if (likely(!(ioflags & IO_INVIS)))
 		xfs_ichgtime(xip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
-		mnt_drop_write(file->f_path.mnt);
-	}
 
 	/*
 	 * If the offset is beyond the size of the file, we have a couple
@@ -715,7 +707,6 @@ start:
 		}
 	}
 
-retry:
 	/* We can write back this queue in page reclaim */
 	current->backing_dev_info = mapping->backing_dev_info;
 
@@ -771,6 +762,17 @@ retry:
 	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
 		ret = wait_on_sync_kiocb(iocb);
 
+	isize = i_size_read(inode);
+	if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
+		*offset = isize;
+
+	if (*offset > xip->i_size) {
+		xfs_ilock(xip, XFS_ILOCK_EXCL);
+		if (*offset > xip->i_size)
+			xip->i_size = *offset;
+		xfs_iunlock(xip, XFS_ILOCK_EXCL);
+	}
+
 	if (ret == -ENOSPC &&
 	    DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
 		xfs_iunlock(xip, iolock);
@@ -784,20 +786,7 @@ retry:
 		xfs_ilock(xip, iolock);
 		if (error)
 			goto out_unlock_internal;
-		pos = xip->i_size;
-		ret = 0;
-		goto retry;
-	}
-
-	isize = i_size_read(inode);
-	if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize))
-		*offset = isize;
-
-	if (*offset > xip->i_size) {
-		xfs_ilock(xip, XFS_ILOCK_EXCL);
-		if (*offset > xip->i_size)
-			xip->i_size = *offset;
-		xfs_iunlock(xip, XFS_ILOCK_EXCL);
+		goto start;
 	}
 
 	error = -ret;
@@ -855,13 +844,7 @@ retry:
 int
 xfs_bdstrat_cb(struct xfs_buf *bp)
 {
-	xfs_mount_t	*mp;
-
-	mp = XFS_BUF_FSPRIVATE3(bp, xfs_mount_t *);
-	if (!XFS_FORCED_SHUTDOWN(mp)) {
-		xfs_buf_iorequest(bp);
-		return 0;
-	} else {
+	if (XFS_FORCED_SHUTDOWN(bp->b_mount)) {
 		xfs_buftrace("XFS__BDSTRAT IOERROR", bp);
 		/*
 		 * Metadata write that didn't get logged but
@@ -874,6 +857,9 @@ xfs_bdstrat_cb(struct xfs_buf *bp)
 		else
 			return (xfs_bioerror(bp));
 	}
+
+	xfs_buf_iorequest(bp);
+	return 0;
 }
 
 /*
diff --git a/fs/xfs/linux-2.6/xfs_stats.c b/fs/xfs/linux-2.6/xfs_stats.c
index 3d5b67c075c..c3526d445f6 100644
--- a/fs/xfs/linux-2.6/xfs_stats.c
+++ b/fs/xfs/linux-2.6/xfs_stats.c
@@ -53,11 +53,15 @@ xfs_read_xfsstats(
 		{ "icluster",		XFSSTAT_END_INODE_CLUSTER	},
 		{ "vnodes",		XFSSTAT_END_VNODE_OPS		},
 		{ "buf",		XFSSTAT_END_BUF			},
+		{ "abtb2",		XFSSTAT_END_ABTB_V2		},
+		{ "abtc2",		XFSSTAT_END_ABTC_V2		},
+		{ "bmbt2",		XFSSTAT_END_BMBT_V2		},
+		{ "ibt2",		XFSSTAT_END_IBT_V2		},
 	};
 
 	/* Loop over all stats groups */
 	for (i=j=len = 0; i < ARRAY_SIZE(xstats); i++) {
-		len += sprintf(buffer + len, xstats[i].desc);
+		len += sprintf(buffer + len, "%s", xstats[i].desc);
 		/* inner loop does each group */
 		while (j < xstats[i].endpoint) {
 			val = 0;
diff --git a/fs/xfs/linux-2.6/xfs_stats.h b/fs/xfs/linux-2.6/xfs_stats.h
index e83820febc9..736854b1ca1 100644
--- a/fs/xfs/linux-2.6/xfs_stats.h
+++ b/fs/xfs/linux-2.6/xfs_stats.h
@@ -118,6 +118,71 @@ struct xfsstats {
 	__uint32_t		xb_page_retries;
 	__uint32_t		xb_page_found;
 	__uint32_t		xb_get_read;
+/* Version 2 btree counters */
+#define XFSSTAT_END_ABTB_V2		(XFSSTAT_END_BUF+15)
+	__uint32_t		xs_abtb_2_lookup;
+	__uint32_t		xs_abtb_2_compare;
+	__uint32_t		xs_abtb_2_insrec;
+	__uint32_t		xs_abtb_2_delrec;
+	__uint32_t		xs_abtb_2_newroot;
+	__uint32_t		xs_abtb_2_killroot;
+	__uint32_t		xs_abtb_2_increment;
+	__uint32_t		xs_abtb_2_decrement;
+	__uint32_t		xs_abtb_2_lshift;
+	__uint32_t		xs_abtb_2_rshift;
+	__uint32_t		xs_abtb_2_split;
+	__uint32_t		xs_abtb_2_join;
+	__uint32_t		xs_abtb_2_alloc;
+	__uint32_t		xs_abtb_2_free;
+	__uint32_t		xs_abtb_2_moves;
+#define XFSSTAT_END_ABTC_V2		(XFSSTAT_END_ABTB_V2+15)
+	__uint32_t		xs_abtc_2_lookup;
+	__uint32_t		xs_abtc_2_compare;
+	__uint32_t		xs_abtc_2_insrec;
+	__uint32_t		xs_abtc_2_delrec;
+	__uint32_t		xs_abtc_2_newroot;
+	__uint32_t		xs_abtc_2_killroot;
+	__uint32_t		xs_abtc_2_increment;
+	__uint32_t		xs_abtc_2_decrement;
+	__uint32_t		xs_abtc_2_lshift;
+	__uint32_t		xs_abtc_2_rshift;
+	__uint32_t		xs_abtc_2_split;
+	__uint32_t		xs_abtc_2_join;
+	__uint32_t		xs_abtc_2_alloc;
+	__uint32_t		xs_abtc_2_free;
+	__uint32_t		xs_abtc_2_moves;
+#define XFSSTAT_END_BMBT_V2		(XFSSTAT_END_ABTC_V2+15)
+	__uint32_t		xs_bmbt_2_lookup;
+	__uint32_t		xs_bmbt_2_compare;
+	__uint32_t		xs_bmbt_2_insrec;
+	__uint32_t		xs_bmbt_2_delrec;
+	__uint32_t		xs_bmbt_2_newroot;
+	__uint32_t		xs_bmbt_2_killroot;
+	__uint32_t		xs_bmbt_2_increment;
+	__uint32_t		xs_bmbt_2_decrement;
+	__uint32_t		xs_bmbt_2_lshift;
+	__uint32_t		xs_bmbt_2_rshift;
+	__uint32_t		xs_bmbt_2_split;
+	__uint32_t		xs_bmbt_2_join;
+	__uint32_t		xs_bmbt_2_alloc;
+	__uint32_t		xs_bmbt_2_free;
+	__uint32_t		xs_bmbt_2_moves;
+#define XFSSTAT_END_IBT_V2		(XFSSTAT_END_BMBT_V2+15)
+	__uint32_t		xs_ibt_2_lookup;
+	__uint32_t		xs_ibt_2_compare;
+	__uint32_t		xs_ibt_2_insrec;
+	__uint32_t		xs_ibt_2_delrec;
+	__uint32_t		xs_ibt_2_newroot;
+	__uint32_t		xs_ibt_2_killroot;
+	__uint32_t		xs_ibt_2_increment;
+	__uint32_t		xs_ibt_2_decrement;
+	__uint32_t		xs_ibt_2_lshift;
+	__uint32_t		xs_ibt_2_rshift;
+	__uint32_t		xs_ibt_2_split;
+	__uint32_t		xs_ibt_2_join;
+	__uint32_t		xs_ibt_2_alloc;
+	__uint32_t		xs_ibt_2_free;
+	__uint32_t		xs_ibt_2_moves;
 /* Extra precision counters */
 	__uint64_t		xs_xstrat_bytes;
 	__uint64_t		xs_write_bytes;
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index 37ebe36056e..95a97108036 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -18,7 +18,6 @@
 #include "xfs.h"
 #include "xfs_bit.h"
 #include "xfs_log.h"
-#include "xfs_clnt.h"
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
@@ -36,6 +35,7 @@
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_btree.h"
+#include "xfs_btree_trace.h"
 #include "xfs_ialloc.h"
 #include "xfs_bmap.h"
 #include "xfs_rtalloc.h"
@@ -48,7 +48,6 @@
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
 #include "xfs_vnodeops.h"
-#include "xfs_vfsops.h"
 #include "xfs_version.h"
 #include "xfs_log_priv.h"
 #include "xfs_trans_priv.h"
@@ -58,6 +57,7 @@
 #include "xfs_extfree_item.h"
 #include "xfs_mru_cache.h"
 #include "xfs_inode_item.h"
+#include "xfs_sync.h"
 
 #include <linux/namei.h>
 #include <linux/init.h>
@@ -70,36 +70,9 @@
 
 static struct quotactl_ops xfs_quotactl_operations;
 static struct super_operations xfs_super_operations;
-static kmem_zone_t *xfs_vnode_zone;
 static kmem_zone_t *xfs_ioend_zone;
 mempool_t *xfs_ioend_pool;
 
-STATIC struct xfs_mount_args *
-xfs_args_allocate(
-	struct super_block	*sb,
-	int			silent)
-{
-	struct xfs_mount_args	*args;
-
-	args = kzalloc(sizeof(struct xfs_mount_args), GFP_KERNEL);
-	if (!args)
-		return NULL;
-
-	args->logbufs = args->logbufsize = -1;
-	strncpy(args->fsname, sb->s_id, MAXNAMELEN);
-
-	/* Copy the already-parsed mount(2) flags we're interested in */
-	if (sb->s_flags & MS_DIRSYNC)
-		args->flags |= XFSMNT_DIRSYNC;
-	if (sb->s_flags & MS_SYNCHRONOUS)
-		args->flags |= XFSMNT_WSYNC;
-	if (silent)
-		args->flags |= XFSMNT_QUIET;
-	args->flags |= XFSMNT_32BITINODES;
-
-	return args;
-}
-
 #define MNTOPT_LOGBUFS	"logbufs"	/* number of XFS log buffers */
 #define MNTOPT_LOGBSIZE	"logbsize"	/* size of XFS log buffers */
 #define MNTOPT_LOGDEV	"logdev"	/* log device */
@@ -188,26 +161,54 @@ suffix_strtoul(char *s, char **endp, unsigned int base)
 	return simple_strtoul((const char *)s, endp, base) << shift_left_factor;
 }
 
+/*
+ * This function fills in xfs_mount_t fields based on mount args.
+ * Note: the superblock has _not_ yet been read in.
+ *
+ * Note that this function leaks the various device name allocations on
+ * failure.  The caller takes care of them.
+ */
 STATIC int
 xfs_parseargs(
 	struct xfs_mount	*mp,
 	char			*options,
-	struct xfs_mount_args	*args,
-	int			update)
+	char			**mtpt)
 {
+	struct super_block	*sb = mp->m_super;
 	char			*this_char, *value, *eov;
-	int			dsunit, dswidth, vol_dsunit, vol_dswidth;
-	int			iosize;
+	int			dsunit = 0;
+	int			dswidth = 0;
+	int			iosize = 0;
 	int			dmapi_implies_ikeep = 1;
+	uchar_t			iosizelog = 0;
+
+	/*
+	 * Copy binary VFS mount flags we are interested in.
+	 */
+	if (sb->s_flags & MS_RDONLY)
+		mp->m_flags |= XFS_MOUNT_RDONLY;
+	if (sb->s_flags & MS_DIRSYNC)
+		mp->m_flags |= XFS_MOUNT_DIRSYNC;
+	if (sb->s_flags & MS_SYNCHRONOUS)
+		mp->m_flags |= XFS_MOUNT_WSYNC;
+
+	/*
+	 * Set some default flags that could be cleared by the mount option
+	 * parsing.
+	 */
+	mp->m_flags |= XFS_MOUNT_BARRIER;
+	mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
+	mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
 
-	args->flags |= XFSMNT_BARRIER;
-	args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
+	/*
+	 * These can be overridden by the mount option parsing.
+	 */
+	mp->m_logbufs = -1;
+	mp->m_logbsize = -1;
 
 	if (!options)
 		goto done;
 
-	iosize = dsunit = dswidth = vol_dsunit = vol_dswidth = 0;
-
 	while ((this_char = strsep(&options, ",")) != NULL) {
 		if (!*this_char)
 			continue;
@@ -221,7 +222,7 @@ xfs_parseargs(
 					this_char);
 				return EINVAL;
 			}
-			args->logbufs = simple_strtoul(value, &eov, 10);
+			mp->m_logbufs = simple_strtoul(value, &eov, 10);
 		} else if (!strcmp(this_char, MNTOPT_LOGBSIZE)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -229,7 +230,7 @@ xfs_parseargs(
 					this_char);
 				return EINVAL;
 			}
-			args->logbufsize = suffix_strtoul(value, &eov, 10);
+			mp->m_logbsize = suffix_strtoul(value, &eov, 10);
 		} else if (!strcmp(this_char, MNTOPT_LOGDEV)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -237,7 +238,9 @@ xfs_parseargs(
 					this_char);
 				return EINVAL;
 			}
-			strncpy(args->logname, value, MAXNAMELEN);
+			mp->m_logname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
+			if (!mp->m_logname)
+				return ENOMEM;
 		} else if (!strcmp(this_char, MNTOPT_MTPT)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -245,7 +248,9 @@ xfs_parseargs(
 					this_char);
 				return EINVAL;
 			}
-			strncpy(args->mtpt, value, MAXNAMELEN);
+			*mtpt = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
+			if (!*mtpt)
+				return ENOMEM;
 		} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -253,7 +258,9 @@ xfs_parseargs(
 					this_char);
 				return EINVAL;
 			}
-			strncpy(args->rtname, value, MAXNAMELEN);
+			mp->m_rtname = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
+			if (!mp->m_rtname)
+				return ENOMEM;
 		} else if (!strcmp(this_char, MNTOPT_BIOSIZE)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -262,8 +269,7 @@ xfs_parseargs(
 				return EINVAL;
 			}
 			iosize = simple_strtoul(value, &eov, 10);
-			args->flags |= XFSMNT_IOSIZE;
-			args->iosizelog = (uint8_t) iosize;
+			iosizelog = ffs(iosize) - 1;
 		} else if (!strcmp(this_char, MNTOPT_ALLOCSIZE)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -272,8 +278,7 @@ xfs_parseargs(
 				return EINVAL;
 			}
 			iosize = suffix_strtoul(value, &eov, 10);
-			args->flags |= XFSMNT_IOSIZE;
-			args->iosizelog = ffs(iosize) - 1;
+			iosizelog = ffs(iosize) - 1;
 		} else if (!strcmp(this_char, MNTOPT_GRPID) ||
 			   !strcmp(this_char, MNTOPT_BSDGROUPS)) {
 			mp->m_flags |= XFS_MOUNT_GRPID;
@@ -281,23 +286,25 @@ xfs_parseargs(
 			   !strcmp(this_char, MNTOPT_SYSVGROUPS)) {
 			mp->m_flags &= ~XFS_MOUNT_GRPID;
 		} else if (!strcmp(this_char, MNTOPT_WSYNC)) {
-			args->flags |= XFSMNT_WSYNC;
+			mp->m_flags |= XFS_MOUNT_WSYNC;
 		} else if (!strcmp(this_char, MNTOPT_OSYNCISOSYNC)) {
-			args->flags |= XFSMNT_OSYNCISOSYNC;
+			mp->m_flags |= XFS_MOUNT_OSYNCISOSYNC;
 		} else if (!strcmp(this_char, MNTOPT_NORECOVERY)) {
-			args->flags |= XFSMNT_NORECOVERY;
+			mp->m_flags |= XFS_MOUNT_NORECOVERY;
 		} else if (!strcmp(this_char, MNTOPT_INO64)) {
-			args->flags |= XFSMNT_INO64;
-#if !XFS_BIG_INUMS
+#if XFS_BIG_INUMS
+			mp->m_flags |= XFS_MOUNT_INO64;
+			mp->m_inoadd = XFS_INO64_OFFSET;
+#else
 			cmn_err(CE_WARN,
 				"XFS: %s option not allowed on this system",
 				this_char);
 			return EINVAL;
 #endif
 		} else if (!strcmp(this_char, MNTOPT_NOALIGN)) {
-			args->flags |= XFSMNT_NOALIGN;
+			mp->m_flags |= XFS_MOUNT_NOALIGN;
 		} else if (!strcmp(this_char, MNTOPT_SWALLOC)) {
-			args->flags |= XFSMNT_SWALLOC;
+			mp->m_flags |= XFS_MOUNT_SWALLOC;
 		} else if (!strcmp(this_char, MNTOPT_SUNIT)) {
 			if (!value || !*value) {
 				cmn_err(CE_WARN,
@@ -315,7 +322,7 @@ xfs_parseargs(
 			}
 			dswidth = simple_strtoul(value, &eov, 10);
 		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
-			args->flags &= ~XFSMNT_32BITINODES;
+			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
 #if !XFS_BIG_INUMS
 			cmn_err(CE_WARN,
 				"XFS: %s option not allowed on this system",
@@ -323,56 +330,61 @@ xfs_parseargs(
 			return EINVAL;
 #endif
 		} else if (!strcmp(this_char, MNTOPT_NOUUID)) {
-			args->flags |= XFSMNT_NOUUID;
+			mp->m_flags |= XFS_MOUNT_NOUUID;
 		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {
-			args->flags |= XFSMNT_BARRIER;
+			mp->m_flags |= XFS_MOUNT_BARRIER;
 		} else if (!strcmp(this_char, MNTOPT_NOBARRIER)) {
-			args->flags &= ~XFSMNT_BARRIER;
+			mp->m_flags &= ~XFS_MOUNT_BARRIER;
 		} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
-			args->flags |= XFSMNT_IKEEP;
+			mp->m_flags |= XFS_MOUNT_IKEEP;
 		} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
 			dmapi_implies_ikeep = 0;
-			args->flags &= ~XFSMNT_IKEEP;
+			mp->m_flags &= ~XFS_MOUNT_IKEEP;
 		} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
-			args->flags2 &= ~XFSMNT2_COMPAT_IOSIZE;
+			mp->m_flags &= ~XFS_MOUNT_COMPAT_IOSIZE;
 		} else if (!strcmp(this_char, MNTOPT_NOLARGEIO)) {
-			args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
+			mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
 		} else if (!strcmp(this_char, MNTOPT_ATTR2)) {
-			args->flags |= XFSMNT_ATTR2;
+			mp->m_flags |= XFS_MOUNT_ATTR2;
 		} else if (!strcmp(this_char, MNTOPT_NOATTR2)) {
-			args->flags &= ~XFSMNT_ATTR2;
-			args->flags |= XFSMNT_NOATTR2;
+			mp->m_flags &= ~XFS_MOUNT_ATTR2;
+			mp->m_flags |= XFS_MOUNT_NOATTR2;
 		} else if (!strcmp(this_char, MNTOPT_FILESTREAM)) {
-			args->flags2 |= XFSMNT2_FILESTREAMS;
+			mp->m_flags |= XFS_MOUNT_FILESTREAMS;
 		} else if (!strcmp(this_char, MNTOPT_NOQUOTA)) {
-			args->flags &= ~(XFSMNT_UQUOTAENF|XFSMNT_UQUOTA);
-			args->flags &= ~(XFSMNT_GQUOTAENF|XFSMNT_GQUOTA);
+			mp->m_qflags &= ~(XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
+					  XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
+					  XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
+					  XFS_UQUOTA_ENFD | XFS_OQUOTA_ENFD);
 		} else if (!strcmp(this_char, MNTOPT_QUOTA) ||
 			   !strcmp(this_char, MNTOPT_UQUOTA) ||
 			   !strcmp(this_char, MNTOPT_USRQUOTA)) {
-			args->flags |= XFSMNT_UQUOTA | XFSMNT_UQUOTAENF;
+			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE |
+					 XFS_UQUOTA_ENFD);
 		} else if (!strcmp(this_char, MNTOPT_QUOTANOENF) ||
 			   !strcmp(this_char, MNTOPT_UQUOTANOENF)) {
-			args->flags |= XFSMNT_UQUOTA;
-			args->flags &= ~XFSMNT_UQUOTAENF;
+			mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
+			mp->m_qflags &= ~XFS_UQUOTA_ENFD;
 		} else if (!strcmp(this_char, MNTOPT_PQUOTA) ||
 			   !strcmp(this_char, MNTOPT_PRJQUOTA)) {
-			args->flags |= XFSMNT_PQUOTA | XFSMNT_PQUOTAENF;
+			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE |
+					 XFS_OQUOTA_ENFD);
 		} else if (!strcmp(this_char, MNTOPT_PQUOTANOENF)) {
-			args->flags |= XFSMNT_PQUOTA;
-			args->flags &= ~XFSMNT_PQUOTAENF;
+			mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
+			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
 		} else if (!strcmp(this_char, MNTOPT_GQUOTA) ||
 			   !strcmp(this_char, MNTOPT_GRPQUOTA)) {
-			args->flags |= XFSMNT_GQUOTA | XFSMNT_GQUOTAENF;
+			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE |
+					 XFS_OQUOTA_ENFD);
 		} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
-			args->flags |= XFSMNT_GQUOTA;
-			args->flags &= ~XFSMNT_GQUOTAENF;
+			mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
+			mp->m_qflags &= ~XFS_OQUOTA_ENFD;
 		} else if (!strcmp(this_char, MNTOPT_DMAPI)) {
-			args->flags |= XFSMNT_DMAPI;
+			mp->m_flags |= XFS_MOUNT_DMAPI;
 		} else if (!strcmp(this_char, MNTOPT_XDSM)) {
-			args->flags |= XFSMNT_DMAPI;
+			mp->m_flags |= XFS_MOUNT_DMAPI;
 		} else if (!strcmp(this_char, MNTOPT_DMI)) {
-			args->flags |= XFSMNT_DMAPI;
+			mp->m_flags |= XFS_MOUNT_DMAPI;
 		} else if (!strcmp(this_char, "ihashsize")) {
 			cmn_err(CE_WARN,
 	"XFS: ihashsize no longer used, option is deprecated.");
@@ -390,27 +402,29 @@ xfs_parseargs(
 		}
 	}
 
-	if (args->flags & XFSMNT_NORECOVERY) {
-		if ((mp->m_flags & XFS_MOUNT_RDONLY) == 0) {
-			cmn_err(CE_WARN,
-				"XFS: no-recovery mounts must be read-only.");
-			return EINVAL;
-		}
+	/*
+	 * no recovery flag requires a read-only mount
+	 */
+	if ((mp->m_flags & XFS_MOUNT_NORECOVERY) &&
+	    !(mp->m_flags & XFS_MOUNT_RDONLY)) {
+		cmn_err(CE_WARN, "XFS: no-recovery mounts must be read-only.");
+		return EINVAL;
 	}
 
-	if ((args->flags & XFSMNT_NOALIGN) && (dsunit || dswidth)) {
+	if ((mp->m_flags & XFS_MOUNT_NOALIGN) && (dsunit || dswidth)) {
 		cmn_err(CE_WARN,
 	"XFS: sunit and swidth options incompatible with the noalign option");
 		return EINVAL;
 	}
 
-	if ((args->flags & XFSMNT_GQUOTA) && (args->flags & XFSMNT_PQUOTA)) {
+	if ((mp->m_qflags & (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE)) &&
+	    (mp->m_qflags & (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE))) {
 		cmn_err(CE_WARN,
 			"XFS: cannot mount with both project and group quota");
 		return EINVAL;
 	}
 
-	if ((args->flags & XFSMNT_DMAPI) && *args->mtpt == '\0') {
+	if ((mp->m_flags & XFS_MOUNT_DMAPI) && (!*mtpt || *mtpt[0] == '\0')) {
 		printk("XFS: %s option needs the mount point option as well\n",
 			MNTOPT_DMAPI);
 		return EINVAL;
@@ -438,27 +452,66 @@ xfs_parseargs(
 	 * Note that if "ikeep" or "noikeep" mount options are
 	 * supplied, then they are honored.
 	 */
-	if ((args->flags & XFSMNT_DMAPI) && dmapi_implies_ikeep)
-		args->flags |= XFSMNT_IKEEP;
+	if ((mp->m_flags & XFS_MOUNT_DMAPI) && dmapi_implies_ikeep)
+		mp->m_flags |= XFS_MOUNT_IKEEP;
 
-	if ((args->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
+done:
+	if (!(mp->m_flags & XFS_MOUNT_NOALIGN)) {
+		/*
+		 * At this point the superblock has not been read
+		 * in, therefore we do not know the block size.
+		 * Before the mount call ends we will convert
+		 * these to FSBs.
+		 */
 		if (dsunit) {
-			args->sunit = dsunit;
-			args->flags |= XFSMNT_RETERR;
-		} else {
-			args->sunit = vol_dsunit;
+			mp->m_dalign = dsunit;
+			mp->m_flags |= XFS_MOUNT_RETERR;
 		}
-		dswidth ? (args->swidth = dswidth) :
-			  (args->swidth = vol_dswidth);
-	} else {
-		args->sunit = args->swidth = 0;
+
+		if (dswidth)
+			mp->m_swidth = dswidth;
+	}
+
+	if (mp->m_logbufs != -1 &&
+	    mp->m_logbufs != 0 &&
+	    (mp->m_logbufs < XLOG_MIN_ICLOGS ||
+	     mp->m_logbufs > XLOG_MAX_ICLOGS)) {
+		cmn_err(CE_WARN,
+			"XFS: invalid logbufs value: %d [not %d-%d]",
+			mp->m_logbufs, XLOG_MIN_ICLOGS, XLOG_MAX_ICLOGS);
+		return XFS_ERROR(EINVAL);
+	}
+	if (mp->m_logbsize != -1 &&
+	    mp->m_logbsize !=  0 &&
+	    (mp->m_logbsize < XLOG_MIN_RECORD_BSIZE ||
+	     mp->m_logbsize > XLOG_MAX_RECORD_BSIZE ||
+	     !is_power_of_2(mp->m_logbsize))) {
+		cmn_err(CE_WARN,
+	"XFS: invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
+			mp->m_logbsize);
+		return XFS_ERROR(EINVAL);
+	}
+
+	mp->m_fsname = kstrndup(sb->s_id, MAXNAMELEN, GFP_KERNEL);
+	if (!mp->m_fsname)
+		return ENOMEM;
+	mp->m_fsname_len = strlen(mp->m_fsname) + 1;
+
+	if (iosizelog) {
+		if (iosizelog > XFS_MAX_IO_LOG ||
+		    iosizelog < XFS_MIN_IO_LOG) {
+			cmn_err(CE_WARN,
+		"XFS: invalid log iosize: %d [not %d-%d]",
+				iosizelog, XFS_MIN_IO_LOG,
+				XFS_MAX_IO_LOG);
+			return XFS_ERROR(EINVAL);
+		}
+
+		mp->m_flags |= XFS_MOUNT_DFLT_IOSIZE;
+		mp->m_readio_log = iosizelog;
+		mp->m_writeio_log = iosizelog;
 	}
 
-done:
-	if (args->flags & XFSMNT_32BITINODES)
-		mp->m_flags |= XFS_MOUNT_SMALL_INUMS;
-	if (args->flags2)
-		args->flags |= XFSMNT_FLAGS2;
 	return 0;
 }
 
@@ -704,8 +757,7 @@ xfs_close_devices(
  */
 STATIC int
 xfs_open_devices(
-	struct xfs_mount	*mp,
-	struct xfs_mount_args	*args)
+	struct xfs_mount	*mp)
 {
 	struct block_device	*ddev = mp->m_super->s_bdev;
 	struct block_device	*logdev = NULL, *rtdev = NULL;
@@ -714,14 +766,14 @@ xfs_open_devices(
 	/*
 	 * Open real time and log devices - order is important.
 	 */
-	if (args->logname[0]) {
-		error = xfs_blkdev_get(mp, args->logname, &logdev);
+	if (mp->m_logname) {
+		error = xfs_blkdev_get(mp, mp->m_logname, &logdev);
 		if (error)
 			goto out;
 	}
 
-	if (args->rtname[0]) {
-		error = xfs_blkdev_get(mp, args->rtname, &rtdev);
+	if (mp->m_rtname) {
+		error = xfs_blkdev_get(mp, mp->m_rtname, &rtdev);
 		if (error)
 			goto out_close_logdev;
 
@@ -813,18 +865,18 @@ xfs_setup_devices(
  */
 void
 xfsaild_wakeup(
-	xfs_mount_t		*mp,
+	struct xfs_ail		*ailp,
 	xfs_lsn_t		threshold_lsn)
 {
-	mp->m_ail.xa_target = threshold_lsn;
-	wake_up_process(mp->m_ail.xa_task);
+	ailp->xa_target = threshold_lsn;
+	wake_up_process(ailp->xa_task);
 }
 
 int
 xfsaild(
 	void	*data)
 {
-	xfs_mount_t	*mp = (xfs_mount_t *)data;
+	struct xfs_ail	*ailp = data;
 	xfs_lsn_t	last_pushed_lsn = 0;
 	long		tout = 0;
 
@@ -836,11 +888,11 @@ xfsaild(
 		/* swsusp */
 		try_to_freeze();
 
-		ASSERT(mp->m_log);
-		if (XFS_FORCED_SHUTDOWN(mp))
+		ASSERT(ailp->xa_mount->m_log);
+		if (XFS_FORCED_SHUTDOWN(ailp->xa_mount))
 			continue;
 
-		tout = xfsaild_push(mp, &last_pushed_lsn);
+		tout = xfsaild_push(ailp, &last_pushed_lsn);
 	}
 
 	return 0;
@@ -848,43 +900,82 @@ xfsaild(
 
 int
 xfsaild_start(
-	xfs_mount_t	*mp)
+	struct xfs_ail	*ailp)
 {
-	mp->m_ail.xa_target = 0;
-	mp->m_ail.xa_task = kthread_run(xfsaild, mp, "xfsaild");
-	if (IS_ERR(mp->m_ail.xa_task))
-		return -PTR_ERR(mp->m_ail.xa_task);
+	ailp->xa_target = 0;
+	ailp->xa_task = kthread_run(xfsaild, ailp, "xfsaild");
+	if (IS_ERR(ailp->xa_task))
+		return -PTR_ERR(ailp->xa_task);
 	return 0;
 }
 
 void
 xfsaild_stop(
-	xfs_mount_t	*mp)
+	struct xfs_ail	*ailp)
 {
-	kthread_stop(mp->m_ail.xa_task);
+	kthread_stop(ailp->xa_task);
 }
 
 
-
+/* Catch misguided souls that try to use this interface on XFS */
 STATIC struct inode *
 xfs_fs_alloc_inode(
 	struct super_block	*sb)
 {
-	return kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
+	BUG();
+	return NULL;
 }
 
+/*
+ * Now that the generic code is guaranteed not to be accessing
+ * the linux inode, we can reclaim the inode.
+ */
 STATIC void
 xfs_fs_destroy_inode(
-	struct inode		*inode)
+	struct inode	*inode)
 {
-	kmem_zone_free(xfs_vnode_zone, inode);
+	xfs_inode_t		*ip = XFS_I(inode);
+
+	XFS_STATS_INC(vn_reclaim);
+	if (xfs_reclaim(ip))
+		panic("%s: cannot reclaim 0x%p\n", __func__, inode);
 }
 
+/*
+ * Slab object creation initialisation for the XFS inode.
+ * This covers only the idempotent fields in the XFS inode;
+ * all other fields need to be initialised on allocation
+ * from the slab. This avoids the need to repeatedly intialise
+ * fields in the xfs inode that left in the initialise state
+ * when freeing the inode.
+ */
 STATIC void
 xfs_fs_inode_init_once(
-	void			*vnode)
+	void			*inode)
 {
-	inode_init_once((struct inode *)vnode);
+	struct xfs_inode	*ip = inode;
+
+	memset(ip, 0, sizeof(struct xfs_inode));
+
+	/* vfs inode */
+	inode_init_once(VFS_I(ip));
+
+	/* xfs inode */
+	atomic_set(&ip->i_iocount, 0);
+	atomic_set(&ip->i_pincount, 0);
+	spin_lock_init(&ip->i_flags_lock);
+	init_waitqueue_head(&ip->i_ipin_wait);
+	/*
+	 * Because we want to use a counting completion, complete
+	 * the flush completion once to allow a single access to
+	 * the flush completion without blocking.
+	 */
+	init_completion(&ip->i_flush);
+	complete(&ip->i_flush);
+
+	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
+		     "xfsino", ip->i_ino);
+	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
 }
 
 /*
@@ -898,21 +989,26 @@ xfs_fs_write_inode(
 	struct inode		*inode,
 	int			sync)
 {
+	struct xfs_inode	*ip = XFS_I(inode);
 	int			error = 0;
 	int			flags = 0;
 
-	xfs_itrace_entry(XFS_I(inode));
+	xfs_itrace_entry(ip);
 	if (sync) {
-		filemap_fdatawait(inode->i_mapping);
+		error = xfs_wait_on_pages(ip, 0, -1);
+		if (error)
+			goto out_error;
 		flags |= FLUSH_SYNC;
 	}
-	error = xfs_inode_flush(XFS_I(inode), flags);
+	error = xfs_inode_flush(ip, flags);
+
+out_error:
 	/*
 	 * if we failed to write out the inode then mark
 	 * it dirty again so we'll try again later.
 	 */
 	if (error)
-		mark_inode_dirty_sync(inode);
+		xfs_mark_inode_dirty_sync(ip);
 
 	return -error;
 }
@@ -923,164 +1019,12 @@ xfs_fs_clear_inode(
 {
 	xfs_inode_t		*ip = XFS_I(inode);
 
-	/*
-	 * ip can be null when xfs_iget_core calls xfs_idestroy if we
-	 * find an inode with di_mode == 0 but without IGET_CREATE set.
-	 */
-	if (ip) {
-		xfs_itrace_entry(ip);
-		XFS_STATS_INC(vn_rele);
-		XFS_STATS_INC(vn_remove);
-		XFS_STATS_INC(vn_reclaim);
-		XFS_STATS_DEC(vn_active);
-
-		xfs_inactive(ip);
-		xfs_iflags_clear(ip, XFS_IMODIFIED);
-		if (xfs_reclaim(ip))
-			panic("%s: cannot reclaim 0x%p\n", __func__, inode);
-	}
-
-	ASSERT(XFS_I(inode) == NULL);
-}
-
-/*
- * Enqueue a work item to be picked up by the vfs xfssyncd thread.
- * Doing this has two advantages:
- * - It saves on stack space, which is tight in certain situations
- * - It can be used (with care) as a mechanism to avoid deadlocks.
- * Flushing while allocating in a full filesystem requires both.
- */
-STATIC void
-xfs_syncd_queue_work(
-	struct xfs_mount *mp,
-	void		*data,
-	void		(*syncer)(struct xfs_mount *, void *))
-{
-	struct bhv_vfs_sync_work *work;
-
-	work = kmem_alloc(sizeof(struct bhv_vfs_sync_work), KM_SLEEP);
-	INIT_LIST_HEAD(&work->w_list);
-	work->w_syncer = syncer;
-	work->w_data = data;
-	work->w_mount = mp;
-	spin_lock(&mp->m_sync_lock);
-	list_add_tail(&work->w_list, &mp->m_sync_list);
-	spin_unlock(&mp->m_sync_lock);
-	wake_up_process(mp->m_sync_task);
-}
-
-/*
- * Flush delayed allocate data, attempting to free up reserved space
- * from existing allocations.  At this point a new allocation attempt
- * has failed with ENOSPC and we are in the process of scratching our
- * heads, looking about for more room...
- */
-STATIC void
-xfs_flush_inode_work(
-	struct xfs_mount *mp,
-	void		*arg)
-{
-	struct inode	*inode = arg;
-	filemap_flush(inode->i_mapping);
-	iput(inode);
-}
-
-void
-xfs_flush_inode(
-	xfs_inode_t	*ip)
-{
-	struct inode	*inode = VFS_I(ip);
-
-	igrab(inode);
-	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inode_work);
-	delay(msecs_to_jiffies(500));
-}
-
-/*
- * This is the "bigger hammer" version of xfs_flush_inode_work...
- * (IOW, "If at first you don't succeed, use a Bigger Hammer").
- */
-STATIC void
-xfs_flush_device_work(
-	struct xfs_mount *mp,
-	void		*arg)
-{
-	struct inode	*inode = arg;
-	sync_blockdev(mp->m_super->s_bdev);
-	iput(inode);
-}
-
-void
-xfs_flush_device(
-	xfs_inode_t	*ip)
-{
-	struct inode	*inode = VFS_I(ip);
+	xfs_itrace_entry(ip);
+	XFS_STATS_INC(vn_rele);
+	XFS_STATS_INC(vn_remove);
+	XFS_STATS_DEC(vn_active);
 
-	igrab(inode);
-	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_device_work);
-	delay(msecs_to_jiffies(500));
-	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
-}
-
-STATIC void
-xfs_sync_worker(
-	struct xfs_mount *mp,
-	void		*unused)
-{
-	int		error;
-
-	if (!(mp->m_flags & XFS_MOUNT_RDONLY))
-		error = xfs_sync(mp, SYNC_FSDATA | SYNC_BDFLUSH | SYNC_ATTR);
-	mp->m_sync_seq++;
-	wake_up(&mp->m_wait_single_sync_task);
-}
-
-STATIC int
-xfssyncd(
-	void			*arg)
-{
-	struct xfs_mount	*mp = arg;
-	long			timeleft;
-	bhv_vfs_sync_work_t	*work, *n;
-	LIST_HEAD		(tmp);
-
-	set_freezable();
-	timeleft = xfs_syncd_centisecs * msecs_to_jiffies(10);
-	for (;;) {
-		timeleft = schedule_timeout_interruptible(timeleft);
-		/* swsusp */
-		try_to_freeze();
-		if (kthread_should_stop() && list_empty(&mp->m_sync_list))
-			break;
-
-		spin_lock(&mp->m_sync_lock);
-		/*
-		 * We can get woken by laptop mode, to do a sync -
-		 * that's the (only!) case where the list would be
-		 * empty with time remaining.
-		 */
-		if (!timeleft || list_empty(&mp->m_sync_list)) {
-			if (!timeleft)
-				timeleft = xfs_syncd_centisecs *
-							msecs_to_jiffies(10);
-			INIT_LIST_HEAD(&mp->m_sync_work.w_list);
-			list_add_tail(&mp->m_sync_work.w_list,
-					&mp->m_sync_list);
-		}
-		list_for_each_entry_safe(work, n, &mp->m_sync_list, w_list)
-			list_move(&work->w_list, &tmp);
-		spin_unlock(&mp->m_sync_lock);
-
-		list_for_each_entry_safe(work, n, &tmp, w_list) {
-			(*work->w_syncer)(mp, work->w_data);
-			list_del(&work->w_list);
-			if (work == &mp->m_sync_work)
-				continue;
-			kmem_free(work);
-		}
-	}
-
-	return 0;
+	xfs_inactive(ip);
 }
 
 STATIC void
@@ -1099,11 +1043,9 @@ xfs_fs_put_super(
 	struct xfs_mount	*mp = XFS_M(sb);
 	struct xfs_inode	*rip = mp->m_rootip;
 	int			unmount_event_flags = 0;
-	int			error;
-
-	kthread_stop(mp->m_sync_task);
 
-	xfs_sync(mp, SYNC_ATTR | SYNC_DELWRI);
+	xfs_syncd_stop(mp);
+	xfs_sync_inodes(mp, SYNC_ATTR|SYNC_DELWRI);
 
 #ifdef HAVE_DMAPI
 	if (mp->m_flags & XFS_MOUNT_DMAPI) {
@@ -1128,18 +1070,6 @@ xfs_fs_put_super(
 	xfs_filestream_unmount(mp);
 
 	XFS_bflush(mp->m_ddev_targp);
-	error = xfs_unmount_flush(mp, 0);
-	WARN_ON(error);
-
-	/*
-	 * If we're forcing a shutdown, typically because of a media error,
-	 * we want to make sure we invalidate dirty pages that belong to
-	 * referenced vnodes as well.
-	 */
-	if (XFS_FORCED_SHUTDOWN(mp)) {
-		error = xfs_sync(mp, SYNC_WAIT | SYNC_CLOSE);
-		ASSERT(error != EFSCORRUPTED);
-	}
 
 	if (mp->m_flags & XFS_MOUNT_DMAPI) {
 		XFS_SEND_UNMOUNT(mp, rip, DM_RIGHT_NULL, 0, 0,
@@ -1161,7 +1091,7 @@ xfs_fs_write_super(
 	struct super_block	*sb)
 {
 	if (!(sb->s_flags & MS_RDONLY))
-		xfs_sync(XFS_M(sb), SYNC_FSDATA);
+		xfs_sync_fsdata(XFS_M(sb), 0);
 	sb->s_dirt = 0;
 }
 
@@ -1172,7 +1102,6 @@ xfs_fs_sync_super(
 {
 	struct xfs_mount	*mp = XFS_M(sb);
 	int			error;
-	int			flags;
 
 	/*
 	 * Treat a sync operation like a freeze.  This is to work
@@ -1186,20 +1115,10 @@ xfs_fs_sync_super(
 	 * dirty the Linux inode until after the transaction I/O
 	 * completes.
 	 */
-	if (wait || unlikely(sb->s_frozen == SB_FREEZE_WRITE)) {
-		/*
-		 * First stage of freeze - no more writers will make progress
-		 * now we are here, so we flush delwri and delalloc buffers
-		 * here, then wait for all I/O to complete.  Data is frozen at
-		 * that point. Metadata is not frozen, transactions can still
-		 * occur here so don't bother flushing the buftarg (i.e
-		 * SYNC_QUIESCE) because it'll just get dirty again.
-		 */
-		flags = SYNC_DATA_QUIESCE;
-	} else
-		flags = SYNC_FSDATA;
-
-	error = xfs_sync(mp, flags);
+	if (wait || unlikely(sb->s_frozen == SB_FREEZE_WRITE))
+		error = xfs_quiesce_data(mp);
+	else
+		error = xfs_sync_fsdata(mp, 0);
 	sb->s_dirt = 0;
 
 	if (unlikely(laptop_mode)) {
@@ -1337,9 +1256,8 @@ xfs_fs_remount(
 
 	/* rw -> ro */
 	if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & MS_RDONLY)) {
-		xfs_filestream_flush(mp);
-		xfs_sync(mp, SYNC_DATA_QUIESCE);
-		xfs_attr_quiesce(mp);
+		xfs_quiesce_data(mp);
+		xfs_quiesce_attr(mp);
 		mp->m_flags |= XFS_MOUNT_RDONLY;
 	}
 
@@ -1348,17 +1266,17 @@ xfs_fs_remount(
 
 /*
  * Second stage of a freeze. The data is already frozen so we only
- * need to take care of themetadata. Once that's done write a dummy
+ * need to take care of the metadata. Once that's done write a dummy
  * record to dirty the log in case of a crash while frozen.
  */
-STATIC void
-xfs_fs_lockfs(
+STATIC int
+xfs_fs_freeze(
 	struct super_block	*sb)
 {
 	struct xfs_mount	*mp = XFS_M(sb);
 
-	xfs_attr_quiesce(mp);
-	xfs_fs_log_dummy(mp);
+	xfs_quiesce_attr(mp);
+	return -xfs_fs_log_dummy(mp);
 }
 
 STATIC int
@@ -1422,175 +1340,28 @@ xfs_fs_setxquota(
 
 /*
  * This function fills in xfs_mount_t fields based on mount args.
- * Note: the superblock has _not_ yet been read in.
- */
-STATIC int
-xfs_start_flags(
-	struct xfs_mount_args	*ap,
-	struct xfs_mount	*mp)
-{
-	int			error;
-
-	/* Values are in BBs */
-	if ((ap->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
-		/*
-		 * At this point the superblock has not been read
-		 * in, therefore we do not know the block size.
-		 * Before the mount call ends we will convert
-		 * these to FSBs.
-		 */
-		mp->m_dalign = ap->sunit;
-		mp->m_swidth = ap->swidth;
-	}
-
-	if (ap->logbufs != -1 &&
-	    ap->logbufs != 0 &&
-	    (ap->logbufs < XLOG_MIN_ICLOGS ||
-	     ap->logbufs > XLOG_MAX_ICLOGS)) {
-		cmn_err(CE_WARN,
-			"XFS: invalid logbufs value: %d [not %d-%d]",
-			ap->logbufs, XLOG_MIN_ICLOGS, XLOG_MAX_ICLOGS);
-		return XFS_ERROR(EINVAL);
-	}
-	mp->m_logbufs = ap->logbufs;
-	if (ap->logbufsize != -1 &&
-	    ap->logbufsize !=  0 &&
-	    (ap->logbufsize < XLOG_MIN_RECORD_BSIZE ||
-	     ap->logbufsize > XLOG_MAX_RECORD_BSIZE ||
-	     !is_power_of_2(ap->logbufsize))) {
-		cmn_err(CE_WARN,
-	"XFS: invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
-			ap->logbufsize);
-		return XFS_ERROR(EINVAL);
-	}
-
-	error = ENOMEM;
-
-	mp->m_logbsize = ap->logbufsize;
-	mp->m_fsname_len = strlen(ap->fsname) + 1;
-
-	mp->m_fsname = kstrdup(ap->fsname, GFP_KERNEL);
-	if (!mp->m_fsname)
-		goto out;
-
-	if (ap->rtname[0]) {
-		mp->m_rtname = kstrdup(ap->rtname, GFP_KERNEL);
-		if (!mp->m_rtname)
-			goto out_free_fsname;
-
-	}
-
-	if (ap->logname[0]) {
-		mp->m_logname = kstrdup(ap->logname, GFP_KERNEL);
-		if (!mp->m_logname)
-			goto out_free_rtname;
-	}
-
-	if (ap->flags & XFSMNT_WSYNC)
-		mp->m_flags |= XFS_MOUNT_WSYNC;
-#if XFS_BIG_INUMS
-	if (ap->flags & XFSMNT_INO64) {
-		mp->m_flags |= XFS_MOUNT_INO64;
-		mp->m_inoadd = XFS_INO64_OFFSET;
-	}
-#endif
-	if (ap->flags & XFSMNT_RETERR)
-		mp->m_flags |= XFS_MOUNT_RETERR;
-	if (ap->flags & XFSMNT_NOALIGN)
-		mp->m_flags |= XFS_MOUNT_NOALIGN;
-	if (ap->flags & XFSMNT_SWALLOC)
-		mp->m_flags |= XFS_MOUNT_SWALLOC;
-	if (ap->flags & XFSMNT_OSYNCISOSYNC)
-		mp->m_flags |= XFS_MOUNT_OSYNCISOSYNC;
-	if (ap->flags & XFSMNT_32BITINODES)
-		mp->m_flags |= XFS_MOUNT_32BITINODES;
-
-	if (ap->flags & XFSMNT_IOSIZE) {
-		if (ap->iosizelog > XFS_MAX_IO_LOG ||
-		    ap->iosizelog < XFS_MIN_IO_LOG) {
-			cmn_err(CE_WARN,
-		"XFS: invalid log iosize: %d [not %d-%d]",
-				ap->iosizelog, XFS_MIN_IO_LOG,
-				XFS_MAX_IO_LOG);
-			return XFS_ERROR(EINVAL);
-		}
-
-		mp->m_flags |= XFS_MOUNT_DFLT_IOSIZE;
-		mp->m_readio_log = mp->m_writeio_log = ap->iosizelog;
-	}
-
-	if (ap->flags & XFSMNT_IKEEP)
-		mp->m_flags |= XFS_MOUNT_IKEEP;
-	if (ap->flags & XFSMNT_DIRSYNC)
-		mp->m_flags |= XFS_MOUNT_DIRSYNC;
-	if (ap->flags & XFSMNT_ATTR2)
-		mp->m_flags |= XFS_MOUNT_ATTR2;
-	if (ap->flags & XFSMNT_NOATTR2)
-		mp->m_flags |= XFS_MOUNT_NOATTR2;
-
-	if (ap->flags2 & XFSMNT2_COMPAT_IOSIZE)
-		mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
-
-	/*
-	 * no recovery flag requires a read-only mount
-	 */
-	if (ap->flags & XFSMNT_NORECOVERY) {
-		if (!(mp->m_flags & XFS_MOUNT_RDONLY)) {
-			cmn_err(CE_WARN,
-	"XFS: tried to mount a FS read-write without recovery!");
-			return XFS_ERROR(EINVAL);
-		}
-		mp->m_flags |= XFS_MOUNT_NORECOVERY;
-	}
-
-	if (ap->flags & XFSMNT_NOUUID)
-		mp->m_flags |= XFS_MOUNT_NOUUID;
-	if (ap->flags & XFSMNT_BARRIER)
-		mp->m_flags |= XFS_MOUNT_BARRIER;
-	else
-		mp->m_flags &= ~XFS_MOUNT_BARRIER;
-
-	if (ap->flags2 & XFSMNT2_FILESTREAMS)
-		mp->m_flags |= XFS_MOUNT_FILESTREAMS;
-
-	if (ap->flags & XFSMNT_DMAPI)
-		mp->m_flags |= XFS_MOUNT_DMAPI;
-	return 0;
-
-
- out_free_rtname:
-	kfree(mp->m_rtname);
- out_free_fsname:
-	kfree(mp->m_fsname);
- out:
-	return error;
-}
-
-/*
- * This function fills in xfs_mount_t fields based on mount args.
  * Note: the superblock _has_ now been read in.
  */
 STATIC int
 xfs_finish_flags(
-	struct xfs_mount_args	*ap,
 	struct xfs_mount	*mp)
 {
 	int			ronly = (mp->m_flags & XFS_MOUNT_RDONLY);
 
-	/* Fail a mount where the logbuf is smaller then the log stripe */
+	/* Fail a mount where the logbuf is smaller than the log stripe */
 	if (xfs_sb_version_haslogv2(&mp->m_sb)) {
-		if ((ap->logbufsize <= 0) &&
-		    (mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE)) {
+		if (mp->m_logbsize <= 0 &&
+		    mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE) {
 			mp->m_logbsize = mp->m_sb.sb_logsunit;
-		} else if (ap->logbufsize > 0 &&
-			   ap->logbufsize < mp->m_sb.sb_logsunit) {
+		} else if (mp->m_logbsize > 0 &&
+			   mp->m_logbsize < mp->m_sb.sb_logsunit) {
 			cmn_err(CE_WARN,
 	"XFS: logbuf size must be greater than or equal to log stripe size");
 			return XFS_ERROR(EINVAL);
 		}
 	} else {
 		/* Fail a mount if the logbuf is larger than 32K */
-		if (ap->logbufsize > XLOG_BIG_RECORD_BSIZE) {
+		if (mp->m_logbsize > XLOG_BIG_RECORD_BSIZE) {
 			cmn_err(CE_WARN,
 	"XFS: logbuf size for version 1 logs must be 16K or 32K");
 			return XFS_ERROR(EINVAL);
@@ -1602,7 +1373,7 @@ xfs_finish_flags(
 	 * told by noattr2 to turn it off
 	 */
 	if (xfs_sb_version_hasattr2(&mp->m_sb) &&
-	    !(ap->flags & XFSMNT_NOATTR2))
+	    !(mp->m_flags & XFS_MOUNT_NOATTR2))
 		mp->m_flags |= XFS_MOUNT_ATTR2;
 
 	/*
@@ -1614,48 +1385,6 @@ xfs_finish_flags(
 		return XFS_ERROR(EROFS);
 	}
 
-	/*
-	 * check for shared mount.
-	 */
-	if (ap->flags & XFSMNT_SHARED) {
-		if (!xfs_sb_version_hasshared(&mp->m_sb))
-			return XFS_ERROR(EINVAL);
-
-		/*
-		 * For IRIX 6.5, shared mounts must have the shared
-		 * version bit set, have the persistent readonly
-		 * field set, must be version 0 and can only be mounted
-		 * read-only.
-		 */
-		if (!ronly || !(mp->m_sb.sb_flags & XFS_SBF_READONLY) ||
-		     (mp->m_sb.sb_shared_vn != 0))
-			return XFS_ERROR(EINVAL);
-
-		mp->m_flags |= XFS_MOUNT_SHARED;
-
-		/*
-		 * Shared XFS V0 can't deal with DMI.  Return EINVAL.
-		 */
-		if (mp->m_sb.sb_shared_vn == 0 && (ap->flags & XFSMNT_DMAPI))
-			return XFS_ERROR(EINVAL);
-	}
-
-	if (ap->flags & XFSMNT_UQUOTA) {
-		mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
-		if (ap->flags & XFSMNT_UQUOTAENF)
-			mp->m_qflags |= XFS_UQUOTA_ENFD;
-	}
-
-	if (ap->flags & XFSMNT_GQUOTA) {
-		mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
-		if (ap->flags & XFSMNT_GQUOTAENF)
-			mp->m_qflags |= XFS_OQUOTA_ENFD;
-	} else if (ap->flags & XFSMNT_PQUOTA) {
-		mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
-		if (ap->flags & XFSMNT_PQUOTAENF)
-			mp->m_qflags |= XFS_OQUOTA_ENFD;
-	}
-
 	return 0;
 }
 
@@ -1667,19 +1396,14 @@ xfs_fs_fill_super(
 {
 	struct inode		*root;
 	struct xfs_mount	*mp = NULL;
-	struct xfs_mount_args	*args;
 	int			flags = 0, error = ENOMEM;
-
-	args = xfs_args_allocate(sb, silent);
-	if (!args)
-		return -ENOMEM;
+	char			*mtpt = NULL;
 
 	mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL);
 	if (!mp)
-		goto out_free_args;
+		goto out;
 
 	spin_lock_init(&mp->m_sb_lock);
-	mutex_init(&mp->m_ilock);
 	mutex_init(&mp->m_growlock);
 	atomic_set(&mp->m_active_trans, 0);
 	INIT_LIST_HEAD(&mp->m_sync_list);
@@ -1689,12 +1413,9 @@ xfs_fs_fill_super(
 	mp->m_super = sb;
 	sb->s_fs_info = mp;
 
-	if (sb->s_flags & MS_RDONLY)
-		mp->m_flags |= XFS_MOUNT_RDONLY;
-
-	error = xfs_parseargs(mp, (char *)data, args, 0);
+	error = xfs_parseargs(mp, (char *)data, &mtpt);
 	if (error)
-		goto out_free_mp;
+		goto out_free_fsname;
 
 	sb_min_blocksize(sb, BBSIZE);
 	sb->s_xattr = xfs_xattr_handlers;
@@ -1702,33 +1423,28 @@ xfs_fs_fill_super(
 	sb->s_qcop = &xfs_quotactl_operations;
 	sb->s_op = &xfs_super_operations;
 
-	error = xfs_dmops_get(mp, args);
+	error = xfs_dmops_get(mp);
 	if (error)
-		goto out_free_mp;
-	error = xfs_qmops_get(mp, args);
+		goto out_free_fsname;
+	error = xfs_qmops_get(mp);
 	if (error)
 		goto out_put_dmops;
 
-	if (args->flags & XFSMNT_QUIET)
+	if (silent)
 		flags |= XFS_MFSI_QUIET;
 
-	error = xfs_open_devices(mp, args);
+	error = xfs_open_devices(mp);
 	if (error)
 		goto out_put_qmops;
 
 	if (xfs_icsb_init_counters(mp))
 		mp->m_flags |= XFS_MOUNT_NO_PERCPU_SB;
 
-	/*
-	 * Setup flags based on mount(2) options and then the superblock
-	 */
-	error = xfs_start_flags(args, mp);
-	if (error)
-		goto out_free_fsname;
 	error = xfs_readsb(mp, flags);
 	if (error)
-		goto out_free_fsname;
-	error = xfs_finish_flags(args, mp);
+		goto out_destroy_counters;
+
+	error = xfs_finish_flags(mp);
 	if (error)
 		goto out_free_sb;
 
@@ -1747,7 +1463,7 @@ xfs_fs_fill_super(
 	if (error)
 		goto out_filestream_unmount;
 
-	XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, args->mtpt, args->fsname);
+	XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, mtpt, mp->m_fsname);
 
 	sb->s_dirt = 1;
 	sb->s_magic = XFS_SB_MAGIC;
@@ -1772,35 +1488,31 @@ xfs_fs_fill_super(
 		goto fail_vnrele;
 	}
 
-	mp->m_sync_work.w_syncer = xfs_sync_worker;
-	mp->m_sync_work.w_mount = mp;
-	mp->m_sync_task = kthread_run(xfssyncd, mp, "xfssyncd");
-	if (IS_ERR(mp->m_sync_task)) {
-		error = -PTR_ERR(mp->m_sync_task);
+	error = xfs_syncd_init(mp);
+	if (error)
 		goto fail_vnrele;
-	}
 
-	xfs_itrace_exit(XFS_I(sb->s_root->d_inode));
+	kfree(mtpt);
 
-	kfree(args);
+	xfs_itrace_exit(XFS_I(sb->s_root->d_inode));
 	return 0;
 
  out_filestream_unmount:
 	xfs_filestream_unmount(mp);
  out_free_sb:
 	xfs_freesb(mp);
- out_free_fsname:
-	xfs_free_fsname(mp);
+ out_destroy_counters:
 	xfs_icsb_destroy_counters(mp);
 	xfs_close_devices(mp);
  out_put_qmops:
 	xfs_qmops_put(mp);
  out_put_dmops:
 	xfs_dmops_put(mp);
- out_free_mp:
+ out_free_fsname:
+	xfs_free_fsname(mp);
+	kfree(mtpt);
 	kfree(mp);
- out_free_args:
-	kfree(args);
+ out:
 	return -error;
 
  fail_vnrele:
@@ -1820,8 +1532,6 @@ xfs_fs_fill_super(
 	xfs_filestream_unmount(mp);
 
 	XFS_bflush(mp->m_ddev_targp);
-	error = xfs_unmount_flush(mp, 0);
-	WARN_ON(error);
 
 	xfs_unmountfs(mp);
 	goto out_free_sb;
@@ -1847,7 +1557,7 @@ static struct super_operations xfs_super_operations = {
 	.put_super		= xfs_fs_put_super,
 	.write_super		= xfs_fs_write_super,
 	.sync_fs		= xfs_fs_sync_super,
-	.write_super_lockfs	= xfs_fs_lockfs,
+	.freeze_fs		= xfs_fs_freeze,
 	.statfs			= xfs_fs_statfs,
 	.remount_fs		= xfs_fs_remount,
 	.show_options		= xfs_fs_show_options,
@@ -1882,10 +1592,19 @@ xfs_alloc_trace_bufs(void)
 	if (!xfs_bmap_trace_buf)
 		goto out_free_alloc_trace;
 #endif
-#ifdef XFS_BMBT_TRACE
+#ifdef XFS_BTREE_TRACE
+	xfs_allocbt_trace_buf = ktrace_alloc(XFS_ALLOCBT_TRACE_SIZE,
+					     KM_MAYFAIL);
+	if (!xfs_allocbt_trace_buf)
+		goto out_free_bmap_trace;
+
+	xfs_inobt_trace_buf = ktrace_alloc(XFS_INOBT_TRACE_SIZE, KM_MAYFAIL);
+	if (!xfs_inobt_trace_buf)
+		goto out_free_allocbt_trace;
+
 	xfs_bmbt_trace_buf = ktrace_alloc(XFS_BMBT_TRACE_SIZE, KM_MAYFAIL);
 	if (!xfs_bmbt_trace_buf)
-		goto out_free_bmap_trace;
+		goto out_free_inobt_trace;
 #endif
 #ifdef XFS_ATTR_TRACE
 	xfs_attr_trace_buf = ktrace_alloc(XFS_ATTR_TRACE_SIZE, KM_MAYFAIL);
@@ -1907,8 +1626,12 @@ xfs_alloc_trace_bufs(void)
 	ktrace_free(xfs_attr_trace_buf);
  out_free_bmbt_trace:
 #endif
-#ifdef XFS_BMBT_TRACE
+#ifdef XFS_BTREE_TRACE
 	ktrace_free(xfs_bmbt_trace_buf);
+ out_free_inobt_trace:
+	ktrace_free(xfs_inobt_trace_buf);
+ out_free_allocbt_trace:
+	ktrace_free(xfs_allocbt_trace_buf);
  out_free_bmap_trace:
 #endif
 #ifdef XFS_BMAP_TRACE
@@ -1931,8 +1654,10 @@ xfs_free_trace_bufs(void)
 #ifdef XFS_ATTR_TRACE
 	ktrace_free(xfs_attr_trace_buf);
 #endif
-#ifdef XFS_BMBT_TRACE
+#ifdef XFS_BTREE_TRACE
 	ktrace_free(xfs_bmbt_trace_buf);
+	ktrace_free(xfs_inobt_trace_buf);
+	ktrace_free(xfs_allocbt_trace_buf);
 #endif
 #ifdef XFS_BMAP_TRACE
 	ktrace_free(xfs_bmap_trace_buf);
@@ -1945,16 +1670,10 @@ xfs_free_trace_bufs(void)
 STATIC int __init
 xfs_init_zones(void)
 {
-	xfs_vnode_zone = kmem_zone_init_flags(sizeof(struct inode), "xfs_vnode",
-					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
-					KM_ZONE_SPREAD,
-					xfs_fs_inode_init_once);
-	if (!xfs_vnode_zone)
-		goto out;
 
 	xfs_ioend_zone = kmem_zone_init(sizeof(xfs_ioend_t), "xfs_ioend");
 	if (!xfs_ioend_zone)
-		goto out_destroy_vnode_zone;
+		goto out;
 
 	xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE,
 						  xfs_ioend_zone);
@@ -1970,6 +1689,7 @@ xfs_init_zones(void)
 						"xfs_bmap_free_item");
 	if (!xfs_bmap_free_item_zone)
 		goto out_destroy_log_ticket_zone;
+
 	xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
 						"xfs_btree_cur");
 	if (!xfs_btree_cur_zone)
@@ -2017,8 +1737,8 @@ xfs_init_zones(void)
 
 	xfs_inode_zone =
 		kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
-					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
-					KM_ZONE_SPREAD, NULL);
+			KM_ZONE_HWALIGN | KM_ZONE_RECLAIM | KM_ZONE_SPREAD,
+			xfs_fs_inode_init_once);
 	if (!xfs_inode_zone)
 		goto out_destroy_efi_zone;
 
@@ -2066,8 +1786,6 @@ xfs_init_zones(void)
 	mempool_destroy(xfs_ioend_pool);
  out_destroy_ioend_zone:
 	kmem_zone_destroy(xfs_ioend_zone);
- out_destroy_vnode_zone:
-	kmem_zone_destroy(xfs_vnode_zone);
  out:
 	return -ENOMEM;
 }
@@ -2092,7 +1810,6 @@ xfs_destroy_zones(void)
 	kmem_zone_destroy(xfs_log_ticket_zone);
 	mempool_destroy(xfs_ioend_pool);
 	kmem_zone_destroy(xfs_ioend_zone);
-	kmem_zone_destroy(xfs_vnode_zone);
 
 }
 
@@ -2100,13 +1817,12 @@ STATIC int __init
 init_xfs_fs(void)
 {
 	int			error;
-	static char		message[] __initdata = KERN_INFO \
-		XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled\n";
 
-	printk(message);
+	printk(KERN_INFO XFS_VERSION_STRING " with "
+			 XFS_BUILD_OPTIONS " enabled\n");
 
 	ktrace_init(64);
-	vn_init();
+	xfs_ioend_init();
 	xfs_dir_startup();
 
 	error = xfs_init_zones();
diff --git a/fs/xfs/linux-2.6/xfs_super.h b/fs/xfs/linux-2.6/xfs_super.h
index fe2ef4e6a0f..d5d776d4cd6 100644
--- a/fs/xfs/linux-2.6/xfs_super.h
+++ b/fs/xfs/linux-2.6/xfs_super.h
@@ -20,24 +20,12 @@
 
 #include <linux/exportfs.h>
 
-#ifdef CONFIG_XFS_DMAPI
-# define vfs_insertdmapi(vfs)	vfs_insertops(vfsp, &xfs_dmops)
-# define vfs_initdmapi()	dmapi_init()
-# define vfs_exitdmapi()	dmapi_uninit()
-#else
-# define vfs_insertdmapi(vfs)	do { } while (0)
-# define vfs_initdmapi()	do { } while (0)
-# define vfs_exitdmapi()	do { } while (0)
-#endif
-
 #ifdef CONFIG_XFS_QUOTA
-# define vfs_insertquota(vfs)	vfs_insertops(vfsp, &xfs_qmops)
 extern void xfs_qm_init(void);
 extern void xfs_qm_exit(void);
 # define vfs_initquota()	xfs_qm_init()
 # define vfs_exitquota()	xfs_qm_exit()
 #else
-# define vfs_insertquota(vfs)	do { } while (0)
 # define vfs_initquota()	do { } while (0)
 # define vfs_exitquota()	do { } while (0)
 #endif
@@ -101,9 +89,6 @@ struct block_device;
 
 extern __uint64_t xfs_max_file_offset(unsigned int);
 
-extern void xfs_flush_inode(struct xfs_inode *);
-extern void xfs_flush_device(struct xfs_inode *);
-
 extern void xfs_blkdev_issue_flush(struct xfs_buftarg *);
 
 extern const struct export_operations xfs_export_operations;
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
new file mode 100644
index 00000000000..2ed035354c2
--- /dev/null
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -0,0 +1,762 @@
+/*
+ * Copyright (c) 2000-2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_types.h"
+#include "xfs_bit.h"
+#include "xfs_log.h"
+#include "xfs_inum.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dir2.h"
+#include "xfs_dmapi.h"
+#include "xfs_mount.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_btree.h"
+#include "xfs_dir2_sf.h"
+#include "xfs_attr_sf.h"
+#include "xfs_inode.h"
+#include "xfs_dinode.h"
+#include "xfs_error.h"
+#include "xfs_mru_cache.h"
+#include "xfs_filestream.h"
+#include "xfs_vnodeops.h"
+#include "xfs_utils.h"
+#include "xfs_buf_item.h"
+#include "xfs_inode_item.h"
+#include "xfs_rw.h"
+
+#include <linux/kthread.h>
+#include <linux/freezer.h>
+
+/*
+ * Sync all the inodes in the given AG according to the
+ * direction given by the flags.
+ */
+STATIC int
+xfs_sync_inodes_ag(
+	xfs_mount_t	*mp,
+	int		ag,
+	int		flags)
+{
+	xfs_perag_t	*pag = &mp->m_perag[ag];
+	int		nr_found;
+	uint32_t	first_index = 0;
+	int		error = 0;
+	int		last_error = 0;
+	int		fflag = XFS_B_ASYNC;
+
+	if (flags & SYNC_DELWRI)
+		fflag = XFS_B_DELWRI;
+	if (flags & SYNC_WAIT)
+		fflag = 0;		/* synchronous overrides all */
+
+	do {
+		struct inode	*inode;
+		xfs_inode_t	*ip = NULL;
+		int		lock_flags = XFS_ILOCK_SHARED;
+
+		/*
+		 * use a gang lookup to find the next inode in the tree
+		 * as the tree is sparse and a gang lookup walks to find
+		 * the number of objects requested.
+		 */
+		read_lock(&pag->pag_ici_lock);
+		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
+				(void**)&ip, first_index, 1);
+
+		if (!nr_found) {
+			read_unlock(&pag->pag_ici_lock);
+			break;
+		}
+
+		/*
+		 * Update the index for the next lookup. Catch overflows
+		 * into the next AG range which can occur if we have inodes
+		 * in the last block of the AG and we are currently
+		 * pointing to the last inode.
+		 */
+		first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+		if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino)) {
+			read_unlock(&pag->pag_ici_lock);
+			break;
+		}
+
+		/* nothing to sync during shutdown */
+		if (XFS_FORCED_SHUTDOWN(mp)) {
+			read_unlock(&pag->pag_ici_lock);
+			return 0;
+		}
+
+		/*
+		 * If we can't get a reference on the inode, it must be
+		 * in reclaim. Leave it for the reclaim code to flush.
+		 */
+		inode = VFS_I(ip);
+		if (!igrab(inode)) {
+			read_unlock(&pag->pag_ici_lock);
+			continue;
+		}
+		read_unlock(&pag->pag_ici_lock);
+
+		/* avoid new or bad inodes */
+		if (is_bad_inode(inode) ||
+		    xfs_iflags_test(ip, XFS_INEW)) {
+			IRELE(ip);
+			continue;
+		}
+
+		/*
+		 * If we have to flush data or wait for I/O completion
+		 * we need to hold the iolock.
+		 */
+		if ((flags & SYNC_DELWRI) && VN_DIRTY(inode)) {
+			xfs_ilock(ip, XFS_IOLOCK_SHARED);
+			lock_flags |= XFS_IOLOCK_SHARED;
+			error = xfs_flush_pages(ip, 0, -1, fflag, FI_NONE);
+			if (flags & SYNC_IOWAIT)
+				xfs_ioend_wait(ip);
+		}
+		xfs_ilock(ip, XFS_ILOCK_SHARED);
+
+		if ((flags & SYNC_ATTR) && !xfs_inode_clean(ip)) {
+			if (flags & SYNC_WAIT) {
+				xfs_iflock(ip);
+				if (!xfs_inode_clean(ip))
+					error = xfs_iflush(ip, XFS_IFLUSH_SYNC);
+				else
+					xfs_ifunlock(ip);
+			} else if (xfs_iflock_nowait(ip)) {
+				if (!xfs_inode_clean(ip))
+					error = xfs_iflush(ip, XFS_IFLUSH_DELWRI);
+				else
+					xfs_ifunlock(ip);
+			}
+		}
+		xfs_iput(ip, lock_flags);
+
+		if (error)
+			last_error = error;
+		/*
+		 * bail out if the filesystem is corrupted.
+		 */
+		if (error == EFSCORRUPTED)
+			return XFS_ERROR(error);
+
+	} while (nr_found);
+
+	return last_error;
+}
+
+int
+xfs_sync_inodes(
+	xfs_mount_t	*mp,
+	int		flags)
+{
+	int		error;
+	int		last_error;
+	int		i;
+	int		lflags = XFS_LOG_FORCE;
+
+	if (mp->m_flags & XFS_MOUNT_RDONLY)
+		return 0;
+	error = 0;
+	last_error = 0;
+
+	if (flags & SYNC_WAIT)
+		lflags |= XFS_LOG_SYNC;
+
+	for (i = 0; i < mp->m_sb.sb_agcount; i++) {
+		if (!mp->m_perag[i].pag_ici_init)
+			continue;
+		error = xfs_sync_inodes_ag(mp, i, flags);
+		if (error)
+			last_error = error;
+		if (error == EFSCORRUPTED)
+			break;
+	}
+	if (flags & SYNC_DELWRI)
+		xfs_log_force(mp, 0, lflags);
+
+	return XFS_ERROR(last_error);
+}
+
+STATIC int
+xfs_commit_dummy_trans(
+	struct xfs_mount	*mp,
+	uint			log_flags)
+{
+	struct xfs_inode	*ip = mp->m_rootip;
+	struct xfs_trans	*tp;
+	int			error;
+
+	/*
+	 * Put a dummy transaction in the log to tell recovery
+	 * that all others are OK.
+	 */
+	tp = xfs_trans_alloc(mp, XFS_TRANS_DUMMY1);
+	error = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
+	if (error) {
+		xfs_trans_cancel(tp, 0);
+		return error;
+	}
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+	xfs_trans_ihold(tp, ip);
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	/* XXX(hch): ignoring the error here.. */
+	error = xfs_trans_commit(tp, 0);
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+
+	xfs_log_force(mp, 0, log_flags);
+	return 0;
+}
+
+int
+xfs_sync_fsdata(
+	struct xfs_mount	*mp,
+	int			flags)
+{
+	struct xfs_buf		*bp;
+	struct xfs_buf_log_item	*bip;
+	int			error = 0;
+
+	/*
+	 * If this is xfssyncd() then only sync the superblock if we can
+	 * lock it without sleeping and it is not pinned.
+	 */
+	if (flags & SYNC_BDFLUSH) {
+		ASSERT(!(flags & SYNC_WAIT));
+
+		bp = xfs_getsb(mp, XFS_BUF_TRYLOCK);
+		if (!bp)
+			goto out;
+
+		bip = XFS_BUF_FSPRIVATE(bp, struct xfs_buf_log_item *);
+		if (!bip || !xfs_buf_item_dirty(bip) || XFS_BUF_ISPINNED(bp))
+			goto out_brelse;
+	} else {
+		bp = xfs_getsb(mp, 0);
+
+		/*
+		 * If the buffer is pinned then push on the log so we won't
+		 * get stuck waiting in the write for someone, maybe
+		 * ourselves, to flush the log.
+		 *
+		 * Even though we just pushed the log above, we did not have
+		 * the superblock buffer locked at that point so it can
+		 * become pinned in between there and here.
+		 */
+		if (XFS_BUF_ISPINNED(bp))
+			xfs_log_force(mp, 0, XFS_LOG_FORCE);
+	}
+
+
+	if (flags & SYNC_WAIT)
+		XFS_BUF_UNASYNC(bp);
+	else
+		XFS_BUF_ASYNC(bp);
+
+	return xfs_bwrite(mp, bp);
+
+ out_brelse:
+	xfs_buf_relse(bp);
+ out:
+	return error;
+}
+
+/*
+ * When remounting a filesystem read-only or freezing the filesystem, we have
+ * two phases to execute. This first phase is syncing the data before we
+ * quiesce the filesystem, and the second is flushing all the inodes out after
+ * we've waited for all the transactions created by the first phase to
+ * complete. The second phase ensures that the inodes are written to their
+ * location on disk rather than just existing in transactions in the log. This
+ * means after a quiesce there is no log replay required to write the inodes to
+ * disk (this is the main difference between a sync and a quiesce).
+ */
+/*
+ * First stage of freeze - no writers will make progress now we are here,
+ * so we flush delwri and delalloc buffers here, then wait for all I/O to
+ * complete.  Data is frozen at that point. Metadata is not frozen,
+ * transactions can still occur here so don't bother flushing the buftarg
+ * because it'll just get dirty again.
+ */
+int
+xfs_quiesce_data(
+	struct xfs_mount	*mp)
+{
+	int error;
+
+	/* push non-blocking */
+	xfs_sync_inodes(mp, SYNC_DELWRI|SYNC_BDFLUSH);
+	XFS_QM_DQSYNC(mp, SYNC_BDFLUSH);
+	xfs_filestream_flush(mp);
+
+	/* push and block */
+	xfs_sync_inodes(mp, SYNC_DELWRI|SYNC_WAIT|SYNC_IOWAIT);
+	XFS_QM_DQSYNC(mp, SYNC_WAIT);
+
+	/* write superblock and hoover up shutdown errors */
+	error = xfs_sync_fsdata(mp, 0);
+
+	/* flush data-only devices */
+	if (mp->m_rtdev_targp)
+		XFS_bflush(mp->m_rtdev_targp);
+
+	return error;
+}
+
+STATIC void
+xfs_quiesce_fs(
+	struct xfs_mount	*mp)
+{
+	int	count = 0, pincount;
+
+	xfs_flush_buftarg(mp->m_ddev_targp, 0);
+	xfs_reclaim_inodes(mp, 0, XFS_IFLUSH_DELWRI_ELSE_ASYNC);
+
+	/*
+	 * This loop must run at least twice.  The first instance of the loop
+	 * will flush most meta data but that will generate more meta data
+	 * (typically directory updates).  Which then must be flushed and
+	 * logged before we can write the unmount record.
+	 */
+	do {
+		xfs_sync_inodes(mp, SYNC_ATTR|SYNC_WAIT);
+		pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1);
+		if (!pincount) {
+			delay(50);
+			count++;
+		}
+	} while (count < 2);
+}
+
+/*
+ * Second stage of a quiesce. The data is already synced, now we have to take
+ * care of the metadata. New transactions are already blocked, so we need to
+ * wait for any remaining transactions to drain out before proceding.
+ */
+void
+xfs_quiesce_attr(
+	struct xfs_mount	*mp)
+{
+	int	error = 0;
+
+	/* wait for all modifications to complete */
+	while (atomic_read(&mp->m_active_trans) > 0)
+		delay(100);
+
+	/* flush inodes and push all remaining buffers out to disk */
+	xfs_quiesce_fs(mp);
+
+	ASSERT_ALWAYS(atomic_read(&mp->m_active_trans) == 0);
+
+	/* Push the superblock and write an unmount record */
+	error = xfs_log_sbcount(mp, 1);
+	if (error)
+		xfs_fs_cmn_err(CE_WARN, mp,
+				"xfs_attr_quiesce: failed to log sb changes. "
+				"Frozen image may not be consistent.");
+	xfs_log_unmount_write(mp);
+	xfs_unmountfs_writesb(mp);
+}
+
+/*
+ * Enqueue a work item to be picked up by the vfs xfssyncd thread.
+ * Doing this has two advantages:
+ * - It saves on stack space, which is tight in certain situations
+ * - It can be used (with care) as a mechanism to avoid deadlocks.
+ * Flushing while allocating in a full filesystem requires both.
+ */
+STATIC void
+xfs_syncd_queue_work(
+	struct xfs_mount *mp,
+	void		*data,
+	void		(*syncer)(struct xfs_mount *, void *))
+{
+	struct bhv_vfs_sync_work *work;
+
+	work = kmem_alloc(sizeof(struct bhv_vfs_sync_work), KM_SLEEP);
+	INIT_LIST_HEAD(&work->w_list);
+	work->w_syncer = syncer;
+	work->w_data = data;
+	work->w_mount = mp;
+	spin_lock(&mp->m_sync_lock);
+	list_add_tail(&work->w_list, &mp->m_sync_list);
+	spin_unlock(&mp->m_sync_lock);
+	wake_up_process(mp->m_sync_task);
+}
+
+/*
+ * Flush delayed allocate data, attempting to free up reserved space
+ * from existing allocations.  At this point a new allocation attempt
+ * has failed with ENOSPC and we are in the process of scratching our
+ * heads, looking about for more room...
+ */
+STATIC void
+xfs_flush_inode_work(
+	struct xfs_mount *mp,
+	void		*arg)
+{
+	struct inode	*inode = arg;
+	filemap_flush(inode->i_mapping);
+	iput(inode);
+}
+
+void
+xfs_flush_inode(
+	xfs_inode_t	*ip)
+{
+	struct inode	*inode = VFS_I(ip);
+
+	igrab(inode);
+	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_inode_work);
+	delay(msecs_to_jiffies(500));
+}
+
+/*
+ * This is the "bigger hammer" version of xfs_flush_inode_work...
+ * (IOW, "If at first you don't succeed, use a Bigger Hammer").
+ */
+STATIC void
+xfs_flush_device_work(
+	struct xfs_mount *mp,
+	void		*arg)
+{
+	struct inode	*inode = arg;
+	sync_blockdev(mp->m_super->s_bdev);
+	iput(inode);
+}
+
+void
+xfs_flush_device(
+	xfs_inode_t	*ip)
+{
+	struct inode	*inode = VFS_I(ip);
+
+	igrab(inode);
+	xfs_syncd_queue_work(ip->i_mount, inode, xfs_flush_device_work);
+	delay(msecs_to_jiffies(500));
+	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
+}
+
+/*
+ * Every sync period we need to unpin all items, reclaim inodes, sync
+ * quota and write out the superblock. We might need to cover the log
+ * to indicate it is idle.
+ */
+STATIC void
+xfs_sync_worker(
+	struct xfs_mount *mp,
+	void		*unused)
+{
+	int		error;
+
+	if (!(mp->m_flags & XFS_MOUNT_RDONLY)) {
+		xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE);
+		xfs_reclaim_inodes(mp, 0, XFS_IFLUSH_DELWRI_ELSE_ASYNC);
+		/* dgc: errors ignored here */
+		error = XFS_QM_DQSYNC(mp, SYNC_BDFLUSH);
+		error = xfs_sync_fsdata(mp, SYNC_BDFLUSH);
+		if (xfs_log_need_covered(mp))
+			error = xfs_commit_dummy_trans(mp, XFS_LOG_FORCE);
+	}
+	mp->m_sync_seq++;
+	wake_up(&mp->m_wait_single_sync_task);
+}
+
+STATIC int
+xfssyncd(
+	void			*arg)
+{
+	struct xfs_mount	*mp = arg;
+	long			timeleft;
+	bhv_vfs_sync_work_t	*work, *n;
+	LIST_HEAD		(tmp);
+
+	set_freezable();
+	timeleft = xfs_syncd_centisecs * msecs_to_jiffies(10);
+	for (;;) {
+		timeleft = schedule_timeout_interruptible(timeleft);
+		/* swsusp */
+		try_to_freeze();
+		if (kthread_should_stop() && list_empty(&mp->m_sync_list))
+			break;
+
+		spin_lock(&mp->m_sync_lock);
+		/*
+		 * We can get woken by laptop mode, to do a sync -
+		 * that's the (only!) case where the list would be
+		 * empty with time remaining.
+		 */
+		if (!timeleft || list_empty(&mp->m_sync_list)) {
+			if (!timeleft)
+				timeleft = xfs_syncd_centisecs *
+							msecs_to_jiffies(10);
+			INIT_LIST_HEAD(&mp->m_sync_work.w_list);
+			list_add_tail(&mp->m_sync_work.w_list,
+					&mp->m_sync_list);
+		}
+		list_for_each_entry_safe(work, n, &mp->m_sync_list, w_list)
+			list_move(&work->w_list, &tmp);
+		spin_unlock(&mp->m_sync_lock);
+
+		list_for_each_entry_safe(work, n, &tmp, w_list) {
+			(*work->w_syncer)(mp, work->w_data);
+			list_del(&work->w_list);
+			if (work == &mp->m_sync_work)
+				continue;
+			kmem_free(work);
+		}
+	}
+
+	return 0;
+}
+
+int
+xfs_syncd_init(
+	struct xfs_mount	*mp)
+{
+	mp->m_sync_work.w_syncer = xfs_sync_worker;
+	mp->m_sync_work.w_mount = mp;
+	mp->m_sync_task = kthread_run(xfssyncd, mp, "xfssyncd");
+	if (IS_ERR(mp->m_sync_task))
+		return -PTR_ERR(mp->m_sync_task);
+	return 0;
+}
+
+void
+xfs_syncd_stop(
+	struct xfs_mount	*mp)
+{
+	kthread_stop(mp->m_sync_task);
+}
+
+int
+xfs_reclaim_inode(
+	xfs_inode_t	*ip,
+	int		locked,
+	int		sync_mode)
+{
+	xfs_perag_t	*pag = xfs_get_perag(ip->i_mount, ip->i_ino);
+
+	/* The hash lock here protects a thread in xfs_iget_core from
+	 * racing with us on linking the inode back with a vnode.
+	 * Once we have the XFS_IRECLAIM flag set it will not touch
+	 * us.
+	 */
+	write_lock(&pag->pag_ici_lock);
+	spin_lock(&ip->i_flags_lock);
+	if (__xfs_iflags_test(ip, XFS_IRECLAIM) ||
+	    !__xfs_iflags_test(ip, XFS_IRECLAIMABLE)) {
+		spin_unlock(&ip->i_flags_lock);
+		write_unlock(&pag->pag_ici_lock);
+		if (locked) {
+			xfs_ifunlock(ip);
+			xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		}
+		return 1;
+	}
+	__xfs_iflags_set(ip, XFS_IRECLAIM);
+	spin_unlock(&ip->i_flags_lock);
+	write_unlock(&pag->pag_ici_lock);
+	xfs_put_perag(ip->i_mount, pag);
+
+	/*
+	 * If the inode is still dirty, then flush it out.  If the inode
+	 * is not in the AIL, then it will be OK to flush it delwri as
+	 * long as xfs_iflush() does not keep any references to the inode.
+	 * We leave that decision up to xfs_iflush() since it has the
+	 * knowledge of whether it's OK to simply do a delwri flush of
+	 * the inode or whether we need to wait until the inode is
+	 * pulled from the AIL.
+	 * We get the flush lock regardless, though, just to make sure
+	 * we don't free it while it is being flushed.
+	 */
+	if (!locked) {
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
+		xfs_iflock(ip);
+	}
+
+	/*
+	 * In the case of a forced shutdown we rely on xfs_iflush() to
+	 * wait for the inode to be unpinned before returning an error.
+	 */
+	if (!is_bad_inode(VFS_I(ip)) && xfs_iflush(ip, sync_mode) == 0) {
+		/* synchronize with xfs_iflush_done */
+		xfs_iflock(ip);
+		xfs_ifunlock(ip);
+	}
+
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	xfs_ireclaim(ip);
+	return 0;
+}
+
+/*
+ * We set the inode flag atomically with the radix tree tag.
+ * Once we get tag lookups on the radix tree, this inode flag
+ * can go away.
+ */
+void
+xfs_inode_set_reclaim_tag(
+	xfs_inode_t	*ip)
+{
+	xfs_mount_t	*mp = ip->i_mount;
+	xfs_perag_t	*pag = xfs_get_perag(mp, ip->i_ino);
+
+	read_lock(&pag->pag_ici_lock);
+	spin_lock(&ip->i_flags_lock);
+	radix_tree_tag_set(&pag->pag_ici_root,
+			XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
+	__xfs_iflags_set(ip, XFS_IRECLAIMABLE);
+	spin_unlock(&ip->i_flags_lock);
+	read_unlock(&pag->pag_ici_lock);
+	xfs_put_perag(mp, pag);
+}
+
+void
+__xfs_inode_clear_reclaim_tag(
+	xfs_mount_t	*mp,
+	xfs_perag_t	*pag,
+	xfs_inode_t	*ip)
+{
+	radix_tree_tag_clear(&pag->pag_ici_root,
+			XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
+}
+
+void
+xfs_inode_clear_reclaim_tag(
+	xfs_inode_t	*ip)
+{
+	xfs_mount_t	*mp = ip->i_mount;
+	xfs_perag_t	*pag = xfs_get_perag(mp, ip->i_ino);
+
+	read_lock(&pag->pag_ici_lock);
+	spin_lock(&ip->i_flags_lock);
+	__xfs_inode_clear_reclaim_tag(mp, pag, ip);
+	spin_unlock(&ip->i_flags_lock);
+	read_unlock(&pag->pag_ici_lock);
+	xfs_put_perag(mp, pag);
+}
+
+
+STATIC void
+xfs_reclaim_inodes_ag(
+	xfs_mount_t	*mp,
+	int		ag,
+	int		noblock,
+	int		mode)
+{
+	xfs_inode_t	*ip = NULL;
+	xfs_perag_t	*pag = &mp->m_perag[ag];
+	int		nr_found;
+	uint32_t	first_index;
+	int		skipped;
+
+restart:
+	first_index = 0;
+	skipped = 0;
+	do {
+		/*
+		 * use a gang lookup to find the next inode in the tree
+		 * as the tree is sparse and a gang lookup walks to find
+		 * the number of objects requested.
+		 */
+		read_lock(&pag->pag_ici_lock);
+		nr_found = radix_tree_gang_lookup_tag(&pag->pag_ici_root,
+					(void**)&ip, first_index, 1,
+					XFS_ICI_RECLAIM_TAG);
+
+		if (!nr_found) {
+			read_unlock(&pag->pag_ici_lock);
+			break;
+		}
+
+		/*
+		 * Update the index for the next lookup. Catch overflows
+		 * into the next AG range which can occur if we have inodes
+		 * in the last block of the AG and we are currently
+		 * pointing to the last inode.
+		 */
+		first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+		if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino)) {
+			read_unlock(&pag->pag_ici_lock);
+			break;
+		}
+
+		/* ignore if already under reclaim */
+		if (xfs_iflags_test(ip, XFS_IRECLAIM)) {
+			read_unlock(&pag->pag_ici_lock);
+			continue;
+		}
+
+		if (noblock) {
+			if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
+				read_unlock(&pag->pag_ici_lock);
+				continue;
+			}
+			if (xfs_ipincount(ip) ||
+			    !xfs_iflock_nowait(ip)) {
+				xfs_iunlock(ip, XFS_ILOCK_EXCL);
+				read_unlock(&pag->pag_ici_lock);
+				continue;
+			}
+		}
+		read_unlock(&pag->pag_ici_lock);
+
+		/*
+		 * hmmm - this is an inode already in reclaim. Do
+		 * we even bother catching it here?
+		 */
+		if (xfs_reclaim_inode(ip, noblock, mode))
+			skipped++;
+	} while (nr_found);
+
+	if (skipped) {
+		delay(1);
+		goto restart;
+	}
+	return;
+
+}
+
+int
+xfs_reclaim_inodes(
+	xfs_mount_t	*mp,
+	int		 noblock,
+	int		mode)
+{
+	int		i;
+
+	for (i = 0; i < mp->m_sb.sb_agcount; i++) {
+		if (!mp->m_perag[i].pag_ici_init)
+			continue;
+		xfs_reclaim_inodes_ag(mp, i, noblock, mode);
+	}
+	return 0;
+}
+
+
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
new file mode 100644
index 00000000000..5f6de1efe1f
--- /dev/null
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -0,0 +1,55 @@
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef XFS_SYNC_H
+#define XFS_SYNC_H 1
+
+struct xfs_mount;
+
+typedef struct bhv_vfs_sync_work {
+	struct list_head	w_list;
+	struct xfs_mount	*w_mount;
+	void			*w_data;	/* syncer routine argument */
+	void			(*w_syncer)(struct xfs_mount *, void *);
+} bhv_vfs_sync_work_t;
+
+#define SYNC_ATTR		0x0001	/* sync attributes */
+#define SYNC_DELWRI		0x0002	/* look at delayed writes */
+#define SYNC_WAIT		0x0004	/* wait for i/o to complete */
+#define SYNC_BDFLUSH		0x0008	/* BDFLUSH is calling -- don't block */
+#define SYNC_IOWAIT		0x0010  /* wait for all I/O to complete */
+
+int xfs_syncd_init(struct xfs_mount *mp);
+void xfs_syncd_stop(struct xfs_mount *mp);
+
+int xfs_sync_inodes(struct xfs_mount *mp, int flags);
+int xfs_sync_fsdata(struct xfs_mount *mp, int flags);
+
+int xfs_quiesce_data(struct xfs_mount *mp);
+void xfs_quiesce_attr(struct xfs_mount *mp);
+
+void xfs_flush_inode(struct xfs_inode *ip);
+void xfs_flush_device(struct xfs_inode *ip);
+
+int xfs_reclaim_inode(struct xfs_inode *ip, int locked, int sync_mode);
+int xfs_reclaim_inodes(struct xfs_mount *mp, int noblock, int mode);
+
+void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
+void xfs_inode_clear_reclaim_tag(struct xfs_inode *ip);
+void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
+				struct xfs_inode *ip);
+#endif
diff --git a/fs/xfs/linux-2.6/xfs_sysctl.c b/fs/xfs/linux-2.6/xfs_sysctl.c
index 7dacb5bbde3..916c0ffb608 100644
--- a/fs/xfs/linux-2.6/xfs_sysctl.c
+++ b/fs/xfs/linux-2.6/xfs_sysctl.c
@@ -56,17 +56,6 @@ xfs_stats_clear_proc_handler(
 
 static ctl_table xfs_table[] = {
 	{
-		.ctl_name	= XFS_RESTRICT_CHOWN,
-		.procname	= "restrict_chown",
-		.data		= &xfs_params.restrict_chown.val,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= &proc_dointvec_minmax,
-		.strategy	= &sysctl_intvec,
-		.extra1		= &xfs_params.restrict_chown.min,
-		.extra2		= &xfs_params.restrict_chown.max
-	},
-	{
 		.ctl_name	= XFS_SGID_INHERIT,
 		.procname	= "irix_sgid_inherit",
 		.data		= &xfs_params.sgid_inherit.val,
diff --git a/fs/xfs/linux-2.6/xfs_sysctl.h b/fs/xfs/linux-2.6/xfs_sysctl.h
index 4aadb8056c3..b9937d450f8 100644
--- a/fs/xfs/linux-2.6/xfs_sysctl.h
+++ b/fs/xfs/linux-2.6/xfs_sysctl.h
@@ -31,7 +31,6 @@ typedef struct xfs_sysctl_val {
 } xfs_sysctl_val_t;
 
 typedef struct xfs_param {
-	xfs_sysctl_val_t restrict_chown;/* Root/non-root can give away files.*/
 	xfs_sysctl_val_t sgid_inherit;	/* Inherit S_ISGID if process' GID is
 					 * not a member of parent dir GID. */
 	xfs_sysctl_val_t symlink_mode;	/* Link creat mode affected by umask */
@@ -68,7 +67,7 @@ typedef struct xfs_param {
 enum {
 	/* XFS_REFCACHE_SIZE = 1 */
 	/* XFS_REFCACHE_PURGE = 2 */
-	XFS_RESTRICT_CHOWN = 3,
+	/* XFS_RESTRICT_CHOWN = 3 */
 	XFS_SGID_INHERIT = 4,
 	XFS_SYMLINK_MODE = 5,
 	XFS_PANIC_MASK = 6,
diff --git a/fs/xfs/linux-2.6/xfs_vfs.h b/fs/xfs/linux-2.6/xfs_vfs.h
deleted file mode 100644
index 7e60c7776b1..00000000000
--- a/fs/xfs/linux-2.6/xfs_vfs.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_VFS_H__
-#define __XFS_VFS_H__
-
-#include <linux/vfs.h>
-#include "xfs_fs.h"
-
-struct inode;
-
-struct fid;
-struct cred;
-struct seq_file;
-struct super_block;
-struct xfs_inode;
-struct xfs_mount;
-struct xfs_mount_args;
-
-typedef struct kstatfs	bhv_statvfs_t;
-
-typedef struct bhv_vfs_sync_work {
-	struct list_head	w_list;
-	struct xfs_mount	*w_mount;
-	void			*w_data;	/* syncer routine argument */
-	void			(*w_syncer)(struct xfs_mount *, void *);
-} bhv_vfs_sync_work_t;
-
-#define SYNC_ATTR		0x0001	/* sync attributes */
-#define SYNC_CLOSE		0x0002	/* close file system down */
-#define SYNC_DELWRI		0x0004	/* look at delayed writes */
-#define SYNC_WAIT		0x0008	/* wait for i/o to complete */
-#define SYNC_BDFLUSH		0x0010	/* BDFLUSH is calling -- don't block */
-#define SYNC_FSDATA		0x0020	/* flush fs data (e.g. superblocks) */
-#define SYNC_REFCACHE		0x0040  /* prune some of the nfs ref cache */
-#define SYNC_REMOUNT		0x0080  /* remount readonly, no dummy LRs */
-#define SYNC_IOWAIT		0x0100  /* wait for all I/O to complete */
-
-/*
- * When remounting a filesystem read-only or freezing the filesystem,
- * we have two phases to execute. This first phase is syncing the data
- * before we quiesce the fielsystem, and the second is flushing all the
- * inodes out after we've waited for all the transactions created by
- * the first phase to complete. The second phase uses SYNC_INODE_QUIESCE
- * to ensure that the inodes are written to their location on disk
- * rather than just existing in transactions in the log. This means
- * after a quiesce there is no log replay required to write the inodes
- * to disk (this is the main difference between a sync and a quiesce).
- */
-#define SYNC_DATA_QUIESCE	(SYNC_DELWRI|SYNC_FSDATA|SYNC_WAIT|SYNC_IOWAIT)
-#define SYNC_INODE_QUIESCE	(SYNC_REMOUNT|SYNC_ATTR|SYNC_WAIT)
-
-#define SHUTDOWN_META_IO_ERROR	0x0001	/* write attempt to metadata failed */
-#define SHUTDOWN_LOG_IO_ERROR	0x0002	/* write attempt to the log failed */
-#define SHUTDOWN_FORCE_UMOUNT	0x0004	/* shutdown from a forced unmount */
-#define SHUTDOWN_CORRUPT_INCORE	0x0008	/* corrupt in-memory data structures */
-#define SHUTDOWN_REMOTE_REQ	0x0010	/* shutdown came from remote cell */
-#define SHUTDOWN_DEVICE_REQ	0x0020	/* failed all paths to the device */
-
-#define xfs_test_for_freeze(mp)		((mp)->m_super->s_frozen)
-#define xfs_wait_for_freeze(mp,l)	vfs_check_frozen((mp)->m_super, (l))
-
-#endif	/* __XFS_VFS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_vnode.c b/fs/xfs/linux-2.6/xfs_vnode.c
deleted file mode 100644
index b52528bbbff..00000000000
--- a/fs/xfs/linux-2.6/xfs_vnode.c
+++ /dev/null
@@ -1,145 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_vnodeops.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_inode.h"
-
-/*
- * And this gunk is needed for xfs_mount.h"
- */
-#include "xfs_log.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_dmapi.h"
-#include "xfs_inum.h"
-#include "xfs_ag.h"
-#include "xfs_mount.h"
-
-
-/*
- * Dedicated vnode inactive/reclaim sync wait queues.
- * Prime number of hash buckets since address is used as the key.
- */
-#define NVSYNC                  37
-#define vptosync(v)             (&vsync[((unsigned long)v) % NVSYNC])
-static wait_queue_head_t vsync[NVSYNC];
-
-void __init
-vn_init(void)
-{
-	int i;
-
-	for (i = 0; i < NVSYNC; i++)
-		init_waitqueue_head(&vsync[i]);
-}
-
-void
-vn_iowait(
-	xfs_inode_t	*ip)
-{
-	wait_queue_head_t *wq = vptosync(ip);
-
-	wait_event(*wq, (atomic_read(&ip->i_iocount) == 0));
-}
-
-void
-vn_iowake(
-	xfs_inode_t	*ip)
-{
-	if (atomic_dec_and_test(&ip->i_iocount))
-		wake_up(vptosync(ip));
-}
-
-/*
- * Volume managers supporting multiple paths can send back ENODEV when the
- * final path disappears.  In this case continuing to fill the page cache
- * with dirty data which cannot be written out is evil, so prevent that.
- */
-void
-vn_ioerror(
-	xfs_inode_t	*ip,
-	int		error,
-	char		*f,
-	int		l)
-{
-	if (unlikely(error == -ENODEV))
-		xfs_do_force_shutdown(ip->i_mount, SHUTDOWN_DEVICE_REQ, f, l);
-}
-
-#ifdef	XFS_INODE_TRACE
-
-/*
- * Reference count of Linux inode if present, -1 if the xfs_inode
- * has no associated Linux inode.
- */
-static inline int xfs_icount(struct xfs_inode *ip)
-{
-	struct inode *vp = VFS_I(ip);
-
-	if (vp)
-		return vn_count(vp);
-	return -1;
-}
-
-#define KTRACE_ENTER(ip, vk, s, line, ra)			\
-	ktrace_enter(	(ip)->i_trace,				\
-/*  0 */		(void *)(__psint_t)(vk),		\
-/*  1 */		(void *)(s),				\
-/*  2 */		(void *)(__psint_t) line,		\
-/*  3 */		(void *)(__psint_t)xfs_icount(ip),	\
-/*  4 */		(void *)(ra),				\
-/*  5 */		NULL,					\
-/*  6 */		(void *)(__psint_t)current_cpu(),	\
-/*  7 */		(void *)(__psint_t)current_pid(),	\
-/*  8 */		(void *)__return_address,		\
-/*  9 */		NULL, NULL, NULL, NULL, NULL, NULL, NULL)
-
-/*
- * Vnode tracing code.
- */
-void
-_xfs_itrace_entry(xfs_inode_t *ip, const char *func, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_ENTRY, func, 0, ra);
-}
-
-void
-_xfs_itrace_exit(xfs_inode_t *ip, const char *func, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_EXIT, func, 0, ra);
-}
-
-void
-xfs_itrace_hold(xfs_inode_t *ip, char *file, int line, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_HOLD, file, line, ra);
-}
-
-void
-_xfs_itrace_ref(xfs_inode_t *ip, char *file, int line, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_REF, file, line, ra);
-}
-
-void
-xfs_itrace_rele(xfs_inode_t *ip, char *file, int line, inst_t *ra)
-{
-	KTRACE_ENTER(ip, INODE_KTRACE_RELE, file, line, ra);
-}
-#endif	/* XFS_INODE_TRACE */
diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h
index 683ce16210f..f65983a230d 100644
--- a/fs/xfs/linux-2.6/xfs_vnode.h
+++ b/fs/xfs/linux-2.6/xfs_vnode.h
@@ -18,7 +18,10 @@
 #ifndef __XFS_VNODE_H__
 #define __XFS_VNODE_H__
 
+#include "xfs_fs.h"
+
 struct file;
+struct xfs_inode;
 struct xfs_iomap;
 struct attrlist_cursor_kern;
 
@@ -51,40 +54,6 @@ struct attrlist_cursor_kern;
 					   Prevent VM access to the pages until
 					   the operation completes. */
 
-
-extern void	vn_init(void);
-
-/*
- * Yeah, these don't take vnode anymore at all, all this should be
- * cleaned up at some point.
- */
-extern void	vn_iowait(struct xfs_inode *ip);
-extern void	vn_iowake(struct xfs_inode *ip);
-extern void	vn_ioerror(struct xfs_inode *ip, int error, char *f, int l);
-
-static inline int vn_count(struct inode *vp)
-{
-	return atomic_read(&vp->i_count);
-}
-
-#define IHOLD(ip) \
-do { \
-	ASSERT(atomic_read(&VFS_I(ip)->i_count) > 0) ; \
-	atomic_inc(&(VFS_I(ip)->i_count)); \
-	xfs_itrace_hold((ip), __FILE__, __LINE__, (inst_t *)__return_address); \
-} while (0)
-
-#define IRELE(ip) \
-do { \
-	xfs_itrace_rele((ip), __FILE__, __LINE__, (inst_t *)__return_address); \
-	iput(VFS_I(ip)); \
-} while (0)
-
-static inline struct inode *vn_grab(struct inode *vp)
-{
-	return igrab(vp);
-}
-
 /*
  * Dealing with bad inodes
  */
@@ -121,39 +90,4 @@ static inline void vn_atime_to_time_t(struct inode *vp, time_t *tt)
 					PAGECACHE_TAG_DIRTY)
 
 
-/*
- * Tracking vnode activity.
- */
-#if defined(XFS_INODE_TRACE)
-
-#define	INODE_TRACE_SIZE	16		/* number of trace entries */
-#define	INODE_KTRACE_ENTRY	1
-#define	INODE_KTRACE_EXIT	2
-#define	INODE_KTRACE_HOLD	3
-#define	INODE_KTRACE_REF	4
-#define	INODE_KTRACE_RELE	5
-
-extern void _xfs_itrace_entry(struct xfs_inode *, const char *, inst_t *);
-extern void _xfs_itrace_exit(struct xfs_inode *, const char *, inst_t *);
-extern void xfs_itrace_hold(struct xfs_inode *, char *, int, inst_t *);
-extern void _xfs_itrace_ref(struct xfs_inode *, char *, int, inst_t *);
-extern void xfs_itrace_rele(struct xfs_inode *, char *, int, inst_t *);
-#define xfs_itrace_entry(ip)	\
-	_xfs_itrace_entry(ip, __func__, (inst_t *)__return_address)
-#define xfs_itrace_exit(ip)	\
-	_xfs_itrace_exit(ip, __func__, (inst_t *)__return_address)
-#define xfs_itrace_exit_tag(ip, tag)	\
-	_xfs_itrace_exit(ip, tag, (inst_t *)__return_address)
-#define xfs_itrace_ref(ip)	\
-	_xfs_itrace_ref(ip, __FILE__, __LINE__, (inst_t *)__return_address)
-
-#else
-#define	xfs_itrace_entry(a)
-#define	xfs_itrace_exit(a)
-#define	xfs_itrace_exit_tag(a, b)
-#define	xfs_itrace_hold(a, b, c, d)
-#define	xfs_itrace_ref(a)
-#define	xfs_itrace_rele(a, b, c, d)
-#endif
-
 #endif	/* __XFS_VNODE_H__ */
diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c
index f2705f2fd43..591ca6602bf 100644
--- a/fs/xfs/quota/xfs_dquot.c
+++ b/fs/xfs/quota/xfs_dquot.c
@@ -101,7 +101,7 @@ xfs_qm_dqinit(
 	if (brandnewdquot) {
 		dqp->dq_flnext = dqp->dq_flprev = dqp;
 		mutex_init(&dqp->q_qlock);
-		sv_init(&dqp->q_pinwait, SV_DEFAULT, "pdq");
+		init_waitqueue_head(&dqp->q_pinwait);
 
 		/*
 		 * Because we want to use a counting completion, complete
@@ -131,7 +131,7 @@ xfs_qm_dqinit(
 		 dqp->q_res_bcount = 0;
 		 dqp->q_res_icount = 0;
 		 dqp->q_res_rtbcount = 0;
-		 dqp->q_pincount = 0;
+		 atomic_set(&dqp->q_pincount, 0);
 		 dqp->q_hash = NULL;
 		 ASSERT(dqp->dq_flnext == dqp->dq_flprev);
 
@@ -1221,16 +1221,14 @@ xfs_qm_dqflush(
 	xfs_dqtrace_entry(dqp, "DQFLUSH");
 
 	/*
-	 * If not dirty, nada.
+	 * If not dirty, or it's pinned and we are not supposed to
+	 * block, nada.
 	 */
-	if (!XFS_DQ_IS_DIRTY(dqp)) {
+	if (!XFS_DQ_IS_DIRTY(dqp) ||
+	    (!(flags & XFS_QMOPT_SYNC) && atomic_read(&dqp->q_pincount) > 0)) {
 		xfs_dqfunlock(dqp);
-		return (0);
+		return 0;
 	}
-
-	/*
-	 * Cant flush a pinned dquot. Wait for it.
-	 */
 	xfs_qm_dqunpin_wait(dqp);
 
 	/*
@@ -1274,10 +1272,8 @@ xfs_qm_dqflush(
 	dqp->dq_flags &= ~(XFS_DQ_DIRTY);
 	mp = dqp->q_mount;
 
-	/* lsn is 64 bits */
-	spin_lock(&mp->m_ail_lock);
-	dqp->q_logitem.qli_flush_lsn = dqp->q_logitem.qli_item.li_lsn;
-	spin_unlock(&mp->m_ail_lock);
+	xfs_trans_ail_copy_lsn(mp->m_ail, &dqp->q_logitem.qli_flush_lsn,
+					&dqp->q_logitem.qli_item.li_lsn);
 
 	/*
 	 * Attach an iodone routine so that we can remove this dquot from the
@@ -1323,8 +1319,10 @@ xfs_qm_dqflush_done(
 	xfs_dq_logitem_t	*qip)
 {
 	xfs_dquot_t		*dqp;
+	struct xfs_ail		*ailp;
 
 	dqp = qip->qli_dquot;
+	ailp = qip->qli_item.li_ailp;
 
 	/*
 	 * We only want to pull the item from the AIL if its
@@ -1337,15 +1335,12 @@ xfs_qm_dqflush_done(
 	if ((qip->qli_item.li_flags & XFS_LI_IN_AIL) &&
 	    qip->qli_item.li_lsn == qip->qli_flush_lsn) {
 
-		spin_lock(&dqp->q_mount->m_ail_lock);
-		/*
-		 * xfs_trans_delete_ail() drops the AIL lock.
-		 */
+		/* xfs_trans_ail_delete() drops the AIL lock. */
+		spin_lock(&ailp->xa_lock);
 		if (qip->qli_item.li_lsn == qip->qli_flush_lsn)
-			xfs_trans_delete_ail(dqp->q_mount,
-					     (xfs_log_item_t*)qip);
+			xfs_trans_ail_delete(ailp, (xfs_log_item_t*)qip);
 		else
-			spin_unlock(&dqp->q_mount->m_ail_lock);
+			spin_unlock(&ailp->xa_lock);
 	}
 
 	/*
@@ -1375,7 +1370,7 @@ xfs_dqunlock(
 	mutex_unlock(&(dqp->q_qlock));
 	if (dqp->q_logitem.qli_dquot == dqp) {
 		/* Once was dqp->q_mount, but might just have been cleared */
-		xfs_trans_unlocked_item(dqp->q_logitem.qli_item.li_mountp,
+		xfs_trans_unlocked_item(dqp->q_logitem.qli_item.li_ailp,
 					(xfs_log_item_t*)&(dqp->q_logitem));
 	}
 }
@@ -1489,7 +1484,7 @@ xfs_qm_dqpurge(
 				"xfs_qm_dqpurge: dquot %p flush failed", dqp);
 		xfs_dqflock(dqp);
 	}
-	ASSERT(dqp->q_pincount == 0);
+	ASSERT(atomic_read(&dqp->q_pincount) == 0);
 	ASSERT(XFS_FORCED_SHUTDOWN(mp) ||
 	       !(dqp->q_logitem.qli_item.li_flags & XFS_LI_IN_AIL));
 
diff --git a/fs/xfs/quota/xfs_dquot.h b/fs/xfs/quota/xfs_dquot.h
index 8958d0faf8d..7e455337e2b 100644
--- a/fs/xfs/quota/xfs_dquot.h
+++ b/fs/xfs/quota/xfs_dquot.h
@@ -83,8 +83,8 @@ typedef struct xfs_dquot {
 	xfs_qcnt_t	 q_res_rtbcount;/* total realtime blks used+reserved */
 	mutex_t		 q_qlock;	/* quota lock */
 	struct completion q_flush;	/* flush completion queue */
-	uint		 q_pincount;	/* pin count for this dquot */
-	sv_t		 q_pinwait;	/* sync var for pinning */
+	atomic_t          q_pincount;	/* dquot pin count */
+	wait_queue_head_t q_pinwait;	/* dquot pinning wait queue */
 #ifdef XFS_DQUOT_TRACE
 	struct ktrace	*q_trace;	/* trace header structure */
 #endif
diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c
index f028644caa5..1728f6a7c4f 100644
--- a/fs/xfs/quota/xfs_dquot_item.c
+++ b/fs/xfs/quota/xfs_dquot_item.c
@@ -88,25 +88,22 @@ xfs_qm_dquot_logitem_format(
 
 /*
  * Increment the pin count of the given dquot.
- * This value is protected by pinlock spinlock in the xQM structure.
  */
 STATIC void
 xfs_qm_dquot_logitem_pin(
 	xfs_dq_logitem_t *logitem)
 {
-	xfs_dquot_t *dqp;
+	xfs_dquot_t *dqp = logitem->qli_dquot;
 
-	dqp = logitem->qli_dquot;
 	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	spin_lock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-	dqp->q_pincount++;
-	spin_unlock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
+	atomic_inc(&dqp->q_pincount);
 }
 
 /*
  * Decrement the pin count of the given dquot, and wake up
  * anyone in xfs_dqwait_unpin() if the count goes to 0.	 The
- * dquot must have been previously pinned with a call to xfs_dqpin().
+ * dquot must have been previously pinned with a call to
+ * xfs_qm_dquot_logitem_pin().
  */
 /* ARGSUSED */
 STATIC void
@@ -114,16 +111,11 @@ xfs_qm_dquot_logitem_unpin(
 	xfs_dq_logitem_t *logitem,
 	int		  stale)
 {
-	xfs_dquot_t *dqp;
+	xfs_dquot_t *dqp = logitem->qli_dquot;
 
-	dqp = logitem->qli_dquot;
-	ASSERT(dqp->q_pincount > 0);
-	spin_lock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-	dqp->q_pincount--;
-	if (dqp->q_pincount == 0) {
-		sv_broadcast(&dqp->q_pinwait);
-	}
-	spin_unlock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
+	ASSERT(atomic_read(&dqp->q_pincount) > 0);
+	if (atomic_dec_and_test(&dqp->q_pincount))
+		wake_up(&dqp->q_pinwait);
 }
 
 /* ARGSUSED */
@@ -193,21 +185,14 @@ xfs_qm_dqunpin_wait(
 	xfs_dquot_t	*dqp)
 {
 	ASSERT(XFS_DQ_IS_LOCKED(dqp));
-	if (dqp->q_pincount == 0) {
+	if (atomic_read(&dqp->q_pincount) == 0)
 		return;
-	}
 
 	/*
 	 * Give the log a push so we don't wait here too long.
 	 */
 	xfs_log_force(dqp->q_mount, (xfs_lsn_t)0, XFS_LOG_FORCE);
-	spin_lock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-	if (dqp->q_pincount == 0) {
-		spin_unlock(&(XFS_DQ_TO_QINF(dqp)->qi_pinlock));
-		return;
-	}
-	sv_wait(&(dqp->q_pinwait), PINOD,
-		&(XFS_DQ_TO_QINF(dqp)->qi_pinlock), s);
+	wait_event(dqp->q_pinwait, (atomic_read(&dqp->q_pincount) == 0));
 }
 
 /*
@@ -310,7 +295,7 @@ xfs_qm_dquot_logitem_trylock(
 	uint			retval;
 
 	dqp = qip->qli_dquot;
-	if (dqp->q_pincount > 0)
+	if (atomic_read(&dqp->q_pincount) > 0)
 		return (XFS_ITEM_PINNED);
 
 	if (! xfs_qm_dqlock_nowait(dqp))
@@ -568,14 +553,16 @@ xfs_qm_qoffend_logitem_committed(
 	xfs_lsn_t lsn)
 {
 	xfs_qoff_logitem_t	*qfs;
+	struct xfs_ail		*ailp;
 
 	qfs = qfe->qql_start_lip;
-	spin_lock(&qfs->qql_item.li_mountp->m_ail_lock);
+	ailp = qfs->qql_item.li_ailp;
+	spin_lock(&ailp->xa_lock);
 	/*
 	 * Delete the qoff-start logitem from the AIL.
-	 * xfs_trans_delete_ail() drops the AIL lock.
+	 * xfs_trans_ail_delete() drops the AIL lock.
 	 */
-	xfs_trans_delete_ail(qfs->qql_item.li_mountp, (xfs_log_item_t *)qfs);
+	xfs_trans_ail_delete(ailp, (xfs_log_item_t *)qfs);
 	kmem_free(qfs);
 	kmem_free(qfe);
 	return (xfs_lsn_t)-1;
diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c
index df0ffef9775..6b13960cf31 100644
--- a/fs/xfs/quota/xfs_qm.c
+++ b/fs/xfs/quota/xfs_qm.c
@@ -20,7 +20,6 @@
 #include "xfs_bit.h"
 #include "xfs_log.h"
 #include "xfs_inum.h"
-#include "xfs_clnt.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
@@ -396,13 +395,10 @@ xfs_qm_mount_quotas(
 /*
  * Called from the vfsops layer.
  */
-int
+void
 xfs_qm_unmount_quotas(
 	xfs_mount_t	*mp)
 {
-	xfs_inode_t	*uqp, *gqp;
-	int		error = 0;
-
 	/*
 	 * Release the dquots that root inode, et al might be holding,
 	 * before we flush quotas and blow away the quotainfo structure.
@@ -415,43 +411,18 @@ xfs_qm_unmount_quotas(
 		xfs_qm_dqdetach(mp->m_rsumip);
 
 	/*
-	 * Flush out the quota inodes.
+	 * Release the quota inodes.
 	 */
-	uqp = gqp = NULL;
 	if (mp->m_quotainfo) {
-		if ((uqp = mp->m_quotainfo->qi_uquotaip) != NULL) {
-			xfs_ilock(uqp, XFS_ILOCK_EXCL);
-			xfs_iflock(uqp);
-			error = xfs_iflush(uqp, XFS_IFLUSH_SYNC);
-			xfs_iunlock(uqp, XFS_ILOCK_EXCL);
-			if (unlikely(error == EFSCORRUPTED)) {
-				XFS_ERROR_REPORT("xfs_qm_unmount_quotas(1)",
-						 XFS_ERRLEVEL_LOW, mp);
-				goto out;
-			}
+		if (mp->m_quotainfo->qi_uquotaip) {
+			IRELE(mp->m_quotainfo->qi_uquotaip);
+			mp->m_quotainfo->qi_uquotaip = NULL;
 		}
-		if ((gqp = mp->m_quotainfo->qi_gquotaip) != NULL) {
-			xfs_ilock(gqp, XFS_ILOCK_EXCL);
-			xfs_iflock(gqp);
-			error = xfs_iflush(gqp, XFS_IFLUSH_SYNC);
-			xfs_iunlock(gqp, XFS_ILOCK_EXCL);
-			if (unlikely(error == EFSCORRUPTED)) {
-				XFS_ERROR_REPORT("xfs_qm_unmount_quotas(2)",
-						 XFS_ERRLEVEL_LOW, mp);
-				goto out;
-			}
+		if (mp->m_quotainfo->qi_gquotaip) {
+			IRELE(mp->m_quotainfo->qi_gquotaip);
+			mp->m_quotainfo->qi_gquotaip = NULL;
 		}
 	}
-	if (uqp) {
-		 IRELE(uqp);
-		 mp->m_quotainfo->qi_uquotaip = NULL;
-	}
-	if (gqp) {
-		IRELE(gqp);
-		mp->m_quotainfo->qi_gquotaip = NULL;
-	}
-out:
-	return XFS_ERROR(error);
 }
 
 /*
@@ -987,14 +958,10 @@ xfs_qm_dqdetach(
 }
 
 /*
- * This is called by VFS_SYNC and flags arg determines the caller,
- * and its motives, as done in xfs_sync.
- *
- * vfs_sync: SYNC_FSDATA|SYNC_ATTR|SYNC_BDFLUSH 0x31
- * syscall sync: SYNC_FSDATA|SYNC_ATTR|SYNC_DELWRI 0x25
- * umountroot : SYNC_WAIT | SYNC_CLOSE | SYNC_ATTR | SYNC_FSDATA
+ * This is called to sync quotas. We can be told to use non-blocking
+ * semantics by either the SYNC_BDFLUSH flag or the absence of the
+ * SYNC_WAIT flag.
  */
-
 int
 xfs_qm_sync(
 	xfs_mount_t	*mp,
@@ -1137,7 +1104,6 @@ xfs_qm_init_quotainfo(
 		return error;
 	}
 
-	spin_lock_init(&qinf->qi_pinlock);
 	xfs_qm_list_init(&qinf->qi_dqlist, "mpdqlist", 0);
 	qinf->qi_dqreclaims = 0;
 
@@ -1234,7 +1200,6 @@ xfs_qm_destroy_quotainfo(
 	 */
 	xfs_qm_rele_quotafs_ref(mp);
 
-	spinlock_destroy(&qi->qi_pinlock);
 	xfs_qm_list_destroy(&qi->qi_dqlist);
 
 	if (qi->qi_uquotaip) {
diff --git a/fs/xfs/quota/xfs_qm.h b/fs/xfs/quota/xfs_qm.h
index 44f25349e47..ddf09166387 100644
--- a/fs/xfs/quota/xfs_qm.h
+++ b/fs/xfs/quota/xfs_qm.h
@@ -106,7 +106,6 @@ typedef struct xfs_qm {
 typedef struct xfs_quotainfo {
 	xfs_inode_t	*qi_uquotaip;	 /* user quota inode */
 	xfs_inode_t	*qi_gquotaip;	 /* group quota inode */
-	spinlock_t	 qi_pinlock;	 /* dquot pinning lock */
 	xfs_dqlist_t	 qi_dqlist;	 /* all dquots in filesys */
 	int		 qi_dqreclaims;	 /* a change here indicates
 					    a removal in the dqlist */
@@ -168,7 +167,7 @@ extern void		xfs_qm_destroy_quotainfo(xfs_mount_t *);
 extern void		xfs_qm_mount_quotas(xfs_mount_t *);
 extern int		xfs_qm_quotacheck(xfs_mount_t *);
 extern void		xfs_qm_unmount_quotadestroy(xfs_mount_t *);
-extern int		xfs_qm_unmount_quotas(xfs_mount_t *);
+extern void		xfs_qm_unmount_quotas(xfs_mount_t *);
 extern int		xfs_qm_write_sb_changes(xfs_mount_t *, __int64_t);
 extern int		xfs_qm_sync(xfs_mount_t *, int);
 
diff --git a/fs/xfs/quota/xfs_qm_bhv.c b/fs/xfs/quota/xfs_qm_bhv.c
index eea2e60b456..bc6c5cca3e1 100644
--- a/fs/xfs/quota/xfs_qm_bhv.c
+++ b/fs/xfs/quota/xfs_qm_bhv.c
@@ -20,7 +20,6 @@
 #include "xfs_bit.h"
 #include "xfs_log.h"
 #include "xfs_inum.h"
-#include "xfs_clnt.h"
 #include "xfs_trans.h"
 #include "xfs_sb.h"
 #include "xfs_ag.h"
@@ -51,7 +50,7 @@
 
 STATIC void
 xfs_fill_statvfs_from_dquot(
-	bhv_statvfs_t		*statp,
+	struct kstatfs		*statp,
 	xfs_disk_dquot_t	*dp)
 {
 	__uint64_t		limit;
@@ -88,7 +87,7 @@ xfs_fill_statvfs_from_dquot(
 STATIC void
 xfs_qm_statvfs(
 	xfs_inode_t		*ip,
-	bhv_statvfs_t		*statp)
+	struct kstatfs		*statp)
 {
 	xfs_mount_t		*mp = ip->i_mount;
 	xfs_dquot_t		*dqp;
diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c
index 1a3b803dfa5..68139b38aed 100644
--- a/fs/xfs/quota/xfs_qm_syscalls.c
+++ b/fs/xfs/quota/xfs_qm_syscalls.c
@@ -127,7 +127,7 @@ xfs_qm_quotactl(
 		break;
 
 	case Q_XQUOTASYNC:
-		return (xfs_sync_inodes(mp, SYNC_DELWRI, NULL));
+		return xfs_sync_inodes(mp, SYNC_DELWRI);
 
 	default:
 		break;
@@ -1022,101 +1022,104 @@ xfs_qm_export_flags(
 
 
 /*
- * Go thru all the inodes in the file system, releasing their dquots.
- * Note that the mount structure gets modified to indicate that quotas are off
- * AFTER this, in the case of quotaoff. This also gets called from
- * xfs_rootumount.
+ * Release all the dquots on the inodes in an AG.
  */
-void
-xfs_qm_dqrele_all_inodes(
-	struct xfs_mount *mp,
-	uint		 flags)
+STATIC void
+xfs_qm_dqrele_inodes_ag(
+	xfs_mount_t	*mp,
+	int		ag,
+	uint		flags)
 {
-	xfs_inode_t	*ip, *topino;
-	uint		ireclaims;
-	struct inode	*vp;
-	boolean_t	vnode_refd;
+	xfs_inode_t	*ip = NULL;
+	xfs_perag_t	*pag = &mp->m_perag[ag];
+	int		first_index = 0;
+	int		nr_found;
 
-	ASSERT(mp->m_quotainfo);
-
-	XFS_MOUNT_ILOCK(mp);
-again:
-	ip = mp->m_inodes;
-	if (ip == NULL) {
-		XFS_MOUNT_IUNLOCK(mp);
-		return;
-	}
 	do {
-		/* Skip markers inserted by xfs_sync */
-		if (ip->i_mount == NULL) {
-			ip = ip->i_mnext;
-			continue;
+		/*
+		 * use a gang lookup to find the next inode in the tree
+		 * as the tree is sparse and a gang lookup walks to find
+		 * the number of objects requested.
+		 */
+		read_lock(&pag->pag_ici_lock);
+		nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
+				(void**)&ip, first_index, 1);
+
+		if (!nr_found) {
+			read_unlock(&pag->pag_ici_lock);
+			break;
 		}
-		/* Root inode, rbmip and rsumip have associated blocks */
+
+		/*
+		 * Update the index for the next lookup. Catch overflows
+		 * into the next AG range which can occur if we have inodes
+		 * in the last block of the AG and we are currently
+		 * pointing to the last inode.
+		 */
+		first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
+		if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino)) {
+			read_unlock(&pag->pag_ici_lock);
+			break;
+		}
+
+		/* skip quota inodes */
 		if (ip == XFS_QI_UQIP(mp) || ip == XFS_QI_GQIP(mp)) {
 			ASSERT(ip->i_udquot == NULL);
 			ASSERT(ip->i_gdquot == NULL);
-			ip = ip->i_mnext;
+			read_unlock(&pag->pag_ici_lock);
 			continue;
 		}
-		vp = VFS_I(ip);
-		if (!vp) {
-			ASSERT(ip->i_udquot == NULL);
-			ASSERT(ip->i_gdquot == NULL);
-			ip = ip->i_mnext;
+
+		/*
+		 * If we can't get a reference on the inode, it must be
+		 * in reclaim. Leave it for the reclaim code to flush.
+		 */
+		if (!igrab(VFS_I(ip))) {
+			read_unlock(&pag->pag_ici_lock);
 			continue;
 		}
-		vnode_refd = B_FALSE;
-		if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL) == 0) {
-			ireclaims = mp->m_ireclaims;
-			topino = mp->m_inodes;
-			vp = vn_grab(vp);
-			if (!vp)
-				goto again;
-
-			XFS_MOUNT_IUNLOCK(mp);
-			/* XXX restart limit ? */
-			xfs_ilock(ip, XFS_ILOCK_EXCL);
-			vnode_refd = B_TRUE;
-		} else {
-			ireclaims = mp->m_ireclaims;
-			topino = mp->m_inodes;
-			XFS_MOUNT_IUNLOCK(mp);
+		read_unlock(&pag->pag_ici_lock);
+
+		/* avoid new inodes though we shouldn't find any here */
+		if (xfs_iflags_test(ip, XFS_INEW)) {
+			IRELE(ip);
+			continue;
 		}
 
-		/*
-		 * We don't keep the mountlock across the dqrele() call,
-		 * since it can take a while..
-		 */
+		xfs_ilock(ip, XFS_ILOCK_EXCL);
 		if ((flags & XFS_UQUOTA_ACCT) && ip->i_udquot) {
 			xfs_qm_dqrele(ip->i_udquot);
 			ip->i_udquot = NULL;
 		}
-		if (flags & (XFS_PQUOTA_ACCT|XFS_GQUOTA_ACCT) && ip->i_gdquot) {
+		if (flags & (XFS_PQUOTA_ACCT|XFS_GQUOTA_ACCT) &&
+		    ip->i_gdquot) {
 			xfs_qm_dqrele(ip->i_gdquot);
 			ip->i_gdquot = NULL;
 		}
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		/*
-		 * Wait until we've dropped the ilock and mountlock to
-		 * do the vn_rele. Or be condemned to an eternity in the
-		 * inactive code in hell.
-		 */
-		if (vnode_refd)
-			IRELE(ip);
-		XFS_MOUNT_ILOCK(mp);
-		/*
-		 * If an inode was inserted or removed, we gotta
-		 * start over again.
-		 */
-		if (topino != mp->m_inodes || mp->m_ireclaims != ireclaims) {
-			/* XXX use a sentinel */
-			goto again;
-		}
-		ip = ip->i_mnext;
-	} while (ip != mp->m_inodes);
+		xfs_iput(ip, XFS_ILOCK_EXCL);
+
+	} while (nr_found);
+}
+
+/*
+ * Go thru all the inodes in the file system, releasing their dquots.
+ * Note that the mount structure gets modified to indicate that quotas are off
+ * AFTER this, in the case of quotaoff. This also gets called from
+ * xfs_rootumount.
+ */
+void
+xfs_qm_dqrele_all_inodes(
+	struct xfs_mount *mp,
+	uint		 flags)
+{
+	int		i;
 
-	XFS_MOUNT_IUNLOCK(mp);
+	ASSERT(mp->m_quotainfo);
+	for (i = 0; i < mp->m_sb.sb_agcount; i++) {
+		if (!mp->m_perag[i].pag_ici_init)
+			continue;
+		xfs_qm_dqrele_inodes_ag(mp, i, flags);
+	}
 }
 
 /*------------------------------------------------------------------------*/
diff --git a/fs/xfs/support/debug.c b/fs/xfs/support/debug.c
index c27abef7b84..ae548296542 100644
--- a/fs/xfs/support/debug.c
+++ b/fs/xfs/support/debug.c
@@ -18,6 +18,13 @@
 #include <xfs.h>
 #include "debug.h"
 
+/* xfs_mount.h drags a lot of crap in, sorry.. */
+#include "xfs_sb.h"
+#include "xfs_inum.h"
+#include "xfs_ag.h"
+#include "xfs_dmapi.h"
+#include "xfs_mount.h"
+
 static char		message[1024];	/* keep it off the stack */
 static DEFINE_SPINLOCK(xfs_err_lock);
 
@@ -55,22 +62,42 @@ cmn_err(register int level, char *fmt, ...)
 }
 
 void
-icmn_err(register int level, char *fmt, va_list ap)
+xfs_fs_vcmn_err(
+	int			level,
+	struct xfs_mount	*mp,
+	char			*fmt,
+	va_list			ap)
 {
-	ulong	flags;
-	int	len;
+	unsigned long		flags;
+	int			len = 0;
 
 	level &= XFS_ERR_MASK;
-	if(level > XFS_MAX_ERR_LEVEL)
+	if (level > XFS_MAX_ERR_LEVEL)
 		level = XFS_MAX_ERR_LEVEL;
+
 	spin_lock_irqsave(&xfs_err_lock,flags);
-	len = vsnprintf(message, sizeof(message), fmt, ap);
+
+	if (mp) {
+		len = sprintf(message, "Filesystem \"%s\": ", mp->m_fsname);
+
+		/*
+		 * Skip the printk if we can't print anything useful
+		 * due to an over-long device name.
+		 */
+		if (len >= sizeof(message))
+			goto out;
+	}
+
+	len = vsnprintf(message + len, sizeof(message) - len, fmt, ap);
 	if (len >= sizeof(message))
 		len = sizeof(message) - 1;
 	if (message[len-1] == '\n')
 		message[len-1] = 0;
+
 	printk("%s%s\n", err_level[level], message);
+ out:
 	spin_unlock_irqrestore(&xfs_err_lock,flags);
+
 	BUG_ON(level == CE_PANIC);
 }
 
@@ -84,5 +111,5 @@ assfail(char *expr, char *file, int line)
 void
 xfs_hex_dump(void *p, int length)
 {
-	print_hex_dump(KERN_ALERT, "", DUMP_PREFIX_OFFSET, 16, 1, p, length, 1);
+	print_hex_dump(KERN_ALERT, "", DUMP_PREFIX_ADDRESS, 16, 1, p, length, 1);
 }
diff --git a/fs/xfs/support/debug.h b/fs/xfs/support/debug.h
index 75845f95081..6f4fd37c67a 100644
--- a/fs/xfs/support/debug.h
+++ b/fs/xfs/support/debug.h
@@ -27,8 +27,6 @@
 #define CE_ALERT        1               /* alert        */
 #define CE_PANIC        0               /* panic        */
 
-extern void icmn_err(int, char *, va_list)
-	__attribute__ ((format (printf, 2, 0)));
 extern void cmn_err(int, char *, ...)
 	__attribute__ ((format (printf, 2, 3)));
 extern void assfail(char *expr, char *f, int l);
diff --git a/fs/xfs/support/ktrace.c b/fs/xfs/support/ktrace.c
index a34ef05489b..2d494c26717 100644
--- a/fs/xfs/support/ktrace.c
+++ b/fs/xfs/support/ktrace.c
@@ -113,21 +113,16 @@ ktrace_alloc(int nentries, unsigned int __nocast sleep)
 void
 ktrace_free(ktrace_t *ktp)
 {
-	int     entries_size;
-
 	if (ktp == (ktrace_t *)NULL)
 		return;
 
 	/*
 	 * Special treatment for the Vnode trace buffer.
 	 */
-	if (ktp->kt_nentries == ktrace_zentries) {
+	if (ktp->kt_nentries == ktrace_zentries)
 		kmem_zone_free(ktrace_ent_zone, ktp->kt_entries);
-	} else {
-		entries_size = (int)(ktp->kt_nentries * sizeof(ktrace_entry_t));
-
+	else
 		kmem_free(ktp->kt_entries);
-	}
 
 	kmem_zone_free(ktrace_hdr_zone, ktp);
 }
diff --git a/fs/xfs/xfs.h b/fs/xfs/xfs.h
index 540e4c98982..17254b529c5 100644
--- a/fs/xfs/xfs.h
+++ b/fs/xfs/xfs.h
@@ -30,7 +30,7 @@
 #define XFS_ATTR_TRACE 1
 #define XFS_BLI_TRACE 1
 #define XFS_BMAP_TRACE 1
-#define XFS_BMBT_TRACE 1
+#define XFS_BTREE_TRACE 1
 #define XFS_DIR2_TRACE 1
 #define XFS_DQUOT_TRACE 1
 #define XFS_ILOCK_TRACE 1
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
index 91d69338d3b..a8cdd73999a 100644
--- a/fs/xfs/xfs_acl.c
+++ b/fs/xfs/xfs_acl.c
@@ -758,7 +758,7 @@ xfs_acl_setmode(
 	if (gap && nomask)
 		iattr.ia_mode |= gap->ae_perm << 3;
 
-	return xfs_setattr(XFS_I(vp), &iattr, 0, sys_cred);
+	return xfs_setattr(XFS_I(vp), &iattr, 0);
 }
 
 /*
diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h
index 61b292a9fb4..f2e21817a22 100644
--- a/fs/xfs/xfs_ag.h
+++ b/fs/xfs/xfs_ag.h
@@ -91,6 +91,8 @@ typedef struct xfs_agf {
 #define	XFS_AGF_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGF_DADDR(mp))
 #define	XFS_BUF_TO_AGF(bp)	((xfs_agf_t *)XFS_BUF_PTR(bp))
 
+extern int xfs_read_agf(struct xfs_mount *mp, struct xfs_trans *tp,
+			xfs_agnumber_t agno, int flags, struct xfs_buf **bpp);
 
 /*
  * Size of the unlinked inode hash table in the agi.
@@ -142,6 +144,9 @@ typedef struct xfs_agi {
 #define	XFS_AGI_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGI_DADDR(mp))
 #define	XFS_BUF_TO_AGI(bp)	((xfs_agi_t *)XFS_BUF_PTR(bp))
 
+extern int xfs_read_agi(struct xfs_mount *mp, struct xfs_trans *tp,
+				xfs_agnumber_t agno, struct xfs_buf **bpp);
+
 /*
  * The third a.g. block contains the a.g. freelist, an array
  * of block pointers to blocks owned by the allocation btree code.
@@ -192,17 +197,23 @@ typedef struct xfs_perag
 	xfs_agino_t	pagi_freecount;	/* number of free inodes */
 	xfs_agino_t	pagi_count;	/* number of allocated inodes */
 	int		pagb_count;	/* pagb slots in use */
+	xfs_perag_busy_t *pagb_list;	/* unstable blocks */
 #ifdef __KERNEL__
 	spinlock_t	pagb_lock;	/* lock for pagb_list */
-#endif
-	xfs_perag_busy_t *pagb_list;	/* unstable blocks */
+
 	atomic_t        pagf_fstrms;    /* # of filestreams active in this AG */
 
 	int		pag_ici_init;	/* incore inode cache initialised */
 	rwlock_t	pag_ici_lock;	/* incore inode lock */
 	struct radix_tree_root pag_ici_root;	/* incore inode cache root */
+#endif
 } xfs_perag_t;
 
+/*
+ * tags for inode radix tree
+ */
+#define XFS_ICI_RECLAIM_TAG	0	/* inode is to be reclaimed */
+
 #define	XFS_AG_MAXLEVELS(mp)		((mp)->m_ag_maxlevels)
 #define	XFS_MIN_FREELIST_RAW(bl,cl,mp)	\
 	(MIN(bl + 1, XFS_AG_MAXLEVELS(mp)) + MIN(cl + 1, XFS_AG_MAXLEVELS(mp)))
diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c
index 1956f83489f..028e44e58ea 100644
--- a/fs/xfs/xfs_alloc.c
+++ b/fs/xfs/xfs_alloc.c
@@ -90,6 +90,92 @@ STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *,
  */
 
 /*
+ * Lookup the record equal to [bno, len] in the btree given by cur.
+ */
+STATIC int				/* error */
+xfs_alloc_lookup_eq(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
+}
+
+/*
+ * Lookup the first record greater than or equal to [bno, len]
+ * in the btree given by cur.
+ */
+STATIC int				/* error */
+xfs_alloc_lookup_ge(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+}
+
+/*
+ * Lookup the first record less than or equal to [bno, len]
+ * in the btree given by cur.
+ */
+STATIC int				/* error */
+xfs_alloc_lookup_le(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len,	/* length of extent */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.a.ar_startblock = bno;
+	cur->bc_rec.a.ar_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
+}
+
+/*
+ * Update the record referred to by cur to the value given
+ * by [bno, len].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int				/* error */
+xfs_alloc_update(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* starting block of extent */
+	xfs_extlen_t		len)	/* length of extent */
+{
+	union xfs_btree_rec	rec;
+
+	rec.alloc.ar_startblock = cpu_to_be32(bno);
+	rec.alloc.ar_blockcount = cpu_to_be32(len);
+	return xfs_btree_update(cur, &rec);
+}
+
+/*
+ * Get the data from the pointed-to record.
+ */
+STATIC int				/* error */
+xfs_alloc_get_rec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		*bno,	/* output: starting block of extent */
+	xfs_extlen_t		*len,	/* output: length of extent */
+	int			*stat)	/* output: success/failure */
+{
+	union xfs_btree_rec	*rec;
+	int			error;
+
+	error = xfs_btree_get_rec(cur, &rec, stat);
+	if (!error && *stat == 1) {
+		*bno = be32_to_cpu(rec->alloc.ar_startblock);
+		*len = be32_to_cpu(rec->alloc.ar_blockcount);
+	}
+	return error;
+}
+
+/*
  * Compute aligned version of the found extent.
  * Takes alignment and min length into account.
  */
@@ -294,21 +380,20 @@ xfs_alloc_fixup_trees(
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
+
 #ifdef DEBUG
-	{
-		xfs_alloc_block_t	*bnoblock;
-		xfs_alloc_block_t	*cntblock;
-
-		if (bno_cur->bc_nlevels == 1 &&
-		    cnt_cur->bc_nlevels == 1) {
-			bnoblock = XFS_BUF_TO_ALLOC_BLOCK(bno_cur->bc_bufs[0]);
-			cntblock = XFS_BUF_TO_ALLOC_BLOCK(cnt_cur->bc_bufs[0]);
-			XFS_WANT_CORRUPTED_RETURN(
-				be16_to_cpu(bnoblock->bb_numrecs) ==
-				be16_to_cpu(cntblock->bb_numrecs));
-		}
+	if (bno_cur->bc_nlevels == 1 && cnt_cur->bc_nlevels == 1) {
+		struct xfs_btree_block	*bnoblock;
+		struct xfs_btree_block	*cntblock;
+
+		bnoblock = XFS_BUF_TO_BLOCK(bno_cur->bc_bufs[0]);
+		cntblock = XFS_BUF_TO_BLOCK(cnt_cur->bc_bufs[0]);
+
+		XFS_WANT_CORRUPTED_RETURN(
+			bnoblock->bb_numrecs == cntblock->bb_numrecs);
 	}
 #endif
+
 	/*
 	 * Deal with all four cases: the allocated record is contained
 	 * within the freespace record, so we can have new freespace
@@ -333,7 +418,7 @@ xfs_alloc_fixup_trees(
 	/*
 	 * Delete the entry from the by-size btree.
 	 */
-	if ((error = xfs_alloc_delete(cnt_cur, &i)))
+	if ((error = xfs_btree_delete(cnt_cur, &i)))
 		return error;
 	XFS_WANT_CORRUPTED_RETURN(i == 1);
 	/*
@@ -343,7 +428,7 @@ xfs_alloc_fixup_trees(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno1, nflen1, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 0);
-		if ((error = xfs_alloc_insert(cnt_cur, &i)))
+		if ((error = xfs_btree_insert(cnt_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
@@ -351,7 +436,7 @@ xfs_alloc_fixup_trees(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno2, nflen2, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 0);
-		if ((error = xfs_alloc_insert(cnt_cur, &i)))
+		if ((error = xfs_btree_insert(cnt_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
@@ -362,7 +447,7 @@ xfs_alloc_fixup_trees(
 		/*
 		 * No remaining freespace, just delete the by-block tree entry.
 		 */
-		if ((error = xfs_alloc_delete(bno_cur, &i)))
+		if ((error = xfs_btree_delete(bno_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	} else {
@@ -379,7 +464,7 @@ xfs_alloc_fixup_trees(
 		if ((error = xfs_alloc_lookup_eq(bno_cur, nfbno2, nflen2, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 0);
-		if ((error = xfs_alloc_insert(bno_cur, &i)))
+		if ((error = xfs_btree_insert(bno_cur, &i)))
 			return error;
 		XFS_WANT_CORRUPTED_RETURN(i == 1);
 	}
@@ -640,8 +725,8 @@ xfs_alloc_ag_vextent_exact(
 	/*
 	 * Allocate/initialize a cursor for the by-number freespace btree.
 	 */
-	bno_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_BNO, NULL, 0);
+	bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_BNO);
 	/*
 	 * Lookup bno and minlen in the btree (minlen is irrelevant, really).
 	 * Look for the closest free block <= bno, it must contain bno
@@ -696,8 +781,8 @@ xfs_alloc_ag_vextent_exact(
 	 * We are allocating agbno for rlen [agbno .. end]
 	 * Allocate/initialize a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_CNT, NULL, 0);
+	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_CNT);
 	ASSERT(args->agbno + args->len <=
 		be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
 	if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen,
@@ -759,8 +844,8 @@ xfs_alloc_ag_vextent_near(
 	/*
 	 * Get a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_CNT, NULL, 0);
+	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_CNT);
 	ltlen = 0;
 	bno_cur_lt = bno_cur_gt = NULL;
 	/*
@@ -818,7 +903,7 @@ xfs_alloc_ag_vextent_near(
 				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 				if (ltlen >= args->minlen)
 					break;
-				if ((error = xfs_alloc_increment(cnt_cur, 0, &i)))
+				if ((error = xfs_btree_increment(cnt_cur, 0, &i)))
 					goto error0;
 			} while (i);
 			ASSERT(ltlen >= args->minlen);
@@ -828,7 +913,7 @@ xfs_alloc_ag_vextent_near(
 		i = cnt_cur->bc_ptrs[0];
 		for (j = 1, blen = 0, bdiff = 0;
 		     !error && j && (blen < args->maxlen || bdiff > 0);
-		     error = xfs_alloc_increment(cnt_cur, 0, &j)) {
+		     error = xfs_btree_increment(cnt_cur, 0, &j)) {
 			/*
 			 * For each entry, decide if it's better than
 			 * the previous best entry.
@@ -886,8 +971,8 @@ xfs_alloc_ag_vextent_near(
 		/*
 		 * Set up a cursor for the by-bno tree.
 		 */
-		bno_cur_lt = xfs_btree_init_cursor(args->mp, args->tp,
-			args->agbp, args->agno, XFS_BTNUM_BNO, NULL, 0);
+		bno_cur_lt = xfs_allocbt_init_cursor(args->mp, args->tp,
+			args->agbp, args->agno, XFS_BTNUM_BNO);
 		/*
 		 * Fix up the btree entries.
 		 */
@@ -914,8 +999,8 @@ xfs_alloc_ag_vextent_near(
 	/*
 	 * Allocate and initialize the cursor for the leftward search.
 	 */
-	bno_cur_lt = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_BNO, NULL, 0);
+	bno_cur_lt = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_BNO);
 	/*
 	 * Lookup <= bno to find the leftward search's starting point.
 	 */
@@ -938,7 +1023,7 @@ xfs_alloc_ag_vextent_near(
 	 * Increment the cursor, so we will point at the entry just right
 	 * of the leftward entry if any, or to the leftmost entry.
 	 */
-	if ((error = xfs_alloc_increment(bno_cur_gt, 0, &i)))
+	if ((error = xfs_btree_increment(bno_cur_gt, 0, &i)))
 		goto error0;
 	if (!i) {
 		/*
@@ -961,7 +1046,7 @@ xfs_alloc_ag_vextent_near(
 					args->minlen, &ltbnoa, &ltlena);
 			if (ltlena >= args->minlen)
 				break;
-			if ((error = xfs_alloc_decrement(bno_cur_lt, 0, &i)))
+			if ((error = xfs_btree_decrement(bno_cur_lt, 0, &i)))
 				goto error0;
 			if (!i) {
 				xfs_btree_del_cursor(bno_cur_lt,
@@ -977,7 +1062,7 @@ xfs_alloc_ag_vextent_near(
 					args->minlen, &gtbnoa, &gtlena);
 			if (gtlena >= args->minlen)
 				break;
-			if ((error = xfs_alloc_increment(bno_cur_gt, 0, &i)))
+			if ((error = xfs_btree_increment(bno_cur_gt, 0, &i)))
 				goto error0;
 			if (!i) {
 				xfs_btree_del_cursor(bno_cur_gt,
@@ -1066,7 +1151,7 @@ xfs_alloc_ag_vextent_near(
 					/*
 					 * Fell off the right end.
 					 */
-					if ((error = xfs_alloc_increment(
+					if ((error = xfs_btree_increment(
 							bno_cur_gt, 0, &i)))
 						goto error0;
 					if (!i) {
@@ -1162,7 +1247,7 @@ xfs_alloc_ag_vextent_near(
 					/*
 					 * Fell off the left end.
 					 */
-					if ((error = xfs_alloc_decrement(
+					if ((error = xfs_btree_decrement(
 							bno_cur_lt, 0, &i)))
 						goto error0;
 					if (!i) {
@@ -1267,8 +1352,8 @@ xfs_alloc_ag_vextent_size(
 	/*
 	 * Allocate and initialize a cursor for the by-size btree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_CNT, NULL, 0);
+	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_CNT);
 	bno_cur = NULL;
 	/*
 	 * Look for an entry >= maxlen+alignment-1 blocks.
@@ -1321,7 +1406,7 @@ xfs_alloc_ag_vextent_size(
 		bestflen = flen;
 		bestfbno = fbno;
 		for (;;) {
-			if ((error = xfs_alloc_decrement(cnt_cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cnt_cur, 0, &i)))
 				goto error0;
 			if (i == 0)
 				break;
@@ -1372,8 +1457,8 @@ xfs_alloc_ag_vextent_size(
 	/*
 	 * Allocate and initialize a cursor for the by-block tree.
 	 */
-	bno_cur = xfs_btree_init_cursor(args->mp, args->tp, args->agbp,
-		args->agno, XFS_BTNUM_BNO, NULL, 0);
+	bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
+		args->agno, XFS_BTNUM_BNO);
 	if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen,
 			rbno, rlen, XFSA_FIXUP_CNT_OK)))
 		goto error0;
@@ -1416,7 +1501,7 @@ xfs_alloc_ag_vextent_small(
 	xfs_extlen_t	flen;
 	int		i;
 
-	if ((error = xfs_alloc_decrement(ccur, 0, &i)))
+	if ((error = xfs_btree_decrement(ccur, 0, &i)))
 		goto error0;
 	if (i) {
 		if ((error = xfs_alloc_get_rec(ccur, &fbno, &flen, &i)))
@@ -1515,8 +1600,7 @@ xfs_free_ag_extent(
 	/*
 	 * Allocate and initialize a cursor for the by-block btree.
 	 */
-	bno_cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_BNO, NULL,
-		0);
+	bno_cur = xfs_allocbt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_BNO);
 	cnt_cur = NULL;
 	/*
 	 * Look for a neighboring block on the left (lower block numbers)
@@ -1549,7 +1633,7 @@ xfs_free_ag_extent(
 	 * Look for a neighboring block on the right (higher block numbers)
 	 * that is contiguous with this space.
 	 */
-	if ((error = xfs_alloc_increment(bno_cur, 0, &haveright)))
+	if ((error = xfs_btree_increment(bno_cur, 0, &haveright)))
 		goto error0;
 	if (haveright) {
 		/*
@@ -1575,8 +1659,7 @@ xfs_free_ag_extent(
 	/*
 	 * Now allocate and initialize a cursor for the by-size tree.
 	 */
-	cnt_cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_CNT, NULL,
-		0);
+	cnt_cur = xfs_allocbt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_CNT);
 	/*
 	 * Have both left and right contiguous neighbors.
 	 * Merge all three into a single free block.
@@ -1588,7 +1671,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
@@ -1597,19 +1680,19 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
 		 * Delete the old by-block entry for the right block.
 		 */
-		if ((error = xfs_alloc_delete(bno_cur, &i)))
+		if ((error = xfs_btree_delete(bno_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
 		 * Move the by-block cursor back to the left neighbor.
 		 */
-		if ((error = xfs_alloc_decrement(bno_cur, 0, &i)))
+		if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 #ifdef DEBUG
@@ -1648,14 +1731,14 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
 		 * Back up the by-block cursor to the left neighbor, and
 		 * update its length.
 		 */
-		if ((error = xfs_alloc_decrement(bno_cur, 0, &i)))
+		if ((error = xfs_btree_decrement(bno_cur, 0, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		nbno = ltbno;
@@ -1674,7 +1757,7 @@ xfs_free_ag_extent(
 		if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_delete(cnt_cur, &i)))
+		if ((error = xfs_btree_delete(cnt_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 		/*
@@ -1693,7 +1776,7 @@ xfs_free_ag_extent(
 	else {
 		nbno = bno;
 		nlen = len;
-		if ((error = xfs_alloc_insert(bno_cur, &i)))
+		if ((error = xfs_btree_insert(bno_cur, &i)))
 			goto error0;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 	}
@@ -1705,7 +1788,7 @@ xfs_free_ag_extent(
 	if ((error = xfs_alloc_lookup_eq(cnt_cur, nbno, nlen, &i)))
 		goto error0;
 	XFS_WANT_CORRUPTED_GOTO(i == 0, error0);
-	if ((error = xfs_alloc_insert(cnt_cur, &i)))
+	if ((error = xfs_btree_insert(cnt_cur, &i)))
 		goto error0;
 	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 	xfs_btree_del_cursor(cnt_cur, XFS_BTREE_NOERROR);
@@ -2150,51 +2233,83 @@ xfs_alloc_put_freelist(
  * Read in the allocation group header (free/alloc section).
  */
 int					/* error */
-xfs_alloc_read_agf(
-	xfs_mount_t	*mp,		/* mount point structure */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	int		flags,		/* XFS_ALLOC_FLAG_... */
-	xfs_buf_t	**bpp)		/* buffer for the ag freelist header */
+xfs_read_agf(
+	struct xfs_mount	*mp,	/* mount point structure */
+	struct xfs_trans	*tp,	/* transaction pointer */
+	xfs_agnumber_t		agno,	/* allocation group number */
+	int			flags,	/* XFS_BUF_ */
+	struct xfs_buf		**bpp)	/* buffer for the ag freelist header */
 {
-	xfs_agf_t	*agf;		/* ag freelist header */
+	struct xfs_agf	*agf;		/* ag freelist header */
 	int		agf_ok;		/* set if agf is consistent */
-	xfs_buf_t	*bp;		/* return value */
-	xfs_perag_t	*pag;		/* per allocation group data */
 	int		error;
 
 	ASSERT(agno != NULLAGNUMBER);
 	error = xfs_trans_read_buf(
 			mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)),
-			XFS_FSS_TO_BB(mp, 1),
-			(flags & XFS_ALLOC_FLAG_TRYLOCK) ? XFS_BUF_TRYLOCK : 0U,
-			&bp);
+			XFS_FSS_TO_BB(mp, 1), flags, bpp);
 	if (error)
 		return error;
-	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
-	if (!bp) {
-		*bpp = NULL;
+	if (!*bpp)
 		return 0;
-	}
+
+	ASSERT(!XFS_BUF_GETERROR(*bpp));
+	agf = XFS_BUF_TO_AGF(*bpp);
+
 	/*
 	 * Validate the magic number of the agf block.
 	 */
-	agf = XFS_BUF_TO_AGF(bp);
 	agf_ok =
 		be32_to_cpu(agf->agf_magicnum) == XFS_AGF_MAGIC &&
 		XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) &&
 		be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) &&
 		be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) &&
 		be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) &&
-		be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp);
+		be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp) &&
+		be32_to_cpu(agf->agf_seqno) == agno;
+	if (xfs_sb_version_haslazysbcount(&mp->m_sb))
+		agf_ok = agf_ok && be32_to_cpu(agf->agf_btreeblks) <=
+						be32_to_cpu(agf->agf_length);
 	if (unlikely(XFS_TEST_ERROR(!agf_ok, mp, XFS_ERRTAG_ALLOC_READ_AGF,
 			XFS_RANDOM_ALLOC_READ_AGF))) {
 		XFS_CORRUPTION_ERROR("xfs_alloc_read_agf",
 				     XFS_ERRLEVEL_LOW, mp, agf);
-		xfs_trans_brelse(tp, bp);
+		xfs_trans_brelse(tp, *bpp);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
+
+	XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_AGF, XFS_AGF_REF);
+	return 0;
+}
+
+/*
+ * Read in the allocation group header (free/alloc section).
+ */
+int					/* error */
+xfs_alloc_read_agf(
+	struct xfs_mount	*mp,	/* mount point structure */
+	struct xfs_trans	*tp,	/* transaction pointer */
+	xfs_agnumber_t		agno,	/* allocation group number */
+	int			flags,	/* XFS_ALLOC_FLAG_... */
+	struct xfs_buf		**bpp)	/* buffer for the ag freelist header */
+{
+	struct xfs_agf		*agf;		/* ag freelist header */
+	struct xfs_perag	*pag;		/* per allocation group data */
+	int			error;
+
+	ASSERT(agno != NULLAGNUMBER);
+
+	error = xfs_read_agf(mp, tp, agno,
+			(flags & XFS_ALLOC_FLAG_TRYLOCK) ? XFS_BUF_TRYLOCK : 0,
+			bpp);
+	if (error)
+		return error;
+	if (!*bpp)
+		return 0;
+	ASSERT(!XFS_BUF_GETERROR(*bpp));
+
+	agf = XFS_BUF_TO_AGF(*bpp);
 	pag = &mp->m_perag[agno];
 	if (!pag->pagf_init) {
 		pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks);
@@ -2213,6 +2328,7 @@ xfs_alloc_read_agf(
 #ifdef DEBUG
 	else if (!XFS_FORCED_SHUTDOWN(mp)) {
 		ASSERT(pag->pagf_freeblks == be32_to_cpu(agf->agf_freeblks));
+		ASSERT(pag->pagf_btreeblks == be32_to_cpu(agf->agf_btreeblks));
 		ASSERT(pag->pagf_flcount == be32_to_cpu(agf->agf_flcount));
 		ASSERT(pag->pagf_longest == be32_to_cpu(agf->agf_longest));
 		ASSERT(pag->pagf_levels[XFS_BTNUM_BNOi] ==
@@ -2221,8 +2337,6 @@ xfs_alloc_read_agf(
 		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]));
 	}
 #endif
-	XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGF, XFS_AGF_REF);
-	*bpp = bp;
 	return 0;
 }
 
diff --git a/fs/xfs/xfs_alloc.h b/fs/xfs/xfs_alloc.h
index 5aec15d0651..588172796f7 100644
--- a/fs/xfs/xfs_alloc.h
+++ b/fs/xfs/xfs_alloc.h
@@ -121,6 +121,19 @@ extern ktrace_t *xfs_alloc_trace_buf;
 #define	XFS_ALLOC_KTRACE_BUSYSEARCH	6
 #endif
 
+void
+xfs_alloc_mark_busy(xfs_trans_t *tp,
+		xfs_agnumber_t agno,
+		xfs_agblock_t bno,
+		xfs_extlen_t len);
+
+void
+xfs_alloc_clear_busy(xfs_trans_t *tp,
+		xfs_agnumber_t ag,
+		int idx);
+
+#endif	/* __KERNEL__ */
+
 /*
  * Compute and fill in value of m_ag_maxlevels.
  */
@@ -196,18 +209,4 @@ xfs_free_extent(
 	xfs_fsblock_t	bno,	/* starting block number of extent */
 	xfs_extlen_t	len);	/* length of extent */
 
-void
-xfs_alloc_mark_busy(xfs_trans_t *tp,
-		xfs_agnumber_t agno,
-		xfs_agblock_t bno,
-		xfs_extlen_t len);
-
-void
-xfs_alloc_clear_busy(xfs_trans_t *tp,
-		xfs_agnumber_t ag,
-		int idx);
-
-
-#endif	/* __KERNEL__ */
-
 #endif	/* __XFS_ALLOC_H__ */
diff --git a/fs/xfs/xfs_alloc_btree.c b/fs/xfs/xfs_alloc_btree.c
index 3ce2645508a..733cb75a8c5 100644
--- a/fs/xfs/xfs_alloc_btree.c
+++ b/fs/xfs/xfs_alloc_btree.c
@@ -35,2177 +35,464 @@
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_btree.h"
+#include "xfs_btree_trace.h"
 #include "xfs_ialloc.h"
 #include "xfs_alloc.h"
 #include "xfs_error.h"
 
-/*
- * Prototypes for internal functions.
- */
 
-STATIC void xfs_alloc_log_block(xfs_trans_t *, xfs_buf_t *, int);
-STATIC void xfs_alloc_log_keys(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_alloc_log_ptrs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_alloc_log_recs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC int xfs_alloc_lshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_alloc_newroot(xfs_btree_cur_t *, int *);
-STATIC int xfs_alloc_rshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_alloc_split(xfs_btree_cur_t *, int, xfs_agblock_t *,
-		xfs_alloc_key_t *, xfs_btree_cur_t **, int *);
-STATIC int xfs_alloc_updkey(xfs_btree_cur_t *, xfs_alloc_key_t *, int);
+STATIC struct xfs_btree_cur *
+xfs_allocbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.a.agbp, cur->bc_private.a.agno,
+			cur->bc_btnum);
+}
 
-/*
- * Internal functions.
- */
+STATIC void
+xfs_allocbt_set_root(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			inc)
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	int			btnum = cur->bc_btnum;
 
-/*
- * Single level of the xfs_alloc_delete record deletion routine.
- * Delete record pointed to by cur/level.
- * Remove the record from its block then rebalance the tree.
- * Return 0 for error, 1 for done, 2 to go on to the next level.
- */
-STATIC int				/* error */
-xfs_alloc_delrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level removing record from */
-	int			*stat)	/* fail/done/go-on */
+	ASSERT(ptr->s != 0);
+
+	agf->agf_roots[btnum] = ptr->s;
+	be32_add_cpu(&agf->agf_levels[btnum], inc);
+	cur->bc_mp->m_perag[seqno].pagf_levels[btnum] += inc;
+
+	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
+}
+
+STATIC int
+xfs_allocbt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			length,
+	int			*stat)
 {
-	xfs_agf_t		*agf;	/* allocation group freelist header */
-	xfs_alloc_block_t	*block;	/* btree block record/key lives in */
-	xfs_agblock_t		bno;	/* btree block number */
-	xfs_buf_t		*bp;	/* buffer for block */
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_alloc_key_t		key;	/* kp points here if block is level 0 */
-	xfs_agblock_t		lbno;	/* left block's block number */
-	xfs_buf_t		*lbp;	/* left block's buffer pointer */
-	xfs_alloc_block_t	*left;	/* left btree block */
-	xfs_alloc_key_t		*lkp=NULL;	/* left block key pointer */
-	xfs_alloc_ptr_t		*lpp=NULL;	/* left block address pointer */
-	int			lrecs=0;	/* number of records in left block */
-	xfs_alloc_rec_t		*lrp;	/* left block record pointer */
-	xfs_mount_t		*mp;	/* mount structure */
-	int			ptr;	/* index in btree block for this rec */
-	xfs_agblock_t		rbno;	/* right block's block number */
-	xfs_buf_t		*rbp;	/* right block's buffer pointer */
-	xfs_alloc_block_t	*right;	/* right btree block */
-	xfs_alloc_key_t		*rkp;	/* right block key pointer */
-	xfs_alloc_ptr_t		*rpp;	/* right block address pointer */
-	int			rrecs=0;	/* number of records in right block */
-	int			numrecs;
-	xfs_alloc_rec_t		*rrp;	/* right block record pointer */
-	xfs_btree_cur_t		*tcur;	/* temporary btree cursor */
+	int			error;
+	xfs_agblock_t		bno;
 
-	/*
-	 * Get the index of the entry being deleted, check for nothing there.
-	 */
-	ptr = cur->bc_ptrs[level];
-	if (ptr == 0) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Get the buffer & block containing the record or key/ptr.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+
+	/* Allocate the new block from the freelist. If we can't, give up.  */
+	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
+				       &bno, 1);
+	if (error) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
 		return error;
-#endif
-	/*
-	 * Fail if we're off the end of the block.
-	 */
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (ptr > numrecs) {
+	}
+
+	if (bno == NULLAGBLOCK) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 		*stat = 0;
 		return 0;
 	}
-	XFS_STATS_INC(xs_abt_delrec);
-	/*
-	 * It's a nonleaf.  Excise the key and ptr being deleted, by
-	 * sliding the entries past them down one.
-	 * Log the changed areas of the block.
-	 */
-	if (level > 0) {
-		lkp = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-		lpp = XFS_ALLOC_PTR_ADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = ptr; i < numrecs; i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(lpp[i]), level)))
-				return error;
-		}
-#endif
-		if (ptr < numrecs) {
-			memmove(&lkp[ptr - 1], &lkp[ptr],
-				(numrecs - ptr) * sizeof(*lkp));
-			memmove(&lpp[ptr - 1], &lpp[ptr],
-				(numrecs - ptr) * sizeof(*lpp));
-			xfs_alloc_log_ptrs(cur, bp, ptr, numrecs - 1);
-			xfs_alloc_log_keys(cur, bp, ptr, numrecs - 1);
-		}
-	}
-	/*
-	 * It's a leaf.  Excise the record being deleted, by sliding the
-	 * entries past it down one.  Log the changed areas of the block.
-	 */
-	else {
-		lrp = XFS_ALLOC_REC_ADDR(block, 1, cur);
-		if (ptr < numrecs) {
-			memmove(&lrp[ptr - 1], &lrp[ptr],
-				(numrecs - ptr) * sizeof(*lrp));
-			xfs_alloc_log_recs(cur, bp, ptr, numrecs - 1);
-		}
-		/*
-		 * If it's the first record in the block, we'll need a key
-		 * structure to pass up to the next level (updkey).
-		 */
-		if (ptr == 1) {
-			key.ar_startblock = lrp->ar_startblock;
-			key.ar_blockcount = lrp->ar_blockcount;
-			lkp = &key;
-		}
-	}
-	/*
-	 * Decrement and log the number of entries in the block.
-	 */
-	numrecs--;
-	block->bb_numrecs = cpu_to_be16(numrecs);
-	xfs_alloc_log_block(cur->bc_tp, bp, XFS_BB_NUMRECS);
-	/*
-	 * See if the longest free extent in the allocation group was
-	 * changed by this operation.  True if it's the by-size btree, and
-	 * this is the leaf level, and there is no right sibling block,
-	 * and this was the last record.
-	 */
-	agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-	mp = cur->bc_mp;
 
-	if (level == 0 &&
-	    cur->bc_btnum == XFS_BTNUM_CNT &&
-	    be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK &&
-	    ptr > numrecs) {
-		ASSERT(ptr == numrecs + 1);
-		/*
-		 * There are still records in the block.  Grab the size
-		 * from the last one.
-		 */
-		if (numrecs) {
-			rrp = XFS_ALLOC_REC_ADDR(block, numrecs, cur);
-			agf->agf_longest = rrp->ar_blockcount;
-		}
-		/*
-		 * No free extents left.
-		 */
-		else
-			agf->agf_longest = 0;
-		mp->m_perag[be32_to_cpu(agf->agf_seqno)].pagf_longest =
-			be32_to_cpu(agf->agf_longest);
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_LONGEST);
-	}
-	/*
-	 * Is this the root level?  If so, we're almost done.
-	 */
-	if (level == cur->bc_nlevels - 1) {
-		/*
-		 * If this is the root level,
-		 * and there's only one entry left,
-		 * and it's NOT the leaf level,
-		 * then we can get rid of this level.
-		 */
-		if (numrecs == 1 && level > 0) {
-			/*
-			 * lpp is still set to the first pointer in the block.
-			 * Make it the new root of the btree.
-			 */
-			bno = be32_to_cpu(agf->agf_roots[cur->bc_btnum]);
-			agf->agf_roots[cur->bc_btnum] = *lpp;
-			be32_add_cpu(&agf->agf_levels[cur->bc_btnum], -1);
-			mp->m_perag[be32_to_cpu(agf->agf_seqno)].pagf_levels[cur->bc_btnum]--;
-			/*
-			 * Put this buffer/block on the ag's freelist.
-			 */
-			error = xfs_alloc_put_freelist(cur->bc_tp,
-					cur->bc_private.a.agbp, NULL, bno, 1);
-			if (error)
-				return error;
-			/*
-			 * Since blocks move to the free list without the
-			 * coordination used in xfs_bmap_finish, we can't allow
-			 * block to be available for reallocation and
-			 * non-transaction writing (user data) until we know
-			 * that the transaction that moved it to the free list
-			 * is permanently on disk. We track the blocks by
-			 * declaring these blocks as "busy"; the busy list is
-			 * maintained on a per-ag basis and each transaction
-			 * records which entries should be removed when the
-			 * iclog commits to disk. If a busy block is
-			 * allocated, the iclog is pushed up to the LSN
-			 * that freed the block.
-			 */
-			xfs_alloc_mark_busy(cur->bc_tp,
-				be32_to_cpu(agf->agf_seqno), bno, 1);
+	xfs_trans_agbtree_delta(cur->bc_tp, 1);
+	new->s = cpu_to_be32(bno);
 
-			xfs_trans_agbtree_delta(cur->bc_tp, -1);
-			xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-				XFS_AGF_ROOTS | XFS_AGF_LEVELS);
-			/*
-			 * Update the cursor so there's one fewer level.
-			 */
-			xfs_btree_setbuf(cur, level, NULL);
-			cur->bc_nlevels--;
-		} else if (level > 0 &&
-			   (error = xfs_alloc_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * If we deleted the leftmost entry in the block, update the
-	 * key values above us in the tree.
-	 */
-	if (ptr == 1 && (error = xfs_alloc_updkey(cur, lkp, level + 1)))
-		return error;
-	/*
-	 * If the number of records remaining in the block is at least
-	 * the minimum, we're done.
-	 */
-	if (numrecs >= XFS_ALLOC_BLOCK_MINRECS(level, cur)) {
-		if (level > 0 && (error = xfs_alloc_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * Otherwise, we have to move some records around to keep the
-	 * tree balanced.  Look at the left and right sibling blocks to
-	 * see if we can re-balance by moving only one record.
-	 */
-	rbno = be32_to_cpu(block->bb_rightsib);
-	lbno = be32_to_cpu(block->bb_leftsib);
-	bno = NULLAGBLOCK;
-	ASSERT(rbno != NULLAGBLOCK || lbno != NULLAGBLOCK);
-	/*
-	 * Duplicate the cursor so our btree manipulations here won't
-	 * disrupt the next level up.
-	 */
-	if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-		return error;
-	/*
-	 * If there's a right sibling, see if it's ok to shift an entry
-	 * out of it.
-	 */
-	if (rbno != NULLAGBLOCK) {
-		/*
-		 * Move the temp cursor to the last entry in the next block.
-		 * Actually any entry but the first would suffice.
-		 */
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_increment(tcur, level, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		/*
-		 * Grab a pointer to the block.
-		 */
-		rbp = tcur->bc_bufs[level];
-		right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-			goto error0;
-#endif
-		/*
-		 * Grab the current block number, for future use.
-		 */
-		bno = be32_to_cpu(right->bb_leftsib);
-		/*
-		 * If right block is full enough so that removing one entry
-		 * won't make it too empty, and left-shifting an entry out
-		 * of right to us works, we're done.
-		 */
-		if (be16_to_cpu(right->bb_numrecs) - 1 >=
-		     XFS_ALLOC_BLOCK_MINRECS(level, cur)) {
-			if ((error = xfs_alloc_lshift(tcur, level, &i)))
-				goto error0;
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_ALLOC_BLOCK_MINRECS(level, cur));
-				xfs_btree_del_cursor(tcur,
-						     XFS_BTREE_NOERROR);
-				if (level > 0 &&
-				    (error = xfs_alloc_decrement(cur, level,
-					    &i)))
-					return error;
-				*stat = 1;
-				return 0;
-			}
-		}
-		/*
-		 * Otherwise, grab the number of records in right for
-		 * future reference, and fix up the temp cursor to point
-		 * to our block again (last record).
-		 */
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if (lbno != NULLAGBLOCK) {
-			i = xfs_btree_firstrec(tcur, level);
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			if ((error = xfs_alloc_decrement(tcur, level, &i)))
-				goto error0;
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		}
-	}
-	/*
-	 * If there's a left sibling, see if it's ok to shift an entry
-	 * out of it.
-	 */
-	if (lbno != NULLAGBLOCK) {
-		/*
-		 * Move the temp cursor to the first entry in the
-		 * previous block.
-		 */
-		i = xfs_btree_firstrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_alloc_decrement(tcur, level, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		xfs_btree_firstrec(tcur, level);
-		/*
-		 * Grab a pointer to the block.
-		 */
-		lbp = tcur->bc_bufs[level];
-		left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-			goto error0;
-#endif
-		/*
-		 * Grab the current block number, for future use.
-		 */
-		bno = be32_to_cpu(left->bb_rightsib);
-		/*
-		 * If left block is full enough so that removing one entry
-		 * won't make it too empty, and right-shifting an entry out
-		 * of left to us works, we're done.
-		 */
-		if (be16_to_cpu(left->bb_numrecs) - 1 >=
-		     XFS_ALLOC_BLOCK_MINRECS(level, cur)) {
-			if ((error = xfs_alloc_rshift(tcur, level, &i)))
-				goto error0;
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_ALLOC_BLOCK_MINRECS(level, cur));
-				xfs_btree_del_cursor(tcur,
-						     XFS_BTREE_NOERROR);
-				if (level == 0)
-					cur->bc_ptrs[0]++;
-				*stat = 1;
-				return 0;
-			}
-		}
-		/*
-		 * Otherwise, grab the number of records in right for
-		 * future reference.
-		 */
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	}
-	/*
-	 * Delete the temp cursor, we're done with it.
-	 */
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	/*
-	 * If here, we need to do a join to keep the tree balanced.
-	 */
-	ASSERT(bno != NULLAGBLOCK);
-	/*
-	 * See if we can join with the left neighbor block.
-	 */
-	if (lbno != NULLAGBLOCK &&
-	    lrecs + numrecs <= XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * Set "right" to be the starting block,
-		 * "left" to be the left neighbor.
-		 */
-		rbno = bno;
-		right = block;
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		rbp = bp;
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, lbno, 0, &lbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-		lrecs = be16_to_cpu(left->bb_numrecs);
-		if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-			return error;
-	}
-	/*
-	 * If that won't work, see if we can join with the right neighbor block.
-	 */
-	else if (rbno != NULLAGBLOCK &&
-		 rrecs + numrecs <= XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * Set "left" to be the starting block,
-		 * "right" to be the right neighbor.
-		 */
-		lbno = bno;
-		left = block;
-		lrecs = be16_to_cpu(left->bb_numrecs);
-		lbp = bp;
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, rbno, 0, &rbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-			return error;
-	}
-	/*
-	 * Otherwise, we can't fix the imbalance.
-	 * Just return.  This is probably a logic error, but it's not fatal.
-	 */
-	else {
-		if (level > 0 && (error = xfs_alloc_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * We're now going to join "left" and "right" by moving all the stuff
-	 * in "right" to "left" and deleting "right".
-	 */
-	if (level > 0) {
-		/*
-		 * It's a non-leaf.  Move keys and pointers.
-		 */
-		lkp = XFS_ALLOC_KEY_ADDR(left, lrecs + 1, cur);
-		lpp = XFS_ALLOC_PTR_ADDR(left, lrecs + 1, cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < rrecs; i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i]), level)))
-				return error;
-		}
-#endif
-		memcpy(lkp, rkp, rrecs * sizeof(*lkp));
-		memcpy(lpp, rpp, rrecs * sizeof(*lpp));
-		xfs_alloc_log_keys(cur, lbp, lrecs + 1, lrecs + rrecs);
-		xfs_alloc_log_ptrs(cur, lbp, lrecs + 1, lrecs + rrecs);
-	} else {
-		/*
-		 * It's a leaf.  Move records.
-		 */
-		lrp = XFS_ALLOC_REC_ADDR(left, lrecs + 1, cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		memcpy(lrp, rrp, rrecs * sizeof(*lrp));
-		xfs_alloc_log_recs(cur, lbp, lrecs + 1, lrecs + rrecs);
-	}
-	/*
-	 * If we joined with the left neighbor, set the buffer in the
-	 * cursor to the left block, and fix up the index.
-	 */
-	if (bp != lbp) {
-		xfs_btree_setbuf(cur, level, lbp);
-		cur->bc_ptrs[level] += lrecs;
-	}
-	/*
-	 * If we joined with the right neighbor and there's a level above
-	 * us, increment the cursor at that level.
-	 */
-	else if (level + 1 < cur->bc_nlevels &&
-		 (error = xfs_alloc_increment(cur, level + 1, &i)))
-		return error;
-	/*
-	 * Fix up the number of records in the surviving block.
-	 */
-	lrecs += rrecs;
-	left->bb_numrecs = cpu_to_be16(lrecs);
-	/*
-	 * Fix up the right block pointer in the surviving block, and log it.
-	 */
-	left->bb_rightsib = right->bb_rightsib;
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	/*
-	 * If there is a right sibling now, make it point to the
-	 * remaining block.
-	 */
-	if (be32_to_cpu(left->bb_rightsib) != NULLAGBLOCK) {
-		xfs_alloc_block_t	*rrblock;
-		xfs_buf_t		*rrbp;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+}
 
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib), 0,
-				&rrbp, XFS_ALLOC_BTREE_REF)))
-			return error;
-		rrblock = XFS_BUF_TO_ALLOC_BLOCK(rrbp);
-		if ((error = xfs_btree_check_sblock(cur, rrblock, level, rrbp)))
-			return error;
-		rrblock->bb_leftsib = cpu_to_be32(lbno);
-		xfs_alloc_log_block(cur->bc_tp, rrbp, XFS_BB_LEFTSIB);
-	}
-	/*
-	 * Free the deleting block by putting it on the freelist.
-	 */
-	error = xfs_alloc_put_freelist(cur->bc_tp,
-					 cur->bc_private.a.agbp, NULL, rbno, 1);
+STATIC int
+xfs_allocbt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	xfs_agblock_t		bno;
+	int			error;
+
+	bno = XFS_DADDR_TO_AGBNO(cur->bc_mp, XFS_BUF_ADDR(bp));
+	error = xfs_alloc_put_freelist(cur->bc_tp, agbp, NULL, bno, 1);
 	if (error)
 		return error;
+
 	/*
-	 * Since blocks move to the free list without the coordination
-	 * used in xfs_bmap_finish, we can't allow block to be available
-	 * for reallocation and non-transaction writing (user data)
-	 * until we know that the transaction that moved it to the free
-	 * list is permanently on disk. We track the blocks by declaring
-	 * these blocks as "busy"; the busy list is maintained on a
-	 * per-ag basis and each transaction records which entries
-	 * should be removed when the iclog commits to disk. If a
-	 * busy block is allocated, the iclog is pushed up to the
+	 * Since blocks move to the free list without the coordination used in
+	 * xfs_bmap_finish, we can't allow block to be available for
+	 * reallocation and non-transaction writing (user data) until we know
+	 * that the transaction that moved it to the free list is permanently
+	 * on disk. We track the blocks by declaring these blocks as "busy";
+	 * the busy list is maintained on a per-ag basis and each transaction
+	 * records which entries should be removed when the iclog commits to
+	 * disk. If a busy block is allocated, the iclog is pushed up to the
 	 * LSN that freed the block.
 	 */
 	xfs_alloc_mark_busy(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1);
 	xfs_trans_agbtree_delta(cur->bc_tp, -1);
-
-	/*
-	 * Adjust the current level's cursor so that we're left referring
-	 * to the right node, after we're done.
-	 * If this leaves the ptr value 0 our caller will fix it up.
-	 */
-	if (level > 0)
-		cur->bc_ptrs[level]--;
-	/*
-	 * Return value means the next level up has something to do.
-	 */
-	*stat = 2;
 	return 0;
-
-error0:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
 }
 
 /*
- * Insert one record/level.  Return information to the caller
- * allowing the next level up to proceed if necessary.
+ * Update the longest extent in the AGF
  */
-STATIC int				/* error */
-xfs_alloc_insrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to insert record at */
-	xfs_agblock_t		*bnop,	/* i/o: block number inserted */
-	xfs_alloc_rec_t		*recp,	/* i/o: record data inserted */
-	xfs_btree_cur_t		**curp,	/* output: new cursor replacing cur */
-	int			*stat)	/* output: success/failure */
+STATIC void
+xfs_allocbt_update_lastrec(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	union xfs_btree_rec	*rec,
+	int			ptr,
+	int			reason)
 {
-	xfs_agf_t		*agf;	/* allocation group freelist header */
-	xfs_alloc_block_t	*block;	/* btree block record/key lives in */
-	xfs_buf_t		*bp;	/* buffer for block */
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_alloc_key_t		key;	/* key value being inserted */
-	xfs_alloc_key_t		*kp;	/* pointer to btree keys */
-	xfs_agblock_t		nbno;	/* block number of allocated block */
-	xfs_btree_cur_t		*ncur;	/* new cursor to be used at next lvl */
-	xfs_alloc_key_t		nkey;	/* new key value, from split */
-	xfs_alloc_rec_t		nrec;	/* new record value, for caller */
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
+	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	__be32			len;
 	int			numrecs;
-	int			optr;	/* old ptr value */
-	xfs_alloc_ptr_t		*pp;	/* pointer to btree addresses */
-	int			ptr;	/* index in btree block for this rec */
-	xfs_alloc_rec_t		*rp;	/* pointer to btree records */
 
-	ASSERT(be32_to_cpu(recp->ar_blockcount) > 0);
+	ASSERT(cur->bc_btnum == XFS_BTNUM_CNT);
+
+	switch (reason) {
+	case LASTREC_UPDATE:
+		/*
+		 * If this is the last leaf block and it's the last record,
+		 * then update the size of the longest extent in the AG.
+		 */
+		if (ptr != xfs_btree_get_numrecs(block))
+			return;
+		len = rec->alloc.ar_blockcount;
+		break;
+	case LASTREC_INSREC:
+		if (be32_to_cpu(rec->alloc.ar_blockcount) <=
+		    be32_to_cpu(agf->agf_longest))
+			return;
+		len = rec->alloc.ar_blockcount;
+		break;
+	case LASTREC_DELREC:
+		numrecs = xfs_btree_get_numrecs(block);
+		if (ptr <= numrecs)
+			return;
+		ASSERT(ptr == numrecs + 1);
 
-	/*
-	 * GCC doesn't understand the (arguably complex) control flow in
-	 * this function and complains about uninitialized structure fields
-	 * without this.
-	 */
-	memset(&nrec, 0, sizeof(nrec));
+		if (numrecs) {
+			xfs_alloc_rec_t *rrp;
 
-	/*
-	 * If we made it to the root level, allocate a new root block
-	 * and we're done.
-	 */
-	if (level >= cur->bc_nlevels) {
-		XFS_STATS_INC(xs_abt_insrec);
-		if ((error = xfs_alloc_newroot(cur, &i)))
-			return error;
-		*bnop = NULLAGBLOCK;
-		*stat = i;
-		return 0;
-	}
-	/*
-	 * Make a key out of the record data to be inserted, and save it.
-	 */
-	key.ar_startblock = recp->ar_startblock;
-	key.ar_blockcount = recp->ar_blockcount;
-	optr = ptr = cur->bc_ptrs[level];
-	/*
-	 * If we're off the left edge, return failure.
-	 */
-	if (ptr == 0) {
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_abt_insrec);
-	/*
-	 * Get pointers to the btree buffer and block.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-	/*
-	 * Check that the new entry is being inserted in the right place.
-	 */
-	if (ptr <= numrecs) {
-		if (level == 0) {
-			rp = XFS_ALLOC_REC_ADDR(block, ptr, cur);
-			xfs_btree_check_rec(cur->bc_btnum, recp, rp);
+			rrp = XFS_ALLOC_REC_ADDR(cur->bc_mp, block, numrecs);
+			len = rrp->ar_blockcount;
 		} else {
-			kp = XFS_ALLOC_KEY_ADDR(block, ptr, cur);
-			xfs_btree_check_key(cur->bc_btnum, &key, kp);
-		}
-	}
-#endif
-	nbno = NULLAGBLOCK;
-	ncur = NULL;
-	/*
-	 * If the block is full, we can't insert the new entry until we
-	 * make the block un-full.
-	 */
-	if (numrecs == XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * First, try shifting an entry to the right neighbor.
-		 */
-		if ((error = xfs_alloc_rshift(cur, level, &i)))
-			return error;
-		if (i) {
-			/* nothing */
-		}
-		/*
-		 * Next, try shifting an entry to the left neighbor.
-		 */
-		else {
-			if ((error = xfs_alloc_lshift(cur, level, &i)))
-				return error;
-			if (i)
-				optr = ptr = cur->bc_ptrs[level];
-			else {
-				/*
-				 * Next, try splitting the current block in
-				 * half. If this works we have to re-set our
-				 * variables because we could be in a
-				 * different block now.
-				 */
-				if ((error = xfs_alloc_split(cur, level, &nbno,
-						&nkey, &ncur, &i)))
-					return error;
-				if (i) {
-					bp = cur->bc_bufs[level];
-					block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-					if ((error =
-						xfs_btree_check_sblock(cur,
-							block, level, bp)))
-						return error;
-#endif
-					ptr = cur->bc_ptrs[level];
-					nrec.ar_startblock = nkey.ar_startblock;
-					nrec.ar_blockcount = nkey.ar_blockcount;
-				}
-				/*
-				 * Otherwise the insert fails.
-				 */
-				else {
-					*stat = 0;
-					return 0;
-				}
-			}
-		}
-	}
-	/*
-	 * At this point we know there's room for our new entry in the block
-	 * we're pointing at.
-	 */
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (level > 0) {
-		/*
-		 * It's a non-leaf entry.  Make a hole for the new data
-		 * in the key and ptr regions of the block.
-		 */
-		kp = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-		pp = XFS_ALLOC_PTR_ADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = numrecs; i >= ptr; i--) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(pp[i - 1]), level)))
-				return error;
+			len = 0;
 		}
-#endif
-		memmove(&kp[ptr], &kp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*kp));
-		memmove(&pp[ptr], &pp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*pp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, *bnop, level)))
-			return error;
-#endif
-		/*
-		 * Now stuff the new data in, bump numrecs and log the new data.
-		 */
-		kp[ptr - 1] = key;
-		pp[ptr - 1] = cpu_to_be32(*bnop);
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_alloc_log_keys(cur, bp, ptr, numrecs);
-		xfs_alloc_log_ptrs(cur, bp, ptr, numrecs);
-#ifdef DEBUG
-		if (ptr < numrecs)
-			xfs_btree_check_key(cur->bc_btnum, kp + ptr - 1,
-				kp + ptr);
-#endif
-	} else {
-		/*
-		 * It's a leaf entry.  Make a hole for the new record.
-		 */
-		rp = XFS_ALLOC_REC_ADDR(block, 1, cur);
-		memmove(&rp[ptr], &rp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*rp));
-		/*
-		 * Now stuff the new record in, bump numrecs
-		 * and log the new data.
-		 */
-		rp[ptr - 1] = *recp;
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_alloc_log_recs(cur, bp, ptr, numrecs);
-#ifdef DEBUG
-		if (ptr < numrecs)
-			xfs_btree_check_rec(cur->bc_btnum, rp + ptr - 1,
-				rp + ptr);
-#endif
-	}
-	/*
-	 * Log the new number of records in the btree header.
-	 */
-	xfs_alloc_log_block(cur->bc_tp, bp, XFS_BB_NUMRECS);
-	/*
-	 * If we inserted at the start of a block, update the parents' keys.
-	 */
-	if (optr == 1 && (error = xfs_alloc_updkey(cur, &key, level + 1)))
-		return error;
-	/*
-	 * Look to see if the longest extent in the allocation group
-	 * needs to be updated.
-	 */
 
-	agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-	if (level == 0 &&
-	    cur->bc_btnum == XFS_BTNUM_CNT &&
-	    be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK &&
-	    be32_to_cpu(recp->ar_blockcount) > be32_to_cpu(agf->agf_longest)) {
-		/*
-		 * If this is a leaf in the by-size btree and there
-		 * is no right sibling block and this block is bigger
-		 * than the previous longest block, update it.
-		 */
-		agf->agf_longest = recp->ar_blockcount;
-		cur->bc_mp->m_perag[be32_to_cpu(agf->agf_seqno)].pagf_longest
-			= be32_to_cpu(recp->ar_blockcount);
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_LONGEST);
+		break;
+	default:
+		ASSERT(0);
+		return;
 	}
-	/*
-	 * Return the new block number, if any.
-	 * If there is one, give back a record value and a cursor too.
-	 */
-	*bnop = nbno;
-	if (nbno != NULLAGBLOCK) {
-		*recp = nrec;
-		*curp = ncur;
-	}
-	*stat = 1;
-	return 0;
+
+	agf->agf_longest = len;
+	cur->bc_mp->m_perag[seqno].pagf_longest = be32_to_cpu(len);
+	xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp, XFS_AGF_LONGEST);
 }
 
-/*
- * Log header fields from a btree block.
- */
-STATIC void
-xfs_alloc_log_block(
-	xfs_trans_t		*tp,	/* transaction pointer */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			fields)	/* mask of fields: XFS_BB_... */
+STATIC int
+xfs_allocbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
 {
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	static const short	offsets[] = {	/* table of offsets */
-		offsetof(xfs_alloc_block_t, bb_magic),
-		offsetof(xfs_alloc_block_t, bb_level),
-		offsetof(xfs_alloc_block_t, bb_numrecs),
-		offsetof(xfs_alloc_block_t, bb_leftsib),
-		offsetof(xfs_alloc_block_t, bb_rightsib),
-		sizeof(xfs_alloc_block_t)
-	};
+	return cur->bc_mp->m_alloc_mnr[level != 0];
+}
 
-	xfs_btree_offsets(fields, offsets, XFS_BB_NUM_BITS, &first, &last);
-	xfs_trans_log_buf(tp, bp, first, last);
+STATIC int
+xfs_allocbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	return cur->bc_mp->m_alloc_mxr[level != 0];
 }
 
-/*
- * Log keys from a btree block (nonleaf).
- */
 STATIC void
-xfs_alloc_log_keys(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			kfirst,	/* index of first key to log */
-	int			klast)	/* index of last key to log */
+xfs_allocbt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
 {
-	xfs_alloc_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	xfs_alloc_key_t		*kp;	/* key pointer in btree block */
-	int			last;	/* last byte offset logged */
+	ASSERT(rec->alloc.ar_startblock != 0);
 
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	kp = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&kp[kfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&kp[klast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	key->alloc.ar_startblock = rec->alloc.ar_startblock;
+	key->alloc.ar_blockcount = rec->alloc.ar_blockcount;
 }
 
-/*
- * Log block pointer fields from a btree block (nonleaf).
- */
 STATIC void
-xfs_alloc_log_ptrs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			pfirst,	/* index of first pointer to log */
-	int			plast)	/* index of last pointer to log */
+xfs_allocbt_init_rec_from_key(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
 {
-	xfs_alloc_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	xfs_alloc_ptr_t		*pp;	/* block-pointer pointer in btree blk */
+	ASSERT(key->alloc.ar_startblock != 0);
 
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	pp = XFS_ALLOC_PTR_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&pp[pfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&pp[plast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	rec->alloc.ar_startblock = key->alloc.ar_startblock;
+	rec->alloc.ar_blockcount = key->alloc.ar_blockcount;
 }
 
-/*
- * Log records from a btree block (leaf).
- */
 STATIC void
-xfs_alloc_log_recs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			rfirst,	/* index of first record to log */
-	int			rlast)	/* index of last record to log */
+xfs_allocbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
 {
-	xfs_alloc_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	xfs_alloc_rec_t		*rp;	/* record pointer for btree block */
-
+	ASSERT(cur->bc_rec.a.ar_startblock != 0);
 
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	rp = XFS_ALLOC_REC_ADDR(block, 1, cur);
-#ifdef DEBUG
-	{
-		xfs_agf_t	*agf;
-		xfs_alloc_rec_t	*p;
-
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		for (p = &rp[rfirst - 1]; p <= &rp[rlast - 1]; p++)
-			ASSERT(be32_to_cpu(p->ar_startblock) +
-			       be32_to_cpu(p->ar_blockcount) <=
-			       be32_to_cpu(agf->agf_length));
-	}
-#endif
-	first = (int)((xfs_caddr_t)&rp[rfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&rp[rlast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	rec->alloc.ar_startblock = cpu_to_be32(cur->bc_rec.a.ar_startblock);
+	rec->alloc.ar_blockcount = cpu_to_be32(cur->bc_rec.a.ar_blockcount);
 }
 
-/*
- * Lookup the record.  The cursor is made to point to it, based on dir.
- * Return 0 if can't find any such record, 1 for success.
- */
-STATIC int				/* error */
-xfs_alloc_lookup(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_lookup_t		dir,	/* <=, ==, or >= */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_allocbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
 {
-	xfs_agblock_t		agbno;	/* a.g. relative btree block number */
-	xfs_agnumber_t		agno;	/* allocation group number */
-	xfs_alloc_block_t	*block=NULL;	/* current btree block */
-	int			diff;	/* difference for the current key */
-	int			error;	/* error return value */
-	int			keyno=0;	/* current key number */
-	int			level;	/* level in the btree */
-	xfs_mount_t		*mp;	/* file system mount point */
-
-	XFS_STATS_INC(xs_abt_lookup);
-	/*
-	 * Get the allocation group header, and the root block number.
-	 */
-	mp = cur->bc_mp;
-
-	{
-		xfs_agf_t	*agf;	/* a.g. freespace header */
-
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		agno = be32_to_cpu(agf->agf_seqno);
-		agbno = be32_to_cpu(agf->agf_roots[cur->bc_btnum]);
-	}
-	/*
-	 * Iterate over each level in the btree, starting at the root.
-	 * For each level above the leaves, find the key we need, based
-	 * on the lookup record, then follow the corresponding block
-	 * pointer down to the next level.
-	 */
-	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
-		xfs_buf_t	*bp;	/* buffer pointer for btree block */
-		xfs_daddr_t	d;	/* disk address of btree block */
-
-		/*
-		 * Get the disk address we're looking for.
-		 */
-		d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-		/*
-		 * If the old buffer at this level is for a different block,
-		 * throw it away, otherwise just use it.
-		 */
-		bp = cur->bc_bufs[level];
-		if (bp && XFS_BUF_ADDR(bp) != d)
-			bp = NULL;
-		if (!bp) {
-			/*
-			 * Need to get a new buffer.  Read it, then
-			 * set it in the cursor, releasing the old one.
-			 */
-			if ((error = xfs_btree_read_bufs(mp, cur->bc_tp, agno,
-					agbno, 0, &bp, XFS_ALLOC_BTREE_REF)))
-				return error;
-			xfs_btree_setbuf(cur, level, bp);
-			/*
-			 * Point to the btree block, now that we have the buffer
-			 */
-			block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-			if ((error = xfs_btree_check_sblock(cur, block, level,
-					bp)))
-				return error;
-		} else
-			block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-		/*
-		 * If we already had a key match at a higher level, we know
-		 * we need to use the first entry in this block.
-		 */
-		if (diff == 0)
-			keyno = 1;
-		/*
-		 * Otherwise we need to search this block.  Do a binary search.
-		 */
-		else {
-			int		high;	/* high entry number */
-			xfs_alloc_key_t	*kkbase=NULL;/* base of keys in block */
-			xfs_alloc_rec_t	*krbase=NULL;/* base of records in block */
-			int		low;	/* low entry number */
-
-			/*
-			 * Get a pointer to keys or records.
-			 */
-			if (level > 0)
-				kkbase = XFS_ALLOC_KEY_ADDR(block, 1, cur);
-			else
-				krbase = XFS_ALLOC_REC_ADDR(block, 1, cur);
-			/*
-			 * Set low and high entry numbers, 1-based.
-			 */
-			low = 1;
-			if (!(high = be16_to_cpu(block->bb_numrecs))) {
-				/*
-				 * If the block is empty, the tree must
-				 * be an empty leaf.
-				 */
-				ASSERT(level == 0 && cur->bc_nlevels == 1);
-				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
-				*stat = 0;
-				return 0;
-			}
-			/*
-			 * Binary search the block.
-			 */
-			while (low <= high) {
-				xfs_extlen_t	blockcount;	/* key value */
-				xfs_agblock_t	startblock;	/* key value */
-
-				XFS_STATS_INC(xs_abt_compare);
-				/*
-				 * keyno is average of low and high.
-				 */
-				keyno = (low + high) >> 1;
-				/*
-				 * Get startblock & blockcount.
-				 */
-				if (level > 0) {
-					xfs_alloc_key_t	*kkp;
-
-					kkp = kkbase + keyno - 1;
-					startblock = be32_to_cpu(kkp->ar_startblock);
-					blockcount = be32_to_cpu(kkp->ar_blockcount);
-				} else {
-					xfs_alloc_rec_t	*krp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
 
-					krp = krbase + keyno - 1;
-					startblock = be32_to_cpu(krp->ar_startblock);
-					blockcount = be32_to_cpu(krp->ar_blockcount);
-				}
-				/*
-				 * Compute difference to get next direction.
-				 */
-				if (cur->bc_btnum == XFS_BTNUM_BNO)
-					diff = (int)startblock -
-					       (int)cur->bc_rec.a.ar_startblock;
-				else if (!(diff = (int)blockcount -
-					    (int)cur->bc_rec.a.ar_blockcount))
-					diff = (int)startblock -
-					    (int)cur->bc_rec.a.ar_startblock;
-				/*
-				 * Less than, move right.
-				 */
-				if (diff < 0)
-					low = keyno + 1;
-				/*
-				 * Greater than, move left.
-				 */
-				else if (diff > 0)
-					high = keyno - 1;
-				/*
-				 * Equal, we're done.
-				 */
-				else
-					break;
-			}
-		}
-		/*
-		 * If there are more levels, set up for the next level
-		 * by getting the block number and filling in the cursor.
-		 */
-		if (level > 0) {
-			/*
-			 * If we moved left, need the previous key number,
-			 * unless there isn't one.
-			 */
-			if (diff > 0 && --keyno < 1)
-				keyno = 1;
-			agbno = be32_to_cpu(*XFS_ALLOC_PTR_ADDR(block, keyno, cur));
-#ifdef DEBUG
-			if ((error = xfs_btree_check_sptr(cur, agbno, level)))
-				return error;
-#endif
-			cur->bc_ptrs[level] = keyno;
-		}
-	}
-	/*
-	 * Done with the search.
-	 * See if we need to adjust the results.
-	 */
-	if (dir != XFS_LOOKUP_LE && diff < 0) {
-		keyno++;
-		/*
-		 * If ge search and we went off the end of the block, but it's
-		 * not the last block, we're in the wrong block.
-		 */
-		if (dir == XFS_LOOKUP_GE &&
-		    keyno > be16_to_cpu(block->bb_numrecs) &&
-		    be32_to_cpu(block->bb_rightsib) != NULLAGBLOCK) {
-			int	i;
+	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
+	ASSERT(agf->agf_roots[cur->bc_btnum] != 0);
 
-			cur->bc_ptrs[0] = keyno;
-			if ((error = xfs_alloc_increment(cur, 0, &i)))
-				return error;
-			XFS_WANT_CORRUPTED_RETURN(i == 1);
-			*stat = 1;
-			return 0;
-		}
-	}
-	else if (dir == XFS_LOOKUP_LE && diff > 0)
-		keyno--;
-	cur->bc_ptrs[0] = keyno;
-	/*
-	 * Return if we succeeded or not.
-	 */
-	if (keyno == 0 || keyno > be16_to_cpu(block->bb_numrecs))
-		*stat = 0;
-	else
-		*stat = ((dir != XFS_LOOKUP_EQ) || (diff == 0));
-	return 0;
+	ptr->s = agf->agf_roots[cur->bc_btnum];
 }
 
-/*
- * Move 1 record left from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int				/* error */
-xfs_alloc_lshift(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to shift record on */
-	int			*stat)	/* success/failure */
+STATIC __int64_t
+xfs_allocbt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
 {
-	int			error;	/* error return value */
-#ifdef DEBUG
-	int			i;	/* loop index */
-#endif
-	xfs_alloc_key_t		key;	/* key value for leaf level upward */
-	xfs_buf_t		*lbp;	/* buffer for left neighbor block */
-	xfs_alloc_block_t	*left;	/* left neighbor btree block */
-	int			nrec;	/* new number of left block entries */
-	xfs_buf_t		*rbp;	/* buffer for right (current) block */
-	xfs_alloc_block_t	*right;	/* right (current) btree block */
-	xfs_alloc_key_t		*rkp=NULL;	/* key pointer for right block */
-	xfs_alloc_ptr_t		*rpp=NULL;	/* address pointer for right block */
-	xfs_alloc_rec_t		*rrp=NULL;	/* record pointer for right block */
+	xfs_alloc_rec_incore_t	*rec = &cur->bc_rec.a;
+	xfs_alloc_key_t		*kp = &key->alloc;
+	__int64_t		diff;
 
-	/*
-	 * Set up variables for this block as "right".
-	 */
-	rbp = cur->bc_bufs[level];
-	right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-		return error;
-#endif
-	/*
-	 * If we've got no left sibling then we can't shift an entry left.
-	 */
-	if (be32_to_cpu(right->bb_leftsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * If the cursor entry is the one that would be moved, don't
-	 * do it... it's too complicated.
-	 */
-	if (cur->bc_ptrs[level] <= 1) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Set up the left neighbor as "left".
-	 */
-	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agno, be32_to_cpu(right->bb_leftsib),
-			0, &lbp, XFS_ALLOC_BTREE_REF)))
-		return error;
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-	/*
-	 * If it's full, it can't take another entry.
-	 */
-	if (be16_to_cpu(left->bb_numrecs) == XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		*stat = 0;
-		return 0;
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		return (__int64_t)be32_to_cpu(kp->ar_startblock) -
+				rec->ar_startblock;
 	}
-	nrec = be16_to_cpu(left->bb_numrecs) + 1;
-	/*
-	 * If non-leaf, copy a key and a ptr to the left block.
-	 */
-	if (level > 0) {
-		xfs_alloc_key_t	*lkp;	/* key pointer for left block */
-		xfs_alloc_ptr_t	*lpp;	/* address pointer for left block */
 
-		lkp = XFS_ALLOC_KEY_ADDR(left, nrec, cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		*lkp = *rkp;
-		xfs_alloc_log_keys(cur, lbp, nrec, nrec);
-		lpp = XFS_ALLOC_PTR_ADDR(left, nrec, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(*rpp), level)))
-			return error;
-#endif
-		*lpp = *rpp;
-		xfs_alloc_log_ptrs(cur, lbp, nrec, nrec);
-		xfs_btree_check_key(cur->bc_btnum, lkp - 1, lkp);
-	}
-	/*
-	 * If leaf, copy a record to the left block.
-	 */
-	else {
-		xfs_alloc_rec_t	*lrp;	/* record pointer for left block */
+	diff = (__int64_t)be32_to_cpu(kp->ar_blockcount) - rec->ar_blockcount;
+	if (diff)
+		return diff;
 
-		lrp = XFS_ALLOC_REC_ADDR(left, nrec, cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		*lrp = *rrp;
-		xfs_alloc_log_recs(cur, lbp, nrec, nrec);
-		xfs_btree_check_rec(cur->bc_btnum, lrp - 1, lrp);
-	}
-	/*
-	 * Bump and log left's numrecs, decrement and log right's numrecs.
-	 */
-	be16_add_cpu(&left->bb_numrecs, 1);
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, -1);
-	xfs_alloc_log_block(cur->bc_tp, rbp, XFS_BB_NUMRECS);
-	/*
-	 * Slide the contents of right down one entry.
-	 */
-	if (level > 0) {
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i + 1]),
-					level)))
-				return error;
-		}
-#endif
-		memmove(rkp, rkp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp, rpp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_alloc_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_alloc_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-	} else {
-		memmove(rrp, rrp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_alloc_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		key.ar_startblock = rrp->ar_startblock;
-		key.ar_blockcount = rrp->ar_blockcount;
-		rkp = &key;
-	}
-	/*
-	 * Update the parent key values of right.
-	 */
-	if ((error = xfs_alloc_updkey(cur, rkp, level + 1)))
-		return error;
-	/*
-	 * Slide the cursor value left one.
-	 */
-	cur->bc_ptrs[level]--;
-	*stat = 1;
-	return 0;
+	return (__int64_t)be32_to_cpu(kp->ar_startblock) - rec->ar_startblock;
 }
 
-/*
- * Allocate a new root block, fill it in.
- */
-STATIC int				/* error */
-xfs_alloc_newroot(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			*stat)	/* success/failure */
+STATIC int
+xfs_allocbt_kill_root(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			level,
+	union xfs_btree_ptr	*newroot)
 {
-	int			error;	/* error return value */
-	xfs_agblock_t		lbno;	/* left block number */
-	xfs_buf_t		*lbp;	/* left btree buffer */
-	xfs_alloc_block_t	*left;	/* left btree block */
-	xfs_mount_t		*mp;	/* mount structure */
-	xfs_agblock_t		nbno;	/* new block number */
-	xfs_buf_t		*nbp;	/* new (root) buffer */
-	xfs_alloc_block_t	*new;	/* new (root) btree block */
-	int			nptr;	/* new value for key index, 1 or 2 */
-	xfs_agblock_t		rbno;	/* right block number */
-	xfs_buf_t		*rbp;	/* right btree buffer */
-	xfs_alloc_block_t	*right;	/* right btree block */
-
-	mp = cur->bc_mp;
+	int			error;
 
-	ASSERT(cur->bc_nlevels < XFS_AG_MAXLEVELS(mp));
-	/*
-	 * Get a buffer from the freelist blocks, for the new root.
-	 */
-	error = xfs_alloc_get_freelist(cur->bc_tp,
-					cur->bc_private.a.agbp, &nbno, 1);
-	if (error)
-		return error;
-	/*
-	 * None available, we fail.
-	 */
-	if (nbno == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	xfs_trans_agbtree_delta(cur->bc_tp, 1);
-	nbp = xfs_btree_get_bufs(mp, cur->bc_tp, cur->bc_private.a.agno, nbno,
-		0);
-	new = XFS_BUF_TO_ALLOC_BLOCK(nbp);
-	/*
-	 * Set the root data in the a.g. freespace structure.
-	 */
-	{
-		xfs_agf_t	*agf;	/* a.g. freespace header */
-		xfs_agnumber_t	seqno;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, killroot);
 
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		agf->agf_roots[cur->bc_btnum] = cpu_to_be32(nbno);
-		be32_add_cpu(&agf->agf_levels[cur->bc_btnum], 1);
-		seqno = be32_to_cpu(agf->agf_seqno);
-		mp->m_perag[seqno].pagf_levels[cur->bc_btnum]++;
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_ROOTS | XFS_AGF_LEVELS);
-	}
 	/*
-	 * At the previous root level there are now two blocks: the old
-	 * root, and the new block generated when it was split.
-	 * We don't know which one the cursor is pointing at, so we
-	 * set up variables "left" and "right" for each case.
+	 * Update the root pointer, decreasing the level by 1 and then
+	 * free the old root.
 	 */
-	lbp = cur->bc_bufs[cur->bc_nlevels - 1];
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, cur->bc_nlevels - 1, lbp)))
+	xfs_allocbt_set_root(cur, newroot, -1);
+	error = xfs_allocbt_free_block(cur, bp);
+	if (error) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
 		return error;
-#endif
-	if (be32_to_cpu(left->bb_rightsib) != NULLAGBLOCK) {
-		/*
-		 * Our block is left, pick up the right block.
-		 */
-		lbno = XFS_DADDR_TO_AGBNO(mp, XFS_BUF_ADDR(lbp));
-		rbno = be32_to_cpu(left->bb_rightsib);
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, rbno, 0, &rbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-		if ((error = xfs_btree_check_sblock(cur, right,
-				cur->bc_nlevels - 1, rbp)))
-			return error;
-		nptr = 1;
-	} else {
-		/*
-		 * Our block is right, pick up the left block.
-		 */
-		rbp = lbp;
-		right = left;
-		rbno = XFS_DADDR_TO_AGBNO(mp, XFS_BUF_ADDR(rbp));
-		lbno = be32_to_cpu(right->bb_leftsib);
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, lbno, 0, &lbp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-		if ((error = xfs_btree_check_sblock(cur, left,
-				cur->bc_nlevels - 1, lbp)))
-			return error;
-		nptr = 2;
 	}
-	/*
-	 * Fill in the new block's btree header and log it.
-	 */
-	new->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
-	new->bb_level = cpu_to_be16(cur->bc_nlevels);
-	new->bb_numrecs = cpu_to_be16(2);
-	new->bb_leftsib = cpu_to_be32(NULLAGBLOCK);
-	new->bb_rightsib = cpu_to_be32(NULLAGBLOCK);
-	xfs_alloc_log_block(cur->bc_tp, nbp, XFS_BB_ALL_BITS);
-	ASSERT(lbno != NULLAGBLOCK && rbno != NULLAGBLOCK);
-	/*
-	 * Fill in the key data in the new root.
-	 */
-	{
-		xfs_alloc_key_t		*kp;	/* btree key pointer */
 
-		kp = XFS_ALLOC_KEY_ADDR(new, 1, cur);
-		if (be16_to_cpu(left->bb_level) > 0) {
-			kp[0] = *XFS_ALLOC_KEY_ADDR(left, 1, cur);
-			kp[1] = *XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		} else {
-			xfs_alloc_rec_t	*rp;	/* btree record pointer */
+	XFS_BTREE_STATS_INC(cur, free);
 
-			rp = XFS_ALLOC_REC_ADDR(left, 1, cur);
-			kp[0].ar_startblock = rp->ar_startblock;
-			kp[0].ar_blockcount = rp->ar_blockcount;
-			rp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-			kp[1].ar_startblock = rp->ar_startblock;
-			kp[1].ar_blockcount = rp->ar_blockcount;
-		}
-	}
-	xfs_alloc_log_keys(cur, nbp, 1, 2);
-	/*
-	 * Fill in the pointer data in the new root.
-	 */
-	{
-		xfs_alloc_ptr_t		*pp;	/* btree address pointer */
+	xfs_btree_setbuf(cur, level, NULL);
+	cur->bc_nlevels--;
 
-		pp = XFS_ALLOC_PTR_ADDR(new, 1, cur);
-		pp[0] = cpu_to_be32(lbno);
-		pp[1] = cpu_to_be32(rbno);
-	}
-	xfs_alloc_log_ptrs(cur, nbp, 1, 2);
-	/*
-	 * Fix up the cursor.
-	 */
-	xfs_btree_setbuf(cur, cur->bc_nlevels, nbp);
-	cur->bc_ptrs[cur->bc_nlevels] = nptr;
-	cur->bc_nlevels++;
-	*stat = 1;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 	return 0;
 }
 
-/*
- * Move 1 record right from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int				/* error */
-xfs_alloc_rshift(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to shift record on */
-	int			*stat)	/* success/failure */
-{
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_alloc_key_t		key;	/* key value for leaf level upward */
-	xfs_buf_t		*lbp;	/* buffer for left (current) block */
-	xfs_alloc_block_t	*left;	/* left (current) btree block */
-	xfs_buf_t		*rbp;	/* buffer for right neighbor block */
-	xfs_alloc_block_t	*right;	/* right neighbor btree block */
-	xfs_alloc_key_t		*rkp;	/* key pointer for right block */
-	xfs_btree_cur_t		*tcur;	/* temporary cursor */
-
-	/*
-	 * Set up variables for this block as "left".
-	 */
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-#endif
-	/*
-	 * If we've got no right sibling then we can't shift an entry right.
-	 */
-	if (be32_to_cpu(left->bb_rightsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * If the cursor entry is the one that would be moved, don't
-	 * do it... it's too complicated.
-	 */
-	if (cur->bc_ptrs[level] >= be16_to_cpu(left->bb_numrecs)) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Set up the right neighbor as "right".
-	 */
-	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib),
-			0, &rbp, XFS_ALLOC_BTREE_REF)))
-		return error;
-	right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-	if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-		return error;
-	/*
-	 * If it's full, it can't take another entry.
-	 */
-	if (be16_to_cpu(right->bb_numrecs) == XFS_ALLOC_BLOCK_MAXRECS(level, cur)) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Make a hole at the start of the right neighbor block, then
-	 * copy the last left block entry to the hole.
-	 */
-	if (level > 0) {
-		xfs_alloc_key_t	*lkp;	/* key pointer for left block */
-		xfs_alloc_ptr_t	*lpp;	/* address pointer for left block */
-		xfs_alloc_ptr_t	*rpp;	/* address pointer for right block */
-
-		lkp = XFS_ALLOC_KEY_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		lpp = XFS_ALLOC_PTR_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
 #ifdef DEBUG
-		for (i = be16_to_cpu(right->bb_numrecs) - 1; i >= 0; i--) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i]), level)))
-				return error;
-		}
-#endif
-		memmove(rkp + 1, rkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp + 1, rpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(*lpp), level)))
-			return error;
-#endif
-		*rkp = *lkp;
-		*rpp = *lpp;
-		xfs_alloc_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_alloc_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_btree_check_key(cur->bc_btnum, rkp, rkp + 1);
+STATIC int
+xfs_allocbt_keys_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		return be32_to_cpu(k1->alloc.ar_startblock) <
+		       be32_to_cpu(k2->alloc.ar_startblock);
 	} else {
-		xfs_alloc_rec_t	*lrp;	/* record pointer for left block */
-		xfs_alloc_rec_t	*rrp;	/* record pointer for right block */
-
-		lrp = XFS_ALLOC_REC_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		memmove(rrp + 1, rrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		*rrp = *lrp;
-		xfs_alloc_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		key.ar_startblock = rrp->ar_startblock;
-		key.ar_blockcount = rrp->ar_blockcount;
-		rkp = &key;
-		xfs_btree_check_rec(cur->bc_btnum, rrp, rrp + 1);
+		return be32_to_cpu(k1->alloc.ar_blockcount) <
+			be32_to_cpu(k2->alloc.ar_blockcount) ||
+			(k1->alloc.ar_blockcount == k2->alloc.ar_blockcount &&
+			 be32_to_cpu(k1->alloc.ar_startblock) <
+			 be32_to_cpu(k2->alloc.ar_startblock));
 	}
-	/*
-	 * Decrement and log left's numrecs, bump and log right's numrecs.
-	 */
-	be16_add_cpu(&left->bb_numrecs, -1);
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, 1);
-	xfs_alloc_log_block(cur->bc_tp, rbp, XFS_BB_NUMRECS);
-	/*
-	 * Using a temporary cursor, update the parent key values of the
-	 * block on the right.
-	 */
-	if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-		return error;
-	i = xfs_btree_lastrec(tcur, level);
-	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_alloc_increment(tcur, level, &i)) ||
-	    (error = xfs_alloc_updkey(tcur, rkp, level + 1)))
-		goto error0;
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	*stat = 1;
-	return 0;
-error0:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
 }
 
-/*
- * Split cur/level block in half.
- * Return new block number and its first record (to be inserted into parent).
- */
-STATIC int				/* error */
-xfs_alloc_split(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to split */
-	xfs_agblock_t		*bnop,	/* output: block number allocated */
-	xfs_alloc_key_t		*keyp,	/* output: first key of new block */
-	xfs_btree_cur_t		**curp,	/* output: new cursor */
-	int			*stat)	/* success/failure */
+STATIC int
+xfs_allocbt_recs_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*r1,
+	union xfs_btree_rec	*r2)
 {
-	int			error;	/* error return value */
-	int			i;	/* loop index/record number */
-	xfs_agblock_t		lbno;	/* left (current) block number */
-	xfs_buf_t		*lbp;	/* buffer for left block */
-	xfs_alloc_block_t	*left;	/* left (current) btree block */
-	xfs_agblock_t		rbno;	/* right (new) block number */
-	xfs_buf_t		*rbp;	/* buffer for right block */
-	xfs_alloc_block_t	*right;	/* right (new) btree block */
-
-	/*
-	 * Allocate the new block from the freelist.
-	 * If we can't do it, we're toast.  Give up.
-	 */
-	error = xfs_alloc_get_freelist(cur->bc_tp,
-					 cur->bc_private.a.agbp, &rbno, 1);
-	if (error)
-		return error;
-	if (rbno == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	xfs_trans_agbtree_delta(cur->bc_tp, 1);
-	rbp = xfs_btree_get_bufs(cur->bc_mp, cur->bc_tp, cur->bc_private.a.agno,
-		rbno, 0);
-	/*
-	 * Set up the new block as "right".
-	 */
-	right = XFS_BUF_TO_ALLOC_BLOCK(rbp);
-	/*
-	 * "Left" is the current (according to the cursor) block.
-	 */
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_ALLOC_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-#endif
-	/*
-	 * Fill in the btree header for the new block.
-	 */
-	right->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
-	right->bb_level = left->bb_level;
-	right->bb_numrecs = cpu_to_be16(be16_to_cpu(left->bb_numrecs) / 2);
-	/*
-	 * Make sure that if there's an odd number of entries now, that
-	 * each new block will have the same number of entries.
-	 */
-	if ((be16_to_cpu(left->bb_numrecs) & 1) &&
-	    cur->bc_ptrs[level] <= be16_to_cpu(right->bb_numrecs) + 1)
-		be16_add_cpu(&right->bb_numrecs, 1);
-	i = be16_to_cpu(left->bb_numrecs) - be16_to_cpu(right->bb_numrecs) + 1;
-	/*
-	 * For non-leaf blocks, copy keys and addresses over to the new block.
-	 */
-	if (level > 0) {
-		xfs_alloc_key_t	*lkp;	/* left btree key pointer */
-		xfs_alloc_ptr_t	*lpp;	/* left btree address pointer */
-		xfs_alloc_key_t	*rkp;	/* right btree key pointer */
-		xfs_alloc_ptr_t	*rpp;	/* right btree address pointer */
-
-		lkp = XFS_ALLOC_KEY_ADDR(left, i, cur);
-		lpp = XFS_ALLOC_PTR_ADDR(left, i, cur);
-		rkp = XFS_ALLOC_KEY_ADDR(right, 1, cur);
-		rpp = XFS_ALLOC_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(lpp[i]), level)))
-				return error;
-		}
-#endif
-		memcpy(rkp, lkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memcpy(rpp, lpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_alloc_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_alloc_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*keyp = *rkp;
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		return be32_to_cpu(r1->alloc.ar_startblock) +
+			be32_to_cpu(r1->alloc.ar_blockcount) <=
+			be32_to_cpu(r2->alloc.ar_startblock);
+	} else {
+		return be32_to_cpu(r1->alloc.ar_blockcount) <
+			be32_to_cpu(r2->alloc.ar_blockcount) ||
+			(r1->alloc.ar_blockcount == r2->alloc.ar_blockcount &&
+			 be32_to_cpu(r1->alloc.ar_startblock) <
+			 be32_to_cpu(r2->alloc.ar_startblock));
 	}
-	/*
-	 * For leaf blocks, copy records over to the new block.
-	 */
-	else {
-		xfs_alloc_rec_t	*lrp;	/* left btree record pointer */
-		xfs_alloc_rec_t	*rrp;	/* right btree record pointer */
+}
+#endif	/* DEBUG */
 
-		lrp = XFS_ALLOC_REC_ADDR(left, i, cur);
-		rrp = XFS_ALLOC_REC_ADDR(right, 1, cur);
-		memcpy(rrp, lrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_alloc_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		keyp->ar_startblock = rrp->ar_startblock;
-		keyp->ar_blockcount = rrp->ar_blockcount;
-	}
-	/*
-	 * Find the left block number by looking in the buffer.
-	 * Adjust numrecs, sibling pointers.
-	 */
-	lbno = XFS_DADDR_TO_AGBNO(cur->bc_mp, XFS_BUF_ADDR(lbp));
-	be16_add_cpu(&left->bb_numrecs, -(be16_to_cpu(right->bb_numrecs)));
-	right->bb_rightsib = left->bb_rightsib;
-	left->bb_rightsib = cpu_to_be32(rbno);
-	right->bb_leftsib = cpu_to_be32(lbno);
-	xfs_alloc_log_block(cur->bc_tp, rbp, XFS_BB_ALL_BITS);
-	xfs_alloc_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	/*
-	 * If there's a block to the new block's right, make that block
-	 * point back to right instead of to left.
-	 */
-	if (be32_to_cpu(right->bb_rightsib) != NULLAGBLOCK) {
-		xfs_alloc_block_t	*rrblock;	/* rr btree block */
-		xfs_buf_t		*rrbp;		/* buffer for rrblock */
+#ifdef XFS_BTREE_TRACE
+ktrace_t	*xfs_allocbt_trace_buf;
 
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, be32_to_cpu(right->bb_rightsib), 0,
-				&rrbp, XFS_ALLOC_BTREE_REF)))
-			return error;
-		rrblock = XFS_BUF_TO_ALLOC_BLOCK(rrbp);
-		if ((error = xfs_btree_check_sblock(cur, rrblock, level, rrbp)))
-			return error;
-		rrblock->bb_leftsib = cpu_to_be32(rbno);
-		xfs_alloc_log_block(cur->bc_tp, rrbp, XFS_BB_LEFTSIB);
-	}
-	/*
-	 * If the cursor is really in the right block, move it there.
-	 * If it's just pointing past the last entry in left, then we'll
-	 * insert there, so don't change anything in that case.
-	 */
-	if (cur->bc_ptrs[level] > be16_to_cpu(left->bb_numrecs) + 1) {
-		xfs_btree_setbuf(cur, level, rbp);
-		cur->bc_ptrs[level] -= be16_to_cpu(left->bb_numrecs);
-	}
-	/*
-	 * If there are more levels, we'll need another cursor which refers to
-	 * the right block, no matter where this cursor was.
-	 */
-	if (level + 1 < cur->bc_nlevels) {
-		if ((error = xfs_btree_dup_cursor(cur, curp)))
-			return error;
-		(*curp)->bc_ptrs[level + 1]++;
-	}
-	*bnop = rbno;
-	*stat = 1;
-	return 0;
+STATIC void
+xfs_allocbt_trace_enter(
+	struct xfs_btree_cur	*cur,
+	const char		*func,
+	char			*s,
+	int			type,
+	int			line,
+	__psunsigned_t		a0,
+	__psunsigned_t		a1,
+	__psunsigned_t		a2,
+	__psunsigned_t		a3,
+	__psunsigned_t		a4,
+	__psunsigned_t		a5,
+	__psunsigned_t		a6,
+	__psunsigned_t		a7,
+	__psunsigned_t		a8,
+	__psunsigned_t		a9,
+	__psunsigned_t		a10)
+{
+	ktrace_enter(xfs_allocbt_trace_buf, (void *)(__psint_t)type,
+		(void *)func, (void *)s, NULL, (void *)cur,
+		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
+		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
+		(void *)a8, (void *)a9, (void *)a10);
 }
 
-/*
- * Update keys at all levels from here to the root along the cursor's path.
- */
-STATIC int				/* error */
-xfs_alloc_updkey(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_alloc_key_t		*keyp,	/* new key value to update to */
-	int			level)	/* starting level for update */
+STATIC void
+xfs_allocbt_trace_cursor(
+	struct xfs_btree_cur	*cur,
+	__uint32_t		*s0,
+	__uint64_t		*l0,
+	__uint64_t		*l1)
 {
-	int			ptr;	/* index of key in block */
-
-	/*
-	 * Go up the tree from this level toward the root.
-	 * At each level, update the key value to the value input.
-	 * Stop when we reach a level where the cursor isn't pointing
-	 * at the first entry in the block.
-	 */
-	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
-		xfs_alloc_block_t	*block;	/* btree block */
-		xfs_buf_t		*bp;	/* buffer for block */
-#ifdef DEBUG
-		int			error;	/* error return value */
-#endif
-		xfs_alloc_key_t		*kp;	/* ptr to btree block keys */
-
-		bp = cur->bc_bufs[level];
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-			return error;
-#endif
-		ptr = cur->bc_ptrs[level];
-		kp = XFS_ALLOC_KEY_ADDR(block, ptr, cur);
-		*kp = *keyp;
-		xfs_alloc_log_keys(cur, bp, ptr, ptr);
-	}
-	return 0;
+	*s0 = cur->bc_private.a.agno;
+	*l0 = cur->bc_rec.a.ar_startblock;
+	*l1 = cur->bc_rec.a.ar_blockcount;
 }
 
-/*
- * Externally visible routines.
- */
-
-/*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int					/* error */
-xfs_alloc_decrement(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree, 0 is leaf */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_allocbt_trace_key(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key,
+	__uint64_t		*l0,
+	__uint64_t		*l1)
 {
-	xfs_alloc_block_t	*block;	/* btree block */
-	int			error;	/* error return value */
-	int			lev;	/* btree level */
-
-	ASSERT(level < cur->bc_nlevels);
-	/*
-	 * Read-ahead to the left at this level.
-	 */
-	xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
-	/*
-	 * Decrement the ptr at this level.  If we're still in the block
-	 * then we're done.
-	 */
-	if (--cur->bc_ptrs[level] > 0) {
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * Get a pointer to the btree block.
-	 */
-	block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[level]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level,
-			cur->bc_bufs[level])))
-		return error;
-#endif
-	/*
-	 * If we just went off the left edge of the tree, return failure.
-	 */
-	if (be32_to_cpu(block->bb_leftsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * March up the tree decrementing pointers.
-	 * Stop when we don't go off the left edge of a block.
-	 */
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		if (--cur->bc_ptrs[lev] > 0)
-			break;
-		/*
-		 * Read-ahead the left block, we're going to read it
-		 * in the next loop.
-		 */
-		xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
-	}
-	/*
-	 * If we went off the root then we are seriously confused.
-	 */
-	ASSERT(lev < cur->bc_nlevels);
-	/*
-	 * Now walk back down the tree, fixing up the cursor's buffer
-	 * pointers and key numbers.
-	 */
-	for (block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[lev]); lev > level; ) {
-		xfs_agblock_t	agbno;	/* block number of btree block */
-		xfs_buf_t	*bp;	/* buffer pointer for block */
-
-		agbno = be32_to_cpu(*XFS_ALLOC_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, agbno, 0, &bp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-		cur->bc_ptrs[lev] = be16_to_cpu(block->bb_numrecs);
-	}
-	*stat = 1;
-	return 0;
+	*l0 = be32_to_cpu(key->alloc.ar_startblock);
+	*l1 = be32_to_cpu(key->alloc.ar_blockcount);
 }
 
-/*
- * Delete the record pointed to by cur.
- * The cursor refers to the place where the record was (could be inserted)
- * when the operation returns.
- */
-int					/* error */
-xfs_alloc_delete(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	int		*stat)		/* success/failure */
+STATIC void
+xfs_allocbt_trace_record(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec,
+	__uint64_t		*l0,
+	__uint64_t		*l1,
+	__uint64_t		*l2)
 {
-	int		error;		/* error return value */
-	int		i;		/* result code */
-	int		level;		/* btree level */
-
-	/*
-	 * Go up the tree, starting at leaf level.
-	 * If 2 is returned then a join was done; go to the next level.
-	 * Otherwise we are done.
-	 */
-	for (level = 0, i = 2; i == 2; level++) {
-		if ((error = xfs_alloc_delrec(cur, level, &i)))
-			return error;
-	}
-	if (i == 0) {
-		for (level = 1; level < cur->bc_nlevels; level++) {
-			if (cur->bc_ptrs[level] == 0) {
-				if ((error = xfs_alloc_decrement(cur, level, &i)))
-					return error;
-				break;
-			}
-		}
-	}
-	*stat = i;
-	return 0;
+	*l0 = be32_to_cpu(rec->alloc.ar_startblock);
+	*l1 = be32_to_cpu(rec->alloc.ar_blockcount);
+	*l2 = 0;
 }
+#endif /* XFS_BTREE_TRACE */
+
+static const struct xfs_btree_ops xfs_allocbt_ops = {
+	.rec_len		= sizeof(xfs_alloc_rec_t),
+	.key_len		= sizeof(xfs_alloc_key_t),
+
+	.dup_cursor		= xfs_allocbt_dup_cursor,
+	.set_root		= xfs_allocbt_set_root,
+	.kill_root		= xfs_allocbt_kill_root,
+	.alloc_block		= xfs_allocbt_alloc_block,
+	.free_block		= xfs_allocbt_free_block,
+	.update_lastrec		= xfs_allocbt_update_lastrec,
+	.get_minrecs		= xfs_allocbt_get_minrecs,
+	.get_maxrecs		= xfs_allocbt_get_maxrecs,
+	.init_key_from_rec	= xfs_allocbt_init_key_from_rec,
+	.init_rec_from_key	= xfs_allocbt_init_rec_from_key,
+	.init_rec_from_cur	= xfs_allocbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_allocbt_init_ptr_from_cur,
+	.key_diff		= xfs_allocbt_key_diff,
 
-/*
- * Get the data from the pointed-to record.
- */
-int					/* error */
-xfs_alloc_get_rec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agblock_t		*bno,	/* output: starting block of extent */
-	xfs_extlen_t		*len,	/* output: length of extent */
-	int			*stat)	/* output: success/failure */
-{
-	xfs_alloc_block_t	*block;	/* btree block */
 #ifdef DEBUG
-	int			error;	/* error return value */
+	.keys_inorder		= xfs_allocbt_keys_inorder,
+	.recs_inorder		= xfs_allocbt_recs_inorder,
 #endif
-	int			ptr;	/* record number */
 
-	ptr = cur->bc_ptrs[0];
-	block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[0]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, 0, cur->bc_bufs[0])))
-		return error;
+#ifdef XFS_BTREE_TRACE
+	.trace_enter		= xfs_allocbt_trace_enter,
+	.trace_cursor		= xfs_allocbt_trace_cursor,
+	.trace_key		= xfs_allocbt_trace_key,
+	.trace_record		= xfs_allocbt_trace_record,
 #endif
-	/*
-	 * Off the right end or left end, return failure.
-	 */
-	if (ptr > be16_to_cpu(block->bb_numrecs) || ptr <= 0) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Point to the record and extract its data.
-	 */
-	{
-		xfs_alloc_rec_t		*rec;	/* record data */
-
-		rec = XFS_ALLOC_REC_ADDR(block, ptr, cur);
-		*bno = be32_to_cpu(rec->ar_startblock);
-		*len = be32_to_cpu(rec->ar_blockcount);
-	}
-	*stat = 1;
-	return 0;
-}
+};
 
 /*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
+ * Allocate a new allocation btree cursor.
  */
-int					/* error */
-xfs_alloc_increment(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree, 0 is leaf */
-	int			*stat)	/* success/failure */
+struct xfs_btree_cur *			/* new alloc btree cursor */
+xfs_allocbt_init_cursor(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_buf		*agbp,		/* buffer for agf structure */
+	xfs_agnumber_t		agno,		/* allocation group number */
+	xfs_btnum_t		btnum)		/* btree identifier */
 {
-	xfs_alloc_block_t	*block;	/* btree block */
-	xfs_buf_t		*bp;	/* tree block buffer */
-	int			error;	/* error return value */
-	int			lev;	/* btree level */
-
-	ASSERT(level < cur->bc_nlevels);
-	/*
-	 * Read-ahead to the right at this level.
-	 */
-	xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
-	/*
-	 * Get a pointer to the btree block.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-#endif
-	/*
-	 * Increment the ptr at this level.  If we're still in the block
-	 * then we're done.
-	 */
-	if (++cur->bc_ptrs[level] <= be16_to_cpu(block->bb_numrecs)) {
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * If we just went off the right edge of the tree, return failure.
-	 */
-	if (be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * March up the tree incrementing pointers.
-	 * Stop when we don't go off the right edge of a block.
-	 */
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		bp = cur->bc_bufs[lev];
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-#endif
-		if (++cur->bc_ptrs[lev] <= be16_to_cpu(block->bb_numrecs))
-			break;
-		/*
-		 * Read-ahead the right block, we're going to read it
-		 * in the next loop.
-		 */
-		xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
-	}
-	/*
-	 * If we went off the root then we are seriously confused.
-	 */
-	ASSERT(lev < cur->bc_nlevels);
-	/*
-	 * Now walk back down the tree, fixing up the cursor's buffer
-	 * pointers and key numbers.
-	 */
-	for (bp = cur->bc_bufs[lev], block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-	     lev > level; ) {
-		xfs_agblock_t	agbno;	/* block number of btree block */
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_btree_cur	*cur;
 
-		agbno = be32_to_cpu(*XFS_ALLOC_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, agbno, 0, &bp,
-				XFS_ALLOC_BTREE_REF)))
-			return error;
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_ALLOC_BLOCK(bp);
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-		cur->bc_ptrs[lev] = 1;
-	}
-	*stat = 1;
-	return 0;
-}
+	ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
 
-/*
- * Insert the current record at the point referenced by cur.
- * The cursor may be inconsistent on return if splits have been done.
- */
-int					/* error */
-xfs_alloc_insert(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	int		*stat)		/* success/failure */
-{
-	int		error;		/* error return value */
-	int		i;		/* result value, 0 for failure */
-	int		level;		/* current level number in btree */
-	xfs_agblock_t	nbno;		/* new block number (split result) */
-	xfs_btree_cur_t	*ncur;		/* new cursor (split result) */
-	xfs_alloc_rec_t	nrec;		/* record being inserted this level */
-	xfs_btree_cur_t	*pcur;		/* previous level's cursor */
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
 
-	level = 0;
-	nbno = NULLAGBLOCK;
-	nrec.ar_startblock = cpu_to_be32(cur->bc_rec.a.ar_startblock);
-	nrec.ar_blockcount = cpu_to_be32(cur->bc_rec.a.ar_blockcount);
-	ncur = NULL;
-	pcur = cur;
-	/*
-	 * Loop going up the tree, starting at the leaf level.
-	 * Stop when we don't get a split block, that must mean that
-	 * the insert is finished with this level.
-	 */
-	do {
-		/*
-		 * Insert nrec/nbno into this level of the tree.
-		 * Note if we fail, nbno will be null.
-		 */
-		if ((error = xfs_alloc_insrec(pcur, level++, &nbno, &nrec, &ncur,
-				&i))) {
-			if (pcur != cur)
-				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
-			return error;
-		}
-		/*
-		 * See if the cursor we just used is trash.
-		 * Can't trash the caller's cursor, but otherwise we should
-		 * if ncur is a new cursor or we're about to be done.
-		 */
-		if (pcur != cur && (ncur || nbno == NULLAGBLOCK)) {
-			cur->bc_nlevels = pcur->bc_nlevels;
-			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
-		}
-		/*
-		 * If we got a new cursor, switch to it.
-		 */
-		if (ncur) {
-			pcur = ncur;
-			ncur = NULL;
-		}
-	} while (nbno != NULLAGBLOCK);
-	*stat = i;
-	return 0;
-}
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_nlevels = be32_to_cpu(agf->agf_levels[btnum]);
+	cur->bc_btnum = btnum;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 
-/*
- * Lookup the record equal to [bno, len] in the btree given by cur.
- */
-int					/* error */
-xfs_alloc_lookup_eq(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	bno,		/* starting block of extent */
-	xfs_extlen_t	len,		/* length of extent */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	return xfs_alloc_lookup(cur, XFS_LOOKUP_EQ, stat);
-}
+	cur->bc_ops = &xfs_allocbt_ops;
+	if (btnum == XFS_BTNUM_CNT)
+		cur->bc_flags = XFS_BTREE_LASTREC_UPDATE;
 
-/*
- * Lookup the first record greater than or equal to [bno, len]
- * in the btree given by cur.
- */
-int					/* error */
-xfs_alloc_lookup_ge(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	bno,		/* starting block of extent */
-	xfs_extlen_t	len,		/* length of extent */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	return xfs_alloc_lookup(cur, XFS_LOOKUP_GE, stat);
-}
+	cur->bc_private.a.agbp = agbp;
+	cur->bc_private.a.agno = agno;
 
-/*
- * Lookup the first record less than or equal to [bno, len]
- * in the btree given by cur.
- */
-int					/* error */
-xfs_alloc_lookup_le(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	bno,		/* starting block of extent */
-	xfs_extlen_t	len,		/* length of extent */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.a.ar_startblock = bno;
-	cur->bc_rec.a.ar_blockcount = len;
-	return xfs_alloc_lookup(cur, XFS_LOOKUP_LE, stat);
+	return cur;
 }
 
 /*
- * Update the record referred to by cur, to the value given by [bno, len].
- * This either works (return 0) or gets an EFSCORRUPTED error.
+ * Calculate number of records in an alloc btree block.
  */
-int					/* error */
-xfs_alloc_update(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agblock_t		bno,	/* starting block of extent */
-	xfs_extlen_t		len)	/* length of extent */
+int
+xfs_allocbt_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	int			leaf)
 {
-	xfs_alloc_block_t	*block;	/* btree block to update */
-	int			error;	/* error return value */
-	int			ptr;	/* current record number (updating) */
+	blocklen -= XFS_ALLOC_BLOCK_LEN(mp);
 
-	ASSERT(len > 0);
-	/*
-	 * Pick up the a.g. freelist struct and the current block.
-	 */
-	block = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[0]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, 0, cur->bc_bufs[0])))
-		return error;
-#endif
-	/*
-	 * Get the address of the rec to be updated.
-	 */
-	ptr = cur->bc_ptrs[0];
-	{
-		xfs_alloc_rec_t		*rp;	/* pointer to updated record */
-
-		rp = XFS_ALLOC_REC_ADDR(block, ptr, cur);
-		/*
-		 * Fill in the new contents and log them.
-		 */
-		rp->ar_startblock = cpu_to_be32(bno);
-		rp->ar_blockcount = cpu_to_be32(len);
-		xfs_alloc_log_recs(cur, cur->bc_bufs[0], ptr, ptr);
-	}
-	/*
-	 * If it's the by-size btree and it's the last leaf block and
-	 * it's the last record... then update the size of the longest
-	 * extent in the a.g., which we cache in the a.g. freelist header.
-	 */
-	if (cur->bc_btnum == XFS_BTNUM_CNT &&
-	    be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK &&
-	    ptr == be16_to_cpu(block->bb_numrecs)) {
-		xfs_agf_t	*agf;	/* a.g. freespace header */
-		xfs_agnumber_t	seqno;
-
-		agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-		seqno = be32_to_cpu(agf->agf_seqno);
-		cur->bc_mp->m_perag[seqno].pagf_longest = len;
-		agf->agf_longest = cpu_to_be32(len);
-		xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp,
-			XFS_AGF_LONGEST);
-	}
-	/*
-	 * Updating first record in leaf. Pass new key value up to our parent.
-	 */
-	if (ptr == 1) {
-		xfs_alloc_key_t	key;	/* key containing [bno, len] */
-
-		key.ar_startblock = cpu_to_be32(bno);
-		key.ar_blockcount = cpu_to_be32(len);
-		if ((error = xfs_alloc_updkey(cur, &key, 1)))
-			return error;
-	}
-	return 0;
+	if (leaf)
+		return blocklen / sizeof(xfs_alloc_rec_t);
+	return blocklen / (sizeof(xfs_alloc_key_t) + sizeof(xfs_alloc_ptr_t));
 }
diff --git a/fs/xfs/xfs_alloc_btree.h b/fs/xfs/xfs_alloc_btree.h
index 5bd1a2c8bd0..a6caa0022c9 100644
--- a/fs/xfs/xfs_alloc_btree.h
+++ b/fs/xfs/xfs_alloc_btree.h
@@ -24,7 +24,6 @@
 
 struct xfs_buf;
 struct xfs_btree_cur;
-struct xfs_btree_sblock;
 struct xfs_mount;
 
 /*
@@ -50,16 +49,6 @@ typedef struct xfs_alloc_rec_incore {
 
 /* btree pointer type */
 typedef __be32 xfs_alloc_ptr_t;
-/* btree block header type */
-typedef	struct xfs_btree_sblock xfs_alloc_block_t;
-
-#define	XFS_BUF_TO_ALLOC_BLOCK(bp)	((xfs_alloc_block_t *)XFS_BUF_PTR(bp))
-
-/*
- * Real block structures have a size equal to the disk block size.
- */
-#define	XFS_ALLOC_BLOCK_MAXRECS(lev,cur) ((cur)->bc_mp->m_alloc_mxr[lev != 0])
-#define	XFS_ALLOC_BLOCK_MINRECS(lev,cur) ((cur)->bc_mp->m_alloc_mnr[lev != 0])
 
 /*
  * Minimum and maximum blocksize and sectorsize.
@@ -83,73 +72,39 @@ typedef	struct xfs_btree_sblock xfs_alloc_block_t;
 #define	XFS_CNT_BLOCK(mp)	((xfs_agblock_t)(XFS_BNO_BLOCK(mp) + 1))
 
 /*
- * Record, key, and pointer address macros for btree blocks.
- */
-#define	XFS_ALLOC_REC_ADDR(bb,i,cur)	\
-	XFS_BTREE_REC_ADDR(xfs_alloc, bb, i)
-
-#define	XFS_ALLOC_KEY_ADDR(bb,i,cur)	\
-	XFS_BTREE_KEY_ADDR(xfs_alloc, bb, i)
-
-#define	XFS_ALLOC_PTR_ADDR(bb,i,cur)	\
-	XFS_BTREE_PTR_ADDR(xfs_alloc, bb, i, XFS_ALLOC_BLOCK_MAXRECS(1, cur))
-
-/*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-extern int xfs_alloc_decrement(struct xfs_btree_cur *cur, int level, int *stat);
-
-/*
- * Delete the record pointed to by cur.
- * The cursor refers to the place where the record was (could be inserted)
- * when the operation returns.
- */
-extern int xfs_alloc_delete(struct xfs_btree_cur *cur, int *stat);
-
-/*
- * Get the data from the pointed-to record.
- */
-extern int xfs_alloc_get_rec(struct xfs_btree_cur *cur,	xfs_agblock_t *bno,
-				xfs_extlen_t *len, int *stat);
-
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-extern int xfs_alloc_increment(struct xfs_btree_cur *cur, int level, int *stat);
-
-/*
- * Insert the current record at the point referenced by cur.
- * The cursor may be inconsistent on return if splits have been done.
- */
-extern int xfs_alloc_insert(struct xfs_btree_cur *cur, int *stat);
-
-/*
- * Lookup the record equal to [bno, len] in the btree given by cur.
- */
-extern int xfs_alloc_lookup_eq(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len, int *stat);
-
-/*
- * Lookup the first record greater than or equal to [bno, len]
- * in the btree given by cur.
- */
-extern int xfs_alloc_lookup_ge(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len, int *stat);
-
-/*
- * Lookup the first record less than or equal to [bno, len]
- * in the btree given by cur.
+ * Btree block header size depends on a superblock flag.
+ *
+ * (not quite yet, but soon)
  */
-extern int xfs_alloc_lookup_le(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len, int *stat);
+#define XFS_ALLOC_BLOCK_LEN(mp)	XFS_BTREE_SBLOCK_LEN
 
 /*
- * Update the record referred to by cur, to the value given by [bno, len].
- * This either works (return 0) or gets an EFSCORRUPTED error.
- */
-extern int xfs_alloc_update(struct xfs_btree_cur *cur, xfs_agblock_t bno,
-				xfs_extlen_t len);
+ * Record, key, and pointer address macros for btree blocks.
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+#define XFS_ALLOC_REC_ADDR(mp, block, index) \
+	((xfs_alloc_rec_t *) \
+		((char *)(block) + \
+		 XFS_ALLOC_BLOCK_LEN(mp) + \
+		 (((index) - 1) * sizeof(xfs_alloc_rec_t))))
+
+#define XFS_ALLOC_KEY_ADDR(mp, block, index) \
+	((xfs_alloc_key_t *) \
+		((char *)(block) + \
+		 XFS_ALLOC_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_alloc_key_t)))
+
+#define XFS_ALLOC_PTR_ADDR(mp, block, index, maxrecs) \
+	((xfs_alloc_ptr_t *) \
+		((char *)(block) + \
+		 XFS_ALLOC_BLOCK_LEN(mp) + \
+		 (maxrecs) * sizeof(xfs_alloc_key_t) + \
+		 ((index) - 1) * sizeof(xfs_alloc_ptr_t)))
+
+extern struct xfs_btree_cur *xfs_allocbt_init_cursor(struct xfs_mount *,
+		struct xfs_trans *, struct xfs_buf *,
+		xfs_agnumber_t, xfs_btnum_t);
+extern int xfs_allocbt_maxrecs(struct xfs_mount *, int, int);
 
 #endif	/* __XFS_ALLOC_BTREE_H__ */
diff --git a/fs/xfs/xfs_arch.h b/fs/xfs/xfs_arch.h
index 0b3b5efe848..53d5e70d136 100644
--- a/fs/xfs/xfs_arch.h
+++ b/fs/xfs/xfs_arch.h
@@ -41,21 +41,36 @@
 #endif
 
 #ifdef XFS_NATIVE_HOST
-#define cpu_to_be16(val)	((__be16)(val))
-#define cpu_to_be32(val)	((__be32)(val))
-#define cpu_to_be64(val)	((__be64)(val))
-#define be16_to_cpu(val)	((__uint16_t)(val))
-#define be32_to_cpu(val)	((__uint32_t)(val))
-#define be64_to_cpu(val)	((__uint64_t)(val))
+#define cpu_to_be16(val)	((__force __be16)(__u16)(val))
+#define cpu_to_be32(val)	((__force __be32)(__u32)(val))
+#define cpu_to_be64(val)	((__force __be64)(__u64)(val))
+#define be16_to_cpu(val)	((__force __u16)(__be16)(val))
+#define be32_to_cpu(val)	((__force __u32)(__be32)(val))
+#define be64_to_cpu(val)	((__force __u64)(__be64)(val))
 #else
-#define cpu_to_be16(val)	(__swab16((__uint16_t)(val)))
-#define cpu_to_be32(val)	(__swab32((__uint32_t)(val)))
-#define cpu_to_be64(val)	(__swab64((__uint64_t)(val)))
-#define be16_to_cpu(val)	(__swab16((__be16)(val)))
-#define be32_to_cpu(val)	(__swab32((__be32)(val)))
-#define be64_to_cpu(val)	(__swab64((__be64)(val)))
+#define cpu_to_be16(val)	((__force __be16)__swab16((__u16)(val)))
+#define cpu_to_be32(val)	((__force __be32)__swab32((__u32)(val)))
+#define cpu_to_be64(val)	((__force __be64)__swab64((__u64)(val)))
+#define be16_to_cpu(val)	(__swab16((__force __u16)(__be16)(val)))
+#define be32_to_cpu(val)	(__swab32((__force __u32)(__be32)(val)))
+#define be64_to_cpu(val)	(__swab64((__force __u64)(__be64)(val)))
 #endif
 
+static inline void be16_add_cpu(__be16 *a, __s16 b)
+{
+	*a = cpu_to_be16(be16_to_cpu(*a) + b);
+}
+
+static inline void be32_add_cpu(__be32 *a, __s32 b)
+{
+	*a = cpu_to_be32(be32_to_cpu(*a) + b);
+}
+
+static inline void be64_add_cpu(__be64 *a, __s64 b)
+{
+	*a = cpu_to_be64(be64_to_cpu(*a) + b);
+}
+
 #endif	/* __KERNEL__ */
 
 /* do we need conversion? */
diff --git a/fs/xfs/xfs_bit.h b/fs/xfs/xfs_bit.h
index 8e0e463dae2..bca7b243c31 100644
--- a/fs/xfs/xfs_bit.h
+++ b/fs/xfs/xfs_bit.h
@@ -61,8 +61,7 @@ static inline int xfs_highbit64(__uint64_t v)
 /* Get low bit set out of 32-bit argument, -1 if none set */
 static inline int xfs_lowbit32(__uint32_t v)
 {
-	unsigned long	t = v;
-	return (v) ? find_first_bit(&t, 32) : -1;
+	return ffs(v) - 1;
 }
 
 /* Get low bit set out of 64-bit argument, -1 if none set */
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index a1aab9275d5..138308e70d1 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -393,8 +393,8 @@ xfs_bmap_count_leaves(
 
 STATIC void
 xfs_bmap_disk_count_leaves(
-	xfs_extnum_t		idx,
-	xfs_bmbt_block_t	*block,
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*block,
 	int			numrecs,
 	int			*count);
 
@@ -402,6 +402,53 @@ xfs_bmap_disk_count_leaves(
  * Bmap internal routines.
  */
 
+STATIC int				/* error */
+xfs_bmbt_lookup_eq(
+	struct xfs_btree_cur	*cur,
+	xfs_fileoff_t		off,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.b.br_startoff = off;
+	cur->bc_rec.b.br_startblock = bno;
+	cur->bc_rec.b.br_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
+}
+
+STATIC int				/* error */
+xfs_bmbt_lookup_ge(
+	struct xfs_btree_cur	*cur,
+	xfs_fileoff_t		off,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.b.br_startoff = off;
+	cur->bc_rec.b.br_startblock = bno;
+	cur->bc_rec.b.br_blockcount = len;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+}
+
+/*
+* Update the record referred to by cur to the value given
+ * by [off, bno, len, state].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int
+xfs_bmbt_update(
+	struct xfs_btree_cur	*cur,
+	xfs_fileoff_t		off,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	xfs_exntst_t		state)
+{
+	union xfs_btree_rec	rec;
+
+	xfs_bmbt_disk_set_allf(&rec.bmbt, off, bno, len, state);
+	return xfs_btree_update(cur, &rec);
+}
+
 /*
  * Called from xfs_bmap_add_attrfork to handle btree format files.
  */
@@ -422,15 +469,14 @@ xfs_bmap_add_attrfork_btree(
 	if (ip->i_df.if_broot_bytes <= XFS_IFORK_DSIZE(ip))
 		*flags |= XFS_ILOG_DBROOT;
 	else {
-		cur = xfs_btree_init_cursor(mp, tp, NULL, 0, XFS_BTNUM_BMAP, ip,
-			XFS_DATA_FORK);
+		cur = xfs_bmbt_init_cursor(mp, tp, ip, XFS_DATA_FORK);
 		cur->bc_private.b.flist = flist;
 		cur->bc_private.b.firstblock = *firstblock;
 		if ((error = xfs_bmbt_lookup_ge(cur, 0, 0, 0, &stat)))
 			goto error0;
 		/* must be at least one entry */
 		XFS_WANT_CORRUPTED_GOTO(stat == 1, error0);
-		if ((error = xfs_bmbt_newroot(cur, flags, &stat)))
+		if ((error = xfs_btree_new_iroot(cur, flags, &stat)))
 			goto error0;
 		if (stat == 0) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
@@ -818,10 +864,10 @@ xfs_bmap_add_extent_delay_real(
 					RIGHT.br_blockcount, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
@@ -931,7 +977,7 @@ xfs_bmap_add_extent_delay_real(
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -1007,7 +1053,7 @@ xfs_bmap_add_extent_delay_real(
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -1097,7 +1143,7 @@ xfs_bmap_add_extent_delay_real(
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -1152,7 +1198,7 @@ xfs_bmap_add_extent_delay_real(
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -1379,16 +1425,16 @@ xfs_bmap_add_extent_unwritten_real(
 					RIGHT.br_blockcount, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
@@ -1428,10 +1474,10 @@ xfs_bmap_add_extent_unwritten_real(
 					&i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, LEFT.br_startoff,
@@ -1471,10 +1517,10 @@ xfs_bmap_add_extent_unwritten_real(
 					RIGHT.br_blockcount, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, new->br_startoff,
@@ -1557,7 +1603,7 @@ xfs_bmap_add_extent_unwritten_real(
 				PREV.br_blockcount - new->br_blockcount,
 				oldext)))
 				goto done;
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			if (xfs_bmbt_update(cur, LEFT.br_startoff,
 				LEFT.br_startblock,
@@ -1605,7 +1651,7 @@ xfs_bmap_add_extent_unwritten_real(
 				oldext)))
 				goto done;
 			cur->bc_rec.b = *new;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -1647,7 +1693,7 @@ xfs_bmap_add_extent_unwritten_real(
 				PREV.br_blockcount - new->br_blockcount,
 				oldext)))
 				goto done;
-			if ((error = xfs_bmbt_increment(cur, 0, &i)))
+			if ((error = xfs_btree_increment(cur, 0, &i)))
 				goto done;
 			if ((error = xfs_bmbt_update(cur, new->br_startoff,
 				new->br_startblock,
@@ -1695,7 +1741,7 @@ xfs_bmap_add_extent_unwritten_real(
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = XFS_EXT_NORM;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -1743,7 +1789,7 @@ xfs_bmap_add_extent_unwritten_real(
 			cur->bc_rec.b = PREV;
 			cur->bc_rec.b.br_blockcount =
 				new->br_startoff - PREV.br_startoff;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			/*
@@ -1758,7 +1804,7 @@ xfs_bmap_add_extent_unwritten_real(
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			/* new middle extent - newext */
 			cur->bc_rec.b.br_state = new->br_state;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -2106,10 +2152,10 @@ xfs_bmap_add_extent_hole_real(
 					right.br_blockcount, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_delete(cur, &i)))
+			if ((error = xfs_btree_delete(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
-			if ((error = xfs_bmbt_decrement(cur, 0, &i)))
+			if ((error = xfs_btree_decrement(cur, 0, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 			if ((error = xfs_bmbt_update(cur, left.br_startoff,
@@ -2218,7 +2264,7 @@ xfs_bmap_add_extent_hole_real(
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 0, done);
 			cur->bc_rec.b.br_state = new->br_state;
-			if ((error = xfs_bmbt_insert(cur, &i)))
+			if ((error = xfs_btree_insert(cur, &i)))
 				goto done;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		}
@@ -2996,24 +3042,24 @@ xfs_bmap_btree_to_extents(
 	int			whichfork)  /* data or attr fork */
 {
 	/* REFERENCED */
-	xfs_bmbt_block_t	*cblock;/* child btree block */
+	struct xfs_btree_block	*cblock;/* child btree block */
 	xfs_fsblock_t		cbno;	/* child block number */
 	xfs_buf_t		*cbp;	/* child block's buffer */
 	int			error;	/* error return value */
 	xfs_ifork_t		*ifp;	/* inode fork data */
 	xfs_mount_t		*mp;	/* mount point structure */
 	__be64			*pp;	/* ptr to block address */
-	xfs_bmbt_block_t	*rblock;/* root btree block */
+	struct xfs_btree_block	*rblock;/* root btree block */
 
+	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
 	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
 	rblock = ifp->if_broot;
 	ASSERT(be16_to_cpu(rblock->bb_level) == 1);
 	ASSERT(be16_to_cpu(rblock->bb_numrecs) == 1);
-	ASSERT(XFS_BMAP_BROOT_MAXRECS(ifp->if_broot_bytes) == 1);
-	mp = ip->i_mount;
-	pp = XFS_BMAP_BROOT_PTR_ADDR(rblock, 1, ifp->if_broot_bytes);
+	ASSERT(xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0) == 1);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, ifp->if_broot_bytes);
 	cbno = be64_to_cpu(*pp);
 	*logflagsp = 0;
 #ifdef DEBUG
@@ -3023,8 +3069,8 @@ xfs_bmap_btree_to_extents(
 	if ((error = xfs_btree_read_bufl(mp, tp, cbno, 0, &cbp,
 			XFS_BMAP_BTREE_REF)))
 		return error;
-	cblock = XFS_BUF_TO_BMBT_BLOCK(cbp);
-	if ((error = xfs_btree_check_lblock(cur, cblock, 0, cbp)))
+	cblock = XFS_BUF_TO_BLOCK(cbp);
+	if ((error = xfs_btree_check_block(cur, cblock, 0, cbp)))
 		return error;
 	xfs_bmap_add_free(cbno, 1, cur->bc_private.b.flist, mp);
 	ip->i_d.di_nblocks--;
@@ -3170,7 +3216,7 @@ xfs_bmap_del_extent(
 			flags |= XFS_ILOG_FEXT(whichfork);
 			break;
 		}
-		if ((error = xfs_bmbt_delete(cur, &i)))
+		if ((error = xfs_btree_delete(cur, &i)))
 			goto done;
 		XFS_WANT_CORRUPTED_GOTO(i == 1, done);
 		break;
@@ -3254,10 +3300,10 @@ xfs_bmap_del_extent(
 						got.br_startblock, temp,
 						got.br_state)))
 					goto done;
-				if ((error = xfs_bmbt_increment(cur, 0, &i)))
+				if ((error = xfs_btree_increment(cur, 0, &i)))
 					goto done;
 				cur->bc_rec.b = new;
-				error = xfs_bmbt_insert(cur, &i);
+				error = xfs_btree_insert(cur, &i);
 				if (error && error != ENOSPC)
 					goto done;
 				/*
@@ -3404,11 +3450,11 @@ xfs_bmap_extents_to_btree(
 	int			*logflagsp,	/* inode logging flags */
 	int			whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_block_t	*ablock;	/* allocated (child) bt block */
+	struct xfs_btree_block	*ablock;	/* allocated (child) bt block */
 	xfs_buf_t		*abp;		/* buffer for ablock */
 	xfs_alloc_arg_t		args;		/* allocation arguments */
 	xfs_bmbt_rec_t		*arp;		/* child record pointer */
-	xfs_bmbt_block_t	*block;		/* btree root block */
+	struct xfs_btree_block	*block;		/* btree root block */
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
 	xfs_bmbt_rec_host_t	*ep;		/* extent record pointer */
 	int			error;		/* error return value */
@@ -3428,6 +3474,7 @@ xfs_bmap_extents_to_btree(
 	 */
 	xfs_iroot_realloc(ip, 1, whichfork);
 	ifp->if_flags |= XFS_IFBROOT;
+
 	/*
 	 * Fill in the root.
 	 */
@@ -3435,14 +3482,14 @@ xfs_bmap_extents_to_btree(
 	block->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
 	block->bb_level = cpu_to_be16(1);
 	block->bb_numrecs = cpu_to_be16(1);
-	block->bb_leftsib = cpu_to_be64(NULLDFSBNO);
-	block->bb_rightsib = cpu_to_be64(NULLDFSBNO);
+	block->bb_u.l.bb_leftsib = cpu_to_be64(NULLDFSBNO);
+	block->bb_u.l.bb_rightsib = cpu_to_be64(NULLDFSBNO);
+
 	/*
 	 * Need a cursor.  Can't allocate until bb_level is filled in.
 	 */
 	mp = ip->i_mount;
-	cur = xfs_btree_init_cursor(mp, tp, NULL, 0, XFS_BTNUM_BMAP, ip,
-		whichfork);
+	cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
 	cur->bc_private.b.firstblock = *firstblock;
 	cur->bc_private.b.flist = flist;
 	cur->bc_private.b.flags = wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
@@ -3489,12 +3536,12 @@ xfs_bmap_extents_to_btree(
 	/*
 	 * Fill in the child block.
 	 */
-	ablock = XFS_BUF_TO_BMBT_BLOCK(abp);
+	ablock = XFS_BUF_TO_BLOCK(abp);
 	ablock->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
 	ablock->bb_level = 0;
-	ablock->bb_leftsib = cpu_to_be64(NULLDFSBNO);
-	ablock->bb_rightsib = cpu_to_be64(NULLDFSBNO);
-	arp = XFS_BMAP_REC_IADDR(ablock, 1, cur);
+	ablock->bb_u.l.bb_leftsib = cpu_to_be64(NULLDFSBNO);
+	ablock->bb_u.l.bb_rightsib = cpu_to_be64(NULLDFSBNO);
+	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 	for (cnt = i = 0; i < nextents; i++) {
 		ep = xfs_iext_get_ext(ifp, i);
@@ -3505,21 +3552,24 @@ xfs_bmap_extents_to_btree(
 		}
 	}
 	ASSERT(cnt == XFS_IFORK_NEXTENTS(ip, whichfork));
-	ablock->bb_numrecs = cpu_to_be16(cnt);
+	xfs_btree_set_numrecs(ablock, cnt);
+
 	/*
 	 * Fill in the root key and pointer.
 	 */
-	kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-	arp = XFS_BMAP_REC_IADDR(ablock, 1, cur);
+	kp = XFS_BMBT_KEY_ADDR(mp, block, 1);
+	arp = XFS_BMBT_REC_ADDR(mp, ablock, 1);
 	kp->br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(arp));
-	pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
+	pp = XFS_BMBT_PTR_ADDR(mp, block, 1, xfs_bmbt_get_maxrecs(cur,
+						be16_to_cpu(block->bb_level)));
 	*pp = cpu_to_be64(args.fsbno);
+
 	/*
 	 * Do all this logging at the end so that
 	 * the root is at the right level.
 	 */
-	xfs_bmbt_log_block(cur, abp, XFS_BB_ALL_BITS);
-	xfs_bmbt_log_recs(cur, abp, 1, be16_to_cpu(ablock->bb_numrecs));
+	xfs_btree_log_block(cur, abp, XFS_BB_ALL_BITS);
+	xfs_btree_log_recs(cur, abp, 1, be16_to_cpu(ablock->bb_numrecs));
 	ASSERT(*curp == NULL);
 	*curp = cur;
 	*logflagsp = XFS_ILOG_CORE | XFS_ILOG_FBROOT(whichfork);
@@ -4176,7 +4226,7 @@ xfs_bmap_compute_maxlevels(
 		maxleafents = MAXAEXTNUM;
 		sz = XFS_BMDR_SPACE_CALC(MINABTPTRS);
 	}
-	maxrootrecs = (int)XFS_BTREE_BLOCK_MAXRECS(sz, xfs_bmdr, 0);
+	maxrootrecs = xfs_bmdr_maxrecs(mp, sz, 0);
 	minleafrecs = mp->m_bmap_dmnr[0];
 	minnoderecs = mp->m_bmap_dmnr[1];
 	maxblocks = (maxleafents + minleafrecs - 1) / minleafrecs;
@@ -4242,9 +4292,15 @@ xfs_bmap_finish(
 	 * We have a new transaction, so we should return committed=1,
 	 * even though we're returning an error.
 	 */
-	if (error) {
+	if (error)
 		return error;
-	}
+
+	/*
+	 * transaction commit worked ok so we can drop the extra ticket
+	 * reference that we gained in xfs_trans_dup()
+	 */
+	xfs_log_ticket_put(ntp->t_ticket);
+
 	if ((error = xfs_trans_reserve(ntp, 0, logres, 0, XFS_TRANS_PERM_LOG_RES,
 			logcount)))
 		return error;
@@ -4474,6 +4530,22 @@ xfs_bmap_one_block(
 	return rval;
 }
 
+STATIC int
+xfs_bmap_sanity_check(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	int			level)
+{
+	struct xfs_btree_block  *block = XFS_BUF_TO_BLOCK(bp);
+
+	if (be32_to_cpu(block->bb_magic) != XFS_BMAP_MAGIC ||
+	    be16_to_cpu(block->bb_level) != level ||
+	    be16_to_cpu(block->bb_numrecs) == 0 ||
+	    be16_to_cpu(block->bb_numrecs) > mp->m_bmap_dmxr[level != 0])
+		return 0;
+	return 1;
+}
+
 /*
  * Read in the extents to if_extents.
  * All inode fields are set up by caller, we just traverse the btree
@@ -4486,7 +4558,7 @@ xfs_bmap_read_extents(
 	xfs_inode_t		*ip,	/* incore inode */
 	int			whichfork) /* data or attr fork */
 {
-	xfs_bmbt_block_t	*block;	/* current btree block */
+	struct xfs_btree_block	*block;	/* current btree block */
 	xfs_fsblock_t		bno;	/* block # of "block" */
 	xfs_buf_t		*bp;	/* buffer for "block" */
 	int			error;	/* error return value */
@@ -4510,7 +4582,7 @@ xfs_bmap_read_extents(
 	 */
 	level = be16_to_cpu(block->bb_level);
 	ASSERT(level > 0);
-	pp = XFS_BMAP_BROOT_PTR_ADDR(block, 1, ifp->if_broot_bytes);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
 	bno = be64_to_cpu(*pp);
 	ASSERT(bno != NULLDFSBNO);
 	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
@@ -4523,13 +4595,13 @@ xfs_bmap_read_extents(
 		if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
 				XFS_BMAP_BTREE_REF)))
 			return error;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 		XFS_WANT_CORRUPTED_GOTO(
-			XFS_BMAP_SANITY_CHECK(mp, block, level),
+			xfs_bmap_sanity_check(mp, bp, level),
 			error0);
 		if (level == 0)
 			break;
-		pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
+		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
 		bno = be64_to_cpu(*pp);
 		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
 		xfs_trans_brelse(tp, bp);
@@ -4549,7 +4621,7 @@ xfs_bmap_read_extents(
 		xfs_extnum_t	start;
 
 
-		num_recs = be16_to_cpu(block->bb_numrecs);
+		num_recs = xfs_btree_get_numrecs(block);
 		if (unlikely(i + num_recs > room)) {
 			ASSERT(i + num_recs <= room);
 			xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
@@ -4561,18 +4633,18 @@ xfs_bmap_read_extents(
 			goto error0;
 		}
 		XFS_WANT_CORRUPTED_GOTO(
-			XFS_BMAP_SANITY_CHECK(mp, block, 0),
+			xfs_bmap_sanity_check(mp, bp, 0),
 			error0);
 		/*
 		 * Read-ahead the next leaf block, if any.
 		 */
-		nextbno = be64_to_cpu(block->bb_rightsib);
+		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
 		if (nextbno != NULLFSBLOCK)
 			xfs_btree_reada_bufl(mp, nextbno, 1);
 		/*
 		 * Copy records into the extent records.
 		 */
-		frp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, 1);
+		frp = XFS_BMBT_REC_ADDR(mp, block, 1);
 		start = i;
 		for (j = 0; j < num_recs; j++, i++, frp++) {
 			xfs_bmbt_rec_host_t *trp = xfs_iext_get_ext(ifp, i);
@@ -4603,7 +4675,7 @@ xfs_bmap_read_extents(
 		if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp,
 				XFS_BMAP_BTREE_REF)))
 			return error;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 	}
 	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
 	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
@@ -5029,8 +5101,7 @@ xfs_bmapi(
 				if (abno == NULLFSBLOCK)
 					break;
 				if ((ifp->if_flags & XFS_IFBROOT) && !cur) {
-					cur = xfs_btree_init_cursor(mp,
-						tp, NULL, 0, XFS_BTNUM_BMAP,
+					cur = xfs_bmbt_init_cursor(mp, tp,
 						ip, whichfork);
 					cur->bc_private.b.firstblock =
 						*firstblock;
@@ -5147,9 +5218,8 @@ xfs_bmapi(
 			 */
 			ASSERT(mval->br_blockcount <= len);
 			if ((ifp->if_flags & XFS_IFBROOT) && !cur) {
-				cur = xfs_btree_init_cursor(mp,
-					tp, NULL, 0, XFS_BTNUM_BMAP,
-					ip, whichfork);
+				cur = xfs_bmbt_init_cursor(mp,
+					tp, ip, whichfork);
 				cur->bc_private.b.firstblock =
 					*firstblock;
 				cur->bc_private.b.flist = flist;
@@ -5440,8 +5510,7 @@ xfs_bunmapi(
 	logflags = 0;
 	if (ifp->if_flags & XFS_IFBROOT) {
 		ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
-		cur = xfs_btree_init_cursor(mp, tp, NULL, 0, XFS_BTNUM_BMAP, ip,
-			whichfork);
+		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
 		cur->bc_private.b.firstblock = *firstblock;
 		cur->bc_private.b.flist = flist;
 		cur->bc_private.b.flags = 0;
@@ -5742,14 +5811,17 @@ error0:
 STATIC int
 xfs_getbmapx_fix_eof_hole(
 	xfs_inode_t		*ip,		/* xfs incore inode pointer */
-	struct getbmap		*out,		/* output structure */
+	struct getbmapx		*out,		/* output structure */
 	int			prealloced,	/* this is a file with
-						* preallocated data space */
+						 * preallocated data space */
 	__int64_t		end,		/* last block requested */
 	xfs_fsblock_t		startblock)
 {
 	__int64_t		fixlen;
 	xfs_mount_t		*mp;		/* file system mount point */
+	xfs_ifork_t		*ifp;		/* inode fork pointer */
+	xfs_extnum_t		lastx;		/* last extent pointer */
+	xfs_fileoff_t		fileblock;
 
 	if (startblock == HOLESTARTBLOCK) {
 		mp = ip->i_mount;
@@ -5763,21 +5835,33 @@ xfs_getbmapx_fix_eof_hole(
 			out->bmv_length = fixlen;
 		}
 	} else {
-		out->bmv_block = XFS_FSB_TO_DB(ip, startblock);
+		if (startblock == DELAYSTARTBLOCK)
+			out->bmv_block = -2;
+		else
+			out->bmv_block = XFS_FSB_TO_DB(ip, startblock);
+		fileblock = XFS_BB_TO_FSB(ip->i_mount, out->bmv_offset);
+		ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+		if (xfs_iext_bno_to_ext(ifp, fileblock, &lastx) &&
+		   (lastx == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))-1))
+			out->bmv_oflags |= BMV_OF_LAST;
 	}
 
 	return 1;
 }
 
 /*
- * Fcntl interface to xfs_bmapi.
+ * Get inode's extents as described in bmv, and format for output.
+ * Calls formatter to fill the user's buffer until all extents
+ * are mapped, until the passed-in bmv->bmv_count slots have
+ * been filled, or until the formatter short-circuits the loop,
+ * if it is tracking filled-in extents on its own.
  */
 int						/* error code */
 xfs_getbmap(
 	xfs_inode_t		*ip,
-	struct getbmap		*bmv,		/* user bmap structure */
-	void			__user *ap,	/* pointer to user's array */
-	int			interface)	/* interface flags */
+	struct getbmapx		*bmv,		/* user bmap structure */
+	xfs_bmap_format_t	formatter,	/* format to user */
+	void			*arg)		/* formatter arg */
 {
 	__int64_t		bmvend;		/* last block requested */
 	int			error;		/* return value */
@@ -5790,19 +5874,17 @@ xfs_getbmap(
 	int			nexleft;	/* # of user extents left */
 	int			subnex;		/* # of bmapi's can do */
 	int			nmap;		/* number of map entries */
-	struct getbmap		out;		/* output structure */
+	struct getbmapx		out;		/* output structure */
 	int			whichfork;	/* data or attr fork */
 	int			prealloced;	/* this is a file with
 						 * preallocated data space */
-	int			sh_unwritten;	/* true, if unwritten */
-						/* extents listed separately */
+	int			iflags;		/* interface flags */
 	int			bmapi_flags;	/* flags for xfs_bmapi */
-	__int32_t		oflags;		/* getbmapx bmv_oflags field */
 
 	mp = ip->i_mount;
+	iflags = bmv->bmv_iflags;
 
-	whichfork = interface & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK;
-	sh_unwritten = (interface & BMV_IF_PREALLOC) != 0;
+	whichfork = iflags & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK;
 
 	/*	If the BMV_IF_NO_DMAPI_READ interface bit specified, do not
 	 *	generate a DMAPI read event.  Otherwise, if the DM_EVENT_READ
@@ -5817,7 +5899,7 @@ xfs_getbmap(
 	 *	could misinterpret holes in a DMAPI file as true holes,
 	 *	when in fact they may represent offline user data.
 	 */
-	if ((interface & BMV_IF_NO_DMAPI_READ) == 0 &&
+	if ((iflags & BMV_IF_NO_DMAPI_READ) == 0 &&
 	    DM_EVENT_ENABLED(ip, DM_EVENT_READ) &&
 	    whichfork == XFS_DATA_FORK) {
 		error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip, 0, 0, 0, NULL);
@@ -5873,8 +5955,9 @@ xfs_getbmap(
 
 	xfs_ilock(ip, XFS_IOLOCK_SHARED);
 
-	if (whichfork == XFS_DATA_FORK &&
-		(ip->i_delayed_blks || ip->i_size > ip->i_d.di_size)) {
+	if (((iflags & BMV_IF_DELALLOC) == 0) &&
+	    (whichfork == XFS_DATA_FORK) &&
+	    (ip->i_delayed_blks || ip->i_size > ip->i_d.di_size)) {
 		/* xfs_fsize_t last_byte = xfs_file_last_byte(ip); */
 		error = xfs_flush_pages(ip, (xfs_off_t)0,
 					       -1, 0, FI_REMAPF);
@@ -5884,7 +5967,8 @@ xfs_getbmap(
 		}
 	}
 
-	ASSERT(whichfork == XFS_ATTR_FORK || ip->i_delayed_blks == 0);
+	ASSERT(whichfork == XFS_ATTR_FORK || (iflags & BMV_IF_DELALLOC) ||
+	       ip->i_delayed_blks == 0);
 
 	lock = xfs_ilock_map_shared(ip);
 
@@ -5896,7 +5980,7 @@ xfs_getbmap(
 		nex = XFS_IFORK_NEXTENTS(ip, whichfork) * 2 + 1;
 
 	bmapi_flags = XFS_BMAPI_AFLAG(whichfork) |
-			((sh_unwritten) ? 0 : XFS_BMAPI_IGSTATE);
+			((iflags & BMV_IF_PREALLOC) ? 0 : XFS_BMAPI_IGSTATE);
 
 	/*
 	 * Allocate enough space to handle "subnex" maps at a time.
@@ -5906,9 +5990,12 @@ xfs_getbmap(
 
 	bmv->bmv_entries = 0;
 
-	if (XFS_IFORK_NEXTENTS(ip, whichfork) == 0) {
-		error = 0;
-		goto unlock_and_return;
+	if ((XFS_IFORK_NEXTENTS(ip, whichfork) == 0)) {
+		if (((iflags & BMV_IF_DELALLOC) == 0) ||
+		    whichfork == XFS_ATTR_FORK) {
+			error = 0;
+			goto unlock_and_return;
+		}
 	}
 
 	nexleft = nex;
@@ -5924,52 +6011,40 @@ xfs_getbmap(
 		ASSERT(nmap <= subnex);
 
 		for (i = 0; i < nmap && nexleft && bmv->bmv_length; i++) {
-			nexleft--;
-			oflags = (map[i].br_state == XFS_EXT_UNWRITTEN) ?
-					BMV_OF_PREALLOC : 0;
+			out.bmv_oflags = 0;
+			if (map[i].br_state == XFS_EXT_UNWRITTEN)
+				out.bmv_oflags |= BMV_OF_PREALLOC;
+			else if (map[i].br_startblock == DELAYSTARTBLOCK)
+				out.bmv_oflags |= BMV_OF_DELALLOC;
 			out.bmv_offset = XFS_FSB_TO_BB(mp, map[i].br_startoff);
 			out.bmv_length = XFS_FSB_TO_BB(mp, map[i].br_blockcount);
-			ASSERT(map[i].br_startblock != DELAYSTARTBLOCK);
+			out.bmv_unused1 = out.bmv_unused2 = 0;
+			ASSERT(((iflags & BMV_IF_DELALLOC) != 0) ||
+			      (map[i].br_startblock != DELAYSTARTBLOCK));
                         if (map[i].br_startblock == HOLESTARTBLOCK &&
 			    whichfork == XFS_ATTR_FORK) {
 				/* came to the end of attribute fork */
+				out.bmv_oflags |= BMV_OF_LAST;
 				goto unlock_and_return;
 			} else {
+				int full = 0;	/* user array is full */
+
 				if (!xfs_getbmapx_fix_eof_hole(ip, &out,
 							prealloced, bmvend,
 							map[i].br_startblock)) {
 					goto unlock_and_return;
 				}
 
-				/* return either getbmap/getbmapx structure. */
-				if (interface & BMV_IF_EXTENDED) {
-					struct	getbmapx	outx;
-
-					GETBMAP_CONVERT(out,outx);
-					outx.bmv_oflags = oflags;
-					outx.bmv_unused1 = outx.bmv_unused2 = 0;
-					if (copy_to_user(ap, &outx,
-							sizeof(outx))) {
-						error = XFS_ERROR(EFAULT);
-						goto unlock_and_return;
-					}
-				} else {
-					if (copy_to_user(ap, &out,
-							sizeof(out))) {
-						error = XFS_ERROR(EFAULT);
-						goto unlock_and_return;
-					}
-				}
+				/* format results & advance arg */
+				error = formatter(&arg, &out, &full);
+				if (error || full)
+					goto unlock_and_return;
+				nexleft--;
 				bmv->bmv_offset =
 					out.bmv_offset + out.bmv_length;
 				bmv->bmv_length = MAX((__int64_t)0,
 					(__int64_t)(bmvend - bmv->bmv_offset));
 				bmv->bmv_entries++;
-				ap = (interface & BMV_IF_EXTENDED) ?
-						(void __user *)
-					((struct getbmapx __user *)ap + 1) :
-						(void __user *)
-					((struct getbmap __user *)ap + 1);
 			}
 		}
 	} while (nmap && nexleft && bmv->bmv_length);
@@ -6131,7 +6206,7 @@ xfs_bmap_get_bp(
 
 void
 xfs_check_block(
-	xfs_bmbt_block_t        *block,
+	struct xfs_btree_block	*block,
 	xfs_mount_t		*mp,
 	int			root,
 	short			sz)
@@ -6143,36 +6218,29 @@ xfs_check_block(
 	ASSERT(be16_to_cpu(block->bb_level) > 0);
 
 	prevp = NULL;
-	for( i = 1; i <= be16_to_cpu(block->bb_numrecs); i++) {
+	for( i = 1; i <= xfs_btree_get_numrecs(block); i++) {
 		dmxr = mp->m_bmap_dmxr[0];
-
-		if (root) {
-			keyp = XFS_BMAP_BROOT_KEY_ADDR(block, i, sz);
-		} else {
-			keyp = XFS_BTREE_KEY_ADDR(xfs_bmbt, block, i);
-		}
+		keyp = XFS_BMBT_KEY_ADDR(mp, block, i);
 
 		if (prevp) {
-			xfs_btree_check_key(XFS_BTNUM_BMAP, prevp, keyp);
+			ASSERT(be64_to_cpu(prevp->br_startoff) <
+			       be64_to_cpu(keyp->br_startoff));
 		}
 		prevp = keyp;
 
 		/*
 		 * Compare the block numbers to see if there are dups.
 		 */
+		if (root)
+			pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, i, sz);
+		else
+			pp = XFS_BMBT_PTR_ADDR(mp, block, i, dmxr);
 
-		if (root) {
-			pp = XFS_BMAP_BROOT_PTR_ADDR(block, i, sz);
-		} else {
-			pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, i, dmxr);
-		}
 		for (j = i+1; j <= be16_to_cpu(block->bb_numrecs); j++) {
-			if (root) {
-				thispa = XFS_BMAP_BROOT_PTR_ADDR(block, j, sz);
-			} else {
-				thispa = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, j,
-							    dmxr);
-			}
+			if (root)
+				thispa = XFS_BMAP_BROOT_PTR_ADDR(mp, block, j, sz);
+			else
+				thispa = XFS_BMBT_PTR_ADDR(mp, block, j, dmxr);
 			if (*thispa == *pp) {
 				cmn_err(CE_WARN, "%s: thispa(%d) == pp(%d) %Ld",
 					__func__, j, i,
@@ -6195,7 +6263,7 @@ xfs_bmap_check_leaf_extents(
 	xfs_inode_t		*ip,		/* incore inode pointer */
 	int			whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_block_t	*block;	/* current btree block */
+	struct xfs_btree_block	*block;	/* current btree block */
 	xfs_fsblock_t		bno;	/* block # of "block" */
 	xfs_buf_t		*bp;	/* buffer for "block" */
 	int			error;	/* error return value */
@@ -6223,7 +6291,7 @@ xfs_bmap_check_leaf_extents(
 	level = be16_to_cpu(block->bb_level);
 	ASSERT(level > 0);
 	xfs_check_block(block, mp, 1, ifp->if_broot_bytes);
-	pp = XFS_BMAP_BROOT_PTR_ADDR(block, 1, ifp->if_broot_bytes);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
 	bno = be64_to_cpu(*pp);
 
 	ASSERT(bno != NULLDFSBNO);
@@ -6245,9 +6313,9 @@ xfs_bmap_check_leaf_extents(
 		if (!bp && (error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
 				XFS_BMAP_BTREE_REF)))
 			goto error_norelse;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 		XFS_WANT_CORRUPTED_GOTO(
-			XFS_BMAP_SANITY_CHECK(mp, block, level),
+			xfs_bmap_sanity_check(mp, bp, level),
 			error0);
 		if (level == 0)
 			break;
@@ -6258,7 +6326,7 @@ xfs_bmap_check_leaf_extents(
 		 */
 
 		xfs_check_block(block, mp, 0, 0);
-		pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
+		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
 		bno = be64_to_cpu(*pp);
 		XFS_WANT_CORRUPTED_GOTO(XFS_FSB_SANITY_CHECK(mp, bno), error0);
 		if (bp_release) {
@@ -6280,13 +6348,13 @@ xfs_bmap_check_leaf_extents(
 		xfs_extnum_t	num_recs;
 
 
-		num_recs = be16_to_cpu(block->bb_numrecs);
+		num_recs = xfs_btree_get_numrecs(block);
 
 		/*
 		 * Read-ahead the next leaf block, if any.
 		 */
 
-		nextbno = be64_to_cpu(block->bb_rightsib);
+		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
 
 		/*
 		 * Check all the extents to make sure they are OK.
@@ -6294,13 +6362,17 @@ xfs_bmap_check_leaf_extents(
 		 * conform with the first entry in this one.
 		 */
 
-		ep = XFS_BTREE_REC_ADDR(xfs_bmbt, block, 1);
+		ep = XFS_BMBT_REC_ADDR(mp, block, 1);
 		if (i) {
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, &last, ep);
+			ASSERT(xfs_bmbt_disk_get_startoff(&last) +
+			       xfs_bmbt_disk_get_blockcount(&last) <=
+			       xfs_bmbt_disk_get_startoff(ep));
 		}
 		for (j = 1; j < num_recs; j++) {
-			nextp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, j + 1);
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, ep, nextp);
+			nextp = XFS_BMBT_REC_ADDR(mp, block, j + 1);
+			ASSERT(xfs_bmbt_disk_get_startoff(ep) +
+			       xfs_bmbt_disk_get_blockcount(ep) <=
+			       xfs_bmbt_disk_get_startoff(nextp));
 			ep = nextp;
 		}
 
@@ -6326,7 +6398,7 @@ xfs_bmap_check_leaf_extents(
 		if (!bp && (error = xfs_btree_read_bufl(mp, NULL, bno, 0, &bp,
 				XFS_BMAP_BTREE_REF)))
 			goto error_norelse;
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 	}
 	if (bp_release) {
 		bp_release = 0;
@@ -6356,7 +6428,7 @@ xfs_bmap_count_blocks(
 	int			whichfork,	/* data or attr fork */
 	int			*count)		/* out: count of blocks */
 {
-	xfs_bmbt_block_t	*block;	/* current btree block */
+	struct xfs_btree_block	*block;	/* current btree block */
 	xfs_fsblock_t		bno;	/* block # of "block" */
 	xfs_ifork_t		*ifp;	/* fork structure */
 	int			level;	/* btree level, for checking */
@@ -6379,7 +6451,7 @@ xfs_bmap_count_blocks(
 	block = ifp->if_broot;
 	level = be16_to_cpu(block->bb_level);
 	ASSERT(level > 0);
-	pp = XFS_BMAP_BROOT_PTR_ADDR(block, 1, ifp->if_broot_bytes);
+	pp = XFS_BMAP_BROOT_PTR_ADDR(mp, block, 1, ifp->if_broot_bytes);
 	bno = be64_to_cpu(*pp);
 	ASSERT(bno != NULLDFSBNO);
 	ASSERT(XFS_FSB_TO_AGNO(mp, bno) < mp->m_sb.sb_agcount);
@@ -6413,29 +6485,29 @@ xfs_bmap_count_tree(
 	__be64			*pp;
 	xfs_fsblock_t           bno = blockno;
 	xfs_fsblock_t		nextbno;
-	xfs_bmbt_block_t        *block, *nextblock;
+	struct xfs_btree_block	*block, *nextblock;
 	int			numrecs;
 
 	if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp, XFS_BMAP_BTREE_REF)))
 		return error;
 	*count += 1;
-	block = XFS_BUF_TO_BMBT_BLOCK(bp);
+	block = XFS_BUF_TO_BLOCK(bp);
 
 	if (--level) {
 		/* Not at node above leafs, count this level of nodes */
-		nextbno = be64_to_cpu(block->bb_rightsib);
+		nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
 		while (nextbno != NULLFSBLOCK) {
 			if ((error = xfs_btree_read_bufl(mp, tp, nextbno,
 				0, &nbp, XFS_BMAP_BTREE_REF)))
 				return error;
 			*count += 1;
-			nextblock = XFS_BUF_TO_BMBT_BLOCK(nbp);
-			nextbno = be64_to_cpu(nextblock->bb_rightsib);
+			nextblock = XFS_BUF_TO_BLOCK(nbp);
+			nextbno = be64_to_cpu(nextblock->bb_u.l.bb_rightsib);
 			xfs_trans_brelse(tp, nbp);
 		}
 
 		/* Dive to the next level */
-		pp = XFS_BTREE_PTR_ADDR(xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
+		pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]);
 		bno = be64_to_cpu(*pp);
 		if (unlikely((error =
 		     xfs_bmap_count_tree(mp, tp, ifp, bno, level, count)) < 0)) {
@@ -6448,9 +6520,9 @@ xfs_bmap_count_tree(
 	} else {
 		/* count all level 1 nodes and their leaves */
 		for (;;) {
-			nextbno = be64_to_cpu(block->bb_rightsib);
+			nextbno = be64_to_cpu(block->bb_u.l.bb_rightsib);
 			numrecs = be16_to_cpu(block->bb_numrecs);
-			xfs_bmap_disk_count_leaves(0, block, numrecs, count);
+			xfs_bmap_disk_count_leaves(mp, block, numrecs, count);
 			xfs_trans_brelse(tp, bp);
 			if (nextbno == NULLFSBLOCK)
 				break;
@@ -6459,7 +6531,7 @@ xfs_bmap_count_tree(
 				XFS_BMAP_BTREE_REF)))
 				return error;
 			*count += 1;
-			block = XFS_BUF_TO_BMBT_BLOCK(bp);
+			block = XFS_BUF_TO_BLOCK(bp);
 		}
 	}
 	return 0;
@@ -6489,8 +6561,8 @@ xfs_bmap_count_leaves(
  */
 STATIC void
 xfs_bmap_disk_count_leaves(
-	xfs_extnum_t		idx,
-	xfs_bmbt_block_t	*block,
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*block,
 	int			numrecs,
 	int			*count)
 {
@@ -6498,7 +6570,7 @@ xfs_bmap_disk_count_leaves(
 	xfs_bmbt_rec_t	*frp;
 
 	for (b = 1; b <= numrecs; b++) {
-		frp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, idx + b);
+		frp = XFS_BMBT_REC_ADDR(mp, block, b);
 		*count += xfs_bmbt_disk_get_blockcount(frp);
 	}
 }
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index 9f3e3a836d1..284571c05ed 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -137,9 +137,7 @@ typedef struct xfs_bmalloca {
 	char			conv;	/* overwriting unwritten extents */
 } xfs_bmalloca_t;
 
-#ifdef __KERNEL__
-
-#if defined(XFS_BMAP_TRACE)
+#if defined(__KERNEL__) && defined(XFS_BMAP_TRACE)
 /*
  * Trace operations for bmap extent tracing
  */
@@ -163,9 +161,12 @@ xfs_bmap_trace_exlist(
 	int			whichfork);	/* data or attr fork */
 #define	XFS_BMAP_TRACE_EXLIST(ip,c,w)	\
 	xfs_bmap_trace_exlist(__func__,ip,c,w)
-#else
+
+#else	/* __KERNEL__ && XFS_BMAP_TRACE */
+
 #define	XFS_BMAP_TRACE_EXLIST(ip,c,w)
-#endif
+
+#endif	/* __KERNEL__ && XFS_BMAP_TRACE */
 
 /*
  * Convert inode from non-attributed to attributed.
@@ -206,20 +207,6 @@ xfs_bmap_compute_maxlevels(
 	int			whichfork);	/* data or attr fork */
 
 /*
- * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi
- * caller.  Frees all the extents that need freeing, which must be done
- * last due to locking considerations.
- *
- * Return 1 if the given transaction was committed and a new one allocated,
- * and 0 otherwise.
- */
-int						/* error */
-xfs_bmap_finish(
-	struct xfs_trans	**tp,		/* transaction pointer addr */
-	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
-	int			*committed);	/* xact committed or not */
-
-/*
  * Returns the file-relative block number of the first unused block in the file.
  * This is the lowest-address hole if the file has holes, else the first block
  * past the end of file.
@@ -344,14 +331,43 @@ xfs_bunmapi(
 	int			*done);		/* set if not done yet */
 
 /*
- * Fcntl interface to xfs_bmapi.
+ * Check an extent list, which has just been read, for
+ * any bit in the extent flag field.
+ */
+int
+xfs_check_nostate_extents(
+	struct xfs_ifork	*ifp,
+	xfs_extnum_t		idx,
+	xfs_extnum_t		num);
+
+#ifdef __KERNEL__
+
+/*
+ * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi
+ * caller.  Frees all the extents that need freeing, which must be done
+ * last due to locking considerations.
+ *
+ * Return 1 if the given transaction was committed and a new one allocated,
+ * and 0 otherwise.
+ */
+int						/* error */
+xfs_bmap_finish(
+	struct xfs_trans	**tp,		/* transaction pointer addr */
+	xfs_bmap_free_t		*flist,		/* i/o: list extents to free */
+	int			*committed);	/* xact committed or not */
+
+/* bmap to userspace formatter - copy to user & advance pointer */
+typedef int (*xfs_bmap_format_t)(void **, struct getbmapx *, int *);
+
+/*
+ * Get inode's extents as described in bmv, and format for output.
  */
 int						/* error code */
 xfs_getbmap(
 	xfs_inode_t		*ip,
-	struct getbmap		*bmv,		/* user bmap structure */
-	void			__user *ap,	/* pointer to user's array */
-	int			iflags);	/* interface flags */
+	struct getbmapx		*bmv,		/* user bmap structure */
+	xfs_bmap_format_t	formatter,	/* format to user */
+	void			*arg);		/* formatter arg */
 
 /*
  * Check if the endoff is outside the last extent. If so the caller will grow
@@ -375,16 +391,6 @@ xfs_bmap_count_blocks(
 	int			*count);
 
 /*
- * Check an extent list, which has just been read, for
- * any bit in the extent flag field.
- */
-int
-xfs_check_nostate_extents(
-	struct xfs_ifork	*ifp,
-	xfs_extnum_t		idx,
-	xfs_extnum_t		num);
-
-/*
  * Search the extent records for the entry containing block bno.
  * If bno lies in a hole, point to the next entry.  If bno lies
  * past eof, *eofp will be set, and *prevp will contain the last
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index 23efad29a5c..8f1ec73725d 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -37,1406 +37,13 @@
 #include "xfs_inode_item.h"
 #include "xfs_alloc.h"
 #include "xfs_btree.h"
+#include "xfs_btree_trace.h"
 #include "xfs_ialloc.h"
 #include "xfs_itable.h"
 #include "xfs_bmap.h"
 #include "xfs_error.h"
 #include "xfs_quota.h"
 
-#if defined(XFS_BMBT_TRACE)
-ktrace_t	*xfs_bmbt_trace_buf;
-#endif
-
-/*
- * Prototypes for internal btree functions.
- */
-
-
-STATIC int xfs_bmbt_killroot(xfs_btree_cur_t *);
-STATIC void xfs_bmbt_log_keys(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_bmbt_log_ptrs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC int xfs_bmbt_lshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_bmbt_rshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_bmbt_split(xfs_btree_cur_t *, int, xfs_fsblock_t *,
-		__uint64_t *, xfs_btree_cur_t **, int *);
-STATIC int xfs_bmbt_updkey(xfs_btree_cur_t *, xfs_bmbt_key_t *, int);
-
-
-#if defined(XFS_BMBT_TRACE)
-
-static char	ARGS[] = "args";
-static char	ENTRY[] = "entry";
-static char	ERROR[] = "error";
-#undef EXIT
-static char	EXIT[] = "exit";
-
-/*
- * Add a trace buffer entry for the arguments given to the routine,
- * generic form.
- */
-STATIC void
-xfs_bmbt_trace_enter(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	char		*s,
-	int		type,
-	int		line,
-	__psunsigned_t	a0,
-	__psunsigned_t	a1,
-	__psunsigned_t	a2,
-	__psunsigned_t	a3,
-	__psunsigned_t	a4,
-	__psunsigned_t	a5,
-	__psunsigned_t	a6,
-	__psunsigned_t	a7,
-	__psunsigned_t	a8,
-	__psunsigned_t	a9,
-	__psunsigned_t	a10)
-{
-	xfs_inode_t	*ip;
-	int		whichfork;
-
-	ip = cur->bc_private.b.ip;
-	whichfork = cur->bc_private.b.whichfork;
-	ktrace_enter(xfs_bmbt_trace_buf,
-		(void *)((__psint_t)type | (whichfork << 8) | (line << 16)),
-		(void *)func, (void *)s, (void *)ip, (void *)cur,
-		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
-		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
-		(void *)a8, (void *)a9, (void *)a10);
-	ASSERT(ip->i_btrace);
-	ktrace_enter(ip->i_btrace,
-		(void *)((__psint_t)type | (whichfork << 8) | (line << 16)),
-		(void *)func, (void *)s, (void *)ip, (void *)cur,
-		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
-		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
-		(void *)a8, (void *)a9, (void *)a10);
-}
-/*
- * Add a trace buffer entry for arguments, for a buffer & 1 integer arg.
- */
-STATIC void
-xfs_bmbt_trace_argbi(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*b,
-	int		i,
-	int		line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGBI, line,
-		(__psunsigned_t)b, i, 0, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for a buffer & 2 integer args.
- */
-STATIC void
-xfs_bmbt_trace_argbii(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*b,
-	int		i0,
-	int		i1,
-	int		line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGBII, line,
-		(__psunsigned_t)b, i0, i1, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for 3 block-length args
- * and an integer arg.
- */
-STATIC void
-xfs_bmbt_trace_argfffi(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	xfs_dfiloff_t		o,
-	xfs_dfsbno_t		b,
-	xfs_dfilblks_t		i,
-	int			j,
-	int			line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGFFFI, line,
-		o >> 32, (int)o, b >> 32, (int)b,
-		i >> 32, (int)i, (int)j, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for one integer arg.
- */
-STATIC void
-xfs_bmbt_trace_argi(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	int		i,
-	int		line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGI, line,
-		i, 0, 0, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for int, fsblock, key.
- */
-STATIC void
-xfs_bmbt_trace_argifk(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	int			i,
-	xfs_fsblock_t		f,
-	xfs_dfiloff_t		o,
-	int			line)
-{
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGIFK, line,
-		i, (xfs_dfsbno_t)f >> 32, (int)f, o >> 32,
-		(int)o, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for int, fsblock, rec.
- */
-STATIC void
-xfs_bmbt_trace_argifr(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	int			i,
-	xfs_fsblock_t		f,
-	xfs_bmbt_rec_t		*r,
-	int			line)
-{
-	xfs_dfsbno_t		b;
-	xfs_dfilblks_t		c;
-	xfs_dfsbno_t		d;
-	xfs_dfiloff_t		o;
-	xfs_bmbt_irec_t		s;
-
-	d = (xfs_dfsbno_t)f;
-	xfs_bmbt_disk_get_all(r, &s);
-	o = (xfs_dfiloff_t)s.br_startoff;
-	b = (xfs_dfsbno_t)s.br_startblock;
-	c = s.br_blockcount;
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGIFR, line,
-		i, d >> 32, (int)d, o >> 32,
-		(int)o, b >> 32, (int)b, c >> 32,
-		(int)c, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for arguments, for int, key.
- */
-STATIC void
-xfs_bmbt_trace_argik(
-	const char		*func,
-	xfs_btree_cur_t		*cur,
-	int			i,
-	xfs_bmbt_key_t		*k,
-	int			line)
-{
-	xfs_dfiloff_t		o;
-
-	o = be64_to_cpu(k->br_startoff);
-	xfs_bmbt_trace_enter(func, cur, ARGS, XFS_BMBT_KTRACE_ARGIFK, line,
-		i, o >> 32, (int)o, 0,
-		0, 0, 0, 0,
-		0, 0, 0);
-}
-
-/*
- * Add a trace buffer entry for the cursor/operation.
- */
-STATIC void
-xfs_bmbt_trace_cursor(
-	const char	*func,
-	xfs_btree_cur_t	*cur,
-	char		*s,
-	int		line)
-{
-	xfs_bmbt_rec_host_t	r;
-
-	xfs_bmbt_set_all(&r, &cur->bc_rec.b);
-	xfs_bmbt_trace_enter(func, cur, s, XFS_BMBT_KTRACE_CUR, line,
-		(cur->bc_nlevels << 24) | (cur->bc_private.b.flags << 16) |
-		cur->bc_private.b.allocated,
-		r.l0 >> 32, (int)r.l0,
-		r.l1 >> 32, (int)r.l1,
-		(unsigned long)cur->bc_bufs[0], (unsigned long)cur->bc_bufs[1],
-		(unsigned long)cur->bc_bufs[2], (unsigned long)cur->bc_bufs[3],
-		(cur->bc_ptrs[0] << 16) | cur->bc_ptrs[1],
-		(cur->bc_ptrs[2] << 16) | cur->bc_ptrs[3]);
-}
-
-#define	XFS_BMBT_TRACE_ARGBI(c,b,i)	\
-	xfs_bmbt_trace_argbi(__func__, c, b, i, __LINE__)
-#define	XFS_BMBT_TRACE_ARGBII(c,b,i,j)	\
-	xfs_bmbt_trace_argbii(__func__, c, b, i, j, __LINE__)
-#define	XFS_BMBT_TRACE_ARGFFFI(c,o,b,i,j)	\
-	xfs_bmbt_trace_argfffi(__func__, c, o, b, i, j, __LINE__)
-#define	XFS_BMBT_TRACE_ARGI(c,i)	\
-	xfs_bmbt_trace_argi(__func__, c, i, __LINE__)
-#define	XFS_BMBT_TRACE_ARGIFK(c,i,f,s)	\
-	xfs_bmbt_trace_argifk(__func__, c, i, f, s, __LINE__)
-#define	XFS_BMBT_TRACE_ARGIFR(c,i,f,r)	\
-	xfs_bmbt_trace_argifr(__func__, c, i, f, r, __LINE__)
-#define	XFS_BMBT_TRACE_ARGIK(c,i,k)	\
-	xfs_bmbt_trace_argik(__func__, c, i, k, __LINE__)
-#define	XFS_BMBT_TRACE_CURSOR(c,s)	\
-	xfs_bmbt_trace_cursor(__func__, c, s, __LINE__)
-#else
-#define	XFS_BMBT_TRACE_ARGBI(c,b,i)
-#define	XFS_BMBT_TRACE_ARGBII(c,b,i,j)
-#define	XFS_BMBT_TRACE_ARGFFFI(c,o,b,i,j)
-#define	XFS_BMBT_TRACE_ARGI(c,i)
-#define	XFS_BMBT_TRACE_ARGIFK(c,i,f,s)
-#define	XFS_BMBT_TRACE_ARGIFR(c,i,f,r)
-#define	XFS_BMBT_TRACE_ARGIK(c,i,k)
-#define	XFS_BMBT_TRACE_CURSOR(c,s)
-#endif	/* XFS_BMBT_TRACE */
-
-
-/*
- * Internal functions.
- */
-
-/*
- * Delete record pointed to by cur/level.
- */
-STATIC int					/* error */
-xfs_bmbt_delrec(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block;		/* bmap btree block */
-	xfs_fsblock_t		bno;		/* fs-relative block number */
-	xfs_buf_t		*bp;		/* buffer for block */
-	int			error;		/* error return value */
-	int			i;		/* loop counter */
-	int			j;		/* temp state */
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_bmbt_key_t		*kp=NULL;	/* pointer to bmap btree key */
-	xfs_fsblock_t		lbno;		/* left sibling block number */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp;		/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	int			lrecs=0;	/* left record count */
-	xfs_bmbt_rec_t		*lrp;		/* left record pointer */
-	xfs_mount_t		*mp;		/* file system mount point */
-	xfs_bmbt_ptr_t		*pp;		/* pointer to bmap block addr */
-	int			ptr;		/* key/record index */
-	xfs_fsblock_t		rbno;		/* right sibling block number */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp;		/* right btree key */
-	xfs_bmbt_rec_t		*rp;		/* pointer to bmap btree rec */
-	xfs_bmbt_ptr_t		*rpp;		/* right address pointer */
-	xfs_bmbt_block_t	*rrblock;	/* right-right btree block */
-	xfs_buf_t		*rrbp;		/* right-right buffer pointer */
-	int			rrecs=0;	/* right record count */
-	xfs_bmbt_rec_t		*rrp;		/* right record pointer */
-	xfs_btree_cur_t		*tcur;		/* temporary btree cursor */
-	int			numrecs;	/* temporary numrec count */
-	int			numlrecs, numrrecs;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	ptr = cur->bc_ptrs[level];
-	tcur = NULL;
-	if (ptr == 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	block = xfs_bmbt_get_block(cur, level, &bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-#endif
-	if (ptr > numrecs) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_bmbt_delrec);
-	if (level > 0) {
-		kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-		pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = ptr; i < numrecs; i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, pp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-		}
-#endif
-		if (ptr < numrecs) {
-			memmove(&kp[ptr - 1], &kp[ptr],
-				(numrecs - ptr) * sizeof(*kp));
-			memmove(&pp[ptr - 1], &pp[ptr],
-				(numrecs - ptr) * sizeof(*pp));
-			xfs_bmbt_log_ptrs(cur, bp, ptr, numrecs - 1);
-			xfs_bmbt_log_keys(cur, bp, ptr, numrecs - 1);
-		}
-	} else {
-		rp = XFS_BMAP_REC_IADDR(block, 1, cur);
-		if (ptr < numrecs) {
-			memmove(&rp[ptr - 1], &rp[ptr],
-				(numrecs - ptr) * sizeof(*rp));
-			xfs_bmbt_log_recs(cur, bp, ptr, numrecs - 1);
-		}
-		if (ptr == 1) {
-			key.br_startoff =
-				cpu_to_be64(xfs_bmbt_disk_get_startoff(rp));
-			kp = &key;
-		}
-	}
-	numrecs--;
-	block->bb_numrecs = cpu_to_be16(numrecs);
-	xfs_bmbt_log_block(cur, bp, XFS_BB_NUMRECS);
-	/*
-	 * We're at the root level.
-	 * First, shrink the root block in-memory.
-	 * Try to get rid of the next level down.
-	 * If we can't then there's nothing left to do.
-	 */
-	if (level == cur->bc_nlevels - 1) {
-		xfs_iroot_realloc(cur->bc_private.b.ip, -1,
-			cur->bc_private.b.whichfork);
-		if ((error = xfs_bmbt_killroot(cur))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &j))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	if (ptr == 1 && (error = xfs_bmbt_updkey(cur, kp, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-	if (numrecs >= XFS_BMAP_BLOCK_IMINRECS(level, cur)) {
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &j))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	rbno = be64_to_cpu(block->bb_rightsib);
-	lbno = be64_to_cpu(block->bb_leftsib);
-	/*
-	 * One child of root, need to get a chance to copy its contents
-	 * into the root and delete it. Can't go up to next level,
-	 * there's nothing to delete there.
-	 */
-	if (lbno == NULLFSBLOCK && rbno == NULLFSBLOCK &&
-	    level == cur->bc_nlevels - 2) {
-		if ((error = xfs_bmbt_killroot(cur))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	ASSERT(rbno != NULLFSBLOCK || lbno != NULLFSBLOCK);
-	if ((error = xfs_btree_dup_cursor(cur, &tcur))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-	bno = NULLFSBLOCK;
-	if (rbno != NULLFSBLOCK) {
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_bmbt_increment(tcur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		rbp = tcur->bc_bufs[level];
-		right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-#endif
-		bno = be64_to_cpu(right->bb_leftsib);
-		if (be16_to_cpu(right->bb_numrecs) - 1 >=
-		    XFS_BMAP_BLOCK_IMINRECS(level, cur)) {
-			if ((error = xfs_bmbt_lshift(tcur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_BMAP_BLOCK_IMINRECS(level, tcur));
-				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-				tcur = NULL;
-				if (level > 0) {
-					if ((error = xfs_bmbt_decrement(cur,
-							level, &i))) {
-						XFS_BMBT_TRACE_CURSOR(cur,
-							ERROR);
-						goto error0;
-					}
-				}
-				XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-				*stat = 1;
-				return 0;
-			}
-		}
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if (lbno != NULLFSBLOCK) {
-			i = xfs_btree_firstrec(tcur, level);
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			if ((error = xfs_bmbt_decrement(tcur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		}
-	}
-	if (lbno != NULLFSBLOCK) {
-		i = xfs_btree_firstrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		/*
-		 * decrement to last in block
-		 */
-		if ((error = xfs_bmbt_decrement(tcur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		i = xfs_btree_firstrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		lbp = tcur->bc_bufs[level];
-		left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-#endif
-		bno = be64_to_cpu(left->bb_rightsib);
-		if (be16_to_cpu(left->bb_numrecs) - 1 >=
-		    XFS_BMAP_BLOCK_IMINRECS(level, cur)) {
-			if ((error = xfs_bmbt_rshift(tcur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_BMAP_BLOCK_IMINRECS(level, tcur));
-				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-				tcur = NULL;
-				if (level == 0)
-					cur->bc_ptrs[0]++;
-				XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-				*stat = 1;
-				return 0;
-			}
-		}
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	}
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	tcur = NULL;
-	mp = cur->bc_mp;
-	ASSERT(bno != NULLFSBLOCK);
-	if (lbno != NULLFSBLOCK &&
-	    lrecs + be16_to_cpu(block->bb_numrecs) <= XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		rbno = bno;
-		right = block;
-		rbp = bp;
-		if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, lbno, 0, &lbp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-		if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-	} else if (rbno != NULLFSBLOCK &&
-		   rrecs + be16_to_cpu(block->bb_numrecs) <=
-		   XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		lbno = bno;
-		left = block;
-		lbp = bp;
-		if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, rbno, 0, &rbp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-		if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	} else {
-		if (level > 0 && (error = xfs_bmbt_decrement(cur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	numlrecs = be16_to_cpu(left->bb_numrecs);
-	numrrecs = be16_to_cpu(right->bb_numrecs);
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, numlrecs + 1, cur);
-		lpp = XFS_BMAP_PTR_IADDR(left, numlrecs + 1, cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < numrrecs; i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, rpp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				goto error0;
-			}
-		}
-#endif
-		memcpy(lkp, rkp, numrrecs * sizeof(*lkp));
-		memcpy(lpp, rpp, numrrecs * sizeof(*lpp));
-		xfs_bmbt_log_keys(cur, lbp, numlrecs + 1, numlrecs + numrrecs);
-		xfs_bmbt_log_ptrs(cur, lbp, numlrecs + 1, numlrecs + numrrecs);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, numlrecs + 1, cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		memcpy(lrp, rrp, numrrecs * sizeof(*lrp));
-		xfs_bmbt_log_recs(cur, lbp, numlrecs + 1, numlrecs + numrrecs);
-	}
-	be16_add_cpu(&left->bb_numrecs, numrrecs);
-	left->bb_rightsib = right->bb_rightsib;
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_RIGHTSIB | XFS_BB_NUMRECS);
-	if (be64_to_cpu(left->bb_rightsib) != NULLDFSBNO) {
-		if ((error = xfs_btree_read_bufl(mp, cur->bc_tp,
-				be64_to_cpu(left->bb_rightsib),
-				0, &rrbp, XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		rrblock = XFS_BUF_TO_BMBT_BLOCK(rrbp);
-		if ((error = xfs_btree_check_lblock(cur, rrblock, level, rrbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			goto error0;
-		}
-		rrblock->bb_leftsib = cpu_to_be64(lbno);
-		xfs_bmbt_log_block(cur, rrbp, XFS_BB_LEFTSIB);
-	}
-	xfs_bmap_add_free(XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(rbp)), 1,
-		cur->bc_private.b.flist, mp);
-	cur->bc_private.b.ip->i_d.di_nblocks--;
-	xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
-	XFS_TRANS_MOD_DQUOT_BYINO(mp, cur->bc_tp, cur->bc_private.b.ip,
-			XFS_TRANS_DQ_BCOUNT, -1L);
-	xfs_trans_binval(cur->bc_tp, rbp);
-	if (bp != lbp) {
-		cur->bc_bufs[level] = lbp;
-		cur->bc_ptrs[level] += lrecs;
-		cur->bc_ra[level] = 0;
-	} else if ((error = xfs_bmbt_increment(cur, level + 1, &i))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		goto error0;
-	}
-	if (level > 0)
-		cur->bc_ptrs[level]--;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 2;
-	return 0;
-
-error0:
-	if (tcur)
-		xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
-}
-
-/*
- * Insert one record/level.  Return information to the caller
- * allowing the next level up to proceed if necessary.
- */
-STATIC int					/* error */
-xfs_bmbt_insrec(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	xfs_fsblock_t		*bnop,
-	xfs_bmbt_rec_t		*recp,
-	xfs_btree_cur_t		**curp,
-	int			*stat)		/* no-go/done/continue */
-{
-	xfs_bmbt_block_t	*block;		/* bmap btree block */
-	xfs_buf_t		*bp;		/* buffer for block */
-	int			error;		/* error return value */
-	int			i;		/* loop index */
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_bmbt_key_t		*kp=NULL;	/* pointer to bmap btree key */
-	int			logflags;	/* inode logging flags */
-	xfs_fsblock_t		nbno;		/* new block number */
-	struct xfs_btree_cur	*ncur;		/* new btree cursor */
-	__uint64_t		startoff;	/* new btree key value */
-	xfs_bmbt_rec_t		nrec;		/* new record count */
-	int			optr;		/* old key/record index */
-	xfs_bmbt_ptr_t		*pp;		/* pointer to bmap block addr */
-	int			ptr;		/* key/record index */
-	xfs_bmbt_rec_t		*rp=NULL;	/* pointer to bmap btree rec */
-	int			numrecs;
-
-	ASSERT(level < cur->bc_nlevels);
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGIFR(cur, level, *bnop, recp);
-	ncur = NULL;
-	key.br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(recp));
-	optr = ptr = cur->bc_ptrs[level];
-	if (ptr == 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	XFS_STATS_INC(xs_bmbt_insrec);
-	block = xfs_bmbt_get_block(cur, level, &bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (ptr <= numrecs) {
-		if (level == 0) {
-			rp = XFS_BMAP_REC_IADDR(block, ptr, cur);
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, recp, rp);
-		} else {
-			kp = XFS_BMAP_KEY_IADDR(block, ptr, cur);
-			xfs_btree_check_key(XFS_BTNUM_BMAP, &key, kp);
-		}
-	}
-#endif
-	nbno = NULLFSBLOCK;
-	if (numrecs == XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		if (numrecs < XFS_BMAP_BLOCK_DMAXRECS(level, cur)) {
-			/*
-			 * A root block, that can be made bigger.
-			 */
-			xfs_iroot_realloc(cur->bc_private.b.ip, 1,
-				cur->bc_private.b.whichfork);
-			block = xfs_bmbt_get_block(cur, level, &bp);
-		} else if (level == cur->bc_nlevels - 1) {
-			if ((error = xfs_bmbt_newroot(cur, &logflags, stat)) ||
-			    *stat == 0) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-			xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
-				logflags);
-			block = xfs_bmbt_get_block(cur, level, &bp);
-		} else {
-			if ((error = xfs_bmbt_rshift(cur, level, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-			if (i) {
-				/* nothing */
-			} else {
-				if ((error = xfs_bmbt_lshift(cur, level, &i))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-				if (i) {
-					optr = ptr = cur->bc_ptrs[level];
-				} else {
-					if ((error = xfs_bmbt_split(cur, level,
-							&nbno, &startoff, &ncur,
-							&i))) {
-						XFS_BMBT_TRACE_CURSOR(cur,
-							ERROR);
-						return error;
-					}
-					if (i) {
-						block = xfs_bmbt_get_block(
-							    cur, level, &bp);
-#ifdef DEBUG
-						if ((error =
-						    xfs_btree_check_lblock(cur,
-							    block, level, bp))) {
-							XFS_BMBT_TRACE_CURSOR(
-								cur, ERROR);
-							return error;
-						}
-#endif
-						ptr = cur->bc_ptrs[level];
-						xfs_bmbt_disk_set_allf(&nrec,
-							startoff, 0, 0,
-							XFS_EXT_NORM);
-					} else {
-						XFS_BMBT_TRACE_CURSOR(cur,
-							EXIT);
-						*stat = 0;
-						return 0;
-					}
-				}
-			}
-		}
-	}
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (level > 0) {
-		kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-		pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = numrecs; i >= ptr; i--) {
-			if ((error = xfs_btree_check_lptr_disk(cur, pp[i - 1],
-					level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memmove(&kp[ptr], &kp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*kp));
-		memmove(&pp[ptr], &pp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*pp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr(cur, *bnop, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		kp[ptr - 1] = key;
-		pp[ptr - 1] = cpu_to_be64(*bnop);
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_bmbt_log_keys(cur, bp, ptr, numrecs);
-		xfs_bmbt_log_ptrs(cur, bp, ptr, numrecs);
-	} else {
-		rp = XFS_BMAP_REC_IADDR(block, 1, cur);
-		memmove(&rp[ptr], &rp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*rp));
-		rp[ptr - 1] = *recp;
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_bmbt_log_recs(cur, bp, ptr, numrecs);
-	}
-	xfs_bmbt_log_block(cur, bp, XFS_BB_NUMRECS);
-#ifdef DEBUG
-	if (ptr < numrecs) {
-		if (level == 0)
-			xfs_btree_check_rec(XFS_BTNUM_BMAP, rp + ptr - 1,
-				rp + ptr);
-		else
-			xfs_btree_check_key(XFS_BTNUM_BMAP, kp + ptr - 1,
-				kp + ptr);
-	}
-#endif
-	if (optr == 1 && (error = xfs_bmbt_updkey(cur, &key, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	*bnop = nbno;
-	if (nbno != NULLFSBLOCK) {
-		*recp = nrec;
-		*curp = ncur;
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-STATIC int
-xfs_bmbt_killroot(
-	xfs_btree_cur_t		*cur)
-{
-	xfs_bmbt_block_t	*block;
-	xfs_bmbt_block_t	*cblock;
-	xfs_buf_t		*cbp;
-	xfs_bmbt_key_t		*ckp;
-	xfs_bmbt_ptr_t		*cpp;
-#ifdef DEBUG
-	int			error;
-#endif
-	int			i;
-	xfs_bmbt_key_t		*kp;
-	xfs_inode_t		*ip;
-	xfs_ifork_t		*ifp;
-	int			level;
-	xfs_bmbt_ptr_t		*pp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	level = cur->bc_nlevels - 1;
-	ASSERT(level >= 1);
-	/*
-	 * Don't deal with the root block needs to be a leaf case.
-	 * We're just going to turn the thing back into extents anyway.
-	 */
-	if (level == 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	block = xfs_bmbt_get_block(cur, level, &cbp);
-	/*
-	 * Give up if the root has multiple children.
-	 */
-	if (be16_to_cpu(block->bb_numrecs) != 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	/*
-	 * Only do this if the next level will fit.
-	 * Then the data must be copied up to the inode,
-	 * instead of freeing the root you free the next level.
-	 */
-	cbp = cur->bc_bufs[level - 1];
-	cblock = XFS_BUF_TO_BMBT_BLOCK(cbp);
-	if (be16_to_cpu(cblock->bb_numrecs) > XFS_BMAP_BLOCK_DMAXRECS(level, cur)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	ASSERT(be64_to_cpu(cblock->bb_leftsib) == NULLDFSBNO);
-	ASSERT(be64_to_cpu(cblock->bb_rightsib) == NULLDFSBNO);
-	ip = cur->bc_private.b.ip;
-	ifp = XFS_IFORK_PTR(ip, cur->bc_private.b.whichfork);
-	ASSERT(XFS_BMAP_BLOCK_IMAXRECS(level, cur) ==
-	       XFS_BMAP_BROOT_MAXRECS(ifp->if_broot_bytes));
-	i = (int)(be16_to_cpu(cblock->bb_numrecs) - XFS_BMAP_BLOCK_IMAXRECS(level, cur));
-	if (i) {
-		xfs_iroot_realloc(ip, i, cur->bc_private.b.whichfork);
-		block = ifp->if_broot;
-	}
-	be16_add_cpu(&block->bb_numrecs, i);
-	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
-	kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-	ckp = XFS_BMAP_KEY_IADDR(cblock, 1, cur);
-	memcpy(kp, ckp, be16_to_cpu(block->bb_numrecs) * sizeof(*kp));
-	pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-	cpp = XFS_BMAP_PTR_IADDR(cblock, 1, cur);
-#ifdef DEBUG
-	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
-		if ((error = xfs_btree_check_lptr_disk(cur, cpp[i], level - 1))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-	}
-#endif
-	memcpy(pp, cpp, be16_to_cpu(block->bb_numrecs) * sizeof(*pp));
-	xfs_bmap_add_free(XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(cbp)), 1,
-			cur->bc_private.b.flist, cur->bc_mp);
-	ip->i_d.di_nblocks--;
-	XFS_TRANS_MOD_DQUOT_BYINO(cur->bc_mp, cur->bc_tp, ip,
-			XFS_TRANS_DQ_BCOUNT, -1L);
-	xfs_trans_binval(cur->bc_tp, cbp);
-	cur->bc_bufs[level - 1] = NULL;
-	be16_add_cpu(&block->bb_level, -1);
-	xfs_trans_log_inode(cur->bc_tp, ip,
-		XFS_ILOG_CORE | XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	cur->bc_nlevels--;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	return 0;
-}
-
-/*
- * Log key values from the btree block.
- */
-STATIC void
-xfs_bmbt_log_keys(
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*bp,
-	int		kfirst,
-	int		klast)
-{
-	xfs_trans_t	*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBII(cur, bp, kfirst, klast);
-	tp = cur->bc_tp;
-	if (bp) {
-		xfs_bmbt_block_t	*block;
-		int			first;
-		xfs_bmbt_key_t		*kp;
-		int			last;
-
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		kp = XFS_BMAP_KEY_DADDR(block, 1, cur);
-		first = (int)((xfs_caddr_t)&kp[kfirst - 1] - (xfs_caddr_t)block);
-		last = (int)(((xfs_caddr_t)&kp[klast] - 1) - (xfs_caddr_t)block);
-		xfs_trans_log_buf(tp, bp, first, last);
-	} else {
-		xfs_inode_t		 *ip;
-
-		ip = cur->bc_private.b.ip;
-		xfs_trans_log_inode(tp, ip,
-			XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-/*
- * Log pointer values from the btree block.
- */
-STATIC void
-xfs_bmbt_log_ptrs(
-	xfs_btree_cur_t	*cur,
-	xfs_buf_t	*bp,
-	int		pfirst,
-	int		plast)
-{
-	xfs_trans_t	*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBII(cur, bp, pfirst, plast);
-	tp = cur->bc_tp;
-	if (bp) {
-		xfs_bmbt_block_t	*block;
-		int			first;
-		int			last;
-		xfs_bmbt_ptr_t		*pp;
-
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		pp = XFS_BMAP_PTR_DADDR(block, 1, cur);
-		first = (int)((xfs_caddr_t)&pp[pfirst - 1] - (xfs_caddr_t)block);
-		last = (int)(((xfs_caddr_t)&pp[plast] - 1) - (xfs_caddr_t)block);
-		xfs_trans_log_buf(tp, bp, first, last);
-	} else {
-		xfs_inode_t		*ip;
-
-		ip = cur->bc_private.b.ip;
-		xfs_trans_log_inode(tp, ip,
-			XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-/*
- * Lookup the record.  The cursor is made to point to it, based on dir.
- */
-STATIC int				/* error */
-xfs_bmbt_lookup(
-	xfs_btree_cur_t		*cur,
-	xfs_lookup_t		dir,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block=NULL;
-	xfs_buf_t		*bp;
-	xfs_daddr_t		d;
-	xfs_sfiloff_t		diff;
-	int			error;		/* error return value */
-	xfs_fsblock_t		fsbno=0;
-	int			high;
-	int			i;
-	int			keyno=0;
-	xfs_bmbt_key_t		*kkbase=NULL;
-	xfs_bmbt_key_t		*kkp;
-	xfs_bmbt_rec_t		*krbase=NULL;
-	xfs_bmbt_rec_t		*krp;
-	int			level;
-	int			low;
-	xfs_mount_t		*mp;
-	xfs_bmbt_ptr_t		*pp;
-	xfs_bmbt_irec_t		*rp;
-	xfs_fileoff_t		startoff;
-	xfs_trans_t		*tp;
-
-	XFS_STATS_INC(xs_bmbt_lookup);
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, (int)dir);
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
-	rp = &cur->bc_rec.b;
-	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
-		if (level < cur->bc_nlevels - 1) {
-			d = XFS_FSB_TO_DADDR(mp, fsbno);
-			bp = cur->bc_bufs[level];
-			if (bp && XFS_BUF_ADDR(bp) != d)
-				bp = NULL;
-			if (!bp) {
-				if ((error = xfs_btree_read_bufl(mp, tp, fsbno,
-						0, &bp, XFS_BMAP_BTREE_REF))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-				xfs_btree_setbuf(cur, level, bp);
-				block = XFS_BUF_TO_BMBT_BLOCK(bp);
-				if ((error = xfs_btree_check_lblock(cur, block,
-						level, bp))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-			} else
-				block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		} else
-			block = xfs_bmbt_get_block(cur, level, &bp);
-		if (diff == 0)
-			keyno = 1;
-		else {
-			if (level > 0)
-				kkbase = XFS_BMAP_KEY_IADDR(block, 1, cur);
-			else
-				krbase = XFS_BMAP_REC_IADDR(block, 1, cur);
-			low = 1;
-			if (!(high = be16_to_cpu(block->bb_numrecs))) {
-				ASSERT(level == 0);
-				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
-				XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-				*stat = 0;
-				return 0;
-			}
-			while (low <= high) {
-				XFS_STATS_INC(xs_bmbt_compare);
-				keyno = (low + high) >> 1;
-				if (level > 0) {
-					kkp = kkbase + keyno - 1;
-					startoff = be64_to_cpu(kkp->br_startoff);
-				} else {
-					krp = krbase + keyno - 1;
-					startoff = xfs_bmbt_disk_get_startoff(krp);
-				}
-				diff = (xfs_sfiloff_t)
-						(startoff - rp->br_startoff);
-				if (diff < 0)
-					low = keyno + 1;
-				else if (diff > 0)
-					high = keyno - 1;
-				else
-					break;
-			}
-		}
-		if (level > 0) {
-			if (diff > 0 && --keyno < 1)
-				keyno = 1;
-			pp = XFS_BMAP_PTR_IADDR(block, keyno, cur);
-			fsbno = be64_to_cpu(*pp);
-#ifdef DEBUG
-			if ((error = xfs_btree_check_lptr(cur, fsbno, level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-#endif
-			cur->bc_ptrs[level] = keyno;
-		}
-	}
-	if (dir != XFS_LOOKUP_LE && diff < 0) {
-		keyno++;
-		/*
-		 * If ge search and we went off the end of the block, but it's
-		 * not the last block, we're in the wrong block.
-		 */
-		if (dir == XFS_LOOKUP_GE && keyno > be16_to_cpu(block->bb_numrecs) &&
-		    be64_to_cpu(block->bb_rightsib) != NULLDFSBNO) {
-			cur->bc_ptrs[0] = keyno;
-			if ((error = xfs_bmbt_increment(cur, 0, &i))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-			XFS_WANT_CORRUPTED_RETURN(i == 1);
-			XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-			*stat = 1;
-			return 0;
-		}
-	}
-	else if (dir == XFS_LOOKUP_LE && diff > 0)
-		keyno--;
-	cur->bc_ptrs[0] = keyno;
-	if (keyno == 0 || keyno > be16_to_cpu(block->bb_numrecs)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-	} else {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = ((dir != XFS_LOOKUP_EQ) || (diff == 0));
-	}
-	return 0;
-}
-
-/*
- * Move 1 record left from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int					/* error */
-xfs_bmbt_lshift(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	int			error;		/* error return value */
-#ifdef DEBUG
-	int			i;		/* loop counter */
-#endif
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp=NULL;	/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	int			lrecs;		/* left record count */
-	xfs_bmbt_rec_t		*lrp=NULL;	/* left record pointer */
-	xfs_mount_t		*mp;		/* file system mount point */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp=NULL;	/* right btree key */
-	xfs_bmbt_ptr_t		*rpp=NULL;	/* right address pointer */
-	xfs_bmbt_rec_t		*rrp=NULL;	/* right record pointer */
-	int			rrecs;		/* right record count */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	if (level == cur->bc_nlevels - 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	rbp = cur->bc_bufs[level];
-	right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (be64_to_cpu(right->bb_leftsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	if (cur->bc_ptrs[level] <= 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	mp = cur->bc_mp;
-	if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, be64_to_cpu(right->bb_leftsib), 0,
-			&lbp, XFS_BMAP_BTREE_REF))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-	if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (be16_to_cpu(left->bb_numrecs) == XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	lrecs = be16_to_cpu(left->bb_numrecs) + 1;
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, lrecs, cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		*lkp = *rkp;
-		xfs_bmbt_log_keys(cur, lbp, lrecs, lrecs);
-		lpp = XFS_BMAP_PTR_IADDR(left, lrecs, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr_disk(cur, *rpp, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		*lpp = *rpp;
-		xfs_bmbt_log_ptrs(cur, lbp, lrecs, lrecs);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, lrecs, cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		*lrp = *rrp;
-		xfs_bmbt_log_recs(cur, lbp, lrecs, lrecs);
-	}
-	left->bb_numrecs = cpu_to_be16(lrecs);
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_NUMRECS);
-#ifdef DEBUG
-	if (level > 0)
-		xfs_btree_check_key(XFS_BTNUM_BMAP, lkp - 1, lkp);
-	else
-		xfs_btree_check_rec(XFS_BTNUM_BMAP, lrp - 1, lrp);
-#endif
-	rrecs = be16_to_cpu(right->bb_numrecs) - 1;
-	right->bb_numrecs = cpu_to_be16(rrecs);
-	xfs_bmbt_log_block(cur, rbp, XFS_BB_NUMRECS);
-	if (level > 0) {
-#ifdef DEBUG
-		for (i = 0; i < rrecs; i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, rpp[i + 1],
-					level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memmove(rkp, rkp + 1, rrecs * sizeof(*rkp));
-		memmove(rpp, rpp + 1, rrecs * sizeof(*rpp));
-		xfs_bmbt_log_keys(cur, rbp, 1, rrecs);
-		xfs_bmbt_log_ptrs(cur, rbp, 1, rrecs);
-	} else {
-		memmove(rrp, rrp + 1, rrecs * sizeof(*rrp));
-		xfs_bmbt_log_recs(cur, rbp, 1, rrecs);
-		key.br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(rrp));
-		rkp = &key;
-	}
-	if ((error = xfs_bmbt_updkey(cur, rkp, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	cur->bc_ptrs[level]--;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Move 1 record right from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int					/* error */
-xfs_bmbt_rshift(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	int			error;		/* error return value */
-	int			i;		/* loop counter */
-	xfs_bmbt_key_t		key;		/* bmap btree key */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp;		/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	xfs_bmbt_rec_t		*lrp;		/* left record pointer */
-	xfs_mount_t		*mp;		/* file system mount point */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp;		/* right btree key */
-	xfs_bmbt_ptr_t		*rpp;		/* right address pointer */
-	xfs_bmbt_rec_t		*rrp=NULL;	/* right record pointer */
-	struct xfs_btree_cur	*tcur;		/* temporary btree cursor */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	if (level == cur->bc_nlevels - 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, left, level, lbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (be64_to_cpu(left->bb_rightsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	if (cur->bc_ptrs[level] >= be16_to_cpu(left->bb_numrecs)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	mp = cur->bc_mp;
-	if ((error = xfs_btree_read_bufl(mp, cur->bc_tp, be64_to_cpu(left->bb_rightsib), 0,
-			&rbp, XFS_BMAP_BTREE_REF))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-	if ((error = xfs_btree_check_lblock(cur, right, level, rbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (be16_to_cpu(right->bb_numrecs) == XFS_BMAP_BLOCK_IMAXRECS(level, cur)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		lpp = XFS_BMAP_PTR_IADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = be16_to_cpu(right->bb_numrecs) - 1; i >= 0; i--) {
-			if ((error = xfs_btree_check_lptr_disk(cur, rpp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memmove(rkp + 1, rkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp + 1, rpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr_disk(cur, *lpp, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		*rkp = *lkp;
-		*rpp = *lpp;
-		xfs_bmbt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_bmbt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		memmove(rrp + 1, rrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		*rrp = *lrp;
-		xfs_bmbt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		key.br_startoff = cpu_to_be64(xfs_bmbt_disk_get_startoff(rrp));
-		rkp = &key;
-	}
-	be16_add_cpu(&left->bb_numrecs, -1);
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, 1);
-#ifdef DEBUG
-	if (level > 0)
-		xfs_btree_check_key(XFS_BTNUM_BMAP, rkp, rkp + 1);
-	else
-		xfs_btree_check_rec(XFS_BTNUM_BMAP, rrp, rrp + 1);
-#endif
-	xfs_bmbt_log_block(cur, rbp, XFS_BB_NUMRECS);
-	if ((error = xfs_btree_dup_cursor(cur, &tcur))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	i = xfs_btree_lastrec(tcur, level);
-	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_bmbt_increment(tcur, level, &i))) {
-		XFS_BMBT_TRACE_CURSOR(tcur, ERROR);
-		goto error1;
-	}
-	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_bmbt_updkey(tcur, rkp, level + 1))) {
-		XFS_BMBT_TRACE_CURSOR(tcur, ERROR);
-		goto error1;
-	}
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-error0:
-	XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-error1:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
-}
-
 /*
  * Determine the extent state.
  */
@@ -1453,229 +60,15 @@ xfs_extent_state(
 	return XFS_EXT_NORM;
 }
 
-
-/*
- * Split cur/level block in half.
- * Return new block number and its first record (to be inserted into parent).
- */
-STATIC int					/* error */
-xfs_bmbt_split(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	xfs_fsblock_t		*bnop,
-	__uint64_t		*startoff,
-	xfs_btree_cur_t		**curp,
-	int			*stat)		/* success/failure */
-{
-	xfs_alloc_arg_t		args;		/* block allocation args */
-	int			error;		/* error return value */
-	int			i;		/* loop counter */
-	xfs_fsblock_t		lbno;		/* left sibling block number */
-	xfs_buf_t		*lbp;		/* left buffer pointer */
-	xfs_bmbt_block_t	*left;		/* left btree block */
-	xfs_bmbt_key_t		*lkp;		/* left btree key */
-	xfs_bmbt_ptr_t		*lpp;		/* left address pointer */
-	xfs_bmbt_rec_t		*lrp;		/* left record pointer */
-	xfs_buf_t		*rbp;		/* right buffer pointer */
-	xfs_bmbt_block_t	*right;		/* right btree block */
-	xfs_bmbt_key_t		*rkp;		/* right btree key */
-	xfs_bmbt_ptr_t		*rpp;		/* right address pointer */
-	xfs_bmbt_block_t	*rrblock;	/* right-right btree block */
-	xfs_buf_t		*rrbp;		/* right-right buffer pointer */
-	xfs_bmbt_rec_t		*rrp;		/* right record pointer */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGIFK(cur, level, *bnop, *startoff);
-	args.tp = cur->bc_tp;
-	args.mp = cur->bc_mp;
-	lbp = cur->bc_bufs[level];
-	lbno = XFS_DADDR_TO_FSB(args.mp, XFS_BUF_ADDR(lbp));
-	left = XFS_BUF_TO_BMBT_BLOCK(lbp);
-	args.fsbno = cur->bc_private.b.firstblock;
-	args.firstblock = args.fsbno;
-	args.minleft = 0;
-	if (args.fsbno == NULLFSBLOCK) {
-		args.fsbno = lbno;
-		args.type = XFS_ALLOCTYPE_START_BNO;
-		/*
-		 * Make sure there is sufficient room left in the AG to
-		 * complete a full tree split for an extent insert.  If
-		 * we are converting the middle part of an extent then
-		 * we may need space for two tree splits.
-		 *
-		 * We are relying on the caller to make the correct block
-		 * reservation for this operation to succeed.  If the
-		 * reservation amount is insufficient then we may fail a
-		 * block allocation here and corrupt the filesystem.
-		 */
-		args.minleft = xfs_trans_get_block_res(args.tp);
-	} else if (cur->bc_private.b.flist->xbf_low)
-		args.type = XFS_ALLOCTYPE_START_BNO;
-	else
-		args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	args.mod = args.alignment = args.total = args.isfl =
-		args.userdata = args.minalignslop = 0;
-	args.minlen = args.maxlen = args.prod = 1;
-	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
-	if (!args.wasdel && xfs_trans_get_block_res(args.tp) == 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return XFS_ERROR(ENOSPC);
-	}
-	if ((error = xfs_alloc_vextent(&args))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (args.fsbno == NULLFSBLOCK && args.minleft) {
-		/*
-		 * Could not find an AG with enough free space to satisfy
-		 * a full btree split.  Try again without minleft and if
-		 * successful activate the lowspace algorithm.
-		 */
-		args.fsbno = 0;
-		args.type = XFS_ALLOCTYPE_FIRST_AG;
-		args.minleft = 0;
-		if ((error = xfs_alloc_vextent(&args))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		cur->bc_private.b.flist->xbf_low = 1;
-	}
-	if (args.fsbno == NULLFSBLOCK) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	ASSERT(args.len == 1);
-	cur->bc_private.b.firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
-	cur->bc_private.b.ip->i_d.di_nblocks++;
-	xfs_trans_log_inode(args.tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
-	XFS_TRANS_MOD_DQUOT_BYINO(args.mp, args.tp, cur->bc_private.b.ip,
-			XFS_TRANS_DQ_BCOUNT, 1L);
-	rbp = xfs_btree_get_bufl(args.mp, args.tp, args.fsbno, 0);
-	right = XFS_BUF_TO_BMBT_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, left, level, rbp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	right->bb_magic = cpu_to_be32(XFS_BMAP_MAGIC);
-	right->bb_level = left->bb_level;
-	right->bb_numrecs = cpu_to_be16(be16_to_cpu(left->bb_numrecs) / 2);
-	if ((be16_to_cpu(left->bb_numrecs) & 1) &&
-	    cur->bc_ptrs[level] <= be16_to_cpu(right->bb_numrecs) + 1)
-		be16_add_cpu(&right->bb_numrecs, 1);
-	i = be16_to_cpu(left->bb_numrecs) - be16_to_cpu(right->bb_numrecs) + 1;
-	if (level > 0) {
-		lkp = XFS_BMAP_KEY_IADDR(left, i, cur);
-		lpp = XFS_BMAP_PTR_IADDR(left, i, cur);
-		rkp = XFS_BMAP_KEY_IADDR(right, 1, cur);
-		rpp = XFS_BMAP_PTR_IADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_lptr_disk(cur, lpp[i], level))) {
-				XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-				return error;
-			}
-		}
-#endif
-		memcpy(rkp, lkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memcpy(rpp, lpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_bmbt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_bmbt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*startoff = be64_to_cpu(rkp->br_startoff);
-	} else {
-		lrp = XFS_BMAP_REC_IADDR(left, i, cur);
-		rrp = XFS_BMAP_REC_IADDR(right, 1, cur);
-		memcpy(rrp, lrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_bmbt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*startoff = xfs_bmbt_disk_get_startoff(rrp);
-	}
-	be16_add_cpu(&left->bb_numrecs, -(be16_to_cpu(right->bb_numrecs)));
-	right->bb_rightsib = left->bb_rightsib;
-	left->bb_rightsib = cpu_to_be64(args.fsbno);
-	right->bb_leftsib = cpu_to_be64(lbno);
-	xfs_bmbt_log_block(cur, rbp, XFS_BB_ALL_BITS);
-	xfs_bmbt_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	if (be64_to_cpu(right->bb_rightsib) != NULLDFSBNO) {
-		if ((error = xfs_btree_read_bufl(args.mp, args.tp,
-				be64_to_cpu(right->bb_rightsib), 0, &rrbp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		rrblock = XFS_BUF_TO_BMBT_BLOCK(rrbp);
-		if ((error = xfs_btree_check_lblock(cur, rrblock, level, rrbp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		rrblock->bb_leftsib = cpu_to_be64(args.fsbno);
-		xfs_bmbt_log_block(cur, rrbp, XFS_BB_LEFTSIB);
-	}
-	if (cur->bc_ptrs[level] > be16_to_cpu(left->bb_numrecs) + 1) {
-		xfs_btree_setbuf(cur, level, rbp);
-		cur->bc_ptrs[level] -= be16_to_cpu(left->bb_numrecs);
-	}
-	if (level + 1 < cur->bc_nlevels) {
-		if ((error = xfs_btree_dup_cursor(cur, curp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		(*curp)->bc_ptrs[level + 1]++;
-	}
-	*bnop = args.fsbno;
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-
-/*
- * Update keys for the record.
- */
-STATIC int
-xfs_bmbt_updkey(
-	xfs_btree_cur_t		*cur,
-	xfs_bmbt_key_t		*keyp,	/* on-disk format */
-	int			level)
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-#ifdef DEBUG
-	int			error;
-#endif
-	xfs_bmbt_key_t		*kp;
-	int			ptr;
-
-	ASSERT(level >= 1);
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGIK(cur, level, keyp);
-	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
-		block = xfs_bmbt_get_block(cur, level, &bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		ptr = cur->bc_ptrs[level];
-		kp = XFS_BMAP_KEY_IADDR(block, ptr, cur);
-		*kp = *keyp;
-		xfs_bmbt_log_keys(cur, bp, ptr, ptr);
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	return 0;
-}
-
 /*
  * Convert on-disk form of btree root to in-memory form.
  */
 void
 xfs_bmdr_to_bmbt(
+	struct xfs_mount	*mp,
 	xfs_bmdr_block_t	*dblock,
 	int			dblocklen,
-	xfs_bmbt_block_t	*rblock,
+	struct xfs_btree_block	*rblock,
 	int			rblocklen)
 {
 	int			dmxr;
@@ -1688,129 +81,19 @@ xfs_bmdr_to_bmbt(
 	rblock->bb_level = dblock->bb_level;
 	ASSERT(be16_to_cpu(rblock->bb_level) > 0);
 	rblock->bb_numrecs = dblock->bb_numrecs;
-	rblock->bb_leftsib = cpu_to_be64(NULLDFSBNO);
-	rblock->bb_rightsib = cpu_to_be64(NULLDFSBNO);
-	dmxr = (int)XFS_BTREE_BLOCK_MAXRECS(dblocklen, xfs_bmdr, 0);
-	fkp = XFS_BTREE_KEY_ADDR(xfs_bmdr, dblock, 1);
-	tkp = XFS_BMAP_BROOT_KEY_ADDR(rblock, 1, rblocklen);
-	fpp = XFS_BTREE_PTR_ADDR(xfs_bmdr, dblock, 1, dmxr);
-	tpp = XFS_BMAP_BROOT_PTR_ADDR(rblock, 1, rblocklen);
+	rblock->bb_u.l.bb_leftsib = cpu_to_be64(NULLDFSBNO);
+	rblock->bb_u.l.bb_rightsib = cpu_to_be64(NULLDFSBNO);
+	dmxr = xfs_bmdr_maxrecs(mp, dblocklen, 0);
+	fkp = XFS_BMDR_KEY_ADDR(dblock, 1);
+	tkp = XFS_BMBT_KEY_ADDR(mp, rblock, 1);
+	fpp = XFS_BMDR_PTR_ADDR(dblock, 1, dmxr);
+	tpp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, rblocklen);
 	dmxr = be16_to_cpu(dblock->bb_numrecs);
 	memcpy(tkp, fkp, sizeof(*fkp) * dmxr);
 	memcpy(tpp, fpp, sizeof(*fpp) * dmxr);
 }
 
 /*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int						/* error */
-xfs_bmbt_decrement(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-	int			error;		/* error return value */
-	xfs_fsblock_t		fsbno;
-	int			lev;
-	xfs_mount_t		*mp;
-	xfs_trans_t		*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	ASSERT(level < cur->bc_nlevels);
-	if (level < cur->bc_nlevels - 1)
-		xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
-	if (--cur->bc_ptrs[level] > 0) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	block = xfs_bmbt_get_block(cur, level, &bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (be64_to_cpu(block->bb_leftsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		if (--cur->bc_ptrs[lev] > 0)
-			break;
-		if (lev < cur->bc_nlevels - 1)
-			xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
-	}
-	if (lev == cur->bc_nlevels) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
-	for (block = xfs_bmbt_get_block(cur, lev, &bp); lev > level; ) {
-		fsbno = be64_to_cpu(*XFS_BMAP_PTR_IADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufl(mp, tp, fsbno, 0, &bp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		if ((error = xfs_btree_check_lblock(cur, block, lev, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		cur->bc_ptrs[lev] = be16_to_cpu(block->bb_numrecs);
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Delete the record pointed to by cur.
- */
-int					/* error */
-xfs_bmbt_delete(
-	xfs_btree_cur_t	*cur,
-	int		*stat)		/* success/failure */
-{
-	int		error;		/* error return value */
-	int		i;
-	int		level;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	for (level = 0, i = 2; i == 2; level++) {
-		if ((error = xfs_bmbt_delrec(cur, level, &i))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-	}
-	if (i == 0) {
-		for (level = 1; level < cur->bc_nlevels; level++) {
-			if (cur->bc_ptrs[level] == 0) {
-				if ((error = xfs_bmbt_decrement(cur, level,
-						&i))) {
-					XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-					return error;
-				}
-				break;
-			}
-		}
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = i;
-	return 0;
-}
-
-/*
  * Convert a compressed bmap extent record to an uncompressed form.
  * This code must be in sync with the routines xfs_bmbt_get_startoff,
  * xfs_bmbt_get_startblock, xfs_bmbt_get_blockcount and xfs_bmbt_get_state.
@@ -1864,31 +147,6 @@ xfs_bmbt_get_all(
 }
 
 /*
- * Get the block pointer for the given level of the cursor.
- * Fill in the buffer pointer, if applicable.
- */
-xfs_bmbt_block_t *
-xfs_bmbt_get_block(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	xfs_buf_t		**bpp)
-{
-	xfs_ifork_t		*ifp;
-	xfs_bmbt_block_t	*rval;
-
-	if (level < cur->bc_nlevels - 1) {
-		*bpp = cur->bc_bufs[level];
-		rval = XFS_BUF_TO_BMBT_BLOCK(*bpp);
-	} else {
-		*bpp = NULL;
-		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
-			cur->bc_private.b.whichfork);
-		rval = ifp->if_broot;
-	}
-	return rval;
-}
-
-/*
  * Extract the blockcount field from an in memory bmap extent record.
  */
 xfs_filblks_t
@@ -1950,7 +208,8 @@ xfs_bmbt_disk_get_all(
 	xfs_bmbt_rec_t	*r,
 	xfs_bmbt_irec_t *s)
 {
-	__xfs_bmbt_get_all(be64_to_cpu(r->l0), be64_to_cpu(r->l1), s);
+	__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
+				get_unaligned_be64(&r->l1), s);
 }
 
 /*
@@ -1974,348 +233,6 @@ xfs_bmbt_disk_get_startoff(
 		 XFS_MASK64LO(64 - BMBT_EXNTFLAG_BITLEN)) >> 9;
 }
 
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int						/* error */
-xfs_bmbt_increment(
-	xfs_btree_cur_t		*cur,
-	int			level,
-	int			*stat)		/* success/failure */
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-	int			error;		/* error return value */
-	xfs_fsblock_t		fsbno;
-	int			lev;
-	xfs_mount_t		*mp;
-	xfs_trans_t		*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGI(cur, level);
-	ASSERT(level < cur->bc_nlevels);
-	if (level < cur->bc_nlevels - 1)
-		xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
-	block = xfs_bmbt_get_block(cur, level, &bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, level, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	if (++cur->bc_ptrs[level] <= be16_to_cpu(block->bb_numrecs)) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 1;
-		return 0;
-	}
-	if (be64_to_cpu(block->bb_rightsib) == NULLDFSBNO) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		block = xfs_bmbt_get_block(cur, lev, &bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lblock(cur, block, lev, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		if (++cur->bc_ptrs[lev] <= be16_to_cpu(block->bb_numrecs))
-			break;
-		if (lev < cur->bc_nlevels - 1)
-			xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
-	}
-	if (lev == cur->bc_nlevels) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	tp = cur->bc_tp;
-	mp = cur->bc_mp;
-	for (block = xfs_bmbt_get_block(cur, lev, &bp); lev > level; ) {
-		fsbno = be64_to_cpu(*XFS_BMAP_PTR_IADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufl(mp, tp, fsbno, 0, &bp,
-				XFS_BMAP_BTREE_REF))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_BMBT_BLOCK(bp);
-		if ((error = xfs_btree_check_lblock(cur, block, lev, bp))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		cur->bc_ptrs[lev] = 1;
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = 1;
-	return 0;
-}
-
-/*
- * Insert the current record at the point referenced by cur.
- *
- * A multi-level split of the tree on insert will invalidate the original
- * cursor.  All callers of this function should assume that the cursor is
- * no longer valid and revalidate it.
- */
-int					/* error */
-xfs_bmbt_insert(
-	xfs_btree_cur_t	*cur,
-	int		*stat)		/* success/failure */
-{
-	int		error;		/* error return value */
-	int		i;
-	int		level;
-	xfs_fsblock_t	nbno;
-	xfs_btree_cur_t	*ncur;
-	xfs_bmbt_rec_t	nrec;
-	xfs_btree_cur_t	*pcur;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	level = 0;
-	nbno = NULLFSBLOCK;
-	xfs_bmbt_disk_set_all(&nrec, &cur->bc_rec.b);
-	ncur = NULL;
-	pcur = cur;
-	do {
-		if ((error = xfs_bmbt_insrec(pcur, level++, &nbno, &nrec, &ncur,
-				&i))) {
-			if (pcur != cur)
-				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if (pcur != cur && (ncur || nbno == NULLFSBLOCK)) {
-			cur->bc_nlevels = pcur->bc_nlevels;
-			cur->bc_private.b.allocated +=
-				pcur->bc_private.b.allocated;
-			pcur->bc_private.b.allocated = 0;
-			ASSERT((cur->bc_private.b.firstblock != NULLFSBLOCK) ||
-			       XFS_IS_REALTIME_INODE(cur->bc_private.b.ip));
-			cur->bc_private.b.firstblock =
-				pcur->bc_private.b.firstblock;
-			ASSERT(cur->bc_private.b.flist ==
-			       pcur->bc_private.b.flist);
-			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
-		}
-		if (ncur) {
-			pcur = ncur;
-			ncur = NULL;
-		}
-	} while (nbno != NULLFSBLOCK);
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*stat = i;
-	return 0;
-error0:
-	XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-	return error;
-}
-
-/*
- * Log fields from the btree block header.
- */
-void
-xfs_bmbt_log_block(
-	xfs_btree_cur_t		*cur,
-	xfs_buf_t		*bp,
-	int			fields)
-{
-	int			first;
-	int			last;
-	xfs_trans_t		*tp;
-	static const short	offsets[] = {
-		offsetof(xfs_bmbt_block_t, bb_magic),
-		offsetof(xfs_bmbt_block_t, bb_level),
-		offsetof(xfs_bmbt_block_t, bb_numrecs),
-		offsetof(xfs_bmbt_block_t, bb_leftsib),
-		offsetof(xfs_bmbt_block_t, bb_rightsib),
-		sizeof(xfs_bmbt_block_t)
-	};
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBI(cur, bp, fields);
-	tp = cur->bc_tp;
-	if (bp) {
-		xfs_btree_offsets(fields, offsets, XFS_BB_NUM_BITS, &first,
-				  &last);
-		xfs_trans_log_buf(tp, bp, first, last);
-	} else
-		xfs_trans_log_inode(tp, cur->bc_private.b.ip,
-			XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-/*
- * Log record values from the btree block.
- */
-void
-xfs_bmbt_log_recs(
-	xfs_btree_cur_t		*cur,
-	xfs_buf_t		*bp,
-	int			rfirst,
-	int			rlast)
-{
-	xfs_bmbt_block_t	*block;
-	int			first;
-	int			last;
-	xfs_bmbt_rec_t		*rp;
-	xfs_trans_t		*tp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGBII(cur, bp, rfirst, rlast);
-	ASSERT(bp);
-	tp = cur->bc_tp;
-	block = XFS_BUF_TO_BMBT_BLOCK(bp);
-	rp = XFS_BMAP_REC_DADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&rp[rfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&rp[rlast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(tp, bp, first, last);
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-}
-
-int					/* error */
-xfs_bmbt_lookup_eq(
-	xfs_btree_cur_t	*cur,
-	xfs_fileoff_t	off,
-	xfs_fsblock_t	bno,
-	xfs_filblks_t	len,
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.b.br_startoff = off;
-	cur->bc_rec.b.br_startblock = bno;
-	cur->bc_rec.b.br_blockcount = len;
-	return xfs_bmbt_lookup(cur, XFS_LOOKUP_EQ, stat);
-}
-
-int					/* error */
-xfs_bmbt_lookup_ge(
-	xfs_btree_cur_t	*cur,
-	xfs_fileoff_t	off,
-	xfs_fsblock_t	bno,
-	xfs_filblks_t	len,
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.b.br_startoff = off;
-	cur->bc_rec.b.br_startblock = bno;
-	cur->bc_rec.b.br_blockcount = len;
-	return xfs_bmbt_lookup(cur, XFS_LOOKUP_GE, stat);
-}
-
-/*
- * Give the bmap btree a new root block.  Copy the old broot contents
- * down into a real block and make the broot point to it.
- */
-int						/* error */
-xfs_bmbt_newroot(
-	xfs_btree_cur_t		*cur,		/* btree cursor */
-	int			*logflags,	/* logging flags for inode */
-	int			*stat)		/* return status - 0 fail */
-{
-	xfs_alloc_arg_t		args;		/* allocation arguments */
-	xfs_bmbt_block_t	*block;		/* bmap btree block */
-	xfs_buf_t		*bp;		/* buffer for block */
-	xfs_bmbt_block_t	*cblock;	/* child btree block */
-	xfs_bmbt_key_t		*ckp;		/* child key pointer */
-	xfs_bmbt_ptr_t		*cpp;		/* child ptr pointer */
-	int			error;		/* error return code */
-#ifdef DEBUG
-	int			i;		/* loop counter */
-#endif
-	xfs_bmbt_key_t		*kp;		/* pointer to bmap btree key */
-	int			level;		/* btree level */
-	xfs_bmbt_ptr_t		*pp;		/* pointer to bmap block addr */
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	level = cur->bc_nlevels - 1;
-	block = xfs_bmbt_get_block(cur, level, &bp);
-	/*
-	 * Copy the root into a real block.
-	 */
-	args.mp = cur->bc_mp;
-	pp = XFS_BMAP_PTR_IADDR(block, 1, cur);
-	args.tp = cur->bc_tp;
-	args.fsbno = cur->bc_private.b.firstblock;
-	args.mod = args.minleft = args.alignment = args.total = args.isfl =
-		args.userdata = args.minalignslop = 0;
-	args.minlen = args.maxlen = args.prod = 1;
-	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
-	args.firstblock = args.fsbno;
-	if (args.fsbno == NULLFSBLOCK) {
-#ifdef DEBUG
-		if ((error = xfs_btree_check_lptr_disk(cur, *pp, level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-#endif
-		args.fsbno = be64_to_cpu(*pp);
-		args.type = XFS_ALLOCTYPE_START_BNO;
-	} else if (cur->bc_private.b.flist->xbf_low)
-		args.type = XFS_ALLOCTYPE_START_BNO;
-	else
-		args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	if ((error = xfs_alloc_vextent(&args))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	if (args.fsbno == NULLFSBLOCK) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		*stat = 0;
-		return 0;
-	}
-	ASSERT(args.len == 1);
-	cur->bc_private.b.firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
-	cur->bc_private.b.ip->i_d.di_nblocks++;
-	XFS_TRANS_MOD_DQUOT_BYINO(args.mp, args.tp, cur->bc_private.b.ip,
-			  XFS_TRANS_DQ_BCOUNT, 1L);
-	bp = xfs_btree_get_bufl(args.mp, cur->bc_tp, args.fsbno, 0);
-	cblock = XFS_BUF_TO_BMBT_BLOCK(bp);
-	*cblock = *block;
-	be16_add_cpu(&block->bb_level, 1);
-	block->bb_numrecs = cpu_to_be16(1);
-	cur->bc_nlevels++;
-	cur->bc_ptrs[level + 1] = 1;
-	kp = XFS_BMAP_KEY_IADDR(block, 1, cur);
-	ckp = XFS_BMAP_KEY_IADDR(cblock, 1, cur);
-	memcpy(ckp, kp, be16_to_cpu(cblock->bb_numrecs) * sizeof(*kp));
-	cpp = XFS_BMAP_PTR_IADDR(cblock, 1, cur);
-#ifdef DEBUG
-	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
-		if ((error = xfs_btree_check_lptr_disk(cur, pp[i], level))) {
-			XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-			return error;
-		}
-	}
-#endif
-	memcpy(cpp, pp, be16_to_cpu(cblock->bb_numrecs) * sizeof(*pp));
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lptr(cur, args.fsbno, level))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	*pp = cpu_to_be64(args.fsbno);
-	xfs_iroot_realloc(cur->bc_private.b.ip, 1 - be16_to_cpu(cblock->bb_numrecs),
-		cur->bc_private.b.whichfork);
-	xfs_btree_setbuf(cur, level, bp);
-	/*
-	 * Do all this logging at the end so that
-	 * the root is at the right level.
-	 */
-	xfs_bmbt_log_block(cur, bp, XFS_BB_ALL_BITS);
-	xfs_bmbt_log_keys(cur, bp, 1, be16_to_cpu(cblock->bb_numrecs));
-	xfs_bmbt_log_ptrs(cur, bp, 1, be16_to_cpu(cblock->bb_numrecs));
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	*logflags |=
-		XFS_ILOG_CORE | XFS_ILOG_FBROOT(cur->bc_private.b.whichfork);
-	*stat = 1;
-	return 0;
-}
 
 /*
  * Set all the fields in a bmap extent record from the arguments.
@@ -2512,7 +429,8 @@ xfs_bmbt_set_state(
  */
 void
 xfs_bmbt_to_bmdr(
-	xfs_bmbt_block_t	*rblock,
+	struct xfs_mount	*mp,
+	struct xfs_btree_block	*rblock,
 	int			rblocklen,
 	xfs_bmdr_block_t	*dblock,
 	int			dblocklen)
@@ -2524,67 +442,22 @@ xfs_bmbt_to_bmdr(
 	__be64			*tpp;
 
 	ASSERT(be32_to_cpu(rblock->bb_magic) == XFS_BMAP_MAGIC);
-	ASSERT(be64_to_cpu(rblock->bb_leftsib) == NULLDFSBNO);
-	ASSERT(be64_to_cpu(rblock->bb_rightsib) == NULLDFSBNO);
+	ASSERT(be64_to_cpu(rblock->bb_u.l.bb_leftsib) == NULLDFSBNO);
+	ASSERT(be64_to_cpu(rblock->bb_u.l.bb_rightsib) == NULLDFSBNO);
 	ASSERT(be16_to_cpu(rblock->bb_level) > 0);
 	dblock->bb_level = rblock->bb_level;
 	dblock->bb_numrecs = rblock->bb_numrecs;
-	dmxr = (int)XFS_BTREE_BLOCK_MAXRECS(dblocklen, xfs_bmdr, 0);
-	fkp = XFS_BMAP_BROOT_KEY_ADDR(rblock, 1, rblocklen);
-	tkp = XFS_BTREE_KEY_ADDR(xfs_bmdr, dblock, 1);
-	fpp = XFS_BMAP_BROOT_PTR_ADDR(rblock, 1, rblocklen);
-	tpp = XFS_BTREE_PTR_ADDR(xfs_bmdr, dblock, 1, dmxr);
+	dmxr = xfs_bmdr_maxrecs(mp, dblocklen, 0);
+	fkp = XFS_BMBT_KEY_ADDR(mp, rblock, 1);
+	tkp = XFS_BMDR_KEY_ADDR(dblock, 1);
+	fpp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, rblocklen);
+	tpp = XFS_BMDR_PTR_ADDR(dblock, 1, dmxr);
 	dmxr = be16_to_cpu(dblock->bb_numrecs);
 	memcpy(tkp, fkp, sizeof(*fkp) * dmxr);
 	memcpy(tpp, fpp, sizeof(*fpp) * dmxr);
 }
 
 /*
- * Update the record to the passed values.
- */
-int
-xfs_bmbt_update(
-	xfs_btree_cur_t		*cur,
-	xfs_fileoff_t		off,
-	xfs_fsblock_t		bno,
-	xfs_filblks_t		len,
-	xfs_exntst_t		state)
-{
-	xfs_bmbt_block_t	*block;
-	xfs_buf_t		*bp;
-	int			error;
-	xfs_bmbt_key_t		key;
-	int			ptr;
-	xfs_bmbt_rec_t		*rp;
-
-	XFS_BMBT_TRACE_CURSOR(cur, ENTRY);
-	XFS_BMBT_TRACE_ARGFFFI(cur, (xfs_dfiloff_t)off, (xfs_dfsbno_t)bno,
-		(xfs_dfilblks_t)len, (int)state);
-	block = xfs_bmbt_get_block(cur, 0, &bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_lblock(cur, block, 0, bp))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-#endif
-	ptr = cur->bc_ptrs[0];
-	rp = XFS_BMAP_REC_IADDR(block, ptr, cur);
-	xfs_bmbt_disk_set_allf(rp, off, bno, len, state);
-	xfs_bmbt_log_recs(cur, bp, ptr, ptr);
-	if (ptr > 1) {
-		XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-		return 0;
-	}
-	key.br_startoff = cpu_to_be64(off);
-	if ((error = xfs_bmbt_updkey(cur, &key, 1))) {
-		XFS_BMBT_TRACE_CURSOR(cur, ERROR);
-		return error;
-	}
-	XFS_BMBT_TRACE_CURSOR(cur, EXIT);
-	return 0;
-}
-
-/*
  * Check extent records, which have just been read, for
  * any bit in the extent flag field. ASSERT on debug
  * kernels, as this condition should not occur.
@@ -2608,3 +481,451 @@ xfs_check_nostate_extents(
 	}
 	return 0;
 }
+
+
+STATIC struct xfs_btree_cur *
+xfs_bmbt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	struct xfs_btree_cur	*new;
+
+	new = xfs_bmbt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.b.ip, cur->bc_private.b.whichfork);
+
+	/*
+	 * Copy the firstblock, flist, and flags values,
+	 * since init cursor doesn't get them.
+	 */
+	new->bc_private.b.firstblock = cur->bc_private.b.firstblock;
+	new->bc_private.b.flist = cur->bc_private.b.flist;
+	new->bc_private.b.flags = cur->bc_private.b.flags;
+
+	return new;
+}
+
+STATIC void
+xfs_bmbt_update_cursor(
+	struct xfs_btree_cur	*src,
+	struct xfs_btree_cur	*dst)
+{
+	ASSERT((dst->bc_private.b.firstblock != NULLFSBLOCK) ||
+	       (dst->bc_private.b.ip->i_d.di_flags & XFS_DIFLAG_REALTIME));
+	ASSERT(dst->bc_private.b.flist == src->bc_private.b.flist);
+
+	dst->bc_private.b.allocated += src->bc_private.b.allocated;
+	dst->bc_private.b.firstblock = src->bc_private.b.firstblock;
+
+	src->bc_private.b.allocated = 0;
+}
+
+STATIC int
+xfs_bmbt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			length,
+	int			*stat)
+{
+	xfs_alloc_arg_t		args;		/* block allocation args */
+	int			error;		/* error return value */
+
+	memset(&args, 0, sizeof(args));
+	args.tp = cur->bc_tp;
+	args.mp = cur->bc_mp;
+	args.fsbno = cur->bc_private.b.firstblock;
+	args.firstblock = args.fsbno;
+
+	if (args.fsbno == NULLFSBLOCK) {
+		args.fsbno = be64_to_cpu(start->l);
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		/*
+		 * Make sure there is sufficient room left in the AG to
+		 * complete a full tree split for an extent insert.  If
+		 * we are converting the middle part of an extent then
+		 * we may need space for two tree splits.
+		 *
+		 * We are relying on the caller to make the correct block
+		 * reservation for this operation to succeed.  If the
+		 * reservation amount is insufficient then we may fail a
+		 * block allocation here and corrupt the filesystem.
+		 */
+		args.minleft = xfs_trans_get_block_res(args.tp);
+	} else if (cur->bc_private.b.flist->xbf_low) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+	} else {
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+	}
+
+	args.minlen = args.maxlen = args.prod = 1;
+	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
+	if (!args.wasdel && xfs_trans_get_block_res(args.tp) == 0) {
+		error = XFS_ERROR(ENOSPC);
+		goto error0;
+	}
+	error = xfs_alloc_vextent(&args);
+	if (error)
+		goto error0;
+
+	if (args.fsbno == NULLFSBLOCK && args.minleft) {
+		/*
+		 * Could not find an AG with enough free space to satisfy
+		 * a full btree split.  Try again without minleft and if
+		 * successful activate the lowspace algorithm.
+		 */
+		args.fsbno = 0;
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.minleft = 0;
+		error = xfs_alloc_vextent(&args);
+		if (error)
+			goto error0;
+		cur->bc_private.b.flist->xbf_low = 1;
+	}
+	if (args.fsbno == NULLFSBLOCK) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+	ASSERT(args.len == 1);
+	cur->bc_private.b.firstblock = args.fsbno;
+	cur->bc_private.b.allocated++;
+	cur->bc_private.b.ip->i_d.di_nblocks++;
+	xfs_trans_log_inode(args.tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
+	XFS_TRANS_MOD_DQUOT_BYINO(args.mp, args.tp, cur->bc_private.b.ip,
+			XFS_TRANS_DQ_BCOUNT, 1L);
+
+	new->l = cpu_to_be64(args.fsbno);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+ error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+STATIC int
+xfs_bmbt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_trans	*tp = cur->bc_tp;
+	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp));
+
+	xfs_bmap_add_free(fsbno, 1, cur->bc_private.b.flist, mp);
+	ip->i_d.di_nblocks--;
+
+	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+	XFS_TRANS_MOD_DQUOT_BYINO(mp, tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
+	xfs_trans_binval(tp, bp);
+	return 0;
+}
+
+STATIC int
+xfs_bmbt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp;
+
+		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
+				    cur->bc_private.b.whichfork);
+
+		return xfs_bmbt_maxrecs(cur->bc_mp,
+					ifp->if_broot_bytes, level == 0) / 2;
+	}
+
+	return cur->bc_mp->m_bmap_dmnr[level != 0];
+}
+
+int
+xfs_bmbt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp;
+
+		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
+				    cur->bc_private.b.whichfork);
+
+		return xfs_bmbt_maxrecs(cur->bc_mp,
+					ifp->if_broot_bytes, level == 0);
+	}
+
+	return cur->bc_mp->m_bmap_dmxr[level != 0];
+
+}
+
+/*
+ * Get the maximum records we could store in the on-disk format.
+ *
+ * For non-root nodes this is equivalent to xfs_bmbt_get_maxrecs, but
+ * for the root node this checks the available space in the dinode fork
+ * so that we can resize the in-memory buffer to match it.  After a
+ * resize to the maximum size this function returns the same value
+ * as xfs_bmbt_get_maxrecs for the root node, too.
+ */
+STATIC int
+xfs_bmbt_get_dmaxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
+{
+	if (level != cur->bc_nlevels - 1)
+		return cur->bc_mp->m_bmap_dmxr[level != 0];
+	return xfs_bmdr_maxrecs(cur->bc_mp, cur->bc_private.b.forksize,
+				level == 0);
+}
+
+STATIC void
+xfs_bmbt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	key->bmbt.br_startoff =
+		cpu_to_be64(xfs_bmbt_disk_get_startoff(&rec->bmbt));
+}
+
+STATIC void
+xfs_bmbt_init_rec_from_key(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
+{
+	ASSERT(key->bmbt.br_startoff != 0);
+
+	xfs_bmbt_disk_set_allf(&rec->bmbt, be64_to_cpu(key->bmbt.br_startoff),
+			       0, 0, XFS_EXT_NORM);
+}
+
+STATIC void
+xfs_bmbt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	xfs_bmbt_disk_set_all(&rec->bmbt, &cur->bc_rec.b);
+}
+
+STATIC void
+xfs_bmbt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	ptr->l = 0;
+}
+
+STATIC __int64_t
+xfs_bmbt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
+{
+	return (__int64_t)be64_to_cpu(key->bmbt.br_startoff) -
+				      cur->bc_rec.b.br_startoff;
+}
+
+#ifdef DEBUG
+STATIC int
+xfs_bmbt_keys_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	return be64_to_cpu(k1->bmbt.br_startoff) <
+		be64_to_cpu(k2->bmbt.br_startoff);
+}
+
+STATIC int
+xfs_bmbt_recs_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*r1,
+	union xfs_btree_rec	*r2)
+{
+	return xfs_bmbt_disk_get_startoff(&r1->bmbt) +
+		xfs_bmbt_disk_get_blockcount(&r1->bmbt) <=
+		xfs_bmbt_disk_get_startoff(&r2->bmbt);
+}
+#endif	/* DEBUG */
+
+#ifdef XFS_BTREE_TRACE
+ktrace_t	*xfs_bmbt_trace_buf;
+
+STATIC void
+xfs_bmbt_trace_enter(
+	struct xfs_btree_cur	*cur,
+	const char		*func,
+	char			*s,
+	int			type,
+	int			line,
+	__psunsigned_t		a0,
+	__psunsigned_t		a1,
+	__psunsigned_t		a2,
+	__psunsigned_t		a3,
+	__psunsigned_t		a4,
+	__psunsigned_t		a5,
+	__psunsigned_t		a6,
+	__psunsigned_t		a7,
+	__psunsigned_t		a8,
+	__psunsigned_t		a9,
+	__psunsigned_t		a10)
+{
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	int			whichfork = cur->bc_private.b.whichfork;
+
+	ktrace_enter(xfs_bmbt_trace_buf,
+		(void *)((__psint_t)type | (whichfork << 8) | (line << 16)),
+		(void *)func, (void *)s, (void *)ip, (void *)cur,
+		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
+		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
+		(void *)a8, (void *)a9, (void *)a10);
+	ktrace_enter(ip->i_btrace,
+		(void *)((__psint_t)type | (whichfork << 8) | (line << 16)),
+		(void *)func, (void *)s, (void *)ip, (void *)cur,
+		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
+		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
+		(void *)a8, (void *)a9, (void *)a10);
+}
+
+STATIC void
+xfs_bmbt_trace_cursor(
+	struct xfs_btree_cur	*cur,
+	__uint32_t		*s0,
+	__uint64_t		*l0,
+	__uint64_t		*l1)
+{
+	struct xfs_bmbt_rec_host r;
+
+	xfs_bmbt_set_all(&r, &cur->bc_rec.b);
+
+	*s0 = (cur->bc_nlevels << 24) |
+	      (cur->bc_private.b.flags << 16) |
+	       cur->bc_private.b.allocated;
+	*l0 = r.l0;
+	*l1 = r.l1;
+}
+
+STATIC void
+xfs_bmbt_trace_key(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key,
+	__uint64_t		*l0,
+	__uint64_t		*l1)
+{
+	*l0 = be64_to_cpu(key->bmbt.br_startoff);
+	*l1 = 0;
+}
+
+STATIC void
+xfs_bmbt_trace_record(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec,
+	__uint64_t		*l0,
+	__uint64_t		*l1,
+	__uint64_t		*l2)
+{
+	struct xfs_bmbt_irec	irec;
+
+	xfs_bmbt_disk_get_all(&rec->bmbt, &irec);
+	*l0 = irec.br_startoff;
+	*l1 = irec.br_startblock;
+	*l2 = irec.br_blockcount;
+}
+#endif /* XFS_BTREE_TRACE */
+
+static const struct xfs_btree_ops xfs_bmbt_ops = {
+	.rec_len		= sizeof(xfs_bmbt_rec_t),
+	.key_len		= sizeof(xfs_bmbt_key_t),
+
+	.dup_cursor		= xfs_bmbt_dup_cursor,
+	.update_cursor		= xfs_bmbt_update_cursor,
+	.alloc_block		= xfs_bmbt_alloc_block,
+	.free_block		= xfs_bmbt_free_block,
+	.get_maxrecs		= xfs_bmbt_get_maxrecs,
+	.get_minrecs		= xfs_bmbt_get_minrecs,
+	.get_dmaxrecs		= xfs_bmbt_get_dmaxrecs,
+	.init_key_from_rec	= xfs_bmbt_init_key_from_rec,
+	.init_rec_from_key	= xfs_bmbt_init_rec_from_key,
+	.init_rec_from_cur	= xfs_bmbt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_bmbt_init_ptr_from_cur,
+	.key_diff		= xfs_bmbt_key_diff,
+
+#ifdef DEBUG
+	.keys_inorder		= xfs_bmbt_keys_inorder,
+	.recs_inorder		= xfs_bmbt_recs_inorder,
+#endif
+
+#ifdef XFS_BTREE_TRACE
+	.trace_enter		= xfs_bmbt_trace_enter,
+	.trace_cursor		= xfs_bmbt_trace_cursor,
+	.trace_key		= xfs_bmbt_trace_key,
+	.trace_record		= xfs_bmbt_trace_record,
+#endif
+};
+
+/*
+ * Allocate a new bmap btree cursor.
+ */
+struct xfs_btree_cur *				/* new bmap btree cursor */
+xfs_bmbt_init_cursor(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_inode	*ip,		/* inode owning the btree */
+	int			whichfork)	/* data or attr fork */
+{
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	struct xfs_btree_cur	*cur;
+
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
+
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
+	cur->bc_btnum = XFS_BTNUM_BMAP;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+
+	cur->bc_ops = &xfs_bmbt_ops;
+	cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE;
+
+	cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, whichfork);
+	cur->bc_private.b.ip = ip;
+	cur->bc_private.b.firstblock = NULLFSBLOCK;
+	cur->bc_private.b.flist = NULL;
+	cur->bc_private.b.allocated = 0;
+	cur->bc_private.b.flags = 0;
+	cur->bc_private.b.whichfork = whichfork;
+
+	return cur;
+}
+
+/*
+ * Calculate number of records in a bmap btree block.
+ */
+int
+xfs_bmbt_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	int			leaf)
+{
+	blocklen -= XFS_BMBT_BLOCK_LEN(mp);
+
+	if (leaf)
+		return blocklen / sizeof(xfs_bmbt_rec_t);
+	return blocklen / (sizeof(xfs_bmbt_key_t) + sizeof(xfs_bmbt_ptr_t));
+}
+
+/*
+ * Calculate number of records in a bmap btree inode root.
+ */
+int
+xfs_bmdr_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	int			leaf)
+{
+	blocklen -= sizeof(xfs_bmdr_block_t);
+
+	if (leaf)
+		return blocklen / sizeof(xfs_bmdr_rec_t);
+	return blocklen / (sizeof(xfs_bmdr_key_t) + sizeof(xfs_bmdr_ptr_t));
+}
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index cd0d4b4bb81..a4555abb662 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -21,9 +21,10 @@
 #define XFS_BMAP_MAGIC	0x424d4150	/* 'BMAP' */
 
 struct xfs_btree_cur;
-struct xfs_btree_lblock;
+struct xfs_btree_block;
 struct xfs_mount;
 struct xfs_inode;
+struct xfs_trans;
 
 /*
  * Bmap root header, on-disk form only.
@@ -145,71 +146,60 @@ typedef struct xfs_bmbt_key {
 /* btree pointer type */
 typedef __be64 xfs_bmbt_ptr_t, xfs_bmdr_ptr_t;
 
-/* btree block header type */
-typedef struct xfs_btree_lblock xfs_bmbt_block_t;
-
-#define XFS_BUF_TO_BMBT_BLOCK(bp)	((xfs_bmbt_block_t *)XFS_BUF_PTR(bp))
-
-#define XFS_BMAP_RBLOCK_DSIZE(lev,cur)	((cur)->bc_private.b.forksize)
-#define XFS_BMAP_RBLOCK_ISIZE(lev,cur)	\
-	((int)XFS_IFORK_PTR((cur)->bc_private.b.ip, \
-		    (cur)->bc_private.b.whichfork)->if_broot_bytes)
-
-#define XFS_BMAP_BLOCK_DMAXRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-		XFS_BTREE_BLOCK_MAXRECS(XFS_BMAP_RBLOCK_DSIZE(lev,cur), \
-			xfs_bmdr, (lev) == 0) : \
-		((cur)->bc_mp->m_bmap_dmxr[(lev) != 0])))
-#define XFS_BMAP_BLOCK_IMAXRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-			XFS_BTREE_BLOCK_MAXRECS(XFS_BMAP_RBLOCK_ISIZE(lev,cur),\
-				xfs_bmbt, (lev) == 0) : \
-			((cur)->bc_mp->m_bmap_dmxr[(lev) != 0])))
-
-#define XFS_BMAP_BLOCK_DMINRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-			XFS_BTREE_BLOCK_MINRECS(XFS_BMAP_RBLOCK_DSIZE(lev,cur),\
-				xfs_bmdr, (lev) == 0) : \
-			((cur)->bc_mp->m_bmap_dmnr[(lev) != 0])))
-#define XFS_BMAP_BLOCK_IMINRECS(lev,cur) \
-	(((lev) == (cur)->bc_nlevels - 1 ? \
-			XFS_BTREE_BLOCK_MINRECS(XFS_BMAP_RBLOCK_ISIZE(lev,cur),\
-				xfs_bmbt, (lev) == 0) : \
-			((cur)->bc_mp->m_bmap_dmnr[(lev) != 0])))
-
-#define XFS_BMAP_REC_DADDR(bb,i,cur)	(XFS_BTREE_REC_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_REC_IADDR(bb,i,cur)	(XFS_BTREE_REC_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_KEY_DADDR(bb,i,cur)	\
-	(XFS_BTREE_KEY_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_KEY_IADDR(bb,i,cur)	\
-	(XFS_BTREE_KEY_ADDR(xfs_bmbt, bb, i))
-
-#define XFS_BMAP_PTR_DADDR(bb,i,cur)	\
-	(XFS_BTREE_PTR_ADDR(xfs_bmbt, bb, i, XFS_BMAP_BLOCK_DMAXRECS(	\
-				be16_to_cpu((bb)->bb_level), cur)))
-#define XFS_BMAP_PTR_IADDR(bb,i,cur)	\
-	(XFS_BTREE_PTR_ADDR(xfs_bmbt, bb, i, XFS_BMAP_BLOCK_IMAXRECS(	\
-				be16_to_cpu((bb)->bb_level), cur)))
+/*
+ * Btree block header size depends on a superblock flag.
+ *
+ * (not quite yet, but soon)
+ */
+#define XFS_BMBT_BLOCK_LEN(mp)	XFS_BTREE_LBLOCK_LEN
+
+#define XFS_BMBT_REC_ADDR(mp, block, index) \
+	((xfs_bmbt_rec_t *) \
+		((char *)(block) + \
+		 XFS_BMBT_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_bmbt_rec_t)))
+
+#define XFS_BMBT_KEY_ADDR(mp, block, index) \
+	((xfs_bmbt_key_t *) \
+		((char *)(block) + \
+		 XFS_BMBT_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_bmbt_key_t)))
+
+#define XFS_BMBT_PTR_ADDR(mp, block, index, maxrecs) \
+	((xfs_bmbt_ptr_t *) \
+		((char *)(block) + \
+		 XFS_BMBT_BLOCK_LEN(mp) + \
+		 (maxrecs) * sizeof(xfs_bmbt_key_t) + \
+		 ((index) - 1) * sizeof(xfs_bmbt_ptr_t)))
+
+#define XFS_BMDR_REC_ADDR(block, index) \
+	((xfs_bmdr_rec_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_bmdr_block) + \
+	         ((index) - 1) * sizeof(xfs_bmdr_rec_t)))
+
+#define XFS_BMDR_KEY_ADDR(block, index) \
+	((xfs_bmdr_key_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_bmdr_block) + \
+		 ((index) - 1) * sizeof(xfs_bmdr_key_t)))
+
+#define XFS_BMDR_PTR_ADDR(block, index, maxrecs) \
+	((xfs_bmdr_ptr_t *) \
+		((char *)(block) + \
+		 sizeof(struct xfs_bmdr_block) + \
+		 (maxrecs) * sizeof(xfs_bmdr_key_t) + \
+		 ((index) - 1) * sizeof(xfs_bmdr_ptr_t)))
 
 /*
  * These are to be used when we know the size of the block and
  * we don't have a cursor.
  */
-#define XFS_BMAP_BROOT_REC_ADDR(bb,i,sz) \
-	(XFS_BTREE_REC_ADDR(xfs_bmbt,bb,i))
-#define XFS_BMAP_BROOT_KEY_ADDR(bb,i,sz) \
-	(XFS_BTREE_KEY_ADDR(xfs_bmbt,bb,i))
-#define XFS_BMAP_BROOT_PTR_ADDR(bb,i,sz) \
-	(XFS_BTREE_PTR_ADDR(xfs_bmbt,bb,i,XFS_BMAP_BROOT_MAXRECS(sz)))
-
-#define XFS_BMAP_BROOT_NUMRECS(bb)	be16_to_cpu((bb)->bb_numrecs)
-#define XFS_BMAP_BROOT_MAXRECS(sz)	XFS_BTREE_BLOCK_MAXRECS(sz,xfs_bmbt,0)
+#define XFS_BMAP_BROOT_PTR_ADDR(mp, bb, i, sz) \
+	XFS_BMBT_PTR_ADDR(mp, bb, i, xfs_bmbt_maxrecs(mp, sz, 0))
 
 #define XFS_BMAP_BROOT_SPACE_CALC(nrecs) \
-	(int)(sizeof(xfs_bmbt_block_t) + \
+	(int)(XFS_BTREE_LBLOCK_LEN + \
 	       ((nrecs) * (sizeof(xfs_bmbt_key_t) + sizeof(xfs_bmbt_ptr_t))))
 
 #define XFS_BMAP_BROOT_SPACE(bb) \
@@ -223,42 +213,12 @@ typedef struct xfs_btree_lblock xfs_bmbt_block_t;
  */
 #define XFS_BM_MAXLEVELS(mp,w)		((mp)->m_bm_maxlevels[(w)])
 
-#define XFS_BMAP_SANITY_CHECK(mp,bb,level) \
-	(be32_to_cpu((bb)->bb_magic) == XFS_BMAP_MAGIC && \
-	 be16_to_cpu((bb)->bb_level) == level && \
-	 be16_to_cpu((bb)->bb_numrecs) > 0 && \
-	 be16_to_cpu((bb)->bb_numrecs) <= (mp)->m_bmap_dmxr[(level) != 0])
-
-
-#ifdef __KERNEL__
-
-#if defined(XFS_BMBT_TRACE)
-/*
- * Trace buffer entry types.
- */
-#define XFS_BMBT_KTRACE_ARGBI	1
-#define XFS_BMBT_KTRACE_ARGBII	2
-#define XFS_BMBT_KTRACE_ARGFFFI 3
-#define XFS_BMBT_KTRACE_ARGI	4
-#define XFS_BMBT_KTRACE_ARGIFK	5
-#define XFS_BMBT_KTRACE_ARGIFR	6
-#define XFS_BMBT_KTRACE_ARGIK	7
-#define XFS_BMBT_KTRACE_CUR	8
-
-#define XFS_BMBT_TRACE_SIZE	4096	/* size of global trace buffer */
-#define XFS_BMBT_KTRACE_SIZE	32	/* size of per-inode trace buffer */
-extern ktrace_t	*xfs_bmbt_trace_buf;
-#endif
-
 /*
  * Prototypes for xfs_bmap.c to call.
  */
-extern void xfs_bmdr_to_bmbt(xfs_bmdr_block_t *, int, xfs_bmbt_block_t *, int);
-extern int xfs_bmbt_decrement(struct xfs_btree_cur *, int, int *);
-extern int xfs_bmbt_delete(struct xfs_btree_cur *, int *);
+extern void xfs_bmdr_to_bmbt(struct xfs_mount *, xfs_bmdr_block_t *, int,
+			struct xfs_btree_block *, int);
 extern void xfs_bmbt_get_all(xfs_bmbt_rec_host_t *r, xfs_bmbt_irec_t *s);
-extern xfs_bmbt_block_t *xfs_bmbt_get_block(struct xfs_btree_cur *cur,
-						int, struct xfs_buf **bpp);
 extern xfs_filblks_t xfs_bmbt_get_blockcount(xfs_bmbt_rec_host_t *r);
 extern xfs_fsblock_t xfs_bmbt_get_startblock(xfs_bmbt_rec_host_t *r);
 extern xfs_fileoff_t xfs_bmbt_get_startoff(xfs_bmbt_rec_host_t *r);
@@ -268,22 +228,6 @@ extern void xfs_bmbt_disk_get_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern xfs_filblks_t xfs_bmbt_disk_get_blockcount(xfs_bmbt_rec_t *r);
 extern xfs_fileoff_t xfs_bmbt_disk_get_startoff(xfs_bmbt_rec_t *r);
 
-extern int xfs_bmbt_increment(struct xfs_btree_cur *, int, int *);
-extern int xfs_bmbt_insert(struct xfs_btree_cur *, int *);
-extern void xfs_bmbt_log_block(struct xfs_btree_cur *, struct xfs_buf *, int);
-extern void xfs_bmbt_log_recs(struct xfs_btree_cur *, struct xfs_buf *, int,
-				int);
-extern int xfs_bmbt_lookup_eq(struct xfs_btree_cur *, xfs_fileoff_t,
-				xfs_fsblock_t, xfs_filblks_t, int *);
-extern int xfs_bmbt_lookup_ge(struct xfs_btree_cur *, xfs_fileoff_t,
-				xfs_fsblock_t, xfs_filblks_t, int *);
-
-/*
- * Give the bmap btree a new root block.  Copy the old broot contents
- * down into a real block and make the broot point to it.
- */
-extern int xfs_bmbt_newroot(struct xfs_btree_cur *cur, int *lflags, int *stat);
-
 extern void xfs_bmbt_set_all(xfs_bmbt_rec_host_t *r, xfs_bmbt_irec_t *s);
 extern void xfs_bmbt_set_allf(xfs_bmbt_rec_host_t *r, xfs_fileoff_t o,
 			xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v);
@@ -296,10 +240,15 @@ extern void xfs_bmbt_disk_set_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
 extern void xfs_bmbt_disk_set_allf(xfs_bmbt_rec_t *r, xfs_fileoff_t o,
 			xfs_fsblock_t b, xfs_filblks_t c, xfs_exntst_t v);
 
-extern void xfs_bmbt_to_bmdr(xfs_bmbt_block_t *, int, xfs_bmdr_block_t *, int);
-extern int xfs_bmbt_update(struct xfs_btree_cur *, xfs_fileoff_t,
-				xfs_fsblock_t, xfs_filblks_t, xfs_exntst_t);
+extern void xfs_bmbt_to_bmdr(struct xfs_mount *, struct xfs_btree_block *, int,
+			xfs_bmdr_block_t *, int);
+
+extern int xfs_bmbt_get_maxrecs(struct xfs_btree_cur *, int level);
+extern int xfs_bmdr_maxrecs(struct xfs_mount *, int blocklen, int leaf);
+extern int xfs_bmbt_maxrecs(struct xfs_mount *, int blocklen, int leaf);
+
+extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
+		struct xfs_trans *, struct xfs_inode *, int);
 
-#endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c
index cc593a84c34..7ed59267420 100644
--- a/fs/xfs/xfs_btree.c
+++ b/fs/xfs/xfs_btree.c
@@ -34,7 +34,9 @@
 #include "xfs_attr_sf.h"
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
+#include "xfs_inode_item.h"
 #include "xfs_btree.h"
+#include "xfs_btree_trace.h"
 #include "xfs_ialloc.h"
 #include "xfs_error.h"
 
@@ -50,135 +52,33 @@ const __uint32_t xfs_magics[XFS_BTNUM_MAX] = {
 	XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC
 };
 
-/*
- * Checking routine: return maxrecs for the block.
- */
-STATIC int				/* number of records fitting in block */
-xfs_btree_maxrecs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block)	/* generic btree block pointer */
-{
-	switch (cur->bc_btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		return (int)XFS_ALLOC_BLOCK_MAXRECS(
-				be16_to_cpu(block->bb_h.bb_level), cur);
-	case XFS_BTNUM_BMAP:
-		return (int)XFS_BMAP_BLOCK_IMAXRECS(
-				be16_to_cpu(block->bb_h.bb_level), cur);
-	case XFS_BTNUM_INO:
-		return (int)XFS_INOBT_BLOCK_MAXRECS(
-				be16_to_cpu(block->bb_h.bb_level), cur);
-	default:
-		ASSERT(0);
-		return 0;
-	}
-}
-
-/*
- * External routines.
- */
-
-#ifdef DEBUG
-/*
- * Debug routine: check that block header is ok.
- */
-void
-xfs_btree_check_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block,	/* generic btree block pointer */
-	int			level,	/* level of the btree block */
-	xfs_buf_t		*bp)	/* buffer containing block, if any */
-{
-	if (XFS_BTREE_LONG_PTRS(cur->bc_btnum))
-		xfs_btree_check_lblock(cur, (xfs_btree_lblock_t *)block, level,
-			bp);
-	else
-		xfs_btree_check_sblock(cur, (xfs_btree_sblock_t *)block, level,
-			bp);
-}
-
-/*
- * Debug routine: check that keys are in the right order.
- */
-void
-xfs_btree_check_key(
-	xfs_btnum_t	btnum,		/* btree identifier */
-	void		*ak1,		/* pointer to left (lower) key */
-	void		*ak2)		/* pointer to right (higher) key */
-{
-	switch (btnum) {
-	case XFS_BTNUM_BNO: {
-		xfs_alloc_key_t	*k1;
-		xfs_alloc_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be32_to_cpu(k1->ar_startblock) < be32_to_cpu(k2->ar_startblock));
-		break;
-	    }
-	case XFS_BTNUM_CNT: {
-		xfs_alloc_key_t	*k1;
-		xfs_alloc_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be32_to_cpu(k1->ar_blockcount) < be32_to_cpu(k2->ar_blockcount) ||
-		       (k1->ar_blockcount == k2->ar_blockcount &&
-			be32_to_cpu(k1->ar_startblock) < be32_to_cpu(k2->ar_startblock)));
-		break;
-	    }
-	case XFS_BTNUM_BMAP: {
-		xfs_bmbt_key_t	*k1;
-		xfs_bmbt_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be64_to_cpu(k1->br_startoff) < be64_to_cpu(k2->br_startoff));
-		break;
-	    }
-	case XFS_BTNUM_INO: {
-		xfs_inobt_key_t	*k1;
-		xfs_inobt_key_t	*k2;
-
-		k1 = ak1;
-		k2 = ak2;
-		ASSERT(be32_to_cpu(k1->ir_startino) < be32_to_cpu(k2->ir_startino));
-		break;
-	    }
-	default:
-		ASSERT(0);
-	}
-}
-#endif	/* DEBUG */
 
-/*
- * Checking routine: check that long form block header is ok.
- */
-/* ARGSUSED */
-int					/* error (0 or EFSCORRUPTED) */
+STATIC int				/* error (0 or EFSCORRUPTED) */
 xfs_btree_check_lblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_lblock_t	*block,	/* btree long form block pointer */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* btree long form block pointer */
 	int			level,	/* level of the btree block */
-	xfs_buf_t		*bp)	/* buffer for block, if any */
+	struct xfs_buf		*bp)	/* buffer for block, if any */
 {
 	int			lblock_ok; /* block passes checks */
-	xfs_mount_t		*mp;	/* file system mount point */
+	struct xfs_mount	*mp;	/* file system mount point */
 
 	mp = cur->bc_mp;
 	lblock_ok =
 		be32_to_cpu(block->bb_magic) == xfs_magics[cur->bc_btnum] &&
 		be16_to_cpu(block->bb_level) == level &&
 		be16_to_cpu(block->bb_numrecs) <=
-			xfs_btree_maxrecs(cur, (xfs_btree_block_t *)block) &&
-		block->bb_leftsib &&
-		(be64_to_cpu(block->bb_leftsib) == NULLDFSBNO ||
-		 XFS_FSB_SANITY_CHECK(mp, be64_to_cpu(block->bb_leftsib))) &&
-		block->bb_rightsib &&
-		(be64_to_cpu(block->bb_rightsib) == NULLDFSBNO ||
-		 XFS_FSB_SANITY_CHECK(mp, be64_to_cpu(block->bb_rightsib)));
-	if (unlikely(XFS_TEST_ERROR(!lblock_ok, mp, XFS_ERRTAG_BTREE_CHECK_LBLOCK,
+			cur->bc_ops->get_maxrecs(cur, level) &&
+		block->bb_u.l.bb_leftsib &&
+		(be64_to_cpu(block->bb_u.l.bb_leftsib) == NULLDFSBNO ||
+		 XFS_FSB_SANITY_CHECK(mp,
+		 	be64_to_cpu(block->bb_u.l.bb_leftsib))) &&
+		block->bb_u.l.bb_rightsib &&
+		(be64_to_cpu(block->bb_u.l.bb_rightsib) == NULLDFSBNO ||
+		 XFS_FSB_SANITY_CHECK(mp,
+		 	be64_to_cpu(block->bb_u.l.bb_rightsib)));
+	if (unlikely(XFS_TEST_ERROR(!lblock_ok, mp,
+			XFS_ERRTAG_BTREE_CHECK_LBLOCK,
 			XFS_RANDOM_BTREE_CHECK_LBLOCK))) {
 		if (bp)
 			xfs_buftrace("LBTREE ERROR", bp);
@@ -189,98 +89,15 @@ xfs_btree_check_lblock(
 	return 0;
 }
 
-/*
- * Checking routine: check that (long) pointer is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_lptr(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_dfsbno_t	ptr,		/* btree block disk address */
-	int		level)		/* btree block level */
-{
-	xfs_mount_t	*mp;		/* file system mount point */
-
-	mp = cur->bc_mp;
-	XFS_WANT_CORRUPTED_RETURN(
-		level > 0 &&
-		ptr != NULLDFSBNO &&
-		XFS_FSB_SANITY_CHECK(mp, ptr));
-	return 0;
-}
-
-#ifdef DEBUG
-/*
- * Debug routine: check that records are in the right order.
- */
-void
-xfs_btree_check_rec(
-	xfs_btnum_t	btnum,		/* btree identifier */
-	void		*ar1,		/* pointer to left (lower) record */
-	void		*ar2)		/* pointer to right (higher) record */
-{
-	switch (btnum) {
-	case XFS_BTNUM_BNO: {
-		xfs_alloc_rec_t	*r1;
-		xfs_alloc_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(be32_to_cpu(r1->ar_startblock) +
-		       be32_to_cpu(r1->ar_blockcount) <=
-		       be32_to_cpu(r2->ar_startblock));
-		break;
-	    }
-	case XFS_BTNUM_CNT: {
-		xfs_alloc_rec_t	*r1;
-		xfs_alloc_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(be32_to_cpu(r1->ar_blockcount) < be32_to_cpu(r2->ar_blockcount) ||
-		       (r1->ar_blockcount == r2->ar_blockcount &&
-			be32_to_cpu(r1->ar_startblock) < be32_to_cpu(r2->ar_startblock)));
-		break;
-	    }
-	case XFS_BTNUM_BMAP: {
-		xfs_bmbt_rec_t	*r1;
-		xfs_bmbt_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(xfs_bmbt_disk_get_startoff(r1) +
-		       xfs_bmbt_disk_get_blockcount(r1) <=
-		       xfs_bmbt_disk_get_startoff(r2));
-		break;
-	    }
-	case XFS_BTNUM_INO: {
-		xfs_inobt_rec_t	*r1;
-		xfs_inobt_rec_t	*r2;
-
-		r1 = ar1;
-		r2 = ar2;
-		ASSERT(be32_to_cpu(r1->ir_startino) + XFS_INODES_PER_CHUNK <=
-		       be32_to_cpu(r2->ir_startino));
-		break;
-	    }
-	default:
-		ASSERT(0);
-	}
-}
-#endif	/* DEBUG */
-
-/*
- * Checking routine: check that block header is ok.
- */
-/* ARGSUSED */
-int					/* error (0 or EFSCORRUPTED) */
+STATIC int				/* error (0 or EFSCORRUPTED) */
 xfs_btree_check_sblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_sblock_t	*block,	/* btree short form block pointer */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* btree short form block pointer */
 	int			level,	/* level of the btree block */
-	xfs_buf_t		*bp)	/* buffer containing block */
+	struct xfs_buf		*bp)	/* buffer containing block */
 {
-	xfs_buf_t		*agbp;	/* buffer for ag. freespace struct */
-	xfs_agf_t		*agf;	/* ag. freespace structure */
+	struct xfs_buf		*agbp;	/* buffer for ag. freespace struct */
+	struct xfs_agf		*agf;	/* ag. freespace structure */
 	xfs_agblock_t		agflen;	/* native ag. freespace length */
 	int			sblock_ok; /* block passes checks */
 
@@ -291,13 +108,13 @@ xfs_btree_check_sblock(
 		be32_to_cpu(block->bb_magic) == xfs_magics[cur->bc_btnum] &&
 		be16_to_cpu(block->bb_level) == level &&
 		be16_to_cpu(block->bb_numrecs) <=
-			xfs_btree_maxrecs(cur, (xfs_btree_block_t *)block) &&
-		(be32_to_cpu(block->bb_leftsib) == NULLAGBLOCK ||
-		 be32_to_cpu(block->bb_leftsib) < agflen) &&
-		block->bb_leftsib &&
-		(be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK ||
-		 be32_to_cpu(block->bb_rightsib) < agflen) &&
-		block->bb_rightsib;
+			cur->bc_ops->get_maxrecs(cur, level) &&
+		(be32_to_cpu(block->bb_u.s.bb_leftsib) == NULLAGBLOCK ||
+		 be32_to_cpu(block->bb_u.s.bb_leftsib) < agflen) &&
+		block->bb_u.s.bb_leftsib &&
+		(be32_to_cpu(block->bb_u.s.bb_rightsib) == NULLAGBLOCK ||
+		 be32_to_cpu(block->bb_u.s.bb_rightsib) < agflen) &&
+		block->bb_u.s.bb_rightsib;
 	if (unlikely(XFS_TEST_ERROR(!sblock_ok, cur->bc_mp,
 			XFS_ERRTAG_BTREE_CHECK_SBLOCK,
 			XFS_RANDOM_BTREE_CHECK_SBLOCK))) {
@@ -311,27 +128,78 @@ xfs_btree_check_sblock(
 }
 
 /*
- * Checking routine: check that (short) pointer is ok.
+ * Debug routine: check that block header is ok.
+ */
+int
+xfs_btree_check_block(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* generic btree block pointer */
+	int			level,	/* level of the btree block */
+	struct xfs_buf		*bp)	/* buffer containing block, if any */
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return xfs_btree_check_lblock(cur, block, level, bp);
+	else
+		return xfs_btree_check_sblock(cur, block, level, bp);
+}
+
+/*
+ * Check that (long) pointer is ok.
  */
 int					/* error (0 or EFSCORRUPTED) */
+xfs_btree_check_lptr(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_dfsbno_t		bno,	/* btree block disk address */
+	int			level)	/* btree block level */
+{
+	XFS_WANT_CORRUPTED_RETURN(
+		level > 0 &&
+		bno != NULLDFSBNO &&
+		XFS_FSB_SANITY_CHECK(cur->bc_mp, bno));
+	return 0;
+}
+
+#ifdef DEBUG
+/*
+ * Check that (short) pointer is ok.
+ */
+STATIC int				/* error (0 or EFSCORRUPTED) */
 xfs_btree_check_sptr(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agblock_t	ptr,		/* btree block disk address */
-	int		level)		/* btree block level */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agblock_t		bno,	/* btree block disk address */
+	int			level)	/* btree block level */
 {
-	xfs_buf_t	*agbp;		/* buffer for ag. freespace struct */
-	xfs_agf_t	*agf;		/* ag. freespace structure */
+	xfs_agblock_t		agblocks = cur->bc_mp->m_sb.sb_agblocks;
 
-	agbp = cur->bc_private.a.agbp;
-	agf = XFS_BUF_TO_AGF(agbp);
 	XFS_WANT_CORRUPTED_RETURN(
 		level > 0 &&
-		ptr != NULLAGBLOCK && ptr != 0 &&
-		ptr < be32_to_cpu(agf->agf_length));
+		bno != NULLAGBLOCK &&
+		bno != 0 &&
+		bno < agblocks);
 	return 0;
 }
 
 /*
+ * Check that block ptr is ok.
+ */
+STATIC int				/* error (0 or EFSCORRUPTED) */
+xfs_btree_check_ptr(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	union xfs_btree_ptr	*ptr,	/* btree block disk address */
+	int			index,	/* offset from ptr to check */
+	int			level)	/* btree block level */
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		return xfs_btree_check_lptr(cur,
+				be64_to_cpu((&ptr->l)[index]), level);
+	} else {
+		return xfs_btree_check_sptr(cur,
+				be32_to_cpu((&ptr->s)[index]), level);
+	}
+}
+#endif
+
+/*
  * Delete the btree cursor.
  */
 void
@@ -387,16 +255,17 @@ xfs_btree_dup_cursor(
 
 	tp = cur->bc_tp;
 	mp = cur->bc_mp;
+
 	/*
 	 * Allocate a new cursor like the old one.
 	 */
-	new = xfs_btree_init_cursor(mp, tp, cur->bc_private.a.agbp,
-		cur->bc_private.a.agno, cur->bc_btnum, cur->bc_private.b.ip,
-		cur->bc_private.b.whichfork);
+	new = cur->bc_ops->dup_cursor(cur);
+
 	/*
 	 * Copy the record currently in the cursor.
 	 */
 	new->bc_rec = cur->bc_rec;
+
 	/*
 	 * For each level current, re-get the buffer and copy the ptr value.
 	 */
@@ -416,46 +285,174 @@ xfs_btree_dup_cursor(
 		} else
 			new->bc_bufs[i] = NULL;
 	}
-	/*
-	 * For bmap btrees, copy the firstblock, flist, and flags values,
-	 * since init cursor doesn't get them.
-	 */
-	if (new->bc_btnum == XFS_BTNUM_BMAP) {
-		new->bc_private.b.firstblock = cur->bc_private.b.firstblock;
-		new->bc_private.b.flist = cur->bc_private.b.flist;
-		new->bc_private.b.flags = cur->bc_private.b.flags;
-	}
 	*ncur = new;
 	return 0;
 }
 
 /*
+ * XFS btree block layout and addressing:
+ *
+ * There are two types of blocks in the btree: leaf and non-leaf blocks.
+ *
+ * The leaf record start with a header then followed by records containing
+ * the values.  A non-leaf block also starts with the same header, and
+ * then first contains lookup keys followed by an equal number of pointers
+ * to the btree blocks at the previous level.
+ *
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ * Leaf:	| header | rec 1 | rec 2 | rec 3 | rec 4 | rec 5 | rec N |
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ *
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ * Non-Leaf:	| header | key 1 | key 2 | key N | ptr 1 | ptr 2 | ptr N |
+ *		+--------+-------+-------+-------+-------+-------+-------+
+ *
+ * The header is called struct xfs_btree_block for reasons better left unknown
+ * and comes in different versions for short (32bit) and long (64bit) block
+ * pointers.  The record and key structures are defined by the btree instances
+ * and opaque to the btree core.  The block pointers are simple disk endian
+ * integers, available in a short (32bit) and long (64bit) variant.
+ *
+ * The helpers below calculate the offset of a given record, key or pointer
+ * into a btree block (xfs_btree_*_offset) or return a pointer to the given
+ * record, key or pointer (xfs_btree_*_addr).  Note that all addressing
+ * inside the btree block is done using indices starting at one, not zero!
+ */
+
+/*
+ * Return size of the btree block header for this btree instance.
+ */
+static inline size_t xfs_btree_block_len(struct xfs_btree_cur *cur)
+{
+	return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+		XFS_BTREE_LBLOCK_LEN :
+		XFS_BTREE_SBLOCK_LEN;
+}
+
+/*
+ * Return size of btree block pointers for this btree instance.
+ */
+static inline size_t xfs_btree_ptr_len(struct xfs_btree_cur *cur)
+{
+	return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+		sizeof(__be64) : sizeof(__be32);
+}
+
+/*
+ * Calculate offset of the n-th record in a btree block.
+ */
+STATIC size_t
+xfs_btree_rec_offset(
+	struct xfs_btree_cur	*cur,
+	int			n)
+{
+	return xfs_btree_block_len(cur) +
+		(n - 1) * cur->bc_ops->rec_len;
+}
+
+/*
+ * Calculate offset of the n-th key in a btree block.
+ */
+STATIC size_t
+xfs_btree_key_offset(
+	struct xfs_btree_cur	*cur,
+	int			n)
+{
+	return xfs_btree_block_len(cur) +
+		(n - 1) * cur->bc_ops->key_len;
+}
+
+/*
+ * Calculate offset of the n-th block pointer in a btree block.
+ */
+STATIC size_t
+xfs_btree_ptr_offset(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	int			level)
+{
+	return xfs_btree_block_len(cur) +
+		cur->bc_ops->get_maxrecs(cur, level) * cur->bc_ops->key_len +
+		(n - 1) * xfs_btree_ptr_len(cur);
+}
+
+/*
+ * Return a pointer to the n-th record in the btree block.
+ */
+STATIC union xfs_btree_rec *
+xfs_btree_rec_addr(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	struct xfs_btree_block	*block)
+{
+	return (union xfs_btree_rec *)
+		((char *)block + xfs_btree_rec_offset(cur, n));
+}
+
+/*
+ * Return a pointer to the n-th key in the btree block.
+ */
+STATIC union xfs_btree_key *
+xfs_btree_key_addr(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	struct xfs_btree_block	*block)
+{
+	return (union xfs_btree_key *)
+		((char *)block + xfs_btree_key_offset(cur, n));
+}
+
+/*
+ * Return a pointer to the n-th block pointer in the btree block.
+ */
+STATIC union xfs_btree_ptr *
+xfs_btree_ptr_addr(
+	struct xfs_btree_cur	*cur,
+	int			n,
+	struct xfs_btree_block	*block)
+{
+	int			level = xfs_btree_get_level(block);
+
+	ASSERT(block->bb_level != 0);
+
+	return (union xfs_btree_ptr *)
+		((char *)block + xfs_btree_ptr_offset(cur, n, level));
+}
+
+/*
+ * Get a the root block which is stored in the inode.
+ *
+ * For now this btree implementation assumes the btree root is always
+ * stored in the if_broot field of an inode fork.
+ */
+STATIC struct xfs_btree_block *
+xfs_btree_get_iroot(
+       struct xfs_btree_cur    *cur)
+{
+       struct xfs_ifork        *ifp;
+
+       ifp = XFS_IFORK_PTR(cur->bc_private.b.ip, cur->bc_private.b.whichfork);
+       return (struct xfs_btree_block *)ifp->if_broot;
+}
+
+/*
  * Retrieve the block pointer from the cursor at the given level.
- * This may be a bmap btree root or from a buffer.
+ * This may be an inode btree root or from a buffer.
  */
-STATIC xfs_btree_block_t *		/* generic btree block pointer */
+STATIC struct xfs_btree_block *		/* generic btree block pointer */
 xfs_btree_get_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
 	int			level,	/* level in btree */
-	xfs_buf_t		**bpp)	/* buffer containing the block */
-{
-	xfs_btree_block_t	*block;	/* return value */
-	xfs_buf_t		*bp;	/* return buffer */
-	xfs_ifork_t		*ifp;	/* inode fork pointer */
-	int			whichfork; /* data or attr fork */
-
-	if (cur->bc_btnum == XFS_BTNUM_BMAP && level == cur->bc_nlevels - 1) {
-		whichfork = cur->bc_private.b.whichfork;
-		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip, whichfork);
-		block = (xfs_btree_block_t *)ifp->if_broot;
-		bp = NULL;
-	} else {
-		bp = cur->bc_bufs[level];
-		block = XFS_BUF_TO_BLOCK(bp);
+	struct xfs_buf		**bpp)	/* buffer containing the block */
+{
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level == cur->bc_nlevels - 1)) {
+		*bpp = NULL;
+		return xfs_btree_get_iroot(cur);
 	}
-	ASSERT(block != NULL);
-	*bpp = bp;
-	return block;
+
+	*bpp = cur->bc_bufs[level];
+	return XFS_BUF_TO_BLOCK(*bpp);
 }
 
 /*
@@ -505,97 +502,6 @@ xfs_btree_get_bufs(
 }
 
 /*
- * Allocate a new btree cursor.
- * The cursor is either for allocation (A) or bmap (B) or inodes (I).
- */
-xfs_btree_cur_t *			/* new btree cursor */
-xfs_btree_init_cursor(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_buf_t	*agbp,		/* (A only) buffer for agf structure */
-					/* (I only) buffer for agi structure */
-	xfs_agnumber_t	agno,		/* (AI only) allocation group number */
-	xfs_btnum_t	btnum,		/* btree identifier */
-	xfs_inode_t	*ip,		/* (B only) inode owning the btree */
-	int		whichfork)	/* (B only) data or attr fork */
-{
-	xfs_agf_t	*agf;		/* (A) allocation group freespace */
-	xfs_agi_t	*agi;		/* (I) allocation group inodespace */
-	xfs_btree_cur_t	*cur;		/* return value */
-	xfs_ifork_t	*ifp;		/* (I) inode fork pointer */
-	int		nlevels=0;	/* number of levels in the btree */
-
-	ASSERT(xfs_btree_cur_zone != NULL);
-	/*
-	 * Allocate a new cursor.
-	 */
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
-	/*
-	 * Deduce the number of btree levels from the arguments.
-	 */
-	switch (btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		agf = XFS_BUF_TO_AGF(agbp);
-		nlevels = be32_to_cpu(agf->agf_levels[btnum]);
-		break;
-	case XFS_BTNUM_BMAP:
-		ifp = XFS_IFORK_PTR(ip, whichfork);
-		nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
-		break;
-	case XFS_BTNUM_INO:
-		agi = XFS_BUF_TO_AGI(agbp);
-		nlevels = be32_to_cpu(agi->agi_level);
-		break;
-	default:
-		ASSERT(0);
-	}
-	/*
-	 * Fill in the common fields.
-	 */
-	cur->bc_tp = tp;
-	cur->bc_mp = mp;
-	cur->bc_nlevels = nlevels;
-	cur->bc_btnum = btnum;
-	cur->bc_blocklog = mp->m_sb.sb_blocklog;
-	/*
-	 * Fill in private fields.
-	 */
-	switch (btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		/*
-		 * Allocation btree fields.
-		 */
-		cur->bc_private.a.agbp = agbp;
-		cur->bc_private.a.agno = agno;
-		break;
-	case XFS_BTNUM_INO:
-		/*
-		 * Inode allocation btree fields.
-		 */
-		cur->bc_private.a.agbp = agbp;
-		cur->bc_private.a.agno = agno;
-		break;
-	case XFS_BTNUM_BMAP:
-		/*
-		 * Bmap btree fields.
-		 */
-		cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, whichfork);
-		cur->bc_private.b.ip = ip;
-		cur->bc_private.b.firstblock = NULLFSBLOCK;
-		cur->bc_private.b.flist = NULL;
-		cur->bc_private.b.allocated = 0;
-		cur->bc_private.b.flags = 0;
-		cur->bc_private.b.whichfork = whichfork;
-		break;
-	default:
-		ASSERT(0);
-	}
-	return cur;
-}
-
-/*
  * Check for the cursor referring to the last block at the given level.
  */
 int					/* 1=is last block, 0=not last block */
@@ -603,12 +509,12 @@ xfs_btree_islastblock(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			level)	/* level to check */
 {
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
+	struct xfs_btree_block	*block;	/* generic btree block pointer */
 	xfs_buf_t		*bp;	/* buffer containing block */
 
 	block = xfs_btree_get_block(cur, level, &bp);
 	xfs_btree_check_block(cur, block, level, bp);
-	if (XFS_BTREE_LONG_PTRS(cur->bc_btnum))
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
 		return be64_to_cpu(block->bb_u.l.bb_rightsib) == NULLDFSBNO;
 	else
 		return be32_to_cpu(block->bb_u.s.bb_rightsib) == NULLAGBLOCK;
@@ -618,12 +524,12 @@ xfs_btree_islastblock(
  * Change the cursor to point to the first record at the given level.
  * Other levels are unaffected.
  */
-int					/* success=1, failure=0 */
+STATIC int				/* success=1, failure=0 */
 xfs_btree_firstrec(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			level)	/* level to change */
 {
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
+	struct xfs_btree_block	*block;	/* generic btree block pointer */
 	xfs_buf_t		*bp;	/* buffer containing block */
 
 	/*
@@ -634,7 +540,7 @@ xfs_btree_firstrec(
 	/*
 	 * It's empty, there is no such record.
 	 */
-	if (!block->bb_h.bb_numrecs)
+	if (!block->bb_numrecs)
 		return 0;
 	/*
 	 * Set the ptr value to 1, that's the first record/key.
@@ -647,12 +553,12 @@ xfs_btree_firstrec(
  * Change the cursor to point to the last record in the current block
  * at the given level.  Other levels are unaffected.
  */
-int					/* success=1, failure=0 */
+STATIC int				/* success=1, failure=0 */
 xfs_btree_lastrec(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			level)	/* level to change */
 {
-	xfs_btree_block_t	*block;	/* generic btree block pointer */
+	struct xfs_btree_block	*block;	/* generic btree block pointer */
 	xfs_buf_t		*bp;	/* buffer containing block */
 
 	/*
@@ -663,12 +569,12 @@ xfs_btree_lastrec(
 	/*
 	 * It's empty, there is no such record.
 	 */
-	if (!block->bb_h.bb_numrecs)
+	if (!block->bb_numrecs)
 		return 0;
 	/*
 	 * Set the ptr value to numrecs, that's the last record/key.
 	 */
-	cur->bc_ptrs[level] = be16_to_cpu(block->bb_h.bb_numrecs);
+	cur->bc_ptrs[level] = be16_to_cpu(block->bb_numrecs);
 	return 1;
 }
 
@@ -817,66 +723,84 @@ xfs_btree_reada_bufs(
 	xfs_baread(mp->m_ddev_targp, d, mp->m_bsize * count);
 }
 
+STATIC int
+xfs_btree_readahead_lblock(
+	struct xfs_btree_cur	*cur,
+	int			lr,
+	struct xfs_btree_block	*block)
+{
+	int			rval = 0;
+	xfs_fsblock_t		left = be64_to_cpu(block->bb_u.l.bb_leftsib);
+	xfs_fsblock_t		right = be64_to_cpu(block->bb_u.l.bb_rightsib);
+
+	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLDFSBNO) {
+		xfs_btree_reada_bufl(cur->bc_mp, left, 1);
+		rval++;
+	}
+
+	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLDFSBNO) {
+		xfs_btree_reada_bufl(cur->bc_mp, right, 1);
+		rval++;
+	}
+
+	return rval;
+}
+
+STATIC int
+xfs_btree_readahead_sblock(
+	struct xfs_btree_cur	*cur,
+	int			lr,
+	struct xfs_btree_block *block)
+{
+	int			rval = 0;
+	xfs_agblock_t		left = be32_to_cpu(block->bb_u.s.bb_leftsib);
+	xfs_agblock_t		right = be32_to_cpu(block->bb_u.s.bb_rightsib);
+
+
+	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLAGBLOCK) {
+		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
+				     left, 1);
+		rval++;
+	}
+
+	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLAGBLOCK) {
+		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
+				     right, 1);
+		rval++;
+	}
+
+	return rval;
+}
+
 /*
  * Read-ahead btree blocks, at the given level.
  * Bits in lr are set from XFS_BTCUR_{LEFT,RIGHT}RA.
  */
-int
-xfs_btree_readahead_core(
-	xfs_btree_cur_t		*cur,		/* btree cursor */
+STATIC int
+xfs_btree_readahead(
+	struct xfs_btree_cur	*cur,		/* btree cursor */
 	int			lev,		/* level in btree */
 	int			lr)		/* left/right bits */
 {
-	xfs_alloc_block_t	*a;
-	xfs_bmbt_block_t	*b;
-	xfs_inobt_block_t	*i;
-	int			rval = 0;
+	struct xfs_btree_block	*block;
+
+	/*
+	 * No readahead needed if we are at the root level and the
+	 * btree root is stored in the inode.
+	 */
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (lev == cur->bc_nlevels - 1))
+		return 0;
+
+	if ((cur->bc_ra[lev] | lr) == cur->bc_ra[lev])
+		return 0;
 
-	ASSERT(cur->bc_bufs[lev] != NULL);
 	cur->bc_ra[lev] |= lr;
-	switch (cur->bc_btnum) {
-	case XFS_BTNUM_BNO:
-	case XFS_BTNUM_CNT:
-		a = XFS_BUF_TO_ALLOC_BLOCK(cur->bc_bufs[lev]);
-		if ((lr & XFS_BTCUR_LEFTRA) && be32_to_cpu(a->bb_leftsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
-				be32_to_cpu(a->bb_leftsib), 1);
-			rval++;
-		}
-		if ((lr & XFS_BTCUR_RIGHTRA) && be32_to_cpu(a->bb_rightsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
-				be32_to_cpu(a->bb_rightsib), 1);
-			rval++;
-		}
-		break;
-	case XFS_BTNUM_BMAP:
-		b = XFS_BUF_TO_BMBT_BLOCK(cur->bc_bufs[lev]);
-		if ((lr & XFS_BTCUR_LEFTRA) && be64_to_cpu(b->bb_leftsib) != NULLDFSBNO) {
-			xfs_btree_reada_bufl(cur->bc_mp, be64_to_cpu(b->bb_leftsib), 1);
-			rval++;
-		}
-		if ((lr & XFS_BTCUR_RIGHTRA) && be64_to_cpu(b->bb_rightsib) != NULLDFSBNO) {
-			xfs_btree_reada_bufl(cur->bc_mp, be64_to_cpu(b->bb_rightsib), 1);
-			rval++;
-		}
-		break;
-	case XFS_BTNUM_INO:
-		i = XFS_BUF_TO_INOBT_BLOCK(cur->bc_bufs[lev]);
-		if ((lr & XFS_BTCUR_LEFTRA) && be32_to_cpu(i->bb_leftsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
-				be32_to_cpu(i->bb_leftsib), 1);
-			rval++;
-		}
-		if ((lr & XFS_BTCUR_RIGHTRA) && be32_to_cpu(i->bb_rightsib) != NULLAGBLOCK) {
-			xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
-				be32_to_cpu(i->bb_rightsib), 1);
-			rval++;
-		}
-		break;
-	default:
-		ASSERT(0);
-	}
-	return rval;
+	block = XFS_BUF_TO_BLOCK(cur->bc_bufs[lev]);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return xfs_btree_readahead_lblock(cur, lr, block);
+	return xfs_btree_readahead_sblock(cur, lr, block);
 }
 
 /*
@@ -889,7 +813,7 @@ xfs_btree_setbuf(
 	int			lev,	/* level in btree */
 	xfs_buf_t		*bp)	/* new buffer to set */
 {
-	xfs_btree_block_t	*b;	/* btree block */
+	struct xfs_btree_block	*b;	/* btree block */
 	xfs_buf_t		*obp;	/* old buffer pointer */
 
 	obp = cur->bc_bufs[lev];
@@ -900,7 +824,7 @@ xfs_btree_setbuf(
 	if (!bp)
 		return;
 	b = XFS_BUF_TO_BLOCK(bp);
-	if (XFS_BTREE_LONG_PTRS(cur->bc_btnum)) {
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
 		if (be64_to_cpu(b->bb_u.l.bb_leftsib) == NULLDFSBNO)
 			cur->bc_ra[lev] |= XFS_BTCUR_LEFTRA;
 		if (be64_to_cpu(b->bb_u.l.bb_rightsib) == NULLDFSBNO)
@@ -912,3 +836,2855 @@ xfs_btree_setbuf(
 			cur->bc_ra[lev] |= XFS_BTCUR_RIGHTRA;
 	}
 }
+
+STATIC int
+xfs_btree_ptr_is_null(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		return be64_to_cpu(ptr->l) == NULLFSBLOCK;
+	else
+		return be32_to_cpu(ptr->s) == NULLAGBLOCK;
+}
+
+STATIC void
+xfs_btree_set_ptr_null(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		ptr->l = cpu_to_be64(NULLFSBLOCK);
+	else
+		ptr->s = cpu_to_be32(NULLAGBLOCK);
+}
+
+/*
+ * Get/set/init sibling pointers
+ */
+STATIC void
+xfs_btree_get_sibling(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	union xfs_btree_ptr	*ptr,
+	int			lr)
+{
+	ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (lr == XFS_BB_RIGHTSIB)
+			ptr->l = block->bb_u.l.bb_rightsib;
+		else
+			ptr->l = block->bb_u.l.bb_leftsib;
+	} else {
+		if (lr == XFS_BB_RIGHTSIB)
+			ptr->s = block->bb_u.s.bb_rightsib;
+		else
+			ptr->s = block->bb_u.s.bb_leftsib;
+	}
+}
+
+STATIC void
+xfs_btree_set_sibling(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	union xfs_btree_ptr	*ptr,
+	int			lr)
+{
+	ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		if (lr == XFS_BB_RIGHTSIB)
+			block->bb_u.l.bb_rightsib = ptr->l;
+		else
+			block->bb_u.l.bb_leftsib = ptr->l;
+	} else {
+		if (lr == XFS_BB_RIGHTSIB)
+			block->bb_u.s.bb_rightsib = ptr->s;
+		else
+			block->bb_u.s.bb_leftsib = ptr->s;
+	}
+}
+
+STATIC void
+xfs_btree_init_block(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			numrecs,
+	struct xfs_btree_block	*new)	/* new block */
+{
+	new->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
+	new->bb_level = cpu_to_be16(level);
+	new->bb_numrecs = cpu_to_be16(numrecs);
+
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		new->bb_u.l.bb_leftsib = cpu_to_be64(NULLFSBLOCK);
+		new->bb_u.l.bb_rightsib = cpu_to_be64(NULLFSBLOCK);
+	} else {
+		new->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK);
+		new->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK);
+	}
+}
+
+/*
+ * Return true if ptr is the last record in the btree and
+ * we need to track updateѕ to this record.  The decision
+ * will be further refined in the update_lastrec method.
+ */
+STATIC int
+xfs_btree_is_lastrec(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_block	*block,
+	int			level)
+{
+	union xfs_btree_ptr	ptr;
+
+	if (level > 0)
+		return 0;
+	if (!(cur->bc_flags & XFS_BTREE_LASTREC_UPDATE))
+		return 0;
+
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+	if (!xfs_btree_ptr_is_null(cur, &ptr))
+		return 0;
+	return 1;
+}
+
+STATIC void
+xfs_btree_buf_to_ptr(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		ptr->l = cpu_to_be64(XFS_DADDR_TO_FSB(cur->bc_mp,
+					XFS_BUF_ADDR(bp)));
+	else {
+		ptr->s = cpu_to_be32(XFS_DADDR_TO_AGBNO(cur->bc_mp,
+					XFS_BUF_ADDR(bp)));
+	}
+}
+
+STATIC xfs_daddr_t
+xfs_btree_ptr_to_daddr(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		ASSERT(be64_to_cpu(ptr->l) != NULLFSBLOCK);
+
+		return XFS_FSB_TO_DADDR(cur->bc_mp, be64_to_cpu(ptr->l));
+	} else {
+		ASSERT(cur->bc_private.a.agno != NULLAGNUMBER);
+		ASSERT(be32_to_cpu(ptr->s) != NULLAGBLOCK);
+
+		return XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_private.a.agno,
+					be32_to_cpu(ptr->s));
+	}
+}
+
+STATIC void
+xfs_btree_set_refs(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	switch (cur->bc_btnum) {
+	case XFS_BTNUM_BNO:
+	case XFS_BTNUM_CNT:
+		XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_MAP, XFS_ALLOC_BTREE_REF);
+		break;
+	case XFS_BTNUM_INO:
+		XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_INOMAP, XFS_INO_BTREE_REF);
+		break;
+	case XFS_BTNUM_BMAP:
+		XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_MAP, XFS_BMAP_BTREE_REF);
+		break;
+	default:
+		ASSERT(0);
+	}
+}
+
+STATIC int
+xfs_btree_get_buf_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			flags,
+	struct xfs_btree_block	**block,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_daddr_t		d;
+
+	/* need to sort out how callers deal with failures first */
+	ASSERT(!(flags & XFS_BUF_TRYLOCK));
+
+	d = xfs_btree_ptr_to_daddr(cur, ptr);
+	*bpp = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d,
+				 mp->m_bsize, flags);
+
+	ASSERT(*bpp);
+	ASSERT(!XFS_BUF_GETERROR(*bpp));
+
+	*block = XFS_BUF_TO_BLOCK(*bpp);
+	return 0;
+}
+
+/*
+ * Read in the buffer at the given ptr and return the buffer and
+ * the block pointer within the buffer.
+ */
+STATIC int
+xfs_btree_read_buf_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			level,
+	int			flags,
+	struct xfs_btree_block	**block,
+	struct xfs_buf		**bpp)
+{
+	struct xfs_mount	*mp = cur->bc_mp;
+	xfs_daddr_t		d;
+	int			error;
+
+	/* need to sort out how callers deal with failures first */
+	ASSERT(!(flags & XFS_BUF_TRYLOCK));
+
+	d = xfs_btree_ptr_to_daddr(cur, ptr);
+	error = xfs_trans_read_buf(mp, cur->bc_tp, mp->m_ddev_targp, d,
+				   mp->m_bsize, flags, bpp);
+	if (error)
+		return error;
+
+	ASSERT(*bpp != NULL);
+	ASSERT(!XFS_BUF_GETERROR(*bpp));
+
+	xfs_btree_set_refs(cur, *bpp);
+	*block = XFS_BUF_TO_BLOCK(*bpp);
+
+	error = xfs_btree_check_block(cur, *block, level, *bpp);
+	if (error)
+		xfs_trans_brelse(cur->bc_tp, *bpp);
+	return error;
+}
+
+/*
+ * Copy keys from one btree block to another.
+ */
+STATIC void
+xfs_btree_copy_keys(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*dst_key,
+	union xfs_btree_key	*src_key,
+	int			numkeys)
+{
+	ASSERT(numkeys >= 0);
+	memcpy(dst_key, src_key, numkeys * cur->bc_ops->key_len);
+}
+
+/*
+ * Copy records from one btree block to another.
+ */
+STATIC void
+xfs_btree_copy_recs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*dst_rec,
+	union xfs_btree_rec	*src_rec,
+	int			numrecs)
+{
+	ASSERT(numrecs >= 0);
+	memcpy(dst_rec, src_rec, numrecs * cur->bc_ops->rec_len);
+}
+
+/*
+ * Copy block pointers from one btree block to another.
+ */
+STATIC void
+xfs_btree_copy_ptrs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*dst_ptr,
+	union xfs_btree_ptr	*src_ptr,
+	int			numptrs)
+{
+	ASSERT(numptrs >= 0);
+	memcpy(dst_ptr, src_ptr, numptrs * xfs_btree_ptr_len(cur));
+}
+
+/*
+ * Shift keys one index left/right inside a single btree block.
+ */
+STATIC void
+xfs_btree_shift_keys(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key,
+	int			dir,
+	int			numkeys)
+{
+	char			*dst_key;
+
+	ASSERT(numkeys >= 0);
+	ASSERT(dir == 1 || dir == -1);
+
+	dst_key = (char *)key + (dir * cur->bc_ops->key_len);
+	memmove(dst_key, key, numkeys * cur->bc_ops->key_len);
+}
+
+/*
+ * Shift records one index left/right inside a single btree block.
+ */
+STATIC void
+xfs_btree_shift_recs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec,
+	int			dir,
+	int			numrecs)
+{
+	char			*dst_rec;
+
+	ASSERT(numrecs >= 0);
+	ASSERT(dir == 1 || dir == -1);
+
+	dst_rec = (char *)rec + (dir * cur->bc_ops->rec_len);
+	memmove(dst_rec, rec, numrecs * cur->bc_ops->rec_len);
+}
+
+/*
+ * Shift block pointers one index left/right inside a single btree block.
+ */
+STATIC void
+xfs_btree_shift_ptrs(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			dir,
+	int			numptrs)
+{
+	char			*dst_ptr;
+
+	ASSERT(numptrs >= 0);
+	ASSERT(dir == 1 || dir == -1);
+
+	dst_ptr = (char *)ptr + (dir * xfs_btree_ptr_len(cur));
+	memmove(dst_ptr, ptr, numptrs * xfs_btree_ptr_len(cur));
+}
+
+/*
+ * Log key values from the btree block.
+ */
+STATIC void
+xfs_btree_log_keys(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			first,
+	int			last)
+{
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBII(cur, bp, first, last);
+
+	if (bp) {
+		xfs_trans_log_buf(cur->bc_tp, bp,
+				  xfs_btree_key_offset(cur, first),
+				  xfs_btree_key_offset(cur, last + 1) - 1);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+				xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Log record values from the btree block.
+ */
+void
+xfs_btree_log_recs(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			first,
+	int			last)
+{
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBII(cur, bp, first, last);
+
+	xfs_trans_log_buf(cur->bc_tp, bp,
+			  xfs_btree_rec_offset(cur, first),
+			  xfs_btree_rec_offset(cur, last + 1) - 1);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Log block pointer fields from a btree block (nonleaf).
+ */
+STATIC void
+xfs_btree_log_ptrs(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_buf		*bp,	/* buffer containing btree block */
+	int			first,	/* index of first pointer to log */
+	int			last)	/* index of last pointer to log */
+{
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBII(cur, bp, first, last);
+
+	if (bp) {
+		struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
+		int			level = xfs_btree_get_level(block);
+
+		xfs_trans_log_buf(cur->bc_tp, bp,
+				xfs_btree_ptr_offset(cur, first, level),
+				xfs_btree_ptr_offset(cur, last + 1, level) - 1);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+			xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Log fields from a btree block header.
+ */
+void
+xfs_btree_log_block(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_buf		*bp,	/* buffer containing btree block */
+	int			fields)	/* mask of fields: XFS_BB_... */
+{
+	int			first;	/* first byte offset logged */
+	int			last;	/* last byte offset logged */
+	static const short	soffsets[] = {	/* table of offsets (short) */
+		offsetof(struct xfs_btree_block, bb_magic),
+		offsetof(struct xfs_btree_block, bb_level),
+		offsetof(struct xfs_btree_block, bb_numrecs),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_leftsib),
+		offsetof(struct xfs_btree_block, bb_u.s.bb_rightsib),
+		XFS_BTREE_SBLOCK_LEN
+	};
+	static const short	loffsets[] = {	/* table of offsets (long) */
+		offsetof(struct xfs_btree_block, bb_magic),
+		offsetof(struct xfs_btree_block, bb_level),
+		offsetof(struct xfs_btree_block, bb_numrecs),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_leftsib),
+		offsetof(struct xfs_btree_block, bb_u.l.bb_rightsib),
+		XFS_BTREE_LBLOCK_LEN
+	};
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGBI(cur, bp, fields);
+
+	if (bp) {
+		xfs_btree_offsets(fields,
+				  (cur->bc_flags & XFS_BTREE_LONG_PTRS) ?
+					loffsets : soffsets,
+				  XFS_BB_NUM_BITS, &first, &last);
+		xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	} else {
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
+			xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+}
+
+/*
+ * Increment cursor by one record at the level.
+ * For nonzero levels the leaf-ward information is untouched.
+ */
+int						/* error */
+xfs_btree_increment(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	struct xfs_btree_block	*block;
+	union xfs_btree_ptr	ptr;
+	struct xfs_buf		*bp;
+	int			error;		/* error return value */
+	int			lev;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	ASSERT(level < cur->bc_nlevels);
+
+	/* Read-ahead to the right at this level. */
+	xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
+
+	/* Get a pointer to the btree block. */
+	block = xfs_btree_get_block(cur, level, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+#endif
+
+	/* We're done if we remain in the block after the increment. */
+	if (++cur->bc_ptrs[level] <= xfs_btree_get_numrecs(block))
+		goto out1;
+
+	/* Fail if we just went off the right edge of the tree. */
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+	if (xfs_btree_ptr_is_null(cur, &ptr))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, increment);
+
+	/*
+	 * March up the tree incrementing pointers.
+	 * Stop when we don't go off the right edge of a block.
+	 */
+	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
+		block = xfs_btree_get_block(cur, lev, &bp);
+
+#ifdef DEBUG
+		error = xfs_btree_check_block(cur, block, lev, bp);
+		if (error)
+			goto error0;
+#endif
+
+		if (++cur->bc_ptrs[lev] <= xfs_btree_get_numrecs(block))
+			break;
+
+		/* Read-ahead the right block for the next loop. */
+		xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
+	}
+
+	/*
+	 * If we went off the root then we are either seriously
+	 * confused or have the tree root in an inode.
+	 */
+	if (lev == cur->bc_nlevels) {
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+			goto out0;
+		ASSERT(0);
+		error = EFSCORRUPTED;
+		goto error0;
+	}
+	ASSERT(lev < cur->bc_nlevels);
+
+	/*
+	 * Now walk back down the tree, fixing up the cursor's buffer
+	 * pointers and key numbers.
+	 */
+	for (block = xfs_btree_get_block(cur, lev, &bp); lev > level; ) {
+		union xfs_btree_ptr	*ptrp;
+
+		ptrp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[lev], block);
+		error = xfs_btree_read_buf_block(cur, ptrp, --lev,
+							0, &block, &bp);
+		if (error)
+			goto error0;
+
+		xfs_btree_setbuf(cur, lev, bp);
+		cur->bc_ptrs[lev] = 1;
+	}
+out1:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Decrement cursor by one record at the level.
+ * For nonzero levels the leaf-ward information is untouched.
+ */
+int						/* error */
+xfs_btree_decrement(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	struct xfs_btree_block	*block;
+	xfs_buf_t		*bp;
+	int			error;		/* error return value */
+	int			lev;
+	union xfs_btree_ptr	ptr;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	ASSERT(level < cur->bc_nlevels);
+
+	/* Read-ahead to the left at this level. */
+	xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
+
+	/* We're done if we remain in the block after the decrement. */
+	if (--cur->bc_ptrs[level] > 0)
+		goto out1;
+
+	/* Get a pointer to the btree block. */
+	block = xfs_btree_get_block(cur, level, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+#endif
+
+	/* Fail if we just went off the left edge of the tree. */
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_LEFTSIB);
+	if (xfs_btree_ptr_is_null(cur, &ptr))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, decrement);
+
+	/*
+	 * March up the tree decrementing pointers.
+	 * Stop when we don't go off the left edge of a block.
+	 */
+	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
+		if (--cur->bc_ptrs[lev] > 0)
+			break;
+		/* Read-ahead the left block for the next loop. */
+		xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
+	}
+
+	/*
+	 * If we went off the root then we are seriously confused.
+	 * or the root of the tree is in an inode.
+	 */
+	if (lev == cur->bc_nlevels) {
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+			goto out0;
+		ASSERT(0);
+		error = EFSCORRUPTED;
+		goto error0;
+	}
+	ASSERT(lev < cur->bc_nlevels);
+
+	/*
+	 * Now walk back down the tree, fixing up the cursor's buffer
+	 * pointers and key numbers.
+	 */
+	for (block = xfs_btree_get_block(cur, lev, &bp); lev > level; ) {
+		union xfs_btree_ptr	*ptrp;
+
+		ptrp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[lev], block);
+		error = xfs_btree_read_buf_block(cur, ptrp, --lev,
+							0, &block, &bp);
+		if (error)
+			goto error0;
+		xfs_btree_setbuf(cur, lev, bp);
+		cur->bc_ptrs[lev] = xfs_btree_get_numrecs(block);
+	}
+out1:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+STATIC int
+xfs_btree_lookup_get_block(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			level,	/* level in the btree */
+	union xfs_btree_ptr	*pp,	/* ptr to btree block */
+	struct xfs_btree_block	**blkp) /* return btree block */
+{
+	struct xfs_buf		*bp;	/* buffer pointer for btree block */
+	int			error = 0;
+
+	/* special case the root block if in an inode */
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level == cur->bc_nlevels - 1)) {
+		*blkp = xfs_btree_get_iroot(cur);
+		return 0;
+	}
+
+	/*
+	 * If the old buffer at this level for the disk address we are
+	 * looking for re-use it.
+	 *
+	 * Otherwise throw it away and get a new one.
+	 */
+	bp = cur->bc_bufs[level];
+	if (bp && XFS_BUF_ADDR(bp) == xfs_btree_ptr_to_daddr(cur, pp)) {
+		*blkp = XFS_BUF_TO_BLOCK(bp);
+		return 0;
+	}
+
+	error = xfs_btree_read_buf_block(cur, pp, level, 0, blkp, &bp);
+	if (error)
+		return error;
+
+	xfs_btree_setbuf(cur, level, bp);
+	return 0;
+}
+
+/*
+ * Get current search key.  For level 0 we don't actually have a key
+ * structure so we make one up from the record.  For all other levels
+ * we just return the right key.
+ */
+STATIC union xfs_btree_key *
+xfs_lookup_get_search_key(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			keyno,
+	struct xfs_btree_block	*block,
+	union xfs_btree_key	*kp)
+{
+	if (level == 0) {
+		cur->bc_ops->init_key_from_rec(kp,
+				xfs_btree_rec_addr(cur, keyno, block));
+		return kp;
+	}
+
+	return xfs_btree_key_addr(cur, keyno, block);
+}
+
+/*
+ * Lookup the record.  The cursor is made to point to it, based on dir.
+ * Return 0 if can't find any such record, 1 for success.
+ */
+int					/* error */
+xfs_btree_lookup(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_lookup_t		dir,	/* <=, ==, or >= */
+	int			*stat)	/* success/failure */
+{
+	struct xfs_btree_block	*block;	/* current btree block */
+	__int64_t		diff;	/* difference for the current key */
+	int			error;	/* error return value */
+	int			keyno;	/* current key number */
+	int			level;	/* level in the btree */
+	union xfs_btree_ptr	*pp;	/* ptr to btree block */
+	union xfs_btree_ptr	ptr;	/* ptr to btree block */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, dir);
+
+	XFS_BTREE_STATS_INC(cur, lookup);
+
+	block = NULL;
+	keyno = 0;
+
+	/* initialise start pointer from cursor */
+	cur->bc_ops->init_ptr_from_cur(cur, &ptr);
+	pp = &ptr;
+
+	/*
+	 * Iterate over each level in the btree, starting at the root.
+	 * For each level above the leaves, find the key we need, based
+	 * on the lookup record, then follow the corresponding block
+	 * pointer down to the next level.
+	 */
+	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
+		/* Get the block we need to do the lookup on. */
+		error = xfs_btree_lookup_get_block(cur, level, pp, &block);
+		if (error)
+			goto error0;
+
+		if (diff == 0) {
+			/*
+			 * If we already had a key match at a higher level, we
+			 * know we need to use the first entry in this block.
+			 */
+			keyno = 1;
+		} else {
+			/* Otherwise search this block. Do a binary search. */
+
+			int	high;	/* high entry number */
+			int	low;	/* low entry number */
+
+			/* Set low and high entry numbers, 1-based. */
+			low = 1;
+			high = xfs_btree_get_numrecs(block);
+			if (!high) {
+				/* Block is empty, must be an empty leaf. */
+				ASSERT(level == 0 && cur->bc_nlevels == 1);
+
+				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
+				XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+				*stat = 0;
+				return 0;
+			}
+
+			/* Binary search the block. */
+			while (low <= high) {
+				union xfs_btree_key	key;
+				union xfs_btree_key	*kp;
+
+				XFS_BTREE_STATS_INC(cur, compare);
+
+				/* keyno is average of low and high. */
+				keyno = (low + high) >> 1;
+
+				/* Get current search key */
+				kp = xfs_lookup_get_search_key(cur, level,
+						keyno, block, &key);
+
+				/*
+				 * Compute difference to get next direction:
+				 *  - less than, move right
+				 *  - greater than, move left
+				 *  - equal, we're done
+				 */
+				diff = cur->bc_ops->key_diff(cur, kp);
+				if (diff < 0)
+					low = keyno + 1;
+				else if (diff > 0)
+					high = keyno - 1;
+				else
+					break;
+			}
+		}
+
+		/*
+		 * If there are more levels, set up for the next level
+		 * by getting the block number and filling in the cursor.
+		 */
+		if (level > 0) {
+			/*
+			 * If we moved left, need the previous key number,
+			 * unless there isn't one.
+			 */
+			if (diff > 0 && --keyno < 1)
+				keyno = 1;
+			pp = xfs_btree_ptr_addr(cur, keyno, block);
+
+#ifdef DEBUG
+			error = xfs_btree_check_ptr(cur, pp, 0, level);
+			if (error)
+				goto error0;
+#endif
+			cur->bc_ptrs[level] = keyno;
+		}
+	}
+
+	/* Done with the search. See if we need to adjust the results. */
+	if (dir != XFS_LOOKUP_LE && diff < 0) {
+		keyno++;
+		/*
+		 * If ge search and we went off the end of the block, but it's
+		 * not the last block, we're in the wrong block.
+		 */
+		xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+		if (dir == XFS_LOOKUP_GE &&
+		    keyno > xfs_btree_get_numrecs(block) &&
+		    !xfs_btree_ptr_is_null(cur, &ptr)) {
+			int	i;
+
+			cur->bc_ptrs[0] = keyno;
+			error = xfs_btree_increment(cur, 0, &i);
+			if (error)
+				goto error0;
+			XFS_WANT_CORRUPTED_RETURN(i == 1);
+			XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+			*stat = 1;
+			return 0;
+		}
+	} else if (dir == XFS_LOOKUP_LE && diff > 0)
+		keyno--;
+	cur->bc_ptrs[0] = keyno;
+
+	/* Return if we succeeded or not. */
+	if (keyno == 0 || keyno > xfs_btree_get_numrecs(block))
+		*stat = 0;
+	else if (dir != XFS_LOOKUP_EQ || diff == 0)
+		*stat = 1;
+	else
+		*stat = 0;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Update keys at all levels from here to the root along the cursor's path.
+ */
+STATIC int
+xfs_btree_updkey(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*keyp,
+	int			level)
+{
+	struct xfs_btree_block	*block;
+	struct xfs_buf		*bp;
+	union xfs_btree_key	*kp;
+	int			ptr;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGIK(cur, level, keyp);
+
+	ASSERT(!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) || level >= 1);
+
+	/*
+	 * Go up the tree from this level toward the root.
+	 * At each level, update the key value to the value input.
+	 * Stop when we reach a level where the cursor isn't pointing
+	 * at the first entry in the block.
+	 */
+	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
+#ifdef DEBUG
+		int		error;
+#endif
+		block = xfs_btree_get_block(cur, level, &bp);
+#ifdef DEBUG
+		error = xfs_btree_check_block(cur, block, level, bp);
+		if (error) {
+			XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+			return error;
+		}
+#endif
+		ptr = cur->bc_ptrs[level];
+		kp = xfs_btree_key_addr(cur, ptr, block);
+		xfs_btree_copy_keys(cur, kp, keyp, 1);
+		xfs_btree_log_keys(cur, bp, ptr, ptr);
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+}
+
+/*
+ * Update the record referred to by cur to the value in the
+ * given record. This either works (return 0) or gets an
+ * EFSCORRUPTED error.
+ */
+int
+xfs_btree_update(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	struct xfs_btree_block	*block;
+	struct xfs_buf		*bp;
+	int			error;
+	int			ptr;
+	union xfs_btree_rec	*rp;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGR(cur, rec);
+
+	/* Pick up the current block. */
+	block = xfs_btree_get_block(cur, 0, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, 0, bp);
+	if (error)
+		goto error0;
+#endif
+	/* Get the address of the rec to be updated. */
+	ptr = cur->bc_ptrs[0];
+	rp = xfs_btree_rec_addr(cur, ptr, block);
+
+	/* Fill in the new contents and log them. */
+	xfs_btree_copy_recs(cur, rp, rec, 1);
+	xfs_btree_log_recs(cur, bp, ptr, ptr);
+
+	/*
+	 * If we are tracking the last record in the tree and
+	 * we are at the far right edge of the tree, update it.
+	 */
+	if (xfs_btree_is_lastrec(cur, block, 0)) {
+		cur->bc_ops->update_lastrec(cur, block, rec,
+					    ptr, LASTREC_UPDATE);
+	}
+
+	/* Updating first rec in leaf. Pass new key value up to our parent. */
+	if (ptr == 1) {
+		union xfs_btree_key	key;
+
+		cur->bc_ops->init_key_from_rec(&key, rec);
+		error = xfs_btree_updkey(cur, &key, 1);
+		if (error)
+			goto error0;
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Move 1 record left from cur/level if possible.
+ * Update cur to reflect the new path.
+ */
+STATIC int					/* error */
+xfs_btree_lshift(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	union xfs_btree_key	key;		/* btree key */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	int			lrecs;		/* left record count */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	int			rrecs;		/* right record count */
+	union xfs_btree_ptr	lptr;		/* left btree pointer */
+	union xfs_btree_key	*rkp = NULL;	/* right btree key */
+	union xfs_btree_ptr	*rpp = NULL;	/* right address pointer */
+	union xfs_btree_rec	*rrp = NULL;	/* right record pointer */
+	int			error;		/* error return value */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == cur->bc_nlevels - 1)
+		goto out0;
+
+	/* Set up variables for this block as "right". */
+	right = xfs_btree_get_block(cur, level, &rbp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, right, level, rbp);
+	if (error)
+		goto error0;
+#endif
+
+	/* If we've got no left sibling then we can't shift an entry left. */
+	xfs_btree_get_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
+	if (xfs_btree_ptr_is_null(cur, &lptr))
+		goto out0;
+
+	/*
+	 * If the cursor entry is the one that would be moved, don't
+	 * do it... it's too complicated.
+	 */
+	if (cur->bc_ptrs[level] <= 1)
+		goto out0;
+
+	/* Set up the left neighbor as "left". */
+	error = xfs_btree_read_buf_block(cur, &lptr, level, 0, &left, &lbp);
+	if (error)
+		goto error0;
+
+	/* If it's full, it can't take another entry. */
+	lrecs = xfs_btree_get_numrecs(left);
+	if (lrecs == cur->bc_ops->get_maxrecs(cur, level))
+		goto out0;
+
+	rrecs = xfs_btree_get_numrecs(right);
+
+	/*
+	 * We add one entry to the left side and remove one for the right side.
+	 * Accout for it here, the changes will be updated on disk and logged
+	 * later.
+	 */
+	lrecs++;
+	rrecs--;
+
+	XFS_BTREE_STATS_INC(cur, lshift);
+	XFS_BTREE_STATS_ADD(cur, moves, 1);
+
+	/*
+	 * If non-leaf, copy a key and a ptr to the left block.
+	 * Log the changes to the left block.
+	 */
+	if (level > 0) {
+		/* It's a non-leaf.  Move keys and pointers. */
+		union xfs_btree_key	*lkp;	/* left btree key */
+		union xfs_btree_ptr	*lpp;	/* left address pointer */
+
+		lkp = xfs_btree_key_addr(cur, lrecs, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+
+		lpp = xfs_btree_ptr_addr(cur, lrecs, left);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+#ifdef DEBUG
+		error = xfs_btree_check_ptr(cur, rpp, 0, level);
+		if (error)
+			goto error0;
+#endif
+		xfs_btree_copy_keys(cur, lkp, rkp, 1);
+		xfs_btree_copy_ptrs(cur, lpp, rpp, 1);
+
+		xfs_btree_log_keys(cur, lbp, lrecs, lrecs);
+		xfs_btree_log_ptrs(cur, lbp, lrecs, lrecs);
+
+		ASSERT(cur->bc_ops->keys_inorder(cur,
+			xfs_btree_key_addr(cur, lrecs - 1, left), lkp));
+	} else {
+		/* It's a leaf.  Move records.  */
+		union xfs_btree_rec	*lrp;	/* left record pointer */
+
+		lrp = xfs_btree_rec_addr(cur, lrecs, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_copy_recs(cur, lrp, rrp, 1);
+		xfs_btree_log_recs(cur, lbp, lrecs, lrecs);
+
+		ASSERT(cur->bc_ops->recs_inorder(cur,
+			xfs_btree_rec_addr(cur, lrecs - 1, left), lrp));
+	}
+
+	xfs_btree_set_numrecs(left, lrecs);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS);
+
+	xfs_btree_set_numrecs(right, rrecs);
+	xfs_btree_log_block(cur, rbp, XFS_BB_NUMRECS);
+
+	/*
+	 * Slide the contents of right down one entry.
+	 */
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs - 1);
+	if (level > 0) {
+		/* It's a nonleaf. operate on keys and ptrs */
+#ifdef DEBUG
+		int			i;		/* loop index */
+
+		for (i = 0; i < rrecs; i++) {
+			error = xfs_btree_check_ptr(cur, rpp, i + 1, level);
+			if (error)
+				goto error0;
+		}
+#endif
+		xfs_btree_shift_keys(cur,
+				xfs_btree_key_addr(cur, 2, right),
+				-1, rrecs);
+		xfs_btree_shift_ptrs(cur,
+				xfs_btree_ptr_addr(cur, 2, right),
+				-1, rrecs);
+
+		xfs_btree_log_keys(cur, rbp, 1, rrecs);
+		xfs_btree_log_ptrs(cur, rbp, 1, rrecs);
+	} else {
+		/* It's a leaf. operate on records */
+		xfs_btree_shift_recs(cur,
+			xfs_btree_rec_addr(cur, 2, right),
+			-1, rrecs);
+		xfs_btree_log_recs(cur, rbp, 1, rrecs);
+
+		/*
+		 * If it's the first record in the block, we'll need a key
+		 * structure to pass up to the next level (updkey).
+		 */
+		cur->bc_ops->init_key_from_rec(&key,
+			xfs_btree_rec_addr(cur, 1, right));
+		rkp = &key;
+	}
+
+	/* Update the parent key values of right. */
+	error = xfs_btree_updkey(cur, rkp, level + 1);
+	if (error)
+		goto error0;
+
+	/* Slide the cursor value left one. */
+	cur->bc_ptrs[level]--;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Move 1 record right from cur/level if possible.
+ * Update cur to reflect the new path.
+ */
+STATIC int					/* error */
+xfs_btree_rshift(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)		/* success/failure */
+{
+	union xfs_btree_key	key;		/* btree key */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	struct xfs_btree_cur	*tcur;		/* temporary btree cursor */
+	union xfs_btree_ptr	rptr;		/* right block pointer */
+	union xfs_btree_key	*rkp;		/* right btree key */
+	int			rrecs;		/* right record count */
+	int			lrecs;		/* left record count */
+	int			error;		/* error return value */
+	int			i;		/* loop counter */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level == cur->bc_nlevels - 1))
+		goto out0;
+
+	/* Set up variables for this block as "left". */
+	left = xfs_btree_get_block(cur, level, &lbp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, left, level, lbp);
+	if (error)
+		goto error0;
+#endif
+
+	/* If we've got no right sibling then we can't shift an entry right. */
+	xfs_btree_get_sibling(cur, left, &rptr, XFS_BB_RIGHTSIB);
+	if (xfs_btree_ptr_is_null(cur, &rptr))
+		goto out0;
+
+	/*
+	 * If the cursor entry is the one that would be moved, don't
+	 * do it... it's too complicated.
+	 */
+	lrecs = xfs_btree_get_numrecs(left);
+	if (cur->bc_ptrs[level] >= lrecs)
+		goto out0;
+
+	/* Set up the right neighbor as "right". */
+	error = xfs_btree_read_buf_block(cur, &rptr, level, 0, &right, &rbp);
+	if (error)
+		goto error0;
+
+	/* If it's full, it can't take another entry. */
+	rrecs = xfs_btree_get_numrecs(right);
+	if (rrecs == cur->bc_ops->get_maxrecs(cur, level))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, rshift);
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs);
+
+	/*
+	 * Make a hole at the start of the right neighbor block, then
+	 * copy the last left block entry to the hole.
+	 */
+	if (level > 0) {
+		/* It's a nonleaf. make a hole in the keys and ptrs */
+		union xfs_btree_key	*lkp;
+		union xfs_btree_ptr	*lpp;
+		union xfs_btree_ptr	*rpp;
+
+		lkp = xfs_btree_key_addr(cur, lrecs, left);
+		lpp = xfs_btree_ptr_addr(cur, lrecs, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+
+#ifdef DEBUG
+		for (i = rrecs - 1; i >= 0; i--) {
+			error = xfs_btree_check_ptr(cur, rpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+
+		xfs_btree_shift_keys(cur, rkp, 1, rrecs);
+		xfs_btree_shift_ptrs(cur, rpp, 1, rrecs);
+
+#ifdef DEBUG
+		error = xfs_btree_check_ptr(cur, lpp, 0, level);
+		if (error)
+			goto error0;
+#endif
+
+		/* Now put the new data in, and log it. */
+		xfs_btree_copy_keys(cur, rkp, lkp, 1);
+		xfs_btree_copy_ptrs(cur, rpp, lpp, 1);
+
+		xfs_btree_log_keys(cur, rbp, 1, rrecs + 1);
+		xfs_btree_log_ptrs(cur, rbp, 1, rrecs + 1);
+
+		ASSERT(cur->bc_ops->keys_inorder(cur, rkp,
+			xfs_btree_key_addr(cur, 2, right)));
+	} else {
+		/* It's a leaf. make a hole in the records */
+		union xfs_btree_rec	*lrp;
+		union xfs_btree_rec	*rrp;
+
+		lrp = xfs_btree_rec_addr(cur, lrecs, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_shift_recs(cur, rrp, 1, rrecs);
+
+		/* Now put the new data in, and log it. */
+		xfs_btree_copy_recs(cur, rrp, lrp, 1);
+		xfs_btree_log_recs(cur, rbp, 1, rrecs + 1);
+
+		cur->bc_ops->init_key_from_rec(&key, rrp);
+		rkp = &key;
+
+		ASSERT(cur->bc_ops->recs_inorder(cur, rrp,
+			xfs_btree_rec_addr(cur, 2, right)));
+	}
+
+	/*
+	 * Decrement and log left's numrecs, bump and log right's numrecs.
+	 */
+	xfs_btree_set_numrecs(left, --lrecs);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS);
+
+	xfs_btree_set_numrecs(right, ++rrecs);
+	xfs_btree_log_block(cur, rbp, XFS_BB_NUMRECS);
+
+	/*
+	 * Using a temporary cursor, update the parent key values of the
+	 * block on the right.
+	 */
+	error = xfs_btree_dup_cursor(cur, &tcur);
+	if (error)
+		goto error0;
+	i = xfs_btree_lastrec(tcur, level);
+	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+	error = xfs_btree_increment(tcur, level, &i);
+	if (error)
+		goto error1;
+
+	error = xfs_btree_updkey(tcur, rkp, level + 1);
+	if (error)
+		goto error1;
+
+	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+
+error1:
+	XFS_BTREE_TRACE_CURSOR(tcur, XBT_ERROR);
+	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/*
+ * Split cur/level block in half.
+ * Return new block number and the key to its first
+ * record (to be inserted into parent).
+ */
+STATIC int					/* error */
+xfs_btree_split(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	union xfs_btree_ptr	*ptrp,
+	union xfs_btree_key	*key,
+	struct xfs_btree_cur	**curp,
+	int			*stat)		/* success/failure */
+{
+	union xfs_btree_ptr	lptr;		/* left sibling block ptr */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	union xfs_btree_ptr	rptr;		/* right sibling block ptr */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	union xfs_btree_ptr	rrptr;		/* right-right sibling ptr */
+	struct xfs_buf		*rrbp;		/* right-right buffer pointer */
+	struct xfs_btree_block	*rrblock;	/* right-right btree block */
+	int			lrecs;
+	int			rrecs;
+	int			src_index;
+	int			error;		/* error return value */
+#ifdef DEBUG
+	int			i;
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGIPK(cur, level, *ptrp, key);
+
+	XFS_BTREE_STATS_INC(cur, split);
+
+	/* Set up left block (current one). */
+	left = xfs_btree_get_block(cur, level, &lbp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, left, level, lbp);
+	if (error)
+		goto error0;
+#endif
+
+	xfs_btree_buf_to_ptr(cur, lbp, &lptr);
+
+	/* Allocate the new block. If we can't do it, we're toast. Give up. */
+	error = cur->bc_ops->alloc_block(cur, &lptr, &rptr, 1, stat);
+	if (error)
+		goto error0;
+	if (*stat == 0)
+		goto out0;
+	XFS_BTREE_STATS_INC(cur, alloc);
+
+	/* Set up the new block as "right". */
+	error = xfs_btree_get_buf_block(cur, &rptr, 0, &right, &rbp);
+	if (error)
+		goto error0;
+
+	/* Fill in the btree header for the new right block. */
+	xfs_btree_init_block(cur, xfs_btree_get_level(left), 0, right);
+
+	/*
+	 * Split the entries between the old and the new block evenly.
+	 * Make sure that if there's an odd number of entries now, that
+	 * each new block will have the same number of entries.
+	 */
+	lrecs = xfs_btree_get_numrecs(left);
+	rrecs = lrecs / 2;
+	if ((lrecs & 1) && cur->bc_ptrs[level] <= rrecs + 1)
+		rrecs++;
+	src_index = (lrecs - rrecs + 1);
+
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs);
+
+	/*
+	 * Copy btree block entries from the left block over to the
+	 * new block, the right. Update the right block and log the
+	 * changes.
+	 */
+	if (level > 0) {
+		/* It's a non-leaf.  Move keys and pointers. */
+		union xfs_btree_key	*lkp;	/* left btree key */
+		union xfs_btree_ptr	*lpp;	/* left address pointer */
+		union xfs_btree_key	*rkp;	/* right btree key */
+		union xfs_btree_ptr	*rpp;	/* right address pointer */
+
+		lkp = xfs_btree_key_addr(cur, src_index, left);
+		lpp = xfs_btree_ptr_addr(cur, src_index, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+
+#ifdef DEBUG
+		for (i = src_index; i < rrecs; i++) {
+			error = xfs_btree_check_ptr(cur, lpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+
+		xfs_btree_copy_keys(cur, rkp, lkp, rrecs);
+		xfs_btree_copy_ptrs(cur, rpp, lpp, rrecs);
+
+		xfs_btree_log_keys(cur, rbp, 1, rrecs);
+		xfs_btree_log_ptrs(cur, rbp, 1, rrecs);
+
+		/* Grab the keys to the entries moved to the right block */
+		xfs_btree_copy_keys(cur, key, rkp, 1);
+	} else {
+		/* It's a leaf.  Move records.  */
+		union xfs_btree_rec	*lrp;	/* left record pointer */
+		union xfs_btree_rec	*rrp;	/* right record pointer */
+
+		lrp = xfs_btree_rec_addr(cur, src_index, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_copy_recs(cur, rrp, lrp, rrecs);
+		xfs_btree_log_recs(cur, rbp, 1, rrecs);
+
+		cur->bc_ops->init_key_from_rec(key,
+			xfs_btree_rec_addr(cur, 1, right));
+	}
+
+
+	/*
+	 * Find the left block number by looking in the buffer.
+	 * Adjust numrecs, sibling pointers.
+	 */
+	xfs_btree_get_sibling(cur, left, &rrptr, XFS_BB_RIGHTSIB);
+	xfs_btree_set_sibling(cur, right, &rrptr, XFS_BB_RIGHTSIB);
+	xfs_btree_set_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
+	xfs_btree_set_sibling(cur, left, &rptr, XFS_BB_RIGHTSIB);
+
+	lrecs -= rrecs;
+	xfs_btree_set_numrecs(left, lrecs);
+	xfs_btree_set_numrecs(right, xfs_btree_get_numrecs(right) + rrecs);
+
+	xfs_btree_log_block(cur, rbp, XFS_BB_ALL_BITS);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
+
+	/*
+	 * If there's a block to the new block's right, make that block
+	 * point back to right instead of to left.
+	 */
+	if (!xfs_btree_ptr_is_null(cur, &rrptr)) {
+		error = xfs_btree_read_buf_block(cur, &rrptr, level,
+							0, &rrblock, &rrbp);
+		if (error)
+			goto error0;
+		xfs_btree_set_sibling(cur, rrblock, &rptr, XFS_BB_LEFTSIB);
+		xfs_btree_log_block(cur, rrbp, XFS_BB_LEFTSIB);
+	}
+	/*
+	 * If the cursor is really in the right block, move it there.
+	 * If it's just pointing past the last entry in left, then we'll
+	 * insert there, so don't change anything in that case.
+	 */
+	if (cur->bc_ptrs[level] > lrecs + 1) {
+		xfs_btree_setbuf(cur, level, rbp);
+		cur->bc_ptrs[level] -= lrecs;
+	}
+	/*
+	 * If there are more levels, we'll need another cursor which refers
+	 * the right block, no matter where this cursor was.
+	 */
+	if (level + 1 < cur->bc_nlevels) {
+		error = xfs_btree_dup_cursor(cur, curp);
+		if (error)
+			goto error0;
+		(*curp)->bc_ptrs[level + 1]++;
+	}
+	*ptrp = rptr;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Copy the old inode root contents into a real block and make the
+ * broot point to it.
+ */
+int						/* error */
+xfs_btree_new_iroot(
+	struct xfs_btree_cur	*cur,		/* btree cursor */
+	int			*logflags,	/* logging flags for inode */
+	int			*stat)		/* return status - 0 fail */
+{
+	struct xfs_buf		*cbp;		/* buffer for cblock */
+	struct xfs_btree_block	*block;		/* btree block */
+	struct xfs_btree_block	*cblock;	/* child btree block */
+	union xfs_btree_key	*ckp;		/* child key pointer */
+	union xfs_btree_ptr	*cpp;		/* child ptr pointer */
+	union xfs_btree_key	*kp;		/* pointer to btree key */
+	union xfs_btree_ptr	*pp;		/* pointer to block addr */
+	union xfs_btree_ptr	nptr;		/* new block addr */
+	int			level;		/* btree level */
+	int			error;		/* error return code */
+#ifdef DEBUG
+	int			i;		/* loop counter */
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, newroot);
+
+	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+
+	level = cur->bc_nlevels - 1;
+
+	block = xfs_btree_get_iroot(cur);
+	pp = xfs_btree_ptr_addr(cur, 1, block);
+
+	/* Allocate the new block. If we can't do it, we're toast. Give up. */
+	error = cur->bc_ops->alloc_block(cur, pp, &nptr, 1, stat);
+	if (error)
+		goto error0;
+	if (*stat == 0) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		return 0;
+	}
+	XFS_BTREE_STATS_INC(cur, alloc);
+
+	/* Copy the root into a real block. */
+	error = xfs_btree_get_buf_block(cur, &nptr, 0, &cblock, &cbp);
+	if (error)
+		goto error0;
+
+	memcpy(cblock, block, xfs_btree_block_len(cur));
+
+	be16_add_cpu(&block->bb_level, 1);
+	xfs_btree_set_numrecs(block, 1);
+	cur->bc_nlevels++;
+	cur->bc_ptrs[level + 1] = 1;
+
+	kp = xfs_btree_key_addr(cur, 1, block);
+	ckp = xfs_btree_key_addr(cur, 1, cblock);
+	xfs_btree_copy_keys(cur, ckp, kp, xfs_btree_get_numrecs(cblock));
+
+	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+#ifdef DEBUG
+	for (i = 0; i < be16_to_cpu(cblock->bb_numrecs); i++) {
+		error = xfs_btree_check_ptr(cur, pp, i, level);
+		if (error)
+			goto error0;
+	}
+#endif
+	xfs_btree_copy_ptrs(cur, cpp, pp, xfs_btree_get_numrecs(cblock));
+
+#ifdef DEBUG
+	error = xfs_btree_check_ptr(cur, &nptr, 0, level);
+	if (error)
+		goto error0;
+#endif
+	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
+
+	xfs_iroot_realloc(cur->bc_private.b.ip,
+			  1 - xfs_btree_get_numrecs(cblock),
+			  cur->bc_private.b.whichfork);
+
+	xfs_btree_setbuf(cur, level, cbp);
+
+	/*
+	 * Do all this logging at the end so that
+	 * the root is at the right level.
+	 */
+	xfs_btree_log_block(cur, cbp, XFS_BB_ALL_BITS);
+	xfs_btree_log_keys(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
+	xfs_btree_log_ptrs(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
+
+	*logflags |=
+		XFS_ILOG_CORE | XFS_ILOG_FBROOT(cur->bc_private.b.whichfork);
+	*stat = 1;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Allocate a new root block, fill it in.
+ */
+STATIC int				/* error */
+xfs_btree_new_root(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			*stat)	/* success/failure */
+{
+	struct xfs_btree_block	*block;	/* one half of the old root block */
+	struct xfs_buf		*bp;	/* buffer containing block */
+	int			error;	/* error return value */
+	struct xfs_buf		*lbp;	/* left buffer pointer */
+	struct xfs_btree_block	*left;	/* left btree block */
+	struct xfs_buf		*nbp;	/* new (root) buffer */
+	struct xfs_btree_block	*new;	/* new (root) btree block */
+	int			nptr;	/* new value for key index, 1 or 2 */
+	struct xfs_buf		*rbp;	/* right buffer pointer */
+	struct xfs_btree_block	*right;	/* right btree block */
+	union xfs_btree_ptr	rptr;
+	union xfs_btree_ptr	lptr;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, newroot);
+
+	/* initialise our start point from the cursor */
+	cur->bc_ops->init_ptr_from_cur(cur, &rptr);
+
+	/* Allocate the new block. If we can't do it, we're toast. Give up. */
+	error = cur->bc_ops->alloc_block(cur, &rptr, &lptr, 1, stat);
+	if (error)
+		goto error0;
+	if (*stat == 0)
+		goto out0;
+	XFS_BTREE_STATS_INC(cur, alloc);
+
+	/* Set up the new block. */
+	error = xfs_btree_get_buf_block(cur, &lptr, 0, &new, &nbp);
+	if (error)
+		goto error0;
+
+	/* Set the root in the holding structure  increasing the level by 1. */
+	cur->bc_ops->set_root(cur, &lptr, 1);
+
+	/*
+	 * At the previous root level there are now two blocks: the old root,
+	 * and the new block generated when it was split.  We don't know which
+	 * one the cursor is pointing at, so we set up variables "left" and
+	 * "right" for each case.
+	 */
+	block = xfs_btree_get_block(cur, cur->bc_nlevels - 1, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, cur->bc_nlevels - 1, bp);
+	if (error)
+		goto error0;
+#endif
+
+	xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
+	if (!xfs_btree_ptr_is_null(cur, &rptr)) {
+		/* Our block is left, pick up the right block. */
+		lbp = bp;
+		xfs_btree_buf_to_ptr(cur, lbp, &lptr);
+		left = block;
+		error = xfs_btree_read_buf_block(cur, &rptr,
+					cur->bc_nlevels - 1, 0, &right, &rbp);
+		if (error)
+			goto error0;
+		bp = rbp;
+		nptr = 1;
+	} else {
+		/* Our block is right, pick up the left block. */
+		rbp = bp;
+		xfs_btree_buf_to_ptr(cur, rbp, &rptr);
+		right = block;
+		xfs_btree_get_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
+		error = xfs_btree_read_buf_block(cur, &lptr,
+					cur->bc_nlevels - 1, 0, &left, &lbp);
+		if (error)
+			goto error0;
+		bp = lbp;
+		nptr = 2;
+	}
+	/* Fill in the new block's btree header and log it. */
+	xfs_btree_init_block(cur, cur->bc_nlevels, 2, new);
+	xfs_btree_log_block(cur, nbp, XFS_BB_ALL_BITS);
+	ASSERT(!xfs_btree_ptr_is_null(cur, &lptr) &&
+			!xfs_btree_ptr_is_null(cur, &rptr));
+
+	/* Fill in the key data in the new root. */
+	if (xfs_btree_get_level(left) > 0) {
+		xfs_btree_copy_keys(cur,
+				xfs_btree_key_addr(cur, 1, new),
+				xfs_btree_key_addr(cur, 1, left), 1);
+		xfs_btree_copy_keys(cur,
+				xfs_btree_key_addr(cur, 2, new),
+				xfs_btree_key_addr(cur, 1, right), 1);
+	} else {
+		cur->bc_ops->init_key_from_rec(
+				xfs_btree_key_addr(cur, 1, new),
+				xfs_btree_rec_addr(cur, 1, left));
+		cur->bc_ops->init_key_from_rec(
+				xfs_btree_key_addr(cur, 2, new),
+				xfs_btree_rec_addr(cur, 1, right));
+	}
+	xfs_btree_log_keys(cur, nbp, 1, 2);
+
+	/* Fill in the pointer data in the new root. */
+	xfs_btree_copy_ptrs(cur,
+		xfs_btree_ptr_addr(cur, 1, new), &lptr, 1);
+	xfs_btree_copy_ptrs(cur,
+		xfs_btree_ptr_addr(cur, 2, new), &rptr, 1);
+	xfs_btree_log_ptrs(cur, nbp, 1, 2);
+
+	/* Fix up the cursor. */
+	xfs_btree_setbuf(cur, cur->bc_nlevels, nbp);
+	cur->bc_ptrs[cur->bc_nlevels] = nptr;
+	cur->bc_nlevels++;
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 0;
+	return 0;
+}
+
+STATIC int
+xfs_btree_make_block_unfull(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			level,	/* btree level */
+	int			numrecs,/* # of recs in block */
+	int			*oindex,/* old tree index */
+	int			*index,	/* new tree index */
+	union xfs_btree_ptr	*nptr,	/* new btree ptr */
+	struct xfs_btree_cur	**ncur,	/* new btree cursor */
+	union xfs_btree_rec	*nrec,	/* new record */
+	int			*stat)
+{
+	union xfs_btree_key	key;	/* new btree key value */
+	int			error = 0;
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == cur->bc_nlevels - 1) {
+	    	struct xfs_inode *ip = cur->bc_private.b.ip;
+
+		if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) {
+			/* A root block that can be made bigger. */
+
+			xfs_iroot_realloc(ip, 1, cur->bc_private.b.whichfork);
+		} else {
+			/* A root block that needs replacing */
+			int	logflags = 0;
+
+			error = xfs_btree_new_iroot(cur, &logflags, stat);
+			if (error || *stat == 0)
+				return error;
+
+			xfs_trans_log_inode(cur->bc_tp, ip, logflags);
+		}
+
+		return 0;
+	}
+
+	/* First, try shifting an entry to the right neighbor. */
+	error = xfs_btree_rshift(cur, level, stat);
+	if (error || *stat)
+		return error;
+
+	/* Next, try shifting an entry to the left neighbor. */
+	error = xfs_btree_lshift(cur, level, stat);
+	if (error)
+		return error;
+
+	if (*stat) {
+		*oindex = *index = cur->bc_ptrs[level];
+		return 0;
+	}
+
+	/*
+	 * Next, try splitting the current block in half.
+	 *
+	 * If this works we have to re-set our variables because we
+	 * could be in a different block now.
+	 */
+	error = xfs_btree_split(cur, level, nptr, &key, ncur, stat);
+	if (error || *stat == 0)
+		return error;
+
+
+	*index = cur->bc_ptrs[level];
+	cur->bc_ops->init_rec_from_key(&key, nrec);
+	return 0;
+}
+
+/*
+ * Insert one record/level.  Return information to the caller
+ * allowing the next level up to proceed if necessary.
+ */
+STATIC int
+xfs_btree_insrec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	int			level,	/* level to insert record at */
+	union xfs_btree_ptr	*ptrp,	/* i/o: block number inserted */
+	union xfs_btree_rec	*recp,	/* i/o: record data inserted */
+	struct xfs_btree_cur	**curp,	/* output: new cursor replacing cur */
+	int			*stat)	/* success/failure */
+{
+	struct xfs_btree_block	*block;	/* btree block */
+	struct xfs_buf		*bp;	/* buffer for block */
+	union xfs_btree_key	key;	/* btree key */
+	union xfs_btree_ptr	nptr;	/* new block ptr */
+	struct xfs_btree_cur	*ncur;	/* new btree cursor */
+	union xfs_btree_rec	nrec;	/* new record count */
+	int			optr;	/* old key/record index */
+	int			ptr;	/* key/record index */
+	int			numrecs;/* number of records */
+	int			error;	/* error return value */
+#ifdef DEBUG
+	int			i;
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGIPR(cur, level, *ptrp, recp);
+
+	ncur = NULL;
+
+	/*
+	 * If we have an external root pointer, and we've made it to the
+	 * root level, allocate a new root block and we're done.
+	 */
+	if (!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    (level >= cur->bc_nlevels)) {
+		error = xfs_btree_new_root(cur, stat);
+		xfs_btree_set_ptr_null(cur, ptrp);
+
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		return error;
+	}
+
+	/* If we're off the left edge, return failure. */
+	ptr = cur->bc_ptrs[level];
+	if (ptr == 0) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+
+	/* Make a key out of the record data to be inserted, and save it. */
+	cur->bc_ops->init_key_from_rec(&key, recp);
+
+	optr = ptr;
+
+	XFS_BTREE_STATS_INC(cur, insrec);
+
+	/* Get pointers to the btree buffer and block. */
+	block = xfs_btree_get_block(cur, level, &bp);
+	numrecs = xfs_btree_get_numrecs(block);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+
+	/* Check that the new entry is being inserted in the right place. */
+	if (ptr <= numrecs) {
+		if (level == 0) {
+			ASSERT(cur->bc_ops->recs_inorder(cur, recp,
+				xfs_btree_rec_addr(cur, ptr, block)));
+		} else {
+			ASSERT(cur->bc_ops->keys_inorder(cur, &key,
+				xfs_btree_key_addr(cur, ptr, block)));
+		}
+	}
+#endif
+
+	/*
+	 * If the block is full, we can't insert the new entry until we
+	 * make the block un-full.
+	 */
+	xfs_btree_set_ptr_null(cur, &nptr);
+	if (numrecs == cur->bc_ops->get_maxrecs(cur, level)) {
+		error = xfs_btree_make_block_unfull(cur, level, numrecs,
+					&optr, &ptr, &nptr, &ncur, &nrec, stat);
+		if (error || *stat == 0)
+			goto error0;
+	}
+
+	/*
+	 * The current block may have changed if the block was
+	 * previously full and we have just made space in it.
+	 */
+	block = xfs_btree_get_block(cur, level, &bp);
+	numrecs = xfs_btree_get_numrecs(block);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		return error;
+#endif
+
+	/*
+	 * At this point we know there's room for our new entry in the block
+	 * we're pointing at.
+	 */
+	XFS_BTREE_STATS_ADD(cur, moves, numrecs - ptr + 1);
+
+	if (level > 0) {
+		/* It's a nonleaf. make a hole in the keys and ptrs */
+		union xfs_btree_key	*kp;
+		union xfs_btree_ptr	*pp;
+
+		kp = xfs_btree_key_addr(cur, ptr, block);
+		pp = xfs_btree_ptr_addr(cur, ptr, block);
+
+#ifdef DEBUG
+		for (i = numrecs - ptr; i >= 0; i--) {
+			error = xfs_btree_check_ptr(cur, pp, i, level);
+			if (error)
+				return error;
+		}
+#endif
+
+		xfs_btree_shift_keys(cur, kp, 1, numrecs - ptr + 1);
+		xfs_btree_shift_ptrs(cur, pp, 1, numrecs - ptr + 1);
+
+#ifdef DEBUG
+		error = xfs_btree_check_ptr(cur, ptrp, 0, level);
+		if (error)
+			goto error0;
+#endif
+
+		/* Now put the new data in, bump numrecs and log it. */
+		xfs_btree_copy_keys(cur, kp, &key, 1);
+		xfs_btree_copy_ptrs(cur, pp, ptrp, 1);
+		numrecs++;
+		xfs_btree_set_numrecs(block, numrecs);
+		xfs_btree_log_ptrs(cur, bp, ptr, numrecs);
+		xfs_btree_log_keys(cur, bp, ptr, numrecs);
+#ifdef DEBUG
+		if (ptr < numrecs) {
+			ASSERT(cur->bc_ops->keys_inorder(cur, kp,
+				xfs_btree_key_addr(cur, ptr + 1, block)));
+		}
+#endif
+	} else {
+		/* It's a leaf. make a hole in the records */
+		union xfs_btree_rec             *rp;
+
+		rp = xfs_btree_rec_addr(cur, ptr, block);
+
+		xfs_btree_shift_recs(cur, rp, 1, numrecs - ptr + 1);
+
+		/* Now put the new data in, bump numrecs and log it. */
+		xfs_btree_copy_recs(cur, rp, recp, 1);
+		xfs_btree_set_numrecs(block, ++numrecs);
+		xfs_btree_log_recs(cur, bp, ptr, numrecs);
+#ifdef DEBUG
+		if (ptr < numrecs) {
+			ASSERT(cur->bc_ops->recs_inorder(cur, rp,
+				xfs_btree_rec_addr(cur, ptr + 1, block)));
+		}
+#endif
+	}
+
+	/* Log the new number of records in the btree header. */
+	xfs_btree_log_block(cur, bp, XFS_BB_NUMRECS);
+
+	/* If we inserted at the start of a block, update the parents' keys. */
+	if (optr == 1) {
+		error = xfs_btree_updkey(cur, &key, level + 1);
+		if (error)
+			goto error0;
+	}
+
+	/*
+	 * If we are tracking the last record in the tree and
+	 * we are at the far right edge of the tree, update it.
+	 */
+	if (xfs_btree_is_lastrec(cur, block, level)) {
+		cur->bc_ops->update_lastrec(cur, block, recp,
+					    ptr, LASTREC_INSREC);
+	}
+
+	/*
+	 * Return the new block number, if any.
+	 * If there is one, give back a record value and a cursor too.
+	 */
+	*ptrp = nptr;
+	if (!xfs_btree_ptr_is_null(cur, &nptr)) {
+		*recp = nrec;
+		*curp = ncur;
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Insert the record at the point referenced by cur.
+ *
+ * A multi-level split of the tree on insert will invalidate the original
+ * cursor.  All callers of this function should assume that the cursor is
+ * no longer valid and revalidate it.
+ */
+int
+xfs_btree_insert(
+	struct xfs_btree_cur	*cur,
+	int			*stat)
+{
+	int			error;	/* error return value */
+	int			i;	/* result value, 0 for failure */
+	int			level;	/* current level number in btree */
+	union xfs_btree_ptr	nptr;	/* new block number (split result) */
+	struct xfs_btree_cur	*ncur;	/* new cursor (split result) */
+	struct xfs_btree_cur	*pcur;	/* previous level's cursor */
+	union xfs_btree_rec	rec;	/* record to insert */
+
+	level = 0;
+	ncur = NULL;
+	pcur = cur;
+
+	xfs_btree_set_ptr_null(cur, &nptr);
+	cur->bc_ops->init_rec_from_cur(cur, &rec);
+
+	/*
+	 * Loop going up the tree, starting at the leaf level.
+	 * Stop when we don't get a split block, that must mean that
+	 * the insert is finished with this level.
+	 */
+	do {
+		/*
+		 * Insert nrec/nptr into this level of the tree.
+		 * Note if we fail, nptr will be null.
+		 */
+		error = xfs_btree_insrec(pcur, level, &nptr, &rec, &ncur, &i);
+		if (error) {
+			if (pcur != cur)
+				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
+			goto error0;
+		}
+
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		level++;
+
+		/*
+		 * See if the cursor we just used is trash.
+		 * Can't trash the caller's cursor, but otherwise we should
+		 * if ncur is a new cursor or we're about to be done.
+		 */
+		if (pcur != cur &&
+		    (ncur || xfs_btree_ptr_is_null(cur, &nptr))) {
+			/* Save the state from the cursor before we trash it */
+			if (cur->bc_ops->update_cursor)
+				cur->bc_ops->update_cursor(pcur, cur);
+			cur->bc_nlevels = pcur->bc_nlevels;
+			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
+		}
+		/* If we got a new cursor, switch to it. */
+		if (ncur) {
+			pcur = ncur;
+			ncur = NULL;
+		}
+	} while (!xfs_btree_ptr_is_null(cur, &nptr));
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = i;
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Try to merge a non-leaf block back into the inode root.
+ *
+ * Note: the killroot names comes from the fact that we're effectively
+ * killing the old root block.  But because we can't just delete the
+ * inode we have to copy the single block it was pointing to into the
+ * inode.
+ */
+int
+xfs_btree_kill_iroot(
+	struct xfs_btree_cur	*cur)
+{
+	int			whichfork = cur->bc_private.b.whichfork;
+	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	struct xfs_btree_block	*block;
+	struct xfs_btree_block	*cblock;
+	union xfs_btree_key	*kp;
+	union xfs_btree_key	*ckp;
+	union xfs_btree_ptr	*pp;
+	union xfs_btree_ptr	*cpp;
+	struct xfs_buf		*cbp;
+	int			level;
+	int			index;
+	int			numrecs;
+#ifdef DEBUG
+	union xfs_btree_ptr	ptr;
+	int			i;
+#endif
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+
+	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_nlevels > 1);
+
+	/*
+	 * Don't deal with the root block needs to be a leaf case.
+	 * We're just going to turn the thing back into extents anyway.
+	 */
+	level = cur->bc_nlevels - 1;
+	if (level == 1)
+		goto out0;
+
+	/*
+	 * Give up if the root has multiple children.
+	 */
+	block = xfs_btree_get_iroot(cur);
+	if (xfs_btree_get_numrecs(block) != 1)
+		goto out0;
+
+	cblock = xfs_btree_get_block(cur, level - 1, &cbp);
+	numrecs = xfs_btree_get_numrecs(cblock);
+
+	/*
+	 * Only do this if the next level will fit.
+	 * Then the data must be copied up to the inode,
+	 * instead of freeing the root you free the next level.
+	 */
+	if (numrecs > cur->bc_ops->get_dmaxrecs(cur, level))
+		goto out0;
+
+	XFS_BTREE_STATS_INC(cur, killroot);
+
+#ifdef DEBUG
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_LEFTSIB);
+	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
+	xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB);
+	ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
+#endif
+
+	index = numrecs - cur->bc_ops->get_maxrecs(cur, level);
+	if (index) {
+		xfs_iroot_realloc(cur->bc_private.b.ip, index,
+				  cur->bc_private.b.whichfork);
+		block = ifp->if_broot;
+	}
+
+	be16_add_cpu(&block->bb_numrecs, index);
+	ASSERT(block->bb_numrecs == cblock->bb_numrecs);
+
+	kp = xfs_btree_key_addr(cur, 1, block);
+	ckp = xfs_btree_key_addr(cur, 1, cblock);
+	xfs_btree_copy_keys(cur, kp, ckp, numrecs);
+
+	pp = xfs_btree_ptr_addr(cur, 1, block);
+	cpp = xfs_btree_ptr_addr(cur, 1, cblock);
+#ifdef DEBUG
+	for (i = 0; i < numrecs; i++) {
+		int		error;
+
+		error = xfs_btree_check_ptr(cur, cpp, i, level - 1);
+		if (error) {
+			XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+			return error;
+		}
+	}
+#endif
+	xfs_btree_copy_ptrs(cur, pp, cpp, numrecs);
+
+	cur->bc_ops->free_block(cur, cbp);
+	XFS_BTREE_STATS_INC(cur, free);
+
+	cur->bc_bufs[level - 1] = NULL;
+	be16_add_cpu(&block->bb_level, -1);
+	xfs_trans_log_inode(cur->bc_tp, ip,
+		XFS_ILOG_CORE | XFS_ILOG_FBROOT(cur->bc_private.b.whichfork));
+	cur->bc_nlevels--;
+out0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	return 0;
+}
+
+STATIC int
+xfs_btree_dec_cursor(
+	struct xfs_btree_cur	*cur,
+	int			level,
+	int			*stat)
+{
+	int			error;
+	int			i;
+
+	if (level > 0) {
+		error = xfs_btree_decrement(cur, level, &i);
+		if (error)
+			return error;
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = 1;
+	return 0;
+}
+
+/*
+ * Single level of the btree record deletion routine.
+ * Delete record pointed to by cur/level.
+ * Remove the record from its block then rebalance the tree.
+ * Return 0 for error, 1 for done, 2 to go on to the next level.
+ */
+STATIC int					/* error */
+xfs_btree_delrec(
+	struct xfs_btree_cur	*cur,		/* btree cursor */
+	int			level,		/* level removing record from */
+	int			*stat)		/* fail/done/go-on */
+{
+	struct xfs_btree_block	*block;		/* btree block */
+	union xfs_btree_ptr	cptr;		/* current block ptr */
+	struct xfs_buf		*bp;		/* buffer for block */
+	int			error;		/* error return value */
+	int			i;		/* loop counter */
+	union xfs_btree_key	key;		/* storage for keyp */
+	union xfs_btree_key	*keyp = &key;	/* passed to the next level */
+	union xfs_btree_ptr	lptr;		/* left sibling block ptr */
+	struct xfs_buf		*lbp;		/* left buffer pointer */
+	struct xfs_btree_block	*left;		/* left btree block */
+	int			lrecs = 0;	/* left record count */
+	int			ptr;		/* key/record index */
+	union xfs_btree_ptr	rptr;		/* right sibling block ptr */
+	struct xfs_buf		*rbp;		/* right buffer pointer */
+	struct xfs_btree_block	*right;		/* right btree block */
+	struct xfs_btree_block	*rrblock;	/* right-right btree block */
+	struct xfs_buf		*rrbp;		/* right-right buffer pointer */
+	int			rrecs = 0;	/* right record count */
+	struct xfs_btree_cur	*tcur;		/* temporary btree cursor */
+	int			numrecs;	/* temporary numrec count */
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_TRACE_ARGI(cur, level);
+
+	tcur = NULL;
+
+	/* Get the index of the entry being deleted, check for nothing there. */
+	ptr = cur->bc_ptrs[level];
+	if (ptr == 0) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+
+	/* Get the buffer & block containing the record or key/ptr. */
+	block = xfs_btree_get_block(cur, level, &bp);
+	numrecs = xfs_btree_get_numrecs(block);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, level, bp);
+	if (error)
+		goto error0;
+#endif
+
+	/* Fail if we're off the end of the block. */
+	if (ptr > numrecs) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+		*stat = 0;
+		return 0;
+	}
+
+	XFS_BTREE_STATS_INC(cur, delrec);
+	XFS_BTREE_STATS_ADD(cur, moves, numrecs - ptr);
+
+	/* Excise the entries being deleted. */
+	if (level > 0) {
+		/* It's a nonleaf. operate on keys and ptrs */
+		union xfs_btree_key	*lkp;
+		union xfs_btree_ptr	*lpp;
+
+		lkp = xfs_btree_key_addr(cur, ptr + 1, block);
+		lpp = xfs_btree_ptr_addr(cur, ptr + 1, block);
+
+#ifdef DEBUG
+		for (i = 0; i < numrecs - ptr; i++) {
+			error = xfs_btree_check_ptr(cur, lpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+
+		if (ptr < numrecs) {
+			xfs_btree_shift_keys(cur, lkp, -1, numrecs - ptr);
+			xfs_btree_shift_ptrs(cur, lpp, -1, numrecs - ptr);
+			xfs_btree_log_keys(cur, bp, ptr, numrecs - 1);
+			xfs_btree_log_ptrs(cur, bp, ptr, numrecs - 1);
+		}
+
+		/*
+		 * If it's the first record in the block, we'll need to pass a
+		 * key up to the next level (updkey).
+		 */
+		if (ptr == 1)
+			keyp = xfs_btree_key_addr(cur, 1, block);
+	} else {
+		/* It's a leaf. operate on records */
+		if (ptr < numrecs) {
+			xfs_btree_shift_recs(cur,
+				xfs_btree_rec_addr(cur, ptr + 1, block),
+				-1, numrecs - ptr);
+			xfs_btree_log_recs(cur, bp, ptr, numrecs - 1);
+		}
+
+		/*
+		 * If it's the first record in the block, we'll need a key
+		 * structure to pass up to the next level (updkey).
+		 */
+		if (ptr == 1) {
+			cur->bc_ops->init_key_from_rec(&key,
+					xfs_btree_rec_addr(cur, 1, block));
+			keyp = &key;
+		}
+	}
+
+	/*
+	 * Decrement and log the number of entries in the block.
+	 */
+	xfs_btree_set_numrecs(block, --numrecs);
+	xfs_btree_log_block(cur, bp, XFS_BB_NUMRECS);
+
+	/*
+	 * If we are tracking the last record in the tree and
+	 * we are at the far right edge of the tree, update it.
+	 */
+	if (xfs_btree_is_lastrec(cur, block, level)) {
+		cur->bc_ops->update_lastrec(cur, block, NULL,
+					    ptr, LASTREC_DELREC);
+	}
+
+	/*
+	 * We're at the root level.  First, shrink the root block in-memory.
+	 * Try to get rid of the next level down.  If we can't then there's
+	 * nothing left to do.
+	 */
+	if (level == cur->bc_nlevels - 1) {
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+			xfs_iroot_realloc(cur->bc_private.b.ip, -1,
+					  cur->bc_private.b.whichfork);
+
+			error = xfs_btree_kill_iroot(cur);
+			if (error)
+				goto error0;
+
+			error = xfs_btree_dec_cursor(cur, level, stat);
+			if (error)
+				goto error0;
+			*stat = 1;
+			return 0;
+		}
+
+		/*
+		 * If this is the root level, and there's only one entry left,
+		 * and it's NOT the leaf level, then we can get rid of this
+		 * level.
+		 */
+		if (numrecs == 1 && level > 0) {
+			union xfs_btree_ptr	*pp;
+			/*
+			 * pp is still set to the first pointer in the block.
+			 * Make it the new root of the btree.
+			 */
+			pp = xfs_btree_ptr_addr(cur, 1, block);
+			error = cur->bc_ops->kill_root(cur, bp, level, pp);
+			if (error)
+				goto error0;
+		} else if (level > 0) {
+			error = xfs_btree_dec_cursor(cur, level, stat);
+			if (error)
+				goto error0;
+		}
+		*stat = 1;
+		return 0;
+	}
+
+	/*
+	 * If we deleted the leftmost entry in the block, update the
+	 * key values above us in the tree.
+	 */
+	if (ptr == 1) {
+		error = xfs_btree_updkey(cur, keyp, level + 1);
+		if (error)
+			goto error0;
+	}
+
+	/*
+	 * If the number of records remaining in the block is at least
+	 * the minimum, we're done.
+	 */
+	if (numrecs >= cur->bc_ops->get_minrecs(cur, level)) {
+		error = xfs_btree_dec_cursor(cur, level, stat);
+		if (error)
+			goto error0;
+		return 0;
+	}
+
+	/*
+	 * Otherwise, we have to move some records around to keep the
+	 * tree balanced.  Look at the left and right sibling blocks to
+	 * see if we can re-balance by moving only one record.
+	 */
+	xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB);
+	xfs_btree_get_sibling(cur, block, &lptr, XFS_BB_LEFTSIB);
+
+	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+		/*
+		 * One child of root, need to get a chance to copy its contents
+		 * into the root and delete it. Can't go up to next level,
+		 * there's nothing to delete there.
+		 */
+		if (xfs_btree_ptr_is_null(cur, &rptr) &&
+		    xfs_btree_ptr_is_null(cur, &lptr) &&
+		    level == cur->bc_nlevels - 2) {
+			error = xfs_btree_kill_iroot(cur);
+			if (!error)
+				error = xfs_btree_dec_cursor(cur, level, stat);
+			if (error)
+				goto error0;
+			return 0;
+		}
+	}
+
+	ASSERT(!xfs_btree_ptr_is_null(cur, &rptr) ||
+	       !xfs_btree_ptr_is_null(cur, &lptr));
+
+	/*
+	 * Duplicate the cursor so our btree manipulations here won't
+	 * disrupt the next level up.
+	 */
+	error = xfs_btree_dup_cursor(cur, &tcur);
+	if (error)
+		goto error0;
+
+	/*
+	 * If there's a right sibling, see if it's ok to shift an entry
+	 * out of it.
+	 */
+	if (!xfs_btree_ptr_is_null(cur, &rptr)) {
+		/*
+		 * Move the temp cursor to the last entry in the next block.
+		 * Actually any entry but the first would suffice.
+		 */
+		i = xfs_btree_lastrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		error = xfs_btree_increment(tcur, level, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		i = xfs_btree_lastrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		/* Grab a pointer to the block. */
+		right = xfs_btree_get_block(tcur, level, &rbp);
+#ifdef DEBUG
+		error = xfs_btree_check_block(tcur, right, level, rbp);
+		if (error)
+			goto error0;
+#endif
+		/* Grab the current block number, for future use. */
+		xfs_btree_get_sibling(tcur, right, &cptr, XFS_BB_LEFTSIB);
+
+		/*
+		 * If right block is full enough so that removing one entry
+		 * won't make it too empty, and left-shifting an entry out
+		 * of right to us works, we're done.
+		 */
+		if (xfs_btree_get_numrecs(right) - 1 >=
+		    cur->bc_ops->get_minrecs(tcur, level)) {
+			error = xfs_btree_lshift(tcur, level, &i);
+			if (error)
+				goto error0;
+			if (i) {
+				ASSERT(xfs_btree_get_numrecs(block) >=
+				       cur->bc_ops->get_minrecs(tcur, level));
+
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+				tcur = NULL;
+
+				error = xfs_btree_dec_cursor(cur, level, stat);
+				if (error)
+					goto error0;
+				return 0;
+			}
+		}
+
+		/*
+		 * Otherwise, grab the number of records in right for
+		 * future reference, and fix up the temp cursor to point
+		 * to our block again (last record).
+		 */
+		rrecs = xfs_btree_get_numrecs(right);
+		if (!xfs_btree_ptr_is_null(cur, &lptr)) {
+			i = xfs_btree_firstrec(tcur, level);
+			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+			error = xfs_btree_decrement(tcur, level, &i);
+			if (error)
+				goto error0;
+			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		}
+	}
+
+	/*
+	 * If there's a left sibling, see if it's ok to shift an entry
+	 * out of it.
+	 */
+	if (!xfs_btree_ptr_is_null(cur, &lptr)) {
+		/*
+		 * Move the temp cursor to the first entry in the
+		 * previous block.
+		 */
+		i = xfs_btree_firstrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		error = xfs_btree_decrement(tcur, level, &i);
+		if (error)
+			goto error0;
+		i = xfs_btree_firstrec(tcur, level);
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		/* Grab a pointer to the block. */
+		left = xfs_btree_get_block(tcur, level, &lbp);
+#ifdef DEBUG
+		error = xfs_btree_check_block(cur, left, level, lbp);
+		if (error)
+			goto error0;
+#endif
+		/* Grab the current block number, for future use. */
+		xfs_btree_get_sibling(tcur, left, &cptr, XFS_BB_RIGHTSIB);
+
+		/*
+		 * If left block is full enough so that removing one entry
+		 * won't make it too empty, and right-shifting an entry out
+		 * of left to us works, we're done.
+		 */
+		if (xfs_btree_get_numrecs(left) - 1 >=
+		    cur->bc_ops->get_minrecs(tcur, level)) {
+			error = xfs_btree_rshift(tcur, level, &i);
+			if (error)
+				goto error0;
+			if (i) {
+				ASSERT(xfs_btree_get_numrecs(block) >=
+				       cur->bc_ops->get_minrecs(tcur, level));
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+				tcur = NULL;
+				if (level == 0)
+					cur->bc_ptrs[0]++;
+				XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+				*stat = 1;
+				return 0;
+			}
+		}
+
+		/*
+		 * Otherwise, grab the number of records in right for
+		 * future reference.
+		 */
+		lrecs = xfs_btree_get_numrecs(left);
+	}
+
+	/* Delete the temp cursor, we're done with it. */
+	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+	tcur = NULL;
+
+	/* If here, we need to do a join to keep the tree balanced. */
+	ASSERT(!xfs_btree_ptr_is_null(cur, &cptr));
+
+	if (!xfs_btree_ptr_is_null(cur, &lptr) &&
+	    lrecs + xfs_btree_get_numrecs(block) <=
+			cur->bc_ops->get_maxrecs(cur, level)) {
+		/*
+		 * Set "right" to be the starting block,
+		 * "left" to be the left neighbor.
+		 */
+		rptr = cptr;
+		right = block;
+		rbp = bp;
+		error = xfs_btree_read_buf_block(cur, &lptr, level,
+							0, &left, &lbp);
+		if (error)
+			goto error0;
+
+	/*
+	 * If that won't work, see if we can join with the right neighbor block.
+	 */
+	} else if (!xfs_btree_ptr_is_null(cur, &rptr) &&
+		   rrecs + xfs_btree_get_numrecs(block) <=
+			cur->bc_ops->get_maxrecs(cur, level)) {
+		/*
+		 * Set "left" to be the starting block,
+		 * "right" to be the right neighbor.
+		 */
+		lptr = cptr;
+		left = block;
+		lbp = bp;
+		error = xfs_btree_read_buf_block(cur, &rptr, level,
+							0, &right, &rbp);
+		if (error)
+			goto error0;
+
+	/*
+	 * Otherwise, we can't fix the imbalance.
+	 * Just return.  This is probably a logic error, but it's not fatal.
+	 */
+	} else {
+		error = xfs_btree_dec_cursor(cur, level, stat);
+		if (error)
+			goto error0;
+		return 0;
+	}
+
+	rrecs = xfs_btree_get_numrecs(right);
+	lrecs = xfs_btree_get_numrecs(left);
+
+	/*
+	 * We're now going to join "left" and "right" by moving all the stuff
+	 * in "right" to "left" and deleting "right".
+	 */
+	XFS_BTREE_STATS_ADD(cur, moves, rrecs);
+	if (level > 0) {
+		/* It's a non-leaf.  Move keys and pointers. */
+		union xfs_btree_key	*lkp;	/* left btree key */
+		union xfs_btree_ptr	*lpp;	/* left address pointer */
+		union xfs_btree_key	*rkp;	/* right btree key */
+		union xfs_btree_ptr	*rpp;	/* right address pointer */
+
+		lkp = xfs_btree_key_addr(cur, lrecs + 1, left);
+		lpp = xfs_btree_ptr_addr(cur, lrecs + 1, left);
+		rkp = xfs_btree_key_addr(cur, 1, right);
+		rpp = xfs_btree_ptr_addr(cur, 1, right);
+#ifdef DEBUG
+		for (i = 1; i < rrecs; i++) {
+			error = xfs_btree_check_ptr(cur, rpp, i, level);
+			if (error)
+				goto error0;
+		}
+#endif
+		xfs_btree_copy_keys(cur, lkp, rkp, rrecs);
+		xfs_btree_copy_ptrs(cur, lpp, rpp, rrecs);
+
+		xfs_btree_log_keys(cur, lbp, lrecs + 1, lrecs + rrecs);
+		xfs_btree_log_ptrs(cur, lbp, lrecs + 1, lrecs + rrecs);
+	} else {
+		/* It's a leaf.  Move records.  */
+		union xfs_btree_rec	*lrp;	/* left record pointer */
+		union xfs_btree_rec	*rrp;	/* right record pointer */
+
+		lrp = xfs_btree_rec_addr(cur, lrecs + 1, left);
+		rrp = xfs_btree_rec_addr(cur, 1, right);
+
+		xfs_btree_copy_recs(cur, lrp, rrp, rrecs);
+		xfs_btree_log_recs(cur, lbp, lrecs + 1, lrecs + rrecs);
+	}
+
+	XFS_BTREE_STATS_INC(cur, join);
+
+	/*
+	 * Fix up the the number of records and right block pointer in the
+	 * surviving block, and log it.
+	 */
+	xfs_btree_set_numrecs(left, lrecs + rrecs);
+	xfs_btree_get_sibling(cur, right, &cptr, XFS_BB_RIGHTSIB),
+	xfs_btree_set_sibling(cur, left, &cptr, XFS_BB_RIGHTSIB);
+	xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
+
+	/* If there is a right sibling, point it to the remaining block. */
+	xfs_btree_get_sibling(cur, left, &cptr, XFS_BB_RIGHTSIB);
+	if (!xfs_btree_ptr_is_null(cur, &cptr)) {
+		error = xfs_btree_read_buf_block(cur, &cptr, level,
+							0, &rrblock, &rrbp);
+		if (error)
+			goto error0;
+		xfs_btree_set_sibling(cur, rrblock, &lptr, XFS_BB_LEFTSIB);
+		xfs_btree_log_block(cur, rrbp, XFS_BB_LEFTSIB);
+	}
+
+	/* Free the deleted block. */
+	error = cur->bc_ops->free_block(cur, rbp);
+	if (error)
+		goto error0;
+	XFS_BTREE_STATS_INC(cur, free);
+
+	/*
+	 * If we joined with the left neighbor, set the buffer in the
+	 * cursor to the left block, and fix up the index.
+	 */
+	if (bp != lbp) {
+		cur->bc_bufs[level] = lbp;
+		cur->bc_ptrs[level] += lrecs;
+		cur->bc_ra[level] = 0;
+	}
+	/*
+	 * If we joined with the right neighbor and there's a level above
+	 * us, increment the cursor at that level.
+	 */
+	else if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) ||
+		   (level + 1 < cur->bc_nlevels)) {
+		error = xfs_btree_increment(cur, level + 1, &i);
+		if (error)
+			goto error0;
+	}
+
+	/*
+	 * Readjust the ptr at this level if it's not a leaf, since it's
+	 * still pointing at the deletion point, which makes the cursor
+	 * inconsistent.  If this makes the ptr 0, the caller fixes it up.
+	 * We can't use decrement because it would change the next level up.
+	 */
+	if (level > 0)
+		cur->bc_ptrs[level]--;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	/* Return value means the next level up has something to do. */
+	*stat = 2;
+	return 0;
+
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	if (tcur)
+		xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/*
+ * Delete the record pointed to by cur.
+ * The cursor refers to the place where the record was (could be inserted)
+ * when the operation returns.
+ */
+int					/* error */
+xfs_btree_delete(
+	struct xfs_btree_cur	*cur,
+	int			*stat)	/* success/failure */
+{
+	int			error;	/* error return value */
+	int			level;
+	int			i;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+
+	/*
+	 * Go up the tree, starting at leaf level.
+	 *
+	 * If 2 is returned then a join was done; go to the next level.
+	 * Otherwise we are done.
+	 */
+	for (level = 0, i = 2; i == 2; level++) {
+		error = xfs_btree_delrec(cur, level, &i);
+		if (error)
+			goto error0;
+	}
+
+	if (i == 0) {
+		for (level = 1; level < cur->bc_nlevels; level++) {
+			if (cur->bc_ptrs[level] == 0) {
+				error = xfs_btree_decrement(cur, level, &i);
+				if (error)
+					goto error0;
+				break;
+			}
+		}
+	}
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+	*stat = i;
+	return 0;
+error0:
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	return error;
+}
+
+/*
+ * Get the data from the pointed-to record.
+ */
+int					/* error */
+xfs_btree_get_rec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	union xfs_btree_rec	**recp,	/* output: btree record */
+	int			*stat)	/* output: success/failure */
+{
+	struct xfs_btree_block	*block;	/* btree block */
+	struct xfs_buf		*bp;	/* buffer pointer */
+	int			ptr;	/* record number */
+#ifdef DEBUG
+	int			error;	/* error return value */
+#endif
+
+	ptr = cur->bc_ptrs[0];
+	block = xfs_btree_get_block(cur, 0, &bp);
+
+#ifdef DEBUG
+	error = xfs_btree_check_block(cur, block, 0, bp);
+	if (error)
+		return error;
+#endif
+
+	/*
+	 * Off the right end or left end, return failure.
+	 */
+	if (ptr > xfs_btree_get_numrecs(block) || ptr <= 0) {
+		*stat = 0;
+		return 0;
+	}
+
+	/*
+	 * Point to the record and extract its data.
+	 */
+	*recp = xfs_btree_rec_addr(cur, ptr, block);
+	*stat = 1;
+	return 0;
+}
diff --git a/fs/xfs/xfs_btree.h b/fs/xfs/xfs_btree.h
index 1f528a2a375..789fffdf8b2 100644
--- a/fs/xfs/xfs_btree.h
+++ b/fs/xfs/xfs_btree.h
@@ -39,39 +39,19 @@ extern kmem_zone_t	*xfs_btree_cur_zone;
 #define	XFS_BTNUM_INO	((xfs_btnum_t)XFS_BTNUM_INOi)
 
 /*
- * Short form header: space allocation btrees.
- */
-typedef struct xfs_btree_sblock {
-	__be32		bb_magic;	/* magic number for block type */
-	__be16		bb_level;	/* 0 is a leaf */
-	__be16		bb_numrecs;	/* current # of data records */
-	__be32		bb_leftsib;	/* left sibling block or NULLAGBLOCK */
-	__be32		bb_rightsib;	/* right sibling block or NULLAGBLOCK */
-} xfs_btree_sblock_t;
-
-/*
- * Long form header: bmap btrees.
- */
-typedef struct xfs_btree_lblock {
-	__be32		bb_magic;	/* magic number for block type */
-	__be16		bb_level;	/* 0 is a leaf */
-	__be16		bb_numrecs;	/* current # of data records */
-	__be64		bb_leftsib;	/* left sibling block or NULLDFSBNO */
-	__be64		bb_rightsib;	/* right sibling block or NULLDFSBNO */
-} xfs_btree_lblock_t;
-
-/*
- * Combined header and structure, used by common code.
+ * Generic btree header.
+ *
+ * This is a comination of the actual format used on disk for short and long
+ * format btrees.  The first three fields are shared by both format, but
+ * the pointers are different and should be used with care.
+ *
+ * To get the size of the actual short or long form headers please use
+ * the size macros below.  Never use sizeof(xfs_btree_block).
  */
-typedef struct xfs_btree_hdr
-{
+struct xfs_btree_block {
 	__be32		bb_magic;	/* magic number for block type */
 	__be16		bb_level;	/* 0 is a leaf */
 	__be16		bb_numrecs;	/* current # of data records */
-} xfs_btree_hdr_t;
-
-typedef struct xfs_btree_block {
-	xfs_btree_hdr_t	bb_h;		/* header */
 	union {
 		struct {
 			__be32		bb_leftsib;
@@ -82,7 +62,36 @@ typedef struct xfs_btree_block {
 			__be64		bb_rightsib;
 		} l;			/* long form pointers */
 	} bb_u;				/* rest */
-} xfs_btree_block_t;
+};
+
+#define XFS_BTREE_SBLOCK_LEN	16	/* size of a short form block */
+#define XFS_BTREE_LBLOCK_LEN	24	/* size of a long form block */
+
+
+/*
+ * Generic key, ptr and record wrapper structures.
+ *
+ * These are disk format structures, and are converted where necessary
+ * by the btree specific code that needs to interpret them.
+ */
+union xfs_btree_ptr {
+	__be32			s;	/* short form ptr */
+	__be64			l;	/* long form ptr */
+};
+
+union xfs_btree_key {
+	xfs_bmbt_key_t		bmbt;
+	xfs_bmdr_key_t		bmbr;	/* bmbt root block */
+	xfs_alloc_key_t		alloc;
+	xfs_inobt_key_t		inobt;
+};
+
+union xfs_btree_rec {
+	xfs_bmbt_rec_t		bmbt;
+	xfs_bmdr_rec_t		bmbr;	/* bmbt root block */
+	xfs_alloc_rec_t		alloc;
+	xfs_inobt_rec_t		inobt;
+};
 
 /*
  * For logging record fields.
@@ -96,46 +105,131 @@ typedef struct xfs_btree_block {
 #define	XFS_BB_ALL_BITS		((1 << XFS_BB_NUM_BITS) - 1)
 
 /*
- * Boolean to select which form of xfs_btree_block_t.bb_u to use.
- */
-#define	XFS_BTREE_LONG_PTRS(btnum)	((btnum) == XFS_BTNUM_BMAP)
-
-/*
  * Magic numbers for btree blocks.
  */
 extern const __uint32_t	xfs_magics[];
 
 /*
- * Maximum and minimum records in a btree block.
- * Given block size, type prefix, and leaf flag (0 or 1).
- * The divisor below is equivalent to lf ? (e1) : (e2) but that produces
- * compiler warnings.
- */
-#define	XFS_BTREE_BLOCK_MAXRECS(bsz,t,lf)	\
-	((int)(((bsz) - (uint)sizeof(t ## _block_t)) / \
-	 (((lf) * (uint)sizeof(t ## _rec_t)) + \
-	  ((1 - (lf)) * \
-	   ((uint)sizeof(t ## _key_t) + (uint)sizeof(t ## _ptr_t))))))
-#define	XFS_BTREE_BLOCK_MINRECS(bsz,t,lf)	\
-	(XFS_BTREE_BLOCK_MAXRECS(bsz,t,lf) / 2)
-
-/*
- * Record, key, and pointer address calculation macros.
- * Given block size, type prefix, block pointer, and index of requested entry
- * (first entry numbered 1).
- */
-#define	XFS_BTREE_REC_ADDR(t,bb,i)	\
-	((t ## _rec_t *)((char *)(bb) + sizeof(t ## _block_t) + \
-	 ((i) - 1) * sizeof(t ## _rec_t)))
-#define	XFS_BTREE_KEY_ADDR(t,bb,i)	\
-	((t ## _key_t *)((char *)(bb) + sizeof(t ## _block_t) + \
-	 ((i) - 1) * sizeof(t ## _key_t)))
-#define	XFS_BTREE_PTR_ADDR(t,bb,i,mxr)	\
-	((t ## _ptr_t *)((char *)(bb) + sizeof(t ## _block_t) + \
-	 (mxr) * sizeof(t ## _key_t) + ((i) - 1) * sizeof(t ## _ptr_t)))
+ * Generic stats interface
+ */
+#define __XFS_BTREE_STATS_INC(type, stat) \
+	XFS_STATS_INC(xs_ ## type ## _2_ ## stat)
+#define XFS_BTREE_STATS_INC(cur, stat)  \
+do {    \
+	switch (cur->bc_btnum) {  \
+	case XFS_BTNUM_BNO: __XFS_BTREE_STATS_INC(abtb, stat); break;	\
+	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_INC(abtc, stat); break;	\
+	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(bmbt, stat); break;	\
+	case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(ibt, stat); break;	\
+	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
+	}       \
+} while (0)
+
+#define __XFS_BTREE_STATS_ADD(type, stat, val) \
+	XFS_STATS_ADD(xs_ ## type ## _2_ ## stat, val)
+#define XFS_BTREE_STATS_ADD(cur, stat, val)  \
+do {    \
+	switch (cur->bc_btnum) {  \
+	case XFS_BTNUM_BNO: __XFS_BTREE_STATS_ADD(abtb, stat, val); break; \
+	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_ADD(abtc, stat, val); break; \
+	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_ADD(bmbt, stat, val); break; \
+	case XFS_BTNUM_INO: __XFS_BTREE_STATS_ADD(ibt, stat, val); break; \
+	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
+	}       \
+} while (0)
 
 #define	XFS_BTREE_MAXLEVELS	8	/* max of all btrees */
 
+struct xfs_btree_ops {
+	/* size of the key and record structures */
+	size_t	key_len;
+	size_t	rec_len;
+
+	/* cursor operations */
+	struct xfs_btree_cur *(*dup_cursor)(struct xfs_btree_cur *);
+	void	(*update_cursor)(struct xfs_btree_cur *src,
+				 struct xfs_btree_cur *dst);
+
+	/* update btree root pointer */
+	void	(*set_root)(struct xfs_btree_cur *cur,
+				union xfs_btree_ptr *nptr, int level_change);
+	int	(*kill_root)(struct xfs_btree_cur *cur, struct xfs_buf *bp,
+				int level, union xfs_btree_ptr *newroot);
+
+	/* block allocation / freeing */
+	int	(*alloc_block)(struct xfs_btree_cur *cur,
+			       union xfs_btree_ptr *start_bno,
+			       union xfs_btree_ptr *new_bno,
+			       int length, int *stat);
+	int	(*free_block)(struct xfs_btree_cur *cur, struct xfs_buf *bp);
+
+	/* update last record information */
+	void	(*update_lastrec)(struct xfs_btree_cur *cur,
+				  struct xfs_btree_block *block,
+				  union xfs_btree_rec *rec,
+				  int ptr, int reason);
+
+	/* records in block/level */
+	int	(*get_minrecs)(struct xfs_btree_cur *cur, int level);
+	int	(*get_maxrecs)(struct xfs_btree_cur *cur, int level);
+
+	/* records on disk.  Matter for the root in inode case. */
+	int	(*get_dmaxrecs)(struct xfs_btree_cur *cur, int level);
+
+	/* init values of btree structures */
+	void	(*init_key_from_rec)(union xfs_btree_key *key,
+				     union xfs_btree_rec *rec);
+	void	(*init_rec_from_key)(union xfs_btree_key *key,
+				     union xfs_btree_rec *rec);
+	void	(*init_rec_from_cur)(struct xfs_btree_cur *cur,
+				     union xfs_btree_rec *rec);
+	void	(*init_ptr_from_cur)(struct xfs_btree_cur *cur,
+				     union xfs_btree_ptr *ptr);
+
+	/* difference between key value and cursor value */
+	__int64_t (*key_diff)(struct xfs_btree_cur *cur,
+			      union xfs_btree_key *key);
+
+#ifdef DEBUG
+	/* check that k1 is lower than k2 */
+	int	(*keys_inorder)(struct xfs_btree_cur *cur,
+				union xfs_btree_key *k1,
+				union xfs_btree_key *k2);
+
+	/* check that r1 is lower than r2 */
+	int	(*recs_inorder)(struct xfs_btree_cur *cur,
+				union xfs_btree_rec *r1,
+				union xfs_btree_rec *r2);
+#endif
+
+	/* btree tracing */
+#ifdef XFS_BTREE_TRACE
+	void		(*trace_enter)(struct xfs_btree_cur *, const char *,
+				       char *, int, int, __psunsigned_t,
+				       __psunsigned_t, __psunsigned_t,
+				       __psunsigned_t, __psunsigned_t,
+				       __psunsigned_t, __psunsigned_t,
+				       __psunsigned_t, __psunsigned_t,
+				       __psunsigned_t, __psunsigned_t);
+	void		(*trace_cursor)(struct xfs_btree_cur *, __uint32_t *,
+					__uint64_t *, __uint64_t *);
+	void		(*trace_key)(struct xfs_btree_cur *,
+				     union xfs_btree_key *, __uint64_t *,
+				     __uint64_t *);
+	void		(*trace_record)(struct xfs_btree_cur *,
+					union xfs_btree_rec *, __uint64_t *,
+					__uint64_t *, __uint64_t *);
+#endif
+};
+
+/*
+ * Reasons for the update_lastrec method to be called.
+ */
+#define LASTREC_UPDATE	0
+#define LASTREC_INSREC	1
+#define LASTREC_DELREC	2
+
+
 /*
  * Btree cursor structure.
  * This collects all information needed by the btree code in one place.
@@ -144,6 +238,8 @@ typedef struct xfs_btree_cur
 {
 	struct xfs_trans	*bc_tp;	/* transaction we're in, if any */
 	struct xfs_mount	*bc_mp;	/* file system mount struct */
+	const struct xfs_btree_ops *bc_ops;
+	uint			bc_flags; /* btree features - below */
 	union {
 		xfs_alloc_rec_incore_t	a;
 		xfs_bmbt_irec_t		b;
@@ -175,94 +271,40 @@ typedef struct xfs_btree_cur
 	}		bc_private;	/* per-btree type data */
 } xfs_btree_cur_t;
 
+/* cursor flags */
+#define XFS_BTREE_LONG_PTRS		(1<<0)	/* pointers are 64bits long */
+#define XFS_BTREE_ROOT_IN_INODE		(1<<1)	/* root may be variable size */
+#define XFS_BTREE_LASTREC_UPDATE	(1<<2)	/* track last rec externally */
+
+
 #define	XFS_BTREE_NOERROR	0
 #define	XFS_BTREE_ERROR		1
 
 /*
  * Convert from buffer to btree block header.
  */
-#define	XFS_BUF_TO_BLOCK(bp)	((xfs_btree_block_t *)XFS_BUF_PTR(bp))
-#define	XFS_BUF_TO_LBLOCK(bp)	((xfs_btree_lblock_t *)XFS_BUF_PTR(bp))
-#define	XFS_BUF_TO_SBLOCK(bp)	((xfs_btree_sblock_t *)XFS_BUF_PTR(bp))
+#define	XFS_BUF_TO_BLOCK(bp)	((struct xfs_btree_block *)XFS_BUF_PTR(bp))
 
 
-#ifdef __KERNEL__
-
-#ifdef DEBUG
 /*
- * Debug routine: check that block header is ok.
+ * Check that block header is ok.
  */
-void
+int
 xfs_btree_check_block(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_block_t	*block,	/* generic btree block pointer */
-	int			level,	/* level of the btree block */
-	struct xfs_buf		*bp);	/* buffer containing block, if any */
-
-/*
- * Debug routine: check that keys are in the right order.
- */
-void
-xfs_btree_check_key(
-	xfs_btnum_t		btnum,	/* btree identifier */
-	void			*ak1,	/* pointer to left (lower) key */
-	void			*ak2);	/* pointer to right (higher) key */
-
-/*
- * Debug routine: check that records are in the right order.
- */
-void
-xfs_btree_check_rec(
-	xfs_btnum_t		btnum,	/* btree identifier */
-	void			*ar1,	/* pointer to left (lower) record */
-	void			*ar2);	/* pointer to right (higher) record */
-#else
-#define	xfs_btree_check_block(a,b,c,d)
-#define	xfs_btree_check_key(a,b,c)
-#define	xfs_btree_check_rec(a,b,c)
-#endif	/* DEBUG */
-
-/*
- * Checking routine: check that long form block header is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_lblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_lblock_t	*block,	/* btree long form block pointer */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	struct xfs_btree_block	*block,	/* generic btree block pointer */
 	int			level,	/* level of the btree block */
 	struct xfs_buf		*bp);	/* buffer containing block, if any */
 
 /*
- * Checking routine: check that (long) pointer is ok.
+ * Check that (long) pointer is ok.
  */
 int					/* error (0 or EFSCORRUPTED) */
 xfs_btree_check_lptr(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
+	struct xfs_btree_cur	*cur,	/* btree cursor */
 	xfs_dfsbno_t		ptr,	/* btree block disk address */
 	int			level);	/* btree block level */
 
-#define xfs_btree_check_lptr_disk(cur, ptr, level) \
-	xfs_btree_check_lptr(cur, be64_to_cpu(ptr), level)
-
-/*
- * Checking routine: check that short form block header is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sblock(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_btree_sblock_t	*block,	/* btree short form block pointer */
-	int			level,	/* level of the btree block */
-	struct xfs_buf		*bp);	/* buffer containing block */
-
-/*
- * Checking routine: check that (short) pointer is ok.
- */
-int					/* error (0 or EFSCORRUPTED) */
-xfs_btree_check_sptr(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agblock_t		ptr,	/* btree block disk address */
-	int			level);	/* btree block level */
-
 /*
  * Delete the btree cursor.
  */
@@ -281,15 +323,6 @@ xfs_btree_dup_cursor(
 	xfs_btree_cur_t		**ncur);/* output cursor */
 
 /*
- * Change the cursor to point to the first record in the current block
- * at the given level.  Other levels are unaffected.
- */
-int					/* success=1, failure=0 */
-xfs_btree_firstrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level);	/* level to change */
-
-/*
  * Get a buffer for the block, return it with no data read.
  * Long-form addressing.
  */
@@ -313,20 +346,6 @@ xfs_btree_get_bufs(
 	uint			lock);	/* lock flags for get_buf */
 
 /*
- * Allocate a new btree cursor.
- * The cursor is either for allocation (A) or bmap (B).
- */
-xfs_btree_cur_t *			/* new btree cursor */
-xfs_btree_init_cursor(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	struct xfs_buf		*agbp,	/* (A only) buffer for agf structure */
-	xfs_agnumber_t		agno,	/* (A only) allocation group number */
-	xfs_btnum_t		btnum,	/* btree identifier */
-	struct xfs_inode	*ip,	/* (B only) inode owning the btree */
-	int			whichfork); /* (B only) data/attr fork */
-
-/*
  * Check for the cursor referring to the last block at the given level.
  */
 int					/* 1=is last block, 0=not last block */
@@ -335,15 +354,6 @@ xfs_btree_islastblock(
 	int			level);	/* level to check */
 
 /*
- * Change the cursor to point to the last record in the current block
- * at the given level.  Other levels are unaffected.
- */
-int					/* success=1, failure=0 */
-xfs_btree_lastrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level);	/* level to change */
-
-/*
  * Compute first and last byte offsets for the fields given.
  * Interprets the offsets table, which contains struct field offsets.
  */
@@ -404,39 +414,53 @@ xfs_btree_reada_bufs(
 	xfs_extlen_t		count);	/* count of filesystem blocks */
 
 /*
- * Read-ahead btree blocks, at the given level.
- * Bits in lr are set from XFS_BTCUR_{LEFT,RIGHT}RA.
+ * Set the buffer for level "lev" in the cursor to bp, releasing
+ * any previous buffer.
  */
-int					/* readahead block count */
-xfs_btree_readahead_core(
+void
+xfs_btree_setbuf(
 	xfs_btree_cur_t		*cur,	/* btree cursor */
 	int			lev,	/* level in btree */
-	int			lr);	/* left/right bits */
+	struct xfs_buf		*bp);	/* new buffer to set */
 
-static inline int			/* readahead block count */
-xfs_btree_readahead(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			lev,	/* level in btree */
-	int			lr)	/* left/right bits */
-{
-	if ((cur->bc_ra[lev] | lr) == cur->bc_ra[lev])
-		return 0;
 
-	return xfs_btree_readahead_core(cur, lev, lr);
-}
+/*
+ * Common btree core entry points.
+ */
+int xfs_btree_increment(struct xfs_btree_cur *, int, int *);
+int xfs_btree_decrement(struct xfs_btree_cur *, int, int *);
+int xfs_btree_lookup(struct xfs_btree_cur *, xfs_lookup_t, int *);
+int xfs_btree_update(struct xfs_btree_cur *, union xfs_btree_rec *);
+int xfs_btree_new_iroot(struct xfs_btree_cur *, int *, int *);
+int xfs_btree_kill_iroot(struct xfs_btree_cur *);
+int xfs_btree_insert(struct xfs_btree_cur *, int *);
+int xfs_btree_delete(struct xfs_btree_cur *, int *);
+int xfs_btree_get_rec(struct xfs_btree_cur *, union xfs_btree_rec **, int *);
 
+/*
+ * Internal btree helpers also used by xfs_bmap.c.
+ */
+void xfs_btree_log_block(struct xfs_btree_cur *, struct xfs_buf *, int);
+void xfs_btree_log_recs(struct xfs_btree_cur *, struct xfs_buf *, int, int);
 
 /*
- * Set the buffer for level "lev" in the cursor to bp, releasing
- * any previous buffer.
+ * Helpers.
  */
-void
-xfs_btree_setbuf(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			lev,	/* level in btree */
-	struct xfs_buf		*bp);	/* new buffer to set */
+static inline int xfs_btree_get_numrecs(struct xfs_btree_block *block)
+{
+	return be16_to_cpu(block->bb_numrecs);
+}
+
+static inline void xfs_btree_set_numrecs(struct xfs_btree_block *block,
+		__uint16_t numrecs)
+{
+	block->bb_numrecs = cpu_to_be16(numrecs);
+}
 
-#endif	/* __KERNEL__ */
+static inline int xfs_btree_get_level(struct xfs_btree_block *block)
+{
+	return be16_to_cpu(block->bb_level);
+}
 
 
 /*
diff --git a/fs/xfs/xfs_btree_trace.c b/fs/xfs/xfs_btree_trace.c
new file mode 100644
index 00000000000..44ff942a0fd
--- /dev/null
+++ b/fs/xfs/xfs_btree_trace.c
@@ -0,0 +1,249 @@
+/*
+ * Copyright (c) 2008 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#include "xfs.h"
+#include "xfs_types.h"
+#include "xfs_inum.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_btree_trace.h"
+
+STATIC void
+xfs_btree_trace_ptr(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	ptr,
+	__psunsigned_t		*high,
+	__psunsigned_t		*low)
+{
+	if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+		__u64 val = be64_to_cpu(ptr.l);
+		*high = val >> 32;
+		*low = (int)val;
+	} else {
+		*high = 0;
+		*low = be32_to_cpu(ptr.s);
+	}
+}
+
+/*
+ * Add a trace buffer entry for arguments, for a buffer & 1 integer arg.
+ */
+void
+xfs_btree_trace_argbi(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*b,
+	int			i,
+	int			line)
+{
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGBI,
+				 line, (__psunsigned_t)b, i, 0, 0, 0, 0, 0,
+				 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for a buffer & 2 integer args.
+ */
+void
+xfs_btree_trace_argbii(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*b,
+	int			i0,
+	int			i1,
+	int			line)
+{
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGBII,
+				 line, (__psunsigned_t)b, i0, i1, 0, 0, 0, 0,
+				 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for 3 block-length args
+ * and an integer arg.
+ */
+void
+xfs_btree_trace_argfffi(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	xfs_dfiloff_t		o,
+	xfs_dfsbno_t		b,
+	xfs_dfilblks_t		i,
+	int			j,
+	int			line)
+{
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGFFFI,
+				 line,
+				 o >> 32, (int)o,
+				 b >> 32, (int)b,
+				 i >> 32, (int)i,
+				 (int)j, 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for one integer arg.
+ */
+void
+xfs_btree_trace_argi(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	int			i,
+	int			line)
+{
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGI,
+				 line, i, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for int, fsblock, key.
+ */
+void
+xfs_btree_trace_argipk(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	int			i,
+	union xfs_btree_ptr	ptr,
+	union xfs_btree_key	*key,
+	int			line)
+{
+	__psunsigned_t		high, low;
+	__uint64_t		l0, l1;
+
+	xfs_btree_trace_ptr(cur, ptr, &high, &low);
+	cur->bc_ops->trace_key(cur, key, &l0, &l1);
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGIPK,
+				 line, i, high, low,
+				 l0 >> 32, (int)l0,
+				 l1 >> 32, (int)l1,
+				 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for int, fsblock, rec.
+ */
+void
+xfs_btree_trace_argipr(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	int			i,
+	union xfs_btree_ptr	ptr,
+	union xfs_btree_rec	*rec,
+	int			line)
+{
+	__psunsigned_t		high, low;
+	__uint64_t		l0, l1, l2;
+
+	xfs_btree_trace_ptr(cur, ptr, &high, &low);
+	cur->bc_ops->trace_record(cur, rec, &l0, &l1, &l2);
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGIPR,
+			      line, i,
+			      high, low,
+			      l0 >> 32, (int)l0,
+			      l1 >> 32, (int)l1,
+			      l2 >> 32, (int)l2,
+			      0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for int, key.
+ */
+void
+xfs_btree_trace_argik(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	int			i,
+	union xfs_btree_key	*key,
+	int			line)
+{
+	__uint64_t		l0, l1;
+
+	cur->bc_ops->trace_key(cur, key, &l0, &l1);
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGIK,
+				 line, i,
+				 l0 >> 32, (int)l0,
+				 l1 >> 32, (int)l1,
+				 0, 0, 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for arguments, for record.
+ */
+void
+xfs_btree_trace_argr(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec,
+	int			line)
+{
+	__uint64_t		l0, l1, l2;
+
+	cur->bc_ops->trace_record(cur, rec, &l0, &l1, &l2);
+	cur->bc_ops->trace_enter(cur, func, XBT_ARGS, XFS_BTREE_KTRACE_ARGR,
+			      line,
+			      l0 >> 32, (int)l0,
+			      l1 >> 32, (int)l1,
+			      l2 >> 32, (int)l2,
+			      0, 0, 0, 0, 0);
+}
+
+/*
+ * Add a trace buffer entry for the cursor/operation.
+ */
+void
+xfs_btree_trace_cursor(
+	const char		*func,
+	struct xfs_btree_cur	*cur,
+	int			type,
+	int			line)
+{
+	__uint32_t		s0;
+	__uint64_t		l0, l1;
+	char			*s;
+
+	switch (type) {
+	case XBT_ARGS:
+		s = "args";
+		break;
+	case XBT_ENTRY:
+		s = "entry";
+		break;
+	case XBT_ERROR:
+		s = "error";
+		break;
+	case XBT_EXIT:
+		s = "exit";
+		break;
+	default:
+		s = "unknown";
+		break;
+	}
+
+	cur->bc_ops->trace_cursor(cur, &s0, &l0, &l1);
+	cur->bc_ops->trace_enter(cur, func, s, XFS_BTREE_KTRACE_CUR, line,
+				 s0,
+				 l0 >> 32, (int)l0,
+				 l1 >> 32, (int)l1,
+				 (__psunsigned_t)cur->bc_bufs[0],
+				 (__psunsigned_t)cur->bc_bufs[1],
+				 (__psunsigned_t)cur->bc_bufs[2],
+				 (__psunsigned_t)cur->bc_bufs[3],
+				 (cur->bc_ptrs[0] << 16) | cur->bc_ptrs[1],
+				 (cur->bc_ptrs[2] << 16) | cur->bc_ptrs[3]);
+}
diff --git a/fs/xfs/xfs_btree_trace.h b/fs/xfs/xfs_btree_trace.h
new file mode 100644
index 00000000000..b3f5eb3c3c6
--- /dev/null
+++ b/fs/xfs/xfs_btree_trace.h
@@ -0,0 +1,116 @@
+/*
+ * Copyright (c) 2008 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+#ifndef __XFS_BTREE_TRACE_H__
+#define	__XFS_BTREE_TRACE_H__
+
+struct xfs_btree_cur;
+struct xfs_buf;
+
+
+/*
+ * Trace hooks.
+ * i,j = integer (32 bit)
+ * b = btree block buffer (xfs_buf_t)
+ * p = btree ptr
+ * r = btree record
+ * k = btree key
+ */
+
+#ifdef XFS_BTREE_TRACE
+
+/*
+ * Trace buffer entry types.
+ */
+#define XFS_BTREE_KTRACE_ARGBI   1
+#define XFS_BTREE_KTRACE_ARGBII  2
+#define XFS_BTREE_KTRACE_ARGFFFI 3
+#define XFS_BTREE_KTRACE_ARGI    4
+#define XFS_BTREE_KTRACE_ARGIPK  5
+#define XFS_BTREE_KTRACE_ARGIPR  6
+#define XFS_BTREE_KTRACE_ARGIK   7
+#define XFS_BTREE_KTRACE_ARGR	 8
+#define XFS_BTREE_KTRACE_CUR     9
+
+/*
+ * Sub-types for cursor traces.
+ */
+#define XBT_ARGS	0
+#define XBT_ENTRY	1
+#define XBT_ERROR	2
+#define XBT_EXIT	3
+
+void xfs_btree_trace_argbi(const char *, struct xfs_btree_cur *,
+		struct xfs_buf *, int, int);
+void xfs_btree_trace_argbii(const char *, struct xfs_btree_cur *,
+		struct xfs_buf *, int, int, int);
+void xfs_btree_trace_argfffi(const char *, struct xfs_btree_cur *,
+		xfs_dfiloff_t, xfs_dfsbno_t, xfs_dfilblks_t, int, int);
+void xfs_btree_trace_argi(const char *, struct xfs_btree_cur *, int, int);
+void xfs_btree_trace_argipk(const char *, struct xfs_btree_cur *, int,
+		union xfs_btree_ptr, union xfs_btree_key *, int);
+void xfs_btree_trace_argipr(const char *, struct xfs_btree_cur *, int,
+		union xfs_btree_ptr, union xfs_btree_rec *, int);
+void xfs_btree_trace_argik(const char *, struct xfs_btree_cur *, int,
+		union xfs_btree_key *, int);
+void xfs_btree_trace_argr(const char *, struct xfs_btree_cur *,
+		union xfs_btree_rec *, int);
+void xfs_btree_trace_cursor(const char *, struct xfs_btree_cur *, int, int);
+
+
+#define XFS_ALLOCBT_TRACE_SIZE	4096	/* size of global trace buffer */
+extern ktrace_t	*xfs_allocbt_trace_buf;
+
+#define XFS_INOBT_TRACE_SIZE	4096	/* size of global trace buffer */
+extern ktrace_t	*xfs_inobt_trace_buf;
+
+#define XFS_BMBT_TRACE_SIZE	4096	/* size of global trace buffer */
+#define XFS_BMBT_KTRACE_SIZE	32	/* size of per-inode trace buffer */
+extern ktrace_t	*xfs_bmbt_trace_buf;
+
+
+#define	XFS_BTREE_TRACE_ARGBI(c, b, i)	\
+	xfs_btree_trace_argbi(__func__, c, b, i, __LINE__)
+#define	XFS_BTREE_TRACE_ARGBII(c, b, i, j)	\
+	xfs_btree_trace_argbii(__func__, c, b, i, j, __LINE__)
+#define	XFS_BTREE_TRACE_ARGFFFI(c, o, b, i, j)	\
+	xfs_btree_trace_argfffi(__func__, c, o, b, i, j, __LINE__)
+#define	XFS_BTREE_TRACE_ARGI(c, i)	\
+	xfs_btree_trace_argi(__func__, c, i, __LINE__)
+#define	XFS_BTREE_TRACE_ARGIPK(c, i, p, k)	\
+	xfs_btree_trace_argipk(__func__, c, i, p, k, __LINE__)
+#define	XFS_BTREE_TRACE_ARGIPR(c, i, p, r)	\
+	xfs_btree_trace_argipr(__func__, c, i, p, r, __LINE__)
+#define	XFS_BTREE_TRACE_ARGIK(c, i, k)	\
+	xfs_btree_trace_argik(__func__, c, i, k, __LINE__)
+#define XFS_BTREE_TRACE_ARGR(c, r)	\
+	xfs_btree_trace_argr(__func__, c, r, __LINE__)
+#define	XFS_BTREE_TRACE_CURSOR(c, t)	\
+	xfs_btree_trace_cursor(__func__, c, t, __LINE__)
+#else
+#define	XFS_BTREE_TRACE_ARGBI(c, b, i)
+#define	XFS_BTREE_TRACE_ARGBII(c, b, i, j)
+#define	XFS_BTREE_TRACE_ARGFFFI(c, o, b, i, j)
+#define	XFS_BTREE_TRACE_ARGI(c, i)
+#define	XFS_BTREE_TRACE_ARGIPK(c, i, p, s)
+#define	XFS_BTREE_TRACE_ARGIPR(c, i, p, r)
+#define	XFS_BTREE_TRACE_ARGIK(c, i, k)
+#define XFS_BTREE_TRACE_ARGR(c, r)
+#define	XFS_BTREE_TRACE_CURSOR(c, t)
+#endif	/* XFS_BTREE_TRACE */
+
+#endif /* __XFS_BTREE_TRACE_H__ */
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 002fc2617c8..92af4098c7e 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -375,7 +375,7 @@ xfs_buf_item_unpin(
 	xfs_buf_log_item_t	*bip,
 	int			stale)
 {
-	xfs_mount_t	*mp;
+	struct xfs_ail	*ailp;
 	xfs_buf_t	*bp;
 	int		freed;
 
@@ -387,7 +387,7 @@ xfs_buf_item_unpin(
 	xfs_buftrace("XFS_UNPIN", bp);
 
 	freed = atomic_dec_and_test(&bip->bli_refcount);
-	mp = bip->bli_item.li_mountp;
+	ailp = bip->bli_item.li_ailp;
 	xfs_bunpin(bp);
 	if (freed && stale) {
 		ASSERT(bip->bli_flags & XFS_BLI_STALE);
@@ -399,17 +399,17 @@ xfs_buf_item_unpin(
 		xfs_buftrace("XFS_UNPIN STALE", bp);
 		/*
 		 * If we get called here because of an IO error, we may
-		 * or may not have the item on the AIL. xfs_trans_delete_ail()
+		 * or may not have the item on the AIL. xfs_trans_ail_delete()
 		 * will take care of that situation.
-		 * xfs_trans_delete_ail() drops the AIL lock.
+		 * xfs_trans_ail_delete() drops the AIL lock.
 		 */
 		if (bip->bli_flags & XFS_BLI_STALE_INODE) {
 			xfs_buf_do_callbacks(bp, (xfs_log_item_t *)bip);
 			XFS_BUF_SET_FSPRIVATE(bp, NULL);
 			XFS_BUF_CLR_IODONE_FUNC(bp);
 		} else {
-			spin_lock(&mp->m_ail_lock);
-			xfs_trans_delete_ail(mp, (xfs_log_item_t *)bip);
+			spin_lock(&ailp->xa_lock);
+			xfs_trans_ail_delete(ailp, (xfs_log_item_t *)bip);
 			xfs_buf_item_relse(bp);
 			ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL);
 		}
@@ -707,8 +707,8 @@ xfs_buf_item_init(
 	 * the first.  If we do already have one, there is
 	 * nothing to do here so return.
 	 */
-	if (XFS_BUF_FSPRIVATE3(bp, xfs_mount_t *) != mp)
-		XFS_BUF_SET_FSPRIVATE3(bp, mp);
+	if (bp->b_mount != mp)
+		bp->b_mount = mp;
 	XFS_BUF_SET_BDSTRAT_FUNC(bp, xfs_bdstrat_cb);
 	if (XFS_BUF_FSPRIVATE(bp, void *) != NULL) {
 		lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
@@ -731,6 +731,7 @@ xfs_buf_item_init(
 	bip->bli_item.li_type = XFS_LI_BUF;
 	bip->bli_item.li_ops = &xfs_buf_item_ops;
 	bip->bli_item.li_mountp = mp;
+	bip->bli_item.li_ailp = mp->m_ail;
 	bip->bli_buf = bp;
 	xfs_buf_hold(bp);
 	bip->bli_format.blf_type = XFS_LI_BUF;
@@ -997,21 +998,7 @@ xfs_buf_iodone_callbacks(
 			xfs_buf_do_callbacks(bp, lip);
 			XFS_BUF_SET_FSPRIVATE(bp, NULL);
 			XFS_BUF_CLR_IODONE_FUNC(bp);
-
-			/*
-			 * XFS_SHUT flag gets set when we go thru the
-			 * entire buffer cache and deliberately start
-			 * throwing away delayed write buffers.
-			 * Since there's no biowait done on those,
-			 * we should just brelse them.
-			 */
-			if (XFS_BUF_ISSHUT(bp)) {
-			    XFS_BUF_UNSHUT(bp);
-				xfs_buf_relse(bp);
-			} else {
-				xfs_biodone(bp);
-			}
-
+			xfs_biodone(bp);
 			return;
 		}
 
@@ -1122,27 +1109,23 @@ xfs_buf_iodone(
 	xfs_buf_t		*bp,
 	xfs_buf_log_item_t	*bip)
 {
-	struct xfs_mount	*mp;
+	struct xfs_ail		*ailp = bip->bli_item.li_ailp;
 
 	ASSERT(bip->bli_buf == bp);
 
 	xfs_buf_rele(bp);
-	mp = bip->bli_item.li_mountp;
 
 	/*
 	 * If we are forcibly shutting down, this may well be
 	 * off the AIL already. That's because we simulate the
 	 * log-committed callbacks to unpin these buffers. Or we may never
 	 * have put this item on AIL because of the transaction was
-	 * aborted forcibly. xfs_trans_delete_ail() takes care of these.
+	 * aborted forcibly. xfs_trans_ail_delete() takes care of these.
 	 *
 	 * Either way, AIL is useless if we're forcing a shutdown.
 	 */
-	spin_lock(&mp->m_ail_lock);
-	/*
-	 * xfs_trans_delete_ail() drops the AIL lock.
-	 */
-	xfs_trans_delete_ail(mp, (xfs_log_item_t *)bip);
+	spin_lock(&ailp->xa_lock);
+	xfs_trans_ail_delete(ailp, (xfs_log_item_t *)bip);
 	xfs_buf_item_free(bip);
 }
 
diff --git a/fs/xfs/xfs_clnt.h b/fs/xfs/xfs_clnt.h
deleted file mode 100644
index d2ce5dd70d8..00000000000
--- a/fs/xfs/xfs_clnt.h
+++ /dev/null
@@ -1,105 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_CLNT_H__
-#define __XFS_CLNT_H__
-
-/*
- * XFS arguments structure, constructed from the arguments we
- * are passed via the mount system call.
- *
- * NOTE: The mount system call is handled differently between
- * Linux and IRIX.  In IRIX we worked work with a binary data
- * structure coming in across the syscall interface from user
- * space (the mount userspace knows about each filesystem type
- * and the set of valid options for it, and converts the users
- * argument string into a binary structure _before_ making the
- * system call), and the ABI issues that this implies.
- *
- * In Linux, we are passed a comma separated set of options;
- * ie. a NULL terminated string of characters.  Userspace mount
- * code does not have any knowledge of mount options expected by
- * each filesystem type and so each filesystem parses its mount
- * options in kernel space.
- *
- * For the Linux port, we kept this structure pretty much intact
- * and use it internally (because the existing code groks it).
- */
-struct xfs_mount_args {
-	int	flags;		/* flags -> see XFSMNT_... macros below */
-	int	flags2;		/* flags -> see XFSMNT2_... macros below */
-	int	logbufs;	/* Number of log buffers, -1 to default */
-	int	logbufsize;	/* Size of log buffers, -1 to default */
-	char	fsname[MAXNAMELEN+1];	/* data device name */
-	char	rtname[MAXNAMELEN+1];	/* realtime device filename */
-	char	logname[MAXNAMELEN+1];	/* journal device filename */
-	char	mtpt[MAXNAMELEN+1];	/* filesystem mount point */
-	int	sunit;		/* stripe unit (BBs) */
-	int	swidth;		/* stripe width (BBs), multiple of sunit */
-	uchar_t iosizelog;	/* log2 of the preferred I/O size */
-	int	ihashsize;	/* inode hash table size (buckets) */
-};
-
-/*
- * XFS mount option flags -- args->flags1
- */
-#define	XFSMNT_ATTR2		0x00000001	/* allow ATTR2 EA format */
-#define	XFSMNT_WSYNC		0x00000002	/* safe mode nfs mount
-						 * compatible */
-#define	XFSMNT_INO64		0x00000004	/* move inode numbers up
-						 * past 2^32 */
-#define XFSMNT_UQUOTA		0x00000008	/* user quota accounting */
-#define XFSMNT_PQUOTA		0x00000010	/* IRIX prj quota accounting */
-#define XFSMNT_UQUOTAENF	0x00000020	/* user quota limit
-						 * enforcement */
-#define XFSMNT_PQUOTAENF	0x00000040	/* IRIX project quota limit
-						 * enforcement */
-#define XFSMNT_QUIET		0x00000080	/* don't report mount errors */
-#define XFSMNT_NOALIGN		0x00000200	/* don't allocate at
-						 * stripe boundaries*/
-#define XFSMNT_RETERR		0x00000400	/* return error to user */
-#define XFSMNT_NORECOVERY	0x00000800	/* no recovery, implies
-						 * read-only mount */
-#define XFSMNT_SHARED		0x00001000	/* shared XFS mount */
-#define XFSMNT_IOSIZE		0x00002000	/* optimize for I/O size */
-#define XFSMNT_OSYNCISOSYNC	0x00004000	/* o_sync is REALLY o_sync */
-						/* (osyncisdsync is default) */
-#define XFSMNT_NOATTR2		0x00008000	/* turn off ATTR2 EA format */
-#define XFSMNT_32BITINODES	0x00200000	/* restrict inodes to 32
-						 * bits of address space */
-#define XFSMNT_GQUOTA		0x00400000	/* group quota accounting */
-#define XFSMNT_GQUOTAENF	0x00800000	/* group quota limit
-						 * enforcement */
-#define XFSMNT_NOUUID		0x01000000	/* Ignore fs uuid */
-#define XFSMNT_DMAPI		0x02000000	/* enable dmapi/xdsm */
-#define XFSMNT_BARRIER		0x04000000	/* use write barriers */
-#define XFSMNT_IKEEP		0x08000000	/* inode cluster delete */
-#define XFSMNT_SWALLOC		0x10000000	/* turn on stripe width
-						 * allocation */
-#define XFSMNT_DIRSYNC		0x40000000	/* sync creat,link,unlink,rename
-						 * symlink,mkdir,rmdir,mknod */
-#define XFSMNT_FLAGS2		0x80000000	/* more flags set in flags2 */
-
-/*
- * XFS mount option flags -- args->flags2
- */
-#define XFSMNT2_COMPAT_IOSIZE	0x00000001	/* don't report large preferred
-						 * I/O size in stat(2) */
-#define XFSMNT2_FILESTREAMS	0x00000002	/* enable the filestreams
-						 * allocator */
-
-#endif	/* __XFS_CLNT_H__ */
diff --git a/fs/xfs/xfs_da_btree.h b/fs/xfs/xfs_da_btree.h
index 8be0b00ede9..70b710c1792 100644
--- a/fs/xfs/xfs_da_btree.h
+++ b/fs/xfs/xfs_da_btree.h
@@ -72,27 +72,7 @@ typedef struct xfs_da_intnode {
 typedef struct xfs_da_node_hdr xfs_da_node_hdr_t;
 typedef struct xfs_da_node_entry xfs_da_node_entry_t;
 
-#define XFS_DA_MAXHASH	((xfs_dahash_t)-1) /* largest valid hash value */
-
 #define	XFS_LBSIZE(mp)	(mp)->m_sb.sb_blocksize
-#define	XFS_LBLOG(mp)	(mp)->m_sb.sb_blocklog
-
-#define	XFS_DA_MAKE_BNOENTRY(mp,bno,entry)	\
-	(((bno) << (mp)->m_dircook_elog) | (entry))
-#define	XFS_DA_MAKE_COOKIE(mp,bno,entry,hash)	\
-	(((xfs_off_t)XFS_DA_MAKE_BNOENTRY(mp, bno, entry) << 32) | (hash))
-#define	XFS_DA_COOKIE_HASH(mp,cookie)		((xfs_dahash_t)cookie)
-#define	XFS_DA_COOKIE_BNO(mp,cookie)		\
-	((((xfs_off_t)(cookie) >> 31) == -1LL ? \
-		(xfs_dablk_t)0 : \
-		(xfs_dablk_t)((xfs_off_t)(cookie) >> \
-				((mp)->m_dircook_elog + 32))))
-#define	XFS_DA_COOKIE_ENTRY(mp,cookie)		\
-	((((xfs_off_t)(cookie) >> 31) == -1LL ?	\
-		(xfs_dablk_t)0 : \
-		(xfs_dablk_t)(((xfs_off_t)(cookie) >> 32) & \
-				((1 << (mp)->m_dircook_elog) - 1))))
-
 
 /*========================================================================
  * Btree searching and modification structure definitions.
@@ -226,9 +206,8 @@ struct xfs_nameops {
 };
 
 
-#ifdef __KERNEL__
 /*========================================================================
- * Function prototypes for the kernel.
+ * Function prototypes.
  *========================================================================*/
 
 /*
@@ -289,6 +268,5 @@ xfs_daddr_t xfs_da_blkno(xfs_dabuf_t *dabuf);
 
 extern struct kmem_zone *xfs_da_state_zone;
 extern struct kmem_zone *xfs_dabuf_zone;
-#endif	/* __KERNEL__ */
 
 #endif	/* __XFS_DA_BTREE_H__ */
diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c
index 75b0cd4da0e..b4c1ee71349 100644
--- a/fs/xfs/xfs_dfrag.c
+++ b/fs/xfs/xfs_dfrag.c
@@ -49,9 +49,8 @@
  */
 int
 xfs_swapext(
-	xfs_swapext_t	__user *sxu)
+	xfs_swapext_t	*sxp)
 {
-	xfs_swapext_t	*sxp;
 	xfs_inode_t     *ip, *tip;
 	struct file	*file, *target_file;
 	int		error = 0;
@@ -62,11 +61,6 @@ xfs_swapext(
 		goto out;
 	}
 
-	if (copy_from_user(sxp, sxu, sizeof(xfs_swapext_t))) {
-		error = XFS_ERROR(EFAULT);
-		goto out_free_sxp;
-	}
-
 	/* Pull information for the target fd */
 	file = fget((int)sxp->sx_fdtarget);
 	if (!file) {
diff --git a/fs/xfs/xfs_dfrag.h b/fs/xfs/xfs_dfrag.h
index da178205be6..4f55a630655 100644
--- a/fs/xfs/xfs_dfrag.h
+++ b/fs/xfs/xfs_dfrag.h
@@ -46,7 +46,7 @@ typedef struct xfs_swapext
 /*
  * Syscall interface for xfs_swapext
  */
-int	xfs_swapext(struct xfs_swapext __user *sx);
+int	xfs_swapext(struct xfs_swapext *sx);
 
 int	xfs_swap_extents(struct xfs_inode *ip, struct xfs_inode *tip,
 		struct xfs_swapext *sxp);
diff --git a/fs/xfs/xfs_dinode.h b/fs/xfs/xfs_dinode.h
index c9065eaf2a4..162e8726df5 100644
--- a/fs/xfs/xfs_dinode.h
+++ b/fs/xfs/xfs_dinode.h
@@ -18,32 +18,29 @@
 #ifndef __XFS_DINODE_H__
 #define	__XFS_DINODE_H__
 
-struct xfs_buf;
-struct xfs_mount;
+#define	XFS_DINODE_MAGIC		0x494e	/* 'IN' */
+#define XFS_DINODE_GOOD_VERSION(v)	(((v) == 1 || (v) == 2))
 
-#define	XFS_DINODE_VERSION_1	1
-#define	XFS_DINODE_VERSION_2	2
-#define XFS_DINODE_GOOD_VERSION(v)	\
-	(((v) == XFS_DINODE_VERSION_1 || (v) == XFS_DINODE_VERSION_2))
-#define	XFS_DINODE_MAGIC	0x494e	/* 'IN' */
-
-/*
- * Disk inode structure.
- * This is just the header; the inode is expanded to fill a variable size
- * with the last field expanding.  It is split into the core and "other"
- * because we only need the core part in the in-core inode.
- */
 typedef struct xfs_timestamp {
 	__be32		t_sec;		/* timestamp seconds */
 	__be32		t_nsec;		/* timestamp nanoseconds */
 } xfs_timestamp_t;
 
 /*
- * Note: Coordinate changes to this structure with the XFS_DI_* #defines
- * below, the offsets table in xfs_ialloc_log_di() and struct xfs_icdinode
- * in xfs_inode.h.
+ * On-disk inode structure.
+ *
+ * This is just the header or "dinode core", the inode is expanded to fill a
+ * variable size the leftover area split into a data and an attribute fork.
+ * The format of the data and attribute fork depends on the format of the
+ * inode as indicated by di_format and di_aformat.  To access the data and
+ * attribute use the XFS_DFORK_PTR, XFS_DFORK_DPTR, and XFS_DFORK_PTR macros
+ * below.
+ *
+ * There is a very similar struct icdinode in xfs_inode which matches the
+ * layout of the first 96 bytes of this structure, but is kept in native
+ * format instead of big endian.
  */
-typedef struct xfs_dinode_core {
+typedef struct xfs_dinode {
 	__be16		di_magic;	/* inode magic # = XFS_DINODE_MAGIC */
 	__be16		di_mode;	/* mode and type of file */
 	__u8		di_version;	/* inode version */
@@ -69,34 +66,12 @@ typedef struct xfs_dinode_core {
 	__be16		di_dmstate;	/* DMIG state info */
 	__be16		di_flags;	/* random flags, XFS_DIFLAG_... */
 	__be32		di_gen;		/* generation number */
-} xfs_dinode_core_t;
 
-#define DI_MAX_FLUSH 0xffff
+	/* di_next_unlinked is the only non-core field in the old dinode */
+	__be32		di_next_unlinked;/* agi unlinked list ptr */
+} __attribute__((packed)) xfs_dinode_t;
 
-typedef struct xfs_dinode
-{
-	xfs_dinode_core_t	di_core;
-	/*
-	 * In adding anything between the core and the union, be
-	 * sure to update the macros like XFS_LITINO below and
-	 * XFS_BMAP_RBLOCK_DSIZE in xfs_bmap_btree.h.
-	 */
-	__be32			di_next_unlinked;/* agi unlinked list ptr */
-	union {
-		xfs_bmdr_block_t di_bmbt;	/* btree root block */
-		xfs_bmbt_rec_32_t di_bmx[1];	/* extent list */
-		xfs_dir2_sf_t	di_dir2sf;	/* shortform directory v2 */
-		char		di_c[1];	/* local contents */
-		__be32		di_dev;		/* device for S_IFCHR/S_IFBLK */
-		uuid_t		di_muuid;	/* mount point value */
-		char		di_symlink[1];	/* local symbolic link */
-	}		di_u;
-	union {
-		xfs_bmdr_block_t di_abmbt;	/* btree root block */
-		xfs_bmbt_rec_32_t di_abmx[1];	/* extent list */
-		xfs_attr_shortform_t di_attrsf;	/* shortform attribute list */
-	}		di_a;
-} xfs_dinode_t;
+#define DI_MAX_FLUSH 0xffff
 
 /*
  * The 32 bit link count in the inode theoretically maxes out at UINT_MAX.
@@ -107,50 +82,14 @@ typedef struct xfs_dinode
 #define	XFS_MAXLINK_1		65535U
 
 /*
- * Bit names for logging disk inodes only
- */
-#define	XFS_DI_MAGIC		0x0000001
-#define	XFS_DI_MODE		0x0000002
-#define	XFS_DI_VERSION		0x0000004
-#define	XFS_DI_FORMAT		0x0000008
-#define	XFS_DI_ONLINK		0x0000010
-#define	XFS_DI_UID		0x0000020
-#define	XFS_DI_GID		0x0000040
-#define	XFS_DI_NLINK		0x0000080
-#define	XFS_DI_PROJID		0x0000100
-#define	XFS_DI_PAD		0x0000200
-#define	XFS_DI_ATIME		0x0000400
-#define	XFS_DI_MTIME		0x0000800
-#define	XFS_DI_CTIME		0x0001000
-#define	XFS_DI_SIZE		0x0002000
-#define	XFS_DI_NBLOCKS		0x0004000
-#define	XFS_DI_EXTSIZE		0x0008000
-#define	XFS_DI_NEXTENTS		0x0010000
-#define	XFS_DI_NAEXTENTS	0x0020000
-#define	XFS_DI_FORKOFF		0x0040000
-#define	XFS_DI_AFORMAT		0x0080000
-#define	XFS_DI_DMEVMASK		0x0100000
-#define	XFS_DI_DMSTATE		0x0200000
-#define	XFS_DI_FLAGS		0x0400000
-#define	XFS_DI_GEN		0x0800000
-#define	XFS_DI_NEXT_UNLINKED	0x1000000
-#define	XFS_DI_U		0x2000000
-#define	XFS_DI_A		0x4000000
-#define	XFS_DI_NUM_BITS		27
-#define	XFS_DI_ALL_BITS		((1 << XFS_DI_NUM_BITS) - 1)
-#define	XFS_DI_CORE_BITS	(XFS_DI_ALL_BITS & ~(XFS_DI_U|XFS_DI_A))
-
-/*
  * Values for di_format
  */
-typedef enum xfs_dinode_fmt
-{
-	XFS_DINODE_FMT_DEV,		/* CHR, BLK: di_dev */
-	XFS_DINODE_FMT_LOCAL,		/* DIR, REG: di_c */
-					/* LNK: di_symlink */
-	XFS_DINODE_FMT_EXTENTS,		/* DIR, REG, LNK: di_bmx */
-	XFS_DINODE_FMT_BTREE,		/* DIR, REG, LNK: di_bmbt */
-	XFS_DINODE_FMT_UUID		/* MNT: di_uuid */
+typedef enum xfs_dinode_fmt {
+	XFS_DINODE_FMT_DEV,		/* xfs_dev_t */
+	XFS_DINODE_FMT_LOCAL,		/* bulk data */
+	XFS_DINODE_FMT_EXTENTS,		/* struct xfs_bmbt_rec */
+	XFS_DINODE_FMT_BTREE,		/* struct xfs_bmdr_block */
+	XFS_DINODE_FMT_UUID		/* uuid_t */
 } xfs_dinode_fmt_t;
 
 /*
@@ -166,13 +105,13 @@ typedef enum xfs_dinode_fmt
  */
 #define	XFS_LITINO(mp)	((mp)->m_litino)
 #define	XFS_BROOT_SIZE_ADJ	\
-	(sizeof(xfs_bmbt_block_t) - sizeof(xfs_bmdr_block_t))
+	(XFS_BTREE_LBLOCK_LEN - sizeof(xfs_bmdr_block_t))
 
 /*
  * Inode data & attribute fork sizes, per inode.
  */
-#define XFS_DFORK_Q(dip)		((dip)->di_core.di_forkoff != 0)
-#define XFS_DFORK_BOFF(dip)		((int)((dip)->di_core.di_forkoff << 3))
+#define XFS_DFORK_Q(dip)		((dip)->di_forkoff != 0)
+#define XFS_DFORK_BOFF(dip)		((int)((dip)->di_forkoff << 3))
 
 #define XFS_DFORK_DSIZE(dip,mp) \
 	(XFS_DFORK_Q(dip) ? \
@@ -187,23 +126,42 @@ typedef enum xfs_dinode_fmt
 		XFS_DFORK_DSIZE(dip, mp) : \
 		XFS_DFORK_ASIZE(dip, mp))
 
-#define XFS_DFORK_DPTR(dip)		    ((dip)->di_u.di_c)
+/*
+ * Return pointers to the data or attribute forks.
+ */
+#define XFS_DFORK_DPTR(dip) \
+	((char *)(dip) + sizeof(struct xfs_dinode))
 #define XFS_DFORK_APTR(dip)	\
-	((dip)->di_u.di_c + XFS_DFORK_BOFF(dip))
+	(XFS_DFORK_DPTR(dip) + XFS_DFORK_BOFF(dip))
 #define XFS_DFORK_PTR(dip,w)	\
 	((w) == XFS_DATA_FORK ? XFS_DFORK_DPTR(dip) : XFS_DFORK_APTR(dip))
+
 #define XFS_DFORK_FORMAT(dip,w) \
 	((w) == XFS_DATA_FORK ? \
-		(dip)->di_core.di_format : \
-		(dip)->di_core.di_aformat)
+		(dip)->di_format : \
+		(dip)->di_aformat)
 #define XFS_DFORK_NEXTENTS(dip,w) \
 	((w) == XFS_DATA_FORK ? \
-	 	be32_to_cpu((dip)->di_core.di_nextents) : \
-	 	be16_to_cpu((dip)->di_core.di_anextents))
+		be32_to_cpu((dip)->di_nextents) : \
+		be16_to_cpu((dip)->di_anextents))
 
 #define	XFS_BUF_TO_DINODE(bp)	((xfs_dinode_t *)XFS_BUF_PTR(bp))
 
 /*
+ * For block and character special files the 32bit dev_t is stored at the
+ * beginning of the data fork.
+ */
+static inline xfs_dev_t xfs_dinode_get_rdev(struct xfs_dinode *dip)
+{
+	return be32_to_cpu(*(__be32 *)XFS_DFORK_DPTR(dip));
+}
+
+static inline void xfs_dinode_put_rdev(struct xfs_dinode *dip, xfs_dev_t rdev)
+{
+	*(__be32 *)XFS_DFORK_DPTR(dip) = cpu_to_be32(rdev);
+}
+
+/*
  * Values for di_flags
  * There should be a one-to-one correspondence between these flags and the
  * XFS_XFLAG_s.
diff --git a/fs/xfs/xfs_dir2_sf.h b/fs/xfs/xfs_dir2_sf.h
index deecc9d238f..6ac44b550d3 100644
--- a/fs/xfs/xfs_dir2_sf.h
+++ b/fs/xfs/xfs_dir2_sf.h
@@ -34,13 +34,6 @@ struct xfs_mount;
 struct xfs_trans;
 
 /*
- * Maximum size of a shortform directory.
- */
-#define	XFS_DIR2_SF_MAX_SIZE	\
-	(XFS_DINODE_MAX_SIZE - (uint)sizeof(xfs_dinode_core_t) - \
-	 (uint)sizeof(xfs_agino_t))
-
-/*
  * Inode number stored as 8 8-bit values.
  */
 typedef	struct { __uint8_t i[8]; } xfs_dir2_ino8_t;
diff --git a/fs/xfs/xfs_dmops.c b/fs/xfs/xfs_dmops.c
index a1e55fb9d5d..e71e2581c0c 100644
--- a/fs/xfs/xfs_dmops.c
+++ b/fs/xfs/xfs_dmops.c
@@ -25,7 +25,6 @@
 #include "xfs_inum.h"
 #include "xfs_ag.h"
 #include "xfs_mount.h"
-#include "xfs_clnt.h"
 
 
 static struct xfs_dmops xfs_dmcore_stub = {
@@ -38,9 +37,9 @@ static struct xfs_dmops xfs_dmcore_stub = {
 };
 
 int
-xfs_dmops_get(struct xfs_mount *mp, struct xfs_mount_args *args)
+xfs_dmops_get(struct xfs_mount *mp)
 {
-	if (args->flags & XFSMNT_DMAPI) {
+	if (mp->m_flags & XFS_MOUNT_DMAPI) {
 		cmn_err(CE_WARN,
 			"XFS: dmapi support not available in this kernel.");
 		return EINVAL;
diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
index f227ecd1a29..92d5cd5bf4f 100644
--- a/fs/xfs/xfs_error.c
+++ b/fs/xfs/xfs_error.c
@@ -153,21 +153,6 @@ xfs_errortag_clearall(xfs_mount_t *mp, int loud)
 }
 #endif /* DEBUG */
 
-static void
-xfs_fs_vcmn_err(int level, xfs_mount_t *mp, char *fmt, va_list ap)
-{
-	if (mp != NULL) {
-		char	*newfmt;
-		int	len = 16 + mp->m_fsname_len + strlen(fmt);
-
-		newfmt = kmem_alloc(len, KM_SLEEP);
-		sprintf(newfmt, "Filesystem \"%s\": %s", mp->m_fsname, fmt);
-		icmn_err(level, newfmt, ap);
-		kmem_free(newfmt);
-	} else {
-		icmn_err(level, fmt, ap);
-	}
-}
 
 void
 xfs_fs_cmn_err(int level, xfs_mount_t *mp, char *fmt, ...)
diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h
index 11543f10b0c..0c93051c465 100644
--- a/fs/xfs/xfs_error.h
+++ b/fs/xfs/xfs_error.h
@@ -159,11 +159,15 @@ extern int xfs_errortag_clearall(xfs_mount_t *mp, int loud);
 #define		XFS_PTAG_FSBLOCK_ZERO		0x00000080
 
 struct xfs_mount;
-/* PRINTFLIKE4 */
+
+extern void xfs_fs_vcmn_err(int level, struct xfs_mount *mp,
+		char *fmt, va_list ap)
+	__attribute__ ((format (printf, 3, 0)));
 extern void xfs_cmn_err(int panic_tag, int level, struct xfs_mount *mp,
-			char *fmt, ...);
-/* PRINTFLIKE3 */
-extern void xfs_fs_cmn_err(int level, struct xfs_mount *mp, char *fmt, ...);
+			char *fmt, ...)
+	__attribute__ ((format (printf, 4, 5)));
+extern void xfs_fs_cmn_err(int level, struct xfs_mount *mp, char *fmt, ...)
+	__attribute__ ((format (printf, 3, 4)));
 
 extern void xfs_hex_dump(void *p, int length);
 
diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c
index 8aa28f751b2..05a4bdd4be3 100644
--- a/fs/xfs/xfs_extfree_item.c
+++ b/fs/xfs/xfs_extfree_item.c
@@ -108,19 +108,16 @@ xfs_efi_item_pin(xfs_efi_log_item_t *efip)
 STATIC void
 xfs_efi_item_unpin(xfs_efi_log_item_t *efip, int stale)
 {
-	xfs_mount_t	*mp;
+	struct xfs_ail		*ailp = efip->efi_item.li_ailp;
 
-	mp = efip->efi_item.li_mountp;
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&ailp->xa_lock);
 	if (efip->efi_flags & XFS_EFI_CANCELED) {
-		/*
-		 * xfs_trans_delete_ail() drops the AIL lock.
-		 */
-		xfs_trans_delete_ail(mp, (xfs_log_item_t *)efip);
+		/* xfs_trans_ail_delete() drops the AIL lock. */
+		xfs_trans_ail_delete(ailp, (xfs_log_item_t *)efip);
 		xfs_efi_item_free(efip);
 	} else {
 		efip->efi_flags |= XFS_EFI_COMMITTED;
-		spin_unlock(&mp->m_ail_lock);
+		spin_unlock(&ailp->xa_lock);
 	}
 }
 
@@ -134,26 +131,23 @@ xfs_efi_item_unpin(xfs_efi_log_item_t *efip, int stale)
 STATIC void
 xfs_efi_item_unpin_remove(xfs_efi_log_item_t *efip, xfs_trans_t *tp)
 {
-	xfs_mount_t	*mp;
+	struct xfs_ail		*ailp = efip->efi_item.li_ailp;
 	xfs_log_item_desc_t	*lidp;
 
-	mp = efip->efi_item.li_mountp;
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&ailp->xa_lock);
 	if (efip->efi_flags & XFS_EFI_CANCELED) {
 		/*
 		 * free the xaction descriptor pointing to this item
 		 */
 		lidp = xfs_trans_find_item(tp, (xfs_log_item_t *) efip);
 		xfs_trans_free_item(tp, lidp);
-		/*
-		 * pull the item off the AIL.
-		 * xfs_trans_delete_ail() drops the AIL lock.
-		 */
-		xfs_trans_delete_ail(mp, (xfs_log_item_t *)efip);
+
+		/* xfs_trans_ail_delete() drops the AIL lock. */
+		xfs_trans_ail_delete(ailp, (xfs_log_item_t *)efip);
 		xfs_efi_item_free(efip);
 	} else {
 		efip->efi_flags |= XFS_EFI_COMMITTED;
-		spin_unlock(&mp->m_ail_lock);
+		spin_unlock(&ailp->xa_lock);
 	}
 }
 
@@ -268,6 +262,7 @@ xfs_efi_init(xfs_mount_t	*mp,
 	efip->efi_item.li_type = XFS_LI_EFI;
 	efip->efi_item.li_ops = &xfs_efi_item_ops;
 	efip->efi_item.li_mountp = mp;
+	efip->efi_item.li_ailp = mp->m_ail;
 	efip->efi_format.efi_nextents = nextents;
 	efip->efi_format.efi_id = (__psint_t)(void*)efip;
 
@@ -345,25 +340,22 @@ void
 xfs_efi_release(xfs_efi_log_item_t	*efip,
 		uint			nextents)
 {
-	xfs_mount_t	*mp;
-	int		extents_left;
+	struct xfs_ail		*ailp = efip->efi_item.li_ailp;
+	int			extents_left;
 
-	mp = efip->efi_item.li_mountp;
 	ASSERT(efip->efi_next_extent > 0);
 	ASSERT(efip->efi_flags & XFS_EFI_COMMITTED);
 
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&ailp->xa_lock);
 	ASSERT(efip->efi_next_extent >= nextents);
 	efip->efi_next_extent -= nextents;
 	extents_left = efip->efi_next_extent;
 	if (extents_left == 0) {
-		/*
-		 * xfs_trans_delete_ail() drops the AIL lock.
-		 */
-		xfs_trans_delete_ail(mp, (xfs_log_item_t *)efip);
+		/* xfs_trans_ail_delete() drops the AIL lock. */
+		xfs_trans_ail_delete(ailp, (xfs_log_item_t *)efip);
 		xfs_efi_item_free(efip);
 	} else {
-		spin_unlock(&mp->m_ail_lock);
+		spin_unlock(&ailp->xa_lock);
 	}
 }
 
@@ -565,6 +557,7 @@ xfs_efd_init(xfs_mount_t	*mp,
 	efdp->efd_item.li_type = XFS_LI_EFD;
 	efdp->efd_item.li_ops = &xfs_efd_item_ops;
 	efdp->efd_item.li_mountp = mp;
+	efdp->efd_item.li_ailp = mp->m_ail;
 	efdp->efd_efip = efip;
 	efdp->efd_format.efd_nextents = nextents;
 	efdp->efd_format.efd_efi_id = efip->efi_format.efi_id;
diff --git a/fs/xfs/xfs_fs.h b/fs/xfs/xfs_fs.h
index 01c0cc88d3f..f7c06fac822 100644
--- a/fs/xfs/xfs_fs.h
+++ b/fs/xfs/xfs_fs.h
@@ -113,22 +113,14 @@ struct getbmapx {
 #define BMV_IF_ATTRFORK		0x1	/* return attr fork rather than data */
 #define BMV_IF_NO_DMAPI_READ	0x2	/* Do not generate DMAPI read event  */
 #define BMV_IF_PREALLOC		0x4	/* rtn status BMV_OF_PREALLOC if req */
-#define BMV_IF_VALID	(BMV_IF_ATTRFORK|BMV_IF_NO_DMAPI_READ|BMV_IF_PREALLOC)
-#ifdef __KERNEL__
-#define BMV_IF_EXTENDED 0x40000000	/* getpmapx if set */
-#endif
+#define BMV_IF_DELALLOC		0x8	/* rtn status BMV_OF_DELALLOC if req */
+#define BMV_IF_VALID	\
+	(BMV_IF_ATTRFORK|BMV_IF_NO_DMAPI_READ|BMV_IF_PREALLOC|BMV_IF_DELALLOC)
 
 /*	bmv_oflags values - returned for for each non-header segment */
 #define BMV_OF_PREALLOC		0x1	/* segment = unwritten pre-allocation */
-
-/*	Convert getbmap <-> getbmapx - move fields from p1 to p2. */
-#define GETBMAP_CONVERT(p1,p2) {	\
-	p2.bmv_offset = p1.bmv_offset;	\
-	p2.bmv_block = p1.bmv_block;	\
-	p2.bmv_length = p1.bmv_length;	\
-	p2.bmv_count = p1.bmv_count;	\
-	p2.bmv_entries = p1.bmv_entries;  }
-
+#define BMV_OF_DELALLOC		0x2	/* segment = delayed allocation */
+#define BMV_OF_LAST		0x4	/* segment is the last in the file */
 
 /*
  * Structure for XFS_IOC_FSSETDM.
@@ -426,10 +418,6 @@ typedef struct xfs_handle {
 #define XFS_IOC_GETXFLAGS	FS_IOC_GETFLAGS
 #define XFS_IOC_SETXFLAGS	FS_IOC_SETFLAGS
 #define XFS_IOC_GETVERSION	FS_IOC_GETVERSION
-/* 32-bit compat counterparts */
-#define XFS_IOC32_GETXFLAGS	FS_IOC32_GETFLAGS
-#define XFS_IOC32_SETXFLAGS	FS_IOC32_SETFLAGS
-#define XFS_IOC32_GETVERSION	FS_IOC32_GETVERSION
 
 /*
  * ioctl commands that replace IRIX fcntl()'s
@@ -477,8 +465,8 @@ typedef struct xfs_handle {
 #define XFS_IOC_ERROR_INJECTION	     _IOW ('X', 116, struct xfs_error_injection)
 #define XFS_IOC_ERROR_CLEARALL	     _IOW ('X', 117, struct xfs_error_injection)
 /*	XFS_IOC_ATTRCTL_BY_HANDLE -- deprecated 118	 */
-#define XFS_IOC_FREEZE		     _IOWR('X', 119, int)
-#define XFS_IOC_THAW		     _IOWR('X', 120, int)
+/*	XFS_IOC_FREEZE		  -- FIFREEZE   119	 */
+/*	XFS_IOC_THAW		  -- FITHAW     120	 */
 #define XFS_IOC_FSSETDM_BY_HANDLE    _IOW ('X', 121, struct xfs_fsop_setdm_handlereq)
 #define XFS_IOC_ATTRLIST_BY_HANDLE   _IOW ('X', 122, struct xfs_fsop_attrlist_handlereq)
 #define XFS_IOC_ATTRMULTI_BY_HANDLE  _IOW ('X', 123, struct xfs_fsop_attrmulti_handlereq)
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 84583cf73db..680d0e0ec93 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -126,7 +126,7 @@ xfs_growfs_data_private(
 	xfs_extlen_t		agsize;
 	xfs_extlen_t		tmpsize;
 	xfs_alloc_rec_t		*arec;
-	xfs_btree_sblock_t	*block;
+	struct xfs_btree_block	*block;
 	xfs_buf_t		*bp;
 	int			bucket;
 	int			dpct;
@@ -251,14 +251,14 @@ xfs_growfs_data_private(
 		bp = xfs_buf_get(mp->m_ddev_targp,
 			XFS_AGB_TO_DADDR(mp, agno, XFS_BNO_BLOCK(mp)),
 			BTOBB(mp->m_sb.sb_blocksize), 0);
-		block = XFS_BUF_TO_SBLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 		memset(block, 0, mp->m_sb.sb_blocksize);
 		block->bb_magic = cpu_to_be32(XFS_ABTB_MAGIC);
 		block->bb_level = 0;
 		block->bb_numrecs = cpu_to_be16(1);
-		block->bb_leftsib = cpu_to_be32(NULLAGBLOCK);
-		block->bb_rightsib = cpu_to_be32(NULLAGBLOCK);
-		arec = XFS_BTREE_REC_ADDR(xfs_alloc, block, 1);
+		block->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK);
+		block->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK);
+		arec = XFS_ALLOC_REC_ADDR(mp, block, 1);
 		arec->ar_startblock = cpu_to_be32(XFS_PREALLOC_BLOCKS(mp));
 		arec->ar_blockcount = cpu_to_be32(
 			agsize - be32_to_cpu(arec->ar_startblock));
@@ -272,14 +272,14 @@ xfs_growfs_data_private(
 		bp = xfs_buf_get(mp->m_ddev_targp,
 			XFS_AGB_TO_DADDR(mp, agno, XFS_CNT_BLOCK(mp)),
 			BTOBB(mp->m_sb.sb_blocksize), 0);
-		block = XFS_BUF_TO_SBLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 		memset(block, 0, mp->m_sb.sb_blocksize);
 		block->bb_magic = cpu_to_be32(XFS_ABTC_MAGIC);
 		block->bb_level = 0;
 		block->bb_numrecs = cpu_to_be16(1);
-		block->bb_leftsib = cpu_to_be32(NULLAGBLOCK);
-		block->bb_rightsib = cpu_to_be32(NULLAGBLOCK);
-		arec = XFS_BTREE_REC_ADDR(xfs_alloc, block, 1);
+		block->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK);
+		block->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK);
+		arec = XFS_ALLOC_REC_ADDR(mp, block, 1);
 		arec->ar_startblock = cpu_to_be32(XFS_PREALLOC_BLOCKS(mp));
 		arec->ar_blockcount = cpu_to_be32(
 			agsize - be32_to_cpu(arec->ar_startblock));
@@ -294,13 +294,13 @@ xfs_growfs_data_private(
 		bp = xfs_buf_get(mp->m_ddev_targp,
 			XFS_AGB_TO_DADDR(mp, agno, XFS_IBT_BLOCK(mp)),
 			BTOBB(mp->m_sb.sb_blocksize), 0);
-		block = XFS_BUF_TO_SBLOCK(bp);
+		block = XFS_BUF_TO_BLOCK(bp);
 		memset(block, 0, mp->m_sb.sb_blocksize);
 		block->bb_magic = cpu_to_be32(XFS_IBT_MAGIC);
 		block->bb_level = 0;
 		block->bb_numrecs = 0;
-		block->bb_leftsib = cpu_to_be32(NULLAGBLOCK);
-		block->bb_rightsib = cpu_to_be32(NULLAGBLOCK);
+		block->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK);
+		block->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK);
 		error = xfs_bwrite(mp, bp);
 		if (error) {
 			goto error0;
@@ -435,6 +435,9 @@ xfs_growfs_data(
 	xfs_growfs_data_t	*in)
 {
 	int error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return XFS_ERROR(EPERM);
 	if (!mutex_trylock(&mp->m_growlock))
 		return XFS_ERROR(EWOULDBLOCK);
 	error = xfs_growfs_data_private(mp, in);
@@ -448,6 +451,9 @@ xfs_growfs_log(
 	xfs_growfs_log_t	*in)
 {
 	int error;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return XFS_ERROR(EPERM);
 	if (!mutex_trylock(&mp->m_growlock))
 		return XFS_ERROR(EWOULDBLOCK);
 	error = xfs_growfs_log_private(mp, in);
@@ -589,17 +595,19 @@ out:
 	return 0;
 }
 
-void
+int
 xfs_fs_log_dummy(
 	xfs_mount_t	*mp)
 {
 	xfs_trans_t	*tp;
 	xfs_inode_t	*ip;
+	int		error;
 
 	tp = _xfs_trans_alloc(mp, XFS_TRANS_DUMMY1);
-	if (xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0)) {
+	error = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0);
+	if (error) {
 		xfs_trans_cancel(tp, 0);
-		return;
+		return error;
 	}
 
 	ip = mp->m_rootip;
@@ -609,9 +617,10 @@ xfs_fs_log_dummy(
 	xfs_trans_ihold(tp, ip);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	xfs_trans_set_sync(tp);
-	xfs_trans_commit(tp, 0);
+	error = xfs_trans_commit(tp, 0);
 
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	return error;
 }
 
 int
diff --git a/fs/xfs/xfs_fsops.h b/fs/xfs/xfs_fsops.h
index 300d0c9d61a..88435e0a77c 100644
--- a/fs/xfs/xfs_fsops.h
+++ b/fs/xfs/xfs_fsops.h
@@ -25,6 +25,6 @@ extern int xfs_fs_counts(xfs_mount_t *mp, xfs_fsop_counts_t *cnt);
 extern int xfs_reserve_blocks(xfs_mount_t *mp, __uint64_t *inval,
 				xfs_fsop_resblks_t *outval);
 extern int xfs_fs_goingdown(xfs_mount_t *mp, __uint32_t inflags);
-extern void xfs_fs_log_dummy(xfs_mount_t *mp);
+extern int xfs_fs_log_dummy(xfs_mount_t *mp);
 
 #endif	/* __XFS_FSOPS_H__ */
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index aad8c5da38a..e6ebbaeb4dc 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -41,68 +41,6 @@
 #include "xfs_error.h"
 #include "xfs_bmap.h"
 
-/*
- * Log specified fields for the inode given by bp and off.
- */
-STATIC void
-xfs_ialloc_log_di(
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_buf_t	*bp,		/* inode buffer */
-	int		off,		/* index of inode in buffer */
-	int		fields)		/* bitmask of fields to log */
-{
-	int			first;		/* first byte number */
-	int			ioffset;	/* off in bytes */
-	int			last;		/* last byte number */
-	xfs_mount_t		*mp;		/* mount point structure */
-	static const short	offsets[] = {	/* field offsets */
-						/* keep in sync with bits */
-		offsetof(xfs_dinode_core_t, di_magic),
-		offsetof(xfs_dinode_core_t, di_mode),
-		offsetof(xfs_dinode_core_t, di_version),
-		offsetof(xfs_dinode_core_t, di_format),
-		offsetof(xfs_dinode_core_t, di_onlink),
-		offsetof(xfs_dinode_core_t, di_uid),
-		offsetof(xfs_dinode_core_t, di_gid),
-		offsetof(xfs_dinode_core_t, di_nlink),
-		offsetof(xfs_dinode_core_t, di_projid),
-		offsetof(xfs_dinode_core_t, di_pad),
-		offsetof(xfs_dinode_core_t, di_atime),
-		offsetof(xfs_dinode_core_t, di_mtime),
-		offsetof(xfs_dinode_core_t, di_ctime),
-		offsetof(xfs_dinode_core_t, di_size),
-		offsetof(xfs_dinode_core_t, di_nblocks),
-		offsetof(xfs_dinode_core_t, di_extsize),
-		offsetof(xfs_dinode_core_t, di_nextents),
-		offsetof(xfs_dinode_core_t, di_anextents),
-		offsetof(xfs_dinode_core_t, di_forkoff),
-		offsetof(xfs_dinode_core_t, di_aformat),
-		offsetof(xfs_dinode_core_t, di_dmevmask),
-		offsetof(xfs_dinode_core_t, di_dmstate),
-		offsetof(xfs_dinode_core_t, di_flags),
-		offsetof(xfs_dinode_core_t, di_gen),
-		offsetof(xfs_dinode_t, di_next_unlinked),
-		offsetof(xfs_dinode_t, di_u),
-		offsetof(xfs_dinode_t, di_a),
-		sizeof(xfs_dinode_t)
-	};
-
-
-	ASSERT(offsetof(xfs_dinode_t, di_core) == 0);
-	ASSERT((fields & (XFS_DI_U|XFS_DI_A)) == 0);
-	mp = tp->t_mountp;
-	/*
-	 * Get the inode-relative first and last bytes for these fields
-	 */
-	xfs_btree_offsets(fields, offsets, XFS_DI_NUM_BITS, &first, &last);
-	/*
-	 * Convert to buffer offsets and log it.
-	 */
-	ioffset = off << mp->m_sb.sb_inodelog;
-	first += ioffset;
-	last += ioffset;
-	xfs_trans_log_buf(tp, bp, first, last);
-}
 
 /*
  * Allocation group level functions.
@@ -119,6 +57,102 @@ xfs_ialloc_cluster_alignment(
 }
 
 /*
+ * Lookup the record equal to ino in the btree given by cur.
+ */
+STATIC int				/* error */
+xfs_inobt_lookup_eq(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agino_t		ino,	/* starting inode of chunk */
+	__int32_t		fcnt,	/* free inode count */
+	xfs_inofree_t		free,	/* free inode mask */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.i.ir_startino = ino;
+	cur->bc_rec.i.ir_freecount = fcnt;
+	cur->bc_rec.i.ir_free = free;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
+}
+
+/*
+ * Lookup the first record greater than or equal to ino
+ * in the btree given by cur.
+ */
+int					/* error */
+xfs_inobt_lookup_ge(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agino_t		ino,	/* starting inode of chunk */
+	__int32_t		fcnt,	/* free inode count */
+	xfs_inofree_t		free,	/* free inode mask */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.i.ir_startino = ino;
+	cur->bc_rec.i.ir_freecount = fcnt;
+	cur->bc_rec.i.ir_free = free;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
+}
+
+/*
+ * Lookup the first record less than or equal to ino
+ * in the btree given by cur.
+ */
+int					/* error */
+xfs_inobt_lookup_le(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agino_t		ino,	/* starting inode of chunk */
+	__int32_t		fcnt,	/* free inode count */
+	xfs_inofree_t		free,	/* free inode mask */
+	int			*stat)	/* success/failure */
+{
+	cur->bc_rec.i.ir_startino = ino;
+	cur->bc_rec.i.ir_freecount = fcnt;
+	cur->bc_rec.i.ir_free = free;
+	return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
+}
+
+/*
+ * Update the record referred to by cur to the value given
+ * by [ino, fcnt, free].
+ * This either works (return 0) or gets an EFSCORRUPTED error.
+ */
+STATIC int				/* error */
+xfs_inobt_update(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agino_t		ino,	/* starting inode of chunk */
+	__int32_t		fcnt,	/* free inode count */
+	xfs_inofree_t		free)	/* free inode mask */
+{
+	union xfs_btree_rec	rec;
+
+	rec.inobt.ir_startino = cpu_to_be32(ino);
+	rec.inobt.ir_freecount = cpu_to_be32(fcnt);
+	rec.inobt.ir_free = cpu_to_be64(free);
+	return xfs_btree_update(cur, &rec);
+}
+
+/*
+ * Get the data from the pointed-to record.
+ */
+int					/* error */
+xfs_inobt_get_rec(
+	struct xfs_btree_cur	*cur,	/* btree cursor */
+	xfs_agino_t		*ino,	/* output: starting inode of chunk */
+	__int32_t		*fcnt,	/* output: number of free inodes */
+	xfs_inofree_t		*free,	/* output: free inode mask */
+	int			*stat)	/* output: success/failure */
+{
+	union xfs_btree_rec	*rec;
+	int			error;
+
+	error = xfs_btree_get_rec(cur, &rec, stat);
+	if (!error && *stat == 1) {
+		*ino = be32_to_cpu(rec->inobt.ir_startino);
+		*fcnt = be32_to_cpu(rec->inobt.ir_freecount);
+		*free = be64_to_cpu(rec->inobt.ir_free);
+	}
+	return error;
+}
+
+/*
  * Allocate new inodes in the allocation group specified by agbp.
  * Return 0 for success, else error code.
  */
@@ -287,9 +321,9 @@ xfs_ialloc_ag_alloc(
 	 * able to use the file system.
 	 */
 	if (xfs_sb_version_hasnlink(&args.mp->m_sb))
-		version = XFS_DINODE_VERSION_2;
+		version = 2;
 	else
-		version = XFS_DINODE_VERSION_1;
+		version = 1;
 
 	/*
 	 * Seed the new inode cluster with a random generation number. This
@@ -310,18 +344,25 @@ xfs_ialloc_ag_alloc(
 					 XFS_BUF_LOCK);
 		ASSERT(fbuf);
 		ASSERT(!XFS_BUF_GETERROR(fbuf));
+
 		/*
-		 * Set initial values for the inodes in this buffer.
+		 * Initialize all inodes in this buffer and then log them.
+		 *
+		 * XXX: It would be much better if we had just one transaction to
+		 *      log a whole cluster of inodes instead of all the indivdual
+		 *      transactions causing a lot of log traffic.
 		 */
 		xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
 		for (i = 0; i < ninodes; i++) {
+			int	ioffset = i << args.mp->m_sb.sb_inodelog;
+			uint	isize = sizeof(struct xfs_dinode);
+
 			free = XFS_MAKE_IPTR(args.mp, fbuf, i);
-			free->di_core.di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
-			free->di_core.di_version = version;
-			free->di_core.di_gen = cpu_to_be32(gen);
+			free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
+			free->di_version = version;
+			free->di_gen = cpu_to_be32(gen);
 			free->di_next_unlinked = cpu_to_be32(NULLAGINO);
-			xfs_ialloc_log_di(tp, fbuf, i,
-				XFS_DI_CORE_BITS | XFS_DI_NEXT_UNLINKED);
+			xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
 		}
 		xfs_trans_inode_alloc_buf(tp, fbuf);
 	}
@@ -335,8 +376,7 @@ xfs_ialloc_ag_alloc(
 	/*
 	 * Insert records describing the new inode chunk into the btree.
 	 */
-	cur = xfs_btree_init_cursor(args.mp, tp, agbp, agno,
-			XFS_BTNUM_INO, (xfs_inode_t *)0, 0);
+	cur = xfs_inobt_init_cursor(args.mp, tp, agbp, agno);
 	for (thisino = newino;
 	     thisino < newino + newlen;
 	     thisino += XFS_INODES_PER_CHUNK) {
@@ -346,7 +386,7 @@ xfs_ialloc_ag_alloc(
 			return error;
 		}
 		ASSERT(i == 0);
-		if ((error = xfs_inobt_insert(cur, &i))) {
+		if ((error = xfs_btree_insert(cur, &i))) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 			return error;
 		}
@@ -676,8 +716,7 @@ nextag:
 	 */
 	agno = tagno;
 	*IO_agbp = NULL;
-	cur = xfs_btree_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno),
-				    XFS_BTNUM_INO, (xfs_inode_t *)0, 0);
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno));
 	/*
 	 * If pagino is 0 (this is the root inode allocation) use newino.
 	 * This must work because we've just allocated some.
@@ -697,7 +736,7 @@ nextag:
 				goto error0;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 			freecount += rec.ir_freecount;
-			if ((error = xfs_inobt_increment(cur, 0, &i)))
+			if ((error = xfs_btree_increment(cur, 0, &i)))
 				goto error0;
 		} while (i == 1);
 
@@ -741,7 +780,7 @@ nextag:
 			/*
 			 * Search left with tcur, back up 1 record.
 			 */
-			if ((error = xfs_inobt_decrement(tcur, 0, &i)))
+			if ((error = xfs_btree_decrement(tcur, 0, &i)))
 				goto error1;
 			doneleft = !i;
 			if (!doneleft) {
@@ -755,7 +794,7 @@ nextag:
 			/*
 			 * Search right with cur, go forward 1 record.
 			 */
-			if ((error = xfs_inobt_increment(cur, 0, &i)))
+			if ((error = xfs_btree_increment(cur, 0, &i)))
 				goto error1;
 			doneright = !i;
 			if (!doneright) {
@@ -817,7 +856,7 @@ nextag:
 				 * further left.
 				 */
 				if (useleft) {
-					if ((error = xfs_inobt_decrement(tcur, 0,
+					if ((error = xfs_btree_decrement(tcur, 0,
 							&i)))
 						goto error1;
 					doneleft = !i;
@@ -837,7 +876,7 @@ nextag:
 				 * further right.
 				 */
 				else {
-					if ((error = xfs_inobt_increment(cur, 0,
+					if ((error = xfs_btree_increment(cur, 0,
 							&i)))
 						goto error1;
 					doneright = !i;
@@ -892,7 +931,7 @@ nextag:
 				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 				if (rec.ir_freecount > 0)
 					break;
-				if ((error = xfs_inobt_increment(cur, 0, &i)))
+				if ((error = xfs_btree_increment(cur, 0, &i)))
 					goto error0;
 				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 			}
@@ -926,7 +965,7 @@ nextag:
 				goto error0;
 			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
 			freecount += rec.ir_freecount;
-			if ((error = xfs_inobt_increment(cur, 0, &i)))
+			if ((error = xfs_btree_increment(cur, 0, &i)))
 				goto error0;
 		} while (i == 1);
 		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
@@ -1022,8 +1061,7 @@ xfs_difree(
 	/*
 	 * Initialize the cursor.
 	 */
-	cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO,
-		(xfs_inode_t *)0, 0);
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
 #ifdef DEBUG
 	if (cur->bc_nlevels == 1) {
 		int freecount = 0;
@@ -1036,7 +1074,7 @@ xfs_difree(
 				goto error0;
 			if (i) {
 				freecount += rec.ir_freecount;
-				if ((error = xfs_inobt_increment(cur, 0, &i)))
+				if ((error = xfs_btree_increment(cur, 0, &i)))
 					goto error0;
 			}
 		} while (i == 1);
@@ -1098,8 +1136,8 @@ xfs_difree(
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, -ilen);
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -(ilen - 1));
 
-		if ((error = xfs_inobt_delete(cur, &i))) {
-			cmn_err(CE_WARN, "xfs_difree: xfs_inobt_delete returned an error %d on %s.\n",
+		if ((error = xfs_btree_delete(cur, &i))) {
+			cmn_err(CE_WARN, "xfs_difree: xfs_btree_delete returned an error %d on %s.\n",
 				error, mp->m_fsname);
 			goto error0;
 		}
@@ -1141,7 +1179,7 @@ xfs_difree(
 				goto error0;
 			if (i) {
 				freecount += rec.ir_freecount;
-				if ((error = xfs_inobt_increment(cur, 0, &i)))
+				if ((error = xfs_btree_increment(cur, 0, &i)))
 					goto error0;
 			}
 		} while (i == 1);
@@ -1158,36 +1196,28 @@ error0:
 }
 
 /*
- * Return the location of the inode in bno/off, for mapping it into a buffer.
+ * Return the location of the inode in imap, for mapping it into a buffer.
  */
-/*ARGSUSED*/
 int
-xfs_dilocate(
-	xfs_mount_t	*mp,	/* file system mount structure */
-	xfs_trans_t	*tp,	/* transaction pointer */
+xfs_imap(
+	xfs_mount_t	 *mp,	/* file system mount structure */
+	xfs_trans_t	 *tp,	/* transaction pointer */
 	xfs_ino_t	ino,	/* inode to locate */
-	xfs_fsblock_t	*bno,	/* output: block containing inode */
-	int		*len,	/* output: num blocks in inode cluster */
-	int		*off,	/* output: index in block of inode */
-	uint		flags)	/* flags concerning inode lookup */
+	struct xfs_imap	*imap,	/* location map structure */
+	uint		flags)	/* flags for inode btree lookup */
 {
 	xfs_agblock_t	agbno;	/* block number of inode in the alloc group */
-	xfs_buf_t	*agbp;	/* agi buffer */
 	xfs_agino_t	agino;	/* inode number within alloc group */
 	xfs_agnumber_t	agno;	/* allocation group number */
 	int		blks_per_cluster; /* num blocks per inode cluster */
 	xfs_agblock_t	chunk_agbno;	/* first block in inode chunk */
-	xfs_agino_t	chunk_agino;	/* first agino in inode chunk */
-	__int32_t	chunk_cnt;	/* count of free inodes in chunk */
-	xfs_inofree_t	chunk_free;	/* mask of free inodes in chunk */
 	xfs_agblock_t	cluster_agbno;	/* first block in inode cluster */
-	xfs_btree_cur_t	*cur;	/* inode btree cursor */
 	int		error;	/* error code */
-	int		i;	/* temp state */
 	int		offset;	/* index of inode in its buffer */
 	int		offset_agbno;	/* blks from chunk start to inode */
 
 	ASSERT(ino != NULLFSINO);
+
 	/*
 	 * Split up the inode number into its parts.
 	 */
@@ -1198,24 +1228,24 @@ xfs_dilocate(
 	    ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
 #ifdef DEBUG
 		/* no diagnostics for bulkstat, ino comes from userspace */
-		if (flags & XFS_IMAP_BULKSTAT)
+		if (flags & XFS_IGET_BULKSTAT)
 			return XFS_ERROR(EINVAL);
 		if (agno >= mp->m_sb.sb_agcount) {
 			xfs_fs_cmn_err(CE_ALERT, mp,
-					"xfs_dilocate: agno (%d) >= "
+					"xfs_imap: agno (%d) >= "
 					"mp->m_sb.sb_agcount (%d)",
 					agno,  mp->m_sb.sb_agcount);
 		}
 		if (agbno >= mp->m_sb.sb_agblocks) {
 			xfs_fs_cmn_err(CE_ALERT, mp,
-					"xfs_dilocate: agbno (0x%llx) >= "
+					"xfs_imap: agbno (0x%llx) >= "
 					"mp->m_sb.sb_agblocks (0x%lx)",
 					(unsigned long long) agbno,
 					(unsigned long) mp->m_sb.sb_agblocks);
 		}
 		if (ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
 			xfs_fs_cmn_err(CE_ALERT, mp,
-					"xfs_dilocate: ino (0x%llx) != "
+					"xfs_imap: ino (0x%llx) != "
 					"XFS_AGINO_TO_INO(mp, agno, agino) "
 					"(0x%llx)",
 					ino, XFS_AGINO_TO_INO(mp, agno, agino));
@@ -1224,65 +1254,89 @@ xfs_dilocate(
 #endif /* DEBUG */
 		return XFS_ERROR(EINVAL);
 	}
-	if ((mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) ||
-	    !(flags & XFS_IMAP_LOOKUP)) {
+
+	/*
+	 * If the inode cluster size is the same as the blocksize or
+	 * smaller we get to the buffer by simple arithmetics.
+	 */
+	if (XFS_INODE_CLUSTER_SIZE(mp) <= mp->m_sb.sb_blocksize) {
 		offset = XFS_INO_TO_OFFSET(mp, ino);
 		ASSERT(offset < mp->m_sb.sb_inopblock);
-		*bno = XFS_AGB_TO_FSB(mp, agno, agbno);
-		*off = offset;
-		*len = 1;
+
+		imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, agbno);
+		imap->im_len = XFS_FSB_TO_BB(mp, 1);
+		imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog);
 		return 0;
 	}
+
 	blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_blocklog;
-	if (*bno != NULLFSBLOCK) {
+
+	/*
+	 * If we get a block number passed from bulkstat we can use it to
+	 * find the buffer easily.
+	 */
+	if (imap->im_blkno) {
 		offset = XFS_INO_TO_OFFSET(mp, ino);
 		ASSERT(offset < mp->m_sb.sb_inopblock);
-		cluster_agbno = XFS_FSB_TO_AGBNO(mp, *bno);
-		*off = ((agbno - cluster_agbno) * mp->m_sb.sb_inopblock) +
-			offset;
-		*len = blks_per_cluster;
+
+		cluster_agbno = XFS_DADDR_TO_AGBNO(mp, imap->im_blkno);
+		offset += (agbno - cluster_agbno) * mp->m_sb.sb_inopblock;
+
+		imap->im_len = XFS_FSB_TO_BB(mp, blks_per_cluster);
+		imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog);
 		return 0;
 	}
+
+	/*
+	 * If the inode chunks are aligned then use simple maths to
+	 * find the location. Otherwise we have to do a btree
+	 * lookup to find the location.
+	 */
 	if (mp->m_inoalign_mask) {
 		offset_agbno = agbno & mp->m_inoalign_mask;
 		chunk_agbno = agbno - offset_agbno;
 	} else {
+		xfs_btree_cur_t	*cur;	/* inode btree cursor */
+		xfs_agino_t	chunk_agino; /* first agino in inode chunk */
+		__int32_t	chunk_cnt; /* count of free inodes in chunk */
+		xfs_inofree_t	chunk_free; /* mask of free inodes in chunk */
+		xfs_buf_t	*agbp;	/* agi buffer */
+		int		i;	/* temp state */
+
 		down_read(&mp->m_peraglock);
 		error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
 		up_read(&mp->m_peraglock);
 		if (error) {
-#ifdef DEBUG
-			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
+			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
 					"xfs_ialloc_read_agi() returned "
 					"error %d, agno %d",
 					error, agno);
-#endif /* DEBUG */
 			return error;
 		}
-		cur = xfs_btree_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO,
-			(xfs_inode_t *)0, 0);
-		if ((error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i))) {
-#ifdef DEBUG
-			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
+
+		cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
+		error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i);
+		if (error) {
+			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
 					"xfs_inobt_lookup_le() failed");
-#endif /* DEBUG */
 			goto error0;
 		}
-		if ((error = xfs_inobt_get_rec(cur, &chunk_agino, &chunk_cnt,
-				&chunk_free, &i))) {
-#ifdef DEBUG
-			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
+
+		error = xfs_inobt_get_rec(cur, &chunk_agino, &chunk_cnt,
+				&chunk_free, &i);
+		if (error) {
+			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
 					"xfs_inobt_get_rec() failed");
-#endif /* DEBUG */
 			goto error0;
 		}
 		if (i == 0) {
 #ifdef DEBUG
-			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_dilocate: "
+			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
 					"xfs_inobt_get_rec() failed");
 #endif /* DEBUG */
 			error = XFS_ERROR(EINVAL);
 		}
+ error0:
 		xfs_trans_brelse(tp, agbp);
 		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 		if (error)
@@ -1290,19 +1344,35 @@ xfs_dilocate(
 		chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_agino);
 		offset_agbno = agbno - chunk_agbno;
 	}
+
 	ASSERT(agbno >= chunk_agbno);
 	cluster_agbno = chunk_agbno +
 		((offset_agbno / blks_per_cluster) * blks_per_cluster);
 	offset = ((agbno - cluster_agbno) * mp->m_sb.sb_inopblock) +
 		XFS_INO_TO_OFFSET(mp, ino);
-	*bno = XFS_AGB_TO_FSB(mp, agno, cluster_agbno);
-	*off = offset;
-	*len = blks_per_cluster;
+
+	imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, cluster_agbno);
+	imap->im_len = XFS_FSB_TO_BB(mp, blks_per_cluster);
+	imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog);
+
+	/*
+	 * If the inode number maps to a block outside the bounds
+	 * of the file system then return NULL rather than calling
+	 * read_buf and panicing when we get an error from the
+	 * driver.
+	 */
+	if ((imap->im_blkno + imap->im_len) >
+	    XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) {
+		xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
+			"(imap->im_blkno (0x%llx) + imap->im_len (0x%llx)) > "
+			" XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks) (0x%llx)",
+			(unsigned long long) imap->im_blkno,
+			(unsigned long long) imap->im_len,
+			XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks));
+		return XFS_ERROR(EINVAL);
+	}
+
 	return 0;
-error0:
-	xfs_trans_brelse(tp, agbp);
-	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
-	return error;
 }
 
 /*
@@ -1370,70 +1440,95 @@ xfs_ialloc_log_agi(
 	xfs_trans_log_buf(tp, bp, first, last);
 }
 
+#ifdef DEBUG
+STATIC void
+xfs_check_agi_unlinked(
+	struct xfs_agi		*agi)
+{
+	int			i;
+
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++)
+		ASSERT(agi->agi_unlinked[i]);
+}
+#else
+#define xfs_check_agi_unlinked(agi)
+#endif
+
 /*
  * Read in the allocation group header (inode allocation section)
  */
 int
-xfs_ialloc_read_agi(
-	xfs_mount_t	*mp,		/* file system mount structure */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_buf_t	**bpp)		/* allocation group hdr buf */
+xfs_read_agi(
+	struct xfs_mount	*mp,	/* file system mount structure */
+	struct xfs_trans	*tp,	/* transaction pointer */
+	xfs_agnumber_t		agno,	/* allocation group number */
+	struct xfs_buf		**bpp)	/* allocation group hdr buf */
 {
-	xfs_agi_t	*agi;		/* allocation group header */
-	int		agi_ok;		/* agi is consistent */
-	xfs_buf_t	*bp;		/* allocation group hdr buf */
-	xfs_perag_t	*pag;		/* per allocation group data */
-	int		error;
+	struct xfs_agi		*agi;	/* allocation group header */
+	int			agi_ok;	/* agi is consistent */
+	int			error;
 
 	ASSERT(agno != NULLAGNUMBER);
-	error = xfs_trans_read_buf(
-			mp, tp, mp->m_ddev_targp,
+
+	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
 			XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-			XFS_FSS_TO_BB(mp, 1), 0, &bp);
+			XFS_FSS_TO_BB(mp, 1), 0, bpp);
 	if (error)
 		return error;
-	ASSERT(bp && !XFS_BUF_GETERROR(bp));
+
+	ASSERT(*bpp && !XFS_BUF_GETERROR(*bpp));
+	agi = XFS_BUF_TO_AGI(*bpp);
 
 	/*
 	 * Validate the magic number of the agi block.
 	 */
-	agi = XFS_BUF_TO_AGI(bp);
-	agi_ok =
-		be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
-		XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum));
+	agi_ok = be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
+		XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum)) &&
+		be32_to_cpu(agi->agi_seqno) == agno;
 	if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IALLOC_READ_AGI,
 			XFS_RANDOM_IALLOC_READ_AGI))) {
-		XFS_CORRUPTION_ERROR("xfs_ialloc_read_agi", XFS_ERRLEVEL_LOW,
+		XFS_CORRUPTION_ERROR("xfs_read_agi", XFS_ERRLEVEL_LOW,
 				     mp, agi);
-		xfs_trans_brelse(tp, bp);
+		xfs_trans_brelse(tp, *bpp);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
+
+	XFS_BUF_SET_VTYPE_REF(*bpp, B_FS_AGI, XFS_AGI_REF);
+
+	xfs_check_agi_unlinked(agi);
+	return 0;
+}
+
+int
+xfs_ialloc_read_agi(
+	struct xfs_mount	*mp,	/* file system mount structure */
+	struct xfs_trans	*tp,	/* transaction pointer */
+	xfs_agnumber_t		agno,	/* allocation group number */
+	struct xfs_buf		**bpp)	/* allocation group hdr buf */
+{
+	struct xfs_agi		*agi;	/* allocation group header */
+	struct xfs_perag	*pag;	/* per allocation group data */
+	int			error;
+
+	error = xfs_read_agi(mp, tp, agno, bpp);
+	if (error)
+		return error;
+
+	agi = XFS_BUF_TO_AGI(*bpp);
 	pag = &mp->m_perag[agno];
+
 	if (!pag->pagi_init) {
 		pag->pagi_freecount = be32_to_cpu(agi->agi_freecount);
 		pag->pagi_count = be32_to_cpu(agi->agi_count);
 		pag->pagi_init = 1;
-	} else {
-		/*
-		 * It's possible for these to be out of sync if
-		 * we are in the middle of a forced shutdown.
-		 */
-		ASSERT(pag->pagi_freecount == be32_to_cpu(agi->agi_freecount) ||
-			XFS_FORCED_SHUTDOWN(mp));
 	}
 
-#ifdef DEBUG
-	{
-		int	i;
-
-		for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++)
-			ASSERT(agi->agi_unlinked[i]);
-	}
-#endif
-
-	XFS_BUF_SET_VTYPE_REF(bp, B_FS_AGI, XFS_AGI_REF);
-	*bpp = bp;
+	/*
+	 * It's possible for these to be out of sync if
+	 * we are in the middle of a forced shutdown.
+	 */
+	ASSERT(pag->pagi_freecount == be32_to_cpu(agi->agi_freecount) ||
+		XFS_FORCED_SHUTDOWN(mp));
 	return 0;
 }
 
diff --git a/fs/xfs/xfs_ialloc.h b/fs/xfs/xfs_ialloc.h
index 4e30ec1d13b..50f558a4e0a 100644
--- a/fs/xfs/xfs_ialloc.h
+++ b/fs/xfs/xfs_ialloc.h
@@ -20,6 +20,7 @@
 
 struct xfs_buf;
 struct xfs_dinode;
+struct xfs_imap;
 struct xfs_mount;
 struct xfs_trans;
 
@@ -56,7 +57,6 @@ static inline int xfs_ialloc_find_free(xfs_inofree_t *fp)
 }
 
 
-#ifdef __KERNEL__
 /*
  * Allocate an inode on disk.
  * Mode is used to tell whether the new inode will need space, and whether
@@ -105,17 +105,14 @@ xfs_difree(
 	xfs_ino_t	*first_ino);	/* first inode in deleted cluster */
 
 /*
- * Return the location of the inode in bno/len/off,
- * for mapping it into a buffer.
+ * Return the location of the inode in imap, for mapping it into a buffer.
  */
 int
-xfs_dilocate(
+xfs_imap(
 	struct xfs_mount *mp,		/* file system mount structure */
 	struct xfs_trans *tp,		/* transaction pointer */
 	xfs_ino_t	ino,		/* inode to locate */
-	xfs_fsblock_t	*bno,		/* output: block containing inode */
-	int		*len,		/* output: num blocks in cluster*/
-	int		*off,		/* output: index in block of inode */
+	struct xfs_imap	*imap,		/* location map structure */
 	uint		flags);		/* flags for inode btree lookup */
 
 /*
@@ -154,6 +151,24 @@ xfs_ialloc_pagi_init(
 	struct xfs_trans *tp,		/* transaction pointer */
         xfs_agnumber_t  agno);		/* allocation group number */
 
-#endif	/* __KERNEL__ */
+/*
+ * Lookup the first record greater than or equal to ino
+ * in the btree given by cur.
+ */
+int xfs_inobt_lookup_ge(struct xfs_btree_cur *cur, xfs_agino_t ino,
+		__int32_t fcnt,	xfs_inofree_t free, int *stat);
+
+/*
+ * Lookup the first record less than or equal to ino
+ * in the btree given by cur.
+ */
+int xfs_inobt_lookup_le(struct xfs_btree_cur *cur, xfs_agino_t ino,
+		__int32_t fcnt,	xfs_inofree_t free, int *stat);
+
+/*
+ * Get the data from the pointed-to record.
+ */
+extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur, xfs_agino_t *ino,
+			     __int32_t *fcnt, xfs_inofree_t *free, int *stat);
 
 #endif	/* __XFS_IALLOC_H__ */
diff --git a/fs/xfs/xfs_ialloc_btree.c b/fs/xfs/xfs_ialloc_btree.c
index 83502f3edef..99f2408e8d8 100644
--- a/fs/xfs/xfs_ialloc_btree.c
+++ b/fs/xfs/xfs_ialloc_btree.c
@@ -35,2044 +35,349 @@
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_btree.h"
+#include "xfs_btree_trace.h"
 #include "xfs_ialloc.h"
 #include "xfs_alloc.h"
 #include "xfs_error.h"
 
-STATIC void xfs_inobt_log_block(xfs_trans_t *, xfs_buf_t *, int);
-STATIC void xfs_inobt_log_keys(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_inobt_log_ptrs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC void xfs_inobt_log_recs(xfs_btree_cur_t *, xfs_buf_t *, int, int);
-STATIC int xfs_inobt_lshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_inobt_newroot(xfs_btree_cur_t *, int *);
-STATIC int xfs_inobt_rshift(xfs_btree_cur_t *, int, int *);
-STATIC int xfs_inobt_split(xfs_btree_cur_t *, int, xfs_agblock_t *,
-		xfs_inobt_key_t *, xfs_btree_cur_t **, int *);
-STATIC int xfs_inobt_updkey(xfs_btree_cur_t *, xfs_inobt_key_t *, int);
 
-/*
- * Single level of the xfs_inobt_delete record deletion routine.
- * Delete record pointed to by cur/level.
- * Remove the record from its block then rebalance the tree.
- * Return 0 for error, 1 for done, 2 to go on to the next level.
- */
-STATIC int				/* error */
-xfs_inobt_delrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level removing record from */
-	int			*stat)	/* fail/done/go-on */
+STATIC int
+xfs_inobt_get_minrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
 {
-	xfs_buf_t		*agbp;	/* buffer for a.g. inode header */
-	xfs_mount_t		*mp;	/* mount structure */
-	xfs_agi_t		*agi;	/* allocation group inode header */
-	xfs_inobt_block_t	*block;	/* btree block record/key lives in */
-	xfs_agblock_t		bno;	/* btree block number */
-	xfs_buf_t		*bp;	/* buffer for block */
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_inobt_key_t		key;	/* kp points here if block is level 0 */
-	xfs_inobt_key_t		*kp = NULL;	/* pointer to btree keys */
-	xfs_agblock_t		lbno;	/* left block's block number */
-	xfs_buf_t		*lbp;	/* left block's buffer pointer */
-	xfs_inobt_block_t	*left;	/* left btree block */
-	xfs_inobt_key_t		*lkp;	/* left block key pointer */
-	xfs_inobt_ptr_t		*lpp;	/* left block address pointer */
-	int			lrecs = 0;	/* number of records in left block */
-	xfs_inobt_rec_t		*lrp;	/* left block record pointer */
-	xfs_inobt_ptr_t		*pp = NULL;	/* pointer to btree addresses */
-	int			ptr;	/* index in btree block for this rec */
-	xfs_agblock_t		rbno;	/* right block's block number */
-	xfs_buf_t		*rbp;	/* right block's buffer pointer */
-	xfs_inobt_block_t	*right;	/* right btree block */
-	xfs_inobt_key_t		*rkp;	/* right block key pointer */
-	xfs_inobt_rec_t		*rp;	/* pointer to btree records */
-	xfs_inobt_ptr_t		*rpp;	/* right block address pointer */
-	int			rrecs = 0;	/* number of records in right block */
-	int			numrecs;
-	xfs_inobt_rec_t		*rrp;	/* right block record pointer */
-	xfs_btree_cur_t		*tcur;	/* temporary btree cursor */
-
-	mp = cur->bc_mp;
-
-	/*
-	 * Get the index of the entry being deleted, check for nothing there.
-	 */
-	ptr = cur->bc_ptrs[level];
-	if (ptr == 0) {
-		*stat = 0;
-		return 0;
-	}
-
-	/*
-	 * Get the buffer & block containing the record or key/ptr.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-#endif
-	/*
-	 * Fail if we're off the end of the block.
-	 */
+	return cur->bc_mp->m_inobt_mnr[level != 0];
+}
 
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (ptr > numrecs) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * It's a nonleaf.  Excise the key and ptr being deleted, by
-	 * sliding the entries past them down one.
-	 * Log the changed areas of the block.
-	 */
-	if (level > 0) {
-		kp = XFS_INOBT_KEY_ADDR(block, 1, cur);
-		pp = XFS_INOBT_PTR_ADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = ptr; i < numrecs; i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(pp[i]), level)))
-				return error;
-		}
-#endif
-		if (ptr < numrecs) {
-			memmove(&kp[ptr - 1], &kp[ptr],
-				(numrecs - ptr) * sizeof(*kp));
-			memmove(&pp[ptr - 1], &pp[ptr],
-				(numrecs - ptr) * sizeof(*kp));
-			xfs_inobt_log_keys(cur, bp, ptr, numrecs - 1);
-			xfs_inobt_log_ptrs(cur, bp, ptr, numrecs - 1);
-		}
-	}
-	/*
-	 * It's a leaf.  Excise the record being deleted, by sliding the
-	 * entries past it down one.  Log the changed areas of the block.
-	 */
-	else {
-		rp = XFS_INOBT_REC_ADDR(block, 1, cur);
-		if (ptr < numrecs) {
-			memmove(&rp[ptr - 1], &rp[ptr],
-				(numrecs - ptr) * sizeof(*rp));
-			xfs_inobt_log_recs(cur, bp, ptr, numrecs - 1);
-		}
-		/*
-		 * If it's the first record in the block, we'll need a key
-		 * structure to pass up to the next level (updkey).
-		 */
-		if (ptr == 1) {
-			key.ir_startino = rp->ir_startino;
-			kp = &key;
-		}
-	}
-	/*
-	 * Decrement and log the number of entries in the block.
-	 */
-	numrecs--;
-	block->bb_numrecs = cpu_to_be16(numrecs);
-	xfs_inobt_log_block(cur->bc_tp, bp, XFS_BB_NUMRECS);
-	/*
-	 * Is this the root level?  If so, we're almost done.
-	 */
-	if (level == cur->bc_nlevels - 1) {
-		/*
-		 * If this is the root level,
-		 * and there's only one entry left,
-		 * and it's NOT the leaf level,
-		 * then we can get rid of this level.
-		 */
-		if (numrecs == 1 && level > 0) {
-			agbp = cur->bc_private.a.agbp;
-			agi = XFS_BUF_TO_AGI(agbp);
-			/*
-			 * pp is still set to the first pointer in the block.
-			 * Make it the new root of the btree.
-			 */
-			bno = be32_to_cpu(agi->agi_root);
-			agi->agi_root = *pp;
-			be32_add_cpu(&agi->agi_level, -1);
-			/*
-			 * Free the block.
-			 */
-			if ((error = xfs_free_extent(cur->bc_tp,
-				XFS_AGB_TO_FSB(mp, cur->bc_private.a.agno, bno), 1)))
-				return error;
-			xfs_trans_binval(cur->bc_tp, bp);
-			xfs_ialloc_log_agi(cur->bc_tp, agbp,
-				XFS_AGI_ROOT | XFS_AGI_LEVEL);
-			/*
-			 * Update the cursor so there's one fewer level.
-			 */
-			cur->bc_bufs[level] = NULL;
-			cur->bc_nlevels--;
-		} else if (level > 0 &&
-			   (error = xfs_inobt_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * If we deleted the leftmost entry in the block, update the
-	 * key values above us in the tree.
-	 */
-	if (ptr == 1 && (error = xfs_inobt_updkey(cur, kp, level + 1)))
-		return error;
-	/*
-	 * If the number of records remaining in the block is at least
-	 * the minimum, we're done.
-	 */
-	if (numrecs >= XFS_INOBT_BLOCK_MINRECS(level, cur)) {
-		if (level > 0 &&
-		    (error = xfs_inobt_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * Otherwise, we have to move some records around to keep the
-	 * tree balanced.  Look at the left and right sibling blocks to
-	 * see if we can re-balance by moving only one record.
-	 */
-	rbno = be32_to_cpu(block->bb_rightsib);
-	lbno = be32_to_cpu(block->bb_leftsib);
-	bno = NULLAGBLOCK;
-	ASSERT(rbno != NULLAGBLOCK || lbno != NULLAGBLOCK);
-	/*
-	 * Duplicate the cursor so our btree manipulations here won't
-	 * disrupt the next level up.
-	 */
-	if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-		return error;
-	/*
-	 * If there's a right sibling, see if it's ok to shift an entry
-	 * out of it.
-	 */
-	if (rbno != NULLAGBLOCK) {
-		/*
-		 * Move the temp cursor to the last entry in the next block.
-		 * Actually any entry but the first would suffice.
-		 */
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		if ((error = xfs_inobt_increment(tcur, level, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		i = xfs_btree_lastrec(tcur, level);
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		/*
-		 * Grab a pointer to the block.
-		 */
-		rbp = tcur->bc_bufs[level];
-		right = XFS_BUF_TO_INOBT_BLOCK(rbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-			goto error0;
-#endif
-		/*
-		 * Grab the current block number, for future use.
-		 */
-		bno = be32_to_cpu(right->bb_leftsib);
-		/*
-		 * If right block is full enough so that removing one entry
-		 * won't make it too empty, and left-shifting an entry out
-		 * of right to us works, we're done.
-		 */
-		if (be16_to_cpu(right->bb_numrecs) - 1 >=
-		     XFS_INOBT_BLOCK_MINRECS(level, cur)) {
-			if ((error = xfs_inobt_lshift(tcur, level, &i)))
-				goto error0;
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_INOBT_BLOCK_MINRECS(level, cur));
-				xfs_btree_del_cursor(tcur,
-						     XFS_BTREE_NOERROR);
-				if (level > 0 &&
-				    (error = xfs_inobt_decrement(cur, level,
-						&i)))
-					return error;
-				*stat = 1;
-				return 0;
-			}
-		}
-		/*
-		 * Otherwise, grab the number of records in right for
-		 * future reference, and fix up the temp cursor to point
-		 * to our block again (last record).
-		 */
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if (lbno != NULLAGBLOCK) {
-			xfs_btree_firstrec(tcur, level);
-			if ((error = xfs_inobt_decrement(tcur, level, &i)))
-				goto error0;
-		}
-	}
-	/*
-	 * If there's a left sibling, see if it's ok to shift an entry
-	 * out of it.
-	 */
-	if (lbno != NULLAGBLOCK) {
-		/*
-		 * Move the temp cursor to the first entry in the
-		 * previous block.
-		 */
-		xfs_btree_firstrec(tcur, level);
-		if ((error = xfs_inobt_decrement(tcur, level, &i)))
-			goto error0;
-		xfs_btree_firstrec(tcur, level);
-		/*
-		 * Grab a pointer to the block.
-		 */
-		lbp = tcur->bc_bufs[level];
-		left = XFS_BUF_TO_INOBT_BLOCK(lbp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-			goto error0;
-#endif
-		/*
-		 * Grab the current block number, for future use.
-		 */
-		bno = be32_to_cpu(left->bb_rightsib);
-		/*
-		 * If left block is full enough so that removing one entry
-		 * won't make it too empty, and right-shifting an entry out
-		 * of left to us works, we're done.
-		 */
-		if (be16_to_cpu(left->bb_numrecs) - 1 >=
-		     XFS_INOBT_BLOCK_MINRECS(level, cur)) {
-			if ((error = xfs_inobt_rshift(tcur, level, &i)))
-				goto error0;
-			if (i) {
-				ASSERT(be16_to_cpu(block->bb_numrecs) >=
-				       XFS_INOBT_BLOCK_MINRECS(level, cur));
-				xfs_btree_del_cursor(tcur,
-						     XFS_BTREE_NOERROR);
-				if (level == 0)
-					cur->bc_ptrs[0]++;
-				*stat = 1;
-				return 0;
-			}
-		}
-		/*
-		 * Otherwise, grab the number of records in right for
-		 * future reference.
-		 */
-		lrecs = be16_to_cpu(left->bb_numrecs);
-	}
-	/*
-	 * Delete the temp cursor, we're done with it.
-	 */
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	/*
-	 * If here, we need to do a join to keep the tree balanced.
-	 */
-	ASSERT(bno != NULLAGBLOCK);
-	/*
-	 * See if we can join with the left neighbor block.
-	 */
-	if (lbno != NULLAGBLOCK &&
-	    lrecs + numrecs <= XFS_INOBT_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * Set "right" to be the starting block,
-		 * "left" to be the left neighbor.
-		 */
-		rbno = bno;
-		right = block;
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		rbp = bp;
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, lbno, 0, &lbp,
-				XFS_INO_BTREE_REF)))
-			return error;
-		left = XFS_BUF_TO_INOBT_BLOCK(lbp);
-		lrecs = be16_to_cpu(left->bb_numrecs);
-		if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-			return error;
-	}
-	/*
-	 * If that won't work, see if we can join with the right neighbor block.
-	 */
-	else if (rbno != NULLAGBLOCK &&
-		 rrecs + numrecs <= XFS_INOBT_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * Set "left" to be the starting block,
-		 * "right" to be the right neighbor.
-		 */
-		lbno = bno;
-		left = block;
-		lrecs = be16_to_cpu(left->bb_numrecs);
-		lbp = bp;
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, rbno, 0, &rbp,
-				XFS_INO_BTREE_REF)))
-			return error;
-		right = XFS_BUF_TO_INOBT_BLOCK(rbp);
-		rrecs = be16_to_cpu(right->bb_numrecs);
-		if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-			return error;
-	}
-	/*
-	 * Otherwise, we can't fix the imbalance.
-	 * Just return.  This is probably a logic error, but it's not fatal.
-	 */
-	else {
-		if (level > 0 && (error = xfs_inobt_decrement(cur, level, &i)))
-			return error;
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * We're now going to join "left" and "right" by moving all the stuff
-	 * in "right" to "left" and deleting "right".
-	 */
-	if (level > 0) {
-		/*
-		 * It's a non-leaf.  Move keys and pointers.
-		 */
-		lkp = XFS_INOBT_KEY_ADDR(left, lrecs + 1, cur);
-		lpp = XFS_INOBT_PTR_ADDR(left, lrecs + 1, cur);
-		rkp = XFS_INOBT_KEY_ADDR(right, 1, cur);
-		rpp = XFS_INOBT_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < rrecs; i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i]), level)))
-				return error;
-		}
-#endif
-		memcpy(lkp, rkp, rrecs * sizeof(*lkp));
-		memcpy(lpp, rpp, rrecs * sizeof(*lpp));
-		xfs_inobt_log_keys(cur, lbp, lrecs + 1, lrecs + rrecs);
-		xfs_inobt_log_ptrs(cur, lbp, lrecs + 1, lrecs + rrecs);
-	} else {
-		/*
-		 * It's a leaf.  Move records.
-		 */
-		lrp = XFS_INOBT_REC_ADDR(left, lrecs + 1, cur);
-		rrp = XFS_INOBT_REC_ADDR(right, 1, cur);
-		memcpy(lrp, rrp, rrecs * sizeof(*lrp));
-		xfs_inobt_log_recs(cur, lbp, lrecs + 1, lrecs + rrecs);
-	}
-	/*
-	 * If we joined with the left neighbor, set the buffer in the
-	 * cursor to the left block, and fix up the index.
-	 */
-	if (bp != lbp) {
-		xfs_btree_setbuf(cur, level, lbp);
-		cur->bc_ptrs[level] += lrecs;
-	}
-	/*
-	 * If we joined with the right neighbor and there's a level above
-	 * us, increment the cursor at that level.
-	 */
-	else if (level + 1 < cur->bc_nlevels &&
-		 (error = xfs_alloc_increment(cur, level + 1, &i)))
-		return error;
-	/*
-	 * Fix up the number of records in the surviving block.
-	 */
-	lrecs += rrecs;
-	left->bb_numrecs = cpu_to_be16(lrecs);
-	/*
-	 * Fix up the right block pointer in the surviving block, and log it.
-	 */
-	left->bb_rightsib = right->bb_rightsib;
-	xfs_inobt_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	/*
-	 * If there is a right sibling now, make it point to the
-	 * remaining block.
-	 */
-	if (be32_to_cpu(left->bb_rightsib) != NULLAGBLOCK) {
-		xfs_inobt_block_t	*rrblock;
-		xfs_buf_t		*rrbp;
+STATIC struct xfs_btree_cur *
+xfs_inobt_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	return xfs_inobt_init_cursor(cur->bc_mp, cur->bc_tp,
+			cur->bc_private.a.agbp, cur->bc_private.a.agno);
+}
 
-		if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-				cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib), 0,
-				&rrbp, XFS_INO_BTREE_REF)))
-			return error;
-		rrblock = XFS_BUF_TO_INOBT_BLOCK(rrbp);
-		if ((error = xfs_btree_check_sblock(cur, rrblock, level, rrbp)))
-			return error;
-		rrblock->bb_leftsib = cpu_to_be32(lbno);
-		xfs_inobt_log_block(cur->bc_tp, rrbp, XFS_BB_LEFTSIB);
-	}
-	/*
-	 * Free the deleting block.
-	 */
-	if ((error = xfs_free_extent(cur->bc_tp, XFS_AGB_TO_FSB(mp,
-				     cur->bc_private.a.agno, rbno), 1)))
-		return error;
-	xfs_trans_binval(cur->bc_tp, rbp);
-	/*
-	 * Readjust the ptr at this level if it's not a leaf, since it's
-	 * still pointing at the deletion point, which makes the cursor
-	 * inconsistent.  If this makes the ptr 0, the caller fixes it up.
-	 * We can't use decrement because it would change the next level up.
-	 */
-	if (level > 0)
-		cur->bc_ptrs[level]--;
-	/*
-	 * Return value means the next level up has something to do.
-	 */
-	*stat = 2;
-	return 0;
+STATIC void
+xfs_inobt_set_root(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*nptr,
+	int			inc)	/* level change */
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
 
-error0:
-	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-	return error;
+	agi->agi_root = nptr->s;
+	be32_add_cpu(&agi->agi_level, inc);
+	xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_ROOT | XFS_AGI_LEVEL);
 }
 
-/*
- * Insert one record/level.  Return information to the caller
- * allowing the next level up to proceed if necessary.
- */
-STATIC int				/* error */
-xfs_inobt_insrec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to insert record at */
-	xfs_agblock_t		*bnop,	/* i/o: block number inserted */
-	xfs_inobt_rec_t		*recp,	/* i/o: record data inserted */
-	xfs_btree_cur_t		**curp,	/* output: new cursor replacing cur */
-	int			*stat)	/* success/failure */
+STATIC int
+xfs_inobt_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start,
+	union xfs_btree_ptr	*new,
+	int			length,
+	int			*stat)
 {
-	xfs_inobt_block_t	*block;	/* btree block record/key lives in */
-	xfs_buf_t		*bp;	/* buffer for block */
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_inobt_key_t		key;	/* key value being inserted */
-	xfs_inobt_key_t		*kp=NULL;	/* pointer to btree keys */
-	xfs_agblock_t		nbno;	/* block number of allocated block */
-	xfs_btree_cur_t		*ncur;	/* new cursor to be used at next lvl */
-	xfs_inobt_key_t		nkey;	/* new key value, from split */
-	xfs_inobt_rec_t		nrec;	/* new record value, for caller */
-	int			numrecs;
-	int			optr;	/* old ptr value */
-	xfs_inobt_ptr_t		*pp;	/* pointer to btree addresses */
-	int			ptr;	/* index in btree block for this rec */
-	xfs_inobt_rec_t		*rp=NULL;	/* pointer to btree records */
+	xfs_alloc_arg_t		args;		/* block allocation args */
+	int			error;		/* error return value */
+	xfs_agblock_t		sbno = be32_to_cpu(start->s);
 
-	/*
-	 * GCC doesn't understand the (arguably complex) control flow in
-	 * this function and complains about uninitialized structure fields
-	 * without this.
-	 */
-	memset(&nrec, 0, sizeof(nrec));
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
 
-	/*
-	 * If we made it to the root level, allocate a new root block
-	 * and we're done.
-	 */
-	if (level >= cur->bc_nlevels) {
-		error = xfs_inobt_newroot(cur, &i);
-		*bnop = NULLAGBLOCK;
-		*stat = i;
+	memset(&args, 0, sizeof(args));
+	args.tp = cur->bc_tp;
+	args.mp = cur->bc_mp;
+	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno, sbno);
+	args.minlen = 1;
+	args.maxlen = 1;
+	args.prod = 1;
+	args.type = XFS_ALLOCTYPE_NEAR_BNO;
+
+	error = xfs_alloc_vextent(&args);
+	if (error) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
 		return error;
 	}
-	/*
-	 * Make a key out of the record data to be inserted, and save it.
-	 */
-	key.ir_startino = recp->ir_startino;
-	optr = ptr = cur->bc_ptrs[level];
-	/*
-	 * If we're off the left edge, return failure.
-	 */
-	if (ptr == 0) {
+	if (args.fsbno == NULLFSBLOCK) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 		*stat = 0;
 		return 0;
 	}
-	/*
-	 * Get pointers to the btree buffer and block.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-	numrecs = be16_to_cpu(block->bb_numrecs);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-	/*
-	 * Check that the new entry is being inserted in the right place.
-	 */
-	if (ptr <= numrecs) {
-		if (level == 0) {
-			rp = XFS_INOBT_REC_ADDR(block, ptr, cur);
-			xfs_btree_check_rec(cur->bc_btnum, recp, rp);
-		} else {
-			kp = XFS_INOBT_KEY_ADDR(block, ptr, cur);
-			xfs_btree_check_key(cur->bc_btnum, &key, kp);
-		}
-	}
-#endif
-	nbno = NULLAGBLOCK;
-	ncur = NULL;
-	/*
-	 * If the block is full, we can't insert the new entry until we
-	 * make the block un-full.
-	 */
-	if (numrecs == XFS_INOBT_BLOCK_MAXRECS(level, cur)) {
-		/*
-		 * First, try shifting an entry to the right neighbor.
-		 */
-		if ((error = xfs_inobt_rshift(cur, level, &i)))
-			return error;
-		if (i) {
-			/* nothing */
-		}
-		/*
-		 * Next, try shifting an entry to the left neighbor.
-		 */
-		else {
-			if ((error = xfs_inobt_lshift(cur, level, &i)))
-				return error;
-			if (i) {
-				optr = ptr = cur->bc_ptrs[level];
-			} else {
-				/*
-				 * Next, try splitting the current block
-				 * in half. If this works we have to
-				 * re-set our variables because
-				 * we could be in a different block now.
-				 */
-				if ((error = xfs_inobt_split(cur, level, &nbno,
-						&nkey, &ncur, &i)))
-					return error;
-				if (i) {
-					bp = cur->bc_bufs[level];
-					block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-					if ((error = xfs_btree_check_sblock(cur,
-							block, level, bp)))
-						return error;
-#endif
-					ptr = cur->bc_ptrs[level];
-					nrec.ir_startino = nkey.ir_startino;
-				} else {
-					/*
-					 * Otherwise the insert fails.
-					 */
-					*stat = 0;
-					return 0;
-				}
-			}
-		}
-	}
-	/*
-	 * At this point we know there's room for our new entry in the block
-	 * we're pointing at.
-	 */
-	numrecs = be16_to_cpu(block->bb_numrecs);
-	if (level > 0) {
-		/*
-		 * It's a non-leaf entry.  Make a hole for the new data
-		 * in the key and ptr regions of the block.
-		 */
-		kp = XFS_INOBT_KEY_ADDR(block, 1, cur);
-		pp = XFS_INOBT_PTR_ADDR(block, 1, cur);
-#ifdef DEBUG
-		for (i = numrecs; i >= ptr; i--) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(pp[i - 1]), level)))
-				return error;
-		}
-#endif
-		memmove(&kp[ptr], &kp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*kp));
-		memmove(&pp[ptr], &pp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*pp));
-		/*
-		 * Now stuff the new data in, bump numrecs and log the new data.
-		 */
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, *bnop, level)))
-			return error;
-#endif
-		kp[ptr - 1] = key;
-		pp[ptr - 1] = cpu_to_be32(*bnop);
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_inobt_log_keys(cur, bp, ptr, numrecs);
-		xfs_inobt_log_ptrs(cur, bp, ptr, numrecs);
-	} else {
-		/*
-		 * It's a leaf entry.  Make a hole for the new record.
-		 */
-		rp = XFS_INOBT_REC_ADDR(block, 1, cur);
-		memmove(&rp[ptr], &rp[ptr - 1],
-			(numrecs - ptr + 1) * sizeof(*rp));
-		/*
-		 * Now stuff the new record in, bump numrecs
-		 * and log the new data.
-		 */
-		rp[ptr - 1] = *recp;
-		numrecs++;
-		block->bb_numrecs = cpu_to_be16(numrecs);
-		xfs_inobt_log_recs(cur, bp, ptr, numrecs);
-	}
-	/*
-	 * Log the new number of records in the btree header.
-	 */
-	xfs_inobt_log_block(cur->bc_tp, bp, XFS_BB_NUMRECS);
-#ifdef DEBUG
-	/*
-	 * Check that the key/record is in the right place, now.
-	 */
-	if (ptr < numrecs) {
-		if (level == 0)
-			xfs_btree_check_rec(cur->bc_btnum, rp + ptr - 1,
-				rp + ptr);
-		else
-			xfs_btree_check_key(cur->bc_btnum, kp + ptr - 1,
-				kp + ptr);
-	}
-#endif
-	/*
-	 * If we inserted at the start of a block, update the parents' keys.
-	 */
-	if (optr == 1 && (error = xfs_inobt_updkey(cur, &key, level + 1)))
-		return error;
-	/*
-	 * Return the new block number, if any.
-	 * If there is one, give back a record value and a cursor too.
-	 */
-	*bnop = nbno;
-	if (nbno != NULLAGBLOCK) {
-		*recp = nrec;
-		*curp = ncur;
-	}
+	ASSERT(args.len == 1);
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
+
+	new->s = cpu_to_be32(XFS_FSB_TO_AGBNO(args.mp, args.fsbno));
 	*stat = 1;
 	return 0;
 }
 
-/*
- * Log header fields from a btree block.
- */
-STATIC void
-xfs_inobt_log_block(
-	xfs_trans_t		*tp,	/* transaction pointer */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			fields)	/* mask of fields: XFS_BB_... */
+STATIC int
+xfs_inobt_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
 {
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	static const short	offsets[] = {	/* table of offsets */
-		offsetof(xfs_inobt_block_t, bb_magic),
-		offsetof(xfs_inobt_block_t, bb_level),
-		offsetof(xfs_inobt_block_t, bb_numrecs),
-		offsetof(xfs_inobt_block_t, bb_leftsib),
-		offsetof(xfs_inobt_block_t, bb_rightsib),
-		sizeof(xfs_inobt_block_t)
-	};
+	xfs_fsblock_t		fsbno;
+	int			error;
 
-	xfs_btree_offsets(fields, offsets, XFS_BB_NUM_BITS, &first, &last);
-	xfs_trans_log_buf(tp, bp, first, last);
+	fsbno = XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(bp));
+	error = xfs_free_extent(cur->bc_tp, fsbno, 1);
+	if (error)
+		return error;
+
+	xfs_trans_binval(cur->bc_tp, bp);
+	return error;
 }
 
-/*
- * Log keys from a btree block (nonleaf).
- */
-STATIC void
-xfs_inobt_log_keys(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			kfirst,	/* index of first key to log */
-	int			klast)	/* index of last key to log */
+STATIC int
+xfs_inobt_get_maxrecs(
+	struct xfs_btree_cur	*cur,
+	int			level)
 {
-	xfs_inobt_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	xfs_inobt_key_t		*kp;	/* key pointer in btree block */
-	int			last;	/* last byte offset logged */
-
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-	kp = XFS_INOBT_KEY_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&kp[kfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&kp[klast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	return cur->bc_mp->m_inobt_mxr[level != 0];
 }
 
-/*
- * Log block pointer fields from a btree block (nonleaf).
- */
 STATIC void
-xfs_inobt_log_ptrs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			pfirst,	/* index of first pointer to log */
-	int			plast)	/* index of last pointer to log */
+xfs_inobt_init_key_from_rec(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
 {
-	xfs_inobt_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	xfs_inobt_ptr_t		*pp;	/* block-pointer pointer in btree blk */
-
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-	pp = XFS_INOBT_PTR_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&pp[pfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&pp[plast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+	key->inobt.ir_startino = rec->inobt.ir_startino;
 }
 
-/*
- * Log records from a btree block (leaf).
- */
 STATIC void
-xfs_inobt_log_recs(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_buf_t		*bp,	/* buffer containing btree block */
-	int			rfirst,	/* index of first record to log */
-	int			rlast)	/* index of last record to log */
+xfs_inobt_init_rec_from_key(
+	union xfs_btree_key	*key,
+	union xfs_btree_rec	*rec)
 {
-	xfs_inobt_block_t	*block;	/* btree block to log from */
-	int			first;	/* first byte offset logged */
-	int			last;	/* last byte offset logged */
-	xfs_inobt_rec_t		*rp;	/* record pointer for btree block */
+	rec->inobt.ir_startino = key->inobt.ir_startino;
+}
 
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-	rp = XFS_INOBT_REC_ADDR(block, 1, cur);
-	first = (int)((xfs_caddr_t)&rp[rfirst - 1] - (xfs_caddr_t)block);
-	last = (int)(((xfs_caddr_t)&rp[rlast] - 1) - (xfs_caddr_t)block);
-	xfs_trans_log_buf(cur->bc_tp, bp, first, last);
+STATIC void
+xfs_inobt_init_rec_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec)
+{
+	rec->inobt.ir_startino = cpu_to_be32(cur->bc_rec.i.ir_startino);
+	rec->inobt.ir_freecount = cpu_to_be32(cur->bc_rec.i.ir_freecount);
+	rec->inobt.ir_free = cpu_to_be64(cur->bc_rec.i.ir_free);
 }
 
 /*
- * Lookup the record.  The cursor is made to point to it, based on dir.
- * Return 0 if can't find any such record, 1 for success.
+ * intial value of ptr for lookup
  */
-STATIC int				/* error */
-xfs_inobt_lookup(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_lookup_t		dir,	/* <=, ==, or >= */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_inobt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
 {
-	xfs_agblock_t		agbno;	/* a.g. relative btree block number */
-	xfs_agnumber_t		agno;	/* allocation group number */
-	xfs_inobt_block_t	*block=NULL;	/* current btree block */
-	__int64_t		diff;	/* difference for the current key */
-	int			error;	/* error return value */
-	int			keyno=0;	/* current key number */
-	int			level;	/* level in the btree */
-	xfs_mount_t		*mp;	/* file system mount point */
-
-	/*
-	 * Get the allocation group header, and the root block number.
-	 */
-	mp = cur->bc_mp;
-	{
-		xfs_agi_t	*agi;	/* a.g. inode header */
-
-		agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp);
-		agno = be32_to_cpu(agi->agi_seqno);
-		agbno = be32_to_cpu(agi->agi_root);
-	}
-	/*
-	 * Iterate over each level in the btree, starting at the root.
-	 * For each level above the leaves, find the key we need, based
-	 * on the lookup record, then follow the corresponding block
-	 * pointer down to the next level.
-	 */
-	for (level = cur->bc_nlevels - 1, diff = 1; level >= 0; level--) {
-		xfs_buf_t	*bp;	/* buffer pointer for btree block */
-		xfs_daddr_t	d;	/* disk address of btree block */
-
-		/*
-		 * Get the disk address we're looking for.
-		 */
-		d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-		/*
-		 * If the old buffer at this level is for a different block,
-		 * throw it away, otherwise just use it.
-		 */
-		bp = cur->bc_bufs[level];
-		if (bp && XFS_BUF_ADDR(bp) != d)
-			bp = NULL;
-		if (!bp) {
-			/*
-			 * Need to get a new buffer.  Read it, then
-			 * set it in the cursor, releasing the old one.
-			 */
-			if ((error = xfs_btree_read_bufs(mp, cur->bc_tp,
-					agno, agbno, 0, &bp, XFS_INO_BTREE_REF)))
-				return error;
-			xfs_btree_setbuf(cur, level, bp);
-			/*
-			 * Point to the btree block, now that we have the buffer
-			 */
-			block = XFS_BUF_TO_INOBT_BLOCK(bp);
-			if ((error = xfs_btree_check_sblock(cur, block, level,
-					bp)))
-				return error;
-		} else
-			block = XFS_BUF_TO_INOBT_BLOCK(bp);
-		/*
-		 * If we already had a key match at a higher level, we know
-		 * we need to use the first entry in this block.
-		 */
-		if (diff == 0)
-			keyno = 1;
-		/*
-		 * Otherwise we need to search this block.  Do a binary search.
-		 */
-		else {
-			int		high;	/* high entry number */
-			xfs_inobt_key_t	*kkbase=NULL;/* base of keys in block */
-			xfs_inobt_rec_t	*krbase=NULL;/* base of records in block */
-			int		low;	/* low entry number */
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp);
 
-			/*
-			 * Get a pointer to keys or records.
-			 */
-			if (level > 0)
-				kkbase = XFS_INOBT_KEY_ADDR(block, 1, cur);
-			else
-				krbase = XFS_INOBT_REC_ADDR(block, 1, cur);
-			/*
-			 * Set low and high entry numbers, 1-based.
-			 */
-			low = 1;
-			if (!(high = be16_to_cpu(block->bb_numrecs))) {
-				/*
-				 * If the block is empty, the tree must
-				 * be an empty leaf.
-				 */
-				ASSERT(level == 0 && cur->bc_nlevels == 1);
-				cur->bc_ptrs[0] = dir != XFS_LOOKUP_LE;
-				*stat = 0;
-				return 0;
-			}
-			/*
-			 * Binary search the block.
-			 */
-			while (low <= high) {
-				xfs_agino_t	startino;	/* key value */
-
-				/*
-				 * keyno is average of low and high.
-				 */
-				keyno = (low + high) >> 1;
-				/*
-				 * Get startino.
-				 */
-				if (level > 0) {
-					xfs_inobt_key_t	*kkp;
-
-					kkp = kkbase + keyno - 1;
-					startino = be32_to_cpu(kkp->ir_startino);
-				} else {
-					xfs_inobt_rec_t	*krp;
-
-					krp = krbase + keyno - 1;
-					startino = be32_to_cpu(krp->ir_startino);
-				}
-				/*
-				 * Compute difference to get next direction.
-				 */
-				diff = (__int64_t)
-					startino - cur->bc_rec.i.ir_startino;
-				/*
-				 * Less than, move right.
-				 */
-				if (diff < 0)
-					low = keyno + 1;
-				/*
-				 * Greater than, move left.
-				 */
-				else if (diff > 0)
-					high = keyno - 1;
-				/*
-				 * Equal, we're done.
-				 */
-				else
-					break;
-			}
-		}
-		/*
-		 * If there are more levels, set up for the next level
-		 * by getting the block number and filling in the cursor.
-		 */
-		if (level > 0) {
-			/*
-			 * If we moved left, need the previous key number,
-			 * unless there isn't one.
-			 */
-			if (diff > 0 && --keyno < 1)
-				keyno = 1;
-			agbno = be32_to_cpu(*XFS_INOBT_PTR_ADDR(block, keyno, cur));
-#ifdef DEBUG
-			if ((error = xfs_btree_check_sptr(cur, agbno, level)))
-				return error;
-#endif
-			cur->bc_ptrs[level] = keyno;
-		}
-	}
-	/*
-	 * Done with the search.
-	 * See if we need to adjust the results.
-	 */
-	if (dir != XFS_LOOKUP_LE && diff < 0) {
-		keyno++;
-		/*
-		 * If ge search and we went off the end of the block, but it's
-		 * not the last block, we're in the wrong block.
-		 */
-		if (dir == XFS_LOOKUP_GE &&
-		    keyno > be16_to_cpu(block->bb_numrecs) &&
-		    be32_to_cpu(block->bb_rightsib) != NULLAGBLOCK) {
-			int	i;
+	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agi->agi_seqno));
 
-			cur->bc_ptrs[0] = keyno;
-			if ((error = xfs_inobt_increment(cur, 0, &i)))
-				return error;
-			ASSERT(i == 1);
-			*stat = 1;
-			return 0;
-		}
-	}
-	else if (dir == XFS_LOOKUP_LE && diff > 0)
-		keyno--;
-	cur->bc_ptrs[0] = keyno;
-	/*
-	 * Return if we succeeded or not.
-	 */
-	if (keyno == 0 || keyno > be16_to_cpu(block->bb_numrecs))
-		*stat = 0;
-	else
-		*stat = ((dir != XFS_LOOKUP_EQ) || (diff == 0));
-	return 0;
+	ptr->s = agi->agi_root;
 }
 
-/*
- * Move 1 record left from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int				/* error */
-xfs_inobt_lshift(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to shift record on */
-	int			*stat)	/* success/failure */
+STATIC __int64_t
+xfs_inobt_key_diff(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key)
 {
-	int			error;	/* error return value */
-#ifdef DEBUG
-	int			i;	/* loop index */
-#endif
-	xfs_inobt_key_t		key;	/* key value for leaf level upward */
-	xfs_buf_t		*lbp;	/* buffer for left neighbor block */
-	xfs_inobt_block_t	*left;	/* left neighbor btree block */
-	xfs_inobt_key_t		*lkp=NULL;	/* key pointer for left block */
-	xfs_inobt_ptr_t		*lpp;	/* address pointer for left block */
-	xfs_inobt_rec_t		*lrp=NULL;	/* record pointer for left block */
-	int			nrec;	/* new number of left block entries */
-	xfs_buf_t		*rbp;	/* buffer for right (current) block */
-	xfs_inobt_block_t	*right;	/* right (current) btree block */
-	xfs_inobt_key_t		*rkp=NULL;	/* key pointer for right block */
-	xfs_inobt_ptr_t		*rpp=NULL;	/* address pointer for right block */
-	xfs_inobt_rec_t		*rrp=NULL;	/* record pointer for right block */
-
-	/*
-	 * Set up variables for this block as "right".
-	 */
-	rbp = cur->bc_bufs[level];
-	right = XFS_BUF_TO_INOBT_BLOCK(rbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-		return error;
-#endif
-	/*
-	 * If we've got no left sibling then we can't shift an entry left.
-	 */
-	if (be32_to_cpu(right->bb_leftsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * If the cursor entry is the one that would be moved, don't
-	 * do it... it's too complicated.
-	 */
-	if (cur->bc_ptrs[level] <= 1) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Set up the left neighbor as "left".
-	 */
-	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agno, be32_to_cpu(right->bb_leftsib),
-			0, &lbp, XFS_INO_BTREE_REF)))
-		return error;
-	left = XFS_BUF_TO_INOBT_BLOCK(lbp);
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-	/*
-	 * If it's full, it can't take another entry.
-	 */
-	if (be16_to_cpu(left->bb_numrecs) == XFS_INOBT_BLOCK_MAXRECS(level, cur)) {
-		*stat = 0;
-		return 0;
-	}
-	nrec = be16_to_cpu(left->bb_numrecs) + 1;
-	/*
-	 * If non-leaf, copy a key and a ptr to the left block.
-	 */
-	if (level > 0) {
-		lkp = XFS_INOBT_KEY_ADDR(left, nrec, cur);
-		rkp = XFS_INOBT_KEY_ADDR(right, 1, cur);
-		*lkp = *rkp;
-		xfs_inobt_log_keys(cur, lbp, nrec, nrec);
-		lpp = XFS_INOBT_PTR_ADDR(left, nrec, cur);
-		rpp = XFS_INOBT_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(*rpp), level)))
-			return error;
-#endif
-		*lpp = *rpp;
-		xfs_inobt_log_ptrs(cur, lbp, nrec, nrec);
-	}
-	/*
-	 * If leaf, copy a record to the left block.
-	 */
-	else {
-		lrp = XFS_INOBT_REC_ADDR(left, nrec, cur);
-		rrp = XFS_INOBT_REC_ADDR(right, 1, cur);
-		*lrp = *rrp;
-		xfs_inobt_log_recs(cur, lbp, nrec, nrec);
-	}
-	/*
-	 * Bump and log left's numrecs, decrement and log right's numrecs.
-	 */
-	be16_add_cpu(&left->bb_numrecs, 1);
-	xfs_inobt_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS);
-#ifdef DEBUG
-	if (level > 0)
-		xfs_btree_check_key(cur->bc_btnum, lkp - 1, lkp);
-	else
-		xfs_btree_check_rec(cur->bc_btnum, lrp - 1, lrp);
-#endif
-	be16_add_cpu(&right->bb_numrecs, -1);
-	xfs_inobt_log_block(cur->bc_tp, rbp, XFS_BB_NUMRECS);
-	/*
-	 * Slide the contents of right down one entry.
-	 */
-	if (level > 0) {
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i + 1]),
-					level)))
-				return error;
-		}
-#endif
-		memmove(rkp, rkp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp, rpp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_inobt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_inobt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-	} else {
-		memmove(rrp, rrp + 1, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_inobt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		key.ir_startino = rrp->ir_startino;
-		rkp = &key;
-	}
-	/*
-	 * Update the parent key values of right.
-	 */
-	if ((error = xfs_inobt_updkey(cur, rkp, level + 1)))
-		return error;
-	/*
-	 * Slide the cursor value left one.
-	 */
-	cur->bc_ptrs[level]--;
-	*stat = 1;
-	return 0;
+	return (__int64_t)be32_to_cpu(key->inobt.ir_startino) -
+			  cur->bc_rec.i.ir_startino;
 }
 
-/*
- * Allocate a new root block, fill it in.
- */
-STATIC int				/* error */
-xfs_inobt_newroot(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			*stat)	/* success/failure */
+STATIC int
+xfs_inobt_kill_root(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp,
+	int			level,
+	union xfs_btree_ptr	*newroot)
 {
-	xfs_agi_t		*agi;	/* a.g. inode header */
-	xfs_alloc_arg_t		args;	/* allocation argument structure */
-	xfs_inobt_block_t	*block;	/* one half of the old root block */
-	xfs_buf_t		*bp;	/* buffer containing block */
-	int			error;	/* error return value */
-	xfs_inobt_key_t		*kp;	/* btree key pointer */
-	xfs_agblock_t		lbno;	/* left block number */
-	xfs_buf_t		*lbp;	/* left buffer pointer */
-	xfs_inobt_block_t	*left;	/* left btree block */
-	xfs_buf_t		*nbp;	/* new (root) buffer */
-	xfs_inobt_block_t	*new;	/* new (root) btree block */
-	int			nptr;	/* new value for key index, 1 or 2 */
-	xfs_inobt_ptr_t		*pp;	/* btree address pointer */
-	xfs_agblock_t		rbno;	/* right block number */
-	xfs_buf_t		*rbp;	/* right buffer pointer */
-	xfs_inobt_block_t	*right;	/* right btree block */
-	xfs_inobt_rec_t		*rp;	/* btree record pointer */
+	int			error;
 
-	ASSERT(cur->bc_nlevels < XFS_IN_MAXLEVELS(cur->bc_mp));
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
+	XFS_BTREE_STATS_INC(cur, killroot);
 
 	/*
-	 * Get a block & a buffer.
+	 * Update the root pointer, decreasing the level by 1 and then
+	 * free the old root.
 	 */
-	agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp);
-	args.tp = cur->bc_tp;
-	args.mp = cur->bc_mp;
-	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno,
-		be32_to_cpu(agi->agi_root));
-	args.mod = args.minleft = args.alignment = args.total = args.wasdel =
-		args.isfl = args.userdata = args.minalignslop = 0;
-	args.minlen = args.maxlen = args.prod = 1;
-	args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	if ((error = xfs_alloc_vextent(&args)))
+	xfs_inobt_set_root(cur, newroot, -1);
+	error = xfs_inobt_free_block(cur, bp);
+	if (error) {
+		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
 		return error;
-	/*
-	 * None available, we fail.
-	 */
-	if (args.fsbno == NULLFSBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	ASSERT(args.len == 1);
-	nbp = xfs_btree_get_bufs(args.mp, args.tp, args.agno, args.agbno, 0);
-	new = XFS_BUF_TO_INOBT_BLOCK(nbp);
-	/*
-	 * Set the root data in the a.g. inode structure.
-	 */
-	agi->agi_root = cpu_to_be32(args.agbno);
-	be32_add_cpu(&agi->agi_level, 1);
-	xfs_ialloc_log_agi(args.tp, cur->bc_private.a.agbp,
-		XFS_AGI_ROOT | XFS_AGI_LEVEL);
-	/*
-	 * At the previous root level there are now two blocks: the old
-	 * root, and the new block generated when it was split.
-	 * We don't know which one the cursor is pointing at, so we
-	 * set up variables "left" and "right" for each case.
-	 */
-	bp = cur->bc_bufs[cur->bc_nlevels - 1];
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, cur->bc_nlevels - 1, bp)))
-		return error;
-#endif
-	if (be32_to_cpu(block->bb_rightsib) != NULLAGBLOCK) {
-		/*
-		 * Our block is left, pick up the right block.
-		 */
-		lbp = bp;
-		lbno = XFS_DADDR_TO_AGBNO(args.mp, XFS_BUF_ADDR(lbp));
-		left = block;
-		rbno = be32_to_cpu(left->bb_rightsib);
-		if ((error = xfs_btree_read_bufs(args.mp, args.tp, args.agno,
-				rbno, 0, &rbp, XFS_INO_BTREE_REF)))
-			return error;
-		bp = rbp;
-		right = XFS_BUF_TO_INOBT_BLOCK(rbp);
-		if ((error = xfs_btree_check_sblock(cur, right,
-				cur->bc_nlevels - 1, rbp)))
-			return error;
-		nptr = 1;
-	} else {
-		/*
-		 * Our block is right, pick up the left block.
-		 */
-		rbp = bp;
-		rbno = XFS_DADDR_TO_AGBNO(args.mp, XFS_BUF_ADDR(rbp));
-		right = block;
-		lbno = be32_to_cpu(right->bb_leftsib);
-		if ((error = xfs_btree_read_bufs(args.mp, args.tp, args.agno,
-				lbno, 0, &lbp, XFS_INO_BTREE_REF)))
-			return error;
-		bp = lbp;
-		left = XFS_BUF_TO_INOBT_BLOCK(lbp);
-		if ((error = xfs_btree_check_sblock(cur, left,
-				cur->bc_nlevels - 1, lbp)))
-			return error;
-		nptr = 2;
-	}
-	/*
-	 * Fill in the new block's btree header and log it.
-	 */
-	new->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
-	new->bb_level = cpu_to_be16(cur->bc_nlevels);
-	new->bb_numrecs = cpu_to_be16(2);
-	new->bb_leftsib = cpu_to_be32(NULLAGBLOCK);
-	new->bb_rightsib = cpu_to_be32(NULLAGBLOCK);
-	xfs_inobt_log_block(args.tp, nbp, XFS_BB_ALL_BITS);
-	ASSERT(lbno != NULLAGBLOCK && rbno != NULLAGBLOCK);
-	/*
-	 * Fill in the key data in the new root.
-	 */
-	kp = XFS_INOBT_KEY_ADDR(new, 1, cur);
-	if (be16_to_cpu(left->bb_level) > 0) {
-		kp[0] = *XFS_INOBT_KEY_ADDR(left, 1, cur);
-		kp[1] = *XFS_INOBT_KEY_ADDR(right, 1, cur);
-	} else {
-		rp = XFS_INOBT_REC_ADDR(left, 1, cur);
-		kp[0].ir_startino = rp->ir_startino;
-		rp = XFS_INOBT_REC_ADDR(right, 1, cur);
-		kp[1].ir_startino = rp->ir_startino;
 	}
-	xfs_inobt_log_keys(cur, nbp, 1, 2);
-	/*
-	 * Fill in the pointer data in the new root.
-	 */
-	pp = XFS_INOBT_PTR_ADDR(new, 1, cur);
-	pp[0] = cpu_to_be32(lbno);
-	pp[1] = cpu_to_be32(rbno);
-	xfs_inobt_log_ptrs(cur, nbp, 1, 2);
-	/*
-	 * Fix up the cursor.
-	 */
-	xfs_btree_setbuf(cur, cur->bc_nlevels, nbp);
-	cur->bc_ptrs[cur->bc_nlevels] = nptr;
-	cur->bc_nlevels++;
-	*stat = 1;
-	return 0;
-}
 
-/*
- * Move 1 record right from cur/level if possible.
- * Update cur to reflect the new path.
- */
-STATIC int				/* error */
-xfs_inobt_rshift(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to shift record on */
-	int			*stat)	/* success/failure */
-{
-	int			error;	/* error return value */
-	int			i;	/* loop index */
-	xfs_inobt_key_t		key;	/* key value for leaf level upward */
-	xfs_buf_t		*lbp;	/* buffer for left (current) block */
-	xfs_inobt_block_t	*left;	/* left (current) btree block */
-	xfs_inobt_key_t		*lkp;	/* key pointer for left block */
-	xfs_inobt_ptr_t		*lpp;	/* address pointer for left block */
-	xfs_inobt_rec_t		*lrp;	/* record pointer for left block */
-	xfs_buf_t		*rbp;	/* buffer for right neighbor block */
-	xfs_inobt_block_t	*right;	/* right neighbor btree block */
-	xfs_inobt_key_t		*rkp;	/* key pointer for right block */
-	xfs_inobt_ptr_t		*rpp;	/* address pointer for right block */
-	xfs_inobt_rec_t		*rrp=NULL;	/* record pointer for right block */
-	xfs_btree_cur_t		*tcur;	/* temporary cursor */
+	XFS_BTREE_STATS_INC(cur, free);
 
-	/*
-	 * Set up variables for this block as "left".
-	 */
-	lbp = cur->bc_bufs[level];
-	left = XFS_BUF_TO_INOBT_BLOCK(lbp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-#endif
-	/*
-	 * If we've got no right sibling then we can't shift an entry right.
-	 */
-	if (be32_to_cpu(left->bb_rightsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * If the cursor entry is the one that would be moved, don't
-	 * do it... it's too complicated.
-	 */
-	if (cur->bc_ptrs[level] >= be16_to_cpu(left->bb_numrecs)) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Set up the right neighbor as "right".
-	 */
-	if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agno, be32_to_cpu(left->bb_rightsib),
-			0, &rbp, XFS_INO_BTREE_REF)))
-		return error;
-	right = XFS_BUF_TO_INOBT_BLOCK(rbp);
-	if ((error = xfs_btree_check_sblock(cur, right, level, rbp)))
-		return error;
-	/*
-	 * If it's full, it can't take another entry.
-	 */
-	if (be16_to_cpu(right->bb_numrecs) == XFS_INOBT_BLOCK_MAXRECS(level, cur)) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Make a hole at the start of the right neighbor block, then
-	 * copy the last left block entry to the hole.
-	 */
-	if (level > 0) {
-		lkp = XFS_INOBT_KEY_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		lpp = XFS_INOBT_PTR_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rkp = XFS_INOBT_KEY_ADDR(right, 1, cur);
-		rpp = XFS_INOBT_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = be16_to_cpu(right->bb_numrecs) - 1; i >= 0; i--) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(rpp[i]), level)))
-				return error;
-		}
-#endif
-		memmove(rkp + 1, rkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memmove(rpp + 1, rpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(*lpp), level)))
-			return error;
-#endif
-		*rkp = *lkp;
-		*rpp = *lpp;
-		xfs_inobt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		xfs_inobt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-	} else {
-		lrp = XFS_INOBT_REC_ADDR(left, be16_to_cpu(left->bb_numrecs), cur);
-		rrp = XFS_INOBT_REC_ADDR(right, 1, cur);
-		memmove(rrp + 1, rrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		*rrp = *lrp;
-		xfs_inobt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs) + 1);
-		key.ir_startino = rrp->ir_startino;
-		rkp = &key;
-	}
-	/*
-	 * Decrement and log left's numrecs, bump and log right's numrecs.
-	 */
-	be16_add_cpu(&left->bb_numrecs, -1);
-	xfs_inobt_log_block(cur->bc_tp, lbp, XFS_BB_NUMRECS);
-	be16_add_cpu(&right->bb_numrecs, 1);
-#ifdef DEBUG
-	if (level > 0)
-		xfs_btree_check_key(cur->bc_btnum, rkp, rkp + 1);
-	else
-		xfs_btree_check_rec(cur->bc_btnum, rrp, rrp + 1);
-#endif
-	xfs_inobt_log_block(cur->bc_tp, rbp, XFS_BB_NUMRECS);
-	/*
-	 * Using a temporary cursor, update the parent key values of the
-	 * block on the right.
-	 */
-	if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-		return error;
-	xfs_btree_lastrec(tcur, level);
-	if ((error = xfs_inobt_increment(tcur, level, &i)) ||
-	    (error = xfs_inobt_updkey(tcur, rkp, level + 1))) {
-		xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
-		return error;
-	}
-	xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
-	*stat = 1;
+	cur->bc_bufs[level] = NULL;
+	cur->bc_nlevels--;
+
+	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 	return 0;
 }
 
-/*
- * Split cur/level block in half.
- * Return new block number and its first record (to be inserted into parent).
- */
-STATIC int				/* error */
-xfs_inobt_split(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level to split */
-	xfs_agblock_t		*bnop,	/* output: block number allocated */
-	xfs_inobt_key_t		*keyp,	/* output: first key of new block */
-	xfs_btree_cur_t		**curp,	/* output: new cursor */
-	int			*stat)	/* success/failure */
-{
-	xfs_alloc_arg_t		args;	/* allocation argument structure */
-	int			error;	/* error return value */
-	int			i;	/* loop index/record number */
-	xfs_agblock_t		lbno;	/* left (current) block number */
-	xfs_buf_t		*lbp;	/* buffer for left block */
-	xfs_inobt_block_t	*left;	/* left (current) btree block */
-	xfs_inobt_key_t		*lkp;	/* left btree key pointer */
-	xfs_inobt_ptr_t		*lpp;	/* left btree address pointer */
-	xfs_inobt_rec_t		*lrp;	/* left btree record pointer */
-	xfs_buf_t		*rbp;	/* buffer for right block */
-	xfs_inobt_block_t	*right;	/* right (new) btree block */
-	xfs_inobt_key_t		*rkp;	/* right btree key pointer */
-	xfs_inobt_ptr_t		*rpp;	/* right btree address pointer */
-	xfs_inobt_rec_t		*rrp;	/* right btree record pointer */
-
-	/*
-	 * Set up left block (current one).
-	 */
-	lbp = cur->bc_bufs[level];
-	args.tp = cur->bc_tp;
-	args.mp = cur->bc_mp;
-	lbno = XFS_DADDR_TO_AGBNO(args.mp, XFS_BUF_ADDR(lbp));
-	/*
-	 * Allocate the new block.
-	 * If we can't do it, we're toast.  Give up.
-	 */
-	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno, lbno);
-	args.mod = args.minleft = args.alignment = args.total = args.wasdel =
-		args.isfl = args.userdata = args.minalignslop = 0;
-	args.minlen = args.maxlen = args.prod = 1;
-	args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	if ((error = xfs_alloc_vextent(&args)))
-		return error;
-	if (args.fsbno == NULLFSBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	ASSERT(args.len == 1);
-	rbp = xfs_btree_get_bufs(args.mp, args.tp, args.agno, args.agbno, 0);
-	/*
-	 * Set up the new block as "right".
-	 */
-	right = XFS_BUF_TO_INOBT_BLOCK(rbp);
-	/*
-	 * "Left" is the current (according to the cursor) block.
-	 */
-	left = XFS_BUF_TO_INOBT_BLOCK(lbp);
 #ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, left, level, lbp)))
-		return error;
-#endif
-	/*
-	 * Fill in the btree header for the new block.
-	 */
-	right->bb_magic = cpu_to_be32(xfs_magics[cur->bc_btnum]);
-	right->bb_level = left->bb_level;
-	right->bb_numrecs = cpu_to_be16(be16_to_cpu(left->bb_numrecs) / 2);
-	/*
-	 * Make sure that if there's an odd number of entries now, that
-	 * each new block will have the same number of entries.
-	 */
-	if ((be16_to_cpu(left->bb_numrecs) & 1) &&
-	    cur->bc_ptrs[level] <= be16_to_cpu(right->bb_numrecs) + 1)
-		be16_add_cpu(&right->bb_numrecs, 1);
-	i = be16_to_cpu(left->bb_numrecs) - be16_to_cpu(right->bb_numrecs) + 1;
-	/*
-	 * For non-leaf blocks, copy keys and addresses over to the new block.
-	 */
-	if (level > 0) {
-		lkp = XFS_INOBT_KEY_ADDR(left, i, cur);
-		lpp = XFS_INOBT_PTR_ADDR(left, i, cur);
-		rkp = XFS_INOBT_KEY_ADDR(right, 1, cur);
-		rpp = XFS_INOBT_PTR_ADDR(right, 1, cur);
-#ifdef DEBUG
-		for (i = 0; i < be16_to_cpu(right->bb_numrecs); i++) {
-			if ((error = xfs_btree_check_sptr(cur, be32_to_cpu(lpp[i]), level)))
-				return error;
-		}
-#endif
-		memcpy(rkp, lkp, be16_to_cpu(right->bb_numrecs) * sizeof(*rkp));
-		memcpy(rpp, lpp, be16_to_cpu(right->bb_numrecs) * sizeof(*rpp));
-		xfs_inobt_log_keys(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		xfs_inobt_log_ptrs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		*keyp = *rkp;
-	}
-	/*
-	 * For leaf blocks, copy records over to the new block.
-	 */
-	else {
-		lrp = XFS_INOBT_REC_ADDR(left, i, cur);
-		rrp = XFS_INOBT_REC_ADDR(right, 1, cur);
-		memcpy(rrp, lrp, be16_to_cpu(right->bb_numrecs) * sizeof(*rrp));
-		xfs_inobt_log_recs(cur, rbp, 1, be16_to_cpu(right->bb_numrecs));
-		keyp->ir_startino = rrp->ir_startino;
-	}
-	/*
-	 * Find the left block number by looking in the buffer.
-	 * Adjust numrecs, sibling pointers.
-	 */
-	be16_add_cpu(&left->bb_numrecs, -(be16_to_cpu(right->bb_numrecs)));
-	right->bb_rightsib = left->bb_rightsib;
-	left->bb_rightsib = cpu_to_be32(args.agbno);
-	right->bb_leftsib = cpu_to_be32(lbno);
-	xfs_inobt_log_block(args.tp, rbp, XFS_BB_ALL_BITS);
-	xfs_inobt_log_block(args.tp, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB);
-	/*
-	 * If there's a block to the new block's right, make that block
-	 * point back to right instead of to left.
-	 */
-	if (be32_to_cpu(right->bb_rightsib) != NULLAGBLOCK) {
-		xfs_inobt_block_t	*rrblock;	/* rr btree block */
-		xfs_buf_t		*rrbp;		/* buffer for rrblock */
-
-		if ((error = xfs_btree_read_bufs(args.mp, args.tp, args.agno,
-				be32_to_cpu(right->bb_rightsib), 0, &rrbp,
-				XFS_INO_BTREE_REF)))
-			return error;
-		rrblock = XFS_BUF_TO_INOBT_BLOCK(rrbp);
-		if ((error = xfs_btree_check_sblock(cur, rrblock, level, rrbp)))
-			return error;
-		rrblock->bb_leftsib = cpu_to_be32(args.agbno);
-		xfs_inobt_log_block(args.tp, rrbp, XFS_BB_LEFTSIB);
-	}
-	/*
-	 * If the cursor is really in the right block, move it there.
-	 * If it's just pointing past the last entry in left, then we'll
-	 * insert there, so don't change anything in that case.
-	 */
-	if (cur->bc_ptrs[level] > be16_to_cpu(left->bb_numrecs) + 1) {
-		xfs_btree_setbuf(cur, level, rbp);
-		cur->bc_ptrs[level] -= be16_to_cpu(left->bb_numrecs);
-	}
-	/*
-	 * If there are more levels, we'll need another cursor which refers
-	 * the right block, no matter where this cursor was.
-	 */
-	if (level + 1 < cur->bc_nlevels) {
-		if ((error = xfs_btree_dup_cursor(cur, curp)))
-			return error;
-		(*curp)->bc_ptrs[level + 1]++;
-	}
-	*bnop = args.agbno;
-	*stat = 1;
-	return 0;
+STATIC int
+xfs_inobt_keys_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*k1,
+	union xfs_btree_key	*k2)
+{
+	return be32_to_cpu(k1->inobt.ir_startino) <
+		be32_to_cpu(k2->inobt.ir_startino);
 }
 
-/*
- * Update keys at all levels from here to the root along the cursor's path.
- */
-STATIC int				/* error */
-xfs_inobt_updkey(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_inobt_key_t		*keyp,	/* new key value to update to */
-	int			level)	/* starting level for update */
+STATIC int
+xfs_inobt_recs_inorder(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*r1,
+	union xfs_btree_rec	*r2)
 {
-	int			ptr;	/* index of key in block */
-
-	/*
-	 * Go up the tree from this level toward the root.
-	 * At each level, update the key value to the value input.
-	 * Stop when we reach a level where the cursor isn't pointing
-	 * at the first entry in the block.
-	 */
-	for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) {
-		xfs_buf_t		*bp;	/* buffer for block */
-		xfs_inobt_block_t	*block;	/* btree block */
-#ifdef DEBUG
-		int			error;	/* error return value */
-#endif
-		xfs_inobt_key_t		*kp;	/* ptr to btree block keys */
-
-		bp = cur->bc_bufs[level];
-		block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-			return error;
-#endif
-		ptr = cur->bc_ptrs[level];
-		kp = XFS_INOBT_KEY_ADDR(block, ptr, cur);
-		*kp = *keyp;
-		xfs_inobt_log_keys(cur, bp, ptr, ptr);
-	}
-	return 0;
+	return be32_to_cpu(r1->inobt.ir_startino) + XFS_INODES_PER_CHUNK <=
+		be32_to_cpu(r2->inobt.ir_startino);
 }
+#endif	/* DEBUG */
 
-/*
- * Externally visible routines.
- */
+#ifdef XFS_BTREE_TRACE
+ktrace_t	*xfs_inobt_trace_buf;
 
-/*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int					/* error */
-xfs_inobt_decrement(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree, 0 is leaf */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_inobt_trace_enter(
+	struct xfs_btree_cur	*cur,
+	const char		*func,
+	char			*s,
+	int			type,
+	int			line,
+	__psunsigned_t		a0,
+	__psunsigned_t		a1,
+	__psunsigned_t		a2,
+	__psunsigned_t		a3,
+	__psunsigned_t		a4,
+	__psunsigned_t		a5,
+	__psunsigned_t		a6,
+	__psunsigned_t		a7,
+	__psunsigned_t		a8,
+	__psunsigned_t		a9,
+	__psunsigned_t		a10)
 {
-	xfs_inobt_block_t	*block;	/* btree block */
-	int			error;
-	int			lev;	/* btree level */
-
-	ASSERT(level < cur->bc_nlevels);
-	/*
-	 * Read-ahead to the left at this level.
-	 */
-	xfs_btree_readahead(cur, level, XFS_BTCUR_LEFTRA);
-	/*
-	 * Decrement the ptr at this level.  If we're still in the block
-	 * then we're done.
-	 */
-	if (--cur->bc_ptrs[level] > 0) {
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * Get a pointer to the btree block.
-	 */
-	block = XFS_BUF_TO_INOBT_BLOCK(cur->bc_bufs[level]);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level,
-			cur->bc_bufs[level])))
-		return error;
-#endif
-	/*
-	 * If we just went off the left edge of the tree, return failure.
-	 */
-	if (be32_to_cpu(block->bb_leftsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * March up the tree decrementing pointers.
-	 * Stop when we don't go off the left edge of a block.
-	 */
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		if (--cur->bc_ptrs[lev] > 0)
-			break;
-		/*
-		 * Read-ahead the left block, we're going to read it
-		 * in the next loop.
-		 */
-		xfs_btree_readahead(cur, lev, XFS_BTCUR_LEFTRA);
-	}
-	/*
-	 * If we went off the root then we are seriously confused.
-	 */
-	ASSERT(lev < cur->bc_nlevels);
-	/*
-	 * Now walk back down the tree, fixing up the cursor's buffer
-	 * pointers and key numbers.
-	 */
-	for (block = XFS_BUF_TO_INOBT_BLOCK(cur->bc_bufs[lev]); lev > level; ) {
-		xfs_agblock_t	agbno;	/* block number of btree block */
-		xfs_buf_t	*bp;	/* buffer containing btree block */
-
-		agbno = be32_to_cpu(*XFS_INOBT_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, agbno, 0, &bp,
-				XFS_INO_BTREE_REF)))
-			return error;
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_INOBT_BLOCK(bp);
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-		cur->bc_ptrs[lev] = be16_to_cpu(block->bb_numrecs);
-	}
-	*stat = 1;
-	return 0;
+	ktrace_enter(xfs_inobt_trace_buf, (void *)(__psint_t)type,
+		(void *)func, (void *)s, NULL, (void *)cur,
+		(void *)a0, (void *)a1, (void *)a2, (void *)a3,
+		(void *)a4, (void *)a5, (void *)a6, (void *)a7,
+		(void *)a8, (void *)a9, (void *)a10);
 }
 
-/*
- * Delete the record pointed to by cur.
- * The cursor refers to the place where the record was (could be inserted)
- * when the operation returns.
- */
-int					/* error */
-xfs_inobt_delete(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	int		*stat)		/* success/failure */
+STATIC void
+xfs_inobt_trace_cursor(
+	struct xfs_btree_cur	*cur,
+	__uint32_t		*s0,
+	__uint64_t		*l0,
+	__uint64_t		*l1)
 {
-	int		error;
-	int		i;		/* result code */
-	int		level;		/* btree level */
-
-	/*
-	 * Go up the tree, starting at leaf level.
-	 * If 2 is returned then a join was done; go to the next level.
-	 * Otherwise we are done.
-	 */
-	for (level = 0, i = 2; i == 2; level++) {
-		if ((error = xfs_inobt_delrec(cur, level, &i)))
-			return error;
-	}
-	if (i == 0) {
-		for (level = 1; level < cur->bc_nlevels; level++) {
-			if (cur->bc_ptrs[level] == 0) {
-				if ((error = xfs_inobt_decrement(cur, level, &i)))
-					return error;
-				break;
-			}
-		}
-	}
-	*stat = i;
-	return 0;
+	*s0 = cur->bc_private.a.agno;
+	*l0 = cur->bc_rec.i.ir_startino;
+	*l1 = cur->bc_rec.i.ir_free;
 }
 
-
-/*
- * Get the data from the pointed-to record.
- */
-int					/* error */
-xfs_inobt_get_rec(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agino_t		*ino,	/* output: starting inode of chunk */
-	__int32_t		*fcnt,	/* output: number of free inodes */
-	xfs_inofree_t		*free,	/* output: free inode mask */
-	int			*stat)	/* output: success/failure */
+STATIC void
+xfs_inobt_trace_key(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_key	*key,
+	__uint64_t		*l0,
+	__uint64_t		*l1)
 {
-	xfs_inobt_block_t	*block;	/* btree block */
-	xfs_buf_t		*bp;	/* buffer containing btree block */
-#ifdef DEBUG
-	int			error;	/* error return value */
-#endif
-	int			ptr;	/* record number */
-	xfs_inobt_rec_t		*rec;	/* record data */
-
-	bp = cur->bc_bufs[0];
-	ptr = cur->bc_ptrs[0];
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, 0, bp)))
-		return error;
-#endif
-	/*
-	 * Off the right end or left end, return failure.
-	 */
-	if (ptr > be16_to_cpu(block->bb_numrecs) || ptr <= 0) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * Point to the record and extract its data.
-	 */
-	rec = XFS_INOBT_REC_ADDR(block, ptr, cur);
-	*ino = be32_to_cpu(rec->ir_startino);
-	*fcnt = be32_to_cpu(rec->ir_freecount);
-	*free = be64_to_cpu(rec->ir_free);
-	*stat = 1;
-	return 0;
+	*l0 = be32_to_cpu(key->inobt.ir_startino);
+	*l1 = 0;
 }
 
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-int					/* error */
-xfs_inobt_increment(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	int			level,	/* level in btree, 0 is leaf */
-	int			*stat)	/* success/failure */
+STATIC void
+xfs_inobt_trace_record(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_rec	*rec,
+	__uint64_t		*l0,
+	__uint64_t		*l1,
+	__uint64_t		*l2)
 {
-	xfs_inobt_block_t	*block;	/* btree block */
-	xfs_buf_t		*bp;	/* buffer containing btree block */
-	int			error;	/* error return value */
-	int			lev;	/* btree level */
+	*l0 = be32_to_cpu(rec->inobt.ir_startino);
+	*l1 = be32_to_cpu(rec->inobt.ir_freecount);
+	*l2 = be64_to_cpu(rec->inobt.ir_free);
+}
+#endif /* XFS_BTREE_TRACE */
+
+static const struct xfs_btree_ops xfs_inobt_ops = {
+	.rec_len		= sizeof(xfs_inobt_rec_t),
+	.key_len		= sizeof(xfs_inobt_key_t),
+
+	.dup_cursor		= xfs_inobt_dup_cursor,
+	.set_root		= xfs_inobt_set_root,
+	.kill_root		= xfs_inobt_kill_root,
+	.alloc_block		= xfs_inobt_alloc_block,
+	.free_block		= xfs_inobt_free_block,
+	.get_minrecs		= xfs_inobt_get_minrecs,
+	.get_maxrecs		= xfs_inobt_get_maxrecs,
+	.init_key_from_rec	= xfs_inobt_init_key_from_rec,
+	.init_rec_from_key	= xfs_inobt_init_rec_from_key,
+	.init_rec_from_cur	= xfs_inobt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_inobt_init_ptr_from_cur,
+	.key_diff		= xfs_inobt_key_diff,
 
-	ASSERT(level < cur->bc_nlevels);
-	/*
-	 * Read-ahead to the right at this level.
-	 */
-	xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA);
-	/*
-	 * Get a pointer to the btree block.
-	 */
-	bp = cur->bc_bufs[level];
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, level, bp)))
-		return error;
-#endif
-	/*
-	 * Increment the ptr at this level.  If we're still in the block
-	 * then we're done.
-	 */
-	if (++cur->bc_ptrs[level] <= be16_to_cpu(block->bb_numrecs)) {
-		*stat = 1;
-		return 0;
-	}
-	/*
-	 * If we just went off the right edge of the tree, return failure.
-	 */
-	if (be32_to_cpu(block->bb_rightsib) == NULLAGBLOCK) {
-		*stat = 0;
-		return 0;
-	}
-	/*
-	 * March up the tree incrementing pointers.
-	 * Stop when we don't go off the right edge of a block.
-	 */
-	for (lev = level + 1; lev < cur->bc_nlevels; lev++) {
-		bp = cur->bc_bufs[lev];
-		block = XFS_BUF_TO_INOBT_BLOCK(bp);
 #ifdef DEBUG
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
+	.keys_inorder		= xfs_inobt_keys_inorder,
+	.recs_inorder		= xfs_inobt_recs_inorder,
 #endif
-		if (++cur->bc_ptrs[lev] <= be16_to_cpu(block->bb_numrecs))
-			break;
-		/*
-		 * Read-ahead the right block, we're going to read it
-		 * in the next loop.
-		 */
-		xfs_btree_readahead(cur, lev, XFS_BTCUR_RIGHTRA);
-	}
-	/*
-	 * If we went off the root then we are seriously confused.
-	 */
-	ASSERT(lev < cur->bc_nlevels);
-	/*
-	 * Now walk back down the tree, fixing up the cursor's buffer
-	 * pointers and key numbers.
-	 */
-	for (bp = cur->bc_bufs[lev], block = XFS_BUF_TO_INOBT_BLOCK(bp);
-	     lev > level; ) {
-		xfs_agblock_t	agbno;	/* block number of btree block */
 
-		agbno = be32_to_cpu(*XFS_INOBT_PTR_ADDR(block, cur->bc_ptrs[lev], cur));
-		if ((error = xfs_btree_read_bufs(cur->bc_mp, cur->bc_tp,
-				cur->bc_private.a.agno, agbno, 0, &bp,
-				XFS_INO_BTREE_REF)))
-			return error;
-		lev--;
-		xfs_btree_setbuf(cur, lev, bp);
-		block = XFS_BUF_TO_INOBT_BLOCK(bp);
-		if ((error = xfs_btree_check_sblock(cur, block, lev, bp)))
-			return error;
-		cur->bc_ptrs[lev] = 1;
-	}
-	*stat = 1;
-	return 0;
-}
+#ifdef XFS_BTREE_TRACE
+	.trace_enter		= xfs_inobt_trace_enter,
+	.trace_cursor		= xfs_inobt_trace_cursor,
+	.trace_key		= xfs_inobt_trace_key,
+	.trace_record		= xfs_inobt_trace_record,
+#endif
+};
 
 /*
- * Insert the current record at the point referenced by cur.
- * The cursor may be inconsistent on return if splits have been done.
+ * Allocate a new inode btree cursor.
  */
-int					/* error */
-xfs_inobt_insert(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	int		*stat)		/* success/failure */
+struct xfs_btree_cur *				/* new inode btree cursor */
+xfs_inobt_init_cursor(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_buf		*agbp,		/* buffer for agi structure */
+	xfs_agnumber_t		agno)		/* allocation group number */
 {
-	int		error;		/* error return value */
-	int		i;		/* result value, 0 for failure */
-	int		level;		/* current level number in btree */
-	xfs_agblock_t	nbno;		/* new block number (split result) */
-	xfs_btree_cur_t	*ncur;		/* new cursor (split result) */
-	xfs_inobt_rec_t	nrec;		/* record being inserted this level */
-	xfs_btree_cur_t	*pcur;		/* previous level's cursor */
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_btree_cur	*cur;
 
-	level = 0;
-	nbno = NULLAGBLOCK;
-	nrec.ir_startino = cpu_to_be32(cur->bc_rec.i.ir_startino);
-	nrec.ir_freecount = cpu_to_be32(cur->bc_rec.i.ir_freecount);
-	nrec.ir_free = cpu_to_be64(cur->bc_rec.i.ir_free);
-	ncur = NULL;
-	pcur = cur;
-	/*
-	 * Loop going up the tree, starting at the leaf level.
-	 * Stop when we don't get a split block, that must mean that
-	 * the insert is finished with this level.
-	 */
-	do {
-		/*
-		 * Insert nrec/nbno into this level of the tree.
-		 * Note if we fail, nbno will be null.
-		 */
-		if ((error = xfs_inobt_insrec(pcur, level++, &nbno, &nrec, &ncur,
-				&i))) {
-			if (pcur != cur)
-				xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR);
-			return error;
-		}
-		/*
-		 * See if the cursor we just used is trash.
-		 * Can't trash the caller's cursor, but otherwise we should
-		 * if ncur is a new cursor or we're about to be done.
-		 */
-		if (pcur != cur && (ncur || nbno == NULLAGBLOCK)) {
-			cur->bc_nlevels = pcur->bc_nlevels;
-			xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR);
-		}
-		/*
-		 * If we got a new cursor, switch to it.
-		 */
-		if (ncur) {
-			pcur = ncur;
-			ncur = NULL;
-		}
-	} while (nbno != NULLAGBLOCK);
-	*stat = i;
-	return 0;
-}
+	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
 
-/*
- * Lookup the record equal to ino in the btree given by cur.
- */
-int					/* error */
-xfs_inobt_lookup_eq(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agino_t	ino,		/* starting inode of chunk */
-	__int32_t	fcnt,		/* free inode count */
-	xfs_inofree_t	free,		/* free inode mask */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.i.ir_startino = ino;
-	cur->bc_rec.i.ir_freecount = fcnt;
-	cur->bc_rec.i.ir_free = free;
-	return xfs_inobt_lookup(cur, XFS_LOOKUP_EQ, stat);
-}
+	cur->bc_tp = tp;
+	cur->bc_mp = mp;
+	cur->bc_nlevels = be32_to_cpu(agi->agi_level);
+	cur->bc_btnum = XFS_BTNUM_INO;
+	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 
-/*
- * Lookup the first record greater than or equal to ino
- * in the btree given by cur.
- */
-int					/* error */
-xfs_inobt_lookup_ge(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agino_t	ino,		/* starting inode of chunk */
-	__int32_t	fcnt,		/* free inode count */
-	xfs_inofree_t	free,		/* free inode mask */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.i.ir_startino = ino;
-	cur->bc_rec.i.ir_freecount = fcnt;
-	cur->bc_rec.i.ir_free = free;
-	return xfs_inobt_lookup(cur, XFS_LOOKUP_GE, stat);
-}
+	cur->bc_ops = &xfs_inobt_ops;
 
-/*
- * Lookup the first record less than or equal to ino
- * in the btree given by cur.
- */
-int					/* error */
-xfs_inobt_lookup_le(
-	xfs_btree_cur_t	*cur,		/* btree cursor */
-	xfs_agino_t	ino,		/* starting inode of chunk */
-	__int32_t	fcnt,		/* free inode count */
-	xfs_inofree_t	free,		/* free inode mask */
-	int		*stat)		/* success/failure */
-{
-	cur->bc_rec.i.ir_startino = ino;
-	cur->bc_rec.i.ir_freecount = fcnt;
-	cur->bc_rec.i.ir_free = free;
-	return xfs_inobt_lookup(cur, XFS_LOOKUP_LE, stat);
+	cur->bc_private.a.agbp = agbp;
+	cur->bc_private.a.agno = agno;
+
+	return cur;
 }
 
 /*
- * Update the record referred to by cur, to the value given
- * by [ino, fcnt, free].
- * This either works (return 0) or gets an EFSCORRUPTED error.
+ * Calculate number of records in an inobt btree block.
  */
-int					/* error */
-xfs_inobt_update(
-	xfs_btree_cur_t		*cur,	/* btree cursor */
-	xfs_agino_t		ino,	/* starting inode of chunk */
-	__int32_t		fcnt,	/* free inode count */
-	xfs_inofree_t		free)	/* free inode mask */
+int
+xfs_inobt_maxrecs(
+	struct xfs_mount	*mp,
+	int			blocklen,
+	int			leaf)
 {
-	xfs_inobt_block_t	*block;	/* btree block to update */
-	xfs_buf_t		*bp;	/* buffer containing btree block */
-	int			error;	/* error return value */
-	int			ptr;	/* current record number (updating) */
-	xfs_inobt_rec_t		*rp;	/* pointer to updated record */
+	blocklen -= XFS_INOBT_BLOCK_LEN(mp);
 
-	/*
-	 * Pick up the current block.
-	 */
-	bp = cur->bc_bufs[0];
-	block = XFS_BUF_TO_INOBT_BLOCK(bp);
-#ifdef DEBUG
-	if ((error = xfs_btree_check_sblock(cur, block, 0, bp)))
-		return error;
-#endif
-	/*
-	 * Get the address of the rec to be updated.
-	 */
-	ptr = cur->bc_ptrs[0];
-	rp = XFS_INOBT_REC_ADDR(block, ptr, cur);
-	/*
-	 * Fill in the new contents and log them.
-	 */
-	rp->ir_startino = cpu_to_be32(ino);
-	rp->ir_freecount = cpu_to_be32(fcnt);
-	rp->ir_free = cpu_to_be64(free);
-	xfs_inobt_log_recs(cur, bp, ptr, ptr);
-	/*
-	 * Updating first record in leaf. Pass new key value up to our parent.
-	 */
-	if (ptr == 1) {
-		xfs_inobt_key_t	key;	/* key containing [ino] */
-
-		key.ir_startino = cpu_to_be32(ino);
-		if ((error = xfs_inobt_updkey(cur, &key, 1)))
-			return error;
-	}
-	return 0;
+	if (leaf)
+		return blocklen / sizeof(xfs_inobt_rec_t);
+	return blocklen / (sizeof(xfs_inobt_key_t) + sizeof(xfs_inobt_ptr_t));
 }
diff --git a/fs/xfs/xfs_ialloc_btree.h b/fs/xfs/xfs_ialloc_btree.h
index 8efc4a5b8b9..37e5dd01a57 100644
--- a/fs/xfs/xfs_ialloc_btree.h
+++ b/fs/xfs/xfs_ialloc_btree.h
@@ -24,7 +24,6 @@
 
 struct xfs_buf;
 struct xfs_btree_cur;
-struct xfs_btree_sblock;
 struct xfs_mount;
 
 /*
@@ -70,11 +69,6 @@ typedef struct xfs_inobt_key {
 /* btree pointer type */
 typedef __be32 xfs_inobt_ptr_t;
 
-/* btree block header type */
-typedef	struct xfs_btree_sblock xfs_inobt_block_t;
-
-#define	XFS_BUF_TO_INOBT_BLOCK(bp)	((xfs_inobt_block_t *)XFS_BUF_PTR(bp))
-
 /*
  * Bit manipulations for ir_free.
  */
@@ -85,14 +79,6 @@ typedef	struct xfs_btree_sblock xfs_inobt_block_t;
 #define	XFS_INOBT_CLR_FREE(rp,i)	((rp)->ir_free &= ~XFS_INOBT_MASK(i))
 
 /*
- * Real block structures have a size equal to the disk block size.
- */
-#define	XFS_INOBT_BLOCK_MAXRECS(lev,cur) ((cur)->bc_mp->m_inobt_mxr[lev != 0])
-#define	XFS_INOBT_BLOCK_MINRECS(lev,cur) ((cur)->bc_mp->m_inobt_mnr[lev != 0])
-#define	XFS_INOBT_IS_LAST_REC(cur)	\
-	((cur)->bc_ptrs[0] == be16_to_cpu(XFS_BUF_TO_INOBT_BLOCK((cur)->bc_bufs[0])->bb_numrecs))
-
-/*
  * Maximum number of inode btree levels.
  */
 #define	XFS_IN_MAXLEVELS(mp)		((mp)->m_in_maxlevels)
@@ -104,75 +90,38 @@ typedef	struct xfs_btree_sblock xfs_inobt_block_t;
 #define	XFS_PREALLOC_BLOCKS(mp)		((xfs_agblock_t)(XFS_IBT_BLOCK(mp) + 1))
 
 /*
- * Record, key, and pointer address macros for btree blocks.
- */
-#define XFS_INOBT_REC_ADDR(bb,i,cur) \
-	(XFS_BTREE_REC_ADDR(xfs_inobt, bb, i))
-
-#define	XFS_INOBT_KEY_ADDR(bb,i,cur) \
-	(XFS_BTREE_KEY_ADDR(xfs_inobt, bb, i))
-
-#define	XFS_INOBT_PTR_ADDR(bb,i,cur) \
-	(XFS_BTREE_PTR_ADDR(xfs_inobt, bb, \
-				i, XFS_INOBT_BLOCK_MAXRECS(1, cur)))
-
-/*
- * Decrement cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-extern int xfs_inobt_decrement(struct xfs_btree_cur *cur, int level, int *stat);
-
-/*
- * Delete the record pointed to by cur.
- * The cursor refers to the place where the record was (could be inserted)
- * when the operation returns.
- */
-extern int xfs_inobt_delete(struct xfs_btree_cur *cur, int *stat);
-
-/*
- * Get the data from the pointed-to record.
- */
-extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur, xfs_agino_t *ino,
-			     __int32_t *fcnt, xfs_inofree_t *free, int *stat);
-
-/*
- * Increment cursor by one record at the level.
- * For nonzero levels the leaf-ward information is untouched.
- */
-extern int xfs_inobt_increment(struct xfs_btree_cur *cur, int level, int *stat);
-
-/*
- * Insert the current record at the point referenced by cur.
- * The cursor may be inconsistent on return if splits have been done.
- */
-extern int xfs_inobt_insert(struct xfs_btree_cur *cur, int *stat);
-
-/*
- * Lookup the record equal to ino in the btree given by cur.
- */
-extern int xfs_inobt_lookup_eq(struct xfs_btree_cur *cur, xfs_agino_t ino,
-				__int32_t fcnt, xfs_inofree_t free, int *stat);
-
-/*
- * Lookup the first record greater than or equal to ino
- * in the btree given by cur.
- */
-extern int xfs_inobt_lookup_ge(struct xfs_btree_cur *cur, xfs_agino_t ino,
-				__int32_t fcnt,	xfs_inofree_t free, int *stat);
-
-/*
- * Lookup the first record less than or equal to ino
- * in the btree given by cur.
+ * Btree block header size depends on a superblock flag.
+ *
+ * (not quite yet, but soon)
  */
-extern int xfs_inobt_lookup_le(struct xfs_btree_cur *cur, xfs_agino_t ino,
-				__int32_t fcnt,	xfs_inofree_t free, int *stat);
+#define XFS_INOBT_BLOCK_LEN(mp)	XFS_BTREE_SBLOCK_LEN
 
 /*
- * Update the record referred to by cur, to the value given
- * by [ino, fcnt, free].
- * This either works (return 0) or gets an EFSCORRUPTED error.
- */
-extern int xfs_inobt_update(struct xfs_btree_cur *cur, xfs_agino_t ino,
-				__int32_t fcnt, xfs_inofree_t free);
+ * Record, key, and pointer address macros for btree blocks.
+ *
+ * (note that some of these may appear unused, but they are used in userspace)
+ */
+#define XFS_INOBT_REC_ADDR(mp, block, index) \
+	((xfs_inobt_rec_t *) \
+		((char *)(block) + \
+		 XFS_INOBT_BLOCK_LEN(mp) + \
+		 (((index) - 1) * sizeof(xfs_inobt_rec_t))))
+
+#define XFS_INOBT_KEY_ADDR(mp, block, index) \
+	((xfs_inobt_key_t *) \
+		((char *)(block) + \
+		 XFS_INOBT_BLOCK_LEN(mp) + \
+		 ((index) - 1) * sizeof(xfs_inobt_key_t)))
+
+#define XFS_INOBT_PTR_ADDR(mp, block, index, maxrecs) \
+	((xfs_inobt_ptr_t *) \
+		((char *)(block) + \
+		 XFS_INOBT_BLOCK_LEN(mp) + \
+		 (maxrecs) * sizeof(xfs_inobt_key_t) + \
+		 ((index) - 1) * sizeof(xfs_inobt_ptr_t)))
+
+extern struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_mount *,
+		struct xfs_trans *, struct xfs_buf *, xfs_agnumber_t);
+extern int xfs_inobt_maxrecs(struct xfs_mount *, int, int);
 
 #endif	/* __XFS_IALLOC_BTREE_H__ */
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index e229e9e001c..e2fb6210d4c 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -38,281 +38,283 @@
 #include "xfs_ialloc.h"
 #include "xfs_quota.h"
 #include "xfs_utils.h"
+#include "xfs_trans_priv.h"
+#include "xfs_inode_item.h"
+#include "xfs_bmap.h"
+#include "xfs_btree_trace.h"
+#include "xfs_dir2_trace.h"
+
 
 /*
- * Look up an inode by number in the given file system.
- * The inode is looked up in the cache held in each AG.
- * If the inode is found in the cache, attach it to the provided
- * vnode.
- *
- * If it is not in core, read it in from the file system's device,
- * add it to the cache and attach the provided vnode.
- *
- * The inode is locked according to the value of the lock_flags parameter.
- * This flag parameter indicates how and if the inode's IO lock and inode lock
- * should be taken.
- *
- * mp -- the mount point structure for the current file system.  It points
- *       to the inode hash table.
- * tp -- a pointer to the current transaction if there is one.  This is
- *       simply passed through to the xfs_iread() call.
- * ino -- the number of the inode desired.  This is the unique identifier
- *        within the file system for the inode being requested.
- * lock_flags -- flags indicating how to lock the inode.  See the comment
- *		 for xfs_ilock() for a list of valid values.
- * bno -- the block number starting the buffer containing the inode,
- *	  if known (as by bulkstat), else 0.
+ * Allocate and initialise an xfs_inode.
  */
-STATIC int
-xfs_iget_core(
-	struct inode	*inode,
-	xfs_mount_t	*mp,
-	xfs_trans_t	*tp,
-	xfs_ino_t	ino,
-	uint		flags,
-	uint		lock_flags,
-	xfs_inode_t	**ipp,
-	xfs_daddr_t	bno)
+STATIC struct xfs_inode *
+xfs_inode_alloc(
+	struct xfs_mount	*mp,
+	xfs_ino_t		ino)
 {
-	struct inode	*old_inode;
-	xfs_inode_t	*ip;
-	xfs_inode_t	*iq;
-	int		error;
-	unsigned long	first_index, mask;
-	xfs_perag_t	*pag;
-	xfs_agino_t	agino;
+	struct xfs_inode	*ip;
 
-	/* the radix tree exists only in inode capable AGs */
-	if (XFS_INO_TO_AGNO(mp, ino) >= mp->m_maxagi)
-		return EINVAL;
+	/*
+	 * if this didn't occur in transactions, we could use
+	 * KM_MAYFAIL and return NULL here on ENOMEM. Set the
+	 * code up to do this anyway.
+	 */
+	ip = kmem_zone_alloc(xfs_inode_zone, KM_SLEEP);
+	if (!ip)
+		return NULL;
 
-	/* get the perag structure and ensure that it's inode capable */
-	pag = xfs_get_perag(mp, ino);
-	if (!pag->pagi_inodeok)
-		return EINVAL;
-	ASSERT(pag->pag_ici_init);
-	agino = XFS_INO_TO_AGINO(mp, ino);
+	ASSERT(atomic_read(&ip->i_iocount) == 0);
+	ASSERT(atomic_read(&ip->i_pincount) == 0);
+	ASSERT(!spin_is_locked(&ip->i_flags_lock));
+	ASSERT(completion_done(&ip->i_flush));
 
-again:
-	read_lock(&pag->pag_ici_lock);
-	ip = radix_tree_lookup(&pag->pag_ici_root, agino);
+	/*
+	 * initialise the VFS inode here to get failures
+	 * out of the way early.
+	 */
+	if (!inode_init_always(mp->m_super, VFS_I(ip))) {
+		kmem_zone_free(xfs_inode_zone, ip);
+		return NULL;
+	}
+
+	/* initialise the xfs inode */
+	ip->i_ino = ino;
+	ip->i_mount = mp;
+	memset(&ip->i_imap, 0, sizeof(struct xfs_imap));
+	ip->i_afp = NULL;
+	memset(&ip->i_df, 0, sizeof(xfs_ifork_t));
+	ip->i_flags = 0;
+	ip->i_update_core = 0;
+	ip->i_update_size = 0;
+	ip->i_delayed_blks = 0;
+	memset(&ip->i_d, 0, sizeof(xfs_icdinode_t));
+	ip->i_size = 0;
+	ip->i_new_size = 0;
+
+	/*
+	 * Initialize inode's trace buffers.
+	 */
+#ifdef	XFS_INODE_TRACE
+	ip->i_trace = ktrace_alloc(INODE_TRACE_SIZE, KM_NOFS);
+#endif
+#ifdef XFS_BMAP_TRACE
+	ip->i_xtrace = ktrace_alloc(XFS_BMAP_KTRACE_SIZE, KM_NOFS);
+#endif
+#ifdef XFS_BTREE_TRACE
+	ip->i_btrace = ktrace_alloc(XFS_BMBT_KTRACE_SIZE, KM_NOFS);
+#endif
+#ifdef XFS_RW_TRACE
+	ip->i_rwtrace = ktrace_alloc(XFS_RW_KTRACE_SIZE, KM_NOFS);
+#endif
+#ifdef XFS_ILOCK_TRACE
+	ip->i_lock_trace = ktrace_alloc(XFS_ILOCK_KTRACE_SIZE, KM_NOFS);
+#endif
+#ifdef XFS_DIR2_TRACE
+	ip->i_dir_trace = ktrace_alloc(XFS_DIR2_KTRACE_SIZE, KM_NOFS);
+#endif
+
+	return ip;
+}
+
+/*
+ * Check the validity of the inode we just found it the cache
+ */
+static int
+xfs_iget_cache_hit(
+	struct xfs_perag	*pag,
+	struct xfs_inode	*ip,
+	int			flags,
+	int			lock_flags) __releases(pag->pag_ici_lock)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	int			error = EAGAIN;
+
+	/*
+	 * If INEW is set this inode is being set up
+	 * If IRECLAIM is set this inode is being torn down
+	 * Pause and try again.
+	 */
+	if (xfs_iflags_test(ip, (XFS_INEW|XFS_IRECLAIM))) {
+		XFS_STATS_INC(xs_ig_frecycle);
+		goto out_error;
+	}
+
+	/* If IRECLAIMABLE is set, we've torn down the vfs inode part */
+	if (xfs_iflags_test(ip, XFS_IRECLAIMABLE)) {
 
-	if (ip != NULL) {
 		/*
-		 * If INEW is set this inode is being set up
-		 * we need to pause and try again.
+		 * If lookup is racing with unlink, then we should return an
+		 * error immediately so we don't remove it from the reclaim
+		 * list and potentially leak the inode.
 		 */
-		if (xfs_iflags_test(ip, XFS_INEW)) {
-			read_unlock(&pag->pag_ici_lock);
-			delay(1);
-			XFS_STATS_INC(xs_ig_frecycle);
-
-			goto again;
+		if ((ip->i_d.di_mode == 0) && !(flags & XFS_IGET_CREATE)) {
+			error = ENOENT;
+			goto out_error;
 		}
 
-		old_inode = ip->i_vnode;
-		if (old_inode == NULL) {
-			/*
-			 * If IRECLAIM is set this inode is
-			 * on its way out of the system,
-			 * we need to pause and try again.
-			 */
-			if (xfs_iflags_test(ip, XFS_IRECLAIM)) {
-				read_unlock(&pag->pag_ici_lock);
-				delay(1);
-				XFS_STATS_INC(xs_ig_frecycle);
-
-				goto again;
-			}
-			ASSERT(xfs_iflags_test(ip, XFS_IRECLAIMABLE));
-
-			/*
-			 * If lookup is racing with unlink, then we
-			 * should return an error immediately so we
-			 * don't remove it from the reclaim list and
-			 * potentially leak the inode.
-			 */
-			if ((ip->i_d.di_mode == 0) &&
-			    !(flags & XFS_IGET_CREATE)) {
-				read_unlock(&pag->pag_ici_lock);
-				xfs_put_perag(mp, pag);
-				return ENOENT;
-			}
-
-			xfs_itrace_exit_tag(ip, "xfs_iget.alloc");
-
-			XFS_STATS_INC(xs_ig_found);
-			xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
-			read_unlock(&pag->pag_ici_lock);
-
-			XFS_MOUNT_ILOCK(mp);
-			list_del_init(&ip->i_reclaim);
-			XFS_MOUNT_IUNLOCK(mp);
-
-			goto finish_inode;
-
-		} else if (inode != old_inode) {
-			/* The inode is being torn down, pause and
-			 * try again.
-			 */
-			if (old_inode->i_state & (I_FREEING | I_CLEAR)) {
-				read_unlock(&pag->pag_ici_lock);
-				delay(1);
-				XFS_STATS_INC(xs_ig_frecycle);
-
-				goto again;
-			}
-/* Chances are the other vnode (the one in the inode) is being torn
-* down right now, and we landed on top of it. Question is, what do
-* we do? Unhook the old inode and hook up the new one?
-*/
-			cmn_err(CE_PANIC,
-		"xfs_iget_core: ambiguous vns: vp/0x%p, invp/0x%p",
-					old_inode, inode);
-		}
+		xfs_itrace_exit_tag(ip, "xfs_iget.alloc");
 
 		/*
-		 * Inode cache hit
+		 * We need to re-initialise the VFS inode as it has been
+		 * 'freed' by the VFS. Do this here so we can deal with
+		 * errors cleanly, then tag it so it can be set up correctly
+		 * later.
 		 */
-		read_unlock(&pag->pag_ici_lock);
-		XFS_STATS_INC(xs_ig_found);
-
-finish_inode:
-		if (ip->i_d.di_mode == 0 && !(flags & XFS_IGET_CREATE)) {
-			xfs_put_perag(mp, pag);
-			return ENOENT;
+		if (!inode_init_always(mp->m_super, VFS_I(ip))) {
+			error = ENOMEM;
+			goto out_error;
 		}
 
-		if (lock_flags != 0)
-			xfs_ilock(ip, lock_flags);
+		/*
+		 * We must set the XFS_INEW flag before clearing the
+		 * XFS_IRECLAIMABLE flag so that if a racing lookup does
+		 * not find the XFS_IRECLAIMABLE above but has the igrab()
+		 * below succeed we can safely check XFS_INEW to detect
+		 * that this inode is still being initialised.
+		 */
+		xfs_iflags_set(ip, XFS_INEW);
+		xfs_iflags_clear(ip, XFS_IRECLAIMABLE);
+
+		/* clear the radix tree reclaim flag as well. */
+		__xfs_inode_clear_reclaim_tag(mp, pag, ip);
+	} else if (!igrab(VFS_I(ip))) {
+		/* If the VFS inode is being torn down, pause and try again. */
+		XFS_STATS_INC(xs_ig_frecycle);
+		goto out_error;
+	} else if (xfs_iflags_test(ip, XFS_INEW)) {
+		/*
+		 * We are racing with another cache hit that is
+		 * currently recycling this inode out of the XFS_IRECLAIMABLE
+		 * state. Wait for the initialisation to complete before
+		 * continuing.
+		 */
+		wait_on_inode(VFS_I(ip));
+	}
 
-		xfs_iflags_clear(ip, XFS_ISTALE);
-		xfs_itrace_exit_tag(ip, "xfs_iget.found");
-		goto return_ip;
+	if (ip->i_d.di_mode == 0 && !(flags & XFS_IGET_CREATE)) {
+		error = ENOENT;
+		iput(VFS_I(ip));
+		goto out_error;
 	}
 
-	/*
-	 * Inode cache miss
-	 */
+	/* We've got a live one. */
 	read_unlock(&pag->pag_ici_lock);
-	XFS_STATS_INC(xs_ig_missed);
 
-	/*
-	 * Read the disk inode attributes into a new inode structure and get
-	 * a new vnode for it. This should also initialize i_ino and i_mount.
-	 */
-	error = xfs_iread(mp, tp, ino, &ip, bno,
-			  (flags & XFS_IGET_BULKSTAT) ? XFS_IMAP_BULKSTAT : 0);
-	if (error) {
-		xfs_put_perag(mp, pag);
-		return error;
-	}
+	if (lock_flags != 0)
+		xfs_ilock(ip, lock_flags);
 
-	xfs_itrace_exit_tag(ip, "xfs_iget.alloc");
+	xfs_iflags_clear(ip, XFS_ISTALE);
+	xfs_itrace_exit_tag(ip, "xfs_iget.found");
+	XFS_STATS_INC(xs_ig_found);
+	return 0;
+
+out_error:
+	read_unlock(&pag->pag_ici_lock);
+	return error;
+}
 
 
-	mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER,
-		     "xfsino", ip->i_ino);
-	mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino);
-	init_waitqueue_head(&ip->i_ipin_wait);
-	atomic_set(&ip->i_pincount, 0);
+static int
+xfs_iget_cache_miss(
+	struct xfs_mount	*mp,
+	struct xfs_perag	*pag,
+	xfs_trans_t		*tp,
+	xfs_ino_t		ino,
+	struct xfs_inode	**ipp,
+	xfs_daddr_t		bno,
+	int			flags,
+	int			lock_flags) __releases(pag->pag_ici_lock)
+{
+	struct xfs_inode	*ip;
+	int			error;
+	unsigned long		first_index, mask;
+	xfs_agino_t		agino = XFS_INO_TO_AGINO(mp, ino);
 
-	/*
-	 * Because we want to use a counting completion, complete
-	 * the flush completion once to allow a single access to
-	 * the flush completion without blocking.
-	 */
-	init_completion(&ip->i_flush);
-	complete(&ip->i_flush);
+	ip = xfs_inode_alloc(mp, ino);
+	if (!ip)
+		return ENOMEM;
 
-	if (lock_flags)
-		xfs_ilock(ip, lock_flags);
+	error = xfs_iread(mp, tp, ip, bno, flags);
+	if (error)
+		goto out_destroy;
+
+	xfs_itrace_exit_tag(ip, "xfs_iget.alloc");
 
 	if ((ip->i_d.di_mode == 0) && !(flags & XFS_IGET_CREATE)) {
-		xfs_idestroy(ip);
-		xfs_put_perag(mp, pag);
-		return ENOENT;
+		error = ENOENT;
+		goto out_destroy;
 	}
 
+	if (lock_flags)
+		xfs_ilock(ip, lock_flags);
+
 	/*
 	 * Preload the radix tree so we can insert safely under the
-	 * write spinlock.
+	 * write spinlock. Note that we cannot sleep inside the preload
+	 * region.
 	 */
 	if (radix_tree_preload(GFP_KERNEL)) {
-		xfs_idestroy(ip);
-		delay(1);
-		goto again;
+		error = EAGAIN;
+		goto out_unlock;
 	}
+
 	mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1);
 	first_index = agino & mask;
 	write_lock(&pag->pag_ici_lock);
-	/*
-	 * insert the new inode
-	 */
+
+	/* insert the new inode */
 	error = radix_tree_insert(&pag->pag_ici_root, agino, ip);
 	if (unlikely(error)) {
-		BUG_ON(error != -EEXIST);
-		write_unlock(&pag->pag_ici_lock);
-		radix_tree_preload_end();
-		xfs_idestroy(ip);
+		WARN_ON(error != -EEXIST);
 		XFS_STATS_INC(xs_ig_dup);
-		goto again;
+		error = EAGAIN;
+		goto out_preload_end;
 	}
 
-	/*
-	 * These values _must_ be set before releasing the radix tree lock!
-	 */
+	/* These values _must_ be set before releasing the radix tree lock! */
 	ip->i_udquot = ip->i_gdquot = NULL;
 	xfs_iflags_set(ip, XFS_INEW);
 
 	write_unlock(&pag->pag_ici_lock);
 	radix_tree_preload_end();
-
-	/*
-	 * Link ip to its mount and thread it on the mount's inode list.
-	 */
-	XFS_MOUNT_ILOCK(mp);
-	if ((iq = mp->m_inodes)) {
-		ASSERT(iq->i_mprev->i_mnext == iq);
-		ip->i_mprev = iq->i_mprev;
-		iq->i_mprev->i_mnext = ip;
-		iq->i_mprev = ip;
-		ip->i_mnext = iq;
-	} else {
-		ip->i_mnext = ip;
-		ip->i_mprev = ip;
-	}
-	mp->m_inodes = ip;
-
-	XFS_MOUNT_IUNLOCK(mp);
-	xfs_put_perag(mp, pag);
-
- return_ip:
-	ASSERT(ip->i_df.if_ext_max ==
-	       XFS_IFORK_DSIZE(ip) / sizeof(xfs_bmbt_rec_t));
-
-	xfs_iflags_set(ip, XFS_IMODIFIED);
 	*ipp = ip;
-
-	/*
-	 * Set up the Linux with the Linux inode.
-	 */
-	ip->i_vnode = inode;
-	inode->i_private = ip;
-
-	/*
-	 * If we have a real type for an on-disk inode, we can set ops(&unlock)
-	 * now.	 If it's a new inode being created, xfs_ialloc will handle it.
-	 */
-	if (ip->i_d.di_mode != 0)
-		xfs_setup_inode(ip);
 	return 0;
-}
 
+out_preload_end:
+	write_unlock(&pag->pag_ici_lock);
+	radix_tree_preload_end();
+out_unlock:
+	if (lock_flags)
+		xfs_iunlock(ip, lock_flags);
+out_destroy:
+	xfs_destroy_inode(ip);
+	return error;
+}
 
 /*
- * The 'normal' internal xfs_iget, if needed it will
- * 'allocate', or 'get', the vnode.
+ * Look up an inode by number in the given file system.
+ * The inode is looked up in the cache held in each AG.
+ * If the inode is found in the cache, initialise the vfs inode
+ * if necessary.
+ *
+ * If it is not in core, read it in from the file system's device,
+ * add it to the cache and initialise the vfs inode.
+ *
+ * The inode is locked according to the value of the lock_flags parameter.
+ * This flag parameter indicates how and if the inode's IO lock and inode lock
+ * should be taken.
+ *
+ * mp -- the mount point structure for the current file system.  It points
+ *       to the inode hash table.
+ * tp -- a pointer to the current transaction if there is one.  This is
+ *       simply passed through to the xfs_iread() call.
+ * ino -- the number of the inode desired.  This is the unique identifier
+ *        within the file system for the inode being requested.
+ * lock_flags -- flags indicating how to lock the inode.  See the comment
+ *		 for xfs_ilock() for a list of valid values.
+ * bno -- the block number starting the buffer containing the inode,
+ *	  if known (as by bulkstat), else 0.
  */
 int
 xfs_iget(
@@ -324,61 +326,64 @@ xfs_iget(
 	xfs_inode_t	**ipp,
 	xfs_daddr_t	bno)
 {
-	struct inode	*inode;
 	xfs_inode_t	*ip;
 	int		error;
+	xfs_perag_t	*pag;
+	xfs_agino_t	agino;
 
-	XFS_STATS_INC(xs_ig_attempts);
+	/* the radix tree exists only in inode capable AGs */
+	if (XFS_INO_TO_AGNO(mp, ino) >= mp->m_maxagi)
+		return EINVAL;
 
-retry:
-	inode = iget_locked(mp->m_super, ino);
-	if (!inode)
-		/* If we got no inode we are out of memory */
-		return ENOMEM;
+	/* get the perag structure and ensure that it's inode capable */
+	pag = xfs_get_perag(mp, ino);
+	if (!pag->pagi_inodeok)
+		return EINVAL;
+	ASSERT(pag->pag_ici_init);
+	agino = XFS_INO_TO_AGINO(mp, ino);
 
-	if (inode->i_state & I_NEW) {
-		XFS_STATS_INC(vn_active);
-		XFS_STATS_INC(vn_alloc);
-
-		error = xfs_iget_core(inode, mp, tp, ino, flags,
-				lock_flags, ipp, bno);
-		if (error) {
-			make_bad_inode(inode);
-			if (inode->i_state & I_NEW)
-				unlock_new_inode(inode);
-			iput(inode);
-		}
-		return error;
+again:
+	error = 0;
+	read_lock(&pag->pag_ici_lock);
+	ip = radix_tree_lookup(&pag->pag_ici_root, agino);
+
+	if (ip) {
+		error = xfs_iget_cache_hit(pag, ip, flags, lock_flags);
+		if (error)
+			goto out_error_or_again;
+	} else {
+		read_unlock(&pag->pag_ici_lock);
+		XFS_STATS_INC(xs_ig_missed);
+
+		error = xfs_iget_cache_miss(mp, pag, tp, ino, &ip, bno,
+							flags, lock_flags);
+		if (error)
+			goto out_error_or_again;
 	}
+	xfs_put_perag(mp, pag);
 
+	*ipp = ip;
+
+	ASSERT(ip->i_df.if_ext_max ==
+	       XFS_IFORK_DSIZE(ip) / sizeof(xfs_bmbt_rec_t));
 	/*
-	 * If the inode is not fully constructed due to
-	 * filehandle mismatches wait for the inode to go
-	 * away and try again.
-	 *
-	 * iget_locked will call __wait_on_freeing_inode
-	 * to wait for the inode to go away.
+	 * If we have a real type for an on-disk inode, we can set ops(&unlock)
+	 * now.	 If it's a new inode being created, xfs_ialloc will handle it.
 	 */
-	if (is_bad_inode(inode)) {
-		iput(inode);
-		delay(1);
-		goto retry;
-	}
+	if (xfs_iflags_test(ip, XFS_INEW) && ip->i_d.di_mode != 0)
+		xfs_setup_inode(ip);
+	return 0;
 
-	ip = XFS_I(inode);
-	if (!ip) {
-		iput(inode);
+out_error_or_again:
+	if (error == EAGAIN) {
 		delay(1);
-		goto retry;
+		goto again;
 	}
-
-	if (lock_flags != 0)
-		xfs_ilock(ip, lock_flags);
-	XFS_STATS_INC(xs_ig_found);
-	*ipp = ip;
-	return 0;
+	xfs_put_perag(mp, pag);
+	return error;
 }
 
+
 /*
  * Look for the inode corresponding to the given ino in the hash table.
  * If it is there and its i_transp pointer matches tp, return it.
@@ -444,99 +449,109 @@ xfs_iput_new(
 	IRELE(ip);
 }
 
-
 /*
- * This routine embodies the part of the reclaim code that pulls
- * the inode from the inode hash table and the mount structure's
- * inode list.
- * This should only be called from xfs_reclaim().
+ * This is called free all the memory associated with an inode.
+ * It must free the inode itself and any buffers allocated for
+ * if_extents/if_data and if_broot.  It must also free the lock
+ * associated with the inode.
+ *
+ * Note: because we don't initialise everything on reallocation out
+ * of the zone, we must ensure we nullify everything correctly before
+ * freeing the structure.
  */
 void
-xfs_ireclaim(xfs_inode_t *ip)
+xfs_ireclaim(
+	struct xfs_inode	*ip)
 {
-	/*
-	 * Remove from old hash list and mount list.
-	 */
-	XFS_STATS_INC(xs_ig_reclaims);
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_perag	*pag;
 
-	xfs_iextract(ip);
-
-	/*
-	 * Here we do a spurious inode lock in order to coordinate with
-	 * xfs_sync().  This is because xfs_sync() references the inodes
-	 * in the mount list without taking references on the corresponding
-	 * vnodes.  We make that OK here by ensuring that we wait until
-	 * the inode is unlocked in xfs_sync() before we go ahead and
-	 * free it.  We get both the regular lock and the io lock because
-	 * the xfs_sync() code may need to drop the regular one but will
-	 * still hold the io lock.
-	 */
-	xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
-
-	/*
-	 * Release dquots (and their references) if any. An inode may escape
-	 * xfs_inactive and get here via vn_alloc->vn_reclaim path.
-	 */
-	XFS_QM_DQDETACH(ip->i_mount, ip);
-
-	/*
-	 * Pull our behavior descriptor from the vnode chain.
-	 */
-	if (ip->i_vnode) {
-		ip->i_vnode->i_private = NULL;
-		ip->i_vnode = NULL;
-	}
+	XFS_STATS_INC(xs_ig_reclaims);
 
 	/*
-	 * Free all memory associated with the inode.
+	 * Remove the inode from the per-AG radix tree.  It doesn't matter
+	 * if it was never added to it because radix_tree_delete can deal
+	 * with that case just fine.
 	 */
-	xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
-	xfs_idestroy(ip);
-}
-
-/*
- * This routine removes an about-to-be-destroyed inode from
- * all of the lists in which it is located with the exception
- * of the behavior chain.
- */
-void
-xfs_iextract(
-	xfs_inode_t	*ip)
-{
-	xfs_mount_t	*mp = ip->i_mount;
-	xfs_perag_t	*pag = xfs_get_perag(mp, ip->i_ino);
-	xfs_inode_t	*iq;
-
+	pag = xfs_get_perag(mp, ip->i_ino);
 	write_lock(&pag->pag_ici_lock);
 	radix_tree_delete(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ip->i_ino));
 	write_unlock(&pag->pag_ici_lock);
 	xfs_put_perag(mp, pag);
 
 	/*
-	 * Remove from mount's inode list.
+	 * Here we do an (almost) spurious inode lock in order to coordinate
+	 * with inode cache radix tree lookups.  This is because the lookup
+	 * can reference the inodes in the cache without taking references.
+	 *
+	 * We make that OK here by ensuring that we wait until the inode is
+	 * unlocked after the lookup before we go ahead and free it.  We get
+	 * both the ilock and the iolock because the code may need to drop the
+	 * ilock one but will still hold the iolock.
 	 */
-	XFS_MOUNT_ILOCK(mp);
-	ASSERT((ip->i_mnext != NULL) && (ip->i_mprev != NULL));
-	iq = ip->i_mnext;
-	iq->i_mprev = ip->i_mprev;
-	ip->i_mprev->i_mnext = iq;
-
+	xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
 	/*
-	 * Fix up the head pointer if it points to the inode being deleted.
+	 * Release dquots (and their references) if any.
 	 */
-	if (mp->m_inodes == ip) {
-		if (ip == iq) {
-			mp->m_inodes = NULL;
-		} else {
-			mp->m_inodes = iq;
-		}
+	XFS_QM_DQDETACH(ip->i_mount, ip);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+
+	switch (ip->i_d.di_mode & S_IFMT) {
+	case S_IFREG:
+	case S_IFDIR:
+	case S_IFLNK:
+		xfs_idestroy_fork(ip, XFS_DATA_FORK);
+		break;
 	}
 
-	/* Deal with the deleted inodes list */
-	list_del_init(&ip->i_reclaim);
+	if (ip->i_afp)
+		xfs_idestroy_fork(ip, XFS_ATTR_FORK);
 
-	mp->m_ireclaims++;
-	XFS_MOUNT_IUNLOCK(mp);
+#ifdef XFS_INODE_TRACE
+	ktrace_free(ip->i_trace);
+#endif
+#ifdef XFS_BMAP_TRACE
+	ktrace_free(ip->i_xtrace);
+#endif
+#ifdef XFS_BTREE_TRACE
+	ktrace_free(ip->i_btrace);
+#endif
+#ifdef XFS_RW_TRACE
+	ktrace_free(ip->i_rwtrace);
+#endif
+#ifdef XFS_ILOCK_TRACE
+	ktrace_free(ip->i_lock_trace);
+#endif
+#ifdef XFS_DIR2_TRACE
+	ktrace_free(ip->i_dir_trace);
+#endif
+	if (ip->i_itemp) {
+		/*
+		 * Only if we are shutting down the fs will we see an
+		 * inode still in the AIL. If it is there, we should remove
+		 * it to prevent a use-after-free from occurring.
+		 */
+		xfs_log_item_t	*lip = &ip->i_itemp->ili_item;
+		struct xfs_ail	*ailp = lip->li_ailp;
+
+		ASSERT(((lip->li_flags & XFS_LI_IN_AIL) == 0) ||
+				       XFS_FORCED_SHUTDOWN(ip->i_mount));
+		if (lip->li_flags & XFS_LI_IN_AIL) {
+			spin_lock(&ailp->xa_lock);
+			if (lip->li_flags & XFS_LI_IN_AIL)
+				xfs_trans_ail_delete(ailp, lip);
+			else
+				spin_unlock(&ailp->xa_lock);
+		}
+		xfs_inode_item_destroy(ip);
+		ip->i_itemp = NULL;
+	}
+	/* asserts to verify all state is correct here */
+	ASSERT(atomic_read(&ip->i_iocount) == 0);
+	ASSERT(atomic_read(&ip->i_pincount) == 0);
+	ASSERT(!spin_is_locked(&ip->i_flags_lock));
+	ASSERT(completion_done(&ip->i_flush));
+	kmem_zone_free(xfs_inode_zone, ip);
 }
 
 /*
@@ -737,7 +752,7 @@ xfs_iunlock(
 		 * it is in the AIL and anyone is waiting on it.  Don't do
 		 * this if the caller has asked us not to.
 		 */
-		xfs_trans_unlocked_item(ip->i_mount,
+		xfs_trans_unlocked_item(ip->i_itemp->ili_item.li_ailp,
 					(xfs_log_item_t*)(ip->i_itemp));
 	}
 	xfs_ilock_trace(ip, 3, lock_flags, (inst_t *)__return_address);
@@ -790,3 +805,51 @@ xfs_isilocked(
 }
 #endif
 
+#ifdef	XFS_INODE_TRACE
+
+#define KTRACE_ENTER(ip, vk, s, line, ra)			\
+	ktrace_enter((ip)->i_trace,				\
+/*  0 */		(void *)(__psint_t)(vk),		\
+/*  1 */		(void *)(s),				\
+/*  2 */		(void *)(__psint_t) line,		\
+/*  3 */		(void *)(__psint_t)atomic_read(&VFS_I(ip)->i_count), \
+/*  4 */		(void *)(ra),				\
+/*  5 */		NULL,					\
+/*  6 */		(void *)(__psint_t)current_cpu(),	\
+/*  7 */		(void *)(__psint_t)current_pid(),	\
+/*  8 */		(void *)__return_address,		\
+/*  9 */		NULL, NULL, NULL, NULL, NULL, NULL, NULL)
+
+/*
+ * Vnode tracing code.
+ */
+void
+_xfs_itrace_entry(xfs_inode_t *ip, const char *func, inst_t *ra)
+{
+	KTRACE_ENTER(ip, INODE_KTRACE_ENTRY, func, 0, ra);
+}
+
+void
+_xfs_itrace_exit(xfs_inode_t *ip, const char *func, inst_t *ra)
+{
+	KTRACE_ENTER(ip, INODE_KTRACE_EXIT, func, 0, ra);
+}
+
+void
+xfs_itrace_hold(xfs_inode_t *ip, char *file, int line, inst_t *ra)
+{
+	KTRACE_ENTER(ip, INODE_KTRACE_HOLD, file, line, ra);
+}
+
+void
+_xfs_itrace_ref(xfs_inode_t *ip, char *file, int line, inst_t *ra)
+{
+	KTRACE_ENTER(ip, INODE_KTRACE_REF, file, line, ra);
+}
+
+void
+xfs_itrace_rele(xfs_inode_t *ip, char *file, int line, inst_t *ra)
+{
+	KTRACE_ENTER(ip, INODE_KTRACE_RELE, file, line, ra);
+}
+#endif	/* XFS_INODE_TRACE */
diff --git a/fs/xfs/xfs_imap.h b/fs/xfs/xfs_imap.h
deleted file mode 100644
index d3645000398..00000000000
--- a/fs/xfs/xfs_imap.h
+++ /dev/null
@@ -1,40 +0,0 @@
-/*
- * Copyright (c) 2000,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_IMAP_H__
-#define	__XFS_IMAP_H__
-
-/*
- * This is the structure passed to xfs_imap() to map
- * an inode number to its on disk location.
- */
-typedef struct xfs_imap {
-	xfs_daddr_t	im_blkno;	/* starting BB of inode chunk */
-	uint		im_len;		/* length in BBs of inode chunk */
-	xfs_agblock_t	im_agblkno;	/* logical block of inode chunk in ag */
-	ushort		im_ioffset;	/* inode offset in block in "inodes" */
-	ushort		im_boffset;	/* inode offset in block in bytes */
-} xfs_imap_t;
-
-#ifdef __KERNEL__
-struct xfs_mount;
-struct xfs_trans;
-int	xfs_imap(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
-		 xfs_imap_t *, uint);
-#endif
-
-#endif	/* __XFS_IMAP_H__ */
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index a391b955df0..5a5e035e5d3 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -23,7 +23,6 @@
 #include "xfs_bit.h"
 #include "xfs_log.h"
 #include "xfs_inum.h"
-#include "xfs_imap.h"
 #include "xfs_trans.h"
 #include "xfs_trans_priv.h"
 #include "xfs_sb.h"
@@ -41,6 +40,7 @@
 #include "xfs_buf_item.h"
 #include "xfs_inode_item.h"
 #include "xfs_btree.h"
+#include "xfs_btree_trace.h"
 #include "xfs_alloc.h"
 #include "xfs_ialloc.h"
 #include "xfs_bmap.h"
@@ -133,10 +133,10 @@ STATIC int
 xfs_imap_to_bp(
 	xfs_mount_t	*mp,
 	xfs_trans_t	*tp,
-	xfs_imap_t	*imap,
+	struct xfs_imap	*imap,
 	xfs_buf_t	**bpp,
 	uint		buf_flags,
-	uint		imap_flags)
+	uint		iget_flags)
 {
 	int		error;
 	int		i;
@@ -173,12 +173,12 @@ xfs_imap_to_bp(
 
 		dip = (xfs_dinode_t *)xfs_buf_offset(bp,
 					(i << mp->m_sb.sb_inodelog));
-		di_ok = be16_to_cpu(dip->di_core.di_magic) == XFS_DINODE_MAGIC &&
-			    XFS_DINODE_GOOD_VERSION(dip->di_core.di_version);
+		di_ok = be16_to_cpu(dip->di_magic) == XFS_DINODE_MAGIC &&
+			    XFS_DINODE_GOOD_VERSION(dip->di_version);
 		if (unlikely(XFS_TEST_ERROR(!di_ok, mp,
 						XFS_ERRTAG_ITOBP_INOTOBP,
 						XFS_RANDOM_ITOBP_INOTOBP))) {
-			if (imap_flags & XFS_IMAP_BULKSTAT) {
+			if (iget_flags & XFS_IGET_BULKSTAT) {
 				xfs_trans_brelse(tp, bp);
 				return XFS_ERROR(EINVAL);
 			}
@@ -190,7 +190,7 @@ xfs_imap_to_bp(
 					"daddr %lld #%d (magic=%x)",
 				XFS_BUFTARG_NAME(mp->m_ddev_targp),
 				(unsigned long long)imap->im_blkno, i,
-				be16_to_cpu(dip->di_core.di_magic));
+				be16_to_cpu(dip->di_magic));
 #endif
 			xfs_trans_brelse(tp, bp);
 			return XFS_ERROR(EFSCORRUPTED);
@@ -221,25 +221,26 @@ xfs_imap_to_bp(
  * Use xfs_imap() to determine the size and location of the
  * buffer to read from disk.
  */
-STATIC int
+int
 xfs_inotobp(
 	xfs_mount_t	*mp,
 	xfs_trans_t	*tp,
 	xfs_ino_t	ino,
 	xfs_dinode_t	**dipp,
 	xfs_buf_t	**bpp,
-	int		*offset)
+	int		*offset,
+	uint		imap_flags)
 {
-	xfs_imap_t	imap;
+	struct xfs_imap	imap;
 	xfs_buf_t	*bp;
 	int		error;
 
 	imap.im_blkno = 0;
-	error = xfs_imap(mp, tp, ino, &imap, XFS_IMAP_LOOKUP);
+	error = xfs_imap(mp, tp, ino, &imap, imap_flags);
 	if (error)
 		return error;
 
-	error = xfs_imap_to_bp(mp, tp, &imap, &bp, XFS_BUF_LOCK, 0);
+	error = xfs_imap_to_bp(mp, tp, &imap, &bp, XFS_BUF_LOCK, imap_flags);
 	if (error)
 		return error;
 
@@ -260,15 +261,11 @@ xfs_inotobp(
  * If a non-zero error is returned, then the contents of bpp and
  * dipp are undefined.
  *
- * If the inode is new and has not yet been initialized, use xfs_imap()
- * to determine the size and location of the buffer to read from disk.
- * If the inode has already been mapped to its buffer and read in once,
- * then use the mapping information stored in the inode rather than
- * calling xfs_imap().  This allows us to avoid the overhead of looking
- * at the inode btree for small block file systems (see xfs_dilocate()).
- * We can tell whether the inode has been mapped in before by comparing
- * its disk block address to 0.  Only uninitialized inodes will have
- * 0 for the disk block address.
+ * The inode is expected to already been mapped to its buffer and read
+ * in once, thus we can use the mapping information stored in the inode
+ * rather than calling xfs_imap().  This allows us to avoid the overhead
+ * of looking at the inode btree for small block file systems
+ * (see xfs_imap()).
  */
 int
 xfs_itobp(
@@ -277,40 +274,14 @@ xfs_itobp(
 	xfs_inode_t	*ip,
 	xfs_dinode_t	**dipp,
 	xfs_buf_t	**bpp,
-	xfs_daddr_t	bno,
-	uint		imap_flags,
 	uint		buf_flags)
 {
-	xfs_imap_t	imap;
 	xfs_buf_t	*bp;
 	int		error;
 
-	if (ip->i_blkno == (xfs_daddr_t)0) {
-		imap.im_blkno = bno;
-		error = xfs_imap(mp, tp, ip->i_ino, &imap,
-					XFS_IMAP_LOOKUP | imap_flags);
-		if (error)
-			return error;
+	ASSERT(ip->i_imap.im_blkno != 0);
 
-		/*
-		 * Fill in the fields in the inode that will be used to
-		 * map the inode to its buffer from now on.
-		 */
-		ip->i_blkno = imap.im_blkno;
-		ip->i_len = imap.im_len;
-		ip->i_boffset = imap.im_boffset;
-	} else {
-		/*
-		 * We've already mapped the inode once, so just use the
-		 * mapping that we saved the first time.
-		 */
-		imap.im_blkno = ip->i_blkno;
-		imap.im_len = ip->i_len;
-		imap.im_boffset = ip->i_boffset;
-	}
-	ASSERT(bno == 0 || bno == imap.im_blkno);
-
-	error = xfs_imap_to_bp(mp, tp, &imap, &bp, buf_flags, imap_flags);
+	error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &bp, buf_flags, 0);
 	if (error)
 		return error;
 
@@ -321,7 +292,7 @@ xfs_itobp(
 		return EAGAIN;
 	}
 
-	*dipp = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset);
+	*dipp = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
 	*bpp = bp;
 	return 0;
 }
@@ -348,26 +319,26 @@ xfs_iformat(
 		XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
 	error = 0;
 
-	if (unlikely(be32_to_cpu(dip->di_core.di_nextents) +
-		     be16_to_cpu(dip->di_core.di_anextents) >
-		     be64_to_cpu(dip->di_core.di_nblocks))) {
+	if (unlikely(be32_to_cpu(dip->di_nextents) +
+		     be16_to_cpu(dip->di_anextents) >
+		     be64_to_cpu(dip->di_nblocks))) {
 		xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 			"corrupt dinode %Lu, extent total = %d, nblocks = %Lu.",
 			(unsigned long long)ip->i_ino,
-			(int)(be32_to_cpu(dip->di_core.di_nextents) +
-			      be16_to_cpu(dip->di_core.di_anextents)),
+			(int)(be32_to_cpu(dip->di_nextents) +
+			      be16_to_cpu(dip->di_anextents)),
 			(unsigned long long)
-				be64_to_cpu(dip->di_core.di_nblocks));
+				be64_to_cpu(dip->di_nblocks));
 		XFS_CORRUPTION_ERROR("xfs_iformat(1)", XFS_ERRLEVEL_LOW,
 				     ip->i_mount, dip);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
 
-	if (unlikely(dip->di_core.di_forkoff > ip->i_mount->m_sb.sb_inodesize)) {
+	if (unlikely(dip->di_forkoff > ip->i_mount->m_sb.sb_inodesize)) {
 		xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 			"corrupt dinode %Lu, forkoff = 0x%x.",
 			(unsigned long long)ip->i_ino,
-			dip->di_core.di_forkoff);
+			dip->di_forkoff);
 		XFS_CORRUPTION_ERROR("xfs_iformat(2)", XFS_ERRLEVEL_LOW,
 				     ip->i_mount, dip);
 		return XFS_ERROR(EFSCORRUPTED);
@@ -378,25 +349,25 @@ xfs_iformat(
 	case S_IFCHR:
 	case S_IFBLK:
 	case S_IFSOCK:
-		if (unlikely(dip->di_core.di_format != XFS_DINODE_FMT_DEV)) {
+		if (unlikely(dip->di_format != XFS_DINODE_FMT_DEV)) {
 			XFS_CORRUPTION_ERROR("xfs_iformat(3)", XFS_ERRLEVEL_LOW,
 					      ip->i_mount, dip);
 			return XFS_ERROR(EFSCORRUPTED);
 		}
 		ip->i_d.di_size = 0;
 		ip->i_size = 0;
-		ip->i_df.if_u2.if_rdev = be32_to_cpu(dip->di_u.di_dev);
+		ip->i_df.if_u2.if_rdev = xfs_dinode_get_rdev(dip);
 		break;
 
 	case S_IFREG:
 	case S_IFLNK:
 	case S_IFDIR:
-		switch (dip->di_core.di_format) {
+		switch (dip->di_format) {
 		case XFS_DINODE_FMT_LOCAL:
 			/*
 			 * no local regular files yet
 			 */
-			if (unlikely((be16_to_cpu(dip->di_core.di_mode) & S_IFMT) == S_IFREG)) {
+			if (unlikely((be16_to_cpu(dip->di_mode) & S_IFMT) == S_IFREG)) {
 				xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 					"corrupt inode %Lu "
 					"(local format for regular file).",
@@ -407,7 +378,7 @@ xfs_iformat(
 				return XFS_ERROR(EFSCORRUPTED);
 			}
 
-			di_size = be64_to_cpu(dip->di_core.di_size);
+			di_size = be64_to_cpu(dip->di_size);
 			if (unlikely(di_size > XFS_DFORK_DSIZE(dip, ip->i_mount))) {
 				xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount,
 					"corrupt inode %Lu "
@@ -449,7 +420,7 @@ xfs_iformat(
 	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP);
 	ip->i_afp->if_ext_max =
 		XFS_IFORK_ASIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
-	switch (dip->di_core.di_aformat) {
+	switch (dip->di_aformat) {
 	case XFS_DINODE_FMT_LOCAL:
 		atp = (xfs_attr_shortform_t *)XFS_DFORK_APTR(dip);
 		size = be16_to_cpu(atp->hdr.totsize);
@@ -621,7 +592,7 @@ xfs_iformat_btree(
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	dfp = (xfs_bmdr_block_t *)XFS_DFORK_PTR(dip, whichfork);
 	size = XFS_BMAP_BROOT_SPACE(dfp);
-	nrecs = XFS_BMAP_BROOT_NUMRECS(dfp);
+	nrecs = be16_to_cpu(dfp->bb_numrecs);
 
 	/*
 	 * blow out if -- fork has less extents than can fit in
@@ -649,8 +620,9 @@ xfs_iformat_btree(
 	 * Copy and convert from the on-disk structure
 	 * to the in-memory structure.
 	 */
-	xfs_bmdr_to_bmbt(dfp, XFS_DFORK_SIZE(dip, ip->i_mount, whichfork),
-		ifp->if_broot, size);
+	xfs_bmdr_to_bmbt(ip->i_mount, dfp,
+			 XFS_DFORK_SIZE(dip, ip->i_mount, whichfork),
+			 ifp->if_broot, size);
 	ifp->if_flags &= ~XFS_IFEXTENTS;
 	ifp->if_flags |= XFS_IFBROOT;
 
@@ -660,7 +632,7 @@ xfs_iformat_btree(
 void
 xfs_dinode_from_disk(
 	xfs_icdinode_t		*to,
-	xfs_dinode_core_t	*from)
+	xfs_dinode_t		*from)
 {
 	to->di_magic = be16_to_cpu(from->di_magic);
 	to->di_mode = be16_to_cpu(from->di_mode);
@@ -694,7 +666,7 @@ xfs_dinode_from_disk(
 
 void
 xfs_dinode_to_disk(
-	xfs_dinode_core_t	*to,
+	xfs_dinode_t		*to,
 	xfs_icdinode_t		*from)
 {
 	to->di_magic = cpu_to_be16(from->di_magic);
@@ -781,93 +753,57 @@ uint
 xfs_dic2xflags(
 	xfs_dinode_t		*dip)
 {
-	xfs_dinode_core_t	*dic = &dip->di_core;
-
-	return _xfs_dic2xflags(be16_to_cpu(dic->di_flags)) |
+	return _xfs_dic2xflags(be16_to_cpu(dip->di_flags)) |
 				(XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
 }
 
 /*
- * Given a mount structure and an inode number, return a pointer
- * to a newly allocated in-core inode corresponding to the given
- * inode number.
- *
- * Initialize the inode's attributes and extent pointers if it
- * already has them (it will not if the inode has no links).
+ * Read the disk inode attributes into the in-core inode structure.
  */
 int
 xfs_iread(
 	xfs_mount_t	*mp,
 	xfs_trans_t	*tp,
-	xfs_ino_t	ino,
-	xfs_inode_t	**ipp,
+	xfs_inode_t	*ip,
 	xfs_daddr_t	bno,
-	uint		imap_flags)
+	uint		iget_flags)
 {
 	xfs_buf_t	*bp;
 	xfs_dinode_t	*dip;
-	xfs_inode_t	*ip;
 	int		error;
 
-	ASSERT(xfs_inode_zone != NULL);
-
-	ip = kmem_zone_zalloc(xfs_inode_zone, KM_SLEEP);
-	ip->i_ino = ino;
-	ip->i_mount = mp;
-	atomic_set(&ip->i_iocount, 0);
-	spin_lock_init(&ip->i_flags_lock);
-
 	/*
-	 * Get pointer's to the on-disk inode and the buffer containing it.
-	 * If the inode number refers to a block outside the file system
-	 * then xfs_itobp() will return NULL.  In this case we should
-	 * return NULL as well.  Set i_blkno to 0 so that xfs_itobp() will
-	 * know that this is a new incore inode.
+	 * Fill in the location information in the in-core inode.
 	 */
-	error = xfs_itobp(mp, tp, ip, &dip, &bp, bno, imap_flags, XFS_BUF_LOCK);
-	if (error) {
-		kmem_zone_free(xfs_inode_zone, ip);
+	ip->i_imap.im_blkno = bno;
+	error = xfs_imap(mp, tp, ip->i_ino, &ip->i_imap, iget_flags);
+	if (error)
 		return error;
-	}
+	ASSERT(bno == 0 || bno == ip->i_imap.im_blkno);
 
 	/*
-	 * Initialize inode's trace buffers.
-	 * Do this before xfs_iformat in case it adds entries.
+	 * Get pointers to the on-disk inode and the buffer containing it.
 	 */
-#ifdef	XFS_INODE_TRACE
-	ip->i_trace = ktrace_alloc(INODE_TRACE_SIZE, KM_NOFS);
-#endif
-#ifdef XFS_BMAP_TRACE
-	ip->i_xtrace = ktrace_alloc(XFS_BMAP_KTRACE_SIZE, KM_NOFS);
-#endif
-#ifdef XFS_BMBT_TRACE
-	ip->i_btrace = ktrace_alloc(XFS_BMBT_KTRACE_SIZE, KM_NOFS);
-#endif
-#ifdef XFS_RW_TRACE
-	ip->i_rwtrace = ktrace_alloc(XFS_RW_KTRACE_SIZE, KM_NOFS);
-#endif
-#ifdef XFS_ILOCK_TRACE
-	ip->i_lock_trace = ktrace_alloc(XFS_ILOCK_KTRACE_SIZE, KM_NOFS);
-#endif
-#ifdef XFS_DIR2_TRACE
-	ip->i_dir_trace = ktrace_alloc(XFS_DIR2_KTRACE_SIZE, KM_NOFS);
-#endif
+	error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &bp,
+			       XFS_BUF_LOCK, iget_flags);
+	if (error)
+		return error;
+	dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
 
 	/*
 	 * If we got something that isn't an inode it means someone
 	 * (nfs or dmi) has a stale handle.
 	 */
-	if (be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC) {
-		kmem_zone_free(xfs_inode_zone, ip);
-		xfs_trans_brelse(tp, bp);
+	if (be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC) {
 #ifdef DEBUG
 		xfs_fs_cmn_err(CE_ALERT, mp, "xfs_iread: "
-				"dip->di_core.di_magic (0x%x) != "
+				"dip->di_magic (0x%x) != "
 				"XFS_DINODE_MAGIC (0x%x)",
-				be16_to_cpu(dip->di_core.di_magic),
+				be16_to_cpu(dip->di_magic),
 				XFS_DINODE_MAGIC);
 #endif /* DEBUG */
-		return XFS_ERROR(EINVAL);
+		error = XFS_ERROR(EINVAL);
+		goto out_brelse;
 	}
 
 	/*
@@ -877,24 +813,22 @@ xfs_iread(
 	 * specific information.
 	 * Otherwise, just get the truly permanent information.
 	 */
-	if (dip->di_core.di_mode) {
-		xfs_dinode_from_disk(&ip->i_d, &dip->di_core);
+	if (dip->di_mode) {
+		xfs_dinode_from_disk(&ip->i_d, dip);
 		error = xfs_iformat(ip, dip);
 		if (error)  {
-			kmem_zone_free(xfs_inode_zone, ip);
-			xfs_trans_brelse(tp, bp);
 #ifdef DEBUG
 			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_iread: "
 					"xfs_iformat() returned error %d",
 					error);
 #endif /* DEBUG */
-			return error;
+			goto out_brelse;
 		}
 	} else {
-		ip->i_d.di_magic = be16_to_cpu(dip->di_core.di_magic);
-		ip->i_d.di_version = dip->di_core.di_version;
-		ip->i_d.di_gen = be32_to_cpu(dip->di_core.di_gen);
-		ip->i_d.di_flushiter = be16_to_cpu(dip->di_core.di_flushiter);
+		ip->i_d.di_magic = be16_to_cpu(dip->di_magic);
+		ip->i_d.di_version = dip->di_version;
+		ip->i_d.di_gen = be32_to_cpu(dip->di_gen);
+		ip->i_d.di_flushiter = be16_to_cpu(dip->di_flushiter);
 		/*
 		 * Make sure to pull in the mode here as well in
 		 * case the inode is released without being used.
@@ -911,8 +845,6 @@ xfs_iread(
 			XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t);
 	}
 
-	INIT_LIST_HEAD(&ip->i_reclaim);
-
 	/*
 	 * The inode format changed when we moved the link count and
 	 * made it 32 bits long.  If this is an old format inode,
@@ -924,7 +856,7 @@ xfs_iread(
 	 * the new format. We don't change the version number so that we
 	 * can distinguish this from a real new format inode.
 	 */
-	if (ip->i_d.di_version == XFS_DINODE_VERSION_1) {
+	if (ip->i_d.di_version == 1) {
 		ip->i_d.di_nlink = ip->i_d.di_onlink;
 		ip->i_d.di_onlink = 0;
 		ip->i_d.di_projid = 0;
@@ -938,7 +870,7 @@ xfs_iread(
 	 * around for a while.  This helps to keep recently accessed
 	 * meta-data in-core longer.
 	 */
-	 XFS_BUF_SET_REF(bp, XFS_INO_REF);
+	XFS_BUF_SET_REF(bp, XFS_INO_REF);
 
 	/*
 	 * Use xfs_trans_brelse() to release the buffer containing the
@@ -953,9 +885,9 @@ xfs_iread(
 	 * to worry about the inode being changed just because we released
 	 * the buffer.
 	 */
+ out_brelse:
 	xfs_trans_brelse(tp, bp);
-	*ipp = ip;
-	return 0;
+	return error;
 }
 
 /*
@@ -1049,6 +981,7 @@ xfs_ialloc(
 	uint		flags;
 	int		error;
 	timespec_t	tv;
+	int		filestreams = 0;
 
 	/*
 	 * Call the space management code to pick
@@ -1056,9 +989,8 @@ xfs_ialloc(
 	 */
 	error = xfs_dialloc(tp, pip ? pip->i_ino : 0, mode, okalloc,
 			    ialloc_context, call_again, &ino);
-	if (error != 0) {
+	if (error)
 		return error;
-	}
 	if (*call_again || ino == NULLFSINO) {
 		*ipp = NULL;
 		return 0;
@@ -1072,9 +1004,8 @@ xfs_ialloc(
 	 */
 	error = xfs_trans_iget(tp->t_mountp, tp, ino,
 				XFS_IGET_CREATE, XFS_ILOCK_EXCL, &ip);
-	if (error != 0) {
+	if (error)
 		return error;
-	}
 	ASSERT(ip != NULL);
 
 	ip->i_d.di_mode = (__uint16_t)mode;
@@ -1093,8 +1024,8 @@ xfs_ialloc(
 	 * here rather than here and in the flush/logging code.
 	 */
 	if (xfs_sb_version_hasnlink(&tp->t_mountp->m_sb) &&
-	    ip->i_d.di_version == XFS_DINODE_VERSION_1) {
-		ip->i_d.di_version = XFS_DINODE_VERSION_2;
+	    ip->i_d.di_version == 1) {
+		ip->i_d.di_version = 2;
 		/*
 		 * We've already zeroed the old link count, the projid field,
 		 * and the pad field.
@@ -1104,7 +1035,7 @@ xfs_ialloc(
 	/*
 	 * Project ids won't be stored on disk if we are using a version 1 inode.
 	 */
-	if ((prid != 0) && (ip->i_d.di_version == XFS_DINODE_VERSION_1))
+	if ((prid != 0) && (ip->i_d.di_version == 1))
 		xfs_bump_ino_vers2(tp, ip);
 
 	if (pip && XFS_INHERIT_GID(pip)) {
@@ -1155,13 +1086,12 @@ xfs_ialloc(
 		flags |= XFS_ILOG_DEV;
 		break;
 	case S_IFREG:
-		if (pip && xfs_inode_is_filestream(pip)) {
-			error = xfs_filestream_associate(pip, ip);
-			if (error < 0)
-				return -error;
-			if (!error)
-				xfs_iflags_set(ip, XFS_IFILESTREAM);
-		}
+		/*
+		 * we can't set up filestreams until after the VFS inode
+		 * is set up properly.
+		 */
+		if (pip && xfs_inode_is_filestream(pip))
+			filestreams = 1;
 		/* fall through */
 	case S_IFDIR:
 		if (pip && (pip->i_d.di_flags & XFS_DIFLAG_ANY)) {
@@ -1227,6 +1157,15 @@ xfs_ialloc(
 	/* now that we have an i_mode we can setup inode ops and unlock */
 	xfs_setup_inode(ip);
 
+	/* now we have set up the vfs inode we can associate the filestream */
+	if (filestreams) {
+		error = xfs_filestream_associate(pip, ip);
+		if (error < 0)
+			return -error;
+		if (!error)
+			xfs_iflags_set(ip, XFS_IFILESTREAM);
+	}
+
 	*ipp = ip;
 	return 0;
 }
@@ -1383,8 +1322,8 @@ xfs_itrunc_trace(
  * direct I/O with the truncate operation.  Also, because we hold
  * the IOLOCK in exclusive mode, we prevent new direct I/Os from being
  * started until the truncate completes and drops the lock. Essentially,
- * the vn_iowait() call forms an I/O barrier that provides strict ordering
- * between direct I/Os and the truncate operation.
+ * the xfs_ioend_wait() call forms an I/O barrier that provides strict
+ * ordering between direct I/Os and the truncate operation.
  *
  * The flags parameter can have either the value XFS_ITRUNC_DEFINITE
  * or XFS_ITRUNC_MAYBE.  The XFS_ITRUNC_MAYBE value should be used
@@ -1415,7 +1354,7 @@ xfs_itruncate_start(
 
 	/* wait for the completion of any pending DIOs */
 	if (new_size == 0 || new_size < ip->i_size)
-		vn_iowait(ip);
+		xfs_ioend_wait(ip);
 
 	/*
 	 * Call toss_pages or flushinval_pages to get rid of pages
@@ -1726,8 +1665,14 @@ xfs_itruncate_finish(
 		xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
 		xfs_trans_ihold(ntp, ip);
 
-		if (!error)
-			error = xfs_trans_reserve(ntp, 0,
+		if (error)
+			return error;
+		/*
+		 * transaction commit worked ok so we can drop the extra ticket
+		 * reference that we gained in xfs_trans_dup()
+		 */
+		xfs_log_ticket_put(ntp->t_ticket);
+		error = xfs_trans_reserve(ntp, 0,
 					XFS_ITRUNCATE_LOG_RES(mp), 0,
 					XFS_TRANS_PERM_LOG_RES,
 					XFS_ITRUNCATE_LOG_COUNT);
@@ -1781,13 +1726,10 @@ xfs_iunlink(
 	xfs_dinode_t	*dip;
 	xfs_buf_t	*agibp;
 	xfs_buf_t	*ibp;
-	xfs_agnumber_t	agno;
-	xfs_daddr_t	agdaddr;
 	xfs_agino_t	agino;
 	short		bucket_index;
 	int		offset;
 	int		error;
-	int		agi_ok;
 
 	ASSERT(ip->i_d.di_nlink == 0);
 	ASSERT(ip->i_d.di_mode != 0);
@@ -1795,31 +1737,15 @@ xfs_iunlink(
 
 	mp = tp->t_mountp;
 
-	agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
-	agdaddr = XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp));
-
 	/*
 	 * Get the agi buffer first.  It ensures lock ordering
 	 * on the list.
 	 */
-	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, agdaddr,
-				   XFS_FSS_TO_BB(mp, 1), 0, &agibp);
+	error = xfs_read_agi(mp, tp, XFS_INO_TO_AGNO(mp, ip->i_ino), &agibp);
 	if (error)
 		return error;
-
-	/*
-	 * Validate the magic number of the agi block.
-	 */
 	agi = XFS_BUF_TO_AGI(agibp);
-	agi_ok =
-		be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
-		XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum));
-	if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IUNLINK,
-			XFS_RANDOM_IUNLINK))) {
-		XFS_CORRUPTION_ERROR("xfs_iunlink", XFS_ERRLEVEL_LOW, mp, agi);
-		xfs_trans_brelse(tp, agibp);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
+
 	/*
 	 * Get the index into the agi hash table for the
 	 * list this inode will go on.
@@ -1837,14 +1763,14 @@ xfs_iunlink(
 		 * Here we put the head pointer into our next pointer,
 		 * and then we fall through to point the head at us.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, XFS_BUF_LOCK);
 		if (error)
 			return error;
 
 		ASSERT(be32_to_cpu(dip->di_next_unlinked) == NULLAGINO);
 		/* both on-disk, don't endian flip twice */
 		dip->di_next_unlinked = agi->agi_unlinked[bucket_index];
-		offset = ip->i_boffset +
+		offset = ip->i_imap.im_boffset +
 			offsetof(xfs_dinode_t, di_next_unlinked);
 		xfs_trans_inode_buf(tp, ibp);
 		xfs_trans_log_buf(tp, ibp, offset,
@@ -1879,7 +1805,6 @@ xfs_iunlink_remove(
 	xfs_buf_t	*agibp;
 	xfs_buf_t	*ibp;
 	xfs_agnumber_t	agno;
-	xfs_daddr_t	agdaddr;
 	xfs_agino_t	agino;
 	xfs_agino_t	next_agino;
 	xfs_buf_t	*last_ibp;
@@ -1887,45 +1812,20 @@ xfs_iunlink_remove(
 	short		bucket_index;
 	int		offset, last_offset = 0;
 	int		error;
-	int		agi_ok;
 
-	/*
-	 * First pull the on-disk inode from the AGI unlinked list.
-	 */
 	mp = tp->t_mountp;
-
 	agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
-	agdaddr = XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp));
 
 	/*
 	 * Get the agi buffer first.  It ensures lock ordering
 	 * on the list.
 	 */
-	error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, agdaddr,
-				   XFS_FSS_TO_BB(mp, 1), 0, &agibp);
-	if (error) {
-		cmn_err(CE_WARN,
-			"xfs_iunlink_remove: xfs_trans_read_buf()  returned an error %d on %s.  Returning error.",
-			error, mp->m_fsname);
+	error = xfs_read_agi(mp, tp, agno, &agibp);
+	if (error)
 		return error;
-	}
-	/*
-	 * Validate the magic number of the agi block.
-	 */
+
 	agi = XFS_BUF_TO_AGI(agibp);
-	agi_ok =
-		be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC &&
-		XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum));
-	if (unlikely(XFS_TEST_ERROR(!agi_ok, mp, XFS_ERRTAG_IUNLINK_REMOVE,
-			XFS_RANDOM_IUNLINK_REMOVE))) {
-		XFS_CORRUPTION_ERROR("xfs_iunlink_remove", XFS_ERRLEVEL_LOW,
-				     mp, agi);
-		xfs_trans_brelse(tp, agibp);
-		cmn_err(CE_WARN,
-			"xfs_iunlink_remove: XFS_TEST_ERROR()  returned an error on %s.  Returning EFSCORRUPTED.",
-			 mp->m_fsname);
-		return XFS_ERROR(EFSCORRUPTED);
-	}
+
 	/*
 	 * Get the index into the agi hash table for the
 	 * list this inode will go on.
@@ -1945,7 +1845,7 @@ xfs_iunlink_remove(
 		 * of dealing with the buffer when there is no need to
 		 * change it.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, XFS_BUF_LOCK);
 		if (error) {
 			cmn_err(CE_WARN,
 				"xfs_iunlink_remove: xfs_itobp()  returned an error %d on %s.  Returning error.",
@@ -1956,7 +1856,7 @@ xfs_iunlink_remove(
 		ASSERT(next_agino != 0);
 		if (next_agino != NULLAGINO) {
 			dip->di_next_unlinked = cpu_to_be32(NULLAGINO);
-			offset = ip->i_boffset +
+			offset = ip->i_imap.im_boffset +
 				offsetof(xfs_dinode_t, di_next_unlinked);
 			xfs_trans_inode_buf(tp, ibp);
 			xfs_trans_log_buf(tp, ibp, offset,
@@ -1992,7 +1892,7 @@ xfs_iunlink_remove(
 			}
 			next_ino = XFS_AGINO_TO_INO(mp, agno, next_agino);
 			error = xfs_inotobp(mp, tp, next_ino, &last_dip,
-					    &last_ibp, &last_offset);
+					    &last_ibp, &last_offset, 0);
 			if (error) {
 				cmn_err(CE_WARN,
 			"xfs_iunlink_remove: xfs_inotobp()  returned an error %d on %s.  Returning error.",
@@ -2007,7 +1907,7 @@ xfs_iunlink_remove(
 		 * Now last_ibp points to the buffer previous to us on
 		 * the unlinked list.  Pull us from the list.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, XFS_BUF_LOCK);
 		if (error) {
 			cmn_err(CE_WARN,
 				"xfs_iunlink_remove: xfs_itobp()  returned an error %d on %s.  Returning error.",
@@ -2019,7 +1919,7 @@ xfs_iunlink_remove(
 		ASSERT(next_agino != agino);
 		if (next_agino != NULLAGINO) {
 			dip->di_next_unlinked = cpu_to_be32(NULLAGINO);
-			offset = ip->i_boffset +
+			offset = ip->i_imap.im_boffset +
 				offsetof(xfs_dinode_t, di_next_unlinked);
 			xfs_trans_inode_buf(tp, ibp);
 			xfs_trans_log_buf(tp, ibp, offset,
@@ -2160,9 +2060,9 @@ xfs_ifree_cluster(
 				iip = (xfs_inode_log_item_t *)lip;
 				ASSERT(iip->ili_logged == 1);
 				lip->li_cb = (void(*)(xfs_buf_t*,xfs_log_item_t*)) xfs_istale_done;
-				spin_lock(&mp->m_ail_lock);
-				iip->ili_flush_lsn = iip->ili_item.li_lsn;
-				spin_unlock(&mp->m_ail_lock);
+				xfs_trans_ail_copy_lsn(mp->m_ail,
+							&iip->ili_flush_lsn,
+							&iip->ili_item.li_lsn);
 				xfs_iflags_set(iip->ili_inode, XFS_ISTALE);
 				pre_flushed++;
 			}
@@ -2183,9 +2083,8 @@ xfs_ifree_cluster(
 			iip->ili_last_fields = iip->ili_format.ilf_fields;
 			iip->ili_format.ilf_fields = 0;
 			iip->ili_logged = 1;
-			spin_lock(&mp->m_ail_lock);
-			iip->ili_flush_lsn = iip->ili_item.li_lsn;
-			spin_unlock(&mp->m_ail_lock);
+			xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
+						&iip->ili_item.li_lsn);
 
 			xfs_buf_attach_iodone(bp,
 				(void(*)(xfs_buf_t*,xfs_log_item_t*))
@@ -2263,7 +2162,7 @@ xfs_ifree(
 
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
-	error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK);
+	error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, XFS_BUF_LOCK);
 	if (error)
 		return error;
 
@@ -2279,7 +2178,7 @@ xfs_ifree(
 	* This is a temporary hack that would require a proper fix
 	* in the future.
 	*/
-	dip->di_core.di_mode = 0;
+	dip->di_mode = 0;
 
 	if (delete) {
 		xfs_ifree_cluster(ip, tp, first_ino);
@@ -2312,9 +2211,10 @@ xfs_iroot_realloc(
 	int			rec_diff,
 	int			whichfork)
 {
+	struct xfs_mount	*mp = ip->i_mount;
 	int			cur_max;
 	xfs_ifork_t		*ifp;
-	xfs_bmbt_block_t	*new_broot;
+	struct xfs_btree_block	*new_broot;
 	int			new_max;
 	size_t			new_size;
 	char			*np;
@@ -2335,8 +2235,7 @@ xfs_iroot_realloc(
 		 */
 		if (ifp->if_broot_bytes == 0) {
 			new_size = (size_t)XFS_BMAP_BROOT_SPACE_CALC(rec_diff);
-			ifp->if_broot = (xfs_bmbt_block_t*)kmem_alloc(new_size,
-								     KM_SLEEP);
+			ifp->if_broot = kmem_alloc(new_size, KM_SLEEP);
 			ifp->if_broot_bytes = (int)new_size;
 			return;
 		}
@@ -2347,18 +2246,16 @@ xfs_iroot_realloc(
 		 * location.  The records don't change location because
 		 * they are kept butted up against the btree block header.
 		 */
-		cur_max = XFS_BMAP_BROOT_MAXRECS(ifp->if_broot_bytes);
+		cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0);
 		new_max = cur_max + rec_diff;
 		new_size = (size_t)XFS_BMAP_BROOT_SPACE_CALC(new_max);
-		ifp->if_broot = (xfs_bmbt_block_t *)
-		  kmem_realloc(ifp->if_broot,
-				new_size,
+		ifp->if_broot = kmem_realloc(ifp->if_broot, new_size,
 				(size_t)XFS_BMAP_BROOT_SPACE_CALC(cur_max), /* old size */
 				KM_SLEEP);
-		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(ifp->if_broot, 1,
-						      ifp->if_broot_bytes);
-		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(ifp->if_broot, 1,
-						      (int)new_size);
+		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
+						     ifp->if_broot_bytes);
+		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
+						     (int)new_size);
 		ifp->if_broot_bytes = (int)new_size;
 		ASSERT(ifp->if_broot_bytes <=
 			XFS_IFORK_SIZE(ip, whichfork) + XFS_BROOT_SIZE_ADJ);
@@ -2372,7 +2269,7 @@ xfs_iroot_realloc(
 	 * records, just get rid of the root and clear the status bit.
 	 */
 	ASSERT((ifp->if_broot != NULL) && (ifp->if_broot_bytes > 0));
-	cur_max = XFS_BMAP_BROOT_MAXRECS(ifp->if_broot_bytes);
+	cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0);
 	new_max = cur_max + rec_diff;
 	ASSERT(new_max >= 0);
 	if (new_max > 0)
@@ -2380,11 +2277,11 @@ xfs_iroot_realloc(
 	else
 		new_size = 0;
 	if (new_size > 0) {
-		new_broot = (xfs_bmbt_block_t *)kmem_alloc(new_size, KM_SLEEP);
+		new_broot = kmem_alloc(new_size, KM_SLEEP);
 		/*
 		 * First copy over the btree block header.
 		 */
-		memcpy(new_broot, ifp->if_broot, sizeof(xfs_bmbt_block_t));
+		memcpy(new_broot, ifp->if_broot, XFS_BTREE_LBLOCK_LEN);
 	} else {
 		new_broot = NULL;
 		ifp->if_flags &= ~XFS_IFBROOT;
@@ -2397,18 +2294,16 @@ xfs_iroot_realloc(
 		/*
 		 * First copy the records.
 		 */
-		op = (char *)XFS_BMAP_BROOT_REC_ADDR(ifp->if_broot, 1,
-						     ifp->if_broot_bytes);
-		np = (char *)XFS_BMAP_BROOT_REC_ADDR(new_broot, 1,
-						     (int)new_size);
+		op = (char *)XFS_BMBT_REC_ADDR(mp, ifp->if_broot, 1);
+		np = (char *)XFS_BMBT_REC_ADDR(mp, new_broot, 1);
 		memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_rec_t));
 
 		/*
 		 * Then copy the pointers.
 		 */
-		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(ifp->if_broot, 1,
+		op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1,
 						     ifp->if_broot_bytes);
-		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(new_broot, 1,
+		np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, new_broot, 1,
 						     (int)new_size);
 		memcpy(np, op, new_max * (uint)sizeof(xfs_dfsbno_t));
 	}
@@ -2511,64 +2406,6 @@ xfs_idata_realloc(
 	ASSERT(ifp->if_bytes <= XFS_IFORK_SIZE(ip, whichfork));
 }
 
-
-
-
-/*
- * Map inode to disk block and offset.
- *
- * mp -- the mount point structure for the current file system
- * tp -- the current transaction
- * ino -- the inode number of the inode to be located
- * imap -- this structure is filled in with the information necessary
- *	 to retrieve the given inode from disk
- * flags -- flags to pass to xfs_dilocate indicating whether or not
- *	 lookups in the inode btree were OK or not
- */
-int
-xfs_imap(
-	xfs_mount_t	*mp,
-	xfs_trans_t	*tp,
-	xfs_ino_t	ino,
-	xfs_imap_t	*imap,
-	uint		flags)
-{
-	xfs_fsblock_t	fsbno;
-	int		len;
-	int		off;
-	int		error;
-
-	fsbno = imap->im_blkno ?
-		XFS_DADDR_TO_FSB(mp, imap->im_blkno) : NULLFSBLOCK;
-	error = xfs_dilocate(mp, tp, ino, &fsbno, &len, &off, flags);
-	if (error)
-		return error;
-
-	imap->im_blkno = XFS_FSB_TO_DADDR(mp, fsbno);
-	imap->im_len = XFS_FSB_TO_BB(mp, len);
-	imap->im_agblkno = XFS_FSB_TO_AGBNO(mp, fsbno);
-	imap->im_ioffset = (ushort)off;
-	imap->im_boffset = (ushort)(off << mp->m_sb.sb_inodelog);
-
-	/*
-	 * If the inode number maps to a block outside the bounds
-	 * of the file system then return NULL rather than calling
-	 * read_buf and panicing when we get an error from the
-	 * driver.
-	 */
-	if ((imap->im_blkno + imap->im_len) >
-	    XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) {
-		xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
-			"(imap->im_blkno (0x%llx) + imap->im_len (0x%llx)) > "
-			" XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks) (0x%llx)",
-			(unsigned long long) imap->im_blkno,
-			(unsigned long long) imap->im_len,
-			XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks));
-		return EINVAL;
-	}
-	return 0;
-}
-
 void
 xfs_idestroy_fork(
 	xfs_inode_t	*ip,
@@ -2613,70 +2450,6 @@ xfs_idestroy_fork(
 }
 
 /*
- * This is called free all the memory associated with an inode.
- * It must free the inode itself and any buffers allocated for
- * if_extents/if_data and if_broot.  It must also free the lock
- * associated with the inode.
- */
-void
-xfs_idestroy(
-	xfs_inode_t	*ip)
-{
-	switch (ip->i_d.di_mode & S_IFMT) {
-	case S_IFREG:
-	case S_IFDIR:
-	case S_IFLNK:
-		xfs_idestroy_fork(ip, XFS_DATA_FORK);
-		break;
-	}
-	if (ip->i_afp)
-		xfs_idestroy_fork(ip, XFS_ATTR_FORK);
-	mrfree(&ip->i_lock);
-	mrfree(&ip->i_iolock);
-
-#ifdef XFS_INODE_TRACE
-	ktrace_free(ip->i_trace);
-#endif
-#ifdef XFS_BMAP_TRACE
-	ktrace_free(ip->i_xtrace);
-#endif
-#ifdef XFS_BMBT_TRACE
-	ktrace_free(ip->i_btrace);
-#endif
-#ifdef XFS_RW_TRACE
-	ktrace_free(ip->i_rwtrace);
-#endif
-#ifdef XFS_ILOCK_TRACE
-	ktrace_free(ip->i_lock_trace);
-#endif
-#ifdef XFS_DIR2_TRACE
-	ktrace_free(ip->i_dir_trace);
-#endif
-	if (ip->i_itemp) {
-		/*
-		 * Only if we are shutting down the fs will we see an
-		 * inode still in the AIL. If it is there, we should remove
-		 * it to prevent a use-after-free from occurring.
-		 */
-		xfs_mount_t	*mp = ip->i_mount;
-		xfs_log_item_t	*lip = &ip->i_itemp->ili_item;
-
-		ASSERT(((lip->li_flags & XFS_LI_IN_AIL) == 0) ||
-				       XFS_FORCED_SHUTDOWN(ip->i_mount));
-		if (lip->li_flags & XFS_LI_IN_AIL) {
-			spin_lock(&mp->m_ail_lock);
-			if (lip->li_flags & XFS_LI_IN_AIL)
-				xfs_trans_delete_ail(mp, lip);
-			else
-				spin_unlock(&mp->m_ail_lock);
-		}
-		xfs_inode_item_destroy(ip);
-	}
-	kmem_zone_free(xfs_inode_zone, ip);
-}
-
-
-/*
  * Increment the pin count of the given buffer.
  * This value is protected by ipinlock spinlock in the mount structure.
  */
@@ -2880,7 +2653,7 @@ xfs_iflush_fork(
 			ASSERT(ifp->if_broot_bytes <=
 			       (XFS_IFORK_SIZE(ip, whichfork) +
 				XFS_BROOT_SIZE_ADJ));
-			xfs_bmbt_to_bmdr(ifp->if_broot, ifp->if_broot_bytes,
+			xfs_bmbt_to_bmdr(mp, ifp->if_broot, ifp->if_broot_bytes,
 				(xfs_bmdr_block_t *)cp,
 				XFS_DFORK_SIZE(dip, mp, whichfork));
 		}
@@ -2889,15 +2662,16 @@ xfs_iflush_fork(
 	case XFS_DINODE_FMT_DEV:
 		if (iip->ili_format.ilf_fields & XFS_ILOG_DEV) {
 			ASSERT(whichfork == XFS_DATA_FORK);
-			dip->di_u.di_dev = cpu_to_be32(ip->i_df.if_u2.if_rdev);
+			xfs_dinode_put_rdev(dip, ip->i_df.if_u2.if_rdev);
 		}
 		break;
 
 	case XFS_DINODE_FMT_UUID:
 		if (iip->ili_format.ilf_fields & XFS_ILOG_UUID) {
 			ASSERT(whichfork == XFS_DATA_FORK);
-			memcpy(&dip->di_u.di_muuid, &ip->i_df.if_u2.if_uuid,
-				sizeof(uuid_t));
+			memcpy(XFS_DFORK_DPTR(dip),
+			       &ip->i_df.if_u2.if_uuid,
+			       sizeof(uuid_t));
 		}
 		break;
 
@@ -3030,7 +2804,6 @@ cluster_corrupt_out:
 			XFS_BUF_CLR_BDSTRAT_FUNC(bp);
 			XFS_BUF_UNDONE(bp);
 			XFS_BUF_STALE(bp);
-			XFS_BUF_SHUT(bp);
 			XFS_BUF_ERROR(bp,EIO);
 			xfs_biodone(bp);
 		} else {
@@ -3172,7 +2945,7 @@ xfs_iflush(
 	/*
 	 * Get the buffer containing the on-disk inode.
 	 */
-	error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0, 0,
+	error = xfs_itobp(mp, NULL, ip, &dip, &bp,
 				noblock ? XFS_BUF_TRYLOCK : XFS_BUF_LOCK);
 	if (error || !bp) {
 		xfs_ifunlock(ip);
@@ -3253,7 +3026,7 @@ xfs_iflush_int(
 	}
 
 	/* set *dip = inode's place in the buffer */
-	dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_boffset);
+	dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
 
 	/*
 	 * Clear i_update_core before copying out the data.
@@ -3275,11 +3048,11 @@ xfs_iflush_int(
 	 */
 	xfs_synchronize_atime(ip);
 
-	if (XFS_TEST_ERROR(be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC,
+	if (XFS_TEST_ERROR(be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC,
 			       mp, XFS_ERRTAG_IFLUSH_1, XFS_RANDOM_IFLUSH_1)) {
 		xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp,
 		    "xfs_iflush: Bad inode %Lu magic number 0x%x, ptr 0x%p",
-			ip->i_ino, be16_to_cpu(dip->di_core.di_magic), dip);
+			ip->i_ino, be16_to_cpu(dip->di_magic), dip);
 		goto corrupt_out;
 	}
 	if (XFS_TEST_ERROR(ip->i_d.di_magic != XFS_DINODE_MAGIC,
@@ -3342,7 +3115,7 @@ xfs_iflush_int(
 	 * because if the inode is dirty at all the core must
 	 * be.
 	 */
-	xfs_dinode_to_disk(&dip->di_core, &ip->i_d);
+	xfs_dinode_to_disk(dip, &ip->i_d);
 
 	/* Wrap, we never let the log put out DI_MAX_FLUSH */
 	if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
@@ -3354,28 +3127,27 @@ xfs_iflush_int(
 	 * convert back to the old inode format.  If the superblock version
 	 * has been updated, then make the conversion permanent.
 	 */
-	ASSERT(ip->i_d.di_version == XFS_DINODE_VERSION_1 ||
-	       xfs_sb_version_hasnlink(&mp->m_sb));
-	if (ip->i_d.di_version == XFS_DINODE_VERSION_1) {
+	ASSERT(ip->i_d.di_version == 1 || xfs_sb_version_hasnlink(&mp->m_sb));
+	if (ip->i_d.di_version == 1) {
 		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
 			/*
 			 * Convert it back.
 			 */
 			ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
-			dip->di_core.di_onlink = cpu_to_be16(ip->i_d.di_nlink);
+			dip->di_onlink = cpu_to_be16(ip->i_d.di_nlink);
 		} else {
 			/*
 			 * The superblock version has already been bumped,
 			 * so just make the conversion to the new inode
 			 * format permanent.
 			 */
-			ip->i_d.di_version = XFS_DINODE_VERSION_2;
-			dip->di_core.di_version =  XFS_DINODE_VERSION_2;
+			ip->i_d.di_version = 2;
+			dip->di_version = 2;
 			ip->i_d.di_onlink = 0;
-			dip->di_core.di_onlink = 0;
+			dip->di_onlink = 0;
 			memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
-			memset(&(dip->di_core.di_pad[0]), 0,
-			      sizeof(dip->di_core.di_pad));
+			memset(&(dip->di_pad[0]), 0,
+			      sizeof(dip->di_pad));
 			ASSERT(ip->i_d.di_projid == 0);
 		}
 	}
@@ -3418,10 +3190,8 @@ xfs_iflush_int(
 		iip->ili_format.ilf_fields = 0;
 		iip->ili_logged = 1;
 
-		ASSERT(sizeof(xfs_lsn_t) == 8);	/* don't lock if it shrinks */
-		spin_lock(&mp->m_ail_lock);
-		iip->ili_flush_lsn = iip->ili_item.li_lsn;
-		spin_unlock(&mp->m_ail_lock);
+		xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
+					&iip->ili_item.li_lsn);
 
 		/*
 		 * Attach the function xfs_iflush_done to the inode's
@@ -3459,45 +3229,8 @@ corrupt_out:
 }
 
 
-/*
- * Flush all inactive inodes in mp.
- */
-void
-xfs_iflush_all(
-	xfs_mount_t	*mp)
-{
-	xfs_inode_t	*ip;
-
- again:
-	XFS_MOUNT_ILOCK(mp);
-	ip = mp->m_inodes;
-	if (ip == NULL)
-		goto out;
-
-	do {
-		/* Make sure we skip markers inserted by sync */
-		if (ip->i_mount == NULL) {
-			ip = ip->i_mnext;
-			continue;
-		}
-
-		if (!VFS_I(ip)) {
-			XFS_MOUNT_IUNLOCK(mp);
-			xfs_finish_reclaim(ip, 0, XFS_IFLUSH_ASYNC);
-			goto again;
-		}
-
-		ASSERT(vn_count(VFS_I(ip)) == 0);
-
-		ip = ip->i_mnext;
-	} while (ip != mp->m_inodes);
- out:
-	XFS_MOUNT_IUNLOCK(mp);
-}
 
 #ifdef XFS_ILOCK_TRACE
-ktrace_t	*xfs_ilock_trace_buf;
-
 void
 xfs_ilock_trace(xfs_inode_t *ip, int lock, unsigned int lockflags, inst_t *ra)
 {
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 6be310d41da..1f175fa34b2 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -19,8 +19,7 @@
 #define	__XFS_INODE_H__
 
 struct xfs_dinode;
-struct xfs_dinode_core;
-
+struct xfs_inode;
 
 /*
  * Fork identifiers.
@@ -63,7 +62,7 @@ typedef struct xfs_ext_irec {
 typedef struct xfs_ifork {
 	int			if_bytes;	/* bytes in if_u1 */
 	int			if_real_bytes;	/* bytes allocated in if_u1 */
-	xfs_bmbt_block_t	*if_broot;	/* file's incore btree root */
+	struct xfs_btree_block	*if_broot;	/* file's incore btree root */
 	short			if_broot_bytes;	/* bytes allocated for root */
 	unsigned char		if_flags;	/* per-fork flags */
 	unsigned char		if_ext_max;	/* max # of extent records */
@@ -84,52 +83,14 @@ typedef struct xfs_ifork {
 } xfs_ifork_t;
 
 /*
- * Flags for xfs_ichgtime().
+ * Inode location information.  Stored in the inode and passed to
+ * xfs_imap_to_bp() to get a buffer and dinode for a given inode.
  */
-#define	XFS_ICHGTIME_MOD	0x1	/* data fork modification timestamp */
-#define	XFS_ICHGTIME_CHG	0x2	/* inode field change timestamp */
-
-/*
- * Per-fork incore inode flags.
- */
-#define	XFS_IFINLINE	0x01	/* Inline data is read in */
-#define	XFS_IFEXTENTS	0x02	/* All extent pointers are read in */
-#define	XFS_IFBROOT	0x04	/* i_broot points to the bmap b-tree root */
-#define	XFS_IFEXTIREC	0x08	/* Indirection array of extent blocks */
-
-/*
- * Flags for xfs_itobp(), xfs_imap() and xfs_dilocate().
- */
-#define XFS_IMAP_LOOKUP		0x1
-#define XFS_IMAP_BULKSTAT	0x2
-
-#ifdef __KERNEL__
-struct bhv_desc;
-struct cred;
-struct ktrace;
-struct xfs_buf;
-struct xfs_bmap_free;
-struct xfs_bmbt_irec;
-struct xfs_bmbt_block;
-struct xfs_inode;
-struct xfs_inode_log_item;
-struct xfs_mount;
-struct xfs_trans;
-struct xfs_dquot;
-
-#if defined(XFS_ILOCK_TRACE)
-#define XFS_ILOCK_KTRACE_SIZE	32
-extern ktrace_t *xfs_ilock_trace_buf;
-extern void xfs_ilock_trace(struct xfs_inode *, int, unsigned int, inst_t *);
-#else
-#define	xfs_ilock_trace(i,n,f,ra)
-#endif
-
-typedef struct dm_attrs_s {
-	__uint32_t	da_dmevmask;	/* DMIG event mask */
-	__uint16_t	da_dmstate;	/* DMIG state info */
-	__uint16_t	da_pad;		/* DMIG extra padding */
-} dm_attrs_t;
+struct xfs_imap {
+	xfs_daddr_t	im_blkno;	/* starting BB of inode chunk */
+	ushort		im_len;		/* length in BBs of inode chunk */
+	ushort		im_boffset;	/* inode offset in block in bytes */
+};
 
 /*
  * This is the xfs in-core inode structure.
@@ -160,7 +121,7 @@ typedef struct xfs_ictimestamp {
 } xfs_ictimestamp_t;
 
 /*
- * NOTE:  This structure must be kept identical to struct xfs_dinode_core
+ * NOTE:  This structure must be kept identical to struct xfs_dinode
  * 	  in xfs_dinode.h except for the endianess annotations.
  */
 typedef struct xfs_icdinode {
@@ -191,27 +152,97 @@ typedef struct xfs_icdinode {
 	__uint32_t	di_gen;		/* generation number */
 } xfs_icdinode_t;
 
-typedef struct {
-	struct xfs_inode	*ip_mnext;	/* next inode in mount list */
-	struct xfs_inode	*ip_mprev;	/* ptr to prev inode */
-	struct xfs_mount	*ip_mount;	/* fs mount struct ptr */
-} xfs_iptr_t;
+/*
+ * Flags for xfs_ichgtime().
+ */
+#define	XFS_ICHGTIME_MOD	0x1	/* data fork modification timestamp */
+#define	XFS_ICHGTIME_CHG	0x2	/* inode field change timestamp */
+
+/*
+ * Per-fork incore inode flags.
+ */
+#define	XFS_IFINLINE	0x01	/* Inline data is read in */
+#define	XFS_IFEXTENTS	0x02	/* All extent pointers are read in */
+#define	XFS_IFBROOT	0x04	/* i_broot points to the bmap b-tree root */
+#define	XFS_IFEXTIREC	0x08	/* Indirection array of extent blocks */
+
+/*
+ * Fork handling.
+ */
+
+#define XFS_IFORK_Q(ip)			((ip)->i_d.di_forkoff != 0)
+#define XFS_IFORK_BOFF(ip)		((int)((ip)->i_d.di_forkoff << 3))
+
+#define XFS_IFORK_PTR(ip,w)		\
+	((w) == XFS_DATA_FORK ? \
+		&(ip)->i_df : \
+		(ip)->i_afp)
+#define XFS_IFORK_DSIZE(ip) \
+	(XFS_IFORK_Q(ip) ? \
+		XFS_IFORK_BOFF(ip) : \
+		XFS_LITINO((ip)->i_mount))
+#define XFS_IFORK_ASIZE(ip) \
+	(XFS_IFORK_Q(ip) ? \
+		XFS_LITINO((ip)->i_mount) - XFS_IFORK_BOFF(ip) : \
+		0)
+#define XFS_IFORK_SIZE(ip,w) \
+	((w) == XFS_DATA_FORK ? \
+		XFS_IFORK_DSIZE(ip) : \
+		XFS_IFORK_ASIZE(ip))
+#define XFS_IFORK_FORMAT(ip,w) \
+	((w) == XFS_DATA_FORK ? \
+		(ip)->i_d.di_format : \
+		(ip)->i_d.di_aformat)
+#define XFS_IFORK_FMT_SET(ip,w,n) \
+	((w) == XFS_DATA_FORK ? \
+		((ip)->i_d.di_format = (n)) : \
+		((ip)->i_d.di_aformat = (n)))
+#define XFS_IFORK_NEXTENTS(ip,w) \
+	((w) == XFS_DATA_FORK ? \
+		(ip)->i_d.di_nextents : \
+		(ip)->i_d.di_anextents)
+#define XFS_IFORK_NEXT_SET(ip,w,n) \
+	((w) == XFS_DATA_FORK ? \
+		((ip)->i_d.di_nextents = (n)) : \
+		((ip)->i_d.di_anextents = (n)))
+
+
+
+#ifdef __KERNEL__
+
+struct bhv_desc;
+struct cred;
+struct ktrace;
+struct xfs_buf;
+struct xfs_bmap_free;
+struct xfs_bmbt_irec;
+struct xfs_inode_log_item;
+struct xfs_mount;
+struct xfs_trans;
+struct xfs_dquot;
+
+#if defined(XFS_ILOCK_TRACE)
+#define XFS_ILOCK_KTRACE_SIZE	32
+extern void xfs_ilock_trace(struct xfs_inode *, int, unsigned int, inst_t *);
+#else
+#define	xfs_ilock_trace(i,n,f,ra)
+#endif
+
+typedef struct dm_attrs_s {
+	__uint32_t	da_dmevmask;	/* DMIG event mask */
+	__uint16_t	da_dmstate;	/* DMIG state info */
+	__uint16_t	da_pad;		/* DMIG extra padding */
+} dm_attrs_t;
 
 typedef struct xfs_inode {
 	/* Inode linking and identification information. */
-	struct xfs_inode	*i_mnext;	/* next inode in mount list */
-	struct xfs_inode	*i_mprev;	/* ptr to prev inode */
 	struct xfs_mount	*i_mount;	/* fs mount struct ptr */
-	struct list_head	i_reclaim;	/* reclaim list */
-	struct inode		*i_vnode;	/* vnode backpointer */
 	struct xfs_dquot	*i_udquot;	/* user dquot */
 	struct xfs_dquot	*i_gdquot;	/* group dquot */
 
 	/* Inode location stuff */
 	xfs_ino_t		i_ino;		/* inode number (agno/agino)*/
-	xfs_daddr_t		i_blkno;	/* blkno of inode buffer */
-	ushort			i_len;		/* len of inode buffer */
-	ushort			i_boffset;	/* off of inode in buffer */
+	struct xfs_imap		i_imap;		/* location for xfs_imap() */
 
 	/* Extent information. */
 	xfs_ifork_t		*i_afp;		/* attribute fork pointer */
@@ -230,7 +261,6 @@ typedef struct xfs_inode {
 	unsigned short		i_flags;	/* see defined flags below */
 	unsigned char		i_update_core;	/* timestamps/size is dirty */
 	unsigned char		i_update_size;	/* di_size field is dirty */
-	unsigned int		i_gen;		/* generation count */
 	unsigned int		i_delayed_blks;	/* count of delay alloc blks */
 
 	xfs_icdinode_t		i_d;		/* most of ondisk inode */
@@ -238,6 +268,10 @@ typedef struct xfs_inode {
 	xfs_fsize_t		i_size;		/* in-memory size */
 	xfs_fsize_t		i_new_size;	/* size when write completes */
 	atomic_t		i_iocount;	/* outstanding I/O count */
+
+	/* VFS inode */
+	struct inode		i_vnode;	/* embedded VFS inode */
+
 	/* Trace buffers per inode. */
 #ifdef XFS_INODE_TRACE
 	struct ktrace		*i_trace;	/* general inode trace */
@@ -245,7 +279,7 @@ typedef struct xfs_inode {
 #ifdef XFS_BMAP_TRACE
 	struct ktrace		*i_xtrace;	/* inode extent list trace */
 #endif
-#ifdef XFS_BMBT_TRACE
+#ifdef XFS_BTREE_TRACE
 	struct ktrace		*i_btrace;	/* inode bmap btree trace */
 #endif
 #ifdef XFS_RW_TRACE
@@ -265,13 +299,30 @@ typedef struct xfs_inode {
 /* Convert from vfs inode to xfs inode */
 static inline struct xfs_inode *XFS_I(struct inode *inode)
 {
-	return (struct xfs_inode *)inode->i_private;
+	return container_of(inode, struct xfs_inode, i_vnode);
 }
 
 /* convert from xfs inode to vfs inode */
 static inline struct inode *VFS_I(struct xfs_inode *ip)
 {
-	return (struct inode *)ip->i_vnode;
+	return &ip->i_vnode;
+}
+
+/*
+ * Get rid of a partially initialized inode.
+ *
+ * We have to go through destroy_inode to make sure allocations
+ * from init_inode_always like the security data are undone.
+ *
+ * We mark the inode bad so that it takes the short cut in
+ * the reclaim path instead of going through the flush path
+ * which doesn't make sense for an inode that has never seen the
+ * light of day.
+ */
+static inline void xfs_destroy_inode(struct xfs_inode *ip)
+{
+	make_bad_inode(VFS_I(ip));
+	return destroy_inode(VFS_I(ip));
 }
 
 /*
@@ -327,65 +378,36 @@ xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags)
 	spin_unlock(&ip->i_flags_lock);
 	return ret;
 }
-#endif	/* __KERNEL__ */
-
 
 /*
- * Fork handling.
+ * Manage the i_flush queue embedded in the inode.  This completion
+ * queue synchronizes processes attempting to flush the in-core
+ * inode back to disk.
  */
+static inline void xfs_iflock(xfs_inode_t *ip)
+{
+	wait_for_completion(&ip->i_flush);
+}
 
-#define XFS_IFORK_Q(ip)			((ip)->i_d.di_forkoff != 0)
-#define XFS_IFORK_BOFF(ip)		((int)((ip)->i_d.di_forkoff << 3))
-
-#define XFS_IFORK_PTR(ip,w)		\
-	((w) == XFS_DATA_FORK ? \
-		&(ip)->i_df : \
-		(ip)->i_afp)
-#define XFS_IFORK_DSIZE(ip) \
-	(XFS_IFORK_Q(ip) ? \
-		XFS_IFORK_BOFF(ip) : \
-		XFS_LITINO((ip)->i_mount))
-#define XFS_IFORK_ASIZE(ip) \
-	(XFS_IFORK_Q(ip) ? \
-		XFS_LITINO((ip)->i_mount) - XFS_IFORK_BOFF(ip) : \
-		0)
-#define XFS_IFORK_SIZE(ip,w) \
-	((w) == XFS_DATA_FORK ? \
-		XFS_IFORK_DSIZE(ip) : \
-		XFS_IFORK_ASIZE(ip))
-#define XFS_IFORK_FORMAT(ip,w) \
-	((w) == XFS_DATA_FORK ? \
-		(ip)->i_d.di_format : \
-		(ip)->i_d.di_aformat)
-#define XFS_IFORK_FMT_SET(ip,w,n) \
-	((w) == XFS_DATA_FORK ? \
-		((ip)->i_d.di_format = (n)) : \
-		((ip)->i_d.di_aformat = (n)))
-#define XFS_IFORK_NEXTENTS(ip,w) \
-	((w) == XFS_DATA_FORK ? \
-		(ip)->i_d.di_nextents : \
-		(ip)->i_d.di_anextents)
-#define XFS_IFORK_NEXT_SET(ip,w,n) \
-	((w) == XFS_DATA_FORK ? \
-		((ip)->i_d.di_nextents = (n)) : \
-		((ip)->i_d.di_anextents = (n)))
+static inline int xfs_iflock_nowait(xfs_inode_t *ip)
+{
+	return try_wait_for_completion(&ip->i_flush);
+}
 
-#ifdef __KERNEL__
+static inline void xfs_ifunlock(xfs_inode_t *ip)
+{
+	complete(&ip->i_flush);
+}
 
 /*
  * In-core inode flags.
  */
-#define XFS_IGRIO	0x0001  /* inode used for guaranteed rate i/o */
-#define XFS_IUIOSZ	0x0002  /* inode i/o sizes have been explicitly set */
-#define XFS_IQUIESCE    0x0004  /* we have started quiescing for this inode */
-#define XFS_IRECLAIM    0x0008  /* we have started reclaiming this inode    */
-#define XFS_ISTALE	0x0010	/* inode has been staled */
-#define XFS_IRECLAIMABLE 0x0020 /* inode can be reclaimed */
-#define XFS_INEW	0x0040
-#define XFS_IFILESTREAM	0x0080	/* inode is in a filestream directory */
-#define XFS_IMODIFIED	0x0100	/* XFS inode state possibly differs */
-				/* to the Linux inode state. */
-#define XFS_ITRUNCATED	0x0200	/* truncated down so flush-on-close */
+#define XFS_IRECLAIM    0x0001  /* we have started reclaiming this inode    */
+#define XFS_ISTALE	0x0002	/* inode has been staled */
+#define XFS_IRECLAIMABLE 0x0004 /* inode can be reclaimed */
+#define XFS_INEW	0x0008	/* inode has just been allocated */
+#define XFS_IFILESTREAM	0x0010	/* inode is in a filestream directory */
+#define XFS_ITRUNCATED	0x0020	/* truncated down so flush-on-close */
 
 /*
  * Flags for inode locking.
@@ -460,16 +482,8 @@ xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags)
 	 ((pip)->i_d.di_mode & S_ISGID))
 
 /*
- * Flags for xfs_iget()
- */
-#define XFS_IGET_CREATE		0x1
-#define XFS_IGET_BULKSTAT	0x2
-
-/*
  * xfs_iget.c prototypes.
  */
-void		xfs_ihash_init(struct xfs_mount *);
-void		xfs_ihash_free(struct xfs_mount *);
 xfs_inode_t	*xfs_inode_incore(struct xfs_mount *, xfs_ino_t,
 				  struct xfs_trans *);
 int		xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
@@ -484,25 +498,13 @@ int		xfs_isilocked(xfs_inode_t *, uint);
 uint		xfs_ilock_map_shared(xfs_inode_t *);
 void		xfs_iunlock_map_shared(xfs_inode_t *, uint);
 void		xfs_ireclaim(xfs_inode_t *);
-int		xfs_finish_reclaim(xfs_inode_t *, int, int);
-int		xfs_finish_reclaim_all(struct xfs_mount *, int);
 
 /*
  * xfs_inode.c prototypes.
  */
-int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
-			  xfs_inode_t *, struct xfs_dinode **, struct xfs_buf **,
-			  xfs_daddr_t, uint, uint);
-int		xfs_iread(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
-			  xfs_inode_t **, xfs_daddr_t, uint);
-int		xfs_iread_extents(struct xfs_trans *, xfs_inode_t *, int);
 int		xfs_ialloc(struct xfs_trans *, xfs_inode_t *, mode_t,
 			   xfs_nlink_t, xfs_dev_t, cred_t *, xfs_prid_t,
 			   int, struct xfs_buf **, boolean_t *, xfs_inode_t **);
-void		xfs_dinode_from_disk(struct xfs_icdinode *,
-				     struct xfs_dinode_core *);
-void		xfs_dinode_to_disk(struct xfs_dinode_core *,
-				   struct xfs_icdinode *);
 
 uint		xfs_ip2xflags(struct xfs_inode *);
 uint		xfs_dic2xflags(struct xfs_dinode *);
@@ -513,17 +515,10 @@ int		xfs_itruncate_finish(struct xfs_trans **, xfs_inode_t *,
 				     xfs_fsize_t, int, int);
 int		xfs_iunlink(struct xfs_trans *, xfs_inode_t *);
 
-void		xfs_idestroy_fork(xfs_inode_t *, int);
-void		xfs_idestroy(xfs_inode_t *);
-void		xfs_idata_realloc(xfs_inode_t *, int, int);
-void		xfs_iextract(xfs_inode_t *);
 void		xfs_iext_realloc(xfs_inode_t *, int, int);
-void		xfs_iroot_realloc(xfs_inode_t *, int, int);
 void		xfs_ipin(xfs_inode_t *);
 void		xfs_iunpin(xfs_inode_t *);
-int		xfs_iextents_copy(xfs_inode_t *, xfs_bmbt_rec_t *, int);
 int		xfs_iflush(xfs_inode_t *, uint);
-void		xfs_iflush_all(struct xfs_mount *);
 void		xfs_ichgtime(xfs_inode_t *, int);
 xfs_fsize_t	xfs_file_last_byte(xfs_inode_t *);
 void		xfs_lock_inodes(xfs_inode_t **, int, uint);
@@ -532,6 +527,77 @@ void		xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint);
 void		xfs_synchronize_atime(xfs_inode_t *);
 void		xfs_mark_inode_dirty_sync(xfs_inode_t *);
 
+#if defined(XFS_INODE_TRACE)
+
+#define	INODE_TRACE_SIZE	16		/* number of trace entries */
+#define	INODE_KTRACE_ENTRY	1
+#define	INODE_KTRACE_EXIT	2
+#define	INODE_KTRACE_HOLD	3
+#define	INODE_KTRACE_REF	4
+#define	INODE_KTRACE_RELE	5
+
+extern void _xfs_itrace_entry(struct xfs_inode *, const char *, inst_t *);
+extern void _xfs_itrace_exit(struct xfs_inode *, const char *, inst_t *);
+extern void xfs_itrace_hold(struct xfs_inode *, char *, int, inst_t *);
+extern void _xfs_itrace_ref(struct xfs_inode *, char *, int, inst_t *);
+extern void xfs_itrace_rele(struct xfs_inode *, char *, int, inst_t *);
+#define xfs_itrace_entry(ip)	\
+	_xfs_itrace_entry(ip, __func__, (inst_t *)__return_address)
+#define xfs_itrace_exit(ip)	\
+	_xfs_itrace_exit(ip, __func__, (inst_t *)__return_address)
+#define xfs_itrace_exit_tag(ip, tag)	\
+	_xfs_itrace_exit(ip, tag, (inst_t *)__return_address)
+#define xfs_itrace_ref(ip)	\
+	_xfs_itrace_ref(ip, __FILE__, __LINE__, (inst_t *)__return_address)
+
+#else
+#define	xfs_itrace_entry(a)
+#define	xfs_itrace_exit(a)
+#define	xfs_itrace_exit_tag(a, b)
+#define	xfs_itrace_hold(a, b, c, d)
+#define	xfs_itrace_ref(a)
+#define	xfs_itrace_rele(a, b, c, d)
+#endif
+
+#define IHOLD(ip) \
+do { \
+	ASSERT(atomic_read(&VFS_I(ip)->i_count) > 0) ; \
+	atomic_inc(&(VFS_I(ip)->i_count)); \
+	xfs_itrace_hold((ip), __FILE__, __LINE__, (inst_t *)__return_address); \
+} while (0)
+
+#define IRELE(ip) \
+do { \
+	xfs_itrace_rele((ip), __FILE__, __LINE__, (inst_t *)__return_address); \
+	iput(VFS_I(ip)); \
+} while (0)
+
+#endif /* __KERNEL__ */
+
+/*
+ * Flags for xfs_iget()
+ */
+#define XFS_IGET_CREATE		0x1
+#define XFS_IGET_BULKSTAT	0x2
+
+int		xfs_inotobp(struct xfs_mount *, struct xfs_trans *,
+			    xfs_ino_t, struct xfs_dinode **,
+			    struct xfs_buf **, int *, uint);
+int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
+			  struct xfs_inode *, struct xfs_dinode **,
+			  struct xfs_buf **, uint);
+int		xfs_iread(struct xfs_mount *, struct xfs_trans *,
+			  struct xfs_inode *, xfs_daddr_t, uint);
+void		xfs_dinode_from_disk(struct xfs_icdinode *,
+				     struct xfs_dinode *);
+void		xfs_dinode_to_disk(struct xfs_dinode *,
+				   struct xfs_icdinode *);
+void		xfs_idestroy_fork(struct xfs_inode *, int);
+void		xfs_idata_realloc(struct xfs_inode *, int, int);
+void		xfs_iroot_realloc(struct xfs_inode *, int, int);
+int		xfs_iread_extents(struct xfs_trans *, struct xfs_inode *, int);
+int		xfs_iextents_copy(struct xfs_inode *, xfs_bmbt_rec_t *, int);
+
 xfs_bmbt_rec_host_t *xfs_iext_get_ext(xfs_ifork_t *, xfs_extnum_t);
 void		xfs_iext_insert(xfs_ifork_t *, xfs_extnum_t, xfs_extnum_t,
 				xfs_bmbt_irec_t *);
@@ -561,7 +627,8 @@ void		xfs_iext_irec_update_extoffs(xfs_ifork_t *, int, int);
 #define xfs_ipincount(ip)	((unsigned int) atomic_read(&ip->i_pincount))
 
 #ifdef DEBUG
-void		xfs_isize_check(struct xfs_mount *, xfs_inode_t *, xfs_fsize_t);
+void		xfs_isize_check(struct xfs_mount *, struct xfs_inode *,
+				xfs_fsize_t);
 #else	/* DEBUG */
 #define xfs_isize_check(mp, ip, isize)
 #endif	/* DEBUG */
@@ -576,26 +643,4 @@ extern struct kmem_zone	*xfs_ifork_zone;
 extern struct kmem_zone	*xfs_inode_zone;
 extern struct kmem_zone	*xfs_ili_zone;
 
-/*
- * Manage the i_flush queue embedded in the inode.  This completion
- * queue synchronizes processes attempting to flush the in-core
- * inode back to disk.
- */
-static inline void xfs_iflock(xfs_inode_t *ip)
-{
-	wait_for_completion(&ip->i_flush);
-}
-
-static inline int xfs_iflock_nowait(xfs_inode_t *ip)
-{
-	return try_wait_for_completion(&ip->i_flush);
-}
-
-static inline void xfs_ifunlock(xfs_inode_t *ip)
-{
-	complete(&ip->i_flush);
-}
-
-#endif	/* __KERNEL__ */
-
 #endif	/* __XFS_INODE_H__ */
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 97c7452e262..977c4aec587 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -281,7 +281,7 @@ xfs_inode_item_format(
 	xfs_mark_inode_dirty_sync(ip);
 
 	vecp->i_addr = (xfs_caddr_t)&ip->i_d;
-	vecp->i_len  = sizeof(xfs_dinode_core_t);
+	vecp->i_len  = sizeof(struct xfs_icdinode);
 	XLOG_VEC_SET_TYPE(vecp, XLOG_REG_TYPE_ICORE);
 	vecp++;
 	nvecs++;
@@ -296,9 +296,8 @@ xfs_inode_item_format(
 	 * has a new version number, then we don't bother converting back.
 	 */
 	mp = ip->i_mount;
-	ASSERT(ip->i_d.di_version == XFS_DINODE_VERSION_1 ||
-	       xfs_sb_version_hasnlink(&mp->m_sb));
-	if (ip->i_d.di_version == XFS_DINODE_VERSION_1) {
+	ASSERT(ip->i_d.di_version == 1 || xfs_sb_version_hasnlink(&mp->m_sb));
+	if (ip->i_d.di_version == 1) {
 		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
 			/*
 			 * Convert it back.
@@ -311,7 +310,7 @@ xfs_inode_item_format(
 			 * so just make the conversion to the new inode
 			 * format permanent.
 			 */
-			ip->i_d.di_version = XFS_DINODE_VERSION_2;
+			ip->i_d.di_version = 2;
 			ip->i_d.di_onlink = 0;
 			memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
 		}
@@ -932,6 +931,7 @@ xfs_inode_item_init(
 	iip->ili_item.li_type = XFS_LI_INODE;
 	iip->ili_item.li_ops = &xfs_inode_item_ops;
 	iip->ili_item.li_mountp = mp;
+	iip->ili_item.li_ailp = mp->m_ail;
 	iip->ili_inode = ip;
 
 	/*
@@ -942,9 +942,9 @@ xfs_inode_item_init(
 
 	iip->ili_format.ilf_type = XFS_LI_INODE;
 	iip->ili_format.ilf_ino = ip->i_ino;
-	iip->ili_format.ilf_blkno = ip->i_blkno;
-	iip->ili_format.ilf_len = ip->i_len;
-	iip->ili_format.ilf_boffset = ip->i_boffset;
+	iip->ili_format.ilf_blkno = ip->i_imap.im_blkno;
+	iip->ili_format.ilf_len = ip->i_imap.im_len;
+	iip->ili_format.ilf_boffset = ip->i_imap.im_boffset;
 }
 
 /*
@@ -976,9 +976,8 @@ xfs_iflush_done(
 	xfs_buf_t		*bp,
 	xfs_inode_log_item_t	*iip)
 {
-	xfs_inode_t	*ip;
-
-	ip = iip->ili_inode;
+	xfs_inode_t		*ip = iip->ili_inode;
+	struct xfs_ail		*ailp = iip->ili_item.li_ailp;
 
 	/*
 	 * We only want to pull the item from the AIL if it is
@@ -991,15 +990,12 @@ xfs_iflush_done(
 	 */
 	if (iip->ili_logged &&
 	    (iip->ili_item.li_lsn == iip->ili_flush_lsn)) {
-		spin_lock(&ip->i_mount->m_ail_lock);
+		spin_lock(&ailp->xa_lock);
 		if (iip->ili_item.li_lsn == iip->ili_flush_lsn) {
-			/*
-			 * xfs_trans_delete_ail() drops the AIL lock.
-			 */
-			xfs_trans_delete_ail(ip->i_mount,
-					     (xfs_log_item_t*)iip);
+			/* xfs_trans_ail_delete() drops the AIL lock. */
+			xfs_trans_ail_delete(ailp, (xfs_log_item_t*)iip);
 		} else {
-			spin_unlock(&ip->i_mount->m_ail_lock);
+			spin_unlock(&ailp->xa_lock);
 		}
 	}
 
@@ -1031,21 +1027,20 @@ void
 xfs_iflush_abort(
 	xfs_inode_t		*ip)
 {
-	xfs_inode_log_item_t	*iip;
+	xfs_inode_log_item_t	*iip = ip->i_itemp;
 	xfs_mount_t		*mp;
 
 	iip = ip->i_itemp;
 	mp = ip->i_mount;
 	if (iip) {
+		struct xfs_ail	*ailp = iip->ili_item.li_ailp;
 		if (iip->ili_item.li_flags & XFS_LI_IN_AIL) {
-			spin_lock(&mp->m_ail_lock);
+			spin_lock(&ailp->xa_lock);
 			if (iip->ili_item.li_flags & XFS_LI_IN_AIL) {
-				/*
-				 * xfs_trans_delete_ail() drops the AIL lock.
-				 */
-				xfs_trans_delete_ail(mp, (xfs_log_item_t *)iip);
+				/* xfs_trans_ail_delete() drops the AIL lock. */
+				xfs_trans_ail_delete(ailp, (xfs_log_item_t *)iip);
 			} else
-				spin_unlock(&mp->m_ail_lock);
+				spin_unlock(&ailp->xa_lock);
 		}
 		iip->ili_logged = 0;
 		/*
diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h
index 40513077ab3..1ff04cc323a 100644
--- a/fs/xfs/xfs_inode_item.h
+++ b/fs/xfs/xfs_inode_item.h
@@ -112,6 +112,24 @@ typedef struct xfs_inode_log_format_64 {
 #define	XFS_ILI_IOLOCKED_ANY   (XFS_ILI_IOLOCKED_EXCL | XFS_ILI_IOLOCKED_SHARED)
 
 
+#define	XFS_ILOG_FBROOT(w)	xfs_ilog_fbroot(w)
+static inline int xfs_ilog_fbroot(int w)
+{
+	return (w == XFS_DATA_FORK ? XFS_ILOG_DBROOT : XFS_ILOG_ABROOT);
+}
+
+#define	XFS_ILOG_FEXT(w)	xfs_ilog_fext(w)
+static inline int xfs_ilog_fext(int w)
+{
+	return (w == XFS_DATA_FORK ? XFS_ILOG_DEXT : XFS_ILOG_AEXT);
+}
+
+#define	XFS_ILOG_FDATA(w)	xfs_ilog_fdata(w)
+static inline int xfs_ilog_fdata(int w)
+{
+	return (w == XFS_DATA_FORK ? XFS_ILOG_DDATA : XFS_ILOG_ADATA);
+}
+
 #ifdef __KERNEL__
 
 struct xfs_buf;
@@ -148,26 +166,6 @@ typedef struct xfs_inode_log_item {
 } xfs_inode_log_item_t;
 
 
-#define	XFS_ILOG_FDATA(w)	xfs_ilog_fdata(w)
-static inline int xfs_ilog_fdata(int w)
-{
-	return (w == XFS_DATA_FORK ? XFS_ILOG_DDATA : XFS_ILOG_ADATA);
-}
-
-#endif	/* __KERNEL__ */
-
-#define	XFS_ILOG_FBROOT(w)	xfs_ilog_fbroot(w)
-static inline int xfs_ilog_fbroot(int w)
-{
-	return (w == XFS_DATA_FORK ? XFS_ILOG_DBROOT : XFS_ILOG_ABROOT);
-}
-
-#define	XFS_ILOG_FEXT(w)	xfs_ilog_fext(w)
-static inline int xfs_ilog_fext(int w)
-{
-	return (w == XFS_DATA_FORK ? XFS_ILOG_DEXT : XFS_ILOG_AEXT);
-}
-
 static inline int xfs_inode_clean(xfs_inode_t *ip)
 {
 	return (!ip->i_itemp ||
@@ -175,9 +173,6 @@ static inline int xfs_inode_clean(xfs_inode_t *ip)
 	       !ip->i_update_core;
 }
 
-
-#ifdef __KERNEL__
-
 extern void xfs_inode_item_init(struct xfs_inode *, struct xfs_mount *);
 extern void xfs_inode_item_destroy(struct xfs_inode *);
 extern void xfs_iflush_done(struct xfs_buf *, xfs_inode_log_item_t *);
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 67f22b2b44b..911062cf73a 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -290,7 +290,6 @@ STATIC int
 xfs_iomap_eof_align_last_fsb(
 	xfs_mount_t	*mp,
 	xfs_inode_t	*ip,
-	xfs_fsize_t	isize,
 	xfs_extlen_t	extsize,
 	xfs_fileoff_t	*last_fsb)
 {
@@ -306,14 +305,14 @@ xfs_iomap_eof_align_last_fsb(
 	 * stripe width and we are allocating past the allocation eof.
 	 */
 	else if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC) &&
-	        (isize >= XFS_FSB_TO_B(mp, mp->m_swidth)))
+	        (ip->i_size >= XFS_FSB_TO_B(mp, mp->m_swidth)))
 		new_last_fsb = roundup_64(*last_fsb, mp->m_swidth);
 	/*
 	 * Roundup the allocation request to a stripe unit (m_dalign) boundary
 	 * if the file size is >= stripe unit size, and we are allocating past
 	 * the allocation eof.
 	 */
-	else if (mp->m_dalign && (isize >= XFS_FSB_TO_B(mp, mp->m_dalign)))
+	else if (mp->m_dalign && (ip->i_size >= XFS_FSB_TO_B(mp, mp->m_dalign)))
 		new_last_fsb = roundup_64(*last_fsb, mp->m_dalign);
 
 	/*
@@ -403,7 +402,6 @@ xfs_iomap_write_direct(
 	xfs_filblks_t	count_fsb, resaligned;
 	xfs_fsblock_t	firstfsb;
 	xfs_extlen_t	extsz, temp;
-	xfs_fsize_t	isize;
 	int		nimaps;
 	int		bmapi_flag;
 	int		quota_flag;
@@ -426,15 +424,10 @@ xfs_iomap_write_direct(
 	rt = XFS_IS_REALTIME_INODE(ip);
 	extsz = xfs_get_extsz_hint(ip);
 
-	isize = ip->i_size;
-	if (ip->i_new_size > isize)
-		isize = ip->i_new_size;
-
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
 	last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count)));
-	if ((offset + count) > isize) {
-		error = xfs_iomap_eof_align_last_fsb(mp, ip, isize, extsz,
-							&last_fsb);
+	if ((offset + count) > ip->i_size) {
+		error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb);
 		if (error)
 			goto error_out;
 	} else {
@@ -559,7 +552,6 @@ STATIC int
 xfs_iomap_eof_want_preallocate(
 	xfs_mount_t	*mp,
 	xfs_inode_t	*ip,
-	xfs_fsize_t	isize,
 	xfs_off_t	offset,
 	size_t		count,
 	int		ioflag,
@@ -573,7 +565,7 @@ xfs_iomap_eof_want_preallocate(
 	int		n, error, imaps;
 
 	*prealloc = 0;
-	if ((ioflag & BMAPI_SYNC) || (offset + count) <= isize)
+	if ((ioflag & BMAPI_SYNC) || (offset + count) <= ip->i_size)
 		return 0;
 
 	/*
@@ -617,7 +609,6 @@ xfs_iomap_write_delay(
 	xfs_fileoff_t	ioalign;
 	xfs_fsblock_t	firstblock;
 	xfs_extlen_t	extsz;
-	xfs_fsize_t	isize;
 	int		nimaps;
 	xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS];
 	int		prealloc, fsynced = 0;
@@ -637,11 +628,7 @@ xfs_iomap_write_delay(
 	offset_fsb = XFS_B_TO_FSBT(mp, offset);
 
 retry:
-	isize = ip->i_size;
-	if (ip->i_new_size > isize)
-		isize = ip->i_new_size;
-
-	error = xfs_iomap_eof_want_preallocate(mp, ip, isize, offset, count,
+	error = xfs_iomap_eof_want_preallocate(mp, ip, offset, count,
 				ioflag, imap, XFS_WRITE_IMAPS, &prealloc);
 	if (error)
 		return error;
@@ -655,8 +642,7 @@ retry:
 	}
 
 	if (prealloc || extsz) {
-		error = xfs_iomap_eof_align_last_fsb(mp, ip, isize, extsz,
-							&last_fsb);
+		error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb);
 		if (error)
 			return error;
 	}
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index cf6754a3c5b..e19d0a8d561 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -69,7 +69,7 @@ xfs_bulkstat_one_iget(
 	}
 
 	ASSERT(ip != NULL);
-	ASSERT(ip->i_blkno != (xfs_daddr_t)0);
+	ASSERT(ip->i_imap.im_blkno != 0);
 
 	dic = &ip->i_d;
 
@@ -125,13 +125,9 @@ STATIC void
 xfs_bulkstat_one_dinode(
 	xfs_mount_t	*mp,		/* mount point for filesystem */
 	xfs_ino_t	ino,		/* inode number to get data for */
-	xfs_dinode_t	*dip,		/* dinode inode pointer */
+	xfs_dinode_t	*dic,		/* dinode inode pointer */
 	xfs_bstat_t	*buf)		/* return buffer */
 {
-	xfs_dinode_core_t *dic;		/* dinode core info pointer */
-
-	dic = &dip->di_core;
-
 	/*
 	 * The inode format changed when we moved the link count and
 	 * made it 32 bits long.  If this is an old format inode,
@@ -143,7 +139,7 @@ xfs_bulkstat_one_dinode(
 	 * the new format. We don't change the version number so that we
 	 * can distinguish this from a real new format inode.
 	 */
-	if (dic->di_version == XFS_DINODE_VERSION_1) {
+	if (dic->di_version == 1) {
 		buf->bs_nlink = be16_to_cpu(dic->di_onlink);
 		buf->bs_projid = 0;
 	} else {
@@ -162,7 +158,7 @@ xfs_bulkstat_one_dinode(
 	buf->bs_mtime.tv_nsec = be32_to_cpu(dic->di_mtime.t_nsec);
 	buf->bs_ctime.tv_sec = be32_to_cpu(dic->di_ctime.t_sec);
 	buf->bs_ctime.tv_nsec = be32_to_cpu(dic->di_ctime.t_nsec);
-	buf->bs_xflags = xfs_dic2xflags(dip);
+	buf->bs_xflags = xfs_dic2xflags(dic);
 	buf->bs_extsize = be32_to_cpu(dic->di_extsize) << mp->m_sb.sb_blocklog;
 	buf->bs_extents = be32_to_cpu(dic->di_nextents);
 	buf->bs_gen = be32_to_cpu(dic->di_gen);
@@ -173,7 +169,7 @@ xfs_bulkstat_one_dinode(
 
 	switch (dic->di_format) {
 	case XFS_DINODE_FMT_DEV:
-		buf->bs_rdev = be32_to_cpu(dip->di_u.di_dev);
+		buf->bs_rdev = xfs_dinode_get_rdev(dic);
 		buf->bs_blksize = BLKDEV_IOSIZE;
 		buf->bs_blocks = 0;
 		break;
@@ -192,27 +188,34 @@ xfs_bulkstat_one_dinode(
 	}
 }
 
+/* Return 0 on success or positive error */
 STATIC int
 xfs_bulkstat_one_fmt(
 	void			__user *ubuffer,
+	int			ubsize,
+	int			*ubused,
 	const xfs_bstat_t	*buffer)
 {
+	if (ubsize < sizeof(*buffer))
+		return XFS_ERROR(ENOMEM);
 	if (copy_to_user(ubuffer, buffer, sizeof(*buffer)))
-		return -EFAULT;
-	return sizeof(*buffer);
+		return XFS_ERROR(EFAULT);
+	if (ubused)
+		*ubused = sizeof(*buffer);
+	return 0;
 }
 
 /*
  * Return stat information for one inode.
  * Return 0 if ok, else errno.
  */
-int		       		/* error status */
-xfs_bulkstat_one(
+int		   	    		/* error status */
+xfs_bulkstat_one_int(
 	xfs_mount_t	*mp,		/* mount point for filesystem */
 	xfs_ino_t	ino,		/* inode number to get data for */
 	void		__user *buffer,	/* buffer to place output in */
 	int		ubsize,		/* size of buffer */
-	void		*private_data,	/* my private data */
+	bulkstat_one_fmt_pf formatter,	/* formatter, copy to user */
 	xfs_daddr_t	bno,		/* starting bno of inode cluster */
 	int		*ubused,	/* bytes used by me */
 	void		*dibuff,	/* on-disk inode buffer */
@@ -221,15 +224,12 @@ xfs_bulkstat_one(
 	xfs_bstat_t	*buf;		/* return buffer */
 	int		error = 0;	/* error value */
 	xfs_dinode_t	*dip;		/* dinode inode pointer */
-	bulkstat_one_fmt_pf formatter = private_data ? : xfs_bulkstat_one_fmt;
 
 	dip = (xfs_dinode_t *)dibuff;
 	*stat = BULKSTAT_RV_NOTHING;
 
 	if (!buffer || xfs_internal_inum(mp, ino))
 		return XFS_ERROR(EINVAL);
-	if (ubsize < sizeof(*buf))
-		return XFS_ERROR(ENOMEM);
 
 	buf = kmem_alloc(sizeof(*buf), KM_SLEEP);
 
@@ -244,21 +244,34 @@ xfs_bulkstat_one(
 		xfs_bulkstat_one_dinode(mp, ino, dip, buf);
 	}
 
-	error = formatter(buffer, buf);
-	if (error < 0)  {
-		error = EFAULT;
+	error = formatter(buffer, ubsize, ubused, buf);
+	if (error)
 		goto out_free;
-	}
 
 	*stat = BULKSTAT_RV_DIDONE;
-	if (ubused)
-		*ubused = error;
 
  out_free:
 	kmem_free(buf);
 	return error;
 }
 
+int
+xfs_bulkstat_one(
+	xfs_mount_t	*mp,		/* mount point for filesystem */
+	xfs_ino_t	ino,		/* inode number to get data for */
+	void		__user *buffer,	/* buffer to place output in */
+	int		ubsize,		/* size of buffer */
+	void		*private_data,	/* my private data */
+	xfs_daddr_t	bno,		/* starting bno of inode cluster */
+	int		*ubused,	/* bytes used by me */
+	void		*dibuff,	/* on-disk inode buffer */
+	int		*stat)		/* BULKSTAT_RV_... */
+{
+	return xfs_bulkstat_one_int(mp, ino, buffer, ubsize,
+				    xfs_bulkstat_one_fmt, bno,
+				    ubused, dibuff, stat);
+}
+
 /*
  * Test to see whether we can use the ondisk inode directly, based
  * on the given bulkstat flags, filling in dipp accordingly.
@@ -287,19 +300,19 @@ xfs_bulkstat_use_dinode(
 	 * to disk yet. This is a temporary hack that would require a proper
 	 * fix in the future.
 	 */
-	if (be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC ||
-	    !XFS_DINODE_GOOD_VERSION(dip->di_core.di_version) ||
-	    !dip->di_core.di_mode)
+	if (be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC ||
+	    !XFS_DINODE_GOOD_VERSION(dip->di_version) ||
+	    !dip->di_mode)
 		return 0;
 	if (flags & BULKSTAT_FG_QUICK) {
 		*dipp = dip;
 		return 1;
 	}
 	/* BULKSTAT_FG_INLINE: if attr fork is local, or not there, use it */
-	aformat = dip->di_core.di_aformat;
+	aformat = dip->di_aformat;
 	if ((XFS_DFORK_Q(dip) == 0) ||
 	    (aformat == XFS_DINODE_FMT_LOCAL) ||
-	    (aformat == XFS_DINODE_FMT_EXTENTS && !dip->di_core.di_anextents)) {
+	    (aformat == XFS_DINODE_FMT_EXTENTS && !dip->di_anextents)) {
 		*dipp = dip;
 		return 1;
 	}
@@ -359,7 +372,6 @@ xfs_bulkstat(
 	int			ubused;	/* bytes used by formatter */
 	xfs_buf_t		*bp;	/* ptr to on-disk inode cluster buf */
 	xfs_dinode_t		*dip;	/* ptr into bp for specific inode */
-	xfs_inode_t		*ip;	/* ptr to in-core inode struct */
 
 	/*
 	 * Get the last inode value, see if there's nothing to do.
@@ -416,8 +428,7 @@ xfs_bulkstat(
 		/*
 		 * Allocate and initialize a btree cursor for ialloc btree.
 		 */
-		cur = xfs_btree_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_INO,
-						(xfs_inode_t *)0, 0);
+		cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
 		irbp = irbuf;
 		irbufend = irbuf + nirbuf;
 		end_of_ag = 0;
@@ -472,7 +483,7 @@ xfs_bulkstat(
 			 * In any case, increment to the next record.
 			 */
 			if (!error)
-				error = xfs_inobt_increment(cur, 0, &tmp);
+				error = xfs_btree_increment(cur, 0, &tmp);
 		} else {
 			/*
 			 * Start of ag.  Lookup the first inode chunk.
@@ -539,7 +550,7 @@ xfs_bulkstat(
 			 * Set agino to after this chunk and bump the cursor.
 			 */
 			agino = gino + XFS_INODES_PER_CHUNK;
-			error = xfs_inobt_increment(cur, 0, &tmp);
+			error = xfs_btree_increment(cur, 0, &tmp);
 			cond_resched();
 		}
 		/*
@@ -586,6 +597,8 @@ xfs_bulkstat(
 
 					if (flags & (BULKSTAT_FG_QUICK |
 						     BULKSTAT_FG_INLINE)) {
+						int offset;
+
 						ino = XFS_AGINO_TO_INO(mp, agno,
 								       agino);
 						bno = XFS_AGB_TO_DADDR(mp, agno,
@@ -594,21 +607,15 @@ xfs_bulkstat(
 						/*
 						 * Get the inode cluster buffer
 						 */
-						ASSERT(xfs_inode_zone != NULL);
-						ip = kmem_zone_zalloc(xfs_inode_zone,
-								      KM_SLEEP);
-						ip->i_ino = ino;
-						ip->i_mount = mp;
-						spin_lock_init(&ip->i_flags_lock);
 						if (bp)
 							xfs_buf_relse(bp);
-						error = xfs_itobp(mp, NULL, ip,
-								&dip, &bp, bno,
-								XFS_IMAP_BULKSTAT,
-								XFS_BUF_LOCK);
+
+						error = xfs_inotobp(mp, NULL, ino, &dip,
+								    &bp, &offset,
+								    XFS_IGET_BULKSTAT);
+
 						if (!error)
-							clustidx = ip->i_boffset / mp->m_sb.sb_inodesize;
-						kmem_zone_free(xfs_inode_zone, ip);
+							clustidx = offset / mp->m_sb.sb_inodesize;
 						if (XFS_TEST_ERROR(error != 0,
 								   mp, XFS_ERRTAG_BULKSTAT_READ_CHUNK,
 								   XFS_RANDOM_BULKSTAT_READ_CHUNK)) {
@@ -842,8 +849,7 @@ xfs_inumbers(
 				agino = 0;
 				continue;
 			}
-			cur = xfs_btree_init_cursor(mp, NULL, agbp, agno,
-				XFS_BTNUM_INO, (xfs_inode_t *)0, 0);
+			cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
 			error = xfs_inobt_lookup_ge(cur, agino, 0, 0, &tmp);
 			if (error) {
 				xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
@@ -887,7 +893,7 @@ xfs_inumbers(
 			bufidx = 0;
 		}
 		if (left) {
-			error = xfs_inobt_increment(cur, 0, &tmp);
+			error = xfs_btree_increment(cur, 0, &tmp);
 			if (error) {
 				xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 				cur = NULL;
diff --git a/fs/xfs/xfs_itable.h b/fs/xfs/xfs_itable.h
index a1f18fce9b7..1fb04e7deb6 100644
--- a/fs/xfs/xfs_itable.h
+++ b/fs/xfs/xfs_itable.h
@@ -71,9 +71,23 @@ xfs_bulkstat_single(
 
 typedef int (*bulkstat_one_fmt_pf)(  /* used size in bytes or negative error */
 	void			__user *ubuffer, /* buffer to write to */
+	int			ubsize,		 /* remaining user buffer sz */
+	int			*ubused,	 /* bytes used by formatter */
 	const xfs_bstat_t	*buffer);        /* buffer to read from */
 
 int
+xfs_bulkstat_one_int(
+	xfs_mount_t		*mp,
+	xfs_ino_t		ino,
+	void			__user *buffer,
+	int			ubsize,
+	bulkstat_one_fmt_pf	formatter,
+	xfs_daddr_t		bno,
+	int			*ubused,
+	void			*dibuff,
+	int			*stat);
+
+int
 xfs_bulkstat_one(
 	xfs_mount_t		*mp,
 	xfs_ino_t		ino,
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index 3608a0f0a5f..f4726f702a9 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -100,12 +100,11 @@ STATIC void xlog_ungrant_log_space(xlog_t	 *log,
 
 
 /* local ticket functions */
-STATIC xlog_ticket_t	*xlog_ticket_get(xlog_t *log,
+STATIC xlog_ticket_t	*xlog_ticket_alloc(xlog_t *log,
 					 int	unit_bytes,
 					 int	count,
 					 char	clientid,
 					 uint	flags);
-STATIC void		xlog_ticket_put(xlog_t *log, xlog_ticket_t *ticket);
 
 #if defined(DEBUG)
 STATIC void	xlog_verify_dest_ptr(xlog_t *log, __psint_t ptr);
@@ -360,7 +359,7 @@ xfs_log_done(xfs_mount_t	*mp,
 		 */
 		xlog_trace_loggrant(log, ticket, "xfs_log_done: (non-permanent)");
 		xlog_ungrant_log_space(log, ticket);
-		xlog_ticket_put(log, ticket);
+		xfs_log_ticket_put(ticket);
 	} else {
 		xlog_trace_loggrant(log, ticket, "xfs_log_done: (permanent)");
 		xlog_regrant_reserve_log_space(log, ticket);
@@ -514,7 +513,7 @@ xfs_log_reserve(xfs_mount_t	 *mp,
 		retval = xlog_regrant_write_log_space(log, internal_ticket);
 	} else {
 		/* may sleep if need to allocate more tickets */
-		internal_ticket = xlog_ticket_get(log, unit_bytes, cnt,
+		internal_ticket = xlog_ticket_alloc(log, unit_bytes, cnt,
 						  client, flags);
 		if (!internal_ticket)
 			return XFS_ERROR(ENOMEM);
@@ -572,12 +571,12 @@ xfs_log_mount(
 	/*
 	 * Initialize the AIL now we have a log.
 	 */
-	spin_lock_init(&mp->m_ail_lock);
 	error = xfs_trans_ail_init(mp);
 	if (error) {
 		cmn_err(CE_WARN, "XFS: AIL initialisation failed: error %d", error);
 		goto error;
 	}
+	mp->m_log->l_ailp = mp->m_ail;
 
 	/*
 	 * skip log recovery on a norecovery mount.  pretend it all
@@ -730,8 +729,8 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 		spin_lock(&log->l_icloglock);
 		iclog = log->l_iclog;
 		atomic_inc(&iclog->ic_refcnt);
-		spin_unlock(&log->l_icloglock);
 		xlog_state_want_sync(log, iclog);
+		spin_unlock(&log->l_icloglock);
 		error = xlog_state_release_iclog(log, iclog);
 
 		spin_lock(&log->l_icloglock);
@@ -749,7 +748,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 		if (tic) {
 			xlog_trace_loggrant(log, tic, "unmount rec");
 			xlog_ungrant_log_space(log, tic);
-			xlog_ticket_put(log, tic);
+			xfs_log_ticket_put(tic);
 		}
 	} else {
 		/*
@@ -768,9 +767,9 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 		spin_lock(&log->l_icloglock);
 		iclog = log->l_iclog;
 		atomic_inc(&iclog->ic_refcnt);
-		spin_unlock(&log->l_icloglock);
 
 		xlog_state_want_sync(log, iclog);
+		spin_unlock(&log->l_icloglock);
 		error =  xlog_state_release_iclog(log, iclog);
 
 		spin_lock(&log->l_icloglock);
@@ -906,7 +905,7 @@ xfs_log_move_tail(xfs_mount_t	*mp,
 int
 xfs_log_need_covered(xfs_mount_t *mp)
 {
-	int		needed = 0, gen;
+	int		needed = 0;
 	xlog_t		*log = mp->m_log;
 
 	if (!xfs_fs_writable(mp))
@@ -915,7 +914,7 @@ xfs_log_need_covered(xfs_mount_t *mp)
 	spin_lock(&log->l_icloglock);
 	if (((log->l_covered_state == XLOG_STATE_COVER_NEED) ||
 		(log->l_covered_state == XLOG_STATE_COVER_NEED2))
-			&& !xfs_trans_first_ail(mp, &gen)
+			&& !xfs_trans_ail_tail(log->l_ailp)
 			&& xlog_iclogs_empty(log)) {
 		if (log->l_covered_state == XLOG_STATE_COVER_NEED)
 			log->l_covered_state = XLOG_STATE_COVER_DONE;
@@ -952,7 +951,7 @@ xlog_assign_tail_lsn(xfs_mount_t *mp)
 	xfs_lsn_t tail_lsn;
 	xlog_t	  *log = mp->m_log;
 
-	tail_lsn = xfs_trans_tail_ail(mp);
+	tail_lsn = xfs_trans_ail_tail(mp->m_ail);
 	spin_lock(&log->l_grant_lock);
 	if (tail_lsn != 0) {
 		log->l_tail_lsn = tail_lsn;
@@ -1030,12 +1029,6 @@ xlog_iodone(xfs_buf_t *bp)
 	ASSERT(XFS_BUF_FSPRIVATE2(bp, unsigned long) == (unsigned long) 2);
 	XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1);
 	aborted = 0;
-
-	/*
-	 * Some versions of cpp barf on the recursive definition of
-	 * ic_log -> hic_fields.ic_log and expand ic_log twice when
-	 * it is passed through two macros.  Workaround broken cpp.
-	 */
 	l = iclog->ic_log;
 
 	/*
@@ -1302,7 +1295,7 @@ xlog_alloc_log(xfs_mount_t	*mp,
 		XFS_BUF_SET_BDSTRAT_FUNC(bp, xlog_bdstrat_cb);
 		XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1);
 		iclog->ic_bp = bp;
-		iclog->hic_data = bp->b_addr;
+		iclog->ic_data = bp->b_addr;
 #ifdef DEBUG
 		log->l_iclog_bak[i] = (xfs_caddr_t)&(iclog->ic_header);
 #endif
@@ -1322,7 +1315,7 @@ xlog_alloc_log(xfs_mount_t	*mp,
 		atomic_set(&iclog->ic_refcnt, 0);
 		spin_lock_init(&iclog->ic_callback_lock);
 		iclog->ic_callback_tail = &(iclog->ic_callback);
-		iclog->ic_datap = (char *)iclog->hic_data + log->l_iclog_hsize;
+		iclog->ic_datap = (char *)iclog->ic_data + log->l_iclog_hsize;
 
 		ASSERT(XFS_BUF_ISBUSY(iclog->ic_bp));
 		ASSERT(XFS_BUF_VALUSEMA(iclog->ic_bp) <= 0);
@@ -1446,7 +1439,7 @@ xlog_grant_push_ail(xfs_mount_t	*mp,
      */
     if (threshold_lsn &&
 	!XLOG_FORCED_SHUTDOWN(log))
-	    xfs_trans_push_ail(mp, threshold_lsn);
+	    xfs_trans_ail_push(log->l_ailp, threshold_lsn);
 }	/* xlog_grant_push_ail */
 
 
@@ -1991,7 +1984,9 @@ xlog_write(xfs_mount_t *	mp,
 		if (iclog->ic_size - log_offset <= sizeof(xlog_op_header_t)) {
 		    xlog_state_finish_copy(log, iclog, record_cnt, data_cnt);
 		    record_cnt = data_cnt = 0;
+		    spin_lock(&log->l_icloglock);
 		    xlog_state_want_sync(log, iclog);
+		    spin_unlock(&log->l_icloglock);
 		    if (commit_iclog) {
 			ASSERT(flags & XLOG_COMMIT_TRANS);
 			*commit_iclog = iclog;
@@ -3200,7 +3195,7 @@ try_again:
 STATIC void
 xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog)
 {
-	spin_lock(&log->l_icloglock);
+	ASSERT(spin_is_locked(&log->l_icloglock));
 
 	if (iclog->ic_state == XLOG_STATE_ACTIVE) {
 		xlog_state_switch_iclogs(log, iclog, 0);
@@ -3208,10 +3203,7 @@ xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog)
 		ASSERT(iclog->ic_state &
 			(XLOG_STATE_WANT_SYNC|XLOG_STATE_IOERROR));
 	}
-
-	spin_unlock(&log->l_icloglock);
-}	/* xlog_state_want_sync */
-
+}
 
 
 /*****************************************************************************
@@ -3222,22 +3214,33 @@ xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog)
  */
 
 /*
- * Free a used ticket.
+ * Free a used ticket when it's refcount falls to zero.
  */
-STATIC void
-xlog_ticket_put(xlog_t		*log,
-		xlog_ticket_t	*ticket)
+void
+xfs_log_ticket_put(
+	xlog_ticket_t	*ticket)
 {
-	sv_destroy(&ticket->t_wait);
-	kmem_zone_free(xfs_log_ticket_zone, ticket);
-}	/* xlog_ticket_put */
+	ASSERT(atomic_read(&ticket->t_ref) > 0);
+	if (atomic_dec_and_test(&ticket->t_ref)) {
+		sv_destroy(&ticket->t_wait);
+		kmem_zone_free(xfs_log_ticket_zone, ticket);
+	}
+}
 
+xlog_ticket_t *
+xfs_log_ticket_get(
+	xlog_ticket_t	*ticket)
+{
+	ASSERT(atomic_read(&ticket->t_ref) > 0);
+	atomic_inc(&ticket->t_ref);
+	return ticket;
+}
 
 /*
  * Allocate and initialise a new log ticket.
  */
 STATIC xlog_ticket_t *
-xlog_ticket_get(xlog_t		*log,
+xlog_ticket_alloc(xlog_t		*log,
 		int		unit_bytes,
 		int		cnt,
 		char		client,
@@ -3308,6 +3311,7 @@ xlog_ticket_get(xlog_t		*log,
 		unit_bytes += 2*BBSIZE;
         }
 
+	atomic_set(&tic->t_ref, 1);
 	tic->t_unit_res		= unit_bytes;
 	tic->t_curr_res		= unit_bytes;
 	tic->t_cnt		= cnt;
@@ -3323,7 +3327,7 @@ xlog_ticket_get(xlog_t		*log,
 	xlog_tic_reset_res(tic);
 
 	return tic;
-}	/* xlog_ticket_get */
+}
 
 
 /******************************************************************************
@@ -3452,7 +3456,7 @@ xlog_verify_iclog(xlog_t	 *log,
 	ptr = iclog->ic_datap;
 	base_ptr = ptr;
 	ophead = (xlog_op_header_t *)ptr;
-	xhdr = (xlog_in_core_2_t *)&iclog->ic_header;
+	xhdr = iclog->ic_data;
 	for (i = 0; i < len; i++) {
 		ophead = (xlog_op_header_t *)ptr;
 
@@ -3558,7 +3562,8 @@ xfs_log_force_umount(
 	if (!log ||
 	    log->l_flags & XLOG_ACTIVE_RECOVERY) {
 		mp->m_flags |= XFS_MOUNT_FS_SHUTDOWN;
-		XFS_BUF_DONE(mp->m_sb_bp);
+		if (mp->m_sb_bp)
+			XFS_BUF_DONE(mp->m_sb_bp);
 		return 0;
 	}
 
@@ -3579,7 +3584,9 @@ xfs_log_force_umount(
 	spin_lock(&log->l_icloglock);
 	spin_lock(&log->l_grant_lock);
 	mp->m_flags |= XFS_MOUNT_FS_SHUTDOWN;
-	XFS_BUF_DONE(mp->m_sb_bp);
+	if (mp->m_sb_bp)
+		XFS_BUF_DONE(mp->m_sb_bp);
+
 	/*
 	 * This flag is sort of redundant because of the mount flag, but
 	 * it's good to maintain the separation between the log and the rest
diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h
index d47b91f1082..8a3e84e900a 100644
--- a/fs/xfs/xfs_log.h
+++ b/fs/xfs/xfs_log.h
@@ -134,6 +134,7 @@ typedef struct xfs_log_callback {
 #ifdef __KERNEL__
 /* Log manager interfaces */
 struct xfs_mount;
+struct xlog_ticket;
 xfs_lsn_t xfs_log_done(struct xfs_mount *mp,
 		       xfs_log_ticket_t ticket,
 		       void		**iclog,
@@ -177,6 +178,9 @@ int	  xfs_log_need_covered(struct xfs_mount *mp);
 
 void	  xlog_iodone(struct xfs_buf *);
 
+struct xlog_ticket * xfs_log_ticket_get(struct xlog_ticket *ticket);
+void	  xfs_log_ticket_put(struct xlog_ticket *ticket);
+
 #endif
 
 
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index e7d8f84443f..654167be0ef 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -245,6 +245,7 @@ typedef struct xlog_ticket {
 	struct xlog_ticket *t_next;	 /*			         :4|8 */
 	struct xlog_ticket *t_prev;	 /*				 :4|8 */
 	xlog_tid_t	   t_tid;	 /* transaction identifier	 : 4  */
+	atomic_t	   t_ref;	 /* ticket reference count       : 4  */
 	int		   t_curr_res;	 /* current reservation in bytes : 4  */
 	int		   t_unit_res;	 /* unit reservation in bytes    : 4  */
 	char		   t_ocnt;	 /* original count		 : 1  */
@@ -309,6 +310,16 @@ typedef struct xlog_rec_ext_header {
 } xlog_rec_ext_header_t;
 
 #ifdef __KERNEL__
+
+/*
+ * Quite misnamed, because this union lays out the actual on-disk log buffer.
+ */
+typedef union xlog_in_core2 {
+	xlog_rec_header_t	hic_header;
+	xlog_rec_ext_header_t	hic_xheader;
+	char			hic_sector[XLOG_HEADER_SIZE];
+} xlog_in_core_2_t;
+
 /*
  * - A log record header is 512 bytes.  There is plenty of room to grow the
  *	xlog_rec_header_t into the reserved space.
@@ -338,7 +349,7 @@ typedef struct xlog_rec_ext_header {
  * We'll put all the read-only and l_icloglock fields in the first cacheline,
  * and move everything else out to subsequent cachelines.
  */
-typedef struct xlog_iclog_fields {
+typedef struct xlog_in_core {
 	sv_t			ic_force_wait;
 	sv_t			ic_write_wait;
 	struct xlog_in_core	*ic_next;
@@ -361,41 +372,11 @@ typedef struct xlog_iclog_fields {
 
 	/* reference counts need their own cacheline */
 	atomic_t		ic_refcnt ____cacheline_aligned_in_smp;
-} xlog_iclog_fields_t;
-
-typedef union xlog_in_core2 {
-	xlog_rec_header_t	hic_header;
-	xlog_rec_ext_header_t	hic_xheader;
-	char			hic_sector[XLOG_HEADER_SIZE];
-} xlog_in_core_2_t;
-
-typedef struct xlog_in_core {
-	xlog_iclog_fields_t	hic_fields;
-	xlog_in_core_2_t	*hic_data;
+	xlog_in_core_2_t	*ic_data;
+#define ic_header	ic_data->hic_header
 } xlog_in_core_t;
 
 /*
- * Defines to save our code from this glop.
- */
-#define	ic_force_wait	hic_fields.ic_force_wait
-#define ic_write_wait	hic_fields.ic_write_wait
-#define	ic_next		hic_fields.ic_next
-#define	ic_prev		hic_fields.ic_prev
-#define	ic_bp		hic_fields.ic_bp
-#define	ic_log		hic_fields.ic_log
-#define	ic_callback	hic_fields.ic_callback
-#define	ic_callback_lock hic_fields.ic_callback_lock
-#define	ic_callback_tail hic_fields.ic_callback_tail
-#define	ic_trace	hic_fields.ic_trace
-#define	ic_size		hic_fields.ic_size
-#define	ic_offset	hic_fields.ic_offset
-#define	ic_refcnt	hic_fields.ic_refcnt
-#define	ic_bwritecnt	hic_fields.ic_bwritecnt
-#define	ic_state	hic_fields.ic_state
-#define ic_datap	hic_fields.ic_datap
-#define ic_header	hic_data->hic_header
-
-/*
  * The reservation head lsn is not made up of a cycle number and block number.
  * Instead, it uses a cycle number and byte number.  Logs don't expect to
  * overflow 31 bits worth of byte offset, so using a byte number will mean
@@ -404,6 +385,7 @@ typedef struct xlog_in_core {
 typedef struct log {
 	/* The following fields don't need locking */
 	struct xfs_mount	*l_mp;	        /* mount point */
+	struct xfs_ail		*l_ailp;	/* AIL log is working with */
 	struct xfs_buf		*l_xbuf;        /* extra buffer for log
 						 * wrapping */
 	struct xfs_buftarg	*l_targ;        /* buftarg of log */
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 70e3ba32e6b..35cca98bd94 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -36,7 +36,6 @@
 #include "xfs_dinode.h"
 #include "xfs_inode.h"
 #include "xfs_inode_item.h"
-#include "xfs_imap.h"
 #include "xfs_alloc.h"
 #include "xfs_ialloc.h"
 #include "xfs_log_priv.h"
@@ -54,10 +53,8 @@ STATIC void	xlog_recover_insert_item_backq(xlog_recover_item_t **q,
 					       xlog_recover_item_t *item);
 #if defined(DEBUG)
 STATIC void	xlog_recover_check_summary(xlog_t *);
-STATIC void	xlog_recover_check_ail(xfs_mount_t *, xfs_log_item_t *, int);
 #else
 #define	xlog_recover_check_summary(log)
-#define	xlog_recover_check_ail(mp, lip, gen)
 #endif
 
 
@@ -270,21 +267,16 @@ STATIC void
 xlog_recover_iodone(
 	struct xfs_buf	*bp)
 {
-	xfs_mount_t	*mp;
-
-	ASSERT(XFS_BUF_FSPRIVATE(bp, void *));
-
 	if (XFS_BUF_GETERROR(bp)) {
 		/*
 		 * We're not going to bother about retrying
 		 * this during recovery. One strike!
 		 */
-		mp = XFS_BUF_FSPRIVATE(bp, xfs_mount_t *);
 		xfs_ioerror_alert("xlog_recover_iodone",
-				  mp, bp, XFS_BUF_ADDR(bp));
-		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+				  bp->b_mount, bp, XFS_BUF_ADDR(bp));
+		xfs_force_shutdown(bp->b_mount, SHUTDOWN_META_IO_ERROR);
 	}
-	XFS_BUF_SET_FSPRIVATE(bp, NULL);
+	bp->b_mount = NULL;
 	XFS_BUF_CLR_IODONE_FUNC(bp);
 	xfs_biodone(bp);
 }
@@ -2228,9 +2220,8 @@ xlog_recover_do_buffer_trans(
 		XFS_BUF_STALE(bp);
 		error = xfs_bwrite(mp, bp);
 	} else {
-		ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL ||
-		       XFS_BUF_FSPRIVATE(bp, xfs_mount_t *) == mp);
-		XFS_BUF_SET_FSPRIVATE(bp, mp);
+		ASSERT(bp->b_mount == NULL || bp->b_mount == mp);
+		bp->b_mount = mp;
 		XFS_BUF_SET_IODONE_FUNC(bp, xlog_recover_iodone);
 		xfs_bdwrite(mp, bp);
 	}
@@ -2247,7 +2238,6 @@ xlog_recover_do_inode_trans(
 	xfs_inode_log_format_t	*in_f;
 	xfs_mount_t		*mp;
 	xfs_buf_t		*bp;
-	xfs_imap_t		imap;
 	xfs_dinode_t		*dip;
 	xfs_ino_t		ino;
 	int			len;
@@ -2275,54 +2265,35 @@ xlog_recover_do_inode_trans(
 	}
 	ino = in_f->ilf_ino;
 	mp = log->l_mp;
-	if (ITEM_TYPE(item) == XFS_LI_INODE) {
-		imap.im_blkno = (xfs_daddr_t)in_f->ilf_blkno;
-		imap.im_len = in_f->ilf_len;
-		imap.im_boffset = in_f->ilf_boffset;
-	} else {
-		/*
-		 * It's an old inode format record.  We don't know where
-		 * its cluster is located on disk, and we can't allow
-		 * xfs_imap() to figure it out because the inode btrees
-		 * are not ready to be used.  Therefore do not pass the
-		 * XFS_IMAP_LOOKUP flag to xfs_imap().  This will give
-		 * us only the single block in which the inode lives
-		 * rather than its cluster, so we must make sure to
-		 * invalidate the buffer when we write it out below.
-		 */
-		imap.im_blkno = 0;
-		error = xfs_imap(log->l_mp, NULL, ino, &imap, 0);
-		if (error)
-			goto error;
-	}
 
 	/*
 	 * Inode buffers can be freed, look out for it,
 	 * and do not replay the inode.
 	 */
-	if (xlog_check_buffer_cancelled(log, imap.im_blkno, imap.im_len, 0)) {
+	if (xlog_check_buffer_cancelled(log, in_f->ilf_blkno,
+					in_f->ilf_len, 0)) {
 		error = 0;
 		goto error;
 	}
 
-	bp = xfs_buf_read_flags(mp->m_ddev_targp, imap.im_blkno, imap.im_len,
-								XFS_BUF_LOCK);
+	bp = xfs_buf_read_flags(mp->m_ddev_targp, in_f->ilf_blkno,
+				in_f->ilf_len, XFS_BUF_LOCK);
 	if (XFS_BUF_ISERROR(bp)) {
 		xfs_ioerror_alert("xlog_recover_do..(read#2)", mp,
-				  bp, imap.im_blkno);
+				  bp, in_f->ilf_blkno);
 		error = XFS_BUF_GETERROR(bp);
 		xfs_buf_relse(bp);
 		goto error;
 	}
 	error = 0;
 	ASSERT(in_f->ilf_fields & XFS_ILOG_CORE);
-	dip = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset);
+	dip = (xfs_dinode_t *)xfs_buf_offset(bp, in_f->ilf_boffset);
 
 	/*
 	 * Make sure the place we're flushing out to really looks
 	 * like an inode!
 	 */
-	if (unlikely(be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC)) {
+	if (unlikely(be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC)) {
 		xfs_buf_relse(bp);
 		xfs_fs_cmn_err(CE_ALERT, mp,
 			"xfs_inode_recover: Bad inode magic number, dino ptr = 0x%p, dino bp = 0x%p, ino = %Ld",
@@ -2345,12 +2316,12 @@ xlog_recover_do_inode_trans(
 	}
 
 	/* Skip replay when the on disk inode is newer than the log one */
-	if (dicp->di_flushiter < be16_to_cpu(dip->di_core.di_flushiter)) {
+	if (dicp->di_flushiter < be16_to_cpu(dip->di_flushiter)) {
 		/*
 		 * Deal with the wrap case, DI_MAX_FLUSH is less
 		 * than smaller numbers
 		 */
-		if (be16_to_cpu(dip->di_core.di_flushiter) == DI_MAX_FLUSH &&
+		if (be16_to_cpu(dip->di_flushiter) == DI_MAX_FLUSH &&
 		    dicp->di_flushiter < (DI_MAX_FLUSH >> 1)) {
 			/* do nothing */
 		} else {
@@ -2410,7 +2381,7 @@ xlog_recover_do_inode_trans(
 		error = EFSCORRUPTED;
 		goto error;
 	}
-	if (unlikely(item->ri_buf[1].i_len > sizeof(xfs_dinode_core_t))) {
+	if (unlikely(item->ri_buf[1].i_len > sizeof(struct xfs_icdinode))) {
 		XFS_CORRUPTION_ERROR("xlog_recover_do_inode_trans(7)",
 				     XFS_ERRLEVEL_LOW, mp, dicp);
 		xfs_buf_relse(bp);
@@ -2422,23 +2393,24 @@ xlog_recover_do_inode_trans(
 	}
 
 	/* The core is in in-core format */
-	xfs_dinode_to_disk(&dip->di_core,
-		(xfs_icdinode_t *)item->ri_buf[1].i_addr);
+	xfs_dinode_to_disk(dip, (xfs_icdinode_t *)item->ri_buf[1].i_addr);
 
 	/* the rest is in on-disk format */
-	if (item->ri_buf[1].i_len > sizeof(xfs_dinode_core_t)) {
-		memcpy((xfs_caddr_t) dip + sizeof(xfs_dinode_core_t),
-			item->ri_buf[1].i_addr + sizeof(xfs_dinode_core_t),
-			item->ri_buf[1].i_len  - sizeof(xfs_dinode_core_t));
+	if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
+		memcpy((xfs_caddr_t) dip + sizeof(struct xfs_icdinode),
+			item->ri_buf[1].i_addr + sizeof(struct xfs_icdinode),
+			item->ri_buf[1].i_len  - sizeof(struct xfs_icdinode));
 	}
 
 	fields = in_f->ilf_fields;
 	switch (fields & (XFS_ILOG_DEV | XFS_ILOG_UUID)) {
 	case XFS_ILOG_DEV:
-		dip->di_u.di_dev = cpu_to_be32(in_f->ilf_u.ilfu_rdev);
+		xfs_dinode_put_rdev(dip, in_f->ilf_u.ilfu_rdev);
 		break;
 	case XFS_ILOG_UUID:
-		dip->di_u.di_muuid = in_f->ilf_u.ilfu_uuid;
+		memcpy(XFS_DFORK_DPTR(dip),
+		       &in_f->ilf_u.ilfu_uuid,
+		       sizeof(uuid_t));
 		break;
 	}
 
@@ -2454,12 +2426,12 @@ xlog_recover_do_inode_trans(
 	switch (fields & XFS_ILOG_DFORK) {
 	case XFS_ILOG_DDATA:
 	case XFS_ILOG_DEXT:
-		memcpy(&dip->di_u, src, len);
+		memcpy(XFS_DFORK_DPTR(dip), src, len);
 		break;
 
 	case XFS_ILOG_DBROOT:
-		xfs_bmbt_to_bmdr((xfs_bmbt_block_t *)src, len,
-				 &(dip->di_u.di_bmbt),
+		xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src, len,
+				 (xfs_bmdr_block_t *)XFS_DFORK_DPTR(dip),
 				 XFS_DFORK_DSIZE(dip, mp));
 		break;
 
@@ -2496,8 +2468,8 @@ xlog_recover_do_inode_trans(
 
 		case XFS_ILOG_ABROOT:
 			dest = XFS_DFORK_APTR(dip);
-			xfs_bmbt_to_bmdr((xfs_bmbt_block_t *)src, len,
-					 (xfs_bmdr_block_t*)dest,
+			xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src,
+					 len, (xfs_bmdr_block_t*)dest,
 					 XFS_DFORK_ASIZE(dip, mp));
 			break;
 
@@ -2512,9 +2484,8 @@ xlog_recover_do_inode_trans(
 
 write_inode_buffer:
 	if (ITEM_TYPE(item) == XFS_LI_INODE) {
-		ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL ||
-		       XFS_BUF_FSPRIVATE(bp, xfs_mount_t *) == mp);
-		XFS_BUF_SET_FSPRIVATE(bp, mp);
+		ASSERT(bp->b_mount == NULL || bp->b_mount == mp);
+		bp->b_mount = mp;
 		XFS_BUF_SET_IODONE_FUNC(bp, xlog_recover_iodone);
 		xfs_bdwrite(mp, bp);
 	} else {
@@ -2645,9 +2616,8 @@ xlog_recover_do_dquot_trans(
 	memcpy(ddq, recddq, item->ri_buf[1].i_len);
 
 	ASSERT(dq_f->qlf_size == 2);
-	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) == NULL ||
-	       XFS_BUF_FSPRIVATE(bp, xfs_mount_t *) == mp);
-	XFS_BUF_SET_FSPRIVATE(bp, mp);
+	ASSERT(bp->b_mount == NULL || bp->b_mount == mp);
+	bp->b_mount = mp;
 	XFS_BUF_SET_IODONE_FUNC(bp, xlog_recover_iodone);
 	xfs_bdwrite(mp, bp);
 
@@ -2689,11 +2659,11 @@ xlog_recover_do_efi_trans(
 	efip->efi_next_extent = efi_formatp->efi_nextents;
 	efip->efi_flags |= XFS_EFI_COMMITTED;
 
-	spin_lock(&mp->m_ail_lock);
+	spin_lock(&log->l_ailp->xa_lock);
 	/*
-	 * xfs_trans_update_ail() drops the AIL lock.
+	 * xfs_trans_ail_update() drops the AIL lock.
 	 */
-	xfs_trans_update_ail(mp, (xfs_log_item_t *)efip, lsn);
+	xfs_trans_ail_update(log->l_ailp, (xfs_log_item_t *)efip, lsn);
 	return 0;
 }
 
@@ -2712,12 +2682,12 @@ xlog_recover_do_efd_trans(
 	xlog_recover_item_t	*item,
 	int			pass)
 {
-	xfs_mount_t		*mp;
 	xfs_efd_log_format_t	*efd_formatp;
 	xfs_efi_log_item_t	*efip = NULL;
 	xfs_log_item_t		*lip;
-	int			gen;
 	__uint64_t		efi_id;
+	struct xfs_ail_cursor	cur;
+	struct xfs_ail		*ailp = log->l_ailp;
 
 	if (pass == XLOG_RECOVER_PASS1) {
 		return;
@@ -2734,25 +2704,26 @@ xlog_recover_do_efd_trans(
 	 * Search for the efi with the id in the efd format structure
 	 * in the AIL.
 	 */
-	mp = log->l_mp;
-	spin_lock(&mp->m_ail_lock);
-	lip = xfs_trans_first_ail(mp, &gen);
+	spin_lock(&ailp->xa_lock);
+	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
 	while (lip != NULL) {
 		if (lip->li_type == XFS_LI_EFI) {
 			efip = (xfs_efi_log_item_t *)lip;
 			if (efip->efi_format.efi_id == efi_id) {
 				/*
-				 * xfs_trans_delete_ail() drops the
+				 * xfs_trans_ail_delete() drops the
 				 * AIL lock.
 				 */
-				xfs_trans_delete_ail(mp, lip);
+				xfs_trans_ail_delete(ailp, lip);
 				xfs_efi_item_free(efip);
-				return;
+				spin_lock(&ailp->xa_lock);
+				break;
 			}
 		}
-		lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
+		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
-	spin_unlock(&mp->m_ail_lock);
+	xfs_trans_ail_cursor_done(ailp, &cur);
+	spin_unlock(&ailp->xa_lock);
 }
 
 /*
@@ -3036,33 +3007,6 @@ abort_error:
 }
 
 /*
- * Verify that once we've encountered something other than an EFI
- * in the AIL that there are no more EFIs in the AIL.
- */
-#if defined(DEBUG)
-STATIC void
-xlog_recover_check_ail(
-	xfs_mount_t		*mp,
-	xfs_log_item_t		*lip,
-	int			gen)
-{
-	int			orig_gen = gen;
-
-	do {
-		ASSERT(lip->li_type != XFS_LI_EFI);
-		lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
-		/*
-		 * The check will be bogus if we restart from the
-		 * beginning of the AIL, so ASSERT that we don't.
-		 * We never should since we're holding the AIL lock
-		 * the entire time.
-		 */
-		ASSERT(gen == orig_gen);
-	} while (lip != NULL);
-}
-#endif	/* DEBUG */
-
-/*
  * When this is called, all of the EFIs which did not have
  * corresponding EFDs should be in the AIL.  What we do now
  * is free the extents associated with each one.
@@ -3086,20 +3030,23 @@ xlog_recover_process_efis(
 {
 	xfs_log_item_t		*lip;
 	xfs_efi_log_item_t	*efip;
-	int			gen;
-	xfs_mount_t		*mp;
 	int			error = 0;
+	struct xfs_ail_cursor	cur;
+	struct xfs_ail		*ailp;
 
-	mp = log->l_mp;
-	spin_lock(&mp->m_ail_lock);
-
-	lip = xfs_trans_first_ail(mp, &gen);
+	ailp = log->l_ailp;
+	spin_lock(&ailp->xa_lock);
+	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
 	while (lip != NULL) {
 		/*
 		 * We're done when we see something other than an EFI.
+		 * There should be no EFIs left in the AIL now.
 		 */
 		if (lip->li_type != XFS_LI_EFI) {
-			xlog_recover_check_ail(mp, lip, gen);
+#ifdef DEBUG
+			for (; lip; lip = xfs_trans_ail_cursor_next(ailp, &cur))
+				ASSERT(lip->li_type != XFS_LI_EFI);
+#endif
 			break;
 		}
 
@@ -3108,18 +3055,20 @@ xlog_recover_process_efis(
 		 */
 		efip = (xfs_efi_log_item_t *)lip;
 		if (efip->efi_flags & XFS_EFI_RECOVERED) {
-			lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
+			lip = xfs_trans_ail_cursor_next(ailp, &cur);
 			continue;
 		}
 
-		spin_unlock(&mp->m_ail_lock);
-		error = xlog_recover_process_efi(mp, efip);
+		spin_unlock(&ailp->xa_lock);
+		error = xlog_recover_process_efi(log->l_mp, efip);
+		spin_lock(&ailp->xa_lock);
 		if (error)
-			return error;
-		spin_lock(&mp->m_ail_lock);
-		lip = xfs_trans_next_ail(mp, lip, &gen, NULL);
+			goto out;
+		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
-	spin_unlock(&mp->m_ail_lock);
+out:
+	xfs_trans_ail_cursor_done(ailp, &cur);
+	spin_unlock(&ailp->xa_lock);
 	return error;
 }
 
@@ -3140,19 +3089,16 @@ xlog_recover_clear_agi_bucket(
 	int		error;
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_CLEAR_AGI_BUCKET);
-	error = xfs_trans_reserve(tp, 0, XFS_CLEAR_AGI_BUCKET_LOG_RES(mp), 0, 0, 0);
-	if (!error)
-		error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
-				   XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-				   XFS_FSS_TO_BB(mp, 1), 0, &agibp);
+	error = xfs_trans_reserve(tp, 0, XFS_CLEAR_AGI_BUCKET_LOG_RES(mp),
+				  0, 0, 0);
 	if (error)
 		goto out_abort;
 
-	error = EINVAL;
-	agi = XFS_BUF_TO_AGI(agibp);
-	if (be32_to_cpu(agi->agi_magicnum) != XFS_AGI_MAGIC)
+	error = xfs_read_agi(mp, tp, agno, &agibp);
+	if (error)
 		goto out_abort;
 
+	agi = XFS_BUF_TO_AGI(agibp);
 	agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
 	offset = offsetof(xfs_agi_t, agi_unlinked) +
 		 (sizeof(xfs_agino_t) * bucket);
@@ -3172,6 +3118,62 @@ out_error:
 	return;
 }
 
+STATIC xfs_agino_t
+xlog_recover_process_one_iunlink(
+	struct xfs_mount		*mp,
+	xfs_agnumber_t			agno,
+	xfs_agino_t			agino,
+	int				bucket)
+{
+	struct xfs_buf			*ibp;
+	struct xfs_dinode		*dip;
+	struct xfs_inode		*ip;
+	xfs_ino_t			ino;
+	int				error;
+
+	ino = XFS_AGINO_TO_INO(mp, agno, agino);
+	error = xfs_iget(mp, NULL, ino, 0, 0, &ip, 0);
+	if (error)
+		goto fail;
+
+	/*
+	 * Get the on disk inode to find the next inode in the bucket.
+	 */
+	error = xfs_itobp(mp, NULL, ip, &dip, &ibp, XFS_BUF_LOCK);
+	if (error)
+		goto fail_iput;
+
+	ASSERT(ip->i_d.di_nlink == 0);
+	ASSERT(ip->i_d.di_mode != 0);
+
+	/* setup for the next pass */
+	agino = be32_to_cpu(dip->di_next_unlinked);
+	xfs_buf_relse(ibp);
+
+	/*
+	 * Prevent any DMAPI event from being sent when the reference on
+	 * the inode is dropped.
+	 */
+	ip->i_d.di_dmevmask = 0;
+
+	IRELE(ip);
+	return agino;
+
+ fail_iput:
+	IRELE(ip);
+ fail:
+	/*
+	 * We can't read in the inode this bucket points to, or this inode
+	 * is messed up.  Just ditch this bucket of inodes.  We will lose
+	 * some inodes and space, but at least we won't hang.
+	 *
+	 * Call xlog_recover_clear_agi_bucket() to perform a transaction to
+	 * clear the inode pointer in the bucket.
+	 */
+	xlog_recover_clear_agi_bucket(mp, agno, bucket);
+	return NULLAGINO;
+}
+
 /*
  * xlog_iunlink_recover
  *
@@ -3192,11 +3194,7 @@ xlog_recover_process_iunlinks(
 	xfs_agnumber_t	agno;
 	xfs_agi_t	*agi;
 	xfs_buf_t	*agibp;
-	xfs_buf_t	*ibp;
-	xfs_dinode_t	*dip;
-	xfs_inode_t	*ip;
 	xfs_agino_t	agino;
-	xfs_ino_t	ino;
 	int		bucket;
 	int		error;
 	uint		mp_dmevmask;
@@ -3213,22 +3211,21 @@ xlog_recover_process_iunlinks(
 		/*
 		 * Find the agi for this ag.
 		 */
-		agibp = xfs_buf_read(mp->m_ddev_targp,
-				XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
-				XFS_FSS_TO_BB(mp, 1), 0);
-		if (XFS_BUF_ISERROR(agibp)) {
-			xfs_ioerror_alert("xlog_recover_process_iunlinks(#1)",
-				log->l_mp, agibp,
-				XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)));
+		error = xfs_read_agi(mp, NULL, agno, &agibp);
+		if (error) {
+			/*
+			 * AGI is b0rked. Don't process it.
+			 *
+			 * We should probably mark the filesystem as corrupt
+			 * after we've recovered all the ag's we can....
+			 */
+			continue;
 		}
 		agi = XFS_BUF_TO_AGI(agibp);
-		ASSERT(XFS_AGI_MAGIC == be32_to_cpu(agi->agi_magicnum));
 
 		for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++) {
-
 			agino = be32_to_cpu(agi->agi_unlinked[bucket]);
 			while (agino != NULLAGINO) {
-
 				/*
 				 * Release the agi buffer so that it can
 				 * be acquired in the normal course of the
@@ -3236,87 +3233,17 @@ xlog_recover_process_iunlinks(
 				 */
 				xfs_buf_relse(agibp);
 
-				ino = XFS_AGINO_TO_INO(mp, agno, agino);
-				error = xfs_iget(mp, NULL, ino, 0, 0, &ip, 0);
-				ASSERT(error || (ip != NULL));
-
-				if (!error) {
-					/*
-					 * Get the on disk inode to find the
-					 * next inode in the bucket.
-					 */
-					error = xfs_itobp(mp, NULL, ip, &dip,
-							&ibp, 0, 0,
-							XFS_BUF_LOCK);
-					ASSERT(error || (dip != NULL));
-				}
-
-				if (!error) {
-					ASSERT(ip->i_d.di_nlink == 0);
-
-					/* setup for the next pass */
-					agino = be32_to_cpu(
-							dip->di_next_unlinked);
-					xfs_buf_relse(ibp);
-					/*
-					 * Prevent any DMAPI event from
-					 * being sent when the
-					 * reference on the inode is
-					 * dropped.
-					 */
-					ip->i_d.di_dmevmask = 0;
-
-					/*
-					 * If this is a new inode, handle
-					 * it specially.  Otherwise,
-					 * just drop our reference to the
-					 * inode.  If there are no
-					 * other references, this will
-					 * send the inode to
-					 * xfs_inactive() which will
-					 * truncate the file and free
-					 * the inode.
-					 */
-					if (ip->i_d.di_mode == 0)
-						xfs_iput_new(ip, 0);
-					else
-						IRELE(ip);
-				} else {
-					/*
-					 * We can't read in the inode
-					 * this bucket points to, or
-					 * this inode is messed up.  Just
-					 * ditch this bucket of inodes.  We
-					 * will lose some inodes and space,
-					 * but at least we won't hang.  Call
-					 * xlog_recover_clear_agi_bucket()
-					 * to perform a transaction to clear
-					 * the inode pointer in the bucket.
-					 */
-					xlog_recover_clear_agi_bucket(mp, agno,
-							bucket);
-
-					agino = NULLAGINO;
-				}
+				agino = xlog_recover_process_one_iunlink(mp,
+							agno, agino, bucket);
 
 				/*
 				 * Reacquire the agibuffer and continue around
-				 * the loop.
+				 * the loop. This should never fail as we know
+				 * the buffer was good earlier on.
 				 */
-				agibp = xfs_buf_read(mp->m_ddev_targp,
-						XFS_AG_DADDR(mp, agno,
-							XFS_AGI_DADDR(mp)),
-						XFS_FSS_TO_BB(mp, 1), 0);
-				if (XFS_BUF_ISERROR(agibp)) {
-					xfs_ioerror_alert(
-				"xlog_recover_process_iunlinks(#2)",
-						log->l_mp, agibp,
-						XFS_AG_DADDR(mp, agno,
-							XFS_AGI_DADDR(mp)));
-				}
+				error = xfs_read_agi(mp, NULL, agno, &agibp);
+				ASSERT(error == 0);
 				agi = XFS_BUF_TO_AGI(agibp);
-				ASSERT(XFS_AGI_MAGIC == be32_to_cpu(
-					agi->agi_magicnum));
 			}
 		}
 
@@ -3367,7 +3294,6 @@ xlog_pack_data(
 	int			size = iclog->ic_offset + roundoff;
 	__be32			cycle_lsn;
 	xfs_caddr_t		dp;
-	xlog_in_core_2_t	*xhdr;
 
 	xlog_pack_data_checksum(log, iclog, size);
 
@@ -3382,7 +3308,8 @@ xlog_pack_data(
 	}
 
 	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) {
-		xhdr = (xlog_in_core_2_t *)&iclog->ic_header;
+		xlog_in_core_2_t *xhdr = iclog->ic_data;
+
 		for ( ; i < BTOBB(size); i++) {
 			j = i / (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
 			k = i % (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
@@ -3440,7 +3367,6 @@ xlog_unpack_data(
 	xlog_t			*log)
 {
 	int			i, j, k;
-	xlog_in_core_2_t	*xhdr;
 
 	for (i = 0; i < BTOBB(be32_to_cpu(rhead->h_len)) &&
 		  i < (XLOG_HEADER_CYCLE_SIZE / BBSIZE); i++) {
@@ -3449,7 +3375,7 @@ xlog_unpack_data(
 	}
 
 	if (xfs_sb_version_haslogv2(&log->l_mp->m_sb)) {
-		xhdr = (xlog_in_core_2_t *)rhead;
+		xlog_in_core_2_t *xhdr = (xlog_in_core_2_t *)rhead;
 		for ( ; i < BTOBB(be32_to_cpu(rhead->h_len)); i++) {
 			j = i / (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
 			k = i % (XLOG_HEADER_CYCLE_SIZE / BBSIZE);
@@ -4003,11 +3929,8 @@ xlog_recover_check_summary(
 {
 	xfs_mount_t	*mp;
 	xfs_agf_t	*agfp;
-	xfs_agi_t	*agip;
 	xfs_buf_t	*agfbp;
 	xfs_buf_t	*agibp;
-	xfs_daddr_t	agfdaddr;
-	xfs_daddr_t	agidaddr;
 	xfs_buf_t	*sbbp;
 #ifdef XFS_LOUD_RECOVERY
 	xfs_sb_t	*sbp;
@@ -4016,6 +3939,7 @@ xlog_recover_check_summary(
 	__uint64_t	freeblks;
 	__uint64_t	itotal;
 	__uint64_t	ifree;
+	int		error;
 
 	mp = log->l_mp;
 
@@ -4023,37 +3947,27 @@ xlog_recover_check_summary(
 	itotal = 0LL;
 	ifree = 0LL;
 	for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) {
-		agfdaddr = XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp));
-		agfbp = xfs_buf_read(mp->m_ddev_targp, agfdaddr,
-				XFS_FSS_TO_BB(mp, 1), 0);
-		if (XFS_BUF_ISERROR(agfbp)) {
-			xfs_ioerror_alert("xlog_recover_check_summary(agf)",
-						mp, agfbp, agfdaddr);
-		}
-		agfp = XFS_BUF_TO_AGF(agfbp);
-		ASSERT(XFS_AGF_MAGIC == be32_to_cpu(agfp->agf_magicnum));
-		ASSERT(XFS_AGF_GOOD_VERSION(be32_to_cpu(agfp->agf_versionnum)));
-		ASSERT(be32_to_cpu(agfp->agf_seqno) == agno);
-
-		freeblks += be32_to_cpu(agfp->agf_freeblks) +
-			    be32_to_cpu(agfp->agf_flcount);
-		xfs_buf_relse(agfbp);
-
-		agidaddr = XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp));
-		agibp = xfs_buf_read(mp->m_ddev_targp, agidaddr,
-				XFS_FSS_TO_BB(mp, 1), 0);
-		if (XFS_BUF_ISERROR(agibp)) {
-			xfs_ioerror_alert("xlog_recover_check_summary(agi)",
-					  mp, agibp, agidaddr);
+		error = xfs_read_agf(mp, NULL, agno, 0, &agfbp);
+		if (error) {
+			xfs_fs_cmn_err(CE_ALERT, mp,
+					"xlog_recover_check_summary(agf)"
+					"agf read failed agno %d error %d",
+							agno, error);
+		} else {
+			agfp = XFS_BUF_TO_AGF(agfbp);
+			freeblks += be32_to_cpu(agfp->agf_freeblks) +
+				    be32_to_cpu(agfp->agf_flcount);
+			xfs_buf_relse(agfbp);
 		}
-		agip = XFS_BUF_TO_AGI(agibp);
-		ASSERT(XFS_AGI_MAGIC == be32_to_cpu(agip->agi_magicnum));
-		ASSERT(XFS_AGI_GOOD_VERSION(be32_to_cpu(agip->agi_versionnum)));
-		ASSERT(be32_to_cpu(agip->agi_seqno) == agno);
 
-		itotal += be32_to_cpu(agip->agi_count);
-		ifree += be32_to_cpu(agip->agi_freecount);
-		xfs_buf_relse(agibp);
+		error = xfs_read_agi(mp, NULL, agno, &agibp);
+		if (!error) {
+			struct xfs_agi	*agi = XFS_BUF_TO_AGI(agibp);
+
+			itotal += be32_to_cpu(agi->agi_count);
+			ifree += be32_to_cpu(agi->agi_freecount);
+			xfs_buf_relse(agibp);
+		}
 	}
 
 	sbbp = xfs_getsb(mp, 0);
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 15f5dd22fbb..3c97c6463a4 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -567,8 +567,6 @@ xfs_readsb(xfs_mount_t *mp, int flags)
 STATIC void
 xfs_mount_common(xfs_mount_t *mp, xfs_sb_t *sbp)
 {
-	int	i;
-
 	mp->m_agfrotor = mp->m_agirotor = 0;
 	spin_lock_init(&mp->m_agirotor_lock);
 	mp->m_maxagi = mp->m_sb.sb_agcount;
@@ -577,12 +575,10 @@ xfs_mount_common(xfs_mount_t *mp, xfs_sb_t *sbp)
 	mp->m_sectbb_log = sbp->sb_sectlog - BBSHIFT;
 	mp->m_agno_log = xfs_highbit32(sbp->sb_agcount - 1) + 1;
 	mp->m_agino_log = sbp->sb_inopblog + sbp->sb_agblklog;
-	mp->m_litino = sbp->sb_inodesize -
-		((uint)sizeof(xfs_dinode_core_t) + (uint)sizeof(xfs_agino_t));
+	mp->m_litino = sbp->sb_inodesize - sizeof(struct xfs_dinode);
 	mp->m_blockmask = sbp->sb_blocksize - 1;
 	mp->m_blockwsize = sbp->sb_blocksize >> XFS_WORDLOG;
 	mp->m_blockwmask = mp->m_blockwsize - 1;
-	INIT_LIST_HEAD(&mp->m_del_inodes);
 
 	/*
 	 * Setup for attributes, in case they get created.
@@ -605,24 +601,20 @@ xfs_mount_common(xfs_mount_t *mp, xfs_sb_t *sbp)
 	}
 	ASSERT(mp->m_attroffset < XFS_LITINO(mp));
 
-	for (i = 0; i < 2; i++) {
-		mp->m_alloc_mxr[i] = XFS_BTREE_BLOCK_MAXRECS(sbp->sb_blocksize,
-			xfs_alloc, i == 0);
-		mp->m_alloc_mnr[i] = XFS_BTREE_BLOCK_MINRECS(sbp->sb_blocksize,
-			xfs_alloc, i == 0);
-	}
-	for (i = 0; i < 2; i++) {
-		mp->m_bmap_dmxr[i] = XFS_BTREE_BLOCK_MAXRECS(sbp->sb_blocksize,
-			xfs_bmbt, i == 0);
-		mp->m_bmap_dmnr[i] = XFS_BTREE_BLOCK_MINRECS(sbp->sb_blocksize,
-			xfs_bmbt, i == 0);
-	}
-	for (i = 0; i < 2; i++) {
-		mp->m_inobt_mxr[i] = XFS_BTREE_BLOCK_MAXRECS(sbp->sb_blocksize,
-			xfs_inobt, i == 0);
-		mp->m_inobt_mnr[i] = XFS_BTREE_BLOCK_MINRECS(sbp->sb_blocksize,
-			xfs_inobt, i == 0);
-	}
+	mp->m_alloc_mxr[0] = xfs_allocbt_maxrecs(mp, sbp->sb_blocksize, 1);
+	mp->m_alloc_mxr[1] = xfs_allocbt_maxrecs(mp, sbp->sb_blocksize, 0);
+	mp->m_alloc_mnr[0] = mp->m_alloc_mxr[0] / 2;
+	mp->m_alloc_mnr[1] = mp->m_alloc_mxr[1] / 2;
+
+	mp->m_inobt_mxr[0] = xfs_inobt_maxrecs(mp, sbp->sb_blocksize, 1);
+	mp->m_inobt_mxr[1] = xfs_inobt_maxrecs(mp, sbp->sb_blocksize, 0);
+	mp->m_inobt_mnr[0] = mp->m_inobt_mxr[0] / 2;
+	mp->m_inobt_mnr[1] = mp->m_inobt_mxr[1] / 2;
+
+	mp->m_bmap_dmxr[0] = xfs_bmbt_maxrecs(mp, sbp->sb_blocksize, 1);
+	mp->m_bmap_dmxr[1] = xfs_bmbt_maxrecs(mp, sbp->sb_blocksize, 0);
+	mp->m_bmap_dmnr[0] = mp->m_bmap_dmxr[0] / 2;
+	mp->m_bmap_dmnr[1] = mp->m_bmap_dmxr[1] / 2;
 
 	mp->m_bsize = XFS_FSB_TO_BB(mp, 1);
 	mp->m_ialloc_inos = (int)MAX((__uint16_t)XFS_INODES_PER_CHUNK,
@@ -1228,6 +1220,16 @@ xfs_unmountfs(
 	__uint64_t		resblks;
 	int			error;
 
+	/*
+	 * Release dquot that rootinode, rbmino and rsumino might be holding,
+	 * and release the quota inodes.
+	 */
+	XFS_QM_UNMOUNT(mp);
+
+	if (mp->m_rbmip)
+		IRELE(mp->m_rbmip);
+	if (mp->m_rsumip)
+		IRELE(mp->m_rsumip);
 	IRELE(mp->m_rootip);
 
 	/*
@@ -1241,7 +1243,7 @@ xfs_unmountfs(
 	 * need to force the log first.
 	 */
 	xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE | XFS_LOG_SYNC);
-	xfs_iflush_all(mp);
+	xfs_reclaim_inodes(mp, 0, XFS_IFLUSH_ASYNC);
 
 	XFS_QM_DQPURGEALL(mp, XFS_QMOPT_QUOTALL | XFS_QMOPT_UMOUNTING);
 
@@ -1288,11 +1290,6 @@ xfs_unmountfs(
 	xfs_unmountfs_wait(mp); 		/* wait for async bufs */
 	xfs_log_unmount(mp);			/* Done! No more fs ops. */
 
-	/*
-	 * All inodes from this mount point should be freed.
-	 */
-	ASSERT(mp->m_inodes == NULL);
-
 	if ((mp->m_flags & XFS_MOUNT_NOUUID) == 0)
 		uuid_table_remove(&mp->m_sb.sb_uuid);
 
@@ -1365,24 +1362,6 @@ xfs_log_sbcount(
 	return error;
 }
 
-STATIC void
-xfs_mark_shared_ro(
-	xfs_mount_t	*mp,
-	xfs_buf_t	*bp)
-{
-	xfs_dsb_t	*sb = XFS_BUF_TO_SBP(bp);
-	__uint16_t	version;
-
-	if (!(sb->sb_flags & XFS_SBF_READONLY))
-		sb->sb_flags |= XFS_SBF_READONLY;
-
-	version = be16_to_cpu(sb->sb_versionnum);
-	if ((version & XFS_SB_VERSION_NUMBITS) != XFS_SB_VERSION_4 ||
-	    !(version & XFS_SB_VERSION_SHAREDBIT))
-		version |= XFS_SB_VERSION_SHAREDBIT;
-	sb->sb_versionnum = cpu_to_be16(version);
-}
-
 int
 xfs_unmountfs_writesb(xfs_mount_t *mp)
 {
@@ -1398,12 +1377,6 @@ xfs_unmountfs_writesb(xfs_mount_t *mp)
 
 		sbp = xfs_getsb(mp, 0);
 
-		/*
-		 * mark shared-readonly if desired
-		 */
-		if (mp->m_mk_sharedro)
-			xfs_mark_shared_ro(mp, sbp);
-
 		XFS_BUF_UNDONE(sbp);
 		XFS_BUF_UNREAD(sbp);
 		XFS_BUF_UNDELAYWRITE(sbp);
@@ -1415,8 +1388,6 @@ xfs_unmountfs_writesb(xfs_mount_t *mp)
 		if (error)
 			xfs_ioerror_alert("xfs_unmountfs_writesb",
 					  mp, sbp, XFS_BUF_ADDR(sbp));
-		if (error && mp->m_mk_sharedro)
-			xfs_fs_cmn_err(CE_ALERT, mp, "Superblock write error detected while unmounting.  Filesystem may not be marked shared readonly");
 		xfs_buf_relse(sbp);
 	}
 	return error;
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index f3c1024b124..c1e02846732 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -18,7 +18,6 @@
 #ifndef __XFS_MOUNT_H__
 #define	__XFS_MOUNT_H__
 
-
 typedef struct xfs_trans_reservations {
 	uint	tr_write;	/* extent alloc trans */
 	uint	tr_itruncate;	/* truncate trans */
@@ -44,14 +43,16 @@ typedef struct xfs_trans_reservations {
 } xfs_trans_reservations_t;
 
 #ifndef __KERNEL__
-/*
- * Moved here from xfs_ag.h to avoid reordering header files
- */
+
 #define XFS_DADDR_TO_AGNO(mp,d) \
 	((xfs_agnumber_t)(XFS_BB_TO_FSBT(mp, d) / (mp)->m_sb.sb_agblocks))
 #define XFS_DADDR_TO_AGBNO(mp,d) \
 	((xfs_agblock_t)(XFS_BB_TO_FSBT(mp, d) % (mp)->m_sb.sb_agblocks))
-#else
+
+#else /* __KERNEL__ */
+
+#include "xfs_sync.h"
+
 struct cred;
 struct log;
 struct xfs_mount_args;
@@ -62,6 +63,7 @@ struct xfs_extdelta;
 struct xfs_swapext;
 struct xfs_mru_cache;
 struct xfs_nameops;
+struct xfs_ail;
 
 /*
  * Prototypes and functions for the Data Migration subsystem.
@@ -115,7 +117,7 @@ struct xfs_quotainfo;
 
 typedef int	(*xfs_qminit_t)(struct xfs_mount *, uint *, uint *);
 typedef int	(*xfs_qmmount_t)(struct xfs_mount *, uint, uint);
-typedef int	(*xfs_qmunmount_t)(struct xfs_mount *);
+typedef void	(*xfs_qmunmount_t)(struct xfs_mount *);
 typedef void	(*xfs_qmdone_t)(struct xfs_mount *);
 typedef void	(*xfs_dqrele_t)(struct xfs_dquot *);
 typedef int	(*xfs_dqattach_t)(struct xfs_inode *, uint);
@@ -132,7 +134,7 @@ typedef struct xfs_dquot * (*xfs_dqvopchown_t)(
 			struct xfs_dquot **, struct xfs_dquot *);
 typedef int	(*xfs_dqvopchownresv_t)(struct xfs_trans *, struct xfs_inode *,
 			struct xfs_dquot *, struct xfs_dquot *, uint);
-typedef void	(*xfs_dqstatvfs_t)(struct xfs_inode *, bhv_statvfs_t *);
+typedef void	(*xfs_dqstatvfs_t)(struct xfs_inode *, struct kstatfs *);
 typedef int	(*xfs_dqsync_t)(struct xfs_mount *, int flags);
 typedef int	(*xfs_quotactl_t)(struct xfs_mount *, int, int, xfs_caddr_t);
 
@@ -223,18 +225,10 @@ extern void	xfs_icsb_sync_counters_locked(struct xfs_mount *, int);
 #define xfs_icsb_sync_counters_locked(mp, flags) do { } while (0)
 #endif
 
-typedef struct xfs_ail {
-	struct list_head	xa_ail;
-	uint			xa_gen;
-	struct task_struct	*xa_task;
-	xfs_lsn_t		xa_target;
-} xfs_ail_t;
-
 typedef struct xfs_mount {
 	struct super_block	*m_super;
 	xfs_tid_t		m_tid;		/* next unused tid for fs */
-	spinlock_t		m_ail_lock;	/* fs AIL mutex */
-	xfs_ail_t		m_ail;		/* fs active log item list */
+	struct xfs_ail		*m_ail;		/* fs active log item list */
 	xfs_sb_t		m_sb;		/* copy of fs superblock */
 	spinlock_t		m_sb_lock;	/* sb counter lock */
 	struct xfs_buf		*m_sb_bp;	/* buffer for superblock */
@@ -247,10 +241,6 @@ typedef struct xfs_mount {
 	xfs_agnumber_t		m_agirotor;	/* last ag dir inode alloced */
 	spinlock_t		m_agirotor_lock;/* .. and lock protecting it */
 	xfs_agnumber_t		m_maxagi;	/* highest inode alloc group */
-	struct xfs_inode	*m_inodes;	/* active inode list */
-	struct list_head	m_del_inodes;	/* inodes to reclaim */
-	mutex_t			m_ilock;	/* inode list mutex */
-	uint			m_ireclaims;	/* count of calls to reclaim*/
 	uint			m_readio_log;	/* min read size log bytes */
 	uint			m_readio_blocks; /* min read size blocks */
 	uint			m_writeio_log;	/* min write size log bytes */
@@ -267,7 +257,6 @@ typedef struct xfs_mount {
 	xfs_buftarg_t		*m_ddev_targp;	/* saves taking the address */
 	xfs_buftarg_t		*m_logdev_targp;/* ptr to log device */
 	xfs_buftarg_t		*m_rtdev_targp;	/* ptr to rt device */
-	__uint8_t		m_dircook_elog;	/* log d-cookie entry bits */
 	__uint8_t		m_blkbit_log;	/* blocklog + NBBY */
 	__uint8_t		m_blkbb_log;	/* blocklog - BBSHIFT */
 	__uint8_t		m_agno_log;	/* log #ag's */
@@ -276,12 +265,12 @@ typedef struct xfs_mount {
 	uint			m_blockmask;	/* sb_blocksize-1 */
 	uint			m_blockwsize;	/* sb_blocksize in words */
 	uint			m_blockwmask;	/* blockwsize-1 */
-	uint			m_alloc_mxr[2];	/* XFS_ALLOC_BLOCK_MAXRECS */
-	uint			m_alloc_mnr[2];	/* XFS_ALLOC_BLOCK_MINRECS */
-	uint			m_bmap_dmxr[2];	/* XFS_BMAP_BLOCK_DMAXRECS */
-	uint			m_bmap_dmnr[2];	/* XFS_BMAP_BLOCK_DMINRECS */
-	uint			m_inobt_mxr[2];	/* XFS_INOBT_BLOCK_MAXRECS */
-	uint			m_inobt_mnr[2];	/* XFS_INOBT_BLOCK_MINRECS */
+	uint			m_alloc_mxr[2];	/* max alloc btree records */
+	uint			m_alloc_mnr[2];	/* min alloc btree records */
+	uint			m_bmap_dmxr[2];	/* max bmap btree records */
+	uint			m_bmap_dmnr[2];	/* min bmap btree records */
+	uint			m_inobt_mxr[2];	/* max inobt btree records */
+	uint			m_inobt_mnr[2];	/* min inobt btree records */
 	uint			m_ag_maxlevels;	/* XFS_AG_MAXLEVELS */
 	uint			m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */
 	uint			m_in_maxlevels;	/* XFS_IN_MAXLEVELS */
@@ -312,9 +301,6 @@ typedef struct xfs_mount {
 	int			m_sinoalign;	/* stripe unit inode alignment */
 	int			m_attr_magicpct;/* 37% of the blocksize */
 	int			m_dir_magicpct;	/* 37% of the dir blocksize */
-	__uint8_t		m_mk_sharedro;	/* mark shared ro on unmount */
-	__uint8_t		m_inode_quiesce;/* call quiesce on new inodes.
-						   field governed by m_ilock */
 	__uint8_t		m_sectbb_log;	/* sectlog - BBSHIFT */
 	const struct xfs_nameops *m_dirnameops;	/* vector of dir name ops */
 	int			m_dirblksize;	/* directory block sz--bytes */
@@ -362,7 +348,6 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_ATTR2		(1ULL << 8)	/* allow use of attr2 format */
 #define XFS_MOUNT_GRPID		(1ULL << 9)	/* group-ID assigned from directory */
 #define XFS_MOUNT_NORECOVERY	(1ULL << 10)	/* no recovery - dirty fs */
-#define XFS_MOUNT_SHARED	(1ULL << 11)	/* shared mount */
 #define XFS_MOUNT_DFLT_IOSIZE	(1ULL << 12)	/* set default i/o size */
 #define XFS_MOUNT_OSYNCISOSYNC	(1ULL << 13)	/* o_sync is REALLY o_sync */
 						/* osyncisdsync is now default*/
@@ -439,6 +424,16 @@ void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
 #define xfs_force_shutdown(m,f)	\
 	xfs_do_force_shutdown(m, f, __FILE__, __LINE__)
 
+#define SHUTDOWN_META_IO_ERROR	0x0001	/* write attempt to metadata failed */
+#define SHUTDOWN_LOG_IO_ERROR	0x0002	/* write attempt to the log failed */
+#define SHUTDOWN_FORCE_UMOUNT	0x0004	/* shutdown from a forced unmount */
+#define SHUTDOWN_CORRUPT_INCORE	0x0008	/* corrupt in-memory data structures */
+#define SHUTDOWN_REMOTE_REQ	0x0010	/* shutdown came from remote cell */
+#define SHUTDOWN_DEVICE_REQ	0x0020	/* failed all paths to the device */
+
+#define xfs_test_for_freeze(mp)		((mp)->m_super->s_frozen)
+#define xfs_wait_for_freeze(mp,l)	vfs_check_frozen((mp)->m_super, (l))
+
 /*
  * Flags for xfs_mountfs
  */
@@ -508,14 +503,12 @@ typedef struct xfs_mod_sb {
 #define	XFS_MOUNT_ILOCK(mp)	mutex_lock(&((mp)->m_ilock))
 #define	XFS_MOUNT_IUNLOCK(mp)	mutex_unlock(&((mp)->m_ilock))
 
-extern void	xfs_mod_sb(xfs_trans_t *, __int64_t);
 extern int	xfs_log_sbcount(xfs_mount_t *, uint);
 extern int	xfs_mountfs(xfs_mount_t *mp);
 extern void	xfs_mountfs_check_barriers(xfs_mount_t *mp);
 
 extern void	xfs_unmountfs(xfs_mount_t *);
 extern int	xfs_unmountfs_writesb(xfs_mount_t *);
-extern int	xfs_unmount_flush(xfs_mount_t *, int);
 extern int	xfs_mod_incore_sb(xfs_mount_t *, xfs_sb_field_t, int64_t, int);
 extern int	xfs_mod_incore_sb_unlocked(xfs_mount_t *, xfs_sb_field_t,
 			int64_t, int);
@@ -525,20 +518,20 @@ extern struct xfs_buf *xfs_getsb(xfs_mount_t *, int);
 extern int	xfs_readsb(xfs_mount_t *, int);
 extern void	xfs_freesb(xfs_mount_t *);
 extern int	xfs_fs_writable(xfs_mount_t *);
-extern int	xfs_syncsub(xfs_mount_t *, int, int *);
-extern int	xfs_sync_inodes(xfs_mount_t *, int, int *);
-extern xfs_agnumber_t	xfs_initialize_perag(xfs_mount_t *, xfs_agnumber_t);
-extern void	xfs_sb_from_disk(struct xfs_sb *, struct xfs_dsb *);
-extern void	xfs_sb_to_disk(struct xfs_dsb *, struct xfs_sb *, __int64_t);
 extern int	xfs_sb_validate_fsb_count(struct xfs_sb *, __uint64_t);
 
-extern int	xfs_dmops_get(struct xfs_mount *, struct xfs_mount_args *);
+extern int	xfs_dmops_get(struct xfs_mount *);
 extern void	xfs_dmops_put(struct xfs_mount *);
-extern int	xfs_qmops_get(struct xfs_mount *, struct xfs_mount_args *);
+extern int	xfs_qmops_get(struct xfs_mount *);
 extern void	xfs_qmops_put(struct xfs_mount *);
 
 extern struct xfs_dmops xfs_dmcore_xfs;
 
 #endif	/* __KERNEL__ */
 
+extern void	xfs_mod_sb(struct xfs_trans *, __int64_t);
+extern xfs_agnumber_t	xfs_initialize_perag(struct xfs_mount *, xfs_agnumber_t);
+extern void	xfs_sb_from_disk(struct xfs_sb *, struct xfs_dsb *);
+extern void	xfs_sb_to_disk(struct xfs_dsb *, struct xfs_sb *, __int64_t);
+
 #endif	/* __XFS_MOUNT_H__ */
diff --git a/fs/xfs/xfs_qmops.c b/fs/xfs/xfs_qmops.c
index a294e58db8d..27f80581520 100644
--- a/fs/xfs/xfs_qmops.c
+++ b/fs/xfs/xfs_qmops.c
@@ -28,7 +28,6 @@
 #include "xfs_mount.h"
 #include "xfs_quota.h"
 #include "xfs_error.h"
-#include "xfs_clnt.h"
 
 
 STATIC struct xfs_dquot *
@@ -131,9 +130,9 @@ static struct xfs_qmops xfs_qmcore_stub = {
 };
 
 int
-xfs_qmops_get(struct xfs_mount *mp, struct xfs_mount_args *args)
+xfs_qmops_get(struct xfs_mount *mp)
 {
-	if (args->flags & (XFSMNT_UQUOTA | XFSMNT_PQUOTA | XFSMNT_GQUOTA)) {
+	if (XFS_IS_QUOTA_RUNNING(mp)) {
 #ifdef CONFIG_XFS_QUOTA
 		mp->m_qm_ops = &xfs_qmcore_xfs;
 #else
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index 12c4ec775af..48965ecaa15 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -84,11 +84,9 @@ typedef struct xfs_dqblk {
 #define XFS_DQ_USER		0x0001		/* a user quota */
 #define XFS_DQ_PROJ		0x0002		/* project quota */
 #define XFS_DQ_GROUP		0x0004		/* a group quota */
-#define XFS_DQ_FLOCKED		0x0008		/* flush lock taken */
-#define XFS_DQ_DIRTY		0x0010		/* dquot is dirty */
-#define XFS_DQ_WANT		0x0020		/* for lookup/reclaim race */
-#define XFS_DQ_INACTIVE		0x0040		/* dq off mplist & hashlist */
-#define XFS_DQ_MARKER		0x0080		/* sentinel */
+#define XFS_DQ_DIRTY		0x0008		/* dquot is dirty */
+#define XFS_DQ_WANT		0x0010		/* for lookup/reclaim race */
+#define XFS_DQ_INACTIVE		0x0020		/* dq off mplist & hashlist */
 
 #define XFS_DQ_ALLTYPES		(XFS_DQ_USER|XFS_DQ_PROJ|XFS_DQ_GROUP)
 
diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c
index c903130be7f..86471bb40fd 100644
--- a/fs/xfs/xfs_rename.c
+++ b/fs/xfs/xfs_rename.c
@@ -42,31 +42,6 @@
 
 
 /*
- * Given an array of up to 4 inode pointers, unlock the pointed to inodes.
- * If there are fewer than 4 entries in the array, the empty entries will
- * be at the end and will have NULL pointers in them.
- */
-STATIC void
-xfs_rename_unlock4(
-	xfs_inode_t	**i_tab,
-	uint		lock_mode)
-{
-	int	i;
-
-	xfs_iunlock(i_tab[0], lock_mode);
-	for (i = 1; i < 4; i++) {
-		if (i_tab[i] == NULL)
-			break;
-
-		/*
-		 * Watch out for duplicate entries in the table.
-		 */
-		if (i_tab[i] != i_tab[i-1])
-			xfs_iunlock(i_tab[i], lock_mode);
-	}
-}
-
-/*
  * Enter all inodes for a rename transaction into a sorted array.
  */
 STATIC void
@@ -205,19 +180,6 @@ xfs_rename(
 	xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL);
 
 	/*
-	 * If we are using project inheritance, we only allow renames
-	 * into our tree when the project IDs are the same; else the
-	 * tree quota mechanism would be circumvented.
-	 */
-	if (unlikely((target_dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
-		     (target_dp->i_d.di_projid != src_ip->i_d.di_projid))) {
-		error = XFS_ERROR(EXDEV);
-		xfs_rename_unlock4(inodes, XFS_ILOCK_EXCL);
-		xfs_trans_cancel(tp, cancel_flags);
-		goto std_return;
-	}
-
-	/*
 	 * Join all the inodes to the transaction. From this point on,
 	 * we can rely on either trans_commit or trans_cancel to unlock
 	 * them.  Note that we need to add a vnode reference to the
@@ -242,6 +204,17 @@ xfs_rename(
 	}
 
 	/*
+	 * If we are using project inheritance, we only allow renames
+	 * into our tree when the project IDs are the same; else the
+	 * tree quota mechanism would be circumvented.
+	 */
+	if (unlikely((target_dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
+		     (target_dp->i_d.di_projid != src_ip->i_d.di_projid))) {
+		error = XFS_ERROR(EXDEV);
+		goto error_return;
+	}
+
+	/*
 	 * Set up the target.
 	 */
 	if (target_ip == NULL) {
@@ -367,19 +340,11 @@ xfs_rename(
 					&first_block, &free_list, spaceres);
 	if (error)
 		goto abort_return;
-	xfs_ichgtime(src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 
-	/*
-	 * Update the generation counts on all the directory inodes
-	 * that we're modifying.
-	 */
-	src_dp->i_gen++;
+	xfs_ichgtime(src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE);
-
-	if (new_parent) {
-		target_dp->i_gen++;
+	if (new_parent)
 		xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE);
-	}
 
 	/*
 	 * If this is a synchronous mount, make sure that the
diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c
index e2f68de1615..edf12c7b834 100644
--- a/fs/xfs/xfs_rtalloc.c
+++ b/fs/xfs/xfs_rtalloc.c
@@ -85,7 +85,6 @@ xfs_growfs_rt_alloc(
 {
 	xfs_fileoff_t	bno;		/* block number in file */
 	xfs_buf_t	*bp;		/* temporary buffer for zeroing */
-	int		cancelflags;	/* flags for xfs_trans_cancel */
 	int		committed;	/* transaction committed flag */
 	xfs_daddr_t	d;		/* disk block address */
 	int		error;		/* error return value */
@@ -96,15 +95,16 @@ xfs_growfs_rt_alloc(
 	xfs_bmbt_irec_t	map;		/* block map output */
 	int		nmap;		/* number of block maps */
 	int		resblks;	/* space reservation */
-	xfs_trans_t	*tp;		/* transaction pointer */
 
 	/*
 	 * Allocate space to the file, as necessary.
 	 */
 	while (oblocks < nblocks) {
+		int		cancelflags = 0;
+		xfs_trans_t	*tp;
+
 		tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFSRT_ALLOC);
 		resblks = XFS_GROWFSRT_SPACE_RES(mp, nblocks - oblocks);
-		cancelflags = 0;
 		/*
 		 * Reserve space & log for one extent added to the file.
 		 */
@@ -171,7 +171,9 @@ xfs_growfs_rt_alloc(
 				mp->m_bsize, 0);
 			if (bp == NULL) {
 				error = XFS_ERROR(EIO);
-				goto error_cancel;
+error_cancel:
+				xfs_trans_cancel(tp, cancelflags);
+				goto error;
 			}
 			memset(XFS_BUF_PTR(bp), 0, mp->m_sb.sb_blocksize);
 			xfs_trans_log_buf(tp, bp, 0, mp->m_sb.sb_blocksize - 1);
@@ -188,8 +190,6 @@ xfs_growfs_rt_alloc(
 		oblocks = map.br_startoff + map.br_blockcount;
 	}
 	return 0;
-error_cancel:
-	xfs_trans_cancel(tp, cancelflags);
 error:
 	return error;
 }
@@ -1856,7 +1856,6 @@ xfs_growfs_rt(
 {
 	xfs_rtblock_t	bmbno;		/* bitmap block number */
 	xfs_buf_t	*bp;		/* temporary buffer */
-	int		cancelflags;	/* flags for xfs_trans_cancel */
 	int		error;		/* error return value */
 	xfs_inode_t	*ip;		/* bitmap inode, used as lock */
 	xfs_mount_t	*nmp;		/* new (fake) mount structure */
@@ -1872,13 +1871,13 @@ xfs_growfs_rt(
 	xfs_extlen_t	rsumblocks;	/* current number of rt summary blks */
 	xfs_sb_t	*sbp;		/* old superblock */
 	xfs_fsblock_t	sumbno;		/* summary block number */
-	xfs_trans_t	*tp;		/* transaction pointer */
 
 	sbp = &mp->m_sb;
-	cancelflags = 0;
 	/*
 	 * Initial error checking.
 	 */
+	if (!capable(CAP_SYS_ADMIN))
+		return XFS_ERROR(EPERM);
 	if (mp->m_rtdev_targp == NULL || mp->m_rbmip == NULL ||
 	    (nrblocks = in->newblocks) <= sbp->sb_rblocks ||
 	    (sbp->sb_rblocks && (in->extsize != sbp->sb_rextsize)))
@@ -1942,6 +1941,9 @@ xfs_growfs_rt(
 		     ((sbp->sb_rextents & ((1 << mp->m_blkbit_log) - 1)) != 0);
 	     bmbno < nrbmblocks;
 	     bmbno++) {
+		xfs_trans_t	*tp;
+		int		cancelflags = 0;
+
 		*nmp = *mp;
 		nsbp = &nmp->m_sb;
 		/*
@@ -1967,16 +1969,15 @@ xfs_growfs_rt(
 		 * Start a transaction, get the log reservation.
 		 */
 		tp = xfs_trans_alloc(mp, XFS_TRANS_GROWFSRT_FREE);
-		cancelflags = 0;
 		if ((error = xfs_trans_reserve(tp, 0,
 				XFS_GROWRTFREE_LOG_RES(nmp), 0, 0, 0)))
-			break;
+			goto error_cancel;
 		/*
 		 * Lock out other callers by grabbing the bitmap inode lock.
 		 */
 		if ((error = xfs_trans_iget(mp, tp, mp->m_sb.sb_rbmino, 0,
 						XFS_ILOCK_EXCL, &ip)))
-			break;
+			goto error_cancel;
 		ASSERT(ip == mp->m_rbmip);
 		/*
 		 * Update the bitmap inode's size.
@@ -1990,7 +1991,7 @@ xfs_growfs_rt(
 		 */
 		if ((error = xfs_trans_iget(mp, tp, mp->m_sb.sb_rsumino, 0,
 						XFS_ILOCK_EXCL, &ip)))
-			break;
+			goto error_cancel;
 		ASSERT(ip == mp->m_rsumip);
 		/*
 		 * Update the summary inode's size.
@@ -2005,7 +2006,7 @@ xfs_growfs_rt(
 		    mp->m_rsumlevels != nmp->m_rsumlevels) {
 			error = xfs_rtcopy_summary(mp, nmp, tp);
 			if (error)
-				break;
+				goto error_cancel;
 		}
 		/*
 		 * Update superblock fields.
@@ -2031,8 +2032,11 @@ xfs_growfs_rt(
 		bp = NULL;
 		error = xfs_rtfree_range(nmp, tp, sbp->sb_rextents,
 			nsbp->sb_rextents - sbp->sb_rextents, &bp, &sumbno);
-		if (error)
+		if (error) {
+error_cancel:
+			xfs_trans_cancel(tp, cancelflags);
 			break;
+		}
 		/*
 		 * Mark more blocks free in the superblock.
 		 */
@@ -2045,15 +2049,10 @@ xfs_growfs_rt(
 		mp->m_rsumsize = nrsumsize;
 
 		error = xfs_trans_commit(tp, 0);
-		if (error) {
-			tp = NULL;
+		if (error)
 			break;
-		}
 	}
 
-	if (error && tp)
-		xfs_trans_cancel(tp, cancelflags);
-
 	/*
 	 * Free the fake mp structure.
 	 */
diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c
index 3a82576dde9..36f3a21c54d 100644
--- a/fs/xfs/xfs_rw.c
+++ b/fs/xfs/xfs_rw.c
@@ -406,7 +406,7 @@ xfs_bwrite(
 	 * XXXsup how does this work for quotas.
 	 */
 	XFS_BUF_SET_BDSTRAT_FUNC(bp, xfs_bdstrat_cb);
-	XFS_BUF_SET_FSPRIVATE3(bp, mp);
+	bp->b_mount = mp;
 	XFS_BUF_WRITE(bp);
 
 	if ((error = XFS_bwrite(bp))) {
diff --git a/fs/xfs/xfs_sb.h b/fs/xfs/xfs_sb.h
index 3f8cf1587f4..1ed71916e4c 100644
--- a/fs/xfs/xfs_sb.h
+++ b/fs/xfs/xfs_sb.h
@@ -79,6 +79,7 @@ struct xfs_mount;
 #define XFS_SB_VERSION2_LAZYSBCOUNTBIT	0x00000002	/* Superblk counters */
 #define XFS_SB_VERSION2_RESERVED4BIT	0x00000004
 #define XFS_SB_VERSION2_ATTR2BIT	0x00000008	/* Inline attr rework */
+#define XFS_SB_VERSION2_PARENTBIT	0x00000010	/* parent pointers */
 
 #define	XFS_SB_VERSION2_OKREALFBITS	\
 	(XFS_SB_VERSION2_LAZYSBCOUNTBIT	| \
@@ -296,30 +297,34 @@ typedef enum {
 
 #define	XFS_SB_VERSION_NUM(sbp)	((sbp)->sb_versionnum & XFS_SB_VERSION_NUMBITS)
 
-#ifdef __KERNEL__
 static inline int xfs_sb_good_version(xfs_sb_t *sbp)
 {
-	return (((sbp->sb_versionnum >= XFS_SB_VERSION_1) && \
-		  (sbp->sb_versionnum <= XFS_SB_VERSION_3)) || \
-		   ((XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		    !((sbp->sb_versionnum & ~XFS_SB_VERSION_OKREALBITS) || \
-		      ((sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT) && \
-		       (sbp->sb_features2 & ~XFS_SB_VERSION2_OKREALBITS))) && \
-	  	    (sbp->sb_shared_vn <= XFS_SB_MAX_SHARED_VN)));
-}
+	/* We always support version 1-3 */
+	if (sbp->sb_versionnum >= XFS_SB_VERSION_1 &&
+	    sbp->sb_versionnum <= XFS_SB_VERSION_3)
+		return 1;
+
+	/* We support version 4 if all feature bits are supported */
+	if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) {
+		if ((sbp->sb_versionnum & ~XFS_SB_VERSION_OKREALBITS) ||
+		    ((sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT) &&
+		     (sbp->sb_features2 & ~XFS_SB_VERSION2_OKREALBITS)))
+			return 0;
+
+#ifdef __KERNEL__
+		if (sbp->sb_shared_vn > XFS_SB_MAX_SHARED_VN)
+			return 0;
 #else
-static inline int xfs_sb_good_version(xfs_sb_t *sbp)
-{
-	return (((sbp->sb_versionnum >= XFS_SB_VERSION_1) && \
-		  (sbp->sb_versionnum <= XFS_SB_VERSION_3)) || \
-		   ((XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		    !((sbp->sb_versionnum & ~XFS_SB_VERSION_OKREALBITS) || \
-		      ((sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT) && \
-		       (sbp->sb_features2 & ~XFS_SB_VERSION2_OKREALBITS))) && \
-		  (!(sbp->sb_versionnum & XFS_SB_VERSION_SHAREDBIT) || \
-		   (sbp->sb_shared_vn <= XFS_SB_MAX_SHARED_VN))));
+		if ((sbp->sb_versionnum & XFS_SB_VERSION_SHAREDBIT) &&
+		    sbp->sb_shared_vn > XFS_SB_MAX_SHARED_VN)
+			return 0;
+#endif
+
+		return 1;
+	}
+
+	return 0;
 }
-#endif /* __KERNEL__ */
 
 /*
  * Detect a mismatched features2 field.  Older kernels read/wrote
@@ -332,123 +337,127 @@ static inline int xfs_sb_has_mismatched_features2(xfs_sb_t *sbp)
 
 static inline unsigned xfs_sb_version_tonew(unsigned v)
 {
-	return ((((v) == XFS_SB_VERSION_1) ? \
-		0 : \
-		(((v) == XFS_SB_VERSION_2) ? \
-			XFS_SB_VERSION_ATTRBIT : \
-			(XFS_SB_VERSION_ATTRBIT | XFS_SB_VERSION_NLINKBIT))) | \
-		XFS_SB_VERSION_4);
+	if (v == XFS_SB_VERSION_1)
+		return XFS_SB_VERSION_4;
+
+	if (v == XFS_SB_VERSION_2)
+		return XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT;
+
+	return XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT |
+		XFS_SB_VERSION_NLINKBIT;
 }
 
 static inline unsigned xfs_sb_version_toold(unsigned v)
 {
-	return (((v) & (XFS_SB_VERSION_QUOTABIT | XFS_SB_VERSION_ALIGNBIT)) ? \
-		0 : \
-		(((v) & XFS_SB_VERSION_NLINKBIT) ? \
-			XFS_SB_VERSION_3 : \
-			(((v) & XFS_SB_VERSION_ATTRBIT) ?  \
-				XFS_SB_VERSION_2 : \
-				XFS_SB_VERSION_1)));
+	if (v & (XFS_SB_VERSION_QUOTABIT | XFS_SB_VERSION_ALIGNBIT))
+		return 0;
+	if (v & XFS_SB_VERSION_NLINKBIT)
+		return XFS_SB_VERSION_3;
+	if (v & XFS_SB_VERSION_ATTRBIT)
+		return XFS_SB_VERSION_2;
+	return XFS_SB_VERSION_1;
 }
 
 static inline int xfs_sb_version_hasattr(xfs_sb_t *sbp)
 {
-	return ((sbp)->sb_versionnum == XFS_SB_VERSION_2) || \
-		 ((sbp)->sb_versionnum == XFS_SB_VERSION_3) || \
-		 ((XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		  ((sbp)->sb_versionnum & XFS_SB_VERSION_ATTRBIT));
+	return sbp->sb_versionnum == XFS_SB_VERSION_2 ||
+		sbp->sb_versionnum == XFS_SB_VERSION_3 ||
+		(XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		 (sbp->sb_versionnum & XFS_SB_VERSION_ATTRBIT));
 }
 
 static inline void xfs_sb_version_addattr(xfs_sb_t *sbp)
 {
-	(sbp)->sb_versionnum = (((sbp)->sb_versionnum == XFS_SB_VERSION_1) ? \
-		XFS_SB_VERSION_2 : \
-		((XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) ? \
-			((sbp)->sb_versionnum | XFS_SB_VERSION_ATTRBIT) : \
-			(XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT)));
+	if (sbp->sb_versionnum == XFS_SB_VERSION_1)
+		sbp->sb_versionnum = XFS_SB_VERSION_2;
+	else if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4)
+		sbp->sb_versionnum |= XFS_SB_VERSION_ATTRBIT;
+	else
+		sbp->sb_versionnum = XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT;
 }
 
 static inline int xfs_sb_version_hasnlink(xfs_sb_t *sbp)
 {
-	return ((sbp)->sb_versionnum == XFS_SB_VERSION_3) || \
-		 ((XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		  ((sbp)->sb_versionnum & XFS_SB_VERSION_NLINKBIT));
+	return sbp->sb_versionnum == XFS_SB_VERSION_3 ||
+		 (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		  (sbp->sb_versionnum & XFS_SB_VERSION_NLINKBIT));
 }
 
 static inline void xfs_sb_version_addnlink(xfs_sb_t *sbp)
 {
-	(sbp)->sb_versionnum = ((sbp)->sb_versionnum <= XFS_SB_VERSION_2 ? \
-		XFS_SB_VERSION_3 : \
-		((sbp)->sb_versionnum | XFS_SB_VERSION_NLINKBIT));
+	if (sbp->sb_versionnum <= XFS_SB_VERSION_2)
+		sbp->sb_versionnum = XFS_SB_VERSION_3;
+	else
+		sbp->sb_versionnum |= XFS_SB_VERSION_NLINKBIT;
 }
 
 static inline int xfs_sb_version_hasquota(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_QUOTABIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_QUOTABIT);
 }
 
 static inline void xfs_sb_version_addquota(xfs_sb_t *sbp)
 {
-	(sbp)->sb_versionnum = \
-		 (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 ? \
-			((sbp)->sb_versionnum | XFS_SB_VERSION_QUOTABIT) : \
-			(xfs_sb_version_tonew((sbp)->sb_versionnum) | \
-			 XFS_SB_VERSION_QUOTABIT));
+	if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4)
+		sbp->sb_versionnum |= XFS_SB_VERSION_QUOTABIT;
+	else
+		sbp->sb_versionnum = xfs_sb_version_tonew(sbp->sb_versionnum) |
+					XFS_SB_VERSION_QUOTABIT;
 }
 
 static inline int xfs_sb_version_hasalign(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_ALIGNBIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_ALIGNBIT);
 }
 
 static inline int xfs_sb_version_hasdalign(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_DALIGNBIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_DALIGNBIT);
 }
 
 static inline int xfs_sb_version_hasshared(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_SHAREDBIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_SHAREDBIT);
 }
 
 static inline int xfs_sb_version_hasdirv2(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_DIRV2BIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_DIRV2BIT);
 }
 
 static inline int xfs_sb_version_haslogv2(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_LOGV2BIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_LOGV2BIT);
 }
 
 static inline int xfs_sb_version_hasextflgbit(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_EXTFLGBIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_EXTFLGBIT);
 }
 
 static inline int xfs_sb_version_hassector(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_SECTORBIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_SECTORBIT);
 }
 
 static inline int xfs_sb_version_hasasciici(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
 		(sbp->sb_versionnum & XFS_SB_VERSION_BORGBIT);
 }
 
 static inline int xfs_sb_version_hasmorebits(xfs_sb_t *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) && \
-		((sbp)->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT);
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
+		(sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT);
 }
 
 /*
@@ -463,22 +472,20 @@ static inline int xfs_sb_version_hasmorebits(xfs_sb_t *sbp)
 
 static inline int xfs_sb_version_haslazysbcount(xfs_sb_t *sbp)
 {
-	return (xfs_sb_version_hasmorebits(sbp) &&	\
-		((sbp)->sb_features2 & XFS_SB_VERSION2_LAZYSBCOUNTBIT));
+	return xfs_sb_version_hasmorebits(sbp) &&
+		(sbp->sb_features2 & XFS_SB_VERSION2_LAZYSBCOUNTBIT);
 }
 
 static inline int xfs_sb_version_hasattr2(xfs_sb_t *sbp)
 {
-	return (xfs_sb_version_hasmorebits(sbp)) &&	\
-		((sbp)->sb_features2 & XFS_SB_VERSION2_ATTR2BIT);
+	return xfs_sb_version_hasmorebits(sbp) &&
+		(sbp->sb_features2 & XFS_SB_VERSION2_ATTR2BIT);
 }
 
 static inline void xfs_sb_version_addattr2(xfs_sb_t *sbp)
 {
-	((sbp)->sb_versionnum =	\
-		((sbp)->sb_versionnum | XFS_SB_VERSION_MOREBITSBIT),	\
-	((sbp)->sb_features2 =	\
-		((sbp)->sb_features2 | XFS_SB_VERSION2_ATTR2BIT)));
+	sbp->sb_versionnum |= XFS_SB_VERSION_MOREBITSBIT;
+	sbp->sb_features2 |= XFS_SB_VERSION2_ATTR2BIT;
 }
 
 static inline void xfs_sb_version_removeattr2(xfs_sb_t *sbp)
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 4e1c22a23be..8570b826fed 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -290,7 +290,7 @@ xfs_trans_dup(
 	ASSERT(tp->t_ticket != NULL);
 
 	ntp->t_flags = XFS_TRANS_PERM_LOG_RES | (tp->t_flags & XFS_TRANS_RESERVE);
-	ntp->t_ticket = tp->t_ticket;
+	ntp->t_ticket = xfs_log_ticket_get(tp->t_ticket);
 	ntp->t_blk_res = tp->t_blk_res - tp->t_blk_res_used;
 	tp->t_blk_res = tp->t_blk_res_used;
 	ntp->t_rtx_res = tp->t_rtx_res - tp->t_rtx_res_used;
@@ -1260,6 +1260,13 @@ xfs_trans_roll(
 	trans = *tpp;
 
 	/*
+	 * transaction commit worked ok so we can drop the extra ticket
+	 * reference that we gained in xfs_trans_dup()
+	 */
+	xfs_log_ticket_put(trans->t_ticket);
+
+
+	/*
 	 * Reserve space in the log for th next transaction.
 	 * This also pushes items in the "AIL", the list of logged items,
 	 * out to disk if they are taking up space at the tail of the log
@@ -1383,11 +1390,12 @@ xfs_trans_chunk_committed(
 	xfs_log_item_desc_t	*lidp;
 	xfs_log_item_t		*lip;
 	xfs_lsn_t		item_lsn;
-	struct xfs_mount	*mp;
 	int			i;
 
 	lidp = licp->lic_descs;
 	for (i = 0; i < licp->lic_unused; i++, lidp++) {
+		struct xfs_ail		*ailp;
+
 		if (xfs_lic_isfree(licp, i)) {
 			continue;
 		}
@@ -1424,19 +1432,19 @@ xfs_trans_chunk_committed(
 		 * This would cause the earlier transaction to fail
 		 * the test below.
 		 */
-		mp = lip->li_mountp;
-		spin_lock(&mp->m_ail_lock);
+		ailp = lip->li_ailp;
+		spin_lock(&ailp->xa_lock);
 		if (XFS_LSN_CMP(item_lsn, lip->li_lsn) > 0) {
 			/*
 			 * This will set the item's lsn to item_lsn
 			 * and update the position of the item in
 			 * the AIL.
 			 *
-			 * xfs_trans_update_ail() drops the AIL lock.
+			 * xfs_trans_ail_update() drops the AIL lock.
 			 */
-			xfs_trans_update_ail(mp, lip, item_lsn);
+			xfs_trans_ail_update(ailp, lip, item_lsn);
 		} else {
-			spin_unlock(&mp->m_ail_lock);
+			spin_unlock(&ailp->xa_lock);
 		}
 
 		/*
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 74c80bd2b0e..d6fe4a88d79 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -18,6 +18,8 @@
 #ifndef	__XFS_TRANS_H__
 #define	__XFS_TRANS_H__
 
+struct xfs_log_item;
+
 /*
  * This is the structure written in the log at the head of
  * every transaction. It identifies the type and id of the
@@ -98,76 +100,6 @@ typedef struct xfs_trans_header {
 #define	XFS_TRANS_TYPE_MAX		41
 /* new transaction types need to be reflected in xfs_logprint(8) */
 
-
-#ifdef __KERNEL__
-struct xfs_buf;
-struct xfs_buftarg;
-struct xfs_efd_log_item;
-struct xfs_efi_log_item;
-struct xfs_inode;
-struct xfs_item_ops;
-struct xfs_log_iovec;
-struct xfs_log_item;
-struct xfs_log_item_desc;
-struct xfs_mount;
-struct xfs_trans;
-struct xfs_dquot_acct;
-
-typedef struct xfs_log_item {
-	struct list_head		li_ail;		/* AIL pointers */
-	xfs_lsn_t			li_lsn;		/* last on-disk lsn */
-	struct xfs_log_item_desc	*li_desc;	/* ptr to current desc*/
-	struct xfs_mount		*li_mountp;	/* ptr to fs mount */
-	uint				li_type;	/* item type */
-	uint				li_flags;	/* misc flags */
-	struct xfs_log_item		*li_bio_list;	/* buffer item list */
-	void				(*li_cb)(struct xfs_buf *,
-						 struct xfs_log_item *);
-							/* buffer item iodone */
-							/* callback func */
-	struct xfs_item_ops		*li_ops;	/* function list */
-} xfs_log_item_t;
-
-#define	XFS_LI_IN_AIL	0x1
-#define XFS_LI_ABORTED	0x2
-
-typedef struct xfs_item_ops {
-	uint (*iop_size)(xfs_log_item_t *);
-	void (*iop_format)(xfs_log_item_t *, struct xfs_log_iovec *);
-	void (*iop_pin)(xfs_log_item_t *);
-	void (*iop_unpin)(xfs_log_item_t *, int);
-	void (*iop_unpin_remove)(xfs_log_item_t *, struct xfs_trans *);
-	uint (*iop_trylock)(xfs_log_item_t *);
-	void (*iop_unlock)(xfs_log_item_t *);
-	xfs_lsn_t (*iop_committed)(xfs_log_item_t *, xfs_lsn_t);
-	void (*iop_push)(xfs_log_item_t *);
-	void (*iop_pushbuf)(xfs_log_item_t *);
-	void (*iop_committing)(xfs_log_item_t *, xfs_lsn_t);
-} xfs_item_ops_t;
-
-#define IOP_SIZE(ip)		(*(ip)->li_ops->iop_size)(ip)
-#define IOP_FORMAT(ip,vp)	(*(ip)->li_ops->iop_format)(ip, vp)
-#define IOP_PIN(ip)		(*(ip)->li_ops->iop_pin)(ip)
-#define IOP_UNPIN(ip, flags)	(*(ip)->li_ops->iop_unpin)(ip, flags)
-#define IOP_UNPIN_REMOVE(ip,tp) (*(ip)->li_ops->iop_unpin_remove)(ip, tp)
-#define IOP_TRYLOCK(ip)		(*(ip)->li_ops->iop_trylock)(ip)
-#define IOP_UNLOCK(ip)		(*(ip)->li_ops->iop_unlock)(ip)
-#define IOP_COMMITTED(ip, lsn)	(*(ip)->li_ops->iop_committed)(ip, lsn)
-#define IOP_PUSH(ip)		(*(ip)->li_ops->iop_push)(ip)
-#define IOP_PUSHBUF(ip)		(*(ip)->li_ops->iop_pushbuf)(ip)
-#define IOP_COMMITTING(ip, lsn) (*(ip)->li_ops->iop_committing)(ip, lsn)
-
-/*
- * Return values for the IOP_TRYLOCK() routines.
- */
-#define	XFS_ITEM_SUCCESS	0
-#define	XFS_ITEM_PINNED		1
-#define	XFS_ITEM_LOCKED		2
-#define	XFS_ITEM_FLUSHING	3
-#define XFS_ITEM_PUSHBUF	4
-
-#endif	/* __KERNEL__ */
-
 /*
  * This structure is used to track log items associated with
  * a transaction.  It points to the log item and keeps some
@@ -176,7 +108,7 @@ typedef struct xfs_item_ops {
  * once we get to commit processing (see xfs_trans_commit()).
  */
 typedef struct xfs_log_item_desc {
-	xfs_log_item_t	*lid_item;
+	struct xfs_log_item	*lid_item;
 	ushort		lid_size;
 	unsigned char	lid_flags;
 	unsigned char	lid_index;
@@ -276,94 +208,6 @@ xfs_lic_desc_to_chunk(xfs_log_item_desc_t *dp)
 		(xfs_caddr_t)(((xfs_log_item_chunk_t*)0)->lic_descs));
 }
 
-#ifdef __KERNEL__
-/*
- * This structure is used to maintain a list of block ranges that have been
- * freed in the transaction.  The ranges are listed in the perag[] busy list
- * between when they're freed and the transaction is committed to disk.
- */
-
-typedef struct xfs_log_busy_slot {
-	xfs_agnumber_t		lbc_ag;
-	ushort			lbc_idx;	/* index in perag.busy[] */
-} xfs_log_busy_slot_t;
-
-#define XFS_LBC_NUM_SLOTS	31
-typedef struct xfs_log_busy_chunk {
-	struct xfs_log_busy_chunk	*lbc_next;
-	uint				lbc_free;	/* free slots bitmask */
-	ushort				lbc_unused;	/* first unused */
-	xfs_log_busy_slot_t		lbc_busy[XFS_LBC_NUM_SLOTS];
-} xfs_log_busy_chunk_t;
-
-#define	XFS_LBC_MAX_SLOT	(XFS_LBC_NUM_SLOTS - 1)
-#define	XFS_LBC_FREEMASK	((1U << XFS_LBC_NUM_SLOTS) - 1)
-
-#define	XFS_LBC_INIT(cp)	((cp)->lbc_free = XFS_LBC_FREEMASK)
-#define	XFS_LBC_CLAIM(cp, slot)	((cp)->lbc_free &= ~(1 << (slot)))
-#define	XFS_LBC_SLOT(cp, slot)	(&((cp)->lbc_busy[(slot)]))
-#define	XFS_LBC_VACANCY(cp)	(((cp)->lbc_free) & XFS_LBC_FREEMASK)
-#define	XFS_LBC_ISFREE(cp, slot) ((cp)->lbc_free & (1 << (slot)))
-
-/*
- * This is the type of function which can be given to xfs_trans_callback()
- * to be called upon the transaction's commit to disk.
- */
-typedef void (*xfs_trans_callback_t)(struct xfs_trans *, void *);
-
-/*
- * This is the structure maintained for every active transaction.
- */
-typedef struct xfs_trans {
-	unsigned int		t_magic;	/* magic number */
-	xfs_log_callback_t	t_logcb;	/* log callback struct */
-	unsigned int		t_type;		/* transaction type */
-	unsigned int		t_log_res;	/* amt of log space resvd */
-	unsigned int		t_log_count;	/* count for perm log res */
-	unsigned int		t_blk_res;	/* # of blocks resvd */
-	unsigned int		t_blk_res_used;	/* # of resvd blocks used */
-	unsigned int		t_rtx_res;	/* # of rt extents resvd */
-	unsigned int		t_rtx_res_used;	/* # of resvd rt extents used */
-	xfs_log_ticket_t	t_ticket;	/* log mgr ticket */
-	xfs_lsn_t		t_lsn;		/* log seq num of start of
-						 * transaction. */
-	xfs_lsn_t		t_commit_lsn;	/* log seq num of end of
-						 * transaction. */
-	struct xfs_mount	*t_mountp;	/* ptr to fs mount struct */
-	struct xfs_dquot_acct   *t_dqinfo;	/* acctg info for dquots */
-	xfs_trans_callback_t	t_callback;	/* transaction callback */
-	void			*t_callarg;	/* callback arg */
-	unsigned int		t_flags;	/* misc flags */
-	int64_t			t_icount_delta;	/* superblock icount change */
-	int64_t			t_ifree_delta;	/* superblock ifree change */
-	int64_t			t_fdblocks_delta; /* superblock fdblocks chg */
-	int64_t			t_res_fdblocks_delta; /* on-disk only chg */
-	int64_t			t_frextents_delta;/* superblock freextents chg*/
-	int64_t			t_res_frextents_delta; /* on-disk only chg */
-#ifdef DEBUG
-	int64_t			t_ag_freeblks_delta; /* debugging counter */
-	int64_t			t_ag_flist_delta; /* debugging counter */
-	int64_t			t_ag_btree_delta; /* debugging counter */
-#endif
-	int64_t			t_dblocks_delta;/* superblock dblocks change */
-	int64_t			t_agcount_delta;/* superblock agcount change */
-	int64_t			t_imaxpct_delta;/* superblock imaxpct change */
-	int64_t			t_rextsize_delta;/* superblock rextsize chg */
-	int64_t			t_rbmblocks_delta;/* superblock rbmblocks chg */
-	int64_t			t_rblocks_delta;/* superblock rblocks change */
-	int64_t			t_rextents_delta;/* superblocks rextents chg */
-	int64_t			t_rextslog_delta;/* superblocks rextslog chg */
-	unsigned int		t_items_free;	/* log item descs free */
-	xfs_log_item_chunk_t	t_items;	/* first log item desc chunk */
-	xfs_trans_header_t	t_header;	/* header for in-log trans */
-	unsigned int		t_busy_free;	/* busy descs free */
-	xfs_log_busy_chunk_t	t_busy;		/* busy/async free blocks */
-	unsigned long		t_pflags;	/* saved process flags state */
-} xfs_trans_t;
-
-#endif	/* __KERNEL__ */
-
-
 #define	XFS_TRANS_MAGIC		0x5452414E	/* 'TRAN' */
 /*
  * Values for t_flags.
@@ -906,6 +750,157 @@ typedef struct xfs_trans {
 #define	XFS_DQUOT_REF		1
 
 #ifdef __KERNEL__
+
+struct xfs_buf;
+struct xfs_buftarg;
+struct xfs_efd_log_item;
+struct xfs_efi_log_item;
+struct xfs_inode;
+struct xfs_item_ops;
+struct xfs_log_iovec;
+struct xfs_log_item_desc;
+struct xfs_mount;
+struct xfs_trans;
+struct xfs_dquot_acct;
+
+typedef struct xfs_log_item {
+	struct list_head		li_ail;		/* AIL pointers */
+	xfs_lsn_t			li_lsn;		/* last on-disk lsn */
+	struct xfs_log_item_desc	*li_desc;	/* ptr to current desc*/
+	struct xfs_mount		*li_mountp;	/* ptr to fs mount */
+	struct xfs_ail			*li_ailp;	/* ptr to AIL */
+	uint				li_type;	/* item type */
+	uint				li_flags;	/* misc flags */
+	struct xfs_log_item		*li_bio_list;	/* buffer item list */
+	void				(*li_cb)(struct xfs_buf *,
+						 struct xfs_log_item *);
+							/* buffer item iodone */
+							/* callback func */
+	struct xfs_item_ops		*li_ops;	/* function list */
+} xfs_log_item_t;
+
+#define	XFS_LI_IN_AIL	0x1
+#define XFS_LI_ABORTED	0x2
+
+typedef struct xfs_item_ops {
+	uint (*iop_size)(xfs_log_item_t *);
+	void (*iop_format)(xfs_log_item_t *, struct xfs_log_iovec *);
+	void (*iop_pin)(xfs_log_item_t *);
+	void (*iop_unpin)(xfs_log_item_t *, int);
+	void (*iop_unpin_remove)(xfs_log_item_t *, struct xfs_trans *);
+	uint (*iop_trylock)(xfs_log_item_t *);
+	void (*iop_unlock)(xfs_log_item_t *);
+	xfs_lsn_t (*iop_committed)(xfs_log_item_t *, xfs_lsn_t);
+	void (*iop_push)(xfs_log_item_t *);
+	void (*iop_pushbuf)(xfs_log_item_t *);
+	void (*iop_committing)(xfs_log_item_t *, xfs_lsn_t);
+} xfs_item_ops_t;
+
+#define IOP_SIZE(ip)		(*(ip)->li_ops->iop_size)(ip)
+#define IOP_FORMAT(ip,vp)	(*(ip)->li_ops->iop_format)(ip, vp)
+#define IOP_PIN(ip)		(*(ip)->li_ops->iop_pin)(ip)
+#define IOP_UNPIN(ip, flags)	(*(ip)->li_ops->iop_unpin)(ip, flags)
+#define IOP_UNPIN_REMOVE(ip,tp) (*(ip)->li_ops->iop_unpin_remove)(ip, tp)
+#define IOP_TRYLOCK(ip)		(*(ip)->li_ops->iop_trylock)(ip)
+#define IOP_UNLOCK(ip)		(*(ip)->li_ops->iop_unlock)(ip)
+#define IOP_COMMITTED(ip, lsn)	(*(ip)->li_ops->iop_committed)(ip, lsn)
+#define IOP_PUSH(ip)		(*(ip)->li_ops->iop_push)(ip)
+#define IOP_PUSHBUF(ip)		(*(ip)->li_ops->iop_pushbuf)(ip)
+#define IOP_COMMITTING(ip, lsn) (*(ip)->li_ops->iop_committing)(ip, lsn)
+
+/*
+ * Return values for the IOP_TRYLOCK() routines.
+ */
+#define	XFS_ITEM_SUCCESS	0
+#define	XFS_ITEM_PINNED		1
+#define	XFS_ITEM_LOCKED		2
+#define	XFS_ITEM_FLUSHING	3
+#define XFS_ITEM_PUSHBUF	4
+
+/*
+ * This structure is used to maintain a list of block ranges that have been
+ * freed in the transaction.  The ranges are listed in the perag[] busy list
+ * between when they're freed and the transaction is committed to disk.
+ */
+
+typedef struct xfs_log_busy_slot {
+	xfs_agnumber_t		lbc_ag;
+	ushort			lbc_idx;	/* index in perag.busy[] */
+} xfs_log_busy_slot_t;
+
+#define XFS_LBC_NUM_SLOTS	31
+typedef struct xfs_log_busy_chunk {
+	struct xfs_log_busy_chunk	*lbc_next;
+	uint				lbc_free;	/* free slots bitmask */
+	ushort				lbc_unused;	/* first unused */
+	xfs_log_busy_slot_t		lbc_busy[XFS_LBC_NUM_SLOTS];
+} xfs_log_busy_chunk_t;
+
+#define	XFS_LBC_MAX_SLOT	(XFS_LBC_NUM_SLOTS - 1)
+#define	XFS_LBC_FREEMASK	((1U << XFS_LBC_NUM_SLOTS) - 1)
+
+#define	XFS_LBC_INIT(cp)	((cp)->lbc_free = XFS_LBC_FREEMASK)
+#define	XFS_LBC_CLAIM(cp, slot)	((cp)->lbc_free &= ~(1 << (slot)))
+#define	XFS_LBC_SLOT(cp, slot)	(&((cp)->lbc_busy[(slot)]))
+#define	XFS_LBC_VACANCY(cp)	(((cp)->lbc_free) & XFS_LBC_FREEMASK)
+#define	XFS_LBC_ISFREE(cp, slot) ((cp)->lbc_free & (1 << (slot)))
+
+/*
+ * This is the type of function which can be given to xfs_trans_callback()
+ * to be called upon the transaction's commit to disk.
+ */
+typedef void (*xfs_trans_callback_t)(struct xfs_trans *, void *);
+
+/*
+ * This is the structure maintained for every active transaction.
+ */
+typedef struct xfs_trans {
+	unsigned int		t_magic;	/* magic number */
+	xfs_log_callback_t	t_logcb;	/* log callback struct */
+	unsigned int		t_type;		/* transaction type */
+	unsigned int		t_log_res;	/* amt of log space resvd */
+	unsigned int		t_log_count;	/* count for perm log res */
+	unsigned int		t_blk_res;	/* # of blocks resvd */
+	unsigned int		t_blk_res_used;	/* # of resvd blocks used */
+	unsigned int		t_rtx_res;	/* # of rt extents resvd */
+	unsigned int		t_rtx_res_used;	/* # of resvd rt extents used */
+	xfs_log_ticket_t	t_ticket;	/* log mgr ticket */
+	xfs_lsn_t		t_lsn;		/* log seq num of start of
+						 * transaction. */
+	xfs_lsn_t		t_commit_lsn;	/* log seq num of end of
+						 * transaction. */
+	struct xfs_mount	*t_mountp;	/* ptr to fs mount struct */
+	struct xfs_dquot_acct   *t_dqinfo;	/* acctg info for dquots */
+	xfs_trans_callback_t	t_callback;	/* transaction callback */
+	void			*t_callarg;	/* callback arg */
+	unsigned int		t_flags;	/* misc flags */
+	int64_t			t_icount_delta;	/* superblock icount change */
+	int64_t			t_ifree_delta;	/* superblock ifree change */
+	int64_t			t_fdblocks_delta; /* superblock fdblocks chg */
+	int64_t			t_res_fdblocks_delta; /* on-disk only chg */
+	int64_t			t_frextents_delta;/* superblock freextents chg*/
+	int64_t			t_res_frextents_delta; /* on-disk only chg */
+#ifdef DEBUG
+	int64_t			t_ag_freeblks_delta; /* debugging counter */
+	int64_t			t_ag_flist_delta; /* debugging counter */
+	int64_t			t_ag_btree_delta; /* debugging counter */
+#endif
+	int64_t			t_dblocks_delta;/* superblock dblocks change */
+	int64_t			t_agcount_delta;/* superblock agcount change */
+	int64_t			t_imaxpct_delta;/* superblock imaxpct change */
+	int64_t			t_rextsize_delta;/* superblock rextsize chg */
+	int64_t			t_rbmblocks_delta;/* superblock rbmblocks chg */
+	int64_t			t_rblocks_delta;/* superblock rblocks change */
+	int64_t			t_rextents_delta;/* superblocks rextents chg */
+	int64_t			t_rextslog_delta;/* superblocks rextslog chg */
+	unsigned int		t_items_free;	/* log item descs free */
+	xfs_log_item_chunk_t	t_items;	/* first log item desc chunk */
+	xfs_trans_header_t	t_header;	/* header for in-log trans */
+	unsigned int		t_busy_free;	/* busy descs free */
+	xfs_log_busy_chunk_t	t_busy;		/* busy/async free blocks */
+	unsigned long		t_pflags;	/* saved process flags state */
+} xfs_trans_t;
+
 /*
  * XFS transaction mechanism exported interfaces that are
  * actually macros.
@@ -928,7 +923,6 @@ typedef struct xfs_trans {
 /*
  * XFS transaction mechanism exported interfaces.
  */
-void		xfs_trans_init(struct xfs_mount *);
 xfs_trans_t	*xfs_trans_alloc(struct xfs_mount *, uint);
 xfs_trans_t	*_xfs_trans_alloc(struct xfs_mount *, uint);
 xfs_trans_t	*xfs_trans_dup(xfs_trans_t *);
@@ -975,13 +969,8 @@ int		_xfs_trans_commit(xfs_trans_t *,
 				  int *);
 #define xfs_trans_commit(tp, flags)	_xfs_trans_commit(tp, flags, NULL)
 void		xfs_trans_cancel(xfs_trans_t *, int);
-int		xfs_trans_roll(struct xfs_trans **, struct xfs_inode *);
 int		xfs_trans_ail_init(struct xfs_mount *);
 void		xfs_trans_ail_destroy(struct xfs_mount *);
-void		xfs_trans_push_ail(struct xfs_mount *, xfs_lsn_t);
-xfs_lsn_t	xfs_trans_tail_ail(struct xfs_mount *);
-void		xfs_trans_unlocked_item(struct xfs_mount *,
-					xfs_log_item_t *);
 xfs_log_busy_slot_t *xfs_trans_add_busy(xfs_trans_t *tp,
 					xfs_agnumber_t ag,
 					xfs_extlen_t idx);
@@ -990,4 +979,7 @@ extern kmem_zone_t	*xfs_trans_zone;
 
 #endif	/* __KERNEL__ */
 
+void		xfs_trans_init(struct xfs_mount *);
+int		xfs_trans_roll(struct xfs_trans **, struct xfs_inode *);
+
 #endif	/* __XFS_TRANS_H__ */
diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index 1f77c00af56..2d47f10f8be 100644
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
+ * Copyright (c) 2008 Dave Chinner
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -28,13 +29,13 @@
 #include "xfs_trans_priv.h"
 #include "xfs_error.h"
 
-STATIC void xfs_ail_insert(xfs_ail_t *, xfs_log_item_t *);
-STATIC xfs_log_item_t * xfs_ail_delete(xfs_ail_t *, xfs_log_item_t *);
-STATIC xfs_log_item_t * xfs_ail_min(xfs_ail_t *);
-STATIC xfs_log_item_t * xfs_ail_next(xfs_ail_t *, xfs_log_item_t *);
+STATIC void xfs_ail_insert(struct xfs_ail *, xfs_log_item_t *);
+STATIC xfs_log_item_t * xfs_ail_delete(struct xfs_ail *, xfs_log_item_t *);
+STATIC xfs_log_item_t * xfs_ail_min(struct xfs_ail *);
+STATIC xfs_log_item_t * xfs_ail_next(struct xfs_ail *, xfs_log_item_t *);
 
 #ifdef DEBUG
-STATIC void xfs_ail_check(xfs_ail_t *, xfs_log_item_t *);
+STATIC void xfs_ail_check(struct xfs_ail *, xfs_log_item_t *);
 #else
 #define	xfs_ail_check(a,l)
 #endif /* DEBUG */
@@ -50,20 +51,20 @@ STATIC void xfs_ail_check(xfs_ail_t *, xfs_log_item_t *);
  * lsn of the last item in the AIL.
  */
 xfs_lsn_t
-xfs_trans_tail_ail(
-	xfs_mount_t	*mp)
+xfs_trans_ail_tail(
+	struct xfs_ail	*ailp)
 {
 	xfs_lsn_t	lsn;
 	xfs_log_item_t	*lip;
 
-	spin_lock(&mp->m_ail_lock);
-	lip = xfs_ail_min(&mp->m_ail);
+	spin_lock(&ailp->xa_lock);
+	lip = xfs_ail_min(ailp);
 	if (lip == NULL) {
 		lsn = (xfs_lsn_t)0;
 	} else {
 		lsn = lip->li_lsn;
 	}
-	spin_unlock(&mp->m_ail_lock);
+	spin_unlock(&ailp->xa_lock);
 
 	return lsn;
 }
@@ -85,16 +86,125 @@ xfs_trans_tail_ail(
  * any of the objects, so the lock is not needed.
  */
 void
-xfs_trans_push_ail(
-	xfs_mount_t		*mp,
-	xfs_lsn_t		threshold_lsn)
+xfs_trans_ail_push(
+	struct xfs_ail	*ailp,
+	xfs_lsn_t	threshold_lsn)
 {
-	xfs_log_item_t		*lip;
+	xfs_log_item_t	*lip;
+
+	lip = xfs_ail_min(ailp);
+	if (lip && !XFS_FORCED_SHUTDOWN(ailp->xa_mount)) {
+		if (XFS_LSN_CMP(threshold_lsn, ailp->xa_target) > 0)
+			xfsaild_wakeup(ailp, threshold_lsn);
+	}
+}
+
+/*
+ * AIL traversal cursor initialisation.
+ *
+ * The cursor keeps track of where our current traversal is up
+ * to by tracking the next ƣtem in the list for us. However, for
+ * this to be safe, removing an object from the AIL needs to invalidate
+ * any cursor that points to it. hence the traversal cursor needs to
+ * be linked to the struct xfs_ail so that deletion can search all the
+ * active cursors for invalidation.
+ *
+ * We don't link the push cursor because it is embedded in the struct
+ * xfs_ail and hence easily findable.
+ */
+STATIC void
+xfs_trans_ail_cursor_init(
+	struct xfs_ail		*ailp,
+	struct xfs_ail_cursor	*cur)
+{
+	cur->item = NULL;
+	if (cur == &ailp->xa_cursors)
+		return;
+
+	cur->next = ailp->xa_cursors.next;
+	ailp->xa_cursors.next = cur;
+}
+
+/*
+ * Set the cursor to the next item, because when we look
+ * up the cursor the current item may have been freed.
+ */
+STATIC void
+xfs_trans_ail_cursor_set(
+	struct xfs_ail		*ailp,
+	struct xfs_ail_cursor	*cur,
+	struct xfs_log_item	*lip)
+{
+	if (lip)
+		cur->item = xfs_ail_next(ailp, lip);
+}
+
+/*
+ * Get the next item in the traversal and advance the cursor.
+ * If the cursor was invalidated (inidicated by a lip of 1),
+ * restart the traversal.
+ */
+struct xfs_log_item *
+xfs_trans_ail_cursor_next(
+	struct xfs_ail		*ailp,
+	struct xfs_ail_cursor	*cur)
+{
+	struct xfs_log_item	*lip = cur->item;
+
+	if ((__psint_t)lip & 1)
+		lip = xfs_ail_min(ailp);
+	xfs_trans_ail_cursor_set(ailp, cur, lip);
+	return lip;
+}
+
+/*
+ * Now that the traversal is complete, we need to remove the cursor
+ * from the list of traversing cursors. Avoid removing the embedded
+ * push cursor, but use the fact it is alway present to make the
+ * list deletion simple.
+ */
+void
+xfs_trans_ail_cursor_done(
+	struct xfs_ail		*ailp,
+	struct xfs_ail_cursor	*done)
+{
+	struct xfs_ail_cursor	*prev = NULL;
+	struct xfs_ail_cursor	*cur;
+
+	done->item = NULL;
+	if (done == &ailp->xa_cursors)
+		return;
+	prev = &ailp->xa_cursors;
+	for (cur = prev->next; cur; prev = cur, cur = prev->next) {
+		if (cur == done) {
+			prev->next = cur->next;
+			break;
+		}
+	}
+	ASSERT(cur);
+}
+
+/*
+ * Invalidate any cursor that is pointing to this item. This is
+ * called when an item is removed from the AIL. Any cursor pointing
+ * to this object is now invalid and the traversal needs to be
+ * terminated so it doesn't reference a freed object. We set the
+ * cursor item to a value of 1 so we can distinguish between an
+ * invalidation and the end of the list when getting the next item
+ * from the cursor.
+ */
+STATIC void
+xfs_trans_ail_cursor_clear(
+	struct xfs_ail		*ailp,
+	struct xfs_log_item	*lip)
+{
+	struct xfs_ail_cursor	*cur;
 
-	lip = xfs_ail_min(&mp->m_ail);
-	if (lip && !XFS_FORCED_SHUTDOWN(mp)) {
-		if (XFS_LSN_CMP(threshold_lsn, mp->m_ail.xa_target) > 0)
-			xfsaild_wakeup(mp, threshold_lsn);
+	/* need to search all cursors */
+	for (cur = &ailp->xa_cursors; cur; cur = cur->next) {
+		if (cur->item == lip)
+			cur->item = (struct xfs_log_item *)
+					((__psint_t)cur->item | 1);
 	}
 }
 
@@ -103,25 +213,27 @@ xfs_trans_push_ail(
  * Return the current tree generation number for use
  * in calls to xfs_trans_next_ail().
  */
-STATIC xfs_log_item_t *
-xfs_trans_first_push_ail(
-	xfs_mount_t	*mp,
-	int		*gen,
-	xfs_lsn_t	lsn)
+xfs_log_item_t *
+xfs_trans_ail_cursor_first(
+	struct xfs_ail		*ailp,
+	struct xfs_ail_cursor	*cur,
+	xfs_lsn_t		lsn)
 {
-	xfs_log_item_t	*lip;
+	xfs_log_item_t		*lip;
 
-	lip = xfs_ail_min(&mp->m_ail);
-	*gen = (int)mp->m_ail.xa_gen;
+	xfs_trans_ail_cursor_init(ailp, cur);
+	lip = xfs_ail_min(ailp);
 	if (lsn == 0)
-		return lip;
+		goto out;
 
-	list_for_each_entry(lip, &mp->m_ail.xa_ail, li_ail) {
+	list_for_each_entry(lip, &ailp->xa_ail, li_ail) {
 		if (XFS_LSN_CMP(lip->li_lsn, lsn) >= 0)
-			return lip;
+			goto out;
 	}
-
-	return NULL;
+	lip = NULL;
+out:
+	xfs_trans_ail_cursor_set(ailp, cur, lip);
+	return lip;
 }
 
 /*
@@ -129,29 +241,29 @@ xfs_trans_first_push_ail(
  */
 long
 xfsaild_push(
-	xfs_mount_t	*mp,
+	struct xfs_ail	*ailp,
 	xfs_lsn_t	*last_lsn)
 {
 	long		tout = 1000; /* milliseconds */
 	xfs_lsn_t	last_pushed_lsn = *last_lsn;
-	xfs_lsn_t	target =  mp->m_ail.xa_target;
+	xfs_lsn_t	target =  ailp->xa_target;
 	xfs_lsn_t	lsn;
 	xfs_log_item_t	*lip;
-	int		gen;
-	int		restarts;
 	int		flush_log, count, stuck;
+	xfs_mount_t	*mp = ailp->xa_mount;
+	struct xfs_ail_cursor	*cur = &ailp->xa_cursors;
 
-#define	XFS_TRANS_PUSH_AIL_RESTARTS	10
-
-	spin_lock(&mp->m_ail_lock);
-	lip = xfs_trans_first_push_ail(mp, &gen, *last_lsn);
+	spin_lock(&ailp->xa_lock);
+	xfs_trans_ail_cursor_init(ailp, cur);
+	lip = xfs_trans_ail_cursor_first(ailp, cur, *last_lsn);
 	if (!lip || XFS_FORCED_SHUTDOWN(mp)) {
 		/*
 		 * AIL is empty or our push has reached the end.
 		 */
-		spin_unlock(&mp->m_ail_lock);
+		xfs_trans_ail_cursor_done(ailp, cur);
+		spin_unlock(&ailp->xa_lock);
 		last_pushed_lsn = 0;
-		goto out;
+		return tout;
 	}
 
 	XFS_STATS_INC(xs_push_ail);
@@ -169,7 +281,7 @@ xfsaild_push(
 	 */
 	tout = 10;
 	lsn = lip->li_lsn;
-	flush_log = stuck = count = restarts = 0;
+	flush_log = stuck = count = 0;
 	while ((XFS_LSN_CMP(lip->li_lsn, target) < 0)) {
 		int	lock_result;
 		/*
@@ -184,7 +296,7 @@ xfsaild_push(
 		 * skip to the next item in the list.
 		 */
 		lock_result = IOP_TRYLOCK(lip);
-		spin_unlock(&mp->m_ail_lock);
+		spin_unlock(&ailp->xa_lock);
 		switch (lock_result) {
 		case XFS_ITEM_SUCCESS:
 			XFS_STATS_INC(xs_push_ail_success);
@@ -221,7 +333,7 @@ xfsaild_push(
 			break;
 		}
 
-		spin_lock(&mp->m_ail_lock);
+		spin_lock(&ailp->xa_lock);
 		/* should we bother continuing? */
 		if (XFS_FORCED_SHUTDOWN(mp))
 			break;
@@ -244,14 +356,13 @@ xfsaild_push(
 		if (stuck > 100)
 			break;
 
-		lip = xfs_trans_next_ail(mp, lip, &gen, &restarts);
+		lip = xfs_trans_ail_cursor_next(ailp, cur);
 		if (lip == NULL)
 			break;
-		if (restarts > XFS_TRANS_PUSH_AIL_RESTARTS)
-			break;
 		lsn = lip->li_lsn;
 	}
-	spin_unlock(&mp->m_ail_lock);
+	xfs_trans_ail_cursor_done(ailp, cur);
+	spin_unlock(&ailp->xa_lock);
 
 	if (flush_log) {
 		/*
@@ -274,8 +385,7 @@ xfsaild_push(
 		 */
 		tout += 20;
 		last_pushed_lsn = 0;
-	} else if ((restarts > XFS_TRANS_PUSH_AIL_RESTARTS) ||
-		   ((stuck * 100) / count > 90)) {
+	} else if ((stuck * 100) / count > 90) {
 		/*
 		 * Either there is a lot of contention on the AIL or we
 		 * are stuck due to operations in progress. "Stuck" in this
@@ -287,7 +397,6 @@ xfsaild_push(
 		 */
 		tout += 10;
 	}
-out:
 	*last_lsn = last_pushed_lsn;
 	return tout;
 }	/* xfsaild_push */
@@ -303,7 +412,7 @@ out:
  */
 void
 xfs_trans_unlocked_item(
-	xfs_mount_t	*mp,
+	struct xfs_ail	*ailp,
 	xfs_log_item_t	*lip)
 {
 	xfs_log_item_t	*min_lip;
@@ -315,7 +424,7 @@ xfs_trans_unlocked_item(
 	 * over some potentially valid data.
 	 */
 	if (!(lip->li_flags & XFS_LI_IN_AIL) ||
-	    XFS_FORCED_SHUTDOWN(mp)) {
+	    XFS_FORCED_SHUTDOWN(ailp->xa_mount)) {
 		return;
 	}
 
@@ -331,10 +440,10 @@ xfs_trans_unlocked_item(
 	 * the call to xfs_log_move_tail() doesn't do anything if there's
 	 * not enough free space to wake people up so we're safe calling it.
 	 */
-	min_lip = xfs_ail_min(&mp->m_ail);
+	min_lip = xfs_ail_min(ailp);
 
 	if (min_lip == lip)
-		xfs_log_move_tail(mp, 1);
+		xfs_log_move_tail(ailp->xa_mount, 1);
 }	/* xfs_trans_unlocked_item */
 
 
@@ -347,41 +456,37 @@ xfs_trans_unlocked_item(
  * we move in the AIL is the minimum one, update the tail lsn in the
  * log manager.
  *
- * Increment the AIL's generation count to indicate that the tree
- * has changed.
- *
  * This function must be called with the AIL lock held.  The lock
  * is dropped before returning.
  */
 void
-xfs_trans_update_ail(
-	xfs_mount_t	*mp,
+xfs_trans_ail_update(
+	struct xfs_ail	*ailp,
 	xfs_log_item_t	*lip,
-	xfs_lsn_t	lsn) __releases(mp->m_ail_lock)
+	xfs_lsn_t	lsn) __releases(ailp->xa_lock)
 {
-	xfs_log_item_t		*dlip=NULL;
+	xfs_log_item_t		*dlip = NULL;
 	xfs_log_item_t		*mlip;	/* ptr to minimum lip */
 
-	mlip = xfs_ail_min(&mp->m_ail);
+	mlip = xfs_ail_min(ailp);
 
 	if (lip->li_flags & XFS_LI_IN_AIL) {
-		dlip = xfs_ail_delete(&mp->m_ail, lip);
+		dlip = xfs_ail_delete(ailp, lip);
 		ASSERT(dlip == lip);
+		xfs_trans_ail_cursor_clear(ailp, dlip);
 	} else {
 		lip->li_flags |= XFS_LI_IN_AIL;
 	}
 
 	lip->li_lsn = lsn;
-
-	xfs_ail_insert(&mp->m_ail, lip);
-	mp->m_ail.xa_gen++;
+	xfs_ail_insert(ailp, lip);
 
 	if (mlip == dlip) {
-		mlip = xfs_ail_min(&mp->m_ail);
-		spin_unlock(&mp->m_ail_lock);
-		xfs_log_move_tail(mp, mlip->li_lsn);
+		mlip = xfs_ail_min(ailp);
+		spin_unlock(&ailp->xa_lock);
+		xfs_log_move_tail(ailp->xa_mount, mlip->li_lsn);
 	} else {
-		spin_unlock(&mp->m_ail_lock);
+		spin_unlock(&ailp->xa_lock);
 	}
 
 
@@ -403,29 +508,30 @@ xfs_trans_update_ail(
  * is dropped before returning.
  */
 void
-xfs_trans_delete_ail(
-	xfs_mount_t	*mp,
-	xfs_log_item_t	*lip) __releases(mp->m_ail_lock)
+xfs_trans_ail_delete(
+	struct xfs_ail	*ailp,
+	xfs_log_item_t	*lip) __releases(ailp->xa_lock)
 {
 	xfs_log_item_t		*dlip;
 	xfs_log_item_t		*mlip;
 
 	if (lip->li_flags & XFS_LI_IN_AIL) {
-		mlip = xfs_ail_min(&mp->m_ail);
-		dlip = xfs_ail_delete(&mp->m_ail, lip);
+		mlip = xfs_ail_min(ailp);
+		dlip = xfs_ail_delete(ailp, lip);
 		ASSERT(dlip == lip);
+		xfs_trans_ail_cursor_clear(ailp, dlip);
 
 
 		lip->li_flags &= ~XFS_LI_IN_AIL;
 		lip->li_lsn = 0;
-		mp->m_ail.xa_gen++;
 
 		if (mlip == dlip) {
-			mlip = xfs_ail_min(&mp->m_ail);
-			spin_unlock(&mp->m_ail_lock);
-			xfs_log_move_tail(mp, (mlip ? mlip->li_lsn : 0));
+			mlip = xfs_ail_min(ailp);
+			spin_unlock(&ailp->xa_lock);
+			xfs_log_move_tail(ailp->xa_mount,
+						(mlip ? mlip->li_lsn : 0));
 		} else {
-			spin_unlock(&mp->m_ail_lock);
+			spin_unlock(&ailp->xa_lock);
 		}
 	}
 	else {
@@ -433,13 +539,13 @@ xfs_trans_delete_ail(
 		 * If the file system is not being shutdown, we are in
 		 * serious trouble if we get to this stage.
 		 */
-		if (XFS_FORCED_SHUTDOWN(mp))
-			spin_unlock(&mp->m_ail_lock);
-		else {
+		struct xfs_mount	*mp = ailp->xa_mount;
+
+		spin_unlock(&ailp->xa_lock);
+		if (!XFS_FORCED_SHUTDOWN(mp)) {
 			xfs_cmn_err(XFS_PTAG_AILDELETE, CE_ALERT, mp,
 		"%s: attempting to delete a log item that is not in the AIL",
 					__func__);
-			spin_unlock(&mp->m_ail_lock);
 			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 		}
 	}
@@ -448,56 +554,6 @@ xfs_trans_delete_ail(
 
 
 /*
- * Return the item in the AIL with the smallest lsn.
- * Return the current tree generation number for use
- * in calls to xfs_trans_next_ail().
- */
-xfs_log_item_t *
-xfs_trans_first_ail(
-	xfs_mount_t	*mp,
-	int		*gen)
-{
-	xfs_log_item_t	*lip;
-
-	lip = xfs_ail_min(&mp->m_ail);
-	*gen = (int)mp->m_ail.xa_gen;
-
-	return lip;
-}
-
-/*
- * If the generation count of the tree has not changed since the
- * caller last took something from the AIL, then return the elmt
- * in the tree which follows the one given.  If the count has changed,
- * then return the minimum elmt of the AIL and bump the restarts counter
- * if one is given.
- */
-xfs_log_item_t *
-xfs_trans_next_ail(
-	xfs_mount_t	*mp,
-	xfs_log_item_t	*lip,
-	int		*gen,
-	int		*restarts)
-{
-	xfs_log_item_t	*nlip;
-
-	ASSERT(mp && lip && gen);
-	if (mp->m_ail.xa_gen == *gen) {
-		nlip = xfs_ail_next(&mp->m_ail, lip);
-	} else {
-		nlip = xfs_ail_min(&mp->m_ail);
-		*gen = (int)mp->m_ail.xa_gen;
-		if (restarts != NULL) {
-			XFS_STATS_INC(xs_push_ail_restarts);
-			(*restarts)++;
-		}
-	}
-
-	return (nlip);
-}
-
-
-/*
  * The active item list (AIL) is a doubly linked list of log
  * items sorted by ascending lsn.  The base of the list is
  * a forw/back pointer pair embedded in the xfs mount structure.
@@ -515,15 +571,35 @@ int
 xfs_trans_ail_init(
 	xfs_mount_t	*mp)
 {
-	INIT_LIST_HEAD(&mp->m_ail.xa_ail);
-	return xfsaild_start(mp);
+	struct xfs_ail	*ailp;
+	int		error;
+
+	ailp = kmem_zalloc(sizeof(struct xfs_ail), KM_MAYFAIL);
+	if (!ailp)
+		return ENOMEM;
+
+	ailp->xa_mount = mp;
+	INIT_LIST_HEAD(&ailp->xa_ail);
+	spin_lock_init(&ailp->xa_lock);
+	error = xfsaild_start(ailp);
+	if (error)
+		goto out_free_ailp;
+	mp->m_ail = ailp;
+	return 0;
+
+out_free_ailp:
+	kmem_free(ailp);
+	return error;
 }
 
 void
 xfs_trans_ail_destroy(
 	xfs_mount_t	*mp)
 {
-	xfsaild_stop(mp);
+	struct xfs_ail	*ailp = mp->m_ail;
+
+	xfsaild_stop(ailp);
+	kmem_free(ailp);
 }
 
 /*
@@ -534,7 +610,7 @@ xfs_trans_ail_destroy(
  */
 STATIC void
 xfs_ail_insert(
-	xfs_ail_t	*ailp,
+	struct xfs_ail	*ailp,
 	xfs_log_item_t	*lip)
 /* ARGSUSED */
 {
@@ -568,7 +644,7 @@ xfs_ail_insert(
 /*ARGSUSED*/
 STATIC xfs_log_item_t *
 xfs_ail_delete(
-	xfs_ail_t	*ailp,
+	struct xfs_ail	*ailp,
 	xfs_log_item_t	*lip)
 /* ARGSUSED */
 {
@@ -585,7 +661,7 @@ xfs_ail_delete(
  */
 STATIC xfs_log_item_t *
 xfs_ail_min(
-	xfs_ail_t	*ailp)
+	struct xfs_ail	*ailp)
 /* ARGSUSED */
 {
 	if (list_empty(&ailp->xa_ail))
@@ -601,7 +677,7 @@ xfs_ail_min(
  */
 STATIC xfs_log_item_t *
 xfs_ail_next(
-	xfs_ail_t	*ailp,
+	struct xfs_ail	*ailp,
 	xfs_log_item_t	*lip)
 /* ARGSUSED */
 {
@@ -617,7 +693,7 @@ xfs_ail_next(
  */
 STATIC void
 xfs_ail_check(
-	xfs_ail_t 	*ailp,
+	struct xfs_ail	*ailp,
 	xfs_log_item_t	*lip)
 {
 	xfs_log_item_t	*prev_lip;
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 4e855b5ced6..8ee2f8c8b0a 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -527,9 +527,8 @@ xfs_trans_brelse(xfs_trans_t	*tp,
 			lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *);
 			if (lip->li_type == XFS_LI_BUF) {
 				bip = XFS_BUF_FSPRIVATE(bp,xfs_buf_log_item_t*);
-				xfs_trans_unlocked_item(
-						bip->bli_item.li_mountp,
-						lip);
+				xfs_trans_unlocked_item(bip->bli_item.li_ailp,
+							lip);
 			}
 		}
 		xfs_buf_relse(bp);
@@ -626,7 +625,7 @@ xfs_trans_brelse(xfs_trans_t	*tp,
 	 * tell the AIL that the buffer is being unlocked.
 	 */
 	if (bip != NULL) {
-		xfs_trans_unlocked_item(bip->bli_item.li_mountp,
+		xfs_trans_unlocked_item(bip->bli_item.li_ailp,
 					(xfs_log_item_t*)bip);
 	}
 
diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c
index 2a1c0f071f9..23d276af2e0 100644
--- a/fs/xfs/xfs_trans_inode.c
+++ b/fs/xfs/xfs_trans_inode.c
@@ -85,7 +85,6 @@ xfs_trans_iget(
 {
 	int			error;
 	xfs_inode_t		*ip;
-	xfs_inode_log_item_t	*iip;
 
 	/*
 	 * If the transaction pointer is NULL, just call the normal
@@ -138,34 +137,7 @@ xfs_trans_iget(
 	}
 	ASSERT(ip != NULL);
 
-	/*
-	 * Get a log_item_desc to point at the new item.
-	 */
-	if (ip->i_itemp == NULL)
-		xfs_inode_item_init(ip, mp);
-	iip = ip->i_itemp;
-	(void) xfs_trans_add_item(tp, (xfs_log_item_t *)(iip));
-
-	xfs_trans_inode_broot_debug(ip);
-
-	/*
-	 * If the IO lock has been acquired, mark that in
-	 * the inode log item so we'll know to unlock it
-	 * when the transaction commits.
-	 */
-	ASSERT(iip->ili_flags == 0);
-	if (lock_flags & XFS_IOLOCK_EXCL) {
-		iip->ili_flags |= XFS_ILI_IOLOCKED_EXCL;
-	} else if (lock_flags & XFS_IOLOCK_SHARED) {
-		iip->ili_flags |= XFS_ILI_IOLOCKED_SHARED;
-	}
-
-	/*
-	 * Initialize i_transp so we can find it with xfs_inode_incore()
-	 * above.
-	 */
-	ip->i_transp = tp;
-
+	xfs_trans_ijoin(tp, ip, lock_flags);
 	*ipp = ip;
 	return 0;
 }
diff --git a/fs/xfs/xfs_trans_item.c b/fs/xfs/xfs_trans_item.c
index 3c666e8317f..e110bf57d7f 100644
--- a/fs/xfs/xfs_trans_item.c
+++ b/fs/xfs/xfs_trans_item.c
@@ -22,6 +22,14 @@
 #include "xfs_inum.h"
 #include "xfs_trans.h"
 #include "xfs_trans_priv.h"
+/* XXX: from here down needed until struct xfs_trans has it's own ailp */
+#include "xfs_bit.h"
+#include "xfs_buf_item.h"
+#include "xfs_sb.h"
+#include "xfs_ag.h"
+#include "xfs_dir2.h"
+#include "xfs_dmapi.h"
+#include "xfs_mount.h"
 
 STATIC int	xfs_trans_unlock_chunk(xfs_log_item_chunk_t *,
 					int, int, xfs_lsn_t);
@@ -79,6 +87,7 @@ xfs_trans_add_item(xfs_trans_t *tp, xfs_log_item_t *lip)
 		lidp->lid_size = 0;
 		lip->li_desc = lidp;
 		lip->li_mountp = tp->t_mountp;
+		lip->li_ailp = tp->t_mountp->m_ail;
 		return lidp;
 	}
 
@@ -120,6 +129,7 @@ xfs_trans_add_item(xfs_trans_t *tp, xfs_log_item_t *lip)
 	lidp->lid_size = 0;
 	lip->li_desc = lidp;
 	lip->li_mountp = tp->t_mountp;
+	lip->li_ailp = tp->t_mountp->m_ail;
 	return lidp;
 }
 
diff --git a/fs/xfs/xfs_trans_priv.h b/fs/xfs/xfs_trans_priv.h
index 3c748c456ed..73e2ad39743 100644
--- a/fs/xfs/xfs_trans_priv.h
+++ b/fs/xfs/xfs_trans_priv.h
@@ -44,25 +44,93 @@ xfs_log_busy_slot_t		*xfs_trans_add_busy(xfs_trans_t *tp,
 						    xfs_extlen_t idx);
 
 /*
- * From xfs_trans_ail.c
+ * AIL traversal cursor.
+ *
+ * Rather than using a generation number for detecting changes in the ail, use
+ * a cursor that is protected by the ail lock. The aild cursor exists in the
+ * struct xfs_ail, but other traversals can declare it on the stack and link it
+ * to the ail list.
+ *
+ * When an object is deleted from or moved int the AIL, the cursor list is
+ * searched to see if the object is a designated cursor item. If it is, it is
+ * deleted from the cursor so that the next time the cursor is used traversal
+ * will return to the start.
+ *
+ * This means a traversal colliding with a removal will cause a restart of the
+ * list scan, rather than any insertion or deletion anywhere in the list. The
+ * low bit of the item pointer is set if the cursor has been invalidated so
+ * that we can tell the difference between invalidation and reaching the end
+ * of the list to trigger traversal restarts.
  */
-void			xfs_trans_update_ail(struct xfs_mount *mp,
-				     struct xfs_log_item *lip, xfs_lsn_t lsn)
-				     __releases(mp->m_ail_lock);
-void			xfs_trans_delete_ail(struct xfs_mount *mp,
-				     struct xfs_log_item *lip)
-				     __releases(mp->m_ail_lock);
-struct xfs_log_item	*xfs_trans_first_ail(struct xfs_mount *, int *);
-struct xfs_log_item	*xfs_trans_next_ail(struct xfs_mount *,
-				     struct xfs_log_item *, int *, int *);
+struct xfs_ail_cursor {
+	struct xfs_ail_cursor	*next;
+	struct xfs_log_item	*item;
+};
 
+/*
+ * Private AIL structures.
+ *
+ * Eventually we need to drive the locking in here as well.
+ */
+struct xfs_ail {
+	struct xfs_mount	*xa_mount;
+	struct list_head	xa_ail;
+	uint			xa_gen;
+	struct task_struct	*xa_task;
+	xfs_lsn_t		xa_target;
+	struct xfs_ail_cursor	xa_cursors;
+	spinlock_t		xa_lock;
+};
 
 /*
- * AIL push thread support
+ * From xfs_trans_ail.c
  */
-long	xfsaild_push(struct xfs_mount *, xfs_lsn_t *);
-void	xfsaild_wakeup(struct xfs_mount *, xfs_lsn_t);
-int	xfsaild_start(struct xfs_mount *);
-void	xfsaild_stop(struct xfs_mount *);
+void			xfs_trans_ail_update(struct xfs_ail *ailp,
+					struct xfs_log_item *lip, xfs_lsn_t lsn)
+					__releases(ailp->xa_lock);
+void			xfs_trans_ail_delete(struct xfs_ail *ailp,
+					struct xfs_log_item *lip)
+					__releases(ailp->xa_lock);
+void			xfs_trans_ail_push(struct xfs_ail *, xfs_lsn_t);
+void			xfs_trans_unlocked_item(struct xfs_ail *,
+					xfs_log_item_t *);
+
+xfs_lsn_t		xfs_trans_ail_tail(struct xfs_ail *ailp);
+
+struct xfs_log_item	*xfs_trans_ail_cursor_first(struct xfs_ail *ailp,
+					struct xfs_ail_cursor *cur,
+					xfs_lsn_t lsn);
+struct xfs_log_item	*xfs_trans_ail_cursor_next(struct xfs_ail *ailp,
+					struct xfs_ail_cursor *cur);
+void			xfs_trans_ail_cursor_done(struct xfs_ail *ailp,
+					struct xfs_ail_cursor *cur);
+
+long	xfsaild_push(struct xfs_ail *, xfs_lsn_t *);
+void	xfsaild_wakeup(struct xfs_ail *, xfs_lsn_t);
+int	xfsaild_start(struct xfs_ail *);
+void	xfsaild_stop(struct xfs_ail *);
 
+#if BITS_PER_LONG != 64
+static inline void
+xfs_trans_ail_copy_lsn(
+	struct xfs_ail	*ailp,
+	xfs_lsn_t	*dst,
+	xfs_lsn_t	*src)
+{
+	ASSERT(sizeof(xfs_lsn_t) == 8);	/* don't lock if it shrinks */
+	spin_lock(&ailp->xa_lock);
+	*dst = *src;
+	spin_unlock(&ailp->xa_lock);
+}
+#else
+static inline void
+xfs_trans_ail_copy_lsn(
+	struct xfs_ail	*ailp,
+	xfs_lsn_t	*dst,
+	xfs_lsn_t	*src)
+{
+	ASSERT(sizeof(xfs_lsn_t) == 8);
+	*dst = *src;
+}
+#endif
 #endif	/* __XFS_TRANS_PRIV_H__ */
diff --git a/fs/xfs/xfs_utils.c b/fs/xfs/xfs_utils.c
index 35d4d414bcc..fcc2285d03e 100644
--- a/fs/xfs/xfs_utils.c
+++ b/fs/xfs/xfs_utils.c
@@ -172,6 +172,12 @@ xfs_dir_ialloc(
 			*ipp = NULL;
 			return code;
 		}
+
+		/*
+		 * transaction commit worked ok so we can drop the extra ticket
+		 * reference that we gained in xfs_trans_dup()
+		 */
+		xfs_log_ticket_put(tp->t_ticket);
 		code = xfs_trans_reserve(tp, 0, log_res, 0,
 					 XFS_TRANS_PERM_LOG_RES, log_count);
 		/*
@@ -268,9 +274,9 @@ xfs_bump_ino_vers2(
 	xfs_mount_t	*mp;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT(ip->i_d.di_version == XFS_DINODE_VERSION_1);
+	ASSERT(ip->i_d.di_version == 1);
 
-	ip->i_d.di_version = XFS_DINODE_VERSION_2;
+	ip->i_d.di_version = 2;
 	ip->i_d.di_onlink = 0;
 	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
 	mp = tp->t_mountp;
@@ -302,7 +308,7 @@ xfs_bumplink(
 	ASSERT(ip->i_d.di_nlink > 0);
 	ip->i_d.di_nlink++;
 	inc_nlink(VFS_I(ip));
-	if ((ip->i_d.di_version == XFS_DINODE_VERSION_1) &&
+	if ((ip->i_d.di_version == 1) &&
 	    (ip->i_d.di_nlink > XFS_MAXLINK_1)) {
 		/*
 		 * The inode has increased its number of links beyond
diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c
deleted file mode 100644
index 439dd3939dd..00000000000
--- a/fs/xfs/xfs_vfsops.c
+++ /dev/null
@@ -1,757 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_types.h"
-#include "xfs_bit.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
-#include "xfs_mount.h"
-#include "xfs_da_btree.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
-#include "xfs_inode_item.h"
-#include "xfs_btree.h"
-#include "xfs_alloc.h"
-#include "xfs_ialloc.h"
-#include "xfs_quota.h"
-#include "xfs_error.h"
-#include "xfs_bmap.h"
-#include "xfs_rw.h"
-#include "xfs_buf_item.h"
-#include "xfs_log_priv.h"
-#include "xfs_dir2_trace.h"
-#include "xfs_extfree_item.h"
-#include "xfs_acl.h"
-#include "xfs_attr.h"
-#include "xfs_clnt.h"
-#include "xfs_mru_cache.h"
-#include "xfs_filestream.h"
-#include "xfs_fsops.h"
-#include "xfs_vnodeops.h"
-#include "xfs_vfsops.h"
-#include "xfs_utils.h"
-
-
-STATIC void
-xfs_quiesce_fs(
-	xfs_mount_t		*mp)
-{
-	int			count = 0, pincount;
-
-	xfs_flush_buftarg(mp->m_ddev_targp, 0);
-	xfs_finish_reclaim_all(mp, 0);
-
-	/* This loop must run at least twice.
-	 * The first instance of the loop will flush
-	 * most meta data but that will generate more
-	 * meta data (typically directory updates).
-	 * Which then must be flushed and logged before
-	 * we can write the unmount record.
-	 */
-	do {
-		xfs_syncsub(mp, SYNC_INODE_QUIESCE, NULL);
-		pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1);
-		if (!pincount) {
-			delay(50);
-			count++;
-		}
-	} while (count < 2);
-}
-
-/*
- * Second stage of a quiesce. The data is already synced, now we have to take
- * care of the metadata. New transactions are already blocked, so we need to
- * wait for any remaining transactions to drain out before proceding.
- */
-void
-xfs_attr_quiesce(
-	xfs_mount_t	*mp)
-{
-	int	error = 0;
-
-	/* wait for all modifications to complete */
-	while (atomic_read(&mp->m_active_trans) > 0)
-		delay(100);
-
-	/* flush inodes and push all remaining buffers out to disk */
-	xfs_quiesce_fs(mp);
-
-	ASSERT_ALWAYS(atomic_read(&mp->m_active_trans) == 0);
-
-	/* Push the superblock and write an unmount record */
-	error = xfs_log_sbcount(mp, 1);
-	if (error)
-		xfs_fs_cmn_err(CE_WARN, mp,
-				"xfs_attr_quiesce: failed to log sb changes. "
-				"Frozen image may not be consistent.");
-	xfs_log_unmount_write(mp);
-	xfs_unmountfs_writesb(mp);
-}
-
-/*
- * xfs_unmount_flush implements a set of flush operation on special
- * inodes, which are needed as a separate set of operations so that
- * they can be called as part of relocation process.
- */
-int
-xfs_unmount_flush(
-	xfs_mount_t	*mp,		/* Mount structure we are getting
-					   rid of. */
-	int             relocation)	/* Called from vfs relocation. */
-{
-	xfs_inode_t	*rip = mp->m_rootip;
-	xfs_inode_t	*rbmip;
-	xfs_inode_t	*rsumip = NULL;
-	int		error;
-
-	xfs_ilock(rip, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
-	xfs_iflock(rip);
-
-	/*
-	 * Flush out the real time inodes.
-	 */
-	if ((rbmip = mp->m_rbmip) != NULL) {
-		xfs_ilock(rbmip, XFS_ILOCK_EXCL);
-		xfs_iflock(rbmip);
-		error = xfs_iflush(rbmip, XFS_IFLUSH_SYNC);
-		xfs_iunlock(rbmip, XFS_ILOCK_EXCL);
-
-		if (error == EFSCORRUPTED)
-			goto fscorrupt_out;
-
-		ASSERT(vn_count(VFS_I(rbmip)) == 1);
-
-		rsumip = mp->m_rsumip;
-		xfs_ilock(rsumip, XFS_ILOCK_EXCL);
-		xfs_iflock(rsumip);
-		error = xfs_iflush(rsumip, XFS_IFLUSH_SYNC);
-		xfs_iunlock(rsumip, XFS_ILOCK_EXCL);
-
-		if (error == EFSCORRUPTED)
-			goto fscorrupt_out;
-
-		ASSERT(vn_count(VFS_I(rsumip)) == 1);
-	}
-
-	/*
-	 * Synchronously flush root inode to disk
-	 */
-	error = xfs_iflush(rip, XFS_IFLUSH_SYNC);
-	if (error == EFSCORRUPTED)
-		goto fscorrupt_out2;
-
-	if (vn_count(VFS_I(rip)) != 1 && !relocation) {
-		xfs_iunlock(rip, XFS_ILOCK_EXCL);
-		return XFS_ERROR(EBUSY);
-	}
-
-	/*
-	 * Release dquot that rootinode, rbmino and rsumino might be holding,
-	 * flush and purge the quota inodes.
-	 */
-	error = XFS_QM_UNMOUNT(mp);
-	if (error == EFSCORRUPTED)
-		goto fscorrupt_out2;
-
-	if (rbmip) {
-		IRELE(rbmip);
-		IRELE(rsumip);
-	}
-
-	xfs_iunlock(rip, XFS_ILOCK_EXCL);
-	return 0;
-
-fscorrupt_out:
-	xfs_ifunlock(rip);
-
-fscorrupt_out2:
-	xfs_iunlock(rip, XFS_ILOCK_EXCL);
-
-	return XFS_ERROR(EFSCORRUPTED);
-}
-
-/*
- * xfs_sync flushes any pending I/O to file system vfsp.
- *
- * This routine is called by vfs_sync() to make sure that things make it
- * out to disk eventually, on sync() system calls to flush out everything,
- * and when the file system is unmounted.  For the vfs_sync() case, all
- * we really need to do is sync out the log to make all of our meta-data
- * updates permanent (except for timestamps).  For calls from pflushd(),
- * dirty pages are kept moving by calling pdflush() on the inodes
- * containing them.  We also flush the inodes that we can lock without
- * sleeping and the superblock if we can lock it without sleeping from
- * vfs_sync() so that items at the tail of the log are always moving out.
- *
- * Flags:
- *      SYNC_BDFLUSH - We're being called from vfs_sync() so we don't want
- *		       to sleep if we can help it.  All we really need
- *		       to do is ensure that the log is synced at least
- *		       periodically.  We also push the inodes and
- *		       superblock if we can lock them without sleeping
- *			and they are not pinned.
- *      SYNC_ATTR    - We need to flush the inodes.  If SYNC_BDFLUSH is not
- *		       set, then we really want to lock each inode and flush
- *		       it.
- *      SYNC_WAIT    - All the flushes that take place in this call should
- *		       be synchronous.
- *      SYNC_DELWRI  - This tells us to push dirty pages associated with
- *		       inodes.  SYNC_WAIT and SYNC_BDFLUSH are used to
- *		       determine if they should be flushed sync, async, or
- *		       delwri.
- *      SYNC_CLOSE   - This flag is passed when the system is being
- *		       unmounted.  We should sync and invalidate everything.
- *      SYNC_FSDATA  - This indicates that the caller would like to make
- *		       sure the superblock is safe on disk.  We can ensure
- *		       this by simply making sure the log gets flushed
- *		       if SYNC_BDFLUSH is set, and by actually writing it
- *		       out otherwise.
- *	SYNC_IOWAIT  - The caller wants us to wait for all data I/O to complete
- *		       before we return (including direct I/O). Forms the drain
- *		       side of the write barrier needed to safely quiesce the
- *		       filesystem.
- *
- */
-int
-xfs_sync(
-	xfs_mount_t	*mp,
-	int		flags)
-{
-	int		error;
-
-	/*
-	 * Get the Quota Manager to flush the dquots.
-	 *
-	 * If XFS quota support is not enabled or this filesystem
-	 * instance does not use quotas XFS_QM_DQSYNC will always
-	 * return zero.
-	 */
-	error = XFS_QM_DQSYNC(mp, flags);
-	if (error) {
-		/*
-		 * If we got an IO error, we will be shutting down.
-		 * So, there's nothing more for us to do here.
-		 */
-		ASSERT(error != EIO || XFS_FORCED_SHUTDOWN(mp));
-		if (XFS_FORCED_SHUTDOWN(mp))
-			return XFS_ERROR(error);
-	}
-
-	if (flags & SYNC_IOWAIT)
-		xfs_filestream_flush(mp);
-
-	return xfs_syncsub(mp, flags, NULL);
-}
-
-/*
- * xfs sync routine for internal use
- *
- * This routine supports all of the flags defined for the generic vfs_sync
- * interface as explained above under xfs_sync.
- *
- */
-int
-xfs_sync_inodes(
-	xfs_mount_t	*mp,
-	int		flags,
-	int             *bypassed)
-{
-	xfs_inode_t	*ip = NULL;
-	struct inode	*vp = NULL;
-	int		error;
-	int		last_error;
-	uint64_t	fflag;
-	uint		lock_flags;
-	uint		base_lock_flags;
-	boolean_t	mount_locked;
-	boolean_t	vnode_refed;
-	int		preempt;
-	xfs_iptr_t	*ipointer;
-#ifdef DEBUG
-	boolean_t	ipointer_in = B_FALSE;
-
-#define IPOINTER_SET	ipointer_in = B_TRUE
-#define IPOINTER_CLR	ipointer_in = B_FALSE
-#else
-#define IPOINTER_SET
-#define IPOINTER_CLR
-#endif
-
-
-/* Insert a marker record into the inode list after inode ip. The list
- * must be locked when this is called. After the call the list will no
- * longer be locked.
- */
-#define IPOINTER_INSERT(ip, mp)	{ \
-		ASSERT(ipointer_in == B_FALSE); \
-		ipointer->ip_mnext = ip->i_mnext; \
-		ipointer->ip_mprev = ip; \
-		ip->i_mnext = (xfs_inode_t *)ipointer; \
-		ipointer->ip_mnext->i_mprev = (xfs_inode_t *)ipointer; \
-		preempt = 0; \
-		XFS_MOUNT_IUNLOCK(mp); \
-		mount_locked = B_FALSE; \
-		IPOINTER_SET; \
-	}
-
-/* Remove the marker from the inode list. If the marker was the only item
- * in the list then there are no remaining inodes and we should zero out
- * the whole list. If we are the current head of the list then move the head
- * past us.
- */
-#define IPOINTER_REMOVE(ip, mp)	{ \
-		ASSERT(ipointer_in == B_TRUE); \
-		if (ipointer->ip_mnext != (xfs_inode_t *)ipointer) { \
-			ip = ipointer->ip_mnext; \
-			ip->i_mprev = ipointer->ip_mprev; \
-			ipointer->ip_mprev->i_mnext = ip; \
-			if (mp->m_inodes == (xfs_inode_t *)ipointer) { \
-				mp->m_inodes = ip; \
-			} \
-		} else { \
-			ASSERT(mp->m_inodes == (xfs_inode_t *)ipointer); \
-			mp->m_inodes = NULL; \
-			ip = NULL; \
-		} \
-		IPOINTER_CLR; \
-	}
-
-#define XFS_PREEMPT_MASK	0x7f
-
-	ASSERT(!(flags & SYNC_BDFLUSH));
-
-	if (bypassed)
-		*bypassed = 0;
-	if (mp->m_flags & XFS_MOUNT_RDONLY)
-		return 0;
-	error = 0;
-	last_error = 0;
-	preempt = 0;
-
-	/* Allocate a reference marker */
-	ipointer = (xfs_iptr_t *)kmem_zalloc(sizeof(xfs_iptr_t), KM_SLEEP);
-
-	fflag = XFS_B_ASYNC;		/* default is don't wait */
-	if (flags & SYNC_DELWRI)
-		fflag = XFS_B_DELWRI;
-	if (flags & SYNC_WAIT)
-		fflag = 0;		/* synchronous overrides all */
-
-	base_lock_flags = XFS_ILOCK_SHARED;
-	if (flags & (SYNC_DELWRI | SYNC_CLOSE)) {
-		/*
-		 * We need the I/O lock if we're going to call any of
-		 * the flush/inval routines.
-		 */
-		base_lock_flags |= XFS_IOLOCK_SHARED;
-	}
-
-	XFS_MOUNT_ILOCK(mp);
-
-	ip = mp->m_inodes;
-
-	mount_locked = B_TRUE;
-	vnode_refed  = B_FALSE;
-
-	IPOINTER_CLR;
-
-	do {
-		ASSERT(ipointer_in == B_FALSE);
-		ASSERT(vnode_refed == B_FALSE);
-
-		lock_flags = base_lock_flags;
-
-		/*
-		 * There were no inodes in the list, just break out
-		 * of the loop.
-		 */
-		if (ip == NULL) {
-			break;
-		}
-
-		/*
-		 * We found another sync thread marker - skip it
-		 */
-		if (ip->i_mount == NULL) {
-			ip = ip->i_mnext;
-			continue;
-		}
-
-		vp = VFS_I(ip);
-
-		/*
-		 * If the vnode is gone then this is being torn down,
-		 * call reclaim if it is flushed, else let regular flush
-		 * code deal with it later in the loop.
-		 */
-
-		if (vp == NULL) {
-			/* Skip ones already in reclaim */
-			if (ip->i_flags & XFS_IRECLAIM) {
-				ip = ip->i_mnext;
-				continue;
-			}
-			if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL) == 0) {
-				ip = ip->i_mnext;
-			} else if ((xfs_ipincount(ip) == 0) &&
-				    xfs_iflock_nowait(ip)) {
-				IPOINTER_INSERT(ip, mp);
-
-				xfs_finish_reclaim(ip, 1,
-						XFS_IFLUSH_DELWRI_ELSE_ASYNC);
-
-				XFS_MOUNT_ILOCK(mp);
-				mount_locked = B_TRUE;
-				IPOINTER_REMOVE(ip, mp);
-			} else {
-				xfs_iunlock(ip, XFS_ILOCK_EXCL);
-				ip = ip->i_mnext;
-			}
-			continue;
-		}
-
-		if (VN_BAD(vp)) {
-			ip = ip->i_mnext;
-			continue;
-		}
-
-		if (XFS_FORCED_SHUTDOWN(mp) && !(flags & SYNC_CLOSE)) {
-			XFS_MOUNT_IUNLOCK(mp);
-			kmem_free(ipointer);
-			return 0;
-		}
-
-		/*
-		 * Try to lock without sleeping.  We're out of order with
-		 * the inode list lock here, so if we fail we need to drop
-		 * the mount lock and try again.  If we're called from
-		 * bdflush() here, then don't bother.
-		 *
-		 * The inode lock here actually coordinates with the
-		 * almost spurious inode lock in xfs_ireclaim() to prevent
-		 * the vnode we handle here without a reference from
-		 * being freed while we reference it.  If we lock the inode
-		 * while it's on the mount list here, then the spurious inode
-		 * lock in xfs_ireclaim() after the inode is pulled from
-		 * the mount list will sleep until we release it here.
-		 * This keeps the vnode from being freed while we reference
-		 * it.
-		 */
-		if (xfs_ilock_nowait(ip, lock_flags) == 0) {
-			if (vp == NULL) {
-				ip = ip->i_mnext;
-				continue;
-			}
-
-			vp = vn_grab(vp);
-			if (vp == NULL) {
-				ip = ip->i_mnext;
-				continue;
-			}
-
-			IPOINTER_INSERT(ip, mp);
-			xfs_ilock(ip, lock_flags);
-
-			ASSERT(vp == VFS_I(ip));
-			ASSERT(ip->i_mount == mp);
-
-			vnode_refed = B_TRUE;
-		}
-
-		/* From here on in the loop we may have a marker record
-		 * in the inode list.
-		 */
-
-		/*
-		 * If we have to flush data or wait for I/O completion
-		 * we need to drop the ilock that we currently hold.
-		 * If we need to drop the lock, insert a marker if we
-		 * have not already done so.
-		 */
-		if ((flags & (SYNC_CLOSE|SYNC_IOWAIT)) ||
-		    ((flags & SYNC_DELWRI) && VN_DIRTY(vp))) {
-			if (mount_locked) {
-				IPOINTER_INSERT(ip, mp);
-			}
-			xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
-			if (flags & SYNC_CLOSE) {
-				/* Shutdown case. Flush and invalidate. */
-				if (XFS_FORCED_SHUTDOWN(mp))
-					xfs_tosspages(ip, 0, -1,
-							     FI_REMAPF);
-				else
-					error = xfs_flushinval_pages(ip,
-							0, -1, FI_REMAPF);
-			} else if ((flags & SYNC_DELWRI) && VN_DIRTY(vp)) {
-				error = xfs_flush_pages(ip, 0,
-							-1, fflag, FI_NONE);
-			}
-
-			/*
-			 * When freezing, we need to wait ensure all I/O (including direct
-			 * I/O) is complete to ensure no further data modification can take
-			 * place after this point
-			 */
-			if (flags & SYNC_IOWAIT)
-				vn_iowait(ip);
-
-			xfs_ilock(ip, XFS_ILOCK_SHARED);
-		}
-
-		if ((flags & SYNC_ATTR) &&
-		    (ip->i_update_core ||
-		     (ip->i_itemp && ip->i_itemp->ili_format.ilf_fields))) {
-			if (mount_locked)
-				IPOINTER_INSERT(ip, mp);
-
-			if (flags & SYNC_WAIT) {
-				xfs_iflock(ip);
-				error = xfs_iflush(ip, XFS_IFLUSH_SYNC);
-
-			/*
-			 * If we can't acquire the flush lock, then the inode
-			 * is already being flushed so don't bother waiting.
-			 *
-			 * If we can lock it then do a delwri flush so we can
-			 * combine multiple inode flushes in each disk write.
-			 */
-			} else if (xfs_iflock_nowait(ip)) {
-				error = xfs_iflush(ip, XFS_IFLUSH_DELWRI);
-			} else if (bypassed) {
-				(*bypassed)++;
-			}
-		}
-
-		if (lock_flags != 0) {
-			xfs_iunlock(ip, lock_flags);
-		}
-
-		if (vnode_refed) {
-			/*
-			 * If we had to take a reference on the vnode
-			 * above, then wait until after we've unlocked
-			 * the inode to release the reference.  This is
-			 * because we can be already holding the inode
-			 * lock when IRELE() calls xfs_inactive().
-			 *
-			 * Make sure to drop the mount lock before calling
-			 * IRELE() so that we don't trip over ourselves if
-			 * we have to go for the mount lock again in the
-			 * inactive code.
-			 */
-			if (mount_locked) {
-				IPOINTER_INSERT(ip, mp);
-			}
-
-			IRELE(ip);
-
-			vnode_refed = B_FALSE;
-		}
-
-		if (error) {
-			last_error = error;
-		}
-
-		/*
-		 * bail out if the filesystem is corrupted.
-		 */
-		if (error == EFSCORRUPTED)  {
-			if (!mount_locked) {
-				XFS_MOUNT_ILOCK(mp);
-				IPOINTER_REMOVE(ip, mp);
-			}
-			XFS_MOUNT_IUNLOCK(mp);
-			ASSERT(ipointer_in == B_FALSE);
-			kmem_free(ipointer);
-			return XFS_ERROR(error);
-		}
-
-		/* Let other threads have a chance at the mount lock
-		 * if we have looped many times without dropping the
-		 * lock.
-		 */
-		if ((++preempt & XFS_PREEMPT_MASK) == 0) {
-			if (mount_locked) {
-				IPOINTER_INSERT(ip, mp);
-			}
-		}
-
-		if (mount_locked == B_FALSE) {
-			XFS_MOUNT_ILOCK(mp);
-			mount_locked = B_TRUE;
-			IPOINTER_REMOVE(ip, mp);
-			continue;
-		}
-
-		ASSERT(ipointer_in == B_FALSE);
-		ip = ip->i_mnext;
-
-	} while (ip != mp->m_inodes);
-
-	XFS_MOUNT_IUNLOCK(mp);
-
-	ASSERT(ipointer_in == B_FALSE);
-
-	kmem_free(ipointer);
-	return XFS_ERROR(last_error);
-}
-
-/*
- * xfs sync routine for internal use
- *
- * This routine supports all of the flags defined for the generic vfs_sync
- * interface as explained above under xfs_sync.
- *
- */
-int
-xfs_syncsub(
-	xfs_mount_t	*mp,
-	int		flags,
-	int             *bypassed)
-{
-	int		error = 0;
-	int		last_error = 0;
-	uint		log_flags = XFS_LOG_FORCE;
-	xfs_buf_t	*bp;
-	xfs_buf_log_item_t	*bip;
-
-	/*
-	 * Sync out the log.  This ensures that the log is periodically
-	 * flushed even if there is not enough activity to fill it up.
-	 */
-	if (flags & SYNC_WAIT)
-		log_flags |= XFS_LOG_SYNC;
-
-	xfs_log_force(mp, (xfs_lsn_t)0, log_flags);
-
-	if (flags & (SYNC_ATTR|SYNC_DELWRI)) {
-		if (flags & SYNC_BDFLUSH)
-			xfs_finish_reclaim_all(mp, 1);
-		else
-			error = xfs_sync_inodes(mp, flags, bypassed);
-	}
-
-	/*
-	 * Flushing out dirty data above probably generated more
-	 * log activity, so if this isn't vfs_sync() then flush
-	 * the log again.
-	 */
-	if (flags & SYNC_DELWRI) {
-		xfs_log_force(mp, (xfs_lsn_t)0, log_flags);
-	}
-
-	if (flags & SYNC_FSDATA) {
-		/*
-		 * If this is vfs_sync() then only sync the superblock
-		 * if we can lock it without sleeping and it is not pinned.
-		 */
-		if (flags & SYNC_BDFLUSH) {
-			bp = xfs_getsb(mp, XFS_BUF_TRYLOCK);
-			if (bp != NULL) {
-				bip = XFS_BUF_FSPRIVATE(bp,xfs_buf_log_item_t*);
-				if ((bip != NULL) &&
-				    xfs_buf_item_dirty(bip)) {
-					if (!(XFS_BUF_ISPINNED(bp))) {
-						XFS_BUF_ASYNC(bp);
-						error = xfs_bwrite(mp, bp);
-					} else {
-						xfs_buf_relse(bp);
-					}
-				} else {
-					xfs_buf_relse(bp);
-				}
-			}
-		} else {
-			bp = xfs_getsb(mp, 0);
-			/*
-			 * If the buffer is pinned then push on the log so
-			 * we won't get stuck waiting in the write for
-			 * someone, maybe ourselves, to flush the log.
-			 * Even though we just pushed the log above, we
-			 * did not have the superblock buffer locked at
-			 * that point so it can become pinned in between
-			 * there and here.
-			 */
-			if (XFS_BUF_ISPINNED(bp))
-				xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE);
-			if (flags & SYNC_WAIT)
-				XFS_BUF_UNASYNC(bp);
-			else
-				XFS_BUF_ASYNC(bp);
-			error = xfs_bwrite(mp, bp);
-		}
-		if (error) {
-			last_error = error;
-		}
-	}
-
-	/*
-	 * Now check to see if the log needs a "dummy" transaction.
-	 */
-	if (!(flags & SYNC_REMOUNT) && xfs_log_need_covered(mp)) {
-		xfs_trans_t *tp;
-		xfs_inode_t *ip;
-
-		/*
-		 * Put a dummy transaction in the log to tell
-		 * recovery that all others are OK.
-		 */
-		tp = xfs_trans_alloc(mp, XFS_TRANS_DUMMY1);
-		if ((error = xfs_trans_reserve(tp, 0,
-				XFS_ICHANGE_LOG_RES(mp),
-				0, 0, 0)))  {
-			xfs_trans_cancel(tp, 0);
-			return error;
-		}
-
-		ip = mp->m_rootip;
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-
-		xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
-		xfs_trans_ihold(tp, ip);
-		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-		error = xfs_trans_commit(tp, 0);
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		xfs_log_force(mp, (xfs_lsn_t)0, log_flags);
-	}
-
-	/*
-	 * When shutting down, we need to insure that the AIL is pushed
-	 * to disk or the filesystem can appear corrupt from the PROM.
-	 */
-	if ((flags & (SYNC_CLOSE|SYNC_WAIT)) == (SYNC_CLOSE|SYNC_WAIT)) {
-		XFS_bflush(mp->m_ddev_targp);
-		if (mp->m_rtdev_targp) {
-			XFS_bflush(mp->m_rtdev_targp);
-		}
-	}
-
-	return XFS_ERROR(last_error);
-}
diff --git a/fs/xfs/xfs_vfsops.h b/fs/xfs/xfs_vfsops.h
deleted file mode 100644
index a74b05087da..00000000000
--- a/fs/xfs/xfs_vfsops.h
+++ /dev/null
@@ -1,16 +0,0 @@
-#ifndef _XFS_VFSOPS_H
-#define _XFS_VFSOPS_H 1
-
-struct cred;
-struct xfs_fid;
-struct inode;
-struct kstatfs;
-struct xfs_mount;
-struct xfs_mount_args;
-
-int xfs_sync(struct xfs_mount *mp, int flags);
-void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
-		int lnnum);
-void xfs_attr_quiesce(struct xfs_mount *mp);
-
-#endif /* _XFS_VFSOPS_H */
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 8b6812f66a1..f07bf8768c3 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -54,33 +54,10 @@
 #include "xfs_vnodeops.h"
 
 int
-xfs_open(
-	xfs_inode_t	*ip)
-{
-	int		mode;
-
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return XFS_ERROR(EIO);
-
-	/*
-	 * If it's a directory with any blocks, read-ahead block 0
-	 * as we're almost certain to have the next operation be a read there.
-	 */
-	if (S_ISDIR(ip->i_d.di_mode) && ip->i_d.di_nextents > 0) {
-		mode = xfs_ilock_map_shared(ip);
-		if (ip->i_d.di_nextents > 0)
-			(void)xfs_da_reada_buf(NULL, ip, 0, XFS_DATA_FORK);
-		xfs_iunlock(ip, mode);
-	}
-	return 0;
-}
-
-int
 xfs_setattr(
 	struct xfs_inode	*ip,
 	struct iattr		*iattr,
-	int			flags,
-	cred_t			*credp)
+	int			flags)
 {
 	xfs_mount_t		*mp = ip->i_mount;
 	struct inode		*inode = VFS_I(ip);
@@ -93,7 +70,6 @@ xfs_setattr(
 	gid_t			gid=0, igid=0;
 	int			timeflags = 0;
 	struct xfs_dquot	*udqp, *gdqp, *olddquot1, *olddquot2;
-	int			file_owner;
 	int			need_iolock = 1;
 
 	xfs_itrace_entry(ip);
@@ -104,6 +80,10 @@ xfs_setattr(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return XFS_ERROR(EIO);
 
+	code = -inode_change_ok(inode, iattr);
+	if (code)
+		return code;
+
 	olddquot1 = olddquot2 = NULL;
 	udqp = gdqp = NULL;
 
@@ -181,62 +161,8 @@ xfs_setattr(
 
 	xfs_ilock(ip, lock_flags);
 
-	/* boolean: are we the file owner? */
-	file_owner = (current_fsuid() == ip->i_d.di_uid);
-
-	/*
-	 * Change various properties of a file.
-	 * Only the owner or users with CAP_FOWNER
-	 * capability may do these things.
-	 */
-	if (mask & (ATTR_MODE|ATTR_UID|ATTR_GID)) {
-		/*
-		 * CAP_FOWNER overrides the following restrictions:
-		 *
-		 * The user ID of the calling process must be equal
-		 * to the file owner ID, except in cases where the
-		 * CAP_FSETID capability is applicable.
-		 */
-		if (!file_owner && !capable(CAP_FOWNER)) {
-			code = XFS_ERROR(EPERM);
-			goto error_return;
-		}
-
-		/*
-		 * CAP_FSETID overrides the following restrictions:
-		 *
-		 * The effective user ID of the calling process shall match
-		 * the file owner when setting the set-user-ID and
-		 * set-group-ID bits on that file.
-		 *
-		 * The effective group ID or one of the supplementary group
-		 * IDs of the calling process shall match the group owner of
-		 * the file when setting the set-group-ID bit on that file
-		 */
-		if (mask & ATTR_MODE) {
-			mode_t m = 0;
-
-			if ((iattr->ia_mode & S_ISUID) && !file_owner)
-				m |= S_ISUID;
-			if ((iattr->ia_mode & S_ISGID) &&
-			    !in_group_p((gid_t)ip->i_d.di_gid))
-				m |= S_ISGID;
-#if 0
-			/* Linux allows this, Irix doesn't. */
-			if ((iattr->ia_mode & S_ISVTX) && !S_ISDIR(ip->i_d.di_mode))
-				m |= S_ISVTX;
-#endif
-			if (m && !capable(CAP_FSETID))
-				iattr->ia_mode &= ~m;
-		}
-	}
-
 	/*
 	 * Change file ownership.  Must be the owner or privileged.
-	 * If the system was configured with the "restricted_chown"
-	 * option, the owner is not permitted to give away the file,
-	 * and can change the group id only to a group of which he
-	 * or she is a member.
 	 */
 	if (mask & (ATTR_UID|ATTR_GID)) {
 		/*
@@ -251,23 +177,6 @@ xfs_setattr(
 		uid = (mask & ATTR_UID) ? iattr->ia_uid : iuid;
 
 		/*
-		 * CAP_CHOWN overrides the following restrictions:
-		 *
-		 * If _POSIX_CHOWN_RESTRICTED is defined, this capability
-		 * shall override the restriction that a process cannot
-		 * change the user ID of a file it owns and the restriction
-		 * that the group ID supplied to the chown() function
-		 * shall be equal to either the group ID or one of the
-		 * supplementary group IDs of the calling process.
-		 */
-		if (restricted_chown &&
-		    (iuid != uid || (igid != gid &&
-				     !in_group_p((gid_t)gid))) &&
-		    !capable(CAP_CHOWN)) {
-			code = XFS_ERROR(EPERM);
-			goto error_return;
-		}
-		/*
 		 * Do a quota reservation only if uid/gid is actually
 		 * going to change.
 		 */
@@ -304,36 +213,22 @@ xfs_setattr(
 			code = XFS_ERROR(EINVAL);
 			goto error_return;
 		}
+
 		/*
 		 * Make sure that the dquots are attached to the inode.
 		 */
-		if ((code = XFS_QM_DQATTACH(mp, ip, XFS_QMOPT_ILOCKED)))
+		code = XFS_QM_DQATTACH(mp, ip, XFS_QMOPT_ILOCKED);
+		if (code)
 			goto error_return;
-	}
-
-	/*
-	 * Change file access or modified times.
-	 */
-	if (mask & (ATTR_ATIME|ATTR_MTIME)) {
-		if (!file_owner) {
-			if ((mask & (ATTR_MTIME_SET|ATTR_ATIME_SET)) &&
-			    !capable(CAP_FOWNER)) {
-				code = XFS_ERROR(EPERM);
-				goto error_return;
-			}
-		}
-	}
 
-	/*
-	 * Now we can make the changes.  Before we join the inode
-	 * to the transaction, if ATTR_SIZE is set then take care of
-	 * the part of the truncation that must be done without the
-	 * inode lock.  This needs to be done before joining the inode
-	 * to the transaction, because the inode cannot be unlocked
-	 * once it is a part of the transaction.
-	 */
-	if (mask & ATTR_SIZE) {
-		code = 0;
+		/*
+		 * Now we can make the changes.  Before we join the inode
+		 * to the transaction, if ATTR_SIZE is set then take care of
+		 * the part of the truncation that must be done without the
+		 * inode lock.  This needs to be done before joining the inode
+		 * to the transaction, because the inode cannot be unlocked
+		 * once it is a part of the transaction.
+		 */
 		if (iattr->ia_size > ip->i_size) {
 			/*
 			 * Do the first part of growing a file: zero any data
@@ -366,7 +261,7 @@ xfs_setattr(
 		}
 
 		/* wait for all I/O to complete */
-		vn_iowait(ip);
+		xfs_ioend_wait(ip);
 
 		if (!code)
 			code = xfs_itruncate_data(ip, iattr->ia_size);
@@ -388,17 +283,10 @@ xfs_setattr(
 		}
 		commit_flags = XFS_TRANS_RELEASE_LOG_RES;
 		xfs_ilock(ip, XFS_ILOCK_EXCL);
-	}
 
-	if (tp) {
 		xfs_trans_ijoin(tp, ip, lock_flags);
 		xfs_trans_ihold(tp, ip);
-	}
 
-	/*
-	 * Truncate file.  Must have write permission and not be a directory.
-	 */
-	if (mask & ATTR_SIZE) {
 		/*
 		 * Only change the c/mtime if we are changing the size
 		 * or we are explicitly asked to change it. This handles
@@ -438,28 +326,13 @@ xfs_setattr(
 			 */
 			xfs_iflags_set(ip, XFS_ITRUNCATED);
 		}
-	}
-
-	/*
-	 * Change file access modes.
-	 */
-	if (mask & ATTR_MODE) {
-		ip->i_d.di_mode &= S_IFMT;
-		ip->i_d.di_mode |= iattr->ia_mode & ~S_IFMT;
-
-		inode->i_mode &= S_IFMT;
-		inode->i_mode |= iattr->ia_mode & ~S_IFMT;
-
-		xfs_trans_log_inode (tp, ip, XFS_ILOG_CORE);
-		timeflags |= XFS_ICHGTIME_CHG;
+	} else if (tp) {
+		xfs_trans_ijoin(tp, ip, lock_flags);
+		xfs_trans_ihold(tp, ip);
 	}
 
 	/*
 	 * Change file ownership.  Must be the owner or privileged.
-	 * If the system was configured with the "restricted_chown"
-	 * option, the owner is not permitted to give away the file,
-	 * and can change the group id only to a group of which he
-	 * or she is a member.
 	 */
 	if (mask & (ATTR_UID|ATTR_GID)) {
 		/*
@@ -503,6 +376,24 @@ xfs_setattr(
 		timeflags |= XFS_ICHGTIME_CHG;
 	}
 
+	/*
+	 * Change file access modes.
+	 */
+	if (mask & ATTR_MODE) {
+		umode_t mode = iattr->ia_mode;
+
+		if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
+			mode &= ~S_ISGID;
+
+		ip->i_d.di_mode &= S_IFMT;
+		ip->i_d.di_mode |= mode & ~S_IFMT;
+
+		inode->i_mode &= S_IFMT;
+		inode->i_mode |= mode & ~S_IFMT;
+
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+		timeflags |= XFS_ICHGTIME_CHG;
+	}
 
 	/*
 	 * Change file access or modified times.
@@ -713,7 +604,7 @@ xfs_fsync(
 		return XFS_ERROR(EIO);
 
 	/* capture size updates in I/O completion before writing the inode. */
-	error = filemap_fdatawait(VFS_I(ip)->i_mapping);
+	error = xfs_wait_on_pages(ip, 0, -1);
 	if (error)
 		return XFS_ERROR(error);
 
@@ -1029,6 +920,12 @@ xfs_inactive_symlink_rmt(
 		goto error0;
 	}
 	/*
+	 * transaction commit worked ok so we can drop the extra ticket
+	 * reference that we gained in xfs_trans_dup()
+	 */
+	xfs_log_ticket_put(tp->t_ticket);
+
+	/*
 	 * Remove the memory for extent descriptions (just bookkeeping).
 	 */
 	if (ip->i_df.if_bytes)
@@ -1625,8 +1522,6 @@ xfs_create(
 		xfs_trans_set_sync(tp);
 	}
 
-	dp->i_gen++;
-
 	/*
 	 * Attach the dquot(s) to the inodes and modify them incore.
 	 * These ids of the inode couldn't have changed since the new
@@ -1993,13 +1888,6 @@ xfs_remove(
 	}
 	xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 
-	/*
-	 * Bump the in memory generation count on the parent
-	 * directory so that other can know that it has changed.
-	 */
-	dp->i_gen++;
-	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
-
 	if (is_dir) {
 		/*
 		 * Drop the link from ip's "..".
@@ -2009,7 +1897,7 @@ xfs_remove(
 			goto out_bmap_cancel;
 
 		/*
-		 * Drop the link from dp to ip.
+		 * Drop the "." link from ip to self.
 		 */
 		error = xfs_droplink(tp, ip);
 		if (error)
@@ -2017,14 +1905,14 @@ xfs_remove(
 	} else {
 		/*
 		 * When removing a non-directory we need to log the parent
-		 * inode here for the i_gen update.  For a directory this is
-		 * done implicitly by the xfs_droplink call for the ".." entry.
+		 * inode here.  For a directory this is done implicitly
+		 * by the xfs_droplink call for the ".." entry.
 		 */
 		xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
 	}
 
 	/*
-	 * Drop the "." link from ip to self.
+	 * Drop the link from dp to ip.
 	 */
 	error = xfs_droplink(tp, ip);
 	if (error)
@@ -2178,7 +2066,6 @@ xfs_link(
 	if (error)
 		goto abort_return;
 	xfs_ichgtime(tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
-	tdp->i_gen++;
 	xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE);
 
 	error = xfs_bumplink(tp, sip);
@@ -2355,18 +2242,10 @@ xfs_mkdir(
 	}
 	xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 
-	/*
-	 * Bump the in memory version number of the parent directory
-	 * so that other processes accessing it will recognize that
-	 * the directory has changed.
-	 */
-	dp->i_gen++;
-
 	error = xfs_dir_init(tp, cdp, dp);
 	if (error)
 		goto error2;
 
-	cdp->i_gen = 1;
 	error = xfs_bumplink(tp, dp);
 	if (error)
 		goto error2;
@@ -2653,13 +2532,6 @@ xfs_symlink(
 	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
 
 	/*
-	 * Bump the in memory version number of the parent directory
-	 * so that other processes accessing it will recognize that
-	 * the directory has changed.
-	 */
-	dp->i_gen++;
-
-	/*
 	 * If this is a synchronous mount, make sure that the
 	 * symlink transaction goes to disk before returning to
 	 * the user.
@@ -2809,7 +2681,7 @@ xfs_reclaim(
 		return 0;
 	}
 
-	vn_iowait(ip);
+	xfs_ioend_wait(ip);
 
 	ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
 
@@ -2833,122 +2705,10 @@ xfs_reclaim(
 	if (!ip->i_update_core && (ip->i_itemp == NULL)) {
 		xfs_ilock(ip, XFS_ILOCK_EXCL);
 		xfs_iflock(ip);
-		return xfs_finish_reclaim(ip, 1, XFS_IFLUSH_DELWRI_ELSE_SYNC);
-	} else {
-		xfs_mount_t	*mp = ip->i_mount;
-
-		/* Protect sync and unpin from us */
-		XFS_MOUNT_ILOCK(mp);
-		spin_lock(&ip->i_flags_lock);
-		__xfs_iflags_set(ip, XFS_IRECLAIMABLE);
-		VFS_I(ip)->i_private = NULL;
-		ip->i_vnode = NULL;
-		spin_unlock(&ip->i_flags_lock);
-		list_add_tail(&ip->i_reclaim, &mp->m_del_inodes);
-		XFS_MOUNT_IUNLOCK(mp);
-	}
-	return 0;
-}
-
-int
-xfs_finish_reclaim(
-	xfs_inode_t	*ip,
-	int		locked,
-	int		sync_mode)
-{
-	xfs_perag_t	*pag = xfs_get_perag(ip->i_mount, ip->i_ino);
-	struct inode	*vp = VFS_I(ip);
-
-	if (vp && VN_BAD(vp))
-		goto reclaim;
-
-	/* The hash lock here protects a thread in xfs_iget_core from
-	 * racing with us on linking the inode back with a vnode.
-	 * Once we have the XFS_IRECLAIM flag set it will not touch
-	 * us.
-	 */
-	write_lock(&pag->pag_ici_lock);
-	spin_lock(&ip->i_flags_lock);
-	if (__xfs_iflags_test(ip, XFS_IRECLAIM) ||
-	    (!__xfs_iflags_test(ip, XFS_IRECLAIMABLE) && vp == NULL)) {
-		spin_unlock(&ip->i_flags_lock);
-		write_unlock(&pag->pag_ici_lock);
-		if (locked) {
-			xfs_ifunlock(ip);
-			xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		}
-		return 1;
-	}
-	__xfs_iflags_set(ip, XFS_IRECLAIM);
-	spin_unlock(&ip->i_flags_lock);
-	write_unlock(&pag->pag_ici_lock);
-	xfs_put_perag(ip->i_mount, pag);
-
-	/*
-	 * If the inode is still dirty, then flush it out.  If the inode
-	 * is not in the AIL, then it will be OK to flush it delwri as
-	 * long as xfs_iflush() does not keep any references to the inode.
-	 * We leave that decision up to xfs_iflush() since it has the
-	 * knowledge of whether it's OK to simply do a delwri flush of
-	 * the inode or whether we need to wait until the inode is
-	 * pulled from the AIL.
-	 * We get the flush lock regardless, though, just to make sure
-	 * we don't free it while it is being flushed.
-	 */
-	if (!locked) {
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		xfs_iflock(ip);
+		xfs_iflags_set(ip, XFS_IRECLAIMABLE);
+		return xfs_reclaim_inode(ip, 1, XFS_IFLUSH_DELWRI_ELSE_SYNC);
 	}
-
-	/*
-	 * In the case of a forced shutdown we rely on xfs_iflush() to
-	 * wait for the inode to be unpinned before returning an error.
-	 */
-	if (xfs_iflush(ip, sync_mode) == 0) {
-		/* synchronize with xfs_iflush_done */
-		xfs_iflock(ip);
-		xfs_ifunlock(ip);
-	}
-
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
- reclaim:
-	xfs_ireclaim(ip);
-	return 0;
-}
-
-int
-xfs_finish_reclaim_all(xfs_mount_t *mp, int noblock)
-{
-	int		purged;
-	xfs_inode_t	*ip, *n;
-	int		done = 0;
-
-	while (!done) {
-		purged = 0;
-		XFS_MOUNT_ILOCK(mp);
-		list_for_each_entry_safe(ip, n, &mp->m_del_inodes, i_reclaim) {
-			if (noblock) {
-				if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL) == 0)
-					continue;
-				if (xfs_ipincount(ip) ||
-				    !xfs_iflock_nowait(ip)) {
-					xfs_iunlock(ip, XFS_ILOCK_EXCL);
-					continue;
-				}
-			}
-			XFS_MOUNT_IUNLOCK(mp);
-			if (xfs_finish_reclaim(ip, noblock,
-					XFS_IFLUSH_DELWRI_ELSE_ASYNC))
-				delay(1);
-			purged = 1;
-			break;
-		}
-
-		done = !purged;
-	}
-
-	XFS_MOUNT_IUNLOCK(mp);
+	xfs_inode_set_reclaim_tag(ip);
 	return 0;
 }
 
@@ -3197,6 +2957,8 @@ xfs_zero_remaining_bytes(
 	bp = xfs_buf_get_noaddr(mp->m_sb.sb_blocksize,
 				XFS_IS_REALTIME_INODE(ip) ?
 				mp->m_rtdev_targp : mp->m_ddev_targp);
+	if (!bp)
+		return XFS_ERROR(ENOMEM);
 
 	for (offset = startoff; offset <= endoff; offset = lastoffset + 1) {
 		offset_fsb = XFS_B_TO_FSBT(mp, offset);
@@ -3312,7 +3074,8 @@ xfs_free_file_space(
 		need_iolock = 0;
 	if (need_iolock) {
 		xfs_ilock(ip, XFS_IOLOCK_EXCL);
-		vn_iowait(ip);	/* wait for the completion of any pending DIOs */
+		/* wait for the completion of any pending DIOs */
+		xfs_ioend_wait(ip);
 	}
 
 	rounding = max_t(uint, 1 << mp->m_sb.sb_blocklog, PAGE_CACHE_SIZE);
@@ -3474,7 +3237,6 @@ xfs_change_file_space(
 	int		cmd,
 	xfs_flock64_t	*bf,
 	xfs_off_t	offset,
-	cred_t		*credp,
 	int		attr_flags)
 {
 	xfs_mount_t	*mp = ip->i_mount;
@@ -3562,7 +3324,7 @@ xfs_change_file_space(
 		iattr.ia_valid = ATTR_SIZE;
 		iattr.ia_size = startoffset;
 
-		error = xfs_setattr(ip, &iattr, attr_flags, credp);
+		error = xfs_setattr(ip, &iattr, attr_flags);
 
 		if (error)
 			return error;
diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h
index 7b0c2ab8833..76df328c61b 100644
--- a/fs/xfs/xfs_vnodeops.h
+++ b/fs/xfs/xfs_vnodeops.h
@@ -14,9 +14,7 @@ struct xfs_inode;
 struct xfs_iomap;
 
 
-int xfs_open(struct xfs_inode *ip);
-int xfs_setattr(struct xfs_inode *ip, struct iattr *vap, int flags,
-		cred_t *credp);
+int xfs_setattr(struct xfs_inode *ip, struct iattr *vap, int flags);
 #define	XFS_ATTR_DMI		0x01	/* invocation from a DMI function */
 #define	XFS_ATTR_NONBLOCK	0x02	/* return EAGAIN if operation would block */
 #define XFS_ATTR_NOLOCK		0x04	/* Don't grab any conflicting locks */
@@ -44,8 +42,7 @@ int xfs_inode_flush(struct xfs_inode *ip, int flags);
 int xfs_set_dmattrs(struct xfs_inode *ip, u_int evmask, u_int16_t state);
 int xfs_reclaim(struct xfs_inode *ip);
 int xfs_change_file_space(struct xfs_inode *ip, int cmd,
-		xfs_flock64_t *bf, xfs_off_t offset,
-		cred_t *credp, int	attr_flags);
+		xfs_flock64_t *bf, xfs_off_t offset, int attr_flags);
 int xfs_rename(struct xfs_inode *src_dp, struct xfs_name *src_name,
 		struct xfs_inode *src_ip, struct xfs_inode *target_dp,
 		struct xfs_name *target_name, struct xfs_inode *target_ip);
@@ -56,8 +53,6 @@ int xfs_attr_set(struct xfs_inode *dp, const char *name, char *value,
 int xfs_attr_remove(struct xfs_inode *dp, const char *name, int flags);
 int xfs_attr_list(struct xfs_inode *dp, char *buffer, int bufsize,
 		int flags, struct attrlist_cursor_kern *cursor);
-int xfs_ioctl(struct xfs_inode *ip, struct file *filp,
-		int ioflags, unsigned int cmd, void __user *arg);
 ssize_t xfs_read(struct xfs_inode *ip, struct kiocb *iocb,
 		const struct iovec *iovp, unsigned int segs,
 		loff_t *offset, int ioflags);
@@ -78,5 +73,6 @@ int xfs_flushinval_pages(struct xfs_inode *ip, xfs_off_t first,
 		xfs_off_t last, int fiopt);
 int xfs_flush_pages(struct xfs_inode *ip, xfs_off_t first,
 		xfs_off_t last, uint64_t flags, int fiopt);
+int xfs_wait_on_pages(struct xfs_inode *ip, xfs_off_t first, xfs_off_t last);
 
 #endif /* _XFS_VNODEOPS_H */