Merge commit 'linus/master' into tracing/kprobes

Conflicts: kernel/trace/Makefile kernel/trace/trace.h kernel/trace/trace_event_types.h kernel/trace/trace_export.c Merge reason: Sync with latest significant tracing core changes.
author: Frederic Weisbecker <fweisbec@gmail.com> 2009-09-23 23:08:43 +0200
committer: Frederic Weisbecker <fweisbec@gmail.com> 2009-09-23 23:08:43 +0200
commit: d7a4b414eed51f1653bb05ebe84122bf9a7ae18b (patch)
tree: bd6603a0c27de4c138a1767871897e9cd3e1a1d2 /fs
parent: 1f0ab40976460bc4673fa204ce917a725185d8f2 (diff)
parent: a724eada8c2a7b62463b73ccf73fd0bb6e928aeb (diff)
352 files changed, 16952 insertions, 8089 deletions
diff --git a/fs/Kconfig b/fs/Kconfig
index 0e7da7bb5d9..d4bf8caad8d 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -43,6 +43,7 @@ source "fs/xfs/Kconfig"
 source "fs/gfs2/Kconfig"
 source "fs/ocfs2/Kconfig"
 source "fs/btrfs/Kconfig"
+source "fs/nilfs2/Kconfig"
 
 endif # BLOCK
 
@@ -108,6 +109,7 @@ source "fs/sysfs/Kconfig"
 
 config TMPFS
 	bool "Virtual memory file system support (former shm fs)"
+	depends on SHMEM
 	help
 	  Tmpfs is a file system which keeps all files in virtual memory.
 
@@ -186,7 +188,6 @@ source "fs/romfs/Kconfig"
 source "fs/sysv/Kconfig"
 source "fs/ufs/Kconfig"
 source "fs/exofs/Kconfig"
-source "fs/nilfs2/Kconfig"
 
 endif # MISC_FILESYSTEMS
 
diff --git a/fs/afs/flock.c b/fs/afs/flock.c
index 3ff8bdd18fb..0931bc1325e 100644
--- a/fs/afs/flock.c
+++ b/fs/afs/flock.c
@@ -21,7 +21,7 @@ static void afs_fl_release_private(struct file_lock *fl);
 static struct workqueue_struct *afs_lock_manager;
 static DEFINE_MUTEX(afs_lock_manager_mutex);
 
-static struct file_lock_operations afs_lock_ops = {
+static const struct file_lock_operations afs_lock_ops = {
 	.fl_copy_lock		= afs_fl_copy_lock,
 	.fl_release_private	= afs_fl_release_private,
 };
diff --git a/fs/afs/proc.c b/fs/afs/proc.c
index 8630615e57f..852739d262a 100644
--- a/fs/afs/proc.c
+++ b/fs/afs/proc.c
@@ -28,7 +28,7 @@ static int afs_proc_cells_show(struct seq_file *m, void *v);
 static ssize_t afs_proc_cells_write(struct file *file, const char __user *buf,
 				    size_t size, loff_t *_pos);
 
-static struct seq_operations afs_proc_cells_ops = {
+static const struct seq_operations afs_proc_cells_ops = {
 	.start	= afs_proc_cells_start,
 	.next	= afs_proc_cells_next,
 	.stop	= afs_proc_cells_stop,
@@ -70,7 +70,7 @@ static void *afs_proc_cell_volumes_next(struct seq_file *p, void *v,
 static void afs_proc_cell_volumes_stop(struct seq_file *p, void *v);
 static int afs_proc_cell_volumes_show(struct seq_file *m, void *v);
 
-static struct seq_operations afs_proc_cell_volumes_ops = {
+static const struct seq_operations afs_proc_cell_volumes_ops = {
 	.start	= afs_proc_cell_volumes_start,
 	.next	= afs_proc_cell_volumes_next,
 	.stop	= afs_proc_cell_volumes_stop,
@@ -95,7 +95,7 @@ static void *afs_proc_cell_vlservers_next(struct seq_file *p, void *v,
 static void afs_proc_cell_vlservers_stop(struct seq_file *p, void *v);
 static int afs_proc_cell_vlservers_show(struct seq_file *m, void *v);
 
-static struct seq_operations afs_proc_cell_vlservers_ops = {
+static const struct seq_operations afs_proc_cell_vlservers_ops = {
 	.start	= afs_proc_cell_vlservers_start,
 	.next	= afs_proc_cell_vlservers_next,
 	.stop	= afs_proc_cell_vlservers_stop,
@@ -119,7 +119,7 @@ static void *afs_proc_cell_servers_next(struct seq_file *p, void *v,
 static void afs_proc_cell_servers_stop(struct seq_file *p, void *v);
 static int afs_proc_cell_servers_show(struct seq_file *m, void *v);
 
-static struct seq_operations afs_proc_cell_servers_ops = {
+static const struct seq_operations afs_proc_cell_servers_ops = {
 	.start	= afs_proc_cell_servers_start,
 	.next	= afs_proc_cell_servers_next,
 	.stop	= afs_proc_cell_servers_stop,
diff --git a/fs/afs/write.c b/fs/afs/write.c
index c2e7a7ff008..c63a3c8beb7 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -712,7 +712,6 @@ int afs_writeback_all(struct afs_vnode *vnode)
 		.bdi		= mapping->backing_dev_info,
 		.sync_mode	= WB_SYNC_ALL,
 		.nr_to_write	= LONG_MAX,
-		.for_writepages = 1,
 		.range_cyclic	= 1,
 	};
 	int ret;
diff --git a/fs/aio.c b/fs/aio.c
index d065b2c3273..02a2c934057 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -24,6 +24,7 @@
 #include <linux/file.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
+#include <linux/mmu_context.h>
 #include <linux/slab.h>
 #include <linux/timer.h>
 #include <linux/aio.h>
@@ -34,7 +35,6 @@
 
 #include <asm/kmap_types.h>
 #include <asm/uaccess.h>
-#include <asm/mmu_context.h>
 
 #if DEBUG > 1
 #define dprintk		printk
@@ -78,6 +78,7 @@ static int __init aio_setup(void)
 
 	return 0;
 }
+__initcall(aio_setup);
 
 static void aio_free_ring(struct kioctx *ctx)
 {
@@ -380,6 +381,7 @@ ssize_t wait_on_sync_kiocb(struct kiocb *iocb)
 	__set_current_state(TASK_RUNNING);
 	return iocb->ki_user_data;
 }
+EXPORT_SYMBOL(wait_on_sync_kiocb);
 
 /* exit_aio: called when the last user of mm goes away.  At this point, 
  * there is no way for any new requests to be submited or any of the 
@@ -573,6 +575,7 @@ int aio_put_req(struct kiocb *req)
 	spin_unlock_irq(&ctx->ctx_lock);
 	return ret;
 }
+EXPORT_SYMBOL(aio_put_req);
 
 static struct kioctx *lookup_ioctx(unsigned long ctx_id)
 {
@@ -595,51 +598,6 @@ static struct kioctx *lookup_ioctx(unsigned long ctx_id)
 }
 
 /*
- * use_mm
- *	Makes the calling kernel thread take on the specified
- *	mm context.
- *	Called by the retry thread execute retries within the
- *	iocb issuer's mm context, so that copy_from/to_user
- *	operations work seamlessly for aio.
- *	(Note: this routine is intended to be called only
- *	from a kernel thread context)
- */
-static void use_mm(struct mm_struct *mm)
-{
-	struct mm_struct *active_mm;
-	struct task_struct *tsk = current;
-
-	task_lock(tsk);
-	active_mm = tsk->active_mm;
-	atomic_inc(&mm->mm_count);
-	tsk->mm = mm;
-	tsk->active_mm = mm;
-	switch_mm(active_mm, mm, tsk);
-	task_unlock(tsk);
-
-	mmdrop(active_mm);
-}
-
-/*
- * unuse_mm
- *	Reverses the effect of use_mm, i.e. releases the
- *	specified mm context which was earlier taken on
- *	by the calling kernel thread
- *	(Note: this routine is intended to be called only
- *	from a kernel thread context)
- */
-static void unuse_mm(struct mm_struct *mm)
-{
-	struct task_struct *tsk = current;
-
-	task_lock(tsk);
-	tsk->mm = NULL;
-	/* active_mm is still 'mm' */
-	enter_lazy_tlb(mm, tsk);
-	task_unlock(tsk);
-}
-
-/*
  * Queue up a kiocb to be retried. Assumes that the kiocb
  * has already been marked as kicked, and places it on
  * the retry run list for the corresponding ioctx, if it
@@ -1037,6 +995,7 @@ put_rq:
 	spin_unlock_irqrestore(&ctx->ctx_lock, flags);
 	return ret;
 }
+EXPORT_SYMBOL(aio_complete);
 
 /* aio_read_evt
  *	Pull an event off of the ioctx's event ring.  Returns the number of 
@@ -1825,9 +1784,3 @@ SYSCALL_DEFINE5(io_getevents, aio_context_t, ctx_id,
 	asmlinkage_protect(5, ret, ctx_id, min_nr, nr, events, timeout);
 	return ret;
 }
-
-__initcall(aio_setup);
-
-EXPORT_SYMBOL(aio_complete);
-EXPORT_SYMBOL(aio_put_req);
-EXPORT_SYMBOL(wait_on_sync_kiocb);
diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c
index 47d4a01c539..d11c51fc2a3 100644
--- a/fs/anon_inodes.c
+++ b/fs/anon_inodes.c
@@ -77,28 +77,24 @@ static const struct address_space_operations anon_aops = {
  *
  * Creates a new file by hooking it on a single inode. This is useful for files
  * that do not need to have a full-fledged inode in order to operate correctly.
- * All the files created with anon_inode_getfd() will share a single inode,
+ * All the files created with anon_inode_getfile() will share a single inode,
  * hence saving memory and avoiding code duplication for the file/inode/dentry
- * setup.  Returns new descriptor or -error.
+ * setup.  Returns the newly created file* or an error pointer.
  */
-int anon_inode_getfd(const char *name, const struct file_operations *fops,
-		     void *priv, int flags)
+struct file *anon_inode_getfile(const char *name,
+				const struct file_operations *fops,
+				void *priv, int flags)
 {
 	struct qstr this;
 	struct dentry *dentry;
 	struct file *file;
-	int error, fd;
+	int error;
 
 	if (IS_ERR(anon_inode_inode))
-		return -ENODEV;
+		return ERR_PTR(-ENODEV);
 
 	if (fops->owner && !try_module_get(fops->owner))
-		return -ENOENT;
-
-	error = get_unused_fd_flags(flags);
-	if (error < 0)
-		goto err_module;
-	fd = error;
+		return ERR_PTR(-ENOENT);
 
 	/*
 	 * Link the inode to a directory entry by creating a unique name
@@ -110,7 +106,7 @@ int anon_inode_getfd(const char *name, const struct file_operations *fops,
 	this.hash = 0;
 	dentry = d_alloc(anon_inode_mnt->mnt_sb->s_root, &this);
 	if (!dentry)
-		goto err_put_unused_fd;
+		goto err_module;
 
 	/*
 	 * We know the anon_inode inode count is always greater than zero,
@@ -136,16 +132,54 @@ int anon_inode_getfd(const char *name, const struct file_operations *fops,
 	file->f_version = 0;
 	file->private_data = priv;
 
+	return file;
+
+err_dput:
+	dput(dentry);
+err_module:
+	module_put(fops->owner);
+	return ERR_PTR(error);
+}
+EXPORT_SYMBOL_GPL(anon_inode_getfile);
+
+/**
+ * anon_inode_getfd - creates a new file instance by hooking it up to an
+ *                    anonymous inode, and a dentry that describe the "class"
+ *                    of the file
+ *
+ * @name:    [in]    name of the "class" of the new file
+ * @fops:    [in]    file operations for the new file
+ * @priv:    [in]    private data for the new file (will be file's private_data)
+ * @flags:   [in]    flags
+ *
+ * Creates a new file by hooking it on a single inode. This is useful for files
+ * that do not need to have a full-fledged inode in order to operate correctly.
+ * All the files created with anon_inode_getfd() will share a single inode,
+ * hence saving memory and avoiding code duplication for the file/inode/dentry
+ * setup.  Returns new descriptor or an error code.
+ */
+int anon_inode_getfd(const char *name, const struct file_operations *fops,
+		     void *priv, int flags)
+{
+	int error, fd;
+	struct file *file;
+
+	error = get_unused_fd_flags(flags);
+	if (error < 0)
+		return error;
+	fd = error;
+
+	file = anon_inode_getfile(name, fops, priv, flags);
+	if (IS_ERR(file)) {
+		error = PTR_ERR(file);
+		goto err_put_unused_fd;
+	}
 	fd_install(fd, file);
 
 	return fd;
 
-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/dirhash.c b/fs/autofs/dirhash.c
index 2316e944a10..e947915109e 100644
--- a/fs/autofs/dirhash.c
+++ b/fs/autofs/dirhash.c
@@ -90,7 +90,7 @@ struct autofs_dir_ent *autofs_expire(struct super_block *sb,
 			DPRINTK(("autofs: not expirable (not a mounted directory): %s\n", ent->name));
 			continue;
 		}
-		while (d_mountpoint(path.dentry) && follow_down(&path));
+		while (d_mountpoint(path.dentry) && follow_down(&path))
 			;
 		umount_ok = may_umount(path.mnt);
 		path_put(&path);
diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c
index 615d5496fe0..dd376c124e7 100644
--- a/fs/befs/linuxvfs.c
+++ b/fs/befs/linuxvfs.c
@@ -842,7 +842,7 @@ befs_fill_super(struct super_block *sb, void *data, int silent)
 	sb->s_magic = BEFS_SUPER_MAGIC;
 	/* Set real blocksize of fs */
 	sb_set_blocksize(sb, (ulong) befs_sb->block_size);
-	sb->s_op = (struct super_operations *) &befs_sops;
+	sb->s_op = &befs_sops;
 	root = befs_iget(sb, iaddr2blockno(sb, &(befs_sb->root_dir)));
 	if (IS_ERR(root)) {
 		ret = PTR_ERR(root);
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index b7c1603cd4b..442d94fe255 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -501,22 +501,22 @@ static unsigned long load_elf_interp(struct elfhdr *interp_elf_ex,
 		}
 	}
 
-	/*
-	 * Now fill out the bss section.  First pad the last page up
-	 * to the page boundary, and then perform a mmap to make sure
-	 * that there are zero-mapped pages up to and including the 
-	 * last bss page.
-	 */
-	if (padzero(elf_bss)) {
-		error = -EFAULT;
-		goto out_close;
-	}
+	if (last_bss > elf_bss) {
+		/*
+		 * Now fill out the bss section.  First pad the last page up
+		 * to the page boundary, and then perform a mmap to make sure
+		 * that there are zero-mapped pages up to and including the
+		 * last bss page.
+		 */
+		if (padzero(elf_bss)) {
+			error = -EFAULT;
+			goto out_close;
+		}
 
-	/* What we have mapped so far */
-	elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1);
+		/* What we have mapped so far */
+		elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1);
 
-	/* Map the last of the bss segment */
-	if (last_bss > elf_bss) {
+		/* Map the last of the bss segment */
 		down_write(&current->mm->mmap_sem);
 		error = do_brk(elf_bss, last_bss - elf_bss);
 		up_write(&current->mm->mmap_sem);
@@ -1280,9 +1280,6 @@ static int writenote(struct memelfnote *men, struct file *file,
 #define DUMP_WRITE(addr, nr)	\
 	if ((size += (nr)) > limit || !dump_write(file, (addr), (nr))) \
 		goto end_coredump;
-#define DUMP_SEEK(off)	\
-	if (!dump_seek(file, (off))) \
-		goto end_coredump;
 
 static void fill_elf_header(struct elfhdr *elf, int segs,
 			    u16 machine, u32 flags, u8 osabi)
@@ -2016,7 +2013,8 @@ static int elf_core_dump(long signr, struct pt_regs *regs, struct file *file, un
 		goto end_coredump;
 
 	/* Align to page */
-	DUMP_SEEK(dataoff - foffset);
+	if (!dump_seek(file, dataoff - foffset))
+		goto end_coredump;
 
 	for (vma = first_vma(current, gate_vma); vma != NULL;
 			vma = next_vma(vma, gate_vma)) {
@@ -2027,33 +2025,19 @@ static int elf_core_dump(long signr, struct pt_regs *regs, struct file *file, un
 
 		for (addr = vma->vm_start; addr < end; addr += PAGE_SIZE) {
 			struct page *page;
-			struct vm_area_struct *tmp_vma;
-
-			if (get_user_pages(current, current->mm, addr, 1, 0, 1,
-						&page, &tmp_vma) <= 0) {
-				DUMP_SEEK(PAGE_SIZE);
-			} else {
-				if (page == ZERO_PAGE(0)) {
-					if (!dump_seek(file, PAGE_SIZE)) {
-						page_cache_release(page);
-						goto end_coredump;
-					}
-				} else {
-					void *kaddr;
-					flush_cache_page(tmp_vma, addr,
-							 page_to_pfn(page));
-					kaddr = kmap(page);
-					if ((size += PAGE_SIZE) > limit ||
-					    !dump_write(file, kaddr,
-					    PAGE_SIZE)) {
-						kunmap(page);
-						page_cache_release(page);
-						goto end_coredump;
-					}
-					kunmap(page);
-				}
+			int stop;
+
+			page = get_dump_page(addr);
+			if (page) {
+				void *kaddr = kmap(page);
+				stop = ((size += PAGE_SIZE) > limit) ||
+					!dump_write(file, kaddr, PAGE_SIZE);
+				kunmap(page);
 				page_cache_release(page);
-			}
+			} else
+				stop = !dump_seek(file, PAGE_SIZE);
+			if (stop)
+				goto end_coredump;
 		}
 	}
 
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index 20fbeced472..76285471073 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -1325,9 +1325,6 @@ static int writenote(struct memelfnote *men, struct file *file)
 #define DUMP_WRITE(addr, nr)	\
 	if ((size += (nr)) > limit || !dump_write(file, (addr), (nr))) \
 		goto end_coredump;
-#define DUMP_SEEK(off)	\
-	if (!dump_seek(file, (off))) \
-		goto end_coredump;
 
 static inline void fill_elf_fdpic_header(struct elfhdr *elf, int segs)
 {
@@ -1518,6 +1515,7 @@ static int elf_fdpic_dump_segments(struct file *file, size_t *size,
 			   unsigned long *limit, unsigned long mm_flags)
 {
 	struct vm_area_struct *vma;
+	int err = 0;
 
 	for (vma = current->mm->mmap; vma; vma = vma->vm_next) {
 		unsigned long addr;
@@ -1525,43 +1523,26 @@ static int elf_fdpic_dump_segments(struct file *file, size_t *size,
 		if (!maydump(vma, mm_flags))
 			continue;
 
-		for (addr = vma->vm_start;
-		     addr < vma->vm_end;
-		     addr += PAGE_SIZE
-		     ) {
-			struct vm_area_struct *vma;
-			struct page *page;
-
-			if (get_user_pages(current, current->mm, addr, 1, 0, 1,
-					   &page, &vma) <= 0) {
-				DUMP_SEEK(file->f_pos + PAGE_SIZE);
-			}
-			else if (page == ZERO_PAGE(0)) {
-				page_cache_release(page);
-				DUMP_SEEK(file->f_pos + PAGE_SIZE);
-			}
-			else {
-				void *kaddr;
-
-				flush_cache_page(vma, addr, page_to_pfn(page));
-				kaddr = kmap(page);
-				if ((*size += PAGE_SIZE) > *limit ||
-				    !dump_write(file, kaddr, PAGE_SIZE)
-				    ) {
-					kunmap(page);
-					page_cache_release(page);
-					return -EIO;
-				}
+		for (addr = vma->vm_start; addr < vma->vm_end;
+							addr += PAGE_SIZE) {
+			struct page *page = get_dump_page(addr);
+			if (page) {
+				void *kaddr = kmap(page);
+				*size += PAGE_SIZE;
+				if (*size > *limit)
+					err = -EFBIG;
+				else if (!dump_write(file, kaddr, PAGE_SIZE))
+					err = -EIO;
 				kunmap(page);
 				page_cache_release(page);
-			}
+			} else if (!dump_seek(file, file->f_pos + PAGE_SIZE))
+				err = -EFBIG;
+			if (err)
+				goto out;
 		}
 	}
-
-	return 0;
-
-end_coredump:
-	return -EFBIG;
+out:
+	return err;
 }
 #endif
 
@@ -1802,7 +1783,8 @@ static int elf_fdpic_core_dump(long signr, struct pt_regs *regs,
 				goto end_coredump;
 	}
 
-	DUMP_SEEK(dataoff);
+	if (!dump_seek(file, dataoff))
+		goto end_coredump;
 
 	if (elf_fdpic_dump_segments(file, &size, &limit, mm_flags) < 0)
 		goto end_coredump;
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 94dfda24c06..5d1ed50bd46 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -420,7 +420,6 @@ static void bdev_destroy_inode(struct inode *inode)
 {
 	struct bdev_inode *bdi = BDEV_I(inode);
 
-	bdi->bdev.bd_inode_backing_dev_info = NULL;
 	kmem_cache_free(bdev_cachep, bdi);
 }
 
@@ -1115,7 +1114,7 @@ EXPORT_SYMBOL(revalidate_disk);
 int check_disk_change(struct block_device *bdev)
 {
 	struct gendisk *disk = bdev->bd_disk;
-	struct block_device_operations * bdops = disk->fops;
+	const struct block_device_operations *bdops = disk->fops;
 
 	if (!bdops->media_changed)
 		return 0;
@@ -1405,6 +1404,33 @@ static long block_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 }
 
 /*
+ * Write data to the block device.  Only intended for the block device itself
+ * and the raw driver which basically is a fake block device.
+ *
+ * Does not take i_mutex for the write and thus is not for general purpose
+ * use.
+ */
+ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
+			 unsigned long nr_segs, loff_t pos)
+{
+	struct file *file = iocb->ki_filp;
+	ssize_t ret;
+
+	BUG_ON(iocb->ki_pos != pos);
+
+	ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
+	if (ret > 0 || ret == -EIOCBQUEUED) {
+		ssize_t err;
+
+		err = generic_write_sync(file, pos, ret);
+		if (err < 0 && ret > 0)
+			ret = err;
+	}
+	return ret;
+}
+EXPORT_SYMBOL_GPL(blkdev_aio_write);
+
+/*
  * Try to release a page associated with block device when the system
  * is under memory pressure.
  */
@@ -1436,7 +1462,7 @@ const struct file_operations def_blk_fops = {
 	.read		= do_sync_read,
 	.write		= do_sync_write,
   	.aio_read	= generic_file_aio_read,
-  	.aio_write	= generic_file_aio_write_nolock,
+	.aio_write	= blkdev_aio_write,
 	.mmap		= generic_file_mmap,
 	.fsync		= block_fsync,
 	.unlocked_ioctl	= block_ioctl,
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index e83be2e4602..6c4173146bb 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -772,7 +772,7 @@ static void btree_invalidatepage(struct page *page, unsigned long offset)
 	}
 }
 
-static struct address_space_operations btree_aops = {
+static const struct address_space_operations btree_aops = {
 	.readpage	= btree_readpage,
 	.writepage	= btree_writepage,
 	.writepages	= btree_writepages,
@@ -1352,6 +1352,7 @@ static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
 {
 	int err;
 
+	bdi->name = "btrfs";
 	bdi->capabilities = BDI_CAP_MAP_COPY;
 	err = bdi_init(bdi);
 	if (err)
@@ -1599,6 +1600,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	sb->s_blocksize = 4096;
 	sb->s_blocksize_bits = blksize_bits(4096);
+	sb->s_bdi = &fs_info->bdi;
 
 	/*
 	 * we set the i_size on the btree inode to the max possible int.
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 72a2b9c28e9..535f85ba104 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -1511,7 +1511,8 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans,
 static void btrfs_issue_discard(struct block_device *bdev,
 				u64 start, u64 len)
 {
-	blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL);
+	blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL,
+			     DISCARD_FL_BARRIER);
 }
 #endif
 
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 59cba180fe8..9096fd0ca3c 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -55,13 +55,13 @@ struct btrfs_iget_args {
 	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 const struct inode_operations btrfs_dir_inode_operations;
+static const struct inode_operations btrfs_symlink_inode_operations;
+static const struct inode_operations btrfs_dir_ro_inode_operations;
+static const struct inode_operations btrfs_special_inode_operations;
+static const struct inode_operations btrfs_file_inode_operations;
+static const struct address_space_operations btrfs_aops;
+static const struct address_space_operations btrfs_symlink_aops;
 static struct file_operations btrfs_dir_file_operations;
 static struct extent_io_ops btrfs_extent_io_ops;
 
@@ -5201,7 +5201,7 @@ static int btrfs_permission(struct inode *inode, int mask)
 	return generic_permission(inode, mask, btrfs_check_acl);
 }
 
-static struct inode_operations btrfs_dir_inode_operations = {
+static const struct inode_operations btrfs_dir_inode_operations = {
 	.getattr	= btrfs_getattr,
 	.lookup		= btrfs_lookup,
 	.create		= btrfs_create,
@@ -5219,7 +5219,7 @@ static struct inode_operations btrfs_dir_inode_operations = {
 	.removexattr	= btrfs_removexattr,
 	.permission	= btrfs_permission,
 };
-static struct inode_operations btrfs_dir_ro_inode_operations = {
+static const struct inode_operations btrfs_dir_ro_inode_operations = {
 	.lookup		= btrfs_lookup,
 	.permission	= btrfs_permission,
 };
@@ -5259,7 +5259,7 @@ static struct extent_io_ops btrfs_extent_io_ops = {
  *
  * For now we're avoiding this by dropping bmap.
  */
-static struct address_space_operations btrfs_aops = {
+static const struct address_space_operations btrfs_aops = {
 	.readpage	= btrfs_readpage,
 	.writepage	= btrfs_writepage,
 	.writepages	= btrfs_writepages,
@@ -5271,14 +5271,14 @@ static struct address_space_operations btrfs_aops = {
 	.set_page_dirty	= btrfs_set_page_dirty,
 };
 
-static struct address_space_operations btrfs_symlink_aops = {
+static const 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 = {
+static const struct inode_operations btrfs_file_inode_operations = {
 	.truncate	= btrfs_truncate,
 	.getattr	= btrfs_getattr,
 	.setattr	= btrfs_setattr,
@@ -5290,7 +5290,7 @@ static struct inode_operations btrfs_file_inode_operations = {
 	.fallocate	= btrfs_fallocate,
 	.fiemap		= btrfs_fiemap,
 };
-static struct inode_operations btrfs_special_inode_operations = {
+static const struct inode_operations btrfs_special_inode_operations = {
 	.getattr	= btrfs_getattr,
 	.setattr	= btrfs_setattr,
 	.permission	= btrfs_permission,
@@ -5299,7 +5299,7 @@ static struct inode_operations btrfs_special_inode_operations = {
 	.listxattr	= btrfs_listxattr,
 	.removexattr	= btrfs_removexattr,
 };
-static struct inode_operations btrfs_symlink_inode_operations = {
+static const struct inode_operations btrfs_symlink_inode_operations = {
 	.readlink	= generic_readlink,
 	.follow_link	= page_follow_link_light,
 	.put_link	= page_put_link,
diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c
index d6f0806c682..7b2f401e604 100644
--- a/fs/btrfs/ordered-data.c
+++ b/fs/btrfs/ordered-data.c
@@ -740,7 +740,6 @@ int btrfs_fdatawrite_range(struct address_space *mapping, loff_t start,
 		.nr_to_write = mapping->nrpages * 2,
 		.range_start = start,
 		.range_end = end,
-		.for_writepages = 1,
 	};
 	return btrfs_writepages(mapping, &wbc);
 }
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 6d6d06cb6df..2db17cd66fc 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -51,7 +51,7 @@
 #include "export.h"
 #include "compression.h"
 
-static struct super_operations btrfs_super_ops;
+static const struct super_operations btrfs_super_ops;
 
 static void btrfs_put_super(struct super_block *sb)
 {
@@ -675,7 +675,7 @@ static int btrfs_unfreeze(struct super_block *sb)
 	return 0;
 }
 
-static struct super_operations btrfs_super_ops = {
+static const struct super_operations btrfs_super_ops = {
 	.delete_inode	= btrfs_delete_inode,
 	.put_super	= btrfs_put_super,
 	.sync_fs	= btrfs_sync_fs,
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index d91b0de7c50..30c0d45c1b5 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -2605,7 +2605,7 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 								extent);
 				cs = btrfs_file_extent_offset(src, extent);
 				cl = btrfs_file_extent_num_bytes(src,
-								extent);;
+								extent);
 				if (btrfs_file_extent_compression(src,
 								  extent)) {
 					cs = 0;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 5dbefd11b4a..5cf405b0828 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -260,7 +260,7 @@ loop_lock:
 		num_run++;
 		batch_run++;
 
-		if (bio_sync(cur))
+		if (bio_rw_flagged(cur, BIO_RW_SYNCIO))
 			num_sync_run++;
 
 		if (need_resched()) {
@@ -2903,7 +2903,7 @@ static noinline int schedule_bio(struct btrfs_root *root,
 	bio->bi_rw |= rw;
 
 	spin_lock(&device->io_lock);
-	if (bio_sync(bio))
+	if (bio_rw_flagged(bio, BIO_RW_SYNCIO))
 		pending_bios = &device->pending_sync_bios;
 	else
 		pending_bios = &device->pending_bios;
diff --git a/fs/buffer.c b/fs/buffer.c
index 28f320fac4d..209f7f15f5f 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -52,6 +52,7 @@ init_buffer(struct buffer_head *bh, bh_end_io_t *handler, void *private)
 	bh->b_end_io = handler;
 	bh->b_private = private;
 }
+EXPORT_SYMBOL(init_buffer);
 
 static int sync_buffer(void *word)
 {
@@ -80,6 +81,7 @@ void unlock_buffer(struct buffer_head *bh)
 	smp_mb__after_clear_bit();
 	wake_up_bit(&bh->b_state, BH_Lock);
 }
+EXPORT_SYMBOL(unlock_buffer);
 
 /*
  * Block until a buffer comes unlocked.  This doesn't stop it
@@ -90,6 +92,7 @@ void __wait_on_buffer(struct buffer_head * bh)
 {
 	wait_on_bit(&bh->b_state, BH_Lock, sync_buffer, TASK_UNINTERRUPTIBLE);
 }
+EXPORT_SYMBOL(__wait_on_buffer);
 
 static void
 __clear_page_buffers(struct page *page)
@@ -144,6 +147,7 @@ void end_buffer_read_sync(struct buffer_head *bh, int uptodate)
 	__end_buffer_read_notouch(bh, uptodate);
 	put_bh(bh);
 }
+EXPORT_SYMBOL(end_buffer_read_sync);
 
 void end_buffer_write_sync(struct buffer_head *bh, int uptodate)
 {
@@ -164,6 +168,7 @@ void end_buffer_write_sync(struct buffer_head *bh, int uptodate)
 	unlock_buffer(bh);
 	put_bh(bh);
 }
+EXPORT_SYMBOL(end_buffer_write_sync);
 
 /*
  * Various filesystems appear to want __find_get_block to be non-blocking.
@@ -272,6 +277,7 @@ void invalidate_bdev(struct block_device *bdev)
 	invalidate_bh_lrus();
 	invalidate_mapping_pages(mapping, 0, -1);
 }
+EXPORT_SYMBOL(invalidate_bdev);
 
 /*
  * Kick pdflush then try to free up some ZONE_NORMAL memory.
@@ -281,7 +287,7 @@ static void free_more_memory(void)
 	struct zone *zone;
 	int nid;
 
-	wakeup_pdflush(1024);
+	wakeup_flusher_threads(1024);
 	yield();
 
 	for_each_online_node(nid) {
@@ -410,6 +416,7 @@ still_busy:
 	local_irq_restore(flags);
 	return;
 }
+EXPORT_SYMBOL(end_buffer_async_write);
 
 /*
  * If a page's buffers are under async readin (end_buffer_async_read
@@ -438,8 +445,8 @@ static void mark_buffer_async_read(struct buffer_head *bh)
 	set_buffer_async_read(bh);
 }
 
-void mark_buffer_async_write_endio(struct buffer_head *bh,
-				   bh_end_io_t *handler)
+static void mark_buffer_async_write_endio(struct buffer_head *bh,
+					  bh_end_io_t *handler)
 {
 	bh->b_end_io = handler;
 	set_buffer_async_write(bh);
@@ -553,7 +560,7 @@ repeat:
 	return err;
 }
 
-void do_thaw_all(struct work_struct *work)
+static void do_thaw_all(struct work_struct *work)
 {
 	struct super_block *sb;
 	char b[BDEVNAME_SIZE];
@@ -1172,6 +1179,7 @@ void mark_buffer_dirty(struct buffer_head *bh)
 		}
 	}
 }
+EXPORT_SYMBOL(mark_buffer_dirty);
 
 /*
  * Decrement a buffer_head's reference count.  If all buffers against a page
@@ -1188,6 +1196,7 @@ void __brelse(struct buffer_head * buf)
 	}
 	WARN(1, KERN_ERR "VFS: brelse: Trying to free free buffer\n");
 }
+EXPORT_SYMBOL(__brelse);
 
 /*
  * bforget() is like brelse(), except it discards any
@@ -1206,6 +1215,7 @@ void __bforget(struct buffer_head *bh)
 	}
 	__brelse(bh);
 }
+EXPORT_SYMBOL(__bforget);
 
 static struct buffer_head *__bread_slow(struct buffer_head *bh)
 {
@@ -2218,6 +2228,7 @@ int block_read_full_page(struct page *page, get_block_t *get_block)
 	}
 	return 0;
 }
+EXPORT_SYMBOL(block_read_full_page);
 
 /* utility function for filesystems that need to do work on expanding
  * truncates.  Uses filesystem pagecache writes to allow the filesystem to
@@ -2252,6 +2263,7 @@ int generic_cont_expand_simple(struct inode *inode, loff_t size)
 out:
 	return err;
 }
+EXPORT_SYMBOL(generic_cont_expand_simple);
 
 static int cont_expand_zero(struct file *file, struct address_space *mapping,
 			    loff_t pos, loff_t *bytes)
@@ -2352,6 +2364,7 @@ int cont_write_begin(struct file *file, struct address_space *mapping,
 out:
 	return err;
 }
+EXPORT_SYMBOL(cont_write_begin);
 
 int block_prepare_write(struct page *page, unsigned from, unsigned to,
 			get_block_t *get_block)
@@ -2362,6 +2375,7 @@ int block_prepare_write(struct page *page, unsigned from, unsigned to,
 		ClearPageUptodate(page);
 	return err;
 }
+EXPORT_SYMBOL(block_prepare_write);
 
 int block_commit_write(struct page *page, unsigned from, unsigned to)
 {
@@ -2369,6 +2383,7 @@ int block_commit_write(struct page *page, unsigned from, unsigned to)
 	__block_commit_write(inode,page,from,to);
 	return 0;
 }
+EXPORT_SYMBOL(block_commit_write);
 
 /*
  * block_page_mkwrite() is not allowed to change the file size as it gets
@@ -2426,6 +2441,7 @@ block_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf,
 out:
 	return ret;
 }
+EXPORT_SYMBOL(block_page_mkwrite);
 
 /*
  * nobh_write_begin()'s prereads are special: the buffer_heads are freed
@@ -2849,6 +2865,7 @@ unlock:
 out:
 	return err;
 }
+EXPORT_SYMBOL(block_truncate_page);
 
 /*
  * The generic ->writepage function for buffer-backed address_spaces
@@ -2890,6 +2907,7 @@ int block_write_full_page_endio(struct page *page, get_block_t *get_block,
 	zero_user_segment(page, offset, PAGE_CACHE_SIZE);
 	return __block_write_full_page(inode, page, get_block, wbc, handler);
 }
+EXPORT_SYMBOL(block_write_full_page_endio);
 
 /*
  * The generic ->writepage function for buffer-backed address_spaces
@@ -2900,7 +2918,7 @@ int block_write_full_page(struct page *page, get_block_t *get_block,
 	return block_write_full_page_endio(page, get_block, wbc,
 					   end_buffer_async_write);
 }
-
+EXPORT_SYMBOL(block_write_full_page);
 
 sector_t generic_block_bmap(struct address_space *mapping, sector_t block,
 			    get_block_t *get_block)
@@ -2913,6 +2931,7 @@ sector_t generic_block_bmap(struct address_space *mapping, sector_t block,
 	get_block(inode, block, &tmp, 0);
 	return tmp.b_blocknr;
 }
+EXPORT_SYMBOL(generic_block_bmap);
 
 static void end_bio_bh_io_sync(struct bio *bio, int err)
 {
@@ -2982,6 +3001,7 @@ int submit_bh(int rw, struct buffer_head * bh)
 	bio_put(bio);
 	return ret;
 }
+EXPORT_SYMBOL(submit_bh);
 
 /**
  * ll_rw_block: low-level access to block devices (DEPRECATED)
@@ -3043,6 +3063,7 @@ void ll_rw_block(int rw, int nr, struct buffer_head *bhs[])
 		unlock_buffer(bh);
 	}
 }
+EXPORT_SYMBOL(ll_rw_block);
 
 /*
  * For a data-integrity writeout, we need to wait upon any in-progress I/O
@@ -3071,6 +3092,7 @@ int sync_dirty_buffer(struct buffer_head *bh)
 	}
 	return ret;
 }
+EXPORT_SYMBOL(sync_dirty_buffer);
 
 /*
  * try_to_free_buffers() checks if all the buffers on this particular page
@@ -3185,6 +3207,7 @@ void block_sync_page(struct page *page)
 	if (mapping)
 		blk_run_backing_dev(mapping->backing_dev_info, page);
 }
+EXPORT_SYMBOL(block_sync_page);
 
 /*
  * There are no bdflush tunables left.  But distributions are
@@ -3361,29 +3384,3 @@ void __init buffer_init(void)
 	max_buffer_heads = nrpages * (PAGE_SIZE / sizeof(struct buffer_head));
 	hotcpu_notifier(buffer_cpu_notify, 0);
 }
-
-EXPORT_SYMBOL(__bforget);
-EXPORT_SYMBOL(__brelse);
-EXPORT_SYMBOL(__wait_on_buffer);
-EXPORT_SYMBOL(block_commit_write);
-EXPORT_SYMBOL(block_prepare_write);
-EXPORT_SYMBOL(block_page_mkwrite);
-EXPORT_SYMBOL(block_read_full_page);
-EXPORT_SYMBOL(block_sync_page);
-EXPORT_SYMBOL(block_truncate_page);
-EXPORT_SYMBOL(block_write_full_page);
-EXPORT_SYMBOL(block_write_full_page_endio);
-EXPORT_SYMBOL(cont_write_begin);
-EXPORT_SYMBOL(end_buffer_read_sync);
-EXPORT_SYMBOL(end_buffer_write_sync);
-EXPORT_SYMBOL(end_buffer_async_write);
-EXPORT_SYMBOL(file_fsync);
-EXPORT_SYMBOL(generic_block_bmap);
-EXPORT_SYMBOL(generic_cont_expand_simple);
-EXPORT_SYMBOL(init_buffer);
-EXPORT_SYMBOL(invalidate_bdev);
-EXPORT_SYMBOL(ll_rw_block);
-EXPORT_SYMBOL(mark_buffer_dirty);
-EXPORT_SYMBOL(submit_bh);
-EXPORT_SYMBOL(sync_dirty_buffer);
-EXPORT_SYMBOL(unlock_buffer);
diff --git a/fs/char_dev.c b/fs/char_dev.c
index a173551e19d..3cbc57f932d 100644
--- a/fs/char_dev.c
+++ b/fs/char_dev.c
@@ -31,6 +31,7 @@
  * - no readahead or I/O queue unplugging required
  */
 struct backing_dev_info directly_mappable_cdev_bdi = {
+	.name = "char",
 	.capabilities	= (
 #ifdef CONFIG_MMU
 		/* permit private copies of the data to be taken */
@@ -237,8 +238,10 @@ int alloc_chrdev_region(dev_t *dev, unsigned baseminor, unsigned count,
 }
 
 /**
- * register_chrdev() - Register a major number for character devices.
+ * __register_chrdev() - create and register a cdev occupying a range of minors
  * @major: major device number or 0 for dynamic allocation
+ * @baseminor: first of the requested range of minor numbers
+ * @count: the number of minor numbers required
  * @name: name of this range of devices
  * @fops: file operations associated with this devices
  *
@@ -254,19 +257,17 @@ int alloc_chrdev_region(dev_t *dev, unsigned baseminor, unsigned count,
  * /dev. It only helps to keep track of the different owners of devices. If
  * your module name has only one type of devices it's ok to use e.g. the name
  * of the module here.
- *
- * This function registers a range of 256 minor numbers. The first minor number
- * is 0.
  */
-int register_chrdev(unsigned int major, const char *name,
-		    const struct file_operations *fops)
+int __register_chrdev(unsigned int major, unsigned int baseminor,
+		      unsigned int count, const char *name,
+		      const struct file_operations *fops)
 {
 	struct char_device_struct *cd;
 	struct cdev *cdev;
 	char *s;
 	int err = -ENOMEM;
 
-	cd = __register_chrdev_region(major, 0, 256, name);
+	cd = __register_chrdev_region(major, baseminor, count, name);
 	if (IS_ERR(cd))
 		return PTR_ERR(cd);
 	
@@ -280,7 +281,7 @@ int register_chrdev(unsigned int major, const char *name,
 	for (s = strchr(kobject_name(&cdev->kobj),'/'); s; s = strchr(s, '/'))
 		*s = '!';
 		
-	err = cdev_add(cdev, MKDEV(cd->major, 0), 256);
+	err = cdev_add(cdev, MKDEV(cd->major, baseminor), count);
 	if (err)
 		goto out;
 
@@ -290,7 +291,7 @@ int register_chrdev(unsigned int major, const char *name,
 out:
 	kobject_put(&cdev->kobj);
 out2:
-	kfree(__unregister_chrdev_region(cd->major, 0, 256));
+	kfree(__unregister_chrdev_region(cd->major, baseminor, count));
 	return err;
 }
 
@@ -316,10 +317,23 @@ void unregister_chrdev_region(dev_t from, unsigned count)
 	}
 }
 
-void unregister_chrdev(unsigned int major, const char *name)
+/**
+ * __unregister_chrdev - unregister and destroy a cdev
+ * @major: major device number
+ * @baseminor: first of the range of minor numbers
+ * @count: the number of minor numbers this cdev is occupying
+ * @name: name of this range of devices
+ *
+ * Unregister and destroy the cdev occupying the region described by
+ * @major, @baseminor and @count.  This function undoes what
+ * __register_chrdev() did.
+ */
+void __unregister_chrdev(unsigned int major, unsigned int baseminor,
+			 unsigned int count, const char *name)
 {
 	struct char_device_struct *cd;
-	cd = __unregister_chrdev_region(major, 0, 256);
+
+	cd = __unregister_chrdev_region(major, baseminor, count);
 	if (cd && cd->cdev)
 		cdev_del(cd->cdev);
 	kfree(cd);
@@ -568,6 +582,6 @@ EXPORT_SYMBOL(cdev_alloc);
 EXPORT_SYMBOL(cdev_del);
 EXPORT_SYMBOL(cdev_add);
 EXPORT_SYMBOL(cdev_index);
-EXPORT_SYMBOL(register_chrdev);
-EXPORT_SYMBOL(unregister_chrdev);
+EXPORT_SYMBOL(__register_chrdev);
+EXPORT_SYMBOL(__unregister_chrdev);
 EXPORT_SYMBOL(directly_mappable_cdev_bdi);
diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES
index e85b1e4389e..145540a316a 100644
--- a/fs/cifs/CHANGES
+++ b/fs/cifs/CHANGES
@@ -3,7 +3,10 @@ Version 1.60
 Fix memory leak in reconnect.  Fix oops in DFS mount error path.
 Set s_maxbytes to smaller (the max that vfs can handle) so that
 sendfile will now work over cifs mounts again.  Add noforcegid
-and noforceuid mount parameters.
+and noforceuid mount parameters. Fix small mem leak when using
+ntlmv2. Fix 2nd mount to same server but with different port to
+be allowed (rather than reusing the 1st port) - only when the
+user explicitly overrides the port on the 2nd mount.
 
 Version 1.59
 ------------
diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c
index 606912d8f2a..fea9e898c4b 100644
--- a/fs/cifs/cifs_dfs_ref.c
+++ b/fs/cifs/cifs_dfs_ref.c
@@ -142,7 +142,7 @@ char *cifs_compose_mount_options(const char *sb_mountdata,
 	rc = dns_resolve_server_name_to_ip(*devname, &srvIP);
 	if (rc != 0) {
 		cERROR(1, ("%s: Failed to resolve server part of %s to IP: %d",
-			  __func__, *devname, rc));;
+			  __func__, *devname, rc));
 		goto compose_mount_options_err;
 	}
 	/* md_len = strlen(...) + 12 for 'sep+prefixpath='
@@ -385,7 +385,7 @@ out_err:
 	goto out;
 }
 
-struct inode_operations cifs_dfs_referral_inode_operations = {
+const struct inode_operations cifs_dfs_referral_inode_operations = {
 	.follow_link = cifs_dfs_follow_mountpoint,
 };
 
diff --git a/fs/cifs/cifs_spnego.c b/fs/cifs/cifs_spnego.c
index 051caecf7d6..8ec7736ce95 100644
--- a/fs/cifs/cifs_spnego.c
+++ b/fs/cifs/cifs_spnego.c
@@ -125,7 +125,7 @@ cifs_get_spnego_key(struct cifsSesInfo *sesInfo)
 	if (server->addr.sockAddr.sin_family == AF_INET)
 		sprintf(dp, "ip4=%pI4", &server->addr.sockAddr.sin_addr);
 	else if (server->addr.sockAddr.sin_family == AF_INET6)
-		sprintf(dp, "ip6=%pi6", &server->addr.sockAddr6.sin6_addr);
+		sprintf(dp, "ip6=%pI6", &server->addr.sockAddr6.sin6_addr);
 	else
 		goto out;
 
diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c
index 6941c22398a..7dfe0842a6f 100644
--- a/fs/cifs/cifsacl.c
+++ b/fs/cifs/cifsacl.c
@@ -607,7 +607,7 @@ static struct cifs_ntsd *get_cifs_acl(struct cifs_sb_info *cifs_sb,
 		return get_cifs_acl_by_path(cifs_sb, path, pacllen);
 
 	pntsd = get_cifs_acl_by_fid(cifs_sb, open_file->netfid, pacllen);
-	atomic_dec(&open_file->wrtPending);
+	cifsFileInfo_put(open_file);
 	return pntsd;
 }
 
@@ -665,7 +665,7 @@ static int set_cifs_acl(struct cifs_ntsd *pnntsd, __u32 acllen,
 		return set_cifs_acl_by_path(cifs_sb, path, pnntsd, acllen);
 
 	rc = set_cifs_acl_by_fid(cifs_sb, open_file->netfid, pnntsd, acllen);
-	atomic_dec(&open_file->wrtPending);
+	cifsFileInfo_put(open_file);
 	return rc;
 }
 
diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c
index 7c9809523f4..7efe1745494 100644
--- a/fs/cifs/cifsencrypt.c
+++ b/fs/cifs/cifsencrypt.c
@@ -373,6 +373,7 @@ calc_exit_2:
 	   compare with the NTLM example */
 	hmac_md5_final(ses->server->ntlmv2_hash, pctxt);
 
+	kfree(pctxt);
 	return rc;
 }
 
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 84b75253b05..d79ce2e95c2 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -50,7 +50,7 @@
 #define CIFS_MAGIC_NUMBER 0xFF534D42	/* the first four bytes of SMB PDUs */
 
 #ifdef CONFIG_CIFS_QUOTA
-static struct quotactl_ops cifs_quotactl_ops;
+static const struct quotactl_ops cifs_quotactl_ops;
 #endif /* QUOTA */
 
 int cifsFYI = 0;
@@ -361,13 +361,10 @@ cifs_show_address(struct seq_file *s, struct TCP_Server_Info *server)
 static int
 cifs_show_options(struct seq_file *s, struct vfsmount *m)
 {
-	struct cifs_sb_info *cifs_sb;
-	struct cifsTconInfo *tcon;
-
-	cifs_sb = CIFS_SB(m->mnt_sb);
-	tcon = cifs_sb->tcon;
+	struct cifs_sb_info *cifs_sb = CIFS_SB(m->mnt_sb);
+	struct cifsTconInfo *tcon = cifs_sb->tcon;
 
-	seq_printf(s, ",unc=%s", cifs_sb->tcon->treeName);
+	seq_printf(s, ",unc=%s", tcon->treeName);
 	if (tcon->ses->userName)
 		seq_printf(s, ",username=%s", tcon->ses->userName);
 	if (tcon->ses->domainName)
@@ -520,7 +517,7 @@ int cifs_xstate_get(struct super_block *sb, struct fs_quota_stat *qstats)
 	return rc;
 }
 
-static struct quotactl_ops cifs_quotactl_ops = {
+static const struct quotactl_ops cifs_quotactl_ops = {
 	.set_xquota	= cifs_xquota_set,
 	.get_xquota	= cifs_xquota_get,
 	.set_xstate	= cifs_xstate_set,
@@ -989,19 +986,19 @@ static int cifs_oplock_thread(void *dummyarg)
 		if (try_to_freeze())
 			continue;
 
-		spin_lock(&GlobalMid_Lock);
-		if (list_empty(&GlobalOplock_Q)) {
-			spin_unlock(&GlobalMid_Lock);
+		spin_lock(&cifs_oplock_lock);
+		if (list_empty(&cifs_oplock_list)) {
+			spin_unlock(&cifs_oplock_lock);
 			set_current_state(TASK_INTERRUPTIBLE);
 			schedule_timeout(39*HZ);
 		} else {
-			oplock_item = list_entry(GlobalOplock_Q.next,
+			oplock_item = list_entry(cifs_oplock_list.next,
 						struct oplock_q_entry, qhead);
 			cFYI(1, ("found oplock item to write out"));
 			pTcon = oplock_item->tcon;
 			inode = oplock_item->pinode;
 			netfid = oplock_item->netfid;
-			spin_unlock(&GlobalMid_Lock);
+			spin_unlock(&cifs_oplock_lock);
 			DeleteOplockQEntry(oplock_item);
 			/* can not grab inode sem here since it would
 				deadlock when oplock received on delete
@@ -1058,7 +1055,7 @@ init_cifs(void)
 	int rc = 0;
 	cifs_proc_init();
 	INIT_LIST_HEAD(&cifs_tcp_ses_list);
-	INIT_LIST_HEAD(&GlobalOplock_Q);
+	INIT_LIST_HEAD(&cifs_oplock_list);
 #ifdef CONFIG_CIFS_EXPERIMENTAL
 	INIT_LIST_HEAD(&GlobalDnotifyReqList);
 	INIT_LIST_HEAD(&GlobalDnotifyRsp_Q);
@@ -1087,6 +1084,7 @@ init_cifs(void)
 	rwlock_init(&GlobalSMBSeslock);
 	rwlock_init(&cifs_tcp_ses_lock);
 	spin_lock_init(&GlobalMid_Lock);
+	spin_lock_init(&cifs_oplock_lock);
 
 	if (cifs_max_pending < 2) {
 		cifs_max_pending = 2;
diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index 6c170948300..ac2b24c192f 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -67,7 +67,7 @@ extern int cifs_setattr(struct dentry *, struct iattr *);
 
 extern const struct inode_operations cifs_file_inode_ops;
 extern const struct inode_operations cifs_symlink_inode_ops;
-extern struct inode_operations cifs_dfs_referral_inode_operations;
+extern const struct inode_operations cifs_dfs_referral_inode_operations;
 
 
 /* Functions related to files and directories */
@@ -113,5 +113,5 @@ extern long cifs_ioctl(struct file *filep, unsigned int cmd, unsigned long arg);
 extern const struct export_operations cifs_export_ops;
 #endif /* EXPERIMENTAL */
 
-#define CIFS_VERSION   "1.60"
+#define CIFS_VERSION   "1.61"
 #endif				/* _CIFSFS_H */
diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h
index 6084d6379c0..6cfc81a3270 100644
--- a/fs/cifs/cifsglob.h
+++ b/fs/cifs/cifsglob.h
@@ -351,11 +351,24 @@ struct cifsFileInfo {
 	bool closePend:1;	/* file is marked to close */
 	bool invalidHandle:1;	/* file closed via session abend */
 	bool messageMode:1;	/* for pipes: message vs byte mode */
-	atomic_t wrtPending;   /* handle in use - defer close */
+	atomic_t count;		/* reference count */
 	struct mutex fh_mutex; /* prevents reopen race after dead ses*/
 	struct cifs_search_info srch_inf;
 };
 
+/* Take a reference on the file private data */
+static inline void cifsFileInfo_get(struct cifsFileInfo *cifs_file)
+{
+	atomic_inc(&cifs_file->count);
+}
+
+/* Release a reference on the file private data */
+static inline void cifsFileInfo_put(struct cifsFileInfo *cifs_file)
+{
+	if (atomic_dec_and_test(&cifs_file->count))
+		kfree(cifs_file);
+}
+
 /*
  * One of these for each file inode
  */
@@ -656,7 +669,11 @@ GLOBAL_EXTERN rwlock_t		cifs_tcp_ses_lock;
  */
 GLOBAL_EXTERN rwlock_t GlobalSMBSeslock;
 
-GLOBAL_EXTERN struct list_head GlobalOplock_Q;
+/* Global list of oplocks */
+GLOBAL_EXTERN struct list_head cifs_oplock_list;
+
+/* Protects the cifs_oplock_list */
+GLOBAL_EXTERN spinlock_t cifs_oplock_lock;
 
 /* Outstanding dir notify requests */
 GLOBAL_EXTERN struct list_head GlobalDnotifyReqList;
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index 1866bc2927d..301e307e127 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -100,110 +100,138 @@ static void mark_open_files_invalid(struct cifsTconInfo *pTcon)
 	   to this tcon */
 }
 
-/* Allocate and return pointer to an SMB request buffer, and set basic
-   SMB information in the SMB header.  If the return code is zero, this
-   function must have filled in request_buf pointer */
+/* reconnect the socket, tcon, and smb session if needed */
 static int
-small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
-		void **request_buf)
+cifs_reconnect_tcon(struct cifsTconInfo *tcon, int smb_command)
 {
 	int rc = 0;
+	struct cifsSesInfo *ses;
+	struct TCP_Server_Info *server;
+	struct nls_table *nls_codepage;
 
-	/* SMBs NegProt, SessSetup, uLogoff do not have tcon yet so
-	   check for tcp and smb session status done differently
-	   for those three - in the calling routine */
-	if (tcon) {
-		if (tcon->tidStatus == CifsExiting) {
-			/* only tree disconnect, open, and write,
-			(and ulogoff which does not have tcon)
-			are allowed as we start force umount */
-			if ((smb_command != SMB_COM_WRITE_ANDX) &&
-			   (smb_command != SMB_COM_OPEN_ANDX) &&
-			   (smb_command != SMB_COM_TREE_DISCONNECT)) {
-				cFYI(1, ("can not send cmd %d while umounting",
-					smb_command));
-				return -ENODEV;
-			}
+	/*
+	 * SMBs NegProt, SessSetup, uLogoff do not have tcon yet so check for
+	 * tcp and smb session status done differently for those three - in the
+	 * calling routine
+	 */
+	if (!tcon)
+		return 0;
+
+	ses = tcon->ses;
+	server = ses->server;
+
+	/*
+	 * only tree disconnect, open, and write, (and ulogoff which does not
+	 * have tcon) are allowed as we start force umount
+	 */
+	if (tcon->tidStatus == CifsExiting) {
+		if (smb_command != SMB_COM_WRITE_ANDX &&
+		    smb_command != SMB_COM_OPEN_ANDX &&
+		    smb_command != SMB_COM_TREE_DISCONNECT) {
+			cFYI(1, ("can not send cmd %d while umounting",
+				smb_command));
+			return -ENODEV;
 		}
-		if ((tcon->ses) && (tcon->ses->status != CifsExiting) &&
-				  (tcon->ses->server)) {
-			struct nls_table *nls_codepage;
-				/* Give Demultiplex thread up to 10 seconds to
-				   reconnect, should be greater than cifs socket
-				   timeout which is 7 seconds */
-			while (tcon->ses->server->tcpStatus ==
-							 CifsNeedReconnect) {
-				wait_event_interruptible_timeout(tcon->ses->server->response_q,
-					(tcon->ses->server->tcpStatus ==
-							CifsGood), 10 * HZ);
-				if (tcon->ses->server->tcpStatus ==
-							CifsNeedReconnect) {
-					/* on "soft" mounts we wait once */
-					if (!tcon->retry ||
-					   (tcon->ses->status == CifsExiting)) {
-						cFYI(1, ("gave up waiting on "
-						      "reconnect in smb_init"));
-						return -EHOSTDOWN;
-					} /* else "hard" mount - keep retrying
-					     until process is killed or server
-					     comes back on-line */
-				} else /* TCP session is reestablished now */
-					break;
-			}
+	}
 
-			nls_codepage = load_nls_default();
-		/* need to prevent multiple threads trying to
-		simultaneously reconnect the same SMB session */
-			down(&tcon->ses->sesSem);
-			if (tcon->ses->need_reconnect)
-				rc = cifs_setup_session(0, tcon->ses,
-							nls_codepage);
-			if (!rc && (tcon->need_reconnect)) {
-				mark_open_files_invalid(tcon);
-				rc = CIFSTCon(0, tcon->ses, tcon->treeName,
-					      tcon, nls_codepage);
-				up(&tcon->ses->sesSem);
-				/* BB FIXME add code to check if wsize needs
-				   update due to negotiated smb buffer size
-				   shrinking */
-				if (rc == 0) {
-					atomic_inc(&tconInfoReconnectCount);
-					/* tell server Unix caps we support */
-					if (tcon->ses->capabilities & CAP_UNIX)
-						reset_cifs_unix_caps(
-						0 /* no xid */,
-						tcon,
-						NULL /* we do not know sb */,
-						NULL /* no vol info */);
-				}
+	if (ses->status == CifsExiting)
+		return -EIO;
 
-				cFYI(1, ("reconnect tcon rc = %d", rc));
-				/* Removed call to reopen open files here.
-				   It is safer (and faster) to reopen files
-				   one at a time as needed in read and write */
-
-				/* Check if handle based operation so we
-				   know whether we can continue or not without
-				   returning to caller to reset file handle */
-				switch (smb_command) {
-					case SMB_COM_READ_ANDX:
-					case SMB_COM_WRITE_ANDX:
-					case SMB_COM_CLOSE:
-					case SMB_COM_FIND_CLOSE2:
-					case SMB_COM_LOCKING_ANDX: {
-						unload_nls(nls_codepage);
-						return -EAGAIN;
-					}
-				}
-			} else {
-				up(&tcon->ses->sesSem);
-			}
-			unload_nls(nls_codepage);
+	/*
+	 * Give demultiplex thread up to 10 seconds to reconnect, should be
+	 * greater than cifs socket timeout which is 7 seconds
+	 */
+	while (server->tcpStatus == CifsNeedReconnect) {
+		wait_event_interruptible_timeout(server->response_q,
+			(server->tcpStatus == CifsGood), 10 * HZ);
 
-		} else {
-			return -EIO;
+		/* is TCP session is reestablished now ?*/
+		if (server->tcpStatus != CifsNeedReconnect)
+			break;
+
+		/*
+		 * on "soft" mounts we wait once. Hard mounts keep
+		 * retrying until process is killed or server comes
+		 * back on-line
+		 */
+		if (!tcon->retry || ses->status == CifsExiting) {
+			cFYI(1, ("gave up waiting on reconnect in smb_init"));
+			return -EHOSTDOWN;
 		}
 	}
+
+	if (!ses->need_reconnect && !tcon->need_reconnect)
+		return 0;
+
+	nls_codepage = load_nls_default();
+
+	/*
+	 * need to prevent multiple threads trying to simultaneously
+	 * reconnect the same SMB session
+	 */
+	down(&ses->sesSem);
+	if (ses->need_reconnect)
+		rc = cifs_setup_session(0, ses, nls_codepage);
+
+	/* do we need to reconnect tcon? */
+	if (rc || !tcon->need_reconnect) {
+		up(&ses->sesSem);
+		goto out;
+	}
+
+	mark_open_files_invalid(tcon);
+	rc = CIFSTCon(0, ses, tcon->treeName, tcon, nls_codepage);
+	up(&ses->sesSem);
+	cFYI(1, ("reconnect tcon rc = %d", rc));
+
+	if (rc)
+		goto out;
+
+	/*
+	 * FIXME: check if wsize needs updated due to negotiated smb buffer
+	 * 	  size shrinking
+	 */
+	atomic_inc(&tconInfoReconnectCount);
+
+	/* tell server Unix caps we support */
+	if (ses->capabilities & CAP_UNIX)
+		reset_cifs_unix_caps(0, tcon, NULL, NULL);
+
+	/*
+	 * Removed call to reopen open files here. It is safer (and faster) to
+	 * reopen files one at a time as needed in read and write.
+	 *
+	 * FIXME: what about file locks? don't we need to reclaim them ASAP?
+	 */
+
+out:
+	/*
+	 * Check if handle based operation so we know whether we can continue
+	 * or not without returning to caller to reset file handle
+	 */
+	switch (smb_command) {
+	case SMB_COM_READ_ANDX:
+	case SMB_COM_WRITE_ANDX:
+	case SMB_COM_CLOSE:
+	case SMB_COM_FIND_CLOSE2:
+	case SMB_COM_LOCKING_ANDX:
+		rc = -EAGAIN;
+	}
+
+	unload_nls(nls_codepage);
+	return rc;
+}
+
+/* Allocate and return pointer to an SMB request buffer, and set basic
+   SMB information in the SMB header.  If the return code is zero, this
+   function must have filled in request_buf pointer */
+static int
+small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
+		void **request_buf)
+{
+	int rc = 0;
+
+	rc = cifs_reconnect_tcon(tcon, smb_command);
 	if (rc)
 		return rc;
 
@@ -256,101 +284,7 @@ smb_init(int smb_command, int wct, struct cifsTconInfo *tcon,
 {
 	int rc = 0;
 
-	/* SMBs NegProt, SessSetup, uLogoff do not have tcon yet so
-	   check for tcp and smb session status done differently
-	   for those three - in the calling routine */
-	if (tcon) {
-		if (tcon->tidStatus == CifsExiting) {
-			/* only tree disconnect, open, and write,
-			  (and ulogoff which does not have tcon)
-			  are allowed as we start force umount */
-			if ((smb_command != SMB_COM_WRITE_ANDX) &&
-			   (smb_command != SMB_COM_OPEN_ANDX) &&
-			   (smb_command != SMB_COM_TREE_DISCONNECT)) {
-				cFYI(1, ("can not send cmd %d while umounting",
-					smb_command));
-				return -ENODEV;
-			}
-		}
-
-		if ((tcon->ses) && (tcon->ses->status != CifsExiting) &&
-				  (tcon->ses->server)) {
-			struct nls_table *nls_codepage;
-				/* Give Demultiplex thread up to 10 seconds to
-				   reconnect, should be greater than cifs socket
-				   timeout which is 7 seconds */
-			while (tcon->ses->server->tcpStatus ==
-							CifsNeedReconnect) {
-				wait_event_interruptible_timeout(tcon->ses->server->response_q,
-					(tcon->ses->server->tcpStatus ==
-							CifsGood), 10 * HZ);
-				if (tcon->ses->server->tcpStatus ==
-						CifsNeedReconnect) {
-					/* on "soft" mounts we wait once */
-					if (!tcon->retry ||
-					   (tcon->ses->status == CifsExiting)) {
-						cFYI(1, ("gave up waiting on "
-						      "reconnect in smb_init"));
-						return -EHOSTDOWN;
-					} /* else "hard" mount - keep retrying
-					     until process is killed or server
-					     comes on-line */
-				} else /* TCP session is reestablished now */
-					break;
-			}
-			nls_codepage = load_nls_default();
-		/* need to prevent multiple threads trying to
-		simultaneously reconnect the same SMB session */
-			down(&tcon->ses->sesSem);
-			if (tcon->ses->need_reconnect)
-				rc = cifs_setup_session(0, tcon->ses,
-							nls_codepage);
-			if (!rc && (tcon->need_reconnect)) {
-				mark_open_files_invalid(tcon);
-				rc = CIFSTCon(0, tcon->ses, tcon->treeName,
-					      tcon, nls_codepage);
-				up(&tcon->ses->sesSem);
-				/* BB FIXME add code to check if wsize needs
-				update due to negotiated smb buffer size
-				shrinking */
-				if (rc == 0) {
-					atomic_inc(&tconInfoReconnectCount);
-					/* tell server Unix caps we support */
-					if (tcon->ses->capabilities & CAP_UNIX)
-						reset_cifs_unix_caps(
-						0 /* no xid */,
-						tcon,
-						NULL /* do not know sb */,
-						NULL /* no vol info */);
-				}
-
-				cFYI(1, ("reconnect tcon rc = %d", rc));
-				/* Removed call to reopen open files here.
-				   It is safer (and faster) to reopen files
-				   one at a time as needed in read and write */
-
-				/* Check if handle based operation so we
-				   know whether we can continue or not without
-				   returning to caller to reset file handle */
-				switch (smb_command) {
-					case SMB_COM_READ_ANDX:
-					case SMB_COM_WRITE_ANDX:
-					case SMB_COM_CLOSE:
-					case SMB_COM_FIND_CLOSE2:
-					case SMB_COM_LOCKING_ANDX: {
-						unload_nls(nls_codepage);
-						return -EAGAIN;
-					}
-				}
-			} else {
-				up(&tcon->ses->sesSem);
-			}
-			unload_nls(nls_codepage);
-
-		} else {
-			return -EIO;
-		}
-	}
+	rc = cifs_reconnect_tcon(tcon, smb_command);
 	if (rc)
 		return rc;
 
@@ -3961,6 +3895,10 @@ parse_DFS_referrals(TRANSACTION2_GET_DFS_REFER_RSP *pSMBr,
 		if (is_unicode) {
 			__le16 *tmp = kmalloc(strlen(searchName)*2 + 2,
 						GFP_KERNEL);
+			if (tmp == NULL) {
+				rc = -ENOMEM;
+				goto parse_DFS_referrals_exit;
+			}
 			cifsConvertToUCS((__le16 *) tmp, searchName,
 					PATH_MAX, nls_codepage, remap);
 			node->path_consumed = cifs_ucs2_bytes(tmp,
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
index 1f3345d7fa7..d49682433c2 100644
--- a/fs/cifs/connect.c
+++ b/fs/cifs/connect.c
@@ -1377,7 +1377,7 @@ cifs_parse_mount_options(char *options, const char *devname,
 }
 
 static struct TCP_Server_Info *
-cifs_find_tcp_session(struct sockaddr_storage *addr)
+cifs_find_tcp_session(struct sockaddr_storage *addr, unsigned short int port)
 {
 	struct list_head *tmp;
 	struct TCP_Server_Info *server;
@@ -1397,16 +1397,37 @@ cifs_find_tcp_session(struct sockaddr_storage *addr)
 		if (server->tcpStatus == CifsNew)
 			continue;
 
-		if (addr->ss_family == AF_INET &&
-		    (addr4->sin_addr.s_addr !=
-		     server->addr.sockAddr.sin_addr.s_addr))
-			continue;
-		else if (addr->ss_family == AF_INET6 &&
-			 (!ipv6_addr_equal(&server->addr.sockAddr6.sin6_addr,
-					   &addr6->sin6_addr) ||
-			  server->addr.sockAddr6.sin6_scope_id !=
-					   addr6->sin6_scope_id))
-			continue;
+		switch (addr->ss_family) {
+		case AF_INET:
+			if (addr4->sin_addr.s_addr ==
+			    server->addr.sockAddr.sin_addr.s_addr) {
+				addr4->sin_port = htons(port);
+				/* user overrode default port? */
+				if (addr4->sin_port) {
+					if (addr4->sin_port !=
+					    server->addr.sockAddr.sin_port)
+						continue;
+				}
+				break;
+			} else
+				continue;
+
+		case AF_INET6:
+			if (ipv6_addr_equal(&addr6->sin6_addr,
+			    &server->addr.sockAddr6.sin6_addr) &&
+			    (addr6->sin6_scope_id ==
+			    server->addr.sockAddr6.sin6_scope_id)) {
+				addr6->sin6_port = htons(port);
+				/* user overrode default port? */
+				if (addr6->sin6_port) {
+					if (addr6->sin6_port !=
+					   server->addr.sockAddr6.sin6_port)
+						continue;
+				}
+				break;
+			} else
+				continue;
+		}
 
 		++server->srv_count;
 		write_unlock(&cifs_tcp_ses_lock);
@@ -1475,7 +1496,7 @@ cifs_get_tcp_session(struct smb_vol *volume_info)
 	}
 
 	/* see if we already have a matching tcp_ses */
-	tcp_ses = cifs_find_tcp_session(&addr);
+	tcp_ses = cifs_find_tcp_session(&addr, volume_info->port);
 	if (tcp_ses)
 		return tcp_ses;
 
@@ -2636,9 +2657,9 @@ CIFSTCon(unsigned int xid, struct cifsSesInfo *ses,
 		return -EIO;
 
 	smb_buffer = cifs_buf_get();
-	if (smb_buffer == NULL) {
+	if (smb_buffer == NULL)
 		return -ENOMEM;
-	}
+
 	smb_buffer_response = smb_buffer;
 
 	header_assemble(smb_buffer, SMB_COM_TREE_CONNECT_ANDX,
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 4326ffd90fa..a6424cfc012 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -153,7 +153,7 @@ cifs_fill_fileinfo(struct inode *newinode, __u16 fileHandle,
 	mutex_init(&pCifsFile->fh_mutex);
 	mutex_init(&pCifsFile->lock_mutex);
 	INIT_LIST_HEAD(&pCifsFile->llist);
-	atomic_set(&pCifsFile->wrtPending, 0);
+	atomic_set(&pCifsFile->count, 1);
 
 	/* set the following in open now
 			pCifsFile->pfile = file; */
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index c34b7f8a217..fa7beac8b80 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -53,11 +53,9 @@ static inline struct cifsFileInfo *cifs_init_private(
 	private_data->pInode = inode;
 	private_data->invalidHandle = false;
 	private_data->closePend = false;
-	/* we have to track num writers to the inode, since writepages
-	does not tell us which handle the write is for so there can
-	be a close (overlapping with write) of the filehandle that
-	cifs_writepages chose to use */
-	atomic_set(&private_data->wrtPending, 0);
+	/* Initialize reference count to one.  The private data is
+	freed on the release of the last reference */
+	atomic_set(&private_data->count, 1);
 
 	return private_data;
 }
@@ -643,7 +641,7 @@ int cifs_close(struct inode *inode, struct file *file)
 			if (!pTcon->need_reconnect) {
 				write_unlock(&GlobalSMBSeslock);
 				timeout = 2;
-				while ((atomic_read(&pSMBFile->wrtPending) != 0)
+				while ((atomic_read(&pSMBFile->count) != 1)
 					&& (timeout <= 2048)) {
 					/* Give write a better chance to get to
 					server ahead of the close.  We do not
@@ -657,8 +655,6 @@ int cifs_close(struct inode *inode, struct file *file)
 					msleep(timeout);
 					timeout *= 4;
 				}
-				if (atomic_read(&pSMBFile->wrtPending))
-					cERROR(1, ("close with pending write"));
 				if (!pTcon->need_reconnect &&
 				    !pSMBFile->invalidHandle)
 					rc = CIFSSMBClose(xid, pTcon,
@@ -681,24 +677,7 @@ int cifs_close(struct inode *inode, struct file *file)
 		list_del(&pSMBFile->flist);
 		list_del(&pSMBFile->tlist);
 		write_unlock(&GlobalSMBSeslock);
-		timeout = 10;
-		/* We waited above to give the SMBWrite a chance to issue
-		   on the wire (so we do not get SMBWrite returning EBADF
-		   if writepages is racing with close.  Note that writepages
-		   does not specify a file handle, so it is possible for a file
-		   to be opened twice, and the application close the "wrong"
-		   file handle - in these cases we delay long enough to allow
-		   the SMBWrite to get on the wire before the SMB Close.
-		   We allow total wait here over 45 seconds, more than
-		   oplock break time, and more than enough to allow any write
-		   to complete on the server, or to time out on the client */
-		while ((atomic_read(&pSMBFile->wrtPending) != 0)
-				&& (timeout <= 50000)) {
-			cERROR(1, ("writes pending, delay free of handle"));
-			msleep(timeout);
-			timeout *= 8;
-		}
-		kfree(file->private_data);
+		cifsFileInfo_put(file->private_data);
 		file->private_data = NULL;
 	} else
 		rc = -EBADF;
@@ -1236,7 +1215,7 @@ struct cifsFileInfo *find_readable_file(struct cifsInodeInfo *cifs_inode)
 			if (!open_file->invalidHandle) {
 				/* found a good file */
 				/* lock it so it will not be closed on us */
-				atomic_inc(&open_file->wrtPending);
+				cifsFileInfo_get(open_file);
 				read_unlock(&GlobalSMBSeslock);
 				return open_file;
 			} /* else might as well continue, and look for
@@ -1276,7 +1255,7 @@ refind_writable:
 		if (open_file->pfile &&
 		    ((open_file->pfile->f_flags & O_RDWR) ||
 		     (open_file->pfile->f_flags & O_WRONLY))) {
-			atomic_inc(&open_file->wrtPending);
+			cifsFileInfo_get(open_file);
 
 			if (!open_file->invalidHandle) {
 				/* found a good writable file */
@@ -1293,7 +1272,7 @@ refind_writable:
 				else { /* start over in case this was deleted */
 				       /* since the list could be modified */
 					read_lock(&GlobalSMBSeslock);
-					atomic_dec(&open_file->wrtPending);
+					cifsFileInfo_put(open_file);
 					goto refind_writable;
 				}
 			}
@@ -1309,7 +1288,7 @@ refind_writable:
 			read_lock(&GlobalSMBSeslock);
 			/* can not use this handle, no write
 			   pending on this one after all */
-			atomic_dec(&open_file->wrtPending);
+			cifsFileInfo_put(open_file);
 
 			if (open_file->closePend) /* list could have changed */
 				goto refind_writable;
@@ -1373,7 +1352,7 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to)
 	if (open_file) {
 		bytes_written = cifs_write(open_file->pfile, write_data,
 					   to-from, &offset);
-		atomic_dec(&open_file->wrtPending);
+		cifsFileInfo_put(open_file);
 		/* Does mm or vfs already set times? */
 		inode->i_atime = inode->i_mtime = current_fs_time(inode->i_sb);
 		if ((bytes_written > 0) && (offset))
@@ -1562,7 +1541,7 @@ retry:
 						   bytes_to_write, offset,
 						   &bytes_written, iov, n_iov,
 						   long_op);
-				atomic_dec(&open_file->wrtPending);
+				cifsFileInfo_put(open_file);
 				cifs_update_eof(cifsi, offset, bytes_written);
 
 				if (rc || bytes_written < bytes_to_write) {
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index 82d83839655..1f09c761931 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -800,7 +800,7 @@ set_via_filehandle:
 	if (open_file == NULL)
 		CIFSSMBClose(xid, pTcon, netfid);
 	else
-		atomic_dec(&open_file->wrtPending);
+		cifsFileInfo_put(open_file);
 out:
 	return rc;
 }
@@ -1635,7 +1635,7 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 		__u32 npid = open_file->pid;
 		rc = CIFSSMBSetFileSize(xid, pTcon, attrs->ia_size, nfid,
 					npid, false);
-		atomic_dec(&open_file->wrtPending);
+		cifsFileInfo_put(open_file);
 		cFYI(1, ("SetFSize for attrs rc = %d", rc));
 		if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) {
 			unsigned int bytes_written;
@@ -1790,7 +1790,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs)
 		u16 nfid = open_file->netfid;
 		u32 npid = open_file->pid;
 		rc = CIFSSMBUnixSetFileInfo(xid, pTcon, args, nfid, npid);
-		atomic_dec(&open_file->wrtPending);
+		cifsFileInfo_put(open_file);
 	} else {
 		rc = CIFSSMBUnixSetPathInfo(xid, pTcon, full_path, args,
 				    cifs_sb->local_nls,
diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c
index 0ad3e2d116a..1da4ab250ea 100644
--- a/fs/cifs/transport.c
+++ b/fs/cifs/transport.c
@@ -119,20 +119,19 @@ AllocOplockQEntry(struct inode *pinode, __u16 fid, struct cifsTconInfo *tcon)
 		temp->pinode = pinode;
 		temp->tcon = tcon;
 		temp->netfid = fid;
-		spin_lock(&GlobalMid_Lock);
-		list_add_tail(&temp->qhead, &GlobalOplock_Q);
-		spin_unlock(&GlobalMid_Lock);
+		spin_lock(&cifs_oplock_lock);
+		list_add_tail(&temp->qhead, &cifs_oplock_list);
+		spin_unlock(&cifs_oplock_lock);
 	}
 	return temp;
-
 }
 
 void DeleteOplockQEntry(struct oplock_q_entry *oplockEntry)
 {
-	spin_lock(&GlobalMid_Lock);
+	spin_lock(&cifs_oplock_lock);
     /* should we check if list empty first? */
 	list_del(&oplockEntry->qhead);
-	spin_unlock(&GlobalMid_Lock);
+	spin_unlock(&cifs_oplock_lock);
 	kmem_cache_free(cifs_oplock_cachep, oplockEntry);
 }
 
@@ -144,14 +143,14 @@ void DeleteTconOplockQEntries(struct cifsTconInfo *tcon)
 	if (tcon == NULL)
 		return;
 
-	spin_lock(&GlobalMid_Lock);
-	list_for_each_entry(temp, &GlobalOplock_Q, qhead) {
+	spin_lock(&cifs_oplock_lock);
+	list_for_each_entry(temp, &cifs_oplock_list, qhead) {
 		if ((temp->tcon) && (temp->tcon == tcon)) {
 			list_del(&temp->qhead);
 			kmem_cache_free(cifs_oplock_cachep, temp);
 		}
 	}
-	spin_unlock(&GlobalMid_Lock);
+	spin_unlock(&cifs_oplock_lock);
 }
 
 static int
diff --git a/fs/compat.c b/fs/compat.c
index 6d6f98fe64a..3aa48834a22 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -100,13 +100,6 @@ asmlinkage long compat_sys_utimensat(unsigned int dfd, char __user *filename, st
 		    get_compat_timespec(&tv[1], &t[1]))
 			return -EFAULT;
 
-		if ((tv[0].tv_nsec == UTIME_OMIT || tv[0].tv_nsec == UTIME_NOW)
-		    && tv[0].tv_sec != 0)
-			return -EINVAL;
-		if ((tv[1].tv_nsec == UTIME_OMIT || tv[1].tv_nsec == UTIME_NOW)
-		    && tv[1].tv_sec != 0)
-			return -EINVAL;
-
 		if (tv[0].tv_nsec == UTIME_OMIT && tv[1].tv_nsec == UTIME_OMIT)
 			return 0;
 	}
diff --git a/fs/configfs/inode.c b/fs/configfs/inode.c
index 4921e7426d9..a2f746066c5 100644
--- a/fs/configfs/inode.c
+++ b/fs/configfs/inode.c
@@ -51,6 +51,7 @@ static const struct address_space_operations configfs_aops = {
 };
 
 static struct backing_dev_info configfs_backing_dev_info = {
+	.name		= "configfs",
 	.ra_pages	= 0,	/* No readahead */
 	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK,
 };
diff --git a/fs/dcache.c b/fs/dcache.c
index 9e5cd3c3a6b..a100fa35a48 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -32,6 +32,7 @@
 #include <linux/swap.h>
 #include <linux/bootmem.h>
 #include <linux/fs_struct.h>
+#include <linux/hardirq.h>
 #include "internal.h"
 
 int sysctl_vfs_cache_pressure __read_mostly = 100;
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index 75efb028974..d5f8c96964b 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -18,14 +18,13 @@
 #include <linux/mount.h>
 #include <linux/tty.h>
 #include <linux/mutex.h>
+#include <linux/magic.h>
 #include <linux/idr.h>
 #include <linux/devpts_fs.h>
 #include <linux/parser.h>
 #include <linux/fsnotify.h>
 #include <linux/seq_file.h>
 
-#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-
diff --git a/fs/dlm/debug_fs.c b/fs/dlm/debug_fs.c
index 1d1d2744223..1c8bb8c3a82 100644
--- a/fs/dlm/debug_fs.c
+++ b/fs/dlm/debug_fs.c
@@ -386,9 +386,9 @@ static int table_seq_show(struct seq_file *seq, void *iter_ptr)
 	return rv;
 }
 
-static struct seq_operations format1_seq_ops;
-static struct seq_operations format2_seq_ops;
-static struct seq_operations format3_seq_ops;
+static const struct seq_operations format1_seq_ops;
+static const struct seq_operations format2_seq_ops;
+static const struct seq_operations format3_seq_ops;
 
 static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 {
@@ -534,21 +534,21 @@ static void table_seq_stop(struct seq_file *seq, void *iter_ptr)
 	}
 }
 
-static struct seq_operations format1_seq_ops = {
+static const struct seq_operations format1_seq_ops = {
 	.start = table_seq_start,
 	.next  = table_seq_next,
 	.stop  = table_seq_stop,
 	.show  = table_seq_show,
 };
 
-static struct seq_operations format2_seq_ops = {
+static const struct seq_operations format2_seq_ops = {
 	.start = table_seq_start,
 	.next  = table_seq_next,
 	.stop  = table_seq_stop,
 	.show  = table_seq_show,
 };
 
-static struct seq_operations format3_seq_ops = {
+static const struct seq_operations format3_seq_ops = {
 	.start = table_seq_start,
 	.next  = table_seq_next,
 	.stop  = table_seq_stop,
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 618a60f0388..240cef14fe5 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -106,6 +106,7 @@ struct connection {
 #define CF_CONNECT_PENDING 3
 #define CF_INIT_PENDING 4
 #define CF_IS_OTHERCON 5
+#define CF_CLOSE 6
 	struct list_head writequeue;  /* List of outgoing writequeue_entries */
 	spinlock_t writequeue_lock;
 	int (*rx_action) (struct connection *);	/* What to do when active */
@@ -299,6 +300,8 @@ static void lowcomms_write_space(struct sock *sk)
 
 static inline void lowcomms_connect_sock(struct connection *con)
 {
+	if (test_bit(CF_CLOSE, &con->flags))
+		return;
 	if (!test_and_set_bit(CF_CONNECT_PENDING, &con->flags))
 		queue_work(send_workqueue, &con->swork);
 }
@@ -926,10 +929,8 @@ static void tcp_connect_to_sock(struct connection *con)
 		goto out_err;
 
 	memset(&saddr, 0, sizeof(saddr));
-	if (dlm_nodeid_to_addr(con->nodeid, &saddr)) {
-		sock_release(sock);
+	if (dlm_nodeid_to_addr(con->nodeid, &saddr))
 		goto out_err;
-	}
 
 	sock->sk->sk_user_data = con;
 	con->rx_action = receive_from_sock;
@@ -1284,7 +1285,6 @@ out:
 static void send_to_sock(struct connection *con)
 {
 	int ret = 0;
-	ssize_t(*sendpage) (struct socket *, struct page *, int, size_t, int);
 	const int msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL;
 	struct writequeue_entry *e;
 	int len, offset;
@@ -1293,8 +1293,6 @@ static void send_to_sock(struct connection *con)
 	if (con->sock == NULL)
 		goto out_connect;
 
-	sendpage = con->sock->ops->sendpage;
-
 	spin_lock(&con->writequeue_lock);
 	for (;;) {
 		e = list_entry(con->writequeue.next, struct writequeue_entry,
@@ -1309,8 +1307,8 @@ static void send_to_sock(struct connection *con)
 
 		ret = 0;
 		if (len) {
-			ret = sendpage(con->sock, e->page, offset, len,
-				       msg_flags);
+			ret = kernel_sendpage(con->sock, e->page, offset, len,
+					      msg_flags);
 			if (ret == -EAGAIN || ret == 0) {
 				cond_resched();
 				goto out;
@@ -1370,6 +1368,13 @@ int dlm_lowcomms_close(int nodeid)
 	log_print("closing connection to node %d", nodeid);
 	con = nodeid2con(nodeid, 0);
 	if (con) {
+		clear_bit(CF_CONNECT_PENDING, &con->flags);
+		clear_bit(CF_WRITE_PENDING, &con->flags);
+		set_bit(CF_CLOSE, &con->flags);
+		if (cancel_work_sync(&con->swork))
+			log_print("canceled swork for node %d", nodeid);
+		if (cancel_work_sync(&con->rwork))
+			log_print("canceled rwork for node %d", nodeid);
 		clean_one_writequeue(con);
 		close_connection(con, true);
 	}
@@ -1395,9 +1400,10 @@ static void process_send_sockets(struct work_struct *work)
 
 	if (test_and_clear_bit(CF_CONNECT_PENDING, &con->flags)) {
 		con->connect_action(con);
+		set_bit(CF_WRITE_PENDING, &con->flags);
 	}
-	clear_bit(CF_WRITE_PENDING, &con->flags);
-	send_to_sock(con);
+	if (test_and_clear_bit(CF_WRITE_PENDING, &con->flags))
+		send_to_sock(con);
 }
 
 
diff --git a/fs/dlm/netlink.c b/fs/dlm/netlink.c
index ccc9d62c462..55ea369f43a 100644
--- a/fs/dlm/netlink.c
+++ b/fs/dlm/netlink.c
@@ -63,7 +63,7 @@ static int send_data(struct sk_buff *skb)
 		return rv;
 	}
 
-	return genlmsg_unicast(skb, listener_nlpid);
+	return genlmsg_unicast(&init_net, skb, listener_nlpid);
 }
 
 static int user_cmd(struct sk_buff *skb, struct genl_info *info)
diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h
index 00b30a2d546..542f625312f 100644
--- a/fs/ecryptfs/ecryptfs_kernel.h
+++ b/fs/ecryptfs/ecryptfs_kernel.h
@@ -582,7 +582,7 @@ extern const struct inode_operations ecryptfs_dir_iops;
 extern const struct inode_operations ecryptfs_symlink_iops;
 extern const struct super_operations ecryptfs_sops;
 extern const struct dentry_operations ecryptfs_dops;
-extern struct address_space_operations ecryptfs_aops;
+extern const struct address_space_operations ecryptfs_aops;
 extern int ecryptfs_verbosity;
 extern unsigned int ecryptfs_message_buf_len;
 extern signed long ecryptfs_message_wait_timeout;
diff --git a/fs/ecryptfs/mmap.c b/fs/ecryptfs/mmap.c
index 5c6bab9786e..05772aeaa8f 100644
--- a/fs/ecryptfs/mmap.c
+++ b/fs/ecryptfs/mmap.c
@@ -545,7 +545,7 @@ static sector_t ecryptfs_bmap(struct address_space *mapping, sector_t block)
 	return rc;
 }
 
-struct address_space_operations ecryptfs_aops = {
+const struct address_space_operations ecryptfs_aops = {
 	.writepage = ecryptfs_writepage,
 	.readpage = ecryptfs_readpage,
 	.write_begin = ecryptfs_write_begin,
diff --git a/fs/eventfd.c b/fs/eventfd.c
index 31d12de83a2..8b47e4200e6 100644
--- a/fs/eventfd.c
+++ b/fs/eventfd.c
@@ -68,11 +68,16 @@ int eventfd_signal(struct eventfd_ctx *ctx, int n)
 }
 EXPORT_SYMBOL_GPL(eventfd_signal);
 
+static void eventfd_free_ctx(struct eventfd_ctx *ctx)
+{
+	kfree(ctx);
+}
+
 static void eventfd_free(struct kref *kref)
 {
 	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);
 
-	kfree(ctx);
+	eventfd_free_ctx(ctx);
 }
 
 /**
@@ -298,9 +303,23 @@ struct eventfd_ctx *eventfd_ctx_fileget(struct file *file)
 }
 EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
 
-SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
+/**
+ * eventfd_file_create - Creates an eventfd file pointer.
+ * @count: Initial eventfd counter value.
+ * @flags: Flags for the eventfd file.
+ *
+ * This function creates an eventfd file pointer, w/out installing it into
+ * the fd table. This is useful when the eventfd file is used during the
+ * initialization of data structures that require extra setup after the eventfd
+ * creation. So the eventfd creation is split into the file pointer creation
+ * phase, and the file descriptor installation phase.
+ * In this way races with userspace closing the newly installed file descriptor
+ * can be avoided.
+ * Returns an eventfd file pointer, or a proper error pointer.
+ */
+struct file *eventfd_file_create(unsigned int count, int flags)
 {
-	int fd;
+	struct file *file;
 	struct eventfd_ctx *ctx;
 
 	/* Check the EFD_* constants for consistency.  */
@@ -308,26 +327,48 @@ SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
 	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
 
 	if (flags & ~EFD_FLAGS_SET)
-		return -EINVAL;
+		return ERR_PTR(-EINVAL);
 
 	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	kref_init(&ctx->kref);
 	init_waitqueue_head(&ctx->wqh);
 	ctx->count = count;
 	ctx->flags = flags;
 
-	/*
-	 * When we call this, the initialization must be complete, since
-	 * anon_inode_getfd() will install the fd.
-	 */
-	fd = anon_inode_getfd("[eventfd]", &eventfd_fops, ctx,
-			      flags & EFD_SHARED_FCNTL_FLAGS);
-	if (fd < 0)
-		kfree(ctx);
+	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx,
+				  flags & EFD_SHARED_FCNTL_FLAGS);
+	if (IS_ERR(file))
+		eventfd_free_ctx(ctx);
+
+	return file;
+}
+
+SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
+{
+	int fd, error;
+	struct file *file;
+
+	error = get_unused_fd_flags(flags & EFD_SHARED_FCNTL_FLAGS);
+	if (error < 0)
+		return error;
+	fd = error;
+
+	file = eventfd_file_create(count, flags);
+	if (IS_ERR(file)) {
+		error = PTR_ERR(file);
+		goto err_put_unused_fd;
+	}
+	fd_install(fd, file);
+
 	return fd;
+
+err_put_unused_fd:
+	put_unused_fd(fd);
+
+	return error;
 }
 
 SYSCALL_DEFINE1(eventfd, unsigned int, count)
diff --git a/fs/exec.c b/fs/exec.c
index 172ceb6edde..5c833c18d0d 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -33,7 +33,7 @@
 #include <linux/string.h>
 #include <linux/init.h>
 #include <linux/pagemap.h>
-#include <linux/perf_counter.h>
+#include <linux/perf_event.h>
 #include <linux/highmem.h>
 #include <linux/spinlock.h>
 #include <linux/key.h>
@@ -845,6 +845,9 @@ static int de_thread(struct task_struct *tsk)
 	sig->notify_count = 0;
 
 no_thread_group:
+	if (current->mm)
+		setmax_mm_hiwater_rss(&sig->maxrss, current->mm);
+
 	exit_itimers(sig);
 	flush_itimer_signals();
 
@@ -923,7 +926,7 @@ void set_task_comm(struct task_struct *tsk, char *buf)
 	task_lock(tsk);
 	strlcpy(tsk->comm, buf, sizeof(tsk->comm));
 	task_unlock(tsk);
-	perf_counter_comm(tsk);
+	perf_event_comm(tsk);
 }
 
 int flush_old_exec(struct linux_binprm * bprm)
@@ -997,7 +1000,7 @@ int flush_old_exec(struct linux_binprm * bprm)
 	 * security domain:
 	 */
 	if (!get_dumpable(current->mm))
-		perf_counter_exit_task(current);
+		perf_event_exit_task(current);
 
 	/* An exec changes our domain. We are no longer part of the thread
 	   group */
@@ -1354,6 +1357,8 @@ int do_execve(char * filename,
 	if (retval < 0)
 		goto out;
 
+	current->stack_start = current->mm->start_stack;
+
 	/* execve succeeded */
 	current->fs->in_exec = 0;
 	current->in_execve = 0;
diff --git a/fs/ext2/acl.c b/fs/ext2/acl.c
index d636e1297ca..a63d44256a7 100644
--- a/fs/ext2/acl.c
+++ b/fs/ext2/acl.c
@@ -230,7 +230,7 @@ ext2_set_acl(struct inode *inode, int type, struct posix_acl *acl)
 	return error;
 }
 
-static int
+int
 ext2_check_acl(struct inode *inode, int mask)
 {
 	struct posix_acl *acl = ext2_get_acl(inode, ACL_TYPE_ACCESS);
@@ -246,12 +246,6 @@ ext2_check_acl(struct inode *inode, int mask)
 	return -EAGAIN;
 }
 
-int
-ext2_permission(struct inode *inode, int mask)
-{
-	return generic_permission(inode, mask, ext2_check_acl);
-}
-
 /*
  * Initialize the ACLs of a new inode. Called from ext2_new_inode.
  *
diff --git a/fs/ext2/acl.h b/fs/ext2/acl.h
index ecefe478898..3ff6cbb9ac4 100644
--- a/fs/ext2/acl.h
+++ b/fs/ext2/acl.h
@@ -54,13 +54,13 @@ static inline int ext2_acl_count(size_t size)
 #ifdef CONFIG_EXT2_FS_POSIX_ACL
 
 /* acl.c */
-extern int ext2_permission (struct inode *, int);
+extern int ext2_check_acl (struct inode *, int);
 extern int ext2_acl_chmod (struct inode *);
 extern int ext2_init_acl (struct inode *, struct inode *);
 
 #else
 #include <linux/sched.h>
-#define ext2_permission NULL
+#define ext2_check_acl	NULL
 #define ext2_get_acl	NULL
 #define ext2_set_acl	NULL
 
diff --git a/fs/ext2/file.c b/fs/ext2/file.c
index 2b9e47dc922..a2f3afd1a1c 100644
--- a/fs/ext2/file.c
+++ b/fs/ext2/file.c
@@ -85,6 +85,6 @@ const struct inode_operations ext2_file_inode_operations = {
 	.removexattr	= generic_removexattr,
 #endif
 	.setattr	= ext2_setattr,
-	.permission	= ext2_permission,
+	.check_acl	= ext2_check_acl,
 	.fiemap		= ext2_fiemap,
 };
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index e27130341d4..1c1638f873a 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -482,7 +482,7 @@ static int ext2_alloc_branch(struct inode *inode,
 		unlock_buffer(bh);
 		mark_buffer_dirty_inode(bh, inode);
 		/* We used to sync bh here if IS_SYNC(inode).
-		 * But we now rely upon generic_osync_inode()
+		 * But we now rely upon generic_write_sync()
 		 * and b_inode_buffers.  But not for directories.
 		 */
 		if (S_ISDIR(inode->i_mode) && IS_DIRSYNC(inode))
diff --git a/fs/ext2/namei.c b/fs/ext2/namei.c
index 78d9b925fc9..dd7175ce560 100644
--- a/fs/ext2/namei.c
+++ b/fs/ext2/namei.c
@@ -70,7 +70,7 @@ static struct dentry *ext2_lookup(struct inode * dir, struct dentry *dentry, str
 			if (PTR_ERR(inode) == -ESTALE) {
 				ext2_error(dir->i_sb, __func__,
 						"deleted inode referenced: %lu",
-						ino);
+						(unsigned long) ino);
 				return ERR_PTR(-EIO);
 			} else {
 				return ERR_CAST(inode);
@@ -400,7 +400,7 @@ const struct inode_operations ext2_dir_inode_operations = {
 	.removexattr	= generic_removexattr,
 #endif
 	.setattr	= ext2_setattr,
-	.permission	= ext2_permission,
+	.check_acl	= ext2_check_acl,
 };
 
 const struct inode_operations ext2_special_inode_operations = {
@@ -411,5 +411,5 @@ const struct inode_operations ext2_special_inode_operations = {
 	.removexattr	= generic_removexattr,
 #endif
 	.setattr	= ext2_setattr,
-	.permission	= ext2_permission,
+	.check_acl	= ext2_check_acl,
 };
diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c
index b72b8588422..c18fbf3e406 100644
--- a/fs/ext2/xip.c
+++ b/fs/ext2/xip.c
@@ -20,7 +20,7 @@ __inode_direct_access(struct inode *inode, sector_t block,
 		      void **kaddr, unsigned long *pfn)
 {
 	struct block_device *bdev = inode->i_sb->s_bdev;
-	struct block_device_operations *ops = bdev->bd_disk->fops;
+	const struct block_device_operations *ops = bdev->bd_disk->fops;
 	sector_t sector;
 
 	sector = block * (PAGE_SIZE / 512); /* ext2 block to bdev sector */
diff --git a/fs/ext3/acl.c b/fs/ext3/acl.c
index e167bae37ef..c9b0df376b5 100644
--- a/fs/ext3/acl.c
+++ b/fs/ext3/acl.c
@@ -238,7 +238,7 @@ ext3_set_acl(handle_t *handle, struct inode *inode, int type,
 	return error;
 }
 
-static int
+int
 ext3_check_acl(struct inode *inode, int mask)
 {
 	struct posix_acl *acl = ext3_get_acl(inode, ACL_TYPE_ACCESS);
@@ -254,12 +254,6 @@ ext3_check_acl(struct inode *inode, int mask)
 	return -EAGAIN;
 }
 
-int
-ext3_permission(struct inode *inode, int mask)
-{
-	return generic_permission(inode, mask, ext3_check_acl);
-}
-
 /*
  * Initialize the ACLs of a new inode. Called from ext3_new_inode.
  *
diff --git a/fs/ext3/acl.h b/fs/ext3/acl.h
index 07d15a3a596..597334626de 100644
--- a/fs/ext3/acl.h
+++ b/fs/ext3/acl.h
@@ -54,13 +54,13 @@ static inline int ext3_acl_count(size_t size)
 #ifdef CONFIG_EXT3_FS_POSIX_ACL
 
 /* acl.c */
-extern int ext3_permission (struct inode *, int);
+extern int ext3_check_acl (struct inode *, int);
 extern int ext3_acl_chmod (struct inode *);
 extern int ext3_init_acl (handle_t *, struct inode *, struct inode *);
 
 #else  /* CONFIG_EXT3_FS_POSIX_ACL */
 #include <linux/sched.h>
-#define ext3_permission NULL
+#define ext3_check_acl NULL
 
 static inline int
 ext3_acl_chmod(struct inode *inode)
diff --git a/fs/ext3/file.c b/fs/ext3/file.c
index 5b49704b231..388bbdfa0b4 100644
--- a/fs/ext3/file.c
+++ b/fs/ext3/file.c
@@ -51,71 +51,12 @@ static int ext3_release_file (struct inode * inode, struct file * filp)
 	return 0;
 }
 
-static ssize_t
-ext3_file_write(struct kiocb *iocb, const struct iovec *iov,
-		unsigned long nr_segs, loff_t pos)
-{
-	struct file *file = iocb->ki_filp;
-	struct inode *inode = file->f_path.dentry->d_inode;
-	ssize_t ret;
-	int err;
-
-	ret = generic_file_aio_write(iocb, iov, nr_segs, pos);
-
-	/*
-	 * Skip flushing if there was an error, or if nothing was written.
-	 */
-	if (ret <= 0)
-		return ret;
-
-	/*
-	 * If the inode is IS_SYNC, or is O_SYNC and we are doing data
-	 * journalling then we need to make sure that we force the transaction
-	 * to disk to keep all metadata uptodate synchronously.
-	 */
-	if (file->f_flags & O_SYNC) {
-		/*
-		 * If we are non-data-journaled, then the dirty data has
-		 * already been flushed to backing store by generic_osync_inode,
-		 * and the inode has been flushed too if there have been any
-		 * modifications other than mere timestamp updates.
-		 *
-		 * Open question --- do we care about flushing timestamps too
-		 * if the inode is IS_SYNC?
-		 */
-		if (!ext3_should_journal_data(inode))
-			return ret;
-
-		goto force_commit;
-	}
-
-	/*
-	 * So we know that there has been no forced data flush.  If the inode
-	 * is marked IS_SYNC, we need to force one ourselves.
-	 */
-	if (!IS_SYNC(inode))
-		return ret;
-
-	/*
-	 * Open question #2 --- should we force data to disk here too?  If we
-	 * don't, the only impact is that data=writeback filesystems won't
-	 * flush data to disk automatically on IS_SYNC, only metadata (but
-	 * historically, that is what ext2 has done.)
-	 */
-
-force_commit:
-	err = ext3_force_commit(inode->i_sb);
-	if (err)
-		return err;
-	return ret;
-}
-
 const struct file_operations ext3_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
 	.aio_read	= generic_file_aio_read,
-	.aio_write	= ext3_file_write,
+	.aio_write	= generic_file_aio_write,
 	.unlocked_ioctl	= ext3_ioctl,
 #ifdef CONFIG_COMPAT
 	.compat_ioctl	= ext3_compat_ioctl,
@@ -137,7 +78,7 @@ const struct inode_operations ext3_file_inode_operations = {
 	.listxattr	= ext3_listxattr,
 	.removexattr	= generic_removexattr,
 #endif
-	.permission	= ext3_permission,
+	.check_acl	= ext3_check_acl,
 	.fiemap		= ext3_fiemap,
 };
 
diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c
index d33634119e1..451d166bbe9 100644
--- a/fs/ext3/fsync.c
+++ b/fs/ext3/fsync.c
@@ -23,6 +23,7 @@
  */
 
 #include <linux/time.h>
+#include <linux/blkdev.h>
 #include <linux/fs.h>
 #include <linux/sched.h>
 #include <linux/writeback.h>
@@ -73,7 +74,7 @@ int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
 	}
 
 	if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
-		goto out;
+		goto flush;
 
 	/*
 	 * The VFS has written the file data.  If the inode is unaltered
@@ -85,7 +86,16 @@ int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
 			.nr_to_write = 0, /* sys_fsync did this */
 		};
 		ret = sync_inode(inode, &wbc);
+		goto out;
 	}
+flush:
+	/*
+	 * In case we didn't commit a transaction, we have to flush
+	 * disk caches manually so that data really is on persistent
+	 * storage
+	 */
+	if (test_opt(inode->i_sb, BARRIER))
+		blkdev_issue_flush(inode->i_sb->s_bdev, NULL);
 out:
 	return ret;
 }
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index b49908a167a..cd098a7b77f 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -172,10 +172,21 @@ static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
  * so before we call here everything must be consistently dirtied against
  * this transaction.
  */
-static int ext3_journal_test_restart(handle_t *handle, struct inode *inode)
+static int truncate_restart_transaction(handle_t *handle, struct inode *inode)
 {
+	int ret;
+
 	jbd_debug(2, "restarting handle %p\n", handle);
-	return ext3_journal_restart(handle, blocks_for_truncate(inode));
+	/*
+	 * Drop truncate_mutex to avoid deadlock with ext3_get_blocks_handle
+	 * At this moment, get_block can be called only for blocks inside
+	 * i_size since page cache has been already dropped and writes are
+	 * blocked by i_mutex. So we can safely drop the truncate_mutex.
+	 */
+	mutex_unlock(&EXT3_I(inode)->truncate_mutex);
+	ret = ext3_journal_restart(handle, blocks_for_truncate(inode));
+	mutex_lock(&EXT3_I(inode)->truncate_mutex);
+	return ret;
 }
 
 /*
@@ -2072,7 +2083,7 @@ static void ext3_clear_blocks(handle_t *handle, struct inode *inode,
 			ext3_journal_dirty_metadata(handle, bh);
 		}
 		ext3_mark_inode_dirty(handle, inode);
-		ext3_journal_test_restart(handle, inode);
+		truncate_restart_transaction(handle, inode);
 		if (bh) {
 			BUFFER_TRACE(bh, "retaking write access");
 			ext3_journal_get_write_access(handle, bh);
@@ -2282,7 +2293,7 @@ static void ext3_free_branches(handle_t *handle, struct inode *inode,
 				return;
 			if (try_to_extend_transaction(handle, inode)) {
 				ext3_mark_inode_dirty(handle, inode);
-				ext3_journal_test_restart(handle, inode);
+				truncate_restart_transaction(handle, inode);
 			}
 
 			ext3_free_blocks(handle, inode, nr, 1);
@@ -2892,6 +2903,10 @@ static int ext3_do_update_inode(handle_t *handle,
 	struct buffer_head *bh = iloc->bh;
 	int err = 0, rc, block;
 
+again:
+	/* we can't allow multiple procs in here at once, its a bit racey */
+	lock_buffer(bh);
+
 	/* For fields not not tracking in the in-memory inode,
 	 * initialise them to zero for new inodes. */
 	if (ei->i_state & EXT3_STATE_NEW)
@@ -2951,16 +2966,20 @@ static int ext3_do_update_inode(handle_t *handle,
 			       /* If this is the first large file
 				* created, add a flag to the superblock.
 				*/
+				unlock_buffer(bh);
 				err = ext3_journal_get_write_access(handle,
 						EXT3_SB(sb)->s_sbh);
 				if (err)
 					goto out_brelse;
+
 				ext3_update_dynamic_rev(sb);
 				EXT3_SET_RO_COMPAT_FEATURE(sb,
 					EXT3_FEATURE_RO_COMPAT_LARGE_FILE);
 				handle->h_sync = 1;
 				err = ext3_journal_dirty_metadata(handle,
 						EXT3_SB(sb)->s_sbh);
+				/* get our lock and start over */
+				goto again;
 			}
 		}
 	}
@@ -2983,6 +3002,7 @@ static int ext3_do_update_inode(handle_t *handle,
 		raw_inode->i_extra_isize = cpu_to_le16(ei->i_extra_isize);
 
 	BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
+	unlock_buffer(bh);
 	rc = ext3_journal_dirty_metadata(handle, bh);
 	if (!err)
 		err = rc;
diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
index 6ff7b973023..aad6400c9b7 100644
--- a/fs/ext3/namei.c
+++ b/fs/ext3/namei.c
@@ -2445,7 +2445,7 @@ const struct inode_operations ext3_dir_inode_operations = {
 	.listxattr	= ext3_listxattr,
 	.removexattr	= generic_removexattr,
 #endif
-	.permission	= ext3_permission,
+	.check_acl	= ext3_check_acl,
 };
 
 const struct inode_operations ext3_special_inode_operations = {
@@ -2456,5 +2456,5 @@ const struct inode_operations ext3_special_inode_operations = {
 	.listxattr	= ext3_listxattr,
 	.removexattr	= generic_removexattr,
 #endif
-	.permission	= ext3_permission,
+	.check_acl	= ext3_check_acl,
 };
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index a8d80a7f110..72743d36050 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -720,7 +720,7 @@ static ssize_t ext3_quota_read(struct super_block *sb, int type, char *data,
 static ssize_t ext3_quota_write(struct super_block *sb, int type,
 				const char *data, size_t len, loff_t off);
 
-static struct dquot_operations ext3_quota_operations = {
+static const struct dquot_operations ext3_quota_operations = {
 	.initialize	= dquot_initialize,
 	.drop		= dquot_drop,
 	.alloc_space	= dquot_alloc_space,
@@ -737,7 +737,7 @@ static struct dquot_operations ext3_quota_operations = {
 	.destroy_dquot	= dquot_destroy,
 };
 
-static struct quotactl_ops ext3_qctl_operations = {
+static const struct quotactl_ops ext3_qctl_operations = {
 	.quota_on	= ext3_quota_on,
 	.quota_off	= vfs_quota_off,
 	.quota_sync	= vfs_quota_sync,
diff --git a/fs/ext4/Kconfig b/fs/ext4/Kconfig
index 418b6f3b0ae..d5c0ea2e8f2 100644
--- a/fs/ext4/Kconfig
+++ b/fs/ext4/Kconfig
@@ -37,7 +37,7 @@ config EXT4DEV_COMPAT
 
 	  To enable backwards compatibility so that systems that are
 	  still expecting to mount ext4 filesystems using ext4dev,
-	  chose Y here.   This feature will go away by 2.6.31, so
+	  choose Y here.   This feature will go away by 2.6.31, so
 	  please arrange to get your userspace programs fixed!
 
 config EXT4_FS_XATTR
@@ -77,3 +77,12 @@ config EXT4_FS_SECURITY
 
 	  If you are not using a security module that requires using
 	  extended attributes for file security labels, say N.
+
+config EXT4_DEBUG
+	bool "EXT4 debugging support"
+	depends on EXT4_FS
+	help
+	  Enables run-time debugging support for the ext4 filesystem.
+
+	  If you select Y here, then you will be able to turn on debugging
+	  with a command such as "echo 1 > /sys/kernel/debug/ext4/mballoc-debug"
diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c
index f6d8967149c..0df88b2a69b 100644
--- a/fs/ext4/acl.c
+++ b/fs/ext4/acl.c
@@ -236,7 +236,7 @@ ext4_set_acl(handle_t *handle, struct inode *inode, int type,
 	return error;
 }
 
-static int
+int
 ext4_check_acl(struct inode *inode, int mask)
 {
 	struct posix_acl *acl = ext4_get_acl(inode, ACL_TYPE_ACCESS);
@@ -252,12 +252,6 @@ ext4_check_acl(struct inode *inode, int mask)
 	return -EAGAIN;
 }
 
-int
-ext4_permission(struct inode *inode, int mask)
-{
-	return generic_permission(inode, mask, ext4_check_acl);
-}
-
 /*
  * Initialize the ACLs of a new inode. Called from ext4_new_inode.
  *
diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h
index 949789d2bba..9d843d5deac 100644
--- a/fs/ext4/acl.h
+++ b/fs/ext4/acl.h
@@ -54,13 +54,13 @@ static inline int ext4_acl_count(size_t size)
 #ifdef CONFIG_EXT4_FS_POSIX_ACL
 
 /* acl.c */
-extern int ext4_permission(struct inode *, int);
+extern int ext4_check_acl(struct inode *, int);
 extern int ext4_acl_chmod(struct inode *);
 extern int ext4_init_acl(handle_t *, struct inode *, struct inode *);
 
 #else  /* CONFIG_EXT4_FS_POSIX_ACL */
 #include <linux/sched.h>
-#define ext4_permission NULL
+#define ext4_check_acl NULL
 
 static inline int
 ext4_acl_chmod(struct inode *inode)
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index e2126d70dff..1d0418980f8 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -478,7 +478,7 @@ void ext4_add_groupblocks(handle_t *handle, struct super_block *sb,
 	 * new bitmap information
 	 */
 	set_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state));
-	ext4_mb_update_group_info(grp, blocks_freed);
+	grp->bb_free += blocks_freed;
 	up_write(&grp->alloc_sem);
 
 	/* We dirtied the bitmap block */
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 9714db393ef..e227eea23f0 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -67,27 +67,29 @@ typedef unsigned int ext4_group_t;
 
 
 /* prefer goal again. length */
-#define EXT4_MB_HINT_MERGE		1
+#define EXT4_MB_HINT_MERGE		0x0001
 /* blocks already reserved */
-#define EXT4_MB_HINT_RESERVED		2
+#define EXT4_MB_HINT_RESERVED		0x0002
 /* metadata is being allocated */
-#define EXT4_MB_HINT_METADATA		4
+#define EXT4_MB_HINT_METADATA		0x0004
 /* first blocks in the file */
-#define EXT4_MB_HINT_FIRST		8
+#define EXT4_MB_HINT_FIRST		0x0008
 /* search for the best chunk */
-#define EXT4_MB_HINT_BEST		16
+#define EXT4_MB_HINT_BEST		0x0010
 /* data is being allocated */
-#define EXT4_MB_HINT_DATA		32
+#define EXT4_MB_HINT_DATA		0x0020
 /* don't preallocate (for tails) */
-#define EXT4_MB_HINT_NOPREALLOC		64
+#define EXT4_MB_HINT_NOPREALLOC		0x0040
 /* allocate for locality group */
-#define EXT4_MB_HINT_GROUP_ALLOC	128
+#define EXT4_MB_HINT_GROUP_ALLOC	0x0080
 /* allocate goal blocks or none */
-#define EXT4_MB_HINT_GOAL_ONLY		256
+#define EXT4_MB_HINT_GOAL_ONLY		0x0100
 /* goal is meaningful */
-#define EXT4_MB_HINT_TRY_GOAL		512
+#define EXT4_MB_HINT_TRY_GOAL		0x0200
 /* blocks already pre-reserved by delayed allocation */
-#define EXT4_MB_DELALLOC_RESERVED      1024
+#define EXT4_MB_DELALLOC_RESERVED	0x0400
+/* We are doing stream allocation */
+#define EXT4_MB_STREAM_ALLOC		0x0800
 
 
 struct ext4_allocation_request {
@@ -112,6 +114,21 @@ struct ext4_allocation_request {
 };
 
 /*
+ * For delayed allocation tracking
+ */
+struct mpage_da_data {
+	struct inode *inode;
+	sector_t b_blocknr;		/* start block number of extent */
+	size_t b_size;			/* size of extent */
+	unsigned long b_state;		/* state of the extent */
+	unsigned long first_page, next_page;	/* extent of pages */
+	struct writeback_control *wbc;
+	int io_done;
+	int pages_written;
+	int retval;
+};
+
+/*
  * Special inodes numbers
  */
 #define	EXT4_BAD_INO		 1	/* Bad blocks inode */
@@ -251,7 +268,6 @@ struct flex_groups {
 #define EXT4_TOPDIR_FL			0x00020000 /* Top of directory hierarchies*/
 #define EXT4_HUGE_FILE_FL               0x00040000 /* Set to each huge file */
 #define EXT4_EXTENTS_FL			0x00080000 /* Inode uses extents */
-#define EXT4_EXT_MIGRATE		0x00100000 /* Inode is migrating */
 #define EXT4_RESERVED_FL		0x80000000 /* reserved for ext4 lib */
 
 #define EXT4_FL_USER_VISIBLE		0x000BDFFF /* User visible flags */
@@ -289,6 +305,7 @@ static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
 #define EXT4_STATE_XATTR		0x00000004 /* has in-inode xattrs */
 #define EXT4_STATE_NO_EXPAND		0x00000008 /* No space for expansion */
 #define EXT4_STATE_DA_ALLOC_CLOSE	0x00000010 /* Alloc DA blks on close */
+#define EXT4_STATE_EXT_MIGRATE		0x00000020 /* Inode is migrating */
 
 /* Used to pass group descriptor data when online resize is done */
 struct ext4_new_group_input {
@@ -386,6 +403,9 @@ struct ext4_mount_options {
 #endif
 };
 
+/* Max physical block we can addres w/o extents */
+#define EXT4_MAX_BLOCK_FILE_PHYS	0xFFFFFFFF
+
 /*
  * Structure of an inode on the disk
  */
@@ -456,7 +476,6 @@ struct move_extent {
 	__u64 len;		/* block length to be moved */
 	__u64 moved_len;	/* moved block length */
 };
-#define MAX_DEFRAG_SIZE         ((1UL<<31) - 1)
 
 #define EXT4_EPOCH_BITS 2
 #define EXT4_EPOCH_MASK ((1 << EXT4_EPOCH_BITS) - 1)
@@ -694,7 +713,6 @@ struct ext4_inode_info {
 #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_JOURNAL_CHECKSUM	0x800000 /* Journal checksums */
 #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT	0x1000000 /* Journal Async Commit */
 #define EXT4_MOUNT_I_VERSION            0x2000000 /* i_version support */
 #define EXT4_MOUNT_DELALLOC		0x8000000 /* Delalloc support */
@@ -841,6 +859,7 @@ struct ext4_sb_info {
 	unsigned long s_gdb_count;	/* Number of group descriptor blocks */
 	unsigned long s_desc_per_block;	/* Number of group descriptors per block */
 	ext4_group_t s_groups_count;	/* Number of groups in the fs */
+	ext4_group_t s_blockfile_groups;/* Groups acceptable for non-extent files */
 	unsigned long s_overhead_last;  /* Last calculated overhead */
 	unsigned long s_blocks_last;    /* Last seen block count */
 	loff_t s_bitmap_maxbytes;	/* max bytes for bitmap files */
@@ -950,6 +969,7 @@ struct ext4_sb_info {
 	atomic_t s_mb_lost_chunks;
 	atomic_t s_mb_preallocated;
 	atomic_t s_mb_discarded;
+	atomic_t s_lock_busy;
 
 	/* locality groups */
 	struct ext4_locality_group *s_locality_groups;
@@ -1340,8 +1360,6 @@ extern void ext4_mb_free_blocks(handle_t *, struct inode *,
 		ext4_fsblk_t, unsigned long, int, unsigned long *);
 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);
@@ -1367,6 +1385,7 @@ extern int ext4_change_inode_journal_flag(struct inode *, int);
 extern int ext4_get_inode_loc(struct inode *, struct ext4_iloc *);
 extern int ext4_can_truncate(struct inode *inode);
 extern void ext4_truncate(struct inode *);
+extern int ext4_truncate_restart_trans(handle_t *, struct inode *, int nblocks);
 extern void ext4_set_inode_flags(struct inode *);
 extern void ext4_get_inode_flags(struct ext4_inode_info *);
 extern int ext4_alloc_da_blocks(struct inode *inode);
@@ -1575,15 +1594,18 @@ static inline void ext4_update_i_disksize(struct inode *inode, loff_t newsize)
 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;
+	ext4_grpblk_t	bb_first_free;	/* first free block */
+	ext4_grpblk_t	bb_free;	/* total free blocks */
+	ext4_grpblk_t	bb_fragments;	/* nr of freespace fragments */
 	struct          list_head bb_prealloc_list;
 #ifdef DOUBLE_CHECK
 	void            *bb_bitmap;
 #endif
 	struct rw_semaphore alloc_sem;
-	unsigned short  bb_counters[];
+	ext4_grpblk_t	bb_counters[];	/* Nr of free power-of-two-block
+					 * regions, index is order.
+					 * bb_counters[3] = 5 means
+					 * 5 free 8-block regions. */
 };
 
 #define EXT4_GROUP_INFO_NEED_INIT_BIT	0
@@ -1591,15 +1613,42 @@ struct ext4_group_info {
 #define EXT4_MB_GRP_NEED_INIT(grp)	\
 	(test_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &((grp)->bb_state)))
 
+#define EXT4_MAX_CONTENTION		8
+#define EXT4_CONTENTION_THRESHOLD	2
+
 static inline spinlock_t *ext4_group_lock_ptr(struct super_block *sb,
 					      ext4_group_t group)
 {
 	return bgl_lock_ptr(EXT4_SB(sb)->s_blockgroup_lock, group);
 }
 
+/*
+ * Returns true if the filesystem is busy enough that attempts to
+ * access the block group locks has run into contention.
+ */
+static inline int ext4_fs_is_busy(struct ext4_sb_info *sbi)
+{
+	return (atomic_read(&sbi->s_lock_busy) > EXT4_CONTENTION_THRESHOLD);
+}
+
 static inline void ext4_lock_group(struct super_block *sb, ext4_group_t group)
 {
-	spin_lock(ext4_group_lock_ptr(sb, group));
+	spinlock_t *lock = ext4_group_lock_ptr(sb, group);
+	if (spin_trylock(lock))
+		/*
+		 * We're able to grab the lock right away, so drop the
+		 * lock contention counter.
+		 */
+		atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, -1, 0);
+	else {
+		/*
+		 * The lock is busy, so bump the contention counter,
+		 * and then wait on the spin lock.
+		 */
+		atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, 1,
+				  EXT4_MAX_CONTENTION);
+		spin_lock(lock);
+	}
 }
 
 static inline void ext4_unlock_group(struct super_block *sb,
diff --git a/fs/ext4/ext4_extents.h b/fs/ext4/ext4_extents.h
index 20a84105a10..61652f1d15e 100644
--- a/fs/ext4/ext4_extents.h
+++ b/fs/ext4/ext4_extents.h
@@ -43,8 +43,7 @@
 #define CHECK_BINSEARCH__
 
 /*
- * If EXT_DEBUG is defined you can use the 'extdebug' mount option
- * to get lots of info about what's going on.
+ * Turn on EXT_DEBUG to get lots of info about extents operations.
  */
 #define EXT_DEBUG__
 #ifdef EXT_DEBUG
@@ -138,6 +137,7 @@ typedef int (*ext_prepare_callback)(struct inode *, struct ext4_ext_path *,
 #define EXT_BREAK      1
 #define EXT_REPEAT     2
 
+/* Maximum logical block in a file; ext4_extent's ee_block is __le32 */
 #define EXT_MAX_BLOCK	0xffffffff
 
 /*
diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c
index eb27fd0f2ee..6a9409920de 100644
--- a/fs/ext4/ext4_jbd2.c
+++ b/fs/ext4/ext4_jbd2.c
@@ -44,7 +44,7 @@ int __ext4_journal_forget(const char *where, handle_t *handle,
 						  handle, err);
 	}
 	else
-		brelse(bh);
+		bforget(bh);
 	return err;
 }
 
@@ -60,7 +60,7 @@ int __ext4_journal_revoke(const char *where, handle_t *handle,
 						  handle, err);
 	}
 	else
-		brelse(bh);
+		bforget(bh);
 	return err;
 }
 
@@ -89,7 +89,10 @@ int __ext4_handle_dirty_metadata(const char *where, handle_t *handle,
 			ext4_journal_abort_handle(where, __func__, bh,
 						  handle, err);
 	} else {
-		mark_buffer_dirty(bh);
+		if (inode && bh)
+			mark_buffer_dirty_inode(bh, inode);
+		else
+			mark_buffer_dirty(bh);
 		if (inode && inode_needs_sync(inode)) {
 			sync_dirty_buffer(bh);
 			if (buffer_req(bh) && !buffer_uptodate(bh)) {
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 73ebfb44ad7..7a383257792 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -93,7 +93,9 @@ static void ext4_idx_store_pblock(struct ext4_extent_idx *ix, ext4_fsblk_t pb)
 	ix->ei_leaf_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff);
 }
 
-static int ext4_ext_journal_restart(handle_t *handle, int needed)
+static int ext4_ext_truncate_extend_restart(handle_t *handle,
+					    struct inode *inode,
+					    int needed)
 {
 	int err;
 
@@ -104,7 +106,14 @@ static int ext4_ext_journal_restart(handle_t *handle, int needed)
 	err = ext4_journal_extend(handle, needed);
 	if (err <= 0)
 		return err;
-	return ext4_journal_restart(handle, needed);
+	err = ext4_truncate_restart_trans(handle, inode, needed);
+	/*
+	 * We have dropped i_data_sem so someone might have cached again
+	 * an extent we are going to truncate.
+	 */
+	ext4_ext_invalidate_cache(inode);
+
+	return err;
 }
 
 /*
@@ -220,57 +229,65 @@ ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
 	return newblock;
 }
 
-static int ext4_ext_space_block(struct inode *inode)
+static inline int ext4_ext_space_block(struct inode *inode, int check)
 {
 	int size;
 
 	size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
 			/ sizeof(struct ext4_extent);
+	if (!check) {
 #ifdef AGGRESSIVE_TEST
-	if (size > 6)
-		size = 6;
+		if (size > 6)
+			size = 6;
 #endif
+	}
 	return size;
 }
 
-static int ext4_ext_space_block_idx(struct inode *inode)
+static inline int ext4_ext_space_block_idx(struct inode *inode, int check)
 {
 	int size;
 
 	size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
 			/ sizeof(struct ext4_extent_idx);
+	if (!check) {
 #ifdef AGGRESSIVE_TEST
-	if (size > 5)
-		size = 5;
+		if (size > 5)
+			size = 5;
 #endif
+	}
 	return size;
 }
 
-static int ext4_ext_space_root(struct inode *inode)
+static inline int ext4_ext_space_root(struct inode *inode, int check)
 {
 	int size;
 
 	size = sizeof(EXT4_I(inode)->i_data);
 	size -= sizeof(struct ext4_extent_header);
 	size /= sizeof(struct ext4_extent);
+	if (!check) {
 #ifdef AGGRESSIVE_TEST
-	if (size > 3)
-		size = 3;
+		if (size > 3)
+			size = 3;
 #endif
+	}
 	return size;
 }
 
-static int ext4_ext_space_root_idx(struct inode *inode)
+static inline int ext4_ext_space_root_idx(struct inode *inode, int check)
 {
 	int size;
 
 	size = sizeof(EXT4_I(inode)->i_data);
 	size -= sizeof(struct ext4_extent_header);
 	size /= sizeof(struct ext4_extent_idx);
+	if (!check) {
 #ifdef AGGRESSIVE_TEST
-	if (size > 4)
-		size = 4;
+		if (size > 4)
+			size = 4;
 #endif
+	}
 	return size;
 }
 
@@ -284,9 +301,9 @@ int ext4_ext_calc_metadata_amount(struct inode *inode, int blocks)
 	int lcap, icap, rcap, leafs, idxs, num;
 	int newextents = blocks;
 
-	rcap = ext4_ext_space_root_idx(inode);
-	lcap = ext4_ext_space_block(inode);
-	icap = ext4_ext_space_block_idx(inode);
+	rcap = ext4_ext_space_root_idx(inode, 0);
+	lcap = ext4_ext_space_block(inode, 0);
+	icap = ext4_ext_space_block_idx(inode, 0);
 
 	/* number of new leaf blocks needed */
 	num = leafs = (newextents + lcap - 1) / lcap;
@@ -311,14 +328,14 @@ ext4_ext_max_entries(struct inode *inode, int depth)
 
 	if (depth == ext_depth(inode)) {
 		if (depth == 0)
-			max = ext4_ext_space_root(inode);
+			max = ext4_ext_space_root(inode, 1);
 		else
-			max = ext4_ext_space_root_idx(inode);
+			max = ext4_ext_space_root_idx(inode, 1);
 	} else {
 		if (depth == 0)
-			max = ext4_ext_space_block(inode);
+			max = ext4_ext_space_block(inode, 1);
 		else
-			max = ext4_ext_space_block_idx(inode);
+			max = ext4_ext_space_block_idx(inode, 1);
 	}
 
 	return max;
@@ -437,8 +454,9 @@ static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
 		  ext_debug("  %d->%llu", le32_to_cpu(path->p_idx->ei_block),
 			    idx_pblock(path->p_idx));
 		} else if (path->p_ext) {
-			ext_debug("  %d:%d:%llu ",
+			ext_debug("  %d:[%d]%d:%llu ",
 				  le32_to_cpu(path->p_ext->ee_block),
+				  ext4_ext_is_uninitialized(path->p_ext),
 				  ext4_ext_get_actual_len(path->p_ext),
 				  ext_pblock(path->p_ext));
 		} else
@@ -460,8 +478,11 @@ static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
 	eh = path[depth].p_hdr;
 	ex = EXT_FIRST_EXTENT(eh);
 
+	ext_debug("Displaying leaf extents for inode %lu\n", inode->i_ino);
+
 	for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
-		ext_debug("%d:%d:%llu ", le32_to_cpu(ex->ee_block),
+		ext_debug("%d:[%d]%d:%llu ", le32_to_cpu(ex->ee_block),
+			  ext4_ext_is_uninitialized(ex),
 			  ext4_ext_get_actual_len(ex), ext_pblock(ex));
 	}
 	ext_debug("\n");
@@ -580,9 +601,10 @@ ext4_ext_binsearch(struct inode *inode,
 	}
 
 	path->p_ext = l - 1;
-	ext_debug("  -> %d:%llu:%d ",
+	ext_debug("  -> %d:%llu:[%d]%d ",
 			le32_to_cpu(path->p_ext->ee_block),
 			ext_pblock(path->p_ext),
+			ext4_ext_is_uninitialized(path->p_ext),
 			ext4_ext_get_actual_len(path->p_ext));
 
 #ifdef CHECK_BINSEARCH
@@ -612,7 +634,7 @@ int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
 	eh->eh_depth = 0;
 	eh->eh_entries = 0;
 	eh->eh_magic = EXT4_EXT_MAGIC;
-	eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode));
+	eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode, 0));
 	ext4_mark_inode_dirty(handle, inode);
 	ext4_ext_invalidate_cache(inode);
 	return 0;
@@ -837,7 +859,7 @@ static int ext4_ext_split(handle_t *handle, struct inode *inode,
 
 	neh = ext_block_hdr(bh);
 	neh->eh_entries = 0;
-	neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode));
+	neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
 	neh->eh_magic = EXT4_EXT_MAGIC;
 	neh->eh_depth = 0;
 	ex = EXT_FIRST_EXTENT(neh);
@@ -850,9 +872,10 @@ static int ext4_ext_split(handle_t *handle, struct inode *inode,
 	path[depth].p_ext++;
 	while (path[depth].p_ext <=
 			EXT_MAX_EXTENT(path[depth].p_hdr)) {
-		ext_debug("move %d:%llu:%d in new leaf %llu\n",
+		ext_debug("move %d:%llu:[%d]%d in new leaf %llu\n",
 				le32_to_cpu(path[depth].p_ext->ee_block),
 				ext_pblock(path[depth].p_ext),
+				ext4_ext_is_uninitialized(path[depth].p_ext),
 				ext4_ext_get_actual_len(path[depth].p_ext),
 				newblock);
 		/*memmove(ex++, path[depth].p_ext++,
@@ -912,7 +935,7 @@ static int ext4_ext_split(handle_t *handle, struct inode *inode,
 		neh = ext_block_hdr(bh);
 		neh->eh_entries = cpu_to_le16(1);
 		neh->eh_magic = EXT4_EXT_MAGIC;
-		neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode));
+		neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
 		neh->eh_depth = cpu_to_le16(depth - i);
 		fidx = EXT_FIRST_INDEX(neh);
 		fidx->ei_block = border;
@@ -1037,9 +1060,9 @@ static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
 	/* old root could have indexes or leaves
 	 * so calculate e_max right way */
 	if (ext_depth(inode))
-	  neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode));
+		neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
 	else
-	  neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode));
+		neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
 	neh->eh_magic = EXT4_EXT_MAGIC;
 	set_buffer_uptodate(bh);
 	unlock_buffer(bh);
@@ -1054,7 +1077,7 @@ static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
 		goto out;
 
 	curp->p_hdr->eh_magic = EXT4_EXT_MAGIC;
-	curp->p_hdr->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode));
+	curp->p_hdr->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode, 0));
 	curp->p_hdr->eh_entries = cpu_to_le16(1);
 	curp->p_idx = EXT_FIRST_INDEX(curp->p_hdr);
 
@@ -1580,9 +1603,11 @@ int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
 
 	/* try to insert block into found extent and return */
 	if (ex && ext4_can_extents_be_merged(inode, ex, newext)) {
-		ext_debug("append %d block to %d:%d (from %llu)\n",
+		ext_debug("append [%d]%d block to %d:[%d]%d (from %llu)\n",
+				ext4_ext_is_uninitialized(newext),
 				ext4_ext_get_actual_len(newext),
 				le32_to_cpu(ex->ee_block),
+				ext4_ext_is_uninitialized(ex),
 				ext4_ext_get_actual_len(ex), ext_pblock(ex));
 		err = ext4_ext_get_access(handle, inode, path + depth);
 		if (err)
@@ -1651,9 +1676,10 @@ has_space:
 
 	if (!nearex) {
 		/* there is no extent in this leaf, create first one */
-		ext_debug("first extent in the leaf: %d:%llu:%d\n",
+		ext_debug("first extent in the leaf: %d:%llu:[%d]%d\n",
 				le32_to_cpu(newext->ee_block),
 				ext_pblock(newext),
+				ext4_ext_is_uninitialized(newext),
 				ext4_ext_get_actual_len(newext));
 		path[depth].p_ext = EXT_FIRST_EXTENT(eh);
 	} else if (le32_to_cpu(newext->ee_block)
@@ -1663,10 +1689,11 @@ has_space:
 			len = EXT_MAX_EXTENT(eh) - nearex;
 			len = (len - 1) * sizeof(struct ext4_extent);
 			len = len < 0 ? 0 : len;
-			ext_debug("insert %d:%llu:%d after: nearest 0x%p, "
+			ext_debug("insert %d:%llu:[%d]%d after: nearest 0x%p, "
 					"move %d from 0x%p to 0x%p\n",
 					le32_to_cpu(newext->ee_block),
 					ext_pblock(newext),
+					ext4_ext_is_uninitialized(newext),
 					ext4_ext_get_actual_len(newext),
 					nearex, len, nearex + 1, nearex + 2);
 			memmove(nearex + 2, nearex + 1, len);
@@ -1676,10 +1703,11 @@ has_space:
 		BUG_ON(newext->ee_block == nearex->ee_block);
 		len = (EXT_MAX_EXTENT(eh) - nearex) * sizeof(struct ext4_extent);
 		len = len < 0 ? 0 : len;
-		ext_debug("insert %d:%llu:%d before: nearest 0x%p, "
+		ext_debug("insert %d:%llu:[%d]%d before: nearest 0x%p, "
 				"move %d from 0x%p to 0x%p\n",
 				le32_to_cpu(newext->ee_block),
 				ext_pblock(newext),
+				ext4_ext_is_uninitialized(newext),
 				ext4_ext_get_actual_len(newext),
 				nearex, len, nearex + 1, nearex + 2);
 		memmove(nearex + 1, nearex, len);
@@ -2094,7 +2122,8 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
 		else
 			uninitialized = 0;
 
-		ext_debug("remove ext %lu:%u\n", ex_ee_block, ex_ee_len);
+		ext_debug("remove ext %u:[%d]%d\n", ex_ee_block,
+			 uninitialized, ex_ee_len);
 		path[depth].p_ext = ex;
 
 		a = ex_ee_block > start ? ex_ee_block : start;
@@ -2138,7 +2167,7 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
 		}
 		credits += 2 * EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb);
 
-		err = ext4_ext_journal_restart(handle, credits);
+		err = ext4_ext_truncate_extend_restart(handle, inode, credits);
 		if (err)
 			goto out;
 
@@ -2327,7 +2356,7 @@ static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
 		if (err == 0) {
 			ext_inode_hdr(inode)->eh_depth = 0;
 			ext_inode_hdr(inode)->eh_max =
-				cpu_to_le16(ext4_ext_space_root(inode));
+				cpu_to_le16(ext4_ext_space_root(inode, 0));
 			err = ext4_ext_dirty(handle, inode, path);
 		}
 	}
@@ -2743,6 +2772,7 @@ insert:
 	} else if (err)
 		goto fix_extent_len;
 out:
+	ext4_ext_show_leaf(inode, path);
 	return err ? err : allocated;
 
 fix_extent_len:
@@ -2786,7 +2816,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	struct ext4_allocation_request ar;
 
 	__clear_bit(BH_New, &bh_result->b_state);
-	ext_debug("blocks %u/%u requested for inode %u\n",
+	ext_debug("blocks %u/%u requested for inode %lu\n",
 			iblock, max_blocks, inode->i_ino);
 
 	/* check in cache */
@@ -2849,7 +2879,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 			newblock = iblock - ee_block + ee_start;
 			/* number of remaining blocks in the extent */
 			allocated = ee_len - (iblock - ee_block);
-			ext_debug("%u fit into %lu:%d -> %llu\n", iblock,
+			ext_debug("%u fit into %u:%d -> %llu\n", iblock,
 					ee_block, ee_len, newblock);
 
 			/* Do not put uninitialized extent in the cache */
@@ -2950,7 +2980,7 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
 	newblock = ext4_mb_new_blocks(handle, &ar, &err);
 	if (!newblock)
 		goto out2;
-	ext_debug("allocate new block: goal %llu, found %llu/%lu\n",
+	ext_debug("allocate new block: goal %llu, found %llu/%u\n",
 		  ar.goal, newblock, allocated);
 
 	/* try to insert new extent into found leaf and return */
diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 3f1873fef1c..5ca3eca70a1 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -58,10 +58,7 @@ static ssize_t
 ext4_file_write(struct kiocb *iocb, const struct iovec *iov,
 		unsigned long nr_segs, loff_t pos)
 {
-	struct file *file = iocb->ki_filp;
-	struct inode *inode = file->f_path.dentry->d_inode;
-	ssize_t ret;
-	int err;
+	struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode;
 
 	/*
 	 * If we have encountered a bitmap-format file, the size limit
@@ -81,53 +78,7 @@ ext4_file_write(struct kiocb *iocb, const struct iovec *iov,
 		}
 	}
 
-	ret = generic_file_aio_write(iocb, iov, nr_segs, pos);
-	/*
-	 * Skip flushing if there was an error, or if nothing was written.
-	 */
-	if (ret <= 0)
-		return ret;
-
-	/*
-	 * If the inode is IS_SYNC, or is O_SYNC and we are doing data
-	 * journalling then we need to make sure that we force the transaction
-	 * to disk to keep all metadata uptodate synchronously.
-	 */
-	if (file->f_flags & O_SYNC) {
-		/*
-		 * If we are non-data-journaled, then the dirty data has
-		 * already been flushed to backing store by generic_osync_inode,
-		 * and the inode has been flushed too if there have been any
-		 * modifications other than mere timestamp updates.
-		 *
-		 * Open question --- do we care about flushing timestamps too
-		 * if the inode is IS_SYNC?
-		 */
-		if (!ext4_should_journal_data(inode))
-			return ret;
-
-		goto force_commit;
-	}
-
-	/*
-	 * So we know that there has been no forced data flush.  If the inode
-	 * is marked IS_SYNC, we need to force one ourselves.
-	 */
-	if (!IS_SYNC(inode))
-		return ret;
-
-	/*
-	 * Open question #2 --- should we force data to disk here too?  If we
-	 * don't, the only impact is that data=writeback filesystems won't
-	 * flush data to disk automatically on IS_SYNC, only metadata (but
-	 * historically, that is what ext2 has done.)
-	 */
-
-force_commit:
-	err = ext4_force_commit(inode->i_sb);
-	if (err)
-		return err;
-	return ret;
+	return generic_file_aio_write(iocb, iov, nr_segs, pos);
 }
 
 static struct vm_operations_struct ext4_file_vm_ops = {
@@ -207,7 +158,7 @@ const struct inode_operations ext4_file_inode_operations = {
 	.listxattr	= ext4_listxattr,
 	.removexattr	= generic_removexattr,
 #endif
-	.permission	= ext4_permission,
+	.check_acl	= ext4_check_acl,
 	.fallocate	= ext4_fallocate,
 	.fiemap		= ext4_fiemap,
 };
diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c
index 83cf6415f59..07475740b51 100644
--- a/fs/ext4/fsync.c
+++ b/fs/ext4/fsync.c
@@ -50,7 +50,7 @@ int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
 {
 	struct inode *inode = dentry->d_inode;
 	journal_t *journal = EXT4_SB(inode->i_sb)->s_journal;
-	int ret = 0;
+	int err, ret = 0;
 
 	J_ASSERT(ext4_journal_current_handle() == NULL);
 
@@ -79,6 +79,9 @@ int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
 		goto out;
 	}
 
+	if (!journal)
+		ret = sync_mapping_buffers(inode->i_mapping);
+
 	if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
 		goto out;
 
@@ -91,10 +94,12 @@ int ext4_sync_file(struct file *file, struct dentry *dentry, int datasync)
 			.sync_mode = WB_SYNC_ALL,
 			.nr_to_write = 0, /* sys_fsync did this */
 		};
-		ret = sync_inode(inode, &wbc);
-		if (journal && (journal->j_flags & JBD2_BARRIER))
-			blkdev_issue_flush(inode->i_sb->s_bdev, NULL);
+		err = sync_inode(inode, &wbc);
+		if (ret == 0)
+			ret = err;
 	}
 out:
+	if (journal && (journal->j_flags & JBD2_BARRIER))
+		blkdev_issue_flush(inode->i_sb->s_bdev, NULL);
 	return ret;
 }
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 29e6dc7299b..f3624ead4f6 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -1189,7 +1189,7 @@ 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, ext4_free_inodes_count(sb, gdp), x);
+			(unsigned long) i, ext4_free_inodes_count(sb, gdp), x);
 		bitmap_count += x;
 	}
 	brelse(bitmap_bh);
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index f9c642b22ef..3a798737e30 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -192,11 +192,24 @@ static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
  * so before we call here everything must be consistently dirtied against
  * this transaction.
  */
-static int ext4_journal_test_restart(handle_t *handle, struct inode *inode)
+ int ext4_truncate_restart_trans(handle_t *handle, struct inode *inode,
+				 int nblocks)
 {
+	int ret;
+
+	/*
+	 * Drop i_data_sem to avoid deadlock with ext4_get_blocks At this
+	 * moment, get_block can be called only for blocks inside i_size since
+	 * page cache has been already dropped and writes are blocked by
+	 * i_mutex. So we can safely drop the i_data_sem here.
+	 */
 	BUG_ON(EXT4_JOURNAL(inode) == NULL);
 	jbd_debug(2, "restarting handle %p\n", handle);
-	return ext4_journal_restart(handle, blocks_for_truncate(inode));
+	up_write(&EXT4_I(inode)->i_data_sem);
+	ret = ext4_journal_restart(handle, blocks_for_truncate(inode));
+	down_write(&EXT4_I(inode)->i_data_sem);
+
+	return ret;
 }
 
 /*
@@ -341,9 +354,7 @@ static int ext4_block_to_path(struct inode *inode,
 	int n = 0;
 	int final = 0;
 
-	if (i_block < 0) {
-		ext4_warning(inode->i_sb, "ext4_block_to_path", "block < 0");
-	} else if (i_block < direct_blocks) {
+	if (i_block < direct_blocks) {
 		offsets[n++] = i_block;
 		final = direct_blocks;
 	} else if ((i_block -= direct_blocks) < indirect_blocks) {
@@ -551,15 +562,21 @@ static ext4_fsblk_t ext4_find_near(struct inode *inode, Indirect *ind)
  *
  *	Normally this function find the preferred place for block allocation,
  *	returns it.
+ *	Because this is only used for non-extent files, we limit the block nr
+ *	to 32 bits.
  */
 static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block,
 				   Indirect *partial)
 {
+	ext4_fsblk_t goal;
+
 	/*
 	 * XXX need to get goal block from mballoc's data structures
 	 */
 
-	return ext4_find_near(inode, partial);
+	goal = ext4_find_near(inode, partial);
+	goal = goal & EXT4_MAX_BLOCK_FILE_PHYS;
+	return goal;
 }
 
 /**
@@ -640,6 +657,8 @@ static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
 		if (*err)
 			goto failed_out;
 
+		BUG_ON(current_block + count > EXT4_MAX_BLOCK_FILE_PHYS);
+
 		target -= count;
 		/* allocate blocks for indirect blocks */
 		while (index < indirect_blks && count) {
@@ -674,6 +693,7 @@ static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
 		ar.flags = EXT4_MB_HINT_DATA;
 
 	current_block = ext4_mb_new_blocks(handle, &ar, err);
+	BUG_ON(current_block + ar.len > EXT4_MAX_BLOCK_FILE_PHYS);
 
 	if (*err && (target == blks)) {
 		/*
@@ -762,8 +782,9 @@ static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
 		BUFFER_TRACE(bh, "call get_create_access");
 		err = ext4_journal_get_create_access(handle, bh);
 		if (err) {
+			/* Don't brelse(bh) here; it's done in
+			 * ext4_journal_forget() below */
 			unlock_buffer(bh);
-			brelse(bh);
 			goto failed;
 		}
 
@@ -1109,16 +1130,15 @@ static void ext4_da_update_reserve_space(struct inode *inode, int used)
 		ext4_discard_preallocations(inode);
 }
 
-static int check_block_validity(struct inode *inode, sector_t logical,
-				sector_t phys, int len)
+static int check_block_validity(struct inode *inode, const char *msg,
+				sector_t logical, sector_t phys, int len)
 {
 	if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), phys, len)) {
-		ext4_error(inode->i_sb, "check_block_validity",
+		ext4_error(inode->i_sb, msg,
 			   "inode #%lu logical block %llu mapped to %llu "
 			   "(size %d)", inode->i_ino,
 			   (unsigned long long) logical,
 			   (unsigned long long) phys, len);
-		WARN_ON(1);
 		return -EIO;
 	}
 	return 0;
@@ -1170,8 +1190,8 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 	up_read((&EXT4_I(inode)->i_data_sem));
 
 	if (retval > 0 && buffer_mapped(bh)) {
-		int ret = check_block_validity(inode, block,
-					       bh->b_blocknr, retval);
+		int ret = check_block_validity(inode, "file system corruption",
+					       block, bh->b_blocknr, retval);
 		if (ret != 0)
 			return ret;
 	}
@@ -1235,8 +1255,7 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 			 * i_data's format changing.  Force the migrate
 			 * to fail by clearing migrate flags
 			 */
-			EXT4_I(inode)->i_flags = EXT4_I(inode)->i_flags &
-							~EXT4_EXT_MIGRATE;
+			EXT4_I(inode)->i_state &= ~EXT4_STATE_EXT_MIGRATE;
 		}
 	}
 
@@ -1252,8 +1271,9 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block,
 
 	up_write((&EXT4_I(inode)->i_data_sem));
 	if (retval > 0 && buffer_mapped(bh)) {
-		int ret = check_block_validity(inode, block,
-					       bh->b_blocknr, retval);
+		int ret = check_block_validity(inode, "file system "
+					       "corruption after allocation",
+					       block, bh->b_blocknr, retval);
 		if (ret != 0)
 			return ret;
 	}
@@ -1863,18 +1883,6 @@ static void ext4_da_page_release_reservation(struct page *page,
  * Delayed allocation stuff
  */
 
-struct mpage_da_data {
-	struct inode *inode;
-	sector_t b_blocknr;		/* start block number of extent */
-	size_t b_size;			/* size of extent */
-	unsigned long b_state;		/* state of the extent */
-	unsigned long first_page, next_page;	/* extent of pages */
-	struct writeback_control *wbc;
-	int io_done;
-	int pages_written;
-	int retval;
-};
-
 /*
  * mpage_da_submit_io - walks through extent of pages and try to write
  * them with writepage() call back
@@ -2329,7 +2337,7 @@ static int __mpage_da_writepage(struct page *page,
 		/*
 		 * Rest of the page in the page_vec
 		 * redirty then and skip then. We will
-		 * try to to write them again after
+		 * try to write them again after
 		 * starting a new transaction
 		 */
 		redirty_page_for_writepage(wbc, page);
@@ -2737,6 +2745,7 @@ static int ext4_da_writepages(struct address_space *mapping,
 	long pages_skipped;
 	int range_cyclic, cycled = 1, io_done = 0;
 	int needed_blocks, ret = 0, nr_to_writebump = 0;
+	loff_t range_start = wbc->range_start;
 	struct ext4_sb_info *sbi = EXT4_SB(mapping->host->i_sb);
 
 	trace_ext4_da_writepages(inode, wbc);
@@ -2850,6 +2859,7 @@ retry:
 			mpd.io_done = 1;
 			ret = MPAGE_DA_EXTENT_TAIL;
 		}
+		trace_ext4_da_write_pages(inode, &mpd);
 		wbc->nr_to_write -= mpd.pages_written;
 
 		ext4_journal_stop(handle);
@@ -2905,6 +2915,7 @@ out_writepages:
 	if (!no_nrwrite_index_update)
 		wbc->no_nrwrite_index_update = 0;
 	wbc->nr_to_write -= nr_to_writebump;
+	wbc->range_start = range_start;
 	trace_ext4_da_writepages_result(inode, wbc, ret, pages_written);
 	return ret;
 }
@@ -3117,6 +3128,8 @@ out:
  */
 int ext4_alloc_da_blocks(struct inode *inode)
 {
+	trace_ext4_alloc_da_blocks(inode);
+
 	if (!EXT4_I(inode)->i_reserved_data_blocks &&
 	    !EXT4_I(inode)->i_reserved_meta_blocks)
 		return 0;
@@ -3659,7 +3672,8 @@ static void ext4_clear_blocks(handle_t *handle, struct inode *inode,
 			ext4_handle_dirty_metadata(handle, inode, bh);
 		}
 		ext4_mark_inode_dirty(handle, inode);
-		ext4_journal_test_restart(handle, inode);
+		ext4_truncate_restart_trans(handle, inode,
+					    blocks_for_truncate(inode));
 		if (bh) {
 			BUFFER_TRACE(bh, "retaking write access");
 			ext4_journal_get_write_access(handle, bh);
@@ -3870,7 +3884,8 @@ static void ext4_free_branches(handle_t *handle, struct inode *inode,
 				return;
 			if (try_to_extend_transaction(handle, inode)) {
 				ext4_mark_inode_dirty(handle, inode);
-				ext4_journal_test_restart(handle, inode);
+				ext4_truncate_restart_trans(handle, inode,
+					    blocks_for_truncate(inode));
 			}
 
 			ext4_free_blocks(handle, inode, nr, 1, 1);
@@ -3958,8 +3973,7 @@ void ext4_truncate(struct inode *inode)
 	if (!ext4_can_truncate(inode))
 		return;
 
-	if (ei->i_disksize && inode->i_size == 0 &&
-	    !test_opt(inode->i_sb, NO_AUTO_DA_ALLOC))
+	if (inode->i_size == 0 && !test_opt(inode->i_sb, NO_AUTO_DA_ALLOC))
 		ei->i_state |= EXT4_STATE_DA_ALLOC_CLOSE;
 
 	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) {
@@ -4533,7 +4547,8 @@ static int ext4_inode_blocks_set(handle_t *handle,
  */
 static int ext4_do_update_inode(handle_t *handle,
 				struct inode *inode,
-				struct ext4_iloc *iloc)
+				struct ext4_iloc *iloc,
+				int do_sync)
 {
 	struct ext4_inode *raw_inode = ext4_raw_inode(iloc);
 	struct ext4_inode_info *ei = EXT4_I(inode);
@@ -4581,8 +4596,7 @@ static int ext4_do_update_inode(handle_t *handle,
 	if (ext4_inode_blocks_set(handle, raw_inode, ei))
 		goto out_brelse;
 	raw_inode->i_dtime = cpu_to_le32(ei->i_dtime);
-	/* clear the migrate flag in the raw_inode */
-	raw_inode->i_flags = cpu_to_le32(ei->i_flags & ~EXT4_EXT_MIGRATE);
+	raw_inode->i_flags = cpu_to_le32(ei->i_flags);
 	if (EXT4_SB(inode->i_sb)->s_es->s_creator_os !=
 	    cpu_to_le32(EXT4_OS_HURD))
 		raw_inode->i_file_acl_high =
@@ -4635,10 +4649,22 @@ 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_handle_dirty_metadata");
-	rc = ext4_handle_dirty_metadata(handle, inode, bh);
-	if (!err)
-		err = rc;
+	/*
+	 * If we're not using a journal and we were called from
+	 * ext4_write_inode() to sync the inode (making do_sync true),
+	 * we can just use sync_dirty_buffer() directly to do our dirty
+	 * work.  Testing s_journal here is a bit redundant but it's
+	 * worth it to avoid potential future trouble.
+	 */
+	if (EXT4_SB(inode->i_sb)->s_journal == NULL && do_sync) {
+		BUFFER_TRACE(bh, "call sync_dirty_buffer");
+		sync_dirty_buffer(bh);
+	} else {
+		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;
 
 out_brelse:
@@ -4684,19 +4710,32 @@ out_brelse:
  */
 int ext4_write_inode(struct inode *inode, int wait)
 {
+	int err;
+
 	if (current->flags & PF_MEMALLOC)
 		return 0;
 
-	if (ext4_journal_current_handle()) {
-		jbd_debug(1, "called recursively, non-PF_MEMALLOC!\n");
-		dump_stack();
-		return -EIO;
-	}
+	if (EXT4_SB(inode->i_sb)->s_journal) {
+		if (ext4_journal_current_handle()) {
+			jbd_debug(1, "called recursively, non-PF_MEMALLOC!\n");
+			dump_stack();
+			return -EIO;
+		}
 
-	if (!wait)
-		return 0;
+		if (!wait)
+			return 0;
+
+		err = ext4_force_commit(inode->i_sb);
+	} else {
+		struct ext4_iloc iloc;
 
-	return ext4_force_commit(inode->i_sb);
+		err = ext4_get_inode_loc(inode, &iloc);
+		if (err)
+			return err;
+		err = ext4_do_update_inode(EXT4_NOJOURNAL_HANDLE,
+					   inode, &iloc, wait);
+	}
+	return err;
 }
 
 /*
@@ -4990,7 +5029,7 @@ int ext4_mark_iloc_dirty(handle_t *handle,
 	get_bh(iloc->bh);
 
 	/* ext4_do_update_inode() does jbd2_journal_dirty_metadata */
-	err = ext4_do_update_inode(handle, inode, iloc);
+	err = ext4_do_update_inode(handle, inode, iloc, 0);
 	put_bh(iloc->bh);
 	return err;
 }
@@ -5281,12 +5320,21 @@ int ext4_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 	else
 		len = PAGE_CACHE_SIZE;
 
+	lock_page(page);
+	/*
+	 * return if we have all the buffers mapped. This avoid
+	 * the need to call write_begin/write_end which does a
+	 * journal_start/journal_stop which can block and take
+	 * long time
+	 */
 	if (page_has_buffers(page)) {
-		/* return if we have all the buffers mapped */
 		if (!walk_page_buffers(NULL, page_buffers(page), 0, len, NULL,
-				       ext4_bh_unmapped))
+					ext4_bh_unmapped)) {
+			unlock_page(page);
 			goto out_unlock;
+		}
 	}
+	unlock_page(page);
 	/*
 	 * OK, we need to fill the hole... Do write_begin write_end
 	 * to do block allocation/reservation.We are not holding
diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index 7050a9cd04a..c1cdf613e72 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -243,10 +243,9 @@ setversion_out:
 					me.donor_start, me.len, &me.moved_len);
 		fput(donor_filp);
 
-		if (!err)
-			if (copy_to_user((struct move_extent *)arg,
-				&me, sizeof(me)))
-				return -EFAULT;
+		if (copy_to_user((struct move_extent *)arg, &me, sizeof(me)))
+			return -EFAULT;
+
 		return err;
 	}
 
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index cd258463e2a..e9c61896d60 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -22,6 +22,7 @@
  */
 
 #include "mballoc.h"
+#include <linux/debugfs.h>
 #include <trace/events/ext4.h>
 
 /*
@@ -622,13 +623,13 @@ static int __mb_check_buddy(struct ext4_buddy *e4b, char *file,
 
 /* FIXME!! need more doc */
 static void ext4_mb_mark_free_simple(struct super_block *sb,
-				void *buddy, unsigned first, int len,
+				void *buddy, ext4_grpblk_t first, ext4_grpblk_t len,
 					struct ext4_group_info *grp)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	unsigned short min;
-	unsigned short max;
-	unsigned short chunk;
+	ext4_grpblk_t min;
+	ext4_grpblk_t max;
+	ext4_grpblk_t chunk;
 	unsigned short border;
 
 	BUG_ON(len > EXT4_BLOCKS_PER_GROUP(sb));
@@ -662,10 +663,10 @@ void ext4_mb_generate_buddy(struct super_block *sb,
 				void *buddy, void *bitmap, ext4_group_t group)
 {
 	struct ext4_group_info *grp = ext4_get_group_info(sb, group);
-	unsigned short max = EXT4_BLOCKS_PER_GROUP(sb);
-	unsigned short i = 0;
-	unsigned short first;
-	unsigned short len;
+	ext4_grpblk_t max = EXT4_BLOCKS_PER_GROUP(sb);
+	ext4_grpblk_t i = 0;
+	ext4_grpblk_t first;
+	ext4_grpblk_t len;
 	unsigned free = 0;
 	unsigned fragments = 0;
 	unsigned long long period = get_cycles();
@@ -743,7 +744,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 	char *data;
 	char *bitmap;
 
-	mb_debug("init page %lu\n", page->index);
+	mb_debug(1, "init page %lu\n", page->index);
 
 	inode = page->mapping->host;
 	sb = inode->i_sb;
@@ -822,7 +823,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 		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 %u\n", first_group + i);
+		mb_debug(1, "read bitmap for group %u\n", first_group + i);
 	}
 
 	/* wait for I/O completion */
@@ -862,12 +863,13 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 		if ((first_block + i) & 1) {
 			/* this is block of buddy */
 			BUG_ON(incore == NULL);
-			mb_debug("put buddy for group %u in page %lu/%x\n",
+			mb_debug(1, "put buddy for group %u in page %lu/%x\n",
 				group, page->index, i * blocksize);
 			grinfo = ext4_get_group_info(sb, group);
 			grinfo->bb_fragments = 0;
 			memset(grinfo->bb_counters, 0,
-			       sizeof(unsigned short)*(sb->s_blocksize_bits+2));
+			       sizeof(*grinfo->bb_counters) *
+				(sb->s_blocksize_bits+2));
 			/*
 			 * incore got set to the group block bitmap below
 			 */
@@ -878,7 +880,7 @@ static int ext4_mb_init_cache(struct page *page, char *incore)
 		} else {
 			/* this is block of bitmap */
 			BUG_ON(incore != NULL);
-			mb_debug("put bitmap for group %u in page %lu/%x\n",
+			mb_debug(1, "put bitmap for group %u in page %lu/%x\n",
 				group, page->index, i * blocksize);
 
 			/* see comments in ext4_mb_put_pa() */
@@ -908,6 +910,100 @@ out:
 	return err;
 }
 
+static noinline_for_stack
+int ext4_mb_init_group(struct super_block *sb, ext4_group_t group)
+{
+
+	int ret = 0;
+	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(1, "init group %u\n", group);
+	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
+	this_grp = ext4_get_group_info(sb, group);
+	/*
+	 * This ensures that we don't reinit the buddy cache
+	 * page which map to the group from which we are already
+	 * allocating. If we are looking at the buddy cache we would
+	 * have taken a reference using ext4_mb_load_buddy and that
+	 * would have taken the alloc_sem lock.
+	 */
+	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_load_buddy(struct super_block *sb, ext4_group_t group,
 					struct ext4_buddy *e4b)
@@ -922,7 +1018,7 @@ ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	struct inode *inode = sbi->s_buddy_cache;
 
-	mb_debug("load group %u\n", group);
+	mb_debug(1, "load group %u\n", group);
 
 	blocks_per_page = PAGE_CACHE_SIZE / sb->s_blocksize;
 	grp = ext4_get_group_info(sb, group);
@@ -941,8 +1037,26 @@ ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
 	 * groups mapped by the page is blocked
 	 * till we are done with allocation
 	 */
+repeat_load_buddy:
 	down_read(e4b->alloc_semp);
 
+	if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) {
+		/* we need to check for group need init flag
+		 * with alloc_semp held so that we can be sure
+		 * that new blocks didn't get added to the group
+		 * when we are loading the buddy cache
+		 */
+		up_read(e4b->alloc_semp);
+		/*
+		 * we need full data about the group
+		 * to make a good selection
+		 */
+		ret = ext4_mb_init_group(sb, group);
+		if (ret)
+			return ret;
+		goto repeat_load_buddy;
+	}
+
 	/*
 	 * the buddy cache inode stores the block bitmap
 	 * and buddy information in consecutive blocks.
@@ -1360,7 +1474,7 @@ static void ext4_mb_use_best_found(struct ext4_allocation_context *ac,
 	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)) {
+	if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) {
 		spin_lock(&sbi->s_md_lock);
 		sbi->s_mb_last_group = ac->ac_f_ex.fe_group;
 		sbi->s_mb_last_start = ac->ac_f_ex.fe_start;
@@ -1837,97 +1951,6 @@ void ext4_mb_put_buddy_cache_lock(struct super_block *sb,
 
 }
 
-static noinline_for_stack
-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)
 {
@@ -1938,11 +1961,14 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac)
 	struct ext4_sb_info *sbi;
 	struct super_block *sb;
 	struct ext4_buddy e4b;
-	loff_t size, isize;
 
 	sb = ac->ac_sb;
 	sbi = EXT4_SB(sb);
 	ngroups = ext4_get_groups_count(sb);
+	/* non-extent files are limited to low blocks/groups */
+	if (!(EXT4_I(ac->ac_inode)->i_flags & EXT4_EXTENTS_FL))
+		ngroups = sbi->s_blockfile_groups;
+
 	BUG_ON(ac->ac_status == AC_STATUS_FOUND);
 
 	/* first, try the goal */
@@ -1974,20 +2000,16 @@ ext4_mb_regular_allocator(struct ext4_allocation_context *ac)
 	}
 
 	bsbits = ac->ac_sb->s_blocksize_bits;
-	/* if stream allocation is enabled, use global goal */
-	size = ac->ac_o_ex.fe_logical + ac->ac_o_ex.fe_len;
-	isize = i_size_read(ac->ac_inode) >> bsbits;
-	if (size < isize)
-		size = isize;
 
-	if (size < sbi->s_mb_stream_request &&
-			(ac->ac_flags & EXT4_MB_HINT_DATA)) {
+	/* if stream allocation is enabled, use global goal */
+	if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) {
 		/* TBD: may be hot point */
 		spin_lock(&sbi->s_md_lock);
 		ac->ac_g_ex.fe_group = sbi->s_mb_last_group;
 		ac->ac_g_ex.fe_start = sbi->s_mb_last_start;
 		spin_unlock(&sbi->s_md_lock);
 	}
+
 	/* Let's just scan groups to find more-less suitable blocks */
 	cr = ac->ac_2order ? 0 : 1;
 	/*
@@ -2015,27 +2037,6 @@ repeat:
 			if (grp->bb_free == 0)
 				continue;
 
-			/*
-			 * if the group is already init we check whether it is
-			 * a good group and if not we don't load the buddy
-			 */
-			if (EXT4_MB_GRP_NEED_INIT(grp)) {
-				/*
-				 * we need full data about the group
-				 * to make a good selection
-				 */
-				err = ext4_mb_init_group(sb, group);
-				if (err)
-					goto out;
-			}
-
-			/*
-			 * If the particular group doesn't satisfy our
-			 * criteria we continue with the next group
-			 */
-			if (!ext4_mb_good_group(ac, group, cr))
-				continue;
-
 			err = ext4_mb_load_buddy(sb, group, &e4b);
 			if (err)
 				goto out;
@@ -2156,7 +2157,7 @@ static int ext4_mb_seq_history_show(struct seq_file *seq, void *v)
 
 	if (v == SEQ_START_TOKEN) {
 		seq_printf(seq, "%-5s %-8s %-23s %-23s %-23s %-5s "
-				"%-5s %-2s %-5s %-5s %-5s %-6s\n",
+				"%-5s %-2s %-6s %-5s %-5s %-6s\n",
 			  "pid", "inode", "original", "goal", "result", "found",
 			   "grps", "cr", "flags", "merge", "tail", "broken");
 		return 0;
@@ -2164,7 +2165,7 @@ 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";
+			"0x%04x %-5s %-5u %-6u\n";
 		sprintf(buf2, "%u/%d/%u@%u", hs->result.fe_group,
 			hs->result.fe_start, hs->result.fe_len,
 			hs->result.fe_logical);
@@ -2205,7 +2206,7 @@ static void ext4_mb_seq_history_stop(struct seq_file *seq, void *v)
 {
 }
 
-static struct seq_operations ext4_mb_seq_history_ops = {
+static const struct seq_operations ext4_mb_seq_history_ops = {
 	.start  = ext4_mb_seq_history_start,
 	.next   = ext4_mb_seq_history_next,
 	.stop   = ext4_mb_seq_history_stop,
@@ -2287,7 +2288,7 @@ static ssize_t ext4_mb_seq_history_write(struct file *file,
 	return count;
 }
 
-static struct file_operations ext4_mb_seq_history_fops = {
+static const struct file_operations ext4_mb_seq_history_fops = {
 	.owner		= THIS_MODULE,
 	.open		= ext4_mb_seq_history_open,
 	.read		= seq_read,
@@ -2328,7 +2329,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v)
 	struct ext4_buddy e4b;
 	struct sg {
 		struct ext4_group_info info;
-		unsigned short counters[16];
+		ext4_grpblk_t counters[16];
 	} sg;
 
 	group--;
@@ -2366,7 +2367,7 @@ static void ext4_mb_seq_groups_stop(struct seq_file *seq, void *v)
 {
 }
 
-static struct seq_operations ext4_mb_seq_groups_ops = {
+static const struct seq_operations ext4_mb_seq_groups_ops = {
 	.start  = ext4_mb_seq_groups_start,
 	.next   = ext4_mb_seq_groups_next,
 	.stop   = ext4_mb_seq_groups_stop,
@@ -2387,7 +2388,7 @@ static int ext4_mb_seq_groups_open(struct inode *inode, struct file *file)
 
 }
 
-static struct file_operations ext4_mb_seq_groups_fops = {
+static const struct file_operations ext4_mb_seq_groups_fops = {
 	.owner		= THIS_MODULE,
 	.open		= ext4_mb_seq_groups_open,
 	.read		= seq_read,
@@ -2532,7 +2533,7 @@ int ext4_mb_add_groupinfo(struct super_block *sb, ext4_group_t group,
 
 	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;;
+	meta_group_info[i]->bb_free_root.rb_node = NULL;
 
 #ifdef DOUBLE_CHECK
 	{
@@ -2558,26 +2559,15 @@ exit_meta_group_info:
 	return -ENOMEM;
 } /* ext4_mb_add_groupinfo */
 
-/*
- * Update an existing group.
- * This function is used for online resize
- */
-void ext4_mb_update_group_info(struct ext4_group_info *grp, ext4_grpblk_t add)
-{
-	grp->bb_free += add;
-}
-
 static int ext4_mb_init_backend(struct super_block *sb)
 {
 	ext4_group_t ngroups = ext4_get_groups_count(sb);
 	ext4_group_t i;
-	int metalen;
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	struct ext4_super_block *es = sbi->s_es;
 	int num_meta_group_infos;
 	int num_meta_group_infos_max;
 	int array_size;
-	struct ext4_group_info **meta_group_info;
 	struct ext4_group_desc *desc;
 
 	/* This is the number of blocks used by GDT */
@@ -2622,22 +2612,6 @@ static int ext4_mb_init_backend(struct super_block *sb)
 		goto err_freesgi;
 	}
 	EXT4_I(sbi->s_buddy_cache)->i_disksize = 0;
-
-	metalen = sizeof(*meta_group_info) << EXT4_DESC_PER_BLOCK_BITS(sb);
-	for (i = 0; i < num_meta_group_infos; i++) {
-		if ((i + 1) == num_meta_group_infos)
-			metalen = sizeof(*meta_group_info) *
-				(ngroups -
-					(i << EXT4_DESC_PER_BLOCK_BITS(sb)));
-		meta_group_info = kmalloc(metalen, GFP_KERNEL);
-		if (meta_group_info == NULL) {
-			printk(KERN_ERR "EXT4-fs: can't allocate mem for a "
-			       "buddy group\n");
-			goto err_freemeta;
-		}
-		sbi->s_group_info[i] = meta_group_info;
-	}
-
 	for (i = 0; i < ngroups; i++) {
 		desc = ext4_get_group_desc(sb, i, NULL);
 		if (desc == NULL) {
@@ -2655,7 +2629,6 @@ err_freebuddy:
 	while (i-- > 0)
 		kfree(ext4_get_group_info(sb, i));
 	i = num_meta_group_infos;
-err_freemeta:
 	while (i-- > 0)
 		kfree(sbi->s_group_info[i]);
 	iput(sbi->s_buddy_cache);
@@ -2672,14 +2645,14 @@ int ext4_mb_init(struct super_block *sb, int needs_recovery)
 	unsigned max;
 	int ret;
 
-	i = (sb->s_blocksize_bits + 2) * sizeof(unsigned short);
+	i = (sb->s_blocksize_bits + 2) * sizeof(*sbi->s_mb_offsets);
 
 	sbi->s_mb_offsets = kmalloc(i, GFP_KERNEL);
 	if (sbi->s_mb_offsets == NULL) {
 		return -ENOMEM;
 	}
 
-	i = (sb->s_blocksize_bits + 2) * sizeof(unsigned int);
+	i = (sb->s_blocksize_bits + 2) * sizeof(*sbi->s_mb_maxs);
 	sbi->s_mb_maxs = kmalloc(i, GFP_KERNEL);
 	if (sbi->s_mb_maxs == NULL) {
 		kfree(sbi->s_mb_offsets);
@@ -2758,7 +2731,7 @@ static void ext4_mb_cleanup_pa(struct ext4_group_info *grp)
 		kmem_cache_free(ext4_pspace_cachep, pa);
 	}
 	if (count)
-		mb_debug("mballoc: %u PAs left\n", count);
+		mb_debug(1, "mballoc: %u PAs left\n", count);
 
 }
 
@@ -2839,7 +2812,7 @@ 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 %u (0x%p):",
+		mb_debug(1, "gonna free %u blocks in group %u (0x%p):",
 			 entry->count, entry->group, entry);
 
 		err = ext4_mb_load_buddy(sb, entry->group, &e4b);
@@ -2874,9 +2847,43 @@ static void release_blocks_on_commit(journal_t *journal, transaction_t *txn)
 		ext4_mb_release_desc(&e4b);
 	}
 
-	mb_debug("freed %u blocks in %u structures\n", count, count2);
+	mb_debug(1, "freed %u blocks in %u structures\n", count, count2);
+}
+
+#ifdef CONFIG_EXT4_DEBUG
+u8 mb_enable_debug __read_mostly;
+
+static struct dentry *debugfs_dir;
+static struct dentry *debugfs_debug;
+
+static void __init ext4_create_debugfs_entry(void)
+{
+	debugfs_dir = debugfs_create_dir("ext4", NULL);
+	if (debugfs_dir)
+		debugfs_debug = debugfs_create_u8("mballoc-debug",
+						  S_IRUGO | S_IWUSR,
+						  debugfs_dir,
+						  &mb_enable_debug);
+}
+
+static void ext4_remove_debugfs_entry(void)
+{
+	debugfs_remove(debugfs_debug);
+	debugfs_remove(debugfs_dir);
 }
 
+#else
+
+static void __init ext4_create_debugfs_entry(void)
+{
+}
+
+static void ext4_remove_debugfs_entry(void)
+{
+}
+
+#endif
+
 int __init init_ext4_mballoc(void)
 {
 	ext4_pspace_cachep =
@@ -2904,6 +2911,7 @@ int __init init_ext4_mballoc(void)
 		kmem_cache_destroy(ext4_ac_cachep);
 		return -ENOMEM;
 	}
+	ext4_create_debugfs_entry();
 	return 0;
 }
 
@@ -2917,6 +2925,7 @@ void exit_ext4_mballoc(void)
 	kmem_cache_destroy(ext4_pspace_cachep);
 	kmem_cache_destroy(ext4_ac_cachep);
 	kmem_cache_destroy(ext4_free_ext_cachep);
+	ext4_remove_debugfs_entry();
 }
 
 
@@ -3061,7 +3070,7 @@ static void ext4_mb_normalize_group_request(struct ext4_allocation_context *ac)
 		ac->ac_g_ex.fe_len = EXT4_SB(sb)->s_stripe;
 	else
 		ac->ac_g_ex.fe_len = EXT4_SB(sb)->s_mb_group_prealloc;
-	mb_debug("#%u: goal %u blocks for locality group\n",
+	mb_debug(1, "#%u: goal %u blocks for locality group\n",
 		current->pid, ac->ac_g_ex.fe_len);
 }
 
@@ -3180,23 +3189,18 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
 		BUG_ON(!(ac->ac_o_ex.fe_logical >= pa_end ||
 			ac->ac_o_ex.fe_logical < pa->pa_lstart));
 
-		/* skip PA normalized request doesn't overlap with */
-		if (pa->pa_lstart >= end) {
-			spin_unlock(&pa->pa_lock);
-			continue;
-		}
-		if (pa_end <= start) {
+		/* skip PAs this normalized request doesn't overlap with */
+		if (pa->pa_lstart >= end || pa_end <= start) {
 			spin_unlock(&pa->pa_lock);
 			continue;
 		}
 		BUG_ON(pa->pa_lstart <= start && pa_end >= end);
 
+		/* adjust start or end to be adjacent to this pa */
 		if (pa_end <= ac->ac_o_ex.fe_logical) {
 			BUG_ON(pa_end < start);
 			start = pa_end;
-		}
-
-		if (pa->pa_lstart > ac->ac_o_ex.fe_logical) {
+		} else if (pa->pa_lstart > ac->ac_o_ex.fe_logical) {
 			BUG_ON(pa->pa_lstart > end);
 			end = pa->pa_lstart;
 		}
@@ -3251,7 +3255,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
 		ac->ac_flags |= EXT4_MB_HINT_TRY_GOAL;
 	}
 
-	mb_debug("goal: %u(was %u) blocks at %u\n", (unsigned) size,
+	mb_debug(1, "goal: %u(was %u) blocks at %u\n", (unsigned) size,
 		(unsigned) orig_size, (unsigned) start);
 }
 
@@ -3300,7 +3304,7 @@ static void ext4_mb_use_inode_pa(struct ext4_allocation_context *ac,
 	BUG_ON(pa->pa_free < len);
 	pa->pa_free -= len;
 
-	mb_debug("use %llu/%u from inode pa %p\n", start, len, pa);
+	mb_debug(1, "use %llu/%u from inode pa %p\n", start, len, pa);
 }
 
 /*
@@ -3324,7 +3328,7 @@ static void ext4_mb_use_group_pa(struct ext4_allocation_context *ac,
 	 * in on-disk bitmap -- see ext4_mb_release_context()
 	 * Other CPUs are prevented from allocating from this pa by lg_mutex
 	 */
-	mb_debug("use %u/%u from group pa %p\n", pa->pa_lstart-len, len, pa);
+	mb_debug(1, "use %u/%u from group pa %p\n", pa->pa_lstart-len, len, pa);
 }
 
 /*
@@ -3382,6 +3386,11 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
 			ac->ac_o_ex.fe_logical >= pa->pa_lstart + pa->pa_len)
 			continue;
 
+		/* non-extent files can't have physical blocks past 2^32 */
+		if (!(EXT4_I(ac->ac_inode)->i_flags & EXT4_EXTENTS_FL) &&
+			pa->pa_pstart + pa->pa_len > EXT4_MAX_BLOCK_FILE_PHYS)
+			continue;
+
 		/* found preallocated blocks, use them */
 		spin_lock(&pa->pa_lock);
 		if (pa->pa_deleted == 0 && pa->pa_free) {
@@ -3503,7 +3512,7 @@ void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap,
 		preallocated += len;
 		count++;
 	}
-	mb_debug("prellocated %u for group %u\n", preallocated, group);
+	mb_debug(1, "prellocated %u for group %u\n", preallocated, group);
 }
 
 static void ext4_mb_pa_callback(struct rcu_head *head)
@@ -3638,7 +3647,7 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
 	pa->pa_deleted = 0;
 	pa->pa_type = MB_INODE_PA;
 
-	mb_debug("new inode pa %p: %llu/%u for %u\n", pa,
+	mb_debug(1, "new inode pa %p: %llu/%u for %u\n", pa,
 			pa->pa_pstart, pa->pa_len, pa->pa_lstart);
 	trace_ext4_mb_new_inode_pa(ac, pa);
 
@@ -3698,7 +3707,7 @@ ext4_mb_new_group_pa(struct ext4_allocation_context *ac)
 	pa->pa_deleted = 0;
 	pa->pa_type = MB_GROUP_PA;
 
-	mb_debug("new group pa %p: %llu/%u for %u\n", pa,
+	mb_debug(1, "new group pa %p: %llu/%u for %u\n", pa,
 			pa->pa_pstart, pa->pa_len, pa->pa_lstart);
 	trace_ext4_mb_new_group_pa(ac, pa);
 
@@ -3777,7 +3786,7 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh,
 		next = mb_find_next_bit(bitmap_bh->b_data, end, bit);
 		start = group * EXT4_BLOCKS_PER_GROUP(sb) + bit +
 				le32_to_cpu(sbi->s_es->s_first_data_block);
-		mb_debug("    free preallocated %u/%u in group %u\n",
+		mb_debug(1, "    free preallocated %u/%u in group %u\n",
 				(unsigned) start, (unsigned) next - bit,
 				(unsigned) group);
 		free += next - bit;
@@ -3868,7 +3877,7 @@ ext4_mb_discard_group_preallocations(struct super_block *sb,
 	int busy = 0;
 	int free = 0;
 
-	mb_debug("discard preallocation for group %u\n", group);
+	mb_debug(1, "discard preallocation for group %u\n", group);
 
 	if (list_empty(&grp->bb_prealloc_list))
 		return 0;
@@ -3992,7 +4001,7 @@ void ext4_discard_preallocations(struct inode *inode)
 		return;
 	}
 
-	mb_debug("discard preallocation for inode %lu\n", inode->i_ino);
+	mb_debug(1, "discard preallocation for inode %lu\n", inode->i_ino);
 	trace_ext4_discard_preallocations(inode);
 
 	INIT_LIST_HEAD(&list);
@@ -4097,7 +4106,7 @@ static void ext4_mb_return_to_preallocation(struct inode *inode,
 {
 	BUG_ON(!list_empty(&EXT4_I(inode)->i_prealloc_list));
 }
-#ifdef MB_DEBUG
+#ifdef CONFIG_EXT4_DEBUG
 static void ext4_mb_show_ac(struct ext4_allocation_context *ac)
 {
 	struct super_block *sb = ac->ac_sb;
@@ -4139,14 +4148,14 @@ static void ext4_mb_show_ac(struct ext4_allocation_context *ac)
 			ext4_get_group_no_and_offset(sb, pa->pa_pstart,
 						     NULL, &start);
 			spin_unlock(&pa->pa_lock);
-			printk(KERN_ERR "PA:%lu:%d:%u \n", i,
-							start, pa->pa_len);
+			printk(KERN_ERR "PA:%u:%d:%u \n", i,
+			       start, pa->pa_len);
 		}
 		ext4_unlock_group(sb, i);
 
 		if (grp->bb_free == 0)
 			continue;
-		printk(KERN_ERR "%lu: %d/%d \n",
+		printk(KERN_ERR "%u: %d/%d \n",
 		       i, grp->bb_free, grp->bb_fragments);
 	}
 	printk(KERN_ERR "\n");
@@ -4174,16 +4183,26 @@ static void ext4_mb_group_or_file(struct ext4_allocation_context *ac)
 	if (!(ac->ac_flags & EXT4_MB_HINT_DATA))
 		return;
 
+	if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
+		return;
+
 	size = ac->ac_o_ex.fe_logical + ac->ac_o_ex.fe_len;
-	isize = i_size_read(ac->ac_inode) >> bsbits;
+	isize = (i_size_read(ac->ac_inode) + ac->ac_sb->s_blocksize - 1)
+		>> bsbits;
 	size = max(size, isize);
 
-	/* don't use group allocation for large files */
-	if (size >= sbi->s_mb_stream_request)
+	if ((size == isize) &&
+	    !ext4_fs_is_busy(sbi) &&
+	    (atomic_read(&ac->ac_inode->i_writecount) == 0)) {
+		ac->ac_flags |= EXT4_MB_HINT_NOPREALLOC;
 		return;
+	}
 
-	if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
+	/* don't use group allocation for large files */
+	if (size >= sbi->s_mb_stream_request) {
+		ac->ac_flags |= EXT4_MB_STREAM_ALLOC;
 		return;
+	}
 
 	BUG_ON(ac->ac_lg != NULL);
 	/*
@@ -4246,7 +4265,7 @@ ext4_mb_initialize_context(struct ext4_allocation_context *ac,
 	 * locality group. this is a policy, actually */
 	ext4_mb_group_or_file(ac);
 
-	mb_debug("init ac: %u blocks @ %u, goal %u, flags %x, 2^%d, "
+	mb_debug(1, "init ac: %u blocks @ %u, goal %u, flags %x, 2^%d, "
 			"left: %u/%u, right %u/%u to %swritable\n",
 			(unsigned) ar->len, (unsigned) ar->logical,
 			(unsigned) ar->goal, ac->ac_flags, ac->ac_2order,
@@ -4268,7 +4287,7 @@ ext4_mb_discard_lg_preallocations(struct super_block *sb,
 	struct ext4_prealloc_space *pa, *tmp;
 	struct ext4_allocation_context *ac;
 
-	mb_debug("discard locality group preallocation\n");
+	mb_debug(1, "discard locality group preallocation\n");
 
 	INIT_LIST_HEAD(&discard_list);
 	ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS);
diff --git a/fs/ext4/mballoc.h b/fs/ext4/mballoc.h
index c96bb19f58f..188d3d709b2 100644
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -37,11 +37,19 @@
 
 /*
  */
-#define MB_DEBUG__
-#ifdef MB_DEBUG
-#define mb_debug(fmt, a...)	printk(fmt, ##a)
+#ifdef CONFIG_EXT4_DEBUG
+extern u8 mb_enable_debug;
+
+#define mb_debug(n, fmt, a...)	                                        \
+	do {								\
+		if ((n) <= mb_enable_debug) {		        	\
+			printk(KERN_DEBUG "(%s, %d): %s: ",		\
+			       __FILE__, __LINE__, __func__);		\
+			printk(fmt, ## a);				\
+		}							\
+	} while (0)
 #else
-#define mb_debug(fmt, a...)
+#define mb_debug(n, fmt, a...)
 #endif
 
 /*
@@ -128,8 +136,8 @@ struct ext4_prealloc_space {
 	unsigned		pa_deleted;
 	ext4_fsblk_t		pa_pstart;	/* phys. block */
 	ext4_lblk_t		pa_lstart;	/* log. block */
-	unsigned short		pa_len;		/* len of preallocated chunk */
-	unsigned short		pa_free;	/* how many blocks are free */
+	ext4_grpblk_t		pa_len;		/* len of preallocated chunk */
+	ext4_grpblk_t		pa_free;	/* how many blocks are free */
 	unsigned short		pa_type;	/* pa type. inode or group */
 	spinlock_t		*pa_obj_lock;
 	struct inode		*pa_inode;	/* hack, for history only */
@@ -144,7 +152,7 @@ struct ext4_free_extent {
 	ext4_lblk_t fe_logical;
 	ext4_grpblk_t fe_start;
 	ext4_group_t fe_group;
-	int fe_len;
+	ext4_grpblk_t fe_len;
 };
 
 /*
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
index 313a50b3974..bf519f239ae 100644
--- a/fs/ext4/migrate.c
+++ b/fs/ext4/migrate.c
@@ -353,17 +353,16 @@ static int ext4_ext_swap_inode_data(handle_t *handle, struct inode *inode,
 
 	down_write(&EXT4_I(inode)->i_data_sem);
 	/*
-	 * if EXT4_EXT_MIGRATE is cleared a block allocation
+	 * if EXT4_STATE_EXT_MIGRATE is cleared a block allocation
 	 * happened after we started the migrate. We need to
 	 * fail the migrate
 	 */
-	if (!(EXT4_I(inode)->i_flags & EXT4_EXT_MIGRATE)) {
+	if (!(EXT4_I(inode)->i_state & EXT4_STATE_EXT_MIGRATE)) {
 		retval = -EAGAIN;
 		up_write(&EXT4_I(inode)->i_data_sem);
 		goto err_out;
 	} else
-		EXT4_I(inode)->i_flags = EXT4_I(inode)->i_flags &
-							~EXT4_EXT_MIGRATE;
+		EXT4_I(inode)->i_state &= ~EXT4_STATE_EXT_MIGRATE;
 	/*
 	 * We have the extent map build with the tmp inode.
 	 * Now copy the i_data across
@@ -517,14 +516,15 @@ int ext4_ext_migrate(struct inode *inode)
 	 * when we add extents we extent the journal
 	 */
 	/*
-	 * Even though we take i_mutex we can still cause block allocation
-	 * via mmap write to holes. If we have allocated new blocks we fail
-	 * migrate.  New block allocation will clear EXT4_EXT_MIGRATE flag.
-	 * The flag is updated with i_data_sem held to prevent racing with
-	 * block allocation.
+	 * Even though we take i_mutex we can still cause block
+	 * allocation via mmap write to holes. If we have allocated
+	 * new blocks we fail migrate.  New block allocation will
+	 * clear EXT4_STATE_EXT_MIGRATE flag.  The flag is updated
+	 * with i_data_sem held to prevent racing with block
+	 * allocation.
 	 */
 	down_read((&EXT4_I(inode)->i_data_sem));
-	EXT4_I(inode)->i_flags = EXT4_I(inode)->i_flags | EXT4_EXT_MIGRATE;
+	EXT4_I(inode)->i_state |= EXT4_STATE_EXT_MIGRATE;
 	up_read((&EXT4_I(inode)->i_data_sem));
 
 	handle = ext4_journal_start(inode, 1);
@@ -618,7 +618,7 @@ err_out:
 	tmp_inode->i_nlink = 0;
 
 	ext4_journal_stop(handle);
-
+	unlock_new_inode(tmp_inode);
 	iput(tmp_inode);
 
 	return retval;
diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index bbf2dd9404d..c07a2915e40 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -19,14 +19,31 @@
 #include "ext4_extents.h"
 #include "ext4.h"
 
-#define get_ext_path(path, inode, block, ret)		\
-	do {								\
-		path = ext4_ext_find_extent(inode, block, path);	\
-		if (IS_ERR(path)) {					\
-			ret = PTR_ERR(path);				\
-			path = NULL;					\
-		}							\
-	} while (0)
+/**
+ * get_ext_path - Find an extent path for designated logical block number.
+ *
+ * @inode:	an inode which is searched
+ * @lblock:	logical block number to find an extent path
+ * @path:	pointer to an extent path pointer (for output)
+ *
+ * ext4_ext_find_extent wrapper. Return 0 on success, or a negative error value
+ * on failure.
+ */
+static inline int
+get_ext_path(struct inode *inode, ext4_lblk_t lblock,
+		struct ext4_ext_path **path)
+{
+	int ret = 0;
+
+	*path = ext4_ext_find_extent(inode, lblock, *path);
+	if (IS_ERR(*path)) {
+		ret = PTR_ERR(*path);
+		*path = NULL;
+	} else if ((*path)[ext_depth(inode)].p_ext == NULL)
+		ret = -ENODATA;
+
+	return ret;
+}
 
 /**
  * copy_extent_status - Copy the extent's initialization status
@@ -113,6 +130,31 @@ mext_next_extent(struct inode *inode, struct ext4_ext_path *path,
 }
 
 /**
+ * mext_check_null_inode - NULL check for two inodes
+ *
+ * If inode1 or inode2 is NULL, return -EIO. Otherwise, return 0.
+ */
+static int
+mext_check_null_inode(struct inode *inode1, struct inode *inode2,
+		const char *function)
+{
+	int ret = 0;
+
+	if (inode1 == NULL) {
+		ext4_error(inode2->i_sb, function,
+			"Both inodes should not be NULL: "
+			"inode1 NULL inode2 %lu", inode2->i_ino);
+		ret = -EIO;
+	} else if (inode2 == NULL) {
+		ext4_error(inode1->i_sb, function,
+			"Both inodes should not be NULL: "
+			"inode1 %lu inode2 NULL", inode1->i_ino);
+		ret = -EIO;
+	}
+	return ret;
+}
+
+/**
  * mext_double_down_read - Acquire two inodes' read semaphore
  *
  * @orig_inode:		original inode structure
@@ -124,8 +166,6 @@ mext_double_down_read(struct inode *orig_inode, struct inode *donor_inode)
 {
 	struct inode *first = orig_inode, *second = donor_inode;
 
-	BUG_ON(orig_inode == NULL || donor_inode == NULL);
-
 	/*
 	 * Use the inode number to provide the stable locking order instead
 	 * of its address, because the C language doesn't guarantee you can
@@ -152,8 +192,6 @@ mext_double_down_write(struct inode *orig_inode, struct inode *donor_inode)
 {
 	struct inode *first = orig_inode, *second = donor_inode;
 
-	BUG_ON(orig_inode == NULL || donor_inode == NULL);
-
 	/*
 	 * Use the inode number to provide the stable locking order instead
 	 * of its address, because the C language doesn't guarantee you can
@@ -178,8 +216,6 @@ mext_double_down_write(struct inode *orig_inode, struct inode *donor_inode)
 static void
 mext_double_up_read(struct inode *orig_inode, struct inode *donor_inode)
 {
-	BUG_ON(orig_inode == NULL || donor_inode == NULL);
-
 	up_read(&EXT4_I(orig_inode)->i_data_sem);
 	up_read(&EXT4_I(donor_inode)->i_data_sem);
 }
@@ -194,8 +230,6 @@ mext_double_up_read(struct inode *orig_inode, struct inode *donor_inode)
 static void
 mext_double_up_write(struct inode *orig_inode, struct inode *donor_inode)
 {
-	BUG_ON(orig_inode == NULL || donor_inode == NULL);
-
 	up_write(&EXT4_I(orig_inode)->i_data_sem);
 	up_write(&EXT4_I(donor_inode)->i_data_sem);
 }
@@ -283,8 +317,8 @@ mext_insert_across_blocks(handle_t *handle, struct inode *orig_inode,
 	}
 
 	if (new_flag) {
-		get_ext_path(orig_path, orig_inode, eblock, err);
-		if (orig_path == NULL)
+		err = get_ext_path(orig_inode, eblock, &orig_path);
+		if (err)
 			goto out;
 
 		if (ext4_ext_insert_extent(handle, orig_inode,
@@ -293,9 +327,9 @@ mext_insert_across_blocks(handle_t *handle, struct inode *orig_inode,
 	}
 
 	if (end_flag) {
-		get_ext_path(orig_path, orig_inode,
-				      le32_to_cpu(end_ext->ee_block) - 1, err);
-		if (orig_path == NULL)
+		err = get_ext_path(orig_inode,
+				le32_to_cpu(end_ext->ee_block) - 1, &orig_path);
+		if (err)
 			goto out;
 
 		if (ext4_ext_insert_extent(handle, orig_inode,
@@ -519,7 +553,15 @@ mext_leaf_block(handle_t *handle, struct inode *orig_inode,
 	 * oext      |-----------|
 	 * new_ext       |-------|
 	 */
-	BUG_ON(le32_to_cpu(oext->ee_block) + oext_alen - 1 < new_ext_end);
+	if (le32_to_cpu(oext->ee_block) + oext_alen - 1 < new_ext_end) {
+		ext4_error(orig_inode->i_sb, __func__,
+			"new_ext_end(%u) should be less than or equal to "
+			"oext->ee_block(%u) + oext_alen(%d) - 1",
+			new_ext_end, le32_to_cpu(oext->ee_block),
+			oext_alen);
+		ret = -EIO;
+		goto out;
+	}
 
 	/*
 	 * Case: new_ext is smaller than original extent
@@ -543,6 +585,7 @@ mext_leaf_block(handle_t *handle, struct inode *orig_inode,
 
 	ret = mext_insert_extents(handle, orig_inode, orig_path, o_start,
 				o_end, &start_ext, &new_ext, &end_ext);
+out:
 	return ret;
 }
 
@@ -554,8 +597,10 @@ mext_leaf_block(handle_t *handle, struct inode *orig_inode,
  * @orig_off:		block offset of original inode
  * @donor_off:		block offset of donor inode
  * @max_count:		the maximun length of extents
+ *
+ * Return 0 on success, or a negative error value on failure.
  */
-static void
+static int
 mext_calc_swap_extents(struct ext4_extent *tmp_dext,
 			      struct ext4_extent *tmp_oext,
 			      ext4_lblk_t orig_off, ext4_lblk_t donor_off,
@@ -564,6 +609,19 @@ mext_calc_swap_extents(struct ext4_extent *tmp_dext,
 	ext4_lblk_t diff, orig_diff;
 	struct ext4_extent dext_old, oext_old;
 
+	BUG_ON(orig_off != donor_off);
+
+	/* original and donor extents have to cover the same block offset */
+	if (orig_off < le32_to_cpu(tmp_oext->ee_block) ||
+	    le32_to_cpu(tmp_oext->ee_block) +
+			ext4_ext_get_actual_len(tmp_oext) - 1 < orig_off)
+		return -ENODATA;
+
+	if (orig_off < le32_to_cpu(tmp_dext->ee_block) ||
+	    le32_to_cpu(tmp_dext->ee_block) +
+			ext4_ext_get_actual_len(tmp_dext) - 1 < orig_off)
+		return -ENODATA;
+
 	dext_old = *tmp_dext;
 	oext_old = *tmp_oext;
 
@@ -591,6 +649,8 @@ mext_calc_swap_extents(struct ext4_extent *tmp_dext,
 
 	copy_extent_status(&oext_old, tmp_dext);
 	copy_extent_status(&dext_old, tmp_oext);
+
+	return 0;
 }
 
 /**
@@ -631,13 +691,13 @@ mext_replace_branches(handle_t *handle, struct inode *orig_inode,
 	mext_double_down_write(orig_inode, donor_inode);
 
 	/* Get the original extent for the block "orig_off" */
-	get_ext_path(orig_path, orig_inode, orig_off, err);
-	if (orig_path == NULL)
+	err = get_ext_path(orig_inode, orig_off, &orig_path);
+	if (err)
 		goto out;
 
 	/* Get the donor extent for the head */
-	get_ext_path(donor_path, donor_inode, donor_off, err);
-	if (donor_path == NULL)
+	err = get_ext_path(donor_inode, donor_off, &donor_path);
+	if (err)
 		goto out;
 	depth = ext_depth(orig_inode);
 	oext = orig_path[depth].p_ext;
@@ -647,13 +707,28 @@ mext_replace_branches(handle_t *handle, struct inode *orig_inode,
 	dext = donor_path[depth].p_ext;
 	tmp_dext = *dext;
 
-	mext_calc_swap_extents(&tmp_dext, &tmp_oext, orig_off,
+	err = mext_calc_swap_extents(&tmp_dext, &tmp_oext, orig_off,
 				      donor_off, count);
+	if (err)
+		goto out;
 
 	/* Loop for the donor extents */
 	while (1) {
 		/* The extent for donor must be found. */
-		BUG_ON(!dext || donor_off != le32_to_cpu(tmp_dext.ee_block));
+		if (!dext) {
+			ext4_error(donor_inode->i_sb, __func__,
+				   "The extent for donor must be found");
+			err = -EIO;
+			goto out;
+		} else if (donor_off != le32_to_cpu(tmp_dext.ee_block)) {
+			ext4_error(donor_inode->i_sb, __func__,
+				"Donor offset(%u) and the first block of donor "
+				"extent(%u) should be equal",
+				donor_off,
+				le32_to_cpu(tmp_dext.ee_block));
+			err = -EIO;
+			goto out;
+		}
 
 		/* Set donor extent to orig extent */
 		err = mext_leaf_block(handle, orig_inode,
@@ -678,8 +753,8 @@ mext_replace_branches(handle_t *handle, struct inode *orig_inode,
 
 		if (orig_path)
 			ext4_ext_drop_refs(orig_path);
-		get_ext_path(orig_path, orig_inode, orig_off, err);
-		if (orig_path == NULL)
+		err = get_ext_path(orig_inode, orig_off, &orig_path);
+		if (err)
 			goto out;
 		depth = ext_depth(orig_inode);
 		oext = orig_path[depth].p_ext;
@@ -692,9 +767,8 @@ mext_replace_branches(handle_t *handle, struct inode *orig_inode,
 
 		if (donor_path)
 			ext4_ext_drop_refs(donor_path);
-		get_ext_path(donor_path, donor_inode,
-				      donor_off, err);
-		if (donor_path == NULL)
+		err = get_ext_path(donor_inode, donor_off, &donor_path);
+		if (err)
 			goto out;
 		depth = ext_depth(donor_inode);
 		dext = donor_path[depth].p_ext;
@@ -705,9 +779,10 @@ mext_replace_branches(handle_t *handle, struct inode *orig_inode,
 		}
 		tmp_dext = *dext;
 
-		mext_calc_swap_extents(&tmp_dext, &tmp_oext, orig_off,
-					      donor_off,
-					      count - replaced_count);
+		err = mext_calc_swap_extents(&tmp_dext, &tmp_oext, orig_off,
+					   donor_off, count - replaced_count);
+		if (err)
+			goto out;
 	}
 
 out:
@@ -740,7 +815,7 @@ out:
  * on success, or a negative error value on failure.
  */
 static int
-move_extent_par_page(struct file *o_filp, struct inode *donor_inode,
+move_extent_per_page(struct file *o_filp, struct inode *donor_inode,
 		  pgoff_t orig_page_offset, int data_offset_in_page,
 		  int block_len_in_page, int uninit)
 {
@@ -871,6 +946,7 @@ out:
 		if (PageLocked(page))
 			unlock_page(page);
 		page_cache_release(page);
+		ext4_journal_stop(handle);
 	}
 out2:
 	ext4_journal_stop(handle);
@@ -897,6 +973,10 @@ mext_check_arguments(struct inode *orig_inode,
 			  struct inode *donor_inode, __u64 orig_start,
 			  __u64 donor_start, __u64 *len, __u64 moved_len)
 {
+	ext4_lblk_t orig_blocks, donor_blocks;
+	unsigned int blkbits = orig_inode->i_blkbits;
+	unsigned int blocksize = 1 << blkbits;
+
 	/* Regular file check */
 	if (!S_ISREG(orig_inode->i_mode) || !S_ISREG(donor_inode->i_mode)) {
 		ext4_debug("ext4 move extent: The argument files should be "
@@ -960,54 +1040,58 @@ mext_check_arguments(struct inode *orig_inode,
 		return -EINVAL;
 	}
 
-	if ((orig_start > MAX_DEFRAG_SIZE) ||
-	    (donor_start > MAX_DEFRAG_SIZE) ||
-	    (*len > MAX_DEFRAG_SIZE) ||
-	    (orig_start + *len > MAX_DEFRAG_SIZE))  {
-		ext4_debug("ext4 move extent: Can't handle over [%lu] blocks "
-			"[ino:orig %lu, donor %lu]\n", MAX_DEFRAG_SIZE,
+	if ((orig_start > EXT_MAX_BLOCK) ||
+	    (donor_start > EXT_MAX_BLOCK) ||
+	    (*len > EXT_MAX_BLOCK) ||
+	    (orig_start + *len > EXT_MAX_BLOCK))  {
+		ext4_debug("ext4 move extent: Can't handle over [%u] blocks "
+			"[ino:orig %lu, donor %lu]\n", EXT_MAX_BLOCK,
 			orig_inode->i_ino, donor_inode->i_ino);
 		return -EINVAL;
 	}
 
 	if (orig_inode->i_size > donor_inode->i_size) {
-		if (orig_start >= donor_inode->i_size) {
+		donor_blocks = (donor_inode->i_size + blocksize - 1) >> blkbits;
+		/* TODO: eliminate this artificial restriction */
+		if (orig_start >= donor_blocks) {
 			ext4_debug("ext4 move extent: orig start offset "
-			"[%llu] should be less than donor file size "
-			"[%lld] [ino:orig %lu, donor_inode %lu]\n",
-			orig_start, donor_inode->i_size,
+			"[%llu] should be less than donor file blocks "
+			"[%u] [ino:orig %lu, donor %lu]\n",
+			orig_start, donor_blocks,
 			orig_inode->i_ino, donor_inode->i_ino);
 			return -EINVAL;
 		}
 
-		if (orig_start + *len > donor_inode->i_size) {
+		/* TODO: eliminate this artificial restriction */
+		if (orig_start + *len > donor_blocks) {
 			ext4_debug("ext4 move extent: End offset [%llu] should "
-				"be less than donor file size [%lld]."
-				"So adjust length from %llu to %lld "
+				"be less than donor file blocks [%u]."
+				"So adjust length from %llu to %llu "
 				"[ino:orig %lu, donor %lu]\n",
-				orig_start + *len, donor_inode->i_size,
-				*len, donor_inode->i_size - orig_start,
+				orig_start + *len, donor_blocks,
+				*len, donor_blocks - orig_start,
 				orig_inode->i_ino, donor_inode->i_ino);
-			*len = donor_inode->i_size - orig_start;
+			*len = donor_blocks - orig_start;
 		}
 	} else {
-		if (orig_start >= orig_inode->i_size) {
+		orig_blocks = (orig_inode->i_size + blocksize - 1) >> blkbits;
+		if (orig_start >= orig_blocks) {
 			ext4_debug("ext4 move extent: start offset [%llu] "
-				"should be less than original file size "
-				"[%lld] [inode:orig %lu, donor %lu]\n",
-				 orig_start, orig_inode->i_size,
+				"should be less than original file blocks "
+				"[%u] [ino:orig %lu, donor %lu]\n",
+				 orig_start, orig_blocks,
 				orig_inode->i_ino, donor_inode->i_ino);
 			return -EINVAL;
 		}
 
-		if (orig_start + *len > orig_inode->i_size) {
+		if (orig_start + *len > orig_blocks) {
 			ext4_debug("ext4 move extent: Adjust length "
-				"from %llu to %lld. Because it should be "
-				"less than original file size "
+				"from %llu to %llu. Because it should be "
+				"less than original file blocks "
 				"[ino:orig %lu, donor %lu]\n",
-				*len, orig_inode->i_size - orig_start,
+				*len, orig_blocks - orig_start,
 				orig_inode->i_ino, donor_inode->i_ino);
-			*len = orig_inode->i_size - orig_start;
+			*len = orig_blocks - orig_start;
 		}
 	}
 
@@ -1027,18 +1111,23 @@ mext_check_arguments(struct inode *orig_inode,
  * @inode1:	the inode structure
  * @inode2:	the inode structure
  *
- * Lock two inodes' i_mutex by i_ino order. This function is moved from
- * fs/inode.c.
+ * Lock two inodes' i_mutex by i_ino order.
+ * If inode1 or inode2 is NULL, return -EIO. Otherwise, return 0.
  */
-static void
+static int
 mext_inode_double_lock(struct inode *inode1, struct inode *inode2)
 {
-	if (inode1 == NULL || inode2 == NULL || inode1 == inode2) {
-		if (inode1)
-			mutex_lock(&inode1->i_mutex);
-		else if (inode2)
-			mutex_lock(&inode2->i_mutex);
-		return;
+	int ret = 0;
+
+	BUG_ON(inode1 == NULL && inode2 == NULL);
+
+	ret = mext_check_null_inode(inode1, inode2, __func__);
+	if (ret < 0)
+		goto out;
+
+	if (inode1 == inode2) {
+		mutex_lock(&inode1->i_mutex);
+		goto out;
 	}
 
 	if (inode1->i_ino < inode2->i_ino) {
@@ -1048,6 +1137,9 @@ mext_inode_double_lock(struct inode *inode1, struct inode *inode2)
 		mutex_lock_nested(&inode2->i_mutex, I_MUTEX_PARENT);
 		mutex_lock_nested(&inode1->i_mutex, I_MUTEX_CHILD);
 	}
+
+out:
+	return ret;
 }
 
 /**
@@ -1056,17 +1148,28 @@ mext_inode_double_lock(struct inode *inode1, struct inode *inode2)
  * @inode1:     the inode that is released first
  * @inode2:     the inode that is released second
  *
- * This function is moved from fs/inode.c.
+ * If inode1 or inode2 is NULL, return -EIO. Otherwise, return 0.
  */
 
-static void
+static int
 mext_inode_double_unlock(struct inode *inode1, struct inode *inode2)
 {
+	int ret = 0;
+
+	BUG_ON(inode1 == NULL && inode2 == NULL);
+
+	ret = mext_check_null_inode(inode1, inode2, __func__);
+	if (ret < 0)
+		goto out;
+
 	if (inode1)
 		mutex_unlock(&inode1->i_mutex);
 
 	if (inode2 && inode2 != inode1)
 		mutex_unlock(&inode2->i_mutex);
+
+out:
+	return ret;
 }
 
 /**
@@ -1123,70 +1226,76 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp,
 	ext4_lblk_t block_end, seq_start, add_blocks, file_end, seq_blocks = 0;
 	ext4_lblk_t rest_blocks;
 	pgoff_t orig_page_offset = 0, seq_end_page;
-	int ret, depth, last_extent = 0;
+	int ret1, ret2, depth, last_extent = 0;
 	int blocks_per_page = PAGE_CACHE_SIZE >> orig_inode->i_blkbits;
 	int data_offset_in_page;
 	int block_len_in_page;
 	int uninit;
 
 	/* protect orig and donor against a truncate */
-	mext_inode_double_lock(orig_inode, donor_inode);
+	ret1 = mext_inode_double_lock(orig_inode, donor_inode);
+	if (ret1 < 0)
+		return ret1;
 
 	mext_double_down_read(orig_inode, donor_inode);
 	/* Check the filesystem environment whether move_extent can be done */
-	ret = mext_check_arguments(orig_inode, donor_inode, orig_start,
+	ret1 = mext_check_arguments(orig_inode, donor_inode, orig_start,
 					donor_start, &len, *moved_len);
 	mext_double_up_read(orig_inode, donor_inode);
-	if (ret)
-		goto out2;
+	if (ret1)
+		goto out;
 
 	file_end = (i_size_read(orig_inode) - 1) >> orig_inode->i_blkbits;
 	block_end = block_start + len - 1;
 	if (file_end < block_end)
 		len -= block_end - file_end;
 
-	get_ext_path(orig_path, orig_inode, block_start, ret);
-	if (orig_path == NULL)
-		goto out2;
+	ret1 = get_ext_path(orig_inode, block_start, &orig_path);
+	if (ret1)
+		goto out;
 
 	/* Get path structure to check the hole */
-	get_ext_path(holecheck_path, orig_inode, block_start, ret);
-	if (holecheck_path == NULL)
+	ret1 = get_ext_path(orig_inode, block_start, &holecheck_path);
+	if (ret1)
 		goto out;
 
 	depth = ext_depth(orig_inode);
 	ext_cur = holecheck_path[depth].p_ext;
-	if (ext_cur == NULL) {
-		ret = -EINVAL;
-		goto out;
-	}
 
 	/*
-	 * Get proper extent whose ee_block is beyond block_start
-	 * if block_start was within the hole.
+	 * Get proper starting location of block replacement if block_start was
+	 * within the hole.
 	 */
 	if (le32_to_cpu(ext_cur->ee_block) +
 		ext4_ext_get_actual_len(ext_cur) - 1 < block_start) {
+		/*
+		 * The hole exists between extents or the tail of
+		 * original file.
+		 */
 		last_extent = mext_next_extent(orig_inode,
 					holecheck_path, &ext_cur);
 		if (last_extent < 0) {
-			ret = last_extent;
+			ret1 = last_extent;
 			goto out;
 		}
 		last_extent = mext_next_extent(orig_inode, orig_path,
 							&ext_dummy);
 		if (last_extent < 0) {
-			ret = last_extent;
+			ret1 = last_extent;
 			goto out;
 		}
-	}
-	seq_start = block_start;
+		seq_start = le32_to_cpu(ext_cur->ee_block);
+	} else if (le32_to_cpu(ext_cur->ee_block) > block_start)
+		/* The hole exists at the beginning of original file. */
+		seq_start = le32_to_cpu(ext_cur->ee_block);
+	else
+		seq_start = block_start;
 
 	/* No blocks within the specified range. */
 	if (le32_to_cpu(ext_cur->ee_block) > block_end) {
 		ext4_debug("ext4 move extent: The specified range of file "
 							"may be the hole\n");
-		ret = -EINVAL;
+		ret1 = -EINVAL;
 		goto out;
 	}
 
@@ -1206,7 +1315,7 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp,
 		last_extent = mext_next_extent(orig_inode, holecheck_path,
 						&ext_cur);
 		if (last_extent < 0) {
-			ret = last_extent;
+			ret1 = last_extent;
 			break;
 		}
 		add_blocks = ext4_ext_get_actual_len(ext_cur);
@@ -1258,16 +1367,23 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp,
 		while (orig_page_offset <= seq_end_page) {
 
 			/* Swap original branches with new branches */
-			ret = move_extent_par_page(o_filp, donor_inode,
+			ret1 = move_extent_per_page(o_filp, donor_inode,
 						orig_page_offset,
 						data_offset_in_page,
 						block_len_in_page, uninit);
-			if (ret < 0)
+			if (ret1 < 0)
 				goto out;
 			orig_page_offset++;
 			/* Count how many blocks we have exchanged */
 			*moved_len += block_len_in_page;
-			BUG_ON(*moved_len > len);
+			if (*moved_len > len) {
+				ext4_error(orig_inode->i_sb, __func__,
+					"We replaced blocks too much! "
+					"sum of replaced: %llu requested: %llu",
+					*moved_len, len);
+				ret1 = -EIO;
+				goto out;
+			}
 
 			data_offset_in_page = 0;
 			rest_blocks -= block_len_in_page;
@@ -1280,17 +1396,16 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp,
 		/* Decrease buffer counter */
 		if (holecheck_path)
 			ext4_ext_drop_refs(holecheck_path);
-		get_ext_path(holecheck_path, orig_inode,
-				      seq_start, ret);
-		if (holecheck_path == NULL)
+		ret1 = get_ext_path(orig_inode, seq_start, &holecheck_path);
+		if (ret1)
 			break;
 		depth = holecheck_path->p_depth;
 
 		/* Decrease buffer counter */
 		if (orig_path)
 			ext4_ext_drop_refs(orig_path);
-		get_ext_path(orig_path, orig_inode, seq_start, ret);
-		if (orig_path == NULL)
+		ret1 = get_ext_path(orig_inode, seq_start, &orig_path);
+		if (ret1)
 			break;
 
 		ext_cur = holecheck_path[depth].p_ext;
@@ -1307,14 +1422,13 @@ out:
 		ext4_ext_drop_refs(holecheck_path);
 		kfree(holecheck_path);
 	}
-out2:
-	mext_inode_double_unlock(orig_inode, donor_inode);
 
-	if (ret)
-		return ret;
+	ret2 = mext_inode_double_unlock(orig_inode, donor_inode);
 
-	/* All of the specified blocks must be exchanged in succeed */
-	BUG_ON(*moved_len != len);
+	if (ret1)
+		return ret1;
+	else if (ret2)
+		return ret2;
 
 	return 0;
 }
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index de04013d16f..42f81d285cd 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1518,8 +1518,12 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 			return retval;
 
 		if (blocks == 1 && !dx_fallback &&
-		    EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_DIR_INDEX))
-			return make_indexed_dir(handle, dentry, inode, bh);
+		    EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_DIR_INDEX)) {
+			retval = make_indexed_dir(handle, dentry, inode, bh);
+			if (retval == -ENOSPC)
+				brelse(bh);
+			return retval;
+		}
 		brelse(bh);
 	}
 	bh = ext4_append(handle, dir, &block, &retval);
@@ -1528,7 +1532,10 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
 	de = (struct ext4_dir_entry_2 *) bh->b_data;
 	de->inode = 0;
 	de->rec_len = ext4_rec_len_to_disk(blocksize, blocksize);
-	return add_dirent_to_buf(handle, dentry, inode, de, bh);
+	retval = add_dirent_to_buf(handle, dentry, inode, de, bh);
+	if (retval == -ENOSPC)
+		brelse(bh);
+	return retval;
 }
 
 /*
@@ -1590,9 +1597,9 @@ static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry,
 			goto cleanup;
 		node2 = (struct dx_node *)(bh2->b_data);
 		entries2 = node2->entries;
+		memset(&node2->fake, 0, sizeof(struct fake_dirent));
 		node2->fake.rec_len = ext4_rec_len_to_disk(sb->s_blocksize,
 							   sb->s_blocksize);
-		node2->fake.inode = 0;
 		BUFFER_TRACE(frame->bh, "get_write_access");
 		err = ext4_journal_get_write_access(handle, frame->bh);
 		if (err)
@@ -1657,7 +1664,8 @@ static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry,
 	if (!de)
 		goto cleanup;
 	err = add_dirent_to_buf(handle, dentry, inode, de, bh);
-	bh = NULL;
+	if (err != -ENOSPC)
+		bh = NULL;
 	goto cleanup;
 
 journal_error:
@@ -2310,7 +2318,7 @@ static int ext4_link(struct dentry *old_dentry,
 	struct inode *inode = old_dentry->d_inode;
 	int err, retries = 0;
 
-	if (EXT4_DIR_LINK_MAX(inode))
+	if (inode->i_nlink >= EXT4_LINK_MAX)
 		return -EMLINK;
 
 	/*
@@ -2413,7 +2421,7 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 			goto end_rename;
 		retval = -EMLINK;
 		if (!new_inode && new_dir != old_dir &&
-				new_dir->i_nlink >= EXT4_LINK_MAX)
+		    EXT4_DIR_LINK_MAX(new_dir))
 			goto end_rename;
 	}
 	if (!new_bh) {
@@ -2536,7 +2544,7 @@ const struct inode_operations ext4_dir_inode_operations = {
 	.listxattr	= ext4_listxattr,
 	.removexattr	= generic_removexattr,
 #endif
-	.permission	= ext4_permission,
+	.check_acl	= ext4_check_acl,
 	.fiemap         = ext4_fiemap,
 };
 
@@ -2548,5 +2556,5 @@ const struct inode_operations ext4_special_inode_operations = {
 	.listxattr	= ext4_listxattr,
 	.removexattr	= generic_removexattr,
 #endif
-	.permission	= ext4_permission,
+	.check_acl	= ext4_check_acl,
 };
diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index 68b0351fc64..3cfc343c41b 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -746,7 +746,6 @@ 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);
@@ -856,7 +855,6 @@ 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));
@@ -875,10 +873,8 @@ int ext4_group_add(struct super_block *sb, struct ext4_new_group_data *input)
 	 * descriptor
 	 */
 	err = ext4_mb_add_groupinfo(sb, input->group, gdp);
-	if (err) {
-		ext4_mb_put_buddy_cache_lock(sb, input->group, num_grp_locked);
+	if (err)
 		goto exit_journal;
-	}
 
 	/*
 	 * Make the new blocks and inodes valid next.  We do this before
@@ -920,7 +916,6 @@ 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_handle_dirty_metadata(handle, NULL, primary);
 
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 8f4f079e6b9..df539ba2777 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -45,6 +45,7 @@
 #include "ext4_jbd2.h"
 #include "xattr.h"
 #include "acl.h"
+#include "mballoc.h"
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/ext4.h>
@@ -344,7 +345,8 @@ static const char *ext4_decode_error(struct super_block *sb, int errno,
 		errstr = "Out of memory";
 		break;
 	case -EROFS:
-		if (!sb || EXT4_SB(sb)->s_journal->j_flags & JBD2_ABORT)
+		if (!sb || (EXT4_SB(sb)->s_journal &&
+			    EXT4_SB(sb)->s_journal->j_flags & JBD2_ABORT))
 			errstr = "Journal has aborted";
 		else
 			errstr = "Readonly filesystem";
@@ -962,7 +964,7 @@ static ssize_t ext4_quota_read(struct super_block *sb, int type, char *data,
 static ssize_t ext4_quota_write(struct super_block *sb, int type,
 				const char *data, size_t len, loff_t off);
 
-static struct dquot_operations ext4_quota_operations = {
+static const struct dquot_operations ext4_quota_operations = {
 	.initialize	= dquot_initialize,
 	.drop		= dquot_drop,
 	.alloc_space	= dquot_alloc_space,
@@ -983,7 +985,7 @@ static struct dquot_operations ext4_quota_operations = {
 	.destroy_dquot	= dquot_destroy,
 };
 
-static struct quotactl_ops ext4_qctl_operations = {
+static const struct quotactl_ops ext4_qctl_operations = {
 	.quota_on	= ext4_quota_on,
 	.quota_off	= vfs_quota_off,
 	.quota_sync	= vfs_quota_sync,
@@ -1279,11 +1281,9 @@ static int parse_options(char *options, struct super_block *sb,
 			*journal_devnum = option;
 			break;
 		case Opt_journal_checksum:
-			set_opt(sbi->s_mount_opt, JOURNAL_CHECKSUM);
-			break;
+			break;	/* Kept for backwards compatibility */
 		case Opt_journal_async_commit:
 			set_opt(sbi->s_mount_opt, JOURNAL_ASYNC_COMMIT);
-			set_opt(sbi->s_mount_opt, JOURNAL_CHECKSUM);
 			break;
 		case Opt_noload:
 			set_opt(sbi->s_mount_opt, NOLOAD);
@@ -1695,12 +1695,12 @@ static int ext4_fill_flex_info(struct super_block *sb)
 		gdp = ext4_get_group_desc(sb, i, NULL);
 
 		flex_group = ext4_flex_group(sbi, i);
-		atomic_set(&sbi->s_flex_groups[flex_group].free_inodes,
-			   ext4_free_inodes_count(sb, gdp));
-		atomic_set(&sbi->s_flex_groups[flex_group].free_blocks,
-			   ext4_free_blks_count(sb, gdp));
-		atomic_set(&sbi->s_flex_groups[flex_group].used_dirs,
-			   ext4_used_dirs_count(sb, gdp));
+		atomic_add(ext4_free_inodes_count(sb, gdp),
+			   &sbi->s_flex_groups[flex_group].free_inodes);
+		atomic_add(ext4_free_blks_count(sb, gdp),
+			   &sbi->s_flex_groups[flex_group].free_blocks);
+		atomic_add(ext4_used_dirs_count(sb, gdp),
+			   &sbi->s_flex_groups[flex_group].used_dirs);
 	}
 
 	return 1;
@@ -2253,6 +2253,49 @@ static struct kobj_type ext4_ktype = {
 	.release	= ext4_sb_release,
 };
 
+/*
+ * Check whether this filesystem can be mounted based on
+ * the features present and the RDONLY/RDWR mount requested.
+ * Returns 1 if this filesystem can be mounted as requested,
+ * 0 if it cannot be.
+ */
+static int ext4_feature_set_ok(struct super_block *sb, int readonly)
+{
+	if (EXT4_HAS_INCOMPAT_FEATURE(sb, ~EXT4_FEATURE_INCOMPAT_SUPP)) {
+		ext4_msg(sb, KERN_ERR,
+			"Couldn't mount because of "
+			"unsupported optional features (%x)",
+			(le32_to_cpu(EXT4_SB(sb)->s_es->s_feature_incompat) &
+			~EXT4_FEATURE_INCOMPAT_SUPP));
+		return 0;
+	}
+
+	if (readonly)
+		return 1;
+
+	/* Check that feature set is OK for a read-write mount */
+	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, ~EXT4_FEATURE_RO_COMPAT_SUPP)) {
+		ext4_msg(sb, KERN_ERR, "couldn't mount RDWR because of "
+			 "unsupported optional features (%x)",
+			 (le32_to_cpu(EXT4_SB(sb)->s_es->s_feature_ro_compat) &
+				~EXT4_FEATURE_RO_COMPAT_SUPP));
+		return 0;
+	}
+	/*
+	 * Large file size enabled file system can only be mounted
+	 * read-write on 32-bit systems if kernel is built with CONFIG_LBDAF
+	 */
+	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_HUGE_FILE)) {
+		if (sizeof(blkcnt_t) < sizeof(u64)) {
+			ext4_msg(sb, KERN_ERR, "Filesystem with huge files "
+				 "cannot be mounted RDWR without "
+				 "CONFIG_LBDAF");
+			return 0;
+		}
+	}
+	return 1;
+}
+
 static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 				__releases(kernel_lock)
 				__acquires(kernel_lock)
@@ -2274,7 +2317,6 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	unsigned int db_count;
 	unsigned int i;
 	int needs_recovery, has_huge_files;
-	int features;
 	__u64 blocks_count;
 	int err;
 	unsigned int journal_ioprio = DEFAULT_JOURNAL_IOPRIO;
@@ -2401,39 +2443,9 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	 * previously didn't change the revision level when setting the flags,
 	 * so there is a chance incompat flags are set on a rev 0 filesystem.
 	 */
-	features = EXT4_HAS_INCOMPAT_FEATURE(sb, ~EXT4_FEATURE_INCOMPAT_SUPP);
-	if (features) {
-		ext4_msg(sb, KERN_ERR,
-			"Couldn't mount because of "
-			"unsupported optional features (%x)",
-			(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) {
-		ext4_msg(sb, KERN_ERR,
-			"Couldn't mount RDWR because of "
-			"unsupported optional features (%x)",
-			(le32_to_cpu(EXT4_SB(sb)->s_es->s_feature_ro_compat) &
-			~EXT4_FEATURE_RO_COMPAT_SUPP));
+	if (!ext4_feature_set_ok(sb, (sb->s_flags & MS_RDONLY)))
 		goto failed_mount;
-	}
-	has_huge_files = EXT4_HAS_RO_COMPAT_FEATURE(sb,
-				    EXT4_FEATURE_RO_COMPAT_HUGE_FILE);
-	if (has_huge_files) {
-		/*
-		 * Large file size enabled file system can only be
-		 * mount if kernel is build with CONFIG_LBDAF
-		 */
-		if (sizeof(root->i_blocks) < sizeof(u64) &&
-				!(sb->s_flags & MS_RDONLY)) {
-			ext4_msg(sb, KERN_ERR, "Filesystem with huge "
-					"files cannot be mounted read-write "
-					"without CONFIG_LBDAF");
-			goto failed_mount;
-		}
-	}
+
 	blocksize = BLOCK_SIZE << le32_to_cpu(es->s_log_block_size);
 
 	if (blocksize < EXT4_MIN_BLOCK_SIZE ||
@@ -2469,6 +2481,8 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		}
 	}
 
+	has_huge_files = EXT4_HAS_RO_COMPAT_FEATURE(sb,
+				EXT4_FEATURE_RO_COMPAT_HUGE_FILE);
 	sbi->s_bitmap_maxbytes = ext4_max_bitmap_size(sb->s_blocksize_bits,
 						      has_huge_files);
 	sb->s_maxbytes = ext4_max_size(sb->s_blocksize_bits, has_huge_files);
@@ -2549,12 +2563,19 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		goto failed_mount;
 	}
 
-	if (ext4_blocks_count(es) >
-		    (sector_t)(~0ULL) >> (sb->s_blocksize_bits - 9)) {
+	/*
+	 * Test whether we have more sectors than will fit in sector_t,
+	 * and whether the max offset is addressable by the page cache.
+	 */
+	if ((ext4_blocks_count(es) >
+	     (sector_t)(~0ULL) >> (sb->s_blocksize_bits - 9)) ||
+	    (ext4_blocks_count(es) >
+	     (pgoff_t)(~0ULL) >> (PAGE_CACHE_SHIFT - sb->s_blocksize_bits))) {
 		ext4_msg(sb, KERN_ERR, "filesystem"
-			" too large to mount safely");
+			 " too large to mount safely on this system");
 		if (sizeof(sector_t) < 8)
 			ext4_msg(sb, KERN_WARNING, "CONFIG_LBDAF not enabled");
+		ret = -EFBIG;
 		goto failed_mount;
 	}
 
@@ -2595,6 +2616,8 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		goto failed_mount;
 	}
 	sbi->s_groups_count = blocks_count;
+	sbi->s_blockfile_groups = min_t(ext4_group_t, sbi->s_groups_count,
+			(EXT4_MAX_BLOCK_FILE_PHYS / EXT4_BLOCKS_PER_GROUP(sb)));
 	db_count = (sbi->s_groups_count + EXT4_DESC_PER_BLOCK(sb) - 1) /
 		   EXT4_DESC_PER_BLOCK(sb);
 	sbi->s_group_desc = kmalloc(db_count * sizeof(struct buffer_head *),
@@ -2729,20 +2752,14 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		goto failed_mount4;
 	}
 
-	if (test_opt(sb, JOURNAL_ASYNC_COMMIT)) {
-		jbd2_journal_set_features(sbi->s_journal,
-				JBD2_FEATURE_COMPAT_CHECKSUM, 0,
+	jbd2_journal_set_features(sbi->s_journal,
+				  JBD2_FEATURE_COMPAT_CHECKSUM, 0, 0);
+	if (test_opt(sb, JOURNAL_ASYNC_COMMIT))
+		jbd2_journal_set_features(sbi->s_journal, 0, 0,
 				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
-	} else if (test_opt(sb, JOURNAL_CHECKSUM)) {
-		jbd2_journal_set_features(sbi->s_journal,
-				JBD2_FEATURE_COMPAT_CHECKSUM, 0, 0);
+	else
 		jbd2_journal_clear_features(sbi->s_journal, 0, 0,
 				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
-	} else {
-		jbd2_journal_clear_features(sbi->s_journal,
-				JBD2_FEATURE_COMPAT_CHECKSUM, 0,
-				JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT);
-	}
 
 	/* We have now updated the journal if required, so we can
 	 * validate the data journaling mode. */
@@ -3208,7 +3225,18 @@ static int ext4_commit_super(struct super_block *sb, int sync)
 		clear_buffer_write_io_error(sbh);
 		set_buffer_uptodate(sbh);
 	}
-	es->s_wtime = cpu_to_le32(get_seconds());
+	/*
+	 * If the file system is mounted read-only, don't update the
+	 * superblock write time.  This avoids updating the superblock
+	 * write time when we are mounting the root file system
+	 * read/only but we need to replay the journal; at that point,
+	 * for people who are east of GMT and who make their clock
+	 * tick in localtime for Windows bug-for-bug compatibility,
+	 * the clock is set in the future, and this will cause e2fsck
+	 * to complain and force a full file system check.
+	 */
+	if (!(sb->s_flags & MS_RDONLY))
+		es->s_wtime = cpu_to_le32(get_seconds());
 	es->s_kbytes_written =
 		cpu_to_le64(EXT4_SB(sb)->s_kbytes_written + 
 			    ((part_stat_read(sb->s_bdev->bd_part, sectors[1]) -
@@ -3477,18 +3505,11 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 			if (sbi->s_journal)
 				ext4_mark_recovery_complete(sb, es);
 		} else {
-			int ret;
-			if ((ret = EXT4_HAS_RO_COMPAT_FEATURE(sb,
-					~EXT4_FEATURE_RO_COMPAT_SUPP))) {
-				ext4_msg(sb, KERN_WARNING, "couldn't "
-				       "remount RDWR because of unsupported "
-				       "optional features (%x)",
-				(le32_to_cpu(sbi->s_es->s_feature_ro_compat) &
-					~EXT4_FEATURE_RO_COMPAT_SUPP));
+			/* Make sure we can mount this feature set readwrite */
+			if (!ext4_feature_set_ok(sb, 0)) {
 				err = -EROFS;
 				goto restore_opts;
 			}
-
 			/*
 			 * Make sure the group descriptor checksums
 			 * are sane.  If they aren't, refuse to remount r/w.
diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index 62b31c24699..fed5b01d7a8 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -810,12 +810,23 @@ inserted:
 			get_bh(new_bh);
 		} else {
 			/* We need to allocate a new block */
-			ext4_fsblk_t goal = ext4_group_first_block_no(sb,
+			ext4_fsblk_t goal, block;
+
+			goal = ext4_group_first_block_no(sb,
 						EXT4_I(inode)->i_block_group);
-			ext4_fsblk_t block = ext4_new_meta_blocks(handle, inode,
+
+			/* non-extent files can't have physical blocks past 2^32 */
+			if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+				goal = goal & EXT4_MAX_BLOCK_FILE_PHYS;
+
+			block = ext4_new_meta_blocks(handle, inode,
 						  goal, NULL, &error);
 			if (error)
 				goto cleanup;
+
+			if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+				BUG_ON(block > EXT4_MAX_BLOCK_FILE_PHYS);
+
 			ea_idebug(inode, "creating block %d", block);
 
 			new_bh = sb_getblk(sb, block);
diff --git a/fs/fat/file.c b/fs/fat/file.c
index f042b965c95..e8c159de236 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -176,8 +176,26 @@ static int fat_cont_expand(struct inode *inode, loff_t size)
 
 	inode->i_ctime = inode->i_mtime = CURRENT_TIME_SEC;
 	mark_inode_dirty(inode);
-	if (IS_SYNC(inode))
-		err = sync_page_range_nolock(inode, mapping, start, count);
+	if (IS_SYNC(inode)) {
+		int err2;
+
+		/*
+		 * Opencode syncing since we don't have a file open to use
+		 * standard fsync path.
+		 */
+		err = filemap_fdatawrite_range(mapping, start,
+					       start + count - 1);
+		err2 = sync_mapping_buffers(mapping);
+		if (!err)
+			err = err2;
+		err2 = write_inode_now(inode, 1);
+		if (!err)
+			err = err2;
+		if (!err) {
+			err =  filemap_fdatawait_range(mapping, start,
+						       start + count - 1);
+		}
+	}
 out:
 	return err;
 }
diff --git a/fs/fat/misc.c b/fs/fat/misc.c
index a6c20473dfd..4e35be873e0 100644
--- a/fs/fat/misc.c
+++ b/fs/fat/misc.c
@@ -119,8 +119,8 @@ int fat_chain_add(struct inode *inode, int new_dclus, int nr_cluster)
 		MSDOS_I(inode)->i_start = new_dclus;
 		MSDOS_I(inode)->i_logstart = new_dclus;
 		/*
-		 * Since generic_osync_inode() synchronize later if
-		 * this is not directory, we don't here.
+		 * Since generic_write_sync() synchronizes regular files later,
+		 * we sync here only directories.
 		 */
 		if (S_ISDIR(inode->i_mode) && IS_DIRSYNC(inode)) {
 			ret = fat_sync_inode(inode);
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index c54226be529..8e1e5e19d21 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -19,171 +19,245 @@
 #include <linux/sched.h>
 #include <linux/fs.h>
 #include <linux/mm.h>
+#include <linux/kthread.h>
+#include <linux/freezer.h>
 #include <linux/writeback.h>
 #include <linux/blkdev.h>
 #include <linux/backing-dev.h>
 #include <linux/buffer_head.h>
 #include "internal.h"
 
+#define inode_to_bdi(inode)	((inode)->i_mapping->backing_dev_info)
 
-/**
- * writeback_acquire - attempt to get exclusive writeback access to a device
- * @bdi: the device's backing_dev_info structure
- *
- * It is a waste of resources to have more than one pdflush thread blocked on
- * a single request queue.  Exclusion at the request_queue level is obtained
- * via a flag in the request_queue's backing_dev_info.state.
- *
- * Non-request_queue-backed address_spaces will share default_backing_dev_info,
- * unless they implement their own.  Which is somewhat inefficient, as this
- * may prevent concurrent writeback against multiple devices.
+/*
+ * We don't actually have pdflush, but this one is exported though /proc...
+ */
+int nr_pdflush_threads;
+
+/*
+ * Passed into wb_writeback(), essentially a subset of writeback_control
+ */
+struct wb_writeback_args {
+	long nr_pages;
+	struct super_block *sb;
+	enum writeback_sync_modes sync_mode;
+	int for_kupdate;
+	int range_cyclic;
+};
+
+/*
+ * Work items for the bdi_writeback threads
  */
-static int writeback_acquire(struct backing_dev_info *bdi)
+struct bdi_work {
+	struct list_head list;		/* pending work list */
+	struct rcu_head rcu_head;	/* for RCU free/clear of work */
+
+	unsigned long seen;		/* threads that have seen this work */
+	atomic_t pending;		/* number of threads still to do work */
+
+	struct wb_writeback_args args;	/* writeback arguments */
+
+	unsigned long state;		/* flag bits, see WS_* */
+};
+
+enum {
+	WS_USED_B = 0,
+	WS_ONSTACK_B,
+};
+
+#define WS_USED (1 << WS_USED_B)
+#define WS_ONSTACK (1 << WS_ONSTACK_B)
+
+static inline bool bdi_work_on_stack(struct bdi_work *work)
+{
+	return test_bit(WS_ONSTACK_B, &work->state);
+}
+
+static inline void bdi_work_init(struct bdi_work *work,
+				 struct wb_writeback_args *args)
 {
-	return !test_and_set_bit(BDI_pdflush, &bdi->state);
+	INIT_RCU_HEAD(&work->rcu_head);
+	work->args = *args;
+	work->state = WS_USED;
 }
 
 /**
  * writeback_in_progress - determine whether there is writeback in progress
  * @bdi: the device's backing_dev_info structure.
  *
- * Determine whether there is writeback in progress against a backing device.
+ * Determine whether there is writeback waiting to be handled against a
+ * backing device.
  */
 int writeback_in_progress(struct backing_dev_info *bdi)
 {
-	return test_bit(BDI_pdflush, &bdi->state);
+	return !list_empty(&bdi->work_list);
 }
 
-/**
- * writeback_release - relinquish exclusive writeback access against a device.
- * @bdi: the device's backing_dev_info structure
- */
-static void writeback_release(struct backing_dev_info *bdi)
+static void bdi_work_clear(struct bdi_work *work)
 {
-	BUG_ON(!writeback_in_progress(bdi));
-	clear_bit(BDI_pdflush, &bdi->state);
+	clear_bit(WS_USED_B, &work->state);
+	smp_mb__after_clear_bit();
+	/*
+	 * work can have disappeared at this point. bit waitq functions
+	 * should be able to tolerate this, provided bdi_sched_wait does
+	 * not dereference it's pointer argument.
+	*/
+	wake_up_bit(&work->state, WS_USED_B);
 }
 
-static noinline void block_dump___mark_inode_dirty(struct inode *inode)
+static void bdi_work_free(struct rcu_head *head)
 {
-	if (inode->i_ino || strcmp(inode->i_sb->s_id, "bdev")) {
-		struct dentry *dentry;
-		const char *name = "?";
+	struct bdi_work *work = container_of(head, struct bdi_work, rcu_head);
 
-		dentry = d_find_alias(inode);
-		if (dentry) {
-			spin_lock(&dentry->d_lock);
-			name = (const char *) dentry->d_name.name;
-		}
-		printk(KERN_DEBUG
-		       "%s(%d): dirtied inode %lu (%s) on %s\n",
-		       current->comm, task_pid_nr(current), inode->i_ino,
-		       name, inode->i_sb->s_id);
-		if (dentry) {
-			spin_unlock(&dentry->d_lock);
-			dput(dentry);
-		}
-	}
+	if (!bdi_work_on_stack(work))
+		kfree(work);
+	else
+		bdi_work_clear(work);
 }
 
-/**
- *	__mark_inode_dirty -	internal function
- *	@inode: inode to mark
- *	@flags: what kind of dirty (i.e. I_DIRTY_SYNC)
- *	Mark an inode as dirty. Callers should use mark_inode_dirty or
- *  	mark_inode_dirty_sync.
- *
- * Put the inode on the super block's dirty list.
- *
- * CAREFUL! We mark it dirty unconditionally, but move it onto the
- * dirty list only if it is hashed or if it refers to a blockdev.
- * If it was not hashed, it will never be added to the dirty list
- * even if it is later hashed, as it will have been marked dirty already.
- *
- * In short, make sure you hash any inodes _before_ you start marking
- * them dirty.
- *
- * This function *must* be atomic for the I_DIRTY_PAGES case -
- * set_page_dirty() is called under spinlock in several places.
- *
- * Note that for blockdevs, inode->dirtied_when represents the dirtying time of
- * the block-special inode (/dev/hda1) itself.  And the ->dirtied_when field of
- * the kernel-internal blockdev inode represents the dirtying time of the
- * blockdev's pages.  This is why for I_DIRTY_PAGES we always use
- * page->mapping->host, so the page-dirtying time is recorded in the internal
- * blockdev inode.
- */
-void __mark_inode_dirty(struct inode *inode, int flags)
+static void wb_work_complete(struct bdi_work *work)
 {
-	struct super_block *sb = inode->i_sb;
+	const enum writeback_sync_modes sync_mode = work->args.sync_mode;
+	int onstack = bdi_work_on_stack(work);
 
 	/*
-	 * Don't do this for I_DIRTY_PAGES - that doesn't actually
-	 * dirty the inode itself
+	 * For allocated work, we can clear the done/seen bit right here.
+	 * For on-stack work, we need to postpone both the clear and free
+	 * to after the RCU grace period, since the stack could be invalidated
+	 * as soon as bdi_work_clear() has done the wakeup.
 	 */
-	if (flags & (I_DIRTY_SYNC | I_DIRTY_DATASYNC)) {
-		if (sb->s_op->dirty_inode)
-			sb->s_op->dirty_inode(inode);
-	}
+	if (!onstack)
+		bdi_work_clear(work);
+	if (sync_mode == WB_SYNC_NONE || onstack)
+		call_rcu(&work->rcu_head, bdi_work_free);
+}
 
+static void wb_clear_pending(struct bdi_writeback *wb, struct bdi_work *work)
+{
 	/*
-	 * make sure that changes are seen by all cpus before we test i_state
-	 * -- mikulas
+	 * The caller has retrieved the work arguments from this work,
+	 * drop our reference. If this is the last ref, delete and free it
 	 */
-	smp_mb();
+	if (atomic_dec_and_test(&work->pending)) {
+		struct backing_dev_info *bdi = wb->bdi;
 
-	/* avoid the locking if we can */
-	if ((inode->i_state & flags) == flags)
-		return;
+		spin_lock(&bdi->wb_lock);
+		list_del_rcu(&work->list);
+		spin_unlock(&bdi->wb_lock);
 
-	if (unlikely(block_dump))
-		block_dump___mark_inode_dirty(inode);
+		wb_work_complete(work);
+	}
+}
 
-	spin_lock(&inode_lock);
-	if ((inode->i_state & flags) != flags) {
-		const int was_dirty = inode->i_state & I_DIRTY;
+static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work)
+{
+	work->seen = bdi->wb_mask;
+	BUG_ON(!work->seen);
+	atomic_set(&work->pending, bdi->wb_cnt);
+	BUG_ON(!bdi->wb_cnt);
 
-		inode->i_state |= flags;
+	/*
+	 * list_add_tail_rcu() contains the necessary barriers to
+	 * make sure the above stores are seen before the item is
+	 * noticed on the list
+	 */
+	spin_lock(&bdi->wb_lock);
+	list_add_tail_rcu(&work->list, &bdi->work_list);
+	spin_unlock(&bdi->wb_lock);
 
-		/*
-		 * If the inode is being synced, just update its dirty state.
-		 * The unlocker will place the inode on the appropriate
-		 * superblock list, based upon its state.
-		 */
-		if (inode->i_state & I_SYNC)
-			goto out;
+	/*
+	 * If the default thread isn't there, make sure we add it. When
+	 * it gets created and wakes up, we'll run this work.
+	 */
+	if (unlikely(list_empty_careful(&bdi->wb_list)))
+		wake_up_process(default_backing_dev_info.wb.task);
+	else {
+		struct bdi_writeback *wb = &bdi->wb;
 
-		/*
-		 * Only add valid (hashed) inodes to the superblock's
-		 * dirty list.  Add blockdev inodes as well.
-		 */
-		if (!S_ISBLK(inode->i_mode)) {
-			if (hlist_unhashed(&inode->i_hash))
-				goto out;
-		}
-		if (inode->i_state & (I_FREEING|I_CLEAR))
-			goto out;
+		if (wb->task)
+			wake_up_process(wb->task);
+	}
+}
 
-		/*
-		 * If the inode was already on s_dirty/s_io/s_more_io, don't
-		 * reposition it (that would break s_dirty time-ordering).
-		 */
-		if (!was_dirty) {
-			inode->dirtied_when = jiffies;
-			list_move(&inode->i_list, &sb->s_dirty);
-		}
+/*
+ * Used for on-stack allocated work items. The caller needs to wait until
+ * the wb threads have acked the work before it's safe to continue.
+ */
+static void bdi_wait_on_work_clear(struct bdi_work *work)
+{
+	wait_on_bit(&work->state, WS_USED_B, bdi_sched_wait,
+		    TASK_UNINTERRUPTIBLE);
+}
+
+static void bdi_alloc_queue_work(struct backing_dev_info *bdi,
+				 struct wb_writeback_args *args)
+{
+	struct bdi_work *work;
+
+	/*
+	 * This is WB_SYNC_NONE writeback, so if allocation fails just
+	 * wakeup the thread for old dirty data writeback
+	 */
+	work = kmalloc(sizeof(*work), GFP_ATOMIC);
+	if (work) {
+		bdi_work_init(work, args);
+		bdi_queue_work(bdi, work);
+	} else {
+		struct bdi_writeback *wb = &bdi->wb;
+
+		if (wb->task)
+			wake_up_process(wb->task);
 	}
-out:
-	spin_unlock(&inode_lock);
 }
 
-EXPORT_SYMBOL(__mark_inode_dirty);
+/**
+ * bdi_sync_writeback - start and wait for writeback
+ * @bdi: the backing device to write from
+ * @sb: write inodes from this super_block
+ *
+ * Description:
+ *   This does WB_SYNC_ALL data integrity writeback and waits for the
+ *   IO to complete. Callers must hold the sb s_umount semaphore for
+ *   reading, to avoid having the super disappear before we are done.
+ */
+static void bdi_sync_writeback(struct backing_dev_info *bdi,
+			       struct super_block *sb)
+{
+	struct wb_writeback_args args = {
+		.sb		= sb,
+		.sync_mode	= WB_SYNC_ALL,
+		.nr_pages	= LONG_MAX,
+		.range_cyclic	= 0,
+	};
+	struct bdi_work work;
 
-static int write_inode(struct inode *inode, int sync)
+	bdi_work_init(&work, &args);
+	work.state |= WS_ONSTACK;
+
+	bdi_queue_work(bdi, &work);
+	bdi_wait_on_work_clear(&work);
+}
+
+/**
+ * bdi_start_writeback - start writeback
+ * @bdi: the backing device to write from
+ * @nr_pages: the number of pages to write
+ *
+ * Description:
+ *   This does WB_SYNC_NONE opportunistic writeback. The IO is only
+ *   started when this function returns, we make no guarentees on
+ *   completion. Caller need not hold sb s_umount semaphore.
+ *
+ */
+void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages)
 {
-	if (inode->i_sb->s_op->write_inode && !is_bad_inode(inode))
-		return inode->i_sb->s_op->write_inode(inode, sync);
-	return 0;
+	struct wb_writeback_args args = {
+		.sync_mode	= WB_SYNC_NONE,
+		.nr_pages	= nr_pages,
+		.range_cyclic	= 1,
+	};
+
+	bdi_alloc_queue_work(bdi, &args);
 }
 
 /*
@@ -191,31 +265,32 @@ static int write_inode(struct inode *inode, int sync)
  * furthest end of its superblock's dirty-inode list.
  *
  * Before stamping the inode's ->dirtied_when, we check to see whether it is
- * already the most-recently-dirtied inode on the s_dirty list.  If that is
+ * already the most-recently-dirtied inode on the b_dirty list.  If that is
  * the case then the inode must have been redirtied while it was being written
  * out and we don't reset its dirtied_when.
  */
 static void redirty_tail(struct inode *inode)
 {
-	struct super_block *sb = inode->i_sb;
+	struct bdi_writeback *wb = &inode_to_bdi(inode)->wb;
 
-	if (!list_empty(&sb->s_dirty)) {
-		struct inode *tail_inode;
+	if (!list_empty(&wb->b_dirty)) {
+		struct inode *tail;
 
-		tail_inode = list_entry(sb->s_dirty.next, struct inode, i_list);
-		if (time_before(inode->dirtied_when,
-				tail_inode->dirtied_when))
+		tail = list_entry(wb->b_dirty.next, struct inode, i_list);
+		if (time_before(inode->dirtied_when, tail->dirtied_when))
 			inode->dirtied_when = jiffies;
 	}
-	list_move(&inode->i_list, &sb->s_dirty);
+	list_move(&inode->i_list, &wb->b_dirty);
 }
 
 /*
- * requeue inode for re-scanning after sb->s_io list is exhausted.
+ * requeue inode for re-scanning after bdi->b_io list is exhausted.
  */
 static void requeue_io(struct inode *inode)
 {
-	list_move(&inode->i_list, &inode->i_sb->s_more_io);
+	struct bdi_writeback *wb = &inode_to_bdi(inode)->wb;
+
+	list_move(&inode->i_list, &wb->b_more_io);
 }
 
 static void inode_sync_complete(struct inode *inode)
@@ -262,20 +337,18 @@ static void move_expired_inodes(struct list_head *delaying_queue,
 /*
  * Queue all expired dirty inodes for io, eldest first.
  */
-static void queue_io(struct super_block *sb,
-				unsigned long *older_than_this)
+static void queue_io(struct bdi_writeback *wb, unsigned long *older_than_this)
 {
-	list_splice_init(&sb->s_more_io, sb->s_io.prev);
-	move_expired_inodes(&sb->s_dirty, &sb->s_io, older_than_this);
+	list_splice_init(&wb->b_more_io, wb->b_io.prev);
+	move_expired_inodes(&wb->b_dirty, &wb->b_io, older_than_this);
 }
 
-int sb_has_dirty_inodes(struct super_block *sb)
+static int write_inode(struct inode *inode, int sync)
 {
-	return !list_empty(&sb->s_dirty) ||
-	       !list_empty(&sb->s_io) ||
-	       !list_empty(&sb->s_more_io);
+	if (inode->i_sb->s_op->write_inode && !is_bad_inode(inode))
+		return inode->i_sb->s_op->write_inode(inode, sync);
+	return 0;
 }
-EXPORT_SYMBOL(sb_has_dirty_inodes);
 
 /*
  * Wait for writeback on an inode to complete.
@@ -322,11 +395,11 @@ writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
 	if (inode->i_state & I_SYNC) {
 		/*
 		 * If this inode is locked for writeback and we are not doing
-		 * writeback-for-data-integrity, move it to s_more_io so that
+		 * writeback-for-data-integrity, move it to b_more_io so that
 		 * writeback can proceed with the other inodes on s_io.
 		 *
 		 * We'll have another go at writing back this inode when we
-		 * completed a full scan of s_io.
+		 * completed a full scan of b_io.
 		 */
 		if (!wait) {
 			requeue_io(inode);
@@ -371,11 +444,11 @@ writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
 			/*
 			 * We didn't write back all the pages.  nfs_writepages()
 			 * sometimes bales out without doing anything. Redirty
-			 * the inode; Move it from s_io onto s_more_io/s_dirty.
+			 * the inode; Move it from b_io onto b_more_io/b_dirty.
 			 */
 			/*
 			 * akpm: if the caller was the kupdate function we put
-			 * this inode at the head of s_dirty so it gets first
+			 * this inode at the head of b_dirty so it gets first
 			 * consideration.  Otherwise, move it to the tail, for
 			 * the reasons described there.  I'm not really sure
 			 * how much sense this makes.  Presumably I had a good
@@ -385,7 +458,7 @@ writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
 			if (wbc->for_kupdate) {
 				/*
 				 * For the kupdate function we move the inode
-				 * to s_more_io so it will get more writeout as
+				 * to b_more_io so it will get more writeout as
 				 * soon as the queue becomes uncongested.
 				 */
 				inode->i_state |= I_DIRTY_PAGES;
@@ -434,50 +507,84 @@ writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
 }
 
 /*
- * Write out a superblock's list of dirty inodes.  A wait will be performed
- * upon no inodes, all inodes or the final one, depending upon sync_mode.
- *
- * If older_than_this is non-NULL, then only write out inodes which
- * had their first dirtying at a time earlier than *older_than_this.
- *
- * If we're a pdflush thread, then implement pdflush collision avoidance
- * against the entire list.
+ * For WB_SYNC_NONE writeback, the caller does not have the sb pinned
+ * before calling writeback. So make sure that we do pin it, so it doesn't
+ * go away while we are writing inodes from it.
  *
- * 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,
- * assume that all inodes are backed by the same queue.
- *
- * FIXME: this linear search could get expensive with many fileystems.  But
- * how to fix?  We need to go from an address_space to all inodes which share
- * a queue with that address_space.  (Easy: have a global "dirty superblocks"
- * list).
- *
- * The inodes to be written are parked on sb->s_io.  They are moved back onto
- * sb->s_dirty as they are selected for writing.  This way, none can be missed
- * on the writer throttling path, and we get decent balancing between many
- * throttled threads: we don't want them all piling up on inode_sync_wait.
+ * Returns 0 if the super was successfully pinned (or pinning wasn't needed),
+ * 1 if we failed.
  */
-void generic_sync_sb_inodes(struct super_block *sb,
+static int pin_sb_for_writeback(struct writeback_control *wbc,
+				   struct inode *inode)
+{
+	struct super_block *sb = inode->i_sb;
+
+	/*
+	 * Caller must already hold the ref for this
+	 */
+	if (wbc->sync_mode == WB_SYNC_ALL) {
+		WARN_ON(!rwsem_is_locked(&sb->s_umount));
+		return 0;
+	}
+
+	spin_lock(&sb_lock);
+	sb->s_count++;
+	if (down_read_trylock(&sb->s_umount)) {
+		if (sb->s_root) {
+			spin_unlock(&sb_lock);
+			return 0;
+		}
+		/*
+		 * umounted, drop rwsem again and fall through to failure
+		 */
+		up_read(&sb->s_umount);
+	}
+
+	sb->s_count--;
+	spin_unlock(&sb_lock);
+	return 1;
+}
+
+static void unpin_sb_for_writeback(struct writeback_control *wbc,
+				   struct inode *inode)
+{
+	struct super_block *sb = inode->i_sb;
+
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		return;
+
+	up_read(&sb->s_umount);
+	put_super(sb);
+}
+
+static void writeback_inodes_wb(struct bdi_writeback *wb,
 				struct writeback_control *wbc)
 {
+	struct super_block *sb = wbc->sb;
+	const int is_blkdev_sb = sb_is_blkdev_sb(sb);
 	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))
-		queue_io(sb, wbc->older_than_this);
 
-	while (!list_empty(&sb->s_io)) {
-		struct inode *inode = list_entry(sb->s_io.prev,
+	if (!wbc->for_kupdate || list_empty(&wb->b_io))
+		queue_io(wb, wbc->older_than_this);
+
+	while (!list_empty(&wb->b_io)) {
+		struct inode *inode = list_entry(wb->b_io.prev,
 						struct inode, i_list);
-		struct address_space *mapping = inode->i_mapping;
-		struct backing_dev_info *bdi = mapping->backing_dev_info;
 		long pages_skipped;
 
-		if (!bdi_cap_writeback_dirty(bdi)) {
+		/*
+		 * super block given and doesn't match, skip this inode
+		 */
+		if (sb && sb != inode->i_sb) {
+			redirty_tail(inode);
+			continue;
+		}
+
+		if (!bdi_cap_writeback_dirty(wb->bdi)) {
 			redirty_tail(inode);
-			if (sb_is_blkdev_sb(sb)) {
+			if (is_blkdev_sb) {
 				/*
 				 * Dirty memory-backed blockdev: the ramdisk
 				 * driver does this.  Skip just this inode
@@ -497,21 +604,14 @@ void generic_sync_sb_inodes(struct super_block *sb,
 			continue;
 		}
 
-		if (wbc->nonblocking && bdi_write_congested(bdi)) {
+		if (wbc->nonblocking && bdi_write_congested(wb->bdi)) {
 			wbc->encountered_congestion = 1;
-			if (!sb_is_blkdev_sb(sb))
+			if (!is_blkdev_sb)
 				break;		/* Skip a congested fs */
 			requeue_io(inode);
 			continue;		/* Skip a congested blockdev */
 		}
 
-		if (wbc->bdi && bdi != wbc->bdi) {
-			if (!sb_is_blkdev_sb(sb))
-				break;		/* fs has the wrong queue */
-			requeue_io(inode);
-			continue;		/* blockdev has wrong queue */
-		}
-
 		/*
 		 * Was this inode dirtied after sync_sb_inodes was called?
 		 * This keeps sync from extra jobs and livelock.
@@ -519,16 +619,16 @@ void generic_sync_sb_inodes(struct super_block *sb,
 		if (inode_dirtied_after(inode, start))
 			break;
 
-		/* Is another pdflush already flushing this queue? */
-		if (current_is_pdflush() && !writeback_acquire(bdi))
-			break;
+		if (pin_sb_for_writeback(wbc, inode)) {
+			requeue_io(inode);
+			continue;
+		}
 
 		BUG_ON(inode->i_state & (I_FREEING | I_CLEAR));
 		__iget(inode);
 		pages_skipped = wbc->pages_skipped;
 		writeback_single_inode(inode, wbc);
-		if (current_is_pdflush())
-			writeback_release(bdi);
+		unpin_sb_for_writeback(wbc, inode);
 		if (wbc->pages_skipped != pages_skipped) {
 			/*
 			 * writeback is not making progress due to locked
@@ -544,144 +644,520 @@ void generic_sync_sb_inodes(struct super_block *sb,
 			wbc->more_io = 1;
 			break;
 		}
-		if (!list_empty(&sb->s_more_io))
+		if (!list_empty(&wb->b_more_io))
 			wbc->more_io = 1;
 	}
 
-	if (sync) {
-		struct inode *inode, *old_inode = NULL;
+	spin_unlock(&inode_lock);
+	/* Leave any unwritten inodes on b_io */
+}
+
+void writeback_inodes_wbc(struct writeback_control *wbc)
+{
+	struct backing_dev_info *bdi = wbc->bdi;
 
+	writeback_inodes_wb(&bdi->wb, wbc);
+}
+
+/*
+ * The maximum number of pages to writeout in a single bdi flush/kupdate
+ * operation.  We do this so we don't hold I_SYNC against an inode for
+ * enormous amounts of time, which would block a userspace task which has
+ * been forced to throttle against that inode.  Also, the code reevaluates
+ * the dirty each time it has written this many pages.
+ */
+#define MAX_WRITEBACK_PAGES     1024
+
+static inline bool over_bground_thresh(void)
+{
+	unsigned long background_thresh, dirty_thresh;
+
+	get_dirty_limits(&background_thresh, &dirty_thresh, NULL, NULL);
+
+	return (global_page_state(NR_FILE_DIRTY) +
+		global_page_state(NR_UNSTABLE_NFS) >= background_thresh);
+}
+
+/*
+ * Explicit flushing or periodic writeback of "old" data.
+ *
+ * Define "old": the first time one of an inode's pages is dirtied, we mark the
+ * dirtying-time in the inode's address_space.  So this periodic writeback code
+ * just walks the superblock inode list, writing back any inodes which are
+ * older than a specific point in time.
+ *
+ * Try to run once per dirty_writeback_interval.  But if a writeback event
+ * takes longer than a dirty_writeback_interval interval, then leave a
+ * one-second gap.
+ *
+ * older_than_this takes precedence over nr_to_write.  So we'll only write back
+ * all dirty pages if they are all attached to "old" mappings.
+ */
+static long wb_writeback(struct bdi_writeback *wb,
+			 struct wb_writeback_args *args)
+{
+	struct writeback_control wbc = {
+		.bdi			= wb->bdi,
+		.sb			= args->sb,
+		.sync_mode		= args->sync_mode,
+		.older_than_this	= NULL,
+		.for_kupdate		= args->for_kupdate,
+		.range_cyclic		= args->range_cyclic,
+	};
+	unsigned long oldest_jif;
+	long wrote = 0;
+
+	if (wbc.for_kupdate) {
+		wbc.older_than_this = &oldest_jif;
+		oldest_jif = jiffies -
+				msecs_to_jiffies(dirty_expire_interval * 10);
+	}
+	if (!wbc.range_cyclic) {
+		wbc.range_start = 0;
+		wbc.range_end = LLONG_MAX;
+	}
+
+	for (;;) {
 		/*
-		 * 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.
+		 * Don't flush anything for non-integrity writeback where
+		 * no nr_pages was given
 		 */
-		list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
-			struct address_space *mapping;
+		if (!args->for_kupdate && args->nr_pages <= 0 &&
+		     args->sync_mode == WB_SYNC_NONE)
+			break;
 
-			if (inode->i_state &
-					(I_FREEING|I_CLEAR|I_WILL_FREE|I_NEW))
-				continue;
-			mapping = inode->i_mapping;
-			if (mapping->nrpages == 0)
+		/*
+		 * If no specific pages were given and this is just a
+		 * periodic background writeout and we are below the
+		 * background dirty threshold, don't do anything
+		 */
+		if (args->for_kupdate && args->nr_pages <= 0 &&
+		    !over_bground_thresh())
+			break;
+
+		wbc.more_io = 0;
+		wbc.encountered_congestion = 0;
+		wbc.nr_to_write = MAX_WRITEBACK_PAGES;
+		wbc.pages_skipped = 0;
+		writeback_inodes_wb(wb, &wbc);
+		args->nr_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write;
+		wrote += MAX_WRITEBACK_PAGES - wbc.nr_to_write;
+
+		/*
+		 * If we ran out of stuff to write, bail unless more_io got set
+		 */
+		if (wbc.nr_to_write > 0 || wbc.pages_skipped > 0) {
+			if (wbc.more_io && !wbc.for_kupdate)
 				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;
+			break;
+		}
+	}
+
+	return wrote;
+}
+
+/*
+ * Return the next bdi_work struct that hasn't been processed by this
+ * wb thread yet. ->seen is initially set for each thread that exists
+ * for this device, when a thread first notices a piece of work it
+ * clears its bit. Depending on writeback type, the thread will notify
+ * completion on either receiving the work (WB_SYNC_NONE) or after
+ * it is done (WB_SYNC_ALL).
+ */
+static struct bdi_work *get_next_work_item(struct backing_dev_info *bdi,
+					   struct bdi_writeback *wb)
+{
+	struct bdi_work *work, *ret = NULL;
+
+	rcu_read_lock();
+
+	list_for_each_entry_rcu(work, &bdi->work_list, list) {
+		if (!test_bit(wb->nr, &work->seen))
+			continue;
+		clear_bit(wb->nr, &work->seen);
+
+		ret = work;
+		break;
+	}
+
+	rcu_read_unlock();
+	return ret;
+}
+
+static long wb_check_old_data_flush(struct bdi_writeback *wb)
+{
+	unsigned long expired;
+	long nr_pages;
+
+	expired = wb->last_old_flush +
+			msecs_to_jiffies(dirty_writeback_interval * 10);
+	if (time_before(jiffies, expired))
+		return 0;
+
+	wb->last_old_flush = jiffies;
+	nr_pages = global_page_state(NR_FILE_DIRTY) +
+			global_page_state(NR_UNSTABLE_NFS) +
+			(inodes_stat.nr_inodes - inodes_stat.nr_unused);
+
+	if (nr_pages) {
+		struct wb_writeback_args args = {
+			.nr_pages	= nr_pages,
+			.sync_mode	= WB_SYNC_NONE,
+			.for_kupdate	= 1,
+			.range_cyclic	= 1,
+		};
+
+		return wb_writeback(wb, &args);
+	}
+
+	return 0;
+}
+
+/*
+ * Retrieve work items and do the writeback they describe
+ */
+long wb_do_writeback(struct bdi_writeback *wb, int force_wait)
+{
+	struct backing_dev_info *bdi = wb->bdi;
+	struct bdi_work *work;
+	long wrote = 0;
 
-			filemap_fdatawait(mapping);
+	while ((work = get_next_work_item(bdi, wb)) != NULL) {
+		struct wb_writeback_args args = work->args;
 
-			cond_resched();
+		/*
+		 * Override sync mode, in case we must wait for completion
+		 */
+		if (force_wait)
+			work->args.sync_mode = args.sync_mode = WB_SYNC_ALL;
 
-			spin_lock(&inode_lock);
+		/*
+		 * If this isn't a data integrity operation, just notify
+		 * that we have seen this work and we are now starting it.
+		 */
+		if (args.sync_mode == WB_SYNC_NONE)
+			wb_clear_pending(wb, work);
+
+		wrote += wb_writeback(wb, &args);
+
+		/*
+		 * This is a data integrity writeback, so only do the
+		 * notification when we have completed the work.
+		 */
+		if (args.sync_mode == WB_SYNC_ALL)
+			wb_clear_pending(wb, work);
+	}
+
+	/*
+	 * Check for periodic writeback, kupdated() style
+	 */
+	wrote += wb_check_old_data_flush(wb);
+
+	return wrote;
+}
+
+/*
+ * Handle writeback of dirty data for the device backed by this bdi. Also
+ * wakes up periodically and does kupdated style flushing.
+ */
+int bdi_writeback_task(struct bdi_writeback *wb)
+{
+	unsigned long last_active = jiffies;
+	unsigned long wait_jiffies = -1UL;
+	long pages_written;
+
+	while (!kthread_should_stop()) {
+		pages_written = wb_do_writeback(wb, 0);
+
+		if (pages_written)
+			last_active = jiffies;
+		else if (wait_jiffies != -1UL) {
+			unsigned long max_idle;
+
+			/*
+			 * Longest period of inactivity that we tolerate. If we
+			 * see dirty data again later, the task will get
+			 * recreated automatically.
+			 */
+			max_idle = max(5UL * 60 * HZ, wait_jiffies);
+			if (time_after(jiffies, max_idle + last_active))
+				break;
 		}
-		spin_unlock(&inode_lock);
-		iput(old_inode);
-	} else
-		spin_unlock(&inode_lock);
 
-	return;		/* Leave any unwritten inodes on s_io */
+		wait_jiffies = msecs_to_jiffies(dirty_writeback_interval * 10);
+		schedule_timeout_interruptible(wait_jiffies);
+		try_to_freeze();
+	}
+
+	return 0;
 }
-EXPORT_SYMBOL_GPL(generic_sync_sb_inodes);
 
-static void sync_sb_inodes(struct super_block *sb,
-				struct writeback_control *wbc)
+/*
+ * Schedule writeback for all backing devices. This does WB_SYNC_NONE
+ * writeback, for integrity writeback see bdi_sync_writeback().
+ */
+static void bdi_writeback_all(struct super_block *sb, long nr_pages)
 {
-	generic_sync_sb_inodes(sb, wbc);
+	struct wb_writeback_args args = {
+		.sb		= sb,
+		.nr_pages	= nr_pages,
+		.sync_mode	= WB_SYNC_NONE,
+	};
+	struct backing_dev_info *bdi;
+
+	rcu_read_lock();
+
+	list_for_each_entry_rcu(bdi, &bdi_list, bdi_list) {
+		if (!bdi_has_dirty_io(bdi))
+			continue;
+
+		bdi_alloc_queue_work(bdi, &args);
+	}
+
+	rcu_read_unlock();
 }
 
 /*
- * Start writeback of dirty pagecache data against all unlocked inodes.
+ * Start writeback of `nr_pages' pages.  If `nr_pages' is zero, write back
+ * the whole world.
+ */
+void wakeup_flusher_threads(long nr_pages)
+{
+	if (nr_pages == 0)
+		nr_pages = global_page_state(NR_FILE_DIRTY) +
+				global_page_state(NR_UNSTABLE_NFS);
+	bdi_writeback_all(NULL, nr_pages);
+}
+
+static noinline void block_dump___mark_inode_dirty(struct inode *inode)
+{
+	if (inode->i_ino || strcmp(inode->i_sb->s_id, "bdev")) {
+		struct dentry *dentry;
+		const char *name = "?";
+
+		dentry = d_find_alias(inode);
+		if (dentry) {
+			spin_lock(&dentry->d_lock);
+			name = (const char *) dentry->d_name.name;
+		}
+		printk(KERN_DEBUG
+		       "%s(%d): dirtied inode %lu (%s) on %s\n",
+		       current->comm, task_pid_nr(current), inode->i_ino,
+		       name, inode->i_sb->s_id);
+		if (dentry) {
+			spin_unlock(&dentry->d_lock);
+			dput(dentry);
+		}
+	}
+}
+
+/**
+ *	__mark_inode_dirty -	internal function
+ *	@inode: inode to mark
+ *	@flags: what kind of dirty (i.e. I_DIRTY_SYNC)
+ *	Mark an inode as dirty. Callers should use mark_inode_dirty or
+ *  	mark_inode_dirty_sync.
+ *
+ * Put the inode on the super block's dirty list.
  *
- * Note:
- * We don't need to grab a reference to superblock here. If it has non-empty
- * ->s_dirty it's hadn't been killed yet and kill_super() won't proceed
- * past sync_inodes_sb() until the ->s_dirty/s_io/s_more_io lists are all
- * empty. Since __sync_single_inode() regains inode_lock before it finally moves
- * inode from superblock lists we are OK.
+ * CAREFUL! We mark it dirty unconditionally, but move it onto the
+ * dirty list only if it is hashed or if it refers to a blockdev.
+ * If it was not hashed, it will never be added to the dirty list
+ * even if it is later hashed, as it will have been marked dirty already.
  *
- * If `older_than_this' is non-zero then only flush inodes which have a
- * flushtime older than *older_than_this.
+ * In short, make sure you hash any inodes _before_ you start marking
+ * them dirty.
  *
- * If `bdi' is non-zero then we will scan the first inode against each
- * superblock until we find the matching ones.  One group will be the dirty
- * inodes against a filesystem.  Then when we hit the dummy blockdev superblock,
- * sync_sb_inodes will seekout the blockdev which matches `bdi'.  Maybe not
- * super-efficient but we're about to do a ton of I/O...
+ * This function *must* be atomic for the I_DIRTY_PAGES case -
+ * set_page_dirty() is called under spinlock in several places.
+ *
+ * Note that for blockdevs, inode->dirtied_when represents the dirtying time of
+ * the block-special inode (/dev/hda1) itself.  And the ->dirtied_when field of
+ * the kernel-internal blockdev inode represents the dirtying time of the
+ * blockdev's pages.  This is why for I_DIRTY_PAGES we always use
+ * page->mapping->host, so the page-dirtying time is recorded in the internal
+ * blockdev inode.
  */
-void
-writeback_inodes(struct writeback_control *wbc)
+void __mark_inode_dirty(struct inode *inode, int flags)
 {
-	struct super_block *sb;
+	struct super_block *sb = inode->i_sb;
 
-	might_sleep();
-	spin_lock(&sb_lock);
-restart:
-	list_for_each_entry_reverse(sb, &super_blocks, s_list) {
-		if (sb_has_dirty_inodes(sb)) {
-			/* we're making our own get_super here */
-			sb->s_count++;
-			spin_unlock(&sb_lock);
-			/*
-			 * If we can't get the readlock, there's no sense in
-			 * waiting around, most of the time the FS is going to
-			 * be unmounted by the time it is released.
-			 */
-			if (down_read_trylock(&sb->s_umount)) {
-				if (sb->s_root)
-					sync_sb_inodes(sb, wbc);
-				up_read(&sb->s_umount);
+	/*
+	 * Don't do this for I_DIRTY_PAGES - that doesn't actually
+	 * dirty the inode itself
+	 */
+	if (flags & (I_DIRTY_SYNC | I_DIRTY_DATASYNC)) {
+		if (sb->s_op->dirty_inode)
+			sb->s_op->dirty_inode(inode);
+	}
+
+	/*
+	 * make sure that changes are seen by all cpus before we test i_state
+	 * -- mikulas
+	 */
+	smp_mb();
+
+	/* avoid the locking if we can */
+	if ((inode->i_state & flags) == flags)
+		return;
+
+	if (unlikely(block_dump))
+		block_dump___mark_inode_dirty(inode);
+
+	spin_lock(&inode_lock);
+	if ((inode->i_state & flags) != flags) {
+		const int was_dirty = inode->i_state & I_DIRTY;
+
+		inode->i_state |= flags;
+
+		/*
+		 * If the inode is being synced, just update its dirty state.
+		 * The unlocker will place the inode on the appropriate
+		 * superblock list, based upon its state.
+		 */
+		if (inode->i_state & I_SYNC)
+			goto out;
+
+		/*
+		 * Only add valid (hashed) inodes to the superblock's
+		 * dirty list.  Add blockdev inodes as well.
+		 */
+		if (!S_ISBLK(inode->i_mode)) {
+			if (hlist_unhashed(&inode->i_hash))
+				goto out;
+		}
+		if (inode->i_state & (I_FREEING|I_CLEAR))
+			goto out;
+
+		/*
+		 * If the inode was already on b_dirty/b_io/b_more_io, don't
+		 * reposition it (that would break b_dirty time-ordering).
+		 */
+		if (!was_dirty) {
+			struct bdi_writeback *wb = &inode_to_bdi(inode)->wb;
+			struct backing_dev_info *bdi = wb->bdi;
+
+			if (bdi_cap_writeback_dirty(bdi) &&
+			    !test_bit(BDI_registered, &bdi->state)) {
+				WARN_ON(1);
+				printk(KERN_ERR "bdi-%s not registered\n",
+								bdi->name);
 			}
-			spin_lock(&sb_lock);
-			if (__put_super_and_need_restart(sb))
-				goto restart;
+
+			inode->dirtied_when = jiffies;
+			list_move(&inode->i_list, &wb->b_dirty);
 		}
-		if (wbc->nr_to_write <= 0)
-			break;
 	}
-	spin_unlock(&sb_lock);
+out:
+	spin_unlock(&inode_lock);
 }
+EXPORT_SYMBOL(__mark_inode_dirty);
 
 /*
- * writeback and wait upon the filesystem's dirty inodes.  The caller will
- * do this in two passes - one to write, and one to wait.
+ * Write out a superblock's list of dirty inodes.  A wait will be performed
+ * upon no inodes, all inodes or the final one, depending upon sync_mode.
+ *
+ * If older_than_this is non-NULL, then only write out inodes which
+ * had their first dirtying at a time earlier than *older_than_this.
+ *
+ * If we're a pdlfush thread, then implement pdflush collision avoidance
+ * against the entire list.
  *
- * A finite limit is set on the number of pages which will be written.
- * To prevent infinite livelock of sys_sync().
+ * 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,
+ * assume that all inodes are backed by the same queue.
  *
- * We add in the number of potentially dirty inodes, because each inode write
- * can dirty pagecache in the underlying blockdev.
+ * The inodes to be written are parked on bdi->b_io.  They are moved back onto
+ * bdi->b_dirty as they are selected for writing.  This way, none can be missed
+ * on the writer throttling path, and we get decent balancing between many
+ * throttled threads: we don't want them all piling up on inode_sync_wait.
  */
-void sync_inodes_sb(struct super_block *sb, int wait)
+static void wait_sb_inodes(struct super_block *sb)
 {
-	struct writeback_control wbc = {
-		.sync_mode	= wait ? WB_SYNC_ALL : WB_SYNC_NONE,
-		.range_start	= 0,
-		.range_end	= LLONG_MAX,
-	};
+	struct inode *inode, *old_inode = NULL;
+
+	/*
+	 * We need to be protected against the filesystem going from
+	 * r/o to r/w or vice versa.
+	 */
+	WARN_ON(!rwsem_is_locked(&sb->s_umount));
+
+	spin_lock(&inode_lock);
+
+	/*
+	 * 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_CLEAR|I_WILL_FREE|I_NEW))
+			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();
 
-	if (!wait) {
-		unsigned long nr_dirty = global_page_state(NR_FILE_DIRTY);
-		unsigned long nr_unstable = global_page_state(NR_UNSTABLE_NFS);
+		spin_lock(&inode_lock);
+	}
+	spin_unlock(&inode_lock);
+	iput(old_inode);
+}
 
-		wbc.nr_to_write = nr_dirty + nr_unstable +
+/**
+ * writeback_inodes_sb	-	writeback dirty inodes from given super_block
+ * @sb: the superblock
+ *
+ * Start writeback on some inodes on this super_block. No guarantees are made
+ * on how many (if any) will be written, and this function does not wait
+ * for IO completion of submitted IO. The number of pages submitted is
+ * returned.
+ */
+void writeback_inodes_sb(struct super_block *sb)
+{
+	unsigned long nr_dirty = global_page_state(NR_FILE_DIRTY);
+	unsigned long nr_unstable = global_page_state(NR_UNSTABLE_NFS);
+	long nr_to_write;
+
+	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);
+	bdi_writeback_all(sb, nr_to_write);
+}
+EXPORT_SYMBOL(writeback_inodes_sb);
+
+/**
+ * sync_inodes_sb	-	sync sb inode pages
+ * @sb: the superblock
+ *
+ * This function writes and waits on any dirty inode belonging to this
+ * super_block. The number of pages synced is returned.
+ */
+void sync_inodes_sb(struct super_block *sb)
+{
+	bdi_sync_writeback(sb->s_bdi, sb);
+	wait_sb_inodes(sb);
 }
+EXPORT_SYMBOL(sync_inodes_sb);
 
 /**
  * write_inode_now	-	write an inode to disk
@@ -737,57 +1213,3 @@ int sync_inode(struct inode *inode, struct writeback_control *wbc)
 	return ret;
 }
 EXPORT_SYMBOL(sync_inode);
-
-/**
- * generic_osync_inode - flush all dirty data for a given inode to disk
- * @inode: inode to write
- * @mapping: the address_space that should be flushed
- * @what:  what to write and wait upon
- *
- * This can be called by file_write functions for files which have the
- * O_SYNC flag set, to flush dirty writes to disk.
- *
- * @what is a bitmask, specifying which part of the inode's data should be
- * written and waited upon.
- *
- *    OSYNC_DATA:     i_mapping's dirty data
- *    OSYNC_METADATA: the buffers at i_mapping->private_list
- *    OSYNC_INODE:    the inode itself
- */
-
-int generic_osync_inode(struct inode *inode, struct address_space *mapping, int what)
-{
-	int err = 0;
-	int need_write_inode_now = 0;
-	int err2;
-
-	if (what & OSYNC_DATA)
-		err = filemap_fdatawrite(mapping);
-	if (what & (OSYNC_METADATA|OSYNC_DATA)) {
-		err2 = sync_mapping_buffers(mapping);
-		if (!err)
-			err = err2;
-	}
-	if (what & OSYNC_DATA) {
-		err2 = filemap_fdatawait(mapping);
-		if (!err)
-			err = err2;
-	}
-
-	spin_lock(&inode_lock);
-	if ((inode->i_state & I_DIRTY) &&
-	    ((what & OSYNC_INODE) || (inode->i_state & I_DIRTY_DATASYNC)))
-		need_write_inode_now = 1;
-	spin_unlock(&inode_lock);
-
-	if (need_write_inode_now) {
-		err2 = write_inode_now(inode, 1);
-		if (!err)
-			err = err2;
-	}
-	else
-		inode_sync_wait(inode);
-
-	return err;
-}
-EXPORT_SYMBOL(generic_osync_inode);
diff --git a/fs/fuse/control.c b/fs/fuse/control.c
index 99c99dfb037..3773fd63d2f 100644
--- a/fs/fuse/control.c
+++ b/fs/fuse/control.c
@@ -61,6 +61,121 @@ static ssize_t fuse_conn_waiting_read(struct file *file, char __user *buf,
 	return simple_read_from_buffer(buf, len, ppos, tmp, size);
 }
 
+static ssize_t fuse_conn_limit_read(struct file *file, char __user *buf,
+				    size_t len, loff_t *ppos, unsigned val)
+{
+	char tmp[32];
+	size_t size = sprintf(tmp, "%u\n", val);
+
+	return simple_read_from_buffer(buf, len, ppos, tmp, size);
+}
+
+static ssize_t fuse_conn_limit_write(struct file *file, const char __user *buf,
+				     size_t count, loff_t *ppos, unsigned *val,
+				     unsigned global_limit)
+{
+	unsigned long t;
+	char tmp[32];
+	unsigned limit = (1 << 16) - 1;
+	int err;
+
+	if (*ppos || count >= sizeof(tmp) - 1)
+		return -EINVAL;
+
+	if (copy_from_user(tmp, buf, count))
+		return -EINVAL;
+
+	tmp[count] = '\0';
+
+	err = strict_strtoul(tmp, 0, &t);
+	if (err)
+		return err;
+
+	if (!capable(CAP_SYS_ADMIN))
+		limit = min(limit, global_limit);
+
+	if (t > limit)
+		return -EINVAL;
+
+	*val = t;
+
+	return count;
+}
+
+static ssize_t fuse_conn_max_background_read(struct file *file,
+					     char __user *buf, size_t len,
+					     loff_t *ppos)
+{
+	struct fuse_conn *fc;
+	unsigned val;
+
+	fc = fuse_ctl_file_conn_get(file);
+	if (!fc)
+		return 0;
+
+	val = fc->max_background;
+	fuse_conn_put(fc);
+
+	return fuse_conn_limit_read(file, buf, len, ppos, val);
+}
+
+static ssize_t fuse_conn_max_background_write(struct file *file,
+					      const char __user *buf,
+					      size_t count, loff_t *ppos)
+{
+	unsigned val;
+	ssize_t ret;
+
+	ret = fuse_conn_limit_write(file, buf, count, ppos, &val,
+				    max_user_bgreq);
+	if (ret > 0) {
+		struct fuse_conn *fc = fuse_ctl_file_conn_get(file);
+		if (fc) {
+			fc->max_background = val;
+			fuse_conn_put(fc);
+		}
+	}
+
+	return ret;
+}
+
+static ssize_t fuse_conn_congestion_threshold_read(struct file *file,
+						   char __user *buf, size_t len,
+						   loff_t *ppos)
+{
+	struct fuse_conn *fc;
+	unsigned val;
+
+	fc = fuse_ctl_file_conn_get(file);
+	if (!fc)
+		return 0;
+
+	val = fc->congestion_threshold;
+	fuse_conn_put(fc);
+
+	return fuse_conn_limit_read(file, buf, len, ppos, val);
+}
+
+static ssize_t fuse_conn_congestion_threshold_write(struct file *file,
+						    const char __user *buf,
+						    size_t count, loff_t *ppos)
+{
+	unsigned val;
+	ssize_t ret;
+
+	ret = fuse_conn_limit_write(file, buf, count, ppos, &val,
+				    max_user_congthresh);
+	if (ret > 0) {
+		struct fuse_conn *fc = fuse_ctl_file_conn_get(file);
+		if (fc) {
+			fc->congestion_threshold = val;
+			fuse_conn_put(fc);
+		}
+	}
+
+	return ret;
+}
+
 static const struct file_operations fuse_ctl_abort_ops = {
 	.open = nonseekable_open,
 	.write = fuse_conn_abort_write,
@@ -71,6 +186,18 @@ static const struct file_operations fuse_ctl_waiting_ops = {
 	.read = fuse_conn_waiting_read,
 };
 
+static const struct file_operations fuse_conn_max_background_ops = {
+	.open = nonseekable_open,
+	.read = fuse_conn_max_background_read,
+	.write = fuse_conn_max_background_write,
+};
+
+static const struct file_operations fuse_conn_congestion_threshold_ops = {
+	.open = nonseekable_open,
+	.read = fuse_conn_congestion_threshold_read,
+	.write = fuse_conn_congestion_threshold_write,
+};
+
 static struct dentry *fuse_ctl_add_dentry(struct dentry *parent,
 					  struct fuse_conn *fc,
 					  const char *name,
@@ -127,9 +254,14 @@ int fuse_ctl_add_conn(struct fuse_conn *fc)
 		goto err;
 
 	if (!fuse_ctl_add_dentry(parent, fc, "waiting", S_IFREG | 0400, 1,
-				NULL, &fuse_ctl_waiting_ops) ||
+				 NULL, &fuse_ctl_waiting_ops) ||
 	    !fuse_ctl_add_dentry(parent, fc, "abort", S_IFREG | 0200, 1,
-				 NULL, &fuse_ctl_abort_ops))
+				 NULL, &fuse_ctl_abort_ops) ||
+	    !fuse_ctl_add_dentry(parent, fc, "max_background", S_IFREG | 0600,
+				 1, NULL, &fuse_conn_max_background_ops) ||
+	    !fuse_ctl_add_dentry(parent, fc, "congestion_threshold",
+				 S_IFREG | 0600, 1, NULL,
+				 &fuse_conn_congestion_threshold_ops))
 		goto err;
 
 	return 0;
@@ -156,7 +288,7 @@ void fuse_ctl_remove_conn(struct fuse_conn *fc)
 		d_drop(dentry);
 		dput(dentry);
 	}
-	fuse_control_sb->s_root->d_inode->i_nlink--;
+	drop_nlink(fuse_control_sb->s_root->d_inode);
 }
 
 static int fuse_ctl_fill_super(struct super_block *sb, void *data, int silent)
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 6484eb75acd..51d9e33d634 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -250,7 +250,7 @@ static void queue_request(struct fuse_conn *fc, struct fuse_req *req)
 
 static void flush_bg_queue(struct fuse_conn *fc)
 {
-	while (fc->active_background < FUSE_MAX_BACKGROUND &&
+	while (fc->active_background < fc->max_background &&
 	       !list_empty(&fc->bg_queue)) {
 		struct fuse_req *req;
 
@@ -280,11 +280,11 @@ __releases(&fc->lock)
 	list_del(&req->intr_entry);
 	req->state = FUSE_REQ_FINISHED;
 	if (req->background) {
-		if (fc->num_background == FUSE_MAX_BACKGROUND) {
+		if (fc->num_background == fc->max_background) {
 			fc->blocked = 0;
 			wake_up_all(&fc->blocked_waitq);
 		}
-		if (fc->num_background == FUSE_CONGESTION_THRESHOLD &&
+		if (fc->num_background == fc->congestion_threshold &&
 		    fc->connected && fc->bdi_initialized) {
 			clear_bdi_congested(&fc->bdi, BLK_RW_SYNC);
 			clear_bdi_congested(&fc->bdi, BLK_RW_ASYNC);
@@ -410,9 +410,9 @@ static void fuse_request_send_nowait_locked(struct fuse_conn *fc,
 {
 	req->background = 1;
 	fc->num_background++;
-	if (fc->num_background == FUSE_MAX_BACKGROUND)
+	if (fc->num_background == fc->max_background)
 		fc->blocked = 1;
-	if (fc->num_background == FUSE_CONGESTION_THRESHOLD &&
+	if (fc->num_background == fc->congestion_threshold &&
 	    fc->bdi_initialized) {
 		set_bdi_congested(&fc->bdi, BLK_RW_SYNC);
 		set_bdi_congested(&fc->bdi, BLK_RW_ASYNC);
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 52b641fc0fa..fc9c79feb5f 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -25,12 +25,6 @@
 /** Max number of pages that can be used in a single read request */
 #define FUSE_MAX_PAGES_PER_REQ 32
 
-/** Maximum number of outstanding background requests */
-#define FUSE_MAX_BACKGROUND 12
-
-/** Congestion starts at 75% of maximum */
-#define FUSE_CONGESTION_THRESHOLD (FUSE_MAX_BACKGROUND * 75 / 100)
-
 /** Bias for fi->writectr, meaning new writepages must not be sent */
 #define FUSE_NOWRITE INT_MIN
 
@@ -38,7 +32,7 @@
 #define FUSE_NAME_MAX 1024
 
 /** Number of dentries for each connection in the control filesystem */
-#define FUSE_CTL_NUM_DENTRIES 3
+#define FUSE_CTL_NUM_DENTRIES 5
 
 /** If the FUSE_DEFAULT_PERMISSIONS flag is given, the filesystem
     module will check permissions based on the file mode.  Otherwise no
@@ -55,6 +49,10 @@ extern struct list_head fuse_conn_list;
 /** Global mutex protecting fuse_conn_list and the control filesystem */
 extern struct mutex fuse_mutex;
 
+/** Module parameters */
+extern unsigned max_user_bgreq;
+extern unsigned max_user_congthresh;
+
 /** FUSE inode */
 struct fuse_inode {
 	/** Inode data */
@@ -349,6 +347,12 @@ struct fuse_conn {
 	/** rbtree of fuse_files waiting for poll events indexed by ph */
 	struct rb_root polled_files;
 
+	/** Maximum number of outstanding background requests */
+	unsigned max_background;
+
+	/** Number of background requests at which congestion starts */
+	unsigned congestion_threshold;
+
 	/** Number of requests currently in the background */
 	unsigned num_background;
 
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index f91ccc4a189..6da947daabd 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -14,6 +14,7 @@
 #include <linux/seq_file.h>
 #include <linux/init.h>
 #include <linux/module.h>
+#include <linux/moduleparam.h>
 #include <linux/parser.h>
 #include <linux/statfs.h>
 #include <linux/random.h>
@@ -28,10 +29,34 @@ static struct kmem_cache *fuse_inode_cachep;
 struct list_head fuse_conn_list;
 DEFINE_MUTEX(fuse_mutex);
 
+static int set_global_limit(const char *val, struct kernel_param *kp);
+
+unsigned max_user_bgreq;
+module_param_call(max_user_bgreq, set_global_limit, param_get_uint,
+		  &max_user_bgreq, 0644);
+__MODULE_PARM_TYPE(max_user_bgreq, "uint");
+MODULE_PARM_DESC(max_user_bgreq,
+ "Global limit for the maximum number of backgrounded requests an "
+ "unprivileged user can set");
+
+unsigned max_user_congthresh;
+module_param_call(max_user_congthresh, set_global_limit, param_get_uint,
+		  &max_user_congthresh, 0644);
+__MODULE_PARM_TYPE(max_user_congthresh, "uint");
+MODULE_PARM_DESC(max_user_congthresh,
+ "Global limit for the maximum congestion threshold an "
+ "unprivileged user can set");
+
 #define FUSE_SUPER_MAGIC 0x65735546
 
 #define FUSE_DEFAULT_BLKSIZE 512
 
+/** Maximum number of outstanding background requests */
+#define FUSE_DEFAULT_MAX_BACKGROUND 12
+
+/** Congestion starts at 75% of maximum */
+#define FUSE_DEFAULT_CONGESTION_THRESHOLD (FUSE_DEFAULT_MAX_BACKGROUND * 3 / 4)
+
 struct fuse_mount_data {
 	int fd;
 	unsigned rootmode;
@@ -517,6 +542,8 @@ void fuse_conn_init(struct fuse_conn *fc)
 	INIT_LIST_HEAD(&fc->bg_queue);
 	INIT_LIST_HEAD(&fc->entry);
 	atomic_set(&fc->num_waiting, 0);
+	fc->max_background = FUSE_DEFAULT_MAX_BACKGROUND;
+	fc->congestion_threshold = FUSE_DEFAULT_CONGESTION_THRESHOLD;
 	fc->khctr = 0;
 	fc->polled_files = RB_ROOT;
 	fc->reqctr = 0;
@@ -727,6 +754,54 @@ static const struct super_operations fuse_super_operations = {
 	.show_options	= fuse_show_options,
 };
 
+static void sanitize_global_limit(unsigned *limit)
+{
+	if (*limit == 0)
+		*limit = ((num_physpages << PAGE_SHIFT) >> 13) /
+			 sizeof(struct fuse_req);
+
+	if (*limit >= 1 << 16)
+		*limit = (1 << 16) - 1;
+}
+
+static int set_global_limit(const char *val, struct kernel_param *kp)
+{
+	int rv;
+
+	rv = param_set_uint(val, kp);
+	if (rv)
+		return rv;
+
+	sanitize_global_limit((unsigned *)kp->arg);
+
+	return 0;
+}
+
+static void process_init_limits(struct fuse_conn *fc, struct fuse_init_out *arg)
+{
+	int cap_sys_admin = capable(CAP_SYS_ADMIN);
+
+	if (arg->minor < 13)
+		return;
+
+	sanitize_global_limit(&max_user_bgreq);
+	sanitize_global_limit(&max_user_congthresh);
+
+	if (arg->max_background) {
+		fc->max_background = arg->max_background;
+
+		if (!cap_sys_admin && fc->max_background > max_user_bgreq)
+			fc->max_background = max_user_bgreq;
+	}
+	if (arg->congestion_threshold) {
+		fc->congestion_threshold = arg->congestion_threshold;
+
+		if (!cap_sys_admin &&
+		    fc->congestion_threshold > max_user_congthresh)
+			fc->congestion_threshold = max_user_congthresh;
+	}
+}
+
 static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
 {
 	struct fuse_init_out *arg = &req->misc.init_out;
@@ -736,6 +811,8 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
 	else {
 		unsigned long ra_pages;
 
+		process_init_limits(fc, arg);
+
 		if (arg->minor >= 6) {
 			ra_pages = arg->max_readahead / PAGE_CACHE_SIZE;
 			if (arg->flags & FUSE_ASYNC_READ)
@@ -801,6 +878,7 @@ static int fuse_bdi_init(struct fuse_conn *fc, struct super_block *sb)
 {
 	int err;
 
+	fc->bdi.name = "fuse";
 	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 */
@@ -893,6 +971,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent)
 	if (err)
 		goto err_put_conn;
 
+	sb->s_bdi = &fc->bdi;
+
 	/* Handle umasking inside the fuse code */
 	if (sb->s_flags & MS_POSIXACL)
 		fc->dont_mask = 1;
@@ -1147,6 +1227,9 @@ static int __init fuse_init(void)
 	if (res)
 		goto err_sysfs_cleanup;
 
+	sanitize_global_limit(&max_user_bgreq);
+	sanitize_global_limit(&max_user_congthresh);
+
 	return 0;
 
  err_sysfs_cleanup:
diff --git a/fs/gfs2/Makefile b/fs/gfs2/Makefile
index 3da2f1f4f73..21f7e46da4c 100644
--- a/fs/gfs2/Makefile
+++ b/fs/gfs2/Makefile
@@ -1,6 +1,6 @@
 EXTRA_CFLAGS := -I$(src)
 obj-$(CONFIG_GFS2_FS) += gfs2.o
-gfs2-y := acl.o bmap.o dir.o eaops.o eattr.o glock.o \
+gfs2-y := acl.o bmap.o dir.o xattr.o glock.o \
 	glops.o inode.o log.o lops.o main.o meta_io.o \
 	aops.o dentry.o export.o file.o \
 	ops_fstype.o ops_inode.o quota.o \
diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c
index fa881bdc3d8..3fc4e3ac7d8 100644
--- a/fs/gfs2/acl.c
+++ b/fs/gfs2/acl.c
@@ -19,8 +19,7 @@
 #include "gfs2.h"
 #include "incore.h"
 #include "acl.h"
-#include "eaops.h"
-#include "eattr.h"
+#include "xattr.h"
 #include "glock.h"
 #include "inode.h"
 #include "meta_io.h"
@@ -31,8 +30,7 @@
 #define ACL_DEFAULT 0
 
 int gfs2_acl_validate_set(struct gfs2_inode *ip, int access,
-		      struct gfs2_ea_request *er,
-		      int *remove, mode_t *mode)
+			  struct gfs2_ea_request *er, int *remove, mode_t *mode)
 {
 	struct posix_acl *acl;
 	int error;
@@ -83,30 +81,20 @@ int gfs2_acl_validate_remove(struct gfs2_inode *ip, int access)
 	return 0;
 }
 
-static int acl_get(struct gfs2_inode *ip, int access, struct posix_acl **acl,
-		   struct gfs2_ea_location *el, char **data, unsigned int *len)
+static int acl_get(struct gfs2_inode *ip, const char *name,
+		   struct posix_acl **acl, struct gfs2_ea_location *el,
+		   char **datap, unsigned int *lenp)
 {
-	struct gfs2_ea_request er;
-	struct gfs2_ea_location el_this;
+	char *data;
+	unsigned int len;
 	int error;
 
+	el->el_bh = NULL;
+
 	if (!ip->i_eattr)
 		return 0;
 
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-	if (access) {
-		er.er_name = GFS2_POSIX_ACL_ACCESS;
-		er.er_name_len = GFS2_POSIX_ACL_ACCESS_LEN;
-	} else {
-		er.er_name = GFS2_POSIX_ACL_DEFAULT;
-		er.er_name_len = GFS2_POSIX_ACL_DEFAULT_LEN;
-	}
-	er.er_type = GFS2_EATYPE_SYS;
-
-	if (!el)
-		el = &el_this;
-
-	error = gfs2_ea_find(ip, &er, el);
+	error = gfs2_ea_find(ip, GFS2_EATYPE_SYS, name, el);
 	if (error)
 		return error;
 	if (!el->el_ea)
@@ -114,32 +102,31 @@ static int acl_get(struct gfs2_inode *ip, int access, struct posix_acl **acl,
 	if (!GFS2_EA_DATA_LEN(el->el_ea))
 		goto out;
 
-	er.er_data_len = GFS2_EA_DATA_LEN(el->el_ea);
-	er.er_data = kmalloc(er.er_data_len, GFP_NOFS);
+	len = GFS2_EA_DATA_LEN(el->el_ea);
+	data = kmalloc(len, GFP_NOFS);
 	error = -ENOMEM;
-	if (!er.er_data)
+	if (!data)
 		goto out;
 
-	error = gfs2_ea_get_copy(ip, el, er.er_data);
-	if (error)
+	error = gfs2_ea_get_copy(ip, el, data, len);
+	if (error < 0)
 		goto out_kfree;
+	error = 0;
 
 	if (acl) {
-		*acl = posix_acl_from_xattr(er.er_data, er.er_data_len);
+		*acl = posix_acl_from_xattr(data, len);
 		if (IS_ERR(*acl))
 			error = PTR_ERR(*acl);
 	}
 
 out_kfree:
-	if (error || !data)
-		kfree(er.er_data);
-	else {
-		*data = er.er_data;
-		*len = er.er_data_len;
+	if (error || !datap) {
+		kfree(data);
+	} else {
+		*datap = data;
+		*lenp = len;
 	}
 out:
-	if (error || el == &el_this)
-		brelse(el->el_bh);
 	return error;
 }
 
@@ -153,10 +140,12 @@ out:
 
 int gfs2_check_acl(struct inode *inode, int mask)
 {
+	struct gfs2_ea_location el;
 	struct posix_acl *acl = NULL;
 	int error;
 
-	error = acl_get(GFS2_I(inode), ACL_ACCESS, &acl, NULL, NULL, NULL);
+	error = acl_get(GFS2_I(inode), GFS2_POSIX_ACL_ACCESS, &acl, &el, NULL, NULL);
+	brelse(el.el_bh);
 	if (error)
 		return error;
 
@@ -196,10 +185,12 @@ static int munge_mode(struct gfs2_inode *ip, mode_t mode)
 
 int gfs2_acl_create(struct gfs2_inode *dip, struct gfs2_inode *ip)
 {
+	struct gfs2_ea_location el;
 	struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode);
 	struct posix_acl *acl = NULL, *clone;
-	struct gfs2_ea_request er;
 	mode_t mode = ip->i_inode.i_mode;
+	char *data = NULL;
+	unsigned int len;
 	int error;
 
 	if (!sdp->sd_args.ar_posix_acl)
@@ -207,11 +198,8 @@ int gfs2_acl_create(struct gfs2_inode *dip, struct gfs2_inode *ip)
 	if (S_ISLNK(ip->i_inode.i_mode))
 		return 0;
 
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-	er.er_type = GFS2_EATYPE_SYS;
-
-	error = acl_get(dip, ACL_DEFAULT, &acl, NULL,
-			&er.er_data, &er.er_data_len);
+	error = acl_get(dip, GFS2_POSIX_ACL_DEFAULT, &acl, &el, &data, &len);
+	brelse(el.el_bh);
 	if (error)
 		return error;
 	if (!acl) {
@@ -229,9 +217,8 @@ int gfs2_acl_create(struct gfs2_inode *dip, struct gfs2_inode *ip)
 	acl = clone;
 
 	if (S_ISDIR(ip->i_inode.i_mode)) {
-		er.er_name = GFS2_POSIX_ACL_DEFAULT;
-		er.er_name_len = GFS2_POSIX_ACL_DEFAULT_LEN;
-		error = gfs2_system_eaops.eo_set(ip, &er);
+		error = gfs2_xattr_set(&ip->i_inode, GFS2_EATYPE_SYS,
+				       GFS2_POSIX_ACL_DEFAULT, data, len, 0);
 		if (error)
 			goto out;
 	}
@@ -239,21 +226,19 @@ int gfs2_acl_create(struct gfs2_inode *dip, struct gfs2_inode *ip)
 	error = posix_acl_create_masq(acl, &mode);
 	if (error < 0)
 		goto out;
-	if (error > 0) {
-		er.er_name = GFS2_POSIX_ACL_ACCESS;
-		er.er_name_len = GFS2_POSIX_ACL_ACCESS_LEN;
-		posix_acl_to_xattr(acl, er.er_data, er.er_data_len);
-		er.er_mode = mode;
-		er.er_flags = GFS2_ERF_MODE;
-		error = gfs2_system_eaops.eo_set(ip, &er);
-		if (error)
-			goto out;
-	} else
-		munge_mode(ip, mode);
+	if (error == 0)
+		goto munge;
 
+	posix_acl_to_xattr(acl, data, len);
+	error = gfs2_xattr_set(&ip->i_inode, GFS2_EATYPE_SYS,
+			       GFS2_POSIX_ACL_ACCESS, data, len, 0);
+	if (error)
+		goto out;
+munge:
+	error = munge_mode(ip, mode);
 out:
 	posix_acl_release(acl);
-	kfree(er.er_data);
+	kfree(data);
 	return error;
 }
 
@@ -265,9 +250,9 @@ int gfs2_acl_chmod(struct gfs2_inode *ip, struct iattr *attr)
 	unsigned int len;
 	int error;
 
-	error = acl_get(ip, ACL_ACCESS, &acl, &el, &data, &len);
+	error = acl_get(ip, GFS2_POSIX_ACL_ACCESS, &acl, &el, &data, &len);
 	if (error)
-		return error;
+		goto out_brelse;
 	if (!acl)
 		return gfs2_setattr_simple(ip, attr);
 
@@ -286,8 +271,9 @@ int gfs2_acl_chmod(struct gfs2_inode *ip, struct iattr *attr)
 
 out:
 	posix_acl_release(acl);
-	brelse(el.el_bh);
 	kfree(data);
+out_brelse:
+	brelse(el.el_bh);
 	return error;
 }
 
diff --git a/fs/gfs2/dentry.c b/fs/gfs2/dentry.c
index 022c66cd560..91beddadd38 100644
--- a/fs/gfs2/dentry.c
+++ b/fs/gfs2/dentry.c
@@ -107,8 +107,26 @@ static int gfs2_dhash(struct dentry *dentry, struct qstr *str)
 	return 0;
 }
 
+static int gfs2_dentry_delete(struct dentry *dentry)
+{
+	struct gfs2_inode *ginode;
+
+	if (!dentry->d_inode)
+		return 0;
+
+	ginode = GFS2_I(dentry->d_inode);
+	if (!ginode->i_iopen_gh.gh_gl)
+		return 0;
+
+	if (test_bit(GLF_DEMOTE, &ginode->i_iopen_gh.gh_gl->gl_flags))
+		return 1;
+
+	return 0;
+}
+
 const struct dentry_operations gfs2_dops = {
 	.d_revalidate = gfs2_drevalidate,
 	.d_hash = gfs2_dhash,
+	.d_delete = gfs2_dentry_delete,
 };
 
diff --git a/fs/gfs2/eaops.c b/fs/gfs2/eaops.c
deleted file mode 100644
index dee9b03e5b3..00000000000
--- a/fs/gfs2/eaops.c
+++ /dev/null
@@ -1,157 +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/slab.h>
-#include <linux/spinlock.h>
-#include <linux/completion.h>
-#include <linux/buffer_head.h>
-#include <linux/capability.h>
-#include <linux/xattr.h>
-#include <linux/gfs2_ondisk.h>
-#include <asm/uaccess.h>
-
-#include "gfs2.h"
-#include "incore.h"
-#include "acl.h"
-#include "eaops.h"
-#include "eattr.h"
-#include "util.h"
-
-/**
- * gfs2_ea_name2type - get the type of the ea, and truncate type from the name
- * @namep: ea name, possibly with type appended
- *
- * Returns: GFS2_EATYPE_XXX
- */
-
-unsigned int gfs2_ea_name2type(const char *name, const char **truncated_name)
-{
-	unsigned int type;
-
-	if (strncmp(name, "system.", 7) == 0) {
-		type = GFS2_EATYPE_SYS;
-		if (truncated_name)
-			*truncated_name = name + sizeof("system.") - 1;
-	} else if (strncmp(name, "user.", 5) == 0) {
-		type = GFS2_EATYPE_USR;
-		if (truncated_name)
-			*truncated_name = name + sizeof("user.") - 1;
-	} else if (strncmp(name, "security.", 9) == 0) {
-		type = GFS2_EATYPE_SECURITY;
-		if (truncated_name)
-			*truncated_name = name + sizeof("security.") - 1;
-	} else {
-		type = GFS2_EATYPE_UNUSED;
-		if (truncated_name)
-			*truncated_name = NULL;
-	}
-
-	return type;
-}
-
-static int system_eo_get(struct gfs2_inode *ip, struct gfs2_ea_request *er)
-{
-	if (!GFS2_ACL_IS_ACCESS(er->er_name, er->er_name_len) &&
-	    !GFS2_ACL_IS_DEFAULT(er->er_name, er->er_name_len) &&
-	    !capable(CAP_SYS_ADMIN))
-		return -EPERM;
-
-	if (GFS2_SB(&ip->i_inode)->sd_args.ar_posix_acl == 0 &&
-	    (GFS2_ACL_IS_ACCESS(er->er_name, er->er_name_len) ||
-	     GFS2_ACL_IS_DEFAULT(er->er_name, er->er_name_len)))
-		return -EOPNOTSUPP;
-
-	return gfs2_ea_get_i(ip, er);
-}
-
-static int system_eo_set(struct gfs2_inode *ip, struct gfs2_ea_request *er)
-{
-	int remove = 0;
-	int error;
-
-	if (GFS2_ACL_IS_ACCESS(er->er_name, er->er_name_len)) {
-		if (!(er->er_flags & GFS2_ERF_MODE)) {
-			er->er_mode = ip->i_inode.i_mode;
-			er->er_flags |= GFS2_ERF_MODE;
-		}
-		error = gfs2_acl_validate_set(ip, 1, er,
-					      &remove, &er->er_mode);
-		if (error)
-			return error;
-		error = gfs2_ea_set_i(ip, er);
-		if (error)
-			return error;
-		if (remove)
-			gfs2_ea_remove_i(ip, er);
-		return 0;
-
-	} else if (GFS2_ACL_IS_DEFAULT(er->er_name, er->er_name_len)) {
-		error = gfs2_acl_validate_set(ip, 0, er,
-					      &remove, NULL);
-		if (error)
-			return error;
-		if (!remove)
-			error = gfs2_ea_set_i(ip, er);
-		else {
-			error = gfs2_ea_remove_i(ip, er);
-			if (error == -ENODATA)
-				error = 0;
-		}
-		return error;
-	}
-
-	return -EPERM;
-}
-
-static int system_eo_remove(struct gfs2_inode *ip, struct gfs2_ea_request *er)
-{
-	if (GFS2_ACL_IS_ACCESS(er->er_name, er->er_name_len)) {
-		int error = gfs2_acl_validate_remove(ip, 1);
-		if (error)
-			return error;
-
-	} else if (GFS2_ACL_IS_DEFAULT(er->er_name, er->er_name_len)) {
-		int error = gfs2_acl_validate_remove(ip, 0);
-		if (error)
-			return error;
-
-	} else
-		return -EPERM;
-
-	return gfs2_ea_remove_i(ip, er);
-}
-
-static const struct gfs2_eattr_operations gfs2_user_eaops = {
-	.eo_get = gfs2_ea_get_i,
-	.eo_set = gfs2_ea_set_i,
-	.eo_remove = gfs2_ea_remove_i,
-	.eo_name = "user",
-};
-
-const struct gfs2_eattr_operations gfs2_system_eaops = {
-	.eo_get = system_eo_get,
-	.eo_set = system_eo_set,
-	.eo_remove = system_eo_remove,
-	.eo_name = "system",
-};
-
-static const struct gfs2_eattr_operations gfs2_security_eaops = {
-	.eo_get = gfs2_ea_get_i,
-	.eo_set = gfs2_ea_set_i,
-	.eo_remove = gfs2_ea_remove_i,
-	.eo_name = "security",
-};
-
-const struct gfs2_eattr_operations *gfs2_ea_ops[] = {
-	NULL,
-	&gfs2_user_eaops,
-	&gfs2_system_eaops,
-	&gfs2_security_eaops,
-};
-
diff --git a/fs/gfs2/eaops.h b/fs/gfs2/eaops.h
deleted file mode 100644
index da2f7fbbb40..00000000000
--- a/fs/gfs2/eaops.h
+++ /dev/null
@@ -1,30 +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 __EAOPS_DOT_H__
-#define __EAOPS_DOT_H__
-
-struct gfs2_ea_request;
-struct gfs2_inode;
-
-struct gfs2_eattr_operations {
-	int (*eo_get) (struct gfs2_inode *ip, struct gfs2_ea_request *er);
-	int (*eo_set) (struct gfs2_inode *ip, struct gfs2_ea_request *er);
-	int (*eo_remove) (struct gfs2_inode *ip, struct gfs2_ea_request *er);
-	char *eo_name;
-};
-
-unsigned int gfs2_ea_name2type(const char *name, const char **truncated_name);
-
-extern const struct gfs2_eattr_operations gfs2_system_eaops;
-
-extern const struct gfs2_eattr_operations *gfs2_ea_ops[];
-
-#endif /* __EAOPS_DOT_H__ */
-
diff --git a/fs/gfs2/export.c b/fs/gfs2/export.c
index 9200ef22171..d15876e9aa2 100644
--- a/fs/gfs2/export.c
+++ b/fs/gfs2/export.c
@@ -143,17 +143,14 @@ static struct dentry *gfs2_get_parent(struct dentry *child)
 }
 
 static struct dentry *gfs2_get_dentry(struct super_block *sb,
-		struct gfs2_inum_host *inum)
+				      struct gfs2_inum_host *inum)
 {
 	struct gfs2_sbd *sdp = sb->s_fs_info;
-	struct gfs2_holder i_gh, ri_gh, rgd_gh;
-	struct gfs2_rgrpd *rgd;
+	struct gfs2_holder i_gh;
 	struct inode *inode;
 	struct dentry *dentry;
 	int error;
 
-	/* System files? */
-
 	inode = gfs2_ilookup(sb, inum->no_addr);
 	if (inode) {
 		if (GFS2_I(inode)->i_no_formal_ino != inum->no_formal_ino) {
@@ -168,29 +165,11 @@ static struct dentry *gfs2_get_dentry(struct super_block *sb,
 	if (error)
 		return ERR_PTR(error);
 
-	error = gfs2_rindex_hold(sdp, &ri_gh);
+	error = gfs2_check_blk_type(sdp, inum->no_addr, GFS2_BLKST_DINODE);
 	if (error)
 		goto fail;
 
-	error = -EINVAL;
-	rgd = gfs2_blk2rgrpd(sdp, inum->no_addr);
-	if (!rgd)
-		goto fail_rindex;
-
-	error = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_SHARED, 0, &rgd_gh);
-	if (error)
-		goto fail_rindex;
-
-	error = -ESTALE;
-	if (gfs2_get_block_type(rgd, inum->no_addr) != GFS2_BLKST_DINODE)
-		goto fail_rgd;
-
-	gfs2_glock_dq_uninit(&rgd_gh);
-	gfs2_glock_dq_uninit(&ri_gh);
-
-	inode = gfs2_inode_lookup(sb, DT_UNKNOWN,
-					inum->no_addr,
-					0, 0);
+	inode = gfs2_inode_lookup(sb, DT_UNKNOWN, inum->no_addr, 0, 0);
 	if (IS_ERR(inode)) {
 		error = PTR_ERR(inode);
 		goto fail;
@@ -224,13 +203,6 @@ out_inode:
 	if (!IS_ERR(dentry))
 		dentry->d_op = &gfs2_dops;
 	return dentry;
-
-fail_rgd:
-	gfs2_glock_dq_uninit(&rgd_gh);
-
-fail_rindex:
-	gfs2_glock_dq_uninit(&ri_gh);
-
 fail:
 	gfs2_glock_dq_uninit(&i_gh);
 	return ERR_PTR(error);
diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 73318a3ce6f..166f38fbd24 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -38,7 +38,6 @@
 #include "rgrp.h"
 #include "trans.h"
 #include "util.h"
-#include "eaops.h"
 
 /**
  * gfs2_llseek - seek to a location in a file
diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h
index 61801ada36f..6edb423f90b 100644
--- a/fs/gfs2/incore.h
+++ b/fs/gfs2/incore.h
@@ -406,6 +406,12 @@ struct gfs2_statfs_change_host {
 #define GFS2_DATA_WRITEBACK	1
 #define GFS2_DATA_ORDERED	2
 
+#define GFS2_ERRORS_DEFAULT     GFS2_ERRORS_WITHDRAW
+#define GFS2_ERRORS_WITHDRAW    0
+#define GFS2_ERRORS_CONTINUE    1 /* place holder for future feature */
+#define GFS2_ERRORS_RO          2 /* place holder for future feature */
+#define GFS2_ERRORS_PANIC       3
+
 struct gfs2_args {
 	char ar_lockproto[GFS2_LOCKNAME_LEN];	/* Name of the Lock Protocol */
 	char ar_locktable[GFS2_LOCKNAME_LEN];	/* Name of the Lock Table */
@@ -422,6 +428,7 @@ struct gfs2_args {
 	unsigned int ar_data:2;			/* ordered/writeback */
 	unsigned int ar_meta:1;			/* mount metafs */
 	unsigned int ar_discard:1;		/* discard requests */
+	unsigned int ar_errors:2;               /* errors=withdraw | panic */
 	int ar_commit;				/* Commit interval */
 };
 
@@ -489,7 +496,6 @@ struct gfs2_sb_host {
  */
 
 struct lm_lockstruct {
-	u32 ls_id;
 	unsigned int ls_jid;
 	unsigned int ls_first;
 	unsigned int ls_first_done;
@@ -541,18 +547,12 @@ struct gfs2_sbd {
 	struct dentry *sd_root_dir;
 
 	struct inode *sd_jindex;
-	struct inode *sd_inum_inode;
 	struct inode *sd_statfs_inode;
-	struct inode *sd_ir_inode;
 	struct inode *sd_sc_inode;
 	struct inode *sd_qc_inode;
 	struct inode *sd_rindex;
 	struct inode *sd_quota_inode;
 
-	/* Inum stuff */
-
-	struct mutex sd_inum_mutex;
-
 	/* StatFS stuff */
 
 	spinlock_t sd_statfs_spin;
@@ -580,7 +580,6 @@ struct gfs2_sbd {
 	struct gfs2_holder sd_journal_gh;
 	struct gfs2_holder sd_jinode_gh;
 
-	struct gfs2_holder sd_ir_gh;
 	struct gfs2_holder sd_sc_gh;
 	struct gfs2_holder sd_qc_gh;
 
diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c
index 2f94bd72369..fb15d3b1f40 100644
--- a/fs/gfs2/inode.c
+++ b/fs/gfs2/inode.c
@@ -24,7 +24,7 @@
 #include "acl.h"
 #include "bmap.h"
 #include "dir.h"
-#include "eattr.h"
+#include "xattr.h"
 #include "glock.h"
 #include "glops.h"
 #include "inode.h"
@@ -519,139 +519,6 @@ out:
 	return inode ? inode : ERR_PTR(error);
 }
 
-static void gfs2_inum_range_in(struct gfs2_inum_range_host *ir, const void *buf)
-{
-	const struct gfs2_inum_range *str = buf;
-
-	ir->ir_start = be64_to_cpu(str->ir_start);
-	ir->ir_length = be64_to_cpu(str->ir_length);
-}
-
-static void gfs2_inum_range_out(const struct gfs2_inum_range_host *ir, void *buf)
-{
-	struct gfs2_inum_range *str = buf;
-
-	str->ir_start = cpu_to_be64(ir->ir_start);
-	str->ir_length = cpu_to_be64(ir->ir_length);
-}
-
-static int pick_formal_ino_1(struct gfs2_sbd *sdp, u64 *formal_ino)
-{
-	struct gfs2_inode *ip = GFS2_I(sdp->sd_ir_inode);
-	struct buffer_head *bh;
-	struct gfs2_inum_range_host ir;
-	int error;
-
-	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
-	if (error)
-		return error;
-	mutex_lock(&sdp->sd_inum_mutex);
-
-	error = gfs2_meta_inode_buffer(ip, &bh);
-	if (error) {
-		mutex_unlock(&sdp->sd_inum_mutex);
-		gfs2_trans_end(sdp);
-		return error;
-	}
-
-	gfs2_inum_range_in(&ir, bh->b_data + sizeof(struct gfs2_dinode));
-
-	if (ir.ir_length) {
-		*formal_ino = ir.ir_start++;
-		ir.ir_length--;
-		gfs2_trans_add_bh(ip->i_gl, bh, 1);
-		gfs2_inum_range_out(&ir,
-				    bh->b_data + sizeof(struct gfs2_dinode));
-		brelse(bh);
-		mutex_unlock(&sdp->sd_inum_mutex);
-		gfs2_trans_end(sdp);
-		return 0;
-	}
-
-	brelse(bh);
-
-	mutex_unlock(&sdp->sd_inum_mutex);
-	gfs2_trans_end(sdp);
-
-	return 1;
-}
-
-static int pick_formal_ino_2(struct gfs2_sbd *sdp, u64 *formal_ino)
-{
-	struct gfs2_inode *ip = GFS2_I(sdp->sd_ir_inode);
-	struct gfs2_inode *m_ip = GFS2_I(sdp->sd_inum_inode);
-	struct gfs2_holder gh;
-	struct buffer_head *bh;
-	struct gfs2_inum_range_host ir;
-	int error;
-
-	error = gfs2_glock_nq_init(m_ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
-	if (error)
-		return error;
-
-	error = gfs2_trans_begin(sdp, 2 * RES_DINODE, 0);
-	if (error)
-		goto out;
-	mutex_lock(&sdp->sd_inum_mutex);
-
-	error = gfs2_meta_inode_buffer(ip, &bh);
-	if (error)
-		goto out_end_trans;
-
-	gfs2_inum_range_in(&ir, bh->b_data + sizeof(struct gfs2_dinode));
-
-	if (!ir.ir_length) {
-		struct buffer_head *m_bh;
-		u64 x, y;
-		__be64 z;
-
-		error = gfs2_meta_inode_buffer(m_ip, &m_bh);
-		if (error)
-			goto out_brelse;
-
-		z = *(__be64 *)(m_bh->b_data + sizeof(struct gfs2_dinode));
-		x = y = be64_to_cpu(z);
-		ir.ir_start = x;
-		ir.ir_length = GFS2_INUM_QUANTUM;
-		x += GFS2_INUM_QUANTUM;
-		if (x < y)
-			gfs2_consist_inode(m_ip);
-		z = cpu_to_be64(x);
-		gfs2_trans_add_bh(m_ip->i_gl, m_bh, 1);
-		*(__be64 *)(m_bh->b_data + sizeof(struct gfs2_dinode)) = z;
-
-		brelse(m_bh);
-	}
-
-	*formal_ino = ir.ir_start++;
-	ir.ir_length--;
-
-	gfs2_trans_add_bh(ip->i_gl, bh, 1);
-	gfs2_inum_range_out(&ir, bh->b_data + sizeof(struct gfs2_dinode));
-
-out_brelse:
-	brelse(bh);
-out_end_trans:
-	mutex_unlock(&sdp->sd_inum_mutex);
-	gfs2_trans_end(sdp);
-out:
-	gfs2_glock_dq_uninit(&gh);
-	return error;
-}
-
-static int pick_formal_ino(struct gfs2_sbd *sdp, u64 *inum)
-{
-	int error;
-
-	error = pick_formal_ino_1(sdp, inum);
-	if (error <= 0)
-		return error;
-
-	error = pick_formal_ino_2(sdp, inum);
-
-	return error;
-}
-
 /**
  * create_ok - OK to create a new on-disk inode here?
  * @dip:  Directory in which dinode is to be created
@@ -731,7 +598,7 @@ static int alloc_dinode(struct gfs2_inode *dip, u64 *no_addr, u64 *generation)
 	if (error)
 		goto out_ipreserv;
 
-	*no_addr = gfs2_alloc_di(dip, generation);
+	error = gfs2_alloc_di(dip, no_addr, generation);
 
 	gfs2_trans_end(sdp);
 
@@ -924,7 +791,6 @@ static int gfs2_security_init(struct gfs2_inode *dip, struct gfs2_inode *ip)
 	size_t len;
 	void *value;
 	char *name;
-	struct gfs2_ea_request er;
 
 	err = security_inode_init_security(&ip->i_inode, &dip->i_inode,
 					   &name, &value, &len);
@@ -935,16 +801,7 @@ static int gfs2_security_init(struct gfs2_inode *dip, struct gfs2_inode *ip)
 		return err;
 	}
 
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-
-	er.er_type = GFS2_EATYPE_SECURITY;
-	er.er_name = name;
-	er.er_data = value;
-	er.er_name_len = strlen(name);
-	er.er_data_len = len;
-
-	err = gfs2_ea_set_i(ip, &er);
-
+	err = gfs2_xattr_set(&ip->i_inode, GFS2_EATYPE_SECURITY, name, value, len, 0);
 	kfree(value);
 	kfree(name);
 
@@ -991,13 +848,10 @@ struct inode *gfs2_createi(struct gfs2_holder *ghs, const struct qstr *name,
 	if (error)
 		goto fail_gunlock;
 
-	error = pick_formal_ino(sdp, &inum.no_formal_ino);
-	if (error)
-		goto fail_gunlock;
-
 	error = alloc_dinode(dip, &inum.no_addr, &generation);
 	if (error)
 		goto fail_gunlock;
+	inum.no_formal_ino = generation;
 
 	error = gfs2_glock_nq_num(sdp, inum.no_addr, &gfs2_inode_glops,
 				  LM_ST_EXCLUSIVE, GL_SKIP, ghs + 1);
@@ -1008,9 +862,8 @@ struct inode *gfs2_createi(struct gfs2_holder *ghs, const struct qstr *name,
 	if (error)
 		goto fail_gunlock2;
 
-	inode = gfs2_inode_lookup(dir->i_sb, IF2DT(mode),
-					inum.no_addr,
-					inum.no_formal_ino, 0);
+	inode = gfs2_inode_lookup(dir->i_sb, IF2DT(mode), inum.no_addr,
+				  inum.no_formal_ino, 0);
 	if (IS_ERR(inode))
 		goto fail_gunlock2;
 
diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c
index 7bc3c45cd67..52fb6c04898 100644
--- a/fs/gfs2/ops_fstype.c
+++ b/fs/gfs2/ops_fstype.c
@@ -84,7 +84,6 @@ static struct gfs2_sbd *init_sbd(struct super_block *sb)
 
 	gfs2_tune_init(&sdp->sd_tune);
 
-	mutex_init(&sdp->sd_inum_mutex);
 	spin_lock_init(&sdp->sd_statfs_spin);
 
 	spin_lock_init(&sdp->sd_rindex_spin);
@@ -833,21 +832,12 @@ static int init_inodes(struct gfs2_sbd *sdp, int undo)
 	if (error)
 		goto fail;
 
-	/* Read in the master inode number inode */
-	sdp->sd_inum_inode = gfs2_lookup_simple(master, "inum");
-	if (IS_ERR(sdp->sd_inum_inode)) {
-		error = PTR_ERR(sdp->sd_inum_inode);
-		fs_err(sdp, "can't read in inum inode: %d\n", error);
-		goto fail_journal;
-	}
-
-
 	/* Read in the master statfs inode */
 	sdp->sd_statfs_inode = gfs2_lookup_simple(master, "statfs");
 	if (IS_ERR(sdp->sd_statfs_inode)) {
 		error = PTR_ERR(sdp->sd_statfs_inode);
 		fs_err(sdp, "can't read in statfs inode: %d\n", error);
-		goto fail_inum;
+		goto fail_journal;
 	}
 
 	/* Read in the resource index inode */
@@ -876,8 +866,6 @@ fail_rindex:
 	iput(sdp->sd_rindex);
 fail_statfs:
 	iput(sdp->sd_statfs_inode);
-fail_inum:
-	iput(sdp->sd_inum_inode);
 fail_journal:
 	init_journal(sdp, UNDO);
 fail:
@@ -905,20 +893,12 @@ static int init_per_node(struct gfs2_sbd *sdp, int undo)
 		return error;
 	}
 
-	sprintf(buf, "inum_range%u", sdp->sd_jdesc->jd_jid);
-	sdp->sd_ir_inode = gfs2_lookup_simple(pn, buf);
-	if (IS_ERR(sdp->sd_ir_inode)) {
-		error = PTR_ERR(sdp->sd_ir_inode);
-		fs_err(sdp, "can't find local \"ir\" file: %d\n", error);
-		goto fail;
-	}
-
 	sprintf(buf, "statfs_change%u", sdp->sd_jdesc->jd_jid);
 	sdp->sd_sc_inode = gfs2_lookup_simple(pn, buf);
 	if (IS_ERR(sdp->sd_sc_inode)) {
 		error = PTR_ERR(sdp->sd_sc_inode);
 		fs_err(sdp, "can't find local \"sc\" file: %d\n", error);
-		goto fail_ir_i;
+		goto fail;
 	}
 
 	sprintf(buf, "quota_change%u", sdp->sd_jdesc->jd_jid);
@@ -932,27 +912,16 @@ static int init_per_node(struct gfs2_sbd *sdp, int undo)
 	iput(pn);
 	pn = NULL;
 
-	ip = GFS2_I(sdp->sd_ir_inode);
-	error = gfs2_glock_nq_init(ip->i_gl,
-				   LM_ST_EXCLUSIVE, 0,
-				   &sdp->sd_ir_gh);
-	if (error) {
-		fs_err(sdp, "can't lock local \"ir\" file: %d\n", error);
-		goto fail_qc_i;
-	}
-
 	ip = GFS2_I(sdp->sd_sc_inode);
-	error = gfs2_glock_nq_init(ip->i_gl,
-				   LM_ST_EXCLUSIVE, 0,
+	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0,
 				   &sdp->sd_sc_gh);
 	if (error) {
 		fs_err(sdp, "can't lock local \"sc\" file: %d\n", error);
-		goto fail_ir_gh;
+		goto fail_qc_i;
 	}
 
 	ip = GFS2_I(sdp->sd_qc_inode);
-	error = gfs2_glock_nq_init(ip->i_gl,
-				   LM_ST_EXCLUSIVE, 0,
+	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0,
 				   &sdp->sd_qc_gh);
 	if (error) {
 		fs_err(sdp, "can't lock local \"qc\" file: %d\n", error);
@@ -965,14 +934,10 @@ fail_qc_gh:
 	gfs2_glock_dq_uninit(&sdp->sd_qc_gh);
 fail_ut_gh:
 	gfs2_glock_dq_uninit(&sdp->sd_sc_gh);
-fail_ir_gh:
-	gfs2_glock_dq_uninit(&sdp->sd_ir_gh);
 fail_qc_i:
 	iput(sdp->sd_qc_inode);
 fail_ut_i:
 	iput(sdp->sd_sc_inode);
-fail_ir_i:
-	iput(sdp->sd_ir_inode);
 fail:
 	if (pn)
 		iput(pn);
@@ -1063,7 +1028,6 @@ static int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent)
 
 	ls->ls_ops = lm;
 	ls->ls_first = 1;
-	ls->ls_id = 0;
 
 	for (options = args->ar_hostdata; (o = strsep(&options, ":")); ) {
 		substring_t tmp[MAX_OPT_ARGS];
@@ -1081,10 +1045,7 @@ static int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent)
 			ls->ls_jid = option;
 			break;
 		case Opt_id:
-			ret = match_int(&tmp[0], &option);
-			if (ret)
-				goto hostdata_error;
-			ls->ls_id = option;
+			/* Obsolete, but left for backward compat purposes */
 			break;
 		case Opt_first:
 			ret = match_int(&tmp[0], &option);
@@ -1133,6 +1094,17 @@ void gfs2_lm_unmount(struct gfs2_sbd *sdp)
 		lm->lm_unmount(sdp);
 }
 
+void gfs2_online_uevent(struct gfs2_sbd *sdp)
+{
+	struct super_block *sb = sdp->sd_vfs;
+	char ro[20];
+	char spectator[20];
+	char *envp[] = { ro, spectator, NULL };
+	sprintf(ro, "RDONLY=%d", (sb->s_flags & MS_RDONLY) ? 1 : 0);
+	sprintf(spectator, "SPECTATOR=%d", sdp->sd_args.ar_spectator ? 1 : 0);
+	kobject_uevent_env(&sdp->sd_kobj, KOBJ_ONLINE, envp);
+}
+
 /**
  * fill_super - Read in superblock
  * @sb: The VFS superblock
@@ -1157,6 +1129,7 @@ static int fill_super(struct super_block *sb, void *data, int silent)
 	sdp->sd_args.ar_quota = GFS2_QUOTA_DEFAULT;
 	sdp->sd_args.ar_data = GFS2_DATA_DEFAULT;
 	sdp->sd_args.ar_commit = 60;
+	sdp->sd_args.ar_errors = GFS2_ERRORS_DEFAULT;
 
 	error = gfs2_mount_args(sdp, &sdp->sd_args, data);
 	if (error) {
@@ -1174,6 +1147,7 @@ static int fill_super(struct super_block *sb, void *data, int silent)
 	sb->s_magic = GFS2_MAGIC;
 	sb->s_op = &gfs2_super_ops;
 	sb->s_export_op = &gfs2_export_ops;
+	sb->s_xattr = gfs2_xattr_handlers;
 	sb->s_time_gran = 1;
 	sb->s_maxbytes = MAX_LFS_FILESIZE;
 
@@ -1236,7 +1210,7 @@ static int fill_super(struct super_block *sb, void *data, int silent)
 	}
 
 	gfs2_glock_dq_uninit(&mount_gh);
-
+	gfs2_online_uevent(sdp);
 	return 0;
 
 fail_threads:
diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c
index f8bd20baf99..c3ac1805405 100644
--- a/fs/gfs2/ops_inode.c
+++ b/fs/gfs2/ops_inode.c
@@ -26,8 +26,7 @@
 #include "acl.h"
 #include "bmap.h"
 #include "dir.h"
-#include "eaops.h"
-#include "eattr.h"
+#include "xattr.h"
 #include "glock.h"
 #include "inode.h"
 #include "meta_io.h"
@@ -349,7 +348,7 @@ static int gfs2_unlink(struct inode *dir, struct dentry *dentry)
 
 	error = gfs2_trans_begin(sdp, 2*RES_DINODE + RES_LEAF + RES_RG_BIT, 0);
 	if (error)
-		goto out_rgrp;
+		goto out_gunlock;
 
 	error = gfs2_dir_del(dip, &dentry->d_name);
         if (error)
@@ -1302,60 +1301,53 @@ static int gfs2_setxattr(struct dentry *dentry, const char *name,
 			 const void *data, size_t size, int flags)
 {
 	struct inode *inode = dentry->d_inode;
-	struct gfs2_ea_request er;
-
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-	er.er_type = gfs2_ea_name2type(name, &er.er_name);
-	if (er.er_type == GFS2_EATYPE_UNUSED)
-		return -EOPNOTSUPP;
-	er.er_data = (char *)data;
-	er.er_name_len = strlen(er.er_name);
-	er.er_data_len = size;
-	er.er_flags = flags;
-
-	gfs2_assert_warn(GFS2_SB(inode), !(er.er_flags & GFS2_ERF_MODE));
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_holder gh;
+	int ret;
 
-	return gfs2_ea_set(GFS2_I(inode), &er);
+	gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
+	ret = gfs2_glock_nq(&gh);
+	if (ret == 0) {
+		ret = generic_setxattr(dentry, name, data, size, flags);
+		gfs2_glock_dq(&gh);
+	}
+	gfs2_holder_uninit(&gh);
+	return ret;
 }
 
 static ssize_t gfs2_getxattr(struct dentry *dentry, const char *name,
 			     void *data, size_t size)
 {
-	struct gfs2_ea_request er;
-
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-	er.er_type = gfs2_ea_name2type(name, &er.er_name);
-	if (er.er_type == GFS2_EATYPE_UNUSED)
-		return -EOPNOTSUPP;
-	er.er_data = data;
-	er.er_name_len = strlen(er.er_name);
-	er.er_data_len = size;
-
-	return gfs2_ea_get(GFS2_I(dentry->d_inode), &er);
-}
-
-static ssize_t gfs2_listxattr(struct dentry *dentry, char *buffer, size_t size)
-{
-	struct gfs2_ea_request er;
-
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-	er.er_data = (size) ? buffer : NULL;
-	er.er_data_len = size;
+	struct inode *inode = dentry->d_inode;
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_holder gh;
+	int ret;
 
-	return gfs2_ea_list(GFS2_I(dentry->d_inode), &er);
+	gfs2_holder_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &gh);
+	ret = gfs2_glock_nq(&gh);
+	if (ret == 0) {
+		ret = generic_getxattr(dentry, name, data, size);
+		gfs2_glock_dq(&gh);
+	}
+	gfs2_holder_uninit(&gh);
+	return ret;
 }
 
 static int gfs2_removexattr(struct dentry *dentry, const char *name)
 {
-	struct gfs2_ea_request er;
-
-	memset(&er, 0, sizeof(struct gfs2_ea_request));
-	er.er_type = gfs2_ea_name2type(name, &er.er_name);
-	if (er.er_type == GFS2_EATYPE_UNUSED)
-		return -EOPNOTSUPP;
-	er.er_name_len = strlen(er.er_name);
+	struct inode *inode = dentry->d_inode;
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_holder gh;
+	int ret;
 
-	return gfs2_ea_remove(GFS2_I(dentry->d_inode), &er);
+	gfs2_holder_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
+	ret = gfs2_glock_nq(&gh);
+	if (ret == 0) {
+		ret = generic_removexattr(dentry, name);
+		gfs2_glock_dq(&gh);
+	}
+	gfs2_holder_uninit(&gh);
+	return ret;
 }
 
 static int gfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index fba795798d3..8f1cfb02a6c 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -179,7 +179,7 @@ static inline u64 gfs2_bit_search(const __le64 *ptr, u64 mask, u8 state)
  * always aligned to a 64 bit boundary.
  *
  * The size of the buffer is in bytes, but is it assumed that it is
- * always ok to to read a complete multiple of 64 bits at the end
+ * always ok to read a complete multiple of 64 bits at the end
  * of the block in case the end is no aligned to a natural boundary.
  *
  * Return: the block number (bitmap buffer scope) that was found
@@ -857,7 +857,8 @@ static void gfs2_rgrp_send_discards(struct gfs2_sbd *sdp, u64 offset,
 					goto start_new_extent;
 				if ((start + nr_sects) != blk) {
 					rv = blkdev_issue_discard(bdev, start,
-							    nr_sects, GFP_NOFS);
+							    nr_sects, GFP_NOFS,
+							    DISCARD_FL_BARRIER);
 					if (rv)
 						goto fail;
 					nr_sects = 0;
@@ -871,7 +872,8 @@ start_new_extent:
 		}
 	}
 	if (nr_sects) {
-		rv = blkdev_issue_discard(bdev, start, nr_sects, GFP_NOFS);
+		rv = blkdev_issue_discard(bdev, start, nr_sects, GFP_NOFS,
+					 DISCARD_FL_BARRIER);
 		if (rv)
 			goto fail;
 	}
@@ -1256,7 +1258,7 @@ void gfs2_inplace_release(struct gfs2_inode *ip)
  * Returns: The block type (GFS2_BLKST_*)
  */
 
-unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block)
+static unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block)
 {
 	struct gfs2_bitmap *bi = NULL;
 	u32 length, rgrp_block, buf_block;
@@ -1459,6 +1461,16 @@ int gfs2_rgrp_dump(struct seq_file *seq, const struct gfs2_glock *gl)
 	return 0;
 }
 
+static void gfs2_rgrp_error(struct gfs2_rgrpd *rgd)
+{
+	struct gfs2_sbd *sdp = rgd->rd_sbd;
+	fs_warn(sdp, "rgrp %llu has an error, marking it readonly until umount\n",
+		(unsigned long long)rgd->rd_addr);
+	fs_warn(sdp, "umount on all nodes and run fsck.gfs2 to fix the error\n");
+	gfs2_rgrp_dump(NULL, rgd->rd_gl);
+	rgd->rd_flags |= GFS2_RDF_ERROR;
+}
+
 /**
  * gfs2_alloc_block - Allocate one or more blocks
  * @ip: the inode to allocate the block for
@@ -1520,22 +1532,20 @@ int gfs2_alloc_block(struct gfs2_inode *ip, u64 *bn, unsigned int *n)
 	return 0;
 
 rgrp_error:
-	fs_warn(sdp, "rgrp %llu has an error, marking it readonly until umount\n",
-	        (unsigned long long)rgd->rd_addr);
-	fs_warn(sdp, "umount on all nodes and run fsck.gfs2 to fix the error\n");
-	gfs2_rgrp_dump(NULL, rgd->rd_gl);
-	rgd->rd_flags |= GFS2_RDF_ERROR;
+	gfs2_rgrp_error(rgd);
 	return -EIO;
 }
 
 /**
  * gfs2_alloc_di - Allocate a dinode
  * @dip: the directory that the inode is going in
+ * @bn: the block number which is allocated
+ * @generation: the generation number of the inode
  *
- * Returns: the block allocated
+ * Returns: 0 on success or error
  */
 
-u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation)
+int gfs2_alloc_di(struct gfs2_inode *dip, u64 *bn, u64 *generation)
 {
 	struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode);
 	struct gfs2_alloc *al = dip->i_alloc;
@@ -1546,16 +1556,21 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation)
 
 	blk = rgblk_search(rgd, rgd->rd_last_alloc,
 			   GFS2_BLKST_FREE, GFS2_BLKST_DINODE, &n);
-	BUG_ON(blk == BFITNOENT);
 
-	rgd->rd_last_alloc = blk;
+	/* Since all blocks are reserved in advance, this shouldn't happen */
+	if (blk == BFITNOENT)
+		goto rgrp_error;
 
+	rgd->rd_last_alloc = blk;
 	block = rgd->rd_data0 + blk;
+	if (rgd->rd_free == 0)
+		goto rgrp_error;
 
-	gfs2_assert_withdraw(sdp, rgd->rd_free);
 	rgd->rd_free--;
 	rgd->rd_dinodes++;
 	*generation = rgd->rd_igeneration++;
+	if (*generation == 0)
+		*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);
 
@@ -1568,7 +1583,12 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation)
 	rgd->rd_free_clone--;
 	spin_unlock(&sdp->sd_rindex_spin);
 	trace_gfs2_block_alloc(dip, block, 1, GFS2_BLKST_DINODE);
-	return block;
+	*bn = block;
+	return 0;
+
+rgrp_error:
+	gfs2_rgrp_error(rgd);
+	return -EIO;
 }
 
 /**
@@ -1676,6 +1696,46 @@ void gfs2_free_di(struct gfs2_rgrpd *rgd, struct gfs2_inode *ip)
 }
 
 /**
+ * gfs2_check_blk_type - Check the type of a block
+ * @sdp: The superblock
+ * @no_addr: The block number to check
+ * @type: The block type we are looking for
+ *
+ * Returns: 0 if the block type matches the expected type
+ *          -ESTALE if it doesn't match
+ *          or -ve errno if something went wrong while checking
+ */
+
+int gfs2_check_blk_type(struct gfs2_sbd *sdp, u64 no_addr, unsigned int type)
+{
+	struct gfs2_rgrpd *rgd;
+	struct gfs2_holder ri_gh, rgd_gh;
+	int error;
+
+	error = gfs2_rindex_hold(sdp, &ri_gh);
+	if (error)
+		goto fail;
+
+	error = -EINVAL;
+	rgd = gfs2_blk2rgrpd(sdp, no_addr);
+	if (!rgd)
+		goto fail_rindex;
+
+	error = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_SHARED, 0, &rgd_gh);
+	if (error)
+		goto fail_rindex;
+
+	if (gfs2_get_block_type(rgd, no_addr) != type)
+		error = -ESTALE;
+
+	gfs2_glock_dq_uninit(&rgd_gh);
+fail_rindex:
+	gfs2_glock_dq_uninit(&ri_gh);
+fail:
+	return error;
+}
+
+/**
  * gfs2_rlist_add - add a RG to a list of RGs
  * @sdp: the filesystem
  * @rlist: the list of resource groups
diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h
index 1e76ff0f3e0..b4106ddaaa9 100644
--- a/fs/gfs2/rgrp.h
+++ b/fs/gfs2/rgrp.h
@@ -44,15 +44,15 @@ gfs2_inplace_reserve_i((ip), __FILE__, __LINE__)
 
 extern void gfs2_inplace_release(struct gfs2_inode *ip);
 
-extern unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block);
-
 extern int gfs2_alloc_block(struct gfs2_inode *ip, u64 *bn, unsigned int *n);
-extern u64 gfs2_alloc_di(struct gfs2_inode *ip, u64 *generation);
+extern int gfs2_alloc_di(struct gfs2_inode *ip, u64 *bn, u64 *generation);
 
 extern void gfs2_free_data(struct gfs2_inode *ip, u64 bstart, u32 blen);
 extern void gfs2_free_meta(struct gfs2_inode *ip, u64 bstart, u32 blen);
 extern void gfs2_free_di(struct gfs2_rgrpd *rgd, struct gfs2_inode *ip);
 extern void gfs2_unlink_di(struct inode *inode);
+extern int gfs2_check_blk_type(struct gfs2_sbd *sdp, u64 no_addr,
+			       unsigned int type);
 
 struct gfs2_rgrp_list {
 	unsigned int rl_rgrps;
diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index f522bb01797..0ec3ec672de 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -38,7 +38,7 @@
 #include "trans.h"
 #include "util.h"
 #include "sys.h"
-#include "eattr.h"
+#include "xattr.h"
 
 #define args_neq(a1, a2, x) ((a1)->ar_##x != (a2)->ar_##x)
 
@@ -68,6 +68,8 @@ enum {
 	Opt_discard,
 	Opt_nodiscard,
 	Opt_commit,
+	Opt_err_withdraw,
+	Opt_err_panic,
 	Opt_error,
 };
 
@@ -97,6 +99,8 @@ static const match_table_t tokens = {
 	{Opt_discard, "discard"},
 	{Opt_nodiscard, "nodiscard"},
 	{Opt_commit, "commit=%d"},
+	{Opt_err_withdraw, "errors=withdraw"},
+	{Opt_err_panic, "errors=panic"},
 	{Opt_error, NULL}
 };
 
@@ -152,6 +156,11 @@ int gfs2_mount_args(struct gfs2_sbd *sdp, struct gfs2_args *args, char *options)
 			args->ar_localcaching = 1;
 			break;
 		case Opt_debug:
+			if (args->ar_errors == GFS2_ERRORS_PANIC) {
+				fs_info(sdp, "-o debug and -o errors=panic "
+				       "are mutually exclusive.\n");
+				return -EINVAL;
+			}
 			args->ar_debug = 1;
 			break;
 		case Opt_nodebug:
@@ -205,6 +214,17 @@ int gfs2_mount_args(struct gfs2_sbd *sdp, struct gfs2_args *args, char *options)
 				return rv ? rv : -EINVAL;
 			}
 			break;
+		case Opt_err_withdraw:
+			args->ar_errors = GFS2_ERRORS_WITHDRAW;
+			break;
+		case Opt_err_panic:
+			if (args->ar_debug) {
+				fs_info(sdp, "-o debug and -o errors=panic "
+					"are mutually exclusive.\n");
+				return -EINVAL;
+			}
+			args->ar_errors = GFS2_ERRORS_PANIC;
+			break;
 		case Opt_error:
 		default:
 			fs_info(sdp, "invalid mount option: %s\n", o);
@@ -768,7 +788,6 @@ restart:
 	/*  Release stuff  */
 
 	iput(sdp->sd_jindex);
-	iput(sdp->sd_inum_inode);
 	iput(sdp->sd_statfs_inode);
 	iput(sdp->sd_rindex);
 	iput(sdp->sd_quota_inode);
@@ -779,10 +798,8 @@ restart:
 	if (!sdp->sd_args.ar_spectator) {
 		gfs2_glock_dq_uninit(&sdp->sd_journal_gh);
 		gfs2_glock_dq_uninit(&sdp->sd_jinode_gh);
-		gfs2_glock_dq_uninit(&sdp->sd_ir_gh);
 		gfs2_glock_dq_uninit(&sdp->sd_sc_gh);
 		gfs2_glock_dq_uninit(&sdp->sd_qc_gh);
-		iput(sdp->sd_ir_inode);
 		iput(sdp->sd_sc_inode);
 		iput(sdp->sd_qc_inode);
 	}
@@ -1084,6 +1101,7 @@ static int gfs2_remount_fs(struct super_block *sb, int *flags, char *data)
 	gt->gt_log_flush_secs = args.ar_commit;
 	spin_unlock(&gt->gt_spin);
 
+	gfs2_online_uevent(sdp);
 	return 0;
 }
 
@@ -1225,6 +1243,22 @@ static int gfs2_show_options(struct seq_file *s, struct vfsmount *mnt)
 	lfsecs = sdp->sd_tune.gt_log_flush_secs;
 	if (lfsecs != 60)
 		seq_printf(s, ",commit=%d", lfsecs);
+	if (args->ar_errors != GFS2_ERRORS_DEFAULT) {
+		const char *state;
+
+		switch (args->ar_errors) {
+		case GFS2_ERRORS_WITHDRAW:
+			state = "withdraw";
+			break;
+		case GFS2_ERRORS_PANIC:
+			state = "panic";
+			break;
+		default:
+			state = "unknown";
+			break;
+		}
+		seq_printf(s, ",errors=%s", state);
+	}
 	return 0;
 }
 
@@ -1252,6 +1286,10 @@ static void gfs2_delete_inode(struct inode *inode)
 		goto out;
 	}
 
+	error = gfs2_check_blk_type(sdp, ip->i_no_addr, GFS2_BLKST_UNLINKED);
+	if (error)
+		goto out_truncate;
+
 	gfs2_glock_dq_wait(&ip->i_iopen_gh);
 	gfs2_holder_reinit(LM_ST_EXCLUSIVE, LM_FLAG_TRY_1CB | GL_NOCACHE, &ip->i_iopen_gh);
 	error = gfs2_glock_nq(&ip->i_iopen_gh);
diff --git a/fs/gfs2/super.h b/fs/gfs2/super.h
index 22e0417ed99..235db368288 100644
--- a/fs/gfs2/super.h
+++ b/fs/gfs2/super.h
@@ -25,7 +25,7 @@ static inline unsigned int gfs2_jindex_size(struct gfs2_sbd *sdp)
 	return x;
 }
 
-void gfs2_jindex_free(struct gfs2_sbd *sdp);
+extern void gfs2_jindex_free(struct gfs2_sbd *sdp);
 
 extern int gfs2_mount_args(struct gfs2_sbd *sdp, struct gfs2_args *args, char *data);
 
@@ -36,7 +36,7 @@ extern int gfs2_lookup_in_master_dir(struct gfs2_sbd *sdp, char *filename,
 				     struct gfs2_inode **ipp);
 
 extern int gfs2_make_fs_rw(struct gfs2_sbd *sdp);
-
+extern void gfs2_online_uevent(struct gfs2_sbd *sdp);
 extern int gfs2_statfs_init(struct gfs2_sbd *sdp);
 extern void gfs2_statfs_change(struct gfs2_sbd *sdp, s64 total, s64 free,
 			       s64 dinodes);
@@ -54,6 +54,7 @@ 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 const struct dentry_operations gfs2_dops;
+extern struct xattr_handler *gfs2_xattr_handlers[];
 
 #endif /* __SUPER_DOT_H__ */
 
diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c
index a7cbfbd340c..446329728d5 100644
--- a/fs/gfs2/sys.c
+++ b/fs/gfs2/sys.c
@@ -16,6 +16,7 @@
 #include <linux/kobject.h>
 #include <asm/uaccess.h>
 #include <linux/gfs2_ondisk.h>
+#include <linux/genhd.h>
 
 #include "gfs2.h"
 #include "incore.h"
@@ -319,12 +320,6 @@ static ssize_t block_store(struct gfs2_sbd *sdp, const char *buf, size_t len)
 	return ret;
 }
 
-static ssize_t lkid_show(struct gfs2_sbd *sdp, char *buf)
-{
-	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
-	return sprintf(buf, "%u\n", ls->ls_id);
-}
-
 static ssize_t lkfirst_show(struct gfs2_sbd *sdp, char *buf)
 {
 	struct lm_lockstruct *ls = &sdp->sd_lockstruct;
@@ -389,7 +384,6 @@ static struct gfs2_attr gdlm_attr_##_name = __ATTR(_name,_mode,_show,_store)
 GDLM_ATTR(proto_name,		0444, proto_name_show,		NULL);
 GDLM_ATTR(block,		0644, block_show,		block_store);
 GDLM_ATTR(withdraw,		0644, withdraw_show,		withdraw_store);
-GDLM_ATTR(id,			0444, lkid_show,		NULL);
 GDLM_ATTR(jid,			0444, jid_show,			NULL);
 GDLM_ATTR(first,		0444, lkfirst_show,		NULL);
 GDLM_ATTR(first_done,		0444, first_done_show,		NULL);
@@ -401,7 +395,6 @@ static struct attribute *lock_module_attrs[] = {
 	&gdlm_attr_proto_name.attr,
 	&gdlm_attr_block.attr,
 	&gdlm_attr_withdraw.attr,
-	&gdlm_attr_id.attr,
 	&gdlm_attr_jid.attr,
 	&gdlm_attr_first.attr,
 	&gdlm_attr_first_done.attr,
@@ -519,7 +512,14 @@ static struct attribute_group lock_module_group = {
 
 int gfs2_sys_fs_add(struct gfs2_sbd *sdp)
 {
+	struct super_block *sb = sdp->sd_vfs;
 	int error;
+	char ro[20];
+	char spectator[20];
+	char *envp[] = { ro, spectator, NULL };
+
+	sprintf(ro, "RDONLY=%d", (sb->s_flags & MS_RDONLY) ? 1 : 0);
+	sprintf(spectator, "SPECTATOR=%d", sdp->sd_args.ar_spectator ? 1 : 0);
 
 	sdp->sd_kobj.kset = gfs2_kset;
 	error = kobject_init_and_add(&sdp->sd_kobj, &gfs2_ktype, NULL,
@@ -535,9 +535,17 @@ int gfs2_sys_fs_add(struct gfs2_sbd *sdp)
 	if (error)
 		goto fail_tune;
 
-	kobject_uevent(&sdp->sd_kobj, KOBJ_ADD);
+	error = sysfs_create_link(&sdp->sd_kobj,
+				  &disk_to_dev(sb->s_bdev->bd_disk)->kobj,
+				  "device");
+	if (error)
+		goto fail_lock_module;
+
+	kobject_uevent_env(&sdp->sd_kobj, KOBJ_ADD, envp);
 	return 0;
 
+fail_lock_module:
+	sysfs_remove_group(&sdp->sd_kobj, &lock_module_group);
 fail_tune:
 	sysfs_remove_group(&sdp->sd_kobj, &tune_group);
 fail_reg:
@@ -549,12 +557,12 @@ fail:
 
 void gfs2_sys_fs_del(struct gfs2_sbd *sdp)
 {
+	sysfs_remove_link(&sdp->sd_kobj, "device");
 	sysfs_remove_group(&sdp->sd_kobj, &tune_group);
 	sysfs_remove_group(&sdp->sd_kobj, &lock_module_group);
 	kobject_put(&sdp->sd_kobj);
 }
 
-
 static int gfs2_uevent(struct kset *kset, struct kobject *kobj,
 		       struct kobj_uevent_env *env)
 {
@@ -563,6 +571,8 @@ static int gfs2_uevent(struct kset *kset, struct kobject *kobj,
 
 	add_uevent_var(env, "LOCKTABLE=%s", sdp->sd_table_name);
 	add_uevent_var(env, "LOCKPROTO=%s", sdp->sd_proto_name);
+	if (!sdp->sd_args.ar_spectator)
+		add_uevent_var(env, "JOURNALID=%u", sdp->sd_lockstruct.ls_jid);
 	if (gfs2_uuid_valid(uuid)) {
 		add_uevent_var(env, "UUID=%02X%02X%02X%02X-%02X%02X-%02X%02X-"
 			       "%02X%02X-%02X%02X%02X%02X%02X%02X",
@@ -578,7 +588,6 @@ static struct kset_uevent_ops gfs2_uevent_ops = {
 	.uevent = gfs2_uevent,
 };
 
-
 int gfs2_sys_init(void)
 {
 	gfs2_kset = kset_create_and_add("gfs2", &gfs2_uevent_ops, fs_kobj);
diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c
index 9d12b1118ba..f6a7efa34eb 100644
--- a/fs/gfs2/util.c
+++ b/fs/gfs2/util.c
@@ -38,24 +38,30 @@ int gfs2_lm_withdraw(struct gfs2_sbd *sdp, char *fmt, ...)
 	const struct lm_lockops *lm = ls->ls_ops;
 	va_list args;
 
-	if (test_and_set_bit(SDF_SHUTDOWN, &sdp->sd_flags))
+	if (sdp->sd_args.ar_errors == GFS2_ERRORS_WITHDRAW &&
+	    test_and_set_bit(SDF_SHUTDOWN, &sdp->sd_flags))
 		return 0;
 
 	va_start(args, fmt);
 	vprintk(fmt, args);
 	va_end(args);
 
-	fs_err(sdp, "about to withdraw this file system\n");
-	BUG_ON(sdp->sd_args.ar_debug);
+	if (sdp->sd_args.ar_errors == GFS2_ERRORS_WITHDRAW) {
+		fs_err(sdp, "about to withdraw this file system\n");
+		BUG_ON(sdp->sd_args.ar_debug);
 
-	kobject_uevent(&sdp->sd_kobj, KOBJ_OFFLINE);
+		kobject_uevent(&sdp->sd_kobj, KOBJ_OFFLINE);
 
-	if (lm->lm_unmount) {
-		fs_err(sdp, "telling LM to unmount\n");
-		lm->lm_unmount(sdp);
+		if (lm->lm_unmount) {
+			fs_err(sdp, "telling LM to unmount\n");
+			lm->lm_unmount(sdp);
+		}
+		fs_err(sdp, "withdrawn\n");
+		dump_stack();
 	}
-	fs_err(sdp, "withdrawn\n");
-	dump_stack();
+
+	if (sdp->sd_args.ar_errors == GFS2_ERRORS_PANIC)
+		panic("GFS2: fsid=%s: panic requested.\n", sdp->sd_fsname);
 
 	return -1;
 }
@@ -93,17 +99,24 @@ int gfs2_assert_warn_i(struct gfs2_sbd *sdp, char *assertion,
 			gfs2_tune_get(sdp, gt_complain_secs) * HZ))
 		return -2;
 
-	printk(KERN_WARNING
-	       "GFS2: fsid=%s: warning: assertion \"%s\" failed\n"
-	       "GFS2: fsid=%s:   function = %s, file = %s, line = %u\n",
-	       sdp->sd_fsname, assertion,
-	       sdp->sd_fsname, function, file, line);
+	if (sdp->sd_args.ar_errors == GFS2_ERRORS_WITHDRAW)
+		printk(KERN_WARNING
+		       "GFS2: fsid=%s: warning: assertion \"%s\" failed\n"
+		       "GFS2: fsid=%s:   function = %s, file = %s, line = %u\n",
+		       sdp->sd_fsname, assertion,
+		       sdp->sd_fsname, function, file, line);
 
 	if (sdp->sd_args.ar_debug)
 		BUG();
 	else
 		dump_stack();
 
+	if (sdp->sd_args.ar_errors == GFS2_ERRORS_PANIC)
+		panic("GFS2: fsid=%s: warning: assertion \"%s\" failed\n"
+		      "GFS2: fsid=%s:   function = %s, file = %s, line = %u\n",
+		      sdp->sd_fsname, assertion,
+		      sdp->sd_fsname, function, file, line);
+
 	sdp->sd_last_warning = jiffies;
 
 	return -1;
diff --git a/fs/gfs2/eattr.c b/fs/gfs2/xattr.c
index 07ea9529add..8a0f8ef6ee2 100644
--- a/fs/gfs2/eattr.c
+++ b/fs/gfs2/xattr.c
@@ -18,8 +18,7 @@
 #include "gfs2.h"
 #include "incore.h"
 #include "acl.h"
-#include "eaops.h"
-#include "eattr.h"
+#include "xattr.h"
 #include "glock.h"
 #include "inode.h"
 #include "meta_io.h"
@@ -38,26 +37,32 @@
  * Returns: 1 if the EA should be stuffed
  */
 
-static int ea_calc_size(struct gfs2_sbd *sdp, struct gfs2_ea_request *er,
+static int ea_calc_size(struct gfs2_sbd *sdp, unsigned int nsize, size_t dsize,
 			unsigned int *size)
 {
-	*size = GFS2_EAREQ_SIZE_STUFFED(er);
-	if (*size <= sdp->sd_jbsize)
+	unsigned int jbsize = sdp->sd_jbsize;
+
+	/* Stuffed */
+	*size = ALIGN(sizeof(struct gfs2_ea_header) + nsize + dsize, 8);
+
+	if (*size <= jbsize)
 		return 1;
 
-	*size = GFS2_EAREQ_SIZE_UNSTUFFED(sdp, er);
+	/* Unstuffed */
+	*size = ALIGN(sizeof(struct gfs2_ea_header) + nsize +
+		      (sizeof(__be64) * DIV_ROUND_UP(dsize, jbsize)), 8);
 
 	return 0;
 }
 
-static int ea_check_size(struct gfs2_sbd *sdp, struct gfs2_ea_request *er)
+static int ea_check_size(struct gfs2_sbd *sdp, unsigned int nsize, size_t dsize)
 {
 	unsigned int size;
 
-	if (er->er_data_len > GFS2_EA_MAX_DATA_LEN)
+	if (dsize > GFS2_EA_MAX_DATA_LEN)
 		return -ERANGE;
 
-	ea_calc_size(sdp, er, &size);
+	ea_calc_size(sdp, nsize, dsize, &size);
 
 	/* This can only happen with 512 byte blocks */
 	if (size > sdp->sd_jbsize)
@@ -151,7 +156,9 @@ out:
 }
 
 struct ea_find {
-	struct gfs2_ea_request *ef_er;
+	int type;
+	const char *name;
+	size_t namel;
 	struct gfs2_ea_location *ef_el;
 };
 
@@ -160,14 +167,13 @@ static int ea_find_i(struct gfs2_inode *ip, struct buffer_head *bh,
 		     void *private)
 {
 	struct ea_find *ef = private;
-	struct gfs2_ea_request *er = ef->ef_er;
 
 	if (ea->ea_type == GFS2_EATYPE_UNUSED)
 		return 0;
 
-	if (ea->ea_type == er->er_type) {
-		if (ea->ea_name_len == er->er_name_len &&
-		    !memcmp(GFS2_EA2NAME(ea), er->er_name, ea->ea_name_len)) {
+	if (ea->ea_type == ef->type) {
+		if (ea->ea_name_len == ef->namel &&
+		    !memcmp(GFS2_EA2NAME(ea), ef->name, ea->ea_name_len)) {
 			struct gfs2_ea_location *el = ef->ef_el;
 			get_bh(bh);
 			el->el_bh = bh;
@@ -180,13 +186,15 @@ static int ea_find_i(struct gfs2_inode *ip, struct buffer_head *bh,
 	return 0;
 }
 
-int gfs2_ea_find(struct gfs2_inode *ip, struct gfs2_ea_request *er,
+int gfs2_ea_find(struct gfs2_inode *ip, int type, const char *name,
 		 struct gfs2_ea_location *el)
 {
 	struct ea_find ef;
 	int error;
 
-	ef.ef_er = er;
+	ef.type = type;
+	ef.name = name;
+	ef.namel = strlen(name);
 	ef.ef_el = el;
 
 	memset(el, 0, sizeof(struct gfs2_ea_location));
@@ -344,6 +352,20 @@ struct ea_list {
 	unsigned int ei_size;
 };
 
+static inline unsigned int gfs2_ea_strlen(struct gfs2_ea_header *ea)
+{
+	switch (ea->ea_type) {
+	case GFS2_EATYPE_USR:
+		return 5 + ea->ea_name_len + 1;
+	case GFS2_EATYPE_SYS:
+		return 7 + ea->ea_name_len + 1;
+	case GFS2_EATYPE_SECURITY:
+		return 9 + ea->ea_name_len + 1;
+	default:
+		return 0;
+	}
+}
+
 static int ea_list_i(struct gfs2_inode *ip, struct buffer_head *bh,
 		     struct gfs2_ea_header *ea, struct gfs2_ea_header *prev,
 		     void *private)
@@ -392,21 +414,25 @@ static int ea_list_i(struct gfs2_inode *ip, struct buffer_head *bh,
 }
 
 /**
- * gfs2_ea_list -
- * @ip:
- * @er:
+ * gfs2_listxattr - List gfs2 extended attributes
+ * @dentry: The dentry whose inode we are interested in
+ * @buffer: The buffer to write the results
+ * @size: The size of the buffer
  *
  * Returns: actual size of data on success, -errno on error
  */
 
-int gfs2_ea_list(struct gfs2_inode *ip, struct gfs2_ea_request *er)
+ssize_t gfs2_listxattr(struct dentry *dentry, char *buffer, size_t size)
 {
+	struct gfs2_inode *ip = GFS2_I(dentry->d_inode);
+	struct gfs2_ea_request er;
 	struct gfs2_holder i_gh;
 	int error;
 
-	if (!er->er_data || !er->er_data_len) {
-		er->er_data = NULL;
-		er->er_data_len = 0;
+	memset(&er, 0, sizeof(struct gfs2_ea_request));
+	if (size) {
+		er.er_data = buffer;
+		er.er_data_len = size;
 	}
 
 	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh);
@@ -414,7 +440,7 @@ int gfs2_ea_list(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 		return error;
 
 	if (ip->i_eattr) {
-		struct ea_list ei = { .ei_er = er, .ei_size = 0 };
+		struct ea_list ei = { .ei_er = &er, .ei_size = 0 };
 
 		error = ea_foreach(ip, ea_list_i, &ei);
 		if (!error)
@@ -491,84 +517,61 @@ out:
 }
 
 int gfs2_ea_get_copy(struct gfs2_inode *ip, struct gfs2_ea_location *el,
-		     char *data)
+		     char *data, size_t size)
 {
+	int ret;
+	size_t len = GFS2_EA_DATA_LEN(el->el_ea);
+	if (len > size)
+		return -ERANGE;
+
 	if (GFS2_EA_IS_STUFFED(el->el_ea)) {
-		memcpy(data, GFS2_EA2DATA(el->el_ea), GFS2_EA_DATA_LEN(el->el_ea));
-		return 0;
-	} else
-		return ea_get_unstuffed(ip, el->el_ea, data);
+		memcpy(data, GFS2_EA2DATA(el->el_ea), len);
+		return len;
+	}
+	ret = ea_get_unstuffed(ip, el->el_ea, data);
+	if (ret < 0)
+		return ret;
+	return len;
 }
 
 /**
- * gfs2_ea_get_i -
- * @ip: The GFS2 inode
- * @er: The request structure
+ * gfs2_xattr_get - Get a GFS2 extended attribute
+ * @inode: The inode
+ * @type: The type of extended attribute
+ * @name: The name of the extended attribute
+ * @buffer: The buffer to write the result into
+ * @size: The size of the buffer
  *
  * Returns: actual size of data on success, -errno on error
  */
 
-int gfs2_ea_get_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
+int gfs2_xattr_get(struct inode *inode, int type, const char *name,
+		   void *buffer, size_t size)
 {
+	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_ea_location el;
 	int error;
 
 	if (!ip->i_eattr)
 		return -ENODATA;
+	if (strlen(name) > GFS2_EA_MAX_NAME_LEN)
+		return -EINVAL;
 
-	error = gfs2_ea_find(ip, er, &el);
+	error = gfs2_ea_find(ip, type, name, &el);
 	if (error)
 		return error;
 	if (!el.el_ea)
 		return -ENODATA;
-
-	if (er->er_data_len) {
-		if (GFS2_EA_DATA_LEN(el.el_ea) > er->er_data_len)
-			error =  -ERANGE;
-		else
-			error = gfs2_ea_get_copy(ip, &el, er->er_data);
-	}
-	if (!error)
+	if (size)
+		error = gfs2_ea_get_copy(ip, &el, buffer, size);
+	else
 		error = GFS2_EA_DATA_LEN(el.el_ea);
-
 	brelse(el.el_bh);
 
 	return error;
 }
 
 /**
- * gfs2_ea_get -
- * @ip: The GFS2 inode
- * @er: The request structure
- *
- * Returns: actual size of data on success, -errno on error
- */
-
-int gfs2_ea_get(struct gfs2_inode *ip, struct gfs2_ea_request *er)
-{
-	struct gfs2_holder i_gh;
-	int error;
-
-	if (!er->er_name_len ||
-	    er->er_name_len > GFS2_EA_MAX_NAME_LEN)
-		return -EINVAL;
-	if (!er->er_data || !er->er_data_len) {
-		er->er_data = NULL;
-		er->er_data_len = 0;
-	}
-
-	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh);
-	if (error)
-		return error;
-
-	error = gfs2_ea_ops[er->er_type]->eo_get(ip, er);
-
-	gfs2_glock_dq_uninit(&i_gh);
-
-	return error;
-}
-
-/**
  * ea_alloc_blk - allocates a new block for extended attributes.
  * @ip: A pointer to the inode that's getting extended attributes
  * @bhp: Pointer to pointer to a struct buffer_head
@@ -713,12 +716,6 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 
 	error = gfs2_meta_inode_buffer(ip, &dibh);
 	if (!error) {
-		if (er->er_flags & GFS2_ERF_MODE) {
-			gfs2_assert_withdraw(GFS2_SB(&ip->i_inode),
-					    (ip->i_inode.i_mode & S_IFMT) ==
-					    (er->er_mode & S_IFMT));
-			ip->i_inode.i_mode = er->er_mode;
-		}
 		ip->i_inode.i_ctime = CURRENT_TIME;
 		gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 		gfs2_dinode_out(ip, dibh->b_data);
@@ -762,15 +759,23 @@ static int ea_init_i(struct gfs2_inode *ip, struct gfs2_ea_request *er,
  * Returns: errno
  */
 
-static int ea_init(struct gfs2_inode *ip, struct gfs2_ea_request *er)
+static int ea_init(struct gfs2_inode *ip, int type, const char *name,
+		   const void *data, size_t size)
 {
+	struct gfs2_ea_request er;
 	unsigned int jbsize = GFS2_SB(&ip->i_inode)->sd_jbsize;
 	unsigned int blks = 1;
 
-	if (GFS2_EAREQ_SIZE_STUFFED(er) > jbsize)
-		blks += DIV_ROUND_UP(er->er_data_len, jbsize);
+	er.er_type = type;
+	er.er_name = name;
+	er.er_name_len = strlen(name);
+	er.er_data = (void *)data;
+	er.er_data_len = size;
+
+	if (GFS2_EAREQ_SIZE_STUFFED(&er) > jbsize)
+		blks += DIV_ROUND_UP(er.er_data_len, jbsize);
 
-	return ea_alloc_skeleton(ip, er, blks, ea_init_i, NULL);
+	return ea_alloc_skeleton(ip, &er, blks, ea_init_i, NULL);
 }
 
 static struct gfs2_ea_header *ea_split_ea(struct gfs2_ea_header *ea)
@@ -848,12 +853,6 @@ static int ea_set_simple_noalloc(struct gfs2_inode *ip, struct buffer_head *bh,
 	error = gfs2_meta_inode_buffer(ip, &dibh);
 	if (error)
 		goto out;
-
-	if (er->er_flags & GFS2_ERF_MODE) {
-		gfs2_assert_withdraw(GFS2_SB(&ip->i_inode),
-			(ip->i_inode.i_mode & S_IFMT) == (er->er_mode & S_IFMT));
-		ip->i_inode.i_mode = er->er_mode;
-	}
 	ip->i_inode.i_ctime = CURRENT_TIME;
 	gfs2_trans_add_bh(ip->i_gl, dibh, 1);
 	gfs2_dinode_out(ip, dibh->b_data);
@@ -894,7 +893,8 @@ static int ea_set_simple(struct gfs2_inode *ip, struct buffer_head *bh,
 	int stuffed;
 	int error;
 
-	stuffed = ea_calc_size(GFS2_SB(&ip->i_inode), es->es_er, &size);
+	stuffed = ea_calc_size(GFS2_SB(&ip->i_inode), es->es_er->er_name_len,
+			       es->es_er->er_data_len, &size);
 
 	if (ea->ea_type == GFS2_EATYPE_UNUSED) {
 		if (GFS2_EA_REC_LEN(ea) < size)
@@ -1005,15 +1005,22 @@ out:
 	return error;
 }
 
-static int ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er,
-		    struct gfs2_ea_location *el)
+static int ea_set_i(struct gfs2_inode *ip, int type, const char *name,
+		    const void *value, size_t size, struct gfs2_ea_location *el)
 {
+	struct gfs2_ea_request er;
 	struct ea_set es;
 	unsigned int blks = 2;
 	int error;
 
+	er.er_type = type;
+	er.er_name = name;
+	er.er_data = (void *)value;
+	er.er_name_len = strlen(name);
+	er.er_data_len = size;
+
 	memset(&es, 0, sizeof(struct ea_set));
-	es.es_er = er;
+	es.es_er = &er;
 	es.es_el = el;
 
 	error = ea_foreach(ip, ea_set_simple, &es);
@@ -1024,10 +1031,10 @@ static int ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er,
 
 	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);
+	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);
 
-	return ea_alloc_skeleton(ip, er, blks, ea_set_block, el);
+	return ea_alloc_skeleton(ip, &er, blks, ea_set_block, el);
 }
 
 static int ea_set_remove_unstuffed(struct gfs2_inode *ip,
@@ -1039,75 +1046,7 @@ static int ea_set_remove_unstuffed(struct gfs2_inode *ip,
 				     GFS2_EA2NEXT(el->el_prev) == el->el_ea);
 	}
 
-	return ea_remove_unstuffed(ip, el->el_bh, el->el_ea, el->el_prev,0);
-}
-
-int gfs2_ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
-{
-	struct gfs2_ea_location el;
-	int error;
-
-	if (!ip->i_eattr) {
-		if (er->er_flags & XATTR_REPLACE)
-			return -ENODATA;
-		return ea_init(ip, er);
-	}
-
-	error = gfs2_ea_find(ip, er, &el);
-	if (error)
-		return error;
-
-	if (el.el_ea) {
-		if (ip->i_diskflags & GFS2_DIF_APPENDONLY) {
-			brelse(el.el_bh);
-			return -EPERM;
-		}
-
-		error = -EEXIST;
-		if (!(er->er_flags & XATTR_CREATE)) {
-			int unstuffed = !GFS2_EA_IS_STUFFED(el.el_ea);
-			error = ea_set_i(ip, er, &el);
-			if (!error && unstuffed)
-				ea_set_remove_unstuffed(ip, &el);
-		}
-
-		brelse(el.el_bh);
-	} else {
-		error = -ENODATA;
-		if (!(er->er_flags & XATTR_REPLACE))
-			error = ea_set_i(ip, er, NULL);
-	}
-
-	return error;
-}
-
-int gfs2_ea_set(struct gfs2_inode *ip, struct gfs2_ea_request *er)
-{
-	struct gfs2_holder i_gh;
-	int error;
-
-	if (!er->er_name_len || er->er_name_len > GFS2_EA_MAX_NAME_LEN)
-		return -EINVAL;
-	if (!er->er_data || !er->er_data_len) {
-		er->er_data = NULL;
-		er->er_data_len = 0;
-	}
-	error = ea_check_size(GFS2_SB(&ip->i_inode), er);
-	if (error)
-		return error;
-
-	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &i_gh);
-	if (error)
-		return error;
-
-	if (IS_IMMUTABLE(&ip->i_inode))
-		error = -EPERM;
-	else
-		error = gfs2_ea_ops[er->er_type]->eo_set(ip, er);
-
-	gfs2_glock_dq_uninit(&i_gh);
-
-	return error;
+	return ea_remove_unstuffed(ip, el->el_bh, el->el_ea, el->el_prev, 0);
 }
 
 static int ea_remove_stuffed(struct gfs2_inode *ip, struct gfs2_ea_location *el)
@@ -1131,8 +1070,9 @@ static int ea_remove_stuffed(struct gfs2_inode *ip, struct gfs2_ea_location *el)
 
 		if (GFS2_EA_IS_LAST(ea))
 			prev->ea_flags |= GFS2_EAFLAG_LAST;
-	} else
+	} else {
 		ea->ea_type = GFS2_EATYPE_UNUSED;
+	}
 
 	error = gfs2_meta_inode_buffer(ip, &dibh);
 	if (!error) {
@@ -1147,15 +1087,29 @@ static int ea_remove_stuffed(struct gfs2_inode *ip, struct gfs2_ea_location *el)
 	return error;
 }
 
-int gfs2_ea_remove_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
+/**
+ * gfs2_xattr_remove - Remove a GFS2 extended attribute
+ * @inode: The inode
+ * @type: The type of the extended attribute
+ * @name: The name of the extended attribute
+ *
+ * This is not called directly by the VFS since we use the (common)
+ * scheme of making a "set with NULL data" mean a remove request. Note
+ * that this is different from a set with zero length data.
+ *
+ * Returns: 0, or errno on failure
+ */
+
+static int gfs2_xattr_remove(struct inode *inode, int type, const char *name)
 {
+	struct gfs2_inode *ip = GFS2_I(inode);
 	struct gfs2_ea_location el;
 	int error;
 
 	if (!ip->i_eattr)
 		return -ENODATA;
 
-	error = gfs2_ea_find(ip, er, &el);
+	error = gfs2_ea_find(ip, type, name, &el);
 	if (error)
 		return error;
 	if (!el.el_ea)
@@ -1164,8 +1118,7 @@ int gfs2_ea_remove_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 	if (GFS2_EA_IS_STUFFED(el.el_ea))
 		error = ea_remove_stuffed(ip, &el);
 	else
-		error = ea_remove_unstuffed(ip, el.el_bh, el.el_ea, el.el_prev,
-					    0);
+		error = ea_remove_unstuffed(ip, el.el_bh, el.el_ea, el.el_prev, 0);
 
 	brelse(el.el_bh);
 
@@ -1173,31 +1126,70 @@ int gfs2_ea_remove_i(struct gfs2_inode *ip, struct gfs2_ea_request *er)
 }
 
 /**
- * gfs2_ea_remove - sets (or creates or replaces) an extended attribute
- * @ip: pointer to the inode of the target file
- * @er: request information
+ * gfs2_xattr_set - Set (or remove) a GFS2 extended attribute
+ * @inode: The inode
+ * @type: The type of the extended attribute
+ * @name: The name of the extended attribute
+ * @value: The value of the extended attribute (NULL for remove)
+ * @size: The size of the @value argument
+ * @flags: Create or Replace
  *
- * Returns: errno
+ * See gfs2_xattr_remove() for details of the removal of xattrs.
+ *
+ * Returns: 0 or errno on failure
  */
 
-int gfs2_ea_remove(struct gfs2_inode *ip, struct gfs2_ea_request *er)
+int gfs2_xattr_set(struct inode *inode, int type, const char *name,
+		   const void *value, size_t size, int flags)
 {
-	struct gfs2_holder i_gh;
+	struct gfs2_sbd *sdp = GFS2_SB(inode);
+	struct gfs2_inode *ip = GFS2_I(inode);
+	struct gfs2_ea_location el;
+	unsigned int namel = strlen(name);
 	int error;
 
-	if (!er->er_name_len || er->er_name_len > GFS2_EA_MAX_NAME_LEN)
-		return -EINVAL;
+	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
+		return -EPERM;
+	if (namel > GFS2_EA_MAX_NAME_LEN)
+		return -ERANGE;
 
-	error = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &i_gh);
+	if (value == NULL)
+		return gfs2_xattr_remove(inode, type, name);
+
+	if (ea_check_size(sdp, namel, size))
+		return -ERANGE;
+
+	if (!ip->i_eattr) {
+		if (flags & XATTR_REPLACE)
+			return -ENODATA;
+		return ea_init(ip, type, name, value, size);
+	}
+
+	error = gfs2_ea_find(ip, type, name, &el);
 	if (error)
 		return error;
 
-	if (IS_IMMUTABLE(&ip->i_inode) || IS_APPEND(&ip->i_inode))
-		error = -EPERM;
-	else
-		error = gfs2_ea_ops[er->er_type]->eo_remove(ip, er);
+	if (el.el_ea) {
+		if (ip->i_diskflags & GFS2_DIF_APPENDONLY) {
+			brelse(el.el_bh);
+			return -EPERM;
+		}
 
-	gfs2_glock_dq_uninit(&i_gh);
+		error = -EEXIST;
+		if (!(flags & XATTR_CREATE)) {
+			int unstuffed = !GFS2_EA_IS_STUFFED(el.el_ea);
+			error = ea_set_i(ip, type, name, value, size, &el);
+			if (!error && unstuffed)
+				ea_set_remove_unstuffed(ip, &el);
+		}
+
+		brelse(el.el_bh);
+		return error;
+	}
+
+	error = -ENODATA;
+	if (!(flags & XATTR_REPLACE))
+		error = ea_set_i(ip, type, name, value, size, NULL);
 
 	return error;
 }
@@ -1503,3 +1495,64 @@ out_alloc:
 	return error;
 }
 
+static int gfs2_xattr_user_get(struct inode *inode, const char *name,
+			       void *buffer, size_t size)
+{
+	return gfs2_xattr_get(inode, GFS2_EATYPE_USR, name, buffer, size);
+}
+
+static int gfs2_xattr_user_set(struct inode *inode, const char *name,
+			       const void *value, size_t size, int flags)
+{
+	return gfs2_xattr_set(inode, GFS2_EATYPE_USR, name, value, size, flags);
+}
+
+static int gfs2_xattr_system_get(struct inode *inode, const char *name,
+				 void *buffer, size_t size)
+{
+	return gfs2_xattr_get(inode, GFS2_EATYPE_SYS, name, buffer, size);
+}
+
+static int gfs2_xattr_system_set(struct inode *inode, const char *name,
+				 const void *value, size_t size, int flags)
+{
+	return gfs2_xattr_set(inode, GFS2_EATYPE_SYS, name, value, size, flags);
+}
+
+static int gfs2_xattr_security_get(struct inode *inode, const char *name,
+				   void *buffer, size_t size)
+{
+	return gfs2_xattr_get(inode, GFS2_EATYPE_SECURITY, name, buffer, size);
+}
+
+static int gfs2_xattr_security_set(struct inode *inode, const char *name,
+				   const void *value, size_t size, int flags)
+{
+	return gfs2_xattr_set(inode, GFS2_EATYPE_SECURITY, name, value, size, flags);
+}
+
+static struct xattr_handler gfs2_xattr_user_handler = {
+	.prefix = XATTR_USER_PREFIX,
+	.get    = gfs2_xattr_user_get,
+	.set    = gfs2_xattr_user_set,
+};
+
+static struct xattr_handler gfs2_xattr_security_handler = {
+	.prefix = XATTR_SECURITY_PREFIX,
+	.get    = gfs2_xattr_security_get,
+	.set    = gfs2_xattr_security_set,
+};
+
+static struct xattr_handler gfs2_xattr_system_handler = {
+	.prefix = XATTR_SYSTEM_PREFIX,
+	.get    = gfs2_xattr_system_get,
+	.set    = gfs2_xattr_system_set,
+};
+
+struct xattr_handler *gfs2_xattr_handlers[] = {
+	&gfs2_xattr_user_handler,
+	&gfs2_xattr_security_handler,
+	&gfs2_xattr_system_handler,
+	NULL,
+};
+
diff --git a/fs/gfs2/eattr.h b/fs/gfs2/xattr.h
index c82dbe01d71..cbdfd774373 100644
--- a/fs/gfs2/eattr.h
+++ b/fs/gfs2/xattr.h
@@ -19,7 +19,7 @@ struct iattr;
 #define GFS2_EA_SIZE(ea) \
 ALIGN(sizeof(struct gfs2_ea_header) + (ea)->ea_name_len + \
       ((GFS2_EA_IS_STUFFED(ea)) ? GFS2_EA_DATA_LEN(ea) : \
-                                  (sizeof(__be64) * (ea)->ea_num_ptrs)), 8)
+				  (sizeof(__be64) * (ea)->ea_num_ptrs)), 8)
 
 #define GFS2_EA_IS_STUFFED(ea) (!(ea)->ea_num_ptrs)
 #define GFS2_EA_IS_LAST(ea) ((ea)->ea_flags & GFS2_EAFLAG_LAST)
@@ -27,10 +27,6 @@ ALIGN(sizeof(struct gfs2_ea_header) + (ea)->ea_name_len + \
 #define GFS2_EAREQ_SIZE_STUFFED(er) \
 ALIGN(sizeof(struct gfs2_ea_header) + (er)->er_name_len + (er)->er_data_len, 8)
 
-#define GFS2_EAREQ_SIZE_UNSTUFFED(sdp, er) \
-ALIGN(sizeof(struct gfs2_ea_header) + (er)->er_name_len + \
-      sizeof(__be64) * DIV_ROUND_UP((er)->er_data_len, (sdp)->sd_jbsize), 8)
-
 #define GFS2_EA2NAME(ea) ((char *)((struct gfs2_ea_header *)(ea) + 1))
 #define GFS2_EA2DATA(ea) (GFS2_EA2NAME(ea) + (ea)->ea_name_len)
 
@@ -43,16 +39,12 @@ ALIGN(sizeof(struct gfs2_ea_header) + (er)->er_name_len + \
 #define GFS2_EA_BH2FIRST(bh) \
 ((struct gfs2_ea_header *)((bh)->b_data + sizeof(struct gfs2_meta_header)))
 
-#define GFS2_ERF_MODE 0x80000000
-
 struct gfs2_ea_request {
 	const char *er_name;
 	char *er_data;
 	unsigned int er_name_len;
 	unsigned int er_data_len;
 	unsigned int er_type; /* GFS2_EATYPE_... */
-	int er_flags;
-	mode_t er_mode;
 };
 
 struct gfs2_ea_location {
@@ -61,40 +53,20 @@ struct gfs2_ea_location {
 	struct gfs2_ea_header *el_prev;
 };
 
-int gfs2_ea_get_i(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-int gfs2_ea_set_i(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-int gfs2_ea_remove_i(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-
-int gfs2_ea_list(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-int gfs2_ea_get(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-int gfs2_ea_set(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-int gfs2_ea_remove(struct gfs2_inode *ip, struct gfs2_ea_request *er);
-
-int gfs2_ea_dealloc(struct gfs2_inode *ip);
+extern int gfs2_xattr_get(struct inode *inode, int type, const char *name,
+			  void *buffer, size_t size);
+extern int gfs2_xattr_set(struct inode *inode, int type, const char *name,
+			  const void *value, size_t size, int flags);
+extern ssize_t gfs2_listxattr(struct dentry *dentry, char *buffer, size_t size);
+extern int gfs2_ea_dealloc(struct gfs2_inode *ip);
 
 /* Exported to acl.c */
 
-int gfs2_ea_find(struct gfs2_inode *ip,
-		 struct gfs2_ea_request *er,
-		 struct gfs2_ea_location *el);
-int gfs2_ea_get_copy(struct gfs2_inode *ip,
-		     struct gfs2_ea_location *el,
-		     char *data);
-int gfs2_ea_acl_chmod(struct gfs2_inode *ip, struct gfs2_ea_location *el,
-		      struct iattr *attr, char *data);
-
-static inline unsigned int gfs2_ea_strlen(struct gfs2_ea_header *ea)
-{
-	switch (ea->ea_type) {
-	case GFS2_EATYPE_USR:
-		return 5 + ea->ea_name_len + 1;
-	case GFS2_EATYPE_SYS:
-		return 7 + ea->ea_name_len + 1;
-	case GFS2_EATYPE_SECURITY:
-		return 9 + ea->ea_name_len + 1;
-	default:
-		return 0;
-	}
-}
+extern int gfs2_ea_find(struct gfs2_inode *ip, int type, const char *name,
+			struct gfs2_ea_location *el);
+extern int gfs2_ea_get_copy(struct gfs2_inode *ip, struct gfs2_ea_location *el,
+			    char *data, size_t size);
+extern int gfs2_ea_acl_chmod(struct gfs2_inode *ip, struct gfs2_ea_location *el,
+			     struct iattr *attr, char *data);
 
 #endif /* __EATTR_DOT_H__ */
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index cb88dac8cca..eba6d552d9c 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -31,12 +31,10 @@
 #include <linux/statfs.h>
 #include <linux/security.h>
 #include <linux/ima.h>
+#include <linux/magic.h>
 
 #include <asm/uaccess.h>
 
-/* some random number */
-#define HUGETLBFS_MAGIC	0x958458f6
-
 static const struct super_operations hugetlbfs_ops;
 static const struct address_space_operations hugetlbfs_aops;
 const struct file_operations hugetlbfs_file_operations;
@@ -44,6 +42,7 @@ static const struct inode_operations hugetlbfs_dir_inode_operations;
 static const struct inode_operations hugetlbfs_inode_operations;
 
 static struct backing_dev_info hugetlbfs_backing_dev_info = {
+	.name		= "hugetlbfs",
 	.ra_pages	= 0,	/* No readahead */
 	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK,
 };
@@ -506,6 +505,13 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb, uid_t uid,
 		inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 		INIT_LIST_HEAD(&inode->i_mapping->private_list);
 		info = HUGETLBFS_I(inode);
+		/*
+		 * The policy is initialized here even if we are creating a
+		 * private inode because initialization simply creates an
+		 * an empty rb tree and calls spin_lock_init(), later when we
+		 * call mpol_free_shared_policy() it will just return because
+		 * the rb tree will still be empty.
+		 */
 		mpol_shared_policy_init(&info->policy, NULL);
 		switch (mode & S_IFMT) {
 		default:
@@ -930,13 +936,19 @@ static struct file_system_type hugetlbfs_fs_type = {
 
 static struct vfsmount *hugetlbfs_vfsmount;
 
-static int can_do_hugetlb_shm(void)
+static int can_do_hugetlb_shm(int creat_flags)
 {
-	return capable(CAP_IPC_LOCK) || in_group_p(sysctl_hugetlb_shm_group);
+	if (creat_flags != HUGETLB_SHMFS_INODE)
+		return 0;
+	if (capable(CAP_IPC_LOCK))
+		return 1;
+	if (in_group_p(sysctl_hugetlb_shm_group))
+		return 1;
+	return 0;
 }
 
 struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag,
-						struct user_struct **user)
+				struct user_struct **user, int creat_flags)
 {
 	int error = -ENOMEM;
 	struct file *file;
@@ -948,7 +960,7 @@ struct file *hugetlb_file_setup(const char *name, size_t size, int acctflag,
 	if (!hugetlbfs_vfsmount)
 		return ERR_PTR(-ENOENT);
 
-	if (!can_do_hugetlb_shm()) {
+	if (!can_do_hugetlb_shm(creat_flags)) {
 		*user = current_user();
 		if (user_shm_lock(size, *user)) {
 			WARN_ONCE(1,
diff --git a/fs/inode.c b/fs/inode.c
index ae7b67e4866..76582b06ab9 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -14,6 +14,7 @@
 #include <linux/module.h>
 #include <linux/backing-dev.h>
 #include <linux/wait.h>
+#include <linux/rwsem.h>
 #include <linux/hash.h>
 #include <linux/swap.h>
 #include <linux/security.h>
@@ -87,14 +88,18 @@ static struct hlist_head *inode_hashtable __read_mostly;
 DEFINE_SPINLOCK(inode_lock);
 
 /*
- * iprune_mutex provides exclusion between the kswapd or try_to_free_pages
+ * iprune_sem provides exclusion between the kswapd or try_to_free_pages
  * icache shrinking path, and the umount path.  Without this exclusion,
  * by the time prune_icache calls iput for the inode whose pages it has
  * been invalidating, or by the time it calls clear_inode & destroy_inode
  * from its final dispose_list, the struct super_block they refer to
  * (for inode->i_sb->s_op) may already have been freed and reused.
+ *
+ * We make this an rwsem because the fastpath is icache shrinking. In
+ * some cases a filesystem may be doing a significant amount of work in
+ * its inode reclaim code, so this should improve parallelism.
  */
-static DEFINE_MUTEX(iprune_mutex);
+static DECLARE_RWSEM(iprune_sem);
 
 /*
  * Statistics gathering..
@@ -123,7 +128,7 @@ static void wake_up_inode(struct inode *inode)
 int 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 inode_operations empty_iops;
 	static const struct file_operations empty_fops;
 	struct address_space *const mapping = &inode->i_data;
 
@@ -182,9 +187,7 @@ int inode_init_always(struct super_block *sb, struct inode *inode)
 	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;
+		bdi = sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
 		mapping->backing_dev_info = bdi;
 	}
 	inode->i_private = NULL;
@@ -383,7 +386,7 @@ static int invalidate_list(struct list_head *head, struct list_head *dispose)
 		/*
 		 * We can reschedule here without worrying about the list's
 		 * consistency because the per-sb list of inodes must not
-		 * change during umount anymore, and because iprune_mutex keeps
+		 * change during umount anymore, and because iprune_sem keeps
 		 * shrink_icache_memory() away.
 		 */
 		cond_resched_lock(&inode_lock);
@@ -422,7 +425,7 @@ int invalidate_inodes(struct super_block *sb)
 	int busy;
 	LIST_HEAD(throw_away);
 
-	mutex_lock(&iprune_mutex);
+	down_write(&iprune_sem);
 	spin_lock(&inode_lock);
 	inotify_unmount_inodes(&sb->s_inodes);
 	fsnotify_unmount_inodes(&sb->s_inodes);
@@ -430,7 +433,7 @@ int invalidate_inodes(struct super_block *sb)
 	spin_unlock(&inode_lock);
 
 	dispose_list(&throw_away);
-	mutex_unlock(&iprune_mutex);
+	up_write(&iprune_sem);
 
 	return busy;
 }
@@ -469,7 +472,7 @@ static void prune_icache(int nr_to_scan)
 	int nr_scanned;
 	unsigned long reap = 0;
 
-	mutex_lock(&iprune_mutex);
+	down_read(&iprune_sem);
 	spin_lock(&inode_lock);
 	for (nr_scanned = 0; nr_scanned < nr_to_scan; nr_scanned++) {
 		struct inode *inode;
@@ -511,7 +514,7 @@ static void prune_icache(int nr_to_scan)
 	spin_unlock(&inode_lock);
 
 	dispose_list(&freeable);
-	mutex_unlock(&iprune_mutex);
+	up_read(&iprune_sem);
 }
 
 /*
@@ -697,13 +700,15 @@ void unlock_new_inode(struct inode *inode)
 	}
 #endif
 	/*
-	 * This is special!  We do not need the spinlock
-	 * when clearing I_LOCK, because we're guaranteed
-	 * that nobody else tries to do anything about the
-	 * state of the inode when it is locked, as we
-	 * just created it (so there can be no old holders
-	 * that haven't tested I_LOCK).
+	 * This is special!  We do not need the spinlock when clearing I_LOCK,
+	 * because we're guaranteed that nobody else tries to do anything about
+	 * the state of the inode when it is locked, as we just created it (so
+	 * there can be no old holders that haven't tested I_LOCK).
+	 * However we must emit the memory barrier so that other CPUs reliably
+	 * see the clearing of I_LOCK after the other inode initialisation has
+	 * completed.
 	 */
+	smp_mb();
 	WARN_ON((inode->i_state & (I_LOCK|I_NEW)) != (I_LOCK|I_NEW));
 	inode->i_state &= ~(I_LOCK|I_NEW);
 	wake_up_inode(inode);
diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c
index 61f32f3868c..b0435dd0654 100644
--- a/fs/jbd/checkpoint.c
+++ b/fs/jbd/checkpoint.c
@@ -456,7 +456,7 @@ int cleanup_journal_tail(journal_t *journal)
 {
 	transaction_t * transaction;
 	tid_t		first_tid;
-	unsigned long	blocknr, freed;
+	unsigned int	blocknr, freed;
 
 	if (is_journal_aborted(journal))
 		return 1;
@@ -502,8 +502,8 @@ int cleanup_journal_tail(journal_t *journal)
 		freed = freed + journal->j_last - journal->j_first;
 
 	jbd_debug(1,
-		  "Cleaning journal tail from %d to %d (offset %lu), "
-		  "freeing %lu\n",
+		  "Cleaning journal tail from %d to %d (offset %u), "
+		  "freeing %u\n",
 		  journal->j_tail_sequence, first_tid, blocknr, freed);
 
 	journal->j_free += freed;
diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c
index 618e21c0b7a..4bd882548c4 100644
--- a/fs/jbd/commit.c
+++ b/fs/jbd/commit.c
@@ -308,7 +308,7 @@ void journal_commit_transaction(journal_t *journal)
 	int bufs;
 	int flags;
 	int err;
-	unsigned long blocknr;
+	unsigned int blocknr;
 	ktime_t start_time;
 	u64 commit_time;
 	char *tagp = NULL;
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index f96f85092d1..bd3c073b485 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -276,7 +276,7 @@ static void journal_kill_thread(journal_t *journal)
 int journal_write_metadata_buffer(transaction_t *transaction,
 				  struct journal_head  *jh_in,
 				  struct journal_head **jh_out,
-				  unsigned long blocknr)
+				  unsigned int blocknr)
 {
 	int need_copy_out = 0;
 	int done_copy_out = 0;
@@ -567,9 +567,9 @@ int log_wait_commit(journal_t *journal, tid_t tid)
  * Log buffer allocation routines:
  */
 
-int journal_next_log_block(journal_t *journal, unsigned long *retp)
+int journal_next_log_block(journal_t *journal, unsigned int *retp)
 {
-	unsigned long blocknr;
+	unsigned int blocknr;
 
 	spin_lock(&journal->j_state_lock);
 	J_ASSERT(journal->j_free > 1);
@@ -590,11 +590,11 @@ int journal_next_log_block(journal_t *journal, unsigned long *retp)
  * this is a no-op.  If needed, we can use j_blk_offset - everything is
  * ready.
  */
-int journal_bmap(journal_t *journal, unsigned long blocknr,
-		 unsigned long *retp)
+int journal_bmap(journal_t *journal, unsigned int blocknr,
+		 unsigned int *retp)
 {
 	int err = 0;
-	unsigned long ret;
+	unsigned int ret;
 
 	if (journal->j_inode) {
 		ret = bmap(journal->j_inode, blocknr);
@@ -604,7 +604,7 @@ int journal_bmap(journal_t *journal, unsigned long blocknr,
 			char b[BDEVNAME_SIZE];
 
 			printk(KERN_ALERT "%s: journal block not found "
-					"at offset %lu on %s\n",
+					"at offset %u on %s\n",
 				__func__,
 				blocknr,
 				bdevname(journal->j_dev, b));
@@ -630,7 +630,7 @@ int journal_bmap(journal_t *journal, unsigned long blocknr,
 struct journal_head *journal_get_descriptor_buffer(journal_t *journal)
 {
 	struct buffer_head *bh;
-	unsigned long blocknr;
+	unsigned int blocknr;
 	int err;
 
 	err = journal_next_log_block(journal, &blocknr);
@@ -774,7 +774,7 @@ journal_t * journal_init_inode (struct inode *inode)
 	journal_t *journal = journal_init_common();
 	int err;
 	int n;
-	unsigned long blocknr;
+	unsigned int blocknr;
 
 	if (!journal)
 		return NULL;
@@ -846,12 +846,12 @@ static void journal_fail_superblock (journal_t *journal)
 static int journal_reset(journal_t *journal)
 {
 	journal_superblock_t *sb = journal->j_superblock;
-	unsigned long first, last;
+	unsigned int first, last;
 
 	first = be32_to_cpu(sb->s_first);
 	last = be32_to_cpu(sb->s_maxlen);
 	if (first + JFS_MIN_JOURNAL_BLOCKS > last + 1) {
-		printk(KERN_ERR "JBD: Journal too short (blocks %lu-%lu).\n",
+		printk(KERN_ERR "JBD: Journal too short (blocks %u-%u).\n",
 		       first, last);
 		journal_fail_superblock(journal);
 		return -EINVAL;
@@ -885,7 +885,7 @@ static int journal_reset(journal_t *journal)
  **/
 int journal_create(journal_t *journal)
 {
-	unsigned long blocknr;
+	unsigned int blocknr;
 	struct buffer_head *bh;
 	journal_superblock_t *sb;
 	int i, err;
@@ -969,14 +969,14 @@ void journal_update_superblock(journal_t *journal, int wait)
 	if (sb->s_start == 0 && journal->j_tail_sequence ==
 				journal->j_transaction_sequence) {
 		jbd_debug(1,"JBD: Skipping superblock update on recovered sb "
-			"(start %ld, seq %d, errno %d)\n",
+			"(start %u, seq %d, errno %d)\n",
 			journal->j_tail, journal->j_tail_sequence,
 			journal->j_errno);
 		goto out;
 	}
 
 	spin_lock(&journal->j_state_lock);
-	jbd_debug(1,"JBD: updating superblock (start %ld, seq %d, errno %d)\n",
+	jbd_debug(1,"JBD: updating superblock (start %u, seq %d, errno %d)\n",
 		  journal->j_tail, journal->j_tail_sequence, journal->j_errno);
 
 	sb->s_sequence = cpu_to_be32(journal->j_tail_sequence);
@@ -1371,7 +1371,7 @@ int journal_flush(journal_t *journal)
 {
 	int err = 0;
 	transaction_t *transaction = NULL;
-	unsigned long old_tail;
+	unsigned int old_tail;
 
 	spin_lock(&journal->j_state_lock);
 
diff --git a/fs/jbd/recovery.c b/fs/jbd/recovery.c
index db5e982c5dd..cb1a49ae605 100644
--- a/fs/jbd/recovery.c
+++ b/fs/jbd/recovery.c
@@ -70,7 +70,7 @@ static int do_readahead(journal_t *journal, unsigned int start)
 {
 	int err;
 	unsigned int max, nbufs, next;
-	unsigned long blocknr;
+	unsigned int blocknr;
 	struct buffer_head *bh;
 
 	struct buffer_head * bufs[MAXBUF];
@@ -132,7 +132,7 @@ static int jread(struct buffer_head **bhp, journal_t *journal,
 		 unsigned int offset)
 {
 	int err;
-	unsigned long blocknr;
+	unsigned int blocknr;
 	struct buffer_head *bh;
 
 	*bhp = NULL;
@@ -314,7 +314,7 @@ static int do_one_pass(journal_t *journal,
 			struct recovery_info *info, enum passtype pass)
 {
 	unsigned int		first_commit_ID, next_commit_ID;
-	unsigned long		next_log_block;
+	unsigned int		next_log_block;
 	int			err, success = 0;
 	journal_superblock_t *	sb;
 	journal_header_t *	tmp;
@@ -367,14 +367,14 @@ static int do_one_pass(journal_t *journal,
 			if (tid_geq(next_commit_ID, info->end_transaction))
 				break;
 
-		jbd_debug(2, "Scanning for sequence ID %u at %lu/%lu\n",
+		jbd_debug(2, "Scanning for sequence ID %u at %u/%u\n",
 			  next_commit_ID, next_log_block, journal->j_last);
 
 		/* Skip over each chunk of the transaction looking
 		 * either the next descriptor block or the final commit
 		 * record. */
 
-		jbd_debug(3, "JBD: checking block %ld\n", next_log_block);
+		jbd_debug(3, "JBD: checking block %u\n", next_log_block);
 		err = jread(&bh, journal, next_log_block);
 		if (err)
 			goto failed;
@@ -429,7 +429,7 @@ static int do_one_pass(journal_t *journal,
 			tagp = &bh->b_data[sizeof(journal_header_t)];
 			while ((tagp - bh->b_data +sizeof(journal_block_tag_t))
 			       <= journal->j_blocksize) {
-				unsigned long io_block;
+				unsigned int io_block;
 
 				tag = (journal_block_tag_t *) tagp;
 				flags = be32_to_cpu(tag->t_flags);
@@ -443,10 +443,10 @@ static int do_one_pass(journal_t *journal,
 					success = err;
 					printk (KERN_ERR
 						"JBD: IO error %d recovering "
-						"block %ld in log\n",
+						"block %u in log\n",
 						err, io_block);
 				} else {
-					unsigned long blocknr;
+					unsigned int blocknr;
 
 					J_ASSERT(obh != NULL);
 					blocknr = be32_to_cpu(tag->t_blocknr);
@@ -581,7 +581,7 @@ static int scan_revoke_records(journal_t *journal, struct buffer_head *bh,
 	max = be32_to_cpu(header->r_count);
 
 	while (offset < max) {
-		unsigned long blocknr;
+		unsigned int blocknr;
 		int err;
 
 		blocknr = be32_to_cpu(* ((__be32 *) (bh->b_data+offset)));
diff --git a/fs/jbd/revoke.c b/fs/jbd/revoke.c
index da6cd9bdaab..ad717328343 100644
--- a/fs/jbd/revoke.c
+++ b/fs/jbd/revoke.c
@@ -101,7 +101,7 @@ struct jbd_revoke_record_s
 {
 	struct list_head  hash;
 	tid_t		  sequence;	/* Used for recovery only */
-	unsigned long	  blocknr;
+	unsigned int	  blocknr;
 };
 
 
@@ -126,7 +126,7 @@ static void flush_descriptor(journal_t *, struct journal_head *, int, int);
 /* Utility functions to maintain the revoke table */
 
 /* Borrowed from buffer.c: this is a tried and tested block hash function */
-static inline int hash(journal_t *journal, unsigned long block)
+static inline int hash(journal_t *journal, unsigned int block)
 {
 	struct jbd_revoke_table_s *table = journal->j_revoke;
 	int hash_shift = table->hash_shift;
@@ -136,7 +136,7 @@ static inline int hash(journal_t *journal, unsigned long block)
 		(block << (hash_shift - 12))) & (table->hash_size - 1);
 }
 
-static int insert_revoke_hash(journal_t *journal, unsigned long blocknr,
+static int insert_revoke_hash(journal_t *journal, unsigned int blocknr,
 			      tid_t seq)
 {
 	struct list_head *hash_list;
@@ -166,7 +166,7 @@ oom:
 /* Find a revoke record in the journal's hash table. */
 
 static struct jbd_revoke_record_s *find_revoke_record(journal_t *journal,
-						      unsigned long blocknr)
+						      unsigned int blocknr)
 {
 	struct list_head *hash_list;
 	struct jbd_revoke_record_s *record;
@@ -332,7 +332,7 @@ void journal_destroy_revoke(journal_t *journal)
  * by one.
  */
 
-int journal_revoke(handle_t *handle, unsigned long blocknr,
+int journal_revoke(handle_t *handle, unsigned int blocknr,
 		   struct buffer_head *bh_in)
 {
 	struct buffer_head *bh = NULL;
@@ -401,7 +401,7 @@ int journal_revoke(handle_t *handle, unsigned long blocknr,
 		}
 	}
 
-	jbd_debug(2, "insert revoke for block %lu, bh_in=%p\n", blocknr, bh_in);
+	jbd_debug(2, "insert revoke for block %u, bh_in=%p\n", blocknr, bh_in);
 	err = insert_revoke_hash(journal, blocknr,
 				handle->h_transaction->t_tid);
 	BUFFER_TRACE(bh_in, "exit");
@@ -644,7 +644,7 @@ static void flush_descriptor(journal_t *journal,
  */
 
 int journal_set_revoke(journal_t *journal,
-		       unsigned long blocknr,
+		       unsigned int blocknr,
 		       tid_t sequence)
 {
 	struct jbd_revoke_record_s *record;
@@ -668,7 +668,7 @@ int journal_set_revoke(journal_t *journal,
  */
 
 int journal_test_revoke(journal_t *journal,
-			unsigned long blocknr,
+			unsigned int blocknr,
 			tid_t sequence)
 {
 	struct jbd_revoke_record_s *record;
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index c03ac11f74b..006f9ad838a 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -56,7 +56,8 @@ get_transaction(journal_t *journal, transaction_t *transaction)
 	spin_lock_init(&transaction->t_handle_lock);
 
 	/* Set up the commit timer for the new transaction. */
-	journal->j_commit_timer.expires = round_jiffies(transaction->t_expires);
+	journal->j_commit_timer.expires =
+				round_jiffies_up(transaction->t_expires);
 	add_timer(&journal->j_commit_timer);
 
 	J_ASSERT(journal->j_running_transaction == NULL);
@@ -228,6 +229,8 @@ repeat_locked:
 		  __log_space_left(journal));
 	spin_unlock(&transaction->t_handle_lock);
 	spin_unlock(&journal->j_state_lock);
+
+	lock_map_acquire(&handle->h_lockdep_map);
 out:
 	if (unlikely(new_transaction))		/* It's usually NULL */
 		kfree(new_transaction);
@@ -292,9 +295,6 @@ handle_t *journal_start(journal_t *journal, int nblocks)
 		handle = ERR_PTR(err);
 		goto out;
 	}
-
-	lock_map_acquire(&handle->h_lockdep_map);
-
 out:
 	return handle;
 }
@@ -416,6 +416,7 @@ int journal_restart(handle_t *handle, int nblocks)
 	__log_start_commit(journal, transaction->t_tid);
 	spin_unlock(&journal->j_state_lock);
 
+	lock_map_release(&handle->h_lockdep_map);
 	handle->h_buffer_credits = nblocks;
 	ret = start_this_handle(journal, handle);
 	return ret;
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 7b4088b2364..26d991ddc1e 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -25,6 +25,7 @@
 #include <linux/writeback.h>
 #include <linux/backing-dev.h>
 #include <linux/bio.h>
+#include <linux/blkdev.h>
 #include <trace/events/jbd2.h>
 
 /*
@@ -133,8 +134,8 @@ static int journal_submit_commit_record(journal_t *journal,
 	bh->b_end_io = journal_end_buffer_io_sync;
 
 	if (journal->j_flags & JBD2_BARRIER &&
-		!JBD2_HAS_INCOMPAT_FEATURE(journal,
-					 JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
+	    !JBD2_HAS_INCOMPAT_FEATURE(journal,
+				       JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
 		set_buffer_ordered(bh);
 		barrier_done = 1;
 	}
@@ -220,7 +221,6 @@ static int journal_submit_inode_data_buffers(struct address_space *mapping)
 		.nr_to_write = mapping->nrpages * 2,
 		.range_start = 0,
 		.range_end = i_size_read(mapping->host),
-		.for_writepages = 1,
 	};
 
 	ret = generic_writepages(mapping, &wbc);
@@ -707,11 +707,13 @@ start_journal_io:
 	/* Done it all: now write the commit record asynchronously. */
 
 	if (JBD2_HAS_INCOMPAT_FEATURE(journal,
-		JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
+				      JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
 		err = journal_submit_commit_record(journal, commit_transaction,
 						 &cbh, crc32_sum);
 		if (err)
 			__jbd2_journal_abort_hard(journal);
+		if (journal->j_flags & JBD2_BARRIER)
+			blkdev_issue_flush(journal->j_dev, NULL);
 	}
 
 	/*
@@ -834,7 +836,7 @@ wait_for_iobuf:
 	jbd_debug(3, "JBD: commit phase 5\n");
 
 	if (!JBD2_HAS_INCOMPAT_FEATURE(journal,
-		JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
+				       JBD2_FEATURE_INCOMPAT_ASYNC_COMMIT)) {
 		err = journal_submit_commit_record(journal, commit_transaction,
 						&cbh, crc32_sum);
 		if (err)
diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index e378cb38397..53b86e16e5f 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -768,7 +768,7 @@ static void jbd2_seq_history_stop(struct seq_file *seq, void *v)
 {
 }
 
-static struct seq_operations jbd2_seq_history_ops = {
+static const struct seq_operations jbd2_seq_history_ops = {
 	.start  = jbd2_seq_history_start,
 	.next   = jbd2_seq_history_next,
 	.stop   = jbd2_seq_history_stop,
@@ -872,7 +872,7 @@ static void jbd2_seq_info_stop(struct seq_file *seq, void *v)
 {
 }
 
-static struct seq_operations jbd2_seq_info_ops = {
+static const struct seq_operations jbd2_seq_info_ops = {
 	.start  = jbd2_seq_info_start,
 	.next   = jbd2_seq_info_next,
 	.stop   = jbd2_seq_info_stop,
@@ -1187,6 +1187,12 @@ static int journal_reset(journal_t *journal)
 
 	first = be32_to_cpu(sb->s_first);
 	last = be32_to_cpu(sb->s_maxlen);
+	if (first + JBD2_MIN_JOURNAL_BLOCKS > last + 1) {
+		printk(KERN_ERR "JBD: Journal too short (blocks %llu-%llu).\n",
+		       first, last);
+		journal_fail_superblock(journal);
+		return -EINVAL;
+	}
 
 	journal->j_first = first;
 	journal->j_last = last;
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index 6213ac728f3..a0512700542 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -57,7 +57,7 @@ jbd2_get_transaction(journal_t *journal, transaction_t *transaction)
 	INIT_LIST_HEAD(&transaction->t_private_list);
 
 	/* Set up the commit timer for the new transaction. */
-	journal->j_commit_timer.expires = round_jiffies(transaction->t_expires);
+	journal->j_commit_timer.expires = round_jiffies_up(transaction->t_expires);
 	add_timer(&journal->j_commit_timer);
 
 	J_ASSERT(journal->j_running_transaction == NULL);
@@ -238,6 +238,8 @@ repeat_locked:
 		  __jbd2_log_space_left(journal));
 	spin_unlock(&transaction->t_handle_lock);
 	spin_unlock(&journal->j_state_lock);
+
+	lock_map_acquire(&handle->h_lockdep_map);
 out:
 	if (unlikely(new_transaction))		/* It's usually NULL */
 		kfree(new_transaction);
@@ -303,8 +305,6 @@ handle_t *jbd2_journal_start(journal_t *journal, int nblocks)
 		handle = ERR_PTR(err);
 		goto out;
 	}
-
-	lock_map_acquire(&handle->h_lockdep_map);
 out:
 	return handle;
 }
@@ -426,6 +426,7 @@ int jbd2_journal_restart(handle_t *handle, int nblocks)
 	__jbd2_log_start_commit(journal, transaction->t_tid);
 	spin_unlock(&journal->j_state_lock);
 
+	lock_map_release(&handle->h_lockdep_map);
 	handle->h_buffer_credits = nblocks;
 	ret = start_this_handle(journal, handle);
 	return ret;
diff --git a/fs/jffs2/acl.c b/fs/jffs2/acl.c
index 8fcb6239218..7edb62e9741 100644
--- a/fs/jffs2/acl.c
+++ b/fs/jffs2/acl.c
@@ -258,7 +258,7 @@ static int jffs2_set_acl(struct inode *inode, int type, struct posix_acl *acl)
 	return rc;
 }
 
-static int jffs2_check_acl(struct inode *inode, int mask)
+int jffs2_check_acl(struct inode *inode, int mask)
 {
 	struct posix_acl *acl;
 	int rc;
@@ -274,11 +274,6 @@ static int jffs2_check_acl(struct inode *inode, int mask)
 	return -EAGAIN;
 }
 
-int jffs2_permission(struct inode *inode, int mask)
-{
-	return generic_permission(inode, mask, jffs2_check_acl);
-}
-
 int jffs2_init_acl_pre(struct inode *dir_i, struct inode *inode, int *i_mode)
 {
 	struct posix_acl *acl, *clone;
diff --git a/fs/jffs2/acl.h b/fs/jffs2/acl.h
index fc929f2a14f..f0ba63e3c36 100644
--- a/fs/jffs2/acl.h
+++ b/fs/jffs2/acl.h
@@ -26,7 +26,7 @@ struct jffs2_acl_header {
 
 #ifdef CONFIG_JFFS2_FS_POSIX_ACL
 
-extern int jffs2_permission(struct inode *, int);
+extern int jffs2_check_acl(struct inode *, int);
 extern int jffs2_acl_chmod(struct inode *);
 extern int jffs2_init_acl_pre(struct inode *, struct inode *, int *);
 extern int jffs2_init_acl_post(struct inode *);
@@ -36,7 +36,7 @@ extern struct xattr_handler jffs2_acl_default_xattr_handler;
 
 #else
 
-#define jffs2_permission			(NULL)
+#define jffs2_check_acl				(NULL)
 #define jffs2_acl_chmod(inode)			(0)
 #define jffs2_init_acl_pre(dir_i,inode,mode)	(0)
 #define jffs2_init_acl_post(inode)		(0)
diff --git a/fs/jffs2/background.c b/fs/jffs2/background.c
index e9580104b6b..3ff50da9478 100644
--- a/fs/jffs2/background.c
+++ b/fs/jffs2/background.c
@@ -15,6 +15,7 @@
 #include <linux/completion.h>
 #include <linux/sched.h>
 #include <linux/freezer.h>
+#include <linux/kthread.h>
 #include "nodelist.h"
 
 
@@ -31,7 +32,7 @@ void jffs2_garbage_collect_trigger(struct jffs2_sb_info *c)
 /* This must only ever be called when no GC thread is currently running */
 int jffs2_start_garbage_collect_thread(struct jffs2_sb_info *c)
 {
-	pid_t pid;
+	struct task_struct *tsk;
 	int ret = 0;
 
 	BUG_ON(c->gc_task);
@@ -39,15 +40,16 @@ int jffs2_start_garbage_collect_thread(struct jffs2_sb_info *c)
 	init_completion(&c->gc_thread_start);
 	init_completion(&c->gc_thread_exit);
 
-	pid = kernel_thread(jffs2_garbage_collect_thread, c, CLONE_FS|CLONE_FILES);
-	if (pid < 0) {
-		printk(KERN_WARNING "fork failed for JFFS2 garbage collect thread: %d\n", -pid);
+	tsk = kthread_run(jffs2_garbage_collect_thread, c, "jffs2_gcd_mtd%d", c->mtd->index);
+	if (IS_ERR(tsk)) {
+		printk(KERN_WARNING "fork failed for JFFS2 garbage collect thread: %ld\n", -PTR_ERR(tsk));
 		complete(&c->gc_thread_exit);
-		ret = pid;
+		ret = PTR_ERR(tsk);
 	} else {
 		/* Wait for it... */
-		D1(printk(KERN_DEBUG "JFFS2: Garbage collect thread is pid %d\n", pid));
+		D1(printk(KERN_DEBUG "JFFS2: Garbage collect thread is pid %d\n", tsk->pid));
 		wait_for_completion(&c->gc_thread_start);
+		ret = tsk->pid;
 	}
 
 	return ret;
@@ -71,7 +73,6 @@ static int jffs2_garbage_collect_thread(void *_c)
 {
 	struct jffs2_sb_info *c = _c;
 
-	daemonize("jffs2_gcd_mtd%d", c->mtd->index);
 	allow_signal(SIGKILL);
 	allow_signal(SIGSTOP);
 	allow_signal(SIGCONT);
@@ -107,6 +108,11 @@ static int jffs2_garbage_collect_thread(void *_c)
 		 * the GC thread get there first. */
 		schedule_timeout_interruptible(msecs_to_jiffies(50));
 
+		if (kthread_should_stop()) {
+			D1(printk(KERN_DEBUG "jffs2_garbage_collect_thread():  kthread_stop() called.\n"));
+			goto die;
+		}
+
 		/* Put_super will send a SIGKILL and then wait on the sem.
 		 */
 		while (signal_pending(current) || freezing(current)) {
diff --git a/fs/jffs2/dir.c b/fs/jffs2/dir.c
index 6f60cc910f4..7aa4417e085 100644
--- a/fs/jffs2/dir.c
+++ b/fs/jffs2/dir.c
@@ -55,7 +55,7 @@ const struct inode_operations jffs2_dir_inode_operations =
 	.rmdir =	jffs2_rmdir,
 	.mknod =	jffs2_mknod,
 	.rename =	jffs2_rename,
-	.permission =	jffs2_permission,
+	.check_acl =	jffs2_check_acl,
 	.setattr =	jffs2_setattr,
 	.setxattr =	jffs2_setxattr,
 	.getxattr =	jffs2_getxattr,
diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index 23c94753986..b7b74e29914 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -56,7 +56,7 @@ const struct file_operations jffs2_file_operations =
 
 const struct inode_operations jffs2_file_inode_operations =
 {
-	.permission =	jffs2_permission,
+	.check_acl =	jffs2_check_acl,
 	.setattr =	jffs2_setattr,
 	.setxattr =	jffs2_setxattr,
 	.getxattr =	jffs2_getxattr,
diff --git a/fs/jffs2/malloc.c b/fs/jffs2/malloc.c
index 9eff2bdae8a..c082868910f 100644
--- a/fs/jffs2/malloc.c
+++ b/fs/jffs2/malloc.c
@@ -39,13 +39,13 @@ int __init jffs2_create_slab_caches(void)
 
 	raw_dirent_slab = kmem_cache_create("jffs2_raw_dirent",
 					    sizeof(struct jffs2_raw_dirent),
-					    0, 0, NULL);
+					    0, SLAB_HWCACHE_ALIGN, NULL);
 	if (!raw_dirent_slab)
 		goto err;
 
 	raw_inode_slab = kmem_cache_create("jffs2_raw_inode",
 					   sizeof(struct jffs2_raw_inode),
-					   0, 0, NULL);
+					   0, SLAB_HWCACHE_ALIGN, NULL);
 	if (!raw_inode_slab)
 		goto err;
 
diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c
index 0035c021395..9a80e8e595d 100644
--- a/fs/jffs2/super.c
+++ b/fs/jffs2/super.c
@@ -123,7 +123,7 @@ static struct dentry *jffs2_get_parent(struct dentry *child)
 	return d_obtain_alias(jffs2_iget(child->d_inode->i_sb, pino));
 }
 
-static struct export_operations jffs2_export_ops = {
+static const struct export_operations jffs2_export_ops = {
 	.get_parent = jffs2_get_parent,
 	.fh_to_dentry = jffs2_fh_to_dentry,
 	.fh_to_parent = jffs2_fh_to_parent,
diff --git a/fs/jffs2/symlink.c b/fs/jffs2/symlink.c
index b7339c3b6ad..4ec11e8bda8 100644
--- a/fs/jffs2/symlink.c
+++ b/fs/jffs2/symlink.c
@@ -21,7 +21,7 @@ const struct inode_operations jffs2_symlink_inode_operations =
 {
 	.readlink =	generic_readlink,
 	.follow_link =	jffs2_follow_link,
-	.permission =	jffs2_permission,
+	.check_acl =	jffs2_check_acl,
 	.setattr =	jffs2_setattr,
 	.setxattr =	jffs2_setxattr,
 	.getxattr =	jffs2_getxattr,
diff --git a/fs/jfs/acl.c b/fs/jfs/acl.c
index a29c7c3e3fb..d66477c3430 100644
--- a/fs/jfs/acl.c
+++ b/fs/jfs/acl.c
@@ -114,7 +114,7 @@ out:
 	return rc;
 }
 
-static int jfs_check_acl(struct inode *inode, int mask)
+int jfs_check_acl(struct inode *inode, int mask)
 {
 	struct posix_acl *acl = jfs_get_acl(inode, ACL_TYPE_ACCESS);
 
@@ -129,11 +129,6 @@ static int jfs_check_acl(struct inode *inode, int mask)
 	return -EAGAIN;
 }
 
-int jfs_permission(struct inode *inode, int mask)
-{
-	return generic_permission(inode, mask, jfs_check_acl);
-}
-
 int jfs_init_acl(tid_t tid, struct inode *inode, struct inode *dir)
 {
 	struct posix_acl *acl = NULL;
diff --git a/fs/jfs/file.c b/fs/jfs/file.c
index 7f6063acaa3..2b70fa78e4a 100644
--- a/fs/jfs/file.c
+++ b/fs/jfs/file.c
@@ -96,7 +96,7 @@ const struct inode_operations jfs_file_inode_operations = {
 	.removexattr	= jfs_removexattr,
 #ifdef CONFIG_JFS_POSIX_ACL
 	.setattr	= jfs_setattr,
-	.permission	= jfs_permission,
+	.check_acl	= jfs_check_acl,
 #endif
 };
 
diff --git a/fs/jfs/jfs_acl.h b/fs/jfs/jfs_acl.h
index 88475f10a38..b07bd417ef8 100644
--- a/fs/jfs/jfs_acl.h
+++ b/fs/jfs/jfs_acl.h
@@ -20,7 +20,7 @@
 
 #ifdef CONFIG_JFS_POSIX_ACL
 
-int jfs_permission(struct inode *, int);
+int jfs_check_acl(struct inode *, int);
 int jfs_init_acl(tid_t, struct inode *, struct inode *);
 int jfs_setattr(struct dentry *, struct iattr *);
 
diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c
index 514ee2edb92..c79a4270f08 100644
--- a/fs/jfs/namei.c
+++ b/fs/jfs/namei.c
@@ -1543,7 +1543,7 @@ const struct inode_operations jfs_dir_inode_operations = {
 	.removexattr	= jfs_removexattr,
 #ifdef CONFIG_JFS_POSIX_ACL
 	.setattr	= jfs_setattr,
-	.permission	= jfs_permission,
+	.check_acl	= jfs_check_acl,
 #endif
 };
 
diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c
index 1f3b0fc0d35..fc9032dc886 100644
--- a/fs/lockd/clntlock.c
+++ b/fs/lockd/clntlock.c
@@ -166,7 +166,7 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 		 */
 		if (fl_blocked->fl_u.nfs_fl.owner->pid != lock->svid)
 			continue;
-		if (!nlm_cmp_addr(nlm_addr(block->b_host), addr))
+		if (!rpc_cmp_addr(nlm_addr(block->b_host), addr))
 			continue;
 		if (nfs_compare_fh(NFS_FH(fl_blocked->fl_file->f_path.dentry->d_inode) ,fh) != 0)
 			continue;
diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
index 4336adba952..c81249fef11 100644
--- a/fs/lockd/clntproc.c
+++ b/fs/lockd/clntproc.c
@@ -458,7 +458,7 @@ static void nlmclnt_locks_release_private(struct file_lock *fl)
 	nlm_put_lockowner(fl->fl_u.nfs_fl.owner);
 }
 
-static struct file_lock_operations nlmclnt_lock_ops = {
+static const struct file_lock_operations nlmclnt_lock_ops = {
 	.fl_copy_lock = nlmclnt_locks_copy_lock,
 	.fl_release_private = nlmclnt_locks_release_private,
 };
diff --git a/fs/lockd/host.c b/fs/lockd/host.c
index 99d737bd432..4600c2037b8 100644
--- a/fs/lockd/host.c
+++ b/fs/lockd/host.c
@@ -87,18 +87,6 @@ static unsigned int nlm_hash_address(const struct sockaddr *sap)
 	return hash & (NLM_HOST_NRHASH - 1);
 }
 
-static void nlm_clear_port(struct sockaddr *sap)
-{
-	switch (sap->sa_family) {
-	case AF_INET:
-		((struct sockaddr_in *)sap)->sin_port = 0;
-		break;
-	case AF_INET6:
-		((struct sockaddr_in6 *)sap)->sin6_port = 0;
-		break;
-	}
-}
-
 /*
  * Common host lookup routine for server & client
  */
@@ -123,7 +111,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 	 */
 	chain = &nlm_hosts[nlm_hash_address(ni->sap)];
 	hlist_for_each_entry(host, pos, chain, h_hash) {
-		if (!nlm_cmp_addr(nlm_addr(host), ni->sap))
+		if (!rpc_cmp_addr(nlm_addr(host), ni->sap))
 			continue;
 
 		/* See if we have an NSM handle for this client */
@@ -137,7 +125,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 		if (host->h_server != ni->server)
 			continue;
 		if (ni->server &&
-		    !nlm_cmp_addr(nlm_srcaddr(host), ni->src_sap))
+		    !rpc_cmp_addr(nlm_srcaddr(host), ni->src_sap))
 			continue;
 
 		/* Move to head of hash chain. */
@@ -177,7 +165,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 	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));
+	rpc_set_port(nlm_addr(host), 0);
 	memcpy(nlm_srcaddr(host), ni->src_sap, ni->src_len);
 	host->h_version    = ni->version;
 	host->h_proto      = ni->protocol;
diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c
index 7fce1b52584..f956651d0f6 100644
--- a/fs/lockd/mon.c
+++ b/fs/lockd/mon.c
@@ -61,43 +61,6 @@ 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 = {
@@ -246,7 +209,7 @@ 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))
+		if (rpc_cmp_addr(nsm_addr(nsm), sap))
 			return nsm;
 	return NULL;
 }
@@ -307,8 +270,11 @@ static struct nsm_handle *nsm_create_handle(const struct sockaddr *sap,
 	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));
+
+	if (rpc_ntop(nsm_addr(new), new->sm_addrbuf,
+					sizeof(new->sm_addrbuf)) == 0)
+		(void)snprintf(new->sm_addrbuf, sizeof(new->sm_addrbuf),
+				"unsupported address family");
 	memcpy(new->sm_name, hostname, hostname_len);
 	new->sm_name[hostname_len] = '\0';
 
diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index e577a78d7ba..d1001790fa9 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -705,7 +705,7 @@ static int nlmsvc_same_owner(struct file_lock *fl1, struct file_lock *fl2)
 	return fl1->fl_owner == fl2->fl_owner && fl1->fl_pid == fl2->fl_pid;
 }
 
-struct lock_manager_operations nlmsvc_lock_operations = {
+const struct lock_manager_operations nlmsvc_lock_operations = {
 	.fl_compare_owner = nlmsvc_same_owner,
 	.fl_notify = nlmsvc_notify_blocked,
 	.fl_grant = nlmsvc_grant_deferred,
diff --git a/fs/lockd/svcsubs.c b/fs/lockd/svcsubs.c
index 9e4d6aab611..ad478da7ca6 100644
--- a/fs/lockd/svcsubs.c
+++ b/fs/lockd/svcsubs.c
@@ -417,7 +417,7 @@ EXPORT_SYMBOL_GPL(nlmsvc_unlock_all_by_sb);
 static int
 nlmsvc_match_ip(void *datap, struct nlm_host *host)
 {
-	return nlm_cmp_addr(nlm_srcaddr(host), datap);
+	return rpc_cmp_addr(nlm_srcaddr(host), datap);
 }
 
 /**
diff --git a/fs/locks.c b/fs/locks.c
index b6440f52178..a8794f233bc 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -434,7 +434,7 @@ static int lease_mylease_callback(struct file_lock *fl, struct file_lock *try)
 	return fl->fl_file == try->fl_file;
 }
 
-static struct lock_manager_operations lease_manager_ops = {
+static const struct lock_manager_operations lease_manager_ops = {
 	.fl_break = lease_break_callback,
 	.fl_release_private = lease_release_private_callback,
 	.fl_mylease = lease_mylease_callback,
@@ -768,7 +768,7 @@ static int flock_lock_file(struct file *filp, struct file_lock *request)
 	 * give it the opportunity to lock the file.
 	 */
 	if (found)
-		cond_resched_bkl();
+		cond_resched();
 
 find_conflict:
 	for_each_lock(inode, before) {
@@ -1591,7 +1591,7 @@ SYSCALL_DEFINE2(flock, unsigned int, fd, unsigned int, cmd)
 	if (can_sleep)
 		lock->fl_flags |= FL_SLEEP;
 
-	error = security_file_lock(filp, cmd);
+	error = security_file_lock(filp, lock->fl_type);
 	if (error)
 		goto out_free;
 
diff --git a/fs/minix/dir.c b/fs/minix/dir.c
index d407e7a0b6f..6198731d7fc 100644
--- a/fs/minix/dir.c
+++ b/fs/minix/dir.c
@@ -308,14 +308,18 @@ int minix_delete_entry(struct minix_dir_entry *de, struct page *page)
 	struct inode *inode = (struct inode*)mapping->host;
 	char *kaddr = page_address(page);
 	loff_t pos = page_offset(page) + (char*)de - kaddr;
-	unsigned len = minix_sb(inode->i_sb)->s_dirsize;
+	struct minix_sb_info *sbi = minix_sb(inode->i_sb);
+	unsigned len = sbi->s_dirsize;
 	int err;
 
 	lock_page(page);
 	err = __minix_write_begin(NULL, mapping, pos, len,
 					AOP_FLAG_UNINTERRUPTIBLE, &page, NULL);
 	if (err == 0) {
-		de->inode = 0;
+		if (sbi->s_version == MINIX_V3)
+			((minix3_dirent *) de)->inode = 0;
+		else
+			de->inode = 0;
 		err = dir_commit_chunk(page, pos, len);
 	} else {
 		unlock_page(page);
@@ -440,7 +444,10 @@ void minix_set_link(struct minix_dir_entry *de, struct page *page,
 	err = __minix_write_begin(NULL, mapping, pos, sbi->s_dirsize,
 					AOP_FLAG_UNINTERRUPTIBLE, &page, NULL);
 	if (err == 0) {
-		de->inode = inode->i_ino;
+		if (sbi->s_version == MINIX_V3)
+			((minix3_dirent *) de)->inode = inode->i_ino;
+		else
+			de->inode = inode->i_ino;
 		err = dir_commit_chunk(page, pos, sbi->s_dirsize);
 	} else {
 		unlock_page(page);
@@ -470,7 +477,14 @@ ino_t minix_inode_by_name(struct dentry *dentry)
 	ino_t res = 0;
 
 	if (de) {
-		res = de->inode;
+		struct address_space *mapping = page->mapping;
+		struct inode *inode = mapping->host;
+		struct minix_sb_info *sbi = minix_sb(inode->i_sb);
+
+		if (sbi->s_version == MINIX_V3)
+			res = ((minix3_dirent *) de)->inode;
+		else
+			res = de->inode;
 		dir_put_page(page);
 	}
 	return res;
diff --git a/fs/namei.c b/fs/namei.c
index f3c5b278895..d11f404667e 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -169,19 +169,10 @@ void putname(const char *name)
 EXPORT_SYMBOL(putname);
 #endif
 
-
-/**
- * generic_permission  -  check for access rights on a Posix-like filesystem
- * @inode:	inode to check access rights for
- * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
- * @check_acl:	optional callback to check for Posix ACLs
- *
- * Used to check for read/write/execute permissions on a file.
- * We use "fsuid" for this, letting us set arbitrary permissions
- * for filesystem access without changing the "normal" uids which
- * are used for other things..
+/*
+ * This does basic POSIX ACL permission checking
  */
-int generic_permission(struct inode *inode, int mask,
+static int acl_permission_check(struct inode *inode, int mask,
 		int (*check_acl)(struct inode *inode, int mask))
 {
 	umode_t			mode = inode->i_mode;
@@ -193,9 +184,7 @@ int generic_permission(struct inode *inode, int mask,
 	else {
 		if (IS_POSIXACL(inode) && (mode & S_IRWXG) && check_acl) {
 			int error = check_acl(inode, mask);
-			if (error == -EACCES)
-				goto check_capabilities;
-			else if (error != -EAGAIN)
+			if (error != -EAGAIN)
 				return error;
 		}
 
@@ -208,8 +197,32 @@ int generic_permission(struct inode *inode, int mask,
 	 */
 	if ((mask & ~mode) == 0)
 		return 0;
+	return -EACCES;
+}
+
+/**
+ * generic_permission  -  check for access rights on a Posix-like filesystem
+ * @inode:	inode to check access rights for
+ * @mask:	right to check for (%MAY_READ, %MAY_WRITE, %MAY_EXEC)
+ * @check_acl:	optional callback to check for Posix ACLs
+ *
+ * Used to check for read/write/execute permissions on a file.
+ * 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 generic_permission(struct inode *inode, int mask,
+		int (*check_acl)(struct inode *inode, int mask))
+{
+	int ret;
+
+	/*
+	 * Do the basic POSIX ACL permission checks.
+	 */
+	ret = acl_permission_check(inode, mask, check_acl);
+	if (ret != -EACCES)
+		return ret;
 
- check_capabilities:
 	/*
 	 * Read/write DACs are always overridable.
 	 * Executable DACs are overridable if at least one exec bit is set.
@@ -262,7 +275,7 @@ int inode_permission(struct inode *inode, int mask)
 	if (inode->i_op->permission)
 		retval = inode->i_op->permission(inode, mask);
 	else
-		retval = generic_permission(inode, mask, NULL);
+		retval = generic_permission(inode, mask, inode->i_op->check_acl);
 
 	if (retval)
 		return retval;
@@ -432,29 +445,22 @@ static struct dentry * cached_lookup(struct dentry * parent, struct qstr * name,
  */
 static int exec_permission_lite(struct inode *inode)
 {
-	umode_t	mode = inode->i_mode;
+	int ret;
 
-	if (inode->i_op->permission)
-		return -EAGAIN;
-
-	if (current_fsuid() == inode->i_uid)
-		mode >>= 6;
-	else if (in_group_p(inode->i_gid))
-		mode >>= 3;
-
-	if (mode & MAY_EXEC)
-		goto ok;
-
-	if ((inode->i_mode & S_IXUGO) && capable(CAP_DAC_OVERRIDE))
-		goto ok;
-
-	if (S_ISDIR(inode->i_mode) && capable(CAP_DAC_OVERRIDE))
+	if (inode->i_op->permission) {
+		ret = inode->i_op->permission(inode, MAY_EXEC);
+		if (!ret)
+			goto ok;
+		return ret;
+	}
+	ret = acl_permission_check(inode, MAY_EXEC, inode->i_op->check_acl);
+	if (!ret)
 		goto ok;
 
-	if (S_ISDIR(inode->i_mode) && capable(CAP_DAC_READ_SEARCH))
+	if (capable(CAP_DAC_OVERRIDE) || capable(CAP_DAC_READ_SEARCH))
 		goto ok;
 
-	return -EACCES;
+	return ret;
 ok:
 	return security_inode_permission(inode, MAY_EXEC);
 }
@@ -853,12 +859,6 @@ static int __link_path_walk(const char *name, struct nameidata *nd)
 
 		nd->flags |= LOOKUP_CONTINUE;
 		err = exec_permission_lite(inode);
-		if (err == -EAGAIN)
-			err = inode_permission(nd->path.dentry->d_inode,
-					       MAY_EXEC);
-		if (!err)
-			err = ima_path_check(&nd->path, MAY_EXEC,
-				             IMA_COUNT_UPDATE);
  		if (err)
 			break;
 
@@ -1533,37 +1533,42 @@ int may_open(struct path *path, int acc_mode, int flag)
 	if (error)
 		return error;
 
-	error = ima_path_check(path,
-			       acc_mode & (MAY_READ | MAY_WRITE | MAY_EXEC),
+	error = ima_path_check(path, acc_mode ?
+			       acc_mode & (MAY_READ | MAY_WRITE | MAY_EXEC) :
+			       ACC_MODE(flag) & (MAY_READ | MAY_WRITE),
 			       IMA_COUNT_UPDATE);
+
 	if (error)
 		return error;
 	/*
 	 * An append-only file must be opened in append mode for writing.
 	 */
 	if (IS_APPEND(inode)) {
+		error = -EPERM;
 		if  ((flag & FMODE_WRITE) && !(flag & O_APPEND))
-			return -EPERM;
+			goto err_out;
 		if (flag & O_TRUNC)
-			return -EPERM;
+			goto err_out;
 	}
 
 	/* O_NOATIME can only be set by the owner or superuser */
 	if (flag & O_NOATIME)
-		if (!is_owner_or_cap(inode))
-			return -EPERM;
+		if (!is_owner_or_cap(inode)) {
+			error = -EPERM;
+			goto err_out;
+		}
 
 	/*
 	 * Ensure there are no outstanding leases on the file.
 	 */
 	error = break_lease(inode, flag);
 	if (error)
-		return error;
+		goto err_out;
 
 	if (flag & O_TRUNC) {
 		error = get_write_access(inode);
 		if (error)
-			return error;
+			goto err_out;
 
 		/*
 		 * Refuse to truncate files with mandatory locks held on them.
@@ -1581,12 +1586,17 @@ int may_open(struct path *path, int acc_mode, int flag)
 		}
 		put_write_access(inode);
 		if (error)
-			return error;
+			goto err_out;
 	} else
 		if (flag & FMODE_WRITE)
 			vfs_dq_init(inode);
 
 	return 0;
+err_out:
+	ima_counts_put(path, acc_mode ?
+		       acc_mode & (MAY_READ | MAY_WRITE | MAY_EXEC) :
+		       ACC_MODE(flag) & (MAY_READ | MAY_WRITE));
+	return error;
 }
 
 /*
diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c
index 9c590722d87..b8b5b30d53f 100644
--- a/fs/ncpfs/dir.c
+++ b/fs/ncpfs/dir.c
@@ -1241,7 +1241,7 @@ ncp_date_unix2dos(int unix_date, __le16 *time, __le16 *date)
 		month = 2;
 	} else {
 		nl_day = (year & 3) || day <= 59 ? day : day - 1;
-		for (month = 0; month < 12; month++)
+		for (month = 1; month < 12; month++)
 			if (day_n[month] > nl_day)
 				break;
 	}
diff --git a/fs/ncpfs/ioctl.c b/fs/ncpfs/ioctl.c
index fa038df63ac..53a7ed7eb9c 100644
--- a/fs/ncpfs/ioctl.c
+++ b/fs/ncpfs/ioctl.c
@@ -442,7 +442,7 @@ static int __ncp_ioctl(struct inode *inode, struct file *filp,
 			if (dentry) {
 				struct inode* s_inode = dentry->d_inode;
 				
-				if (inode) {
+				if (s_inode) {
 					NCP_FINFO(s_inode)->volNumber = vnum;
 					NCP_FINFO(s_inode)->dirEntNum = de;
 					NCP_FINFO(s_inode)->DosDirNum = dosde;
diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile
index 845159814de..da7fda639ea 100644
--- a/fs/nfs/Makefile
+++ b/fs/nfs/Makefile
@@ -6,7 +6,8 @@ obj-$(CONFIG_NFS_FS) += nfs.o
 
 nfs-y 			:= client.o dir.o file.o getroot.o inode.o super.o nfs2xdr.o \
 			   direct.o pagelist.o proc.o read.o symlink.o unlink.o \
-			   write.o namespace.o mount_clnt.o
+			   write.o namespace.o mount_clnt.o \
+			   dns_resolve.o cache_lib.o
 nfs-$(CONFIG_ROOT_NFS)	+= nfsroot.o
 nfs-$(CONFIG_NFS_V3)	+= nfs3proc.o nfs3xdr.o
 nfs-$(CONFIG_NFS_V3_ACL)	+= nfs3acl.o
diff --git a/fs/nfs/cache_lib.c b/fs/nfs/cache_lib.c
new file mode 100644
index 00000000000..b4ffd0146ea
--- /dev/null
+++ b/fs/nfs/cache_lib.c
@@ -0,0 +1,140 @@
+/*
+ * linux/fs/nfs/cache_lib.c
+ *
+ * Helper routines for the NFS client caches
+ *
+ * Copyright (c) 2009 Trond Myklebust <Trond.Myklebust@netapp.com>
+ */
+#include <linux/kmod.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/sunrpc/cache.h>
+#include <linux/sunrpc/rpc_pipe_fs.h>
+
+#include "cache_lib.h"
+
+#define NFS_CACHE_UPCALL_PATHLEN 256
+#define NFS_CACHE_UPCALL_TIMEOUT 15
+
+static char nfs_cache_getent_prog[NFS_CACHE_UPCALL_PATHLEN] =
+				"/sbin/nfs_cache_getent";
+static unsigned long nfs_cache_getent_timeout = NFS_CACHE_UPCALL_TIMEOUT;
+
+module_param_string(cache_getent, nfs_cache_getent_prog,
+		sizeof(nfs_cache_getent_prog), 0600);
+MODULE_PARM_DESC(cache_getent, "Path to the client cache upcall program");
+module_param_named(cache_getent_timeout, nfs_cache_getent_timeout, ulong, 0600);
+MODULE_PARM_DESC(cache_getent_timeout, "Timeout (in seconds) after which "
+		"the cache upcall is assumed to have failed");
+
+int nfs_cache_upcall(struct cache_detail *cd, char *entry_name)
+{
+	static char *envp[] = { "HOME=/",
+		"TERM=linux",
+		"PATH=/sbin:/usr/sbin:/bin:/usr/bin",
+		NULL
+	};
+	char *argv[] = {
+		nfs_cache_getent_prog,
+		cd->name,
+		entry_name,
+		NULL
+	};
+	int ret = -EACCES;
+
+	if (nfs_cache_getent_prog[0] == '\0')
+		goto out;
+	ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_EXEC);
+	/*
+	 * Disable the upcall mechanism if we're getting an ENOENT or
+	 * EACCES error. The admin can re-enable it on the fly by using
+	 * sysfs to set the 'cache_getent' parameter once the problem
+	 * has been fixed.
+	 */
+	if (ret == -ENOENT || ret == -EACCES)
+		nfs_cache_getent_prog[0] = '\0';
+out:
+	return ret > 0 ? 0 : ret;
+}
+
+/*
+ * Deferred request handling
+ */
+void nfs_cache_defer_req_put(struct nfs_cache_defer_req *dreq)
+{
+	if (atomic_dec_and_test(&dreq->count))
+		kfree(dreq);
+}
+
+static void nfs_dns_cache_revisit(struct cache_deferred_req *d, int toomany)
+{
+	struct nfs_cache_defer_req *dreq;
+
+	dreq = container_of(d, struct nfs_cache_defer_req, deferred_req);
+
+	complete_all(&dreq->completion);
+	nfs_cache_defer_req_put(dreq);
+}
+
+static struct cache_deferred_req *nfs_dns_cache_defer(struct cache_req *req)
+{
+	struct nfs_cache_defer_req *dreq;
+
+	dreq = container_of(req, struct nfs_cache_defer_req, req);
+	dreq->deferred_req.revisit = nfs_dns_cache_revisit;
+	atomic_inc(&dreq->count);
+
+	return &dreq->deferred_req;
+}
+
+struct nfs_cache_defer_req *nfs_cache_defer_req_alloc(void)
+{
+	struct nfs_cache_defer_req *dreq;
+
+	dreq = kzalloc(sizeof(*dreq), GFP_KERNEL);
+	if (dreq) {
+		init_completion(&dreq->completion);
+		atomic_set(&dreq->count, 1);
+		dreq->req.defer = nfs_dns_cache_defer;
+	}
+	return dreq;
+}
+
+int nfs_cache_wait_for_upcall(struct nfs_cache_defer_req *dreq)
+{
+	if (wait_for_completion_timeout(&dreq->completion,
+			nfs_cache_getent_timeout * HZ) == 0)
+		return -ETIMEDOUT;
+	return 0;
+}
+
+int nfs_cache_register(struct cache_detail *cd)
+{
+	struct nameidata nd;
+	struct vfsmount *mnt;
+	int ret;
+
+	mnt = rpc_get_mount();
+	if (IS_ERR(mnt))
+		return PTR_ERR(mnt);
+	ret = vfs_path_lookup(mnt->mnt_root, mnt, "/cache", 0, &nd);
+	if (ret)
+		goto err;
+	ret = sunrpc_cache_register_pipefs(nd.path.dentry,
+			cd->name, 0600, cd);
+	path_put(&nd.path);
+	if (!ret)
+		return ret;
+err:
+	rpc_put_mount();
+	return ret;
+}
+
+void nfs_cache_unregister(struct cache_detail *cd)
+{
+	sunrpc_cache_unregister_pipefs(cd);
+	rpc_put_mount();
+}
+
diff --git a/fs/nfs/cache_lib.h b/fs/nfs/cache_lib.h
new file mode 100644
index 00000000000..76f856e284e
--- /dev/null
+++ b/fs/nfs/cache_lib.h
@@ -0,0 +1,27 @@
+/*
+ * Helper routines for the NFS client caches
+ *
+ * Copyright (c) 2009 Trond Myklebust <Trond.Myklebust@netapp.com>
+ */
+
+#include <linux/completion.h>
+#include <linux/sunrpc/cache.h>
+#include <asm/atomic.h>
+
+/*
+ * Deferred request handling
+ */
+struct nfs_cache_defer_req {
+	struct cache_req req;
+	struct cache_deferred_req deferred_req;
+	struct completion completion;
+	atomic_t count;
+};
+
+extern int nfs_cache_upcall(struct cache_detail *cd, char *entry_name);
+extern struct nfs_cache_defer_req *nfs_cache_defer_req_alloc(void);
+extern void nfs_cache_defer_req_put(struct nfs_cache_defer_req *dreq);
+extern int nfs_cache_wait_for_upcall(struct nfs_cache_defer_req *dreq);
+
+extern int nfs_cache_register(struct cache_detail *cd);
+extern void nfs_cache_unregister(struct cache_detail *cd);
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index 7f604c7941f..293fa0528a6 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -43,21 +43,29 @@ static struct svc_program nfs4_callback_program;
 unsigned int nfs_callback_set_tcpport;
 unsigned short nfs_callback_tcpport;
 unsigned short nfs_callback_tcpport6;
-static const int nfs_set_port_min = 0;
-static const int nfs_set_port_max = 65535;
+#define NFS_CALLBACK_MAXPORTNR (65535U)
 
-static int param_set_port(const char *val, struct kernel_param *kp)
+static int param_set_portnr(const char *val, struct kernel_param *kp)
 {
-	char *endp;
-	int num = simple_strtol(val, &endp, 0);
-	if (endp == val || *endp || num < nfs_set_port_min || num > nfs_set_port_max)
+	unsigned long num;
+	int ret;
+
+	if (!val)
+		return -EINVAL;
+	ret = strict_strtoul(val, 0, &num);
+	if (ret == -EINVAL || num > NFS_CALLBACK_MAXPORTNR)
 		return -EINVAL;
-	*((int *)kp->arg) = num;
+	*((unsigned int *)kp->arg) = num;
 	return 0;
 }
 
-module_param_call(callback_tcpport, param_set_port, param_get_int,
-		 &nfs_callback_set_tcpport, 0644);
+static int param_get_portnr(char *buffer, struct kernel_param *kp)
+{
+	return param_get_uint(buffer, kp);
+}
+#define param_check_portnr(name, p) __param_check(name, p, unsigned int);
+
+module_param_named(callback_tcpport, nfs_callback_set_tcpport, portnr, 0644);
 
 /*
  * This is the NFSv4 callback kernel thread.
diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c
index e5a2dac5f71..76b0aa0f73b 100644
--- a/fs/nfs/callback_xdr.c
+++ b/fs/nfs/callback_xdr.c
@@ -222,7 +222,7 @@ static unsigned decode_sessionid(struct xdr_stream *xdr,
 
 	p = read_buf(xdr, len);
 	if (unlikely(p == NULL))
-		return htonl(NFS4ERR_RESOURCE);;
+		return htonl(NFS4ERR_RESOURCE);
 
 	memcpy(sid->data, p, len);
 	return 0;
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index 8d25ccb2d51..152025358da 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -809,6 +809,9 @@ static int nfs_init_server(struct nfs_server *server,
 	/* Initialise the client representation from the mount data */
 	server->flags = data->flags;
 	server->options = data->options;
+	server->caps |= NFS_CAP_HARDLINKS|NFS_CAP_SYMLINKS|NFS_CAP_FILEID|
+		NFS_CAP_MODE|NFS_CAP_NLINK|NFS_CAP_OWNER|NFS_CAP_OWNER_GROUP|
+		NFS_CAP_ATIME|NFS_CAP_CTIME|NFS_CAP_MTIME;
 
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);
@@ -879,6 +882,7 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, struct nfs_fsinfo *
 		server->rsize = NFS_MAX_FILE_IO_SIZE;
 	server->rpages = (server->rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
 
+	server->backing_dev_info.name = "nfs";
 	server->backing_dev_info.ra_pages = server->rpages * NFS_MAX_READAHEAD;
 
 	if (server->wsize > max_rpc_payload)
@@ -929,10 +933,6 @@ static int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *mntfh, str
 		goto out_error;
 
 	nfs_server_set_fsinfo(server, &fsinfo);
-	error = bdi_init(&server->backing_dev_info);
-	if (error)
-		goto out_error;
-
 
 	/* Get some general file system info */
 	if (server->namelen == 0) {
@@ -991,6 +991,12 @@ static struct nfs_server *nfs_alloc_server(void)
 		return NULL;
 	}
 
+	if (bdi_init(&server->backing_dev_info)) {
+		nfs_free_iostats(server->io_stats);
+		kfree(server);
+		return NULL;
+	}
+
 	return server;
 }
 
@@ -1074,10 +1080,6 @@ struct nfs_server *nfs_create_server(const struct nfs_parsed_mount_data *data,
 		(unsigned long long) server->fsid.major,
 		(unsigned long long) server->fsid.minor);
 
-	BUG_ON(!server->nfs_client);
-	BUG_ON(!server->nfs_client->rpc_ops);
-	BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops);
-
 	spin_lock(&nfs_client_lock);
 	list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks);
 	list_add_tail(&server->master_link, &nfs_volume_list);
@@ -1274,7 +1276,7 @@ static int nfs4_init_server(struct nfs_server *server,
 
 	/* Initialise the client representation from the mount data */
 	server->flags = data->flags;
-	server->caps |= NFS_CAP_ATOMIC_OPEN;
+	server->caps |= NFS_CAP_ATOMIC_OPEN|NFS_CAP_CHANGE_ATTR;
 	server->options = data->options;
 
 	/* Get a client record */
@@ -1359,10 +1361,6 @@ struct nfs_server *nfs4_create_server(const struct nfs_parsed_mount_data *data,
 	if (server->namelen == 0 || server->namelen > NFS4_MAXNAMLEN)
 		server->namelen = NFS4_MAXNAMLEN;
 
-	BUG_ON(!server->nfs_client);
-	BUG_ON(!server->nfs_client->rpc_ops);
-	BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops);
-
 	spin_lock(&nfs_client_lock);
 	list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks);
 	list_add_tail(&server->master_link, &nfs_volume_list);
@@ -1400,7 +1398,7 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 
 	/* Initialise the client representation from the parent server */
 	nfs_server_copy_userdata(server, parent_server);
-	server->caps |= NFS_CAP_ATOMIC_OPEN;
+	server->caps |= NFS_CAP_ATOMIC_OPEN|NFS_CAP_CHANGE_ATTR;
 
 	/* Get a client representation.
 	 * Note: NFSv4 always uses TCP, */
@@ -1533,7 +1531,7 @@ static void *nfs_server_list_next(struct seq_file *p, void *v, loff_t *pos);
 static void nfs_server_list_stop(struct seq_file *p, void *v);
 static int nfs_server_list_show(struct seq_file *m, void *v);
 
-static struct seq_operations nfs_server_list_ops = {
+static const struct seq_operations nfs_server_list_ops = {
 	.start	= nfs_server_list_start,
 	.next	= nfs_server_list_next,
 	.stop	= nfs_server_list_stop,
@@ -1554,7 +1552,7 @@ static void *nfs_volume_list_next(struct seq_file *p, void *v, loff_t *pos);
 static void nfs_volume_list_stop(struct seq_file *p, void *v);
 static int nfs_volume_list_show(struct seq_file *m, void *v);
 
-static struct seq_operations nfs_volume_list_ops = {
+static const struct seq_operations nfs_volume_list_ops = {
 	.start	= nfs_volume_list_start,
 	.next	= nfs_volume_list_next,
 	.stop	= nfs_volume_list_stop,
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index e4e089a8f29..6c3210099d5 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -934,9 +934,6 @@ out:
  * back into its cache.  We let the server do generic write
  * parameter checking and report problems.
  *
- * We also avoid an unnecessary invocation of generic_osync_inode(),
- * as it is fairly meaningless to sync the metadata of an NFS file.
- *
  * We eliminate local atime updates, see direct read above.
  *
  * We avoid unnecessary page cache invalidations for normal cached
diff --git a/fs/nfs/dns_resolve.c b/fs/nfs/dns_resolve.c
new file mode 100644
index 00000000000..f4d54ba97cc
--- /dev/null
+++ b/fs/nfs/dns_resolve.c
@@ -0,0 +1,335 @@
+/*
+ * linux/fs/nfs/dns_resolve.c
+ *
+ * Copyright (c) 2009 Trond Myklebust <Trond.Myklebust@netapp.com>
+ *
+ * Resolves DNS hostnames into valid ip addresses
+ */
+
+#include <linux/hash.h>
+#include <linux/string.h>
+#include <linux/kmod.h>
+#include <linux/module.h>
+#include <linux/socket.h>
+#include <linux/seq_file.h>
+#include <linux/inet.h>
+#include <linux/sunrpc/clnt.h>
+#include <linux/sunrpc/cache.h>
+#include <linux/sunrpc/svcauth.h>
+
+#include "dns_resolve.h"
+#include "cache_lib.h"
+
+#define NFS_DNS_HASHBITS 4
+#define NFS_DNS_HASHTBL_SIZE (1 << NFS_DNS_HASHBITS)
+
+static struct cache_head *nfs_dns_table[NFS_DNS_HASHTBL_SIZE];
+
+struct nfs_dns_ent {
+	struct cache_head h;
+
+	char *hostname;
+	size_t namelen;
+
+	struct sockaddr_storage addr;
+	size_t addrlen;
+};
+
+
+static void nfs_dns_ent_init(struct cache_head *cnew,
+		struct cache_head *ckey)
+{
+	struct nfs_dns_ent *new;
+	struct nfs_dns_ent *key;
+
+	new = container_of(cnew, struct nfs_dns_ent, h);
+	key = container_of(ckey, struct nfs_dns_ent, h);
+
+	kfree(new->hostname);
+	new->hostname = kstrndup(key->hostname, key->namelen, GFP_KERNEL);
+	if (new->hostname) {
+		new->namelen = key->namelen;
+		memcpy(&new->addr, &key->addr, key->addrlen);
+		new->addrlen = key->addrlen;
+	} else {
+		new->namelen = 0;
+		new->addrlen = 0;
+	}
+}
+
+static void nfs_dns_ent_put(struct kref *ref)
+{
+	struct nfs_dns_ent *item;
+
+	item = container_of(ref, struct nfs_dns_ent, h.ref);
+	kfree(item->hostname);
+	kfree(item);
+}
+
+static struct cache_head *nfs_dns_ent_alloc(void)
+{
+	struct nfs_dns_ent *item = kmalloc(sizeof(*item), GFP_KERNEL);
+
+	if (item != NULL) {
+		item->hostname = NULL;
+		item->namelen = 0;
+		item->addrlen = 0;
+		return &item->h;
+	}
+	return NULL;
+};
+
+static unsigned int nfs_dns_hash(const struct nfs_dns_ent *key)
+{
+	return hash_str(key->hostname, NFS_DNS_HASHBITS);
+}
+
+static void nfs_dns_request(struct cache_detail *cd,
+		struct cache_head *ch,
+		char **bpp, int *blen)
+{
+	struct nfs_dns_ent *key = container_of(ch, struct nfs_dns_ent, h);
+
+	qword_add(bpp, blen, key->hostname);
+	(*bpp)[-1] = '\n';
+}
+
+static int nfs_dns_upcall(struct cache_detail *cd,
+		struct cache_head *ch)
+{
+	struct nfs_dns_ent *key = container_of(ch, struct nfs_dns_ent, h);
+	int ret;
+
+	ret = nfs_cache_upcall(cd, key->hostname);
+	if (ret)
+		ret = sunrpc_cache_pipe_upcall(cd, ch, nfs_dns_request);
+	return ret;
+}
+
+static int nfs_dns_match(struct cache_head *ca,
+		struct cache_head *cb)
+{
+	struct nfs_dns_ent *a;
+	struct nfs_dns_ent *b;
+
+	a = container_of(ca, struct nfs_dns_ent, h);
+	b = container_of(cb, struct nfs_dns_ent, h);
+
+	if (a->namelen == 0 || a->namelen != b->namelen)
+		return 0;
+	return memcmp(a->hostname, b->hostname, a->namelen) == 0;
+}
+
+static int nfs_dns_show(struct seq_file *m, struct cache_detail *cd,
+		struct cache_head *h)
+{
+	struct nfs_dns_ent *item;
+	long ttl;
+
+	if (h == NULL) {
+		seq_puts(m, "# ip address      hostname        ttl\n");
+		return 0;
+	}
+	item = container_of(h, struct nfs_dns_ent, h);
+	ttl = (long)item->h.expiry_time - (long)get_seconds();
+	if (ttl < 0)
+		ttl = 0;
+
+	if (!test_bit(CACHE_NEGATIVE, &h->flags)) {
+		char buf[INET6_ADDRSTRLEN+IPV6_SCOPE_ID_LEN+1];
+
+		rpc_ntop((struct sockaddr *)&item->addr, buf, sizeof(buf));
+		seq_printf(m, "%15s ", buf);
+	} else
+		seq_puts(m, "<none>          ");
+	seq_printf(m, "%15s %ld\n", item->hostname, ttl);
+	return 0;
+}
+
+struct nfs_dns_ent *nfs_dns_lookup(struct cache_detail *cd,
+		struct nfs_dns_ent *key)
+{
+	struct cache_head *ch;
+
+	ch = sunrpc_cache_lookup(cd,
+			&key->h,
+			nfs_dns_hash(key));
+	if (!ch)
+		return NULL;
+	return container_of(ch, struct nfs_dns_ent, h);
+}
+
+struct nfs_dns_ent *nfs_dns_update(struct cache_detail *cd,
+		struct nfs_dns_ent *new,
+		struct nfs_dns_ent *key)
+{
+	struct cache_head *ch;
+
+	ch = sunrpc_cache_update(cd,
+			&new->h, &key->h,
+			nfs_dns_hash(key));
+	if (!ch)
+		return NULL;
+	return container_of(ch, struct nfs_dns_ent, h);
+}
+
+static int nfs_dns_parse(struct cache_detail *cd, char *buf, int buflen)
+{
+	char buf1[NFS_DNS_HOSTNAME_MAXLEN+1];
+	struct nfs_dns_ent key, *item;
+	unsigned long ttl;
+	ssize_t len;
+	int ret = -EINVAL;
+
+	if (buf[buflen-1] != '\n')
+		goto out;
+	buf[buflen-1] = '\0';
+
+	len = qword_get(&buf, buf1, sizeof(buf1));
+	if (len <= 0)
+		goto out;
+	key.addrlen = rpc_pton(buf1, len,
+			(struct sockaddr *)&key.addr,
+			sizeof(key.addr));
+
+	len = qword_get(&buf, buf1, sizeof(buf1));
+	if (len <= 0)
+		goto out;
+
+	key.hostname = buf1;
+	key.namelen = len;
+	memset(&key.h, 0, sizeof(key.h));
+
+	ttl = get_expiry(&buf);
+	if (ttl == 0)
+		goto out;
+	key.h.expiry_time = ttl + get_seconds();
+
+	ret = -ENOMEM;
+	item = nfs_dns_lookup(cd, &key);
+	if (item == NULL)
+		goto out;
+
+	if (key.addrlen == 0)
+		set_bit(CACHE_NEGATIVE, &key.h.flags);
+
+	item = nfs_dns_update(cd, &key, item);
+	if (item == NULL)
+		goto out;
+
+	ret = 0;
+	cache_put(&item->h, cd);
+out:
+	return ret;
+}
+
+static struct cache_detail nfs_dns_resolve = {
+	.owner = THIS_MODULE,
+	.hash_size = NFS_DNS_HASHTBL_SIZE,
+	.hash_table = nfs_dns_table,
+	.name = "dns_resolve",
+	.cache_put = nfs_dns_ent_put,
+	.cache_upcall = nfs_dns_upcall,
+	.cache_parse = nfs_dns_parse,
+	.cache_show = nfs_dns_show,
+	.match = nfs_dns_match,
+	.init = nfs_dns_ent_init,
+	.update = nfs_dns_ent_init,
+	.alloc = nfs_dns_ent_alloc,
+};
+
+static int do_cache_lookup(struct cache_detail *cd,
+		struct nfs_dns_ent *key,
+		struct nfs_dns_ent **item,
+		struct nfs_cache_defer_req *dreq)
+{
+	int ret = -ENOMEM;
+
+	*item = nfs_dns_lookup(cd, key);
+	if (*item) {
+		ret = cache_check(cd, &(*item)->h, &dreq->req);
+		if (ret)
+			*item = NULL;
+	}
+	return ret;
+}
+
+static int do_cache_lookup_nowait(struct cache_detail *cd,
+		struct nfs_dns_ent *key,
+		struct nfs_dns_ent **item)
+{
+	int ret = -ENOMEM;
+
+	*item = nfs_dns_lookup(cd, key);
+	if (!*item)
+		goto out_err;
+	ret = -ETIMEDOUT;
+	if (!test_bit(CACHE_VALID, &(*item)->h.flags)
+			|| (*item)->h.expiry_time < get_seconds()
+			|| cd->flush_time > (*item)->h.last_refresh)
+		goto out_put;
+	ret = -ENOENT;
+	if (test_bit(CACHE_NEGATIVE, &(*item)->h.flags))
+		goto out_put;
+	return 0;
+out_put:
+	cache_put(&(*item)->h, cd);
+out_err:
+	*item = NULL;
+	return ret;
+}
+
+static int do_cache_lookup_wait(struct cache_detail *cd,
+		struct nfs_dns_ent *key,
+		struct nfs_dns_ent **item)
+{
+	struct nfs_cache_defer_req *dreq;
+	int ret = -ENOMEM;
+
+	dreq = nfs_cache_defer_req_alloc();
+	if (!dreq)
+		goto out;
+	ret = do_cache_lookup(cd, key, item, dreq);
+	if (ret == -EAGAIN) {
+		ret = nfs_cache_wait_for_upcall(dreq);
+		if (!ret)
+			ret = do_cache_lookup_nowait(cd, key, item);
+	}
+	nfs_cache_defer_req_put(dreq);
+out:
+	return ret;
+}
+
+ssize_t nfs_dns_resolve_name(char *name, size_t namelen,
+		struct sockaddr *sa, size_t salen)
+{
+	struct nfs_dns_ent key = {
+		.hostname = name,
+		.namelen = namelen,
+	};
+	struct nfs_dns_ent *item = NULL;
+	ssize_t ret;
+
+	ret = do_cache_lookup_wait(&nfs_dns_resolve, &key, &item);
+	if (ret == 0) {
+		if (salen >= item->addrlen) {
+			memcpy(sa, &item->addr, item->addrlen);
+			ret = item->addrlen;
+		} else
+			ret = -EOVERFLOW;
+		cache_put(&item->h, &nfs_dns_resolve);
+	} else if (ret == -ENOENT)
+		ret = -ESRCH;
+	return ret;
+}
+
+int nfs_dns_resolver_init(void)
+{
+	return nfs_cache_register(&nfs_dns_resolve);
+}
+
+void nfs_dns_resolver_destroy(void)
+{
+	nfs_cache_unregister(&nfs_dns_resolve);
+}
+
diff --git a/fs/nfs/dns_resolve.h b/fs/nfs/dns_resolve.h
new file mode 100644
index 00000000000..a3f0938babf
--- /dev/null
+++ b/fs/nfs/dns_resolve.h
@@ -0,0 +1,14 @@
+/*
+ * Resolve DNS hostnames into valid ip addresses
+ */
+#ifndef __LINUX_FS_NFS_DNS_RESOLVE_H
+#define __LINUX_FS_NFS_DNS_RESOLVE_H
+
+#define NFS_DNS_HOSTNAME_MAXLEN	(128)
+
+extern int nfs_dns_resolver_init(void);
+extern void nfs_dns_resolver_destroy(void);
+extern ssize_t nfs_dns_resolve_name(char *name, size_t namelen,
+		struct sockaddr *sa, size_t salen);
+
+#endif
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 05062329b67..5021b75d2d1 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -328,6 +328,42 @@ nfs_file_fsync(struct file *file, struct dentry *dentry, int datasync)
 }
 
 /*
+ * Decide whether a read/modify/write cycle may be more efficient
+ * then a modify/write/read cycle when writing to a page in the
+ * page cache.
+ *
+ * The modify/write/read cycle may occur if a page is read before
+ * being completely filled by the writer.  In this situation, the
+ * page must be completely written to stable storage on the server
+ * before it can be refilled by reading in the page from the server.
+ * This can lead to expensive, small, FILE_SYNC mode writes being
+ * done.
+ *
+ * It may be more efficient to read the page first if the file is
+ * open for reading in addition to writing, the page is not marked
+ * as Uptodate, it is not dirty or waiting to be committed,
+ * indicating that it was previously allocated and then modified,
+ * that there were valid bytes of data in that range of the file,
+ * and that the new data won't completely replace the old data in
+ * that range of the file.
+ */
+static int nfs_want_read_modify_write(struct file *file, struct page *page,
+			loff_t pos, unsigned len)
+{
+	unsigned int pglen = nfs_page_length(page);
+	unsigned int offset = pos & (PAGE_CACHE_SIZE - 1);
+	unsigned int end = offset + len;
+
+	if ((file->f_mode & FMODE_READ) &&	/* open for read? */
+	    !PageUptodate(page) &&		/* Uptodate? */
+	    !PagePrivate(page) &&		/* i/o request already? */
+	    pglen &&				/* valid bytes of file? */
+	    (end < pglen || offset))		/* replace all valid bytes? */
+		return 1;
+	return 0;
+}
+
+/*
  * This does the "real" work of the write. We must allocate and lock the
  * page to be sent back to the generic routine, which then copies the
  * data from user space.
@@ -340,15 +376,16 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping,
 			struct page **pagep, void **fsdata)
 {
 	int ret;
-	pgoff_t index;
+	pgoff_t index = pos >> PAGE_CACHE_SHIFT;
 	struct page *page;
-	index = pos >> PAGE_CACHE_SHIFT;
+	int once_thru = 0;
 
 	dfprintk(PAGECACHE, "NFS: write_begin(%s/%s(%ld), %u@%lld)\n",
 		file->f_path.dentry->d_parent->d_name.name,
 		file->f_path.dentry->d_name.name,
 		mapping->host->i_ino, len, (long long) pos);
 
+start:
 	/*
 	 * Prevent starvation issues if someone is doing a consistency
 	 * sync-to-disk
@@ -367,6 +404,13 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping,
 	if (ret) {
 		unlock_page(page);
 		page_cache_release(page);
+	} else if (!once_thru &&
+		   nfs_want_read_modify_write(file, page, pos, len)) {
+		once_thru = 1;
+		ret = nfs_readpage(file, page);
+		page_cache_release(page);
+		if (!ret)
+			goto start;
 	}
 	return ret;
 }
@@ -479,6 +523,7 @@ const struct address_space_operations nfs_file_aops = {
 	.invalidatepage = nfs_invalidate_page,
 	.releasepage = nfs_release_page,
 	.direct_IO = nfs_direct_IO,
+	.migratepage = nfs_migrate_page,
 	.launder_page = nfs_launder_page,
 };
 
diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c
index 86147b0ab2c..21a84d45916 100644
--- a/fs/nfs/idmap.c
+++ b/fs/nfs/idmap.c
@@ -101,7 +101,7 @@ static void idmap_pipe_destroy_msg(struct rpc_pipe_msg *);
 
 static unsigned int fnvhash32(const void *, size_t);
 
-static struct rpc_pipe_ops idmap_upcall_ops = {
+static const struct rpc_pipe_ops idmap_upcall_ops = {
 	.upcall		= idmap_pipe_upcall,
 	.downcall	= idmap_pipe_downcall,
 	.destroy_msg	= idmap_pipe_destroy_msg,
@@ -119,8 +119,8 @@ nfs_idmap_new(struct nfs_client *clp)
 	if (idmap == NULL)
 		return -ENOMEM;
 
-	idmap->idmap_dentry = rpc_mkpipe(clp->cl_rpcclient->cl_dentry, "idmap",
-					 idmap, &idmap_upcall_ops, 0);
+	idmap->idmap_dentry = rpc_mkpipe(clp->cl_rpcclient->cl_path.dentry,
+			"idmap", idmap, &idmap_upcall_ops, 0);
 	if (IS_ERR(idmap->idmap_dentry)) {
 		error = PTR_ERR(idmap->idmap_dentry);
 		kfree(idmap);
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index bd7938eda6a..060022b4651 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -46,6 +46,7 @@
 #include "iostat.h"
 #include "internal.h"
 #include "fscache.h"
+#include "dns_resolve.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
 
@@ -286,6 +287,11 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
 		/* We can't support update_atime(), since the server will reset it */
 		inode->i_flags |= S_NOATIME|S_NOCMTIME;
 		inode->i_mode = fattr->mode;
+		if ((fattr->valid & NFS_ATTR_FATTR_MODE) == 0
+				&& nfs_server_capable(inode, NFS_CAP_MODE))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL;
 		/* Why so? Because we want revalidate for devices/FIFOs, and
 		 * that's precisely what we have in nfs_file_inode_operations.
 		 */
@@ -330,20 +336,46 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
 		nfsi->attr_gencount = fattr->gencount;
 		if (fattr->valid & NFS_ATTR_FATTR_ATIME)
 			inode->i_atime = fattr->atime;
+		else if (nfs_server_capable(inode, NFS_CAP_ATIME))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
 		if (fattr->valid & NFS_ATTR_FATTR_MTIME)
 			inode->i_mtime = fattr->mtime;
+		else if (nfs_server_capable(inode, NFS_CAP_MTIME))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_DATA;
 		if (fattr->valid & NFS_ATTR_FATTR_CTIME)
 			inode->i_ctime = fattr->ctime;
+		else if (nfs_server_capable(inode, NFS_CAP_CTIME))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL;
 		if (fattr->valid & NFS_ATTR_FATTR_CHANGE)
 			nfsi->change_attr = fattr->change_attr;
+		else if (nfs_server_capable(inode, NFS_CAP_CHANGE_ATTR))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_DATA;
 		if (fattr->valid & NFS_ATTR_FATTR_SIZE)
 			inode->i_size = nfs_size_to_loff_t(fattr->size);
+		else
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_DATA
+				| NFS_INO_REVAL_PAGECACHE;
 		if (fattr->valid & NFS_ATTR_FATTR_NLINK)
 			inode->i_nlink = fattr->nlink;
+		else if (nfs_server_capable(inode, NFS_CAP_NLINK))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
 		if (fattr->valid & NFS_ATTR_FATTR_OWNER)
 			inode->i_uid = fattr->uid;
+		else if (nfs_server_capable(inode, NFS_CAP_OWNER))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL;
 		if (fattr->valid & NFS_ATTR_FATTR_GROUP)
 			inode->i_gid = fattr->gid;
+		else if (nfs_server_capable(inode, NFS_CAP_OWNER_GROUP))
+			nfsi->cache_validity |= NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL;
 		if (fattr->valid & NFS_ATTR_FATTR_BLOCKS_USED)
 			inode->i_blocks = fattr->du.nfs2.blocks;
 		if (fattr->valid & NFS_ATTR_FATTR_SPACE_USED) {
@@ -1145,6 +1177,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 	loff_t cur_isize, new_isize;
 	unsigned long invalid = 0;
 	unsigned long now = jiffies;
+	unsigned long save_cache_validity;
 
 	dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n",
 			__func__, inode->i_sb->s_id, inode->i_ino,
@@ -1171,10 +1204,11 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 	 */
 	nfsi->read_cache_jiffies = fattr->time_start;
 
-	if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) || (fattr->valid & (NFS_ATTR_FATTR_MTIME|NFS_ATTR_FATTR_CTIME)))
-	    nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR
-		    | NFS_INO_INVALID_ATIME
-		    | NFS_INO_REVAL_PAGECACHE);
+	save_cache_validity = nfsi->cache_validity;
+	nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR
+			| NFS_INO_INVALID_ATIME
+			| NFS_INO_REVAL_FORCED
+			| NFS_INO_REVAL_PAGECACHE);
 
 	/* Do atomic weak cache consistency updates */
 	nfs_wcc_update_inode(inode, fattr);
@@ -1189,7 +1223,8 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 				nfs_force_lookup_revalidate(inode);
 			nfsi->change_attr = fattr->change_attr;
 		}
-	}
+	} else if (server->caps & NFS_CAP_CHANGE_ATTR)
+		invalid |= save_cache_validity;
 
 	if (fattr->valid & NFS_ATTR_FATTR_MTIME) {
 		/* NFSv2/v3: Check if the mtime agrees */
@@ -1201,7 +1236,12 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 				nfs_force_lookup_revalidate(inode);
 			memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
 		}
-	}
+	} else if (server->caps & NFS_CAP_MTIME)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_DATA
+				| NFS_INO_REVAL_PAGECACHE
+				| NFS_INO_REVAL_FORCED);
+
 	if (fattr->valid & NFS_ATTR_FATTR_CTIME) {
 		/* If ctime has changed we should definitely clear access+acl caches */
 		if (!timespec_equal(&inode->i_ctime, &fattr->ctime)) {
@@ -1215,7 +1255,11 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 			}
 			memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
 		}
-	}
+	} else if (server->caps & NFS_CAP_CTIME)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_REVAL_FORCED);
 
 	/* Check if our cached file size is stale */
 	if (fattr->valid & NFS_ATTR_FATTR_SIZE) {
@@ -1231,30 +1275,50 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 			dprintk("NFS: isize change on server for file %s/%ld\n",
 					inode->i_sb->s_id, inode->i_ino);
 		}
-	}
+	} else
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_REVAL_PAGECACHE
+				| NFS_INO_REVAL_FORCED);
 
 
 	if (fattr->valid & NFS_ATTR_FATTR_ATIME)
 		memcpy(&inode->i_atime, &fattr->atime, sizeof(inode->i_atime));
+	else if (server->caps & NFS_CAP_ATIME)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATIME
+				| NFS_INO_REVAL_FORCED);
 
 	if (fattr->valid & NFS_ATTR_FATTR_MODE) {
 		if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)) {
 			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 			inode->i_mode = fattr->mode;
 		}
-	}
+	} else if (server->caps & NFS_CAP_MODE)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_REVAL_FORCED);
+
 	if (fattr->valid & NFS_ATTR_FATTR_OWNER) {
 		if (inode->i_uid != fattr->uid) {
 			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 			inode->i_uid = fattr->uid;
 		}
-	}
+	} else if (server->caps & NFS_CAP_OWNER)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_REVAL_FORCED);
+
 	if (fattr->valid & NFS_ATTR_FATTR_GROUP) {
 		if (inode->i_gid != fattr->gid) {
 			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
 			inode->i_gid = fattr->gid;
 		}
-	}
+	} else if (server->caps & NFS_CAP_OWNER_GROUP)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_REVAL_FORCED);
 
 	if (fattr->valid & NFS_ATTR_FATTR_NLINK) {
 		if (inode->i_nlink != fattr->nlink) {
@@ -1263,7 +1327,9 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 				invalid |= NFS_INO_INVALID_DATA;
 			inode->i_nlink = fattr->nlink;
 		}
-	}
+	} else if (server->caps & NFS_CAP_NLINK)
+		invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR
+				| NFS_INO_REVAL_FORCED);
 
 	if (fattr->valid & NFS_ATTR_FATTR_SPACE_USED) {
 		/*
@@ -1293,9 +1359,8 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 				|| S_ISLNK(inode->i_mode)))
 		invalid &= ~NFS_INO_INVALID_DATA;
 	if (!nfs_have_delegation(inode, FMODE_READ) ||
-			(nfsi->cache_validity & NFS_INO_REVAL_FORCED))
+			(save_cache_validity & NFS_INO_REVAL_FORCED))
 		nfsi->cache_validity |= invalid;
-	nfsi->cache_validity &= ~NFS_INO_REVAL_FORCED;
 
 	return 0;
  out_changed:
@@ -1442,6 +1507,10 @@ static int __init init_nfs_fs(void)
 {
 	int err;
 
+	err = nfs_dns_resolver_init();
+	if (err < 0)
+		goto out8;
+
 	err = nfs_fscache_register();
 	if (err < 0)
 		goto out7;
@@ -1500,6 +1569,8 @@ out5:
 out6:
 	nfs_fscache_unregister();
 out7:
+	nfs_dns_resolver_destroy();
+out8:
 	return err;
 }
 
@@ -1511,6 +1582,7 @@ static void __exit exit_nfs_fs(void)
 	nfs_destroy_inodecache();
 	nfs_destroy_nfspagecache();
 	nfs_fscache_unregister();
+	nfs_dns_resolver_destroy();
 #ifdef CONFIG_PROC_FS
 	rpc_proc_unregister("nfs");
 #endif
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 7dd90a6769d..e21b1bb9972 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -49,6 +49,11 @@ struct nfs_clone_mount {
 #define NFS_MAX_SECFLAVORS	(12)
 
 /*
+ * Value used if the user did not specify a port value.
+ */
+#define NFS_UNSPEC_PORT		(-1)
+
+/*
  * In-kernel mount arguments
  */
 struct nfs_parsed_mount_data {
@@ -63,6 +68,7 @@ struct nfs_parsed_mount_data {
 	unsigned int		auth_flavor_len;
 	rpc_authflavor_t	auth_flavors[1];
 	char			*client_address;
+	unsigned int		version;
 	unsigned int		minorversion;
 	char			*fscache_uniq;
 
@@ -71,7 +77,7 @@ struct nfs_parsed_mount_data {
 		size_t			addrlen;
 		char			*hostname;
 		u32			version;
-		unsigned short		port;
+		int			port;
 		unsigned short		protocol;
 	} mount_server;
 
@@ -80,7 +86,7 @@ struct nfs_parsed_mount_data {
 		size_t			addrlen;
 		char			*hostname;
 		char			*export_path;
-		unsigned short		port;
+		int			port;
 		unsigned short		protocol;
 	} nfs_server;
 
@@ -102,6 +108,7 @@ struct nfs_mount_request {
 };
 
 extern int nfs_mount(struct nfs_mount_request *info);
+extern void nfs_umount(const struct nfs_mount_request *info);
 
 /* client.c */
 extern struct rpc_program nfs_program;
@@ -213,7 +220,6 @@ void nfs_zap_acl_cache(struct inode *inode);
 extern int nfs_wait_bit_killable(void *word);
 
 /* super.c */
-void nfs_parse_ip_address(char *, size_t, struct sockaddr *, size_t *);
 extern struct file_system_type nfs_xdev_fs_type;
 #ifdef CONFIG_NFS_V4
 extern struct file_system_type nfs4_xdev_fs_type;
@@ -248,6 +254,12 @@ extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
 
 /* write.c */
 extern void nfs_write_prepare(struct rpc_task *task, void *calldata);
+#ifdef CONFIG_MIGRATION
+extern int nfs_migrate_page(struct address_space *,
+		struct page *, struct page *);
+#else
+#define nfs_migrate_page NULL
+#endif
 
 /* nfs4proc.c */
 extern int _nfs4_call_sync(struct nfs_server *server,
@@ -368,24 +380,3 @@ unsigned int nfs_page_array_len(unsigned int base, size_t len)
 	return ((unsigned long)len + (unsigned long)base +
 		PAGE_SIZE - 1) >> PAGE_SHIFT;
 }
-
-#define IPV6_SCOPE_DELIMITER	'%'
-
-/*
- * Set the port number in an address.  Be agnostic about the address
- * family.
- */
-static inline void nfs_set_port(struct sockaddr *sap, unsigned short port)
-{
-	struct sockaddr_in *ap = (struct sockaddr_in *)sap;
-	struct sockaddr_in6 *ap6 = (struct sockaddr_in6 *)sap;
-
-	switch (sap->sa_family) {
-	case AF_INET:
-		ap->sin_port = htons(port);
-		break;
-	case AF_INET6:
-		ap6->sin6_port = htons(port);
-		break;
-	}
-}
diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c
index 38ef9eaec40..0adefc40cc8 100644
--- a/fs/nfs/mount_clnt.c
+++ b/fs/nfs/mount_clnt.c
@@ -209,6 +209,71 @@ out_mnt_err:
 	goto out;
 }
 
+/**
+ * nfs_umount - Notify a server that we have unmounted this export
+ * @info: pointer to umount request arguments
+ *
+ * MOUNTPROC_UMNT is advisory, so we set a short timeout, and always
+ * use UDP.
+ */
+void nfs_umount(const struct nfs_mount_request *info)
+{
+	static const struct rpc_timeout nfs_umnt_timeout = {
+		.to_initval = 1 * HZ,
+		.to_maxval = 3 * HZ,
+		.to_retries = 2,
+	};
+	struct rpc_create_args args = {
+		.protocol	= IPPROTO_UDP,
+		.address	= info->sap,
+		.addrsize	= info->salen,
+		.timeout	= &nfs_umnt_timeout,
+		.servername	= info->hostname,
+		.program	= &mnt_program,
+		.version	= info->version,
+		.authflavor	= RPC_AUTH_UNIX,
+		.flags		= RPC_CLNT_CREATE_NOPING,
+	};
+	struct mountres	result;
+	struct rpc_message msg	= {
+		.rpc_argp	= info->dirpath,
+		.rpc_resp	= &result,
+	};
+	struct rpc_clnt *clnt;
+	int status;
+
+	if (info->noresvport)
+		args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
+
+	clnt = rpc_create(&args);
+	if (unlikely(IS_ERR(clnt)))
+		goto out_clnt_err;
+
+	dprintk("NFS: sending UMNT request for %s:%s\n",
+		(info->hostname ? info->hostname : "server"), info->dirpath);
+
+	if (info->version == NFS_MNT3_VERSION)
+		msg.rpc_proc = &clnt->cl_procinfo[MOUNTPROC3_UMNT];
+	else
+		msg.rpc_proc = &clnt->cl_procinfo[MOUNTPROC_UMNT];
+
+	status = rpc_call_sync(clnt, &msg, 0);
+	rpc_shutdown_client(clnt);
+
+	if (unlikely(status < 0))
+		goto out_call_err;
+
+	return;
+
+out_clnt_err:
+	dprintk("NFS: failed to create UMNT RPC client, status=%ld\n",
+			PTR_ERR(clnt));
+	return;
+
+out_call_err:
+	dprintk("NFS: UMNT request failed, status=%d\n", status);
+}
+
 /*
  * XDR encode/decode functions for MOUNT
  */
@@ -258,7 +323,7 @@ static int decode_status(struct xdr_stream *xdr, struct mountres *res)
 		return -EIO;
 	status = ntohl(*p);
 
-	for (i = 0; i <= ARRAY_SIZE(mnt_errtbl); i++) {
+	for (i = 0; i < ARRAY_SIZE(mnt_errtbl); i++) {
 		if (mnt_errtbl[i].status == status) {
 			res->errno = mnt_errtbl[i].errno;
 			return 0;
@@ -309,7 +374,7 @@ static int decode_fhs_status(struct xdr_stream *xdr, struct mountres *res)
 		return -EIO;
 	status = ntohl(*p);
 
-	for (i = 0; i <= ARRAY_SIZE(mnt3_errtbl); i++) {
+	for (i = 0; i < ARRAY_SIZE(mnt3_errtbl); i++) {
 		if (mnt3_errtbl[i].status == status) {
 			res->errno = mnt3_errtbl[i].errno;
 			return 0;
@@ -407,6 +472,13 @@ static struct rpc_procinfo mnt_procedures[] = {
 		.p_statidx	= MOUNTPROC_MNT,
 		.p_name		= "MOUNT",
 	},
+	[MOUNTPROC_UMNT] = {
+		.p_proc		= MOUNTPROC_UMNT,
+		.p_encode	= (kxdrproc_t)mnt_enc_dirpath,
+		.p_arglen	= MNT_enc_dirpath_sz,
+		.p_statidx	= MOUNTPROC_UMNT,
+		.p_name		= "UMOUNT",
+	},
 };
 
 static struct rpc_procinfo mnt3_procedures[] = {
@@ -419,6 +491,13 @@ static struct rpc_procinfo mnt3_procedures[] = {
 		.p_statidx	= MOUNTPROC3_MNT,
 		.p_name		= "MOUNT",
 	},
+	[MOUNTPROC3_UMNT] = {
+		.p_proc		= MOUNTPROC3_UMNT,
+		.p_encode	= (kxdrproc_t)mnt_enc_dirpath,
+		.p_arglen	= MNT_enc_dirpath_sz,
+		.p_statidx	= MOUNTPROC3_UMNT,
+		.p_name		= "UMOUNT",
+	},
 };
 
 
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index d0cc5ce0edf..ee6a13f0544 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -299,7 +299,6 @@ static void nfs3_free_createdata(struct nfs3_createdata *data)
 
 /*
  * Create a regular file.
- * For now, we don't implement O_EXCL.
  */
 static int
 nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c
index 2a2a0a7143a..2636c26d56f 100644
--- a/fs/nfs/nfs4namespace.c
+++ b/fs/nfs/nfs4namespace.c
@@ -17,6 +17,7 @@
 #include <linux/inet.h>
 #include "internal.h"
 #include "nfs4_fs.h"
+#include "dns_resolve.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
 
@@ -95,6 +96,20 @@ static int nfs4_validate_fspath(const struct vfsmount *mnt_parent,
 	return 0;
 }
 
+static size_t nfs_parse_server_name(char *string, size_t len,
+		struct sockaddr *sa, size_t salen)
+{
+	ssize_t ret;
+
+	ret = rpc_pton(string, len, sa, salen);
+	if (ret == 0) {
+		ret = nfs_dns_resolve_name(string, len, sa, salen);
+		if (ret < 0)
+			ret = 0;
+	}
+	return ret;
+}
+
 static struct vfsmount *try_location(struct nfs_clone_mount *mountdata,
 				     char *page, char *page2,
 				     const struct nfs4_fs_location *location)
@@ -121,11 +136,12 @@ static struct vfsmount *try_location(struct nfs_clone_mount *mountdata,
 
 		if (memchr(buf->data, IPV6_SCOPE_DELIMITER, buf->len))
 			continue;
-		nfs_parse_ip_address(buf->data, buf->len,
-				mountdata->addr, &mountdata->addrlen);
-		if (mountdata->addr->sa_family == AF_UNSPEC)
+		mountdata->addrlen = nfs_parse_server_name(buf->data,
+				buf->len,
+				mountdata->addr, mountdata->addrlen);
+		if (mountdata->addrlen == 0)
 			continue;
-		nfs_set_port(mountdata->addr, NFS_PORT);
+		rpc_set_port(mountdata->addr, NFS_PORT);
 
 		memcpy(page2, buf->data, buf->len);
 		page2[buf->len] = '\0';
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 6917311f201..be6544aef41 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -61,6 +61,8 @@
 #define NFS4_POLL_RETRY_MIN	(HZ/10)
 #define NFS4_POLL_RETRY_MAX	(15*HZ)
 
+#define NFS4_MAX_LOOP_ON_RECOVER (10)
+
 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 *);
@@ -426,17 +428,19 @@ out:
 static int nfs4_recover_session(struct nfs4_session *session)
 {
 	struct nfs_client *clp = session->clp;
+	unsigned int loop;
 	int ret;
 
-	for (;;) {
+	for (loop = NFS4_MAX_LOOP_ON_RECOVER; loop != 0; loop--) {
 		ret = nfs4_wait_clnt_recover(clp);
 		if (ret != 0)
-				return ret;
+			break;
 		if (!test_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state))
 			break;
 		nfs4_schedule_state_manager(clp);
+		ret = -EIO;
 	}
-	return 0;
+	return ret;
 }
 
 static int nfs41_setup_sequence(struct nfs4_session *session,
@@ -1444,18 +1448,20 @@ static int _nfs4_proc_open(struct nfs4_opendata *data)
 static int nfs4_recover_expired_lease(struct nfs_server *server)
 {
 	struct nfs_client *clp = server->nfs_client;
+	unsigned int loop;
 	int ret;
 
-	for (;;) {
+	for (loop = NFS4_MAX_LOOP_ON_RECOVER; loop != 0; loop--) {
 		ret = nfs4_wait_clnt_recover(clp);
 		if (ret != 0)
-			return ret;
+			break;
 		if (!test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) &&
 		    !test_bit(NFS4CLNT_CHECK_LEASE,&clp->cl_state))
 			break;
 		nfs4_schedule_state_recovery(clp);
+		ret = -EIO;
 	}
-	return 0;
+	return ret;
 }
 
 /*
@@ -1997,12 +2003,34 @@ static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *f
 	status = nfs4_call_sync(server, &msg, &args, &res, 0);
 	if (status == 0) {
 		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
+		server->caps &= ~(NFS_CAP_ACLS|NFS_CAP_HARDLINKS|
+				NFS_CAP_SYMLINKS|NFS_CAP_FILEID|
+				NFS_CAP_MODE|NFS_CAP_NLINK|NFS_CAP_OWNER|
+				NFS_CAP_OWNER_GROUP|NFS_CAP_ATIME|
+				NFS_CAP_CTIME|NFS_CAP_MTIME);
 		if (res.attr_bitmask[0] & FATTR4_WORD0_ACL)
 			server->caps |= NFS_CAP_ACLS;
 		if (res.has_links != 0)
 			server->caps |= NFS_CAP_HARDLINKS;
 		if (res.has_symlinks != 0)
 			server->caps |= NFS_CAP_SYMLINKS;
+		if (res.attr_bitmask[0] & FATTR4_WORD0_FILEID)
+			server->caps |= NFS_CAP_FILEID;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_MODE)
+			server->caps |= NFS_CAP_MODE;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_NUMLINKS)
+			server->caps |= NFS_CAP_NLINK;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_OWNER)
+			server->caps |= NFS_CAP_OWNER;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_OWNER_GROUP)
+			server->caps |= NFS_CAP_OWNER_GROUP;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_TIME_ACCESS)
+			server->caps |= NFS_CAP_ATIME;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_TIME_METADATA)
+			server->caps |= NFS_CAP_CTIME;
+		if (res.attr_bitmask[1] & FATTR4_WORD1_TIME_MODIFY)
+			server->caps |= NFS_CAP_MTIME;
+
 		memcpy(server->cache_consistency_bitmask, res.attr_bitmask, sizeof(server->cache_consistency_bitmask));
 		server->cache_consistency_bitmask[0] &= FATTR4_WORD0_CHANGE|FATTR4_WORD0_SIZE;
 		server->cache_consistency_bitmask[1] &= FATTR4_WORD1_TIME_METADATA|FATTR4_WORD1_TIME_MODIFY;
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 1434080aefe..2ef4fecf398 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -638,7 +638,7 @@ static void nfs4_fl_release_lock(struct file_lock *fl)
 	nfs4_put_lock_state(fl->fl_u.nfs4_fl.owner);
 }
 
-static struct file_lock_operations nfs4_fl_lock_ops = {
+static const struct file_lock_operations nfs4_fl_lock_ops = {
 	.fl_copy_lock = nfs4_fl_copy_lock,
 	.fl_release_private = nfs4_fl_release_lock,
 };
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 617273e7d47..cfc30d362f9 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -702,29 +702,12 @@ struct compound_hdr {
 	u32		minorversion;
 };
 
-/*
- * START OF "GENERIC" ENCODE ROUTINES.
- *   These may look a little ugly since they are imported from a "generic"
- * set of XDR encode/decode routines which are intended to be shared by
- * all of our NFSv4 implementations (OpenBSD, MacOS X...).
- *
- * If the pain of reading these is too great, it should be a straightforward
- * task to translate them into Linux-specific versions which are more
- * consistent with the style used in NFSv2/v3...
- */
-#define WRITE32(n)               *p++ = htonl(n)
-#define WRITE64(n)               do {				\
-	*p++ = htonl((uint32_t)((n) >> 32));				\
-	*p++ = htonl((uint32_t)(n));					\
-} while (0)
-#define WRITEMEM(ptr,nbytes)     do {				\
-	p = xdr_encode_opaque_fixed(p, ptr, nbytes);		\
-} while (0)
-
-#define RESERVE_SPACE(nbytes)	do {				\
-	p = xdr_reserve_space(xdr, nbytes);			\
-	BUG_ON(!p);						\
-} while (0)
+static __be32 *reserve_space(struct xdr_stream *xdr, size_t nbytes)
+{
+	__be32 *p = xdr_reserve_space(xdr, nbytes);
+	BUG_ON(!p);
+	return p;
+}
 
 static void encode_string(struct xdr_stream *xdr, unsigned int len, const char *str)
 {
@@ -749,12 +732,11 @@ static void encode_compound_hdr(struct xdr_stream *xdr,
 
 	dprintk("encode_compound: tag=%.*s\n", (int)hdr->taglen, hdr->tag);
 	BUG_ON(hdr->taglen > NFS4_MAXTAGLEN);
-	RESERVE_SPACE(12+(XDR_QUADLEN(hdr->taglen)<<2));
-	WRITE32(hdr->taglen);
-	WRITEMEM(hdr->tag, hdr->taglen);
-	WRITE32(hdr->minorversion);
+	p = reserve_space(xdr, 4 + hdr->taglen + 8);
+	p = xdr_encode_opaque(p, hdr->tag, hdr->taglen);
+	*p++ = cpu_to_be32(hdr->minorversion);
 	hdr->nops_p = p;
-	WRITE32(hdr->nops);
+	*p = cpu_to_be32(hdr->nops);
 }
 
 static void encode_nops(struct compound_hdr *hdr)
@@ -829,55 +811,53 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const
 		len += 16;
 	else if (iap->ia_valid & ATTR_MTIME)
 		len += 4;
-	RESERVE_SPACE(len);
+	p = reserve_space(xdr, len);
 
 	/*
 	 * We write the bitmap length now, but leave the bitmap and the attribute
 	 * buffer length to be backfilled at the end of this routine.
 	 */
-	WRITE32(2);
+	*p++ = cpu_to_be32(2);
 	q = p;
 	p += 3;
 
 	if (iap->ia_valid & ATTR_SIZE) {
 		bmval0 |= FATTR4_WORD0_SIZE;
-		WRITE64(iap->ia_size);
+		p = xdr_encode_hyper(p, iap->ia_size);
 	}
 	if (iap->ia_valid & ATTR_MODE) {
 		bmval1 |= FATTR4_WORD1_MODE;
-		WRITE32(iap->ia_mode & S_IALLUGO);
+		*p++ = cpu_to_be32(iap->ia_mode & S_IALLUGO);
 	}
 	if (iap->ia_valid & ATTR_UID) {
 		bmval1 |= FATTR4_WORD1_OWNER;
-		WRITE32(owner_namelen);
-		WRITEMEM(owner_name, owner_namelen);
+		p = xdr_encode_opaque(p, owner_name, owner_namelen);
 	}
 	if (iap->ia_valid & ATTR_GID) {
 		bmval1 |= FATTR4_WORD1_OWNER_GROUP;
-		WRITE32(owner_grouplen);
-		WRITEMEM(owner_group, owner_grouplen);
+		p = xdr_encode_opaque(p, owner_group, owner_grouplen);
 	}
 	if (iap->ia_valid & ATTR_ATIME_SET) {
 		bmval1 |= FATTR4_WORD1_TIME_ACCESS_SET;
-		WRITE32(NFS4_SET_TO_CLIENT_TIME);
-		WRITE32(0);
-		WRITE32(iap->ia_mtime.tv_sec);
-		WRITE32(iap->ia_mtime.tv_nsec);
+		*p++ = cpu_to_be32(NFS4_SET_TO_CLIENT_TIME);
+		*p++ = cpu_to_be32(0);
+		*p++ = cpu_to_be32(iap->ia_mtime.tv_sec);
+		*p++ = cpu_to_be32(iap->ia_mtime.tv_nsec);
 	}
 	else if (iap->ia_valid & ATTR_ATIME) {
 		bmval1 |= FATTR4_WORD1_TIME_ACCESS_SET;
-		WRITE32(NFS4_SET_TO_SERVER_TIME);
+		*p++ = cpu_to_be32(NFS4_SET_TO_SERVER_TIME);
 	}
 	if (iap->ia_valid & ATTR_MTIME_SET) {
 		bmval1 |= FATTR4_WORD1_TIME_MODIFY_SET;
-		WRITE32(NFS4_SET_TO_CLIENT_TIME);
-		WRITE32(0);
-		WRITE32(iap->ia_mtime.tv_sec);
-		WRITE32(iap->ia_mtime.tv_nsec);
+		*p++ = cpu_to_be32(NFS4_SET_TO_CLIENT_TIME);
+		*p++ = cpu_to_be32(0);
+		*p++ = cpu_to_be32(iap->ia_mtime.tv_sec);
+		*p++ = cpu_to_be32(iap->ia_mtime.tv_nsec);
 	}
 	else if (iap->ia_valid & ATTR_MTIME) {
 		bmval1 |= FATTR4_WORD1_TIME_MODIFY_SET;
-		WRITE32(NFS4_SET_TO_SERVER_TIME);
+		*p++ = cpu_to_be32(NFS4_SET_TO_SERVER_TIME);
 	}
 
 	/*
@@ -891,7 +871,7 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const
 	len = (char *)p - (char *)q - 12;
 	*q++ = htonl(bmval0);
 	*q++ = htonl(bmval1);
-	*q++ = htonl(len);
+	*q = htonl(len);
 
 /* out: */
 }
@@ -900,9 +880,9 @@ static void encode_access(struct xdr_stream *xdr, u32 access, struct compound_hd
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8);
-	WRITE32(OP_ACCESS);
-	WRITE32(access);
+	p = reserve_space(xdr, 8);
+	*p++ = cpu_to_be32(OP_ACCESS);
+	*p = cpu_to_be32(access);
 	hdr->nops++;
 	hdr->replen += decode_access_maxsz;
 }
@@ -911,10 +891,10 @@ static void encode_close(struct xdr_stream *xdr, const struct nfs_closeargs *arg
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8+NFS4_STATEID_SIZE);
-	WRITE32(OP_CLOSE);
-	WRITE32(arg->seqid->sequence->counter);
-	WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, 8+NFS4_STATEID_SIZE);
+	*p++ = cpu_to_be32(OP_CLOSE);
+	*p++ = cpu_to_be32(arg->seqid->sequence->counter);
+	xdr_encode_opaque_fixed(p, arg->stateid->data, NFS4_STATEID_SIZE);
 	hdr->nops++;
 	hdr->replen += decode_close_maxsz;
 }
@@ -923,10 +903,10 @@ static void encode_commit(struct xdr_stream *xdr, const struct nfs_writeargs *ar
 {
 	__be32 *p;
 
-	RESERVE_SPACE(16);
-	WRITE32(OP_COMMIT);
-	WRITE64(args->offset);
-	WRITE32(args->count);
+	p = reserve_space(xdr, 16);
+	*p++ = cpu_to_be32(OP_COMMIT);
+	p = xdr_encode_hyper(p, args->offset);
+	*p = cpu_to_be32(args->count);
 	hdr->nops++;
 	hdr->replen += decode_commit_maxsz;
 }
@@ -935,30 +915,28 @@ static void encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8);
-	WRITE32(OP_CREATE);
-	WRITE32(create->ftype);
+	p = reserve_space(xdr, 8);
+	*p++ = cpu_to_be32(OP_CREATE);
+	*p = cpu_to_be32(create->ftype);
 
 	switch (create->ftype) {
 	case NF4LNK:
-		RESERVE_SPACE(4);
-		WRITE32(create->u.symlink.len);
+		p = reserve_space(xdr, 4);
+		*p = cpu_to_be32(create->u.symlink.len);
 		xdr_write_pages(xdr, create->u.symlink.pages, 0, create->u.symlink.len);
 		break;
 
 	case NF4BLK: case NF4CHR:
-		RESERVE_SPACE(8);
-		WRITE32(create->u.device.specdata1);
-		WRITE32(create->u.device.specdata2);
+		p = reserve_space(xdr, 8);
+		*p++ = cpu_to_be32(create->u.device.specdata1);
+		*p = cpu_to_be32(create->u.device.specdata2);
 		break;
 
 	default:
 		break;
 	}
 
-	RESERVE_SPACE(4 + create->name->len);
-	WRITE32(create->name->len);
-	WRITEMEM(create->name->name, create->name->len);
+	encode_string(xdr, create->name->len, create->name->name);
 	hdr->nops++;
 	hdr->replen += decode_create_maxsz;
 
@@ -969,10 +947,10 @@ static void encode_getattr_one(struct xdr_stream *xdr, uint32_t bitmap, struct c
 {
 	__be32 *p;
 
-	RESERVE_SPACE(12);
-	WRITE32(OP_GETATTR);
-	WRITE32(1);
-	WRITE32(bitmap);
+	p = reserve_space(xdr, 12);
+	*p++ = cpu_to_be32(OP_GETATTR);
+	*p++ = cpu_to_be32(1);
+	*p = cpu_to_be32(bitmap);
 	hdr->nops++;
 	hdr->replen += decode_getattr_maxsz;
 }
@@ -981,11 +959,11 @@ static void encode_getattr_two(struct xdr_stream *xdr, uint32_t bm0, uint32_t bm
 {
 	__be32 *p;
 
-	RESERVE_SPACE(16);
-	WRITE32(OP_GETATTR);
-	WRITE32(2);
-	WRITE32(bm0);
-	WRITE32(bm1);
+	p = reserve_space(xdr, 16);
+	*p++ = cpu_to_be32(OP_GETATTR);
+	*p++ = cpu_to_be32(2);
+	*p++ = cpu_to_be32(bm0);
+	*p = cpu_to_be32(bm1);
 	hdr->nops++;
 	hdr->replen += decode_getattr_maxsz;
 }
@@ -1012,8 +990,8 @@ static void encode_getfh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_GETFH);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_GETFH);
 	hdr->nops++;
 	hdr->replen += decode_getfh_maxsz;
 }
@@ -1022,10 +1000,9 @@ static void encode_link(struct xdr_stream *xdr, const struct qstr *name, struct
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8 + name->len);
-	WRITE32(OP_LINK);
-	WRITE32(name->len);
-	WRITEMEM(name->name, name->len);
+	p = reserve_space(xdr, 8 + name->len);
+	*p++ = cpu_to_be32(OP_LINK);
+	xdr_encode_opaque(p, name->name, name->len);
 	hdr->nops++;
 	hdr->replen += decode_link_maxsz;
 }
@@ -1052,27 +1029,27 @@ static void encode_lock(struct xdr_stream *xdr, const struct nfs_lock_args *args
 {
 	__be32 *p;
 
-	RESERVE_SPACE(32);
-	WRITE32(OP_LOCK);
-	WRITE32(nfs4_lock_type(args->fl, args->block));
-	WRITE32(args->reclaim);
-	WRITE64(args->fl->fl_start);
-	WRITE64(nfs4_lock_length(args->fl));
-	WRITE32(args->new_lock_owner);
+	p = reserve_space(xdr, 32);
+	*p++ = cpu_to_be32(OP_LOCK);
+	*p++ = cpu_to_be32(nfs4_lock_type(args->fl, args->block));
+	*p++ = cpu_to_be32(args->reclaim);
+	p = xdr_encode_hyper(p, args->fl->fl_start);
+	p = xdr_encode_hyper(p, nfs4_lock_length(args->fl));
+	*p = cpu_to_be32(args->new_lock_owner);
 	if (args->new_lock_owner){
-		RESERVE_SPACE(4+NFS4_STATEID_SIZE+32);
-		WRITE32(args->open_seqid->sequence->counter);
-		WRITEMEM(args->open_stateid->data, NFS4_STATEID_SIZE);
-		WRITE32(args->lock_seqid->sequence->counter);
-		WRITE64(args->lock_owner.clientid);
-		WRITE32(16);
-		WRITEMEM("lock id:", 8);
-		WRITE64(args->lock_owner.id);
+		p = reserve_space(xdr, 4+NFS4_STATEID_SIZE+32);
+		*p++ = cpu_to_be32(args->open_seqid->sequence->counter);
+		p = xdr_encode_opaque_fixed(p, args->open_stateid->data, NFS4_STATEID_SIZE);
+		*p++ = cpu_to_be32(args->lock_seqid->sequence->counter);
+		p = xdr_encode_hyper(p, args->lock_owner.clientid);
+		*p++ = cpu_to_be32(16);
+		p = xdr_encode_opaque_fixed(p, "lock id:", 8);
+		xdr_encode_hyper(p, args->lock_owner.id);
 	}
 	else {
-		RESERVE_SPACE(NFS4_STATEID_SIZE+4);
-		WRITEMEM(args->lock_stateid->data, NFS4_STATEID_SIZE);
-		WRITE32(args->lock_seqid->sequence->counter);
+		p = reserve_space(xdr, NFS4_STATEID_SIZE+4);
+		p = xdr_encode_opaque_fixed(p, args->lock_stateid->data, NFS4_STATEID_SIZE);
+		*p = cpu_to_be32(args->lock_seqid->sequence->counter);
 	}
 	hdr->nops++;
 	hdr->replen += decode_lock_maxsz;
@@ -1082,15 +1059,15 @@ static void encode_lockt(struct xdr_stream *xdr, const struct nfs_lockt_args *ar
 {
 	__be32 *p;
 
-	RESERVE_SPACE(52);
-	WRITE32(OP_LOCKT);
-	WRITE32(nfs4_lock_type(args->fl, 0));
-	WRITE64(args->fl->fl_start);
-	WRITE64(nfs4_lock_length(args->fl));
-	WRITE64(args->lock_owner.clientid);
-	WRITE32(16);
-	WRITEMEM("lock id:", 8);
-	WRITE64(args->lock_owner.id);
+	p = reserve_space(xdr, 52);
+	*p++ = cpu_to_be32(OP_LOCKT);
+	*p++ = cpu_to_be32(nfs4_lock_type(args->fl, 0));
+	p = xdr_encode_hyper(p, args->fl->fl_start);
+	p = xdr_encode_hyper(p, nfs4_lock_length(args->fl));
+	p = xdr_encode_hyper(p, args->lock_owner.clientid);
+	*p++ = cpu_to_be32(16);
+	p = xdr_encode_opaque_fixed(p, "lock id:", 8);
+	xdr_encode_hyper(p, args->lock_owner.id);
 	hdr->nops++;
 	hdr->replen += decode_lockt_maxsz;
 }
@@ -1099,13 +1076,13 @@ static void encode_locku(struct xdr_stream *xdr, const struct nfs_locku_args *ar
 {
 	__be32 *p;
 
-	RESERVE_SPACE(12+NFS4_STATEID_SIZE+16);
-	WRITE32(OP_LOCKU);
-	WRITE32(nfs4_lock_type(args->fl, 0));
-	WRITE32(args->seqid->sequence->counter);
-	WRITEMEM(args->stateid->data, NFS4_STATEID_SIZE);
-	WRITE64(args->fl->fl_start);
-	WRITE64(nfs4_lock_length(args->fl));
+	p = reserve_space(xdr, 12+NFS4_STATEID_SIZE+16);
+	*p++ = cpu_to_be32(OP_LOCKU);
+	*p++ = cpu_to_be32(nfs4_lock_type(args->fl, 0));
+	*p++ = cpu_to_be32(args->seqid->sequence->counter);
+	p = xdr_encode_opaque_fixed(p, args->stateid->data, NFS4_STATEID_SIZE);
+	p = xdr_encode_hyper(p, args->fl->fl_start);
+	xdr_encode_hyper(p, nfs4_lock_length(args->fl));
 	hdr->nops++;
 	hdr->replen += decode_locku_maxsz;
 }
@@ -1115,10 +1092,9 @@ static void encode_lookup(struct xdr_stream *xdr, const struct qstr *name, struc
 	int len = name->len;
 	__be32 *p;
 
-	RESERVE_SPACE(8 + len);
-	WRITE32(OP_LOOKUP);
-	WRITE32(len);
-	WRITEMEM(name->name, len);
+	p = reserve_space(xdr, 8 + len);
+	*p++ = cpu_to_be32(OP_LOOKUP);
+	xdr_encode_opaque(p, name->name, len);
 	hdr->nops++;
 	hdr->replen += decode_lookup_maxsz;
 }
@@ -1127,21 +1103,21 @@ static void encode_share_access(struct xdr_stream *xdr, fmode_t fmode)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8);
+	p = reserve_space(xdr, 8);
 	switch (fmode & (FMODE_READ|FMODE_WRITE)) {
 	case FMODE_READ:
-		WRITE32(NFS4_SHARE_ACCESS_READ);
+		*p++ = cpu_to_be32(NFS4_SHARE_ACCESS_READ);
 		break;
 	case FMODE_WRITE:
-		WRITE32(NFS4_SHARE_ACCESS_WRITE);
+		*p++ = cpu_to_be32(NFS4_SHARE_ACCESS_WRITE);
 		break;
 	case FMODE_READ|FMODE_WRITE:
-		WRITE32(NFS4_SHARE_ACCESS_BOTH);
+		*p++ = cpu_to_be32(NFS4_SHARE_ACCESS_BOTH);
 		break;
 	default:
-		WRITE32(0);
+		*p++ = cpu_to_be32(0);
 	}
-	WRITE32(0);		/* for linux, share_deny = 0 always */
+	*p = cpu_to_be32(0);		/* for linux, share_deny = 0 always */
 }
 
 static inline void encode_openhdr(struct xdr_stream *xdr, const struct nfs_openargs *arg)
@@ -1151,29 +1127,29 @@ static inline void encode_openhdr(struct xdr_stream *xdr, const struct nfs_opena
  * opcode 4, seqid 4, share_access 4, share_deny 4, clientid 8, ownerlen 4,
  * owner 4 = 32
  */
-	RESERVE_SPACE(8);
-	WRITE32(OP_OPEN);
-	WRITE32(arg->seqid->sequence->counter);
+	p = reserve_space(xdr, 8);
+	*p++ = cpu_to_be32(OP_OPEN);
+	*p = cpu_to_be32(arg->seqid->sequence->counter);
 	encode_share_access(xdr, arg->fmode);
-	RESERVE_SPACE(28);
-	WRITE64(arg->clientid);
-	WRITE32(16);
-	WRITEMEM("open id:", 8);
-	WRITE64(arg->id);
+	p = reserve_space(xdr, 28);
+	p = xdr_encode_hyper(p, arg->clientid);
+	*p++ = cpu_to_be32(16);
+	p = xdr_encode_opaque_fixed(p, "open id:", 8);
+	xdr_encode_hyper(p, arg->id);
 }
 
 static inline void encode_createmode(struct xdr_stream *xdr, const struct nfs_openargs *arg)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
+	p = reserve_space(xdr, 4);
 	switch(arg->open_flags & O_EXCL) {
 	case 0:
-		WRITE32(NFS4_CREATE_UNCHECKED);
+		*p = cpu_to_be32(NFS4_CREATE_UNCHECKED);
 		encode_attrs(xdr, arg->u.attrs, arg->server);
 		break;
 	default:
-		WRITE32(NFS4_CREATE_EXCLUSIVE);
+		*p = cpu_to_be32(NFS4_CREATE_EXCLUSIVE);
 		encode_nfs4_verifier(xdr, &arg->u.verifier);
 	}
 }
@@ -1182,14 +1158,14 @@ static void encode_opentype(struct xdr_stream *xdr, const struct nfs_openargs *a
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
+	p = reserve_space(xdr, 4);
 	switch (arg->open_flags & O_CREAT) {
 	case 0:
-		WRITE32(NFS4_OPEN_NOCREATE);
+		*p = cpu_to_be32(NFS4_OPEN_NOCREATE);
 		break;
 	default:
 		BUG_ON(arg->claim != NFS4_OPEN_CLAIM_NULL);
-		WRITE32(NFS4_OPEN_CREATE);
+		*p = cpu_to_be32(NFS4_OPEN_CREATE);
 		encode_createmode(xdr, arg);
 	}
 }
@@ -1198,16 +1174,16 @@ static inline void encode_delegation_type(struct xdr_stream *xdr, fmode_t delega
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
+	p = reserve_space(xdr, 4);
 	switch (delegation_type) {
 	case 0:
-		WRITE32(NFS4_OPEN_DELEGATE_NONE);
+		*p = cpu_to_be32(NFS4_OPEN_DELEGATE_NONE);
 		break;
 	case FMODE_READ:
-		WRITE32(NFS4_OPEN_DELEGATE_READ);
+		*p = cpu_to_be32(NFS4_OPEN_DELEGATE_READ);
 		break;
 	case FMODE_WRITE|FMODE_READ:
-		WRITE32(NFS4_OPEN_DELEGATE_WRITE);
+		*p = cpu_to_be32(NFS4_OPEN_DELEGATE_WRITE);
 		break;
 	default:
 		BUG();
@@ -1218,8 +1194,8 @@ static inline void encode_claim_null(struct xdr_stream *xdr, const struct qstr *
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(NFS4_OPEN_CLAIM_NULL);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(NFS4_OPEN_CLAIM_NULL);
 	encode_string(xdr, name->len, name->name);
 }
 
@@ -1227,8 +1203,8 @@ static inline void encode_claim_previous(struct xdr_stream *xdr, fmode_t type)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(NFS4_OPEN_CLAIM_PREVIOUS);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(NFS4_OPEN_CLAIM_PREVIOUS);
 	encode_delegation_type(xdr, type);
 }
 
@@ -1236,9 +1212,9 @@ static inline void encode_claim_delegate_cur(struct xdr_stream *xdr, const struc
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4+NFS4_STATEID_SIZE);
-	WRITE32(NFS4_OPEN_CLAIM_DELEGATE_CUR);
-	WRITEMEM(stateid->data, NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, 4+NFS4_STATEID_SIZE);
+	*p++ = cpu_to_be32(NFS4_OPEN_CLAIM_DELEGATE_CUR);
+	xdr_encode_opaque_fixed(p, stateid->data, NFS4_STATEID_SIZE);
 	encode_string(xdr, name->len, name->name);
 }
 
@@ -1267,10 +1243,10 @@ static void encode_open_confirm(struct xdr_stream *xdr, const struct nfs_open_co
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4+NFS4_STATEID_SIZE+4);
-	WRITE32(OP_OPEN_CONFIRM);
-	WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE);
-	WRITE32(arg->seqid->sequence->counter);
+	p = reserve_space(xdr, 4+NFS4_STATEID_SIZE+4);
+	*p++ = cpu_to_be32(OP_OPEN_CONFIRM);
+	p = xdr_encode_opaque_fixed(p, arg->stateid->data, NFS4_STATEID_SIZE);
+	*p = cpu_to_be32(arg->seqid->sequence->counter);
 	hdr->nops++;
 	hdr->replen += decode_open_confirm_maxsz;
 }
@@ -1279,10 +1255,10 @@ static void encode_open_downgrade(struct xdr_stream *xdr, const struct nfs_close
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4+NFS4_STATEID_SIZE+4);
-	WRITE32(OP_OPEN_DOWNGRADE);
-	WRITEMEM(arg->stateid->data, NFS4_STATEID_SIZE);
-	WRITE32(arg->seqid->sequence->counter);
+	p = reserve_space(xdr, 4+NFS4_STATEID_SIZE+4);
+	*p++ = cpu_to_be32(OP_OPEN_DOWNGRADE);
+	p = xdr_encode_opaque_fixed(p, arg->stateid->data, NFS4_STATEID_SIZE);
+	*p = cpu_to_be32(arg->seqid->sequence->counter);
 	encode_share_access(xdr, arg->fmode);
 	hdr->nops++;
 	hdr->replen += decode_open_downgrade_maxsz;
@@ -1294,10 +1270,9 @@ encode_putfh(struct xdr_stream *xdr, const struct nfs_fh *fh, struct compound_hd
 	int len = fh->size;
 	__be32 *p;
 
-	RESERVE_SPACE(8 + len);
-	WRITE32(OP_PUTFH);
-	WRITE32(len);
-	WRITEMEM(fh->data, len);
+	p = reserve_space(xdr, 8 + len);
+	*p++ = cpu_to_be32(OP_PUTFH);
+	xdr_encode_opaque(p, fh->data, len);
 	hdr->nops++;
 	hdr->replen += decode_putfh_maxsz;
 }
@@ -1306,8 +1281,8 @@ static void encode_putrootfh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_PUTROOTFH);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_PUTROOTFH);
 	hdr->nops++;
 	hdr->replen += decode_putrootfh_maxsz;
 }
@@ -1317,26 +1292,26 @@ static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context
 	nfs4_stateid stateid;
 	__be32 *p;
 
-	RESERVE_SPACE(NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, NFS4_STATEID_SIZE);
 	if (ctx->state != NULL) {
 		nfs4_copy_stateid(&stateid, ctx->state, ctx->lockowner);
-		WRITEMEM(stateid.data, NFS4_STATEID_SIZE);
+		xdr_encode_opaque_fixed(p, stateid.data, NFS4_STATEID_SIZE);
 	} else
-		WRITEMEM(zero_stateid.data, NFS4_STATEID_SIZE);
+		xdr_encode_opaque_fixed(p, zero_stateid.data, NFS4_STATEID_SIZE);
 }
 
 static void encode_read(struct xdr_stream *xdr, const struct nfs_readargs *args, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_READ);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_READ);
 
 	encode_stateid(xdr, args->context);
 
-	RESERVE_SPACE(12);
-	WRITE64(args->offset);
-	WRITE32(args->count);
+	p = reserve_space(xdr, 12);
+	p = xdr_encode_hyper(p, args->offset);
+	*p = cpu_to_be32(args->count);
 	hdr->nops++;
 	hdr->replen += decode_read_maxsz;
 }
@@ -1349,20 +1324,20 @@ static void encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg
 	};
 	__be32 *p;
 
-	RESERVE_SPACE(12+NFS4_VERIFIER_SIZE+20);
-	WRITE32(OP_READDIR);
-	WRITE64(readdir->cookie);
-	WRITEMEM(readdir->verifier.data, NFS4_VERIFIER_SIZE);
-	WRITE32(readdir->count >> 1);  /* We're not doing readdirplus */
-	WRITE32(readdir->count);
-	WRITE32(2);
+	p = reserve_space(xdr, 12+NFS4_VERIFIER_SIZE+20);
+	*p++ = cpu_to_be32(OP_READDIR);
+	p = xdr_encode_hyper(p, readdir->cookie);
+	p = xdr_encode_opaque_fixed(p, readdir->verifier.data, NFS4_VERIFIER_SIZE);
+	*p++ = cpu_to_be32(readdir->count >> 1);  /* We're not doing readdirplus */
+	*p++ = cpu_to_be32(readdir->count);
+	*p++ = cpu_to_be32(2);
 	/* Switch to mounted_on_fileid if the server supports it */
 	if (readdir->bitmask[1] & FATTR4_WORD1_MOUNTED_ON_FILEID)
 		attrs[0] &= ~FATTR4_WORD0_FILEID;
 	else
 		attrs[1] &= ~FATTR4_WORD1_MOUNTED_ON_FILEID;
-	WRITE32(attrs[0] & readdir->bitmask[0]);
-	WRITE32(attrs[1] & readdir->bitmask[1]);
+	*p++ = cpu_to_be32(attrs[0] & readdir->bitmask[0]);
+	*p = cpu_to_be32(attrs[1] & readdir->bitmask[1]);
 	hdr->nops++;
 	hdr->replen += decode_readdir_maxsz;
 	dprintk("%s: cookie = %Lu, verifier = %08x:%08x, bitmap = %08x:%08x\n",
@@ -1378,8 +1353,8 @@ static void encode_readlink(struct xdr_stream *xdr, const struct nfs4_readlink *
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_READLINK);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_READLINK);
 	hdr->nops++;
 	hdr->replen += decode_readlink_maxsz;
 }
@@ -1388,10 +1363,9 @@ static void encode_remove(struct xdr_stream *xdr, const struct qstr *name, struc
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8 + name->len);
-	WRITE32(OP_REMOVE);
-	WRITE32(name->len);
-	WRITEMEM(name->name, name->len);
+	p = reserve_space(xdr, 8 + name->len);
+	*p++ = cpu_to_be32(OP_REMOVE);
+	xdr_encode_opaque(p, name->name, name->len);
 	hdr->nops++;
 	hdr->replen += decode_remove_maxsz;
 }
@@ -1400,14 +1374,10 @@ static void encode_rename(struct xdr_stream *xdr, const struct qstr *oldname, co
 {
 	__be32 *p;
 
-	RESERVE_SPACE(8 + oldname->len);
-	WRITE32(OP_RENAME);
-	WRITE32(oldname->len);
-	WRITEMEM(oldname->name, oldname->len);
-
-	RESERVE_SPACE(4 + newname->len);
-	WRITE32(newname->len);
-	WRITEMEM(newname->name, newname->len);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_RENAME);
+	encode_string(xdr, oldname->len, oldname->name);
+	encode_string(xdr, newname->len, newname->name);
 	hdr->nops++;
 	hdr->replen += decode_rename_maxsz;
 }
@@ -1416,9 +1386,9 @@ static void encode_renew(struct xdr_stream *xdr, const struct nfs_client *client
 {
 	__be32 *p;
 
-	RESERVE_SPACE(12);
-	WRITE32(OP_RENEW);
-	WRITE64(client_stateid->cl_clientid);
+	p = reserve_space(xdr, 12);
+	*p++ = cpu_to_be32(OP_RENEW);
+	xdr_encode_hyper(p, client_stateid->cl_clientid);
 	hdr->nops++;
 	hdr->replen += decode_renew_maxsz;
 }
@@ -1428,8 +1398,8 @@ encode_restorefh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_RESTOREFH);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_RESTOREFH);
 	hdr->nops++;
 	hdr->replen += decode_restorefh_maxsz;
 }
@@ -1439,16 +1409,16 @@ encode_setacl(struct xdr_stream *xdr, struct nfs_setaclargs *arg, struct compoun
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4+NFS4_STATEID_SIZE);
-	WRITE32(OP_SETATTR);
-	WRITEMEM(zero_stateid.data, NFS4_STATEID_SIZE);
-	RESERVE_SPACE(2*4);
-	WRITE32(1);
-	WRITE32(FATTR4_WORD0_ACL);
+	p = reserve_space(xdr, 4+NFS4_STATEID_SIZE);
+	*p++ = cpu_to_be32(OP_SETATTR);
+	xdr_encode_opaque_fixed(p, zero_stateid.data, NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, 2*4);
+	*p++ = cpu_to_be32(1);
+	*p = cpu_to_be32(FATTR4_WORD0_ACL);
 	if (arg->acl_len % 4)
 		return -EINVAL;
-	RESERVE_SPACE(4);
-	WRITE32(arg->acl_len);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(arg->acl_len);
 	xdr_write_pages(xdr, arg->acl_pages, arg->acl_pgbase, arg->acl_len);
 	hdr->nops++;
 	hdr->replen += decode_setacl_maxsz;
@@ -1460,8 +1430,8 @@ encode_savefh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_SAVEFH);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_SAVEFH);
 	hdr->nops++;
 	hdr->replen += decode_savefh_maxsz;
 }
@@ -1470,9 +1440,9 @@ static void encode_setattr(struct xdr_stream *xdr, const struct nfs_setattrargs
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4+NFS4_STATEID_SIZE);
-	WRITE32(OP_SETATTR);
-	WRITEMEM(arg->stateid.data, NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, 4+NFS4_STATEID_SIZE);
+	*p++ = cpu_to_be32(OP_SETATTR);
+	xdr_encode_opaque_fixed(p, arg->stateid.data, NFS4_STATEID_SIZE);
 	hdr->nops++;
 	hdr->replen += decode_setattr_maxsz;
 	encode_attrs(xdr, arg->iap, server);
@@ -1482,17 +1452,17 @@ static void encode_setclientid(struct xdr_stream *xdr, const struct nfs4_setclie
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4 + NFS4_VERIFIER_SIZE);
-	WRITE32(OP_SETCLIENTID);
-	WRITEMEM(setclientid->sc_verifier->data, NFS4_VERIFIER_SIZE);
+	p = reserve_space(xdr, 4 + NFS4_VERIFIER_SIZE);
+	*p++ = cpu_to_be32(OP_SETCLIENTID);
+	xdr_encode_opaque_fixed(p, setclientid->sc_verifier->data, NFS4_VERIFIER_SIZE);
 
 	encode_string(xdr, setclientid->sc_name_len, setclientid->sc_name);
-	RESERVE_SPACE(4);
-	WRITE32(setclientid->sc_prog);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(setclientid->sc_prog);
 	encode_string(xdr, setclientid->sc_netid_len, setclientid->sc_netid);
 	encode_string(xdr, setclientid->sc_uaddr_len, setclientid->sc_uaddr);
-	RESERVE_SPACE(4);
-	WRITE32(setclientid->sc_cb_ident);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(setclientid->sc_cb_ident);
 	hdr->nops++;
 	hdr->replen += decode_setclientid_maxsz;
 }
@@ -1501,10 +1471,10 @@ static void encode_setclientid_confirm(struct xdr_stream *xdr, const struct nfs_
 {
 	__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);
+	p = reserve_space(xdr, 12 + NFS4_VERIFIER_SIZE);
+	*p++ = cpu_to_be32(OP_SETCLIENTID_CONFIRM);
+	p = xdr_encode_hyper(p, client_state->cl_clientid);
+	xdr_encode_opaque_fixed(p, client_state->cl_confirm.data, NFS4_VERIFIER_SIZE);
 	hdr->nops++;
 	hdr->replen += decode_setclientid_confirm_maxsz;
 }
@@ -1513,15 +1483,15 @@ static void encode_write(struct xdr_stream *xdr, const struct nfs_writeargs *arg
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_WRITE);
+	p = reserve_space(xdr, 4);
+	*p = cpu_to_be32(OP_WRITE);
 
 	encode_stateid(xdr, args->context);
 
-	RESERVE_SPACE(16);
-	WRITE64(args->offset);
-	WRITE32(args->stable);
-	WRITE32(args->count);
+	p = reserve_space(xdr, 16);
+	p = xdr_encode_hyper(p, args->offset);
+	*p++ = cpu_to_be32(args->stable);
+	*p = cpu_to_be32(args->count);
 
 	xdr_write_pages(xdr, args->pages, args->pgbase, args->count);
 	hdr->nops++;
@@ -1532,10 +1502,10 @@ static void encode_delegreturn(struct xdr_stream *xdr, const nfs4_stateid *state
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4+NFS4_STATEID_SIZE);
+	p = reserve_space(xdr, 4+NFS4_STATEID_SIZE);
 
-	WRITE32(OP_DELEGRETURN);
-	WRITEMEM(stateid->data, NFS4_STATEID_SIZE);
+	*p++ = cpu_to_be32(OP_DELEGRETURN);
+	xdr_encode_opaque_fixed(p, stateid->data, NFS4_STATEID_SIZE);
 	hdr->nops++;
 	hdr->replen += decode_delegreturn_maxsz;
 }
@@ -1548,16 +1518,16 @@ static void encode_exchange_id(struct xdr_stream *xdr,
 {
 	__be32 *p;
 
-	RESERVE_SPACE(4 + sizeof(args->verifier->data));
-	WRITE32(OP_EXCHANGE_ID);
-	WRITEMEM(args->verifier->data, sizeof(args->verifier->data));
+	p = reserve_space(xdr, 4 + sizeof(args->verifier->data));
+	*p++ = cpu_to_be32(OP_EXCHANGE_ID);
+	xdr_encode_opaque_fixed(p, args->verifier->data, sizeof(args->verifier->data));
 
 	encode_string(xdr, args->id_len, args->id);
 
-	RESERVE_SPACE(12);
-	WRITE32(args->flags);
-	WRITE32(0);	/* zero length state_protect4_a */
-	WRITE32(0);	/* zero length implementation id array */
+	p = reserve_space(xdr, 12);
+	*p++ = cpu_to_be32(args->flags);
+	*p++ = cpu_to_be32(0);	/* zero length state_protect4_a */
+	*p = cpu_to_be32(0);	/* zero length implementation id array */
 	hdr->nops++;
 	hdr->replen += decode_exchange_id_maxsz;
 }
@@ -1571,55 +1541,43 @@ static void encode_create_session(struct xdr_stream *xdr,
 	uint32_t len;
 	struct nfs_client *clp = args->client;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_CREATE_SESSION);
-
-	RESERVE_SPACE(8);
-	WRITE64(clp->cl_ex_clid);
+	len = scnprintf(machine_name, sizeof(machine_name), "%s",
+			clp->cl_ipaddr);
 
-	RESERVE_SPACE(8);
-	WRITE32(clp->cl_seqid);			/*Sequence id */
-	WRITE32(args->flags);			/*flags */
+	p = reserve_space(xdr, 20 + 2*28 + 20 + len + 12);
+	*p++ = cpu_to_be32(OP_CREATE_SESSION);
+	p = xdr_encode_hyper(p, clp->cl_ex_clid);
+	*p++ = cpu_to_be32(clp->cl_seqid);			/*Sequence id */
+	*p++ = cpu_to_be32(args->flags);			/*flags */
 
-	RESERVE_SPACE(2*28);			/* 2 channel_attrs */
 	/* Fore Channel */
-	WRITE32(args->fc_attrs.headerpadsz);	/* header padding size */
-	WRITE32(args->fc_attrs.max_rqst_sz);	/* max req size */
-	WRITE32(args->fc_attrs.max_resp_sz);	/* max resp size */
-	WRITE32(args->fc_attrs.max_resp_sz_cached);	/* Max resp sz cached */
-	WRITE32(args->fc_attrs.max_ops);	/* max operations */
-	WRITE32(args->fc_attrs.max_reqs);	/* max requests */
-	WRITE32(0);				/* rdmachannel_attrs */
+	*p++ = cpu_to_be32(args->fc_attrs.headerpadsz);	/* header padding size */
+	*p++ = cpu_to_be32(args->fc_attrs.max_rqst_sz);	/* max req size */
+	*p++ = cpu_to_be32(args->fc_attrs.max_resp_sz);	/* max resp size */
+	*p++ = cpu_to_be32(args->fc_attrs.max_resp_sz_cached);	/* Max resp sz cached */
+	*p++ = cpu_to_be32(args->fc_attrs.max_ops);	/* max operations */
+	*p++ = cpu_to_be32(args->fc_attrs.max_reqs);	/* max requests */
+	*p++ = cpu_to_be32(0);				/* rdmachannel_attrs */
 
 	/* Back Channel */
-	WRITE32(args->fc_attrs.headerpadsz);	/* header padding size */
-	WRITE32(args->bc_attrs.max_rqst_sz);	/* max req size */
-	WRITE32(args->bc_attrs.max_resp_sz);	/* max resp size */
-	WRITE32(args->bc_attrs.max_resp_sz_cached);	/* Max resp sz cached */
-	WRITE32(args->bc_attrs.max_ops);	/* max operations */
-	WRITE32(args->bc_attrs.max_reqs);	/* max requests */
-	WRITE32(0);				/* rdmachannel_attrs */
-
-	RESERVE_SPACE(4);
-	WRITE32(args->cb_program);		/* cb_program */
-
-	RESERVE_SPACE(4);			/* # of security flavors */
-	WRITE32(1);
-
-	RESERVE_SPACE(4);
-	WRITE32(RPC_AUTH_UNIX);			/* auth_sys */
+	*p++ = cpu_to_be32(args->fc_attrs.headerpadsz);	/* header padding size */
+	*p++ = cpu_to_be32(args->bc_attrs.max_rqst_sz);	/* max req size */
+	*p++ = cpu_to_be32(args->bc_attrs.max_resp_sz);	/* max resp size */
+	*p++ = cpu_to_be32(args->bc_attrs.max_resp_sz_cached);	/* Max resp sz cached */
+	*p++ = cpu_to_be32(args->bc_attrs.max_ops);	/* max operations */
+	*p++ = cpu_to_be32(args->bc_attrs.max_reqs);	/* max requests */
+	*p++ = cpu_to_be32(0);				/* rdmachannel_attrs */
+
+	*p++ = cpu_to_be32(args->cb_program);		/* cb_program */
+	*p++ = cpu_to_be32(1);
+	*p++ = cpu_to_be32(RPC_AUTH_UNIX);			/* auth_sys */
 
 	/* authsys_parms rfc1831 */
-	RESERVE_SPACE(4);
-	WRITE32((u32)clp->cl_boot_time.tv_nsec);	/* stamp */
-	len = scnprintf(machine_name, sizeof(machine_name), "%s",
-			clp->cl_ipaddr);
-	RESERVE_SPACE(16 + len);
-	WRITE32(len);
-	WRITEMEM(machine_name, len);
-	WRITE32(0);				/* UID */
-	WRITE32(0);				/* GID */
-	WRITE32(0);				/* No more gids */
+	*p++ = cpu_to_be32((u32)clp->cl_boot_time.tv_nsec);	/* stamp */
+	p = xdr_encode_opaque(p, machine_name, len);
+	*p++ = cpu_to_be32(0);				/* UID */
+	*p++ = cpu_to_be32(0);				/* GID */
+	*p = cpu_to_be32(0);				/* No more gids */
 	hdr->nops++;
 	hdr->replen += decode_create_session_maxsz;
 }
@@ -1629,9 +1587,9 @@ static void encode_destroy_session(struct xdr_stream *xdr,
 				   struct compound_hdr *hdr)
 {
 	__be32 *p;
-	RESERVE_SPACE(4 + NFS4_MAX_SESSIONID_LEN);
-	WRITE32(OP_DESTROY_SESSION);
-	WRITEMEM(session->sess_id.data, NFS4_MAX_SESSIONID_LEN);
+	p = reserve_space(xdr, 4 + NFS4_MAX_SESSIONID_LEN);
+	*p++ = cpu_to_be32(OP_DESTROY_SESSION);
+	xdr_encode_opaque_fixed(p, session->sess_id.data, NFS4_MAX_SESSIONID_LEN);
 	hdr->nops++;
 	hdr->replen += decode_destroy_session_maxsz;
 }
@@ -1655,8 +1613,8 @@ static void encode_sequence(struct xdr_stream *xdr,
 	WARN_ON(args->sa_slotid == NFS4_MAX_SLOT_TABLE);
 	slot = tp->slots + args->sa_slotid;
 
-	RESERVE_SPACE(4);
-	WRITE32(OP_SEQUENCE);
+	p = reserve_space(xdr, 4 + NFS4_MAX_SESSIONID_LEN + 16);
+	*p++ = cpu_to_be32(OP_SEQUENCE);
 
 	/*
 	 * Sessionid + seqid + slotid + max slotid + cache_this
@@ -1670,12 +1628,11 @@ static void encode_sequence(struct xdr_stream *xdr,
 		((u32 *)session->sess_id.data)[3],
 		slot->seq_nr, args->sa_slotid,
 		tp->highest_used_slotid, args->sa_cache_this);
-	RESERVE_SPACE(NFS4_MAX_SESSIONID_LEN + 16);
-	WRITEMEM(session->sess_id.data, NFS4_MAX_SESSIONID_LEN);
-	WRITE32(slot->seq_nr);
-	WRITE32(args->sa_slotid);
-	WRITE32(tp->highest_used_slotid);
-	WRITE32(args->sa_cache_this);
+	p = xdr_encode_opaque_fixed(p, session->sess_id.data, NFS4_MAX_SESSIONID_LEN);
+	*p++ = cpu_to_be32(slot->seq_nr);
+	*p++ = cpu_to_be32(args->sa_slotid);
+	*p++ = cpu_to_be32(tp->highest_used_slotid);
+	*p = cpu_to_be32(args->sa_cache_this);
 	hdr->nops++;
 	hdr->replen += decode_sequence_maxsz;
 #endif /* CONFIG_NFS_V4_1 */
@@ -2466,68 +2423,53 @@ static int nfs4_xdr_enc_get_lease_time(struct rpc_rqst *req, uint32_t *p,
 }
 #endif /* CONFIG_NFS_V4_1 */
 
-/*
- * START OF "GENERIC" DECODE ROUTINES.
- *   These may look a little ugly since they are imported from a "generic"
- * set of XDR encode/decode routines which are intended to be shared by
- * all of our NFSv4 implementations (OpenBSD, MacOS X...).
- *
- * If the pain of reading these is too great, it should be a straightforward
- * task to translate them into Linux-specific versions which are more
- * consistent with the style used in NFSv2/v3...
- */
-#define READ32(x)         (x) = ntohl(*p++)
-#define READ64(x)         do {			\
-	(x) = (u64)ntohl(*p++) << 32;		\
-	(x) |= ntohl(*p++);			\
-} while (0)
-#define READTIME(x)       do {			\
-	p++;					\
-	(x.tv_sec) = ntohl(*p++);		\
-	(x.tv_nsec) = ntohl(*p++);		\
-} while (0)
-#define COPYMEM(x,nbytes) do {			\
-	memcpy((x), p, nbytes);			\
-	p += XDR_QUADLEN(nbytes);		\
-} while (0)
-
-#define READ_BUF(nbytes)  do { \
-	p = xdr_inline_decode(xdr, nbytes); \
-	if (unlikely(!p)) { \
-		dprintk("nfs: %s: prematurely hit end of receive" \
-				" buffer\n", __func__); \
-		dprintk("nfs: %s: xdr->p=%p, bytes=%u, xdr->end=%p\n", \
-				__func__, xdr->p, nbytes, xdr->end); \
-		return -EIO; \
-	} \
-} while (0)
+static void print_overflow_msg(const char *func, const struct xdr_stream *xdr)
+{
+	dprintk("nfs: %s: prematurely hit end of receive buffer. "
+		"Remaining buffer length is %tu words.\n",
+		func, xdr->end - xdr->p);
+}
 
 static int decode_opaque_inline(struct xdr_stream *xdr, unsigned int *len, char **string)
 {
 	__be32 *p;
 
-	READ_BUF(4);
-	READ32(*len);
-	READ_BUF(*len);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	*len = be32_to_cpup(p);
+	p = xdr_inline_decode(xdr, *len);
+	if (unlikely(!p))
+		goto out_overflow;
 	*string = (char *)p;
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_compound_hdr(struct xdr_stream *xdr, struct compound_hdr *hdr)
 {
 	__be32 *p;
 
-	READ_BUF(8);
-	READ32(hdr->status);
-	READ32(hdr->taglen);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	hdr->status = be32_to_cpup(p++);
+	hdr->taglen = be32_to_cpup(p);
 
-	READ_BUF(hdr->taglen + 4);
+	p = xdr_inline_decode(xdr, hdr->taglen + 4);
+	if (unlikely(!p))
+		goto out_overflow;
 	hdr->tag = (char *)p;
 	p += XDR_QUADLEN(hdr->taglen);
-	READ32(hdr->nops);
+	hdr->nops = be32_to_cpup(p);
 	if (unlikely(hdr->nops < 1))
 		return nfs4_stat_to_errno(hdr->status);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected)
@@ -2536,18 +2478,23 @@ static int decode_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected)
 	uint32_t opnum;
 	int32_t nfserr;
 
-	READ_BUF(8);
-	READ32(opnum);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	opnum = be32_to_cpup(p++);
 	if (opnum != expected) {
 		dprintk("nfs: Server returned operation"
 			" %d but we issued a request for %d\n",
 				opnum, expected);
 		return -EIO;
 	}
-	READ32(nfserr);
+	nfserr = be32_to_cpup(p);
 	if (nfserr != NFS_OK)
 		return nfs4_stat_to_errno(nfserr);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 /* Dummy routine */
@@ -2557,8 +2504,11 @@ static int decode_ace(struct xdr_stream *xdr, void *ace, struct nfs_client *clp)
 	unsigned int strlen;
 	char *str;
 
-	READ_BUF(12);
-	return decode_opaque_inline(xdr, &strlen, &str);
+	p = xdr_inline_decode(xdr, 12);
+	if (likely(p))
+		return decode_opaque_inline(xdr, &strlen, &str);
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_bitmap(struct xdr_stream *xdr, uint32_t *bitmap)
@@ -2566,27 +2516,39 @@ static int decode_attr_bitmap(struct xdr_stream *xdr, uint32_t *bitmap)
 	uint32_t bmlen;
 	__be32 *p;
 
-	READ_BUF(4);
-	READ32(bmlen);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	bmlen = be32_to_cpup(p);
 
 	bitmap[0] = bitmap[1] = 0;
-	READ_BUF((bmlen << 2));
+	p = xdr_inline_decode(xdr, (bmlen << 2));
+	if (unlikely(!p))
+		goto out_overflow;
 	if (bmlen > 0) {
-		READ32(bitmap[0]);
+		bitmap[0] = be32_to_cpup(p++);
 		if (bmlen > 1)
-			READ32(bitmap[1]);
+			bitmap[1] = be32_to_cpup(p);
 	}
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static inline int decode_attr_length(struct xdr_stream *xdr, uint32_t *attrlen, __be32 **savep)
 {
 	__be32 *p;
 
-	READ_BUF(4);
-	READ32(*attrlen);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	*attrlen = be32_to_cpup(p);
 	*savep = xdr->p;
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_supported(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *bitmask)
@@ -2609,8 +2571,10 @@ static int decode_attr_type(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_TYPE - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_TYPE)) {
-		READ_BUF(4);
-		READ32(*type);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*type = be32_to_cpup(p);
 		if (*type < NF4REG || *type > NF4NAMEDATTR) {
 			dprintk("%s: bad type %d\n", __func__, *type);
 			return -EIO;
@@ -2620,6 +2584,9 @@ static int decode_attr_type(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *
 	}
 	dprintk("%s: type=0%o\n", __func__, nfs_type2fmt[*type]);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_change(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *change)
@@ -2631,14 +2598,19 @@ static int decode_attr_change(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_CHANGE - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_CHANGE)) {
-		READ_BUF(8);
-		READ64(*change);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, change);
 		bitmap[0] &= ~FATTR4_WORD0_CHANGE;
 		ret = NFS_ATTR_FATTR_CHANGE;
 	}
 	dprintk("%s: change attribute=%Lu\n", __func__,
 			(unsigned long long)*change);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_size(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *size)
@@ -2650,13 +2622,18 @@ static int decode_attr_size(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_SIZE - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_SIZE)) {
-		READ_BUF(8);
-		READ64(*size);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, size);
 		bitmap[0] &= ~FATTR4_WORD0_SIZE;
 		ret = NFS_ATTR_FATTR_SIZE;
 	}
 	dprintk("%s: file size=%Lu\n", __func__, (unsigned long long)*size);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_link_support(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *res)
@@ -2667,12 +2644,17 @@ static int decode_attr_link_support(struct xdr_stream *xdr, uint32_t *bitmap, ui
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_LINK_SUPPORT - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_LINK_SUPPORT)) {
-		READ_BUF(4);
-		READ32(*res);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*res = be32_to_cpup(p);
 		bitmap[0] &= ~FATTR4_WORD0_LINK_SUPPORT;
 	}
 	dprintk("%s: link support=%s\n", __func__, *res == 0 ? "false" : "true");
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_symlink_support(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *res)
@@ -2683,12 +2665,17 @@ static int decode_attr_symlink_support(struct xdr_stream *xdr, uint32_t *bitmap,
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_SYMLINK_SUPPORT - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_SYMLINK_SUPPORT)) {
-		READ_BUF(4);
-		READ32(*res);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*res = be32_to_cpup(p);
 		bitmap[0] &= ~FATTR4_WORD0_SYMLINK_SUPPORT;
 	}
 	dprintk("%s: symlink support=%s\n", __func__, *res == 0 ? "false" : "true");
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_fsid(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs_fsid *fsid)
@@ -2701,9 +2688,11 @@ static int decode_attr_fsid(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_FSID - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_FSID)) {
-		READ_BUF(16);
-		READ64(fsid->major);
-		READ64(fsid->minor);
+		p = xdr_inline_decode(xdr, 16);
+		if (unlikely(!p))
+			goto out_overflow;
+		p = xdr_decode_hyper(p, &fsid->major);
+		xdr_decode_hyper(p, &fsid->minor);
 		bitmap[0] &= ~FATTR4_WORD0_FSID;
 		ret = NFS_ATTR_FATTR_FSID;
 	}
@@ -2711,6 +2700,9 @@ static int decode_attr_fsid(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs
 			(unsigned long long)fsid->major,
 			(unsigned long long)fsid->minor);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_lease_time(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *res)
@@ -2721,12 +2713,17 @@ static int decode_attr_lease_time(struct xdr_stream *xdr, uint32_t *bitmap, uint
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_LEASE_TIME - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_LEASE_TIME)) {
-		READ_BUF(4);
-		READ32(*res);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*res = be32_to_cpup(p);
 		bitmap[0] &= ~FATTR4_WORD0_LEASE_TIME;
 	}
 	dprintk("%s: file size=%u\n", __func__, (unsigned int)*res);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_aclsupport(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *res)
@@ -2737,12 +2734,17 @@ static int decode_attr_aclsupport(struct xdr_stream *xdr, uint32_t *bitmap, uint
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_ACLSUPPORT - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_ACLSUPPORT)) {
-		READ_BUF(4);
-		READ32(*res);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*res = be32_to_cpup(p);
 		bitmap[0] &= ~FATTR4_WORD0_ACLSUPPORT;
 	}
 	dprintk("%s: ACLs supported=%u\n", __func__, (unsigned int)*res);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_fileid(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *fileid)
@@ -2754,13 +2756,18 @@ static int decode_attr_fileid(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_FILEID - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_FILEID)) {
-		READ_BUF(8);
-		READ64(*fileid);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, fileid);
 		bitmap[0] &= ~FATTR4_WORD0_FILEID;
 		ret = NFS_ATTR_FATTR_FILEID;
 	}
 	dprintk("%s: fileid=%Lu\n", __func__, (unsigned long long)*fileid);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_mounted_on_fileid(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *fileid)
@@ -2772,13 +2779,18 @@ static int decode_attr_mounted_on_fileid(struct xdr_stream *xdr, uint32_t *bitma
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_MOUNTED_ON_FILEID - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_MOUNTED_ON_FILEID)) {
-		READ_BUF(8);
-		READ64(*fileid);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, fileid);
 		bitmap[1] &= ~FATTR4_WORD1_MOUNTED_ON_FILEID;
 		ret = NFS_ATTR_FATTR_FILEID;
 	}
 	dprintk("%s: fileid=%Lu\n", __func__, (unsigned long long)*fileid);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_files_avail(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *res)
@@ -2790,12 +2802,17 @@ static int decode_attr_files_avail(struct xdr_stream *xdr, uint32_t *bitmap, uin
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_FILES_AVAIL - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_FILES_AVAIL)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[0] &= ~FATTR4_WORD0_FILES_AVAIL;
 	}
 	dprintk("%s: files avail=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_files_free(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *res)
@@ -2807,12 +2824,17 @@ static int decode_attr_files_free(struct xdr_stream *xdr, uint32_t *bitmap, uint
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_FILES_FREE - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_FILES_FREE)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[0] &= ~FATTR4_WORD0_FILES_FREE;
 	}
 	dprintk("%s: files free=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_files_total(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *res)
@@ -2824,12 +2846,17 @@ static int decode_attr_files_total(struct xdr_stream *xdr, uint32_t *bitmap, uin
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_FILES_TOTAL - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_FILES_TOTAL)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[0] &= ~FATTR4_WORD0_FILES_TOTAL;
 	}
 	dprintk("%s: files total=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_pathname(struct xdr_stream *xdr, struct nfs4_pathname *path)
@@ -2838,8 +2865,10 @@ static int decode_pathname(struct xdr_stream *xdr, struct nfs4_pathname *path)
 	__be32 *p;
 	int status = 0;
 
-	READ_BUF(4);
-	READ32(n);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	n = be32_to_cpup(p);
 	if (n == 0)
 		goto root_path;
 	dprintk("path ");
@@ -2873,6 +2902,9 @@ out_eio:
 	dprintk(" status %d", status);
 	status = -EIO;
 	goto out;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_fs_locations(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs4_fs_locations *res)
@@ -2890,8 +2922,10 @@ static int decode_attr_fs_locations(struct xdr_stream *xdr, uint32_t *bitmap, st
 	status = decode_pathname(xdr, &res->fs_path);
 	if (unlikely(status != 0))
 		goto out;
-	READ_BUF(4);
-	READ32(n);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	n = be32_to_cpup(p);
 	if (n <= 0)
 		goto out_eio;
 	res->nlocations = 0;
@@ -2899,8 +2933,10 @@ static int decode_attr_fs_locations(struct xdr_stream *xdr, uint32_t *bitmap, st
 		u32 m;
 		struct nfs4_fs_location *loc = &res->locations[res->nlocations];
 
-		READ_BUF(4);
-		READ32(m);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		m = be32_to_cpup(p);
 
 		loc->nservers = 0;
 		dprintk("%s: servers ", __func__);
@@ -2939,6 +2975,8 @@ static int decode_attr_fs_locations(struct xdr_stream *xdr, uint32_t *bitmap, st
 out:
 	dprintk("%s: fs_locations done, error = %d\n", __func__, status);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
 out_eio:
 	status = -EIO;
 	goto out;
@@ -2953,12 +2991,17 @@ static int decode_attr_maxfilesize(struct xdr_stream *xdr, uint32_t *bitmap, uin
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_MAXFILESIZE - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_MAXFILESIZE)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[0] &= ~FATTR4_WORD0_MAXFILESIZE;
 	}
 	dprintk("%s: maxfilesize=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_maxlink(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *maxlink)
@@ -2970,12 +3013,17 @@ static int decode_attr_maxlink(struct xdr_stream *xdr, uint32_t *bitmap, uint32_
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_MAXLINK - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_MAXLINK)) {
-		READ_BUF(4);
-		READ32(*maxlink);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*maxlink = be32_to_cpup(p);
 		bitmap[0] &= ~FATTR4_WORD0_MAXLINK;
 	}
 	dprintk("%s: maxlink=%u\n", __func__, *maxlink);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_maxname(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *maxname)
@@ -2987,12 +3035,17 @@ static int decode_attr_maxname(struct xdr_stream *xdr, uint32_t *bitmap, uint32_
 	if (unlikely(bitmap[0] & (FATTR4_WORD0_MAXNAME - 1U)))
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_MAXNAME)) {
-		READ_BUF(4);
-		READ32(*maxname);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*maxname = be32_to_cpup(p);
 		bitmap[0] &= ~FATTR4_WORD0_MAXNAME;
 	}
 	dprintk("%s: maxname=%u\n", __func__, *maxname);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_maxread(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *res)
@@ -3005,8 +3058,10 @@ static int decode_attr_maxread(struct xdr_stream *xdr, uint32_t *bitmap, uint32_
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_MAXREAD)) {
 		uint64_t maxread;
-		READ_BUF(8);
-		READ64(maxread);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, &maxread);
 		if (maxread > 0x7FFFFFFF)
 			maxread = 0x7FFFFFFF;
 		*res = (uint32_t)maxread;
@@ -3014,6 +3069,9 @@ static int decode_attr_maxread(struct xdr_stream *xdr, uint32_t *bitmap, uint32_
 	}
 	dprintk("%s: maxread=%lu\n", __func__, (unsigned long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_maxwrite(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *res)
@@ -3026,8 +3084,10 @@ static int decode_attr_maxwrite(struct xdr_stream *xdr, uint32_t *bitmap, uint32
 		return -EIO;
 	if (likely(bitmap[0] & FATTR4_WORD0_MAXWRITE)) {
 		uint64_t maxwrite;
-		READ_BUF(8);
-		READ64(maxwrite);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, &maxwrite);
 		if (maxwrite > 0x7FFFFFFF)
 			maxwrite = 0x7FFFFFFF;
 		*res = (uint32_t)maxwrite;
@@ -3035,6 +3095,9 @@ static int decode_attr_maxwrite(struct xdr_stream *xdr, uint32_t *bitmap, uint32
 	}
 	dprintk("%s: maxwrite=%lu\n", __func__, (unsigned long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_mode(struct xdr_stream *xdr, uint32_t *bitmap, umode_t *mode)
@@ -3047,14 +3110,19 @@ static int decode_attr_mode(struct xdr_stream *xdr, uint32_t *bitmap, umode_t *m
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_MODE - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_MODE)) {
-		READ_BUF(4);
-		READ32(tmp);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		tmp = be32_to_cpup(p);
 		*mode = tmp & ~S_IFMT;
 		bitmap[1] &= ~FATTR4_WORD1_MODE;
 		ret = NFS_ATTR_FATTR_MODE;
 	}
 	dprintk("%s: file mode=0%o\n", __func__, (unsigned int)*mode);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_nlink(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t *nlink)
@@ -3066,16 +3134,22 @@ static int decode_attr_nlink(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_NUMLINKS - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_NUMLINKS)) {
-		READ_BUF(4);
-		READ32(*nlink);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		*nlink = be32_to_cpup(p);
 		bitmap[1] &= ~FATTR4_WORD1_NUMLINKS;
 		ret = NFS_ATTR_FATTR_NLINK;
 	}
 	dprintk("%s: nlink=%u\n", __func__, (unsigned int)*nlink);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
-static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs_client *clp, uint32_t *uid)
+static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap,
+		struct nfs_client *clp, uint32_t *uid, int may_sleep)
 {
 	uint32_t len;
 	__be32 *p;
@@ -3085,10 +3159,16 @@ static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap, struct nf
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_OWNER - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_OWNER)) {
-		READ_BUF(4);
-		READ32(len);
-		READ_BUF(len);
-		if (len < XDR_MAX_NETOBJ) {
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		len = be32_to_cpup(p);
+		p = xdr_inline_decode(xdr, len);
+		if (unlikely(!p))
+			goto out_overflow;
+		if (!may_sleep) {
+			/* do nothing */
+		} else if (len < XDR_MAX_NETOBJ) {
 			if (nfs_map_name_to_uid(clp, (char *)p, len, uid) == 0)
 				ret = NFS_ATTR_FATTR_OWNER;
 			else
@@ -3101,9 +3181,13 @@ static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap, struct nf
 	}
 	dprintk("%s: uid=%d\n", __func__, (int)*uid);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
-static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs_client *clp, uint32_t *gid)
+static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap,
+		struct nfs_client *clp, uint32_t *gid, int may_sleep)
 {
 	uint32_t len;
 	__be32 *p;
@@ -3113,10 +3197,16 @@ static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap, struct nf
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_OWNER_GROUP - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_OWNER_GROUP)) {
-		READ_BUF(4);
-		READ32(len);
-		READ_BUF(len);
-		if (len < XDR_MAX_NETOBJ) {
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		len = be32_to_cpup(p);
+		p = xdr_inline_decode(xdr, len);
+		if (unlikely(!p))
+			goto out_overflow;
+		if (!may_sleep) {
+			/* do nothing */
+		} else if (len < XDR_MAX_NETOBJ) {
 			if (nfs_map_group_to_gid(clp, (char *)p, len, gid) == 0)
 				ret = NFS_ATTR_FATTR_GROUP;
 			else
@@ -3129,6 +3219,9 @@ static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap, struct nf
 	}
 	dprintk("%s: gid=%d\n", __func__, (int)*gid);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_rdev(struct xdr_stream *xdr, uint32_t *bitmap, dev_t *rdev)
@@ -3143,9 +3236,11 @@ static int decode_attr_rdev(struct xdr_stream *xdr, uint32_t *bitmap, dev_t *rde
 	if (likely(bitmap[1] & FATTR4_WORD1_RAWDEV)) {
 		dev_t tmp;
 
-		READ_BUF(8);
-		READ32(major);
-		READ32(minor);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		major = be32_to_cpup(p++);
+		minor = be32_to_cpup(p);
 		tmp = MKDEV(major, minor);
 		if (MAJOR(tmp) == major && MINOR(tmp) == minor)
 			*rdev = tmp;
@@ -3154,6 +3249,9 @@ static int decode_attr_rdev(struct xdr_stream *xdr, uint32_t *bitmap, dev_t *rde
 	}
 	dprintk("%s: rdev=(0x%x:0x%x)\n", __func__, major, minor);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_space_avail(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *res)
@@ -3165,12 +3263,17 @@ static int decode_attr_space_avail(struct xdr_stream *xdr, uint32_t *bitmap, uin
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_SPACE_AVAIL - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_SPACE_AVAIL)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[1] &= ~FATTR4_WORD1_SPACE_AVAIL;
 	}
 	dprintk("%s: space avail=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_space_free(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *res)
@@ -3182,12 +3285,17 @@ static int decode_attr_space_free(struct xdr_stream *xdr, uint32_t *bitmap, uint
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_SPACE_FREE - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_SPACE_FREE)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[1] &= ~FATTR4_WORD1_SPACE_FREE;
 	}
 	dprintk("%s: space free=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_space_total(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *res)
@@ -3199,12 +3307,17 @@ static int decode_attr_space_total(struct xdr_stream *xdr, uint32_t *bitmap, uin
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_SPACE_TOTAL - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_SPACE_TOTAL)) {
-		READ_BUF(8);
-		READ64(*res);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, res);
 		bitmap[1] &= ~FATTR4_WORD1_SPACE_TOTAL;
 	}
 	dprintk("%s: space total=%Lu\n", __func__, (unsigned long long)*res);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_space_used(struct xdr_stream *xdr, uint32_t *bitmap, uint64_t *used)
@@ -3216,14 +3329,19 @@ static int decode_attr_space_used(struct xdr_stream *xdr, uint32_t *bitmap, uint
 	if (unlikely(bitmap[1] & (FATTR4_WORD1_SPACE_USED - 1U)))
 		return -EIO;
 	if (likely(bitmap[1] & FATTR4_WORD1_SPACE_USED)) {
-		READ_BUF(8);
-		READ64(*used);
+		p = xdr_inline_decode(xdr, 8);
+		if (unlikely(!p))
+			goto out_overflow;
+		xdr_decode_hyper(p, used);
 		bitmap[1] &= ~FATTR4_WORD1_SPACE_USED;
 		ret = NFS_ATTR_FATTR_SPACE_USED;
 	}
 	dprintk("%s: space used=%Lu\n", __func__,
 			(unsigned long long)*used);
 	return ret;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_time(struct xdr_stream *xdr, struct timespec *time)
@@ -3232,12 +3350,17 @@ static int decode_attr_time(struct xdr_stream *xdr, struct timespec *time)
 	uint64_t sec;
 	uint32_t nsec;
 
-	READ_BUF(12);
-	READ64(sec);
-	READ32(nsec);
+	p = xdr_inline_decode(xdr, 12);
+	if (unlikely(!p))
+		goto out_overflow;
+	p = xdr_decode_hyper(p, &sec);
+	nsec = be32_to_cpup(p);
 	time->tv_sec = (time_t)sec;
 	time->tv_nsec = (long)nsec;
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_attr_time_access(struct xdr_stream *xdr, uint32_t *bitmap, struct timespec *time)
@@ -3315,11 +3438,16 @@ static int decode_change_info(struct xdr_stream *xdr, struct nfs4_change_info *c
 {
 	__be32 *p;
 
-	READ_BUF(20);
-	READ32(cinfo->atomic);
-	READ64(cinfo->before);
-	READ64(cinfo->after);
+	p = xdr_inline_decode(xdr, 20);
+	if (unlikely(!p))
+		goto out_overflow;
+	cinfo->atomic = be32_to_cpup(p++);
+	p = xdr_decode_hyper(p, &cinfo->before);
+	xdr_decode_hyper(p, &cinfo->after);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_access(struct xdr_stream *xdr, struct nfs4_accessres *access)
@@ -3331,40 +3459,62 @@ static int decode_access(struct xdr_stream *xdr, struct nfs4_accessres *access)
 	status = decode_op_hdr(xdr, OP_ACCESS);
 	if (status)
 		return status;
-	READ_BUF(8);
-	READ32(supp);
-	READ32(acc);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	supp = be32_to_cpup(p++);
+	acc = be32_to_cpup(p);
 	access->supported = supp;
 	access->access = acc;
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
-static int decode_close(struct xdr_stream *xdr, struct nfs_closeres *res)
+static int decode_opaque_fixed(struct xdr_stream *xdr, void *buf, size_t len)
 {
 	__be32 *p;
+
+	p = xdr_inline_decode(xdr, len);
+	if (likely(p)) {
+		memcpy(buf, p, len);
+		return 0;
+	}
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
+}
+
+static int decode_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid)
+{
+	return decode_opaque_fixed(xdr, stateid->data, NFS4_STATEID_SIZE);
+}
+
+static int decode_close(struct xdr_stream *xdr, struct nfs_closeres *res)
+{
 	int status;
 
 	status = decode_op_hdr(xdr, OP_CLOSE);
 	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)
+		status = decode_stateid(xdr, &res->stateid);
+	return status;
+}
+
+static int decode_verifier(struct xdr_stream *xdr, void *verifier)
+{
+	return decode_opaque_fixed(xdr, verifier, 8);
 }
 
 static int decode_commit(struct xdr_stream *xdr, struct nfs_writeres *res)
 {
-	__be32 *p;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_COMMIT);
-	if (status)
-		return status;
-	READ_BUF(8);
-	COPYMEM(res->verf->verifier, 8);
-	return 0;
+	if (!status)
+		status = decode_verifier(xdr, res->verf->verifier);
+	return status;
 }
 
 static int decode_create(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
@@ -3378,10 +3528,16 @@ static int decode_create(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
 		return status;
 	if ((status = decode_change_info(xdr, cinfo)))
 		return status;
-	READ_BUF(4);
-	READ32(bmlen);
-	READ_BUF(bmlen << 2);
-	return 0;
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	bmlen = be32_to_cpup(p);
+	p = xdr_inline_decode(xdr, bmlen << 2);
+	if (likely(p))
+		return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_server_caps(struct xdr_stream *xdr, struct nfs4_server_caps_res *res)
@@ -3466,7 +3622,8 @@ xdr_error:
 	return status;
 }
 
-static int decode_getfattr(struct xdr_stream *xdr, struct nfs_fattr *fattr, const struct nfs_server *server)
+static int decode_getfattr(struct xdr_stream *xdr, struct nfs_fattr *fattr,
+		const struct nfs_server *server, int may_sleep)
 {
 	__be32 *savep;
 	uint32_t attrlen,
@@ -3538,12 +3695,14 @@ static int decode_getfattr(struct xdr_stream *xdr, struct nfs_fattr *fattr, cons
 		goto xdr_error;
 	fattr->valid |= status;
 
-	status = decode_attr_owner(xdr, bitmap, server->nfs_client, &fattr->uid);
+	status = decode_attr_owner(xdr, bitmap, server->nfs_client,
+			&fattr->uid, may_sleep);
 	if (status < 0)
 		goto xdr_error;
 	fattr->valid |= status;
 
-	status = decode_attr_group(xdr, bitmap, server->nfs_client, &fattr->gid);
+	status = decode_attr_group(xdr, bitmap, server->nfs_client,
+			&fattr->gid, may_sleep);
 	if (status < 0)
 		goto xdr_error;
 	fattr->valid |= status;
@@ -3633,14 +3792,21 @@ static int decode_getfh(struct xdr_stream *xdr, struct nfs_fh *fh)
 	if (status)
 		return status;
 
-	READ_BUF(4);
-	READ32(len);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	len = be32_to_cpup(p);
 	if (len > NFS4_FHSIZE)
 		return -EIO;
 	fh->size = len;
-	READ_BUF(len);
-	COPYMEM(fh->data, len);
+	p = xdr_inline_decode(xdr, len);
+	if (unlikely(!p))
+		goto out_overflow;
+	memcpy(fh->data, p, len);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_link(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
@@ -3662,10 +3828,12 @@ static int decode_lock_denied (struct xdr_stream *xdr, struct file_lock *fl)
 	__be32 *p;
 	uint32_t namelen, type;
 
-	READ_BUF(32);
-	READ64(offset);
-	READ64(length);
-	READ32(type);
+	p = xdr_inline_decode(xdr, 32);
+	if (unlikely(!p))
+		goto out_overflow;
+	p = xdr_decode_hyper(p, &offset);
+	p = xdr_decode_hyper(p, &length);
+	type = be32_to_cpup(p++);
 	if (fl != NULL) {
 		fl->fl_start = (loff_t)offset;
 		fl->fl_end = fl->fl_start + (loff_t)length - 1;
@@ -3676,23 +3844,27 @@ static int decode_lock_denied (struct xdr_stream *xdr, struct file_lock *fl)
 			fl->fl_type = F_RDLCK;
 		fl->fl_pid = 0;
 	}
-	READ64(clientid);
-	READ32(namelen);
-	READ_BUF(namelen);
-	return -NFS4ERR_DENIED;
+	p = xdr_decode_hyper(p, &clientid);
+	namelen = be32_to_cpup(p);
+	p = xdr_inline_decode(xdr, namelen);
+	if (likely(p))
+		return -NFS4ERR_DENIED;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_lock(struct xdr_stream *xdr, struct nfs_lock_res *res)
 {
-	__be32 *p;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_LOCK);
 	if (status == -EIO)
 		goto out;
 	if (status == 0) {
-		READ_BUF(NFS4_STATEID_SIZE);
-		COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
+		status = decode_stateid(xdr, &res->stateid);
+		if (unlikely(status))
+			goto out;
 	} else if (status == -NFS4ERR_DENIED)
 		status = decode_lock_denied(xdr, NULL);
 	if (res->open_seqid != NULL)
@@ -3713,16 +3885,13 @@ static int decode_lockt(struct xdr_stream *xdr, struct nfs_lockt_res *res)
 
 static int decode_locku(struct xdr_stream *xdr, struct nfs_locku_res *res)
 {
-	__be32 *p;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_LOCKU);
 	if (status != -EIO)
 		nfs_increment_lock_seqid(status, res->seqid);
-	if (status == 0) {
-		READ_BUF(NFS4_STATEID_SIZE);
-		COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
-	}
+	if (status == 0)
+		status = decode_stateid(xdr, &res->stateid);
 	return status;
 }
 
@@ -3737,34 +3906,46 @@ static int decode_space_limit(struct xdr_stream *xdr, u64 *maxsize)
 	__be32 *p;
 	uint32_t limit_type, nblocks, blocksize;
 
-	READ_BUF(12);
-	READ32(limit_type);
+	p = xdr_inline_decode(xdr, 12);
+	if (unlikely(!p))
+		goto out_overflow;
+	limit_type = be32_to_cpup(p++);
 	switch (limit_type) {
 	case 1:
-		READ64(*maxsize);
+		xdr_decode_hyper(p, maxsize);
 		break;
 	case 2:
-		READ32(nblocks);
-		READ32(blocksize);
+		nblocks = be32_to_cpup(p++);
+		blocksize = be32_to_cpup(p);
 		*maxsize = (uint64_t)nblocks * (uint64_t)blocksize;
 	}
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_delegation(struct xdr_stream *xdr, struct nfs_openres *res)
 {
 	__be32 *p;
 	uint32_t delegation_type;
+	int status;
 
-	READ_BUF(4);
-	READ32(delegation_type);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	delegation_type = be32_to_cpup(p);
 	if (delegation_type == NFS4_OPEN_DELEGATE_NONE) {
 		res->delegation_type = 0;
 		return 0;
 	}
-	READ_BUF(NFS4_STATEID_SIZE+4);
-	COPYMEM(res->delegation.data, NFS4_STATEID_SIZE);
-	READ32(res->do_recall);
+	status = decode_stateid(xdr, &res->delegation);
+	if (unlikely(status))
+		return status;
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	res->do_recall = be32_to_cpup(p);
 
 	switch (delegation_type) {
 	case NFS4_OPEN_DELEGATE_READ:
@@ -3776,6 +3957,9 @@ static int decode_delegation(struct xdr_stream *xdr, struct nfs_openres *res)
 				return -EIO;
 	}
 	return decode_ace(xdr, NULL, res->server->nfs_client);
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res)
@@ -3787,23 +3971,27 @@ static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res)
 	status = decode_op_hdr(xdr, OP_OPEN);
 	if (status != -EIO)
 		nfs_increment_open_seqid(status, res->seqid);
-	if (status)
+	if (!status)
+		status = decode_stateid(xdr, &res->stateid);
+	if (unlikely(status))
 		return status;
-	READ_BUF(NFS4_STATEID_SIZE);
-	COPYMEM(res->stateid.data, NFS4_STATEID_SIZE);
 
 	decode_change_info(xdr, &res->cinfo);
 
-	READ_BUF(8);
-	READ32(res->rflags);
-	READ32(bmlen);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	res->rflags = be32_to_cpup(p++);
+	bmlen = be32_to_cpup(p);
 	if (bmlen > 10)
 		goto xdr_error;
 
-	READ_BUF(bmlen << 2);
+	p = xdr_inline_decode(xdr, bmlen << 2);
+	if (unlikely(!p))
+		goto out_overflow;
 	savewords = min_t(uint32_t, bmlen, NFS4_BITMAP_SIZE);
 	for (i = 0; i < savewords; ++i)
-		READ32(res->attrset[i]);
+		res->attrset[i] = be32_to_cpup(p++);
 	for (; i < NFS4_BITMAP_SIZE; i++)
 		res->attrset[i] = 0;
 
@@ -3811,36 +3999,33 @@ static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res)
 xdr_error:
 	dprintk("%s: Bitmap too large! Length = %u\n", __func__, bmlen);
 	return -EIO;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_open_confirm(struct xdr_stream *xdr, struct nfs_open_confirmres *res)
 {
-	__be32 *p;
 	int status;
 
 	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)
+		status = decode_stateid(xdr, &res->stateid);
+	return status;
 }
 
 static int decode_open_downgrade(struct xdr_stream *xdr, struct nfs_closeres *res)
 {
-	__be32 *p;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_OPEN_DOWNGRADE);
 	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)
+		status = decode_stateid(xdr, &res->stateid);
+	return status;
 }
 
 static int decode_putfh(struct xdr_stream *xdr)
@@ -3863,9 +4048,11 @@ static int decode_read(struct xdr_stream *xdr, struct rpc_rqst *req, struct nfs_
 	status = decode_op_hdr(xdr, OP_READ);
 	if (status)
 		return status;
-	READ_BUF(8);
-	READ32(eof);
-	READ32(count);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	eof = be32_to_cpup(p++);
+	count = be32_to_cpup(p);
 	hdrlen = (u8 *) p - (u8 *) iov->iov_base;
 	recvd = req->rq_rcv_buf.len - hdrlen;
 	if (count > recvd) {
@@ -3878,6 +4065,9 @@ static int decode_read(struct xdr_stream *xdr, struct rpc_rqst *req, struct nfs_
 	res->eof = eof;
 	res->count = count;
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct nfs4_readdir_res *readdir)
@@ -3892,17 +4082,17 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n
 	int		status;
 
 	status = decode_op_hdr(xdr, OP_READDIR);
-	if (status)
+	if (!status)
+		status = decode_verifier(xdr, readdir->verifier.data);
+	if (unlikely(status))
 		return status;
-	READ_BUF(8);
-	COPYMEM(readdir->verifier.data, 8);
 	dprintk("%s: verifier = %08x:%08x\n",
 			__func__,
 			((u32 *)readdir->verifier.data)[0],
 			((u32 *)readdir->verifier.data)[1]);
 
 
-	hdrlen = (char *) p - (char *) iov->iov_base;
+	hdrlen = (char *) xdr->p - (char *) iov->iov_base;
 	recvd = rcvbuf->len - hdrlen;
 	if (pglen > recvd)
 		pglen = recvd;
@@ -3990,8 +4180,10 @@ static int decode_readlink(struct xdr_stream *xdr, struct rpc_rqst *req)
 		return status;
 
 	/* Convert length of symlink */
-	READ_BUF(4);
-	READ32(len);
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	len = be32_to_cpup(p);
 	if (len >= rcvbuf->page_len || len <= 0) {
 		dprintk("nfs: server returned giant symlink!\n");
 		return -ENAMETOOLONG;
@@ -4015,6 +4207,9 @@ static int decode_readlink(struct xdr_stream *xdr, struct rpc_rqst *req)
 	kaddr[len+rcvbuf->page_base] = '\0';
 	kunmap_atomic(kaddr, KM_USER0);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_remove(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
@@ -4112,10 +4307,16 @@ static int decode_setattr(struct xdr_stream *xdr)
 	status = decode_op_hdr(xdr, OP_SETATTR);
 	if (status)
 		return status;
-	READ_BUF(4);
-	READ32(bmlen);
-	READ_BUF(bmlen << 2);
-	return 0;
+	p = xdr_inline_decode(xdr, 4);
+	if (unlikely(!p))
+		goto out_overflow;
+	bmlen = be32_to_cpup(p);
+	p = xdr_inline_decode(xdr, bmlen << 2);
+	if (likely(p))
+		return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_setclientid(struct xdr_stream *xdr, struct nfs_client *clp)
@@ -4124,35 +4325,50 @@ static int decode_setclientid(struct xdr_stream *xdr, struct nfs_client *clp)
 	uint32_t opnum;
 	int32_t nfserr;
 
-	READ_BUF(8);
-	READ32(opnum);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	opnum = be32_to_cpup(p++);
 	if (opnum != OP_SETCLIENTID) {
 		dprintk("nfs: decode_setclientid: Server returned operation"
 			" %d\n", opnum);
 		return -EIO;
 	}
-	READ32(nfserr);
+	nfserr = be32_to_cpup(p);
 	if (nfserr == NFS_OK) {
-		READ_BUF(8 + NFS4_VERIFIER_SIZE);
-		READ64(clp->cl_clientid);
-		COPYMEM(clp->cl_confirm.data, NFS4_VERIFIER_SIZE);
+		p = xdr_inline_decode(xdr, 8 + NFS4_VERIFIER_SIZE);
+		if (unlikely(!p))
+			goto out_overflow;
+		p = xdr_decode_hyper(p, &clp->cl_clientid);
+		memcpy(clp->cl_confirm.data, p, NFS4_VERIFIER_SIZE);
 	} else if (nfserr == NFSERR_CLID_INUSE) {
 		uint32_t len;
 
 		/* skip netid string */
-		READ_BUF(4);
-		READ32(len);
-		READ_BUF(len);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		len = be32_to_cpup(p);
+		p = xdr_inline_decode(xdr, len);
+		if (unlikely(!p))
+			goto out_overflow;
 
 		/* skip uaddr string */
-		READ_BUF(4);
-		READ32(len);
-		READ_BUF(len);
+		p = xdr_inline_decode(xdr, 4);
+		if (unlikely(!p))
+			goto out_overflow;
+		len = be32_to_cpup(p);
+		p = xdr_inline_decode(xdr, len);
+		if (unlikely(!p))
+			goto out_overflow;
 		return -NFSERR_CLID_INUSE;
 	} else
 		return nfs4_stat_to_errno(nfserr);
 
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_setclientid_confirm(struct xdr_stream *xdr)
@@ -4169,11 +4385,16 @@ static int decode_write(struct xdr_stream *xdr, struct nfs_writeres *res)
 	if (status)
 		return status;
 
-	READ_BUF(16);
-	READ32(res->count);
-	READ32(res->verf->committed);
-	COPYMEM(res->verf->verifier, 8);
+	p = xdr_inline_decode(xdr, 16);
+	if (unlikely(!p))
+		goto out_overflow;
+	res->count = be32_to_cpup(p++);
+	res->verf->committed = be32_to_cpup(p++);
+	memcpy(res->verf->verifier, p, 8);
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_delegreturn(struct xdr_stream *xdr)
@@ -4187,6 +4408,7 @@ static int decode_exchange_id(struct xdr_stream *xdr,
 {
 	__be32 *p;
 	uint32_t dummy;
+	char *dummy_str;
 	int status;
 	struct nfs_client *clp = res->client;
 
@@ -4194,36 +4416,45 @@ static int decode_exchange_id(struct xdr_stream *xdr,
 	if (status)
 		return status;
 
-	READ_BUF(8);
-	READ64(clp->cl_ex_clid);
-	READ_BUF(12);
-	READ32(clp->cl_seqid);
-	READ32(clp->cl_exchange_flags);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	xdr_decode_hyper(p, &clp->cl_ex_clid);
+	p = xdr_inline_decode(xdr, 12);
+	if (unlikely(!p))
+		goto out_overflow;
+	clp->cl_seqid = be32_to_cpup(p++);
+	clp->cl_exchange_flags = be32_to_cpup(p++);
 
 	/* We ask for SP4_NONE */
-	READ32(dummy);
+	dummy = be32_to_cpup(p);
 	if (dummy != SP4_NONE)
 		return -EIO;
 
 	/* Throw away minor_id */
-	READ_BUF(8);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
 
 	/* Throw away Major id */
-	READ_BUF(4);
-	READ32(dummy);
-	READ_BUF(dummy);
+	status = decode_opaque_inline(xdr, &dummy, &dummy_str);
+	if (unlikely(status))
+		return status;
 
 	/* Throw away server_scope */
-	READ_BUF(4);
-	READ32(dummy);
-	READ_BUF(dummy);
+	status = decode_opaque_inline(xdr, &dummy, &dummy_str);
+	if (unlikely(status))
+		return status;
 
 	/* Throw away Implementation id array */
-	READ_BUF(4);
-	READ32(dummy);
-	READ_BUF(dummy);
+	status = decode_opaque_inline(xdr, &dummy, &dummy_str);
+	if (unlikely(status))
+		return status;
 
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_chan_attrs(struct xdr_stream *xdr,
@@ -4232,22 +4463,35 @@ static int decode_chan_attrs(struct xdr_stream *xdr,
 	__be32 *p;
 	u32 nr_attrs;
 
-	READ_BUF(28);
-	READ32(attrs->headerpadsz);
-	READ32(attrs->max_rqst_sz);
-	READ32(attrs->max_resp_sz);
-	READ32(attrs->max_resp_sz_cached);
-	READ32(attrs->max_ops);
-	READ32(attrs->max_reqs);
-	READ32(nr_attrs);
+	p = xdr_inline_decode(xdr, 28);
+	if (unlikely(!p))
+		goto out_overflow;
+	attrs->headerpadsz = be32_to_cpup(p++);
+	attrs->max_rqst_sz = be32_to_cpup(p++);
+	attrs->max_resp_sz = be32_to_cpup(p++);
+	attrs->max_resp_sz_cached = be32_to_cpup(p++);
+	attrs->max_ops = be32_to_cpup(p++);
+	attrs->max_reqs = be32_to_cpup(p++);
+	nr_attrs = be32_to_cpup(p);
 	if (unlikely(nr_attrs > 1)) {
 		printk(KERN_WARNING "%s: Invalid rdma channel attrs count %u\n",
 			__func__, nr_attrs);
 		return -EINVAL;
 	}
-	if (nr_attrs == 1)
-		READ_BUF(4); /* skip rdma_attrs */
+	if (nr_attrs == 1) {
+		p = xdr_inline_decode(xdr, 4); /* skip rdma_attrs */
+		if (unlikely(!p))
+			goto out_overflow;
+	}
 	return 0;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
+}
+
+static int decode_sessionid(struct xdr_stream *xdr, struct nfs4_sessionid *sid)
+{
+	return decode_opaque_fixed(xdr, sid->data, NFS4_MAX_SESSIONID_LEN);
 }
 
 static int decode_create_session(struct xdr_stream *xdr,
@@ -4259,24 +4503,26 @@ static int decode_create_session(struct xdr_stream *xdr,
 	struct nfs4_session *session = clp->cl_session;
 
 	status = decode_op_hdr(xdr, OP_CREATE_SESSION);
-
-	if (status)
+	if (!status)
+		status = decode_sessionid(xdr, &session->sess_id);
+	if (unlikely(status))
 		return status;
 
-	/* sessionid */
-	READ_BUF(NFS4_MAX_SESSIONID_LEN);
-	COPYMEM(&session->sess_id, NFS4_MAX_SESSIONID_LEN);
-
 	/* seqid, flags */
-	READ_BUF(8);
-	READ32(clp->cl_seqid);
-	READ32(session->flags);
+	p = xdr_inline_decode(xdr, 8);
+	if (unlikely(!p))
+		goto out_overflow;
+	clp->cl_seqid = be32_to_cpup(p++);
+	session->flags = be32_to_cpup(p);
 
 	/* Channel attributes */
 	status = decode_chan_attrs(xdr, &session->fc_attrs);
 	if (!status)
 		status = decode_chan_attrs(xdr, &session->bc_attrs);
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	return -EIO;
 }
 
 static int decode_destroy_session(struct xdr_stream *xdr, void *dummy)
@@ -4300,7 +4546,9 @@ static int decode_sequence(struct xdr_stream *xdr,
 		return 0;
 
 	status = decode_op_hdr(xdr, OP_SEQUENCE);
-	if (status)
+	if (!status)
+		status = decode_sessionid(xdr, &id);
+	if (unlikely(status))
 		goto out_err;
 
 	/*
@@ -4309,36 +4557,43 @@ static int decode_sequence(struct xdr_stream *xdr,
 	 */
 	status = -ESERVERFAULT;
 
-	slot = &res->sr_session->fc_slot_table.slots[res->sr_slotid];
-	READ_BUF(NFS4_MAX_SESSIONID_LEN + 20);
-	COPYMEM(id.data, NFS4_MAX_SESSIONID_LEN);
 	if (memcmp(id.data, res->sr_session->sess_id.data,
 		   NFS4_MAX_SESSIONID_LEN)) {
 		dprintk("%s Invalid session id\n", __func__);
 		goto out_err;
 	}
+
+	p = xdr_inline_decode(xdr, 20);
+	if (unlikely(!p))
+		goto out_overflow;
+
 	/* seqid */
-	READ32(dummy);
+	slot = &res->sr_session->fc_slot_table.slots[res->sr_slotid];
+	dummy = be32_to_cpup(p++);
 	if (dummy != slot->seq_nr) {
 		dprintk("%s Invalid sequence number\n", __func__);
 		goto out_err;
 	}
 	/* slot id */
-	READ32(dummy);
+	dummy = be32_to_cpup(p++);
 	if (dummy != res->sr_slotid) {
 		dprintk("%s Invalid slot id\n", __func__);
 		goto out_err;
 	}
 	/* highest slot id - currently not processed */
-	READ32(dummy);
+	dummy = be32_to_cpup(p++);
 	/* target highest slot id - currently not processed */
-	READ32(dummy);
+	dummy = be32_to_cpup(p++);
 	/* result flags - currently not processed */
-	READ32(dummy);
+	dummy = be32_to_cpup(p);
 	status = 0;
 out_err:
 	res->sr_status = status;
 	return status;
+out_overflow:
+	print_overflow_msg(__func__, xdr);
+	status = -EIO;
+	goto out_err;
 #else  /* CONFIG_NFS_V4_1 */
 	return 0;
 #endif /* CONFIG_NFS_V4_1 */
@@ -4370,7 +4625,8 @@ static int nfs4_xdr_dec_open_downgrade(struct rpc_rqst *rqstp, __be32 *p, struct
 	status = decode_open_downgrade(&xdr, res);
 	if (status != 0)
 		goto out;
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4397,7 +4653,8 @@ static int nfs4_xdr_dec_access(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_ac
 	status = decode_access(&xdr, res);
 	if (status != 0)
 		goto out;
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4424,7 +4681,8 @@ static int nfs4_xdr_dec_lookup(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_lo
 		goto out;
 	if ((status = decode_getfh(&xdr, res->fh)) != 0)
 		goto out;
-	status = decode_getfattr(&xdr, res->fattr, res->server);
+	status = decode_getfattr(&xdr, res->fattr, res->server
+			,!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4448,7 +4706,8 @@ static int nfs4_xdr_dec_lookup_root(struct rpc_rqst *rqstp, __be32 *p, struct nf
 	if ((status = decode_putrootfh(&xdr)) != 0)
 		goto out;
 	if ((status = decode_getfh(&xdr, res->fh)) == 0)
-		status = decode_getfattr(&xdr, res->fattr, res->server);
+		status = decode_getfattr(&xdr, res->fattr, res->server,
+				!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4473,7 +4732,8 @@ static int nfs4_xdr_dec_remove(struct rpc_rqst *rqstp, __be32 *p, struct nfs_rem
 		goto out;
 	if ((status = decode_remove(&xdr, &res->cinfo)) != 0)
 		goto out;
-	decode_getfattr(&xdr, &res->dir_attr, res->server);
+	decode_getfattr(&xdr, &res->dir_attr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4503,11 +4763,13 @@ static int nfs4_xdr_dec_rename(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_re
 	if ((status = decode_rename(&xdr, &res->old_cinfo, &res->new_cinfo)) != 0)
 		goto out;
 	/* Current FH is target directory */
-	if (decode_getfattr(&xdr, res->new_fattr, res->server) != 0)
+	if (decode_getfattr(&xdr, res->new_fattr, res->server,
+				!RPC_IS_ASYNC(rqstp->rq_task)) != 0)
 		goto out;
 	if ((status = decode_restorefh(&xdr)) != 0)
 		goto out;
-	decode_getfattr(&xdr, res->old_fattr, res->server);
+	decode_getfattr(&xdr, res->old_fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4540,11 +4802,13 @@ static int nfs4_xdr_dec_link(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_link
 	 * Note order: OP_LINK leaves the directory as the current
 	 *             filehandle.
 	 */
-	if (decode_getfattr(&xdr, res->dir_attr, res->server) != 0)
+	if (decode_getfattr(&xdr, res->dir_attr, res->server,
+				!RPC_IS_ASYNC(rqstp->rq_task)) != 0)
 		goto out;
 	if ((status = decode_restorefh(&xdr)) != 0)
 		goto out;
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4573,11 +4837,13 @@ static int nfs4_xdr_dec_create(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_cr
 		goto out;
 	if ((status = decode_getfh(&xdr, res->fh)) != 0)
 		goto out;
-	if (decode_getfattr(&xdr, res->fattr, res->server) != 0)
+	if (decode_getfattr(&xdr, res->fattr, res->server,
+				!RPC_IS_ASYNC(rqstp->rq_task)) != 0)
 		goto out;
 	if ((status = decode_restorefh(&xdr)) != 0)
 		goto out;
-	decode_getfattr(&xdr, res->dir_fattr, res->server);
+	decode_getfattr(&xdr, res->dir_fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4609,7 +4875,8 @@ static int nfs4_xdr_dec_getattr(struct rpc_rqst *rqstp, __be32 *p, struct nfs4_g
 	status = decode_putfh(&xdr);
 	if (status)
 		goto out;
-	status = decode_getfattr(&xdr, res->fattr, res->server);
+	status = decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4716,7 +4983,8 @@ static int nfs4_xdr_dec_close(struct rpc_rqst *rqstp, __be32 *p, struct nfs_clos
 	 * 	an ESTALE error. Shouldn't be a problem,
 	 * 	though, since fattr->valid will remain unset.
 	 */
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4748,11 +5016,13 @@ static int nfs4_xdr_dec_open(struct rpc_rqst *rqstp, __be32 *p, struct nfs_openr
 		goto out;
 	if (decode_getfh(&xdr, &res->fh) != 0)
 		goto out;
-	if (decode_getfattr(&xdr, res->f_attr, res->server) != 0)
+	if (decode_getfattr(&xdr, res->f_attr, res->server,
+				!RPC_IS_ASYNC(rqstp->rq_task)) != 0)
 		goto out;
 	if (decode_restorefh(&xdr) != 0)
 		goto out;
-	decode_getfattr(&xdr, res->dir_attr, res->server);
+	decode_getfattr(&xdr, res->dir_attr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4800,7 +5070,8 @@ static int nfs4_xdr_dec_open_noattr(struct rpc_rqst *rqstp, __be32 *p, struct nf
 	status = decode_open(&xdr, res);
 	if (status)
 		goto out;
-	decode_getfattr(&xdr, res->f_attr, res->server);
+	decode_getfattr(&xdr, res->f_attr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -4827,7 +5098,8 @@ static int nfs4_xdr_dec_setattr(struct rpc_rqst *rqstp, __be32 *p, struct nfs_se
 	status = decode_setattr(&xdr);
 	if (status)
 		goto out;
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -5001,7 +5273,8 @@ static int nfs4_xdr_dec_write(struct rpc_rqst *rqstp, __be32 *p, struct nfs_writ
 	status = decode_write(&xdr, res);
 	if (status)
 		goto out;
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 	if (!status)
 		status = res->count;
 out:
@@ -5030,7 +5303,8 @@ static int nfs4_xdr_dec_commit(struct rpc_rqst *rqstp, __be32 *p, struct nfs_wri
 	status = decode_commit(&xdr, res);
 	if (status)
 		goto out;
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -5194,7 +5468,8 @@ static int nfs4_xdr_dec_delegreturn(struct rpc_rqst *rqstp, __be32 *p, struct nf
 	if (status != 0)
 		goto out;
 	status = decode_delegreturn(&xdr);
-	decode_getfattr(&xdr, res->fattr, res->server);
+	decode_getfattr(&xdr, res->fattr, res->server,
+			!RPC_IS_ASYNC(rqstp->rq_task));
 out:
 	return status;
 }
@@ -5222,7 +5497,8 @@ static int nfs4_xdr_dec_fs_locations(struct rpc_rqst *req, __be32 *p,
 		goto out;
 	xdr_enter_page(&xdr, PAGE_SIZE);
 	status = decode_getfattr(&xdr, &res->fs_locations->fattr,
-				 res->fs_locations->server);
+				 res->fs_locations->server,
+				 !RPC_IS_ASYNC(req->rq_task));
 out:
 	return status;
 }
diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index 0b4cbdc60ab..f1cc0587cfe 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -73,7 +73,7 @@ enum {
 	Opt_cto, Opt_nocto,
 	Opt_ac, Opt_noac,
 	Opt_lock, Opt_nolock,
-	Opt_v2, Opt_v3,
+	Opt_v2, Opt_v3, Opt_v4,
 	Opt_udp, Opt_tcp, Opt_rdma,
 	Opt_acl, Opt_noacl,
 	Opt_rdirplus, Opt_nordirplus,
@@ -127,6 +127,7 @@ static const match_table_t nfs_mount_option_tokens = {
 	{ Opt_nolock, "nolock" },
 	{ Opt_v2, "v2" },
 	{ Opt_v3, "v3" },
+	{ Opt_v4, "v4" },
 	{ Opt_udp, "udp" },
 	{ Opt_tcp, "tcp" },
 	{ Opt_rdma, "rdma" },
@@ -158,7 +159,7 @@ static const match_table_t nfs_mount_option_tokens = {
 	{ Opt_mountvers, "mountvers=%s" },
 	{ Opt_nfsvers, "nfsvers=%s" },
 	{ Opt_nfsvers, "vers=%s" },
-	{ Opt_minorversion, "minorversion=%u" },
+	{ Opt_minorversion, "minorversion=%s" },
 
 	{ Opt_sec, "sec=%s" },
 	{ Opt_proto, "proto=%s" },
@@ -272,6 +273,10 @@ static const struct super_operations nfs_sops = {
 };
 
 #ifdef CONFIG_NFS_V4
+static int nfs4_validate_text_mount_data(void *options,
+	struct nfs_parsed_mount_data *args, const char *dev_name);
+static int nfs4_try_mount(int flags, const char *dev_name,
+	struct nfs_parsed_mount_data *data, struct vfsmount *mnt);
 static int nfs4_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt);
 static int nfs4_remote_get_sb(struct file_system_type *fs_type,
@@ -742,127 +747,23 @@ static int nfs_verify_server_address(struct sockaddr *addr)
 	}
 	}
 
+	dfprintk(MOUNT, "NFS: Invalid IP address specified\n");
 	return 0;
 }
 
-static void nfs_parse_ipv4_address(char *string, size_t str_len,
-				   struct sockaddr *sap, size_t *addr_len)
-{
-	struct sockaddr_in *sin = (struct sockaddr_in *)sap;
-	u8 *addr = (u8 *)&sin->sin_addr.s_addr;
-
-	if (str_len <= INET_ADDRSTRLEN) {
-		dfprintk(MOUNT, "NFS: parsing IPv4 address %*s\n",
-				(int)str_len, string);
-
-		sin->sin_family = AF_INET;
-		*addr_len = sizeof(*sin);
-		if (in4_pton(string, str_len, addr, '\0', NULL))
-			return;
-	}
-
-	sap->sa_family = AF_UNSPEC;
-	*addr_len = 0;
-}
-
-#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
-static int nfs_parse_ipv6_scope_id(const char *string, const size_t str_len,
-				   const char *delim,
-				   struct sockaddr_in6 *sin6)
-{
-	char *p;
-	size_t len;
-
-	if ((string + str_len) == delim)
-		return 1;
-
-	if (*delim != IPV6_SCOPE_DELIMITER)
-		return 0;
-
-	if (!(ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_LINKLOCAL))
-		return 0;
-
-	len = (string + str_len) - delim - 1;
-	p = kstrndup(delim + 1, len, GFP_KERNEL);
-	if (p) {
-		unsigned long scope_id = 0;
-		struct net_device *dev;
-
-		dev = dev_get_by_name(&init_net, p);
-		if (dev != NULL) {
-			scope_id = dev->ifindex;
-			dev_put(dev);
-		} else {
-			if (strict_strtoul(p, 10, &scope_id) == 0) {
-				kfree(p);
-				return 0;
-			}
-		}
-
-		kfree(p);
-
-		sin6->sin6_scope_id = scope_id;
-		dfprintk(MOUNT, "NFS: IPv6 scope ID = %lu\n", scope_id);
-		return 1;
-	}
-
-	return 0;
-}
-
-static void nfs_parse_ipv6_address(char *string, size_t str_len,
-				   struct sockaddr *sap, size_t *addr_len)
-{
-	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap;
-	u8 *addr = (u8 *)&sin6->sin6_addr.in6_u;
-	const char *delim;
-
-	if (str_len <= INET6_ADDRSTRLEN) {
-		dfprintk(MOUNT, "NFS: parsing IPv6 address %*s\n",
-				(int)str_len, string);
-
-		sin6->sin6_family = AF_INET6;
-		*addr_len = sizeof(*sin6);
-		if (in6_pton(string, str_len, addr,
-					IPV6_SCOPE_DELIMITER, &delim) != 0) {
-			if (nfs_parse_ipv6_scope_id(string, str_len,
-							delim, sin6) != 0)
-				return;
-		}
-	}
-
-	sap->sa_family = AF_UNSPEC;
-	*addr_len = 0;
-}
-#else
-static void nfs_parse_ipv6_address(char *string, size_t str_len,
-				   struct sockaddr *sap, size_t *addr_len)
-{
-	sap->sa_family = AF_UNSPEC;
-	*addr_len = 0;
-}
-#endif
-
 /*
- * Construct a sockaddr based on the contents of a string that contains
- * an IP address in presentation format.
- *
- * If there is a problem constructing the new sockaddr, set the address
- * family to AF_UNSPEC.
+ * Select between a default port value and a user-specified port value.
+ * If a zero value is set, then autobind will be used.
  */
-void nfs_parse_ip_address(char *string, size_t str_len,
-				 struct sockaddr *sap, size_t *addr_len)
+static void nfs_set_default_port(struct sockaddr *sap, const int parsed_port,
+				 const unsigned short default_port)
 {
-	unsigned int i, colons;
+	unsigned short port = default_port;
 
-	colons = 0;
-	for (i = 0; i < str_len; i++)
-		if (string[i] == ':')
-			colons++;
+	if (parsed_port != NFS_UNSPEC_PORT)
+		port = parsed_port;
 
-	if (colons >= 2)
-		nfs_parse_ipv6_address(string, str_len, sap, addr_len);
-	else
-		nfs_parse_ipv4_address(string, str_len, sap, addr_len);
+	rpc_set_port(sap, port);
 }
 
 /*
@@ -904,8 +805,6 @@ static void nfs_set_mount_transport_protocol(struct nfs_parsed_mount_data *mnt)
 
 /*
  * Parse the value of the 'sec=' option.
- *
- * The flavor_len setting is for v4 mounts.
  */
 static int nfs_parse_security_flavors(char *value,
 				      struct nfs_parsed_mount_data *mnt)
@@ -916,53 +815,43 @@ static int nfs_parse_security_flavors(char *value,
 
 	switch (match_token(value, nfs_secflavor_tokens, args)) {
 	case Opt_sec_none:
-		mnt->auth_flavor_len = 0;
 		mnt->auth_flavors[0] = RPC_AUTH_NULL;
 		break;
 	case Opt_sec_sys:
-		mnt->auth_flavor_len = 0;
 		mnt->auth_flavors[0] = RPC_AUTH_UNIX;
 		break;
 	case Opt_sec_krb5:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_KRB5;
 		break;
 	case Opt_sec_krb5i:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_KRB5I;
 		break;
 	case Opt_sec_krb5p:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_KRB5P;
 		break;
 	case Opt_sec_lkey:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_LKEY;
 		break;
 	case Opt_sec_lkeyi:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_LKEYI;
 		break;
 	case Opt_sec_lkeyp:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_LKEYP;
 		break;
 	case Opt_sec_spkm:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_SPKM;
 		break;
 	case Opt_sec_spkmi:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_SPKMI;
 		break;
 	case Opt_sec_spkmp:
-		mnt->auth_flavor_len = 1;
 		mnt->auth_flavors[0] = RPC_AUTH_GSS_SPKMP;
 		break;
 	default:
 		return 0;
 	}
 
+	mnt->auth_flavor_len = 1;
 	return 1;
 }
 
@@ -1001,7 +890,6 @@ static int nfs_parse_mount_options(char *raw,
 	while ((p = strsep(&raw, ",")) != NULL) {
 		substring_t args[MAX_OPT_ARGS];
 		unsigned long option;
-		int int_option;
 		int token;
 
 		if (!*p)
@@ -1047,10 +935,18 @@ static int nfs_parse_mount_options(char *raw,
 			break;
 		case Opt_v2:
 			mnt->flags &= ~NFS_MOUNT_VER3;
+			mnt->version = 2;
 			break;
 		case Opt_v3:
 			mnt->flags |= NFS_MOUNT_VER3;
+			mnt->version = 3;
 			break;
+#ifdef CONFIG_NFS_V4
+		case Opt_v4:
+			mnt->flags &= ~NFS_MOUNT_VER3;
+			mnt->version = 4;
+			break;
+#endif
 		case Opt_udp:
 			mnt->flags &= ~NFS_MOUNT_TCP;
 			mnt->nfs_server.protocol = XPRT_TRANSPORT_UDP;
@@ -1264,20 +1160,33 @@ static int nfs_parse_mount_options(char *raw,
 			switch (option) {
 			case NFS2_VERSION:
 				mnt->flags &= ~NFS_MOUNT_VER3;
+				mnt->version = 2;
 				break;
 			case NFS3_VERSION:
 				mnt->flags |= NFS_MOUNT_VER3;
+				mnt->version = 3;
 				break;
+#ifdef CONFIG_NFS_V4
+			case NFS4_VERSION:
+				mnt->flags &= ~NFS_MOUNT_VER3;
+				mnt->version = 4;
+				break;
+#endif
 			default:
 				goto out_invalid_value;
 			}
 			break;
 		case Opt_minorversion:
-			if (match_int(args, &int_option))
-				return 0;
-			if (int_option < 0 || int_option > NFS4_MAX_MINOR_VERSION)
-				return 0;
-			mnt->minorversion = int_option;
+			string = match_strdup(args);
+			if (string == NULL)
+				goto out_nomem;
+			rc = strict_strtoul(string, 10, &option);
+			kfree(string);
+			if (rc != 0)
+				goto out_invalid_value;
+			if (option > NFS4_MAX_MINOR_VERSION)
+				goto out_invalid_value;
+			mnt->minorversion = option;
 			break;
 
 		/*
@@ -1352,11 +1261,14 @@ static int nfs_parse_mount_options(char *raw,
 			string = match_strdup(args);
 			if (string == NULL)
 				goto out_nomem;
-			nfs_parse_ip_address(string, strlen(string),
-					     (struct sockaddr *)
-						&mnt->nfs_server.address,
-					     &mnt->nfs_server.addrlen);
+			mnt->nfs_server.addrlen =
+				rpc_pton(string, strlen(string),
+					(struct sockaddr *)
+					&mnt->nfs_server.address,
+					sizeof(mnt->nfs_server.address));
 			kfree(string);
+			if (mnt->nfs_server.addrlen == 0)
+				goto out_invalid_address;
 			break;
 		case Opt_clientaddr:
 			string = match_strdup(args);
@@ -1376,11 +1288,14 @@ static int nfs_parse_mount_options(char *raw,
 			string = match_strdup(args);
 			if (string == NULL)
 				goto out_nomem;
-			nfs_parse_ip_address(string, strlen(string),
-					     (struct sockaddr *)
-						&mnt->mount_server.address,
-					     &mnt->mount_server.addrlen);
+			mnt->mount_server.addrlen =
+				rpc_pton(string, strlen(string),
+					(struct sockaddr *)
+					&mnt->mount_server.address,
+					sizeof(mnt->mount_server.address));
 			kfree(string);
+			if (mnt->mount_server.addrlen == 0)
+				goto out_invalid_address;
 			break;
 		case Opt_lookupcache:
 			string = match_strdup(args);
@@ -1432,8 +1347,11 @@ static int nfs_parse_mount_options(char *raw,
 
 	return 1;
 
+out_invalid_address:
+	printk(KERN_INFO "NFS: bad IP address specified: %s\n", p);
+	return 0;
 out_invalid_value:
-	printk(KERN_INFO "NFS: bad mount option value specified: %s \n", p);
+	printk(KERN_INFO "NFS: bad mount option value specified: %s\n", p);
 	return 0;
 out_nomem:
 	printk(KERN_INFO "NFS: not enough memory to parse option\n");
@@ -1445,13 +1363,60 @@ out_security_failure:
 }
 
 /*
+ * Match the requested auth flavors with the list returned by
+ * the server.  Returns zero and sets the mount's authentication
+ * flavor on success; returns -EACCES if server does not support
+ * the requested flavor.
+ */
+static int nfs_walk_authlist(struct nfs_parsed_mount_data *args,
+			     struct nfs_mount_request *request)
+{
+	unsigned int i, j, server_authlist_len = *(request->auth_flav_len);
+
+	/*
+	 * Certain releases of Linux's mountd return an empty
+	 * flavor list.  To prevent behavioral regression with
+	 * these servers (ie. rejecting mounts that used to
+	 * succeed), revert to pre-2.6.32 behavior (no checking)
+	 * if the returned flavor list is empty.
+	 */
+	if (server_authlist_len == 0)
+		return 0;
+
+	/*
+	 * We avoid sophisticated negotiating here, as there are
+	 * plenty of cases where we can get it wrong, providing
+	 * either too little or too much security.
+	 *
+	 * RFC 2623, section 2.7 suggests we SHOULD prefer the
+	 * flavor listed first.  However, some servers list
+	 * AUTH_NULL first.  Our caller plants AUTH_SYS, the
+	 * preferred default, in args->auth_flavors[0] if user
+	 * didn't specify sec= mount option.
+	 */
+	for (i = 0; i < args->auth_flavor_len; i++)
+		for (j = 0; j < server_authlist_len; j++)
+			if (args->auth_flavors[i] == request->auth_flavs[j]) {
+				dfprintk(MOUNT, "NFS: using auth flavor %d\n",
+					request->auth_flavs[j]);
+				args->auth_flavors[0] = request->auth_flavs[j];
+				return 0;
+			}
+
+	dfprintk(MOUNT, "NFS: server does not support requested auth flavor\n");
+	nfs_umount(request);
+	return -EACCES;
+}
+
+/*
  * Use the remote server's MOUNT service to request the NFS file handle
  * corresponding to the provided path.
  */
 static int nfs_try_mount(struct nfs_parsed_mount_data *args,
 			 struct nfs_fh *root_fh)
 {
-	unsigned int auth_flavor_len = 0;
+	rpc_authflavor_t server_authlist[NFS_MAX_SECFLAVORS];
+	unsigned int server_authlist_len = ARRAY_SIZE(server_authlist);
 	struct nfs_mount_request request = {
 		.sap		= (struct sockaddr *)
 						&args->mount_server.address,
@@ -1459,7 +1424,8 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args,
 		.protocol	= args->mount_server.protocol,
 		.fh		= root_fh,
 		.noresvport	= args->flags & NFS_MOUNT_NORESVPORT,
-		.auth_flav_len	= &auth_flavor_len,
+		.auth_flav_len	= &server_authlist_len,
+		.auth_flavs	= server_authlist,
 	};
 	int status;
 
@@ -1485,23 +1451,25 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args,
 		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(request.sap, args->mount_server.port);
+	nfs_set_default_port(request.sap, args->mount_server.port, 0);
 
 	/*
 	 * Now ask the mount server to map our export path
 	 * to a file handle.
 	 */
 	status = nfs_mount(&request);
-	if (status == 0)
-		return 0;
+	if (status != 0) {
+		dfprintk(MOUNT, "NFS: unable to mount server %s, error %d\n",
+				request.hostname, status);
+		return status;
+	}
 
-	dfprintk(MOUNT, "NFS: unable to mount server %s, error %d\n",
-			request.hostname, status);
-	return status;
+	/*
+	 * MNTv1 (NFSv2) does not support auth flavor negotiation.
+	 */
+	if (args->mount_server.version != NFS_MNT3_VERSION)
+		return 0;
+	return nfs_walk_authlist(args, &request);
 }
 
 static int nfs_parse_simple_hostname(const char *dev_name,
@@ -1661,6 +1629,7 @@ static int nfs_validate_mount_data(void *options,
 				   const char *dev_name)
 {
 	struct nfs_mount_data *data = (struct nfs_mount_data *)options;
+	struct sockaddr *sap = (struct sockaddr *)&args->nfs_server.address;
 
 	if (data == NULL)
 		goto out_no_data;
@@ -1672,10 +1641,12 @@ static int nfs_validate_mount_data(void *options,
 	args->acregmax		= NFS_DEF_ACREGMAX;
 	args->acdirmin		= NFS_DEF_ACDIRMIN;
 	args->acdirmax		= NFS_DEF_ACDIRMAX;
-	args->mount_server.port	= 0;	/* autobind unless user sets port */
-	args->nfs_server.port	= 0;	/* autobind unless user sets port */
+	args->mount_server.port	= NFS_UNSPEC_PORT;
+	args->nfs_server.port	= NFS_UNSPEC_PORT;
 	args->nfs_server.protocol = XPRT_TRANSPORT_TCP;
 	args->auth_flavors[0]	= RPC_AUTH_UNIX;
+	args->auth_flavor_len	= 1;
+	args->minorversion	= 0;
 
 	switch (data->version) {
 	case 1:
@@ -1697,8 +1668,11 @@ static int nfs_validate_mount_data(void *options,
 			if (data->root.size > NFS3_FHSIZE || data->root.size == 0)
 				goto out_invalid_fh;
 			mntfh->size = data->root.size;
-		} else
+			args->version = 3;
+		} else {
 			mntfh->size = NFS2_FHSIZE;
+			args->version = 2;
+		}
 
 
 		memcpy(mntfh->data, data->root.data, mntfh->size);
@@ -1720,11 +1694,9 @@ static int nfs_validate_mount_data(void *options,
 		args->acdirmin		= data->acdirmin;
 		args->acdirmax		= data->acdirmax;
 
-		memcpy(&args->nfs_server.address, &data->addr,
-		       sizeof(data->addr));
+		memcpy(sap, &data->addr, sizeof(data->addr));
 		args->nfs_server.addrlen = sizeof(data->addr);
-		if (!nfs_verify_server_address((struct sockaddr *)
-						&args->nfs_server.address))
+		if (!nfs_verify_server_address(sap))
 			goto out_no_address;
 
 		if (!(data->flags & NFS_MOUNT_TCP))
@@ -1772,12 +1744,18 @@ static int nfs_validate_mount_data(void *options,
 		if (nfs_parse_mount_options((char *)options, args) == 0)
 			return -EINVAL;
 
-		if (!nfs_verify_server_address((struct sockaddr *)
-						&args->nfs_server.address))
+		if (!nfs_verify_server_address(sap))
 			goto out_no_address;
 
-		nfs_set_port((struct sockaddr *)&args->nfs_server.address,
-				args->nfs_server.port);
+		if (args->version == 4)
+#ifdef CONFIG_NFS_V4
+			return nfs4_validate_text_mount_data(options,
+							     args, dev_name);
+#else
+			goto out_v4_not_compiled;
+#endif
+
+		nfs_set_default_port(sap, args->nfs_server.port, 0);
 
 		nfs_set_mount_transport_protocol(args);
 
@@ -1825,6 +1803,12 @@ out_v3_not_compiled:
 	return -EPROTONOSUPPORT;
 #endif /* !CONFIG_NFS_V3 */
 
+#ifndef CONFIG_NFS_V4
+out_v4_not_compiled:
+	dfprintk(MOUNT, "NFS: NFSv4 is not compiled into kernel\n");
+	return -EPROTONOSUPPORT;
+#endif /* !CONFIG_NFS_V4 */
+
 out_nomem:
 	dfprintk(MOUNT, "NFS: not enough memory to handle mount options\n");
 	return -ENOMEM;
@@ -1934,6 +1918,8 @@ static inline void nfs_initialise_sb(struct super_block *sb)
 	if (server->flags & NFS_MOUNT_NOAC)
 		sb->s_flags |= MS_SYNCHRONOUS;
 
+	sb->s_bdi = &server->backing_dev_info;
+
 	nfs_super_set_maxbytes(sb, server->maxfilesize);
 }
 
@@ -2120,6 +2106,14 @@ static int nfs_get_sb(struct file_system_type *fs_type,
 	if (error < 0)
 		goto out;
 
+#ifdef CONFIG_NFS_V4
+	if (data->version == 4) {
+		error = nfs4_try_mount(flags, dev_name, data, mnt);
+		kfree(data->client_address);
+		goto out;
+	}
+#endif	/* CONFIG_NFS_V4 */
+
 	/* Get a volume representation */
 	server = nfs_create_server(data, mntfh);
 	if (IS_ERR(server)) {
@@ -2196,8 +2190,8 @@ static void nfs_kill_super(struct super_block *s)
 {
 	struct nfs_server *server = NFS_SB(s);
 
-	bdi_unregister(&server->backing_dev_info);
 	kill_anon_super(s);
+	bdi_unregister(&server->backing_dev_info);
 	nfs_fscache_release_super_cookie(s);
 	nfs_free_server(server);
 }
@@ -2317,6 +2311,43 @@ static void nfs4_validate_mount_flags(struct nfs_parsed_mount_data *args)
 	args->flags &= ~(NFS_MOUNT_NONLM|NFS_MOUNT_NOACL|NFS_MOUNT_VER3);
 }
 
+static int nfs4_validate_text_mount_data(void *options,
+					 struct nfs_parsed_mount_data *args,
+					 const char *dev_name)
+{
+	struct sockaddr *sap = (struct sockaddr *)&args->nfs_server.address;
+
+	nfs_set_default_port(sap, args->nfs_server.port, NFS_PORT);
+
+	nfs_validate_transport_protocol(args);
+
+	nfs4_validate_mount_flags(args);
+
+	if (args->version != 4) {
+		dfprintk(MOUNT,
+			 "NFS4: Illegal mount version\n");
+		return -EINVAL;
+	}
+
+	if (args->auth_flavor_len > 1) {
+		dfprintk(MOUNT,
+			 "NFS4: Too many RPC auth flavours specified\n");
+		return -EINVAL;
+	}
+
+	if (args->client_address == NULL) {
+		dfprintk(MOUNT,
+			 "NFS4: mount program didn't pass callback address\n");
+		return -EINVAL;
+	}
+
+	return nfs_parse_devname(dev_name,
+				   &args->nfs_server.hostname,
+				   NFS4_MAXNAMLEN,
+				   &args->nfs_server.export_path,
+				   NFS4_MAXPATHLEN);
+}
+
 /*
  * Validate NFSv4 mount options
  */
@@ -2324,7 +2355,7 @@ static int nfs4_validate_mount_data(void *options,
 				    struct nfs_parsed_mount_data *args,
 				    const char *dev_name)
 {
-	struct sockaddr_in *ap;
+	struct sockaddr *sap = (struct sockaddr *)&args->nfs_server.address;
 	struct nfs4_mount_data *data = (struct nfs4_mount_data *)options;
 	char *c;
 
@@ -2337,23 +2368,22 @@ static int nfs4_validate_mount_data(void *options,
 	args->acregmax		= NFS_DEF_ACREGMAX;
 	args->acdirmin		= NFS_DEF_ACDIRMIN;
 	args->acdirmax		= NFS_DEF_ACDIRMAX;
-	args->nfs_server.port	= NFS_PORT; /* 2049 unless user set port= */
+	args->nfs_server.port	= NFS_UNSPEC_PORT;
 	args->auth_flavors[0]	= RPC_AUTH_UNIX;
-	args->auth_flavor_len	= 0;
+	args->auth_flavor_len	= 1;
+	args->version		= 4;
 	args->minorversion	= 0;
 
 	switch (data->version) {
 	case 1:
-		ap = (struct sockaddr_in *)&args->nfs_server.address;
 		if (data->host_addrlen > sizeof(args->nfs_server.address))
 			goto out_no_address;
 		if (data->host_addrlen == 0)
 			goto out_no_address;
 		args->nfs_server.addrlen = data->host_addrlen;
-		if (copy_from_user(ap, data->host_addr, data->host_addrlen))
+		if (copy_from_user(sap, data->host_addr, data->host_addrlen))
 			return -EFAULT;
-		if (!nfs_verify_server_address((struct sockaddr *)
-						&args->nfs_server.address))
+		if (!nfs_verify_server_address(sap))
 			goto out_no_address;
 
 		if (data->auth_flavourlen) {
@@ -2399,39 +2429,14 @@ static int nfs4_validate_mount_data(void *options,
 		nfs_validate_transport_protocol(args);
 
 		break;
-	default: {
-		int status;
-
+	default:
 		if (nfs_parse_mount_options((char *)options, args) == 0)
 			return -EINVAL;
 
-		if (!nfs_verify_server_address((struct sockaddr *)
-						&args->nfs_server.address))
+		if (!nfs_verify_server_address(sap))
 			return -EINVAL;
 
-		nfs_set_port((struct sockaddr *)&args->nfs_server.address,
-				args->nfs_server.port);
-
-		nfs_validate_transport_protocol(args);
-
-		nfs4_validate_mount_flags(args);
-
-		if (args->auth_flavor_len > 1)
-			goto out_inval_auth;
-
-		if (args->client_address == NULL)
-			goto out_no_client_address;
-
-		status = nfs_parse_devname(dev_name,
-					   &args->nfs_server.hostname,
-					   NFS4_MAXNAMLEN,
-					   &args->nfs_server.export_path,
-					   NFS4_MAXPATHLEN);
-		if (status < 0)
-			return status;
-
-		break;
-		}
+		return nfs4_validate_text_mount_data(options, args, dev_name);
 	}
 
 	return 0;
@@ -2448,10 +2453,6 @@ out_inval_auth:
 out_no_address:
 	dfprintk(MOUNT, "NFS4: mount program didn't pass remote address\n");
 	return -EINVAL;
-
-out_no_client_address:
-	dfprintk(MOUNT, "NFS4: mount program didn't pass callback address\n");
-	return -EINVAL;
 }
 
 /*
@@ -2618,6 +2619,34 @@ out_err:
 	return ret;
 }
 
+static int nfs4_try_mount(int flags, const char *dev_name,
+			 struct nfs_parsed_mount_data *data,
+			 struct vfsmount *mnt)
+{
+	char *export_path;
+	struct vfsmount *root_mnt;
+	int error;
+
+	dfprintk(MOUNT, "--> nfs4_try_mount()\n");
+
+	export_path = data->nfs_server.export_path;
+	data->nfs_server.export_path = "/";
+	root_mnt = nfs_do_root_mount(&nfs4_remote_fs_type, flags, data,
+			data->nfs_server.hostname);
+	data->nfs_server.export_path = export_path;
+
+	error = PTR_ERR(root_mnt);
+	if (IS_ERR(root_mnt))
+		goto out;
+
+	error = nfs_follow_remote_path(root_mnt, export_path, mnt);
+
+out:
+	dfprintk(MOUNT, "<-- nfs4_try_mount() = %d%s\n", error,
+			error != 0 ? " [error]" : "");
+	return error;
+}
+
 /*
  * Get the superblock for an NFS4 mountpoint
  */
@@ -2625,8 +2654,6 @@ static int nfs4_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt)
 {
 	struct nfs_parsed_mount_data *data;
-	char *export_path;
-	struct vfsmount *root_mnt;
 	int error = -ENOMEM;
 
 	data = kzalloc(sizeof(*data), GFP_KERNEL);
@@ -2638,17 +2665,7 @@ static int nfs4_get_sb(struct file_system_type *fs_type,
 	if (error < 0)
 		goto out;
 
-	export_path = data->nfs_server.export_path;
-	data->nfs_server.export_path = "/";
-	root_mnt = nfs_do_root_mount(&nfs4_remote_fs_type, flags, data,
-			data->nfs_server.hostname);
-	data->nfs_server.export_path = export_path;
-
-	error = PTR_ERR(root_mnt);
-	if (IS_ERR(root_mnt))
-		goto out;
-
-	error = nfs_follow_remote_path(root_mnt, export_path, mnt);
+	error = nfs4_try_mount(flags, dev_name, data, mnt);
 
 out:
 	kfree(data->client_address);
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index a34fae21fe1..53eb26c16b5 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -13,6 +13,7 @@
 #include <linux/file.h>
 #include <linux/writeback.h>
 #include <linux/swap.h>
+#include <linux/migrate.h>
 
 #include <linux/sunrpc/clnt.h>
 #include <linux/nfs_fs.h>
@@ -26,6 +27,7 @@
 #include "internal.h"
 #include "iostat.h"
 #include "nfs4_fs.h"
+#include "fscache.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PAGECACHE
 
@@ -218,24 +220,17 @@ static void nfs_end_page_writeback(struct page *page)
 		clear_bdi_congested(&nfss->backing_dev_info, BLK_RW_ASYNC);
 }
 
-/*
- * Find an associated nfs write request, and prepare to flush it out
- * May return an error if the user signalled nfs_wait_on_request().
- */
-static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
-				struct page *page)
+static struct nfs_page *nfs_find_and_lock_request(struct page *page)
 {
 	struct inode *inode = page->mapping->host;
 	struct nfs_page *req;
 	int ret;
 
 	spin_lock(&inode->i_lock);
-	for(;;) {
+	for (;;) {
 		req = nfs_page_find_request_locked(page);
-		if (req == NULL) {
-			spin_unlock(&inode->i_lock);
-			return 0;
-		}
+		if (req == NULL)
+			break;
 		if (nfs_set_page_tag_locked(req))
 			break;
 		/* Note: If we hold the page lock, as is the case in nfs_writepage,
@@ -247,23 +242,40 @@ static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
 		ret = nfs_wait_on_request(req);
 		nfs_release_request(req);
 		if (ret != 0)
-			return ret;
+			return ERR_PTR(ret);
 		spin_lock(&inode->i_lock);
 	}
-	if (test_bit(PG_CLEAN, &req->wb_flags)) {
-		spin_unlock(&inode->i_lock);
-		BUG();
-	}
-	if (nfs_set_page_writeback(page) != 0) {
-		spin_unlock(&inode->i_lock);
-		BUG();
-	}
 	spin_unlock(&inode->i_lock);
+	return req;
+}
+
+/*
+ * Find an associated nfs write request, and prepare to flush it out
+ * May return an error if the user signalled nfs_wait_on_request().
+ */
+static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
+				struct page *page)
+{
+	struct nfs_page *req;
+	int ret = 0;
+
+	req = nfs_find_and_lock_request(page);
+	if (!req)
+		goto out;
+	ret = PTR_ERR(req);
+	if (IS_ERR(req))
+		goto out;
+
+	ret = nfs_set_page_writeback(page);
+	BUG_ON(ret != 0);
+	BUG_ON(test_bit(PG_CLEAN, &req->wb_flags));
+
 	if (!nfs_pageio_add_request(pgio, req)) {
 		nfs_redirty_request(req);
-		return pgio->pg_error;
+		ret = pgio->pg_error;
 	}
-	return 0;
+out:
+	return ret;
 }
 
 static int nfs_do_writepage(struct page *page, struct writeback_control *wbc, struct nfs_pageio_descriptor *pgio)
@@ -1478,7 +1490,6 @@ static int nfs_write_mapping(struct address_space *mapping, int how)
 		.nr_to_write = LONG_MAX,
 		.range_start = 0,
 		.range_end = LLONG_MAX,
-		.for_writepages = 1,
 	};
 
 	return __nfs_write_mapping(mapping, &wbc, how);
@@ -1580,6 +1591,41 @@ int nfs_wb_page(struct inode *inode, struct page* page)
 	return nfs_wb_page_priority(inode, page, FLUSH_STABLE);
 }
 
+#ifdef CONFIG_MIGRATION
+int nfs_migrate_page(struct address_space *mapping, struct page *newpage,
+		struct page *page)
+{
+	struct nfs_page *req;
+	int ret;
+
+	if (PageFsCache(page))
+		nfs_fscache_release_page(page, GFP_KERNEL);
+
+	req = nfs_find_and_lock_request(page);
+	ret = PTR_ERR(req);
+	if (IS_ERR(req))
+		goto out;
+
+	ret = migrate_page(mapping, newpage, page);
+	if (!req)
+		goto out;
+	if (ret)
+		goto out_unlock;
+	page_cache_get(newpage);
+	req->wb_page = newpage;
+	SetPagePrivate(newpage);
+	set_page_private(newpage, page_private(page));
+	ClearPagePrivate(page);
+	set_page_private(page, 0);
+	page_cache_release(page);
+out_unlock:
+	nfs_clear_page_tag_locked(req);
+	nfs_release_request(req);
+out:
+	return ret;
+}
+#endif
+
 int __init nfs_init_writepagecache(void)
 {
 	nfs_wdata_cachep = kmem_cache_create("nfs_write_data",
diff --git a/fs/nfsd/auth.c b/fs/nfsd/auth.c
index 5573508f707..36fcabbf518 100644
--- a/fs/nfsd/auth.c
+++ b/fs/nfsd/auth.c
@@ -34,6 +34,8 @@ int nfsd_setuser(struct svc_rqst *rqstp, struct svc_export *exp)
 	int flags = nfsexp_flags(rqstp, exp);
 	int ret;
 
+	validate_process_creds();
+
 	/* discard any old override before preparing the new set */
 	revert_creds(get_cred(current->real_cred));
 	new = prepare_creds();
@@ -86,8 +88,10 @@ int nfsd_setuser(struct svc_rqst *rqstp, struct svc_export *exp)
 	else
 		new->cap_effective = cap_raise_nfsd_set(new->cap_effective,
 							new->cap_permitted);
+	validate_process_creds();
 	put_cred(override_creds(new));
 	put_cred(new);
+	validate_process_creds();
 	return 0;
 
 oom:
diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c
index b92a27629fb..c1c9e035d4a 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -85,6 +85,11 @@ static void expkey_request(struct cache_detail *cd,
 	(*bpp)[-1] = '\n';
 }
 
+static int expkey_upcall(struct cache_detail *cd, struct cache_head *h)
+{
+	return sunrpc_cache_pipe_upcall(cd, h, expkey_request);
+}
+
 static struct svc_expkey *svc_expkey_update(struct svc_expkey *new, struct svc_expkey *old);
 static struct svc_expkey *svc_expkey_lookup(struct svc_expkey *);
 static struct cache_detail svc_expkey_cache;
@@ -259,7 +264,7 @@ static struct cache_detail svc_expkey_cache = {
 	.hash_table	= expkey_table,
 	.name		= "nfsd.fh",
 	.cache_put	= expkey_put,
-	.cache_request	= expkey_request,
+	.cache_upcall	= expkey_upcall,
 	.cache_parse	= expkey_parse,
 	.cache_show	= expkey_show,
 	.match		= expkey_match,
@@ -355,6 +360,11 @@ static void svc_export_request(struct cache_detail *cd,
 	(*bpp)[-1] = '\n';
 }
 
+static int svc_export_upcall(struct cache_detail *cd, struct cache_head *h)
+{
+	return sunrpc_cache_pipe_upcall(cd, h, svc_export_request);
+}
+
 static struct svc_export *svc_export_update(struct svc_export *new,
 					    struct svc_export *old);
 static struct svc_export *svc_export_lookup(struct svc_export *);
@@ -724,7 +734,7 @@ struct cache_detail svc_export_cache = {
 	.hash_table	= export_table,
 	.name		= "nfsd.export",
 	.cache_put	= svc_export_put,
-	.cache_request	= svc_export_request,
+	.cache_upcall	= svc_export_upcall,
 	.cache_parse	= svc_export_parse,
 	.cache_show	= svc_export_show,
 	.match		= svc_export_match,
@@ -1331,6 +1341,8 @@ exp_pseudoroot(struct svc_rqst *rqstp, struct svc_fh *fhp)
 	if (rv)
 		goto out;
 	rv = check_nfsd_access(exp, rqstp);
+	if (rv)
+		fh_put(fhp);
 out:
 	exp_put(exp);
 	return rv;
@@ -1505,7 +1517,7 @@ static int e_show(struct seq_file *m, void *p)
 	return svc_export_show(m, &svc_export_cache, cp);
 }
 
-struct seq_operations nfs_exports_op = {
+const struct seq_operations nfs_exports_op = {
 	.start	= e_start,
 	.next	= e_next,
 	.stop	= e_stop,
diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c
index 01d4ec1c88e..edf926e1062 100644
--- a/fs/nfsd/nfs3xdr.c
+++ b/fs/nfsd/nfs3xdr.c
@@ -814,17 +814,6 @@ encode_entry_baggage(struct nfsd3_readdirres *cd, __be32 *p, const char *name,
 	return p;
 }
 
-static __be32 *
-encode_entryplus_baggage(struct nfsd3_readdirres *cd, __be32 *p,
-		struct svc_fh *fhp)
-{
-	p = encode_post_op_attr(cd->rqstp, p, fhp);
-	*p++ = xdr_one;			/* yes, a file handle follows */
-	p = encode_fh(p, fhp);
-	fh_put(fhp);
-	return p;
-}
-
 static int
 compose_entry_fh(struct nfsd3_readdirres *cd, struct svc_fh *fhp,
 		const char *name, int namlen)
@@ -836,29 +825,54 @@ compose_entry_fh(struct nfsd3_readdirres *cd, struct svc_fh *fhp,
 	dparent = cd->fh.fh_dentry;
 	exp  = cd->fh.fh_export;
 
-	fh_init(fhp, NFS3_FHSIZE);
 	if (isdotent(name, namlen)) {
 		if (namlen == 2) {
 			dchild = dget_parent(dparent);
 			if (dchild == dparent) {
 				/* filesystem root - cannot return filehandle for ".." */
 				dput(dchild);
-				return 1;
+				return -ENOENT;
 			}
 		} else
 			dchild = dget(dparent);
 	} else
 		dchild = lookup_one_len(name, dparent, namlen);
 	if (IS_ERR(dchild))
-		return 1;
-	if (d_mountpoint(dchild) ||
-	    fh_compose(fhp, exp, dchild, &cd->fh) != 0 ||
-	    !dchild->d_inode)
-		rv = 1;
+		return -ENOENT;
+	rv = -ENOENT;
+	if (d_mountpoint(dchild))
+		goto out;
+	rv = fh_compose(fhp, exp, dchild, &cd->fh);
+	if (rv)
+		goto out;
+	if (!dchild->d_inode)
+		goto out;
+	rv = 0;
+out:
 	dput(dchild);
 	return rv;
 }
 
+__be32 *encode_entryplus_baggage(struct nfsd3_readdirres *cd, __be32 *p, const char *name, int namlen)
+{
+	struct svc_fh	fh;
+	int err;
+
+	fh_init(&fh, NFS3_FHSIZE);
+	err = compose_entry_fh(cd, &fh, name, namlen);
+	if (err) {
+		*p++ = 0;
+		*p++ = 0;
+		goto out;
+	}
+	p = encode_post_op_attr(cd->rqstp, p, &fh);
+	*p++ = xdr_one;			/* yes, a file handle follows */
+	p = encode_fh(p, &fh);
+out:
+	fh_put(&fh);
+	return p;
+}
+
 /*
  * Encode a directory entry. This one works for both normal readdir
  * and readdirplus.
@@ -929,16 +943,8 @@ encode_entry(struct readdir_cd *ccd, const char *name, int namlen,
 
 		p = encode_entry_baggage(cd, p, name, namlen, ino);
 
-		/* throw in readdirplus baggage */
-		if (plus) {
-			struct svc_fh	fh;
-
-			if (compose_entry_fh(cd, &fh, name, namlen) > 0) {
-				*p++ = 0;
-				*p++ = 0;
-			} else
-				p = encode_entryplus_baggage(cd, p, &fh);
-		}
+		if (plus)
+			p = encode_entryplus_baggage(cd, p, name, namlen);
 		num_entry_words = p - cd->buffer;
 	} else if (cd->rqstp->rq_respages[pn+1] != NULL) {
 		/* temporarily encode entry into next page, then move back to
@@ -951,17 +957,8 @@ encode_entry(struct readdir_cd *ccd, const char *name, int namlen,
 
 		p1 = encode_entry_baggage(cd, p1, name, namlen, ino);
 
-		/* throw in readdirplus baggage */
-		if (plus) {
-			struct svc_fh	fh;
-
-			if (compose_entry_fh(cd, &fh, name, namlen) > 0) {
-				/* zero out the filehandle */
-				*p1++ = 0;
-				*p1++ = 0;
-			} else
-				p1 = encode_entryplus_baggage(cd, p1, &fh);
-		}
+		if (plus)
+			p = encode_entryplus_baggage(cd, p1, name, namlen);
 
 		/* determine entry word length and lengths to go in pages */
 		num_entry_words = p1 - tmp;
diff --git a/fs/nfsd/nfs4acl.c b/fs/nfsd/nfs4acl.c
index 54b8b4140c8..725d02f210e 100644
--- a/fs/nfsd/nfs4acl.c
+++ b/fs/nfsd/nfs4acl.c
@@ -321,7 +321,7 @@ _posix_to_nfsv4_one(struct posix_acl *pacl, struct nfs4_acl *acl,
 	deny = ~pas.group & pas.other;
 	if (deny) {
 		ace->type = NFS4_ACE_ACCESS_DENIED_ACE_TYPE;
-		ace->flag = eflag | NFS4_ACE_IDENTIFIER_GROUP;
+		ace->flag = eflag;
 		ace->access_mask = deny_mask_from_posix(deny, flags);
 		ace->whotype = NFS4_ACL_WHO_GROUP;
 		ace++;
@@ -335,7 +335,7 @@ _posix_to_nfsv4_one(struct posix_acl *pacl, struct nfs4_acl *acl,
 		if (deny) {
 			ace->type = NFS4_ACE_ACCESS_DENIED_ACE_TYPE;
 			ace->flag = eflag | NFS4_ACE_IDENTIFIER_GROUP;
-			ace->access_mask = mask_from_posix(deny, flags);
+			ace->access_mask = deny_mask_from_posix(deny, flags);
 			ace->whotype = NFS4_ACL_WHO_NAMED;
 			ace->who = pa->e_id;
 			ace++;
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 3fd23f7acec..24e8d78f8dd 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -43,25 +43,30 @@
 #include <linux/sunrpc/xdr.h>
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/clnt.h>
+#include <linux/sunrpc/svcsock.h>
 #include <linux/nfsd/nfsd.h>
 #include <linux/nfsd/state.h>
 #include <linux/sunrpc/sched.h>
 #include <linux/nfs4.h>
+#include <linux/sunrpc/xprtsock.h>
 
 #define NFSDDBG_FACILITY                NFSDDBG_PROC
 
 #define NFSPROC4_CB_NULL 0
 #define NFSPROC4_CB_COMPOUND 1
+#define NFS4_STATEID_SIZE 16
 
 /* Index of predefined Linux callback client operations */
 
 enum {
-        NFSPROC4_CLNT_CB_NULL = 0,
+	NFSPROC4_CLNT_CB_NULL = 0,
 	NFSPROC4_CLNT_CB_RECALL,
+	NFSPROC4_CLNT_CB_SEQUENCE,
 };
 
 enum nfs_cb_opnum4 {
 	OP_CB_RECALL            = 4,
+	OP_CB_SEQUENCE          = 11,
 };
 
 #define NFS4_MAXTAGLEN		20
@@ -70,17 +75,29 @@ enum nfs_cb_opnum4 {
 #define NFS4_dec_cb_null_sz		0
 #define cb_compound_enc_hdr_sz		4
 #define cb_compound_dec_hdr_sz		(3 + (NFS4_MAXTAGLEN >> 2))
+#define sessionid_sz			(NFS4_MAX_SESSIONID_LEN >> 2)
+#define cb_sequence_enc_sz		(sessionid_sz + 4 +             \
+					1 /* no referring calls list yet */)
+#define cb_sequence_dec_sz		(op_dec_sz + sessionid_sz + 4)
+
 #define op_enc_sz			1
 #define op_dec_sz			2
 #define enc_nfs4_fh_sz			(1 + (NFS4_FHSIZE >> 2))
 #define enc_stateid_sz			(NFS4_STATEID_SIZE >> 2)
 #define NFS4_enc_cb_recall_sz		(cb_compound_enc_hdr_sz +       \
+					cb_sequence_enc_sz +            \
 					1 + enc_stateid_sz +            \
 					enc_nfs4_fh_sz)
 
 #define NFS4_dec_cb_recall_sz		(cb_compound_dec_hdr_sz  +      \
+					cb_sequence_dec_sz +            \
 					op_dec_sz)
 
+struct nfs4_rpc_args {
+	void				*args_op;
+	struct nfsd4_cb_sequence	args_seq;
+};
+
 /*
 * Generic encode routines from fs/nfs/nfs4xdr.c
 */
@@ -137,11 +154,13 @@ xdr_error:                                      \
 } while (0)
 
 struct nfs4_cb_compound_hdr {
-	int		status;
-	u32		ident;
+	/* args */
+	u32		ident;	/* minorversion 0 only */
 	u32		nops;
 	__be32		*nops_p;
 	u32		minorversion;
+	/* res */
+	int		status;
 	u32		taglen;
 	char		*tag;
 };
@@ -238,6 +257,27 @@ encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp,
 	hdr->nops++;
 }
 
+static void
+encode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_cb_sequence *args,
+		   struct nfs4_cb_compound_hdr *hdr)
+{
+	__be32 *p;
+
+	if (hdr->minorversion == 0)
+		return;
+
+	RESERVE_SPACE(1 + NFS4_MAX_SESSIONID_LEN + 20);
+
+	WRITE32(OP_CB_SEQUENCE);
+	WRITEMEM(args->cbs_clp->cl_sessionid.data, NFS4_MAX_SESSIONID_LEN);
+	WRITE32(args->cbs_clp->cl_cb_seq_nr);
+	WRITE32(0);		/* slotid, always 0 */
+	WRITE32(0);		/* highest slotid always 0 */
+	WRITE32(0);		/* cachethis always 0 */
+	WRITE32(0); /* FIXME: support referring_call_lists */
+	hdr->nops++;
+}
+
 static int
 nfs4_xdr_enc_cb_null(struct rpc_rqst *req, __be32 *p)
 {
@@ -249,15 +289,19 @@ nfs4_xdr_enc_cb_null(struct rpc_rqst *req, __be32 *p)
 }
 
 static int
-nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, __be32 *p, struct nfs4_delegation *args)
+nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, __be32 *p,
+		struct nfs4_rpc_args *rpc_args)
 {
 	struct xdr_stream xdr;
+	struct nfs4_delegation *args = rpc_args->args_op;
 	struct nfs4_cb_compound_hdr hdr = {
 		.ident = args->dl_ident,
+		.minorversion = rpc_args->args_seq.cbs_minorversion,
 	};
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_cb_compound_hdr(&xdr, &hdr);
+	encode_cb_sequence(&xdr, &rpc_args->args_seq, &hdr);
 	encode_cb_recall(&xdr, args, &hdr);
 	encode_cb_nops(&hdr);
 	return 0;
@@ -299,6 +343,57 @@ decode_cb_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected)
 	return 0;
 }
 
+/*
+ * Our current back channel implmentation supports a single backchannel
+ * with a single slot.
+ */
+static int
+decode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_cb_sequence *res,
+		   struct rpc_rqst *rqstp)
+{
+	struct nfs4_sessionid id;
+	int status;
+	u32 dummy;
+	__be32 *p;
+
+	if (res->cbs_minorversion == 0)
+		return 0;
+
+	status = decode_cb_op_hdr(xdr, OP_CB_SEQUENCE);
+	if (status)
+		return status;
+
+	/*
+	 * If the server returns different values for sessionID, slotID or
+	 * sequence number, the server is looney tunes.
+	 */
+	status = -ESERVERFAULT;
+
+	READ_BUF(NFS4_MAX_SESSIONID_LEN + 16);
+	memcpy(id.data, p, NFS4_MAX_SESSIONID_LEN);
+	p += XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN);
+	if (memcmp(id.data, res->cbs_clp->cl_sessionid.data,
+		   NFS4_MAX_SESSIONID_LEN)) {
+		dprintk("%s Invalid session id\n", __func__);
+		goto out;
+	}
+	READ32(dummy);
+	if (dummy != res->cbs_clp->cl_cb_seq_nr) {
+		dprintk("%s Invalid sequence number\n", __func__);
+		goto out;
+	}
+	READ32(dummy); 	/* slotid must be 0 */
+	if (dummy != 0) {
+		dprintk("%s Invalid slotid\n", __func__);
+		goto out;
+	}
+	/* FIXME: process highest slotid and target highest slotid */
+	status = 0;
+out:
+	return status;
+}
+
+
 static int
 nfs4_xdr_dec_cb_null(struct rpc_rqst *req, __be32 *p)
 {
@@ -306,7 +401,8 @@ nfs4_xdr_dec_cb_null(struct rpc_rqst *req, __be32 *p)
 }
 
 static int
-nfs4_xdr_dec_cb_recall(struct rpc_rqst *rqstp, __be32 *p)
+nfs4_xdr_dec_cb_recall(struct rpc_rqst *rqstp, __be32 *p,
+		struct nfsd4_cb_sequence *seq)
 {
 	struct xdr_stream xdr;
 	struct nfs4_cb_compound_hdr hdr;
@@ -316,6 +412,11 @@ nfs4_xdr_dec_cb_recall(struct rpc_rqst *rqstp, __be32 *p)
 	status = decode_cb_compound_hdr(&xdr, &hdr);
 	if (status)
 		goto out;
+	if (seq) {
+		status = decode_cb_sequence(&xdr, seq, rqstp);
+		if (status)
+			goto out;
+	}
 	status = decode_cb_op_hdr(&xdr, OP_CB_RECALL);
 out:
 	return status;
@@ -377,16 +478,15 @@ static int max_cb_time(void)
 
 int setup_callback_client(struct nfs4_client *clp)
 {
-	struct sockaddr_in	addr;
 	struct nfs4_cb_conn *cb = &clp->cl_cb_conn;
 	struct rpc_timeout	timeparms = {
 		.to_initval	= max_cb_time(),
 		.to_retries	= 0,
 	};
 	struct rpc_create_args args = {
-		.protocol	= IPPROTO_TCP,
-		.address	= (struct sockaddr *)&addr,
-		.addrsize	= sizeof(addr),
+		.protocol	= XPRT_TRANSPORT_TCP,
+		.address	= (struct sockaddr *) &cb->cb_addr,
+		.addrsize	= cb->cb_addrlen,
 		.timeout	= &timeparms,
 		.program	= &cb_program,
 		.prognumber	= cb->cb_prog,
@@ -399,13 +499,10 @@ int setup_callback_client(struct nfs4_client *clp)
 
 	if (!clp->cl_principal && (clp->cl_flavor >= RPC_AUTH_GSS_KRB5))
 		return -EINVAL;
-
-	/* Initialize address */
-	memset(&addr, 0, sizeof(addr));
-	addr.sin_family = AF_INET;
-	addr.sin_port = htons(cb->cb_port);
-	addr.sin_addr.s_addr = htonl(cb->cb_addr);
-
+	if (cb->cb_minorversion) {
+		args.bc_xprt = clp->cl_cb_xprt;
+		args.protocol = XPRT_TRANSPORT_BC_TCP;
+	}
 	/* Create RPC client */
 	client = rpc_create(&args);
 	if (IS_ERR(client)) {
@@ -439,42 +536,29 @@ static const struct rpc_call_ops nfsd4_cb_probe_ops = {
 	.rpc_call_done = nfsd4_cb_probe_done,
 };
 
-static struct rpc_cred *lookup_cb_cred(struct nfs4_cb_conn *cb)
-{
-	struct auth_cred acred = {
-		.machine_cred = 1
-	};
+static struct rpc_cred *callback_cred;
 
-	/*
-	 * Note in the gss case this doesn't actually have to wait for a
-	 * gss upcall (or any calls to the client); this just creates a
-	 * non-uptodate cred which the rpc state machine will fill in with
-	 * a refresh_upcall later.
-	 */
-	return rpcauth_lookup_credcache(cb->cb_client->cl_auth, &acred,
-							RPCAUTH_LOOKUP_NEW);
+int set_callback_cred(void)
+{
+	callback_cred = rpc_lookup_machine_cred();
+	if (!callback_cred)
+		return -ENOMEM;
+	return 0;
 }
 
+
 void do_probe_callback(struct nfs4_client *clp)
 {
 	struct nfs4_cb_conn *cb = &clp->cl_cb_conn;
 	struct rpc_message msg = {
 		.rpc_proc       = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL],
 		.rpc_argp       = clp,
+		.rpc_cred	= callback_cred
 	};
-	struct rpc_cred *cred;
 	int status;
 
-	cred = lookup_cb_cred(cb);
-	if (IS_ERR(cred)) {
-		status = PTR_ERR(cred);
-		goto out;
-	}
-	cb->cb_cred = cred;
-	msg.rpc_cred = cb->cb_cred;
 	status = rpc_call_async(cb->cb_client, &msg, RPC_TASK_SOFT,
 				&nfsd4_cb_probe_ops, (void *)clp);
-out:
 	if (status) {
 		warn_no_callback_path(clp, status);
 		put_nfs4_client(clp);
@@ -503,11 +587,95 @@ nfsd4_probe_callback(struct nfs4_client *clp)
 	do_probe_callback(clp);
 }
 
+/*
+ * There's currently a single callback channel slot.
+ * If the slot is available, then mark it busy.  Otherwise, set the
+ * thread for sleeping on the callback RPC wait queue.
+ */
+static int nfsd41_cb_setup_sequence(struct nfs4_client *clp,
+		struct rpc_task *task)
+{
+	struct nfs4_rpc_args *args = task->tk_msg.rpc_argp;
+	u32 *ptr = (u32 *)clp->cl_sessionid.data;
+	int status = 0;
+
+	dprintk("%s: %u:%u:%u:%u\n", __func__,
+		ptr[0], ptr[1], ptr[2], ptr[3]);
+
+	if (test_and_set_bit(0, &clp->cl_cb_slot_busy) != 0) {
+		rpc_sleep_on(&clp->cl_cb_waitq, task, NULL);
+		dprintk("%s slot is busy\n", __func__);
+		status = -EAGAIN;
+		goto out;
+	}
+
+	/*
+	 * We'll need the clp during XDR encoding and decoding,
+	 * and the sequence during decoding to verify the reply
+	 */
+	args->args_seq.cbs_clp = clp;
+	task->tk_msg.rpc_resp = &args->args_seq;
+
+out:
+	dprintk("%s status=%d\n", __func__, status);
+	return status;
+}
+
+/*
+ * TODO: cb_sequence should support referring call lists, cachethis, multiple
+ * slots, and mark callback channel down on communication errors.
+ */
+static void nfsd4_cb_prepare(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_delegation *dp = calldata;
+	struct nfs4_client *clp = dp->dl_client;
+	struct nfs4_rpc_args *args = task->tk_msg.rpc_argp;
+	u32 minorversion = clp->cl_cb_conn.cb_minorversion;
+	int status = 0;
+
+	args->args_seq.cbs_minorversion = minorversion;
+	if (minorversion) {
+		status = nfsd41_cb_setup_sequence(clp, task);
+		if (status) {
+			if (status != -EAGAIN) {
+				/* terminate rpc task */
+				task->tk_status = status;
+				task->tk_action = NULL;
+			}
+			return;
+		}
+	}
+	rpc_call_start(task);
+}
+
+static void nfsd4_cb_done(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_delegation *dp = calldata;
+	struct nfs4_client *clp = dp->dl_client;
+
+	dprintk("%s: minorversion=%d\n", __func__,
+		clp->cl_cb_conn.cb_minorversion);
+
+	if (clp->cl_cb_conn.cb_minorversion) {
+		/* No need for lock, access serialized in nfsd4_cb_prepare */
+		++clp->cl_cb_seq_nr;
+		clear_bit(0, &clp->cl_cb_slot_busy);
+		rpc_wake_up_next(&clp->cl_cb_waitq);
+		dprintk("%s: freed slot, new seqid=%d\n", __func__,
+			clp->cl_cb_seq_nr);
+
+		/* We're done looking into the sequence information */
+		task->tk_msg.rpc_resp = NULL;
+	}
+}
+
 static void nfsd4_cb_recall_done(struct rpc_task *task, void *calldata)
 {
 	struct nfs4_delegation *dp = calldata;
 	struct nfs4_client *clp = dp->dl_client;
 
+	nfsd4_cb_done(task, calldata);
+
 	switch (task->tk_status) {
 	case -EIO:
 		/* Network partition? */
@@ -520,16 +688,19 @@ static void nfsd4_cb_recall_done(struct rpc_task *task, void *calldata)
 		break;
 	default:
 		/* success, or error we can't handle */
-		return;
+		goto done;
 	}
 	if (dp->dl_retries--) {
 		rpc_delay(task, 2*HZ);
 		task->tk_status = 0;
 		rpc_restart_call(task);
+		return;
 	} else {
 		atomic_set(&clp->cl_cb_conn.cb_set, 0);
 		warn_no_callback_path(clp, task->tk_status);
 	}
+done:
+	kfree(task->tk_msg.rpc_argp);
 }
 
 static void nfsd4_cb_recall_release(void *calldata)
@@ -542,6 +713,7 @@ static void nfsd4_cb_recall_release(void *calldata)
 }
 
 static const struct rpc_call_ops nfsd4_cb_recall_ops = {
+	.rpc_call_prepare = nfsd4_cb_prepare,
 	.rpc_call_done = nfsd4_cb_recall_done,
 	.rpc_release = nfsd4_cb_recall_release,
 };
@@ -554,17 +726,24 @@ nfsd4_cb_recall(struct nfs4_delegation *dp)
 {
 	struct nfs4_client *clp = dp->dl_client;
 	struct rpc_clnt *clnt = clp->cl_cb_conn.cb_client;
+	struct nfs4_rpc_args *args;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_RECALL],
-		.rpc_argp = dp,
-		.rpc_cred = clp->cl_cb_conn.cb_cred
+		.rpc_cred = callback_cred
 	};
-	int status;
+	int status = -ENOMEM;
 
+	args = kzalloc(sizeof(*args), GFP_KERNEL);
+	if (!args)
+		goto out;
+	args->args_op = dp;
+	msg.rpc_argp = args;
 	dp->dl_retries = 1;
 	status = rpc_call_async(clnt, &msg, RPC_TASK_SOFT,
 				&nfsd4_cb_recall_ops, dp);
+out:
 	if (status) {
+		kfree(args);
 		put_nfs4_client(clp);
 		nfs4_put_delegation(dp);
 	}
diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c
index 5b398421b05..cdfa86fa147 100644
--- a/fs/nfsd/nfs4idmap.c
+++ b/fs/nfsd/nfs4idmap.c
@@ -146,6 +146,12 @@ idtoname_request(struct cache_detail *cd, struct cache_head *ch, char **bpp,
 }
 
 static int
+idtoname_upcall(struct cache_detail *cd, struct cache_head *ch)
+{
+	return sunrpc_cache_pipe_upcall(cd, ch, idtoname_request);
+}
+
+static int
 idtoname_match(struct cache_head *ca, struct cache_head *cb)
 {
 	struct ent *a = container_of(ca, struct ent, h);
@@ -175,10 +181,10 @@ idtoname_show(struct seq_file *m, struct cache_detail *cd, struct cache_head *h)
 }
 
 static void
-warn_no_idmapd(struct cache_detail *detail)
+warn_no_idmapd(struct cache_detail *detail, int has_died)
 {
 	printk("nfsd: nfsv4 idmapping failing: has idmapd %s?\n",
-			detail->last_close? "died" : "not been started");
+			has_died ? "died" : "not been started");
 }
 
 
@@ -192,7 +198,7 @@ static struct cache_detail idtoname_cache = {
 	.hash_table	= idtoname_table,
 	.name		= "nfs4.idtoname",
 	.cache_put	= ent_put,
-	.cache_request	= idtoname_request,
+	.cache_upcall	= idtoname_upcall,
 	.cache_parse	= idtoname_parse,
 	.cache_show	= idtoname_show,
 	.warn_no_listener = warn_no_idmapd,
@@ -325,6 +331,12 @@ nametoid_request(struct cache_detail *cd, struct cache_head *ch, char **bpp,
 }
 
 static int
+nametoid_upcall(struct cache_detail *cd, struct cache_head *ch)
+{
+	return sunrpc_cache_pipe_upcall(cd, ch, nametoid_request);
+}
+
+static int
 nametoid_match(struct cache_head *ca, struct cache_head *cb)
 {
 	struct ent *a = container_of(ca, struct ent, h);
@@ -363,7 +375,7 @@ static struct cache_detail nametoid_cache = {
 	.hash_table	= nametoid_table,
 	.name		= "nfs4.nametoid",
 	.cache_put	= ent_put,
-	.cache_request	= nametoid_request,
+	.cache_upcall	= nametoid_upcall,
 	.cache_parse	= nametoid_parse,
 	.cache_show	= nametoid_show,
 	.warn_no_listener = warn_no_idmapd,
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 7c8801769a3..bebc0c2e1b0 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -68,7 +68,6 @@ check_attr_support(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		   u32 *bmval, u32 *writable)
 {
 	struct dentry *dentry = cstate->current_fh.fh_dentry;
-	struct svc_export *exp = cstate->current_fh.fh_export;
 
 	/*
 	 * Check about attributes are supported by the NFSv4 server or not.
@@ -80,17 +79,13 @@ check_attr_support(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		return nfserr_attrnotsupp;
 
 	/*
-	 * Check FATTR4_WORD0_ACL & FATTR4_WORD0_FS_LOCATIONS can be supported
+	 * Check FATTR4_WORD0_ACL can be supported
 	 * in current environment or not.
 	 */
 	if (bmval[0] & FATTR4_WORD0_ACL) {
 		if (!IS_POSIXACL(dentry->d_inode))
 			return nfserr_attrnotsupp;
 	}
-	if (bmval[0] & FATTR4_WORD0_FS_LOCATIONS) {
-		if (exp->ex_fslocs.locations == NULL)
-			return nfserr_attrnotsupp;
-	}
 
 	/*
 	 * According to spec, read-only attributes return ERR_INVAL.
@@ -123,6 +118,35 @@ nfsd4_check_open_attributes(struct svc_rqst *rqstp,
 	return status;
 }
 
+static int
+is_create_with_attrs(struct nfsd4_open *open)
+{
+	return open->op_create == NFS4_OPEN_CREATE
+		&& (open->op_createmode == NFS4_CREATE_UNCHECKED
+		    || open->op_createmode == NFS4_CREATE_GUARDED
+		    || open->op_createmode == NFS4_CREATE_EXCLUSIVE4_1);
+}
+
+/*
+ * if error occurs when setting the acl, just clear the acl bit
+ * in the returned attr bitmap.
+ */
+static void
+do_set_nfs4_acl(struct svc_rqst *rqstp, struct svc_fh *fhp,
+		struct nfs4_acl *acl, u32 *bmval)
+{
+	__be32 status;
+
+	status = nfsd4_set_nfs4_acl(rqstp, fhp, acl);
+	if (status)
+		/*
+		 * We should probably fail the whole open at this point,
+		 * but we've already created the file, so it's too late;
+		 * So this seems the least of evils:
+		 */
+		bmval[0] &= ~FATTR4_WORD0_ACL;
+}
+
 static inline void
 fh_dup2(struct svc_fh *dst, struct svc_fh *src)
 {
@@ -206,6 +230,9 @@ do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_o
 	if (status)
 		goto out;
 
+	if (is_create_with_attrs(open) && open->op_acl != NULL)
+		do_set_nfs4_acl(rqstp, &resfh, open->op_acl, open->op_bmval);
+
 	set_change_info(&open->op_cinfo, current_fh);
 	fh_dup2(current_fh, &resfh);
 
@@ -536,12 +563,17 @@ nfsd4_create(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		status = nfserr_badtype;
 	}
 
-	if (!status) {
-		fh_unlock(&cstate->current_fh);
-		set_change_info(&create->cr_cinfo, &cstate->current_fh);
-		fh_dup2(&cstate->current_fh, &resfh);
-	}
+	if (status)
+		goto out;
+
+	if (create->cr_acl != NULL)
+		do_set_nfs4_acl(rqstp, &resfh, create->cr_acl,
+				create->cr_bmval);
 
+	fh_unlock(&cstate->current_fh);
+	set_change_info(&create->cr_cinfo, &cstate->current_fh);
+	fh_dup2(&cstate->current_fh, &resfh);
+out:
 	fh_put(&resfh);
 	return status;
 }
@@ -947,34 +979,6 @@ static struct nfsd4_operation nfsd4_ops[];
 static const char *nfsd4_op_name(unsigned opnum);
 
 /*
- * This is a replay of a compound for which no cache entry pages
- * were used. Encode the sequence operation, and if cachethis is FALSE
- * encode the uncache rep error on the next operation.
- */
-static __be32
-nfsd4_enc_uncached_replay(struct nfsd4_compoundargs *args,
-			 struct nfsd4_compoundres *resp)
-{
-	struct nfsd4_op *op;
-
-	dprintk("--> %s resp->opcnt %d ce_cachethis %u \n", __func__,
-		resp->opcnt, resp->cstate.slot->sl_cache_entry.ce_cachethis);
-
-	/* Encode the replayed sequence operation */
-	BUG_ON(resp->opcnt != 1);
-	op = &args->ops[resp->opcnt - 1];
-	nfsd4_encode_operation(resp, op);
-
-	/*return nfserr_retry_uncached_rep in next operation. */
-	if (resp->cstate.slot->sl_cache_entry.ce_cachethis == 0) {
-		op = &args->ops[resp->opcnt++];
-		op->status = nfserr_retry_uncached_rep;
-		nfsd4_encode_operation(resp, op);
-	}
-	return op->status;
-}
-
-/*
  * Enforce NFSv4.1 COMPOUND ordering rules.
  *
  * TODO:
@@ -1083,13 +1087,10 @@ nfsd4_proc_compound(struct svc_rqst *rqstp,
 			BUG_ON(op->status == nfs_ok);
 
 encode_op:
-		/* Only from SEQUENCE or CREATE_SESSION */
+		/* Only from SEQUENCE */
 		if (resp->cstate.status == nfserr_replay_cache) {
 			dprintk("%s NFS4.1 replay from cache\n", __func__);
-			if (nfsd4_not_cached(resp))
-				status = nfsd4_enc_uncached_replay(args, resp);
-			else
-				status = op->status;
+			status = op->status;
 			goto out;
 		}
 		if (op->status == nfserr_replay_me) {
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 980a216a48c..2153f9bdbeb 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -55,6 +55,7 @@
 #include <linux/lockd/bind.h>
 #include <linux/module.h>
 #include <linux/sunrpc/svcauth_gss.h>
+#include <linux/sunrpc/clnt.h>
 
 #define NFSDDBG_FACILITY                NFSDDBG_PROC
 
@@ -413,36 +414,65 @@ gen_sessionid(struct nfsd4_session *ses)
 }
 
 /*
- * Give the client the number of slots it requests bound by
- * NFSD_MAX_SLOTS_PER_SESSION and by sv_drc_max_pages.
+ * The protocol defines ca_maxresponssize_cached to include the size of
+ * the rpc header, but all we need to cache is the data starting after
+ * the end of the initial SEQUENCE operation--the rest we regenerate
+ * each time.  Therefore we can advertise a ca_maxresponssize_cached
+ * value that is the number of bytes in our cache plus a few additional
+ * bytes.  In order to stay on the safe side, and not promise more than
+ * we can cache, those additional bytes must be the minimum possible: 24
+ * bytes of rpc header (xid through accept state, with AUTH_NULL
+ * verifier), 12 for the compound header (with zero-length tag), and 44
+ * for the SEQUENCE op response:
+ */
+#define NFSD_MIN_HDR_SEQ_SZ  (24 + 12 + 44)
+
+/*
+ * Give the client the number of ca_maxresponsesize_cached slots it
+ * requests, of size bounded by NFSD_SLOT_CACHE_SIZE,
+ * NFSD_MAX_MEM_PER_SESSION, and nfsd_drc_max_mem. Do not allow more
+ * than NFSD_MAX_SLOTS_PER_SESSION.
  *
- * If we run out of pages (sv_drc_pages_used == sv_drc_max_pages) we
- * should (up to a point) re-negotiate active sessions and reduce their
- * slot usage to make rooom for new connections. For now we just fail the
- * create session.
+ * If we run out of reserved DRC memory we should (up to a point)
+ * re-negotiate active sessions and reduce their slot usage to make
+ * rooom for new connections. For now we just fail the create session.
  */
-static int set_forechannel_maxreqs(struct nfsd4_channel_attrs *fchan)
+static int set_forechannel_drc_size(struct nfsd4_channel_attrs *fchan)
 {
-	int status = 0, np = fchan->maxreqs * NFSD_PAGES_PER_SLOT;
+	int mem, size = fchan->maxresp_cached;
 
 	if (fchan->maxreqs < 1)
 		return nfserr_inval;
-	else if (fchan->maxreqs > NFSD_MAX_SLOTS_PER_SESSION)
-		fchan->maxreqs = NFSD_MAX_SLOTS_PER_SESSION;
 
-	spin_lock(&nfsd_serv->sv_lock);
-	if (np + nfsd_serv->sv_drc_pages_used > nfsd_serv->sv_drc_max_pages)
-		np = nfsd_serv->sv_drc_max_pages - nfsd_serv->sv_drc_pages_used;
-	nfsd_serv->sv_drc_pages_used += np;
-	spin_unlock(&nfsd_serv->sv_lock);
+	if (size < NFSD_MIN_HDR_SEQ_SZ)
+		size = NFSD_MIN_HDR_SEQ_SZ;
+	size -= NFSD_MIN_HDR_SEQ_SZ;
+	if (size > NFSD_SLOT_CACHE_SIZE)
+		size = NFSD_SLOT_CACHE_SIZE;
+
+	/* bound the maxreqs by NFSD_MAX_MEM_PER_SESSION */
+	mem = fchan->maxreqs * size;
+	if (mem > NFSD_MAX_MEM_PER_SESSION) {
+		fchan->maxreqs = NFSD_MAX_MEM_PER_SESSION / size;
+		if (fchan->maxreqs > NFSD_MAX_SLOTS_PER_SESSION)
+			fchan->maxreqs = NFSD_MAX_SLOTS_PER_SESSION;
+		mem = fchan->maxreqs * size;
+	}
 
-	if (np <= 0) {
-		status = nfserr_resource;
-		fchan->maxreqs = 0;
-	} else
-		fchan->maxreqs = np / NFSD_PAGES_PER_SLOT;
+	spin_lock(&nfsd_drc_lock);
+	/* bound the total session drc memory ussage */
+	if (mem + nfsd_drc_mem_used > nfsd_drc_max_mem) {
+		fchan->maxreqs = (nfsd_drc_max_mem - nfsd_drc_mem_used) / size;
+		mem = fchan->maxreqs * size;
+	}
+	nfsd_drc_mem_used += mem;
+	spin_unlock(&nfsd_drc_lock);
 
-	return status;
+	if (fchan->maxreqs == 0)
+		return nfserr_serverfault;
+
+	fchan->maxresp_cached = size + NFSD_MIN_HDR_SEQ_SZ;
+	return 0;
 }
 
 /*
@@ -466,36 +496,41 @@ static int init_forechannel_attrs(struct svc_rqst *rqstp,
 		fchan->maxresp_sz = maxcount;
 	session_fchan->maxresp_sz = fchan->maxresp_sz;
 
-	/* Set the max response cached size our default which is
-	 * a multiple of PAGE_SIZE and small */
-	session_fchan->maxresp_cached = NFSD_PAGES_PER_SLOT * PAGE_SIZE;
-	fchan->maxresp_cached = session_fchan->maxresp_cached;
-
 	/* Use the client's maxops if possible */
 	if (fchan->maxops > NFSD_MAX_OPS_PER_COMPOUND)
 		fchan->maxops = NFSD_MAX_OPS_PER_COMPOUND;
 	session_fchan->maxops = fchan->maxops;
 
-	/* try to use the client requested number of slots */
-	if (fchan->maxreqs > NFSD_MAX_SLOTS_PER_SESSION)
-		fchan->maxreqs = NFSD_MAX_SLOTS_PER_SESSION;
-
 	/* FIXME: Error means no more DRC pages so the server should
 	 * recover pages from existing sessions. For now fail session
 	 * creation.
 	 */
-	status = set_forechannel_maxreqs(fchan);
+	status = set_forechannel_drc_size(fchan);
 
+	session_fchan->maxresp_cached = fchan->maxresp_cached;
 	session_fchan->maxreqs = fchan->maxreqs;
+
+	dprintk("%s status %d\n", __func__, status);
 	return status;
 }
 
+static void
+free_session_slots(struct nfsd4_session *ses)
+{
+	int i;
+
+	for (i = 0; i < ses->se_fchannel.maxreqs; i++)
+		kfree(ses->se_slots[i]);
+}
+
 static int
 alloc_init_session(struct svc_rqst *rqstp, struct nfs4_client *clp,
 		   struct nfsd4_create_session *cses)
 {
 	struct nfsd4_session *new, tmp;
-	int idx, status = nfserr_resource, slotsize;
+	struct nfsd4_slot *sp;
+	int idx, slotsize, cachesize, i;
+	int status;
 
 	memset(&tmp, 0, sizeof(tmp));
 
@@ -506,14 +541,27 @@ alloc_init_session(struct svc_rqst *rqstp, struct nfs4_client *clp,
 	if (status)
 		goto out;
 
-	/* allocate struct nfsd4_session and slot table in one piece */
-	slotsize = tmp.se_fchannel.maxreqs * sizeof(struct nfsd4_slot);
+	BUILD_BUG_ON(NFSD_MAX_SLOTS_PER_SESSION * sizeof(struct nfsd4_slot)
+		     + sizeof(struct nfsd4_session) > PAGE_SIZE);
+
+	status = nfserr_serverfault;
+	/* allocate struct nfsd4_session and slot table pointers in one piece */
+	slotsize = tmp.se_fchannel.maxreqs * sizeof(struct nfsd4_slot *);
 	new = kzalloc(sizeof(*new) + slotsize, GFP_KERNEL);
 	if (!new)
 		goto out;
 
 	memcpy(new, &tmp, sizeof(*new));
 
+	/* allocate each struct nfsd4_slot and data cache in one piece */
+	cachesize = new->se_fchannel.maxresp_cached - NFSD_MIN_HDR_SEQ_SZ;
+	for (i = 0; i < new->se_fchannel.maxreqs; i++) {
+		sp = kzalloc(sizeof(*sp) + cachesize, GFP_KERNEL);
+		if (!sp)
+			goto out_free;
+		new->se_slots[i] = sp;
+	}
+
 	new->se_client = clp;
 	gen_sessionid(new);
 	idx = hash_sessionid(&new->se_sessionid);
@@ -530,6 +578,10 @@ alloc_init_session(struct svc_rqst *rqstp, struct nfs4_client *clp,
 	status = nfs_ok;
 out:
 	return status;
+out_free:
+	free_session_slots(new);
+	kfree(new);
+	goto out;
 }
 
 /* caller must hold sessionid_lock */
@@ -572,19 +624,16 @@ release_session(struct nfsd4_session *ses)
 	nfsd4_put_session(ses);
 }
 
-static void nfsd4_release_respages(struct page **respages, short resused);
-
 void
 free_session(struct kref *kref)
 {
 	struct nfsd4_session *ses;
-	int i;
 
 	ses = container_of(kref, struct nfsd4_session, se_ref);
-	for (i = 0; i < ses->se_fchannel.maxreqs; i++) {
-		struct nfsd4_cache_entry *e = &ses->se_slots[i].sl_cache_entry;
-		nfsd4_release_respages(e->ce_respages, e->ce_resused);
-	}
+	spin_lock(&nfsd_drc_lock);
+	nfsd_drc_mem_used -= ses->se_fchannel.maxreqs * NFSD_SLOT_CACHE_SIZE;
+	spin_unlock(&nfsd_drc_lock);
+	free_session_slots(ses);
 	kfree(ses);
 }
 
@@ -647,18 +696,14 @@ shutdown_callback_client(struct nfs4_client *clp)
 		clp->cl_cb_conn.cb_client = NULL;
 		rpc_shutdown_client(clnt);
 	}
-	if (clp->cl_cb_conn.cb_cred) {
-		put_rpccred(clp->cl_cb_conn.cb_cred);
-		clp->cl_cb_conn.cb_cred = NULL;
-	}
 }
 
 static inline void
 free_client(struct nfs4_client *clp)
 {
 	shutdown_callback_client(clp);
-	nfsd4_release_respages(clp->cl_slot.sl_cache_entry.ce_respages,
-			     clp->cl_slot.sl_cache_entry.ce_resused);
+	if (clp->cl_cb_xprt)
+		svc_xprt_put(clp->cl_cb_xprt);
 	if (clp->cl_cred.cr_group_info)
 		put_group_info(clp->cl_cred.cr_group_info);
 	kfree(clp->cl_principal);
@@ -714,25 +759,6 @@ expire_client(struct nfs4_client *clp)
 	put_nfs4_client(clp);
 }
 
-static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir)
-{
-	struct nfs4_client *clp;
-
-	clp = alloc_client(name);
-	if (clp == NULL)
-		return NULL;
-	memcpy(clp->cl_recdir, recdir, HEXDIR_LEN);
-	atomic_set(&clp->cl_count, 1);
-	atomic_set(&clp->cl_cb_conn.cb_set, 0);
-	INIT_LIST_HEAD(&clp->cl_idhash);
-	INIT_LIST_HEAD(&clp->cl_strhash);
-	INIT_LIST_HEAD(&clp->cl_openowners);
-	INIT_LIST_HEAD(&clp->cl_delegations);
-	INIT_LIST_HEAD(&clp->cl_sessions);
-	INIT_LIST_HEAD(&clp->cl_lru);
-	return clp;
-}
-
 static void copy_verf(struct nfs4_client *target, nfs4_verifier *source)
 {
 	memcpy(target->cl_verifier.data, source->data,
@@ -795,6 +821,46 @@ static void gen_confirm(struct nfs4_client *clp)
 	*p++ = i++;
 }
 
+static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir,
+		struct svc_rqst *rqstp, nfs4_verifier *verf)
+{
+	struct nfs4_client *clp;
+	struct sockaddr *sa = svc_addr(rqstp);
+	char *princ;
+
+	clp = alloc_client(name);
+	if (clp == NULL)
+		return NULL;
+
+	princ = svc_gss_principal(rqstp);
+	if (princ) {
+		clp->cl_principal = kstrdup(princ, GFP_KERNEL);
+		if (clp->cl_principal == NULL) {
+			free_client(clp);
+			return NULL;
+		}
+	}
+
+	memcpy(clp->cl_recdir, recdir, HEXDIR_LEN);
+	atomic_set(&clp->cl_count, 1);
+	atomic_set(&clp->cl_cb_conn.cb_set, 0);
+	INIT_LIST_HEAD(&clp->cl_idhash);
+	INIT_LIST_HEAD(&clp->cl_strhash);
+	INIT_LIST_HEAD(&clp->cl_openowners);
+	INIT_LIST_HEAD(&clp->cl_delegations);
+	INIT_LIST_HEAD(&clp->cl_sessions);
+	INIT_LIST_HEAD(&clp->cl_lru);
+	clear_bit(0, &clp->cl_cb_slot_busy);
+	rpc_init_wait_queue(&clp->cl_cb_waitq, "Backchannel slot table");
+	copy_verf(clp, verf);
+	rpc_copy_addr((struct sockaddr *) &clp->cl_addr, sa);
+	clp->cl_flavor = rqstp->rq_flavor;
+	copy_cred(&clp->cl_cred, &rqstp->rq_cred);
+	gen_confirm(clp);
+
+	return clp;
+}
+
 static int check_name(struct xdr_netobj name)
 {
 	if (name.len == 0) 
@@ -902,93 +968,40 @@ find_unconfirmed_client_by_str(const char *dname, unsigned int hashval,
 	return NULL;
 }
 
-/* a helper function for parse_callback */
-static int
-parse_octet(unsigned int *lenp, char **addrp)
-{
-	unsigned int len = *lenp;
-	char *p = *addrp;
-	int n = -1;
-	char c;
-
-	for (;;) {
-		if (!len)
-			break;
-		len--;
-		c = *p++;
-		if (c == '.')
-			break;
-		if ((c < '0') || (c > '9')) {
-			n = -1;
-			break;
-		}
-		if (n < 0)
-			n = 0;
-		n = (n * 10) + (c - '0');
-		if (n > 255) {
-			n = -1;
-			break;
-		}
-	}
-	*lenp = len;
-	*addrp = p;
-	return n;
-}
-
-/* parse and set the setclientid ipv4 callback address */
-static int
-parse_ipv4(unsigned int addr_len, char *addr_val, unsigned int *cbaddrp, unsigned short *cbportp)
-{
-	int temp = 0;
-	u32 cbaddr = 0;
-	u16 cbport = 0;
-	u32 addrlen = addr_len;
-	char *addr = addr_val;
-	int i, shift;
-
-	/* ipaddress */
-	shift = 24;
-	for(i = 4; i > 0  ; i--) {
-		if ((temp = parse_octet(&addrlen, &addr)) < 0) {
-			return 0;
-		}
-		cbaddr |= (temp << shift);
-		if (shift > 0)
-		shift -= 8;
-	}
-	*cbaddrp = cbaddr;
-
-	/* port */
-	shift = 8;
-	for(i = 2; i > 0  ; i--) {
-		if ((temp = parse_octet(&addrlen, &addr)) < 0) {
-			return 0;
-		}
-		cbport |= (temp << shift);
-		if (shift > 0)
-			shift -= 8;
-	}
-	*cbportp = cbport;
-	return 1;
-}
-
 static void
-gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se)
+gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se, u32 scopeid)
 {
 	struct nfs4_cb_conn *cb = &clp->cl_cb_conn;
-
-	/* Currently, we only support tcp for the callback channel */
-	if ((se->se_callback_netid_len != 3) || memcmp((char *)se->se_callback_netid_val, "tcp", 3))
+	unsigned short expected_family;
+
+	/* Currently, we only support tcp and tcp6 for the callback channel */
+	if (se->se_callback_netid_len == 3 &&
+	    !memcmp(se->se_callback_netid_val, "tcp", 3))
+		expected_family = AF_INET;
+	else if (se->se_callback_netid_len == 4 &&
+		 !memcmp(se->se_callback_netid_val, "tcp6", 4))
+		expected_family = AF_INET6;
+	else
 		goto out_err;
 
-	if ( !(parse_ipv4(se->se_callback_addr_len, se->se_callback_addr_val,
-	                 &cb->cb_addr, &cb->cb_port)))
+	cb->cb_addrlen = rpc_uaddr2sockaddr(se->se_callback_addr_val,
+					    se->se_callback_addr_len,
+					    (struct sockaddr *) &cb->cb_addr,
+					    sizeof(cb->cb_addr));
+
+	if (!cb->cb_addrlen || cb->cb_addr.ss_family != expected_family)
 		goto out_err;
+
+	if (cb->cb_addr.ss_family == AF_INET6)
+		((struct sockaddr_in6 *) &cb->cb_addr)->sin6_scope_id = scopeid;
+
 	cb->cb_minorversion = 0;
 	cb->cb_prog = se->se_callback_prog;
 	cb->cb_ident = se->se_callback_ident;
 	return;
 out_err:
+	cb->cb_addr.ss_family = AF_UNSPEC;
+	cb->cb_addrlen = 0;
 	dprintk(KERN_INFO "NFSD: this client (clientid %08x/%08x) "
 		"will not receive delegations\n",
 		clp->cl_clientid.cl_boot, clp->cl_clientid.cl_id);
@@ -996,175 +1009,87 @@ out_err:
 	return;
 }
 
-void
-nfsd4_set_statp(struct svc_rqst *rqstp, __be32 *statp)
-{
-	struct nfsd4_compoundres *resp = rqstp->rq_resp;
-
-	resp->cstate.statp = statp;
-}
-
 /*
- * Dereference the result pages.
+ * Cache a reply. nfsd4_check_drc_limit() has bounded the cache size.
  */
-static void
-nfsd4_release_respages(struct page **respages, short resused)
+void
+nfsd4_store_cache_entry(struct nfsd4_compoundres *resp)
 {
-	int i;
+	struct nfsd4_slot *slot = resp->cstate.slot;
+	unsigned int base;
 
-	dprintk("--> %s\n", __func__);
-	for (i = 0; i < resused; i++) {
-		if (!respages[i])
-			continue;
-		put_page(respages[i]);
-		respages[i] = NULL;
-	}
-}
+	dprintk("--> %s slot %p\n", __func__, slot);
 
-static void
-nfsd4_copy_pages(struct page **topages, struct page **frompages, short count)
-{
-	int i;
+	slot->sl_opcnt = resp->opcnt;
+	slot->sl_status = resp->cstate.status;
 
-	for (i = 0; i < count; i++) {
-		topages[i] = frompages[i];
-		if (!topages[i])
-			continue;
-		get_page(topages[i]);
+	if (nfsd4_not_cached(resp)) {
+		slot->sl_datalen = 0;
+		return;
 	}
+	slot->sl_datalen = (char *)resp->p - (char *)resp->cstate.datap;
+	base = (char *)resp->cstate.datap -
+					(char *)resp->xbuf->head[0].iov_base;
+	if (read_bytes_from_xdr_buf(resp->xbuf, base, slot->sl_data,
+				    slot->sl_datalen))
+		WARN("%s: sessions DRC could not cache compound\n", __func__);
+	return;
 }
 
 /*
- * Cache the reply pages up to NFSD_PAGES_PER_SLOT + 1, clearing the previous
- * pages. We add a page to NFSD_PAGES_PER_SLOT for the case where the total
- * length of the XDR response is less than se_fmaxresp_cached
- * (NFSD_PAGES_PER_SLOT * PAGE_SIZE) but the xdr_buf pages is used for a
- * of the reply (e.g. readdir).
+ * Encode the replay sequence operation from the slot values.
+ * If cachethis is FALSE encode the uncached rep error on the next
+ * operation which sets resp->p and increments resp->opcnt for
+ * nfs4svc_encode_compoundres.
  *
- * Store the base and length of the rq_req.head[0] page
- * of the NFSv4.1 data, just past the rpc header.
  */
-void
-nfsd4_store_cache_entry(struct nfsd4_compoundres *resp)
+static __be32
+nfsd4_enc_sequence_replay(struct nfsd4_compoundargs *args,
+			  struct nfsd4_compoundres *resp)
 {
-	struct nfsd4_cache_entry *entry = &resp->cstate.slot->sl_cache_entry;
-	struct svc_rqst *rqstp = resp->rqstp;
-	struct nfsd4_compoundargs *args = rqstp->rq_argp;
-	struct nfsd4_op *op = &args->ops[resp->opcnt];
-	struct kvec *resv = &rqstp->rq_res.head[0];
-
-	dprintk("--> %s entry %p\n", __func__, entry);
-
-	/* Don't cache a failed OP_SEQUENCE. */
-	if (resp->opcnt == 1 && op->opnum == OP_SEQUENCE && resp->cstate.status)
-		return;
+	struct nfsd4_op *op;
+	struct nfsd4_slot *slot = resp->cstate.slot;
 
-	nfsd4_release_respages(entry->ce_respages, entry->ce_resused);
-	entry->ce_opcnt = resp->opcnt;
-	entry->ce_status = resp->cstate.status;
+	dprintk("--> %s resp->opcnt %d cachethis %u \n", __func__,
+		resp->opcnt, resp->cstate.slot->sl_cachethis);
 
-	/*
-	 * Don't need a page to cache just the sequence operation - the slot
-	 * does this for us!
-	 */
+	/* Encode the replayed sequence operation */
+	op = &args->ops[resp->opcnt - 1];
+	nfsd4_encode_operation(resp, op);
 
-	if (nfsd4_not_cached(resp)) {
-		entry->ce_resused = 0;
-		entry->ce_rpchdrlen = 0;
-		dprintk("%s Just cache SEQUENCE. ce_cachethis %d\n", __func__,
-			resp->cstate.slot->sl_cache_entry.ce_cachethis);
-		return;
-	}
-	entry->ce_resused = rqstp->rq_resused;
-	if (entry->ce_resused > NFSD_PAGES_PER_SLOT + 1)
-		entry->ce_resused = NFSD_PAGES_PER_SLOT + 1;
-	nfsd4_copy_pages(entry->ce_respages, rqstp->rq_respages,
-			 entry->ce_resused);
-	entry->ce_datav.iov_base = resp->cstate.statp;
-	entry->ce_datav.iov_len = resv->iov_len - ((char *)resp->cstate.statp -
-				(char *)page_address(rqstp->rq_respages[0]));
-	/* Current request rpc header length*/
-	entry->ce_rpchdrlen = (char *)resp->cstate.statp -
-				(char *)page_address(rqstp->rq_respages[0]);
-}
-
-/*
- * We keep the rpc header, but take the nfs reply from the replycache.
- */
-static int
-nfsd41_copy_replay_data(struct nfsd4_compoundres *resp,
-			struct nfsd4_cache_entry *entry)
-{
-	struct svc_rqst *rqstp = resp->rqstp;
-	struct kvec *resv = &resp->rqstp->rq_res.head[0];
-	int len;
-
-	/* Current request rpc header length*/
-	len = (char *)resp->cstate.statp -
-			(char *)page_address(rqstp->rq_respages[0]);
-	if (entry->ce_datav.iov_len + len > PAGE_SIZE) {
-		dprintk("%s v41 cached reply too large (%Zd).\n", __func__,
-			entry->ce_datav.iov_len);
-		return 0;
+	/* Return nfserr_retry_uncached_rep in next operation. */
+	if (args->opcnt > 1 && slot->sl_cachethis == 0) {
+		op = &args->ops[resp->opcnt++];
+		op->status = nfserr_retry_uncached_rep;
+		nfsd4_encode_operation(resp, op);
 	}
-	/* copy the cached reply nfsd data past the current rpc header */
-	memcpy((char *)resv->iov_base + len, entry->ce_datav.iov_base,
-		entry->ce_datav.iov_len);
-	resv->iov_len = len + entry->ce_datav.iov_len;
-	return 1;
+	return op->status;
 }
 
 /*
- * Keep the first page of the replay. Copy the NFSv4.1 data from the first
- * cached page.  Replace any futher replay pages from the cache.
+ * The sequence operation is not cached because we can use the slot and
+ * session values.
  */
 __be32
 nfsd4_replay_cache_entry(struct nfsd4_compoundres *resp,
 			 struct nfsd4_sequence *seq)
 {
-	struct nfsd4_cache_entry *entry = &resp->cstate.slot->sl_cache_entry;
+	struct nfsd4_slot *slot = resp->cstate.slot;
 	__be32 status;
 
-	dprintk("--> %s entry %p\n", __func__, entry);
-
-	/*
-	 * If this is just the sequence operation, we did not keep
-	 * a page in the cache entry because we can just use the
-	 * slot info stored in struct nfsd4_sequence that was checked
-	 * against the slot in nfsd4_sequence().
-	 *
-	 * This occurs when seq->cachethis is FALSE, or when the client
-	 * session inactivity timer fires and a solo sequence operation
-	 * is sent (lease renewal).
-	 */
-	if (seq && nfsd4_not_cached(resp)) {
-		seq->maxslots = resp->cstate.session->se_fchannel.maxreqs;
-		return nfs_ok;
-	}
-
-	if (!nfsd41_copy_replay_data(resp, entry)) {
-		/*
-		 * Not enough room to use the replay rpc header, send the
-		 * cached header. Release all the allocated result pages.
-		 */
-		svc_free_res_pages(resp->rqstp);
-		nfsd4_copy_pages(resp->rqstp->rq_respages, entry->ce_respages,
-			entry->ce_resused);
-	} else {
-		/* Release all but the first allocated result page */
+	dprintk("--> %s slot %p\n", __func__, slot);
 
-		resp->rqstp->rq_resused--;
-		svc_free_res_pages(resp->rqstp);
+	/* Either returns 0 or nfserr_retry_uncached */
+	status = nfsd4_enc_sequence_replay(resp->rqstp->rq_argp, resp);
+	if (status == nfserr_retry_uncached_rep)
+		return status;
 
-		nfsd4_copy_pages(&resp->rqstp->rq_respages[1],
-				 &entry->ce_respages[1],
-				 entry->ce_resused - 1);
-	}
+	/* The sequence operation has been encoded, cstate->datap set. */
+	memcpy(resp->cstate.datap, slot->sl_data, slot->sl_datalen);
 
-	resp->rqstp->rq_resused = entry->ce_resused;
-	resp->opcnt = entry->ce_opcnt;
-	resp->cstate.iovlen = entry->ce_datav.iov_len + entry->ce_rpchdrlen;
-	status = entry->ce_status;
+	resp->opcnt = slot->sl_opcnt;
+	resp->p = resp->cstate.datap + XDR_QUADLEN(slot->sl_datalen);
+	status = slot->sl_status;
 
 	return status;
 }
@@ -1194,13 +1119,15 @@ nfsd4_exchange_id(struct svc_rqst *rqstp,
 	int status;
 	unsigned int		strhashval;
 	char			dname[HEXDIR_LEN];
+	char			addr_str[INET6_ADDRSTRLEN];
 	nfs4_verifier		verf = exid->verifier;
-	u32			ip_addr = svc_addr_in(rqstp)->sin_addr.s_addr;
+	struct sockaddr		*sa = svc_addr(rqstp);
 
+	rpc_ntop(sa, addr_str, sizeof(addr_str));
 	dprintk("%s rqstp=%p exid=%p clname.len=%u clname.data=%p "
-		" ip_addr=%u flags %x, spa_how %d\n",
+		"ip_addr=%s flags %x, spa_how %d\n",
 		__func__, rqstp, exid, exid->clname.len, exid->clname.data,
-		ip_addr, exid->flags, exid->spa_how);
+		addr_str, exid->flags, exid->spa_how);
 
 	if (!check_name(exid->clname) || (exid->flags & ~EXCHGID4_FLAG_MASK_A))
 		return nfserr_inval;
@@ -1281,28 +1208,23 @@ nfsd4_exchange_id(struct svc_rqst *rqstp,
 
 out_new:
 	/* Normal case */
-	new = create_client(exid->clname, dname);
+	new = create_client(exid->clname, dname, rqstp, &verf);
 	if (new == NULL) {
-		status = nfserr_resource;
+		status = nfserr_serverfault;
 		goto out;
 	}
 
-	copy_verf(new, &verf);
-	copy_cred(&new->cl_cred, &rqstp->rq_cred);
-	new->cl_addr = ip_addr;
 	gen_clid(new);
-	gen_confirm(new);
 	add_to_unconfirmed(new, strhashval);
 out_copy:
 	exid->clientid.cl_boot = new->cl_clientid.cl_boot;
 	exid->clientid.cl_id = new->cl_clientid.cl_id;
 
-	new->cl_slot.sl_seqid = 0;
 	exid->seqid = 1;
 	nfsd4_set_ex_flags(new, exid);
 
 	dprintk("nfsd4_exchange_id seqid %d flags %x\n",
-		new->cl_slot.sl_seqid, new->cl_exchange_flags);
+		new->cl_cs_slot.sl_seqid, new->cl_exchange_flags);
 	status = nfs_ok;
 
 out:
@@ -1313,40 +1235,60 @@ error:
 }
 
 static int
-check_slot_seqid(u32 seqid, struct nfsd4_slot *slot)
+check_slot_seqid(u32 seqid, u32 slot_seqid, int slot_inuse)
 {
-	dprintk("%s enter. seqid %d slot->sl_seqid %d\n", __func__, seqid,
-		slot->sl_seqid);
+	dprintk("%s enter. seqid %d slot_seqid %d\n", __func__, seqid,
+		slot_seqid);
 
 	/* The slot is in use, and no response has been sent. */
-	if (slot->sl_inuse) {
-		if (seqid == slot->sl_seqid)
+	if (slot_inuse) {
+		if (seqid == slot_seqid)
 			return nfserr_jukebox;
 		else
 			return nfserr_seq_misordered;
 	}
 	/* Normal */
-	if (likely(seqid == slot->sl_seqid + 1))
+	if (likely(seqid == slot_seqid + 1))
 		return nfs_ok;
 	/* Replay */
-	if (seqid == slot->sl_seqid)
+	if (seqid == slot_seqid)
 		return nfserr_replay_cache;
 	/* Wraparound */
-	if (seqid == 1 && (slot->sl_seqid + 1) == 0)
+	if (seqid == 1 && (slot_seqid + 1) == 0)
 		return nfs_ok;
 	/* Misordered replay or misordered new request */
 	return nfserr_seq_misordered;
 }
 
+/*
+ * Cache the create session result into the create session single DRC
+ * slot cache by saving the xdr structure. sl_seqid has been set.
+ * Do this for solo or embedded create session operations.
+ */
+static void
+nfsd4_cache_create_session(struct nfsd4_create_session *cr_ses,
+			   struct nfsd4_clid_slot *slot, int nfserr)
+{
+	slot->sl_status = nfserr;
+	memcpy(&slot->sl_cr_ses, cr_ses, sizeof(*cr_ses));
+}
+
+static __be32
+nfsd4_replay_create_session(struct nfsd4_create_session *cr_ses,
+			    struct nfsd4_clid_slot *slot)
+{
+	memcpy(cr_ses, &slot->sl_cr_ses, sizeof(*cr_ses));
+	return slot->sl_status;
+}
+
 __be32
 nfsd4_create_session(struct svc_rqst *rqstp,
 		     struct nfsd4_compound_state *cstate,
 		     struct nfsd4_create_session *cr_ses)
 {
-	u32 ip_addr = svc_addr_in(rqstp)->sin_addr.s_addr;
-	struct nfsd4_compoundres *resp = rqstp->rq_resp;
+	struct sockaddr *sa = svc_addr(rqstp);
 	struct nfs4_client *conf, *unconf;
-	struct nfsd4_slot *slot = NULL;
+	struct nfsd4_clid_slot *cs_slot = NULL;
 	int status = 0;
 
 	nfs4_lock_state();
@@ -1354,40 +1296,38 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 	conf = find_confirmed_client(&cr_ses->clientid);
 
 	if (conf) {
-		slot = &conf->cl_slot;
-		status = check_slot_seqid(cr_ses->seqid, slot);
+		cs_slot = &conf->cl_cs_slot;
+		status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0);
 		if (status == nfserr_replay_cache) {
 			dprintk("Got a create_session replay! seqid= %d\n",
-				slot->sl_seqid);
-			cstate->slot = slot;
-			cstate->status = status;
+				cs_slot->sl_seqid);
 			/* Return the cached reply status */
-			status = nfsd4_replay_cache_entry(resp, NULL);
+			status = nfsd4_replay_create_session(cr_ses, cs_slot);
 			goto out;
-		} else if (cr_ses->seqid != conf->cl_slot.sl_seqid + 1) {
+		} else if (cr_ses->seqid != cs_slot->sl_seqid + 1) {
 			status = nfserr_seq_misordered;
 			dprintk("Sequence misordered!\n");
 			dprintk("Expected seqid= %d but got seqid= %d\n",
-				slot->sl_seqid, cr_ses->seqid);
+				cs_slot->sl_seqid, cr_ses->seqid);
 			goto out;
 		}
-		conf->cl_slot.sl_seqid++;
+		cs_slot->sl_seqid++;
 	} else if (unconf) {
 		if (!same_creds(&unconf->cl_cred, &rqstp->rq_cred) ||
-		    (ip_addr != unconf->cl_addr)) {
+		    !rpc_cmp_addr(sa, (struct sockaddr *) &unconf->cl_addr)) {
 			status = nfserr_clid_inuse;
 			goto out;
 		}
 
-		slot = &unconf->cl_slot;
-		status = check_slot_seqid(cr_ses->seqid, slot);
+		cs_slot = &unconf->cl_cs_slot;
+		status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0);
 		if (status) {
 			/* an unconfirmed replay returns misordered */
 			status = nfserr_seq_misordered;
-			goto out;
+			goto out_cache;
 		}
 
-		slot->sl_seqid++; /* from 0 to 1 */
+		cs_slot->sl_seqid++; /* from 0 to 1 */
 		move_to_confirmed(unconf);
 
 		/*
@@ -1396,6 +1336,19 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 		cr_ses->flags &= ~SESSION4_PERSIST;
 		cr_ses->flags &= ~SESSION4_RDMA;
 
+		if (cr_ses->flags & SESSION4_BACK_CHAN) {
+			unconf->cl_cb_xprt = rqstp->rq_xprt;
+			svc_xprt_get(unconf->cl_cb_xprt);
+			rpc_copy_addr(
+				(struct sockaddr *)&unconf->cl_cb_conn.cb_addr,
+				sa);
+			unconf->cl_cb_conn.cb_addrlen = svc_addr_len(sa);
+			unconf->cl_cb_conn.cb_minorversion =
+				cstate->minorversion;
+			unconf->cl_cb_conn.cb_prog = cr_ses->callback_prog;
+			unconf->cl_cb_seq_nr = 1;
+			nfsd4_probe_callback(unconf);
+		}
 		conf = unconf;
 	} else {
 		status = nfserr_stale_clientid;
@@ -1408,12 +1361,11 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 
 	memcpy(cr_ses->sessionid.data, conf->cl_sessionid.data,
 	       NFS4_MAX_SESSIONID_LEN);
-	cr_ses->seqid = slot->sl_seqid;
+	cr_ses->seqid = cs_slot->sl_seqid;
 
-	slot->sl_inuse = true;
-	cstate->slot = slot;
-	/* Ensure a page is used for the cache */
-	slot->sl_cache_entry.ce_cachethis = 1;
+out_cache:
+	/* cache solo and embedded create sessions under the state lock */
+	nfsd4_cache_create_session(cr_ses, cs_slot, status);
 out:
 	nfs4_unlock_state();
 	dprintk("%s returns %d\n", __func__, ntohl(status));
@@ -1478,18 +1430,23 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 	if (seq->slotid >= session->se_fchannel.maxreqs)
 		goto out;
 
-	slot = &session->se_slots[seq->slotid];
+	slot = session->se_slots[seq->slotid];
 	dprintk("%s: slotid %d\n", __func__, seq->slotid);
 
-	status = check_slot_seqid(seq->seqid, slot);
+	/* We do not negotiate the number of slots yet, so set the
+	 * maxslots to the session maxreqs which is used to encode
+	 * sr_highest_slotid and the sr_target_slot id to maxslots */
+	seq->maxslots = session->se_fchannel.maxreqs;
+
+	status = check_slot_seqid(seq->seqid, slot->sl_seqid, slot->sl_inuse);
 	if (status == nfserr_replay_cache) {
 		cstate->slot = slot;
 		cstate->session = session;
 		/* Return the cached reply status and set cstate->status
-		 * for nfsd4_svc_encode_compoundres processing */
+		 * for nfsd4_proc_compound processing */
 		status = nfsd4_replay_cache_entry(resp, seq);
 		cstate->status = nfserr_replay_cache;
-		goto replay_cache;
+		goto out;
 	}
 	if (status)
 		goto out;
@@ -1497,23 +1454,23 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 	/* Success! bump slot seqid */
 	slot->sl_inuse = true;
 	slot->sl_seqid = seq->seqid;
-	slot->sl_cache_entry.ce_cachethis = seq->cachethis;
-	/* Always set the cache entry cachethis for solo sequence */
-	if (nfsd4_is_solo_sequence(resp))
-		slot->sl_cache_entry.ce_cachethis = 1;
+	slot->sl_cachethis = seq->cachethis;
 
 	cstate->slot = slot;
 	cstate->session = session;
 
-replay_cache:
-	/* Renew the clientid on success and on replay.
-	 * Hold a session reference until done processing the compound:
+	/* Hold a session reference until done processing the compound:
 	 * nfsd4_put_session called only if the cstate slot is set.
 	 */
-	renew_client(session->se_client);
 	nfsd4_get_session(session);
 out:
 	spin_unlock(&sessionid_lock);
+	/* Renew the clientid on success and on replay */
+	if (cstate->session) {
+		nfs4_lock_state();
+		renew_client(session->se_client);
+		nfs4_unlock_state();
+	}
 	dprintk("%s: return %d\n", __func__, ntohl(status));
 	return status;
 }
@@ -1522,7 +1479,7 @@ __be32
 nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		  struct nfsd4_setclientid *setclid)
 {
-	struct sockaddr_in	*sin = svc_addr_in(rqstp);
+	struct sockaddr		*sa = svc_addr(rqstp);
 	struct xdr_netobj 	clname = { 
 		.len = setclid->se_namelen,
 		.data = setclid->se_name,
@@ -1531,7 +1488,6 @@ 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))
@@ -1554,8 +1510,11 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		/* RFC 3530 14.2.33 CASE 0: */
 		status = nfserr_clid_inuse;
 		if (!same_creds(&conf->cl_cred, &rqstp->rq_cred)) {
-			dprintk("NFSD: setclientid: string in use by client"
-				" at %pI4\n", &conf->cl_addr);
+			char addr_str[INET6_ADDRSTRLEN];
+			rpc_ntop((struct sockaddr *) &conf->cl_addr, addr_str,
+				 sizeof(addr_str));
+			dprintk("NFSD: setclientid: string in use by client "
+				"at %s\n", addr_str);
 			goto out;
 		}
 	}
@@ -1573,7 +1532,7 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		 */
 		if (unconf)
 			expire_client(unconf);
-		new = create_client(clname, dname);
+		new = create_client(clname, dname, rqstp, &clverifier);
 		if (new == NULL)
 			goto out;
 		gen_clid(new);
@@ -1590,7 +1549,7 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 			 */
 			expire_client(unconf);
 		}
-		new = create_client(clname, dname);
+		new = create_client(clname, dname, rqstp, &clverifier);
 		if (new == NULL)
 			goto out;
 		copy_clid(new, conf);
@@ -1600,7 +1559,7 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		 * probable client reboot; state will be removed if
 		 * confirmed.
 		 */
-		new = create_client(clname, dname);
+		new = create_client(clname, dname, rqstp, &clverifier);
 		if (new == NULL)
 			goto out;
 		gen_clid(new);
@@ -1611,25 +1570,12 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 		 * confirmed.
 		 */
 		expire_client(unconf);
-		new = create_client(clname, dname);
+		new = create_client(clname, dname, rqstp, &clverifier);
 		if (new == NULL)
 			goto out;
 		gen_clid(new);
 	}
-	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);
+	gen_callback(new, setclid, rpc_get_scope_id(sa));
 	add_to_unconfirmed(new, strhashval);
 	setclid->se_clientid.cl_boot = new->cl_clientid.cl_boot;
 	setclid->se_clientid.cl_id = new->cl_clientid.cl_id;
@@ -1651,7 +1597,7 @@ nfsd4_setclientid_confirm(struct svc_rqst *rqstp,
 			 struct nfsd4_compound_state *cstate,
 			 struct nfsd4_setclientid_confirm *setclientid_confirm)
 {
-	struct sockaddr_in *sin = svc_addr_in(rqstp);
+	struct sockaddr *sa = svc_addr(rqstp);
 	struct nfs4_client *conf, *unconf;
 	nfs4_verifier confirm = setclientid_confirm->sc_confirm; 
 	clientid_t * clid = &setclientid_confirm->sc_clientid;
@@ -1670,9 +1616,9 @@ nfsd4_setclientid_confirm(struct svc_rqst *rqstp,
 	unconf = find_unconfirmed_client(clid);
 
 	status = nfserr_clid_inuse;
-	if (conf && conf->cl_addr != sin->sin_addr.s_addr)
+	if (conf && !rpc_cmp_addr((struct sockaddr *) &conf->cl_addr, sa))
 		goto out;
-	if (unconf && unconf->cl_addr != sin->sin_addr.s_addr)
+	if (unconf && !rpc_cmp_addr((struct sockaddr *) &unconf->cl_addr, sa))
 		goto out;
 
 	/*
@@ -2163,7 +2109,7 @@ int nfsd_change_deleg_cb(struct file_lock **onlist, int arg)
 		return -EAGAIN;
 }
 
-static struct lock_manager_operations nfsd_lease_mng_ops = {
+static const struct lock_manager_operations nfsd_lease_mng_ops = {
 	.fl_break = nfsd_break_deleg_cb,
 	.fl_release_private = nfsd_release_deleg_cb,
 	.fl_copy_lock = nfsd_copy_lock_deleg_cb,
@@ -3368,7 +3314,7 @@ nfs4_transform_lock_offset(struct file_lock *lock)
 
 /* Hack!: For now, we're defining this just so we can use a pointer to it
  * as a unique cookie to identify our (NFSv4's) posix locks. */
-static struct lock_manager_operations nfsd_posix_mng_ops  = {
+static const struct lock_manager_operations nfsd_posix_mng_ops  = {
 };
 
 static inline void
@@ -4072,7 +4018,7 @@ set_max_delegations(void)
 
 /* initialization to perform when the nfsd service is started: */
 
-static void
+static int
 __nfs4_state_start(void)
 {
 	unsigned long grace_time;
@@ -4084,19 +4030,26 @@ __nfs4_state_start(void)
 	printk(KERN_INFO "NFSD: starting %ld-second grace period\n",
 	       grace_time/HZ);
 	laundry_wq = create_singlethread_workqueue("nfsd4");
+	if (laundry_wq == NULL)
+		return -ENOMEM;
 	queue_delayed_work(laundry_wq, &laundromat_work, grace_time);
 	set_max_delegations();
+	return set_callback_cred();
 }
 
-void
+int
 nfs4_state_start(void)
 {
+	int ret;
+
 	if (nfs4_init)
-		return;
+		return 0;
 	nfsd4_load_reboot_recovery_data();
-	__nfs4_state_start();
+	ret = __nfs4_state_start();
+	if (ret)
+		return ret;
 	nfs4_init = 1;
-	return;
+	return 0;
 }
 
 time_t
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index 2dcc7feaa6f..0fbd50cee1f 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -1599,7 +1599,8 @@ static __be32 nfsd4_encode_fs_location4(struct nfsd4_fs_location *location,
 static char *nfsd4_path(struct svc_rqst *rqstp, struct svc_export *exp, __be32 *stat)
 {
 	struct svc_fh tmp_fh;
-	char *path, *rootpath;
+	char *path = NULL, *rootpath;
+	size_t rootlen;
 
 	fh_init(&tmp_fh, NFS4_FHSIZE);
 	*stat = exp_pseudoroot(rqstp, &tmp_fh);
@@ -1609,14 +1610,18 @@ static char *nfsd4_path(struct svc_rqst *rqstp, struct svc_export *exp, __be32 *
 
 	path = exp->ex_pathname;
 
-	if (strncmp(path, rootpath, strlen(rootpath))) {
+	rootlen = strlen(rootpath);
+	if (strncmp(path, rootpath, rootlen)) {
 		dprintk("nfsd: fs_locations failed;"
 			"%s is not contained in %s\n", path, rootpath);
 		*stat = nfserr_notsupp;
-		return NULL;
+		path = NULL;
+		goto out;
 	}
-
-	return path + strlen(rootpath);
+	path += rootlen;
+out:
+	fh_put(&tmp_fh);
+	return path;
 }
 
 /*
@@ -1793,11 +1798,6 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp,
 				goto out_nfserr;
 		}
 	}
-	if (bmval0 & FATTR4_WORD0_FS_LOCATIONS) {
-		if (exp->ex_fslocs.locations == NULL) {
-			bmval0 &= ~FATTR4_WORD0_FS_LOCATIONS;
-		}
-	}
 	if ((buflen -= 16) < 0)
 		goto out_resource;
 
@@ -1825,8 +1825,6 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp,
 			goto out_resource;
 		if (!aclsupport)
 			word0 &= ~FATTR4_WORD0_ACL;
-		if (!exp->ex_fslocs.locations)
-			word0 &= ~FATTR4_WORD0_FS_LOCATIONS;
 		if (!word2) {
 			WRITE32(2);
 			WRITE32(word0);
@@ -3064,6 +3062,7 @@ nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr,
 	WRITE32(0);
 
 	ADJUST_ARGS();
+	resp->cstate.datap = p; /* DRC cache data pointer */
 	return 0;
 }
 
@@ -3166,7 +3165,7 @@ static int nfsd4_check_drc_limit(struct nfsd4_compoundres *resp)
 		return status;
 
 	session = resp->cstate.session;
-	if (session == NULL || slot->sl_cache_entry.ce_cachethis == 0)
+	if (session == NULL || slot->sl_cachethis == 0)
 		return status;
 
 	if (resp->opcnt >= args->opcnt)
@@ -3291,6 +3290,7 @@ nfs4svc_encode_compoundres(struct svc_rqst *rqstp, __be32 *p, struct nfsd4_compo
 	/*
 	 * All that remains is to write the tag and operation count...
 	 */
+	struct nfsd4_compound_state *cs = &resp->cstate;
 	struct kvec *iov;
 	p = resp->tagp;
 	*p++ = htonl(resp->taglen);
@@ -3304,17 +3304,11 @@ nfs4svc_encode_compoundres(struct svc_rqst *rqstp, __be32 *p, struct nfsd4_compo
 		iov = &rqstp->rq_res.head[0];
 	iov->iov_len = ((char*)resp->p) - (char*)iov->iov_base;
 	BUG_ON(iov->iov_len > PAGE_SIZE);
-	if (nfsd4_has_session(&resp->cstate)) {
-		if (resp->cstate.status == nfserr_replay_cache &&
-				!nfsd4_not_cached(resp)) {
-			iov->iov_len = resp->cstate.iovlen;
-		} else {
-			nfsd4_store_cache_entry(resp);
-			dprintk("%s: SET SLOT STATE TO AVAILABLE\n", __func__);
-			resp->cstate.slot->sl_inuse = 0;
-		}
-		if (resp->cstate.session)
-			nfsd4_put_session(resp->cstate.session);
+	if (nfsd4_has_session(cs) && cs->status != nfserr_replay_cache) {
+		nfsd4_store_cache_entry(resp);
+		dprintk("%s: SET SLOT STATE TO AVAILABLE\n", __func__);
+		resp->cstate.slot->sl_inuse = false;
+		nfsd4_put_session(resp->cstate.session);
 	}
 	return 1;
 }
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index 6d0847562d8..00388d2a3c9 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -37,6 +37,7 @@
 #include <linux/nfsd/xdr.h>
 #include <linux/nfsd/syscall.h>
 #include <linux/lockd/lockd.h>
+#include <linux/sunrpc/clnt.h>
 
 #include <asm/uaccess.h>
 #include <net/ipv6.h>
@@ -173,12 +174,13 @@ static const struct file_operations exports_operations = {
 };
 
 extern int nfsd_pool_stats_open(struct inode *inode, struct file *file);
+extern int nfsd_pool_stats_release(struct inode *inode, struct file *file);
 
 static struct file_operations pool_stats_operations = {
 	.open		= nfsd_pool_stats_open,
 	.read		= seq_read,
 	.llseek		= seq_lseek,
-	.release	= seq_release,
+	.release	= nfsd_pool_stats_release,
 	.owner		= THIS_MODULE,
 };
 
@@ -490,22 +492,18 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size)
  *
  * Input:
  *			buf:	'\n'-terminated C string containing a
- *				presentation format IPv4 address
+ *				presentation format IP 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,
-	};
-	int b1, b2, b3, b4;
-	char c;
+	struct sockaddr_storage address;
+	struct sockaddr *sap = (struct sockaddr *)&address;
+	size_t salen = sizeof(address);
 	char *fo_path;
 
 	/* sanity check */
@@ -519,14 +517,10 @@ static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size)
 	if (qword_get(&buf, fo_path, size) < 0)
 		return -EINVAL;
 
-	/* get ipv4 address */
-	if (sscanf(fo_path, "%u.%u.%u.%u%c", &b1, &b2, &b3, &b4, &c) != 4)
-		return -EINVAL;
-	if (b1 > 255 || b2 > 255 || b3 > 255 || b4 > 255)
+	if (rpc_pton(fo_path, size, sap, salen) == 0)
 		return -EINVAL;
-	sin.sin_addr.s_addr = htonl((b1 << 24) | (b2 << 16) | (b3 << 8) | b4);
 
-	return nlmsvc_unlock_all_by_ip((struct sockaddr *)&sin);
+	return nlmsvc_unlock_all_by_ip(sap);
 }
 
 /**
@@ -783,10 +777,7 @@ static ssize_t write_pool_threads(struct file *file, char *buf, size_t size)
 		size -= len;
 		mesg += len;
 	}
-
-	mutex_unlock(&nfsd_mutex);
-	return (mesg-buf);
-
+	rv = mesg - buf;
 out_free:
 	kfree(nthreads);
 	mutex_unlock(&nfsd_mutex);
diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
index 8847f3fbfc1..01965b2f3a7 100644
--- a/fs/nfsd/nfsfh.c
+++ b/fs/nfsd/nfsfh.c
@@ -397,44 +397,51 @@ static inline void _fh_update_old(struct dentry *dentry,
 		fh->ofh_dirino = 0;
 }
 
-__be32
-fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
-	   struct svc_fh *ref_fh)
+static bool is_root_export(struct svc_export *exp)
 {
-	/* ref_fh is a reference file handle.
-	 * if it is non-null and for the same filesystem, then we should compose
-	 * a filehandle which is of the same version, where possible.
-	 * Currently, that means that if ref_fh->fh_handle.fh_version == 0xca
-	 * Then create a 32byte filehandle using nfs_fhbase_old
-	 *
-	 */
+	return exp->ex_path.dentry == exp->ex_path.dentry->d_sb->s_root;
+}
 
-	u8 version;
-	u8 fsid_type = 0;
-	struct inode * inode = dentry->d_inode;
-	struct dentry *parent = dentry->d_parent;
-	__u32 *datap;
-	dev_t ex_dev = exp->ex_path.dentry->d_inode->i_sb->s_dev;
-	int root_export = (exp->ex_path.dentry == exp->ex_path.dentry->d_sb->s_root);
+static struct super_block *exp_sb(struct svc_export *exp)
+{
+	return exp->ex_path.dentry->d_inode->i_sb;
+}
 
-	dprintk("nfsd: fh_compose(exp %02x:%02x/%ld %s/%s, ino=%ld)\n",
-		MAJOR(ex_dev), MINOR(ex_dev),
-		(long) exp->ex_path.dentry->d_inode->i_ino,
-		parent->d_name.name, dentry->d_name.name,
-		(inode ? inode->i_ino : 0));
+static bool fsid_type_ok_for_exp(u8 fsid_type, struct svc_export *exp)
+{
+	switch (fsid_type) {
+	case FSID_DEV:
+		if (!old_valid_dev(exp_sb(exp)->s_dev))
+			return 0;
+		/* FALL THROUGH */
+	case FSID_MAJOR_MINOR:
+	case FSID_ENCODE_DEV:
+		return exp_sb(exp)->s_type->fs_flags & FS_REQUIRES_DEV;
+	case FSID_NUM:
+		return exp->ex_flags & NFSEXP_FSID;
+	case FSID_UUID8:
+	case FSID_UUID16:
+		if (!is_root_export(exp))
+			return 0;
+		/* fall through */
+	case FSID_UUID4_INUM:
+	case FSID_UUID16_INUM:
+		return exp->ex_uuid != NULL;
+	}
+	return 1;
+}
 
-	/* Choose filehandle version and fsid type based on
-	 * the reference filehandle (if it is in the same export)
-	 * or the export options.
-	 */
- retry:
+
+static void set_version_and_fsid_type(struct svc_fh *fhp, struct svc_export *exp, struct svc_fh *ref_fh)
+{
+	u8 version;
+	u8 fsid_type;
+retry:
 	version = 1;
 	if (ref_fh && ref_fh->fh_export == exp) {
 		version = ref_fh->fh_handle.fh_version;
 		fsid_type = ref_fh->fh_handle.fh_fsid_type;
 
-		if (ref_fh == fhp)
-			fh_put(ref_fh);
 		ref_fh = NULL;
 
 		switch (version) {
@@ -447,58 +454,66 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
 			goto retry;
 		}
 
-		/* Need to check that this type works for this
-		 * export point.  As the fsid -> filesystem mapping
-		 * was guided by user-space, there is no guarantee
-		 * that the filesystem actually supports that fsid
-		 * type. If it doesn't we loop around again without
-		 * ref_fh set.
+		/*
+		 * As the fsid -> filesystem mapping was guided by
+		 * user-space, there is no guarantee that the filesystem
+		 * actually supports that fsid type. If it doesn't we
+		 * loop around again without ref_fh set.
 		 */
-		switch(fsid_type) {
-		case FSID_DEV:
-			if (!old_valid_dev(ex_dev))
-				goto retry;
-			/* FALL THROUGH */
-		case FSID_MAJOR_MINOR:
-		case FSID_ENCODE_DEV:
-			if (!(exp->ex_path.dentry->d_inode->i_sb->s_type->fs_flags
-			      & FS_REQUIRES_DEV))
-				goto retry;
-			break;
-		case FSID_NUM:
-			if (! (exp->ex_flags & NFSEXP_FSID))
-				goto retry;
-			break;
-		case FSID_UUID8:
-		case FSID_UUID16:
-			if (!root_export)
-				goto retry;
-			/* fall through */
-		case FSID_UUID4_INUM:
-		case FSID_UUID16_INUM:
-			if (exp->ex_uuid == NULL)
-				goto retry;
-			break;
-		}
+		if (!fsid_type_ok_for_exp(fsid_type, exp))
+			goto retry;
 	} else if (exp->ex_flags & NFSEXP_FSID) {
 		fsid_type = FSID_NUM;
 	} else if (exp->ex_uuid) {
 		if (fhp->fh_maxsize >= 64) {
-			if (root_export)
+			if (is_root_export(exp))
 				fsid_type = FSID_UUID16;
 			else
 				fsid_type = FSID_UUID16_INUM;
 		} else {
-			if (root_export)
+			if (is_root_export(exp))
 				fsid_type = FSID_UUID8;
 			else
 				fsid_type = FSID_UUID4_INUM;
 		}
-	} else if (!old_valid_dev(ex_dev))
+	} else if (!old_valid_dev(exp_sb(exp)->s_dev))
 		/* for newer device numbers, we must use a newer fsid format */
 		fsid_type = FSID_ENCODE_DEV;
 	else
 		fsid_type = FSID_DEV;
+	fhp->fh_handle.fh_version = version;
+	if (version)
+		fhp->fh_handle.fh_fsid_type = fsid_type;
+}
+
+__be32
+fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
+	   struct svc_fh *ref_fh)
+{
+	/* ref_fh is a reference file handle.
+	 * if it is non-null and for the same filesystem, then we should compose
+	 * a filehandle which is of the same version, where possible.
+	 * Currently, that means that if ref_fh->fh_handle.fh_version == 0xca
+	 * Then create a 32byte filehandle using nfs_fhbase_old
+	 *
+	 */
+
+	struct inode * inode = dentry->d_inode;
+	struct dentry *parent = dentry->d_parent;
+	__u32 *datap;
+	dev_t ex_dev = exp_sb(exp)->s_dev;
+
+	dprintk("nfsd: fh_compose(exp %02x:%02x/%ld %s/%s, ino=%ld)\n",
+		MAJOR(ex_dev), MINOR(ex_dev),
+		(long) exp->ex_path.dentry->d_inode->i_ino,
+		parent->d_name.name, dentry->d_name.name,
+		(inode ? inode->i_ino : 0));
+
+	/* Choose filehandle version and fsid type based on
+	 * the reference filehandle (if it is in the same export)
+	 * or the export options.
+	 */
+	 set_version_and_fsid_type(fhp, exp, ref_fh);
 
 	if (ref_fh == fhp)
 		fh_put(ref_fh);
@@ -516,7 +531,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
 	fhp->fh_export = exp;
 	cache_get(&exp->h);
 
-	if (version == 0xca) {
+	if (fhp->fh_handle.fh_version == 0xca) {
 		/* old style filehandle please */
 		memset(&fhp->fh_handle.fh_base, 0, NFS_FHSIZE);
 		fhp->fh_handle.fh_size = NFS_FHSIZE;
@@ -530,22 +545,22 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
 			_fh_update_old(dentry, exp, &fhp->fh_handle);
 	} else {
 		int len;
-		fhp->fh_handle.fh_version = 1;
 		fhp->fh_handle.fh_auth_type = 0;
 		datap = fhp->fh_handle.fh_auth+0;
-		fhp->fh_handle.fh_fsid_type = fsid_type;
-		mk_fsid(fsid_type, datap, ex_dev,
+		mk_fsid(fhp->fh_handle.fh_fsid_type, datap, ex_dev,
 			exp->ex_path.dentry->d_inode->i_ino,
 			exp->ex_fsid, exp->ex_uuid);
 
-		len = key_len(fsid_type);
+		len = key_len(fhp->fh_handle.fh_fsid_type);
 		datap += len/4;
 		fhp->fh_handle.fh_size = 4 + len;
 
 		if (inode)
 			_fh_update(fhp, exp, dentry);
-		if (fhp->fh_handle.fh_fileid_type == 255)
+		if (fhp->fh_handle.fh_fileid_type == 255) {
+			fh_put(fhp);
 			return nfserr_opnotsupp;
+		}
 	}
 
 	return 0;
@@ -639,8 +654,7 @@ enum fsid_source fsid_source(struct svc_fh *fhp)
 	case FSID_DEV:
 	case FSID_ENCODE_DEV:
 	case FSID_MAJOR_MINOR:
-		if (fhp->fh_export->ex_path.dentry->d_inode->i_sb->s_type->fs_flags
-		    & FS_REQUIRES_DEV)
+		if (exp_sb(fhp->fh_export)->s_type->fs_flags & FS_REQUIRES_DEV)
 			return FSIDSOURCE_DEV;
 		break;
 	case FSID_NUM:
diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index 492c79b7800..67ea83eedd4 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -34,6 +34,7 @@
 #include <linux/nfsd/syscall.h>
 #include <linux/lockd/bind.h>
 #include <linux/nfsacl.h>
+#include <linux/seq_file.h>
 
 #define NFSDDBG_FACILITY	NFSDDBG_SVC
 
@@ -66,6 +67,16 @@ struct timeval			nfssvc_boot;
 DEFINE_MUTEX(nfsd_mutex);
 struct svc_serv 		*nfsd_serv;
 
+/*
+ * nfsd_drc_lock protects nfsd_drc_max_pages and nfsd_drc_pages_used.
+ * nfsd_drc_max_pages limits the total amount of memory available for
+ * version 4.1 DRC caches.
+ * nfsd_drc_pages_used tracks the current version 4.1 DRC memory usage.
+ */
+spinlock_t	nfsd_drc_lock;
+unsigned int	nfsd_drc_max_mem;
+unsigned int	nfsd_drc_mem_used;
+
 #if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
 static struct svc_stat	nfsd_acl_svcstats;
 static struct svc_version *	nfsd_acl_version[] = {
@@ -235,13 +246,12 @@ void nfsd_reset_versions(void)
  */
 static void set_max_drc(void)
 {
-	/* The percent of nr_free_buffer_pages used by the V4.1 server DRC */
-	#define NFSD_DRC_SIZE_SHIFT	7
-	nfsd_serv->sv_drc_max_pages = nr_free_buffer_pages()
-						>> NFSD_DRC_SIZE_SHIFT;
-	nfsd_serv->sv_drc_pages_used = 0;
-	dprintk("%s svc_drc_max_pages %u\n", __func__,
-		nfsd_serv->sv_drc_max_pages);
+	#define NFSD_DRC_SIZE_SHIFT	10
+	nfsd_drc_max_mem = (nr_free_buffer_pages()
+					>> NFSD_DRC_SIZE_SHIFT) * PAGE_SIZE;
+	nfsd_drc_mem_used = 0;
+	spin_lock_init(&nfsd_drc_lock);
+	dprintk("%s nfsd_drc_max_mem %u \n", __func__, nfsd_drc_max_mem);
 }
 
 int nfsd_create_serv(void)
@@ -401,7 +411,9 @@ nfsd_svc(unsigned short port, int nrservs)
 	error =	nfsd_racache_init(2*nrservs);
 	if (error<0)
 		goto out;
-	nfs4_state_start();
+	error = nfs4_state_start();
+	if (error)
+		goto out;
 
 	nfsd_reset_versions();
 
@@ -496,7 +508,9 @@ nfsd(void *vrqstp)
 		/* Lock the export hash tables for reading. */
 		exp_readlock();
 
+		validate_process_creds();
 		svc_process(rqstp);
+		validate_process_creds();
 
 		/* Unlock export hash tables */
 		exp_readunlock();
@@ -567,10 +581,6 @@ nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp)
 		+ rqstp->rq_res.head[0].iov_len;
 	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
 
-	/* NFSv4.1 DRC requires statp */
-	if (rqstp->rq_vers == 4)
-		nfsd4_set_statp(rqstp, statp);
-
 	/* Now call the procedure handler, and encode NFS status. */
 	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
 	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
@@ -605,7 +615,25 @@ nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp)
 
 int nfsd_pool_stats_open(struct inode *inode, struct file *file)
 {
-	if (nfsd_serv == NULL)
+	int ret;
+	mutex_lock(&nfsd_mutex);
+	if (nfsd_serv == NULL) {
+		mutex_unlock(&nfsd_mutex);
 		return -ENODEV;
-	return svc_pool_stats_open(nfsd_serv, file);
+	}
+	/* bump up the psudo refcount while traversing */
+	svc_get(nfsd_serv);
+	ret = svc_pool_stats_open(nfsd_serv, file);
+	mutex_unlock(&nfsd_mutex);
+	return ret;
+}
+
+int nfsd_pool_stats_release(struct inode *inode, struct file *file)
+{
+	int ret = seq_release(inode, file);
+	mutex_lock(&nfsd_mutex);
+	/* this function really, really should have been called svc_put() */
+	svc_destroy(nfsd_serv);
+	mutex_unlock(&nfsd_mutex);
+	return ret;
 }
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 23341c1063b..a293f027326 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -89,6 +89,12 @@ struct raparm_hbucket {
 #define RAPARM_HASH_MASK	(RAPARM_HASH_SIZE-1)
 static struct raparm_hbucket	raparm_hash[RAPARM_HASH_SIZE];
 
+static inline int
+nfsd_v4client(struct svc_rqst *rq)
+{
+    return rq->rq_prog == NFS_PROGRAM && rq->rq_vers == 4;
+}
+
 /* 
  * Called from nfsd_lookup and encode_dirent. Check if we have crossed 
  * a mount point.
@@ -115,7 +121,8 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
 		path_put(&path);
 		goto out;
 	}
-	if ((exp->ex_flags & NFSEXP_CROSSMOUNT) || EX_NOHIDE(exp2)) {
+	if (nfsd_v4client(rqstp) ||
+		(exp->ex_flags & NFSEXP_CROSSMOUNT) || EX_NOHIDE(exp2)) {
 		/* successfully crossed mount point */
 		/*
 		 * This is subtle: path.dentry is *not* on path.mnt
@@ -684,6 +691,8 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 	__be32		err;
 	int		host_err;
 
+	validate_process_creds();
+
 	/*
 	 * If we get here, then the client has already done an "open",
 	 * and (hopefully) checked permission - so allow OWNER_OVERRIDE
@@ -740,6 +749,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 out_nfserr:
 	err = nfserrno(host_err);
 out:
+	validate_process_creds();
 	return err;
 }
 
diff --git a/fs/nilfs2/Kconfig b/fs/nilfs2/Kconfig
index 72da095d400..251da07b2a1 100644
--- a/fs/nilfs2/Kconfig
+++ b/fs/nilfs2/Kconfig
@@ -1,6 +1,6 @@
 config NILFS2_FS
 	tristate "NILFS2 file system support (EXPERIMENTAL)"
-	depends on BLOCK && EXPERIMENTAL
+	depends on EXPERIMENTAL
 	select CRC32
 	help
 	  NILFS2 is a log-structured file system (LFS) supporting continuous
diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c
index 99d58a028b9..08834df6ec6 100644
--- a/fs/nilfs2/bmap.c
+++ b/fs/nilfs2/bmap.c
@@ -36,6 +36,26 @@ struct inode *nilfs_bmap_get_dat(const struct nilfs_bmap *bmap)
 	return nilfs_dat_inode(NILFS_I_NILFS(bmap->b_inode));
 }
 
+/**
+ * nilfs_bmap_lookup_at_level - find a data block or node block
+ * @bmap: bmap
+ * @key: key
+ * @level: level
+ * @ptrp: place to store the value associated to @key
+ *
+ * Description: nilfs_bmap_lookup_at_level() finds a record whose key
+ * matches @key in the block at @level of the bmap.
+ *
+ * Return Value: On success, 0 is returned and the record associated with @key
+ * is stored in the place pointed by @ptrp. On error, one of the following
+ * negative error codes is returned.
+ *
+ * %-EIO - I/O error.
+ *
+ * %-ENOMEM - Insufficient amount of memory available.
+ *
+ * %-ENOENT - A record associated with @key does not exist.
+ */
 int nilfs_bmap_lookup_at_level(struct nilfs_bmap *bmap, __u64 key, int level,
 			       __u64 *ptrp)
 {
@@ -69,39 +89,6 @@ int nilfs_bmap_lookup_contig(struct nilfs_bmap *bmap, __u64 key, __u64 *ptrp,
 	return ret;
 }
 
-/**
- * nilfs_bmap_lookup - find a record
- * @bmap: bmap
- * @key: key
- * @recp: pointer to record
- *
- * Description: nilfs_bmap_lookup() finds a record whose key matches @key in
- * @bmap.
- *
- * Return Value: On success, 0 is returned and the record associated with @key
- * is stored in the place pointed by @recp. On error, one of the following
- * negative error codes is returned.
- *
- * %-EIO - I/O error.
- *
- * %-ENOMEM - Insufficient amount of memory available.
- *
- * %-ENOENT - A record associated with @key does not exist.
- */
-int nilfs_bmap_lookup(struct nilfs_bmap *bmap,
-		      unsigned long key,
-		      unsigned long *recp)
-{
-	__u64 ptr;
-	int ret;
-
-	/* XXX: use macro for level 1 */
-	ret = nilfs_bmap_lookup_at_level(bmap, key, 1, &ptr);
-	if (recp != NULL)
-		*recp = ptr;
-	return ret;
-}
-
 static int nilfs_bmap_do_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr)
 {
 	__u64 keys[NILFS_BMAP_SMALL_HIGH + 1];
@@ -469,104 +456,6 @@ __u64 nilfs_bmap_find_target_in_group(const struct nilfs_bmap *bmap)
 		(entries_per_group / NILFS_BMAP_GROUP_DIV);
 }
 
-int nilfs_bmap_prepare_alloc_v(struct nilfs_bmap *bmap,
-				 union nilfs_bmap_ptr_req *req)
-{
-	return nilfs_dat_prepare_alloc(nilfs_bmap_get_dat(bmap), &req->bpr_req);
-}
-
-void nilfs_bmap_commit_alloc_v(struct nilfs_bmap *bmap,
-				 union nilfs_bmap_ptr_req *req)
-{
-	nilfs_dat_commit_alloc(nilfs_bmap_get_dat(bmap), &req->bpr_req);
-}
-
-void nilfs_bmap_abort_alloc_v(struct nilfs_bmap *bmap,
-			      union nilfs_bmap_ptr_req *req)
-{
-	nilfs_dat_abort_alloc(nilfs_bmap_get_dat(bmap), &req->bpr_req);
-}
-
-int nilfs_bmap_start_v(struct nilfs_bmap *bmap, union nilfs_bmap_ptr_req *req,
-		       sector_t blocknr)
-{
-	struct inode *dat = nilfs_bmap_get_dat(bmap);
-	int ret;
-
-	ret = nilfs_dat_prepare_start(dat, &req->bpr_req);
-	if (likely(!ret))
-		nilfs_dat_commit_start(dat, &req->bpr_req, blocknr);
-	return ret;
-}
-
-int nilfs_bmap_prepare_end_v(struct nilfs_bmap *bmap,
-			     union nilfs_bmap_ptr_req *req)
-{
-	return nilfs_dat_prepare_end(nilfs_bmap_get_dat(bmap), &req->bpr_req);
-}
-
-void nilfs_bmap_commit_end_v(struct nilfs_bmap *bmap,
-			     union nilfs_bmap_ptr_req *req)
-{
-	nilfs_dat_commit_end(nilfs_bmap_get_dat(bmap), &req->bpr_req,
-			     bmap->b_ptr_type == NILFS_BMAP_PTR_VS);
-}
-
-void nilfs_bmap_abort_end_v(struct nilfs_bmap *bmap,
-			    union nilfs_bmap_ptr_req *req)
-{
-	nilfs_dat_abort_end(nilfs_bmap_get_dat(bmap), &req->bpr_req);
-}
-
-int nilfs_bmap_move_v(const struct nilfs_bmap *bmap, __u64 vblocknr,
-		      sector_t blocknr)
-{
-	return nilfs_dat_move(nilfs_bmap_get_dat(bmap), vblocknr, blocknr);
-}
-
-int nilfs_bmap_mark_dirty(const struct nilfs_bmap *bmap, __u64 vblocknr)
-{
-	return nilfs_dat_mark_dirty(nilfs_bmap_get_dat(bmap), vblocknr);
-}
-
-int nilfs_bmap_prepare_update_v(struct nilfs_bmap *bmap,
-				union nilfs_bmap_ptr_req *oldreq,
-				union nilfs_bmap_ptr_req *newreq)
-{
-	struct inode *dat = nilfs_bmap_get_dat(bmap);
-	int ret;
-
-	ret = nilfs_dat_prepare_end(dat, &oldreq->bpr_req);
-	if (ret < 0)
-		return ret;
-	ret = nilfs_dat_prepare_alloc(dat, &newreq->bpr_req);
-	if (ret < 0)
-		nilfs_dat_abort_end(dat, &oldreq->bpr_req);
-
-	return ret;
-}
-
-void nilfs_bmap_commit_update_v(struct nilfs_bmap *bmap,
-				union nilfs_bmap_ptr_req *oldreq,
-				union nilfs_bmap_ptr_req *newreq)
-{
-	struct inode *dat = nilfs_bmap_get_dat(bmap);
-
-	nilfs_dat_commit_end(dat, &oldreq->bpr_req,
-			     bmap->b_ptr_type == NILFS_BMAP_PTR_VS);
-	nilfs_dat_commit_alloc(dat, &newreq->bpr_req);
-}
-
-void nilfs_bmap_abort_update_v(struct nilfs_bmap *bmap,
-			       union nilfs_bmap_ptr_req *oldreq,
-			       union nilfs_bmap_ptr_req *newreq)
-{
-	struct inode *dat = nilfs_bmap_get_dat(bmap);
-
-	nilfs_dat_abort_end(dat, &oldreq->bpr_req);
-	nilfs_dat_abort_alloc(dat, &newreq->bpr_req);
-}
-
 static struct lock_class_key nilfs_bmap_dat_lock_key;
 static struct lock_class_key nilfs_bmap_mdt_lock_key;
 
diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h
index b2890cdcef1..9980d7dbab9 100644
--- a/fs/nilfs2/bmap.h
+++ b/fs/nilfs2/bmap.h
@@ -28,6 +28,7 @@
 #include <linux/buffer_head.h>
 #include <linux/nilfs2_fs.h>
 #include "alloc.h"
+#include "dat.h"
 
 #define NILFS_BMAP_INVALID_PTR	0
 
@@ -141,7 +142,6 @@ struct nilfs_bmap {
 int nilfs_bmap_test_and_clear_dirty(struct nilfs_bmap *);
 int nilfs_bmap_read(struct nilfs_bmap *, struct nilfs_inode *);
 void nilfs_bmap_write(struct nilfs_bmap *, struct nilfs_inode *);
-int nilfs_bmap_lookup(struct nilfs_bmap *, unsigned long, unsigned long *);
 int nilfs_bmap_lookup_contig(struct nilfs_bmap *, __u64, __u64 *, unsigned);
 int nilfs_bmap_insert(struct nilfs_bmap *, unsigned long, unsigned long);
 int nilfs_bmap_delete(struct nilfs_bmap *, unsigned long);
@@ -160,90 +160,76 @@ void nilfs_bmap_init_gcdat(struct nilfs_bmap *, struct nilfs_bmap *);
 void nilfs_bmap_commit_gcdat(struct nilfs_bmap *, struct nilfs_bmap *);
 
 
+static inline int nilfs_bmap_lookup(struct nilfs_bmap *bmap, __u64 key,
+				    __u64 *ptr)
+{
+	return nilfs_bmap_lookup_at_level(bmap, key, 1, ptr);
+}
+
 /*
  * Internal use only
  */
 struct inode *nilfs_bmap_get_dat(const struct nilfs_bmap *);
-int nilfs_bmap_prepare_alloc_v(struct nilfs_bmap *,
-			       union nilfs_bmap_ptr_req *);
-void nilfs_bmap_commit_alloc_v(struct nilfs_bmap *,
-			       union nilfs_bmap_ptr_req *);
-void nilfs_bmap_abort_alloc_v(struct nilfs_bmap *,
-			      union nilfs_bmap_ptr_req *);
 
 static inline int nilfs_bmap_prepare_alloc_ptr(struct nilfs_bmap *bmap,
-					       union nilfs_bmap_ptr_req *req)
+					       union nilfs_bmap_ptr_req *req,
+					       struct inode *dat)
 {
-	if (NILFS_BMAP_USE_VBN(bmap))
-		return nilfs_bmap_prepare_alloc_v(bmap, req);
+	if (dat)
+		return nilfs_dat_prepare_alloc(dat, &req->bpr_req);
 	/* ignore target ptr */
 	req->bpr_ptr = bmap->b_last_allocated_ptr++;
 	return 0;
 }
 
 static inline void nilfs_bmap_commit_alloc_ptr(struct nilfs_bmap *bmap,
-					       union nilfs_bmap_ptr_req *req)
+					       union nilfs_bmap_ptr_req *req,
+					       struct inode *dat)
 {
-	if (NILFS_BMAP_USE_VBN(bmap))
-		nilfs_bmap_commit_alloc_v(bmap, req);
+	if (dat)
+		nilfs_dat_commit_alloc(dat, &req->bpr_req);
 }
 
 static inline void nilfs_bmap_abort_alloc_ptr(struct nilfs_bmap *bmap,
-					      union nilfs_bmap_ptr_req *req)
+					      union nilfs_bmap_ptr_req *req,
+					      struct inode *dat)
 {
-	if (NILFS_BMAP_USE_VBN(bmap))
-		nilfs_bmap_abort_alloc_v(bmap, req);
+	if (dat)
+		nilfs_dat_abort_alloc(dat, &req->bpr_req);
 	else
 		bmap->b_last_allocated_ptr--;
 }
 
-int nilfs_bmap_prepare_end_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *);
-void nilfs_bmap_commit_end_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *);
-void nilfs_bmap_abort_end_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *);
-
 static inline int nilfs_bmap_prepare_end_ptr(struct nilfs_bmap *bmap,
-					     union nilfs_bmap_ptr_req *req)
+					     union nilfs_bmap_ptr_req *req,
+					     struct inode *dat)
 {
-	return NILFS_BMAP_USE_VBN(bmap) ?
-		nilfs_bmap_prepare_end_v(bmap, req) : 0;
+	return dat ? nilfs_dat_prepare_end(dat, &req->bpr_req) : 0;
 }
 
 static inline void nilfs_bmap_commit_end_ptr(struct nilfs_bmap *bmap,
-					     union nilfs_bmap_ptr_req *req)
+					     union nilfs_bmap_ptr_req *req,
+					     struct inode *dat)
 {
-	if (NILFS_BMAP_USE_VBN(bmap))
-		nilfs_bmap_commit_end_v(bmap, req);
+	if (dat)
+		nilfs_dat_commit_end(dat, &req->bpr_req,
+				     bmap->b_ptr_type == NILFS_BMAP_PTR_VS);
 }
 
 static inline void nilfs_bmap_abort_end_ptr(struct nilfs_bmap *bmap,
-					    union nilfs_bmap_ptr_req *req)
+					    union nilfs_bmap_ptr_req *req,
+					    struct inode *dat)
 {
-	if (NILFS_BMAP_USE_VBN(bmap))
-		nilfs_bmap_abort_end_v(bmap, req);
+	if (dat)
+		nilfs_dat_abort_end(dat, &req->bpr_req);
 }
 
-int nilfs_bmap_start_v(struct nilfs_bmap *, union nilfs_bmap_ptr_req *,
-		       sector_t);
-int nilfs_bmap_move_v(const struct nilfs_bmap *, __u64, sector_t);
-int nilfs_bmap_mark_dirty(const struct nilfs_bmap *, __u64);
-
-
 __u64 nilfs_bmap_data_get_key(const struct nilfs_bmap *,
 			      const struct buffer_head *);
 
 __u64 nilfs_bmap_find_target_seq(const struct nilfs_bmap *, __u64);
 __u64 nilfs_bmap_find_target_in_group(const struct nilfs_bmap *);
 
-int nilfs_bmap_prepare_update_v(struct nilfs_bmap *,
-				union nilfs_bmap_ptr_req *,
-				union nilfs_bmap_ptr_req *);
-void nilfs_bmap_commit_update_v(struct nilfs_bmap *,
-				union nilfs_bmap_ptr_req *,
-				union nilfs_bmap_ptr_req *);
-void nilfs_bmap_abort_update_v(struct nilfs_bmap *,
-			       union nilfs_bmap_ptr_req *,
-			       union nilfs_bmap_ptr_req *);
-
 void nilfs_bmap_add_blocks(const struct nilfs_bmap *, int);
 void nilfs_bmap_sub_blocks(const struct nilfs_bmap *, int);
 
diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c
index c668bca579c..6a2711f4c32 100644
--- a/fs/nilfs2/btnode.c
+++ b/fs/nilfs2/btnode.c
@@ -46,7 +46,7 @@ void nilfs_btnode_cache_init_once(struct address_space *btnc)
 	INIT_LIST_HEAD(&btnc->i_mmap_nonlinear);
 }
 
-static struct address_space_operations def_btnode_aops = {
+static const struct address_space_operations def_btnode_aops = {
 	.sync_page		= block_sync_page,
 };
 
diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c
index aa412724b64..e25b507a474 100644
--- a/fs/nilfs2/btree.c
+++ b/fs/nilfs2/btree.c
@@ -71,21 +71,17 @@ void nilfs_btree_path_cache_destroy(void)
 	kmem_cache_destroy(nilfs_btree_path_cache);
 }
 
-static inline struct nilfs_btree_path *
-nilfs_btree_alloc_path(const struct nilfs_btree *btree)
+static inline struct nilfs_btree_path *nilfs_btree_alloc_path(void)
 {
-	return (struct nilfs_btree_path *)
-		kmem_cache_alloc(nilfs_btree_path_cache, GFP_NOFS);
+	return kmem_cache_alloc(nilfs_btree_path_cache, GFP_NOFS);
 }
 
-static inline void nilfs_btree_free_path(const struct nilfs_btree *btree,
-					 struct nilfs_btree_path *path)
+static inline void nilfs_btree_free_path(struct nilfs_btree_path *path)
 {
 	kmem_cache_free(nilfs_btree_path_cache, path);
 }
 
-static void nilfs_btree_init_path(const struct nilfs_btree *btree,
-				  struct nilfs_btree_path *path)
+static void nilfs_btree_init_path(struct nilfs_btree_path *path)
 {
 	int level;
 
@@ -101,26 +97,13 @@ static void nilfs_btree_init_path(const struct nilfs_btree *btree,
 	}
 }
 
-static void nilfs_btree_clear_path(const struct nilfs_btree *btree,
-				   struct nilfs_btree_path *path)
+static void nilfs_btree_release_path(struct nilfs_btree_path *path)
 {
 	int level;
 
-	for (level = NILFS_BTREE_LEVEL_DATA;
-	     level < NILFS_BTREE_LEVEL_MAX;
-	     level++) {
-		if (path[level].bp_bh != NULL) {
-			brelse(path[level].bp_bh);
-			path[level].bp_bh = NULL;
-		}
-		/* sib_bh is released or deleted by prepare or commit
-		 * operations. */
-		path[level].bp_sib_bh = NULL;
-		path[level].bp_index = 0;
-		path[level].bp_oldreq.bpr_ptr = NILFS_BMAP_INVALID_PTR;
-		path[level].bp_newreq.bpr_ptr = NILFS_BMAP_INVALID_PTR;
-		path[level].bp_op = NULL;
-	}
+	for (level = NILFS_BTREE_LEVEL_DATA; level < NILFS_BTREE_LEVEL_MAX;
+	     level++)
+		brelse(path[level].bp_bh);
 }
 
 /*
@@ -148,129 +131,110 @@ static int nilfs_btree_get_new_block(const struct nilfs_btree *btree,
 }
 
 static inline int
-nilfs_btree_node_get_flags(const struct nilfs_btree *btree,
-			   const struct nilfs_btree_node *node)
+nilfs_btree_node_get_flags(const struct nilfs_btree_node *node)
 {
 	return node->bn_flags;
 }
 
 static inline void
-nilfs_btree_node_set_flags(struct nilfs_btree *btree,
-			   struct nilfs_btree_node *node,
-			   int flags)
+nilfs_btree_node_set_flags(struct nilfs_btree_node *node, int flags)
 {
 	node->bn_flags = flags;
 }
 
-static inline int nilfs_btree_node_root(const struct nilfs_btree *btree,
-					const struct nilfs_btree_node *node)
+static inline int nilfs_btree_node_root(const struct nilfs_btree_node *node)
 {
-	return nilfs_btree_node_get_flags(btree, node) & NILFS_BTREE_NODE_ROOT;
+	return nilfs_btree_node_get_flags(node) & NILFS_BTREE_NODE_ROOT;
 }
 
 static inline int
-nilfs_btree_node_get_level(const struct nilfs_btree *btree,
-			   const struct nilfs_btree_node *node)
+nilfs_btree_node_get_level(const struct nilfs_btree_node *node)
 {
 	return node->bn_level;
 }
 
 static inline void
-nilfs_btree_node_set_level(struct nilfs_btree *btree,
-			   struct nilfs_btree_node *node,
-			   int level)
+nilfs_btree_node_set_level(struct nilfs_btree_node *node, int level)
 {
 	node->bn_level = level;
 }
 
 static inline int
-nilfs_btree_node_get_nchildren(const struct nilfs_btree *btree,
-			       const struct nilfs_btree_node *node)
+nilfs_btree_node_get_nchildren(const struct nilfs_btree_node *node)
 {
 	return le16_to_cpu(node->bn_nchildren);
 }
 
 static inline void
-nilfs_btree_node_set_nchildren(struct nilfs_btree *btree,
-			       struct nilfs_btree_node *node,
-			       int nchildren)
+nilfs_btree_node_set_nchildren(struct nilfs_btree_node *node, int nchildren)
 {
 	node->bn_nchildren = cpu_to_le16(nchildren);
 }
 
-static inline int
-nilfs_btree_node_size(const struct nilfs_btree *btree)
+static inline int nilfs_btree_node_size(const struct nilfs_btree *btree)
 {
 	return 1 << btree->bt_bmap.b_inode->i_blkbits;
 }
 
 static inline int
-nilfs_btree_node_nchildren_min(const struct nilfs_btree *btree,
-			       const struct nilfs_btree_node *node)
+nilfs_btree_node_nchildren_min(const struct nilfs_btree_node *node,
+			       const struct nilfs_btree *btree)
 {
-	return nilfs_btree_node_root(btree, node) ?
+	return nilfs_btree_node_root(node) ?
 		NILFS_BTREE_ROOT_NCHILDREN_MIN :
 		NILFS_BTREE_NODE_NCHILDREN_MIN(nilfs_btree_node_size(btree));
 }
 
 static inline int
-nilfs_btree_node_nchildren_max(const struct nilfs_btree *btree,
-			       const struct nilfs_btree_node *node)
+nilfs_btree_node_nchildren_max(const struct nilfs_btree_node *node,
+			       const struct nilfs_btree *btree)
 {
-	return nilfs_btree_node_root(btree, node) ?
+	return nilfs_btree_node_root(node) ?
 		NILFS_BTREE_ROOT_NCHILDREN_MAX :
 		NILFS_BTREE_NODE_NCHILDREN_MAX(nilfs_btree_node_size(btree));
 }
 
 static inline __le64 *
-nilfs_btree_node_dkeys(const struct nilfs_btree *btree,
-		       const struct nilfs_btree_node *node)
+nilfs_btree_node_dkeys(const struct nilfs_btree_node *node)
 {
 	return (__le64 *)((char *)(node + 1) +
-			  (nilfs_btree_node_root(btree, node) ?
+			  (nilfs_btree_node_root(node) ?
 			   0 : NILFS_BTREE_NODE_EXTRA_PAD_SIZE));
 }
 
 static inline __le64 *
-nilfs_btree_node_dptrs(const struct nilfs_btree *btree,
-		       const struct nilfs_btree_node *node)
+nilfs_btree_node_dptrs(const struct nilfs_btree_node *node,
+		       const struct nilfs_btree *btree)
 {
-	return (__le64 *)(nilfs_btree_node_dkeys(btree, node) +
-			  nilfs_btree_node_nchildren_max(btree, node));
+	return (__le64 *)(nilfs_btree_node_dkeys(node) +
+			  nilfs_btree_node_nchildren_max(node, btree));
 }
 
 static inline __u64
-nilfs_btree_node_get_key(const struct nilfs_btree *btree,
-			 const struct nilfs_btree_node *node, int index)
+nilfs_btree_node_get_key(const struct nilfs_btree_node *node, int index)
 {
-	return nilfs_bmap_dkey_to_key(*(nilfs_btree_node_dkeys(btree, node) +
-					index));
+	return nilfs_bmap_dkey_to_key(*(nilfs_btree_node_dkeys(node) + index));
 }
 
 static inline void
-nilfs_btree_node_set_key(struct nilfs_btree *btree,
-			 struct nilfs_btree_node *node, int index, __u64 key)
+nilfs_btree_node_set_key(struct nilfs_btree_node *node, int index, __u64 key)
 {
-	*(nilfs_btree_node_dkeys(btree, node) + index) =
-		nilfs_bmap_key_to_dkey(key);
+	*(nilfs_btree_node_dkeys(node) + index) = nilfs_bmap_key_to_dkey(key);
 }
 
 static inline __u64
 nilfs_btree_node_get_ptr(const struct nilfs_btree *btree,
-			 const struct nilfs_btree_node *node,
-			 int index)
+			 const struct nilfs_btree_node *node, int index)
 {
-	return nilfs_bmap_dptr_to_ptr(*(nilfs_btree_node_dptrs(btree, node) +
+	return nilfs_bmap_dptr_to_ptr(*(nilfs_btree_node_dptrs(node, btree) +
 					index));
 }
 
 static inline void
 nilfs_btree_node_set_ptr(struct nilfs_btree *btree,
-			 struct nilfs_btree_node *node,
-			 int index,
-			 __u64 ptr)
+			 struct nilfs_btree_node *node, int index, __u64 ptr)
 {
-	*(nilfs_btree_node_dptrs(btree, node) + index) =
+	*(nilfs_btree_node_dptrs(node, btree) + index) =
 		nilfs_bmap_ptr_to_dptr(ptr);
 }
 
@@ -283,12 +247,12 @@ static void nilfs_btree_node_init(struct nilfs_btree *btree,
 	__le64 *dptrs;
 	int i;
 
-	nilfs_btree_node_set_flags(btree, node, flags);
-	nilfs_btree_node_set_level(btree, node, level);
-	nilfs_btree_node_set_nchildren(btree, node, nchildren);
+	nilfs_btree_node_set_flags(node, flags);
+	nilfs_btree_node_set_level(node, level);
+	nilfs_btree_node_set_nchildren(node, nchildren);
 
-	dkeys = nilfs_btree_node_dkeys(btree, node);
-	dptrs = nilfs_btree_node_dptrs(btree, node);
+	dkeys = nilfs_btree_node_dkeys(node);
+	dptrs = nilfs_btree_node_dptrs(node, btree);
 	for (i = 0; i < nchildren; i++) {
 		dkeys[i] = nilfs_bmap_key_to_dkey(keys[i]);
 		dptrs[i] = nilfs_bmap_ptr_to_dptr(ptrs[i]);
@@ -305,13 +269,13 @@ static void nilfs_btree_node_move_left(struct nilfs_btree *btree,
 	__le64 *ldptrs, *rdptrs;
 	int lnchildren, rnchildren;
 
-	ldkeys = nilfs_btree_node_dkeys(btree, left);
-	ldptrs = nilfs_btree_node_dptrs(btree, left);
-	lnchildren = nilfs_btree_node_get_nchildren(btree, left);
+	ldkeys = nilfs_btree_node_dkeys(left);
+	ldptrs = nilfs_btree_node_dptrs(left, btree);
+	lnchildren = nilfs_btree_node_get_nchildren(left);
 
-	rdkeys = nilfs_btree_node_dkeys(btree, right);
-	rdptrs = nilfs_btree_node_dptrs(btree, right);
-	rnchildren = nilfs_btree_node_get_nchildren(btree, right);
+	rdkeys = nilfs_btree_node_dkeys(right);
+	rdptrs = nilfs_btree_node_dptrs(right, btree);
+	rnchildren = nilfs_btree_node_get_nchildren(right);
 
 	memcpy(ldkeys + lnchildren, rdkeys, n * sizeof(*rdkeys));
 	memcpy(ldptrs + lnchildren, rdptrs, n * sizeof(*rdptrs));
@@ -320,8 +284,8 @@ static void nilfs_btree_node_move_left(struct nilfs_btree *btree,
 
 	lnchildren += n;
 	rnchildren -= n;
-	nilfs_btree_node_set_nchildren(btree, left, lnchildren);
-	nilfs_btree_node_set_nchildren(btree, right, rnchildren);
+	nilfs_btree_node_set_nchildren(left, lnchildren);
+	nilfs_btree_node_set_nchildren(right, rnchildren);
 }
 
 /* Assume that the buffer heads corresponding to left and right are locked. */
@@ -334,13 +298,13 @@ static void nilfs_btree_node_move_right(struct nilfs_btree *btree,
 	__le64 *ldptrs, *rdptrs;
 	int lnchildren, rnchildren;
 
-	ldkeys = nilfs_btree_node_dkeys(btree, left);
-	ldptrs = nilfs_btree_node_dptrs(btree, left);
-	lnchildren = nilfs_btree_node_get_nchildren(btree, left);
+	ldkeys = nilfs_btree_node_dkeys(left);
+	ldptrs = nilfs_btree_node_dptrs(left, btree);
+	lnchildren = nilfs_btree_node_get_nchildren(left);
 
-	rdkeys = nilfs_btree_node_dkeys(btree, right);
-	rdptrs = nilfs_btree_node_dptrs(btree, right);
-	rnchildren = nilfs_btree_node_get_nchildren(btree, right);
+	rdkeys = nilfs_btree_node_dkeys(right);
+	rdptrs = nilfs_btree_node_dptrs(right, btree);
+	rnchildren = nilfs_btree_node_get_nchildren(right);
 
 	memmove(rdkeys + n, rdkeys, rnchildren * sizeof(*rdkeys));
 	memmove(rdptrs + n, rdptrs, rnchildren * sizeof(*rdptrs));
@@ -349,8 +313,8 @@ static void nilfs_btree_node_move_right(struct nilfs_btree *btree,
 
 	lnchildren -= n;
 	rnchildren += n;
-	nilfs_btree_node_set_nchildren(btree, left, lnchildren);
-	nilfs_btree_node_set_nchildren(btree, right, rnchildren);
+	nilfs_btree_node_set_nchildren(left, lnchildren);
+	nilfs_btree_node_set_nchildren(right, rnchildren);
 }
 
 /* Assume that the buffer head corresponding to node is locked. */
@@ -362,9 +326,9 @@ static void nilfs_btree_node_insert(struct nilfs_btree *btree,
 	__le64 *dptrs;
 	int nchildren;
 
-	dkeys = nilfs_btree_node_dkeys(btree, node);
-	dptrs = nilfs_btree_node_dptrs(btree, node);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
+	dkeys = nilfs_btree_node_dkeys(node);
+	dptrs = nilfs_btree_node_dptrs(node, btree);
+	nchildren = nilfs_btree_node_get_nchildren(node);
 	if (index < nchildren) {
 		memmove(dkeys + index + 1, dkeys + index,
 			(nchildren - index) * sizeof(*dkeys));
@@ -374,7 +338,7 @@ static void nilfs_btree_node_insert(struct nilfs_btree *btree,
 	dkeys[index] = nilfs_bmap_key_to_dkey(key);
 	dptrs[index] = nilfs_bmap_ptr_to_dptr(ptr);
 	nchildren++;
-	nilfs_btree_node_set_nchildren(btree, node, nchildren);
+	nilfs_btree_node_set_nchildren(node, nchildren);
 }
 
 /* Assume that the buffer head corresponding to node is locked. */
@@ -388,11 +352,11 @@ static void nilfs_btree_node_delete(struct nilfs_btree *btree,
 	__le64 *dptrs;
 	int nchildren;
 
-	dkeys = nilfs_btree_node_dkeys(btree, node);
-	dptrs = nilfs_btree_node_dptrs(btree, node);
+	dkeys = nilfs_btree_node_dkeys(node);
+	dptrs = nilfs_btree_node_dptrs(node, btree);
 	key = nilfs_bmap_dkey_to_key(dkeys[index]);
 	ptr = nilfs_bmap_dptr_to_ptr(dptrs[index]);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
+	nchildren = nilfs_btree_node_get_nchildren(node);
 	if (keyp != NULL)
 		*keyp = key;
 	if (ptrp != NULL)
@@ -405,11 +369,10 @@ static void nilfs_btree_node_delete(struct nilfs_btree *btree,
 			(nchildren - index - 1) * sizeof(*dptrs));
 	}
 	nchildren--;
-	nilfs_btree_node_set_nchildren(btree, node, nchildren);
+	nilfs_btree_node_set_nchildren(node, nchildren);
 }
 
-static int nilfs_btree_node_lookup(const struct nilfs_btree *btree,
-				   const struct nilfs_btree_node *node,
+static int nilfs_btree_node_lookup(const struct nilfs_btree_node *node,
 				   __u64 key, int *indexp)
 {
 	__u64 nkey;
@@ -417,12 +380,12 @@ static int nilfs_btree_node_lookup(const struct nilfs_btree *btree,
 
 	/* binary search */
 	low = 0;
-	high = nilfs_btree_node_get_nchildren(btree, node) - 1;
+	high = nilfs_btree_node_get_nchildren(node) - 1;
 	index = 0;
 	s = 0;
 	while (low <= high) {
 		index = (low + high) / 2;
-		nkey = nilfs_btree_node_get_key(btree, node, index);
+		nkey = nilfs_btree_node_get_key(node, index);
 		if (nkey == key) {
 			s = 0;
 			goto out;
@@ -436,9 +399,8 @@ static int nilfs_btree_node_lookup(const struct nilfs_btree *btree,
 	}
 
 	/* adjust index */
-	if (nilfs_btree_node_get_level(btree, node) >
-	    NILFS_BTREE_LEVEL_NODE_MIN) {
-		if ((s > 0) && (index > 0))
+	if (nilfs_btree_node_get_level(node) > NILFS_BTREE_LEVEL_NODE_MIN) {
+		if (s > 0 && index > 0)
 			index--;
 	} else if (s < 0)
 		index++;
@@ -456,25 +418,20 @@ nilfs_btree_get_root(const struct nilfs_btree *btree)
 }
 
 static inline struct nilfs_btree_node *
-nilfs_btree_get_nonroot_node(const struct nilfs_btree *btree,
-			     const struct nilfs_btree_path *path,
-			     int level)
+nilfs_btree_get_nonroot_node(const struct nilfs_btree_path *path, int level)
 {
 	return (struct nilfs_btree_node *)path[level].bp_bh->b_data;
 }
 
 static inline struct nilfs_btree_node *
-nilfs_btree_get_sib_node(const struct nilfs_btree *btree,
-			 const struct nilfs_btree_path *path,
-			 int level)
+nilfs_btree_get_sib_node(const struct nilfs_btree_path *path, int level)
 {
 	return (struct nilfs_btree_node *)path[level].bp_sib_bh->b_data;
 }
 
 static inline int nilfs_btree_height(const struct nilfs_btree *btree)
 {
-	return nilfs_btree_node_get_level(btree, nilfs_btree_get_root(btree))
-		+ 1;
+	return nilfs_btree_node_get_level(nilfs_btree_get_root(btree)) + 1;
 }
 
 static inline struct nilfs_btree_node *
@@ -484,7 +441,7 @@ nilfs_btree_get_node(const struct nilfs_btree *btree,
 {
 	return (level == nilfs_btree_height(btree) - 1) ?
 		nilfs_btree_get_root(btree) :
-		nilfs_btree_get_nonroot_node(btree, path, level);
+		nilfs_btree_get_nonroot_node(path, level);
 }
 
 static int nilfs_btree_do_lookup(const struct nilfs_btree *btree,
@@ -496,12 +453,11 @@ static int nilfs_btree_do_lookup(const struct nilfs_btree *btree,
 	int level, index, found, ret;
 
 	node = nilfs_btree_get_root(btree);
-	level = nilfs_btree_node_get_level(btree, node);
-	if ((level < minlevel) ||
-	    (nilfs_btree_node_get_nchildren(btree, node) <= 0))
+	level = nilfs_btree_node_get_level(node);
+	if (level < minlevel || nilfs_btree_node_get_nchildren(node) <= 0)
 		return -ENOENT;
 
-	found = nilfs_btree_node_lookup(btree, node, key, &index);
+	found = nilfs_btree_node_lookup(node, key, &index);
 	ptr = nilfs_btree_node_get_ptr(btree, node, index);
 	path[level].bp_bh = NULL;
 	path[level].bp_index = index;
@@ -510,14 +466,13 @@ static int nilfs_btree_do_lookup(const struct nilfs_btree *btree,
 		ret = nilfs_btree_get_block(btree, ptr, &path[level].bp_bh);
 		if (ret < 0)
 			return ret;
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
-		BUG_ON(level != nilfs_btree_node_get_level(btree, node));
+		node = nilfs_btree_get_nonroot_node(path, level);
+		BUG_ON(level != nilfs_btree_node_get_level(node));
 		if (!found)
-			found = nilfs_btree_node_lookup(btree, node, key,
-							&index);
+			found = nilfs_btree_node_lookup(node, key, &index);
 		else
 			index = 0;
-		if (index < nilfs_btree_node_nchildren_max(btree, node))
+		if (index < nilfs_btree_node_nchildren_max(node, btree))
 			ptr = nilfs_btree_node_get_ptr(btree, node, index);
 		else {
 			WARN_ON(found || level != NILFS_BTREE_LEVEL_NODE_MIN);
@@ -544,10 +499,10 @@ static int nilfs_btree_do_lookup_last(const struct nilfs_btree *btree,
 	int index, level, ret;
 
 	node = nilfs_btree_get_root(btree);
-	index = nilfs_btree_node_get_nchildren(btree, node) - 1;
+	index = nilfs_btree_node_get_nchildren(node) - 1;
 	if (index < 0)
 		return -ENOENT;
-	level = nilfs_btree_node_get_level(btree, node);
+	level = nilfs_btree_node_get_level(node);
 	ptr = nilfs_btree_node_get_ptr(btree, node, index);
 	path[level].bp_bh = NULL;
 	path[level].bp_index = index;
@@ -556,15 +511,15 @@ static int nilfs_btree_do_lookup_last(const struct nilfs_btree *btree,
 		ret = nilfs_btree_get_block(btree, ptr, &path[level].bp_bh);
 		if (ret < 0)
 			return ret;
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
-		BUG_ON(level != nilfs_btree_node_get_level(btree, node));
-		index = nilfs_btree_node_get_nchildren(btree, node) - 1;
+		node = nilfs_btree_get_nonroot_node(path, level);
+		BUG_ON(level != nilfs_btree_node_get_level(node));
+		index = nilfs_btree_node_get_nchildren(node) - 1;
 		ptr = nilfs_btree_node_get_ptr(btree, node, index);
 		path[level].bp_index = index;
 	}
 
 	if (keyp != NULL)
-		*keyp = nilfs_btree_node_get_key(btree, node, index);
+		*keyp = nilfs_btree_node_get_key(node, index);
 	if (ptrp != NULL)
 		*ptrp = ptr;
 
@@ -580,18 +535,18 @@ static int nilfs_btree_lookup(const struct nilfs_bmap *bmap,
 	int ret;
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 
 	ret = nilfs_btree_do_lookup(btree, path, key, &ptr, level);
 
 	if (ptrp != NULL)
 		*ptrp = ptr;
 
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 
 	return ret;
 }
@@ -608,10 +563,10 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *bmap,
 	int level = NILFS_BTREE_LEVEL_NODE_MIN;
 	int ret, cnt, index, maxlevel;
 
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 	ret = nilfs_btree_do_lookup(btree, path, key, &ptr, level);
 	if (ret < 0)
 		goto out;
@@ -631,8 +586,8 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *bmap,
 	node = nilfs_btree_get_node(btree, path, level);
 	index = path[level].bp_index + 1;
 	for (;;) {
-		while (index < nilfs_btree_node_get_nchildren(btree, node)) {
-			if (nilfs_btree_node_get_key(btree, node, index) !=
+		while (index < nilfs_btree_node_get_nchildren(node)) {
+			if (nilfs_btree_node_get_key(node, index) !=
 			    key + cnt)
 				goto end;
 			ptr2 = nilfs_btree_node_get_ptr(btree, node, index);
@@ -653,8 +608,8 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *bmap,
 		/* look-up right sibling node */
 		node = nilfs_btree_get_node(btree, path, level + 1);
 		index = path[level + 1].bp_index + 1;
-		if (index >= nilfs_btree_node_get_nchildren(btree, node) ||
-		    nilfs_btree_node_get_key(btree, node, index) != key + cnt)
+		if (index >= nilfs_btree_node_get_nchildren(node) ||
+		    nilfs_btree_node_get_key(node, index) != key + cnt)
 			break;
 		ptr2 = nilfs_btree_node_get_ptr(btree, node, index);
 		path[level + 1].bp_index = index;
@@ -664,7 +619,7 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *bmap,
 		ret = nilfs_btree_get_block(btree, ptr2, &path[level].bp_bh);
 		if (ret < 0)
 			goto out;
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
+		node = nilfs_btree_get_nonroot_node(path, level);
 		index = 0;
 		path[level].bp_index = index;
 	}
@@ -672,8 +627,8 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *bmap,
 	*ptrp = ptr;
 	ret = cnt;
  out:
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 	return ret;
 }
 
@@ -685,9 +640,7 @@ static void nilfs_btree_promote_key(struct nilfs_btree *btree,
 		do {
 			lock_buffer(path[level].bp_bh);
 			nilfs_btree_node_set_key(
-				btree,
-				nilfs_btree_get_nonroot_node(
-					btree, path, level),
+				nilfs_btree_get_nonroot_node(path, level),
 				path[level].bp_index, key);
 			if (!buffer_dirty(path[level].bp_bh))
 				nilfs_btnode_mark_dirty(path[level].bp_bh);
@@ -698,8 +651,7 @@ static void nilfs_btree_promote_key(struct nilfs_btree *btree,
 
 	/* root */
 	if (level == nilfs_btree_height(btree) - 1) {
-		nilfs_btree_node_set_key(btree,
-					 nilfs_btree_get_root(btree),
+		nilfs_btree_node_set_key(nilfs_btree_get_root(btree),
 					 path[level].bp_index, key);
 	}
 }
@@ -712,7 +664,7 @@ static void nilfs_btree_do_insert(struct nilfs_btree *btree,
 
 	if (level < nilfs_btree_height(btree) - 1) {
 		lock_buffer(path[level].bp_bh);
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
+		node = nilfs_btree_get_nonroot_node(path, level);
 		nilfs_btree_node_insert(btree, node, *keyp, *ptrp,
 					path[level].bp_index);
 		if (!buffer_dirty(path[level].bp_bh))
@@ -721,8 +673,8 @@ static void nilfs_btree_do_insert(struct nilfs_btree *btree,
 
 		if (path[level].bp_index == 0)
 			nilfs_btree_promote_key(btree, path, level + 1,
-						nilfs_btree_node_get_key(
-							btree, node, 0));
+						nilfs_btree_node_get_key(node,
+									 0));
 	} else {
 		node = nilfs_btree_get_root(btree);
 		nilfs_btree_node_insert(btree, node, *keyp, *ptrp,
@@ -740,10 +692,10 @@ static void nilfs_btree_carry_left(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	left = nilfs_btree_get_sib_node(btree, path, level);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
-	lnchildren = nilfs_btree_node_get_nchildren(btree, left);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	left = nilfs_btree_get_sib_node(path, level);
+	nchildren = nilfs_btree_node_get_nchildren(node);
+	lnchildren = nilfs_btree_node_get_nchildren(left);
 	move = 0;
 
 	n = (nchildren + lnchildren + 1) / 2 - lnchildren;
@@ -764,7 +716,7 @@ static void nilfs_btree_carry_left(struct nilfs_btree *btree,
 	unlock_buffer(path[level].bp_sib_bh);
 
 	nilfs_btree_promote_key(btree, path, level + 1,
-				nilfs_btree_node_get_key(btree, node, 0));
+				nilfs_btree_node_get_key(node, 0));
 
 	if (move) {
 		brelse(path[level].bp_bh);
@@ -791,10 +743,10 @@ static void nilfs_btree_carry_right(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	right = nilfs_btree_get_sib_node(btree, path, level);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
-	rnchildren = nilfs_btree_node_get_nchildren(btree, right);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	right = nilfs_btree_get_sib_node(path, level);
+	nchildren = nilfs_btree_node_get_nchildren(node);
+	rnchildren = nilfs_btree_node_get_nchildren(right);
 	move = 0;
 
 	n = (nchildren + rnchildren + 1) / 2 - rnchildren;
@@ -816,15 +768,14 @@ static void nilfs_btree_carry_right(struct nilfs_btree *btree,
 
 	path[level + 1].bp_index++;
 	nilfs_btree_promote_key(btree, path, level + 1,
-				nilfs_btree_node_get_key(btree, right, 0));
+				nilfs_btree_node_get_key(right, 0));
 	path[level + 1].bp_index--;
 
 	if (move) {
 		brelse(path[level].bp_bh);
 		path[level].bp_bh = path[level].bp_sib_bh;
 		path[level].bp_sib_bh = NULL;
-		path[level].bp_index -=
-			nilfs_btree_node_get_nchildren(btree, node);
+		path[level].bp_index -= nilfs_btree_node_get_nchildren(node);
 		path[level + 1].bp_index++;
 	} else {
 		brelse(path[level].bp_sib_bh);
@@ -846,9 +797,9 @@ static void nilfs_btree_split(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	right = nilfs_btree_get_sib_node(btree, path, level);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	right = nilfs_btree_get_sib_node(path, level);
+	nchildren = nilfs_btree_node_get_nchildren(node);
 	move = 0;
 
 	n = (nchildren + 1) / 2;
@@ -867,16 +818,15 @@ static void nilfs_btree_split(struct nilfs_btree *btree,
 	unlock_buffer(path[level].bp_bh);
 	unlock_buffer(path[level].bp_sib_bh);
 
-	newkey = nilfs_btree_node_get_key(btree, right, 0);
+	newkey = nilfs_btree_node_get_key(right, 0);
 	newptr = path[level].bp_newreq.bpr_ptr;
 
 	if (move) {
-		path[level].bp_index -=
-			nilfs_btree_node_get_nchildren(btree, node);
+		path[level].bp_index -= nilfs_btree_node_get_nchildren(node);
 		nilfs_btree_node_insert(btree, right, *keyp, *ptrp,
 					path[level].bp_index);
 
-		*keyp = nilfs_btree_node_get_key(btree, right, 0);
+		*keyp = nilfs_btree_node_get_key(right, 0);
 		*ptrp = path[level].bp_newreq.bpr_ptr;
 
 		brelse(path[level].bp_bh);
@@ -885,7 +835,7 @@ static void nilfs_btree_split(struct nilfs_btree *btree,
 	} else {
 		nilfs_btree_do_insert(btree, path, level, keyp, ptrp);
 
-		*keyp = nilfs_btree_node_get_key(btree, right, 0);
+		*keyp = nilfs_btree_node_get_key(right, 0);
 		*ptrp = path[level].bp_newreq.bpr_ptr;
 
 		brelse(path[level].bp_sib_bh);
@@ -905,12 +855,12 @@ static void nilfs_btree_grow(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_sib_bh);
 
 	root = nilfs_btree_get_root(btree);
-	child = nilfs_btree_get_sib_node(btree, path, level);
+	child = nilfs_btree_get_sib_node(path, level);
 
-	n = nilfs_btree_node_get_nchildren(btree, root);
+	n = nilfs_btree_node_get_nchildren(root);
 
 	nilfs_btree_node_move_right(btree, root, child, n);
-	nilfs_btree_node_set_level(btree, root, level + 1);
+	nilfs_btree_node_set_level(root, level + 1);
 
 	if (!buffer_dirty(path[level].bp_sib_bh))
 		nilfs_btnode_mark_dirty(path[level].bp_sib_bh);
@@ -922,7 +872,7 @@ static void nilfs_btree_grow(struct nilfs_btree *btree,
 
 	nilfs_btree_do_insert(btree, path, level, keyp, ptrp);
 
-	*keyp = nilfs_btree_node_get_key(btree, child, 0);
+	*keyp = nilfs_btree_node_get_key(child, 0);
 	*ptrp = path[level].bp_newreq.bpr_ptr;
 }
 
@@ -990,26 +940,29 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 	struct nilfs_btree_node *node, *parent, *sib;
 	__u64 sibptr;
 	int pindex, level, ret;
+	struct inode *dat = NULL;
 
 	stats->bs_nblocks = 0;
 	level = NILFS_BTREE_LEVEL_DATA;
 
 	/* allocate a new ptr for data block */
-	if (NILFS_BMAP_USE_VBN(&btree->bt_bmap))
+	if (NILFS_BMAP_USE_VBN(&btree->bt_bmap)) {
 		path[level].bp_newreq.bpr_ptr =
 			nilfs_btree_find_target_v(btree, path, key);
+		dat = nilfs_bmap_get_dat(&btree->bt_bmap);
+	}
 
 	ret = nilfs_bmap_prepare_alloc_ptr(&btree->bt_bmap,
-					   &path[level].bp_newreq);
+					   &path[level].bp_newreq, dat);
 	if (ret < 0)
 		goto err_out_data;
 
 	for (level = NILFS_BTREE_LEVEL_NODE_MIN;
 	     level < nilfs_btree_height(btree) - 1;
 	     level++) {
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
-		if (nilfs_btree_node_get_nchildren(btree, node) <
-		    nilfs_btree_node_nchildren_max(btree, node)) {
+		node = nilfs_btree_get_nonroot_node(path, level);
+		if (nilfs_btree_node_get_nchildren(node) <
+		    nilfs_btree_node_nchildren_max(node, btree)) {
 			path[level].bp_op = nilfs_btree_do_insert;
 			stats->bs_nblocks++;
 			goto out;
@@ -1026,8 +979,8 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 			if (ret < 0)
 				goto err_out_child_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
-			if (nilfs_btree_node_get_nchildren(btree, sib) <
-			    nilfs_btree_node_nchildren_max(btree, sib)) {
+			if (nilfs_btree_node_get_nchildren(sib) <
+			    nilfs_btree_node_nchildren_max(sib, btree)) {
 				path[level].bp_sib_bh = bh;
 				path[level].bp_op = nilfs_btree_carry_left;
 				stats->bs_nblocks++;
@@ -1038,15 +991,15 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 
 		/* right sibling */
 		if (pindex <
-		    nilfs_btree_node_get_nchildren(btree, parent) - 1) {
+		    nilfs_btree_node_get_nchildren(parent) - 1) {
 			sibptr = nilfs_btree_node_get_ptr(btree, parent,
 							  pindex + 1);
 			ret = nilfs_btree_get_block(btree, sibptr, &bh);
 			if (ret < 0)
 				goto err_out_child_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
-			if (nilfs_btree_node_get_nchildren(btree, sib) <
-			    nilfs_btree_node_nchildren_max(btree, sib)) {
+			if (nilfs_btree_node_get_nchildren(sib) <
+			    nilfs_btree_node_nchildren_max(sib, btree)) {
 				path[level].bp_sib_bh = bh;
 				path[level].bp_op = nilfs_btree_carry_right;
 				stats->bs_nblocks++;
@@ -1059,7 +1012,7 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 		path[level].bp_newreq.bpr_ptr =
 			path[level - 1].bp_newreq.bpr_ptr + 1;
 		ret = nilfs_bmap_prepare_alloc_ptr(&btree->bt_bmap,
-						   &path[level].bp_newreq);
+						   &path[level].bp_newreq, dat);
 		if (ret < 0)
 			goto err_out_child_node;
 		ret = nilfs_btree_get_new_block(btree,
@@ -1081,8 +1034,8 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 
 	/* root */
 	node = nilfs_btree_get_root(btree);
-	if (nilfs_btree_node_get_nchildren(btree, node) <
-	    nilfs_btree_node_nchildren_max(btree, node)) {
+	if (nilfs_btree_node_get_nchildren(node) <
+	    nilfs_btree_node_nchildren_max(node, btree)) {
 		path[level].bp_op = nilfs_btree_do_insert;
 		stats->bs_nblocks++;
 		goto out;
@@ -1091,7 +1044,7 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 	/* grow */
 	path[level].bp_newreq.bpr_ptr = path[level - 1].bp_newreq.bpr_ptr + 1;
 	ret = nilfs_bmap_prepare_alloc_ptr(&btree->bt_bmap,
-					   &path[level].bp_newreq);
+					   &path[level].bp_newreq, dat);
 	if (ret < 0)
 		goto err_out_child_node;
 	ret = nilfs_btree_get_new_block(btree, path[level].bp_newreq.bpr_ptr,
@@ -1119,16 +1072,18 @@ static int nilfs_btree_prepare_insert(struct nilfs_btree *btree,
 
 	/* error */
  err_out_curr_node:
-	nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap, &path[level].bp_newreq);
+	nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap, &path[level].bp_newreq,
+				   dat);
  err_out_child_node:
 	for (level--; level > NILFS_BTREE_LEVEL_DATA; level--) {
 		nilfs_btnode_delete(path[level].bp_sib_bh);
 		nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap,
-					   &path[level].bp_newreq);
+					   &path[level].bp_newreq, dat);
 
 	}
 
-	nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap, &path[level].bp_newreq);
+	nilfs_bmap_abort_alloc_ptr(&btree->bt_bmap, &path[level].bp_newreq,
+				   dat);
  err_out_data:
 	*levelp = level;
 	stats->bs_nblocks = 0;
@@ -1139,16 +1094,19 @@ static void nilfs_btree_commit_insert(struct nilfs_btree *btree,
 				      struct nilfs_btree_path *path,
 				      int maxlevel, __u64 key, __u64 ptr)
 {
+	struct inode *dat = NULL;
 	int level;
 
 	set_buffer_nilfs_volatile((struct buffer_head *)((unsigned long)ptr));
 	ptr = path[NILFS_BTREE_LEVEL_DATA].bp_newreq.bpr_ptr;
-	if (NILFS_BMAP_USE_VBN(&btree->bt_bmap))
+	if (NILFS_BMAP_USE_VBN(&btree->bt_bmap)) {
 		nilfs_btree_set_target_v(btree, key, ptr);
+		dat = nilfs_bmap_get_dat(&btree->bt_bmap);
+	}
 
 	for (level = NILFS_BTREE_LEVEL_NODE_MIN; level <= maxlevel; level++) {
 		nilfs_bmap_commit_alloc_ptr(&btree->bt_bmap,
-					    &path[level - 1].bp_newreq);
+					    &path[level - 1].bp_newreq, dat);
 		path[level].bp_op(btree, path, level, &key, &ptr);
 	}
 
@@ -1164,10 +1122,10 @@ static int nilfs_btree_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr)
 	int level, ret;
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 
 	ret = nilfs_btree_do_lookup(btree, path, key, NULL,
 				    NILFS_BTREE_LEVEL_NODE_MIN);
@@ -1184,8 +1142,8 @@ static int nilfs_btree_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr)
 	nilfs_bmap_add_blocks(bmap, stats.bs_nblocks);
 
  out:
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 	return ret;
 }
 
@@ -1197,7 +1155,7 @@ static void nilfs_btree_do_delete(struct nilfs_btree *btree,
 
 	if (level < nilfs_btree_height(btree) - 1) {
 		lock_buffer(path[level].bp_bh);
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
+		node = nilfs_btree_get_nonroot_node(path, level);
 		nilfs_btree_node_delete(btree, node, keyp, ptrp,
 					path[level].bp_index);
 		if (!buffer_dirty(path[level].bp_bh))
@@ -1205,7 +1163,7 @@ static void nilfs_btree_do_delete(struct nilfs_btree *btree,
 		unlock_buffer(path[level].bp_bh);
 		if (path[level].bp_index == 0)
 			nilfs_btree_promote_key(btree, path, level + 1,
-				nilfs_btree_node_get_key(btree, node, 0));
+				nilfs_btree_node_get_key(node, 0));
 	} else {
 		node = nilfs_btree_get_root(btree);
 		nilfs_btree_node_delete(btree, node, keyp, ptrp,
@@ -1225,10 +1183,10 @@ static void nilfs_btree_borrow_left(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	left = nilfs_btree_get_sib_node(btree, path, level);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
-	lnchildren = nilfs_btree_node_get_nchildren(btree, left);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	left = nilfs_btree_get_sib_node(path, level);
+	nchildren = nilfs_btree_node_get_nchildren(node);
+	lnchildren = nilfs_btree_node_get_nchildren(left);
 
 	n = (nchildren + lnchildren) / 2 - nchildren;
 
@@ -1243,7 +1201,7 @@ static void nilfs_btree_borrow_left(struct nilfs_btree *btree,
 	unlock_buffer(path[level].bp_sib_bh);
 
 	nilfs_btree_promote_key(btree, path, level + 1,
-				nilfs_btree_node_get_key(btree, node, 0));
+				nilfs_btree_node_get_key(node, 0));
 
 	brelse(path[level].bp_sib_bh);
 	path[level].bp_sib_bh = NULL;
@@ -1262,10 +1220,10 @@ static void nilfs_btree_borrow_right(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	right = nilfs_btree_get_sib_node(btree, path, level);
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
-	rnchildren = nilfs_btree_node_get_nchildren(btree, right);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	right = nilfs_btree_get_sib_node(path, level);
+	nchildren = nilfs_btree_node_get_nchildren(node);
+	rnchildren = nilfs_btree_node_get_nchildren(right);
 
 	n = (nchildren + rnchildren) / 2 - nchildren;
 
@@ -1281,7 +1239,7 @@ static void nilfs_btree_borrow_right(struct nilfs_btree *btree,
 
 	path[level + 1].bp_index++;
 	nilfs_btree_promote_key(btree, path, level + 1,
-				nilfs_btree_node_get_key(btree, right, 0));
+				nilfs_btree_node_get_key(right, 0));
 	path[level + 1].bp_index--;
 
 	brelse(path[level].bp_sib_bh);
@@ -1300,10 +1258,10 @@ static void nilfs_btree_concat_left(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	left = nilfs_btree_get_sib_node(btree, path, level);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	left = nilfs_btree_get_sib_node(path, level);
 
-	n = nilfs_btree_node_get_nchildren(btree, node);
+	n = nilfs_btree_node_get_nchildren(node);
 
 	nilfs_btree_node_move_left(btree, left, node, n);
 
@@ -1316,7 +1274,7 @@ static void nilfs_btree_concat_left(struct nilfs_btree *btree,
 	nilfs_btnode_delete(path[level].bp_bh);
 	path[level].bp_bh = path[level].bp_sib_bh;
 	path[level].bp_sib_bh = NULL;
-	path[level].bp_index += nilfs_btree_node_get_nchildren(btree, left);
+	path[level].bp_index += nilfs_btree_node_get_nchildren(left);
 }
 
 static void nilfs_btree_concat_right(struct nilfs_btree *btree,
@@ -1331,10 +1289,10 @@ static void nilfs_btree_concat_right(struct nilfs_btree *btree,
 	lock_buffer(path[level].bp_bh);
 	lock_buffer(path[level].bp_sib_bh);
 
-	node = nilfs_btree_get_nonroot_node(btree, path, level);
-	right = nilfs_btree_get_sib_node(btree, path, level);
+	node = nilfs_btree_get_nonroot_node(path, level);
+	right = nilfs_btree_get_sib_node(path, level);
 
-	n = nilfs_btree_node_get_nchildren(btree, right);
+	n = nilfs_btree_node_get_nchildren(right);
 
 	nilfs_btree_node_move_left(btree, node, right, n);
 
@@ -1360,11 +1318,11 @@ static void nilfs_btree_shrink(struct nilfs_btree *btree,
 
 	lock_buffer(path[level].bp_bh);
 	root = nilfs_btree_get_root(btree);
-	child = nilfs_btree_get_nonroot_node(btree, path, level);
+	child = nilfs_btree_get_nonroot_node(path, level);
 
 	nilfs_btree_node_delete(btree, root, NULL, NULL, 0);
-	nilfs_btree_node_set_level(btree, root, level);
-	n = nilfs_btree_node_get_nchildren(btree, child);
+	nilfs_btree_node_set_level(root, level);
+	n = nilfs_btree_node_get_nchildren(child);
 	nilfs_btree_node_move_left(btree, root, child, n);
 	unlock_buffer(path[level].bp_bh);
 
@@ -1376,7 +1334,8 @@ static void nilfs_btree_shrink(struct nilfs_btree *btree,
 static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 				      struct nilfs_btree_path *path,
 				      int *levelp,
-				      struct nilfs_bmap_stats *stats)
+				      struct nilfs_bmap_stats *stats,
+				      struct inode *dat)
 {
 	struct buffer_head *bh;
 	struct nilfs_btree_node *node, *parent, *sib;
@@ -1388,17 +1347,17 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 	for (level = NILFS_BTREE_LEVEL_NODE_MIN;
 	     level < nilfs_btree_height(btree) - 1;
 	     level++) {
-		node = nilfs_btree_get_nonroot_node(btree, path, level);
+		node = nilfs_btree_get_nonroot_node(path, level);
 		path[level].bp_oldreq.bpr_ptr =
 			nilfs_btree_node_get_ptr(btree, node,
 						 path[level].bp_index);
 		ret = nilfs_bmap_prepare_end_ptr(&btree->bt_bmap,
-						 &path[level].bp_oldreq);
+						 &path[level].bp_oldreq, dat);
 		if (ret < 0)
 			goto err_out_child_node;
 
-		if (nilfs_btree_node_get_nchildren(btree, node) >
-		    nilfs_btree_node_nchildren_min(btree, node)) {
+		if (nilfs_btree_node_get_nchildren(node) >
+		    nilfs_btree_node_nchildren_min(node, btree)) {
 			path[level].bp_op = nilfs_btree_do_delete;
 			stats->bs_nblocks++;
 			goto out;
@@ -1415,8 +1374,8 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 			if (ret < 0)
 				goto err_out_curr_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
-			if (nilfs_btree_node_get_nchildren(btree, sib) >
-			    nilfs_btree_node_nchildren_min(btree, sib)) {
+			if (nilfs_btree_node_get_nchildren(sib) >
+			    nilfs_btree_node_nchildren_min(sib, btree)) {
 				path[level].bp_sib_bh = bh;
 				path[level].bp_op = nilfs_btree_borrow_left;
 				stats->bs_nblocks++;
@@ -1428,7 +1387,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 				/* continue; */
 			}
 		} else if (pindex <
-			   nilfs_btree_node_get_nchildren(btree, parent) - 1) {
+			   nilfs_btree_node_get_nchildren(parent) - 1) {
 			/* right sibling */
 			sibptr = nilfs_btree_node_get_ptr(btree, parent,
 							  pindex + 1);
@@ -1436,8 +1395,8 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 			if (ret < 0)
 				goto err_out_curr_node;
 			sib = (struct nilfs_btree_node *)bh->b_data;
-			if (nilfs_btree_node_get_nchildren(btree, sib) >
-			    nilfs_btree_node_nchildren_min(btree, sib)) {
+			if (nilfs_btree_node_get_nchildren(sib) >
+			    nilfs_btree_node_nchildren_min(sib, btree)) {
 				path[level].bp_sib_bh = bh;
 				path[level].bp_op = nilfs_btree_borrow_right;
 				stats->bs_nblocks++;
@@ -1452,7 +1411,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 			/* no siblings */
 			/* the only child of the root node */
 			WARN_ON(level != nilfs_btree_height(btree) - 2);
-			if (nilfs_btree_node_get_nchildren(btree, node) - 1 <=
+			if (nilfs_btree_node_get_nchildren(node) - 1 <=
 			    NILFS_BTREE_ROOT_NCHILDREN_MAX) {
 				path[level].bp_op = nilfs_btree_shrink;
 				stats->bs_nblocks += 2;
@@ -1471,7 +1430,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 		nilfs_btree_node_get_ptr(btree, node, path[level].bp_index);
 
 	ret = nilfs_bmap_prepare_end_ptr(&btree->bt_bmap,
-					 &path[level].bp_oldreq);
+					 &path[level].bp_oldreq, dat);
 	if (ret < 0)
 		goto err_out_child_node;
 
@@ -1486,12 +1445,12 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 
 	/* error */
  err_out_curr_node:
-	nilfs_bmap_abort_end_ptr(&btree->bt_bmap, &path[level].bp_oldreq);
+	nilfs_bmap_abort_end_ptr(&btree->bt_bmap, &path[level].bp_oldreq, dat);
  err_out_child_node:
 	for (level--; level >= NILFS_BTREE_LEVEL_NODE_MIN; level--) {
 		brelse(path[level].bp_sib_bh);
 		nilfs_bmap_abort_end_ptr(&btree->bt_bmap,
-					 &path[level].bp_oldreq);
+					 &path[level].bp_oldreq, dat);
 	}
 	*levelp = level;
 	stats->bs_nblocks = 0;
@@ -1500,13 +1459,13 @@ static int nilfs_btree_prepare_delete(struct nilfs_btree *btree,
 
 static void nilfs_btree_commit_delete(struct nilfs_btree *btree,
 				      struct nilfs_btree_path *path,
-				      int maxlevel)
+				      int maxlevel, struct inode *dat)
 {
 	int level;
 
 	for (level = NILFS_BTREE_LEVEL_NODE_MIN; level <= maxlevel; level++) {
 		nilfs_bmap_commit_end_ptr(&btree->bt_bmap,
-					  &path[level].bp_oldreq);
+					  &path[level].bp_oldreq, dat);
 		path[level].bp_op(btree, path, level, NULL, NULL);
 	}
 
@@ -1520,27 +1479,32 @@ static int nilfs_btree_delete(struct nilfs_bmap *bmap, __u64 key)
 	struct nilfs_btree *btree;
 	struct nilfs_btree_path *path;
 	struct nilfs_bmap_stats stats;
+	struct inode *dat;
 	int level, ret;
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 	ret = nilfs_btree_do_lookup(btree, path, key, NULL,
 				    NILFS_BTREE_LEVEL_NODE_MIN);
 	if (ret < 0)
 		goto out;
 
-	ret = nilfs_btree_prepare_delete(btree, path, &level, &stats);
+
+	dat = NILFS_BMAP_USE_VBN(&btree->bt_bmap) ?
+		nilfs_bmap_get_dat(&btree->bt_bmap) : NULL;
+
+	ret = nilfs_btree_prepare_delete(btree, path, &level, &stats, dat);
 	if (ret < 0)
 		goto out;
-	nilfs_btree_commit_delete(btree, path, level);
+	nilfs_btree_commit_delete(btree, path, level, dat);
 	nilfs_bmap_sub_blocks(bmap, stats.bs_nblocks);
 
 out:
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 	return ret;
 }
 
@@ -1551,15 +1515,15 @@ static int nilfs_btree_last_key(const struct nilfs_bmap *bmap, __u64 *keyp)
 	int ret;
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 
 	ret = nilfs_btree_do_lookup_last(btree, path, keyp, NULL);
 
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 
 	return ret;
 }
@@ -1581,7 +1545,7 @@ static int nilfs_btree_check_delete(struct nilfs_bmap *bmap, __u64 key)
 		node = root;
 		break;
 	case 3:
-		nchildren = nilfs_btree_node_get_nchildren(btree, root);
+		nchildren = nilfs_btree_node_get_nchildren(root);
 		if (nchildren > 1)
 			return 0;
 		ptr = nilfs_btree_node_get_ptr(btree, root, nchildren - 1);
@@ -1594,10 +1558,10 @@ static int nilfs_btree_check_delete(struct nilfs_bmap *bmap, __u64 key)
 		return 0;
 	}
 
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
-	maxkey = nilfs_btree_node_get_key(btree, node, nchildren - 1);
+	nchildren = nilfs_btree_node_get_nchildren(node);
+	maxkey = nilfs_btree_node_get_key(node, nchildren - 1);
 	nextmaxkey = (nchildren > 1) ?
-		nilfs_btree_node_get_key(btree, node, nchildren - 2) : 0;
+		nilfs_btree_node_get_key(node, nchildren - 2) : 0;
 	if (bh != NULL)
 		brelse(bh);
 
@@ -1623,7 +1587,7 @@ static int nilfs_btree_gather_data(struct nilfs_bmap *bmap,
 		node = root;
 		break;
 	case 3:
-		nchildren = nilfs_btree_node_get_nchildren(btree, root);
+		nchildren = nilfs_btree_node_get_nchildren(root);
 		WARN_ON(nchildren > 1);
 		ptr = nilfs_btree_node_get_ptr(btree, root, nchildren - 1);
 		ret = nilfs_btree_get_block(btree, ptr, &bh);
@@ -1636,11 +1600,11 @@ static int nilfs_btree_gather_data(struct nilfs_bmap *bmap,
 		return -EINVAL;
 	}
 
-	nchildren = nilfs_btree_node_get_nchildren(btree, node);
+	nchildren = nilfs_btree_node_get_nchildren(node);
 	if (nchildren < nitems)
 		nitems = nchildren;
-	dkeys = nilfs_btree_node_dkeys(btree, node);
-	dptrs = nilfs_btree_node_dptrs(btree, node);
+	dkeys = nilfs_btree_node_dkeys(node);
+	dptrs = nilfs_btree_node_dptrs(node, btree);
 	for (i = 0; i < nitems; i++) {
 		keys[i] = nilfs_bmap_dkey_to_key(dkeys[i]);
 		ptrs[i] = nilfs_bmap_dptr_to_ptr(dptrs[i]);
@@ -1660,18 +1624,20 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *bmap, __u64 key,
 				       struct nilfs_bmap_stats *stats)
 {
 	struct buffer_head *bh;
-	struct nilfs_btree *btree;
+	struct nilfs_btree *btree = (struct nilfs_btree *)bmap;
+	struct inode *dat = NULL;
 	int ret;
 
-	btree = (struct nilfs_btree *)bmap;
 	stats->bs_nblocks = 0;
 
 	/* for data */
 	/* cannot find near ptr */
-	if (NILFS_BMAP_USE_VBN(bmap))
+	if (NILFS_BMAP_USE_VBN(bmap)) {
 		dreq->bpr_ptr = nilfs_btree_find_target_v(btree, NULL, key);
+		dat = nilfs_bmap_get_dat(bmap);
+	}
 
-	ret = nilfs_bmap_prepare_alloc_ptr(bmap, dreq);
+	ret = nilfs_bmap_prepare_alloc_ptr(bmap, dreq, dat);
 	if (ret < 0)
 		return ret;
 
@@ -1679,7 +1645,7 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *bmap, __u64 key,
 	stats->bs_nblocks++;
 	if (nreq != NULL) {
 		nreq->bpr_ptr = dreq->bpr_ptr + 1;
-		ret = nilfs_bmap_prepare_alloc_ptr(bmap, nreq);
+		ret = nilfs_bmap_prepare_alloc_ptr(bmap, nreq, dat);
 		if (ret < 0)
 			goto err_out_dreq;
 
@@ -1696,9 +1662,9 @@ nilfs_btree_prepare_convert_and_insert(struct nilfs_bmap *bmap, __u64 key,
 
 	/* error */
  err_out_nreq:
-	nilfs_bmap_abort_alloc_ptr(bmap, nreq);
+	nilfs_bmap_abort_alloc_ptr(bmap, nreq, dat);
  err_out_dreq:
-	nilfs_bmap_abort_alloc_ptr(bmap, dreq);
+	nilfs_bmap_abort_alloc_ptr(bmap, dreq, dat);
 	stats->bs_nblocks = 0;
 	return ret;
 
@@ -1713,8 +1679,9 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 				      union nilfs_bmap_ptr_req *nreq,
 				      struct buffer_head *bh)
 {
-	struct nilfs_btree *btree;
+	struct nilfs_btree *btree = (struct nilfs_btree *)bmap;
 	struct nilfs_btree_node *node;
+	struct inode *dat;
 	__u64 tmpptr;
 
 	/* free resources */
@@ -1725,11 +1692,11 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 	set_buffer_nilfs_volatile((struct buffer_head *)((unsigned long)ptr));
 
 	/* convert and insert */
-	btree = (struct nilfs_btree *)bmap;
+	dat = NILFS_BMAP_USE_VBN(bmap) ? nilfs_bmap_get_dat(bmap) : NULL;
 	nilfs_btree_init(bmap);
 	if (nreq != NULL) {
-		nilfs_bmap_commit_alloc_ptr(bmap, dreq);
-		nilfs_bmap_commit_alloc_ptr(bmap, nreq);
+		nilfs_bmap_commit_alloc_ptr(bmap, dreq, dat);
+		nilfs_bmap_commit_alloc_ptr(bmap, nreq, dat);
 
 		/* create child node at level 1 */
 		lock_buffer(bh);
@@ -1751,7 +1718,7 @@ nilfs_btree_commit_convert_and_insert(struct nilfs_bmap *bmap,
 		nilfs_btree_node_init(btree, node, NILFS_BTREE_NODE_ROOT,
 				      2, 1, &keys[0], &tmpptr);
 	} else {
-		nilfs_bmap_commit_alloc_ptr(bmap, dreq);
+		nilfs_bmap_commit_alloc_ptr(bmap, dreq, dat);
 
 		/* create root node at level 1 */
 		node = nilfs_btree_get_root(btree);
@@ -1822,7 +1789,7 @@ static int nilfs_btree_propagate_p(struct nilfs_btree *btree,
 
 static int nilfs_btree_prepare_update_v(struct nilfs_btree *btree,
 					struct nilfs_btree_path *path,
-					int level)
+					int level, struct inode *dat)
 {
 	struct nilfs_btree_node *parent;
 	int ret;
@@ -1832,9 +1799,8 @@ static int nilfs_btree_prepare_update_v(struct nilfs_btree *btree,
 		nilfs_btree_node_get_ptr(btree, parent,
 					 path[level + 1].bp_index);
 	path[level].bp_newreq.bpr_ptr = path[level].bp_oldreq.bpr_ptr + 1;
-	ret = nilfs_bmap_prepare_update_v(&btree->bt_bmap,
-					  &path[level].bp_oldreq,
-					  &path[level].bp_newreq);
+	ret = nilfs_dat_prepare_update(dat, &path[level].bp_oldreq.bpr_req,
+				       &path[level].bp_newreq.bpr_req);
 	if (ret < 0)
 		return ret;
 
@@ -1846,9 +1812,9 @@ static int nilfs_btree_prepare_update_v(struct nilfs_btree *btree,
 			&NILFS_BMAP_I(&btree->bt_bmap)->i_btnode_cache,
 			&path[level].bp_ctxt);
 		if (ret < 0) {
-			nilfs_bmap_abort_update_v(&btree->bt_bmap,
-						  &path[level].bp_oldreq,
-						  &path[level].bp_newreq);
+			nilfs_dat_abort_update(dat,
+					       &path[level].bp_oldreq.bpr_req,
+					       &path[level].bp_newreq.bpr_req);
 			return ret;
 		}
 	}
@@ -1858,13 +1824,13 @@ static int nilfs_btree_prepare_update_v(struct nilfs_btree *btree,
 
 static void nilfs_btree_commit_update_v(struct nilfs_btree *btree,
 					struct nilfs_btree_path *path,
-					int level)
+					int level, struct inode *dat)
 {
 	struct nilfs_btree_node *parent;
 
-	nilfs_bmap_commit_update_v(&btree->bt_bmap,
-				   &path[level].bp_oldreq,
-				   &path[level].bp_newreq);
+	nilfs_dat_commit_update(dat, &path[level].bp_oldreq.bpr_req,
+				&path[level].bp_newreq.bpr_req,
+				btree->bt_bmap.b_ptr_type == NILFS_BMAP_PTR_VS);
 
 	if (buffer_nilfs_node(path[level].bp_bh)) {
 		nilfs_btnode_commit_change_key(
@@ -1881,11 +1847,10 @@ static void nilfs_btree_commit_update_v(struct nilfs_btree *btree,
 
 static void nilfs_btree_abort_update_v(struct nilfs_btree *btree,
 				       struct nilfs_btree_path *path,
-				       int level)
+				       int level, struct inode *dat)
 {
-	nilfs_bmap_abort_update_v(&btree->bt_bmap,
-				  &path[level].bp_oldreq,
-				  &path[level].bp_newreq);
+	nilfs_dat_abort_update(dat, &path[level].bp_oldreq.bpr_req,
+			       &path[level].bp_newreq.bpr_req);
 	if (buffer_nilfs_node(path[level].bp_bh))
 		nilfs_btnode_abort_change_key(
 			&NILFS_BMAP_I(&btree->bt_bmap)->i_btnode_cache,
@@ -1894,14 +1859,14 @@ static void nilfs_btree_abort_update_v(struct nilfs_btree *btree,
 
 static int nilfs_btree_prepare_propagate_v(struct nilfs_btree *btree,
 					   struct nilfs_btree_path *path,
-					   int minlevel,
-					   int *maxlevelp)
+					   int minlevel, int *maxlevelp,
+					   struct inode *dat)
 {
 	int level, ret;
 
 	level = minlevel;
 	if (!buffer_nilfs_volatile(path[level].bp_bh)) {
-		ret = nilfs_btree_prepare_update_v(btree, path, level);
+		ret = nilfs_btree_prepare_update_v(btree, path, level, dat);
 		if (ret < 0)
 			return ret;
 	}
@@ -1909,7 +1874,7 @@ static int nilfs_btree_prepare_propagate_v(struct nilfs_btree *btree,
 	       !buffer_dirty(path[level].bp_bh)) {
 
 		WARN_ON(buffer_nilfs_volatile(path[level].bp_bh));
-		ret = nilfs_btree_prepare_update_v(btree, path, level);
+		ret = nilfs_btree_prepare_update_v(btree, path, level, dat);
 		if (ret < 0)
 			goto out;
 	}
@@ -1921,39 +1886,40 @@ static int nilfs_btree_prepare_propagate_v(struct nilfs_btree *btree,
 	/* error */
  out:
 	while (--level > minlevel)
-		nilfs_btree_abort_update_v(btree, path, level);
+		nilfs_btree_abort_update_v(btree, path, level, dat);
 	if (!buffer_nilfs_volatile(path[level].bp_bh))
-		nilfs_btree_abort_update_v(btree, path, level);
+		nilfs_btree_abort_update_v(btree, path, level, dat);
 	return ret;
 }
 
 static void nilfs_btree_commit_propagate_v(struct nilfs_btree *btree,
 					   struct nilfs_btree_path *path,
-					   int minlevel,
-					   int maxlevel,
-					   struct buffer_head *bh)
+					   int minlevel, int maxlevel,
+					   struct buffer_head *bh,
+					   struct inode *dat)
 {
 	int level;
 
 	if (!buffer_nilfs_volatile(path[minlevel].bp_bh))
-		nilfs_btree_commit_update_v(btree, path, minlevel);
+		nilfs_btree_commit_update_v(btree, path, minlevel, dat);
 
 	for (level = minlevel + 1; level <= maxlevel; level++)
-		nilfs_btree_commit_update_v(btree, path, level);
+		nilfs_btree_commit_update_v(btree, path, level, dat);
 }
 
 static int nilfs_btree_propagate_v(struct nilfs_btree *btree,
 				   struct nilfs_btree_path *path,
-				   int level,
-				   struct buffer_head *bh)
+				   int level, struct buffer_head *bh)
 {
 	int maxlevel, ret;
 	struct nilfs_btree_node *parent;
+	struct inode *dat = nilfs_bmap_get_dat(&btree->bt_bmap);
 	__u64 ptr;
 
 	get_bh(bh);
 	path[level].bp_bh = bh;
-	ret = nilfs_btree_prepare_propagate_v(btree, path, level, &maxlevel);
+	ret = nilfs_btree_prepare_propagate_v(btree, path, level, &maxlevel,
+					      dat);
 	if (ret < 0)
 		goto out;
 
@@ -1961,12 +1927,12 @@ static int nilfs_btree_propagate_v(struct nilfs_btree *btree,
 		parent = nilfs_btree_get_node(btree, path, level + 1);
 		ptr = nilfs_btree_node_get_ptr(btree, parent,
 					       path[level + 1].bp_index);
-		ret = nilfs_bmap_mark_dirty(&btree->bt_bmap, ptr);
+		ret = nilfs_dat_mark_dirty(dat, ptr);
 		if (ret < 0)
 			goto out;
 	}
 
-	nilfs_btree_commit_propagate_v(btree, path, level, maxlevel, bh);
+	nilfs_btree_commit_propagate_v(btree, path, level, maxlevel, bh, dat);
 
  out:
 	brelse(path[level].bp_bh);
@@ -1986,15 +1952,15 @@ static int nilfs_btree_propagate(const struct nilfs_bmap *bmap,
 	WARN_ON(!buffer_dirty(bh));
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 
 	if (buffer_nilfs_node(bh)) {
 		node = (struct nilfs_btree_node *)bh->b_data;
-		key = nilfs_btree_node_get_key(btree, node, 0);
-		level = nilfs_btree_node_get_level(btree, node);
+		key = nilfs_btree_node_get_key(node, 0);
+		level = nilfs_btree_node_get_level(node);
 	} else {
 		key = nilfs_bmap_data_get_key(bmap, bh);
 		level = NILFS_BTREE_LEVEL_DATA;
@@ -2013,8 +1979,8 @@ static int nilfs_btree_propagate(const struct nilfs_bmap *bmap,
 		nilfs_btree_propagate_p(btree, path, level, bh);
 
  out:
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 
 	return ret;
 }
@@ -2022,7 +1988,7 @@ static int nilfs_btree_propagate(const struct nilfs_bmap *bmap,
 static int nilfs_btree_propagate_gc(const struct nilfs_bmap *bmap,
 				    struct buffer_head *bh)
 {
-	return nilfs_bmap_mark_dirty(bmap, bh->b_blocknr);
+	return nilfs_dat_mark_dirty(nilfs_bmap_get_dat(bmap), bh->b_blocknr);
 }
 
 static void nilfs_btree_add_dirty_buffer(struct nilfs_btree *btree,
@@ -2037,12 +2003,12 @@ static void nilfs_btree_add_dirty_buffer(struct nilfs_btree *btree,
 
 	get_bh(bh);
 	node = (struct nilfs_btree_node *)bh->b_data;
-	key = nilfs_btree_node_get_key(btree, node, 0);
-	level = nilfs_btree_node_get_level(btree, node);
+	key = nilfs_btree_node_get_key(node, 0);
+	level = nilfs_btree_node_get_level(node);
 	list_for_each(head, &lists[level]) {
 		cbh = list_entry(head, struct buffer_head, b_assoc_buffers);
 		cnode = (struct nilfs_btree_node *)cbh->b_data;
-		ckey = nilfs_btree_node_get_key(btree, cnode, 0);
+		ckey = nilfs_btree_node_get_key(cnode, 0);
 		if (key < ckey)
 			break;
 	}
@@ -2120,8 +2086,7 @@ static int nilfs_btree_assign_p(struct nilfs_btree *btree,
 	nilfs_btree_node_set_ptr(btree, parent,
 				 path[level + 1].bp_index, blocknr);
 
-	key = nilfs_btree_node_get_key(btree, parent,
-				       path[level + 1].bp_index);
+	key = nilfs_btree_node_get_key(parent, path[level + 1].bp_index);
 	/* on-disk format */
 	binfo->bi_dat.bi_blkoff = nilfs_bmap_key_to_dkey(key);
 	binfo->bi_dat.bi_level = level;
@@ -2137,6 +2102,7 @@ static int nilfs_btree_assign_v(struct nilfs_btree *btree,
 				union nilfs_binfo *binfo)
 {
 	struct nilfs_btree_node *parent;
+	struct inode *dat = nilfs_bmap_get_dat(&btree->bt_bmap);
 	__u64 key;
 	__u64 ptr;
 	union nilfs_bmap_ptr_req req;
@@ -2146,12 +2112,12 @@ static int nilfs_btree_assign_v(struct nilfs_btree *btree,
 	ptr = nilfs_btree_node_get_ptr(btree, parent,
 				       path[level + 1].bp_index);
 	req.bpr_ptr = ptr;
-	ret = nilfs_bmap_start_v(&btree->bt_bmap, &req, blocknr);
-	if (unlikely(ret < 0))
+	ret = nilfs_dat_prepare_start(dat, &req.bpr_req);
+	if (ret < 0)
 		return ret;
+	nilfs_dat_commit_start(dat, &req.bpr_req, blocknr);
 
-	key = nilfs_btree_node_get_key(btree, parent,
-				       path[level + 1].bp_index);
+	key = nilfs_btree_node_get_key(parent, path[level + 1].bp_index);
 	/* on-disk format */
 	binfo->bi_v.bi_vblocknr = nilfs_bmap_ptr_to_dptr(ptr);
 	binfo->bi_v.bi_blkoff = nilfs_bmap_key_to_dkey(key);
@@ -2171,15 +2137,15 @@ static int nilfs_btree_assign(struct nilfs_bmap *bmap,
 	int level, ret;
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 
 	if (buffer_nilfs_node(*bh)) {
 		node = (struct nilfs_btree_node *)(*bh)->b_data;
-		key = nilfs_btree_node_get_key(btree, node, 0);
-		level = nilfs_btree_node_get_level(btree, node);
+		key = nilfs_btree_node_get_key(node, 0);
+		level = nilfs_btree_node_get_level(node);
 	} else {
 		key = nilfs_bmap_data_get_key(bmap, *bh);
 		level = NILFS_BTREE_LEVEL_DATA;
@@ -2196,8 +2162,8 @@ static int nilfs_btree_assign(struct nilfs_bmap *bmap,
 		nilfs_btree_assign_p(btree, path, level, bh, blocknr, binfo);
 
  out:
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 
 	return ret;
 }
@@ -2207,19 +2173,18 @@ static int nilfs_btree_assign_gc(struct nilfs_bmap *bmap,
 				 sector_t blocknr,
 				 union nilfs_binfo *binfo)
 {
-	struct nilfs_btree *btree;
 	struct nilfs_btree_node *node;
 	__u64 key;
 	int ret;
 
-	btree = (struct nilfs_btree *)bmap;
-	ret = nilfs_bmap_move_v(bmap, (*bh)->b_blocknr, blocknr);
+	ret = nilfs_dat_move(nilfs_bmap_get_dat(bmap), (*bh)->b_blocknr,
+			     blocknr);
 	if (ret < 0)
 		return ret;
 
 	if (buffer_nilfs_node(*bh)) {
 		node = (struct nilfs_btree_node *)(*bh)->b_data;
-		key = nilfs_btree_node_get_key(btree, node, 0);
+		key = nilfs_btree_node_get_key(node, 0);
 	} else
 		key = nilfs_bmap_data_get_key(bmap, *bh);
 
@@ -2239,10 +2204,10 @@ static int nilfs_btree_mark(struct nilfs_bmap *bmap, __u64 key, int level)
 	int ret;
 
 	btree = (struct nilfs_btree *)bmap;
-	path = nilfs_btree_alloc_path(btree);
+	path = nilfs_btree_alloc_path();
 	if (path == NULL)
 		return -ENOMEM;
-	nilfs_btree_init_path(btree, path);
+	nilfs_btree_init_path(path);
 
 	ret = nilfs_btree_do_lookup(btree, path, key, &ptr, level + 1);
 	if (ret < 0) {
@@ -2262,8 +2227,8 @@ static int nilfs_btree_mark(struct nilfs_bmap *bmap, __u64 key, int level)
 		nilfs_bmap_set_dirty(&btree->bt_bmap);
 
  out:
-	nilfs_btree_clear_path(btree, path);
-	nilfs_btree_free_path(btree, path);
+	nilfs_btree_release_path(path);
+	nilfs_btree_free_path(path);
 	return ret;
 }
 
diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c
index aec942cf79e..1c6cfb59128 100644
--- a/fs/nilfs2/cpfile.c
+++ b/fs/nilfs2/cpfile.c
@@ -815,8 +815,10 @@ int nilfs_cpfile_is_snapshot(struct inode *cpfile, __u64 cno)
 	void *kaddr;
 	int ret;
 
-	if (cno == 0)
-		return -ENOENT; /* checkpoint number 0 is invalid */
+	/* CP number is invalid if it's zero or larger than the
+	largest	exist one.*/
+	if (cno == 0 || cno >= nilfs_mdt_cno(cpfile))
+		return -ENOENT;
 	down_read(&NILFS_MDT(cpfile)->mi_sem);
 
 	ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &bh);
@@ -824,7 +826,10 @@ int nilfs_cpfile_is_snapshot(struct inode *cpfile, __u64 cno)
 		goto out;
 	kaddr = kmap_atomic(bh->b_page, KM_USER0);
 	cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, bh, kaddr);
-	ret = nilfs_checkpoint_snapshot(cp);
+	if (nilfs_checkpoint_invalid(cp))
+		ret = -ENOENT;
+	else
+		ret = nilfs_checkpoint_snapshot(cp);
 	kunmap_atomic(kaddr, KM_USER0);
 	brelse(bh);
 
diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h
index 788a4595019..debea896e70 100644
--- a/fs/nilfs2/cpfile.h
+++ b/fs/nilfs2/cpfile.h
@@ -27,8 +27,6 @@
 #include <linux/buffer_head.h>
 #include <linux/nilfs2_fs.h>
 
-#define NILFS_CPFILE_GFP	NILFS_MDT_GFP
-
 
 int nilfs_cpfile_get_checkpoint(struct inode *, __u64, int,
 				struct nilfs_checkpoint **,
diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c
index 8927ca27e6f..1ff8e15bd36 100644
--- a/fs/nilfs2/dat.c
+++ b/fs/nilfs2/dat.c
@@ -109,12 +109,6 @@ void nilfs_dat_commit_free(struct inode *dat, struct nilfs_palloc_req *req)
 	nilfs_palloc_commit_free_entry(dat, req);
 }
 
-void nilfs_dat_abort_free(struct inode *dat, struct nilfs_palloc_req *req)
-{
-	nilfs_dat_abort_entry(dat, req);
-	nilfs_palloc_abort_free_entry(dat, req);
-}
-
 int nilfs_dat_prepare_start(struct inode *dat, struct nilfs_palloc_req *req)
 {
 	int ret;
@@ -140,11 +134,6 @@ void nilfs_dat_commit_start(struct inode *dat, struct nilfs_palloc_req *req,
 	nilfs_dat_commit_entry(dat, req);
 }
 
-void nilfs_dat_abort_start(struct inode *dat, struct nilfs_palloc_req *req)
-{
-	nilfs_dat_abort_entry(dat, req);
-}
-
 int nilfs_dat_prepare_end(struct inode *dat, struct nilfs_palloc_req *req)
 {
 	struct nilfs_dat_entry *entry;
@@ -222,6 +211,37 @@ void nilfs_dat_abort_end(struct inode *dat, struct nilfs_palloc_req *req)
 	nilfs_dat_abort_entry(dat, req);
 }
 
+int nilfs_dat_prepare_update(struct inode *dat,
+			     struct nilfs_palloc_req *oldreq,
+			     struct nilfs_palloc_req *newreq)
+{
+	int ret;
+
+	ret = nilfs_dat_prepare_end(dat, oldreq);
+	if (!ret) {
+		ret = nilfs_dat_prepare_alloc(dat, newreq);
+		if (ret < 0)
+			nilfs_dat_abort_end(dat, oldreq);
+	}
+	return ret;
+}
+
+void nilfs_dat_commit_update(struct inode *dat,
+			     struct nilfs_palloc_req *oldreq,
+			     struct nilfs_palloc_req *newreq, int dead)
+{
+	nilfs_dat_commit_end(dat, oldreq, dead);
+	nilfs_dat_commit_alloc(dat, newreq);
+}
+
+void nilfs_dat_abort_update(struct inode *dat,
+			    struct nilfs_palloc_req *oldreq,
+			    struct nilfs_palloc_req *newreq)
+{
+	nilfs_dat_abort_end(dat, oldreq);
+	nilfs_dat_abort_alloc(dat, newreq);
+}
+
 /**
  * nilfs_dat_mark_dirty -
  * @dat: DAT file inode
diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h
index d328b81eead..406070d3ff4 100644
--- a/fs/nilfs2/dat.h
+++ b/fs/nilfs2/dat.h
@@ -27,7 +27,6 @@
 #include <linux/buffer_head.h>
 #include <linux/fs.h>
 
-#define NILFS_DAT_GFP	NILFS_MDT_GFP
 
 struct nilfs_palloc_req;
 
@@ -39,10 +38,15 @@ void nilfs_dat_abort_alloc(struct inode *, struct nilfs_palloc_req *);
 int nilfs_dat_prepare_start(struct inode *, struct nilfs_palloc_req *);
 void nilfs_dat_commit_start(struct inode *, struct nilfs_palloc_req *,
 			    sector_t);
-void nilfs_dat_abort_start(struct inode *, struct nilfs_palloc_req *);
 int nilfs_dat_prepare_end(struct inode *, struct nilfs_palloc_req *);
 void nilfs_dat_commit_end(struct inode *, struct nilfs_palloc_req *, int);
 void nilfs_dat_abort_end(struct inode *, struct nilfs_palloc_req *);
+int nilfs_dat_prepare_update(struct inode *, struct nilfs_palloc_req *,
+			     struct nilfs_palloc_req *);
+void nilfs_dat_commit_update(struct inode *, struct nilfs_palloc_req *,
+			     struct nilfs_palloc_req *, int);
+void nilfs_dat_abort_update(struct inode *, struct nilfs_palloc_req *,
+			    struct nilfs_palloc_req *);
 
 int nilfs_dat_mark_dirty(struct inode *, __u64);
 int nilfs_dat_freev(struct inode *, __u64 *, size_t);
diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c
index 342d9765df8..d369ac71827 100644
--- a/fs/nilfs2/direct.c
+++ b/fs/nilfs2/direct.c
@@ -125,106 +125,64 @@ static void nilfs_direct_set_target_v(struct nilfs_direct *direct,
 	direct->d_bmap.b_last_allocated_ptr = ptr;
 }
 
-static int nilfs_direct_prepare_insert(struct nilfs_direct *direct,
-				       __u64 key,
-				       union nilfs_bmap_ptr_req *req,
-				       struct nilfs_bmap_stats *stats)
-{
-	int ret;
-
-	if (NILFS_BMAP_USE_VBN(&direct->d_bmap))
-		req->bpr_ptr = nilfs_direct_find_target_v(direct, key);
-	ret = nilfs_bmap_prepare_alloc_ptr(&direct->d_bmap, req);
-	if (ret < 0)
-		return ret;
-
-	stats->bs_nblocks = 1;
-	return 0;
-}
-
-static void nilfs_direct_commit_insert(struct nilfs_direct *direct,
-				       union nilfs_bmap_ptr_req *req,
-				       __u64 key, __u64 ptr)
-{
-	struct buffer_head *bh;
-
-	/* ptr must be a pointer to a buffer head. */
-	bh = (struct buffer_head *)((unsigned long)ptr);
-	set_buffer_nilfs_volatile(bh);
-
-	nilfs_bmap_commit_alloc_ptr(&direct->d_bmap, req);
-	nilfs_direct_set_ptr(direct, key, req->bpr_ptr);
-
-	if (!nilfs_bmap_dirty(&direct->d_bmap))
-		nilfs_bmap_set_dirty(&direct->d_bmap);
-
-	if (NILFS_BMAP_USE_VBN(&direct->d_bmap))
-		nilfs_direct_set_target_v(direct, key, req->bpr_ptr);
-}
-
 static int nilfs_direct_insert(struct nilfs_bmap *bmap, __u64 key, __u64 ptr)
 {
-	struct nilfs_direct *direct;
+	struct nilfs_direct *direct = (struct nilfs_direct *)bmap;
 	union nilfs_bmap_ptr_req req;
-	struct nilfs_bmap_stats stats;
+	struct inode *dat = NULL;
+	struct buffer_head *bh;
 	int ret;
 
-	direct = (struct nilfs_direct *)bmap;
 	if (key > NILFS_DIRECT_KEY_MAX)
 		return -ENOENT;
 	if (nilfs_direct_get_ptr(direct, key) != NILFS_BMAP_INVALID_PTR)
 		return -EEXIST;
 
-	ret = nilfs_direct_prepare_insert(direct, key, &req, &stats);
-	if (ret < 0)
-		return ret;
-	nilfs_direct_commit_insert(direct, &req, key, ptr);
-	nilfs_bmap_add_blocks(bmap, stats.bs_nblocks);
+	if (NILFS_BMAP_USE_VBN(bmap)) {
+		req.bpr_ptr = nilfs_direct_find_target_v(direct, key);
+		dat = nilfs_bmap_get_dat(bmap);
+	}
+	ret = nilfs_bmap_prepare_alloc_ptr(bmap, &req, dat);
+	if (!ret) {
+		/* ptr must be a pointer to a buffer head. */
+		bh = (struct buffer_head *)((unsigned long)ptr);
+		set_buffer_nilfs_volatile(bh);
 
-	return 0;
-}
+		nilfs_bmap_commit_alloc_ptr(bmap, &req, dat);
+		nilfs_direct_set_ptr(direct, key, req.bpr_ptr);
 
-static int nilfs_direct_prepare_delete(struct nilfs_direct *direct,
-				       union nilfs_bmap_ptr_req *req,
-				       __u64 key,
-				       struct nilfs_bmap_stats *stats)
-{
-	int ret;
+		if (!nilfs_bmap_dirty(bmap))
+			nilfs_bmap_set_dirty(bmap);
 
-	req->bpr_ptr = nilfs_direct_get_ptr(direct, key);
-	ret = nilfs_bmap_prepare_end_ptr(&direct->d_bmap, req);
-	if (!ret)
-		stats->bs_nblocks = 1;
-	return ret;
-}
+		if (NILFS_BMAP_USE_VBN(bmap))
+			nilfs_direct_set_target_v(direct, key, req.bpr_ptr);
 
-static void nilfs_direct_commit_delete(struct nilfs_direct *direct,
-				       union nilfs_bmap_ptr_req *req,
-				       __u64 key)
-{
-	nilfs_bmap_commit_end_ptr(&direct->d_bmap, req);
-	nilfs_direct_set_ptr(direct, key, NILFS_BMAP_INVALID_PTR);
+		nilfs_bmap_add_blocks(bmap, 1);
+	}
+	return ret;
 }
 
 static int nilfs_direct_delete(struct nilfs_bmap *bmap, __u64 key)
 {
-	struct nilfs_direct *direct;
+	struct nilfs_direct *direct = (struct nilfs_direct *)bmap;
 	union nilfs_bmap_ptr_req req;
-	struct nilfs_bmap_stats stats;
+	struct inode *dat;
 	int ret;
 
-	direct = (struct nilfs_direct *)bmap;
-	if ((key > NILFS_DIRECT_KEY_MAX) ||
+	if (key > NILFS_DIRECT_KEY_MAX ||
 	    nilfs_direct_get_ptr(direct, key) == NILFS_BMAP_INVALID_PTR)
 		return -ENOENT;
 
-	ret = nilfs_direct_prepare_delete(direct, &req, key, &stats);
-	if (ret < 0)
-		return ret;
-	nilfs_direct_commit_delete(direct, &req, key);
-	nilfs_bmap_sub_blocks(bmap, stats.bs_nblocks);
+	dat = NILFS_BMAP_USE_VBN(bmap) ? nilfs_bmap_get_dat(bmap) : NULL;
+	req.bpr_ptr = nilfs_direct_get_ptr(direct, key);
 
-	return 0;
+	ret = nilfs_bmap_prepare_end_ptr(bmap, &req, dat);
+	if (!ret) {
+		nilfs_bmap_commit_end_ptr(bmap, &req, dat);
+		nilfs_direct_set_ptr(direct, key, NILFS_BMAP_INVALID_PTR);
+		nilfs_bmap_sub_blocks(bmap, 1);
+	}
+	return ret;
 }
 
 static int nilfs_direct_last_key(const struct nilfs_bmap *bmap, __u64 *keyp)
@@ -310,59 +268,56 @@ int nilfs_direct_delete_and_convert(struct nilfs_bmap *bmap,
 	return 0;
 }
 
-static int nilfs_direct_propagate_v(struct nilfs_direct *direct,
-				    struct buffer_head *bh)
+static int nilfs_direct_propagate(const struct nilfs_bmap *bmap,
+				  struct buffer_head *bh)
 {
-	union nilfs_bmap_ptr_req oldreq, newreq;
+	struct nilfs_direct *direct = (struct nilfs_direct *)bmap;
+	struct nilfs_palloc_req oldreq, newreq;
+	struct inode *dat;
 	__u64 key;
 	__u64 ptr;
 	int ret;
 
-	key = nilfs_bmap_data_get_key(&direct->d_bmap, bh);
+	if (!NILFS_BMAP_USE_VBN(bmap))
+		return 0;
+
+	dat = nilfs_bmap_get_dat(bmap);
+	key = nilfs_bmap_data_get_key(bmap, bh);
 	ptr = nilfs_direct_get_ptr(direct, key);
 	if (!buffer_nilfs_volatile(bh)) {
-		oldreq.bpr_ptr = ptr;
-		newreq.bpr_ptr = ptr;
-		ret = nilfs_bmap_prepare_update_v(&direct->d_bmap, &oldreq,
-						  &newreq);
+		oldreq.pr_entry_nr = ptr;
+		newreq.pr_entry_nr = ptr;
+		ret = nilfs_dat_prepare_update(dat, &oldreq, &newreq);
 		if (ret < 0)
 			return ret;
-		nilfs_bmap_commit_update_v(&direct->d_bmap, &oldreq, &newreq);
+		nilfs_dat_commit_update(dat, &oldreq, &newreq,
+					bmap->b_ptr_type == NILFS_BMAP_PTR_VS);
 		set_buffer_nilfs_volatile(bh);
-		nilfs_direct_set_ptr(direct, key, newreq.bpr_ptr);
+		nilfs_direct_set_ptr(direct, key, newreq.pr_entry_nr);
 	} else
-		ret = nilfs_bmap_mark_dirty(&direct->d_bmap, ptr);
+		ret = nilfs_dat_mark_dirty(dat, ptr);
 
 	return ret;
 }
 
-static int nilfs_direct_propagate(const struct nilfs_bmap *bmap,
-				  struct buffer_head *bh)
-{
-	struct nilfs_direct *direct = (struct nilfs_direct *)bmap;
-
-	return NILFS_BMAP_USE_VBN(bmap) ?
-		nilfs_direct_propagate_v(direct, bh) : 0;
-}
-
 static int nilfs_direct_assign_v(struct nilfs_direct *direct,
 				 __u64 key, __u64 ptr,
 				 struct buffer_head **bh,
 				 sector_t blocknr,
 				 union nilfs_binfo *binfo)
 {
+	struct inode *dat = nilfs_bmap_get_dat(&direct->d_bmap);
 	union nilfs_bmap_ptr_req req;
 	int ret;
 
 	req.bpr_ptr = ptr;
-	ret = nilfs_bmap_start_v(&direct->d_bmap, &req, blocknr);
-	if (unlikely(ret < 0))
-		return ret;
-
-	binfo->bi_v.bi_vblocknr = nilfs_bmap_ptr_to_dptr(ptr);
-	binfo->bi_v.bi_blkoff = nilfs_bmap_key_to_dkey(key);
-
-	return 0;
+	ret = nilfs_dat_prepare_start(dat, &req.bpr_req);
+	if (!ret) {
+		nilfs_dat_commit_start(dat, &req.bpr_req, blocknr);
+		binfo->bi_v.bi_vblocknr = nilfs_bmap_ptr_to_dptr(ptr);
+		binfo->bi_v.bi_blkoff = nilfs_bmap_key_to_dkey(key);
+	}
+	return ret;
 }
 
 static int nilfs_direct_assign_p(struct nilfs_direct *direct,
diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c
index 6bd84a0d823..fc8278c77cd 100644
--- a/fs/nilfs2/file.c
+++ b/fs/nilfs2/file.c
@@ -151,7 +151,7 @@ struct file_operations nilfs_file_operations = {
 	.splice_read	= generic_file_splice_read,
 };
 
-struct inode_operations nilfs_file_inode_operations = {
+const struct inode_operations nilfs_file_inode_operations = {
 	.truncate	= nilfs_truncate,
 	.setattr	= nilfs_setattr,
 	.permission     = nilfs_permission,
diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c
index 1b3c2bb20da..e6de0a27ab5 100644
--- a/fs/nilfs2/gcinode.c
+++ b/fs/nilfs2/gcinode.c
@@ -52,7 +52,7 @@
 #include "dat.h"
 #include "ifile.h"
 
-static struct address_space_operations def_gcinode_aops = {
+static const struct address_space_operations def_gcinode_aops = {
 	.sync_page		= block_sync_page,
 };
 
diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h
index 5d30a35679b..ecc3ba76db4 100644
--- a/fs/nilfs2/ifile.h
+++ b/fs/nilfs2/ifile.h
@@ -31,7 +31,6 @@
 #include "mdt.h"
 #include "alloc.h"
 
-#define NILFS_IFILE_GFP  NILFS_MDT_GFP
 
 static inline struct nilfs_inode *
 nilfs_ifile_map_inode(struct inode *ifile, ino_t ino, struct buffer_head *ibh)
diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index fe9d8f2a13f..2d2c501deb5 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -238,7 +238,7 @@ nilfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	return size;
 }
 
-struct address_space_operations nilfs_aops = {
+const struct address_space_operations nilfs_aops = {
 	.writepage		= nilfs_writepage,
 	.readpage		= nilfs_readpage,
 	.sync_page		= block_sync_page,
@@ -430,7 +430,8 @@ static int __nilfs_read_inode(struct super_block *sb, unsigned long ino,
 
 	raw_inode = nilfs_ifile_map_inode(sbi->s_ifile, ino, bh);
 
-	if (nilfs_read_inode_common(inode, raw_inode))
+	err = nilfs_read_inode_common(inode, raw_inode);
+	if (err)
 		goto failed_unmap;
 
 	if (S_ISREG(inode->i_mode)) {
diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c
index 6ea5f872e2d..6572ea4bc4d 100644
--- a/fs/nilfs2/ioctl.c
+++ b/fs/nilfs2/ioctl.c
@@ -442,12 +442,6 @@ int nilfs_ioctl_prepare_clean_segments(struct the_nilfs *nilfs,
 	const char *msg;
 	int ret;
 
-	ret = nilfs_ioctl_move_blocks(nilfs, &argv[0], kbufs[0]);
-	if (ret < 0) {
-		msg = "cannot read source blocks";
-		goto failed;
-	}
-
 	ret = nilfs_ioctl_delete_checkpoints(nilfs, &argv[1], kbufs[1]);
 	if (ret < 0) {
 		/*
@@ -548,7 +542,25 @@ static int nilfs_ioctl_clean_segments(struct inode *inode, struct file *filp,
 		}
 	}
 
-	ret = nilfs_clean_segments(inode->i_sb, argv, kbufs);
+	/*
+	 * nilfs_ioctl_move_blocks() will call nilfs_gc_iget(),
+	 * which will operates an inode list without blocking.
+	 * To protect the list from concurrent operations,
+	 * nilfs_ioctl_move_blocks should be atomic operation.
+	 */
+	if (test_and_set_bit(THE_NILFS_GC_RUNNING, &nilfs->ns_flags)) {
+		ret = -EBUSY;
+		goto out_free;
+	}
+
+	ret = nilfs_ioctl_move_blocks(nilfs, &argv[0], kbufs[0]);
+	if (ret < 0)
+		printk(KERN_ERR "NILFS: GC failed during preparation: "
+			"cannot read source blocks: err=%d\n", ret);
+	else
+		ret = nilfs_clean_segments(inode->i_sb, argv, kbufs);
+
+	clear_nilfs_gc_running(nilfs);
 
  out_free:
 	while (--n >= 0)
diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index 2dfd47714ae..b18c4998f8d 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -103,15 +103,12 @@ static int nilfs_mdt_create_block(struct inode *inode, unsigned long block,
 		goto failed_unlock;
 
 	err = -EEXIST;
-	if (buffer_uptodate(bh) || buffer_mapped(bh))
+	if (buffer_uptodate(bh))
 		goto failed_bh;
-#if 0
-	/* The uptodate flag is not protected by the page lock, but
-	   the mapped flag is.  Thus, we don't have to wait the buffer. */
+
 	wait_on_buffer(bh);
 	if (buffer_uptodate(bh))
 		goto failed_bh;
-#endif
 
 	bh->b_bdev = nilfs->ns_bdev;
 	err = nilfs_mdt_insert_new_block(inode, block, bh, init_block);
@@ -139,7 +136,7 @@ nilfs_mdt_submit_block(struct inode *inode, unsigned long blkoff,
 		       int mode, struct buffer_head **out_bh)
 {
 	struct buffer_head *bh;
-	unsigned long blknum = 0;
+	__u64 blknum = 0;
 	int ret = -ENOMEM;
 
 	bh = nilfs_grab_buffer(inode, inode->i_mapping, blkoff, 0);
@@ -162,17 +159,15 @@ nilfs_mdt_submit_block(struct inode *inode, unsigned long blkoff,
 		unlock_buffer(bh);
 		goto out;
 	}
-	if (!buffer_mapped(bh)) { /* unused buffer */
-		ret = nilfs_bmap_lookup(NILFS_I(inode)->i_bmap, blkoff,
-					&blknum);
-		if (unlikely(ret)) {
-			unlock_buffer(bh);
-			goto failed_bh;
-		}
-		bh->b_bdev = NILFS_MDT(inode)->mi_nilfs->ns_bdev;
-		bh->b_blocknr = blknum;
-		set_buffer_mapped(bh);
+
+	ret = nilfs_bmap_lookup(NILFS_I(inode)->i_bmap, blkoff, &blknum);
+	if (unlikely(ret)) {
+		unlock_buffer(bh);
+		goto failed_bh;
 	}
+	bh->b_bdev = NILFS_MDT(inode)->mi_nilfs->ns_bdev;
+	bh->b_blocknr = (sector_t)blknum;
+	set_buffer_mapped(bh);
 
 	bh->b_end_io = end_buffer_read_sync;
 	get_bh(bh);
@@ -402,6 +397,7 @@ nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
 	struct inode *inode = container_of(page->mapping,
 					   struct inode, i_data);
 	struct super_block *sb = inode->i_sb;
+	struct the_nilfs *nilfs = NILFS_MDT(inode)->mi_nilfs;
 	struct nilfs_sb_info *writer = NULL;
 	int err = 0;
 
@@ -411,9 +407,10 @@ nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
 	if (page->mapping->assoc_mapping)
 		return 0; /* Do not request flush for shadow page cache */
 	if (!sb) {
-		writer = nilfs_get_writer(NILFS_MDT(inode)->mi_nilfs);
+		down_read(&nilfs->ns_writer_sem);
+		writer = nilfs->ns_writer;
 		if (!writer) {
-			nilfs_put_writer(NILFS_MDT(inode)->mi_nilfs);
+			up_read(&nilfs->ns_writer_sem);
 			return -EROFS;
 		}
 		sb = writer->s_super;
@@ -425,17 +422,17 @@ nilfs_mdt_write_page(struct page *page, struct writeback_control *wbc)
 		nilfs_flush_segment(sb, inode->i_ino);
 
 	if (writer)
-		nilfs_put_writer(NILFS_MDT(inode)->mi_nilfs);
+		up_read(&nilfs->ns_writer_sem);
 	return err;
 }
 
 
-static struct address_space_operations def_mdt_aops = {
+static const struct address_space_operations def_mdt_aops = {
 	.writepage		= nilfs_mdt_write_page,
 	.sync_page		= block_sync_page,
 };
 
-static struct inode_operations def_mdt_iops;
+static const struct inode_operations def_mdt_iops;
 static struct file_operations def_mdt_fops;
 
 /*
@@ -516,9 +513,10 @@ nilfs_mdt_new_common(struct the_nilfs *nilfs, struct super_block *sb,
 }
 
 struct inode *nilfs_mdt_new(struct the_nilfs *nilfs, struct super_block *sb,
-			    ino_t ino, gfp_t gfp_mask)
+			    ino_t ino)
 {
-	struct inode *inode = nilfs_mdt_new_common(nilfs, sb, ino, gfp_mask);
+	struct inode *inode = nilfs_mdt_new_common(nilfs, sb, ino,
+						   NILFS_MDT_GFP);
 
 	if (!inode)
 		return NULL;
diff --git a/fs/nilfs2/mdt.h b/fs/nilfs2/mdt.h
index df683e0bca6..431599733c9 100644
--- a/fs/nilfs2/mdt.h
+++ b/fs/nilfs2/mdt.h
@@ -74,8 +74,7 @@ int nilfs_mdt_forget_block(struct inode *, unsigned long);
 int nilfs_mdt_mark_block_dirty(struct inode *, unsigned long);
 int nilfs_mdt_fetch_dirty(struct inode *);
 
-struct inode *nilfs_mdt_new(struct the_nilfs *, struct super_block *, ino_t,
-			    gfp_t);
+struct inode *nilfs_mdt_new(struct the_nilfs *, struct super_block *, ino_t);
 struct inode *nilfs_mdt_new_common(struct the_nilfs *, struct super_block *,
 				   ino_t, gfp_t);
 void nilfs_mdt_destroy(struct inode *);
diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c
index df70dadb336..ed02e886fa7 100644
--- a/fs/nilfs2/namei.c
+++ b/fs/nilfs2/namei.c
@@ -448,7 +448,7 @@ out:
 	return err;
 }
 
-struct inode_operations nilfs_dir_inode_operations = {
+const struct inode_operations nilfs_dir_inode_operations = {
 	.create		= nilfs_create,
 	.lookup		= nilfs_lookup,
 	.link		= nilfs_link,
@@ -462,12 +462,12 @@ struct inode_operations nilfs_dir_inode_operations = {
 	.permission	= nilfs_permission,
 };
 
-struct inode_operations nilfs_special_inode_operations = {
+const struct inode_operations nilfs_special_inode_operations = {
 	.setattr	= nilfs_setattr,
 	.permission	= nilfs_permission,
 };
 
-struct inode_operations nilfs_symlink_inode_operations = {
+const struct inode_operations nilfs_symlink_inode_operations = {
 	.readlink	= generic_readlink,
 	.follow_link	= page_follow_link_light,
 	.put_link	= page_put_link,
diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h
index 724c63766e8..bad7368782d 100644
--- a/fs/nilfs2/nilfs.h
+++ b/fs/nilfs2/nilfs.h
@@ -295,12 +295,12 @@ void nilfs_clear_gcdat_inode(struct the_nilfs *);
  * Inodes and files operations
  */
 extern struct file_operations nilfs_dir_operations;
-extern struct inode_operations nilfs_file_inode_operations;
+extern const struct inode_operations nilfs_file_inode_operations;
 extern struct file_operations nilfs_file_operations;
-extern struct address_space_operations nilfs_aops;
-extern struct inode_operations nilfs_dir_inode_operations;
-extern struct inode_operations nilfs_special_inode_operations;
-extern struct inode_operations nilfs_symlink_inode_operations;
+extern const struct address_space_operations nilfs_aops;
+extern const struct inode_operations nilfs_dir_inode_operations;
+extern const struct inode_operations nilfs_special_inode_operations;
+extern const struct inode_operations nilfs_symlink_inode_operations;
 
 /*
  * filesystem type
diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c
index d80cc71be74..6dc83591d11 100644
--- a/fs/nilfs2/recovery.c
+++ b/fs/nilfs2/recovery.c
@@ -552,7 +552,8 @@ static int recover_dsync_blocks(struct nilfs_sb_info *sbi,
 		printk(KERN_WARNING
 		       "NILFS warning: error recovering data block "
 		       "(err=%d, ino=%lu, block-offset=%llu)\n",
-		       err, rb->ino, (unsigned long long)rb->blkoff);
+		       err, (unsigned long)rb->ino,
+		       (unsigned long long)rb->blkoff);
 		if (!err2)
 			err2 = err;
  next:
diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c
index 9e3fe17bb96..e6d9e37fa24 100644
--- a/fs/nilfs2/segbuf.c
+++ b/fs/nilfs2/segbuf.c
@@ -316,10 +316,10 @@ static struct bio *nilfs_alloc_seg_bio(struct super_block *sb, sector_t start,
 {
 	struct bio *bio;
 
-	bio = bio_alloc(GFP_NOWAIT, nr_vecs);
+	bio = bio_alloc(GFP_NOIO, nr_vecs);
 	if (bio == NULL) {
 		while (!bio && (nr_vecs >>= 1))
-			bio = bio_alloc(GFP_NOWAIT, nr_vecs);
+			bio = bio_alloc(GFP_NOIO, nr_vecs);
 	}
 	if (likely(bio)) {
 		bio->bi_bdev = sb->s_bdev;
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 51ff3d0a4ee..683df89dbae 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -2501,7 +2501,8 @@ static int nilfs_segctor_construct(struct nilfs_sc_info *sci,
 		if (test_bit(NILFS_SC_SUPER_ROOT, &sci->sc_flags) &&
 		    nilfs_discontinued(nilfs)) {
 			down_write(&nilfs->ns_sem);
-			req->sb_err = nilfs_commit_super(sbi, 0);
+			req->sb_err = nilfs_commit_super(sbi,
+					nilfs_altsb_need_update(nilfs));
 			up_write(&nilfs->ns_sem);
 		}
 	}
@@ -2689,6 +2690,7 @@ static int nilfs_segctor_thread(void *arg)
 	} else {
 		DEFINE_WAIT(wait);
 		int should_sleep = 1;
+		struct the_nilfs *nilfs;
 
 		prepare_to_wait(&sci->sc_wait_daemon, &wait,
 				TASK_INTERRUPTIBLE);
@@ -2709,6 +2711,9 @@ static int nilfs_segctor_thread(void *arg)
 		finish_wait(&sci->sc_wait_daemon, &wait);
 		timeout = ((sci->sc_state & NILFS_SEGCTOR_COMMIT) &&
 			   time_after_eq(jiffies, sci->sc_timer->expires));
+		nilfs = sci->sc_sbi->s_nilfs;
+		if (sci->sc_super->s_dirt && nilfs_sb_need_update(nilfs))
+			set_nilfs_discontinued(nilfs);
 	}
 	goto loop;
 
diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h
index a2c4d76c336..0e99e5c0bd0 100644
--- a/fs/nilfs2/sufile.h
+++ b/fs/nilfs2/sufile.h
@@ -28,7 +28,6 @@
 #include <linux/nilfs2_fs.h>
 #include "mdt.h"
 
-#define NILFS_SUFILE_GFP	NILFS_MDT_GFP
 
 static inline unsigned long nilfs_sufile_get_nsegments(struct inode *sufile)
 {
diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c
index 151964f0de4..644e66727dd 100644
--- a/fs/nilfs2/super.c
+++ b/fs/nilfs2/super.c
@@ -50,6 +50,8 @@
 #include <linux/writeback.h>
 #include <linux/kobject.h>
 #include <linux/exportfs.h>
+#include <linux/seq_file.h>
+#include <linux/mount.h>
 #include "nilfs.h"
 #include "mdt.h"
 #include "alloc.h"
@@ -65,7 +67,6 @@ MODULE_DESCRIPTION("A New Implementation of the Log-structured Filesystem "
 		   "(NILFS)");
 MODULE_LICENSE("GPL");
 
-static void nilfs_write_super(struct super_block *sb);
 static int nilfs_remount(struct super_block *sb, int *flags, char *data);
 
 /**
@@ -311,9 +312,6 @@ static void nilfs_put_super(struct super_block *sb)
 
 	lock_kernel();
 
-	if (sb->s_dirt)
-		nilfs_write_super(sb);
-
 	nilfs_detach_segment_constructor(sbi);
 
 	if (!(sb->s_flags & MS_RDONLY)) {
@@ -336,63 +334,21 @@ static void nilfs_put_super(struct super_block *sb)
 	unlock_kernel();
 }
 
-/**
- * nilfs_write_super - write super block(s) of NILFS
- * @sb: super_block
- *
- * nilfs_write_super() gets a fs-dependent lock, writes super block(s), and
- * clears s_dirt.  This function is called in the section protected by
- * lock_super().
- *
- * The s_dirt flag is managed by each filesystem and we protect it by ns_sem
- * of the struct the_nilfs.  Lock order must be as follows:
- *
- *   1. lock_super()
- *   2.    down_write(&nilfs->ns_sem)
- *
- * Inside NILFS, locking ns_sem is enough to protect s_dirt and the buffer
- * of the super block (nilfs->ns_sbp[]).
- *
- * In most cases, VFS functions call lock_super() before calling these
- * methods.  So we must be careful not to bring on deadlocks when using
- * lock_super();  see generic_shutdown_super(), write_super(), and so on.
- *
- * Note that order of lock_kernel() and lock_super() depends on contexts
- * of VFS.  We should also note that lock_kernel() can be used in its
- * protective section and only the outermost one has an effect.
- */
-static void nilfs_write_super(struct super_block *sb)
+static int nilfs_sync_fs(struct super_block *sb, int wait)
 {
 	struct nilfs_sb_info *sbi = NILFS_SB(sb);
 	struct the_nilfs *nilfs = sbi->s_nilfs;
-
-	down_write(&nilfs->ns_sem);
-	if (!(sb->s_flags & MS_RDONLY)) {
-		struct nilfs_super_block **sbp = nilfs->ns_sbp;
-		u64 t = get_seconds();
-		int dupsb;
-
-		if (!nilfs_discontinued(nilfs) && t >= nilfs->ns_sbwtime[0] &&
-		    t < nilfs->ns_sbwtime[0] + NILFS_SB_FREQ) {
-			up_write(&nilfs->ns_sem);
-			return;
-		}
-		dupsb = sbp[1] && t > nilfs->ns_sbwtime[1] + NILFS_ALTSB_FREQ;
-		nilfs_commit_super(sbi, dupsb);
-	}
-	sb->s_dirt = 0;
-	up_write(&nilfs->ns_sem);
-}
-
-static int nilfs_sync_fs(struct super_block *sb, int wait)
-{
 	int err = 0;
 
-	nilfs_write_super(sb);
-
 	/* This function is called when super block should be written back */
 	if (wait)
 		err = nilfs_construct_segment(sb);
+
+	down_write(&nilfs->ns_sem);
+	if (sb->s_dirt)
+		nilfs_commit_super(sbi, 1);
+	up_write(&nilfs->ns_sem);
+
 	return err;
 }
 
@@ -407,8 +363,7 @@ int nilfs_attach_checkpoint(struct nilfs_sb_info *sbi, __u64 cno)
 	list_add(&sbi->s_list, &nilfs->ns_supers);
 	up_write(&nilfs->ns_super_sem);
 
-	sbi->s_ifile = nilfs_mdt_new(
-		nilfs, sbi->s_super, NILFS_IFILE_INO, NILFS_IFILE_GFP);
+	sbi->s_ifile = nilfs_mdt_new(nilfs, sbi->s_super, NILFS_IFILE_INO);
 	if (!sbi->s_ifile)
 		return -ENOMEM;
 
@@ -529,7 +484,27 @@ static int nilfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	return 0;
 }
 
-static struct super_operations nilfs_sops = {
+static int nilfs_show_options(struct seq_file *seq, struct vfsmount *vfs)
+{
+	struct super_block *sb = vfs->mnt_sb;
+	struct nilfs_sb_info *sbi = NILFS_SB(sb);
+
+	if (!nilfs_test_opt(sbi, BARRIER))
+		seq_printf(seq, ",barrier=off");
+	if (nilfs_test_opt(sbi, SNAPSHOT))
+		seq_printf(seq, ",cp=%llu",
+			   (unsigned long long int)sbi->s_snapshot_cno);
+	if (nilfs_test_opt(sbi, ERRORS_RO))
+		seq_printf(seq, ",errors=remount-ro");
+	if (nilfs_test_opt(sbi, ERRORS_PANIC))
+		seq_printf(seq, ",errors=panic");
+	if (nilfs_test_opt(sbi, STRICT_ORDER))
+		seq_printf(seq, ",order=strict");
+
+	return 0;
+}
+
+static const struct super_operations nilfs_sops = {
 	.alloc_inode    = nilfs_alloc_inode,
 	.destroy_inode  = nilfs_destroy_inode,
 	.dirty_inode    = nilfs_dirty_inode,
@@ -538,7 +513,7 @@ static struct super_operations nilfs_sops = {
 	/* .drop_inode	  = nilfs_drop_inode, */
 	.delete_inode   = nilfs_delete_inode,
 	.put_super      = nilfs_put_super,
-	.write_super    = nilfs_write_super,
+	/* .write_super    = nilfs_write_super, */
 	.sync_fs        = nilfs_sync_fs,
 	/* .write_super_lockfs */
 	/* .unlockfs */
@@ -546,7 +521,7 @@ static struct super_operations nilfs_sops = {
 	.remount_fs     = nilfs_remount,
 	.clear_inode    = nilfs_clear_inode,
 	/* .umount_begin */
-	/* .show_options */
+	.show_options = nilfs_show_options
 };
 
 static struct inode *
@@ -585,7 +560,7 @@ nilfs_fh_to_parent(struct super_block *sb, struct fid *fid, int fh_len,
 				    nilfs_nfs_get_inode);
 }
 
-static struct export_operations nilfs_export_ops = {
+static const struct export_operations nilfs_export_ops = {
 	.fh_to_dentry = nilfs_fh_to_dentry,
 	.fh_to_parent = nilfs_fh_to_parent,
 	.get_parent = nilfs_get_parent,
@@ -816,10 +791,15 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent,
 
 	if (sb->s_flags & MS_RDONLY) {
 		if (nilfs_test_opt(sbi, SNAPSHOT)) {
+			down_read(&nilfs->ns_segctor_sem);
 			err = nilfs_cpfile_is_snapshot(nilfs->ns_cpfile,
 						       sbi->s_snapshot_cno);
-			if (err < 0)
+			up_read(&nilfs->ns_segctor_sem);
+			if (err < 0) {
+				if (err == -ENOENT)
+					err = -EINVAL;
 				goto failed_sbi;
+			}
 			if (!err) {
 				printk(KERN_ERR
 				       "NILFS: The specified checkpoint is "
@@ -1127,10 +1107,6 @@ nilfs_get_sb(struct file_system_type *fs_type, int flags,
 	 */
 	sd.sbi = nilfs_find_sbinfo(nilfs, !(flags & MS_RDONLY), sd.cno);
 
-	if (!sd.cno)
-		/* trying to get the latest checkpoint.  */
-		sd.cno = nilfs_last_cno(nilfs);
-
 	/*
 	 * Get super block instance holding the nilfs_sb_info struct.
 	 * A new instance is allocated if no existing mount is present or
diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c
index 8b888982571..ad391a8c3e7 100644
--- a/fs/nilfs2/the_nilfs.c
+++ b/fs/nilfs2/the_nilfs.c
@@ -68,12 +68,11 @@ static struct the_nilfs *alloc_nilfs(struct block_device *bdev)
 
 	nilfs->ns_bdev = bdev;
 	atomic_set(&nilfs->ns_count, 1);
-	atomic_set(&nilfs->ns_writer_refcount, -1);
 	atomic_set(&nilfs->ns_ndirtyblks, 0);
 	init_rwsem(&nilfs->ns_sem);
 	init_rwsem(&nilfs->ns_super_sem);
 	mutex_init(&nilfs->ns_mount_mutex);
-	mutex_init(&nilfs->ns_writer_mutex);
+	init_rwsem(&nilfs->ns_writer_sem);
 	INIT_LIST_HEAD(&nilfs->ns_list);
 	INIT_LIST_HEAD(&nilfs->ns_supers);
 	spin_lock_init(&nilfs->ns_last_segment_lock);
@@ -188,23 +187,19 @@ static int nilfs_load_super_root(struct the_nilfs *nilfs,
 	inode_size = nilfs->ns_inode_size;
 
 	err = -ENOMEM;
-	nilfs->ns_dat = nilfs_mdt_new(
-		nilfs, NULL, NILFS_DAT_INO, NILFS_DAT_GFP);
+	nilfs->ns_dat = nilfs_mdt_new(nilfs, NULL, NILFS_DAT_INO);
 	if (unlikely(!nilfs->ns_dat))
 		goto failed;
 
-	nilfs->ns_gc_dat = nilfs_mdt_new(
-		nilfs, NULL, NILFS_DAT_INO, NILFS_DAT_GFP);
+	nilfs->ns_gc_dat = nilfs_mdt_new(nilfs, NULL, NILFS_DAT_INO);
 	if (unlikely(!nilfs->ns_gc_dat))
 		goto failed_dat;
 
-	nilfs->ns_cpfile = nilfs_mdt_new(
-		nilfs, NULL, NILFS_CPFILE_INO, NILFS_CPFILE_GFP);
+	nilfs->ns_cpfile = nilfs_mdt_new(nilfs, NULL, NILFS_CPFILE_INO);
 	if (unlikely(!nilfs->ns_cpfile))
 		goto failed_gc_dat;
 
-	nilfs->ns_sufile = nilfs_mdt_new(
-		nilfs, NULL, NILFS_SUFILE_INO, NILFS_SUFILE_GFP);
+	nilfs->ns_sufile = nilfs_mdt_new(nilfs, NULL, NILFS_SUFILE_INO);
 	if (unlikely(!nilfs->ns_sufile))
 		goto failed_cpfile;
 
@@ -596,9 +591,7 @@ int init_nilfs(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi, char *data)
 
 	nilfs->ns_mount_state = le16_to_cpu(sbp->s_state);
 
-	bdi = nilfs->ns_bdev->bd_inode_backing_dev_info;
-	if (!bdi)
-		bdi = nilfs->ns_bdev->bd_inode->i_mapping->backing_dev_info;
+	bdi = nilfs->ns_bdev->bd_inode->i_mapping->backing_dev_info;
 	nilfs->ns_bdi = bdi ? : &default_backing_dev_info;
 
 	/* Finding last segment */
diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h
index 1b9caafb866..20abd55881e 100644
--- a/fs/nilfs2/the_nilfs.h
+++ b/fs/nilfs2/the_nilfs.h
@@ -37,6 +37,7 @@ enum {
 	THE_NILFS_LOADED,       /* Roll-back/roll-forward has done and
 				   the latest checkpoint was loaded */
 	THE_NILFS_DISCONTINUED,	/* 'next' pointer chain has broken */
+	THE_NILFS_GC_RUNNING,	/* gc process is running */
 };
 
 /**
@@ -50,8 +51,7 @@ enum {
  * @ns_sem: semaphore for shared states
  * @ns_super_sem: semaphore for global operations across super block instances
  * @ns_mount_mutex: mutex protecting mount process of nilfs
- * @ns_writer_mutex: mutex protecting ns_writer attach/detach
- * @ns_writer_refcount: number of referrers on ns_writer
+ * @ns_writer_sem: semaphore protecting ns_writer attach/detach
  * @ns_current: back pointer to current mount
  * @ns_sbh: buffer heads of on-disk super blocks
  * @ns_sbp: pointers to super block data
@@ -100,8 +100,7 @@ struct the_nilfs {
 	struct rw_semaphore	ns_sem;
 	struct rw_semaphore	ns_super_sem;
 	struct mutex		ns_mount_mutex;
-	struct mutex		ns_writer_mutex;
-	atomic_t		ns_writer_refcount;
+	struct rw_semaphore	ns_writer_sem;
 
 	/*
 	 * components protected by ns_super_sem
@@ -197,11 +196,26 @@ static inline int nilfs_##name(struct the_nilfs *nilfs)			\
 THE_NILFS_FNS(INIT, init)
 THE_NILFS_FNS(LOADED, loaded)
 THE_NILFS_FNS(DISCONTINUED, discontinued)
+THE_NILFS_FNS(GC_RUNNING, gc_running)
 
 /* Minimum interval of periodical update of superblocks (in seconds) */
 #define NILFS_SB_FREQ		10
 #define NILFS_ALTSB_FREQ	60  /* spare superblock */
 
+static inline int nilfs_sb_need_update(struct the_nilfs *nilfs)
+{
+	u64 t = get_seconds();
+	return t < nilfs->ns_sbwtime[0] ||
+		 t > nilfs->ns_sbwtime[0] + NILFS_SB_FREQ;
+}
+
+static inline int nilfs_altsb_need_update(struct the_nilfs *nilfs)
+{
+	u64 t = get_seconds();
+	struct nilfs_super_block **sbp = nilfs->ns_sbp;
+	return sbp[1] && t > nilfs->ns_sbwtime[1] + NILFS_ALTSB_FREQ;
+}
+
 void nilfs_set_last_segment(struct the_nilfs *, sector_t, u64, __u64);
 struct the_nilfs *find_or_create_nilfs(struct block_device *);
 void put_nilfs(struct the_nilfs *);
@@ -221,34 +235,21 @@ static inline void get_nilfs(struct the_nilfs *nilfs)
 	atomic_inc(&nilfs->ns_count);
 }
 
-static inline struct nilfs_sb_info *nilfs_get_writer(struct the_nilfs *nilfs)
-{
-	if (atomic_inc_and_test(&nilfs->ns_writer_refcount))
-		mutex_lock(&nilfs->ns_writer_mutex);
-	return nilfs->ns_writer;
-}
-
-static inline void nilfs_put_writer(struct the_nilfs *nilfs)
-{
-	if (atomic_add_negative(-1, &nilfs->ns_writer_refcount))
-		mutex_unlock(&nilfs->ns_writer_mutex);
-}
-
 static inline void
 nilfs_attach_writer(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi)
 {
-	mutex_lock(&nilfs->ns_writer_mutex);
+	down_write(&nilfs->ns_writer_sem);
 	nilfs->ns_writer = sbi;
-	mutex_unlock(&nilfs->ns_writer_mutex);
+	up_write(&nilfs->ns_writer_sem);
 }
 
 static inline void
 nilfs_detach_writer(struct the_nilfs *nilfs, struct nilfs_sb_info *sbi)
 {
-	mutex_lock(&nilfs->ns_writer_mutex);
+	down_write(&nilfs->ns_writer_sem);
 	if (sbi == nilfs->ns_writer)
 		nilfs->ns_writer = NULL;
-	mutex_unlock(&nilfs->ns_writer_mutex);
+	up_write(&nilfs->ns_writer_sem);
 }
 
 static inline void nilfs_put_sbinfo(struct nilfs_sb_info *sbi)
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index 3140a4429af..663c0e341f8 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -2076,14 +2076,6 @@ err_out:
 	*ppos = pos;
 	if (cached_page)
 		page_cache_release(cached_page);
-	/* For now, when the user asks for O_SYNC, we actually give O_DSYNC. */
-	if (likely(!status)) {
-		if (unlikely((file->f_flags & O_SYNC) || IS_SYNC(vi))) {
-			if (!mapping->a_ops->writepage || !is_sync_kiocb(iocb))
-				status = generic_osync_inode(vi, mapping,
-						OSYNC_METADATA|OSYNC_DATA);
-		}
-  	}
 	pagevec_lru_add_file(&lru_pvec);
 	ntfs_debug("Done.  Returning %s (written 0x%lx, status %li).",
 			written ? "written" : "status", (unsigned long)written,
@@ -2145,8 +2137,8 @@ static ssize_t ntfs_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
 	mutex_lock(&inode->i_mutex);
 	ret = ntfs_file_aio_write_nolock(iocb, iov, nr_segs, &iocb->ki_pos);
 	mutex_unlock(&inode->i_mutex);
-	if (ret > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
-		int err = sync_page_range(inode, mapping, pos, ret);
+	if (ret > 0) {
+		int err = generic_write_sync(file, pos, ret);
 		if (err < 0)
 			ret = err;
 	}
@@ -2154,46 +2146,6 @@ static ssize_t ntfs_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
 }
 
 /**
- * ntfs_file_writev -
- *
- * Basically the same as generic_file_writev() except that it ends up calling
- * ntfs_file_aio_write_nolock() instead of __generic_file_aio_write_nolock().
- */
-static ssize_t ntfs_file_writev(struct file *file, const struct iovec *iov,
-		unsigned long nr_segs, loff_t *ppos)
-{
-	struct address_space *mapping = file->f_mapping;
-	struct inode *inode = mapping->host;
-	struct kiocb kiocb;
-	ssize_t ret;
-
-	mutex_lock(&inode->i_mutex);
-	init_sync_kiocb(&kiocb, file);
-	ret = ntfs_file_aio_write_nolock(&kiocb, iov, nr_segs, ppos);
-	if (ret == -EIOCBQUEUED)
-		ret = wait_on_sync_kiocb(&kiocb);
-	mutex_unlock(&inode->i_mutex);
-	if (ret > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
-		int err = sync_page_range(inode, mapping, *ppos - ret, ret);
-		if (err < 0)
-			ret = err;
-	}
-	return ret;
-}
-
-/**
- * ntfs_file_write - simple wrapper for ntfs_file_writev()
- */
-static ssize_t ntfs_file_write(struct file *file, const char __user *buf,
-		size_t count, loff_t *ppos)
-{
-	struct iovec local_iov = { .iov_base = (void __user *)buf,
-				   .iov_len = count };
-
-	return ntfs_file_writev(file, &local_iov, 1, ppos);
-}
-
-/**
  * ntfs_file_fsync - sync a file to disk
  * @filp:	file to be synced
  * @dentry:	dentry describing the file to sync
@@ -2255,7 +2207,7 @@ const struct file_operations ntfs_file_ops = {
 	.read		= do_sync_read,		 /* Read from file. */
 	.aio_read	= generic_file_aio_read, /* Async read from file. */
 #ifdef NTFS_RW
-	.write		= ntfs_file_write,	 /* Write to file. */
+	.write		= do_sync_write,	 /* Write to file. */
 	.aio_write	= ntfs_file_aio_write,	 /* Async write to file. */
 	/*.release	= ,*/			 /* Last file is closed.  See
 						    fs/ext2/file.c::
diff --git a/fs/ntfs/layout.h b/fs/ntfs/layout.h
index 50931b1ce4b..8b2549f672b 100644
--- a/fs/ntfs/layout.h
+++ b/fs/ntfs/layout.h
@@ -829,7 +829,7 @@ enum {
 	/* Note, FILE_ATTR_VALID_SET_FLAGS masks out the old DOS VolId, the
 	   F_A_DEVICE, F_A_DIRECTORY, F_A_SPARSE_FILE, F_A_REPARSE_POINT,
 	   F_A_COMPRESSED, and F_A_ENCRYPTED and preserves the rest.  This mask
-	   is used to to obtain all flags that are valid for setting. */
+	   is used to obtain all flags that are valid for setting. */
 	/*
 	 * The flag FILE_ATTR_DUP_FILENAME_INDEX_PRESENT is present in all
 	 * FILENAME_ATTR attributes but not in the STANDARD_INFORMATION
diff --git a/fs/ntfs/malloc.h b/fs/ntfs/malloc.h
index cd0be3f5c3c..a44b14cbcee 100644
--- a/fs/ntfs/malloc.h
+++ b/fs/ntfs/malloc.h
@@ -47,7 +47,7 @@ static inline void *__ntfs_malloc(unsigned long size, gfp_t gfp_mask)
 		return kmalloc(PAGE_SIZE, gfp_mask & ~__GFP_HIGHMEM);
 		/* return (void *)__get_free_page(gfp_mask); */
 	}
-	if (likely(size >> PAGE_SHIFT < num_physpages))
+	if (likely((size >> PAGE_SHIFT) < totalram_pages))
 		return __vmalloc(size, gfp_mask, PAGE_KERNEL);
 	return NULL;
 }
diff --git a/fs/ntfs/mft.c b/fs/ntfs/mft.c
index 23bf68453d7..1caa0ef0b2b 100644
--- a/fs/ntfs/mft.c
+++ b/fs/ntfs/mft.c
@@ -384,13 +384,12 @@ unm_err_out:
  * it is dirty in the inode meta data rather than the data page cache of the
  * inode, and thus there are no data pages that need writing out.  Therefore, a
  * full mark_inode_dirty() is overkill.  A mark_inode_dirty_sync(), on the
- * other hand, is not sufficient, because I_DIRTY_DATASYNC needs to be set to
- * ensure ->write_inode is called from generic_osync_inode() and this needs to
- * happen or the file data would not necessarily hit the device synchronously,
- * even though the vfs inode has the O_SYNC flag set.  Also, I_DIRTY_DATASYNC
- * simply "feels" better than just I_DIRTY_SYNC, since the file data has not
- * actually hit the block device yet, which is not what I_DIRTY_SYNC on its own
- * would suggest.
+ * other hand, is not sufficient, because ->write_inode needs to be called even
+ * in case of fdatasync. This needs to happen or the file data would not
+ * necessarily hit the device synchronously, even though the vfs inode has the
+ * O_SYNC flag set.  Also, I_DIRTY_DATASYNC simply "feels" better than just
+ * I_DIRTY_SYNC, since the file data has not actually hit the block device yet,
+ * which is not what I_DIRTY_SYNC on its own would suggest.
  */
 void __mark_mft_record_dirty(ntfs_inode *ni)
 {
diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile
index 01596079dd6..31f25ce32c9 100644
--- a/fs/ocfs2/Makefile
+++ b/fs/ocfs2/Makefile
@@ -28,6 +28,7 @@ ocfs2-objs := \
 	locks.o			\
 	mmap.o 			\
 	namei.o 		\
+	refcounttree.o		\
 	resize.o		\
 	slot_map.o 		\
 	suballoc.o 		\
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index ab513ddaeff..38a42f5d59f 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -49,10 +49,21 @@
 #include "super.h"
 #include "uptodate.h"
 #include "xattr.h"
+#include "refcounttree.h"
 
 #include "buffer_head_io.h"
 
+enum ocfs2_contig_type {
+	CONTIG_NONE = 0,
+	CONTIG_LEFT,
+	CONTIG_RIGHT,
+	CONTIG_LEFTRIGHT,
+};
 
+static enum ocfs2_contig_type
+	ocfs2_extent_rec_contig(struct super_block *sb,
+				struct ocfs2_extent_rec *ext,
+				struct ocfs2_extent_rec *insert_rec);
 /*
  * Operations for a specific extent tree type.
  *
@@ -79,18 +90,30 @@ struct ocfs2_extent_tree_operations {
 	 * that value.  new_clusters is the delta, and must be
 	 * added to the total.  Required.
 	 */
-	void (*eo_update_clusters)(struct inode *inode,
-				   struct ocfs2_extent_tree *et,
+	void (*eo_update_clusters)(struct ocfs2_extent_tree *et,
 				   u32 new_clusters);
 
 	/*
+	 * If this extent tree is supported by an extent map, insert
+	 * a record into the map.
+	 */
+	void (*eo_extent_map_insert)(struct ocfs2_extent_tree *et,
+				     struct ocfs2_extent_rec *rec);
+
+	/*
+	 * If this extent tree is supported by an extent map, truncate the
+	 * map to clusters,
+	 */
+	void (*eo_extent_map_truncate)(struct ocfs2_extent_tree *et,
+				       u32 clusters);
+
+	/*
 	 * If ->eo_insert_check() exists, it is called before rec is
 	 * inserted into the extent tree.  It is optional.
 	 */
-	int (*eo_insert_check)(struct inode *inode,
-			       struct ocfs2_extent_tree *et,
+	int (*eo_insert_check)(struct ocfs2_extent_tree *et,
 			       struct ocfs2_extent_rec *rec);
-	int (*eo_sanity_check)(struct inode *inode, struct ocfs2_extent_tree *et);
+	int (*eo_sanity_check)(struct ocfs2_extent_tree *et);
 
 	/*
 	 * --------------------------------------------------------------
@@ -109,8 +132,17 @@ struct ocfs2_extent_tree_operations {
 	 * it exists.  If it does not, et->et_max_leaf_clusters is set
 	 * to 0 (unlimited).  Optional.
 	 */
-	void (*eo_fill_max_leaf_clusters)(struct inode *inode,
-					  struct ocfs2_extent_tree *et);
+	void (*eo_fill_max_leaf_clusters)(struct ocfs2_extent_tree *et);
+
+	/*
+	 * ->eo_extent_contig test whether the 2 ocfs2_extent_rec
+	 * are contiguous or not. Optional. Don't need to set it if use
+	 * ocfs2_extent_rec as the tree leaf.
+	 */
+	enum ocfs2_contig_type
+		(*eo_extent_contig)(struct ocfs2_extent_tree *et,
+				    struct ocfs2_extent_rec *ext,
+				    struct ocfs2_extent_rec *insert_rec);
 };
 
 
@@ -121,19 +153,22 @@ struct ocfs2_extent_tree_operations {
 static u64 ocfs2_dinode_get_last_eb_blk(struct ocfs2_extent_tree *et);
 static void ocfs2_dinode_set_last_eb_blk(struct ocfs2_extent_tree *et,
 					 u64 blkno);
-static void ocfs2_dinode_update_clusters(struct inode *inode,
-					 struct ocfs2_extent_tree *et,
+static void ocfs2_dinode_update_clusters(struct ocfs2_extent_tree *et,
 					 u32 clusters);
-static int ocfs2_dinode_insert_check(struct inode *inode,
-				     struct ocfs2_extent_tree *et,
+static void ocfs2_dinode_extent_map_insert(struct ocfs2_extent_tree *et,
+					   struct ocfs2_extent_rec *rec);
+static void ocfs2_dinode_extent_map_truncate(struct ocfs2_extent_tree *et,
+					     u32 clusters);
+static int ocfs2_dinode_insert_check(struct ocfs2_extent_tree *et,
 				     struct ocfs2_extent_rec *rec);
-static int ocfs2_dinode_sanity_check(struct inode *inode,
-				     struct ocfs2_extent_tree *et);
+static int ocfs2_dinode_sanity_check(struct ocfs2_extent_tree *et);
 static void ocfs2_dinode_fill_root_el(struct ocfs2_extent_tree *et);
 static struct ocfs2_extent_tree_operations ocfs2_dinode_et_ops = {
 	.eo_set_last_eb_blk	= ocfs2_dinode_set_last_eb_blk,
 	.eo_get_last_eb_blk	= ocfs2_dinode_get_last_eb_blk,
 	.eo_update_clusters	= ocfs2_dinode_update_clusters,
+	.eo_extent_map_insert	= ocfs2_dinode_extent_map_insert,
+	.eo_extent_map_truncate	= ocfs2_dinode_extent_map_truncate,
 	.eo_insert_check	= ocfs2_dinode_insert_check,
 	.eo_sanity_check	= ocfs2_dinode_sanity_check,
 	.eo_fill_root_el	= ocfs2_dinode_fill_root_el,
@@ -156,40 +191,53 @@ static u64 ocfs2_dinode_get_last_eb_blk(struct ocfs2_extent_tree *et)
 	return le64_to_cpu(di->i_last_eb_blk);
 }
 
-static void ocfs2_dinode_update_clusters(struct inode *inode,
-					 struct ocfs2_extent_tree *et,
+static void ocfs2_dinode_update_clusters(struct ocfs2_extent_tree *et,
 					 u32 clusters)
 {
+	struct ocfs2_inode_info *oi = cache_info_to_inode(et->et_ci);
 	struct ocfs2_dinode *di = et->et_object;
 
 	le32_add_cpu(&di->i_clusters, clusters);
-	spin_lock(&OCFS2_I(inode)->ip_lock);
-	OCFS2_I(inode)->ip_clusters = le32_to_cpu(di->i_clusters);
-	spin_unlock(&OCFS2_I(inode)->ip_lock);
+	spin_lock(&oi->ip_lock);
+	oi->ip_clusters = le32_to_cpu(di->i_clusters);
+	spin_unlock(&oi->ip_lock);
 }
 
-static int ocfs2_dinode_insert_check(struct inode *inode,
-				     struct ocfs2_extent_tree *et,
+static void ocfs2_dinode_extent_map_insert(struct ocfs2_extent_tree *et,
+					   struct ocfs2_extent_rec *rec)
+{
+	struct inode *inode = &cache_info_to_inode(et->et_ci)->vfs_inode;
+
+	ocfs2_extent_map_insert_rec(inode, rec);
+}
+
+static void ocfs2_dinode_extent_map_truncate(struct ocfs2_extent_tree *et,
+					     u32 clusters)
+{
+	struct inode *inode = &cache_info_to_inode(et->et_ci)->vfs_inode;
+
+	ocfs2_extent_map_trunc(inode, clusters);
+}
+
+static int ocfs2_dinode_insert_check(struct ocfs2_extent_tree *et,
 				     struct ocfs2_extent_rec *rec)
 {
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_inode_info *oi = cache_info_to_inode(et->et_ci);
+	struct ocfs2_super *osb = OCFS2_SB(oi->vfs_inode.i_sb);
 
-	BUG_ON(OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL);
+	BUG_ON(oi->ip_dyn_features & OCFS2_INLINE_DATA_FL);
 	mlog_bug_on_msg(!ocfs2_sparse_alloc(osb) &&
-			(OCFS2_I(inode)->ip_clusters !=
-			 le32_to_cpu(rec->e_cpos)),
+			(oi->ip_clusters != le32_to_cpu(rec->e_cpos)),
 			"Device %s, asking for sparse allocation: inode %llu, "
 			"cpos %u, clusters %u\n",
 			osb->dev_str,
-			(unsigned long long)OCFS2_I(inode)->ip_blkno,
-			rec->e_cpos,
-			OCFS2_I(inode)->ip_clusters);
+			(unsigned long long)oi->ip_blkno,
+			rec->e_cpos, oi->ip_clusters);
 
 	return 0;
 }
 
-static int ocfs2_dinode_sanity_check(struct inode *inode,
-				     struct ocfs2_extent_tree *et)
+static int ocfs2_dinode_sanity_check(struct ocfs2_extent_tree *et)
 {
 	struct ocfs2_dinode *di = et->et_object;
 
@@ -229,8 +277,7 @@ static u64 ocfs2_xattr_value_get_last_eb_blk(struct ocfs2_extent_tree *et)
 	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,
+static void ocfs2_xattr_value_update_clusters(struct ocfs2_extent_tree *et,
 					      u32 clusters)
 {
 	struct ocfs2_xattr_value_buf *vb = et->et_object;
@@ -252,12 +299,11 @@ static void ocfs2_xattr_tree_fill_root_el(struct ocfs2_extent_tree *et)
 	et->et_root_el = &xb->xb_attrs.xb_root.xt_list;
 }
 
-static void ocfs2_xattr_tree_fill_max_leaf_clusters(struct inode *inode,
-						    struct ocfs2_extent_tree *et)
+static void ocfs2_xattr_tree_fill_max_leaf_clusters(struct ocfs2_extent_tree *et)
 {
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
 	et->et_max_leaf_clusters =
-		ocfs2_clusters_for_bytes(inode->i_sb,
-					 OCFS2_MAX_XATTR_TREE_LEAF_SIZE);
+		ocfs2_clusters_for_bytes(sb, OCFS2_MAX_XATTR_TREE_LEAF_SIZE);
 }
 
 static void ocfs2_xattr_tree_set_last_eb_blk(struct ocfs2_extent_tree *et,
@@ -277,8 +323,7 @@ static u64 ocfs2_xattr_tree_get_last_eb_blk(struct ocfs2_extent_tree *et)
 	return le64_to_cpu(xt->xt_last_eb_blk);
 }
 
-static void ocfs2_xattr_tree_update_clusters(struct inode *inode,
-					     struct ocfs2_extent_tree *et,
+static void ocfs2_xattr_tree_update_clusters(struct ocfs2_extent_tree *et,
 					     u32 clusters)
 {
 	struct ocfs2_xattr_block *xb = et->et_object;
@@ -309,8 +354,7 @@ static u64 ocfs2_dx_root_get_last_eb_blk(struct ocfs2_extent_tree *et)
 	return le64_to_cpu(dx_root->dr_last_eb_blk);
 }
 
-static void ocfs2_dx_root_update_clusters(struct inode *inode,
-					  struct ocfs2_extent_tree *et,
+static void ocfs2_dx_root_update_clusters(struct ocfs2_extent_tree *et,
 					  u32 clusters)
 {
 	struct ocfs2_dx_root_block *dx_root = et->et_object;
@@ -318,8 +362,7 @@ static void ocfs2_dx_root_update_clusters(struct inode *inode,
 	le32_add_cpu(&dx_root->dr_clusters, clusters);
 }
 
-static int ocfs2_dx_root_sanity_check(struct inode *inode,
-				      struct ocfs2_extent_tree *et)
+static int ocfs2_dx_root_sanity_check(struct ocfs2_extent_tree *et)
 {
 	struct ocfs2_dx_root_block *dx_root = et->et_object;
 
@@ -343,8 +386,54 @@ static struct ocfs2_extent_tree_operations ocfs2_dx_root_et_ops = {
 	.eo_fill_root_el	= ocfs2_dx_root_fill_root_el,
 };
 
+static void ocfs2_refcount_tree_fill_root_el(struct ocfs2_extent_tree *et)
+{
+	struct ocfs2_refcount_block *rb = et->et_object;
+
+	et->et_root_el = &rb->rf_list;
+}
+
+static void ocfs2_refcount_tree_set_last_eb_blk(struct ocfs2_extent_tree *et,
+						u64 blkno)
+{
+	struct ocfs2_refcount_block *rb = et->et_object;
+
+	rb->rf_last_eb_blk = cpu_to_le64(blkno);
+}
+
+static u64 ocfs2_refcount_tree_get_last_eb_blk(struct ocfs2_extent_tree *et)
+{
+	struct ocfs2_refcount_block *rb = et->et_object;
+
+	return le64_to_cpu(rb->rf_last_eb_blk);
+}
+
+static void ocfs2_refcount_tree_update_clusters(struct ocfs2_extent_tree *et,
+						u32 clusters)
+{
+	struct ocfs2_refcount_block *rb = et->et_object;
+
+	le32_add_cpu(&rb->rf_clusters, clusters);
+}
+
+static enum ocfs2_contig_type
+ocfs2_refcount_tree_extent_contig(struct ocfs2_extent_tree *et,
+				  struct ocfs2_extent_rec *ext,
+				  struct ocfs2_extent_rec *insert_rec)
+{
+	return CONTIG_NONE;
+}
+
+static struct ocfs2_extent_tree_operations ocfs2_refcount_tree_et_ops = {
+	.eo_set_last_eb_blk	= ocfs2_refcount_tree_set_last_eb_blk,
+	.eo_get_last_eb_blk	= ocfs2_refcount_tree_get_last_eb_blk,
+	.eo_update_clusters	= ocfs2_refcount_tree_update_clusters,
+	.eo_fill_root_el	= ocfs2_refcount_tree_fill_root_el,
+	.eo_extent_contig	= ocfs2_refcount_tree_extent_contig,
+};
+
 static void __ocfs2_init_extent_tree(struct ocfs2_extent_tree *et,
-				     struct inode *inode,
+				     struct ocfs2_caching_info *ci,
 				     struct buffer_head *bh,
 				     ocfs2_journal_access_func access,
 				     void *obj,
@@ -352,6 +441,7 @@ static void __ocfs2_init_extent_tree(struct ocfs2_extent_tree *et,
 {
 	et->et_ops = ops;
 	et->et_root_bh = bh;
+	et->et_ci = ci;
 	et->et_root_journal_access = access;
 	if (!obj)
 		obj = (void *)bh->b_data;
@@ -361,41 +451,49 @@ static void __ocfs2_init_extent_tree(struct ocfs2_extent_tree *et,
 	if (!et->et_ops->eo_fill_max_leaf_clusters)
 		et->et_max_leaf_clusters = 0;
 	else
-		et->et_ops->eo_fill_max_leaf_clusters(inode, et);
+		et->et_ops->eo_fill_max_leaf_clusters(et);
 }
 
 void ocfs2_init_dinode_extent_tree(struct ocfs2_extent_tree *et,
-				   struct inode *inode,
+				   struct ocfs2_caching_info *ci,
 				   struct buffer_head *bh)
 {
-	__ocfs2_init_extent_tree(et, inode, bh, ocfs2_journal_access_di,
+	__ocfs2_init_extent_tree(et, ci, 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 ocfs2_caching_info *ci,
 				       struct buffer_head *bh)
 {
-	__ocfs2_init_extent_tree(et, inode, bh, ocfs2_journal_access_xb,
+	__ocfs2_init_extent_tree(et, ci, 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 ocfs2_caching_info *ci,
 					struct ocfs2_xattr_value_buf *vb)
 {
-	__ocfs2_init_extent_tree(et, inode, vb->vb_bh, vb->vb_access, vb,
+	__ocfs2_init_extent_tree(et, ci, vb->vb_bh, vb->vb_access, vb,
 				 &ocfs2_xattr_value_et_ops);
 }
 
 void ocfs2_init_dx_root_extent_tree(struct ocfs2_extent_tree *et,
-				    struct inode *inode,
+				    struct ocfs2_caching_info *ci,
 				    struct buffer_head *bh)
 {
-	__ocfs2_init_extent_tree(et, inode, bh, ocfs2_journal_access_dr,
+	__ocfs2_init_extent_tree(et, ci, bh, ocfs2_journal_access_dr,
 				 NULL, &ocfs2_dx_root_et_ops);
 }
 
+void ocfs2_init_refcount_extent_tree(struct ocfs2_extent_tree *et,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *bh)
+{
+	__ocfs2_init_extent_tree(et, ci, bh, ocfs2_journal_access_rb,
+				 NULL, &ocfs2_refcount_tree_et_ops);
+}
+
 static inline void ocfs2_et_set_last_eb_blk(struct ocfs2_extent_tree *et,
 					    u64 new_last_eb_blk)
 {
@@ -407,78 +505,71 @@ static inline u64 ocfs2_et_get_last_eb_blk(struct ocfs2_extent_tree *et)
 	return et->et_ops->eo_get_last_eb_blk(et);
 }
 
-static inline void ocfs2_et_update_clusters(struct inode *inode,
-					    struct ocfs2_extent_tree *et,
+static inline void ocfs2_et_update_clusters(struct ocfs2_extent_tree *et,
 					    u32 clusters)
 {
-	et->et_ops->eo_update_clusters(inode, et, clusters);
+	et->et_ops->eo_update_clusters(et, clusters);
+}
+
+static inline void ocfs2_et_extent_map_insert(struct ocfs2_extent_tree *et,
+					      struct ocfs2_extent_rec *rec)
+{
+	if (et->et_ops->eo_extent_map_insert)
+		et->et_ops->eo_extent_map_insert(et, rec);
+}
+
+static inline void ocfs2_et_extent_map_truncate(struct ocfs2_extent_tree *et,
+						u32 clusters)
+{
+	if (et->et_ops->eo_extent_map_truncate)
+		et->et_ops->eo_extent_map_truncate(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,
+	return et->et_root_journal_access(handle, et->et_ci, et->et_root_bh,
 					  type);
 }
 
-static inline int ocfs2_et_insert_check(struct inode *inode,
-					struct ocfs2_extent_tree *et,
+static inline enum ocfs2_contig_type
+	ocfs2_et_extent_contig(struct ocfs2_extent_tree *et,
+			       struct ocfs2_extent_rec *rec,
+			       struct ocfs2_extent_rec *insert_rec)
+{
+	if (et->et_ops->eo_extent_contig)
+		return et->et_ops->eo_extent_contig(et, rec, insert_rec);
+
+	return ocfs2_extent_rec_contig(
+				ocfs2_metadata_cache_get_super(et->et_ci),
+				rec, insert_rec);
+}
+
+static inline int ocfs2_et_insert_check(struct ocfs2_extent_tree *et,
 					struct ocfs2_extent_rec *rec)
 {
 	int ret = 0;
 
 	if (et->et_ops->eo_insert_check)
-		ret = et->et_ops->eo_insert_check(inode, et, rec);
+		ret = et->et_ops->eo_insert_check(et, rec);
 	return ret;
 }
 
-static inline int ocfs2_et_sanity_check(struct inode *inode,
-					struct ocfs2_extent_tree *et)
+static inline int ocfs2_et_sanity_check(struct ocfs2_extent_tree *et)
 {
 	int ret = 0;
 
 	if (et->et_ops->eo_sanity_check)
-		ret = et->et_ops->eo_sanity_check(inode, et);
+		ret = et->et_ops->eo_sanity_check(et);
 	return ret;
 }
 
 static void ocfs2_free_truncate_context(struct ocfs2_truncate_context *tc);
 static int ocfs2_cache_extent_block_free(struct ocfs2_cached_dealloc_ctxt *ctxt,
 					 struct ocfs2_extent_block *eb);
-
-/*
- * Structures which describe a path through a btree, and functions to
- * manipulate them.
- *
- * The idea here is to be as generic as possible with the tree
- * manipulation code.
- */
-struct ocfs2_path_item {
-	struct buffer_head		*bh;
-	struct ocfs2_extent_list	*el;
-};
-
-#define OCFS2_MAX_PATH_DEPTH	5
-
-struct ocfs2_path {
-	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)
-
-static int ocfs2_find_path(struct inode *inode, struct ocfs2_path *path,
-			   u32 cpos);
-static void ocfs2_adjust_rightmost_records(struct inode *inode,
-					   handle_t *handle,
+static void ocfs2_adjust_rightmost_records(handle_t *handle,
+					   struct ocfs2_extent_tree *et,
 					   struct ocfs2_path *path,
 					   struct ocfs2_extent_rec *insert_rec);
 /*
@@ -486,7 +577,7 @@ static void ocfs2_adjust_rightmost_records(struct inode *inode,
  * to build another path. Generally, this involves freeing the buffer
  * heads.
  */
-static void ocfs2_reinit_path(struct ocfs2_path *path, int keep_root)
+void ocfs2_reinit_path(struct ocfs2_path *path, int keep_root)
 {
 	int i, start = 0, depth = 0;
 	struct ocfs2_path_item *node;
@@ -515,7 +606,7 @@ static void ocfs2_reinit_path(struct ocfs2_path *path, int keep_root)
 	path->p_tree_depth = depth;
 }
 
-static void ocfs2_free_path(struct ocfs2_path *path)
+void ocfs2_free_path(struct ocfs2_path *path)
 {
 	if (path) {
 		ocfs2_reinit_path(path, 0);
@@ -613,13 +704,13 @@ static struct ocfs2_path *ocfs2_new_path(struct buffer_head *root_bh,
 	return path;
 }
 
-static struct ocfs2_path *ocfs2_new_path_from_path(struct ocfs2_path *path)
+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)
+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);
@@ -632,10 +723,10 @@ static struct ocfs2_path *ocfs2_new_path_from_et(struct ocfs2_extent_tree *et)
  * 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)
+int ocfs2_path_bh_journal_access(handle_t *handle,
+				 struct ocfs2_caching_info *ci,
+				 struct ocfs2_path *path,
+				 int idx)
 {
 	ocfs2_journal_access_func access = path_root_access(path);
 
@@ -645,15 +736,16 @@ static int ocfs2_path_bh_journal_access(handle_t *handle,
 	if (idx)
 		access = ocfs2_journal_access_eb;
 
-	return access(handle, inode, path->p_node[idx].bh,
+	return access(handle, ci, path->p_node[idx].bh,
 		      OCFS2_JOURNAL_ACCESS_WRITE);
 }
 
 /*
  * Convenience function to journal all components in a path.
  */
-static int ocfs2_journal_access_path(struct inode *inode, handle_t *handle,
-				     struct ocfs2_path *path)
+int ocfs2_journal_access_path(struct ocfs2_caching_info *ci,
+			      handle_t *handle,
+			      struct ocfs2_path *path)
 {
 	int i, ret = 0;
 
@@ -661,7 +753,7 @@ 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_path_bh_journal_access(handle, inode, path, i);
+		ret = ocfs2_path_bh_journal_access(handle, ci, path, i);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -702,17 +794,9 @@ int ocfs2_search_extent_list(struct ocfs2_extent_list *el, u32 v_cluster)
 	return ret;
 }
 
-enum ocfs2_contig_type {
-	CONTIG_NONE = 0,
-	CONTIG_LEFT,
-	CONTIG_RIGHT,
-	CONTIG_LEFTRIGHT,
-};
-
-
 /*
  * NOTE: ocfs2_block_extent_contig(), ocfs2_extents_adjacent() and
- * ocfs2_extent_contig only work properly against leaf nodes!
+ * ocfs2_extent_rec_contig only work properly against leaf nodes!
  */
 static int ocfs2_block_extent_contig(struct super_block *sb,
 				     struct ocfs2_extent_rec *ext,
@@ -738,9 +822,9 @@ static int ocfs2_extents_adjacent(struct ocfs2_extent_rec *left,
 }
 
 static enum ocfs2_contig_type
-	ocfs2_extent_contig(struct inode *inode,
-			    struct ocfs2_extent_rec *ext,
-			    struct ocfs2_extent_rec *insert_rec)
+	ocfs2_extent_rec_contig(struct super_block *sb,
+				struct ocfs2_extent_rec *ext,
+				struct ocfs2_extent_rec *insert_rec)
 {
 	u64 blkno = le64_to_cpu(insert_rec->e_blkno);
 
@@ -753,12 +837,12 @@ static enum ocfs2_contig_type
 		return CONTIG_NONE;
 
 	if (ocfs2_extents_adjacent(ext, insert_rec) &&
-	    ocfs2_block_extent_contig(inode->i_sb, ext, blkno))
+	    ocfs2_block_extent_contig(sb, ext, blkno))
 			return CONTIG_RIGHT;
 
 	blkno = le64_to_cpu(ext->e_blkno);
 	if (ocfs2_extents_adjacent(insert_rec, ext) &&
-	    ocfs2_block_extent_contig(inode->i_sb, insert_rec, blkno))
+	    ocfs2_block_extent_contig(sb, insert_rec, blkno))
 		return CONTIG_LEFT;
 
 	return CONTIG_NONE;
@@ -853,13 +937,13 @@ static int ocfs2_validate_extent_block(struct super_block *sb,
 	return 0;
 }
 
-int ocfs2_read_extent_block(struct inode *inode, u64 eb_blkno,
+int ocfs2_read_extent_block(struct ocfs2_caching_info *ci, u64 eb_blkno,
 			    struct buffer_head **bh)
 {
 	int rc;
 	struct buffer_head *tmp = *bh;
 
-	rc = ocfs2_read_block(inode, eb_blkno, &tmp,
+	rc = ocfs2_read_block(ci, eb_blkno, &tmp,
 			      ocfs2_validate_extent_block);
 
 	/* If ocfs2_read_block() got us a new bh, pass it up. */
@@ -874,7 +958,6 @@ int ocfs2_read_extent_block(struct inode *inode, u64 eb_blkno,
  * How many free extents have we got before we need more meta data?
  */
 int ocfs2_num_free_extents(struct ocfs2_super *osb,
-			   struct inode *inode,
 			   struct ocfs2_extent_tree *et)
 {
 	int retval;
@@ -889,7 +972,8 @@ 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_extent_block(inode, last_eb_blk, &eb_bh);
+		retval = ocfs2_read_extent_block(et->et_ci, last_eb_blk,
+						 &eb_bh);
 		if (retval < 0) {
 			mlog_errno(retval);
 			goto bail;
@@ -913,9 +997,8 @@ bail:
  * sets h_signature, h_blkno, h_suballoc_bit, h_suballoc_slot, and
  * l_count for you
  */
-static int ocfs2_create_new_meta_bhs(struct ocfs2_super *osb,
-				     handle_t *handle,
-				     struct inode *inode,
+static int ocfs2_create_new_meta_bhs(handle_t *handle,
+				     struct ocfs2_extent_tree *et,
 				     int wanted,
 				     struct ocfs2_alloc_context *meta_ac,
 				     struct buffer_head *bhs[])
@@ -924,6 +1007,8 @@ static int ocfs2_create_new_meta_bhs(struct ocfs2_super *osb,
 	u16 suballoc_bit_start;
 	u32 num_got;
 	u64 first_blkno;
+	struct ocfs2_super *osb =
+		OCFS2_SB(ocfs2_metadata_cache_get_super(et->et_ci));
 	struct ocfs2_extent_block *eb;
 
 	mlog_entry_void();
@@ -949,9 +1034,10 @@ static int ocfs2_create_new_meta_bhs(struct ocfs2_super *osb,
 				mlog_errno(status);
 				goto bail;
 			}
-			ocfs2_set_new_buffer_uptodate(inode, bhs[i]);
+			ocfs2_set_new_buffer_uptodate(et->et_ci, bhs[i]);
 
-			status = ocfs2_journal_access_eb(handle, inode, bhs[i],
+			status = ocfs2_journal_access_eb(handle, et->et_ci,
+							 bhs[i],
 							 OCFS2_JOURNAL_ACCESS_CREATE);
 			if (status < 0) {
 				mlog_errno(status);
@@ -1023,7 +1109,6 @@ static inline u32 ocfs2_sum_rightmost_rec(struct ocfs2_extent_list  *el)
  * extent block's rightmost record.
  */
 static int ocfs2_adjust_rightmost_branch(handle_t *handle,
-					 struct inode *inode,
 					 struct ocfs2_extent_tree *et)
 {
 	int status;
@@ -1037,7 +1122,7 @@ static int ocfs2_adjust_rightmost_branch(handle_t *handle,
 		return status;
 	}
 
-	status = ocfs2_find_path(inode, path, UINT_MAX);
+	status = ocfs2_find_path(et->et_ci, path, UINT_MAX);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out;
@@ -1050,7 +1135,7 @@ static int ocfs2_adjust_rightmost_branch(handle_t *handle,
 		goto out;
 	}
 
-	status = ocfs2_journal_access_path(inode, handle, path);
+	status = ocfs2_journal_access_path(et->et_ci, handle, path);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out;
@@ -1059,7 +1144,7 @@ static int ocfs2_adjust_rightmost_branch(handle_t *handle,
 	el = path_leaf_el(path);
 	rec = &el->l_recs[le32_to_cpu(el->l_next_free_rec) - 1];
 
-	ocfs2_adjust_rightmost_records(inode, handle, path, rec);
+	ocfs2_adjust_rightmost_records(handle, et, path, rec);
 
 out:
 	ocfs2_free_path(path);
@@ -1068,7 +1153,7 @@ out:
 
 /*
  * Add an entire tree branch to our inode. eb_bh is the extent block
- * to start at, if we don't want to start the branch at the dinode
+ * to start at, if we don't want to start the branch at the root
  * structure.
  *
  * last_eb_bh is required as we have to update it's next_leaf pointer
@@ -1077,9 +1162,7 @@ out:
  * the new branch will be 'empty' in the sense that every block will
  * contain a single record with cluster count == 0.
  */
-static int ocfs2_add_branch(struct ocfs2_super *osb,
-			    handle_t *handle,
-			    struct inode *inode,
+static int ocfs2_add_branch(handle_t *handle,
 			    struct ocfs2_extent_tree *et,
 			    struct buffer_head *eb_bh,
 			    struct buffer_head **last_eb_bh,
@@ -1123,7 +1206,7 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 	if (root_end > new_cpos) {
 		mlog(0, "adjust the cluster end from %u to %u\n",
 		     root_end, new_cpos);
-		status = ocfs2_adjust_rightmost_branch(handle, inode, et);
+		status = ocfs2_adjust_rightmost_branch(handle, et);
 		if (status) {
 			mlog_errno(status);
 			goto bail;
@@ -1139,7 +1222,7 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_create_new_meta_bhs(osb, handle, inode, new_blocks,
+	status = ocfs2_create_new_meta_bhs(handle, et, new_blocks,
 					   meta_ac, new_eb_bhs);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1161,7 +1244,7 @@ static int ocfs2_add_branch(struct ocfs2_super *osb,
 		BUG_ON(!OCFS2_IS_VALID_EXTENT_BLOCK(eb));
 		eb_el = &eb->h_list;
 
-		status = ocfs2_journal_access_eb(handle, inode, bh,
+		status = ocfs2_journal_access_eb(handle, et->et_ci, bh,
 						 OCFS2_JOURNAL_ACCESS_CREATE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -1201,20 +1284,20 @@ 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_eb(handle, inode, *last_eb_bh,
+	status = ocfs2_journal_access_eb(handle, et->et_ci, *last_eb_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
-	status = ocfs2_et_root_journal_access(handle, inode, et,
+	status = ocfs2_et_root_journal_access(handle, et,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 	if (eb_bh) {
-		status = ocfs2_journal_access_eb(handle, inode, eb_bh,
+		status = ocfs2_journal_access_eb(handle, et->et_ci, eb_bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -1274,9 +1357,7 @@ bail:
  * returns back the new extent block so you can add a branch to it
  * after this call.
  */
-static int ocfs2_shift_tree_depth(struct ocfs2_super *osb,
-				  handle_t *handle,
-				  struct inode *inode,
+static int ocfs2_shift_tree_depth(handle_t *handle,
 				  struct ocfs2_extent_tree *et,
 				  struct ocfs2_alloc_context *meta_ac,
 				  struct buffer_head **ret_new_eb_bh)
@@ -1290,7 +1371,7 @@ static int ocfs2_shift_tree_depth(struct ocfs2_super *osb,
 
 	mlog_entry_void();
 
-	status = ocfs2_create_new_meta_bhs(osb, handle, inode, 1, meta_ac,
+	status = ocfs2_create_new_meta_bhs(handle, et, 1, meta_ac,
 					   &new_eb_bh);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1304,7 +1385,7 @@ static int ocfs2_shift_tree_depth(struct ocfs2_super *osb,
 	eb_el = &eb->h_list;
 	root_el = et->et_root_el;
 
-	status = ocfs2_journal_access_eb(handle, inode, new_eb_bh,
+	status = ocfs2_journal_access_eb(handle, et->et_ci, new_eb_bh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1323,7 +1404,7 @@ static int ocfs2_shift_tree_depth(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_et_root_journal_access(handle, inode, et,
+	status = ocfs2_et_root_journal_access(handle, et,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1379,9 +1460,7 @@ bail:
  *
  * return status < 0 indicates an error.
  */
-static int ocfs2_find_branch_target(struct ocfs2_super *osb,
-				    struct inode *inode,
-				    struct ocfs2_extent_tree *et,
+static int ocfs2_find_branch_target(struct ocfs2_extent_tree *et,
 				    struct buffer_head **target_bh)
 {
 	int status = 0, i;
@@ -1399,19 +1478,21 @@ static int ocfs2_find_branch_target(struct ocfs2_super *osb,
 
 	while(le16_to_cpu(el->l_tree_depth) > 1) {
 		if (le16_to_cpu(el->l_next_free_rec) == 0) {
-			ocfs2_error(inode->i_sb, "Dinode %llu has empty "
+			ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+				    "Owner %llu has empty "
 				    "extent list (next_free_rec == 0)",
-				    (unsigned long long)OCFS2_I(inode)->ip_blkno);
+				    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci));
 			status = -EIO;
 			goto bail;
 		}
 		i = le16_to_cpu(el->l_next_free_rec) - 1;
 		blkno = le64_to_cpu(el->l_recs[i].e_blkno);
 		if (!blkno) {
-			ocfs2_error(inode->i_sb, "Dinode %llu has extent "
+			ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+				    "Owner %llu has extent "
 				    "list where extent # %d has no physical "
 				    "block start",
-				    (unsigned long long)OCFS2_I(inode)->ip_blkno, i);
+				    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci), i);
 			status = -EIO;
 			goto bail;
 		}
@@ -1419,7 +1500,7 @@ static int ocfs2_find_branch_target(struct ocfs2_super *osb,
 		brelse(bh);
 		bh = NULL;
 
-		status = ocfs2_read_extent_block(inode, blkno, &bh);
+		status = ocfs2_read_extent_block(et->et_ci, blkno, &bh);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -1460,20 +1541,18 @@ bail:
  *
  * *last_eb_bh will be updated by ocfs2_add_branch().
  */
-static int ocfs2_grow_tree(struct inode *inode, handle_t *handle,
-			   struct ocfs2_extent_tree *et, int *final_depth,
-			   struct buffer_head **last_eb_bh,
+static int ocfs2_grow_tree(handle_t *handle, struct ocfs2_extent_tree *et,
+			   int *final_depth, struct buffer_head **last_eb_bh,
 			   struct ocfs2_alloc_context *meta_ac)
 {
 	int ret, shift;
 	struct ocfs2_extent_list *el = et->et_root_el;
 	int depth = le16_to_cpu(el->l_tree_depth);
-	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct buffer_head *bh = NULL;
 
 	BUG_ON(meta_ac == NULL);
 
-	shift = ocfs2_find_branch_target(osb, inode, et, &bh);
+	shift = ocfs2_find_branch_target(et, &bh);
 	if (shift < 0) {
 		ret = shift;
 		mlog_errno(ret);
@@ -1490,8 +1569,7 @@ static int ocfs2_grow_tree(struct inode *inode, handle_t *handle,
 		/* ocfs2_shift_tree_depth will return us a buffer with
 		 * the new extent block (so we can pass that to
 		 * ocfs2_add_branch). */
-		ret = ocfs2_shift_tree_depth(osb, handle, inode, et,
-					     meta_ac, &bh);
+		ret = ocfs2_shift_tree_depth(handle, et, meta_ac, &bh);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -1517,7 +1595,7 @@ static int ocfs2_grow_tree(struct inode *inode, handle_t *handle,
 	/* call ocfs2_add_branch to add the final part of the tree with
 	 * the new data. */
 	mlog(0, "add branch. bh = %p\n", bh);
-	ret = ocfs2_add_branch(osb, handle, inode, et, bh, last_eb_bh,
+	ret = ocfs2_add_branch(handle, et, bh, last_eb_bh,
 			       meta_ac);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -1687,7 +1765,7 @@ set_and_inc:
  *
  * The array index of the subtree root is passed back.
  */
-static int ocfs2_find_subtree_root(struct inode *inode,
+static int ocfs2_find_subtree_root(struct ocfs2_extent_tree *et,
 				   struct ocfs2_path *left,
 				   struct ocfs2_path *right)
 {
@@ -1705,10 +1783,10 @@ static int ocfs2_find_subtree_root(struct inode *inode,
 		 * The caller didn't pass two adjacent paths.
 		 */
 		mlog_bug_on_msg(i > left->p_tree_depth,
-				"Inode %lu, left depth %u, right depth %u\n"
+				"Owner %llu, left depth %u, right depth %u\n"
 				"left leaf blk %llu, right leaf blk %llu\n",
-				inode->i_ino, left->p_tree_depth,
-				right->p_tree_depth,
+				(unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+				left->p_tree_depth, right->p_tree_depth,
 				(unsigned long long)path_leaf_bh(left)->b_blocknr,
 				(unsigned long long)path_leaf_bh(right)->b_blocknr);
 	} while (left->p_node[i].bh->b_blocknr ==
@@ -1725,7 +1803,7 @@ typedef void (path_insert_t)(void *, struct buffer_head *);
  * This code can be called with a cpos larger than the tree, in which
  * case it will return the rightmost path.
  */
-static int __ocfs2_find_path(struct inode *inode,
+static int __ocfs2_find_path(struct ocfs2_caching_info *ci,
 			     struct ocfs2_extent_list *root_el, u32 cpos,
 			     path_insert_t *func, void *data)
 {
@@ -1736,15 +1814,14 @@ static int __ocfs2_find_path(struct inode *inode,
 	struct ocfs2_extent_block *eb;
 	struct ocfs2_extent_list *el;
 	struct ocfs2_extent_rec *rec;
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 
 	el = root_el;
 	while (el->l_tree_depth) {
 		if (le16_to_cpu(el->l_next_free_rec) == 0) {
-			ocfs2_error(inode->i_sb,
-				    "Inode %llu has empty extent list at "
+			ocfs2_error(ocfs2_metadata_cache_get_super(ci),
+				    "Owner %llu has empty extent list at "
 				    "depth %u\n",
-				    (unsigned long long)oi->ip_blkno,
+				    (unsigned long long)ocfs2_metadata_cache_owner(ci),
 				    le16_to_cpu(el->l_tree_depth));
 			ret = -EROFS;
 			goto out;
@@ -1767,10 +1844,10 @@ static int __ocfs2_find_path(struct inode *inode,
 
 		blkno = le64_to_cpu(el->l_recs[i].e_blkno);
 		if (blkno == 0) {
-			ocfs2_error(inode->i_sb,
-				    "Inode %llu has bad blkno in extent list "
+			ocfs2_error(ocfs2_metadata_cache_get_super(ci),
+				    "Owner %llu has bad blkno in extent list "
 				    "at depth %u (index %d)\n",
-				    (unsigned long long)oi->ip_blkno,
+				    (unsigned long long)ocfs2_metadata_cache_owner(ci),
 				    le16_to_cpu(el->l_tree_depth), i);
 			ret = -EROFS;
 			goto out;
@@ -1778,7 +1855,7 @@ static int __ocfs2_find_path(struct inode *inode,
 
 		brelse(bh);
 		bh = NULL;
-		ret = ocfs2_read_extent_block(inode, blkno, &bh);
+		ret = ocfs2_read_extent_block(ci, blkno, &bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -1789,10 +1866,10 @@ static int __ocfs2_find_path(struct inode *inode,
 
 		if (le16_to_cpu(el->l_next_free_rec) >
 		    le16_to_cpu(el->l_count)) {
-			ocfs2_error(inode->i_sb,
-				    "Inode %llu has bad count in extent list "
+			ocfs2_error(ocfs2_metadata_cache_get_super(ci),
+				    "Owner %llu has bad count in extent list "
 				    "at block %llu (next free=%u, count=%u)\n",
-				    (unsigned long long)oi->ip_blkno,
+				    (unsigned long long)ocfs2_metadata_cache_owner(ci),
 				    (unsigned long long)bh->b_blocknr,
 				    le16_to_cpu(el->l_next_free_rec),
 				    le16_to_cpu(el->l_count));
@@ -1836,14 +1913,14 @@ static void find_path_ins(void *data, struct buffer_head *bh)
 	ocfs2_path_insert_eb(fp->path, fp->index, bh);
 	fp->index++;
 }
-static int ocfs2_find_path(struct inode *inode, struct ocfs2_path *path,
-			   u32 cpos)
+int ocfs2_find_path(struct ocfs2_caching_info *ci,
+		    struct ocfs2_path *path, u32 cpos)
 {
 	struct find_path_data data;
 
 	data.index = 1;
 	data.path = path;
-	return __ocfs2_find_path(inode, path_root_el(path), cpos,
+	return __ocfs2_find_path(ci, path_root_el(path), cpos,
 				 find_path_ins, &data);
 }
 
@@ -1868,13 +1945,14 @@ static void find_leaf_ins(void *data, struct buffer_head *bh)
  *
  * This function doesn't handle non btree extent lists.
  */
-int ocfs2_find_leaf(struct inode *inode, struct ocfs2_extent_list *root_el,
-		    u32 cpos, struct buffer_head **leaf_bh)
+int ocfs2_find_leaf(struct ocfs2_caching_info *ci,
+		    struct ocfs2_extent_list *root_el, u32 cpos,
+		    struct buffer_head **leaf_bh)
 {
 	int ret;
 	struct buffer_head *bh = NULL;
 
-	ret = __ocfs2_find_path(inode, root_el, cpos, find_leaf_ins, &bh);
+	ret = __ocfs2_find_path(ci, root_el, cpos, find_leaf_ins, &bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -1980,7 +2058,7 @@ static void ocfs2_adjust_root_records(struct ocfs2_extent_list *root_el,
  *   - When we've adjusted the last extent record in the left path leaf and the
  *     1st extent record in the right path leaf during cross extent block merge.
  */
-static void ocfs2_complete_edge_insert(struct inode *inode, handle_t *handle,
+static void ocfs2_complete_edge_insert(handle_t *handle,
 				       struct ocfs2_path *left_path,
 				       struct ocfs2_path *right_path,
 				       int subtree_index)
@@ -2058,8 +2136,8 @@ static void ocfs2_complete_edge_insert(struct inode *inode, handle_t *handle,
 		mlog_errno(ret);
 }
 
-static int ocfs2_rotate_subtree_right(struct inode *inode,
-				      handle_t *handle,
+static int ocfs2_rotate_subtree_right(handle_t *handle,
+				      struct ocfs2_extent_tree *et,
 				      struct ocfs2_path *left_path,
 				      struct ocfs2_path *right_path,
 				      int subtree_index)
@@ -2075,10 +2153,10 @@ static int ocfs2_rotate_subtree_right(struct inode *inode,
 	left_el = path_leaf_el(left_path);
 
 	if (left_el->l_next_free_rec != left_el->l_count) {
-		ocfs2_error(inode->i_sb,
+		ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
 			    "Inode %llu has non-full interior leaf node %llu"
 			    "(next free = %u)",
-			    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+			    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 			    (unsigned long long)left_leaf_bh->b_blocknr,
 			    le16_to_cpu(left_el->l_next_free_rec));
 		return -EROFS;
@@ -2094,7 +2172,7 @@ 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_path_bh_journal_access(handle, inode, right_path,
+	ret = ocfs2_path_bh_journal_access(handle, et->et_ci, right_path,
 					   subtree_index);
 	if (ret) {
 		mlog_errno(ret);
@@ -2102,14 +2180,14 @@ static int ocfs2_rotate_subtree_right(struct inode *inode,
 	}
 
 	for(i = subtree_index + 1; i < path_num_items(right_path); i++) {
-		ret = ocfs2_path_bh_journal_access(handle, inode,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 						   right_path, i);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_path_bh_journal_access(handle, inode,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 						   left_path, i);
 		if (ret) {
 			mlog_errno(ret);
@@ -2123,7 +2201,7 @@ static int ocfs2_rotate_subtree_right(struct inode *inode,
 	/* This is a code error, not a disk corruption. */
 	mlog_bug_on_msg(!right_el->l_next_free_rec, "Inode %llu: Rotate fails "
 			"because rightmost leaf block %llu is empty\n",
-			(unsigned long long)OCFS2_I(inode)->ip_blkno,
+			(unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 			(unsigned long long)right_leaf_bh->b_blocknr);
 
 	ocfs2_create_empty_extent(right_el);
@@ -2157,8 +2235,8 @@ static int ocfs2_rotate_subtree_right(struct inode *inode,
 		goto out;
 	}
 
-	ocfs2_complete_edge_insert(inode, handle, left_path, right_path,
-				subtree_index);
+	ocfs2_complete_edge_insert(handle, left_path, right_path,
+				   subtree_index);
 
 out:
 	return ret;
@@ -2248,10 +2326,18 @@ static int ocfs2_extend_rotate_transaction(handle_t *handle, int subtree_depth,
 					   int op_credits,
 					   struct ocfs2_path *path)
 {
+	int ret;
 	int credits = (path->p_tree_depth - subtree_depth) * 2 + 1 + op_credits;
 
-	if (handle->h_buffer_credits < credits)
-		return ocfs2_extend_trans(handle, credits);
+	if (handle->h_buffer_credits < credits) {
+		ret = ocfs2_extend_trans(handle,
+					 credits - handle->h_buffer_credits);
+		if (ret)
+			return ret;
+
+		if (unlikely(handle->h_buffer_credits < credits))
+			return ocfs2_extend_trans(handle, credits);
+	}
 
 	return 0;
 }
@@ -2321,8 +2407,8 @@ static int ocfs2_leftmost_rec_contains(struct ocfs2_extent_list *el, u32 cpos)
  *   *ret_left_path will contain a valid path which can be passed to
  *   ocfs2_insert_path().
  */
-static int ocfs2_rotate_tree_right(struct inode *inode,
-				   handle_t *handle,
+static int ocfs2_rotate_tree_right(handle_t *handle,
+				   struct ocfs2_extent_tree *et,
 				   enum ocfs2_split_type split,
 				   u32 insert_cpos,
 				   struct ocfs2_path *right_path,
@@ -2331,6 +2417,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 	int ret, start, orig_credits = handle->h_buffer_credits;
 	u32 cpos;
 	struct ocfs2_path *left_path = NULL;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
 
 	*ret_left_path = NULL;
 
@@ -2341,7 +2428,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb, right_path, &cpos);
+	ret = ocfs2_find_cpos_for_left_leaf(sb, right_path, &cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -2379,7 +2466,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 		mlog(0, "Rotating a tree: ins. cpos: %u, left path cpos: %u\n",
 		     insert_cpos, cpos);
 
-		ret = ocfs2_find_path(inode, left_path, cpos);
+		ret = ocfs2_find_path(et->et_ci, left_path, cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2387,10 +2474,11 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 
 		mlog_bug_on_msg(path_leaf_bh(left_path) ==
 				path_leaf_bh(right_path),
-				"Inode %lu: error during insert of %u "
+				"Owner %llu: error during insert of %u "
 				"(left path cpos %u) results in two identical "
 				"paths ending at %llu\n",
-				inode->i_ino, insert_cpos, cpos,
+				(unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+				insert_cpos, cpos,
 				(unsigned long long)
 				path_leaf_bh(left_path)->b_blocknr);
 
@@ -2416,7 +2504,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 			goto out_ret_path;
 		}
 
-		start = ocfs2_find_subtree_root(inode, left_path, right_path);
+		start = ocfs2_find_subtree_root(et, left_path, right_path);
 
 		mlog(0, "Subtree root at index %d (blk %llu, depth %d)\n",
 		     start,
@@ -2430,7 +2518,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 			goto out;
 		}
 
-		ret = ocfs2_rotate_subtree_right(inode, handle, left_path,
+		ret = ocfs2_rotate_subtree_right(handle, et, left_path,
 						 right_path, start);
 		if (ret) {
 			mlog_errno(ret);
@@ -2462,8 +2550,7 @@ static int ocfs2_rotate_tree_right(struct inode *inode,
 		 */
 		ocfs2_mv_path(right_path, left_path);
 
-		ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb, right_path,
-						    &cpos);
+		ret = ocfs2_find_cpos_for_left_leaf(sb, right_path, &cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2477,7 +2564,8 @@ out_ret_path:
 	return ret;
 }
 
-static int ocfs2_update_edge_lengths(struct inode *inode, handle_t *handle,
+static int ocfs2_update_edge_lengths(handle_t *handle,
+				     struct ocfs2_extent_tree *et,
 				     int subtree_index, struct ocfs2_path *path)
 {
 	int i, idx, ret;
@@ -2502,7 +2590,7 @@ static int ocfs2_update_edge_lengths(struct inode *inode, handle_t *handle,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_path(inode, handle, path);
+	ret = ocfs2_journal_access_path(et->et_ci, handle, path);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -2532,7 +2620,8 @@ out:
 	return ret;
 }
 
-static void ocfs2_unlink_path(struct inode *inode, handle_t *handle,
+static void ocfs2_unlink_path(handle_t *handle,
+			      struct ocfs2_extent_tree *et,
 			      struct ocfs2_cached_dealloc_ctxt *dealloc,
 			      struct ocfs2_path *path, int unlink_start)
 {
@@ -2554,12 +2643,12 @@ static void ocfs2_unlink_path(struct inode *inode, handle_t *handle,
 			mlog(ML_ERROR,
 			     "Inode %llu, attempted to remove extent block "
 			     "%llu with %u records\n",
-			     (unsigned long long)OCFS2_I(inode)->ip_blkno,
+			     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 			     (unsigned long long)le64_to_cpu(eb->h_blkno),
 			     le16_to_cpu(el->l_next_free_rec));
 
 			ocfs2_journal_dirty(handle, bh);
-			ocfs2_remove_from_cache(inode, bh);
+			ocfs2_remove_from_cache(et->et_ci, bh);
 			continue;
 		}
 
@@ -2572,11 +2661,12 @@ static void ocfs2_unlink_path(struct inode *inode, handle_t *handle,
 		if (ret)
 			mlog_errno(ret);
 
-		ocfs2_remove_from_cache(inode, bh);
+		ocfs2_remove_from_cache(et->et_ci, bh);
 	}
 }
 
-static void ocfs2_unlink_subtree(struct inode *inode, handle_t *handle,
+static void ocfs2_unlink_subtree(handle_t *handle,
+				 struct ocfs2_extent_tree *et,
 				 struct ocfs2_path *left_path,
 				 struct ocfs2_path *right_path,
 				 int subtree_index,
@@ -2607,17 +2697,17 @@ static void ocfs2_unlink_subtree(struct inode *inode, handle_t *handle,
 	ocfs2_journal_dirty(handle, root_bh);
 	ocfs2_journal_dirty(handle, path_leaf_bh(left_path));
 
-	ocfs2_unlink_path(inode, handle, dealloc, right_path,
+	ocfs2_unlink_path(handle, et, dealloc, right_path,
 			  subtree_index + 1);
 }
 
-static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
+static int ocfs2_rotate_subtree_left(handle_t *handle,
+				     struct ocfs2_extent_tree *et,
 				     struct ocfs2_path *left_path,
 				     struct ocfs2_path *right_path,
 				     int subtree_index,
 				     struct ocfs2_cached_dealloc_ctxt *dealloc,
-				     int *deleted,
-				     struct ocfs2_extent_tree *et)
+				     int *deleted)
 {
 	int ret, i, del_right_subtree = 0, right_has_empty = 0;
 	struct buffer_head *root_bh, *et_root_bh = path_root_bh(right_path);
@@ -2653,7 +2743,7 @@ 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_eb(handle, inode,
+			ret = ocfs2_journal_access_eb(handle, et->et_ci,
 						      path_leaf_bh(right_path),
 						      OCFS2_JOURNAL_ACCESS_WRITE);
 			if (ret) {
@@ -2672,7 +2762,7 @@ 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_et_root_journal_access(handle, inode, et,
+		ret = ocfs2_et_root_journal_access(handle, et,
 						   OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
@@ -2688,7 +2778,7 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 	 */
 	BUG_ON(right_has_empty && !del_right_subtree);
 
-	ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+	ret = ocfs2_path_bh_journal_access(handle, et->et_ci, right_path,
 					   subtree_index);
 	if (ret) {
 		mlog_errno(ret);
@@ -2696,14 +2786,14 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 	}
 
 	for(i = subtree_index + 1; i < path_num_items(right_path); i++) {
-		ret = ocfs2_path_bh_journal_access(handle, inode,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 						   right_path, i);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_path_bh_journal_access(handle, inode,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 						   left_path, i);
 		if (ret) {
 			mlog_errno(ret);
@@ -2740,9 +2830,9 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 		mlog_errno(ret);
 
 	if (del_right_subtree) {
-		ocfs2_unlink_subtree(inode, handle, left_path, right_path,
+		ocfs2_unlink_subtree(handle, et, left_path, right_path,
 				     subtree_index, dealloc);
-		ret = ocfs2_update_edge_lengths(inode, handle, subtree_index,
+		ret = ocfs2_update_edge_lengths(handle, et, subtree_index,
 						left_path);
 		if (ret) {
 			mlog_errno(ret);
@@ -2766,7 +2856,7 @@ static int ocfs2_rotate_subtree_left(struct inode *inode, handle_t *handle,
 
 		*deleted = 1;
 	} else
-		ocfs2_complete_edge_insert(inode, handle, left_path, right_path,
+		ocfs2_complete_edge_insert(handle, left_path, right_path,
 					   subtree_index);
 
 out:
@@ -2852,8 +2942,8 @@ out:
 	return ret;
 }
 
-static int ocfs2_rotate_rightmost_leaf_left(struct inode *inode,
-					    handle_t *handle,
+static int ocfs2_rotate_rightmost_leaf_left(handle_t *handle,
+					    struct ocfs2_extent_tree *et,
 					    struct ocfs2_path *path)
 {
 	int ret;
@@ -2863,7 +2953,7 @@ static int ocfs2_rotate_rightmost_leaf_left(struct inode *inode,
 	if (!ocfs2_is_empty_extent(&el->l_recs[0]))
 		return 0;
 
-	ret = ocfs2_path_bh_journal_access(handle, inode, path,
+	ret = ocfs2_path_bh_journal_access(handle, et->et_ci, path,
 					   path_num_items(path) - 1);
 	if (ret) {
 		mlog_errno(ret);
@@ -2880,24 +2970,24 @@ out:
 	return ret;
 }
 
-static int __ocfs2_rotate_tree_left(struct inode *inode,
-				    handle_t *handle, int orig_credits,
+static int __ocfs2_rotate_tree_left(handle_t *handle,
+				    struct ocfs2_extent_tree *et,
+				    int orig_credits,
 				    struct ocfs2_path *path,
 				    struct ocfs2_cached_dealloc_ctxt *dealloc,
-				    struct ocfs2_path **empty_extent_path,
-				    struct ocfs2_extent_tree *et)
+				    struct ocfs2_path **empty_extent_path)
 {
 	int ret, subtree_root, deleted;
 	u32 right_cpos;
 	struct ocfs2_path *left_path = NULL;
 	struct ocfs2_path *right_path = NULL;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
 
 	BUG_ON(!ocfs2_is_empty_extent(&(path_leaf_el(path)->l_recs[0])));
 
 	*empty_extent_path = NULL;
 
-	ret = ocfs2_find_cpos_for_right_leaf(inode->i_sb, path,
-					     &right_cpos);
+	ret = ocfs2_find_cpos_for_right_leaf(sb, path, &right_cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -2920,13 +3010,13 @@ static int __ocfs2_rotate_tree_left(struct inode *inode,
 	}
 
 	while (right_cpos) {
-		ret = ocfs2_find_path(inode, right_path, right_cpos);
+		ret = ocfs2_find_path(et->et_ci, right_path, right_cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		subtree_root = ocfs2_find_subtree_root(inode, left_path,
+		subtree_root = ocfs2_find_subtree_root(et, left_path,
 						       right_path);
 
 		mlog(0, "Subtree root at index %d (blk %llu, depth %d)\n",
@@ -2946,16 +3036,16 @@ 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_path_bh_journal_access(handle, inode,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 						   left_path, 0);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_rotate_subtree_left(inode, handle, left_path,
+		ret = ocfs2_rotate_subtree_left(handle, et, left_path,
 						right_path, subtree_root,
-						dealloc, &deleted, et);
+						dealloc, &deleted);
 		if (ret == -EAGAIN) {
 			/*
 			 * The rotation has to temporarily stop due to
@@ -2982,7 +3072,7 @@ static int __ocfs2_rotate_tree_left(struct inode *inode,
 
 		ocfs2_mv_path(left_path, right_path);
 
-		ret = ocfs2_find_cpos_for_right_leaf(inode->i_sb, left_path,
+		ret = ocfs2_find_cpos_for_right_leaf(sb, left_path,
 						     &right_cpos);
 		if (ret) {
 			mlog_errno(ret);
@@ -2997,10 +3087,10 @@ out:
 	return ret;
 }
 
-static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
+static int ocfs2_remove_rightmost_path(handle_t *handle,
+				struct ocfs2_extent_tree *et,
 				struct ocfs2_path *path,
-				struct ocfs2_cached_dealloc_ctxt *dealloc,
-				struct ocfs2_extent_tree *et)
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
 {
 	int ret, subtree_index;
 	u32 cpos;
@@ -3009,7 +3099,7 @@ static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
 	struct ocfs2_extent_list *el;
 
 
-	ret = ocfs2_et_sanity_check(inode, et);
+	ret = ocfs2_et_sanity_check(et);
 	if (ret)
 		goto out;
 	/*
@@ -3024,13 +3114,14 @@ static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_path(inode, handle, path);
+	ret = ocfs2_journal_access_path(et->et_ci, handle, path);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb, path, &cpos);
+	ret = ocfs2_find_cpos_for_left_leaf(ocfs2_metadata_cache_get_super(et->et_ci),
+					    path, &cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3048,23 +3139,23 @@ static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
 			goto out;
 		}
 
-		ret = ocfs2_find_path(inode, left_path, cpos);
+		ret = ocfs2_find_path(et->et_ci, left_path, cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_journal_access_path(inode, handle, left_path);
+		ret = ocfs2_journal_access_path(et->et_ci, handle, left_path);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		subtree_index = ocfs2_find_subtree_root(inode, left_path, path);
+		subtree_index = ocfs2_find_subtree_root(et, left_path, path);
 
-		ocfs2_unlink_subtree(inode, handle, left_path, path,
+		ocfs2_unlink_subtree(handle, et, left_path, path,
 				     subtree_index, dealloc);
-		ret = ocfs2_update_edge_lengths(inode, handle, subtree_index,
+		ret = ocfs2_update_edge_lengths(handle, et, subtree_index,
 						left_path);
 		if (ret) {
 			mlog_errno(ret);
@@ -3078,10 +3169,10 @@ static int ocfs2_remove_rightmost_path(struct inode *inode, handle_t *handle,
 		 * 'path' is also the leftmost path which
 		 * means it must be the only one. This gets
 		 * handled differently because we want to
-		 * revert the inode back to having extents
+		 * revert the root back to having extents
 		 * in-line.
 		 */
-		ocfs2_unlink_path(inode, handle, dealloc, path, 1);
+		ocfs2_unlink_path(handle, et, dealloc, path, 1);
 
 		el = et->et_root_el;
 		el->l_tree_depth = 0;
@@ -3114,10 +3205,10 @@ out:
  * the rightmost tree leaf record is removed so the caller is
  * responsible for detecting and correcting that.
  */
-static int ocfs2_rotate_tree_left(struct inode *inode, handle_t *handle,
+static int ocfs2_rotate_tree_left(handle_t *handle,
+				  struct ocfs2_extent_tree *et,
 				  struct ocfs2_path *path,
-				  struct ocfs2_cached_dealloc_ctxt *dealloc,
-				  struct ocfs2_extent_tree *et)
+				  struct ocfs2_cached_dealloc_ctxt *dealloc)
 {
 	int ret, orig_credits = handle->h_buffer_credits;
 	struct ocfs2_path *tmp_path = NULL, *restart_path = NULL;
@@ -3134,8 +3225,7 @@ rightmost_no_delete:
 		 * Inline extents. This is trivially handled, so do
 		 * it up front.
 		 */
-		ret = ocfs2_rotate_rightmost_leaf_left(inode, handle,
-						       path);
+		ret = ocfs2_rotate_rightmost_leaf_left(handle, et, path);
 		if (ret)
 			mlog_errno(ret);
 		goto out;
@@ -3151,7 +3241,7 @@ rightmost_no_delete:
 	 *
 	 *  1) is handled via ocfs2_rotate_rightmost_leaf_left()
 	 *  2a) we need the left branch so that we can update it with the unlink
-	 *  2b) we need to bring the inode back to inline extents.
+	 *  2b) we need to bring the root back to inline extents.
 	 */
 
 	eb = (struct ocfs2_extent_block *)path_leaf_bh(path)->b_data;
@@ -3167,9 +3257,9 @@ rightmost_no_delete:
 
 		if (le16_to_cpu(el->l_next_free_rec) == 0) {
 			ret = -EIO;
-			ocfs2_error(inode->i_sb,
-				    "Inode %llu has empty extent block at %llu",
-				    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+			ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+				    "Owner %llu has empty extent block at %llu",
+				    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 				    (unsigned long long)le64_to_cpu(eb->h_blkno));
 			goto out;
 		}
@@ -3183,8 +3273,8 @@ rightmost_no_delete:
 		 * nonempty list.
 		 */
 
-		ret = ocfs2_remove_rightmost_path(inode, handle, path,
-						  dealloc, et);
+		ret = ocfs2_remove_rightmost_path(handle, et, path,
+						  dealloc);
 		if (ret)
 			mlog_errno(ret);
 		goto out;
@@ -3195,8 +3285,8 @@ rightmost_no_delete:
 	 * and restarting from there.
 	 */
 try_rotate:
-	ret = __ocfs2_rotate_tree_left(inode, handle, orig_credits, path,
-				       dealloc, &restart_path, et);
+	ret = __ocfs2_rotate_tree_left(handle, et, orig_credits, path,
+				       dealloc, &restart_path);
 	if (ret && ret != -EAGAIN) {
 		mlog_errno(ret);
 		goto out;
@@ -3206,9 +3296,9 @@ try_rotate:
 		tmp_path = restart_path;
 		restart_path = NULL;
 
-		ret = __ocfs2_rotate_tree_left(inode, handle, orig_credits,
+		ret = __ocfs2_rotate_tree_left(handle, et, orig_credits,
 					       tmp_path, dealloc,
-					       &restart_path, et);
+					       &restart_path);
 		if (ret && ret != -EAGAIN) {
 			mlog_errno(ret);
 			goto out;
@@ -3259,7 +3349,7 @@ static void ocfs2_cleanup_merge(struct ocfs2_extent_list *el,
 	}
 }
 
-static int ocfs2_get_right_path(struct inode *inode,
+static int ocfs2_get_right_path(struct ocfs2_extent_tree *et,
 				struct ocfs2_path *left_path,
 				struct ocfs2_path **ret_right_path)
 {
@@ -3276,8 +3366,8 @@ static int ocfs2_get_right_path(struct inode *inode,
 	left_el = path_leaf_el(left_path);
 	BUG_ON(left_el->l_next_free_rec != left_el->l_count);
 
-	ret = ocfs2_find_cpos_for_right_leaf(inode->i_sb, left_path,
-					     &right_cpos);
+	ret = ocfs2_find_cpos_for_right_leaf(ocfs2_metadata_cache_get_super(et->et_ci),
+					     left_path, &right_cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3293,7 +3383,7 @@ static int ocfs2_get_right_path(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_find_path(inode, right_path, right_cpos);
+	ret = ocfs2_find_path(et->et_ci, right_path, right_cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3313,9 +3403,9 @@ out:
  * For index == l_count - 1, the "next" means the 1st extent rec of the
  * next extent block.
  */
-static int ocfs2_merge_rec_right(struct inode *inode,
-				 struct ocfs2_path *left_path,
+static int ocfs2_merge_rec_right(struct ocfs2_path *left_path,
 				 handle_t *handle,
+				 struct ocfs2_extent_tree *et,
 				 struct ocfs2_extent_rec *split_rec,
 				 int index)
 {
@@ -3336,7 +3426,7 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 	if (index == le16_to_cpu(el->l_next_free_rec) - 1 &&
 	    le16_to_cpu(el->l_next_free_rec) == le16_to_cpu(el->l_count)) {
 		/* we meet with a cross extent block merge. */
-		ret = ocfs2_get_right_path(inode, left_path, &right_path);
+		ret = ocfs2_get_right_path(et, left_path, &right_path);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3355,8 +3445,8 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 		       le16_to_cpu(left_rec->e_leaf_clusters) !=
 		       le32_to_cpu(right_rec->e_cpos));
 
-		subtree_index = ocfs2_find_subtree_root(inode,
-							left_path, right_path);
+		subtree_index = ocfs2_find_subtree_root(et, left_path,
+							right_path);
 
 		ret = ocfs2_extend_rotate_transaction(handle, subtree_index,
 						      handle->h_buffer_credits,
@@ -3369,7 +3459,7 @@ 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_path_bh_journal_access(handle, inode, right_path,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci, right_path,
 						   subtree_index);
 		if (ret) {
 			mlog_errno(ret);
@@ -3378,14 +3468,14 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 
 		for (i = subtree_index + 1;
 		     i < path_num_items(right_path); i++) {
-			ret = ocfs2_path_bh_journal_access(handle, inode,
+			ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 							   right_path, i);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
 			}
 
-			ret = ocfs2_path_bh_journal_access(handle, inode,
+			ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 							   left_path, i);
 			if (ret) {
 				mlog_errno(ret);
@@ -3398,7 +3488,7 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 		right_rec = &el->l_recs[index + 1];
 	}
 
-	ret = ocfs2_path_bh_journal_access(handle, inode, left_path,
+	ret = ocfs2_path_bh_journal_access(handle, et->et_ci, left_path,
 					   path_num_items(left_path) - 1);
 	if (ret) {
 		mlog_errno(ret);
@@ -3409,7 +3499,8 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 
 	le32_add_cpu(&right_rec->e_cpos, -split_clusters);
 	le64_add_cpu(&right_rec->e_blkno,
-		     -ocfs2_clusters_to_blocks(inode->i_sb, split_clusters));
+		     -ocfs2_clusters_to_blocks(ocfs2_metadata_cache_get_super(et->et_ci),
+					       split_clusters));
 	le16_add_cpu(&right_rec->e_leaf_clusters, split_clusters);
 
 	ocfs2_cleanup_merge(el, index);
@@ -3423,8 +3514,8 @@ static int ocfs2_merge_rec_right(struct inode *inode,
 		if (ret)
 			mlog_errno(ret);
 
-		ocfs2_complete_edge_insert(inode, handle, left_path,
-					   right_path, subtree_index);
+		ocfs2_complete_edge_insert(handle, left_path, right_path,
+					   subtree_index);
 	}
 out:
 	if (right_path)
@@ -3432,7 +3523,7 @@ out:
 	return ret;
 }
 
-static int ocfs2_get_left_path(struct inode *inode,
+static int ocfs2_get_left_path(struct ocfs2_extent_tree *et,
 			       struct ocfs2_path *right_path,
 			       struct ocfs2_path **ret_left_path)
 {
@@ -3445,7 +3536,7 @@ static int ocfs2_get_left_path(struct inode *inode,
 	/* This function shouldn't be called for non-trees. */
 	BUG_ON(right_path->p_tree_depth == 0);
 
-	ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb,
+	ret = ocfs2_find_cpos_for_left_leaf(ocfs2_metadata_cache_get_super(et->et_ci),
 					    right_path, &left_cpos);
 	if (ret) {
 		mlog_errno(ret);
@@ -3462,7 +3553,7 @@ static int ocfs2_get_left_path(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_find_path(inode, left_path, left_cpos);
+	ret = ocfs2_find_path(et->et_ci, left_path, left_cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3485,12 +3576,11 @@ out:
  * remove the rightmost leaf extent block in the right_path and change
  * the right path to indicate the new rightmost path.
  */
-static int ocfs2_merge_rec_left(struct inode *inode,
-				struct ocfs2_path *right_path,
+static int ocfs2_merge_rec_left(struct ocfs2_path *right_path,
 				handle_t *handle,
+				struct ocfs2_extent_tree *et,
 				struct ocfs2_extent_rec *split_rec,
 				struct ocfs2_cached_dealloc_ctxt *dealloc,
-				struct ocfs2_extent_tree *et,
 				int index)
 {
 	int ret, i, subtree_index = 0, has_empty_extent = 0;
@@ -3508,7 +3598,7 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 	right_rec = &el->l_recs[index];
 	if (index == 0) {
 		/* we meet with a cross extent block merge. */
-		ret = ocfs2_get_left_path(inode, right_path, &left_path);
+		ret = ocfs2_get_left_path(et, right_path, &left_path);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3524,8 +3614,8 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 		       le16_to_cpu(left_rec->e_leaf_clusters) !=
 		       le32_to_cpu(split_rec->e_cpos));
 
-		subtree_index = ocfs2_find_subtree_root(inode,
-							left_path, right_path);
+		subtree_index = ocfs2_find_subtree_root(et, left_path,
+							right_path);
 
 		ret = ocfs2_extend_rotate_transaction(handle, subtree_index,
 						      handle->h_buffer_credits,
@@ -3538,7 +3628,7 @@ 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_path_bh_journal_access(handle, inode, right_path,
+		ret = ocfs2_path_bh_journal_access(handle, et->et_ci, right_path,
 						   subtree_index);
 		if (ret) {
 			mlog_errno(ret);
@@ -3547,14 +3637,14 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 
 		for (i = subtree_index + 1;
 		     i < path_num_items(right_path); i++) {
-			ret = ocfs2_path_bh_journal_access(handle, inode,
+			ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 							   right_path, i);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
 			}
 
-			ret = ocfs2_path_bh_journal_access(handle, inode,
+			ret = ocfs2_path_bh_journal_access(handle, et->et_ci,
 							   left_path, i);
 			if (ret) {
 				mlog_errno(ret);
@@ -3567,7 +3657,7 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 			has_empty_extent = 1;
 	}
 
-	ret = ocfs2_path_bh_journal_access(handle, inode, right_path,
+	ret = ocfs2_path_bh_journal_access(handle, et->et_ci, right_path,
 					   path_num_items(right_path) - 1);
 	if (ret) {
 		mlog_errno(ret);
@@ -3586,7 +3676,8 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 
 	le32_add_cpu(&right_rec->e_cpos, split_clusters);
 	le64_add_cpu(&right_rec->e_blkno,
-		     ocfs2_clusters_to_blocks(inode->i_sb, split_clusters));
+		     ocfs2_clusters_to_blocks(ocfs2_metadata_cache_get_super(et->et_ci),
+					      split_clusters));
 	le16_add_cpu(&right_rec->e_leaf_clusters, -split_clusters);
 
 	ocfs2_cleanup_merge(el, index);
@@ -3608,9 +3699,9 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 		if (le16_to_cpu(right_rec->e_leaf_clusters) == 0 &&
 		    le16_to_cpu(el->l_next_free_rec) == 1) {
 
-			ret = ocfs2_remove_rightmost_path(inode, handle,
+			ret = ocfs2_remove_rightmost_path(handle, et,
 							  right_path,
-							  dealloc, et);
+							  dealloc);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -3622,7 +3713,7 @@ static int ocfs2_merge_rec_left(struct inode *inode,
 			ocfs2_mv_path(right_path, left_path);
 			left_path = NULL;
 		} else
-			ocfs2_complete_edge_insert(inode, handle, left_path,
+			ocfs2_complete_edge_insert(handle, left_path,
 						   right_path, subtree_index);
 	}
 out:
@@ -3631,15 +3722,13 @@ out:
 	return ret;
 }
 
-static int ocfs2_try_to_merge_extent(struct inode *inode,
-				     handle_t *handle,
+static int ocfs2_try_to_merge_extent(handle_t *handle,
+				     struct ocfs2_extent_tree *et,
 				     struct ocfs2_path *path,
 				     int split_index,
 				     struct ocfs2_extent_rec *split_rec,
 				     struct ocfs2_cached_dealloc_ctxt *dealloc,
-				     struct ocfs2_merge_ctxt *ctxt,
-				     struct ocfs2_extent_tree *et)
-
+				     struct ocfs2_merge_ctxt *ctxt)
 {
 	int ret = 0;
 	struct ocfs2_extent_list *el = path_leaf_el(path);
@@ -3655,8 +3744,7 @@ static int ocfs2_try_to_merge_extent(struct inode *inode,
 		 * extents - having more than one in a leaf is
 		 * illegal.
 		 */
-		ret = ocfs2_rotate_tree_left(inode, handle, path,
-					     dealloc, et);
+		ret = ocfs2_rotate_tree_left(handle, et, path, dealloc);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3685,8 +3773,7 @@ static int ocfs2_try_to_merge_extent(struct inode *inode,
 		 * prevoius extent block. It is more efficient and easier
 		 * if we do merge_right first and merge_left later.
 		 */
-		ret = ocfs2_merge_rec_right(inode, path,
-					    handle, split_rec,
+		ret = ocfs2_merge_rec_right(path, handle, et, split_rec,
 					    split_index);
 		if (ret) {
 			mlog_errno(ret);
@@ -3699,8 +3786,7 @@ static int ocfs2_try_to_merge_extent(struct inode *inode,
 		BUG_ON(!ocfs2_is_empty_extent(&el->l_recs[0]));
 
 		/* The merge left us with an empty extent, remove it. */
-		ret = ocfs2_rotate_tree_left(inode, handle, path,
-					     dealloc, et);
+		ret = ocfs2_rotate_tree_left(handle, et, path, dealloc);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3712,18 +3798,15 @@ static int ocfs2_try_to_merge_extent(struct inode *inode,
 		 * Note that we don't pass split_rec here on purpose -
 		 * we've merged it into the rec already.
 		 */
-		ret = ocfs2_merge_rec_left(inode, path,
-					   handle, rec,
-					   dealloc, et,
-					   split_index);
+		ret = ocfs2_merge_rec_left(path, handle, et, rec,
+					   dealloc, split_index);
 
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		ret = ocfs2_rotate_tree_left(inode, handle, path,
-					     dealloc, et);
+		ret = ocfs2_rotate_tree_left(handle, et, path, dealloc);
 		/*
 		 * Error from this last rotate is not critical, so
 		 * print but don't bubble it up.
@@ -3740,19 +3823,16 @@ static int ocfs2_try_to_merge_extent(struct inode *inode,
 		 * the record on the left (hence the left merge).
 		 */
 		if (ctxt->c_contig_type == CONTIG_RIGHT) {
-			ret = ocfs2_merge_rec_left(inode,
-						   path,
-						   handle, split_rec,
-						   dealloc, et,
+			ret = ocfs2_merge_rec_left(path, handle, et,
+						   split_rec, dealloc,
 						   split_index);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
 			}
 		} else {
-			ret = ocfs2_merge_rec_right(inode,
-						    path,
-						    handle, split_rec,
+			ret = ocfs2_merge_rec_right(path, handle,
+						    et, split_rec,
 						    split_index);
 			if (ret) {
 				mlog_errno(ret);
@@ -3765,8 +3845,8 @@ static int ocfs2_try_to_merge_extent(struct inode *inode,
 			 * The merge may have left an empty extent in
 			 * our leaf. Try to rotate it away.
 			 */
-			ret = ocfs2_rotate_tree_left(inode, handle, path,
-						     dealloc, et);
+			ret = ocfs2_rotate_tree_left(handle, et, path,
+						     dealloc);
 			if (ret)
 				mlog_errno(ret);
 			ret = 0;
@@ -3812,10 +3892,10 @@ static void ocfs2_subtract_from_rec(struct super_block *sb,
  * list. If this leaf is part of an allocation tree, it is assumed
  * that the tree above has been prepared.
  */
-static void ocfs2_insert_at_leaf(struct ocfs2_extent_rec *insert_rec,
+static void ocfs2_insert_at_leaf(struct ocfs2_extent_tree *et,
+				 struct ocfs2_extent_rec *insert_rec,
 				 struct ocfs2_extent_list *el,
-				 struct ocfs2_insert_type *insert,
-				 struct inode *inode)
+				 struct ocfs2_insert_type *insert)
 {
 	int i = insert->ins_contig_index;
 	unsigned int range;
@@ -3827,7 +3907,8 @@ static void ocfs2_insert_at_leaf(struct ocfs2_extent_rec *insert_rec,
 		i = ocfs2_search_extent_list(el, le32_to_cpu(insert_rec->e_cpos));
 		BUG_ON(i == -1);
 		rec = &el->l_recs[i];
-		ocfs2_subtract_from_rec(inode->i_sb, insert->ins_split, rec,
+		ocfs2_subtract_from_rec(ocfs2_metadata_cache_get_super(et->et_ci),
+					insert->ins_split, rec,
 					insert_rec);
 		goto rotate;
 	}
@@ -3869,10 +3950,10 @@ static void ocfs2_insert_at_leaf(struct ocfs2_extent_rec *insert_rec,
 
 		mlog_bug_on_msg(le16_to_cpu(el->l_next_free_rec) >=
 				le16_to_cpu(el->l_count),
-				"inode %lu, depth %u, count %u, next free %u, "
+				"owner %llu, depth %u, count %u, next free %u, "
 				"rec.cpos %u, rec.clusters %u, "
 				"insert.cpos %u, insert.clusters %u\n",
-				inode->i_ino,
+				ocfs2_metadata_cache_owner(et->et_ci),
 				le16_to_cpu(el->l_tree_depth),
 				le16_to_cpu(el->l_count),
 				le16_to_cpu(el->l_next_free_rec),
@@ -3900,8 +3981,8 @@ rotate:
 	ocfs2_rotate_leaf(el, insert_rec);
 }
 
-static void ocfs2_adjust_rightmost_records(struct inode *inode,
-					   handle_t *handle,
+static void ocfs2_adjust_rightmost_records(handle_t *handle,
+					   struct ocfs2_extent_tree *et,
 					   struct ocfs2_path *path,
 					   struct ocfs2_extent_rec *insert_rec)
 {
@@ -3919,9 +4000,9 @@ static void ocfs2_adjust_rightmost_records(struct inode *inode,
 
 		next_free = le16_to_cpu(el->l_next_free_rec);
 		if (next_free == 0) {
-			ocfs2_error(inode->i_sb,
-				    "Dinode %llu has a bad extent list",
-				    (unsigned long long)OCFS2_I(inode)->ip_blkno);
+			ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+				    "Owner %llu has a bad extent list",
+				    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci));
 			ret = -EIO;
 			return;
 		}
@@ -3941,7 +4022,8 @@ static void ocfs2_adjust_rightmost_records(struct inode *inode,
 	}
 }
 
-static int ocfs2_append_rec_to_path(struct inode *inode, handle_t *handle,
+static int ocfs2_append_rec_to_path(handle_t *handle,
+				    struct ocfs2_extent_tree *et,
 				    struct ocfs2_extent_rec *insert_rec,
 				    struct ocfs2_path *right_path,
 				    struct ocfs2_path **ret_left_path)
@@ -3969,8 +4051,8 @@ static int ocfs2_append_rec_to_path(struct inode *inode, handle_t *handle,
 	    (next_free == 1 && ocfs2_is_empty_extent(&el->l_recs[0]))) {
 		u32 left_cpos;
 
-		ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb, right_path,
-						    &left_cpos);
+		ret = ocfs2_find_cpos_for_left_leaf(ocfs2_metadata_cache_get_super(et->et_ci),
+						    right_path, &left_cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -3992,7 +4074,8 @@ static int ocfs2_append_rec_to_path(struct inode *inode, handle_t *handle,
 				goto out;
 			}
 
-			ret = ocfs2_find_path(inode, left_path, left_cpos);
+			ret = ocfs2_find_path(et->et_ci, left_path,
+					      left_cpos);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -4005,13 +4088,13 @@ static int ocfs2_append_rec_to_path(struct inode *inode, handle_t *handle,
 		}
 	}
 
-	ret = ocfs2_journal_access_path(inode, handle, right_path);
+	ret = ocfs2_journal_access_path(et->et_ci, handle, right_path);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ocfs2_adjust_rightmost_records(inode, handle, right_path, insert_rec);
+	ocfs2_adjust_rightmost_records(handle, et, right_path, insert_rec);
 
 	*ret_left_path = left_path;
 	ret = 0;
@@ -4022,7 +4105,7 @@ out:
 	return ret;
 }
 
-static void ocfs2_split_record(struct inode *inode,
+static void ocfs2_split_record(struct ocfs2_extent_tree *et,
 			       struct ocfs2_path *left_path,
 			       struct ocfs2_path *right_path,
 			       struct ocfs2_extent_rec *split_rec,
@@ -4095,7 +4178,8 @@ static void ocfs2_split_record(struct inode *inode,
 	}
 
 	rec = &el->l_recs[index];
-	ocfs2_subtract_from_rec(inode->i_sb, split, rec, split_rec);
+	ocfs2_subtract_from_rec(ocfs2_metadata_cache_get_super(et->et_ci),
+				split, rec, split_rec);
 	ocfs2_rotate_leaf(insert_el, split_rec);
 }
 
@@ -4107,8 +4191,8 @@ static void ocfs2_split_record(struct inode *inode,
  * in. left_path should only be passed in if we need to update that
  * portion of the tree after an edge insert.
  */
-static int ocfs2_insert_path(struct inode *inode,
-			     handle_t *handle,
+static int ocfs2_insert_path(handle_t *handle,
+			     struct ocfs2_extent_tree *et,
 			     struct ocfs2_path *left_path,
 			     struct ocfs2_path *right_path,
 			     struct ocfs2_extent_rec *insert_rec,
@@ -4134,7 +4218,7 @@ static int ocfs2_insert_path(struct inode *inode,
 			goto out;
 		}
 
-		ret = ocfs2_journal_access_path(inode, handle, left_path);
+		ret = ocfs2_journal_access_path(et->et_ci, handle, left_path);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -4145,7 +4229,7 @@ static int ocfs2_insert_path(struct inode *inode,
 	 * Pass both paths to the journal. The majority of inserts
 	 * will be touching all components anyway.
 	 */
-	ret = ocfs2_journal_access_path(inode, handle, right_path);
+	ret = ocfs2_journal_access_path(et->et_ci, handle, right_path);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -4157,7 +4241,7 @@ static int ocfs2_insert_path(struct inode *inode,
 		 * of splits, but it's easier to just let one separate
 		 * function sort it all out.
 		 */
-		ocfs2_split_record(inode, left_path, right_path,
+		ocfs2_split_record(et, left_path, right_path,
 				   insert_rec, insert->ins_split);
 
 		/*
@@ -4171,8 +4255,8 @@ static int ocfs2_insert_path(struct inode *inode,
 			if (ret)
 				mlog_errno(ret);
 	} else
-		ocfs2_insert_at_leaf(insert_rec, path_leaf_el(right_path),
-				     insert, inode);
+		ocfs2_insert_at_leaf(et, insert_rec, path_leaf_el(right_path),
+				     insert);
 
 	ret = ocfs2_journal_dirty(handle, leaf_bh);
 	if (ret)
@@ -4185,10 +4269,10 @@ static int ocfs2_insert_path(struct inode *inode,
 		 *
 		 * XXX: Should we extend the transaction here?
 		 */
-		subtree_index = ocfs2_find_subtree_root(inode, left_path,
+		subtree_index = ocfs2_find_subtree_root(et, left_path,
 							right_path);
-		ocfs2_complete_edge_insert(inode, handle, left_path,
-					   right_path, subtree_index);
+		ocfs2_complete_edge_insert(handle, left_path, right_path,
+					   subtree_index);
 	}
 
 	ret = 0;
@@ -4196,8 +4280,7 @@ out:
 	return ret;
 }
 
-static int ocfs2_do_insert_extent(struct inode *inode,
-				  handle_t *handle,
+static int ocfs2_do_insert_extent(handle_t *handle,
 				  struct ocfs2_extent_tree *et,
 				  struct ocfs2_extent_rec *insert_rec,
 				  struct ocfs2_insert_type *type)
@@ -4210,7 +4293,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 
 	el = et->et_root_el;
 
-	ret = ocfs2_et_root_journal_access(handle, inode, et,
+	ret = ocfs2_et_root_journal_access(handle, et,
 					   OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -4218,7 +4301,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 	}
 
 	if (le16_to_cpu(el->l_tree_depth) == 0) {
-		ocfs2_insert_at_leaf(insert_rec, el, type, inode);
+		ocfs2_insert_at_leaf(et, insert_rec, el, type);
 		goto out_update_clusters;
 	}
 
@@ -4241,7 +4324,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 		cpos = UINT_MAX;
 	}
 
-	ret = ocfs2_find_path(inode, right_path, cpos);
+	ret = ocfs2_find_path(et->et_ci, right_path, cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -4260,7 +4343,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 	 * can wind up skipping both of these two special cases...
 	 */
 	if (rotate) {
-		ret = ocfs2_rotate_tree_right(inode, handle, type->ins_split,
+		ret = ocfs2_rotate_tree_right(handle, et, type->ins_split,
 					      le32_to_cpu(insert_rec->e_cpos),
 					      right_path, &left_path);
 		if (ret) {
@@ -4272,7 +4355,7 @@ 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_et_root_journal_access(handle, inode, et,
+		ret = ocfs2_et_root_journal_access(handle, et,
 						   OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
@@ -4280,7 +4363,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 		}
 	} else if (type->ins_appending == APPEND_TAIL
 		   && type->ins_contig != CONTIG_LEFT) {
-		ret = ocfs2_append_rec_to_path(inode, handle, insert_rec,
+		ret = ocfs2_append_rec_to_path(handle, et, insert_rec,
 					       right_path, &left_path);
 		if (ret) {
 			mlog_errno(ret);
@@ -4288,7 +4371,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 		}
 	}
 
-	ret = ocfs2_insert_path(inode, handle, left_path, right_path,
+	ret = ocfs2_insert_path(handle, et, left_path, right_path,
 				insert_rec, type);
 	if (ret) {
 		mlog_errno(ret);
@@ -4297,7 +4380,7 @@ static int ocfs2_do_insert_extent(struct inode *inode,
 
 out_update_clusters:
 	if (type->ins_split == SPLIT_NONE)
-		ocfs2_et_update_clusters(inode, et,
+		ocfs2_et_update_clusters(et,
 					 le16_to_cpu(insert_rec->e_leaf_clusters));
 
 	ret = ocfs2_journal_dirty(handle, et->et_root_bh);
@@ -4312,7 +4395,8 @@ out:
 }
 
 static enum ocfs2_contig_type
-ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
+ocfs2_figure_merge_contig_type(struct ocfs2_extent_tree *et,
+			       struct ocfs2_path *path,
 			       struct ocfs2_extent_list *el, int index,
 			       struct ocfs2_extent_rec *split_rec)
 {
@@ -4324,12 +4408,12 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 	struct ocfs2_path *left_path = NULL, *right_path = NULL;
 	struct buffer_head *bh;
 	struct ocfs2_extent_block *eb;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
 
 	if (index > 0) {
 		rec = &el->l_recs[index - 1];
 	} else if (path->p_tree_depth > 0) {
-		status = ocfs2_find_cpos_for_left_leaf(inode->i_sb,
-						       path, &left_cpos);
+		status = ocfs2_find_cpos_for_left_leaf(sb, path, &left_cpos);
 		if (status)
 			goto out;
 
@@ -4338,7 +4422,8 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 			if (!left_path)
 				goto out;
 
-			status = ocfs2_find_path(inode, left_path, left_cpos);
+			status = ocfs2_find_path(et->et_ci, left_path,
+						 left_cpos);
 			if (status)
 				goto out;
 
@@ -4348,7 +4433,7 @@ 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_error(inode->i_sb,
+				ocfs2_error(sb,
 					    "Extent block #%llu has an "
 					    "invalid l_next_free_rec of "
 					    "%d.  It should have "
@@ -4373,7 +4458,7 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 			if (split_rec->e_cpos == el->l_recs[index].e_cpos)
 				ret = CONTIG_RIGHT;
 		} else {
-			ret = ocfs2_extent_contig(inode, rec, split_rec);
+			ret = ocfs2_et_extent_contig(et, rec, split_rec);
 		}
 	}
 
@@ -4382,8 +4467,7 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 		rec = &el->l_recs[index + 1];
 	else if (le16_to_cpu(el->l_next_free_rec) == le16_to_cpu(el->l_count) &&
 		 path->p_tree_depth > 0) {
-		status = ocfs2_find_cpos_for_right_leaf(inode->i_sb,
-							path, &right_cpos);
+		status = ocfs2_find_cpos_for_right_leaf(sb, path, &right_cpos);
 		if (status)
 			goto out;
 
@@ -4394,7 +4478,7 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 		if (!right_path)
 			goto out;
 
-		status = ocfs2_find_path(inode, right_path, right_cpos);
+		status = ocfs2_find_path(et->et_ci, right_path, right_cpos);
 		if (status)
 			goto out;
 
@@ -4404,7 +4488,7 @@ 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_error(inode->i_sb,
+				ocfs2_error(sb,
 					    "Extent block #%llu has an "
 					    "invalid l_next_free_rec of %d",
 					    (unsigned long long)le64_to_cpu(eb->h_blkno),
@@ -4419,7 +4503,7 @@ ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path,
 	if (rec) {
 		enum ocfs2_contig_type contig_type;
 
-		contig_type = ocfs2_extent_contig(inode, rec, split_rec);
+		contig_type = ocfs2_et_extent_contig(et, rec, split_rec);
 
 		if (contig_type == CONTIG_LEFT && ret == CONTIG_RIGHT)
 			ret = CONTIG_LEFTRIGHT;
@@ -4436,11 +4520,10 @@ out:
 	return ret;
 }
 
-static void ocfs2_figure_contig_type(struct inode *inode,
+static void ocfs2_figure_contig_type(struct ocfs2_extent_tree *et,
 				     struct ocfs2_insert_type *insert,
 				     struct ocfs2_extent_list *el,
-				     struct ocfs2_extent_rec *insert_rec,
-				     struct ocfs2_extent_tree *et)
+				     struct ocfs2_extent_rec *insert_rec)
 {
 	int i;
 	enum ocfs2_contig_type contig_type = CONTIG_NONE;
@@ -4448,8 +4531,8 @@ static void ocfs2_figure_contig_type(struct inode *inode,
 	BUG_ON(le16_to_cpu(el->l_tree_depth) != 0);
 
 	for(i = 0; i < le16_to_cpu(el->l_next_free_rec); i++) {
-		contig_type = ocfs2_extent_contig(inode, &el->l_recs[i],
-						  insert_rec);
+		contig_type = ocfs2_et_extent_contig(et, &el->l_recs[i],
+						     insert_rec);
 		if (contig_type != CONTIG_NONE) {
 			insert->ins_contig_index = i;
 			break;
@@ -4530,8 +4613,7 @@ set_tail_append:
  * All of the information is stored on the ocfs2_insert_type
  * structure.
  */
-static int ocfs2_figure_insert_type(struct inode *inode,
-				    struct ocfs2_extent_tree *et,
+static int ocfs2_figure_insert_type(struct ocfs2_extent_tree *et,
 				    struct buffer_head **last_eb_bh,
 				    struct ocfs2_extent_rec *insert_rec,
 				    int *free_records,
@@ -4555,7 +4637,7 @@ static int ocfs2_figure_insert_type(struct inode *inode,
 		 * ocfs2_figure_insert_type() and ocfs2_add_branch()
 		 * may want it later.
 		 */
-		ret = ocfs2_read_extent_block(inode,
+		ret = ocfs2_read_extent_block(et->et_ci,
 					      ocfs2_et_get_last_eb_blk(et),
 					      &bh);
 		if (ret) {
@@ -4578,7 +4660,7 @@ static int ocfs2_figure_insert_type(struct inode *inode,
 		le16_to_cpu(el->l_next_free_rec);
 
 	if (!insert->ins_tree_depth) {
-		ocfs2_figure_contig_type(inode, insert, el, insert_rec, et);
+		ocfs2_figure_contig_type(et, insert, el, insert_rec);
 		ocfs2_figure_appending_type(insert, el, insert_rec);
 		return 0;
 	}
@@ -4596,7 +4678,7 @@ static int ocfs2_figure_insert_type(struct inode *inode,
 	 * us the rightmost tree path. This is accounted for below in
 	 * the appending code.
 	 */
-	ret = ocfs2_find_path(inode, path, le32_to_cpu(insert_rec->e_cpos));
+	ret = ocfs2_find_path(et->et_ci, path, le32_to_cpu(insert_rec->e_cpos));
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -4612,7 +4694,7 @@ static int ocfs2_figure_insert_type(struct inode *inode,
          *     into two types of appends: simple record append, or a
          *     rotate inside the tail leaf.
 	 */
-	ocfs2_figure_contig_type(inode, insert, el, insert_rec, et);
+	ocfs2_figure_contig_type(et, insert, el, insert_rec);
 
 	/*
 	 * The insert code isn't quite ready to deal with all cases of
@@ -4657,13 +4739,11 @@ out:
 }
 
 /*
- * Insert an extent into an inode btree.
+ * Insert an extent into a btree.
  *
- * The caller needs to update fe->i_clusters
+ * The caller needs to update the owning btree's cluster count.
  */
-int ocfs2_insert_extent(struct ocfs2_super *osb,
-			handle_t *handle,
-			struct inode *inode,
+int ocfs2_insert_extent(handle_t *handle,
 			struct ocfs2_extent_tree *et,
 			u32 cpos,
 			u64 start_blk,
@@ -4677,21 +4757,22 @@ int ocfs2_insert_extent(struct ocfs2_super *osb,
 	struct ocfs2_insert_type insert = {0, };
 	struct ocfs2_extent_rec rec;
 
-	mlog(0, "add %u clusters at position %u to inode %llu\n",
-	     new_clusters, cpos, (unsigned long long)OCFS2_I(inode)->ip_blkno);
+	mlog(0, "add %u clusters at position %u to owner %llu\n",
+	     new_clusters, cpos,
+	     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci));
 
 	memset(&rec, 0, sizeof(rec));
 	rec.e_cpos = cpu_to_le32(cpos);
 	rec.e_blkno = cpu_to_le64(start_blk);
 	rec.e_leaf_clusters = cpu_to_le16(new_clusters);
 	rec.e_flags = flags;
-	status = ocfs2_et_insert_check(inode, et, &rec);
+	status = ocfs2_et_insert_check(et, &rec);
 	if (status) {
 		mlog_errno(status);
 		goto bail;
 	}
 
-	status = ocfs2_figure_insert_type(inode, et, &last_eb_bh, &rec,
+	status = ocfs2_figure_insert_type(et, &last_eb_bh, &rec,
 					  &free_records, &insert);
 	if (status < 0) {
 		mlog_errno(status);
@@ -4705,7 +4786,7 @@ int ocfs2_insert_extent(struct ocfs2_super *osb,
 	     free_records, insert.ins_tree_depth);
 
 	if (insert.ins_contig == CONTIG_NONE && free_records == 0) {
-		status = ocfs2_grow_tree(inode, handle, et,
+		status = ocfs2_grow_tree(handle, et,
 					 &insert.ins_tree_depth, &last_eb_bh,
 					 meta_ac);
 		if (status) {
@@ -4715,11 +4796,11 @@ int ocfs2_insert_extent(struct ocfs2_super *osb,
 	}
 
 	/* Finally, we can add clusters. This might rotate the tree for us. */
-	status = ocfs2_do_insert_extent(inode, handle, et, &rec, &insert);
+	status = ocfs2_do_insert_extent(handle, et, &rec, &insert);
 	if (status < 0)
 		mlog_errno(status);
-	else if (et->et_ops == &ocfs2_dinode_et_ops)
-		ocfs2_extent_map_insert_rec(inode, &rec);
+	else
+		ocfs2_et_extent_map_insert(et, &rec);
 
 bail:
 	brelse(last_eb_bh);
@@ -4735,13 +4816,11 @@ bail:
  * it is not limited to the file storage. Any extent tree can use this
  * function if it implements the proper ocfs2_extent_tree.
  */
-int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
-				struct inode *inode,
+int ocfs2_add_clusters_in_btree(handle_t *handle,
+				struct ocfs2_extent_tree *et,
 				u32 *logical_offset,
 				u32 clusters_to_add,
 				int mark_unwritten,
-				struct ocfs2_extent_tree *et,
-				handle_t *handle,
 				struct ocfs2_alloc_context *data_ac,
 				struct ocfs2_alloc_context *meta_ac,
 				enum ocfs2_alloc_restarted *reason_ret)
@@ -4752,13 +4831,15 @@ int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
 	u32 bit_off, num_bits;
 	u64 block;
 	u8 flags = 0;
+	struct ocfs2_super *osb =
+		OCFS2_SB(ocfs2_metadata_cache_get_super(et->et_ci));
 
 	BUG_ON(!clusters_to_add);
 
 	if (mark_unwritten)
 		flags = OCFS2_EXT_UNWRITTEN;
 
-	free_extents = ocfs2_num_free_extents(osb, inode, et);
+	free_extents = ocfs2_num_free_extents(osb, et);
 	if (free_extents < 0) {
 		status = free_extents;
 		mlog_errno(status);
@@ -4795,7 +4876,7 @@ 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 tree root */
-	status = ocfs2_et_root_journal_access(handle, inode, et,
+	status = ocfs2_et_root_journal_access(handle, et,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -4803,10 +4884,10 @@ int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
 	}
 
 	block = ocfs2_clusters_to_blocks(osb->sb, bit_off);
-	mlog(0, "Allocating %u clusters at block %u for inode %llu\n",
-	     num_bits, bit_off, (unsigned long long)OCFS2_I(inode)->ip_blkno);
-	status = ocfs2_insert_extent(osb, handle, inode, et,
-				     *logical_offset, block,
+	mlog(0, "Allocating %u clusters at block %u for owner %llu\n",
+	     num_bits, bit_off,
+	     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci));
+	status = ocfs2_insert_extent(handle, et, *logical_offset, block,
 				     num_bits, flags, meta_ac);
 	if (status < 0) {
 		mlog_errno(status);
@@ -4856,10 +4937,9 @@ static void ocfs2_make_right_split_rec(struct super_block *sb,
 	split_rec->e_flags = rec->e_flags;
 }
 
-static int ocfs2_split_and_insert(struct inode *inode,
-				  handle_t *handle,
-				  struct ocfs2_path *path,
+static int ocfs2_split_and_insert(handle_t *handle,
 				  struct ocfs2_extent_tree *et,
+				  struct ocfs2_path *path,
 				  struct buffer_head **last_eb_bh,
 				  int split_index,
 				  struct ocfs2_extent_rec *orig_split_rec,
@@ -4892,7 +4972,7 @@ leftright:
 
 	if (le16_to_cpu(rightmost_el->l_next_free_rec) ==
 	    le16_to_cpu(rightmost_el->l_count)) {
-		ret = ocfs2_grow_tree(inode, handle, et,
+		ret = ocfs2_grow_tree(handle, et,
 				      &depth, last_eb_bh, meta_ac);
 		if (ret) {
 			mlog_errno(ret);
@@ -4921,8 +5001,8 @@ leftright:
 		 */
 		insert.ins_split = SPLIT_RIGHT;
 
-		ocfs2_make_right_split_rec(inode->i_sb, &tmprec, insert_range,
-					   &rec);
+		ocfs2_make_right_split_rec(ocfs2_metadata_cache_get_super(et->et_ci),
+					   &tmprec, insert_range, &rec);
 
 		split_rec = tmprec;
 
@@ -4930,7 +5010,7 @@ leftright:
 		do_leftright = 1;
 	}
 
-	ret = ocfs2_do_insert_extent(inode, handle, et, &split_rec, &insert);
+	ret = ocfs2_do_insert_extent(handle, et, &split_rec, &insert);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -4946,7 +5026,7 @@ leftright:
 		ocfs2_reinit_path(path, 1);
 
 		cpos = le32_to_cpu(split_rec.e_cpos);
-		ret = ocfs2_find_path(inode, path, cpos);
+		ret = ocfs2_find_path(et->et_ci, path, cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -4961,8 +5041,8 @@ out:
 	return ret;
 }
 
-static int ocfs2_replace_extent_rec(struct inode *inode,
-				    handle_t *handle,
+static int ocfs2_replace_extent_rec(handle_t *handle,
+				    struct ocfs2_extent_tree *et,
 				    struct ocfs2_path *path,
 				    struct ocfs2_extent_list *el,
 				    int split_index,
@@ -4970,7 +5050,7 @@ static int ocfs2_replace_extent_rec(struct inode *inode,
 {
 	int ret;
 
-	ret = ocfs2_path_bh_journal_access(handle, inode, path,
+	ret = ocfs2_path_bh_journal_access(handle, et->et_ci, path,
 					   path_num_items(path) - 1);
 	if (ret) {
 		mlog_errno(ret);
@@ -4985,9 +5065,8 @@ out:
 }
 
 /*
- * Mark part or all of the extent record at split_index in the leaf
- * pointed to by path as written. This removes the unwritten
- * extent flag.
+ * Split part or all of the extent record at split_index in the leaf
+ * pointed to by path. Merge with the contiguous extent record if needed.
  *
  * Care is taken to handle contiguousness so as to not grow the tree.
  *
@@ -5004,14 +5083,13 @@ out:
  * have been brought into cache (and pinned via the journal), so the
  * extra overhead is not expressed in terms of disk reads.
  */
-static int __ocfs2_mark_extent_written(struct inode *inode,
-				       struct ocfs2_extent_tree *et,
-				       handle_t *handle,
-				       struct ocfs2_path *path,
-				       int split_index,
-				       struct ocfs2_extent_rec *split_rec,
-				       struct ocfs2_alloc_context *meta_ac,
-				       struct ocfs2_cached_dealloc_ctxt *dealloc)
+int ocfs2_split_extent(handle_t *handle,
+		       struct ocfs2_extent_tree *et,
+		       struct ocfs2_path *path,
+		       int split_index,
+		       struct ocfs2_extent_rec *split_rec,
+		       struct ocfs2_alloc_context *meta_ac,
+		       struct ocfs2_cached_dealloc_ctxt *dealloc)
 {
 	int ret = 0;
 	struct ocfs2_extent_list *el = path_leaf_el(path);
@@ -5020,12 +5098,6 @@ static int __ocfs2_mark_extent_written(struct inode *inode,
 	struct ocfs2_merge_ctxt ctxt;
 	struct ocfs2_extent_list *rightmost_el;
 
-	if (!(rec->e_flags & OCFS2_EXT_UNWRITTEN)) {
-		ret = -EIO;
-		mlog_errno(ret);
-		goto out;
-	}
-
 	if (le32_to_cpu(rec->e_cpos) > le32_to_cpu(split_rec->e_cpos) ||
 	    ((le32_to_cpu(rec->e_cpos) + le16_to_cpu(rec->e_leaf_clusters)) <
 	     (le32_to_cpu(split_rec->e_cpos) + le16_to_cpu(split_rec->e_leaf_clusters)))) {
@@ -5034,19 +5106,19 @@ static int __ocfs2_mark_extent_written(struct inode *inode,
 		goto out;
 	}
 
-	ctxt.c_contig_type = ocfs2_figure_merge_contig_type(inode, path, el,
+	ctxt.c_contig_type = ocfs2_figure_merge_contig_type(et, path, el,
 							    split_index,
 							    split_rec);
 
 	/*
 	 * The core merge / split code wants to know how much room is
-	 * left in this inodes allocation tree, so we pass the
+	 * left in this allocation tree, so we pass the
 	 * rightmost extent list.
 	 */
 	if (path->p_tree_depth) {
 		struct ocfs2_extent_block *eb;
 
-		ret = ocfs2_read_extent_block(inode,
+		ret = ocfs2_read_extent_block(et->et_ci,
 					      ocfs2_et_get_last_eb_blk(et),
 					      &last_eb_bh);
 		if (ret) {
@@ -5073,19 +5145,18 @@ static int __ocfs2_mark_extent_written(struct inode *inode,
 
 	if (ctxt.c_contig_type == CONTIG_NONE) {
 		if (ctxt.c_split_covers_rec)
-			ret = ocfs2_replace_extent_rec(inode, handle,
-						       path, el,
+			ret = ocfs2_replace_extent_rec(handle, et, path, el,
 						       split_index, split_rec);
 		else
-			ret = ocfs2_split_and_insert(inode, handle, path, et,
+			ret = ocfs2_split_and_insert(handle, et, path,
 						     &last_eb_bh, split_index,
 						     split_rec, meta_ac);
 		if (ret)
 			mlog_errno(ret);
 	} else {
-		ret = ocfs2_try_to_merge_extent(inode, handle, path,
+		ret = ocfs2_try_to_merge_extent(handle, et, path,
 						split_index, split_rec,
-						dealloc, &ctxt, et);
+						dealloc, &ctxt);
 		if (ret)
 			mlog_errno(ret);
 	}
@@ -5096,46 +5167,31 @@ out:
 }
 
 /*
- * Mark the already-existing extent at cpos as written for len clusters.
+ * Change the flags of the already-existing extent at cpos for len clusters.
+ *
+ * new_flags: the flags we want to set.
+ * clear_flags: the flags we want to clear.
+ * phys: the new physical offset we want this new extent starts from.
  *
  * If the existing extent is larger than the request, initiate a
  * split. An attempt will be made at merging with adjacent extents.
  *
  * The caller is responsible for passing down meta_ac if we'll need it.
  */
-int ocfs2_mark_extent_written(struct inode *inode,
-			      struct ocfs2_extent_tree *et,
-			      handle_t *handle, u32 cpos, u32 len, u32 phys,
-			      struct ocfs2_alloc_context *meta_ac,
-			      struct ocfs2_cached_dealloc_ctxt *dealloc)
+int ocfs2_change_extent_flag(handle_t *handle,
+			     struct ocfs2_extent_tree *et,
+			     u32 cpos, u32 len, u32 phys,
+			     struct ocfs2_alloc_context *meta_ac,
+			     struct ocfs2_cached_dealloc_ctxt *dealloc,
+			     int new_flags, int clear_flags)
 {
 	int ret, index;
-	u64 start_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys);
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
+	u64 start_blkno = ocfs2_clusters_to_blocks(sb, phys);
 	struct ocfs2_extent_rec split_rec;
 	struct ocfs2_path *left_path = NULL;
 	struct ocfs2_extent_list *el;
-
-	mlog(0, "Inode %lu cpos %u, len %u, phys %u (%llu)\n",
-	     inode->i_ino, cpos, len, phys, (unsigned long long)start_blkno);
-
-	if (!ocfs2_writes_unwritten_extents(OCFS2_SB(inode->i_sb))) {
-		ocfs2_error(inode->i_sb, "Inode %llu has unwritten extents "
-			    "that are being written to, but the feature bit "
-			    "is not set in the super block.",
-			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
-		ret = -EROFS;
-		goto out;
-	}
-
-	/*
-	 * XXX: This should be fixed up so that we just re-insert the
-	 * next extent records.
-	 *
-	 * XXX: This is a hack on the extent tree, maybe it should be
-	 * an op?
-	 */
-	if (et->et_ops == &ocfs2_dinode_et_ops)
-		ocfs2_extent_map_trunc(inode, 0);
+	struct ocfs2_extent_rec *rec;
 
 	left_path = ocfs2_new_path_from_et(et);
 	if (!left_path) {
@@ -5144,7 +5200,7 @@ int ocfs2_mark_extent_written(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_find_path(inode, left_path, cpos);
+	ret = ocfs2_find_path(et->et_ci, left_path, cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -5153,34 +5209,102 @@ int ocfs2_mark_extent_written(struct inode *inode,
 
 	index = ocfs2_search_extent_list(el, cpos);
 	if (index == -1 || index >= le16_to_cpu(el->l_next_free_rec)) {
-		ocfs2_error(inode->i_sb,
-			    "Inode %llu has an extent at cpos %u which can no "
+		ocfs2_error(sb,
+			    "Owner %llu has an extent at cpos %u which can no "
 			    "longer be found.\n",
-			    (unsigned long long)OCFS2_I(inode)->ip_blkno, cpos);
+			     (unsigned long long)
+			     ocfs2_metadata_cache_owner(et->et_ci), cpos);
 		ret = -EROFS;
 		goto out;
 	}
 
+	ret = -EIO;
+	rec = &el->l_recs[index];
+	if (new_flags && (rec->e_flags & new_flags)) {
+		mlog(ML_ERROR, "Owner %llu tried to set %d flags on an "
+		     "extent that already had them",
+		     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+		     new_flags);
+		goto out;
+	}
+
+	if (clear_flags && !(rec->e_flags & clear_flags)) {
+		mlog(ML_ERROR, "Owner %llu tried to clear %d flags on an "
+		     "extent that didn't have them",
+		     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+		     clear_flags);
+		goto out;
+	}
+
 	memset(&split_rec, 0, sizeof(struct ocfs2_extent_rec));
 	split_rec.e_cpos = cpu_to_le32(cpos);
 	split_rec.e_leaf_clusters = cpu_to_le16(len);
 	split_rec.e_blkno = cpu_to_le64(start_blkno);
-	split_rec.e_flags = path_leaf_el(left_path)->l_recs[index].e_flags;
-	split_rec.e_flags &= ~OCFS2_EXT_UNWRITTEN;
-
-	ret = __ocfs2_mark_extent_written(inode, et, handle, left_path,
-					  index, &split_rec, meta_ac,
-					  dealloc);
+	split_rec.e_flags = rec->e_flags;
+	if (new_flags)
+		split_rec.e_flags |= new_flags;
+	if (clear_flags)
+		split_rec.e_flags &= ~clear_flags;
+
+	ret = ocfs2_split_extent(handle, et, left_path,
+				 index, &split_rec, meta_ac,
+				 dealloc);
 	if (ret)
 		mlog_errno(ret);
 
 out:
 	ocfs2_free_path(left_path);
 	return ret;
+
 }
 
-static int ocfs2_split_tree(struct inode *inode, struct ocfs2_extent_tree *et,
-			    handle_t *handle, struct ocfs2_path *path,
+/*
+ * Mark the already-existing extent at cpos as written for len clusters.
+ * This removes the unwritten extent flag.
+ *
+ * If the existing extent is larger than the request, initiate a
+ * split. An attempt will be made at merging with adjacent extents.
+ *
+ * The caller is responsible for passing down meta_ac if we'll need it.
+ */
+int ocfs2_mark_extent_written(struct inode *inode,
+			      struct ocfs2_extent_tree *et,
+			      handle_t *handle, u32 cpos, u32 len, u32 phys,
+			      struct ocfs2_alloc_context *meta_ac,
+			      struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret;
+
+	mlog(0, "Inode %lu cpos %u, len %u, phys clusters %u\n",
+	     inode->i_ino, cpos, len, phys);
+
+	if (!ocfs2_writes_unwritten_extents(OCFS2_SB(inode->i_sb))) {
+		ocfs2_error(inode->i_sb, "Inode %llu has unwritten extents "
+			    "that are being written to, but the feature bit "
+			    "is not set in the super block.",
+			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
+		ret = -EROFS;
+		goto out;
+	}
+
+	/*
+	 * XXX: This should be fixed up so that we just re-insert the
+	 * next extent records.
+	 */
+	ocfs2_et_extent_map_truncate(et, 0);
+
+	ret = ocfs2_change_extent_flag(handle, et, cpos,
+				       len, phys, meta_ac, dealloc,
+				       0, OCFS2_EXT_UNWRITTEN);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	return ret;
+}
+
+static int ocfs2_split_tree(handle_t *handle, struct ocfs2_extent_tree *et,
+			    struct ocfs2_path *path,
 			    int index, u32 new_range,
 			    struct ocfs2_alloc_context *meta_ac)
 {
@@ -5197,11 +5321,12 @@ static int ocfs2_split_tree(struct inode *inode, struct ocfs2_extent_tree *et,
 	 */
 	el = path_leaf_el(path);
 	rec = &el->l_recs[index];
-	ocfs2_make_right_split_rec(inode->i_sb, &split_rec, new_range, rec);
+	ocfs2_make_right_split_rec(ocfs2_metadata_cache_get_super(et->et_ci),
+				   &split_rec, new_range, rec);
 
 	depth = path->p_tree_depth;
 	if (depth > 0) {
-		ret = ocfs2_read_extent_block(inode,
+		ret = ocfs2_read_extent_block(et->et_ci,
 					      ocfs2_et_get_last_eb_blk(et),
 					      &last_eb_bh);
 		if (ret < 0) {
@@ -5224,7 +5349,7 @@ static int ocfs2_split_tree(struct inode *inode, struct ocfs2_extent_tree *et,
 
 	if (le16_to_cpu(rightmost_el->l_next_free_rec) ==
 	    le16_to_cpu(rightmost_el->l_count)) {
-		ret = ocfs2_grow_tree(inode, handle, et, &depth, &last_eb_bh,
+		ret = ocfs2_grow_tree(handle, et, &depth, &last_eb_bh,
 				      meta_ac);
 		if (ret) {
 			mlog_errno(ret);
@@ -5238,7 +5363,7 @@ static int ocfs2_split_tree(struct inode *inode, struct ocfs2_extent_tree *et,
 	insert.ins_split = SPLIT_RIGHT;
 	insert.ins_tree_depth = depth;
 
-	ret = ocfs2_do_insert_extent(inode, handle, et, &split_rec, &insert);
+	ret = ocfs2_do_insert_extent(handle, et, &split_rec, &insert);
 	if (ret)
 		mlog_errno(ret);
 
@@ -5247,23 +5372,23 @@ out:
 	return ret;
 }
 
-static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
+static int ocfs2_truncate_rec(handle_t *handle,
+			      struct ocfs2_extent_tree *et,
 			      struct ocfs2_path *path, int index,
 			      struct ocfs2_cached_dealloc_ctxt *dealloc,
-			      u32 cpos, u32 len,
-			      struct ocfs2_extent_tree *et)
+			      u32 cpos, u32 len)
 {
 	int ret;
 	u32 left_cpos, rec_range, trunc_range;
 	int wants_rotate = 0, is_rightmost_tree_rec = 0;
-	struct super_block *sb = inode->i_sb;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
 	struct ocfs2_path *left_path = NULL;
 	struct ocfs2_extent_list *el = path_leaf_el(path);
 	struct ocfs2_extent_rec *rec;
 	struct ocfs2_extent_block *eb;
 
 	if (ocfs2_is_empty_extent(&el->l_recs[0]) && index > 0) {
-		ret = ocfs2_rotate_tree_left(inode, handle, path, dealloc, et);
+		ret = ocfs2_rotate_tree_left(handle, et, path, dealloc);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -5295,14 +5420,13 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 		 * by this leaf and the one to it's left.
 		 *
 		 * There are two cases we can skip:
-		 *   1) Path is the leftmost one in our inode tree.
+		 *   1) Path is the leftmost one in our btree.
 		 *   2) The leaf is rightmost and will be empty after
 		 *      we remove the extent record - the rotate code
 		 *      knows how to update the newly formed edge.
 		 */
 
-		ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb, path,
-						    &left_cpos);
+		ret = ocfs2_find_cpos_for_left_leaf(sb, path, &left_cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -5316,7 +5440,8 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 				goto out;
 			}
 
-			ret = ocfs2_find_path(inode, left_path, left_cpos);
+			ret = ocfs2_find_path(et->et_ci, left_path,
+					      left_cpos);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -5332,13 +5457,13 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_path(inode, handle, path);
+	ret = ocfs2_journal_access_path(et->et_ci, handle, path);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_path(inode, handle, left_path);
+	ret = ocfs2_journal_access_path(et->et_ci, handle, left_path);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -5361,7 +5486,7 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 			 * be deleted by the rotate code.
 			 */
 			rec = &el->l_recs[next_free - 1];
-			ocfs2_adjust_rightmost_records(inode, handle, path,
+			ocfs2_adjust_rightmost_records(handle, et, path,
 						       rec);
 		}
 	} else if (le32_to_cpu(rec->e_cpos) == cpos) {
@@ -5373,11 +5498,12 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 		/* Remove rightmost portion of the record */
 		le16_add_cpu(&rec->e_leaf_clusters, -len);
 		if (is_rightmost_tree_rec)
-			ocfs2_adjust_rightmost_records(inode, handle, path, rec);
+			ocfs2_adjust_rightmost_records(handle, et, path, rec);
 	} else {
 		/* Caller should have trapped this. */
-		mlog(ML_ERROR, "Inode %llu: Invalid record truncate: (%u, %u) "
-		     "(%u, %u)\n", (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		mlog(ML_ERROR, "Owner %llu: Invalid record truncate: (%u, %u) "
+		     "(%u, %u)\n",
+		     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 		     le32_to_cpu(rec->e_cpos),
 		     le16_to_cpu(rec->e_leaf_clusters), cpos, len);
 		BUG();
@@ -5386,14 +5512,14 @@ static int ocfs2_truncate_rec(struct inode *inode, handle_t *handle,
 	if (left_path) {
 		int subtree_index;
 
-		subtree_index = ocfs2_find_subtree_root(inode, left_path, path);
-		ocfs2_complete_edge_insert(inode, handle, left_path, path,
+		subtree_index = ocfs2_find_subtree_root(et, left_path, path);
+		ocfs2_complete_edge_insert(handle, left_path, path,
 					   subtree_index);
 	}
 
 	ocfs2_journal_dirty(handle, path_leaf_bh(path));
 
-	ret = ocfs2_rotate_tree_left(inode, handle, path, dealloc, et);
+	ret = ocfs2_rotate_tree_left(handle, et, path, dealloc);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -5404,9 +5530,9 @@ out:
 	return ret;
 }
 
-int ocfs2_remove_extent(struct inode *inode,
+int ocfs2_remove_extent(handle_t *handle,
 			struct ocfs2_extent_tree *et,
-			u32 cpos, u32 len, handle_t *handle,
+			u32 cpos, u32 len,
 			struct ocfs2_alloc_context *meta_ac,
 			struct ocfs2_cached_dealloc_ctxt *dealloc)
 {
@@ -5416,7 +5542,11 @@ int ocfs2_remove_extent(struct inode *inode,
 	struct ocfs2_extent_list *el;
 	struct ocfs2_path *path = NULL;
 
-	ocfs2_extent_map_trunc(inode, 0);
+	/*
+	 * XXX: Why are we truncating to 0 instead of wherever this
+	 * affects us?
+	 */
+	ocfs2_et_extent_map_truncate(et, 0);
 
 	path = ocfs2_new_path_from_et(et);
 	if (!path) {
@@ -5425,7 +5555,7 @@ int ocfs2_remove_extent(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_find_path(inode, path, cpos);
+	ret = ocfs2_find_path(et->et_ci, path, cpos);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -5434,10 +5564,11 @@ int ocfs2_remove_extent(struct inode *inode,
 	el = path_leaf_el(path);
 	index = ocfs2_search_extent_list(el, cpos);
 	if (index == -1 || index >= le16_to_cpu(el->l_next_free_rec)) {
-		ocfs2_error(inode->i_sb,
-			    "Inode %llu has an extent at cpos %u which can no "
+		ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+			    "Owner %llu has an extent at cpos %u which can no "
 			    "longer be found.\n",
-			    (unsigned long long)OCFS2_I(inode)->ip_blkno, cpos);
+			    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+			    cpos);
 		ret = -EROFS;
 		goto out;
 	}
@@ -5464,20 +5595,21 @@ int ocfs2_remove_extent(struct inode *inode,
 
 	BUG_ON(cpos < le32_to_cpu(rec->e_cpos) || trunc_range > rec_range);
 
-	mlog(0, "Inode %llu, remove (cpos %u, len %u). Existing index %d "
+	mlog(0, "Owner %llu, remove (cpos %u, len %u). Existing index %d "
 	     "(cpos %u, len %u)\n",
-	     (unsigned long long)OCFS2_I(inode)->ip_blkno, cpos, len, index,
+	     (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+	     cpos, len, index,
 	     le32_to_cpu(rec->e_cpos), ocfs2_rec_clusters(el, rec));
 
 	if (le32_to_cpu(rec->e_cpos) == cpos || rec_range == trunc_range) {
-		ret = ocfs2_truncate_rec(inode, handle, path, index, dealloc,
-					 cpos, len, et);
+		ret = ocfs2_truncate_rec(handle, et, path, index, dealloc,
+					 cpos, len);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 	} else {
-		ret = ocfs2_split_tree(inode, et, handle, path, index,
+		ret = ocfs2_split_tree(handle, et, path, index,
 				       trunc_range, meta_ac);
 		if (ret) {
 			mlog_errno(ret);
@@ -5490,7 +5622,7 @@ int ocfs2_remove_extent(struct inode *inode,
 		 */
 		ocfs2_reinit_path(path, 1);
 
-		ret = ocfs2_find_path(inode, path, cpos);
+		ret = ocfs2_find_path(et->et_ci, path, cpos);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -5499,9 +5631,9 @@ int ocfs2_remove_extent(struct inode *inode,
 		el = path_leaf_el(path);
 		index = ocfs2_search_extent_list(el, cpos);
 		if (index == -1 || index >= le16_to_cpu(el->l_next_free_rec)) {
-			ocfs2_error(inode->i_sb,
-				    "Inode %llu: split at cpos %u lost record.",
-				    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+			ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+				    "Owner %llu: split at cpos %u lost record.",
+				    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 				    cpos);
 			ret = -EROFS;
 			goto out;
@@ -5515,18 +5647,18 @@ int ocfs2_remove_extent(struct inode *inode,
 		rec_range = le32_to_cpu(rec->e_cpos) +
 			ocfs2_rec_clusters(el, rec);
 		if (rec_range != trunc_range) {
-			ocfs2_error(inode->i_sb,
-				    "Inode %llu: error after split at cpos %u"
+			ocfs2_error(ocfs2_metadata_cache_get_super(et->et_ci),
+				    "Owner %llu: error after split at cpos %u"
 				    "trunc len %u, existing record is (%u,%u)",
-				    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+				    (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
 				    cpos, len, le32_to_cpu(rec->e_cpos),
 				    ocfs2_rec_clusters(el, rec));
 			ret = -EROFS;
 			goto out;
 		}
 
-		ret = ocfs2_truncate_rec(inode, handle, path, index, dealloc,
-					 cpos, len, et);
+		ret = ocfs2_truncate_rec(handle, et, path, index, dealloc,
+					 cpos, len);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -5573,7 +5705,7 @@ int ocfs2_remove_btree_range(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_et_root_journal_access(handle, inode, et,
+	ret = ocfs2_et_root_journal_access(handle, et,
 					   OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -5583,14 +5715,13 @@ int ocfs2_remove_btree_range(struct inode *inode,
 	vfs_dq_free_space_nodirty(inode,
 				  ocfs2_clusters_to_bytes(inode->i_sb, len));
 
-	ret = ocfs2_remove_extent(inode, et, cpos, len, handle, meta_ac,
-				  dealloc);
+	ret = ocfs2_remove_extent(handle, et, cpos, len, meta_ac, dealloc);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_commit;
 	}
 
-	ocfs2_et_update_clusters(inode, et, -len);
+	ocfs2_et_update_clusters(et, -len);
 
 	ret = ocfs2_journal_dirty(handle, et->et_root_bh);
 	if (ret) {
@@ -5690,7 +5821,7 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_journal_access_di(handle, tl_inode, tl_bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(tl_inode), tl_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -5752,7 +5883,7 @@ 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_di(handle, tl_inode, tl_bh,
+		status = ocfs2_journal_access_di(handle, INODE_CACHE(tl_inode), tl_bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -6010,7 +6141,7 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
 		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);
+		status = ocfs2_write_block(osb, tl_bh, INODE_CACHE(tl_inode));
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -6400,9 +6531,9 @@ ocfs2_find_per_slot_free_list(int type,
 	return fl;
 }
 
-static int ocfs2_cache_block_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
-				     int type, int slot, u64 blkno,
-				     unsigned int bit)
+int ocfs2_cache_block_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
+			      int type, int slot, u64 blkno,
+			      unsigned int bit)
 {
 	int ret;
 	struct ocfs2_per_slot_free_list *fl;
@@ -6518,7 +6649,7 @@ static int ocfs2_find_new_last_ext_blk(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_find_leaf(inode, path_root_el(path), cpos, &bh);
+	ret = ocfs2_find_leaf(INODE_CACHE(inode), path_root_el(path), cpos, &bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -6551,7 +6682,7 @@ out:
  */
 static int ocfs2_trim_tree(struct inode *inode, struct ocfs2_path *path,
 			   handle_t *handle, struct ocfs2_truncate_context *tc,
-			   u32 clusters_to_del, u64 *delete_start)
+			   u32 clusters_to_del, u64 *delete_start, u8 *flags)
 {
 	int ret, i, index = path->p_tree_depth;
 	u32 new_edge = 0;
@@ -6561,6 +6692,7 @@ static int ocfs2_trim_tree(struct inode *inode, struct ocfs2_path *path,
 	struct ocfs2_extent_rec *rec;
 
 	*delete_start = 0;
+	*flags = 0;
 
 	while (index >= 0) {
 		bh = path->p_node[index].bh;
@@ -6648,6 +6780,7 @@ find_tail_record:
 			*delete_start = le64_to_cpu(rec->e_blkno)
 				+ ocfs2_clusters_to_blocks(inode->i_sb,
 					le16_to_cpu(rec->e_leaf_clusters));
+			*flags = rec->e_flags;
 
 			/*
 			 * If it's now empty, remove this record.
@@ -6719,7 +6852,7 @@ delete:
 
 			mlog(0, "deleting this extent block.\n");
 
-			ocfs2_remove_from_cache(inode, bh);
+			ocfs2_remove_from_cache(INODE_CACHE(inode), bh);
 
 			BUG_ON(ocfs2_rec_clusters(el, &el->l_recs[0]));
 			BUG_ON(le32_to_cpu(el->l_recs[0].e_cpos));
@@ -6747,7 +6880,8 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 			     struct buffer_head *fe_bh,
 			     handle_t *handle,
 			     struct ocfs2_truncate_context *tc,
-			     struct ocfs2_path *path)
+			     struct ocfs2_path *path,
+			     struct ocfs2_alloc_context *meta_ac)
 {
 	int status;
 	struct ocfs2_dinode *fe;
@@ -6755,6 +6889,7 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 	struct ocfs2_extent_list *el;
 	struct buffer_head *last_eb_bh = NULL;
 	u64 delete_blk = 0;
+	u8 rec_flags;
 
 	fe = (struct ocfs2_dinode *) fe_bh->b_data;
 
@@ -6769,14 +6904,14 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 	 * Each component will be touched, so we might as well journal
 	 * here to avoid having to handle errors later.
 	 */
-	status = ocfs2_journal_access_path(inode, handle, path);
+	status = ocfs2_journal_access_path(INODE_CACHE(inode), handle, path);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
 	}
 
 	if (last_eb_bh) {
-		status = ocfs2_journal_access_eb(handle, inode, last_eb_bh,
+		status = ocfs2_journal_access_eb(handle, INODE_CACHE(inode), last_eb_bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -6810,7 +6945,7 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 	inode->i_blocks = ocfs2_inode_sector_count(inode);
 
 	status = ocfs2_trim_tree(inode, path, handle, tc,
-				 clusters_to_del, &delete_blk);
+				 clusters_to_del, &delete_blk, &rec_flags);
 	if (status) {
 		mlog_errno(status);
 		goto bail;
@@ -6842,8 +6977,16 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 	}
 
 	if (delete_blk) {
-		status = ocfs2_truncate_log_append(osb, handle, delete_blk,
-						   clusters_to_del);
+		if (rec_flags & OCFS2_EXT_REFCOUNTED)
+			status = ocfs2_decrease_refcount(inode, handle,
+					ocfs2_blocks_to_clusters(osb->sb,
+								 delete_blk),
+					clusters_to_del, meta_ac,
+					&tc->tc_dealloc, 1);
+		else
+			status = ocfs2_truncate_log_append(osb, handle,
+							   delete_blk,
+							   clusters_to_del);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
@@ -6863,9 +7006,9 @@ static int ocfs2_zero_func(handle_t *handle, struct buffer_head *bh)
 	return 0;
 }
 
-static void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
-				     unsigned int from, unsigned int to,
-				     struct page *page, int zero, u64 *phys)
+void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
+			      unsigned int from, unsigned int to,
+			      struct page *page, int zero, u64 *phys)
 {
 	int ret, partial = 0;
 
@@ -6933,20 +7076,16 @@ out:
 		ocfs2_unlock_and_free_pages(pages, numpages);
 }
 
-static int ocfs2_grab_eof_pages(struct inode *inode, loff_t start, loff_t end,
-				struct page **pages, int *num)
+int ocfs2_grab_pages(struct inode *inode, loff_t start, loff_t end,
+		     struct page **pages, int *num)
 {
 	int numpages, ret = 0;
-	struct super_block *sb = inode->i_sb;
 	struct address_space *mapping = inode->i_mapping;
 	unsigned long index;
 	loff_t last_page_bytes;
 
 	BUG_ON(start > end);
 
-	BUG_ON(start >> OCFS2_SB(sb)->s_clustersize_bits !=
-	       (end - 1) >> OCFS2_SB(sb)->s_clustersize_bits);
-
 	numpages = 0;
 	last_page_bytes = PAGE_ALIGN(end);
 	index = start >> PAGE_CACHE_SHIFT;
@@ -6974,6 +7113,17 @@ out:
 	return ret;
 }
 
+static int ocfs2_grab_eof_pages(struct inode *inode, loff_t start, loff_t end,
+				struct page **pages, int *num)
+{
+	struct super_block *sb = inode->i_sb;
+
+	BUG_ON(start >> OCFS2_SB(sb)->s_clustersize_bits !=
+	       (end - 1) >> OCFS2_SB(sb)->s_clustersize_bits);
+
+	return ocfs2_grab_pages(inode, start, end, pages, num);
+}
+
 /*
  * Zero the area past i_size but still within an allocated
  * cluster. This avoids exposing nonzero data on subsequent file
@@ -7138,7 +7288,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 		goto out_unlock;
 	}
 
-	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -7218,9 +7368,8 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode,
 		 * this proves to be false, we could always re-build
 		 * the in-inode data from our pages.
 		 */
-		ocfs2_init_dinode_extent_tree(&et, inode, di_bh);
-		ret = ocfs2_insert_extent(osb, handle, inode, &et,
-					  0, block, 1, 0, NULL);
+		ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), di_bh);
+		ret = ocfs2_insert_extent(handle, &et, 0, block, 1, 0, NULL);
 		if (ret) {
 			mlog_errno(ret);
 			goto out_commit;
@@ -7262,11 +7411,14 @@ int ocfs2_commit_truncate(struct ocfs2_super *osb,
 {
 	int status, i, credits, tl_sem = 0;
 	u32 clusters_to_del, new_highest_cpos, range;
+	u64 blkno = 0;
 	struct ocfs2_extent_list *el;
 	handle_t *handle = NULL;
 	struct inode *tl_inode = osb->osb_tl_inode;
 	struct ocfs2_path *path = NULL;
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)fe_bh->b_data;
+	struct ocfs2_alloc_context *meta_ac = NULL;
+	struct ocfs2_refcount_tree *ref_tree = NULL;
 
 	mlog_entry_void();
 
@@ -7292,10 +7444,12 @@ start:
 		goto bail;
 	}
 
+	credits = 0;
+
 	/*
 	 * Truncate always works against the rightmost tree branch.
 	 */
-	status = ocfs2_find_path(inode, path, UINT_MAX);
+	status = ocfs2_find_path(INODE_CACHE(inode), path, UINT_MAX);
 	if (status) {
 		mlog_errno(status);
 		goto bail;
@@ -7332,10 +7486,15 @@ start:
 		clusters_to_del = 0;
 	} else if (le32_to_cpu(el->l_recs[i].e_cpos) >= new_highest_cpos) {
 		clusters_to_del = ocfs2_rec_clusters(el, &el->l_recs[i]);
+		blkno = le64_to_cpu(el->l_recs[i].e_blkno);
 	} else if (range > new_highest_cpos) {
 		clusters_to_del = (ocfs2_rec_clusters(el, &el->l_recs[i]) +
 				   le32_to_cpu(el->l_recs[i].e_cpos)) -
 				  new_highest_cpos;
+		blkno = le64_to_cpu(el->l_recs[i].e_blkno) +
+			ocfs2_clusters_to_blocks(inode->i_sb,
+				ocfs2_rec_clusters(el, &el->l_recs[i]) -
+				clusters_to_del);
 	} else {
 		status = 0;
 		goto bail;
@@ -7344,6 +7503,29 @@ start:
 	mlog(0, "clusters_to_del = %u in this pass, tail blk=%llu\n",
 	     clusters_to_del, (unsigned long long)path_leaf_bh(path)->b_blocknr);
 
+	if (el->l_recs[i].e_flags & OCFS2_EXT_REFCOUNTED && clusters_to_del) {
+		BUG_ON(!(OCFS2_I(inode)->ip_dyn_features &
+			 OCFS2_HAS_REFCOUNT_FL));
+
+		status = ocfs2_lock_refcount_tree(osb,
+						le64_to_cpu(di->i_refcount_loc),
+						1, &ref_tree, NULL);
+		if (status) {
+			mlog_errno(status);
+			goto bail;
+		}
+
+		status = ocfs2_prepare_refcount_change_for_del(inode, fe_bh,
+							       blkno,
+							       clusters_to_del,
+							       &credits,
+							       &meta_ac);
+		if (status < 0) {
+			mlog_errno(status);
+			goto bail;
+		}
+	}
+
 	mutex_lock(&tl_inode->i_mutex);
 	tl_sem = 1;
 	/* ocfs2_truncate_log_needs_flush guarantees us at least one
@@ -7357,7 +7539,7 @@ start:
 		}
 	}
 
-	credits = ocfs2_calc_tree_trunc_credits(osb->sb, clusters_to_del,
+	credits += ocfs2_calc_tree_trunc_credits(osb->sb, clusters_to_del,
 						(struct ocfs2_dinode *)fe_bh->b_data,
 						el);
 	handle = ocfs2_start_trans(osb, credits);
@@ -7369,7 +7551,7 @@ start:
 	}
 
 	status = ocfs2_do_truncate(osb, clusters_to_del, inode, fe_bh, handle,
-				   tc, path);
+				   tc, path, meta_ac);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -7383,6 +7565,16 @@ start:
 
 	ocfs2_reinit_path(path, 1);
 
+	if (meta_ac) {
+		ocfs2_free_alloc_context(meta_ac);
+		meta_ac = NULL;
+	}
+
+	if (ref_tree) {
+		ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+		ref_tree = NULL;
+	}
+
 	/*
 	 * The check above will catch the case where we've truncated
 	 * away all allocation.
@@ -7399,6 +7591,12 @@ bail:
 	if (handle)
 		ocfs2_commit_trans(osb, handle);
 
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+
+	if (ref_tree)
+		ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+
 	ocfs2_run_deallocs(osb, &tc->tc_dealloc);
 
 	ocfs2_free_path(path);
@@ -7445,7 +7643,7 @@ 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_extent_block(inode,
+		status = ocfs2_read_extent_block(INODE_CACHE(inode),
 						 le64_to_cpu(fe->i_last_eb_blk),
 						 &last_eb_bh);
 		if (status < 0) {
@@ -7507,7 +7705,7 @@ int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h
index 353254ba29e..9c122d57446 100644
--- a/fs/ocfs2/alloc.h
+++ b/fs/ocfs2/alloc.h
@@ -45,7 +45,8 @@
  *
  * 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.  With metadata ecc, we now call different journal_access
+ * functions.  It needs the ocfs2_caching_info structure associated with
+ * I/O on the tree.  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
@@ -56,6 +57,7 @@ struct ocfs2_extent_tree {
 	struct ocfs2_extent_tree_operations	*et_ops;
 	struct buffer_head			*et_root_bh;
 	struct ocfs2_extent_list		*et_root_el;
+	struct ocfs2_caching_info		*et_ci;
 	ocfs2_journal_access_func		et_root_journal_access;
 	void					*et_object;
 	unsigned int				et_max_leaf_clusters;
@@ -66,31 +68,32 @@ struct ocfs2_extent_tree {
  * specified object buffer.
  */
 void ocfs2_init_dinode_extent_tree(struct ocfs2_extent_tree *et,
-				   struct inode *inode,
+				   struct ocfs2_caching_info *ci,
 				   struct buffer_head *bh);
 void ocfs2_init_xattr_tree_extent_tree(struct ocfs2_extent_tree *et,
-				       struct inode *inode,
+				       struct ocfs2_caching_info *ci,
 				       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 ocfs2_caching_info *ci,
 					struct ocfs2_xattr_value_buf *vb);
 void ocfs2_init_dx_root_extent_tree(struct ocfs2_extent_tree *et,
-				    struct inode *inode,
+				    struct ocfs2_caching_info *ci,
 				    struct buffer_head *bh);
+void ocfs2_init_refcount_extent_tree(struct ocfs2_extent_tree *et,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *bh);
 
 /*
  * 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,
+int ocfs2_read_extent_block(struct ocfs2_caching_info *ci, u64 eb_blkno,
 			    struct buffer_head **bh);
 
 struct ocfs2_alloc_context;
-int ocfs2_insert_extent(struct ocfs2_super *osb,
-			handle_t *handle,
-			struct inode *inode,
+int ocfs2_insert_extent(handle_t *handle,
 			struct ocfs2_extent_tree *et,
 			u32 cpos,
 			u64 start_blk,
@@ -103,25 +106,36 @@ enum ocfs2_alloc_restarted {
 	RESTART_TRANS,
 	RESTART_META
 };
-int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
-				struct inode *inode,
+int ocfs2_add_clusters_in_btree(handle_t *handle,
+				struct ocfs2_extent_tree *et,
 				u32 *logical_offset,
 				u32 clusters_to_add,
 				int mark_unwritten,
-				struct ocfs2_extent_tree *et,
-				handle_t *handle,
 				struct ocfs2_alloc_context *data_ac,
 				struct ocfs2_alloc_context *meta_ac,
 				enum ocfs2_alloc_restarted *reason_ret);
 struct ocfs2_cached_dealloc_ctxt;
+struct ocfs2_path;
+int ocfs2_split_extent(handle_t *handle,
+		       struct ocfs2_extent_tree *et,
+		       struct ocfs2_path *path,
+		       int split_index,
+		       struct ocfs2_extent_rec *split_rec,
+		       struct ocfs2_alloc_context *meta_ac,
+		       struct ocfs2_cached_dealloc_ctxt *dealloc);
 int ocfs2_mark_extent_written(struct inode *inode,
 			      struct ocfs2_extent_tree *et,
 			      handle_t *handle, u32 cpos, u32 len, u32 phys,
 			      struct ocfs2_alloc_context *meta_ac,
 			      struct ocfs2_cached_dealloc_ctxt *dealloc);
-int ocfs2_remove_extent(struct inode *inode,
-			struct ocfs2_extent_tree *et,
-			u32 cpos, u32 len, handle_t *handle,
+int ocfs2_change_extent_flag(handle_t *handle,
+			     struct ocfs2_extent_tree *et,
+			     u32 cpos, u32 len, u32 phys,
+			     struct ocfs2_alloc_context *meta_ac,
+			     struct ocfs2_cached_dealloc_ctxt *dealloc,
+			     int new_flags, int clear_flags);
+int ocfs2_remove_extent(handle_t *handle, struct ocfs2_extent_tree *et,
+			u32 cpos, u32 len,
 			struct ocfs2_alloc_context *meta_ac,
 			struct ocfs2_cached_dealloc_ctxt *dealloc);
 int ocfs2_remove_btree_range(struct inode *inode,
@@ -130,7 +144,6 @@ int ocfs2_remove_btree_range(struct inode *inode,
 			     struct ocfs2_cached_dealloc_ctxt *dealloc);
 
 int ocfs2_num_free_extents(struct ocfs2_super *osb,
-			   struct inode *inode,
 			   struct ocfs2_extent_tree *et);
 
 /*
@@ -195,6 +208,9 @@ static inline void ocfs2_init_dealloc_ctxt(struct ocfs2_cached_dealloc_ctxt *c)
 }
 int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
 				u64 blkno, unsigned int bit);
+int ocfs2_cache_block_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
+			      int type, int slot, u64 blkno,
+			      unsigned int bit);
 static inline int ocfs2_dealloc_has_cluster(struct ocfs2_cached_dealloc_ctxt *c)
 {
 	return c->c_global_allocator != NULL;
@@ -222,8 +238,9 @@ int ocfs2_commit_truncate(struct ocfs2_super *osb,
 int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh,
 			  unsigned int start, unsigned int end, int trunc);
 
-int ocfs2_find_leaf(struct inode *inode, struct ocfs2_extent_list *root_el,
-		    u32 cpos, struct buffer_head **leaf_bh);
+int ocfs2_find_leaf(struct ocfs2_caching_info *ci,
+		    struct ocfs2_extent_list *root_el, u32 cpos,
+		    struct buffer_head **leaf_bh);
 int ocfs2_search_extent_list(struct ocfs2_extent_list *el, u32 v_cluster);
 
 /*
@@ -254,4 +271,50 @@ static inline int ocfs2_is_empty_extent(struct ocfs2_extent_rec *rec)
 	return !rec->e_leaf_clusters;
 }
 
+int ocfs2_grab_pages(struct inode *inode, loff_t start, loff_t end,
+		     struct page **pages, int *num);
+void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
+			      unsigned int from, unsigned int to,
+			      struct page *page, int zero, u64 *phys);
+/*
+ * Structures which describe a path through a btree, and functions to
+ * manipulate them.
+ *
+ * The idea here is to be as generic as possible with the tree
+ * manipulation code.
+ */
+struct ocfs2_path_item {
+	struct buffer_head		*bh;
+	struct ocfs2_extent_list	*el;
+};
+
+#define OCFS2_MAX_PATH_DEPTH	5
+
+struct ocfs2_path {
+	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)
+
+void ocfs2_reinit_path(struct ocfs2_path *path, int keep_root);
+void ocfs2_free_path(struct ocfs2_path *path);
+int ocfs2_find_path(struct ocfs2_caching_info *ci,
+		    struct ocfs2_path *path,
+		    u32 cpos);
+struct ocfs2_path *ocfs2_new_path_from_path(struct ocfs2_path *path);
+struct ocfs2_path *ocfs2_new_path_from_et(struct ocfs2_extent_tree *et);
+int ocfs2_path_bh_journal_access(handle_t *handle,
+				 struct ocfs2_caching_info *ci,
+				 struct ocfs2_path *path,
+				 int idx);
+int ocfs2_journal_access_path(struct ocfs2_caching_info *ci,
+			      handle_t *handle,
+			      struct ocfs2_path *path);
 #endif /* OCFS2_ALLOC_H */
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 8a1e61545f4..72e76062a90 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -44,6 +44,7 @@
 #include "suballoc.h"
 #include "super.h"
 #include "symlink.h"
+#include "refcounttree.h"
 
 #include "buffer_head_io.h"
 
@@ -126,8 +127,8 @@ bail:
 	return err;
 }
 
-static int ocfs2_get_block(struct inode *inode, sector_t iblock,
-			   struct buffer_head *bh_result, int create)
+int ocfs2_get_block(struct inode *inode, sector_t iblock,
+		    struct buffer_head *bh_result, int create)
 {
 	int err = 0;
 	unsigned int ext_flags;
@@ -590,6 +591,8 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
 		goto bail;
 	}
 
+	/* We should already CoW the refcounted extent. */
+	BUG_ON(ext_flags & OCFS2_EXT_REFCOUNTED);
 	/*
 	 * get_more_blocks() expects us to describe a hole by clearing
 	 * the mapped bit on bh_result().
@@ -687,6 +690,10 @@ static ssize_t ocfs2_direct_IO(int rw,
 	if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
 		return 0;
 
+	/* Fallback to buffered I/O if we are appending. */
+	if (i_size_read(inode) <= offset)
+		return 0;
+
 	ret = blockdev_direct_IO_no_locking(rw, iocb, inode,
 					    inode->i_sb->s_bdev, iov, offset,
 					    nr_segs, 
@@ -1259,7 +1266,8 @@ static int ocfs2_write_cluster(struct address_space *mapping,
 			goto out;
 		}
 	} else if (unwritten) {
-		ocfs2_init_dinode_extent_tree(&et, inode, wc->w_di_bh);
+		ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode),
+					      wc->w_di_bh);
 		ret = ocfs2_mark_extent_written(inode, &et,
 						wc->w_handle, cpos, 1, phys,
 						meta_ac, &wc->w_dealloc);
@@ -1448,6 +1456,9 @@ static int ocfs2_populate_write_desc(struct inode *inode,
 				goto out;
 			}
 
+			/* We should already CoW the refcountd extent. */
+			BUG_ON(ext_flags & OCFS2_EXT_REFCOUNTED);
+
 			/*
 			 * Assume worst case - that we're writing in
 			 * the middle of the extent.
@@ -1528,7 +1539,7 @@ static int ocfs2_write_begin_inline(struct address_space *mapping,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_di(handle, inode, wc->w_di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), wc->w_di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		ocfs2_commit_trans(osb, handle);
@@ -1699,6 +1710,19 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 		goto out;
 	}
 
+	ret = ocfs2_check_range_for_refcount(inode, pos, len);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto out;
+	} else if (ret == 1) {
+		ret = ocfs2_refcount_cow(inode, di_bh,
+					 wc->w_cpos, wc->w_clen, UINT_MAX);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+	}
+
 	ret = ocfs2_populate_write_desc(inode, wc, &clusters_to_alloc,
 					&extents_to_split);
 	if (ret) {
@@ -1726,7 +1750,8 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 		     (long long)i_size_read(inode), le32_to_cpu(di->i_clusters),
 		     clusters_to_alloc, extents_to_split);
 
-		ocfs2_init_dinode_extent_tree(&et, inode, wc->w_di_bh);
+		ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode),
+					      wc->w_di_bh);
 		ret = ocfs2_lock_allocators(inode, &et,
 					    clusters_to_alloc, extents_to_split,
 					    &data_ac, &meta_ac);
@@ -1773,7 +1798,7 @@ int ocfs2_write_begin_nolock(struct address_space *mapping,
 	 * We don't want this to fail in ocfs2_write_end(), so do it
 	 * here.
 	 */
-	ret = ocfs2_journal_access_di(handle, inode, wc->w_di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), wc->w_di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h
index 503e49232e1..c48e93ffc51 100644
--- a/fs/ocfs2/aops.h
+++ b/fs/ocfs2/aops.h
@@ -57,6 +57,8 @@ int ocfs2_read_inline_data(struct inode *inode, struct page *page,
 			   struct buffer_head *di_bh);
 int ocfs2_size_fits_inline_data(struct buffer_head *di_bh, u64 new_size);
 
+int ocfs2_get_block(struct inode *inode, sector_t iblock,
+		    struct buffer_head *bh_result, int create);
 /* all ocfs2_dio_end_io()'s fault */
 #define ocfs2_iocb_is_rw_locked(iocb) \
 	test_bit(0, (unsigned long *)&iocb->private)
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index 15c8e6deee2..d43d34a1dd3 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -52,12 +52,12 @@ enum ocfs2_state_bits {
 BUFFER_FNS(NeedsValidate, needs_validate);
 
 int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
-		      struct inode *inode)
+		      struct ocfs2_caching_info *ci)
 {
 	int ret = 0;
 
-	mlog_entry("(bh->b_blocknr = %llu, inode=%p)\n",
-		   (unsigned long long)bh->b_blocknr, inode);
+	mlog_entry("(bh->b_blocknr = %llu, ci=%p)\n",
+		   (unsigned long long)bh->b_blocknr, ci);
 
 	BUG_ON(bh->b_blocknr < OCFS2_SUPER_BLOCK_BLKNO);
 	BUG_ON(buffer_jbd(bh));
@@ -70,7 +70,7 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
 		goto out;
 	}
 
-	mutex_lock(&OCFS2_I(inode)->ip_io_mutex);
+	ocfs2_metadata_cache_io_lock(ci);
 
 	lock_buffer(bh);
 	set_buffer_uptodate(bh);
@@ -85,7 +85,7 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
 	wait_on_buffer(bh);
 
 	if (buffer_uptodate(bh)) {
-		ocfs2_set_buffer_uptodate(inode, bh);
+		ocfs2_set_buffer_uptodate(ci, bh);
 	} else {
 		/* We don't need to remove the clustered uptodate
 		 * information for this bh as it's not marked locally
@@ -94,7 +94,7 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
 		put_bh(bh);
 	}
 
-	mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
+	ocfs2_metadata_cache_io_unlock(ci);
 out:
 	mlog_exit(ret);
 	return ret;
@@ -177,7 +177,7 @@ bail:
 	return status;
 }
 
-int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
+int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
 		      struct buffer_head *bhs[], int flags,
 		      int (*validate)(struct super_block *sb,
 				      struct buffer_head *bh))
@@ -185,11 +185,12 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 	int status = 0;
 	int i, ignore_cache = 0;
 	struct buffer_head *bh;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
 
-	mlog_entry("(inode=%p, block=(%llu), nr=(%d), flags=%d)\n",
-		   inode, (unsigned long long)block, nr, flags);
+	mlog_entry("(ci=%p, block=(%llu), nr=(%d), flags=%d)\n",
+		   ci, (unsigned long long)block, nr, flags);
 
-	BUG_ON(!inode);
+	BUG_ON(!ci);
 	BUG_ON((flags & OCFS2_BH_READAHEAD) &&
 	       (flags & OCFS2_BH_IGNORE_CACHE));
 
@@ -212,12 +213,12 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 		goto bail;
 	}
 
-	mutex_lock(&OCFS2_I(inode)->ip_io_mutex);
+	ocfs2_metadata_cache_io_lock(ci);
 	for (i = 0 ; i < nr ; i++) {
 		if (bhs[i] == NULL) {
-			bhs[i] = sb_getblk(inode->i_sb, block++);
+			bhs[i] = sb_getblk(sb, block++);
 			if (bhs[i] == NULL) {
-				mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
+				ocfs2_metadata_cache_io_unlock(ci);
 				status = -EIO;
 				mlog_errno(status);
 				goto bail;
@@ -250,11 +251,11 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 		 *    before our is-it-in-flight check.
 		 */
 
-		if (!ignore_cache && !ocfs2_buffer_uptodate(inode, bh)) {
+		if (!ignore_cache && !ocfs2_buffer_uptodate(ci, bh)) {
 			mlog(ML_UPTODATE,
-			     "bh (%llu), inode %llu not uptodate\n",
+			     "bh (%llu), owner %llu not uptodate\n",
 			     (unsigned long long)bh->b_blocknr,
-			     (unsigned long long)OCFS2_I(inode)->ip_blkno);
+			     (unsigned long long)ocfs2_metadata_cache_owner(ci));
 			/* We're using ignore_cache here to say
 			 * "go to disk" */
 			ignore_cache = 1;
@@ -283,7 +284,7 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 			 * previously submitted request than we are
 			 * done here. */
 			if ((flags & OCFS2_BH_READAHEAD)
-			    && ocfs2_buffer_read_ahead(inode, bh))
+			    && ocfs2_buffer_read_ahead(ci, bh))
 				continue;
 
 			lock_buffer(bh);
@@ -305,7 +306,7 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 			 * buffer lock. */
 			if (!(flags & OCFS2_BH_IGNORE_CACHE)
 			    && !(flags & OCFS2_BH_READAHEAD)
-			    && ocfs2_buffer_uptodate(inode, bh)) {
+			    && ocfs2_buffer_uptodate(ci, bh)) {
 				unlock_buffer(bh);
 				continue;
 			}
@@ -327,7 +328,7 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 
 		if (!(flags & OCFS2_BH_READAHEAD)) {
 			/* We know this can't have changed as we hold the
-			 * inode sem. Avoid doing any work on the bh if the
+			 * owner sem. Avoid doing any work on the bh if the
 			 * journal has it. */
 			if (!buffer_jbd(bh))
 				wait_on_buffer(bh);
@@ -351,7 +352,7 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 				 * that better not have changed */
 				BUG_ON(buffer_jbd(bh));
 				clear_buffer_needs_validate(bh);
-				status = validate(inode->i_sb, bh);
+				status = validate(sb, bh);
 				if (status) {
 					put_bh(bh);
 					bhs[i] = NULL;
@@ -363,9 +364,9 @@ int ocfs2_read_blocks(struct inode *inode, u64 block, int nr,
 		/* Always set the buffer in the cache, even if it was
 		 * a forced read, or read-ahead which hasn't yet
 		 * completed. */
-		ocfs2_set_buffer_uptodate(inode, bh);
+		ocfs2_set_buffer_uptodate(ci, bh);
 	}
-	mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
+	ocfs2_metadata_cache_io_unlock(ci);
 
 	mlog(ML_BH_IO, "block=(%llu), nr=(%d), cached=%s, flags=0x%x\n", 
 	     (unsigned long long)block, nr,
@@ -399,7 +400,7 @@ static void ocfs2_check_super_or_backup(struct super_block *sb,
 
 /*
  * Write super block and backups doesn't need to collaborate with journal,
- * so we don't need to lock ip_io_mutex and inode doesn't need to bea passed
+ * so we don't need to lock ip_io_mutex and ci doesn't need to bea passed
  * into this function.
  */
 int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
diff --git a/fs/ocfs2/buffer_head_io.h b/fs/ocfs2/buffer_head_io.h
index c75d682dadd..b97bcc6dde7 100644
--- a/fs/ocfs2/buffer_head_io.h
+++ b/fs/ocfs2/buffer_head_io.h
@@ -33,7 +33,7 @@ void ocfs2_end_buffer_io_sync(struct buffer_head *bh,
 
 int ocfs2_write_block(struct ocfs2_super          *osb,
 		      struct buffer_head  *bh,
-		      struct inode        *inode);
+		      struct ocfs2_caching_info   *ci);
 int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block,
 			   unsigned int nr, struct buffer_head *bhs[]);
 
@@ -44,7 +44,7 @@ int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block,
  * 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,
+int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
 		      struct buffer_head *bhs[], int flags,
 		      int (*validate)(struct super_block *sb,
 				      struct buffer_head *bh));
@@ -55,7 +55,7 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
 #define OCFS2_BH_IGNORE_CACHE      1
 #define OCFS2_BH_READAHEAD         8
 
-static inline int ocfs2_read_block(struct inode *inode, u64 off,
+static inline int ocfs2_read_block(struct ocfs2_caching_info *ci, u64 off,
 				   struct buffer_head **bh,
 				   int (*validate)(struct super_block *sb,
 						   struct buffer_head *bh))
@@ -68,7 +68,7 @@ static inline int ocfs2_read_block(struct inode *inode, u64 off,
 		goto bail;
 	}
 
-	status = ocfs2_read_blocks(inode, off, 1, bh, 0, validate);
+	status = ocfs2_read_blocks(ci, off, 1, bh, 0, validate);
 
 bail:
 	return status;
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 96df5416993..1cd2934de61 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -111,6 +111,7 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
 	define_mask(EXPORT),
 	define_mask(XATTR),
 	define_mask(QUOTA),
+	define_mask(REFCOUNT),
 	define_mask(ERROR),
 	define_mask(NOTICE),
 	define_mask(KTHREAD),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 696c32e5071..9b4d11726cf 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -113,6 +113,7 @@
 #define ML_EXPORT	0x0000000010000000ULL /* ocfs2 export operations */
 #define ML_XATTR	0x0000000020000000ULL /* ocfs2 extended attributes */
 #define ML_QUOTA	0x0000000040000000ULL /* ocfs2 quota operations */
+#define ML_REFCOUNT	0x0000000080000000ULL /* refcount tree 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/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c
index f8424874fa0..cfb2be708ab 100644
--- a/fs/ocfs2/cluster/netdebug.c
+++ b/fs/ocfs2/cluster/netdebug.c
@@ -163,7 +163,7 @@ static void nst_seq_stop(struct seq_file *seq, void *v)
 {
 }
 
-static struct seq_operations nst_seq_ops = {
+static const struct seq_operations nst_seq_ops = {
 	.start = nst_seq_start,
 	.next = nst_seq_next,
 	.stop = nst_seq_stop,
@@ -344,7 +344,7 @@ static void sc_seq_stop(struct seq_file *seq, void *v)
 {
 }
 
-static struct seq_operations sc_seq_ops = {
+static const struct seq_operations sc_seq_ops = {
 	.start = sc_seq_start,
 	.next = sc_seq_next,
 	.stop = sc_seq_stop,
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index b358f3bf896..28c3ec23879 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -176,7 +176,7 @@ static int ocfs2_dx_dir_link_trailer(struct inode *dir, handle_t *handle,
 	struct ocfs2_dx_root_block *dx_root;
 	struct ocfs2_dir_block_trailer *trailer;
 
-	ret = ocfs2_journal_access_dr(handle, dir, dx_root_bh,
+	ret = ocfs2_journal_access_dr(handle, INODE_CACHE(dir), dx_root_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -564,7 +564,8 @@ static int ocfs2_read_dir_block_direct(struct inode *dir, u64 phys,
 	int ret;
 	struct buffer_head *tmp = *bh;
 
-	ret = ocfs2_read_block(dir, phys, &tmp, ocfs2_validate_dir_block);
+	ret = ocfs2_read_block(INODE_CACHE(dir), phys, &tmp,
+			       ocfs2_validate_dir_block);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -622,7 +623,8 @@ static int ocfs2_read_dx_root(struct inode *dir, struct ocfs2_dinode *di,
 	u64 blkno = le64_to_cpu(di->i_dx_root);
 	struct buffer_head *tmp = *dx_root_bh;
 
-	ret = ocfs2_read_block(dir, blkno, &tmp, ocfs2_validate_dx_root);
+	ret = ocfs2_read_block(INODE_CACHE(dir), blkno, &tmp,
+			       ocfs2_validate_dx_root);
 
 	/* If ocfs2_read_block() got us a new bh, pass it up. */
 	if (!ret && !*dx_root_bh)
@@ -662,7 +664,8 @@ static int ocfs2_read_dx_leaf(struct inode *dir, u64 blkno,
 	int ret;
 	struct buffer_head *tmp = *dx_leaf_bh;
 
-	ret = ocfs2_read_block(dir, blkno, &tmp, ocfs2_validate_dx_leaf);
+	ret = ocfs2_read_block(INODE_CACHE(dir), blkno, &tmp,
+			       ocfs2_validate_dx_leaf);
 
 	/* If ocfs2_read_block() got us a new bh, pass it up. */
 	if (!ret && !*dx_leaf_bh)
@@ -680,7 +683,7 @@ static int ocfs2_read_dx_leaves(struct inode *dir, u64 start, int num,
 {
 	int ret;
 
-	ret = ocfs2_read_blocks(dir, start, num, dx_leaf_bhs, 0,
+	ret = ocfs2_read_blocks(INODE_CACHE(dir), start, num, dx_leaf_bhs, 0,
 				ocfs2_validate_dx_leaf);
 	if (ret)
 		mlog_errno(ret);
@@ -802,7 +805,8 @@ static int ocfs2_dx_dir_lookup_rec(struct inode *inode,
 	struct ocfs2_extent_rec *rec = NULL;
 
 	if (el->l_tree_depth) {
-		ret = ocfs2_find_leaf(inode, el, major_hash, &eb_bh);
+		ret = ocfs2_find_leaf(INODE_CACHE(inode), el, major_hash,
+				      &eb_bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -1133,7 +1137,8 @@ int ocfs2_update_entry(struct inode *dir, handle_t *handle,
 	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);
+	ret = access(handle, INODE_CACHE(dir), de_bh,
+		     OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -1176,7 +1181,7 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
 			goto bail;
 		}
 		if (de == de_del)  {
-			status = access(handle, dir, bh,
+			status = access(handle, INODE_CACHE(dir), bh,
 					OCFS2_JOURNAL_ACCESS_WRITE);
 			if (status < 0) {
 				status = -EIO;
@@ -1326,7 +1331,7 @@ static int ocfs2_delete_entry_dx(handle_t *handle, struct inode *dir,
 	 * the entry count needs to be updated. Also, we might be
 	 * adding to the start of the free list.
 	 */
-	ret = ocfs2_journal_access_dr(handle, dir, dx_root_bh,
+	ret = ocfs2_journal_access_dr(handle, INODE_CACHE(dir), dx_root_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1334,7 +1339,7 @@ static int ocfs2_delete_entry_dx(handle_t *handle, struct inode *dir,
 	}
 
 	if (!ocfs2_dx_root_inline(dx_root)) {
-		ret = ocfs2_journal_access_dl(handle, dir,
+		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
 					      lookup->dl_dx_leaf_bh,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
@@ -1493,7 +1498,7 @@ static int __ocfs2_dx_dir_leaf_insert(struct inode *dir, handle_t *handle,
 	int ret;
 	struct ocfs2_dx_leaf *dx_leaf;
 
-	ret = ocfs2_journal_access_dl(handle, dir, dx_leaf_bh,
+	ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir), dx_leaf_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1523,7 +1528,7 @@ static int ocfs2_dx_dir_insert(struct inode *dir, handle_t *handle,
 	struct ocfs2_dx_root_block *dx_root;
 	struct buffer_head *dx_root_bh = lookup->dl_dx_root_bh;
 
-	ret = ocfs2_journal_access_dr(handle, dir, dx_root_bh,
+	ret = ocfs2_journal_access_dr(handle, INODE_CACHE(dir), dx_root_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1645,11 +1650,13 @@ int __ocfs2_add_entry(handle_t *handle,
 		 */
 		if (ocfs2_free_list_at_root(lookup)) {
 			bh = lookup->dl_dx_root_bh;
-			retval = ocfs2_journal_access_dr(handle, dir, bh,
+			retval = ocfs2_journal_access_dr(handle,
+						 INODE_CACHE(dir), bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		} else {
 			bh = lookup->dl_prev_leaf_bh;
-			retval = ocfs2_journal_access_db(handle, dir, bh,
+			retval = ocfs2_journal_access_db(handle,
+						 INODE_CACHE(dir), bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		}
 		if (retval) {
@@ -1700,11 +1707,13 @@ int __ocfs2_add_entry(handle_t *handle,
 			}
 
 			if (insert_bh == parent_fe_bh)
-				status = ocfs2_journal_access_di(handle, dir,
+				status = ocfs2_journal_access_di(handle,
+								 INODE_CACHE(dir),
 								 insert_bh,
 								 OCFS2_JOURNAL_ACCESS_WRITE);
 			else {
-				status = ocfs2_journal_access_db(handle, dir,
+				status = ocfs2_journal_access_db(handle,
+								 INODE_CACHE(dir),
 								 insert_bh,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 
@@ -2280,7 +2289,7 @@ 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_di(handle, inode, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -2332,9 +2341,9 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	ocfs2_set_new_buffer_uptodate(inode, new_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), new_bh);
 
-	status = ocfs2_journal_access_db(handle, inode, new_bh,
+	status = ocfs2_journal_access_db(handle, INODE_CACHE(inode), new_bh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -2418,9 +2427,9 @@ static int ocfs2_dx_dir_attach_index(struct ocfs2_super *osb,
 		ret = -EIO;
 		goto out;
 	}
-	ocfs2_set_new_buffer_uptodate(dir, dx_root_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(dir), dx_root_bh);
 
-	ret = ocfs2_journal_access_dr(handle, dir, dx_root_bh,
+	ret = ocfs2_journal_access_dr(handle, INODE_CACHE(dir), dx_root_bh,
 				      OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -2454,7 +2463,7 @@ static int ocfs2_dx_dir_attach_index(struct ocfs2_super *osb,
 	if (ret)
 		mlog_errno(ret);
 
-	ret = ocfs2_journal_access_di(handle, dir, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(dir), di_bh,
 				      OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret) {
 		mlog_errno(ret);
@@ -2495,9 +2504,9 @@ static int ocfs2_dx_dir_format_cluster(struct ocfs2_super *osb,
 		}
 		dx_leaves[i] = bh;
 
-		ocfs2_set_new_buffer_uptodate(dir, bh);
+		ocfs2_set_new_buffer_uptodate(INODE_CACHE(dir), bh);
 
-		ret = ocfs2_journal_access_dl(handle, dir, bh,
+		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir), bh,
 					      OCFS2_JOURNAL_ACCESS_CREATE);
 		if (ret < 0) {
 			mlog_errno(ret);
@@ -2582,7 +2591,6 @@ static int ocfs2_dx_dir_new_cluster(struct inode *dir,
 {
 	int ret;
 	u64 phys_blkno;
-	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
 
 	ret = __ocfs2_dx_dir_new_cluster(dir, cpos, handle, data_ac, dx_leaves,
 					 num_dx_leaves, &phys_blkno);
@@ -2591,7 +2599,7 @@ static int ocfs2_dx_dir_new_cluster(struct inode *dir,
 		goto out;
 	}
 
-	ret = ocfs2_insert_extent(osb, handle, dir, et, cpos, phys_blkno, 1, 0,
+	ret = ocfs2_insert_extent(handle, et, cpos, phys_blkno, 1, 0,
 				  meta_ac);
 	if (ret)
 		mlog_errno(ret);
@@ -2895,7 +2903,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	struct ocfs2_extent_tree dx_et;
 	int did_quota = 0, bytes_allocated = 0;
 
-	ocfs2_init_dinode_extent_tree(&et, dir, di_bh);
+	ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(dir), di_bh);
 
 	alloc = ocfs2_clusters_for_bytes(sb, bytes);
 	dx_alloc = 0;
@@ -3005,9 +3013,9 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 		goto out_commit;
 	}
 
-	ocfs2_set_new_buffer_uptodate(dir, dirdata_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(dir), dirdata_bh);
 
-	ret = ocfs2_journal_access_db(handle, dir, dirdata_bh,
+	ret = ocfs2_journal_access_db(handle, INODE_CACHE(dir), dirdata_bh,
 				      OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret) {
 		mlog_errno(ret);
@@ -3060,7 +3068,7 @@ 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_di(handle, dir, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(dir), di_bh,
 				      OCFS2_JOURNAL_ACCESS_CREATE);
 	if (ret) {
 		mlog_errno(ret);
@@ -3085,7 +3093,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 	 * This should never fail as our extent list is empty and all
 	 * related blocks have been journaled already.
 	 */
-	ret = ocfs2_insert_extent(osb, handle, dir, &et, 0, blkno, len,
+	ret = ocfs2_insert_extent(handle, &et, 0, blkno, len,
 				  0, NULL);
 	if (ret) {
 		mlog_errno(ret);
@@ -3117,8 +3125,10 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 			ocfs2_dx_dir_index_root_block(dir, dx_root_bh,
 						      dirdata_bh);
 		} else {
-			ocfs2_init_dx_root_extent_tree(&dx_et, dir, dx_root_bh);
-			ret = ocfs2_insert_extent(osb, handle, dir, &dx_et, 0,
+			ocfs2_init_dx_root_extent_tree(&dx_et,
+						       INODE_CACHE(dir),
+						       dx_root_bh);
+			ret = ocfs2_insert_extent(handle, &dx_et, 0,
 						  dx_insert_blkno, 1, 0, NULL);
 			if (ret)
 				mlog_errno(ret);
@@ -3138,7 +3148,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh,
 		}
 		blkno = ocfs2_clusters_to_blocks(dir->i_sb, bit_off);
 
-		ret = ocfs2_insert_extent(osb, handle, dir, &et, 1,
+		ret = ocfs2_insert_extent(handle, &et, 1,
 					  blkno, len, 0, NULL);
 		if (ret) {
 			mlog_errno(ret);
@@ -3337,8 +3347,9 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb,
 	spin_lock(&OCFS2_I(dir)->ip_lock);
 	if (dir_i_size == ocfs2_clusters_to_bytes(sb, OCFS2_I(dir)->ip_clusters)) {
 		spin_unlock(&OCFS2_I(dir)->ip_lock);
-		ocfs2_init_dinode_extent_tree(&et, dir, parent_fe_bh);
-		num_free_extents = ocfs2_num_free_extents(osb, dir, &et);
+		ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(dir),
+					      parent_fe_bh);
+		num_free_extents = ocfs2_num_free_extents(osb, &et);
 		if (num_free_extents < 0) {
 			status = num_free_extents;
 			mlog_errno(status);
@@ -3387,9 +3398,9 @@ do_extend:
 		goto bail;
 	}
 
-	ocfs2_set_new_buffer_uptodate(dir, new_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(dir), new_bh);
 
-	status = ocfs2_journal_access_db(handle, dir, new_bh,
+	status = ocfs2_journal_access_db(handle, INODE_CACHE(dir), new_bh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -3829,7 +3840,7 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 	     (unsigned long long)OCFS2_I(dir)->ip_blkno,
 	     (unsigned long long)leaf_blkno, insert_hash);
 
-	ocfs2_init_dx_root_extent_tree(&et, dir, dx_root_bh);
+	ocfs2_init_dx_root_extent_tree(&et, INODE_CACHE(dir), dx_root_bh);
 
 	dx_root = (struct ocfs2_dx_root_block *)dx_root_bh->b_data;
 	/*
@@ -3885,7 +3896,7 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 	}
 	did_quota = 1;
 
-	ret = ocfs2_journal_access_dl(handle, dir, dx_leaf_bh,
+	ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir), dx_leaf_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -3949,7 +3960,8 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
 	}
 
 	for (i = 0; i < num_dx_leaves; i++) {
-		ret = ocfs2_journal_access_dl(handle, dir, orig_dx_leaves[i],
+		ret = ocfs2_journal_access_dl(handle, INODE_CACHE(dir),
+					      orig_dx_leaves[i],
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
@@ -4165,7 +4177,7 @@ static int ocfs2_expand_inline_dx_root(struct inode *dir,
 	 * failure to add the dx_root_bh to the journal won't result
 	 * us losing clusters.
 	 */
-	ret = ocfs2_journal_access_dr(handle, dir, dx_root_bh,
+	ret = ocfs2_journal_access_dr(handle, INODE_CACHE(dir), dx_root_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -4207,9 +4219,8 @@ static int ocfs2_expand_inline_dx_root(struct inode *dir,
 
 	/* This should never fail considering we start with an empty
 	 * dx_root. */
-	ocfs2_init_dx_root_extent_tree(&et, dir, dx_root_bh);
-	ret = ocfs2_insert_extent(osb, handle, dir, &et, 0,
-				  insert_blkno, 1, 0, NULL);
+	ocfs2_init_dx_root_extent_tree(&et, INODE_CACHE(dir), dx_root_bh);
+	ret = ocfs2_insert_extent(handle, &et, 0, insert_blkno, 1, 0, NULL);
 	if (ret)
 		mlog_errno(ret);
 	did_quota = 0;
@@ -4469,7 +4480,7 @@ static int ocfs2_dx_dir_remove_index(struct inode *dir,
 		goto out_unlock;
 	}
 
-	ret = ocfs2_journal_access_di(handle, dir, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(dir), di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -4532,7 +4543,7 @@ int ocfs2_dx_dir_truncate(struct inode *dir, struct buffer_head *di_bh)
 	if (ocfs2_dx_root_inline(dx_root))
 		goto remove_index;
 
-	ocfs2_init_dx_root_extent_tree(&et, dir, dx_root_bh);
+	ocfs2_init_dx_root_extent_tree(&et, INODE_CACHE(dir), dx_root_bh);
 
 	/* XXX: What if dr_clusters is too large? */
 	while (le32_to_cpu(dx_root->dr_clusters)) {
@@ -4565,7 +4576,7 @@ remove_index:
 		goto out;
 	}
 
-	ocfs2_remove_from_cache(dir, dx_root_bh);
+	ocfs2_remove_from_cache(INODE_CACHE(dir), dx_root_bh);
 out:
 	ocfs2_schedule_truncate_log_flush(osb, 1);
 	ocfs2_run_deallocs(osb, &dealloc);
diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c
index df52f706f66..c5c88124096 100644
--- a/fs/ocfs2/dlm/dlmdebug.c
+++ b/fs/ocfs2/dlm/dlmdebug.c
@@ -683,7 +683,7 @@ static int lockres_seq_show(struct seq_file *s, void *v)
 	return 0;
 }
 
-static struct seq_operations debug_lockres_ops = {
+static const struct seq_operations debug_lockres_ops = {
 	.start =	lockres_seq_start,
 	.stop =		lockres_seq_stop,
 	.next =		lockres_seq_next,
diff --git a/fs/ocfs2/dlm/dlmfs.c b/fs/ocfs2/dlm/dlmfs.c
index 1c9efb406a9..02bf17808bd 100644
--- a/fs/ocfs2/dlm/dlmfs.c
+++ b/fs/ocfs2/dlm/dlmfs.c
@@ -325,6 +325,7 @@ clear_fields:
 }
 
 static struct backing_dev_info dlmfs_backing_dev_info = {
+	.name		= "ocfs2-dlmfs",
 	.ra_pages	= 0,	/* No readahead */
 	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK,
 };
diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c
index d490b66ad9d..98569e86c61 100644
--- a/fs/ocfs2/dlm/dlmthread.c
+++ b/fs/ocfs2/dlm/dlmthread.c
@@ -212,14 +212,18 @@ static int dlm_purge_lockres(struct dlm_ctxt *dlm,
 		spin_lock(&dlm->spinlock);
 	}
 
+	spin_lock(&res->spinlock);
 	if (!list_empty(&res->purge)) {
 		mlog(0, "removing lockres %.*s:%p from purgelist, "
 		     "master = %d\n", res->lockname.len, res->lockname.name,
 		     res, master);
 		list_del_init(&res->purge);
+		spin_unlock(&res->spinlock);
 		dlm_lockres_put(res);
 		dlm->purge_count--;
-	}
+	} else
+		spin_unlock(&res->spinlock);
+
 	__dlm_unhash_lockres(res);
 
 	/* lockres is not in the hash now.  drop the flag and wake up
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 110bb57c46a..0d38d67194c 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -53,6 +53,7 @@
 #include "super.h"
 #include "uptodate.h"
 #include "quota.h"
+#include "refcounttree.h"
 
 #include "buffer_head_io.h"
 
@@ -110,6 +111,11 @@ static void ocfs2_dentry_post_unlock(struct ocfs2_super *osb,
 
 static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres);
 
+static int ocfs2_check_refcount_downconvert(struct ocfs2_lock_res *lockres,
+					    int new_level);
+static int ocfs2_refcount_convert_worker(struct ocfs2_lock_res *lockres,
+					 int blocking);
+
 #define mlog_meta_lvb(__level, __lockres) ocfs2_dump_meta_lvb_info(__level, __PRETTY_FUNCTION__, __LINE__, __lockres)
 
 /* This aids in debugging situations where a bad LVB might be involved. */
@@ -278,6 +284,12 @@ static struct ocfs2_lock_res_ops ocfs2_qinfo_lops = {
 	.flags		= LOCK_TYPE_REQUIRES_REFRESH | LOCK_TYPE_USES_LVB,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_refcount_block_lops = {
+	.check_downconvert = ocfs2_check_refcount_downconvert,
+	.downconvert_worker = ocfs2_refcount_convert_worker,
+	.flags		= 0,
+};
+
 static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres)
 {
 	return lockres->l_type == OCFS2_LOCK_TYPE_META ||
@@ -306,6 +318,12 @@ static inline struct ocfs2_mem_dqinfo *ocfs2_lock_res_qinfo(struct ocfs2_lock_re
 	return (struct ocfs2_mem_dqinfo *)lockres->l_priv;
 }
 
+static inline struct ocfs2_refcount_tree *
+ocfs2_lock_res_refcount_tree(struct ocfs2_lock_res *res)
+{
+	return container_of(res, struct ocfs2_refcount_tree, rf_lockres);
+}
+
 static inline struct ocfs2_super *ocfs2_get_lockres_osb(struct ocfs2_lock_res *lockres)
 {
 	if (lockres->l_ops->get_osb)
@@ -693,6 +711,17 @@ void ocfs2_qinfo_lock_res_init(struct ocfs2_lock_res *lockres,
 				   info);
 }
 
+void ocfs2_refcount_lock_res_init(struct ocfs2_lock_res *lockres,
+				  struct ocfs2_super *osb, u64 ref_blkno,
+				  unsigned int generation)
+{
+	ocfs2_lock_res_init_once(lockres);
+	ocfs2_build_lock_name(OCFS2_LOCK_TYPE_REFCOUNT, ref_blkno,
+			      generation, lockres->l_name);
+	ocfs2_lock_res_init_common(osb, lockres, OCFS2_LOCK_TYPE_REFCOUNT,
+				   &ocfs2_refcount_block_lops, osb);
+}
+
 void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
 {
 	mlog_entry_void();
@@ -1548,8 +1577,10 @@ int ocfs2_rw_lock(struct inode *inode, int write)
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
-	if (ocfs2_mount_local(osb))
+	if (ocfs2_mount_local(osb)) {
+		mlog_exit(0);
 		return 0;
+	}
 
 	lockres = &OCFS2_I(inode)->ip_rw_lockres;
 
@@ -2127,7 +2158,7 @@ static int ocfs2_inode_lock_update(struct inode *inode,
 
 	/* This will discard any caching information we might have had
 	 * for the inode metadata. */
-	ocfs2_metadata_cache_purge(inode);
+	ocfs2_metadata_cache_purge(INODE_CACHE(inode));
 
 	ocfs2_extent_map_trunc(inode, 0);
 
@@ -3009,6 +3040,7 @@ static void ocfs2_unlock_ast(void *opaque, int error)
 		     "unlock_action %d\n", error, lockres->l_name,
 		     lockres->l_unlock_action);
 		spin_unlock_irqrestore(&lockres->l_lock, flags);
+		mlog_exit_void();
 		return;
 	}
 
@@ -3495,11 +3527,11 @@ out:
 	return UNBLOCK_CONTINUE;
 }
 
-static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
-					int new_level)
+static int ocfs2_ci_checkpointed(struct ocfs2_caching_info *ci,
+				 struct ocfs2_lock_res *lockres,
+				 int new_level)
 {
-	struct inode *inode = ocfs2_lock_res_inode(lockres);
-	int checkpointed = ocfs2_inode_fully_checkpointed(inode);
+	int checkpointed = ocfs2_ci_fully_checkpointed(ci);
 
 	BUG_ON(new_level != DLM_LOCK_NL && new_level != DLM_LOCK_PR);
 	BUG_ON(lockres->l_level != DLM_LOCK_EX && !checkpointed);
@@ -3507,10 +3539,18 @@ static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
 	if (checkpointed)
 		return 1;
 
-	ocfs2_start_checkpoint(OCFS2_SB(inode->i_sb));
+	ocfs2_start_checkpoint(OCFS2_SB(ocfs2_metadata_cache_get_super(ci)));
 	return 0;
 }
 
+static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
+					int new_level)
+{
+	struct inode *inode = ocfs2_lock_res_inode(lockres);
+
+	return ocfs2_ci_checkpointed(INODE_CACHE(inode), lockres, new_level);
+}
+
 static void ocfs2_set_meta_lvb(struct ocfs2_lock_res *lockres)
 {
 	struct inode *inode = ocfs2_lock_res_inode(lockres);
@@ -3640,6 +3680,26 @@ static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
 	return UNBLOCK_CONTINUE_POST;
 }
 
+static int ocfs2_check_refcount_downconvert(struct ocfs2_lock_res *lockres,
+					    int new_level)
+{
+	struct ocfs2_refcount_tree *tree =
+				ocfs2_lock_res_refcount_tree(lockres);
+
+	return ocfs2_ci_checkpointed(&tree->rf_ci, lockres, new_level);
+}
+
+static int ocfs2_refcount_convert_worker(struct ocfs2_lock_res *lockres,
+					 int blocking)
+{
+	struct ocfs2_refcount_tree *tree =
+				ocfs2_lock_res_refcount_tree(lockres);
+
+	ocfs2_metadata_cache_purge(&tree->rf_ci);
+
+	return UNBLOCK_CONTINUE;
+}
+
 static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres)
 {
 	struct ocfs2_qinfo_lvb *lvb;
@@ -3752,6 +3812,37 @@ bail:
 	return status;
 }
 
+int ocfs2_refcount_lock(struct ocfs2_refcount_tree *ref_tree, int ex)
+{
+	int status;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	struct ocfs2_lock_res *lockres = &ref_tree->rf_lockres;
+	struct ocfs2_super *osb = lockres->l_priv;
+
+
+	if (ocfs2_is_hard_readonly(osb))
+		return -EROFS;
+
+	if (ocfs2_mount_local(osb))
+		return 0;
+
+	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
+	if (status < 0)
+		mlog_errno(status);
+
+	return status;
+}
+
+void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex)
+{
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	struct ocfs2_lock_res *lockres = &ref_tree->rf_lockres;
+	struct ocfs2_super *osb = lockres->l_priv;
+
+	if (!ocfs2_mount_local(osb))
+		ocfs2_cluster_unlock(osb, lockres, level);
+}
+
 /*
  * 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 7553836931d..d1ce48e1b3d 100644
--- a/fs/ocfs2/dlmglue.h
+++ b/fs/ocfs2/dlmglue.h
@@ -101,6 +101,9 @@ void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 struct ocfs2_mem_dqinfo;
 void ocfs2_qinfo_lock_res_init(struct ocfs2_lock_res *lockres,
                                struct ocfs2_mem_dqinfo *info);
+void ocfs2_refcount_lock_res_init(struct ocfs2_lock_res *lockres,
+				  struct ocfs2_super *osb, u64 ref_blkno,
+				  unsigned int generation);
 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);
@@ -148,6 +151,9 @@ 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);
+struct ocfs2_refcount_tree;
+int ocfs2_refcount_lock(struct ocfs2_refcount_tree *ref_tree, int ex);
+void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex);
 
 
 void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres);
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index f2bb1a04d25..843db64e9d4 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_extent_block(inode, last_eb_blk, &eb_bh);
+	ret = ocfs2_read_extent_block(INODE_CACHE(inode), last_eb_blk, &eb_bh);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -353,11 +353,11 @@ static int ocfs2_search_for_hole_index(struct ocfs2_extent_list *el,
  * eb_bh is NULL. Otherwise, eb_bh should point to the extent block
  * containing el.
  */
-static int ocfs2_figure_hole_clusters(struct inode *inode,
-				      struct ocfs2_extent_list *el,
-				      struct buffer_head *eb_bh,
-				      u32 v_cluster,
-				      u32 *num_clusters)
+int ocfs2_figure_hole_clusters(struct ocfs2_caching_info *ci,
+			       struct ocfs2_extent_list *el,
+			       struct buffer_head *eb_bh,
+			       u32 v_cluster,
+			       u32 *num_clusters)
 {
 	int ret, i;
 	struct buffer_head *next_eb_bh = NULL;
@@ -375,7 +375,7 @@ 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_extent_block(inode,
+		ret = ocfs2_read_extent_block(ci,
 					      le64_to_cpu(eb->h_next_leaf_blk),
 					      &next_eb_bh);
 		if (ret) {
@@ -428,7 +428,8 @@ static int ocfs2_get_clusters_nocache(struct inode *inode,
 	tree_height = le16_to_cpu(el->l_tree_depth);
 
 	if (tree_height > 0) {
-		ret = ocfs2_find_leaf(inode, el, v_cluster, &eb_bh);
+		ret = ocfs2_find_leaf(INODE_CACHE(inode), el, v_cluster,
+				      &eb_bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -455,7 +456,8 @@ static int ocfs2_get_clusters_nocache(struct inode *inode,
 		 * field.
 		 */
 		if (hole_len) {
-			ret = ocfs2_figure_hole_clusters(inode, el, eb_bh,
+			ret = ocfs2_figure_hole_clusters(INODE_CACHE(inode),
+							 el, eb_bh,
 							 v_cluster, &len);
 			if (ret) {
 				mlog_errno(ret);
@@ -539,7 +541,8 @@ static void ocfs2_relative_extent_offsets(struct super_block *sb,
 
 int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster,
 			     u32 *p_cluster, u32 *num_clusters,
-			     struct ocfs2_extent_list *el)
+			     struct ocfs2_extent_list *el,
+			     unsigned int *extent_flags)
 {
 	int ret = 0, i;
 	struct buffer_head *eb_bh = NULL;
@@ -548,7 +551,8 @@ int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster,
 	u32 coff;
 
 	if (el->l_tree_depth) {
-		ret = ocfs2_find_leaf(inode, el, v_cluster, &eb_bh);
+		ret = ocfs2_find_leaf(INODE_CACHE(inode), el, v_cluster,
+				      &eb_bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -590,6 +594,9 @@ int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster,
 		*p_cluster = *p_cluster + coff;
 		if (num_clusters)
 			*num_clusters = ocfs2_rec_clusters(el, rec) - coff;
+
+		if (extent_flags)
+			*extent_flags = rec->e_flags;
 	}
 out:
 	if (eb_bh)
@@ -862,8 +869,8 @@ int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr,
 			BUG_ON(bhs[done + i]->b_blocknr != (p_block + i));
 		}
 
-		rc = ocfs2_read_blocks(inode, p_block, count, bhs + done,
-				       flags, validate);
+		rc = ocfs2_read_blocks(INODE_CACHE(inode), p_block, count,
+				       bhs + done, flags, validate);
 		if (rc) {
 			mlog_errno(rc);
 			break;
diff --git a/fs/ocfs2/extent_map.h b/fs/ocfs2/extent_map.h
index b7dd9731b46..e79d41c2c90 100644
--- a/fs/ocfs2/extent_map.h
+++ b/fs/ocfs2/extent_map.h
@@ -55,12 +55,18 @@ int ocfs2_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 
 int ocfs2_xattr_get_clusters(struct inode *inode, u32 v_cluster,
 			     u32 *p_cluster, u32 *num_clusters,
-			     struct ocfs2_extent_list *el);
+			     struct ocfs2_extent_list *el,
+			     unsigned int *extent_flags);
 
 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 ocfs2_figure_hole_clusters(struct ocfs2_caching_info *ci,
+			       struct ocfs2_extent_list *el,
+			       struct buffer_head *eb_bh,
+			       u32 v_cluster,
+			       u32 *num_clusters);
 static inline int ocfs2_read_virt_block(struct inode *inode, u64 v_block,
 					struct buffer_head **bh,
 					int (*validate)(struct super_block *sb,
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index aa501d3f93f..89fc8ee1f5a 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -59,6 +59,7 @@
 #include "xattr.h"
 #include "acl.h"
 #include "quota.h"
+#include "refcounttree.h"
 
 #include "buffer_head_io.h"
 
@@ -259,7 +260,7 @@ int ocfs2_update_inode_atime(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_di(handle, inode, bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -334,6 +335,39 @@ out:
 	return ret;
 }
 
+static int ocfs2_cow_file_pos(struct inode *inode,
+			      struct buffer_head *fe_bh,
+			      u64 offset)
+{
+	int status;
+	u32 phys, cpos = offset >> OCFS2_SB(inode->i_sb)->s_clustersize_bits;
+	unsigned int num_clusters = 0;
+	unsigned int ext_flags = 0;
+
+	/*
+	 * If the new offset is aligned to the range of the cluster, there is
+	 * no space for ocfs2_zero_range_for_truncate to fill, so no need to
+	 * CoW either.
+	 */
+	if ((offset & (OCFS2_SB(inode->i_sb)->s_clustersize - 1)) == 0)
+		return 0;
+
+	status = ocfs2_get_clusters(inode, cpos, &phys,
+				    &num_clusters, &ext_flags);
+	if (status) {
+		mlog_errno(status);
+		goto out;
+	}
+
+	if (!(ext_flags & OCFS2_EXT_REFCOUNTED))
+		goto out;
+
+	return ocfs2_refcount_cow(inode, fe_bh, cpos, 1, cpos+1);
+
+out:
+	return status;
+}
+
 static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb,
 				     struct inode *inode,
 				     struct buffer_head *fe_bh,
@@ -346,6 +380,17 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb,
 
 	mlog_entry_void();
 
+	/*
+	 * We need to CoW the cluster contains the offset if it is reflinked
+	 * since we will call ocfs2_zero_range_for_truncate later which will
+	 * write "0" from offset to the end of the cluster.
+	 */
+	status = ocfs2_cow_file_pos(inode, fe_bh, new_i_size);
+	if (status) {
+		mlog_errno(status);
+		return status;
+	}
+
 	/* TODO: This needs to actually orphan the inode in this
 	 * transaction. */
 
@@ -356,7 +401,7 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb,
 		goto out;
 	}
 
-	status = ocfs2_journal_access_di(handle, inode, fe_bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), fe_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -486,6 +531,8 @@ bail_unlock_sem:
 	up_write(&OCFS2_I(inode)->ip_alloc_sem);
 
 bail:
+	if (!status && OCFS2_I(inode)->ip_clusters == 0)
+		status = ocfs2_try_remove_refcount_tree(inode, di_bh);
 
 	mlog_exit(status);
 	return status;
@@ -515,11 +562,10 @@ int ocfs2_add_inode_data(struct ocfs2_super *osb,
 	int ret;
 	struct ocfs2_extent_tree et;
 
-	ocfs2_init_dinode_extent_tree(&et, inode, fe_bh);
-	ret = ocfs2_add_clusters_in_btree(osb, inode, logical_offset,
-					   clusters_to_add, mark_unwritten,
-					   &et, handle,
-					   data_ac, meta_ac, reason_ret);
+	ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), fe_bh);
+	ret = ocfs2_add_clusters_in_btree(handle, &et, logical_offset,
+					  clusters_to_add, mark_unwritten,
+					  data_ac, meta_ac, reason_ret);
 
 	return ret;
 }
@@ -564,7 +610,7 @@ restart_all:
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     (long long)i_size_read(inode), le32_to_cpu(fe->i_clusters),
 	     clusters_to_add);
-	ocfs2_init_dinode_extent_tree(&et, inode, bh);
+	ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), bh);
 	status = ocfs2_lock_allocators(inode, &et, clusters_to_add, 0,
 				       &data_ac, &meta_ac);
 	if (status) {
@@ -593,7 +639,7 @@ restarted_transaction:
 	/* 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_di(handle, inode, bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1131,7 +1177,7 @@ static int __ocfs2_write_remove_suid(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_di(handle, inode, bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -1395,7 +1441,7 @@ static int ocfs2_remove_inode_range(struct inode *inode,
 	struct address_space *mapping = inode->i_mapping;
 	struct ocfs2_extent_tree et;
 
-	ocfs2_init_dinode_extent_tree(&et, inode, di_bh);
+	ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), di_bh);
 	ocfs2_init_dealloc_ctxt(&dealloc);
 
 	if (byte_len == 0)
@@ -1657,6 +1703,70 @@ static long ocfs2_fallocate(struct inode *inode, int mode, loff_t offset,
 					 OCFS2_IOC_RESVSP64, &sr, change_size);
 }
 
+int ocfs2_check_range_for_refcount(struct inode *inode, loff_t pos,
+				   size_t count)
+{
+	int ret = 0;
+	unsigned int extent_flags;
+	u32 cpos, clusters, extent_len, phys_cpos;
+	struct super_block *sb = inode->i_sb;
+
+	if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb)) ||
+	    !(OCFS2_I(inode)->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL))
+		return 0;
+
+	cpos = pos >> OCFS2_SB(sb)->s_clustersize_bits;
+	clusters = ocfs2_clusters_for_bytes(sb, pos + count) - cpos;
+
+	while (clusters) {
+		ret = ocfs2_get_clusters(inode, cpos, &phys_cpos, &extent_len,
+					 &extent_flags);
+		if (ret < 0) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		if (phys_cpos && (extent_flags & OCFS2_EXT_REFCOUNTED)) {
+			ret = 1;
+			break;
+		}
+
+		if (extent_len > clusters)
+			extent_len = clusters;
+
+		clusters -= extent_len;
+		cpos += extent_len;
+	}
+out:
+	return ret;
+}
+
+static int ocfs2_prepare_inode_for_refcount(struct inode *inode,
+					    loff_t pos, size_t count,
+					    int *meta_level)
+{
+	int ret;
+	struct buffer_head *di_bh = NULL;
+	u32 cpos = pos >> OCFS2_SB(inode->i_sb)->s_clustersize_bits;
+	u32 clusters =
+		ocfs2_clusters_for_bytes(inode->i_sb, pos + count) - cpos;
+
+	ret = ocfs2_inode_lock(inode, &di_bh, 1);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	*meta_level = 1;
+
+	ret = ocfs2_refcount_cow(inode, di_bh, cpos, clusters, UINT_MAX);
+	if (ret)
+		mlog_errno(ret);
+out:
+	brelse(di_bh);
+	return ret;
+}
+
 static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
 					 loff_t *ppos,
 					 size_t count,
@@ -1713,6 +1823,22 @@ static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
 
 		end = saved_pos + count;
 
+		ret = ocfs2_check_range_for_refcount(inode, saved_pos, count);
+		if (ret == 1) {
+			ocfs2_inode_unlock(inode, meta_level);
+			meta_level = -1;
+
+			ret = ocfs2_prepare_inode_for_refcount(inode,
+							       saved_pos,
+							       count,
+							       &meta_level);
+		}
+
+		if (ret < 0) {
+			mlog_errno(ret);
+			goto out_unlock;
+		}
+
 		/*
 		 * Skip the O_DIRECT checks if we don't need
 		 * them.
@@ -1759,7 +1885,8 @@ static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
 		*ppos = saved_pos;
 
 out_unlock:
-	ocfs2_inode_unlock(inode, meta_level);
+	if (meta_level >= 0)
+		ocfs2_inode_unlock(inode, meta_level);
 
 out:
 	return ret;
@@ -1871,8 +1998,7 @@ relock:
 			goto out_dio;
 		}
 	} else {
-		written = generic_file_aio_write_nolock(iocb, iov, nr_segs,
-							*ppos);
+		written = __generic_file_aio_write(iocb, iov, nr_segs, ppos);
 	}
 
 out_dio:
@@ -1880,18 +2006,21 @@ out_dio:
 	BUG_ON(ret == -EIOCBQUEUED && !(file->f_flags & O_DIRECT));
 
 	if ((file->f_flags & O_SYNC && !direct_io) || IS_SYNC(inode)) {
-		/*
-		 * The generic write paths have handled getting data
-		 * to disk, but since we don't make use of the dirty
-		 * inode list, a manual journal commit is necessary
-		 * here.
-		 */
-		if (old_size != i_size_read(inode) ||
-		    old_clusters != OCFS2_I(inode)->ip_clusters) {
+		ret = filemap_fdatawrite_range(file->f_mapping, pos,
+					       pos + count - 1);
+		if (ret < 0)
+			written = ret;
+
+		if (!ret && (old_size != i_size_read(inode) ||
+		    old_clusters != OCFS2_I(inode)->ip_clusters)) {
 			ret = jbd2_journal_force_commit(osb->journal->j_journal);
 			if (ret < 0)
 				written = ret;
 		}
+
+		if (!ret)
+			ret = filemap_fdatawait_range(file->f_mapping, pos,
+						      pos + count - 1);
 	}
 
 	/* 
@@ -1991,31 +2120,16 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe,
 
 	if (ret > 0) {
 		unsigned long nr_pages;
+		int err;
 
-		*ppos += ret;
 		nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
 
-		/*
-		 * If file or inode is SYNC and we actually wrote some data,
-		 * sync it.
-		 */
-		if (unlikely((out->f_flags & O_SYNC) || IS_SYNC(inode))) {
-			int err;
-
-			mutex_lock(&inode->i_mutex);
-			err = ocfs2_rw_lock(inode, 1);
-			if (err < 0) {
-				mlog_errno(err);
-			} else {
-				err = generic_osync_inode(inode, mapping,
-						  OSYNC_METADATA|OSYNC_DATA);
-				ocfs2_rw_unlock(inode, 1);
-			}
-			mutex_unlock(&inode->i_mutex);
+		err = generic_write_sync(out, *ppos, ret);
+		if (err)
+			ret = err;
+		else
+			*ppos += ret;
 
-			if (err)
-				ret = err;
-		}
 		balance_dirty_pages_ratelimited_nr(mapping, nr_pages);
 	}
 
diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h
index 172f9fbc9fc..d66cf4f7c70 100644
--- a/fs/ocfs2/file.h
+++ b/fs/ocfs2/file.h
@@ -69,4 +69,6 @@ int ocfs2_update_inode_atime(struct inode *inode,
 int ocfs2_change_file_space(struct file *file, unsigned int cmd,
 			    struct ocfs2_space_resv *sr);
 
+int ocfs2_check_range_for_refcount(struct inode *inode, loff_t pos,
+				   size_t count);
 #endif /* OCFS2_FILE_H */
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 4dc8890ba31..0297fb8982b 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -53,6 +53,7 @@
 #include "sysfile.h"
 #include "uptodate.h"
 #include "xattr.h"
+#include "refcounttree.h"
 
 #include "buffer_head_io.h"
 
@@ -562,7 +563,8 @@ static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
 			goto out;
 		}
 
-		status = ocfs2_journal_access_di(handle, inode, fe_bh,
+		status = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
+						 fe_bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -646,7 +648,7 @@ static int ocfs2_remove_inode(struct inode *inode,
 	}
 
 	/* set the inodes dtime */
-	status = ocfs2_journal_access_di(handle, inode, di_bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -662,7 +664,7 @@ static int ocfs2_remove_inode(struct inode *inode,
 		goto bail_commit;
 	}
 
-	ocfs2_remove_from_cache(inode, di_bh);
+	ocfs2_remove_from_cache(INODE_CACHE(inode), di_bh);
 	vfs_dq_free_inode(inode);
 
 	status = ocfs2_free_dinode(handle, inode_alloc_inode,
@@ -781,6 +783,12 @@ static int ocfs2_wipe_inode(struct inode *inode,
 		goto bail_unlock_dir;
 	}
 
+	status = ocfs2_remove_refcount_tree(inode, di_bh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto bail_unlock_dir;
+	}
+
 	status = ocfs2_remove_inode(inode, di_bh, orphan_dir_inode,
 				    orphan_dir_bh);
 	if (status < 0)
@@ -1112,13 +1120,14 @@ void ocfs2_clear_inode(struct inode *inode)
 	ocfs2_lock_res_free(&oi->ip_inode_lockres);
 	ocfs2_lock_res_free(&oi->ip_open_lockres);
 
-	ocfs2_metadata_cache_purge(inode);
+	ocfs2_metadata_cache_exit(INODE_CACHE(inode));
 
-	mlog_bug_on_msg(oi->ip_metadata_cache.ci_num_cached,
+	mlog_bug_on_msg(INODE_CACHE(inode)->ci_num_cached,
 			"Clear inode of %llu, inode has %u cache items\n",
-			(unsigned long long)oi->ip_blkno, oi->ip_metadata_cache.ci_num_cached);
+			(unsigned long long)oi->ip_blkno,
+			INODE_CACHE(inode)->ci_num_cached);
 
-	mlog_bug_on_msg(!(oi->ip_flags & OCFS2_INODE_CACHE_INLINE),
+	mlog_bug_on_msg(!(INODE_CACHE(inode)->ci_flags & OCFS2_CACHE_FL_INLINE),
 			"Clear inode of %llu, inode has a bad flag\n",
 			(unsigned long long)oi->ip_blkno);
 
@@ -1145,9 +1154,7 @@ void ocfs2_clear_inode(struct inode *inode)
 			(unsigned long long)oi->ip_blkno, oi->ip_open_count);
 
 	/* Clear all other flags. */
-	oi->ip_flags = OCFS2_INODE_CACHE_INLINE;
-	oi->ip_created_trans = 0;
-	oi->ip_last_trans = 0;
+	oi->ip_flags = 0;
 	oi->ip_dir_start_lookup = 0;
 	oi->ip_blkno = 0ULL;
 
@@ -1239,7 +1246,7 @@ 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_di(handle, inode, bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1380,8 +1387,8 @@ int ocfs2_read_inode_block_full(struct inode *inode, struct buffer_head **bh,
 	int rc;
 	struct buffer_head *tmp = *bh;
 
-	rc = ocfs2_read_blocks(inode, OCFS2_I(inode)->ip_blkno, 1, &tmp,
-			       flags, ocfs2_validate_inode_block);
+	rc = ocfs2_read_blocks(INODE_CACHE(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)
@@ -1394,3 +1401,56 @@ int ocfs2_read_inode_block(struct inode *inode, struct buffer_head **bh)
 {
 	return ocfs2_read_inode_block_full(inode, bh, 0);
 }
+
+
+static u64 ocfs2_inode_cache_owner(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
+
+	return oi->ip_blkno;
+}
+
+static struct super_block *ocfs2_inode_cache_get_super(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
+
+	return oi->vfs_inode.i_sb;
+}
+
+static void ocfs2_inode_cache_lock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
+
+	spin_lock(&oi->ip_lock);
+}
+
+static void ocfs2_inode_cache_unlock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
+
+	spin_unlock(&oi->ip_lock);
+}
+
+static void ocfs2_inode_cache_io_lock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
+
+	mutex_lock(&oi->ip_io_mutex);
+}
+
+static void ocfs2_inode_cache_io_unlock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_inode_info *oi = cache_info_to_inode(ci);
+
+	mutex_unlock(&oi->ip_io_mutex);
+}
+
+const struct ocfs2_caching_operations ocfs2_inode_caching_ops = {
+	.co_owner		= ocfs2_inode_cache_owner,
+	.co_get_super		= ocfs2_inode_cache_get_super,
+	.co_cache_lock		= ocfs2_inode_cache_lock,
+	.co_cache_unlock	= ocfs2_inode_cache_unlock,
+	.co_io_lock		= ocfs2_inode_cache_io_lock,
+	.co_io_unlock		= ocfs2_inode_cache_io_unlock,
+};
+
diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h
index ea71525aad4..ba4fe07b293 100644
--- a/fs/ocfs2/inode.h
+++ b/fs/ocfs2/inode.h
@@ -60,12 +60,6 @@ struct ocfs2_inode_info
 
 	u32				ip_dir_start_lookup;
 
-	/* next two are protected by trans_inc_lock */
-	/* which transaction were we created on? Zero if none. */
-	unsigned long			ip_created_trans;
-	/* last transaction we were a part of. */
-	unsigned long			ip_last_trans;
-
 	struct ocfs2_caching_info	ip_metadata_cache;
 
 	struct ocfs2_extent_map		ip_extent_map;
@@ -106,8 +100,6 @@ struct ocfs2_inode_info
 #define OCFS2_INODE_MAYBE_ORPHANED	0x00000020
 /* Does someone have the file open O_DIRECT */
 #define OCFS2_INODE_OPEN_DIRECT		0x00000040
-/* Indicates that the metadata cache should be used as an array. */
-#define OCFS2_INODE_CACHE_INLINE	0x00000080
 
 static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode)
 {
@@ -120,6 +112,12 @@ static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode)
 extern struct kmem_cache *ocfs2_inode_cache;
 
 extern const struct address_space_operations ocfs2_aops;
+extern const struct ocfs2_caching_operations ocfs2_inode_caching_ops;
+
+static inline struct ocfs2_caching_info *INODE_CACHE(struct inode *inode)
+{
+	return &OCFS2_I(inode)->ip_metadata_cache;
+}
 
 void ocfs2_clear_inode(struct inode *inode);
 void ocfs2_delete_inode(struct inode *inode);
@@ -172,4 +170,10 @@ 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);
+
+static inline struct ocfs2_inode_info *cache_info_to_inode(struct ocfs2_caching_info *ci)
+{
+	return container_of(ci, struct ocfs2_inode_info, ip_metadata_cache);
+}
+
 #endif /* OCFS2_INODE_H */
diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c
index 467b413bec2..31fbb061951 100644
--- a/fs/ocfs2/ioctl.c
+++ b/fs/ocfs2/ioctl.c
@@ -21,6 +21,7 @@
 #include "ocfs2_fs.h"
 #include "ioctl.h"
 #include "resize.h"
+#include "refcounttree.h"
 
 #include <linux/ext2_fs.h>
 
@@ -115,6 +116,9 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 	int status;
 	struct ocfs2_space_resv sr;
 	struct ocfs2_new_group_input input;
+	struct reflink_arguments args;
+	const char *old_path, *new_path;
+	bool preserve;
 
 	switch (cmd) {
 	case OCFS2_IOC_GETFLAGS:
@@ -160,6 +164,15 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			return -EFAULT;
 
 		return ocfs2_group_add(inode, &input);
+	case OCFS2_IOC_REFLINK:
+		if (copy_from_user(&args, (struct reflink_arguments *)arg,
+				   sizeof(args)))
+			return -EFAULT;
+		old_path = (const char *)(unsigned long)args.old_path;
+		new_path = (const char *)(unsigned long)args.new_path;
+		preserve = (args.preserve != 0);
+
+		return ocfs2_reflink_ioctl(inode, old_path, new_path, preserve);
 	default:
 		return -ENOTTY;
 	}
@@ -182,6 +195,7 @@ long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 	case OCFS2_IOC_GROUP_EXTEND:
 	case OCFS2_IOC_GROUP_ADD:
 	case OCFS2_IOC_GROUP_ADD64:
+	case OCFS2_IOC_REFLINK:
 		break;
 	default:
 		return -ENOIOCTLCMD;
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index c48b93ac6b6..54c16b66327 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -48,6 +48,7 @@
 #include "slot_map.h"
 #include "super.h"
 #include "sysfile.h"
+#include "uptodate.h"
 #include "quota.h"
 
 #include "buffer_head_io.h"
@@ -554,6 +555,14 @@ static struct ocfs2_triggers eb_triggers = {
 	.ot_offset	= offsetof(struct ocfs2_extent_block, h_check),
 };
 
+static struct ocfs2_triggers rb_triggers = {
+	.ot_triggers = {
+		.t_commit = ocfs2_commit_trigger,
+		.t_abort = ocfs2_abort_trigger,
+	},
+	.ot_offset	= offsetof(struct ocfs2_refcount_block, rf_check),
+};
+
 static struct ocfs2_triggers gd_triggers = {
 	.ot_triggers = {
 		.t_commit = ocfs2_commit_trigger,
@@ -601,14 +610,16 @@ static struct ocfs2_triggers dl_triggers = {
 };
 
 static int __ocfs2_journal_access(handle_t *handle,
-				  struct inode *inode,
+				  struct ocfs2_caching_info *ci,
 				  struct buffer_head *bh,
 				  struct ocfs2_triggers *triggers,
 				  int type)
 {
 	int status;
+	struct ocfs2_super *osb =
+		OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
 
-	BUG_ON(!inode);
+	BUG_ON(!ci || !ci->ci_ops);
 	BUG_ON(!handle);
 	BUG_ON(!bh);
 
@@ -627,15 +638,15 @@ static int __ocfs2_journal_access(handle_t *handle,
 		BUG();
 	}
 
-	/* Set the current transaction information on the inode so
+	/* Set the current transaction information on the ci so
 	 * that the locking code knows whether it can drop it's locks
-	 * on this inode or not. We're protected from the commit
+	 * on this ci or not. We're protected from the commit
 	 * thread updating the current transaction id until
 	 * ocfs2_commit_trans() because ocfs2_start_trans() took
 	 * j_trans_barrier for us. */
-	ocfs2_set_inode_lock_trans(OCFS2_SB(inode->i_sb)->journal, inode);
+	ocfs2_set_ci_lock_trans(osb->journal, ci);
 
-	mutex_lock(&OCFS2_I(inode)->ip_io_mutex);
+	ocfs2_metadata_cache_io_lock(ci);
 	switch (type) {
 	case OCFS2_JOURNAL_ACCESS_CREATE:
 	case OCFS2_JOURNAL_ACCESS_WRITE:
@@ -650,9 +661,9 @@ static 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)
+	if (!status && ocfs2_meta_ecc(osb) && triggers)
 		jbd2_journal_set_triggers(bh, &triggers->ot_triggers);
-	mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
+	ocfs2_metadata_cache_io_unlock(ci);
 
 	if (status < 0)
 		mlog(ML_ERROR, "Error %d getting %d access to buffer!\n",
@@ -662,66 +673,65 @@ static 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)
+int ocfs2_journal_access_di(handle_t *handle, struct ocfs2_caching_info *ci,
+			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &di_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &di_triggers, type);
 }
 
-int ocfs2_journal_access_eb(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_eb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &eb_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &eb_triggers, type);
 }
 
-int ocfs2_journal_access_gd(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_rb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &gd_triggers,
+	return __ocfs2_journal_access(handle, ci, bh, &rb_triggers,
 				      type);
 }
 
-int ocfs2_journal_access_db(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_gd(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &db_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &gd_triggers, type);
 }
 
-int ocfs2_journal_access_xb(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_db(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &xb_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &db_triggers, type);
 }
 
-int ocfs2_journal_access_dq(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_xb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &dq_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &xb_triggers, type);
 }
 
-int ocfs2_journal_access_dr(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_dq(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &dr_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &dq_triggers, type);
 }
 
-int ocfs2_journal_access_dl(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_dr(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, &dl_triggers,
-				      type);
+	return __ocfs2_journal_access(handle, ci, bh, &dr_triggers, type);
+}
+
+int ocfs2_journal_access_dl(handle_t *handle, struct ocfs2_caching_info *ci,
+			    struct buffer_head *bh, int type)
+{
+	return __ocfs2_journal_access(handle, ci, bh, &dl_triggers, type);
 }
 
-int ocfs2_journal_access(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access(handle_t *handle, struct ocfs2_caching_info *ci,
 			 struct buffer_head *bh, int type)
 {
-	return __ocfs2_journal_access(handle, inode, bh, NULL, type);
+	return __ocfs2_journal_access(handle, ci, bh, NULL, type);
 }
 
 int ocfs2_journal_dirty(handle_t *handle,
@@ -898,7 +908,7 @@ static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
 		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);
+	status = ocfs2_write_block(osb, bh, INODE_CACHE(journal->j_inode));
 	if (status < 0)
 		mlog_errno(status);
 
@@ -1642,7 +1652,7 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
 					ocfs2_get_recovery_generation(fe);
 
 	ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
-	status = ocfs2_write_block(osb, bh, inode);
+	status = ocfs2_write_block(osb, bh, INODE_CACHE(inode));
 	if (status < 0)
 		mlog_errno(status);
 
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index 2c3222aec62..3f74e09b0d8 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -90,56 +90,66 @@ static inline unsigned long ocfs2_inc_trans_id(struct ocfs2_journal *j)
 	return old_id;
 }
 
-static inline void ocfs2_set_inode_lock_trans(struct ocfs2_journal *journal,
-					      struct inode *inode)
+static inline void ocfs2_set_ci_lock_trans(struct ocfs2_journal *journal,
+					   struct ocfs2_caching_info *ci)
 {
 	spin_lock(&trans_inc_lock);
-	OCFS2_I(inode)->ip_last_trans = journal->j_trans_id;
+	ci->ci_last_trans = journal->j_trans_id;
 	spin_unlock(&trans_inc_lock);
 }
 
 /* Used to figure out whether it's safe to drop a metadata lock on an
- * inode. Returns true if all the inodes changes have been
+ * cached object. Returns true if all the object's changes have been
  * checkpointed to disk. You should be holding the spinlock on the
  * metadata lock while calling this to be sure that nobody can take
  * the lock and put it on another transaction. */
-static inline int ocfs2_inode_fully_checkpointed(struct inode *inode)
+static inline int ocfs2_ci_fully_checkpointed(struct ocfs2_caching_info *ci)
 {
 	int ret;
-	struct ocfs2_journal *journal = OCFS2_SB(inode->i_sb)->journal;
+	struct ocfs2_journal *journal =
+		OCFS2_SB(ocfs2_metadata_cache_get_super(ci))->journal;
 
 	spin_lock(&trans_inc_lock);
-	ret = time_after(journal->j_trans_id, OCFS2_I(inode)->ip_last_trans);
+	ret = time_after(journal->j_trans_id, ci->ci_last_trans);
 	spin_unlock(&trans_inc_lock);
 	return ret;
 }
 
-/* convenience function to check if an inode is still new (has never
- * hit disk) Will do you a favor and set created_trans = 0 when you've
- * been checkpointed.  returns '1' if the inode is still new. */
-static inline int ocfs2_inode_is_new(struct inode *inode)
+/* convenience function to check if an object backed by struct
+ * ocfs2_caching_info  is still new (has never hit disk) Will do you a
+ * favor and set created_trans = 0 when you've
+ * been checkpointed.  returns '1' if the ci is still new. */
+static inline int ocfs2_ci_is_new(struct ocfs2_caching_info *ci)
 {
 	int ret;
+	struct ocfs2_journal *journal =
+		OCFS2_SB(ocfs2_metadata_cache_get_super(ci))->journal;
 
+	spin_lock(&trans_inc_lock);
+	ret = !(time_after(journal->j_trans_id, ci->ci_created_trans));
+	if (!ret)
+		ci->ci_created_trans = 0;
+	spin_unlock(&trans_inc_lock);
+	return ret;
+}
+
+/* Wrapper for inodes so we can check system files */
+static inline int ocfs2_inode_is_new(struct inode *inode)
+{
 	/* System files are never "new" as they're written out by
 	 * mkfs. This helps us early during mount, before we have the
 	 * journal open and j_trans_id could be junk. */
 	if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE)
 		return 0;
-	spin_lock(&trans_inc_lock);
-	ret = !(time_after(OCFS2_SB(inode->i_sb)->journal->j_trans_id,
-			   OCFS2_I(inode)->ip_created_trans));
-	if (!ret)
-		OCFS2_I(inode)->ip_created_trans = 0;
-	spin_unlock(&trans_inc_lock);
-	return ret;
+
+	return ocfs2_ci_is_new(INODE_CACHE(inode));
 }
 
-static inline void ocfs2_inode_set_new(struct ocfs2_super *osb,
-				       struct inode *inode)
+static inline void ocfs2_ci_set_new(struct ocfs2_super *osb,
+				    struct ocfs2_caching_info *ci)
 {
 	spin_lock(&trans_inc_lock);
-	OCFS2_I(inode)->ip_created_trans = osb->journal->j_trans_id;
+	ci->ci_created_trans = osb->journal->j_trans_id;
 	spin_unlock(&trans_inc_lock);
 }
 
@@ -200,7 +210,7 @@ static inline void ocfs2_checkpoint_inode(struct inode *inode)
 	if (ocfs2_mount_local(osb))
 		return;
 
-	if (!ocfs2_inode_fully_checkpointed(inode)) {
+	if (!ocfs2_ci_fully_checkpointed(INODE_CACHE(inode))) {
 		/* WARNING: This only kicks off a single
 		 * checkpoint. If someone races you and adds more
 		 * metadata to the journal, you won't know, and will
@@ -210,7 +220,7 @@ static inline void ocfs2_checkpoint_inode(struct inode *inode)
 		ocfs2_start_checkpoint(osb);
 
 		wait_event(osb->journal->j_checkpointed,
-			   ocfs2_inode_fully_checkpointed(inode));
+			   ocfs2_ci_fully_checkpointed(INODE_CACHE(inode)));
 	}
 }
 
@@ -266,31 +276,34 @@ int			     ocfs2_extend_trans(handle_t *handle, int nblocks);
 
 
 /* ocfs2_inode */
-int ocfs2_journal_access_di(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_di(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* ocfs2_extent_block */
-int ocfs2_journal_access_eb(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_eb(handle_t *handle, struct ocfs2_caching_info *ci,
+			    struct buffer_head *bh, int type);
+/* ocfs2_refcount_block */
+int ocfs2_journal_access_rb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* ocfs2_group_desc */
-int ocfs2_journal_access_gd(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_gd(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* ocfs2_xattr_block */
-int ocfs2_journal_access_xb(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_xb(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* quota blocks */
-int ocfs2_journal_access_dq(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_dq(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* dirblock */
-int ocfs2_journal_access_db(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_db(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* ocfs2_dx_root_block */
-int ocfs2_journal_access_dr(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_dr(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* ocfs2_dx_leaf */
-int ocfs2_journal_access_dl(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access_dl(handle_t *handle, struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh, int type);
 /* Anything that has no ecc */
-int ocfs2_journal_access(handle_t *handle, struct inode *inode,
+int ocfs2_journal_access(handle_t *handle, struct ocfs2_caching_info *ci,
 			 struct buffer_head *bh, int type);
 
 /*
@@ -477,6 +490,23 @@ static inline int ocfs2_calc_dxi_expand_credits(struct super_block *sb)
 	return credits;
 }
 
+/* inode update, new refcount block and its allocation credits. */
+#define OCFS2_REFCOUNT_TREE_CREATE_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1 \
+					    + OCFS2_SUBALLOC_ALLOC)
+
+/* inode and the refcount block update. */
+#define OCFS2_REFCOUNT_TREE_SET_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1)
+
+/*
+ * inode and the refcount block update.
+ * It doesn't include the credits for sub alloc change.
+ * So if we need to free the bit, OCFS2_SUBALLOC_FREE needs to be added.
+ */
+#define OCFS2_REFCOUNT_TREE_REMOVE_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1)
+
+/* 2 metadata alloc, 2 new blocks and root refcount block */
+#define OCFS2_EXPAND_REFCOUNT_TREE_CREDITS (OCFS2_SUBALLOC_ALLOC * 2 + 3)
+
 /*
  * Please note that the caller must make sure that root_el is the root
  * of extent tree. So for an inode, it should be &fe->id2.i_list. Otherwise
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index bac7e6abaf4..ac10f83edb9 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -297,8 +297,8 @@ void ocfs2_shutdown_local_alloc(struct ocfs2_super *osb)
 	}
 	memcpy(alloc_copy, alloc, bh->b_size);
 
-	status = ocfs2_journal_access_di(handle, local_alloc_inode, bh,
-					 OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(local_alloc_inode),
+					 bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_commit;
@@ -392,7 +392,7 @@ int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
 	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);
+	status = ocfs2_write_block(osb, alloc_bh, INODE_CACHE(inode));
 	if (status < 0)
 		mlog_errno(status);
 
@@ -678,7 +678,8 @@ int ocfs2_claim_local_alloc_bits(struct ocfs2_super *osb,
 	 * delete bits from it! */
 	*num_bits = bits_wanted;
 
-	status = ocfs2_journal_access_di(handle, local_alloc_inode,
+	status = ocfs2_journal_access_di(handle,
+					 INODE_CACHE(local_alloc_inode),
 					 osb->local_alloc_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
@@ -1156,7 +1157,8 @@ 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_di(handle, local_alloc_inode,
+	status = ocfs2_journal_access_di(handle,
+					 INODE_CACHE(local_alloc_inode),
 					 osb->local_alloc_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index 8601f934010..f010b22b1c4 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -69,7 +69,6 @@
 static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 			      struct inode *dir,
 			      struct inode *inode,
-			      struct dentry *dentry,
 			      dev_t dev,
 			      struct buffer_head **new_fe_bh,
 			      struct buffer_head *parent_fe_bh,
@@ -78,7 +77,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 
 static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
 				    struct inode **ret_orphan_dir,
-				    struct inode *inode,
+				    u64 blkno,
 				    char *name,
 				    struct ocfs2_dir_lookup_result *lookup);
 
@@ -358,8 +357,12 @@ static int ocfs2_mknod(struct inode *dir,
 	}
 	did_quota_inode = 1;
 
+	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);
+
 	/* do the real work now. */
-	status = ocfs2_mknod_locked(osb, dir, inode, dentry, dev,
+	status = ocfs2_mknod_locked(osb, dir, inode, dev,
 				    &new_fe_bh, parent_fe_bh, handle,
 				    inode_ac);
 	if (status < 0) {
@@ -375,7 +378,8 @@ static int ocfs2_mknod(struct inode *dir,
 			goto leave;
 		}
 
-		status = ocfs2_journal_access_di(handle, dir, parent_fe_bh,
+		status = ocfs2_journal_access_di(handle, INODE_CACHE(dir),
+						 parent_fe_bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -465,7 +469,6 @@ leave:
 static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 			      struct inode *dir,
 			      struct inode *inode,
-			      struct dentry *dentry,
 			      dev_t dev,
 			      struct buffer_head **new_fe_bh,
 			      struct buffer_head *parent_fe_bh,
@@ -479,10 +482,6 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	u16 suballoc_bit;
 	u16 feat;
 
-	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;
 
 	status = ocfs2_claim_new_inode(osb, handle, dir, parent_fe_bh,
@@ -507,9 +506,10 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 		mlog_errno(status);
 		goto leave;
 	}
-	ocfs2_set_new_buffer_uptodate(inode, *new_fe_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), *new_fe_bh);
 
-	status = ocfs2_journal_access_di(handle, inode, *new_fe_bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
+					 *new_fe_bh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -565,7 +565,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 	}
 
 	ocfs2_populate_inode(inode, fe, 1);
-	ocfs2_inode_set_new(osb, inode);
+	ocfs2_ci_set_new(osb, INODE_CACHE(inode));
 	if (!ocfs2_mount_local(osb)) {
 		status = ocfs2_create_new_inode_locks(inode);
 		if (status < 0)
@@ -682,7 +682,7 @@ static int ocfs2_link(struct dentry *old_dentry,
 		goto out_unlock_inode;
 	}
 
-	err = ocfs2_journal_access_di(handle, inode, fe_bh,
+	err = ocfs2_journal_access_di(handle, INODE_CACHE(inode), fe_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (err < 0) {
 		mlog_errno(err);
@@ -850,7 +850,8 @@ static int ocfs2_unlink(struct inode *dir,
 	}
 
 	if (inode_is_unlinkable(inode)) {
-		status = ocfs2_prepare_orphan_dir(osb, &orphan_dir, inode,
+		status = ocfs2_prepare_orphan_dir(osb, &orphan_dir,
+						  OCFS2_I(inode)->ip_blkno,
 						  orphan_name, &orphan_insert);
 		if (status < 0) {
 			mlog_errno(status);
@@ -866,7 +867,7 @@ static int ocfs2_unlink(struct inode *dir,
 		goto leave;
 	}
 
-	status = ocfs2_journal_access_di(handle, inode, fe_bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), fe_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1241,9 +1242,8 @@ static int ocfs2_rename(struct inode *old_dir,
 
 		if (S_ISDIR(new_inode->i_mode) || (new_inode->i_nlink == 1)) {
 			status = ocfs2_prepare_orphan_dir(osb, &orphan_dir,
-							  new_inode,
-							  orphan_name,
-							  &orphan_insert);
+						OCFS2_I(new_inode)->ip_blkno,
+						orphan_name, &orphan_insert);
 			if (status < 0) {
 				mlog_errno(status);
 				goto bail;
@@ -1284,7 +1284,8 @@ static int ocfs2_rename(struct inode *old_dir,
 				goto bail;
 			}
 		}
-		status = ocfs2_journal_access_di(handle, new_inode, newfe_bh,
+		status = ocfs2_journal_access_di(handle, INODE_CACHE(new_inode),
+						 newfe_bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -1331,7 +1332,8 @@ static int ocfs2_rename(struct inode *old_dir,
 	old_inode->i_ctime = CURRENT_TIME;
 	mark_inode_dirty(old_inode);
 
-	status = ocfs2_journal_access_di(handle, old_inode, old_inode_bh,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(old_inode),
+					 old_inode_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status >= 0) {
 		old_di = (struct ocfs2_dinode *) old_inode_bh->b_data;
@@ -1407,9 +1409,10 @@ 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_di(handle, old_dir,
-						      old_dir_bh,
-						      OCFS2_JOURNAL_ACCESS_WRITE);
+			status = ocfs2_journal_access_di(handle,
+							 INODE_CACHE(old_dir),
+							 old_dir_bh,
+							 OCFS2_JOURNAL_ACCESS_WRITE);
 			fe = (struct ocfs2_dinode *) old_dir_bh->b_data;
 			ocfs2_set_links_count(fe, old_dir->i_nlink);
 			status = ocfs2_journal_dirty(handle, old_dir_bh);
@@ -1527,9 +1530,11 @@ static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
 			mlog_errno(status);
 			goto bail;
 		}
-		ocfs2_set_new_buffer_uptodate(inode, bhs[virtual]);
+		ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode),
+					      bhs[virtual]);
 
-		status = ocfs2_journal_access(handle, inode, bhs[virtual],
+		status = ocfs2_journal_access(handle, INODE_CACHE(inode),
+					      bhs[virtual],
 					      OCFS2_JOURNAL_ACCESS_CREATE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -1692,7 +1697,11 @@ static int ocfs2_symlink(struct inode *dir,
 	}
 	did_quota_inode = 1;
 
-	status = ocfs2_mknod_locked(osb, dir, inode, dentry,
+	mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry,
+		   inode->i_mode, dentry->d_name.len,
+		   dentry->d_name.name);
+
+	status = ocfs2_mknod_locked(osb, dir, inode,
 				    0, &new_fe_bh, parent_fe_bh, handle,
 				    inode_ac);
 	if (status < 0) {
@@ -1842,7 +1851,7 @@ bail:
 
 static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
 				    struct inode **ret_orphan_dir,
-				    struct inode *inode,
+				    u64 blkno,
 				    char *name,
 				    struct ocfs2_dir_lookup_result *lookup)
 {
@@ -1850,7 +1859,7 @@ static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
 	struct buffer_head *orphan_dir_bh = NULL;
 	int status = 0;
 
-	status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, name);
+	status = ocfs2_blkno_stringify(blkno, name);
 	if (status < 0) {
 		mlog_errno(status);
 		return status;
@@ -1917,7 +1926,9 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
 		goto leave;
 	}
 
-	status = ocfs2_journal_access_di(handle, orphan_dir_inode, orphan_dir_bh,
+	status = ocfs2_journal_access_di(handle,
+					 INODE_CACHE(orphan_dir_inode),
+					 orphan_dir_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -2002,7 +2013,9 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
 		goto leave;
 	}
 
-	status = ocfs2_journal_access_di(handle,orphan_dir_inode,  orphan_dir_bh,
+	status = ocfs2_journal_access_di(handle,
+					 INODE_CACHE(orphan_dir_inode),
+					 orphan_dir_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -2028,6 +2041,274 @@ leave:
 	return status;
 }
 
+int ocfs2_create_inode_in_orphan(struct inode *dir,
+				 int mode,
+				 struct inode **new_inode)
+{
+	int status, did_quota_inode = 0;
+	struct inode *inode = NULL;
+	struct inode *orphan_dir = NULL;
+	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
+	struct ocfs2_dinode *di = NULL;
+	handle_t *handle = NULL;
+	char orphan_name[OCFS2_ORPHAN_NAMELEN + 1];
+	struct buffer_head *parent_di_bh = NULL;
+	struct buffer_head *new_di_bh = NULL;
+	struct ocfs2_alloc_context *inode_ac = NULL;
+	struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
+
+	status = ocfs2_inode_lock(dir, &parent_di_bh, 1);
+	if (status < 0) {
+		if (status != -ENOENT)
+			mlog_errno(status);
+		return status;
+	}
+
+	/*
+	 * We give the orphan dir the root blkno to fake an orphan name,
+	 * and allocate enough space for our insertion.
+	 */
+	status = ocfs2_prepare_orphan_dir(osb, &orphan_dir,
+					  osb->root_blkno,
+					  orphan_name, &orphan_insert);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	/* reserve an inode spot */
+	status = ocfs2_reserve_new_inode(osb, &inode_ac);
+	if (status < 0) {
+		if (status != -ENOSPC)
+			mlog_errno(status);
+		goto leave;
+	}
+
+	inode = ocfs2_get_init_inode(dir, mode);
+	if (!inode) {
+		status = -ENOMEM;
+		mlog_errno(status);
+		goto leave;
+	}
+
+	handle = ocfs2_start_trans(osb, ocfs2_mknod_credits(osb->sb, 0, 0));
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		handle = NULL;
+		mlog_errno(status);
+		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, inode,
+				    0, &new_di_bh, parent_di_bh, handle,
+				    inode_ac);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, orphan_name);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	di = (struct ocfs2_dinode *)new_di_bh->b_data;
+	status = ocfs2_orphan_add(osb, handle, inode, di, orphan_name,
+				  &orphan_insert, orphan_dir);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	/* get open lock so that only nodes can't remove it from orphan dir. */
+	status = ocfs2_open_lock(inode);
+	if (status < 0)
+		mlog_errno(status);
+
+leave:
+	if (status < 0 && did_quota_inode)
+		vfs_dq_free_inode(inode);
+	if (handle)
+		ocfs2_commit_trans(osb, handle);
+
+	if (orphan_dir) {
+		/* This was locked for us in ocfs2_prepare_orphan_dir() */
+		ocfs2_inode_unlock(orphan_dir, 1);
+		mutex_unlock(&orphan_dir->i_mutex);
+		iput(orphan_dir);
+	}
+
+	if (status == -ENOSPC)
+		mlog(0, "Disk is full\n");
+
+	if ((status < 0) && inode) {
+		clear_nlink(inode);
+		iput(inode);
+	}
+
+	if (inode_ac)
+		ocfs2_free_alloc_context(inode_ac);
+
+	brelse(new_di_bh);
+
+	if (!status)
+		*new_inode = inode;
+
+	ocfs2_free_dir_lookup_result(&orphan_insert);
+
+	ocfs2_inode_unlock(dir, 1);
+	brelse(parent_di_bh);
+	return status;
+}
+
+int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
+				   struct inode *inode,
+				   struct dentry *dentry)
+{
+	int status = 0;
+	struct buffer_head *parent_di_bh = NULL;
+	handle_t *handle = NULL;
+	struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
+	struct ocfs2_dinode *dir_di, *di;
+	struct inode *orphan_dir_inode = NULL;
+	struct buffer_head *orphan_dir_bh = NULL;
+	struct buffer_head *di_bh = NULL;
+	struct ocfs2_dir_lookup_result lookup = { NULL, };
+
+	mlog_entry("(0x%p, 0x%p, %.*s')\n", dir, dentry,
+		   dentry->d_name.len, dentry->d_name.name);
+
+	status = ocfs2_inode_lock(dir, &parent_di_bh, 1);
+	if (status < 0) {
+		if (status != -ENOENT)
+			mlog_errno(status);
+		return status;
+	}
+
+	dir_di = (struct ocfs2_dinode *) parent_di_bh->b_data;
+	if (!dir_di->i_links_count) {
+		/* can't make a file in a deleted directory. */
+		status = -ENOENT;
+		goto leave;
+	}
+
+	status = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
+					   dentry->d_name.len);
+	if (status)
+		goto leave;
+
+	/* get a spot inside the dir. */
+	status = ocfs2_prepare_dir_for_insert(osb, dir, parent_di_bh,
+					      dentry->d_name.name,
+					      dentry->d_name.len, &lookup);
+	if (status < 0) {
+		mlog_errno(status);
+		goto leave;
+	}
+
+	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
+						       ORPHAN_DIR_SYSTEM_INODE,
+						       osb->slot_num);
+	if (!orphan_dir_inode) {
+		status = -EEXIST;
+		mlog_errno(status);
+		goto leave;
+	}
+
+	mutex_lock(&orphan_dir_inode->i_mutex);
+
+	status = ocfs2_inode_lock(orphan_dir_inode, &orphan_dir_bh, 1);
+	if (status < 0) {
+		mlog_errno(status);
+		mutex_unlock(&orphan_dir_inode->i_mutex);
+		iput(orphan_dir_inode);
+		goto leave;
+	}
+
+	status = ocfs2_read_inode_block(inode, &di_bh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto orphan_unlock;
+	}
+
+	handle = ocfs2_start_trans(osb, ocfs2_rename_credits(osb->sb));
+	if (IS_ERR(handle)) {
+		status = PTR_ERR(handle);
+		handle = NULL;
+		mlog_errno(status);
+		goto orphan_unlock;
+	}
+
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
+					 di_bh, OCFS2_JOURNAL_ACCESS_WRITE);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_commit;
+	}
+
+	status = ocfs2_orphan_del(osb, handle, orphan_dir_inode, inode,
+				  orphan_dir_bh);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_commit;
+	}
+
+	di = (struct ocfs2_dinode *)di_bh->b_data;
+	le32_add_cpu(&di->i_flags, -OCFS2_ORPHANED_FL);
+	di->i_orphaned_slot = 0;
+	ocfs2_journal_dirty(handle, di_bh);
+
+	status = ocfs2_add_entry(handle, dentry, inode,
+				 OCFS2_I(inode)->ip_blkno, parent_di_bh,
+				 &lookup);
+	if (status < 0) {
+		mlog_errno(status);
+		goto out_commit;
+	}
+
+	status = ocfs2_dentry_attach_lock(dentry, inode,
+					  OCFS2_I(dir)->ip_blkno);
+	if (status) {
+		mlog_errno(status);
+		goto out_commit;
+	}
+
+	insert_inode_hash(inode);
+	dentry->d_op = &ocfs2_dentry_ops;
+	d_instantiate(dentry, inode);
+	status = 0;
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+orphan_unlock:
+	ocfs2_inode_unlock(orphan_dir_inode, 1);
+	mutex_unlock(&orphan_dir_inode->i_mutex);
+	iput(orphan_dir_inode);
+leave:
+
+	ocfs2_inode_unlock(dir, 1);
+
+	brelse(di_bh);
+	brelse(parent_di_bh);
+	brelse(orphan_dir_bh);
+
+	ocfs2_free_dir_lookup_result(&lookup);
+
+	mlog_exit(status);
+
+	return status;
+}
+
 const struct inode_operations ocfs2_dir_iops = {
 	.create		= ocfs2_create,
 	.lookup		= ocfs2_lookup,
diff --git a/fs/ocfs2/namei.h b/fs/ocfs2/namei.h
index 688aef64c87..e5d059d4f11 100644
--- a/fs/ocfs2/namei.h
+++ b/fs/ocfs2/namei.h
@@ -35,5 +35,11 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
 		     struct inode *orphan_dir_inode,
 		     struct inode *inode,
 		     struct buffer_head *orphan_dir_bh);
+int ocfs2_create_inode_in_orphan(struct inode *dir,
+				 int mode,
+				 struct inode **new_inode);
+int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
+				   struct inode *new_inode,
+				   struct dentry *new_dentry);
 
 #endif /* OCFS2_NAMEI_H */
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index 39e1d5a3950..eae40460242 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -51,20 +51,51 @@
 /* For struct ocfs2_blockcheck_stats */
 #include "blockcheck.h"
 
+
+/* Caching of metadata buffers */
+
 /* Most user visible OCFS2 inodes will have very few pieces of
  * metadata, but larger files (including bitmaps, etc) must be taken
  * into account when designing an access scheme. We allow a small
  * amount of inlined blocks to be stored on an array and grow the
  * structure into a rb tree when necessary. */
-#define OCFS2_INODE_MAX_CACHE_ARRAY 2
+#define OCFS2_CACHE_INFO_MAX_ARRAY 2
+
+/* Flags for ocfs2_caching_info */
+
+enum ocfs2_caching_info_flags {
+	/* Indicates that the metadata cache is using the inline array */
+	OCFS2_CACHE_FL_INLINE	= 1<<1,
+};
 
+struct ocfs2_caching_operations;
 struct ocfs2_caching_info {
+	/*
+	 * The parent structure provides the locks, but because the
+	 * parent structure can differ, it provides locking operations
+	 * to struct ocfs2_caching_info.
+	 */
+	const struct ocfs2_caching_operations *ci_ops;
+
+	/* next two are protected by trans_inc_lock */
+	/* which transaction were we created on? Zero if none. */
+	unsigned long		ci_created_trans;
+	/* last transaction we were a part of. */
+	unsigned long		ci_last_trans;
+
+	/* Cache structures */
+	unsigned int		ci_flags;
 	unsigned int		ci_num_cached;
 	union {
-		sector_t	ci_array[OCFS2_INODE_MAX_CACHE_ARRAY];
+	sector_t	ci_array[OCFS2_CACHE_INFO_MAX_ARRAY];
 		struct rb_root	ci_tree;
 	} ci_cache;
 };
+/*
+ * Need this prototype here instead of in uptodate.h because journal.h
+ * uses it.
+ */
+struct super_block *ocfs2_metadata_cache_get_super(struct ocfs2_caching_info *ci);
 
 /* this limits us to 256 nodes
  * if we need more, we can do a kmalloc for the map */
@@ -377,12 +408,17 @@ struct ocfs2_super
 
 	/* the group we used to allocate inodes. */
 	u64				osb_inode_alloc_group;
+
+	/* rb tree root for refcount lock. */
+	struct rb_root	osb_rf_lock_tree;
+	struct ocfs2_refcount_tree *osb_ref_tree_lru;
 };
 
 #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,
+typedef int (*ocfs2_journal_access_func)(handle_t *handle,
+					 struct ocfs2_caching_info *ci,
 					 struct buffer_head *bh, int type);
 
 static inline int ocfs2_should_order_data(struct inode *inode)
@@ -480,6 +516,13 @@ static inline void ocfs2_add_links_count(struct ocfs2_dinode *di, int n)
 	ocfs2_set_links_count(di, links);
 }
 
+static inline int ocfs2_refcount_tree(struct ocfs2_super *osb)
+{
+	if (osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_REFCOUNT_TREE)
+		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
@@ -578,6 +621,9 @@ static inline int ocfs2_uses_extended_slot_map(struct ocfs2_super *osb)
 #define OCFS2_IS_VALID_DX_LEAF(ptr)					\
 	(!strcmp((ptr)->dl_signature, OCFS2_DX_LEAF_SIGNATURE))
 
+#define OCFS2_IS_VALID_REFCOUNT_BLOCK(ptr)				\
+	(!strcmp((ptr)->rf_signature, OCFS2_REFCOUNT_BLOCK_SIGNATURE))
+
 static inline unsigned long ino_from_blkno(struct super_block *sb,
 					   u64 blkno)
 {
diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h
index 7ab6e9e5e77..e9431e4a5e7 100644
--- a/fs/ocfs2/ocfs2_fs.h
+++ b/fs/ocfs2/ocfs2_fs.h
@@ -68,6 +68,7 @@
 #define OCFS2_DIR_TRAILER_SIGNATURE	"DIRTRL1"
 #define OCFS2_DX_ROOT_SIGNATURE		"DXDIR01"
 #define OCFS2_DX_LEAF_SIGNATURE		"DXLEAF1"
+#define OCFS2_REFCOUNT_BLOCK_SIGNATURE	"REFCNT1"
 
 /* Compatibility flags */
 #define OCFS2_HAS_COMPAT_FEATURE(sb,mask)			\
@@ -98,7 +99,8 @@
 					 | OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK \
 					 | OCFS2_FEATURE_INCOMPAT_XATTR \
 					 | OCFS2_FEATURE_INCOMPAT_META_ECC \
-					 | OCFS2_FEATURE_INCOMPAT_INDEXED_DIRS)
+					 | OCFS2_FEATURE_INCOMPAT_INDEXED_DIRS \
+					 | OCFS2_FEATURE_INCOMPAT_REFCOUNT_TREE)
 #define OCFS2_FEATURE_RO_COMPAT_SUPP	(OCFS2_FEATURE_RO_COMPAT_UNWRITTEN \
 					 | OCFS2_FEATURE_RO_COMPAT_USRQUOTA \
 					 | OCFS2_FEATURE_RO_COMPAT_GRPQUOTA)
@@ -160,6 +162,9 @@
 /* Metadata checksum and error correction */
 #define OCFS2_FEATURE_INCOMPAT_META_ECC		0x0800
 
+/* Refcount tree support */
+#define OCFS2_FEATURE_INCOMPAT_REFCOUNT_TREE	0x1000
+
 /*
  * backup superblock flag is used to indicate that this volume
  * has backup superblocks.
@@ -223,6 +228,7 @@
 #define OCFS2_HAS_XATTR_FL	(0x0002)
 #define OCFS2_INLINE_XATTR_FL	(0x0004)
 #define OCFS2_INDEXED_DIR_FL	(0x0008)
+#define OCFS2_HAS_REFCOUNT_FL   (0x0010)
 
 /* Inode attributes, keep in sync with EXT2 */
 #define OCFS2_SECRM_FL		(0x00000001)	/* Secure deletion */
@@ -241,8 +247,11 @@
 /*
  * Extent record flags (e_node.leaf.flags)
  */
-#define OCFS2_EXT_UNWRITTEN	(0x01)	/* Extent is allocated but
-					 * unwritten */
+#define OCFS2_EXT_UNWRITTEN		(0x01)	/* Extent is allocated but
+						 * unwritten */
+#define OCFS2_EXT_REFCOUNTED		(0x02)  /* Extent is reference
+						 * counted in an associated
+						 * refcount tree */
 
 /*
  * ioctl commands
@@ -292,6 +301,15 @@ struct ocfs2_new_group_input {
 #define OCFS2_IOC_GROUP_ADD	_IOW('o', 2,struct ocfs2_new_group_input)
 #define OCFS2_IOC_GROUP_ADD64	_IOW('o', 3,struct ocfs2_new_group_input)
 
+/* Used to pass 2 file names to reflink. */
+struct reflink_arguments {
+	__u64 old_path;
+	__u64 new_path;
+	__u64 preserve;
+};
+#define OCFS2_IOC_REFLINK	_IOW('o', 4, struct reflink_arguments)
+
+
 /*
  * Journal Flags (ocfs2_dinode.id1.journal1.i_flags)
  */
@@ -717,7 +735,8 @@ struct ocfs2_dinode {
 	__le64 i_xattr_loc;
 /*80*/	struct ocfs2_block_check i_check;	/* Error checking */
 /*88*/	__le64 i_dx_root;		/* Pointer to dir index root block */
-	__le64 i_reserved2[5];
+/*90*/	__le64 i_refcount_loc;
+	__le64 i_reserved2[4];
 /*B8*/	union {
 		__le64 i_pad1;		/* Generic way to refer to this
 					   64bit union */
@@ -901,6 +920,60 @@ struct ocfs2_group_desc
 /*40*/	__u8    bg_bitmap[0];
 };
 
+struct ocfs2_refcount_rec {
+/*00*/	__le64 r_cpos;		/* Physical offset, in clusters */
+	__le32 r_clusters;	/* Clusters covered by this extent */
+	__le32 r_refcount;	/* Reference count of this extent */
+/*10*/
+};
+#define OCFS2_32BIT_POS_MASK		(0xffffffffULL)
+
+#define OCFS2_REFCOUNT_LEAF_FL          (0x00000001)
+#define OCFS2_REFCOUNT_TREE_FL          (0x00000002)
+
+struct ocfs2_refcount_list {
+/*00*/	__le16 rl_count;	/* Maximum number of entries possible
+				   in rl_records */
+	__le16 rl_used;		/* Current number of used records */
+	__le32 rl_reserved2;
+	__le64 rl_reserved1;	/* Pad to sizeof(ocfs2_refcount_record) */
+/*10*/	struct ocfs2_refcount_rec rl_recs[0];	/* Refcount records */
+};
+
+
+struct ocfs2_refcount_block {
+/*00*/	__u8 rf_signature[8];		/* Signature for verification */
+	__le16 rf_suballoc_slot;	/* Slot suballocator this block
+					   belongs to */
+	__le16 rf_suballoc_bit;		/* Bit offset in suballocator
+					   block group */
+	__le32 rf_fs_generation;	/* Must match superblock */
+/*10*/	__le64 rf_blkno;		/* Offset on disk, in blocks */
+	__le64 rf_parent;		/* Parent block, only valid if
+					   OCFS2_REFCOUNT_LEAF_FL is set in
+					   rf_flags */
+/*20*/	struct ocfs2_block_check rf_check;	/* Error checking */
+	__le64 rf_last_eb_blk;		/* Pointer to last extent block */
+/*30*/	__le32 rf_count;		/* Number of inodes sharing this
+					   refcount tree */
+	__le32 rf_flags;		/* See the flags above */
+	__le32 rf_clusters;		/* clusters covered by refcount tree. */
+	__le32 rf_cpos;			/* cluster offset in refcount tree.*/
+/*40*/	__le32 rf_generation;		/* generation number. all be the same
+					 * for the same refcount tree. */
+	__le32 rf_reserved0;
+	__le64 rf_reserved1[7];
+/*80*/	union {
+		struct ocfs2_refcount_list rf_records;  /* List of refcount
+							  records */
+		struct ocfs2_extent_list rf_list;	/* Extent record list,
+							only valid if
+							OCFS2_REFCOUNT_TREE_FL
+							is set in rf_flags */
+	};
+/* Actual on-disk size is one block */
+};
+
 /*
  * On disk extended attribute structure for OCFS2.
  */
@@ -1312,6 +1385,32 @@ static inline u16 ocfs2_xattr_recs_per_xb(struct super_block *sb)
 
 	return size / sizeof(struct ocfs2_extent_rec);
 }
+
+static inline u16 ocfs2_extent_recs_per_rb(struct super_block *sb)
+{
+	int size;
+
+	size = sb->s_blocksize -
+		offsetof(struct ocfs2_refcount_block, rf_list.l_recs);
+
+	return size / sizeof(struct ocfs2_extent_rec);
+}
+
+static inline u16 ocfs2_refcount_recs_per_rb(struct super_block *sb)
+{
+	int size;
+
+	size = sb->s_blocksize -
+		offsetof(struct ocfs2_refcount_block, rf_records.rl_recs);
+
+	return size / sizeof(struct ocfs2_refcount_rec);
+}
+
+static inline u32
+ocfs2_get_ref_rec_low_cpos(const struct ocfs2_refcount_rec *rec)
+{
+	return le64_to_cpu(rec->r_cpos) & OCFS2_32BIT_POS_MASK;
+}
 #else
 static inline int ocfs2_fast_symlink_chars(int blocksize)
 {
diff --git a/fs/ocfs2/ocfs2_lockid.h b/fs/ocfs2/ocfs2_lockid.h
index c212cf5a2bd..d277aabf5df 100644
--- a/fs/ocfs2/ocfs2_lockid.h
+++ b/fs/ocfs2/ocfs2_lockid.h
@@ -49,6 +49,7 @@ enum ocfs2_lock_type {
 	OCFS2_LOCK_TYPE_QINFO,
 	OCFS2_LOCK_TYPE_NFS_SYNC,
 	OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+	OCFS2_LOCK_TYPE_REFCOUNT,
 	OCFS2_NUM_LOCK_TYPES
 };
 
@@ -89,6 +90,9 @@ static inline char ocfs2_lock_type_char(enum ocfs2_lock_type type)
 		case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
 			c = 'P';
 			break;
+		case OCFS2_LOCK_TYPE_REFCOUNT:
+			c = 'T';
+			break;
 		default:
 			c = '\0';
 	}
@@ -110,6 +114,7 @@ static char *ocfs2_lock_type_strings[] = {
 	[OCFS2_LOCK_TYPE_QINFO] = "Quota",
 	[OCFS2_LOCK_TYPE_NFS_SYNC] = "NFSSync",
 	[OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
+	[OCFS2_LOCK_TYPE_REFCOUNT] = "Refcount",
 };
 
 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
index 3fb96fcd4c8..e5df9d170b0 100644
--- a/fs/ocfs2/quota.h
+++ b/fs/ocfs2/quota.h
@@ -109,7 +109,7 @@ 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 const struct dquot_operations ocfs2_quota_operations;
 extern struct quota_format_type ocfs2_quota_format;
 
 int ocfs2_quota_setup(void);
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index 44f2a5e1d04..b437dc0c4ca 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -154,7 +154,7 @@ static int ocfs2_get_quota_block(struct inode *inode, int block,
 		err = -EIO;
 		mlog_errno(err);
 	}
-	return err;;
+	return err;
 }
 
 /* Read data from global quotafile - avoid pagecache and such because we cannot
@@ -253,8 +253,9 @@ ssize_t ocfs2_quota_write(struct super_block *sb, int type,
 	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);
+	ocfs2_set_buffer_uptodate(INODE_CACHE(gqinode), bh);
+	err = ocfs2_journal_access_dq(handle, INODE_CACHE(gqinode), bh,
+				      ja_type);
 	if (err < 0) {
 		brelse(bh);
 		goto out;
@@ -849,7 +850,7 @@ static void ocfs2_destroy_dquot(struct dquot *dquot)
 	kmem_cache_free(ocfs2_dquot_cachep, dquot);
 }
 
-struct dquot_operations ocfs2_quota_operations = {
+const struct dquot_operations ocfs2_quota_operations = {
 	.initialize	= dquot_initialize,
 	.drop		= dquot_drop,
 	.alloc_space	= dquot_alloc_space,
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index bdb09cb6e1f..1a2c50a759f 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -108,7 +108,7 @@ static int ocfs2_modify_bh(struct inode *inode, struct buffer_head *bh,
 		mlog_errno(status);
 		return status;
 	}
-	status = ocfs2_journal_access_dq(handle, inode, bh,
+	status = ocfs2_journal_access_dq(handle, INODE_CACHE(inode), bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -510,7 +510,8 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
 				goto out_commit;
 			}
 			/* Release local quota file entry */
-			status = ocfs2_journal_access_dq(handle, lqinode,
+			status = ocfs2_journal_access_dq(handle,
+					INODE_CACHE(lqinode),
 					qbh, OCFS2_JOURNAL_ACCESS_WRITE);
 			if (status < 0) {
 				mlog_errno(status);
@@ -619,7 +620,8 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb,
 			mlog_errno(status);
 			goto out_bh;
 		}
-		status = ocfs2_journal_access_dq(handle, lqinode, bh,
+		status = ocfs2_journal_access_dq(handle, INODE_CACHE(lqinode),
+						 bh,
 						 OCFS2_JOURNAL_ACCESS_WRITE);
 		if (status < 0) {
 			mlog_errno(status);
@@ -993,8 +995,8 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
 		goto out_trans;
 	}
 	dchunk = (struct ocfs2_local_disk_chunk *)bh->b_data;
-	ocfs2_set_new_buffer_uptodate(lqinode, bh);
-	status = ocfs2_journal_access_dq(handle, lqinode, bh,
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(lqinode), bh);
+	status = ocfs2_journal_access_dq(handle, INODE_CACHE(lqinode), bh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1027,8 +1029,8 @@ static struct ocfs2_quota_chunk *ocfs2_local_quota_add_chunk(
 		mlog_errno(status);
 		goto out_trans;
 	}
-	ocfs2_set_new_buffer_uptodate(lqinode, dbh);
-	status = ocfs2_journal_access_dq(handle, lqinode, dbh,
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(lqinode), dbh);
+	status = ocfs2_journal_access_dq(handle, INODE_CACHE(lqinode), dbh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1131,7 +1133,7 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
 		mlog_errno(status);
 		goto out;
 	}
-	ocfs2_set_new_buffer_uptodate(lqinode, bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(lqinode), bh);
 
 	/* Local quota info, chunk header and the new block we initialize */
 	handle = ocfs2_start_trans(OCFS2_SB(sb),
@@ -1143,7 +1145,7 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
 		goto out;
 	}
 	/* Zero created block */
-	status = ocfs2_journal_access_dq(handle, lqinode, bh,
+	status = ocfs2_journal_access_dq(handle, INODE_CACHE(lqinode), bh,
 				 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1158,7 +1160,8 @@ static struct ocfs2_quota_chunk *ocfs2_extend_local_quota_file(
 		goto out_trans;
 	}
 	/* Update chunk header */
-	status = ocfs2_journal_access_dq(handle, lqinode, chunk->qc_headerbh,
+	status = ocfs2_journal_access_dq(handle, INODE_CACHE(lqinode),
+					 chunk->qc_headerbh,
 				 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1292,7 +1295,8 @@ static int ocfs2_local_release_dquot(struct dquot *dquot)
 		goto out;
 	}
 
-	status = ocfs2_journal_access_dq(handle, sb_dqopt(sb)->files[type],
+	status = ocfs2_journal_access_dq(handle,
+			INODE_CACHE(sb_dqopt(sb)->files[type]),
 			od->dq_chunk->qc_headerbh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c
new file mode 100644
index 00000000000..60287fc56bc
--- /dev/null
+++ b/fs/ocfs2/refcounttree.c
@@ -0,0 +1,4313 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * refcounttree.c
+ *
+ * Copyright (C) 2009 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/sort.h>
+#define MLOG_MASK_PREFIX ML_REFCOUNT
+#include <cluster/masklog.h>
+#include "ocfs2.h"
+#include "inode.h"
+#include "alloc.h"
+#include "suballoc.h"
+#include "journal.h"
+#include "uptodate.h"
+#include "super.h"
+#include "buffer_head_io.h"
+#include "blockcheck.h"
+#include "refcounttree.h"
+#include "sysfile.h"
+#include "dlmglue.h"
+#include "extent_map.h"
+#include "aops.h"
+#include "xattr.h"
+#include "namei.h"
+
+#include <linux/bio.h>
+#include <linux/blkdev.h>
+#include <linux/gfp.h>
+#include <linux/slab.h>
+#include <linux/writeback.h>
+#include <linux/pagevec.h>
+#include <linux/swap.h>
+#include <linux/security.h>
+#include <linux/fsnotify.h>
+#include <linux/quotaops.h>
+#include <linux/namei.h>
+#include <linux/mount.h>
+
+struct ocfs2_cow_context {
+	struct inode *inode;
+	u32 cow_start;
+	u32 cow_len;
+	struct ocfs2_extent_tree data_et;
+	struct ocfs2_refcount_tree *ref_tree;
+	struct buffer_head *ref_root_bh;
+	struct ocfs2_alloc_context *meta_ac;
+	struct ocfs2_alloc_context *data_ac;
+	struct ocfs2_cached_dealloc_ctxt dealloc;
+	void *cow_object;
+	struct ocfs2_post_refcount *post_refcount;
+	int extra_credits;
+	int (*get_clusters)(struct ocfs2_cow_context *context,
+			    u32 v_cluster, u32 *p_cluster,
+			    u32 *num_clusters,
+			    unsigned int *extent_flags);
+	int (*cow_duplicate_clusters)(handle_t *handle,
+				      struct ocfs2_cow_context *context,
+				      u32 cpos, u32 old_cluster,
+				      u32 new_cluster, u32 new_len);
+};
+
+static inline struct ocfs2_refcount_tree *
+cache_info_to_refcount(struct ocfs2_caching_info *ci)
+{
+	return container_of(ci, struct ocfs2_refcount_tree, rf_ci);
+}
+
+static int ocfs2_validate_refcount_block(struct super_block *sb,
+					 struct buffer_head *bh)
+{
+	int rc;
+	struct ocfs2_refcount_block *rb =
+		(struct ocfs2_refcount_block *)bh->b_data;
+
+	mlog(0, "Validating refcount 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, &rb->rf_check);
+	if (rc) {
+		mlog(ML_ERROR, "Checksum failed for refcount block %llu\n",
+		     (unsigned long long)bh->b_blocknr);
+		return rc;
+	}
+
+
+	if (!OCFS2_IS_VALID_REFCOUNT_BLOCK(rb)) {
+		ocfs2_error(sb,
+			    "Refcount block #%llu has bad signature %.*s",
+			    (unsigned long long)bh->b_blocknr, 7,
+			    rb->rf_signature);
+		return -EINVAL;
+	}
+
+	if (le64_to_cpu(rb->rf_blkno) != bh->b_blocknr) {
+		ocfs2_error(sb,
+			    "Refcount block #%llu has an invalid rf_blkno "
+			    "of %llu",
+			    (unsigned long long)bh->b_blocknr,
+			    (unsigned long long)le64_to_cpu(rb->rf_blkno));
+		return -EINVAL;
+	}
+
+	if (le32_to_cpu(rb->rf_fs_generation) != OCFS2_SB(sb)->fs_generation) {
+		ocfs2_error(sb,
+			    "Refcount block #%llu has an invalid "
+			    "rf_fs_generation of #%u",
+			    (unsigned long long)bh->b_blocknr,
+			    le32_to_cpu(rb->rf_fs_generation));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int ocfs2_read_refcount_block(struct ocfs2_caching_info *ci,
+				     u64 rb_blkno,
+				     struct buffer_head **bh)
+{
+	int rc;
+	struct buffer_head *tmp = *bh;
+
+	rc = ocfs2_read_block(ci, rb_blkno, &tmp,
+			      ocfs2_validate_refcount_block);
+
+	/* If ocfs2_read_block() got us a new bh, pass it up. */
+	if (!rc && !*bh)
+		*bh = tmp;
+
+	return rc;
+}
+
+static u64 ocfs2_refcount_cache_owner(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_refcount_tree *rf = cache_info_to_refcount(ci);
+
+	return rf->rf_blkno;
+}
+
+static struct super_block *
+ocfs2_refcount_cache_get_super(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_refcount_tree *rf = cache_info_to_refcount(ci);
+
+	return rf->rf_sb;
+}
+
+static void ocfs2_refcount_cache_lock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_refcount_tree *rf = cache_info_to_refcount(ci);
+
+	spin_lock(&rf->rf_lock);
+}
+
+static void ocfs2_refcount_cache_unlock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_refcount_tree *rf = cache_info_to_refcount(ci);
+
+	spin_unlock(&rf->rf_lock);
+}
+
+static void ocfs2_refcount_cache_io_lock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_refcount_tree *rf = cache_info_to_refcount(ci);
+
+	mutex_lock(&rf->rf_io_mutex);
+}
+
+static void ocfs2_refcount_cache_io_unlock(struct ocfs2_caching_info *ci)
+{
+	struct ocfs2_refcount_tree *rf = cache_info_to_refcount(ci);
+
+	mutex_unlock(&rf->rf_io_mutex);
+}
+
+static const struct ocfs2_caching_operations ocfs2_refcount_caching_ops = {
+	.co_owner		= ocfs2_refcount_cache_owner,
+	.co_get_super		= ocfs2_refcount_cache_get_super,
+	.co_cache_lock		= ocfs2_refcount_cache_lock,
+	.co_cache_unlock	= ocfs2_refcount_cache_unlock,
+	.co_io_lock		= ocfs2_refcount_cache_io_lock,
+	.co_io_unlock		= ocfs2_refcount_cache_io_unlock,
+};
+
+static struct ocfs2_refcount_tree *
+ocfs2_find_refcount_tree(struct ocfs2_super *osb, u64 blkno)
+{
+	struct rb_node *n = osb->osb_rf_lock_tree.rb_node;
+	struct ocfs2_refcount_tree *tree = NULL;
+
+	while (n) {
+		tree = rb_entry(n, struct ocfs2_refcount_tree, rf_node);
+
+		if (blkno < tree->rf_blkno)
+			n = n->rb_left;
+		else if (blkno > tree->rf_blkno)
+			n = n->rb_right;
+		else
+			return tree;
+	}
+
+	return NULL;
+}
+
+/* osb_lock is already locked. */
+static void ocfs2_insert_refcount_tree(struct ocfs2_super *osb,
+				       struct ocfs2_refcount_tree *new)
+{
+	u64 rf_blkno = new->rf_blkno;
+	struct rb_node *parent = NULL;
+	struct rb_node **p = &osb->osb_rf_lock_tree.rb_node;
+	struct ocfs2_refcount_tree *tmp;
+
+	while (*p) {
+		parent = *p;
+
+		tmp = rb_entry(parent, struct ocfs2_refcount_tree,
+			       rf_node);
+
+		if (rf_blkno < tmp->rf_blkno)
+			p = &(*p)->rb_left;
+		else if (rf_blkno > tmp->rf_blkno)
+			p = &(*p)->rb_right;
+		else {
+			/* This should never happen! */
+			mlog(ML_ERROR, "Duplicate refcount block %llu found!\n",
+			     (unsigned long long)rf_blkno);
+			BUG();
+		}
+	}
+
+	rb_link_node(&new->rf_node, parent, p);
+	rb_insert_color(&new->rf_node, &osb->osb_rf_lock_tree);
+}
+
+static void ocfs2_free_refcount_tree(struct ocfs2_refcount_tree *tree)
+{
+	ocfs2_metadata_cache_exit(&tree->rf_ci);
+	ocfs2_simple_drop_lockres(OCFS2_SB(tree->rf_sb), &tree->rf_lockres);
+	ocfs2_lock_res_free(&tree->rf_lockres);
+	kfree(tree);
+}
+
+static inline void
+ocfs2_erase_refcount_tree_from_list_no_lock(struct ocfs2_super *osb,
+					struct ocfs2_refcount_tree *tree)
+{
+	rb_erase(&tree->rf_node, &osb->osb_rf_lock_tree);
+	if (osb->osb_ref_tree_lru && osb->osb_ref_tree_lru == tree)
+		osb->osb_ref_tree_lru = NULL;
+}
+
+static void ocfs2_erase_refcount_tree_from_list(struct ocfs2_super *osb,
+					struct ocfs2_refcount_tree *tree)
+{
+	spin_lock(&osb->osb_lock);
+	ocfs2_erase_refcount_tree_from_list_no_lock(osb, tree);
+	spin_unlock(&osb->osb_lock);
+}
+
+void ocfs2_kref_remove_refcount_tree(struct kref *kref)
+{
+	struct ocfs2_refcount_tree *tree =
+		container_of(kref, struct ocfs2_refcount_tree, rf_getcnt);
+
+	ocfs2_free_refcount_tree(tree);
+}
+
+static inline void
+ocfs2_refcount_tree_get(struct ocfs2_refcount_tree *tree)
+{
+	kref_get(&tree->rf_getcnt);
+}
+
+static inline void
+ocfs2_refcount_tree_put(struct ocfs2_refcount_tree *tree)
+{
+	kref_put(&tree->rf_getcnt, ocfs2_kref_remove_refcount_tree);
+}
+
+static inline void ocfs2_init_refcount_tree_ci(struct ocfs2_refcount_tree *new,
+					       struct super_block *sb)
+{
+	ocfs2_metadata_cache_init(&new->rf_ci, &ocfs2_refcount_caching_ops);
+	mutex_init(&new->rf_io_mutex);
+	new->rf_sb = sb;
+	spin_lock_init(&new->rf_lock);
+}
+
+static inline void ocfs2_init_refcount_tree_lock(struct ocfs2_super *osb,
+					struct ocfs2_refcount_tree *new,
+					u64 rf_blkno, u32 generation)
+{
+	init_rwsem(&new->rf_sem);
+	ocfs2_refcount_lock_res_init(&new->rf_lockres, osb,
+				     rf_blkno, generation);
+}
+
+static struct ocfs2_refcount_tree*
+ocfs2_allocate_refcount_tree(struct ocfs2_super *osb, u64 rf_blkno)
+{
+	struct ocfs2_refcount_tree *new;
+
+	new = kzalloc(sizeof(struct ocfs2_refcount_tree), GFP_NOFS);
+	if (!new)
+		return NULL;
+
+	new->rf_blkno = rf_blkno;
+	kref_init(&new->rf_getcnt);
+	ocfs2_init_refcount_tree_ci(new, osb->sb);
+
+	return new;
+}
+
+static int ocfs2_get_refcount_tree(struct ocfs2_super *osb, u64 rf_blkno,
+				   struct ocfs2_refcount_tree **ret_tree)
+{
+	int ret = 0;
+	struct ocfs2_refcount_tree *tree, *new = NULL;
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_refcount_block *ref_rb;
+
+	spin_lock(&osb->osb_lock);
+	if (osb->osb_ref_tree_lru &&
+	    osb->osb_ref_tree_lru->rf_blkno == rf_blkno)
+		tree = osb->osb_ref_tree_lru;
+	else
+		tree = ocfs2_find_refcount_tree(osb, rf_blkno);
+	if (tree)
+		goto out;
+
+	spin_unlock(&osb->osb_lock);
+
+	new = ocfs2_allocate_refcount_tree(osb, rf_blkno);
+	if (!new) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		return ret;
+	}
+	/*
+	 * We need the generation to create the refcount tree lock and since
+	 * it isn't changed during the tree modification, we are safe here to
+	 * read without protection.
+	 * We also have to purge the cache after we create the lock since the
+	 * refcount block may have the stale data. It can only be trusted when
+	 * we hold the refcount lock.
+	 */
+	ret = ocfs2_read_refcount_block(&new->rf_ci, rf_blkno, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		ocfs2_metadata_cache_exit(&new->rf_ci);
+		kfree(new);
+		return ret;
+	}
+
+	ref_rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	new->rf_generation = le32_to_cpu(ref_rb->rf_generation);
+	ocfs2_init_refcount_tree_lock(osb, new, rf_blkno,
+				      new->rf_generation);
+	ocfs2_metadata_cache_purge(&new->rf_ci);
+
+	spin_lock(&osb->osb_lock);
+	tree = ocfs2_find_refcount_tree(osb, rf_blkno);
+	if (tree)
+		goto out;
+
+	ocfs2_insert_refcount_tree(osb, new);
+
+	tree = new;
+	new = NULL;
+
+out:
+	*ret_tree = tree;
+
+	osb->osb_ref_tree_lru = tree;
+
+	spin_unlock(&osb->osb_lock);
+
+	if (new)
+		ocfs2_free_refcount_tree(new);
+
+	brelse(ref_root_bh);
+	return ret;
+}
+
+static int ocfs2_get_refcount_block(struct inode *inode, u64 *ref_blkno)
+{
+	int ret;
+	struct buffer_head *di_bh = NULL;
+	struct ocfs2_dinode *di;
+
+	ret = ocfs2_read_inode_block(inode, &di_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	BUG_ON(!(OCFS2_I(inode)->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL));
+
+	di = (struct ocfs2_dinode *)di_bh->b_data;
+	*ref_blkno = le64_to_cpu(di->i_refcount_loc);
+	brelse(di_bh);
+out:
+	return ret;
+}
+
+static int __ocfs2_lock_refcount_tree(struct ocfs2_super *osb,
+				      struct ocfs2_refcount_tree *tree, int rw)
+{
+	int ret;
+
+	ret = ocfs2_refcount_lock(tree, rw);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (rw)
+		down_write(&tree->rf_sem);
+	else
+		down_read(&tree->rf_sem);
+
+out:
+	return ret;
+}
+
+/*
+ * Lock the refcount tree pointed by ref_blkno and return the tree.
+ * In most case, we lock the tree and read the refcount block.
+ * So read it here if the caller really needs it.
+ *
+ * If the tree has been re-created by other node, it will free the
+ * old one and re-create it.
+ */
+int ocfs2_lock_refcount_tree(struct ocfs2_super *osb,
+			     u64 ref_blkno, int rw,
+			     struct ocfs2_refcount_tree **ret_tree,
+			     struct buffer_head **ref_bh)
+{
+	int ret, delete_tree = 0;
+	struct ocfs2_refcount_tree *tree = NULL;
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_refcount_block *rb;
+
+again:
+	ret = ocfs2_get_refcount_tree(osb, ref_blkno, &tree);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	ocfs2_refcount_tree_get(tree);
+
+	ret = __ocfs2_lock_refcount_tree(osb, tree, rw);
+	if (ret) {
+		mlog_errno(ret);
+		ocfs2_refcount_tree_put(tree);
+		goto out;
+	}
+
+	ret = ocfs2_read_refcount_block(&tree->rf_ci, tree->rf_blkno,
+					&ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		ocfs2_unlock_refcount_tree(osb, tree, rw);
+		ocfs2_refcount_tree_put(tree);
+		goto out;
+	}
+
+	rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	/*
+	 * If the refcount block has been freed and re-created, we may need
+	 * to recreate the refcount tree also.
+	 *
+	 * Here we just remove the tree from the rb-tree, and the last
+	 * kref holder will unlock and delete this refcount_tree.
+	 * Then we goto "again" and ocfs2_get_refcount_tree will create
+	 * the new refcount tree for us.
+	 */
+	if (tree->rf_generation != le32_to_cpu(rb->rf_generation)) {
+		if (!tree->rf_removed) {
+			ocfs2_erase_refcount_tree_from_list(osb, tree);
+			tree->rf_removed = 1;
+			delete_tree = 1;
+		}
+
+		ocfs2_unlock_refcount_tree(osb, tree, rw);
+		/*
+		 * We get an extra reference when we create the refcount
+		 * tree, so another put will destroy it.
+		 */
+		if (delete_tree)
+			ocfs2_refcount_tree_put(tree);
+		brelse(ref_root_bh);
+		ref_root_bh = NULL;
+		goto again;
+	}
+
+	*ret_tree = tree;
+	if (ref_bh) {
+		*ref_bh = ref_root_bh;
+		ref_root_bh = NULL;
+	}
+out:
+	brelse(ref_root_bh);
+	return ret;
+}
+
+int ocfs2_lock_refcount_tree_by_inode(struct inode *inode, int rw,
+				      struct ocfs2_refcount_tree **ret_tree,
+				      struct buffer_head **ref_bh)
+{
+	int ret;
+	u64 ref_blkno;
+
+	ret = ocfs2_get_refcount_block(inode, &ref_blkno);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	return ocfs2_lock_refcount_tree(OCFS2_SB(inode->i_sb), ref_blkno,
+					rw, ret_tree, ref_bh);
+}
+
+void ocfs2_unlock_refcount_tree(struct ocfs2_super *osb,
+				struct ocfs2_refcount_tree *tree, int rw)
+{
+	if (rw)
+		up_write(&tree->rf_sem);
+	else
+		up_read(&tree->rf_sem);
+
+	ocfs2_refcount_unlock(tree, rw);
+	ocfs2_refcount_tree_put(tree);
+}
+
+void ocfs2_purge_refcount_trees(struct ocfs2_super *osb)
+{
+	struct rb_node *node;
+	struct ocfs2_refcount_tree *tree;
+	struct rb_root *root = &osb->osb_rf_lock_tree;
+
+	while ((node = rb_last(root)) != NULL) {
+		tree = rb_entry(node, struct ocfs2_refcount_tree, rf_node);
+
+		mlog(0, "Purge tree %llu\n",
+		     (unsigned long long) tree->rf_blkno);
+
+		rb_erase(&tree->rf_node, root);
+		ocfs2_free_refcount_tree(tree);
+	}
+}
+
+/*
+ * Create a refcount tree for an inode.
+ * We take for granted that the inode is already locked.
+ */
+static int ocfs2_create_refcount_tree(struct inode *inode,
+				      struct buffer_head *di_bh)
+{
+	int ret;
+	handle_t *handle = NULL;
+	struct ocfs2_alloc_context *meta_ac = NULL;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct buffer_head *new_bh = NULL;
+	struct ocfs2_refcount_block *rb;
+	struct ocfs2_refcount_tree *new_tree = NULL, *tree = NULL;
+	u16 suballoc_bit_start;
+	u32 num_got;
+	u64 first_blkno;
+
+	BUG_ON(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL);
+
+	mlog(0, "create tree for inode %lu\n", inode->i_ino);
+
+	ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	handle = ocfs2_start_trans(osb, OCFS2_REFCOUNT_TREE_CREATE_CREDITS);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ret = ocfs2_claim_metadata(osb, handle, meta_ac, 1,
+				   &suballoc_bit_start, &num_got,
+				   &first_blkno);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	new_tree = ocfs2_allocate_refcount_tree(osb, first_blkno);
+	if (!new_tree) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	new_bh = sb_getblk(inode->i_sb, first_blkno);
+	ocfs2_set_new_buffer_uptodate(&new_tree->rf_ci, new_bh);
+
+	ret = ocfs2_journal_access_rb(handle, &new_tree->rf_ci, new_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	/* Initialize ocfs2_refcount_block. */
+	rb = (struct ocfs2_refcount_block *)new_bh->b_data;
+	memset(rb, 0, inode->i_sb->s_blocksize);
+	strcpy((void *)rb, OCFS2_REFCOUNT_BLOCK_SIGNATURE);
+	rb->rf_suballoc_slot = cpu_to_le16(osb->slot_num);
+	rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
+	rb->rf_fs_generation = cpu_to_le32(osb->fs_generation);
+	rb->rf_blkno = cpu_to_le64(first_blkno);
+	rb->rf_count = cpu_to_le32(1);
+	rb->rf_records.rl_count =
+			cpu_to_le16(ocfs2_refcount_recs_per_rb(osb->sb));
+	spin_lock(&osb->osb_lock);
+	rb->rf_generation = osb->s_next_generation++;
+	spin_unlock(&osb->osb_lock);
+
+	ocfs2_journal_dirty(handle, new_bh);
+
+	spin_lock(&oi->ip_lock);
+	oi->ip_dyn_features |= OCFS2_HAS_REFCOUNT_FL;
+	di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
+	di->i_refcount_loc = cpu_to_le64(first_blkno);
+	spin_unlock(&oi->ip_lock);
+
+	mlog(0, "created tree for inode %lu, refblock %llu\n",
+	     inode->i_ino, (unsigned long long)first_blkno);
+
+	ocfs2_journal_dirty(handle, di_bh);
+
+	/*
+	 * We have to init the tree lock here since it will use
+	 * the generation number to create it.
+	 */
+	new_tree->rf_generation = le32_to_cpu(rb->rf_generation);
+	ocfs2_init_refcount_tree_lock(osb, new_tree, first_blkno,
+				      new_tree->rf_generation);
+
+	spin_lock(&osb->osb_lock);
+	tree = ocfs2_find_refcount_tree(osb, first_blkno);
+
+	/*
+	 * We've just created a new refcount tree in this block.  If
+	 * we found a refcount tree on the ocfs2_super, it must be
+	 * one we just deleted.  We free the old tree before
+	 * inserting the new tree.
+	 */
+	BUG_ON(tree && tree->rf_generation == new_tree->rf_generation);
+	if (tree)
+		ocfs2_erase_refcount_tree_from_list_no_lock(osb, tree);
+	ocfs2_insert_refcount_tree(osb, new_tree);
+	spin_unlock(&osb->osb_lock);
+	new_tree = NULL;
+	if (tree)
+		ocfs2_refcount_tree_put(tree);
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+
+out:
+	if (new_tree) {
+		ocfs2_metadata_cache_exit(&new_tree->rf_ci);
+		kfree(new_tree);
+	}
+
+	brelse(new_bh);
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+
+	return ret;
+}
+
+static int ocfs2_set_refcount_tree(struct inode *inode,
+				   struct buffer_head *di_bh,
+				   u64 refcount_loc)
+{
+	int ret;
+	handle_t *handle = NULL;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_refcount_block *rb;
+	struct ocfs2_refcount_tree *ref_tree;
+
+	BUG_ON(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL);
+
+	ret = ocfs2_lock_refcount_tree(osb, refcount_loc, 1,
+				       &ref_tree, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	handle = ocfs2_start_trans(osb, OCFS2_REFCOUNT_TREE_SET_CREDITS);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ret = ocfs2_journal_access_rb(handle, &ref_tree->rf_ci, ref_root_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	le32_add_cpu(&rb->rf_count, 1);
+
+	ocfs2_journal_dirty(handle, ref_root_bh);
+
+	spin_lock(&oi->ip_lock);
+	oi->ip_dyn_features |= OCFS2_HAS_REFCOUNT_FL;
+	di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
+	di->i_refcount_loc = cpu_to_le64(refcount_loc);
+	spin_unlock(&oi->ip_lock);
+	ocfs2_journal_dirty(handle, di_bh);
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+out:
+	ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+	brelse(ref_root_bh);
+
+	return ret;
+}
+
+int ocfs2_remove_refcount_tree(struct inode *inode, struct buffer_head *di_bh)
+{
+	int ret, delete_tree = 0;
+	handle_t *handle = NULL;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_refcount_block *rb;
+	struct inode *alloc_inode = NULL;
+	struct buffer_head *alloc_bh = NULL;
+	struct buffer_head *blk_bh = NULL;
+	struct ocfs2_refcount_tree *ref_tree;
+	int credits = OCFS2_REFCOUNT_TREE_REMOVE_CREDITS;
+	u64 blk = 0, bg_blkno = 0, ref_blkno = le64_to_cpu(di->i_refcount_loc);
+	u16 bit = 0;
+
+	if (!(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL))
+		return 0;
+
+	BUG_ON(!ref_blkno);
+	ret = ocfs2_lock_refcount_tree(osb, ref_blkno, 1, &ref_tree, &blk_bh);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	rb = (struct ocfs2_refcount_block *)blk_bh->b_data;
+
+	/*
+	 * If we are the last user, we need to free the block.
+	 * So lock the allocator ahead.
+	 */
+	if (le32_to_cpu(rb->rf_count) == 1) {
+		blk = le64_to_cpu(rb->rf_blkno);
+		bit = le16_to_cpu(rb->rf_suballoc_bit);
+		bg_blkno = ocfs2_which_suballoc_group(blk, bit);
+
+		alloc_inode = ocfs2_get_system_file_inode(osb,
+					EXTENT_ALLOC_SYSTEM_INODE,
+					le16_to_cpu(rb->rf_suballoc_slot));
+		if (!alloc_inode) {
+			ret = -ENOMEM;
+			mlog_errno(ret);
+			goto out;
+		}
+		mutex_lock(&alloc_inode->i_mutex);
+
+		ret = ocfs2_inode_lock(alloc_inode, &alloc_bh, 1);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_mutex;
+		}
+
+		credits += OCFS2_SUBALLOC_FREE;
+	}
+
+	handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out_unlock;
+	}
+
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ret = ocfs2_journal_access_rb(handle, &ref_tree->rf_ci, blk_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	spin_lock(&oi->ip_lock);
+	oi->ip_dyn_features &= ~OCFS2_HAS_REFCOUNT_FL;
+	di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
+	di->i_refcount_loc = 0;
+	spin_unlock(&oi->ip_lock);
+	ocfs2_journal_dirty(handle, di_bh);
+
+	le32_add_cpu(&rb->rf_count , -1);
+	ocfs2_journal_dirty(handle, blk_bh);
+
+	if (!rb->rf_count) {
+		delete_tree = 1;
+		ocfs2_erase_refcount_tree_from_list(osb, ref_tree);
+		ret = ocfs2_free_suballoc_bits(handle, alloc_inode,
+					       alloc_bh, bit, bg_blkno, 1);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+out_unlock:
+	if (alloc_inode) {
+		ocfs2_inode_unlock(alloc_inode, 1);
+		brelse(alloc_bh);
+	}
+out_mutex:
+	if (alloc_inode) {
+		mutex_unlock(&alloc_inode->i_mutex);
+		iput(alloc_inode);
+	}
+out:
+	ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+	if (delete_tree)
+		ocfs2_refcount_tree_put(ref_tree);
+	brelse(blk_bh);
+
+	return ret;
+}
+
+static void ocfs2_find_refcount_rec_in_rl(struct ocfs2_caching_info *ci,
+					  struct buffer_head *ref_leaf_bh,
+					  u64 cpos, unsigned int len,
+					  struct ocfs2_refcount_rec *ret_rec,
+					  int *index)
+{
+	int i = 0;
+	struct ocfs2_refcount_block *rb =
+		(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_refcount_rec *rec = NULL;
+
+	for (; i < le16_to_cpu(rb->rf_records.rl_used); i++) {
+		rec = &rb->rf_records.rl_recs[i];
+
+		if (le64_to_cpu(rec->r_cpos) +
+		    le32_to_cpu(rec->r_clusters) <= cpos)
+			continue;
+		else if (le64_to_cpu(rec->r_cpos) > cpos)
+			break;
+
+		/* ok, cpos fail in this rec. Just return. */
+		if (ret_rec)
+			*ret_rec = *rec;
+		goto out;
+	}
+
+	if (ret_rec) {
+		/* We meet with a hole here, so fake the rec. */
+		ret_rec->r_cpos = cpu_to_le64(cpos);
+		ret_rec->r_refcount = 0;
+		if (i < le16_to_cpu(rb->rf_records.rl_used) &&
+		    le64_to_cpu(rec->r_cpos) < cpos + len)
+			ret_rec->r_clusters =
+				cpu_to_le32(le64_to_cpu(rec->r_cpos) - cpos);
+		else
+			ret_rec->r_clusters = cpu_to_le32(len);
+	}
+
+out:
+	*index = i;
+}
+
+/*
+ * Try to remove refcount tree. The mechanism is:
+ * 1) Check whether i_clusters == 0, if no, exit.
+ * 2) check whether we have i_xattr_loc in dinode. if yes, exit.
+ * 3) Check whether we have inline xattr stored outside, if yes, exit.
+ * 4) Remove the tree.
+ */
+int ocfs2_try_remove_refcount_tree(struct inode *inode,
+				   struct buffer_head *di_bh)
+{
+	int ret;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+
+	down_write(&oi->ip_xattr_sem);
+	down_write(&oi->ip_alloc_sem);
+
+	if (oi->ip_clusters)
+		goto out;
+
+	if ((oi->ip_dyn_features & OCFS2_HAS_XATTR_FL) && di->i_xattr_loc)
+		goto out;
+
+	if (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL &&
+	    ocfs2_has_inline_xattr_value_outside(inode, di))
+		goto out;
+
+	ret = ocfs2_remove_refcount_tree(inode, di_bh);
+	if (ret)
+		mlog_errno(ret);
+out:
+	up_write(&oi->ip_alloc_sem);
+	up_write(&oi->ip_xattr_sem);
+	return 0;
+}
+
+/*
+ * Given a cpos and len, try to find the refcount record which contains cpos.
+ * 1. If cpos can be found in one refcount record, return the record.
+ * 2. If cpos can't be found, return a fake record which start from cpos
+ *    and end at a small value between cpos+len and start of the next record.
+ *    This fake record has r_refcount = 0.
+ */
+static int ocfs2_get_refcount_rec(struct ocfs2_caching_info *ci,
+				  struct buffer_head *ref_root_bh,
+				  u64 cpos, unsigned int len,
+				  struct ocfs2_refcount_rec *ret_rec,
+				  int *index,
+				  struct buffer_head **ret_bh)
+{
+	int ret = 0, i, found;
+	u32 low_cpos;
+	struct ocfs2_extent_list *el;
+	struct ocfs2_extent_rec *tmp, *rec = NULL;
+	struct ocfs2_extent_block *eb;
+	struct buffer_head *eb_bh = NULL, *ref_leaf_bh = NULL;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_root_bh->b_data;
+
+	if (!(le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL)) {
+		ocfs2_find_refcount_rec_in_rl(ci, ref_root_bh, cpos, len,
+					      ret_rec, index);
+		*ret_bh = ref_root_bh;
+		get_bh(ref_root_bh);
+		return 0;
+	}
+
+	el = &rb->rf_list;
+	low_cpos = cpos & OCFS2_32BIT_POS_MASK;
+
+	if (el->l_tree_depth) {
+		ret = ocfs2_find_leaf(ci, el, low_cpos, &eb_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		eb = (struct ocfs2_extent_block *) eb_bh->b_data;
+		el = &eb->h_list;
+
+		if (el->l_tree_depth) {
+			ocfs2_error(sb,
+			"refcount tree %llu has non zero tree "
+			"depth in leaf btree tree block %llu\n",
+			(unsigned long long)ocfs2_metadata_cache_owner(ci),
+			(unsigned long long)eb_bh->b_blocknr);
+			ret = -EROFS;
+			goto out;
+		}
+	}
+
+	found = 0;
+	for (i = le16_to_cpu(el->l_next_free_rec) - 1; i >= 0; i--) {
+		rec = &el->l_recs[i];
+
+		if (le32_to_cpu(rec->e_cpos) <= low_cpos) {
+			found = 1;
+			break;
+		}
+	}
+
+	/* adjust len when we have ocfs2_extent_rec after it. */
+	if (found && i < le16_to_cpu(el->l_next_free_rec) - 1) {
+		tmp = &el->l_recs[i+1];
+
+		if (le32_to_cpu(tmp->e_cpos) < cpos + len)
+			len = le32_to_cpu(tmp->e_cpos) - cpos;
+	}
+
+	ret = ocfs2_read_refcount_block(ci, le64_to_cpu(rec->e_blkno),
+					&ref_leaf_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ocfs2_find_refcount_rec_in_rl(ci, ref_leaf_bh, cpos, len,
+				      ret_rec, index);
+	*ret_bh = ref_leaf_bh;
+out:
+	brelse(eb_bh);
+	return ret;
+}
+
+enum ocfs2_ref_rec_contig {
+	REF_CONTIG_NONE = 0,
+	REF_CONTIG_LEFT,
+	REF_CONTIG_RIGHT,
+	REF_CONTIG_LEFTRIGHT,
+};
+
+static enum ocfs2_ref_rec_contig
+	ocfs2_refcount_rec_adjacent(struct ocfs2_refcount_block *rb,
+				    int index)
+{
+	if ((rb->rf_records.rl_recs[index].r_refcount ==
+	    rb->rf_records.rl_recs[index + 1].r_refcount) &&
+	    (le64_to_cpu(rb->rf_records.rl_recs[index].r_cpos) +
+	    le32_to_cpu(rb->rf_records.rl_recs[index].r_clusters) ==
+	    le64_to_cpu(rb->rf_records.rl_recs[index + 1].r_cpos)))
+		return REF_CONTIG_RIGHT;
+
+	return REF_CONTIG_NONE;
+}
+
+static enum ocfs2_ref_rec_contig
+	ocfs2_refcount_rec_contig(struct ocfs2_refcount_block *rb,
+				  int index)
+{
+	enum ocfs2_ref_rec_contig ret = REF_CONTIG_NONE;
+
+	if (index < le16_to_cpu(rb->rf_records.rl_used) - 1)
+		ret = ocfs2_refcount_rec_adjacent(rb, index);
+
+	if (index > 0) {
+		enum ocfs2_ref_rec_contig tmp;
+
+		tmp = ocfs2_refcount_rec_adjacent(rb, index - 1);
+
+		if (tmp == REF_CONTIG_RIGHT) {
+			if (ret == REF_CONTIG_RIGHT)
+				ret = REF_CONTIG_LEFTRIGHT;
+			else
+				ret = REF_CONTIG_LEFT;
+		}
+	}
+
+	return ret;
+}
+
+static void ocfs2_rotate_refcount_rec_left(struct ocfs2_refcount_block *rb,
+					   int index)
+{
+	BUG_ON(rb->rf_records.rl_recs[index].r_refcount !=
+	       rb->rf_records.rl_recs[index+1].r_refcount);
+
+	le32_add_cpu(&rb->rf_records.rl_recs[index].r_clusters,
+		     le32_to_cpu(rb->rf_records.rl_recs[index+1].r_clusters));
+
+	if (index < le16_to_cpu(rb->rf_records.rl_used) - 2)
+		memmove(&rb->rf_records.rl_recs[index + 1],
+			&rb->rf_records.rl_recs[index + 2],
+			sizeof(struct ocfs2_refcount_rec) *
+			(le16_to_cpu(rb->rf_records.rl_used) - index - 2));
+
+	memset(&rb->rf_records.rl_recs[le16_to_cpu(rb->rf_records.rl_used) - 1],
+	       0, sizeof(struct ocfs2_refcount_rec));
+	le16_add_cpu(&rb->rf_records.rl_used, -1);
+}
+
+/*
+ * Merge the refcount rec if we are contiguous with the adjacent recs.
+ */
+static void ocfs2_refcount_rec_merge(struct ocfs2_refcount_block *rb,
+				     int index)
+{
+	enum ocfs2_ref_rec_contig contig =
+				ocfs2_refcount_rec_contig(rb, index);
+
+	if (contig == REF_CONTIG_NONE)
+		return;
+
+	if (contig == REF_CONTIG_LEFT || contig == REF_CONTIG_LEFTRIGHT) {
+		BUG_ON(index == 0);
+		index--;
+	}
+
+	ocfs2_rotate_refcount_rec_left(rb, index);
+
+	if (contig == REF_CONTIG_LEFTRIGHT)
+		ocfs2_rotate_refcount_rec_left(rb, index);
+}
+
+/*
+ * Change the refcount indexed by "index" in ref_bh.
+ * If refcount reaches 0, remove it.
+ */
+static int ocfs2_change_refcount_rec(handle_t *handle,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *ref_leaf_bh,
+				     int index, int merge, int change)
+{
+	int ret;
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_refcount_list *rl = &rb->rf_records;
+	struct ocfs2_refcount_rec *rec = &rl->rl_recs[index];
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_leaf_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mlog(0, "change index %d, old count %u, change %d\n", index,
+	     le32_to_cpu(rec->r_refcount), change);
+	le32_add_cpu(&rec->r_refcount, change);
+
+	if (!rec->r_refcount) {
+		if (index != le16_to_cpu(rl->rl_used) - 1) {
+			memmove(rec, rec + 1,
+				(le16_to_cpu(rl->rl_used) - index - 1) *
+				sizeof(struct ocfs2_refcount_rec));
+			memset(&rl->rl_recs[le16_to_cpu(rl->rl_used) - 1],
+			       0, sizeof(struct ocfs2_refcount_rec));
+		}
+
+		le16_add_cpu(&rl->rl_used, -1);
+	} else if (merge)
+		ocfs2_refcount_rec_merge(rb, index);
+
+	ret = ocfs2_journal_dirty(handle, ref_leaf_bh);
+	if (ret)
+		mlog_errno(ret);
+out:
+	return ret;
+}
+
+static int ocfs2_expand_inline_ref_root(handle_t *handle,
+					struct ocfs2_caching_info *ci,
+					struct buffer_head *ref_root_bh,
+					struct buffer_head **ref_leaf_bh,
+					struct ocfs2_alloc_context *meta_ac)
+{
+	int ret;
+	u16 suballoc_bit_start;
+	u32 num_got;
+	u64 blkno;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	struct buffer_head *new_bh = NULL;
+	struct ocfs2_refcount_block *new_rb;
+	struct ocfs2_refcount_block *root_rb =
+			(struct ocfs2_refcount_block *)ref_root_bh->b_data;
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_root_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_claim_metadata(OCFS2_SB(sb), handle, meta_ac, 1,
+				   &suballoc_bit_start, &num_got,
+				   &blkno);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	new_bh = sb_getblk(sb, blkno);
+	if (new_bh == NULL) {
+		ret = -EIO;
+		mlog_errno(ret);
+		goto out;
+	}
+	ocfs2_set_new_buffer_uptodate(ci, new_bh);
+
+	ret = ocfs2_journal_access_rb(handle, ci, new_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/*
+	 * Initialize ocfs2_refcount_block.
+	 * It should contain the same information as the old root.
+	 * so just memcpy it and change the corresponding field.
+	 */
+	memcpy(new_bh->b_data, ref_root_bh->b_data, sb->s_blocksize);
+
+	new_rb = (struct ocfs2_refcount_block *)new_bh->b_data;
+	new_rb->rf_suballoc_slot = cpu_to_le16(OCFS2_SB(sb)->slot_num);
+	new_rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
+	new_rb->rf_blkno = cpu_to_le64(blkno);
+	new_rb->rf_cpos = cpu_to_le32(0);
+	new_rb->rf_parent = cpu_to_le64(ref_root_bh->b_blocknr);
+	new_rb->rf_flags = cpu_to_le32(OCFS2_REFCOUNT_LEAF_FL);
+	ocfs2_journal_dirty(handle, new_bh);
+
+	/* Now change the root. */
+	memset(&root_rb->rf_list, 0, sb->s_blocksize -
+	       offsetof(struct ocfs2_refcount_block, rf_list));
+	root_rb->rf_list.l_count = cpu_to_le16(ocfs2_extent_recs_per_rb(sb));
+	root_rb->rf_clusters = cpu_to_le32(1);
+	root_rb->rf_list.l_next_free_rec = cpu_to_le16(1);
+	root_rb->rf_list.l_recs[0].e_blkno = cpu_to_le64(blkno);
+	root_rb->rf_list.l_recs[0].e_leaf_clusters = cpu_to_le16(1);
+	root_rb->rf_flags = cpu_to_le32(OCFS2_REFCOUNT_TREE_FL);
+
+	ocfs2_journal_dirty(handle, ref_root_bh);
+
+	mlog(0, "new leaf block %llu, used %u\n", (unsigned long long)blkno,
+	     le16_to_cpu(new_rb->rf_records.rl_used));
+
+	*ref_leaf_bh = new_bh;
+	new_bh = NULL;
+out:
+	brelse(new_bh);
+	return ret;
+}
+
+static int ocfs2_refcount_rec_no_intersect(struct ocfs2_refcount_rec *prev,
+					   struct ocfs2_refcount_rec *next)
+{
+	if (ocfs2_get_ref_rec_low_cpos(prev) + le32_to_cpu(prev->r_clusters) <=
+		ocfs2_get_ref_rec_low_cpos(next))
+		return 1;
+
+	return 0;
+}
+
+static int cmp_refcount_rec_by_low_cpos(const void *a, const void *b)
+{
+	const struct ocfs2_refcount_rec *l = a, *r = b;
+	u32 l_cpos = ocfs2_get_ref_rec_low_cpos(l);
+	u32 r_cpos = ocfs2_get_ref_rec_low_cpos(r);
+
+	if (l_cpos > r_cpos)
+		return 1;
+	if (l_cpos < r_cpos)
+		return -1;
+	return 0;
+}
+
+static int cmp_refcount_rec_by_cpos(const void *a, const void *b)
+{
+	const struct ocfs2_refcount_rec *l = a, *r = b;
+	u64 l_cpos = le64_to_cpu(l->r_cpos);
+	u64 r_cpos = le64_to_cpu(r->r_cpos);
+
+	if (l_cpos > r_cpos)
+		return 1;
+	if (l_cpos < r_cpos)
+		return -1;
+	return 0;
+}
+
+static void swap_refcount_rec(void *a, void *b, int size)
+{
+	struct ocfs2_refcount_rec *l = a, *r = b, tmp;
+
+	tmp = *(struct ocfs2_refcount_rec *)l;
+	*(struct ocfs2_refcount_rec *)l =
+			*(struct ocfs2_refcount_rec *)r;
+	*(struct ocfs2_refcount_rec *)r = tmp;
+}
+
+/*
+ * The refcount cpos are ordered by their 64bit cpos,
+ * But we will use the low 32 bit to be the e_cpos in the b-tree.
+ * So we need to make sure that this pos isn't intersected with others.
+ *
+ * Note: The refcount block is already sorted by their low 32 bit cpos,
+ *       So just try the middle pos first, and we will exit when we find
+ *       the good position.
+ */
+static int ocfs2_find_refcount_split_pos(struct ocfs2_refcount_list *rl,
+					 u32 *split_pos, int *split_index)
+{
+	int num_used = le16_to_cpu(rl->rl_used);
+	int delta, middle = num_used / 2;
+
+	for (delta = 0; delta < middle; delta++) {
+		/* Let's check delta earlier than middle */
+		if (ocfs2_refcount_rec_no_intersect(
+					&rl->rl_recs[middle - delta - 1],
+					&rl->rl_recs[middle - delta])) {
+			*split_index = middle - delta;
+			break;
+		}
+
+		/* For even counts, don't walk off the end */
+		if ((middle + delta + 1) == num_used)
+			continue;
+
+		/* Now try delta past middle */
+		if (ocfs2_refcount_rec_no_intersect(
+					&rl->rl_recs[middle + delta],
+					&rl->rl_recs[middle + delta + 1])) {
+			*split_index = middle + delta + 1;
+			break;
+		}
+	}
+
+	if (delta >= middle)
+		return -ENOSPC;
+
+	*split_pos = ocfs2_get_ref_rec_low_cpos(&rl->rl_recs[*split_index]);
+	return 0;
+}
+
+static int ocfs2_divide_leaf_refcount_block(struct buffer_head *ref_leaf_bh,
+					    struct buffer_head *new_bh,
+					    u32 *split_cpos)
+{
+	int split_index = 0, num_moved, ret;
+	u32 cpos = 0;
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_refcount_list *rl = &rb->rf_records;
+	struct ocfs2_refcount_block *new_rb =
+			(struct ocfs2_refcount_block *)new_bh->b_data;
+	struct ocfs2_refcount_list *new_rl = &new_rb->rf_records;
+
+	mlog(0, "split old leaf refcount block %llu, count = %u, used = %u\n",
+	     (unsigned long long)ref_leaf_bh->b_blocknr,
+	     le32_to_cpu(rl->rl_count), le32_to_cpu(rl->rl_used));
+
+	/*
+	 * XXX: Improvement later.
+	 * If we know all the high 32 bit cpos is the same, no need to sort.
+	 *
+	 * In order to make the whole process safe, we do:
+	 * 1. sort the entries by their low 32 bit cpos first so that we can
+	 *    find the split cpos easily.
+	 * 2. call ocfs2_insert_extent to insert the new refcount block.
+	 * 3. move the refcount rec to the new block.
+	 * 4. sort the entries by their 64 bit cpos.
+	 * 5. dirty the new_rb and rb.
+	 */
+	sort(&rl->rl_recs, le16_to_cpu(rl->rl_used),
+	     sizeof(struct ocfs2_refcount_rec),
+	     cmp_refcount_rec_by_low_cpos, swap_refcount_rec);
+
+	ret = ocfs2_find_refcount_split_pos(rl, &cpos, &split_index);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	new_rb->rf_cpos = cpu_to_le32(cpos);
+
+	/* move refcount records starting from split_index to the new block. */
+	num_moved = le16_to_cpu(rl->rl_used) - split_index;
+	memcpy(new_rl->rl_recs, &rl->rl_recs[split_index],
+	       num_moved * sizeof(struct ocfs2_refcount_rec));
+
+	/*ok, remove the entries we just moved over to the other block. */
+	memset(&rl->rl_recs[split_index], 0,
+	       num_moved * sizeof(struct ocfs2_refcount_rec));
+
+	/* change old and new rl_used accordingly. */
+	le16_add_cpu(&rl->rl_used, -num_moved);
+	new_rl->rl_used = cpu_to_le32(num_moved);
+
+	sort(&rl->rl_recs, le16_to_cpu(rl->rl_used),
+	     sizeof(struct ocfs2_refcount_rec),
+	     cmp_refcount_rec_by_cpos, swap_refcount_rec);
+
+	sort(&new_rl->rl_recs, le16_to_cpu(new_rl->rl_used),
+	     sizeof(struct ocfs2_refcount_rec),
+	     cmp_refcount_rec_by_cpos, swap_refcount_rec);
+
+	*split_cpos = cpos;
+	return 0;
+}
+
+static int ocfs2_new_leaf_refcount_block(handle_t *handle,
+					 struct ocfs2_caching_info *ci,
+					 struct buffer_head *ref_root_bh,
+					 struct buffer_head *ref_leaf_bh,
+					 struct ocfs2_alloc_context *meta_ac)
+{
+	int ret;
+	u16 suballoc_bit_start;
+	u32 num_got, new_cpos;
+	u64 blkno;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	struct ocfs2_refcount_block *root_rb =
+			(struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	struct buffer_head *new_bh = NULL;
+	struct ocfs2_refcount_block *new_rb;
+	struct ocfs2_extent_tree ref_et;
+
+	BUG_ON(!(le32_to_cpu(root_rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL));
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_root_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_leaf_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_claim_metadata(OCFS2_SB(sb), handle, meta_ac, 1,
+				   &suballoc_bit_start, &num_got,
+				   &blkno);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	new_bh = sb_getblk(sb, blkno);
+	if (new_bh == NULL) {
+		ret = -EIO;
+		mlog_errno(ret);
+		goto out;
+	}
+	ocfs2_set_new_buffer_uptodate(ci, new_bh);
+
+	ret = ocfs2_journal_access_rb(handle, ci, new_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/* Initialize ocfs2_refcount_block. */
+	new_rb = (struct ocfs2_refcount_block *)new_bh->b_data;
+	memset(new_rb, 0, sb->s_blocksize);
+	strcpy((void *)new_rb, OCFS2_REFCOUNT_BLOCK_SIGNATURE);
+	new_rb->rf_suballoc_slot = cpu_to_le16(OCFS2_SB(sb)->slot_num);
+	new_rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
+	new_rb->rf_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+	new_rb->rf_blkno = cpu_to_le64(blkno);
+	new_rb->rf_parent = cpu_to_le64(ref_root_bh->b_blocknr);
+	new_rb->rf_flags = cpu_to_le32(OCFS2_REFCOUNT_LEAF_FL);
+	new_rb->rf_records.rl_count =
+				cpu_to_le16(ocfs2_refcount_recs_per_rb(sb));
+	new_rb->rf_generation = root_rb->rf_generation;
+
+	ret = ocfs2_divide_leaf_refcount_block(ref_leaf_bh, new_bh, &new_cpos);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ocfs2_journal_dirty(handle, ref_leaf_bh);
+	ocfs2_journal_dirty(handle, new_bh);
+
+	ocfs2_init_refcount_extent_tree(&ref_et, ci, ref_root_bh);
+
+	mlog(0, "insert new leaf block %llu at %u\n",
+	     (unsigned long long)new_bh->b_blocknr, new_cpos);
+
+	/* Insert the new leaf block with the specific offset cpos. */
+	ret = ocfs2_insert_extent(handle, &ref_et, new_cpos, new_bh->b_blocknr,
+				  1, 0, meta_ac);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	brelse(new_bh);
+	return ret;
+}
+
+static int ocfs2_expand_refcount_tree(handle_t *handle,
+				      struct ocfs2_caching_info *ci,
+				      struct buffer_head *ref_root_bh,
+				      struct buffer_head *ref_leaf_bh,
+				      struct ocfs2_alloc_context *meta_ac)
+{
+	int ret;
+	struct buffer_head *expand_bh = NULL;
+
+	if (ref_root_bh == ref_leaf_bh) {
+		/*
+		 * the old root bh hasn't been expanded to a b-tree,
+		 * so expand it first.
+		 */
+		ret = ocfs2_expand_inline_ref_root(handle, ci, ref_root_bh,
+						   &expand_bh, meta_ac);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+	} else {
+		expand_bh = ref_leaf_bh;
+		get_bh(expand_bh);
+	}
+
+
+	/* Now add a new refcount block into the tree.*/
+	ret = ocfs2_new_leaf_refcount_block(handle, ci, ref_root_bh,
+					    expand_bh, meta_ac);
+	if (ret)
+		mlog_errno(ret);
+out:
+	brelse(expand_bh);
+	return ret;
+}
+
+/*
+ * Adjust the extent rec in b-tree representing ref_leaf_bh.
+ *
+ * Only called when we have inserted a new refcount rec at index 0
+ * which means ocfs2_extent_rec.e_cpos may need some change.
+ */
+static int ocfs2_adjust_refcount_rec(handle_t *handle,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *ref_root_bh,
+				     struct buffer_head *ref_leaf_bh,
+				     struct ocfs2_refcount_rec *rec)
+{
+	int ret = 0, i;
+	u32 new_cpos, old_cpos;
+	struct ocfs2_path *path = NULL;
+	struct ocfs2_extent_tree et;
+	struct ocfs2_refcount_block *rb =
+		(struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	struct ocfs2_extent_list *el;
+
+	if (!(le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL))
+		goto out;
+
+	rb = (struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	old_cpos = le32_to_cpu(rb->rf_cpos);
+	new_cpos = le64_to_cpu(rec->r_cpos) & OCFS2_32BIT_POS_MASK;
+	if (old_cpos <= new_cpos)
+		goto out;
+
+	ocfs2_init_refcount_extent_tree(&et, ci, ref_root_bh);
+
+	path = ocfs2_new_path_from_et(&et);
+	if (!path) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_find_path(ci, path, old_cpos);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/*
+	 * 2 more credits, one for the leaf refcount block, one for
+	 * the extent block contains the extent rec.
+	 */
+	ret = ocfs2_extend_trans(handle, handle->h_buffer_credits + 2);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_leaf_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_eb(handle, ci, path_leaf_bh(path),
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/* change the leaf extent block first. */
+	el = path_leaf_el(path);
+
+	for (i = 0; i < le16_to_cpu(el->l_next_free_rec); i++)
+		if (le32_to_cpu(el->l_recs[i].e_cpos) == old_cpos)
+			break;
+
+	BUG_ON(i == le16_to_cpu(el->l_next_free_rec));
+
+	el->l_recs[i].e_cpos = cpu_to_le32(new_cpos);
+
+	/* change the r_cpos in the leaf block. */
+	rb->rf_cpos = cpu_to_le32(new_cpos);
+
+	ocfs2_journal_dirty(handle, path_leaf_bh(path));
+	ocfs2_journal_dirty(handle, ref_leaf_bh);
+
+out:
+	ocfs2_free_path(path);
+	return ret;
+}
+
+static int ocfs2_insert_refcount_rec(handle_t *handle,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *ref_root_bh,
+				     struct buffer_head *ref_leaf_bh,
+				     struct ocfs2_refcount_rec *rec,
+				     int index, int merge,
+				     struct ocfs2_alloc_context *meta_ac)
+{
+	int ret;
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_refcount_list *rf_list = &rb->rf_records;
+	struct buffer_head *new_bh = NULL;
+
+	BUG_ON(le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL);
+
+	if (rf_list->rl_used == rf_list->rl_count) {
+		u64 cpos = le64_to_cpu(rec->r_cpos);
+		u32 len = le32_to_cpu(rec->r_clusters);
+
+		ret = ocfs2_expand_refcount_tree(handle, ci, ref_root_bh,
+						 ref_leaf_bh, meta_ac);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
+					     cpos, len, NULL, &index,
+					     &new_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		ref_leaf_bh = new_bh;
+		rb = (struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+		rf_list = &rb->rf_records;
+	}
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_leaf_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (index < le16_to_cpu(rf_list->rl_used))
+		memmove(&rf_list->rl_recs[index + 1],
+			&rf_list->rl_recs[index],
+			(le16_to_cpu(rf_list->rl_used) - index) *
+			 sizeof(struct ocfs2_refcount_rec));
+
+	mlog(0, "insert refcount record start %llu, len %u, count %u "
+	     "to leaf block %llu at index %d\n",
+	     (unsigned long long)le64_to_cpu(rec->r_cpos),
+	     le32_to_cpu(rec->r_clusters), le32_to_cpu(rec->r_refcount),
+	     (unsigned long long)ref_leaf_bh->b_blocknr, index);
+
+	rf_list->rl_recs[index] = *rec;
+
+	le16_add_cpu(&rf_list->rl_used, 1);
+
+	if (merge)
+		ocfs2_refcount_rec_merge(rb, index);
+
+	ret = ocfs2_journal_dirty(handle, ref_leaf_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (index == 0) {
+		ret = ocfs2_adjust_refcount_rec(handle, ci,
+						ref_root_bh,
+						ref_leaf_bh, rec);
+		if (ret)
+			mlog_errno(ret);
+	}
+out:
+	brelse(new_bh);
+	return ret;
+}
+
+/*
+ * Split the refcount_rec indexed by "index" in ref_leaf_bh.
+ * This is much simple than our b-tree code.
+ * split_rec is the new refcount rec we want to insert.
+ * If split_rec->r_refcount > 0, we are changing the refcount(in case we
+ * increase refcount or decrease a refcount to non-zero).
+ * If split_rec->r_refcount == 0, we are punching a hole in current refcount
+ * rec( in case we decrease a refcount to zero).
+ */
+static int ocfs2_split_refcount_rec(handle_t *handle,
+				    struct ocfs2_caching_info *ci,
+				    struct buffer_head *ref_root_bh,
+				    struct buffer_head *ref_leaf_bh,
+				    struct ocfs2_refcount_rec *split_rec,
+				    int index, int merge,
+				    struct ocfs2_alloc_context *meta_ac,
+				    struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret, recs_need;
+	u32 len;
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_refcount_list *rf_list = &rb->rf_records;
+	struct ocfs2_refcount_rec *orig_rec = &rf_list->rl_recs[index];
+	struct ocfs2_refcount_rec *tail_rec = NULL;
+	struct buffer_head *new_bh = NULL;
+
+	BUG_ON(le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL);
+
+	mlog(0, "original r_pos %llu, cluster %u, split %llu, cluster %u\n",
+	     le64_to_cpu(orig_rec->r_cpos), le32_to_cpu(orig_rec->r_clusters),
+	     le64_to_cpu(split_rec->r_cpos),
+	     le32_to_cpu(split_rec->r_clusters));
+
+	/*
+	 * If we just need to split the header or tail clusters,
+	 * no more recs are needed, just split is OK.
+	 * Otherwise we at least need one new recs.
+	 */
+	if (!split_rec->r_refcount &&
+	    (split_rec->r_cpos == orig_rec->r_cpos ||
+	     le64_to_cpu(split_rec->r_cpos) +
+	     le32_to_cpu(split_rec->r_clusters) ==
+	     le64_to_cpu(orig_rec->r_cpos) + le32_to_cpu(orig_rec->r_clusters)))
+		recs_need = 0;
+	else
+		recs_need = 1;
+
+	/*
+	 * We need one more rec if we split in the middle and the new rec have
+	 * some refcount in it.
+	 */
+	if (split_rec->r_refcount &&
+	    (split_rec->r_cpos != orig_rec->r_cpos &&
+	     le64_to_cpu(split_rec->r_cpos) +
+	     le32_to_cpu(split_rec->r_clusters) !=
+	     le64_to_cpu(orig_rec->r_cpos) + le32_to_cpu(orig_rec->r_clusters)))
+		recs_need++;
+
+	/* If the leaf block don't have enough record, expand it. */
+	if (le16_to_cpu(rf_list->rl_used) + recs_need > rf_list->rl_count) {
+		struct ocfs2_refcount_rec tmp_rec;
+		u64 cpos = le64_to_cpu(orig_rec->r_cpos);
+		len = le32_to_cpu(orig_rec->r_clusters);
+		ret = ocfs2_expand_refcount_tree(handle, ci, ref_root_bh,
+						 ref_leaf_bh, meta_ac);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		/*
+		 * We have to re-get it since now cpos may be moved to
+		 * another leaf block.
+		 */
+		ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
+					     cpos, len, &tmp_rec, &index,
+					     &new_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		ref_leaf_bh = new_bh;
+		rb = (struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+		rf_list = &rb->rf_records;
+		orig_rec = &rf_list->rl_recs[index];
+	}
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_leaf_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/*
+	 * We have calculated out how many new records we need and store
+	 * in recs_need, so spare enough space first by moving the records
+	 * after "index" to the end.
+	 */
+	if (index != le16_to_cpu(rf_list->rl_used) - 1)
+		memmove(&rf_list->rl_recs[index + 1 + recs_need],
+			&rf_list->rl_recs[index + 1],
+			(le16_to_cpu(rf_list->rl_used) - index - 1) *
+			 sizeof(struct ocfs2_refcount_rec));
+
+	len = (le64_to_cpu(orig_rec->r_cpos) +
+	      le32_to_cpu(orig_rec->r_clusters)) -
+	      (le64_to_cpu(split_rec->r_cpos) +
+	      le32_to_cpu(split_rec->r_clusters));
+
+	/*
+	 * If we have "len", the we will split in the tail and move it
+	 * to the end of the space we have just spared.
+	 */
+	if (len) {
+		tail_rec = &rf_list->rl_recs[index + recs_need];
+
+		memcpy(tail_rec, orig_rec, sizeof(struct ocfs2_refcount_rec));
+		le64_add_cpu(&tail_rec->r_cpos,
+			     le32_to_cpu(tail_rec->r_clusters) - len);
+		tail_rec->r_clusters = le32_to_cpu(len);
+	}
+
+	/*
+	 * If the split pos isn't the same as the original one, we need to
+	 * split in the head.
+	 *
+	 * Note: We have the chance that split_rec.r_refcount = 0,
+	 * recs_need = 0 and len > 0, which means we just cut the head from
+	 * the orig_rec and in that case we have done some modification in
+	 * orig_rec above, so the check for r_cpos is faked.
+	 */
+	if (split_rec->r_cpos != orig_rec->r_cpos && tail_rec != orig_rec) {
+		len = le64_to_cpu(split_rec->r_cpos) -
+		      le64_to_cpu(orig_rec->r_cpos);
+		orig_rec->r_clusters = cpu_to_le32(len);
+		index++;
+	}
+
+	le16_add_cpu(&rf_list->rl_used, recs_need);
+
+	if (split_rec->r_refcount) {
+		rf_list->rl_recs[index] = *split_rec;
+		mlog(0, "insert refcount record start %llu, len %u, count %u "
+		     "to leaf block %llu at index %d\n",
+		     (unsigned long long)le64_to_cpu(split_rec->r_cpos),
+		     le32_to_cpu(split_rec->r_clusters),
+		     le32_to_cpu(split_rec->r_refcount),
+		     (unsigned long long)ref_leaf_bh->b_blocknr, index);
+
+		if (merge)
+			ocfs2_refcount_rec_merge(rb, index);
+	}
+
+	ret = ocfs2_journal_dirty(handle, ref_leaf_bh);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	brelse(new_bh);
+	return ret;
+}
+
+static int __ocfs2_increase_refcount(handle_t *handle,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *ref_root_bh,
+				     u64 cpos, u32 len, int merge,
+				     struct ocfs2_alloc_context *meta_ac,
+				     struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret = 0, index;
+	struct buffer_head *ref_leaf_bh = NULL;
+	struct ocfs2_refcount_rec rec;
+	unsigned int set_len = 0;
+
+	mlog(0, "Tree owner %llu, add refcount start %llu, len %u\n",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
+	     (unsigned long long)cpos, len);
+
+	while (len) {
+		ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
+					     cpos, len, &rec, &index,
+					     &ref_leaf_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		set_len = le32_to_cpu(rec.r_clusters);
+
+		/*
+		 * Here we may meet with 3 situations:
+		 *
+		 * 1. If we find an already existing record, and the length
+		 *    is the same, cool, we just need to increase the r_refcount
+		 *    and it is OK.
+		 * 2. If we find a hole, just insert it with r_refcount = 1.
+		 * 3. If we are in the middle of one extent record, split
+		 *    it.
+		 */
+		if (rec.r_refcount && le64_to_cpu(rec.r_cpos) == cpos &&
+		    set_len <= len) {
+			mlog(0, "increase refcount rec, start %llu, len %u, "
+			     "count %u\n", (unsigned long long)cpos, set_len,
+			     le32_to_cpu(rec.r_refcount));
+			ret = ocfs2_change_refcount_rec(handle, ci,
+							ref_leaf_bh, index,
+							merge, 1);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+		} else if (!rec.r_refcount) {
+			rec.r_refcount = cpu_to_le32(1);
+
+			mlog(0, "insert refcount rec, start %llu, len %u\n",
+			     (unsigned long long)le64_to_cpu(rec.r_cpos),
+			     set_len);
+			ret = ocfs2_insert_refcount_rec(handle, ci, ref_root_bh,
+							ref_leaf_bh,
+							&rec, index,
+							merge, meta_ac);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+		} else  {
+			set_len = min((u64)(cpos + len),
+				      le64_to_cpu(rec.r_cpos) + set_len) - cpos;
+			rec.r_cpos = cpu_to_le64(cpos);
+			rec.r_clusters = cpu_to_le32(set_len);
+			le32_add_cpu(&rec.r_refcount, 1);
+
+			mlog(0, "split refcount rec, start %llu, "
+			     "len %u, count %u\n",
+			     (unsigned long long)le64_to_cpu(rec.r_cpos),
+			     set_len, le32_to_cpu(rec.r_refcount));
+			ret = ocfs2_split_refcount_rec(handle, ci,
+						       ref_root_bh, ref_leaf_bh,
+						       &rec, index, merge,
+						       meta_ac, dealloc);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+		}
+
+		cpos += set_len;
+		len -= set_len;
+		brelse(ref_leaf_bh);
+		ref_leaf_bh = NULL;
+	}
+
+out:
+	brelse(ref_leaf_bh);
+	return ret;
+}
+
+static int ocfs2_remove_refcount_extent(handle_t *handle,
+				struct ocfs2_caching_info *ci,
+				struct buffer_head *ref_root_bh,
+				struct buffer_head *ref_leaf_bh,
+				struct ocfs2_alloc_context *meta_ac,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_extent_tree et;
+
+	BUG_ON(rb->rf_records.rl_used);
+
+	ocfs2_init_refcount_extent_tree(&et, ci, ref_root_bh);
+	ret = ocfs2_remove_extent(handle, &et, le32_to_cpu(rb->rf_cpos),
+				  1, meta_ac, dealloc);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ocfs2_remove_from_cache(ci, ref_leaf_bh);
+
+	/*
+	 * add the freed block to the dealloc so that it will be freed
+	 * when we run dealloc.
+	 */
+	ret = ocfs2_cache_block_dealloc(dealloc, EXTENT_ALLOC_SYSTEM_INODE,
+					le16_to_cpu(rb->rf_suballoc_slot),
+					le64_to_cpu(rb->rf_blkno),
+					le16_to_cpu(rb->rf_suballoc_bit));
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_rb(handle, ci, ref_root_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+
+	le32_add_cpu(&rb->rf_clusters, -1);
+
+	/*
+	 * check whether we need to restore the root refcount block if
+	 * there is no leaf extent block at atll.
+	 */
+	if (!rb->rf_list.l_next_free_rec) {
+		BUG_ON(rb->rf_clusters);
+
+		mlog(0, "reset refcount tree root %llu to be a record block.\n",
+		     (unsigned long long)ref_root_bh->b_blocknr);
+
+		rb->rf_flags = 0;
+		rb->rf_parent = 0;
+		rb->rf_cpos = 0;
+		memset(&rb->rf_records, 0, sb->s_blocksize -
+		       offsetof(struct ocfs2_refcount_block, rf_records));
+		rb->rf_records.rl_count =
+				cpu_to_le16(ocfs2_refcount_recs_per_rb(sb));
+	}
+
+	ocfs2_journal_dirty(handle, ref_root_bh);
+
+out:
+	return ret;
+}
+
+int ocfs2_increase_refcount(handle_t *handle,
+			    struct ocfs2_caching_info *ci,
+			    struct buffer_head *ref_root_bh,
+			    u64 cpos, u32 len,
+			    struct ocfs2_alloc_context *meta_ac,
+			    struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	return __ocfs2_increase_refcount(handle, ci, ref_root_bh,
+					 cpos, len, 1,
+					 meta_ac, dealloc);
+}
+
+static int ocfs2_decrease_refcount_rec(handle_t *handle,
+				struct ocfs2_caching_info *ci,
+				struct buffer_head *ref_root_bh,
+				struct buffer_head *ref_leaf_bh,
+				int index, u64 cpos, unsigned int len,
+				struct ocfs2_alloc_context *meta_ac,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret;
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+	struct ocfs2_refcount_rec *rec = &rb->rf_records.rl_recs[index];
+
+	BUG_ON(cpos < le64_to_cpu(rec->r_cpos));
+	BUG_ON(cpos + len >
+	       le64_to_cpu(rec->r_cpos) + le32_to_cpu(rec->r_clusters));
+
+	if (cpos == le64_to_cpu(rec->r_cpos) &&
+	    len == le32_to_cpu(rec->r_clusters))
+		ret = ocfs2_change_refcount_rec(handle, ci,
+						ref_leaf_bh, index, 1, -1);
+	else {
+		struct ocfs2_refcount_rec split = *rec;
+		split.r_cpos = cpu_to_le64(cpos);
+		split.r_clusters = cpu_to_le32(len);
+
+		le32_add_cpu(&split.r_refcount, -1);
+
+		mlog(0, "split refcount rec, start %llu, "
+		     "len %u, count %u, original start %llu, len %u\n",
+		     (unsigned long long)le64_to_cpu(split.r_cpos),
+		     len, le32_to_cpu(split.r_refcount),
+		     (unsigned long long)le64_to_cpu(rec->r_cpos),
+		     le32_to_cpu(rec->r_clusters));
+		ret = ocfs2_split_refcount_rec(handle, ci,
+					       ref_root_bh, ref_leaf_bh,
+					       &split, index, 1,
+					       meta_ac, dealloc);
+	}
+
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/* Remove the leaf refcount block if it contains no refcount record. */
+	if (!rb->rf_records.rl_used && ref_leaf_bh != ref_root_bh) {
+		ret = ocfs2_remove_refcount_extent(handle, ci, ref_root_bh,
+						   ref_leaf_bh, meta_ac,
+						   dealloc);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+out:
+	return ret;
+}
+
+static int __ocfs2_decrease_refcount(handle_t *handle,
+				     struct ocfs2_caching_info *ci,
+				     struct buffer_head *ref_root_bh,
+				     u64 cpos, u32 len,
+				     struct ocfs2_alloc_context *meta_ac,
+				     struct ocfs2_cached_dealloc_ctxt *dealloc,
+				     int delete)
+{
+	int ret = 0, index = 0;
+	struct ocfs2_refcount_rec rec;
+	unsigned int r_count = 0, r_len;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	struct buffer_head *ref_leaf_bh = NULL;
+
+	mlog(0, "Tree owner %llu, decrease refcount start %llu, "
+	     "len %u, delete %u\n",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
+	     (unsigned long long)cpos, len, delete);
+
+	while (len) {
+		ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
+					     cpos, len, &rec, &index,
+					     &ref_leaf_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		r_count = le32_to_cpu(rec.r_refcount);
+		BUG_ON(r_count == 0);
+		if (!delete)
+			BUG_ON(r_count > 1);
+
+		r_len = min((u64)(cpos + len), le64_to_cpu(rec.r_cpos) +
+			      le32_to_cpu(rec.r_clusters)) - cpos;
+
+		ret = ocfs2_decrease_refcount_rec(handle, ci, ref_root_bh,
+						  ref_leaf_bh, index,
+						  cpos, r_len,
+						  meta_ac, dealloc);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		if (le32_to_cpu(rec.r_refcount) == 1 && delete) {
+			ret = ocfs2_cache_cluster_dealloc(dealloc,
+					  ocfs2_clusters_to_blocks(sb, cpos),
+							  r_len);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+		}
+
+		cpos += r_len;
+		len -= r_len;
+		brelse(ref_leaf_bh);
+		ref_leaf_bh = NULL;
+	}
+
+out:
+	brelse(ref_leaf_bh);
+	return ret;
+}
+
+/* Caller must hold refcount tree lock. */
+int ocfs2_decrease_refcount(struct inode *inode,
+			    handle_t *handle, u32 cpos, u32 len,
+			    struct ocfs2_alloc_context *meta_ac,
+			    struct ocfs2_cached_dealloc_ctxt *dealloc,
+			    int delete)
+{
+	int ret;
+	u64 ref_blkno;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_refcount_tree *tree;
+
+	BUG_ON(!(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL));
+
+	ret = ocfs2_get_refcount_block(inode, &ref_blkno);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_get_refcount_tree(OCFS2_SB(inode->i_sb), ref_blkno, &tree);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_read_refcount_block(&tree->rf_ci, tree->rf_blkno,
+					&ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = __ocfs2_decrease_refcount(handle, &tree->rf_ci, ref_root_bh,
+					cpos, len, meta_ac, dealloc, delete);
+	if (ret)
+		mlog_errno(ret);
+out:
+	brelse(ref_root_bh);
+	return ret;
+}
+
+/*
+ * Mark the already-existing extent at cpos as refcounted for len clusters.
+ * This adds the refcount extent flag.
+ *
+ * If the existing extent is larger than the request, initiate a
+ * split. An attempt will be made at merging with adjacent extents.
+ *
+ * The caller is responsible for passing down meta_ac if we'll need it.
+ */
+static int ocfs2_mark_extent_refcounted(struct inode *inode,
+				struct ocfs2_extent_tree *et,
+				handle_t *handle, u32 cpos,
+				u32 len, u32 phys,
+				struct ocfs2_alloc_context *meta_ac,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret;
+
+	mlog(0, "Inode %lu refcount tree cpos %u, len %u, phys cluster %u\n",
+	     inode->i_ino, cpos, len, phys);
+
+	if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) {
+		ocfs2_error(inode->i_sb, "Inode %lu want to use refcount "
+			    "tree, but the feature bit is not set in the "
+			    "super block.", inode->i_ino);
+		ret = -EROFS;
+		goto out;
+	}
+
+	ret = ocfs2_change_extent_flag(handle, et, cpos,
+				       len, phys, meta_ac, dealloc,
+				       OCFS2_EXT_REFCOUNTED, 0);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	return ret;
+}
+
+/*
+ * Given some contiguous physical clusters, calculate what we need
+ * for modifying their refcount.
+ */
+static int ocfs2_calc_refcount_meta_credits(struct super_block *sb,
+					    struct ocfs2_caching_info *ci,
+					    struct buffer_head *ref_root_bh,
+					    u64 start_cpos,
+					    u32 clusters,
+					    int *meta_add,
+					    int *credits)
+{
+	int ret = 0, index, ref_blocks = 0, recs_add = 0;
+	u64 cpos = start_cpos;
+	struct ocfs2_refcount_block *rb;
+	struct ocfs2_refcount_rec rec;
+	struct buffer_head *ref_leaf_bh = NULL, *prev_bh = NULL;
+	u32 len;
+
+	mlog(0, "start_cpos %llu, clusters %u\n",
+	     (unsigned long long)start_cpos, clusters);
+	while (clusters) {
+		ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
+					     cpos, clusters, &rec,
+					     &index, &ref_leaf_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		if (ref_leaf_bh != prev_bh) {
+			/*
+			 * Now we encounter a new leaf block, so calculate
+			 * whether we need to extend the old leaf.
+			 */
+			if (prev_bh) {
+				rb = (struct ocfs2_refcount_block *)
+							prev_bh->b_data;
+
+				if (le64_to_cpu(rb->rf_records.rl_used) +
+				    recs_add >
+				    le16_to_cpu(rb->rf_records.rl_count))
+					ref_blocks++;
+			}
+
+			recs_add = 0;
+			*credits += 1;
+			brelse(prev_bh);
+			prev_bh = ref_leaf_bh;
+			get_bh(prev_bh);
+		}
+
+		rb = (struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+
+		mlog(0, "recs_add %d,cpos %llu, clusters %u, rec->r_cpos %llu,"
+		     "rec->r_clusters %u, rec->r_refcount %u, index %d\n",
+		     recs_add, (unsigned long long)cpos, clusters,
+		     (unsigned long long)le64_to_cpu(rec.r_cpos),
+		     le32_to_cpu(rec.r_clusters),
+		     le32_to_cpu(rec.r_refcount), index);
+
+		len = min((u64)cpos + clusters, le64_to_cpu(rec.r_cpos) +
+			  le32_to_cpu(rec.r_clusters)) - cpos;
+		/*
+		 * If the refcount rec already exist, cool. We just need
+		 * to check whether there is a split. Otherwise we just need
+		 * to increase the refcount.
+		 * If we will insert one, increases recs_add.
+		 *
+		 * We record all the records which will be inserted to the
+		 * same refcount block, so that we can tell exactly whether
+		 * we need a new refcount block or not.
+		 */
+		if (rec.r_refcount) {
+			/* Check whether we need a split at the beginning. */
+			if (cpos == start_cpos &&
+			    cpos != le64_to_cpu(rec.r_cpos))
+				recs_add++;
+
+			/* Check whether we need a split in the end. */
+			if (cpos + clusters < le64_to_cpu(rec.r_cpos) +
+			    le32_to_cpu(rec.r_clusters))
+				recs_add++;
+		} else
+			recs_add++;
+
+		brelse(ref_leaf_bh);
+		ref_leaf_bh = NULL;
+		clusters -= len;
+		cpos += len;
+	}
+
+	if (prev_bh) {
+		rb = (struct ocfs2_refcount_block *)prev_bh->b_data;
+
+		if (le64_to_cpu(rb->rf_records.rl_used) + recs_add >
+		    le16_to_cpu(rb->rf_records.rl_count))
+			ref_blocks++;
+
+		*credits += 1;
+	}
+
+	if (!ref_blocks)
+		goto out;
+
+	mlog(0, "we need ref_blocks %d\n", ref_blocks);
+	*meta_add += ref_blocks;
+	*credits += ref_blocks;
+
+	/*
+	 * So we may need ref_blocks to insert into the tree.
+	 * That also means we need to change the b-tree and add that number
+	 * of records since we never merge them.
+	 * We need one more block for expansion since the new created leaf
+	 * block is also full and needs split.
+	 */
+	rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	if (le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL) {
+		struct ocfs2_extent_tree et;
+
+		ocfs2_init_refcount_extent_tree(&et, ci, ref_root_bh);
+		*meta_add += ocfs2_extend_meta_needed(et.et_root_el);
+		*credits += ocfs2_calc_extend_credits(sb,
+						      et.et_root_el,
+						      ref_blocks);
+	} else {
+		*credits += OCFS2_EXPAND_REFCOUNT_TREE_CREDITS;
+		*meta_add += 1;
+	}
+
+out:
+	brelse(ref_leaf_bh);
+	brelse(prev_bh);
+	return ret;
+}
+
+/*
+ * For refcount tree, we will decrease some contiguous clusters
+ * refcount count, so just go through it to see how many blocks
+ * we gonna touch and whether we need to create new blocks.
+ *
+ * Normally the refcount blocks store these refcount should be
+ * continguous also, so that we can get the number easily.
+ * As for meta_ac, we will at most add split 2 refcount record and
+ * 2 more refcount block, so just check it in a rough way.
+ *
+ * Caller must hold refcount tree lock.
+ */
+int ocfs2_prepare_refcount_change_for_del(struct inode *inode,
+					  struct buffer_head *di_bh,
+					  u64 phys_blkno,
+					  u32 clusters,
+					  int *credits,
+					  struct ocfs2_alloc_context **meta_ac)
+{
+	int ret, ref_blocks = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_refcount_tree *tree;
+	u64 start_cpos = ocfs2_blocks_to_clusters(inode->i_sb, phys_blkno);
+
+	if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) {
+		ocfs2_error(inode->i_sb, "Inode %lu want to use refcount "
+			    "tree, but the feature bit is not set in the "
+			    "super block.", inode->i_ino);
+		ret = -EROFS;
+		goto out;
+	}
+
+	BUG_ON(!(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL));
+
+	ret = ocfs2_get_refcount_tree(OCFS2_SB(inode->i_sb),
+				      le64_to_cpu(di->i_refcount_loc), &tree);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_read_refcount_block(&tree->rf_ci,
+					le64_to_cpu(di->i_refcount_loc),
+					&ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_calc_refcount_meta_credits(inode->i_sb,
+					       &tree->rf_ci,
+					       ref_root_bh,
+					       start_cpos, clusters,
+					       &ref_blocks, credits);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mlog(0, "reserve new metadata %d, credits = %d\n",
+	     ref_blocks, *credits);
+
+	if (ref_blocks) {
+		ret = ocfs2_reserve_new_metadata_blocks(OCFS2_SB(inode->i_sb),
+							ref_blocks, meta_ac);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+out:
+	brelse(ref_root_bh);
+	return ret;
+}
+
+#define	MAX_CONTIG_BYTES	1048576
+
+static inline unsigned int ocfs2_cow_contig_clusters(struct super_block *sb)
+{
+	return ocfs2_clusters_for_bytes(sb, MAX_CONTIG_BYTES);
+}
+
+static inline unsigned int ocfs2_cow_contig_mask(struct super_block *sb)
+{
+	return ~(ocfs2_cow_contig_clusters(sb) - 1);
+}
+
+/*
+ * Given an extent that starts at 'start' and an I/O that starts at 'cpos',
+ * find an offset (start + (n * contig_clusters)) that is closest to cpos
+ * while still being less than or equal to it.
+ *
+ * The goal is to break the extent at a multiple of contig_clusters.
+ */
+static inline unsigned int ocfs2_cow_align_start(struct super_block *sb,
+						 unsigned int start,
+						 unsigned int cpos)
+{
+	BUG_ON(start > cpos);
+
+	return start + ((cpos - start) & ocfs2_cow_contig_mask(sb));
+}
+
+/*
+ * Given a cluster count of len, pad it out so that it is a multiple
+ * of contig_clusters.
+ */
+static inline unsigned int ocfs2_cow_align_length(struct super_block *sb,
+						  unsigned int len)
+{
+	unsigned int padded =
+		(len + (ocfs2_cow_contig_clusters(sb) - 1)) &
+		ocfs2_cow_contig_mask(sb);
+
+	/* Did we wrap? */
+	if (padded < len)
+		padded = UINT_MAX;
+
+	return padded;
+}
+
+/*
+ * Calculate out the start and number of virtual clusters we need to to CoW.
+ *
+ * cpos is vitual start cluster position we want to do CoW in a
+ * file and write_len is the cluster length.
+ * max_cpos is the place where we want to stop CoW intentionally.
+ *
+ * Normal we will start CoW from the beginning of extent record cotaining cpos.
+ * We try to break up extents on boundaries of MAX_CONTIG_BYTES so that we
+ * get good I/O from the resulting extent tree.
+ */
+static int ocfs2_refcount_cal_cow_clusters(struct inode *inode,
+					   struct ocfs2_extent_list *el,
+					   u32 cpos,
+					   u32 write_len,
+					   u32 max_cpos,
+					   u32 *cow_start,
+					   u32 *cow_len)
+{
+	int ret = 0;
+	int tree_height = le16_to_cpu(el->l_tree_depth), i;
+	struct buffer_head *eb_bh = NULL;
+	struct ocfs2_extent_block *eb = NULL;
+	struct ocfs2_extent_rec *rec;
+	unsigned int want_clusters, rec_end = 0;
+	int contig_clusters = ocfs2_cow_contig_clusters(inode->i_sb);
+	int leaf_clusters;
+
+	BUG_ON(cpos + write_len > max_cpos);
+
+	if (tree_height > 0) {
+		ret = ocfs2_find_leaf(INODE_CACHE(inode), el, cpos, &eb_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		eb = (struct ocfs2_extent_block *) eb_bh->b_data;
+		el = &eb->h_list;
+
+		if (el->l_tree_depth) {
+			ocfs2_error(inode->i_sb,
+				    "Inode %lu has non zero tree depth in "
+				    "leaf block %llu\n", inode->i_ino,
+				    (unsigned long long)eb_bh->b_blocknr);
+			ret = -EROFS;
+			goto out;
+		}
+	}
+
+	*cow_len = 0;
+	for (i = 0; i < le16_to_cpu(el->l_next_free_rec); i++) {
+		rec = &el->l_recs[i];
+
+		if (ocfs2_is_empty_extent(rec)) {
+			mlog_bug_on_msg(i != 0, "Inode %lu has empty record in "
+					"index %d\n", inode->i_ino, i);
+			continue;
+		}
+
+		if (le32_to_cpu(rec->e_cpos) +
+		    le16_to_cpu(rec->e_leaf_clusters) <= cpos)
+			continue;
+
+		if (*cow_len == 0) {
+			/*
+			 * We should find a refcounted record in the
+			 * first pass.
+			 */
+			BUG_ON(!(rec->e_flags & OCFS2_EXT_REFCOUNTED));
+			*cow_start = le32_to_cpu(rec->e_cpos);
+		}
+
+		/*
+		 * If we encounter a hole, a non-refcounted record or
+		 * pass the max_cpos, stop the search.
+		 */
+		if ((!(rec->e_flags & OCFS2_EXT_REFCOUNTED)) ||
+		    (*cow_len && rec_end != le32_to_cpu(rec->e_cpos)) ||
+		    (max_cpos <= le32_to_cpu(rec->e_cpos)))
+			break;
+
+		leaf_clusters = le16_to_cpu(rec->e_leaf_clusters);
+		rec_end = le32_to_cpu(rec->e_cpos) + leaf_clusters;
+		if (rec_end > max_cpos) {
+			rec_end = max_cpos;
+			leaf_clusters = rec_end - le32_to_cpu(rec->e_cpos);
+		}
+
+		/*
+		 * How many clusters do we actually need from
+		 * this extent?  First we see how many we actually
+		 * need to complete the write.  If that's smaller
+		 * than contig_clusters, we try for contig_clusters.
+		 */
+		if (!*cow_len)
+			want_clusters = write_len;
+		else
+			want_clusters = (cpos + write_len) -
+				(*cow_start + *cow_len);
+		if (want_clusters < contig_clusters)
+			want_clusters = contig_clusters;
+
+		/*
+		 * If the write does not cover the whole extent, we
+		 * need to calculate how we're going to split the extent.
+		 * We try to do it on contig_clusters boundaries.
+		 *
+		 * Any extent smaller than contig_clusters will be
+		 * CoWed in its entirety.
+		 */
+		if (leaf_clusters <= contig_clusters)
+			*cow_len += leaf_clusters;
+		else if (*cow_len || (*cow_start == cpos)) {
+			/*
+			 * This extent needs to be CoW'd from its
+			 * beginning, so all we have to do is compute
+			 * how many clusters to grab.  We align
+			 * want_clusters to the edge of contig_clusters
+			 * to get better I/O.
+			 */
+			want_clusters = ocfs2_cow_align_length(inode->i_sb,
+							       want_clusters);
+
+			if (leaf_clusters < want_clusters)
+				*cow_len += leaf_clusters;
+			else
+				*cow_len += want_clusters;
+		} else if ((*cow_start + contig_clusters) >=
+			   (cpos + write_len)) {
+			/*
+			 * Breaking off contig_clusters at the front
+			 * of the extent will cover our write.  That's
+			 * easy.
+			 */
+			*cow_len = contig_clusters;
+		} else if ((rec_end - cpos) <= contig_clusters) {
+			/*
+			 * Breaking off contig_clusters at the tail of
+			 * this extent will cover cpos.
+			 */
+			*cow_start = rec_end - contig_clusters;
+			*cow_len = contig_clusters;
+		} else if ((rec_end - cpos) <= want_clusters) {
+			/*
+			 * While we can't fit the entire write in this
+			 * extent, we know that the write goes from cpos
+			 * to the end of the extent.  Break that off.
+			 * We try to break it at some multiple of
+			 * contig_clusters from the front of the extent.
+			 * Failing that (ie, cpos is within
+			 * contig_clusters of the front), we'll CoW the
+			 * entire extent.
+			 */
+			*cow_start = ocfs2_cow_align_start(inode->i_sb,
+							   *cow_start, cpos);
+			*cow_len = rec_end - *cow_start;
+		} else {
+			/*
+			 * Ok, the entire write lives in the middle of
+			 * this extent.  Let's try to slice the extent up
+			 * nicely.  Optimally, our CoW region starts at
+			 * m*contig_clusters from the beginning of the
+			 * extent and goes for n*contig_clusters,
+			 * covering the entire write.
+			 */
+			*cow_start = ocfs2_cow_align_start(inode->i_sb,
+							   *cow_start, cpos);
+
+			want_clusters = (cpos + write_len) - *cow_start;
+			want_clusters = ocfs2_cow_align_length(inode->i_sb,
+							       want_clusters);
+			if (*cow_start + want_clusters <= rec_end)
+				*cow_len = want_clusters;
+			else
+				*cow_len = rec_end - *cow_start;
+		}
+
+		/* Have we covered our entire write yet? */
+		if ((*cow_start + *cow_len) >= (cpos + write_len))
+			break;
+
+		/*
+		 * If we reach the end of the extent block and don't get enough
+		 * clusters, continue with the next extent block if possible.
+		 */
+		if (i + 1 == le16_to_cpu(el->l_next_free_rec) &&
+		    eb && eb->h_next_leaf_blk) {
+			brelse(eb_bh);
+			eb_bh = NULL;
+
+			ret = ocfs2_read_extent_block(INODE_CACHE(inode),
+					       le64_to_cpu(eb->h_next_leaf_blk),
+					       &eb_bh);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+
+			eb = (struct ocfs2_extent_block *) eb_bh->b_data;
+			el = &eb->h_list;
+			i = -1;
+		}
+	}
+
+out:
+	brelse(eb_bh);
+	return ret;
+}
+
+/*
+ * Prepare meta_ac, data_ac and calculate credits when we want to add some
+ * num_clusters in data_tree "et" and change the refcount for the old
+ * clusters(starting form p_cluster) in the refcount tree.
+ *
+ * Note:
+ * 1. since we may split the old tree, so we at most will need num_clusters + 2
+ *    more new leaf records.
+ * 2. In some case, we may not need to reserve new clusters(e.g, reflink), so
+ *    just give data_ac = NULL.
+ */
+static int ocfs2_lock_refcount_allocators(struct super_block *sb,
+					u32 p_cluster, u32 num_clusters,
+					struct ocfs2_extent_tree *et,
+					struct ocfs2_caching_info *ref_ci,
+					struct buffer_head *ref_root_bh,
+					struct ocfs2_alloc_context **meta_ac,
+					struct ocfs2_alloc_context **data_ac,
+					int *credits)
+{
+	int ret = 0, meta_add = 0;
+	int num_free_extents = ocfs2_num_free_extents(OCFS2_SB(sb), et);
+
+	if (num_free_extents < 0) {
+		ret = num_free_extents;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (num_free_extents < num_clusters + 2)
+		meta_add =
+			ocfs2_extend_meta_needed(et->et_root_el);
+
+	*credits += ocfs2_calc_extend_credits(sb, et->et_root_el,
+					      num_clusters + 2);
+
+	ret = ocfs2_calc_refcount_meta_credits(sb, ref_ci, ref_root_bh,
+					       p_cluster, num_clusters,
+					       &meta_add, credits);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mlog(0, "reserve new metadata %d, clusters %u, credits = %d\n",
+	     meta_add, num_clusters, *credits);
+	ret = ocfs2_reserve_new_metadata_blocks(OCFS2_SB(sb), meta_add,
+						meta_ac);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (data_ac) {
+		ret = ocfs2_reserve_clusters(OCFS2_SB(sb), num_clusters,
+					     data_ac);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+out:
+	if (ret) {
+		if (*meta_ac) {
+			ocfs2_free_alloc_context(*meta_ac);
+			*meta_ac = NULL;
+		}
+	}
+
+	return ret;
+}
+
+static int ocfs2_clear_cow_buffer(handle_t *handle, struct buffer_head *bh)
+{
+	BUG_ON(buffer_dirty(bh));
+
+	clear_buffer_mapped(bh);
+
+	return 0;
+}
+
+static int ocfs2_duplicate_clusters_by_page(handle_t *handle,
+					    struct ocfs2_cow_context *context,
+					    u32 cpos, u32 old_cluster,
+					    u32 new_cluster, u32 new_len)
+{
+	int ret = 0, partial;
+	struct ocfs2_caching_info *ci = context->data_et.et_ci;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	u64 new_block = ocfs2_clusters_to_blocks(sb, new_cluster);
+	struct page *page;
+	pgoff_t page_index;
+	unsigned int from, to;
+	loff_t offset, end, map_end;
+	struct address_space *mapping = context->inode->i_mapping;
+
+	mlog(0, "old_cluster %u, new %u, len %u at offset %u\n", old_cluster,
+	     new_cluster, new_len, cpos);
+
+	offset = ((loff_t)cpos) << OCFS2_SB(sb)->s_clustersize_bits;
+	end = offset + (new_len << OCFS2_SB(sb)->s_clustersize_bits);
+
+	while (offset < end) {
+		page_index = offset >> PAGE_CACHE_SHIFT;
+		map_end = (page_index + 1) << PAGE_CACHE_SHIFT;
+		if (map_end > end)
+			map_end = end;
+
+		/* from, to is the offset within the page. */
+		from = offset & (PAGE_CACHE_SIZE - 1);
+		to = PAGE_CACHE_SIZE;
+		if (map_end & (PAGE_CACHE_SIZE - 1))
+			to = map_end & (PAGE_CACHE_SIZE - 1);
+
+		page = grab_cache_page(mapping, page_index);
+
+		/* This page can't be dirtied before we CoW it out. */
+		BUG_ON(PageDirty(page));
+
+		if (!PageUptodate(page)) {
+			ret = block_read_full_page(page, ocfs2_get_block);
+			if (ret) {
+				mlog_errno(ret);
+				goto unlock;
+			}
+			lock_page(page);
+		}
+
+		if (page_has_buffers(page)) {
+			ret = walk_page_buffers(handle, page_buffers(page),
+						from, to, &partial,
+						ocfs2_clear_cow_buffer);
+			if (ret) {
+				mlog_errno(ret);
+				goto unlock;
+			}
+		}
+
+		ocfs2_map_and_dirty_page(context->inode,
+					 handle, from, to,
+					 page, 0, &new_block);
+		mark_page_accessed(page);
+unlock:
+		unlock_page(page);
+		page_cache_release(page);
+		page = NULL;
+		offset = map_end;
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+static int ocfs2_duplicate_clusters_by_jbd(handle_t *handle,
+					   struct ocfs2_cow_context *context,
+					   u32 cpos, u32 old_cluster,
+					   u32 new_cluster, u32 new_len)
+{
+	int ret = 0;
+	struct super_block *sb = context->inode->i_sb;
+	struct ocfs2_caching_info *ci = context->data_et.et_ci;
+	int i, blocks = ocfs2_clusters_to_blocks(sb, new_len);
+	u64 old_block = ocfs2_clusters_to_blocks(sb, old_cluster);
+	u64 new_block = ocfs2_clusters_to_blocks(sb, new_cluster);
+	struct ocfs2_super *osb = OCFS2_SB(sb);
+	struct buffer_head *old_bh = NULL;
+	struct buffer_head *new_bh = NULL;
+
+	mlog(0, "old_cluster %u, new %u, len %u\n", old_cluster,
+	     new_cluster, new_len);
+
+	for (i = 0; i < blocks; i++, old_block++, new_block++) {
+		new_bh = sb_getblk(osb->sb, new_block);
+		if (new_bh == NULL) {
+			ret = -EIO;
+			mlog_errno(ret);
+			break;
+		}
+
+		ocfs2_set_new_buffer_uptodate(ci, new_bh);
+
+		ret = ocfs2_read_block(ci, old_block, &old_bh, NULL);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		ret = ocfs2_journal_access(handle, ci, new_bh,
+					   OCFS2_JOURNAL_ACCESS_CREATE);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		memcpy(new_bh->b_data, old_bh->b_data, sb->s_blocksize);
+		ret = ocfs2_journal_dirty(handle, new_bh);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		brelse(new_bh);
+		brelse(old_bh);
+		new_bh = NULL;
+		old_bh = NULL;
+	}
+
+	brelse(new_bh);
+	brelse(old_bh);
+	return ret;
+}
+
+static int ocfs2_clear_ext_refcount(handle_t *handle,
+				    struct ocfs2_extent_tree *et,
+				    u32 cpos, u32 p_cluster, u32 len,
+				    unsigned int ext_flags,
+				    struct ocfs2_alloc_context *meta_ac,
+				    struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret, index;
+	struct ocfs2_extent_rec replace_rec;
+	struct ocfs2_path *path = NULL;
+	struct ocfs2_extent_list *el;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
+	u64 ino = ocfs2_metadata_cache_owner(et->et_ci);
+
+	mlog(0, "inode %llu cpos %u, len %u, p_cluster %u, ext_flags %u\n",
+	     (unsigned long long)ino, cpos, len, p_cluster, ext_flags);
+
+	memset(&replace_rec, 0, sizeof(replace_rec));
+	replace_rec.e_cpos = cpu_to_le32(cpos);
+	replace_rec.e_leaf_clusters = cpu_to_le16(len);
+	replace_rec.e_blkno = cpu_to_le64(ocfs2_clusters_to_blocks(sb,
+								   p_cluster));
+	replace_rec.e_flags = ext_flags;
+	replace_rec.e_flags &= ~OCFS2_EXT_REFCOUNTED;
+
+	path = ocfs2_new_path_from_et(et);
+	if (!path) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_find_path(et->et_ci, path, cpos);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	el = path_leaf_el(path);
+
+	index = ocfs2_search_extent_list(el, cpos);
+	if (index == -1 || index >= le16_to_cpu(el->l_next_free_rec)) {
+		ocfs2_error(sb,
+			    "Inode %llu has an extent at cpos %u which can no "
+			    "longer be found.\n",
+			    (unsigned long long)ino, cpos);
+		ret = -EROFS;
+		goto out;
+	}
+
+	ret = ocfs2_split_extent(handle, et, path, index,
+				 &replace_rec, meta_ac, dealloc);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	ocfs2_free_path(path);
+	return ret;
+}
+
+static int ocfs2_replace_clusters(handle_t *handle,
+				  struct ocfs2_cow_context *context,
+				  u32 cpos, u32 old,
+				  u32 new, u32 len,
+				  unsigned int ext_flags)
+{
+	int ret;
+	struct ocfs2_caching_info *ci = context->data_et.et_ci;
+	u64 ino = ocfs2_metadata_cache_owner(ci);
+
+	mlog(0, "inode %llu, cpos %u, old %u, new %u, len %u, ext_flags %u\n",
+	     (unsigned long long)ino, cpos, old, new, len, ext_flags);
+
+	/*If the old clusters is unwritten, no need to duplicate. */
+	if (!(ext_flags & OCFS2_EXT_UNWRITTEN)) {
+		ret = context->cow_duplicate_clusters(handle, context, cpos,
+						      old, new, len);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+	}
+
+	ret = ocfs2_clear_ext_refcount(handle, &context->data_et,
+				       cpos, new, len, ext_flags,
+				       context->meta_ac, &context->dealloc);
+	if (ret)
+		mlog_errno(ret);
+out:
+	return ret;
+}
+
+static int ocfs2_cow_sync_writeback(struct super_block *sb,
+				    struct ocfs2_cow_context *context,
+				    u32 cpos, u32 num_clusters)
+{
+	int ret = 0;
+	loff_t offset, end, map_end;
+	pgoff_t page_index;
+	struct page *page;
+
+	if (ocfs2_should_order_data(context->inode))
+		return 0;
+
+	offset = ((loff_t)cpos) << OCFS2_SB(sb)->s_clustersize_bits;
+	end = offset + (num_clusters << OCFS2_SB(sb)->s_clustersize_bits);
+
+	ret = filemap_fdatawrite_range(context->inode->i_mapping,
+				       offset, end - 1);
+	if (ret < 0) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	while (offset < end) {
+		page_index = offset >> PAGE_CACHE_SHIFT;
+		map_end = (page_index + 1) << PAGE_CACHE_SHIFT;
+		if (map_end > end)
+			map_end = end;
+
+		page = grab_cache_page(context->inode->i_mapping, page_index);
+		BUG_ON(!page);
+
+		wait_on_page_writeback(page);
+		if (PageError(page)) {
+			ret = -EIO;
+			mlog_errno(ret);
+		} else
+			mark_page_accessed(page);
+
+		unlock_page(page);
+		page_cache_release(page);
+		page = NULL;
+		offset = map_end;
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+static int ocfs2_di_get_clusters(struct ocfs2_cow_context *context,
+				 u32 v_cluster, u32 *p_cluster,
+				 u32 *num_clusters,
+				 unsigned int *extent_flags)
+{
+	return ocfs2_get_clusters(context->inode, v_cluster, p_cluster,
+				  num_clusters, extent_flags);
+}
+
+static int ocfs2_make_clusters_writable(struct super_block *sb,
+					struct ocfs2_cow_context *context,
+					u32 cpos, u32 p_cluster,
+					u32 num_clusters, unsigned int e_flags)
+{
+	int ret, delete, index, credits =  0;
+	u32 new_bit, new_len;
+	unsigned int set_len;
+	struct ocfs2_super *osb = OCFS2_SB(sb);
+	handle_t *handle;
+	struct buffer_head *ref_leaf_bh = NULL;
+	struct ocfs2_caching_info *ref_ci = &context->ref_tree->rf_ci;
+	struct ocfs2_refcount_rec rec;
+
+	mlog(0, "cpos %u, p_cluster %u, num_clusters %u, e_flags %u\n",
+	     cpos, p_cluster, num_clusters, e_flags);
+
+	ret = ocfs2_lock_refcount_allocators(sb, p_cluster, num_clusters,
+					     &context->data_et,
+					     ref_ci,
+					     context->ref_root_bh,
+					     &context->meta_ac,
+					     &context->data_ac, &credits);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	if (context->post_refcount)
+		credits += context->post_refcount->credits;
+
+	credits += context->extra_credits;
+	handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	while (num_clusters) {
+		ret = ocfs2_get_refcount_rec(ref_ci, context->ref_root_bh,
+					     p_cluster, num_clusters,
+					     &rec, &index, &ref_leaf_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_commit;
+		}
+
+		BUG_ON(!rec.r_refcount);
+		set_len = min((u64)p_cluster + num_clusters,
+			      le64_to_cpu(rec.r_cpos) +
+			      le32_to_cpu(rec.r_clusters)) - p_cluster;
+
+		/*
+		 * There are many different situation here.
+		 * 1. If refcount == 1, remove the flag and don't COW.
+		 * 2. If refcount > 1, allocate clusters.
+		 *    Here we may not allocate r_len once at a time, so continue
+		 *    until we reach num_clusters.
+		 */
+		if (le32_to_cpu(rec.r_refcount) == 1) {
+			delete = 0;
+			ret = ocfs2_clear_ext_refcount(handle,
+						       &context->data_et,
+						       cpos, p_cluster,
+						       set_len, e_flags,
+						       context->meta_ac,
+						       &context->dealloc);
+			if (ret) {
+				mlog_errno(ret);
+				goto out_commit;
+			}
+		} else {
+			delete = 1;
+
+			ret = __ocfs2_claim_clusters(osb, handle,
+						     context->data_ac,
+						     1, set_len,
+						     &new_bit, &new_len);
+			if (ret) {
+				mlog_errno(ret);
+				goto out_commit;
+			}
+
+			ret = ocfs2_replace_clusters(handle, context,
+						     cpos, p_cluster, new_bit,
+						     new_len, e_flags);
+			if (ret) {
+				mlog_errno(ret);
+				goto out_commit;
+			}
+			set_len = new_len;
+		}
+
+		ret = __ocfs2_decrease_refcount(handle, ref_ci,
+						context->ref_root_bh,
+						p_cluster, set_len,
+						context->meta_ac,
+						&context->dealloc, delete);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_commit;
+		}
+
+		cpos += set_len;
+		p_cluster += set_len;
+		num_clusters -= set_len;
+		brelse(ref_leaf_bh);
+		ref_leaf_bh = NULL;
+	}
+
+	/* handle any post_cow action. */
+	if (context->post_refcount && context->post_refcount->func) {
+		ret = context->post_refcount->func(context->inode, handle,
+						context->post_refcount->para);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_commit;
+		}
+	}
+
+	/*
+	 * Here we should write the new page out first if we are
+	 * in write-back mode.
+	 */
+	if (context->get_clusters == ocfs2_di_get_clusters) {
+		ret = ocfs2_cow_sync_writeback(sb, context, cpos, num_clusters);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+
+out:
+	if (context->data_ac) {
+		ocfs2_free_alloc_context(context->data_ac);
+		context->data_ac = NULL;
+	}
+	if (context->meta_ac) {
+		ocfs2_free_alloc_context(context->meta_ac);
+		context->meta_ac = NULL;
+	}
+	brelse(ref_leaf_bh);
+
+	return ret;
+}
+
+static int ocfs2_replace_cow(struct ocfs2_cow_context *context)
+{
+	int ret = 0;
+	struct inode *inode = context->inode;
+	u32 cow_start = context->cow_start, cow_len = context->cow_len;
+	u32 p_cluster, num_clusters;
+	unsigned int ext_flags;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+
+	if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) {
+		ocfs2_error(inode->i_sb, "Inode %lu want to use refcount "
+			    "tree, but the feature bit is not set in the "
+			    "super block.", inode->i_ino);
+		return -EROFS;
+	}
+
+	ocfs2_init_dealloc_ctxt(&context->dealloc);
+
+	while (cow_len) {
+		ret = context->get_clusters(context, cow_start, &p_cluster,
+					    &num_clusters, &ext_flags);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		BUG_ON(!(ext_flags & OCFS2_EXT_REFCOUNTED));
+
+		if (cow_len < num_clusters)
+			num_clusters = cow_len;
+
+		ret = ocfs2_make_clusters_writable(inode->i_sb, context,
+						   cow_start, p_cluster,
+						   num_clusters, ext_flags);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		cow_len -= num_clusters;
+		cow_start += num_clusters;
+	}
+
+	if (ocfs2_dealloc_has_cluster(&context->dealloc)) {
+		ocfs2_schedule_truncate_log_flush(osb, 1);
+		ocfs2_run_deallocs(osb, &context->dealloc);
+	}
+
+	return ret;
+}
+
+/*
+ * Starting at cpos, try to CoW write_len clusters.  Don't CoW
+ * past max_cpos.  This will stop when it runs into a hole or an
+ * unrefcounted extent.
+ */
+static int ocfs2_refcount_cow_hunk(struct inode *inode,
+				   struct buffer_head *di_bh,
+				   u32 cpos, u32 write_len, u32 max_cpos)
+{
+	int ret;
+	u32 cow_start = 0, cow_len = 0;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_refcount_tree *ref_tree;
+	struct ocfs2_cow_context *context = NULL;
+
+	BUG_ON(!(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL));
+
+	ret = ocfs2_refcount_cal_cow_clusters(inode, &di->id2.i_list,
+					      cpos, write_len, max_cpos,
+					      &cow_start, &cow_len);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mlog(0, "CoW inode %lu, cpos %u, write_len %u, cow_start %u, "
+	     "cow_len %u\n", inode->i_ino,
+	     cpos, write_len, cow_start, cow_len);
+
+	BUG_ON(cow_len == 0);
+
+	context = kzalloc(sizeof(struct ocfs2_cow_context), GFP_NOFS);
+	if (!context) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_lock_refcount_tree(osb, le64_to_cpu(di->i_refcount_loc),
+				       1, &ref_tree, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	context->inode = inode;
+	context->cow_start = cow_start;
+	context->cow_len = cow_len;
+	context->ref_tree = ref_tree;
+	context->ref_root_bh = ref_root_bh;
+	context->cow_duplicate_clusters = ocfs2_duplicate_clusters_by_page;
+	context->get_clusters = ocfs2_di_get_clusters;
+
+	ocfs2_init_dinode_extent_tree(&context->data_et,
+				      INODE_CACHE(inode), di_bh);
+
+	ret = ocfs2_replace_cow(context);
+	if (ret)
+		mlog_errno(ret);
+
+	/*
+	 * truncate the extent map here since no matter whether we meet with
+	 * any error during the action, we shouldn't trust cached extent map
+	 * any more.
+	 */
+	ocfs2_extent_map_trunc(inode, cow_start);
+
+	ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+	brelse(ref_root_bh);
+out:
+	kfree(context);
+	return ret;
+}
+
+/*
+ * CoW any and all clusters between cpos and cpos+write_len.
+ * Don't CoW past max_cpos.  If this returns successfully, all
+ * clusters between cpos and cpos+write_len are safe to modify.
+ */
+int ocfs2_refcount_cow(struct inode *inode,
+		       struct buffer_head *di_bh,
+		       u32 cpos, u32 write_len, u32 max_cpos)
+{
+	int ret = 0;
+	u32 p_cluster, num_clusters;
+	unsigned int ext_flags;
+
+	while (write_len) {
+		ret = ocfs2_get_clusters(inode, cpos, &p_cluster,
+					 &num_clusters, &ext_flags);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		if (write_len < num_clusters)
+			num_clusters = write_len;
+
+		if (ext_flags & OCFS2_EXT_REFCOUNTED) {
+			ret = ocfs2_refcount_cow_hunk(inode, di_bh, cpos,
+						      num_clusters, max_cpos);
+			if (ret) {
+				mlog_errno(ret);
+				break;
+			}
+		}
+
+		write_len -= num_clusters;
+		cpos += num_clusters;
+	}
+
+	return ret;
+}
+
+static int ocfs2_xattr_value_get_clusters(struct ocfs2_cow_context *context,
+					  u32 v_cluster, u32 *p_cluster,
+					  u32 *num_clusters,
+					  unsigned int *extent_flags)
+{
+	struct inode *inode = context->inode;
+	struct ocfs2_xattr_value_root *xv = context->cow_object;
+
+	return ocfs2_xattr_get_clusters(inode, v_cluster, p_cluster,
+					num_clusters, &xv->xr_list,
+					extent_flags);
+}
+
+/*
+ * Given a xattr value root, calculate the most meta/credits we need for
+ * refcount tree change if we truncate it to 0.
+ */
+int ocfs2_refcounted_xattr_delete_need(struct inode *inode,
+				       struct ocfs2_caching_info *ref_ci,
+				       struct buffer_head *ref_root_bh,
+				       struct ocfs2_xattr_value_root *xv,
+				       int *meta_add, int *credits)
+{
+	int ret = 0, index, ref_blocks = 0;
+	u32 p_cluster, num_clusters;
+	u32 cpos = 0, clusters = le32_to_cpu(xv->xr_clusters);
+	struct ocfs2_refcount_block *rb;
+	struct ocfs2_refcount_rec rec;
+	struct buffer_head *ref_leaf_bh = NULL;
+
+	while (cpos < clusters) {
+		ret = ocfs2_xattr_get_clusters(inode, cpos, &p_cluster,
+					       &num_clusters, &xv->xr_list,
+					       NULL);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+
+		cpos += num_clusters;
+
+		while (num_clusters) {
+			ret = ocfs2_get_refcount_rec(ref_ci, ref_root_bh,
+						     p_cluster, num_clusters,
+						     &rec, &index,
+						     &ref_leaf_bh);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+
+			BUG_ON(!rec.r_refcount);
+
+			rb = (struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
+
+			/*
+			 * We really don't know whether the other clusters is in
+			 * this refcount block or not, so just take the worst
+			 * case that all the clusters are in this block and each
+			 * one will split a refcount rec, so totally we need
+			 * clusters * 2 new refcount rec.
+			 */
+			if (le64_to_cpu(rb->rf_records.rl_used) + clusters * 2 >
+			    le16_to_cpu(rb->rf_records.rl_count))
+				ref_blocks++;
+
+			*credits += 1;
+			brelse(ref_leaf_bh);
+			ref_leaf_bh = NULL;
+
+			if (num_clusters <= le32_to_cpu(rec.r_clusters))
+				break;
+			else
+				num_clusters -= le32_to_cpu(rec.r_clusters);
+			p_cluster += num_clusters;
+		}
+	}
+
+	*meta_add += ref_blocks;
+	if (!ref_blocks)
+		goto out;
+
+	rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+	if (le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL)
+		*credits += OCFS2_EXPAND_REFCOUNT_TREE_CREDITS;
+	else {
+		struct ocfs2_extent_tree et;
+
+		ocfs2_init_refcount_extent_tree(&et, ref_ci, ref_root_bh);
+		*credits += ocfs2_calc_extend_credits(inode->i_sb,
+						      et.et_root_el,
+						      ref_blocks);
+	}
+
+out:
+	brelse(ref_leaf_bh);
+	return ret;
+}
+
+/*
+ * Do CoW for xattr.
+ */
+int ocfs2_refcount_cow_xattr(struct inode *inode,
+			     struct ocfs2_dinode *di,
+			     struct ocfs2_xattr_value_buf *vb,
+			     struct ocfs2_refcount_tree *ref_tree,
+			     struct buffer_head *ref_root_bh,
+			     u32 cpos, u32 write_len,
+			     struct ocfs2_post_refcount *post)
+{
+	int ret;
+	struct ocfs2_xattr_value_root *xv = vb->vb_xv;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_cow_context *context = NULL;
+	u32 cow_start, cow_len;
+
+	BUG_ON(!(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL));
+
+	ret = ocfs2_refcount_cal_cow_clusters(inode, &xv->xr_list,
+					      cpos, write_len, UINT_MAX,
+					      &cow_start, &cow_len);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	BUG_ON(cow_len == 0);
+
+	context = kzalloc(sizeof(struct ocfs2_cow_context), GFP_NOFS);
+	if (!context) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	context->inode = inode;
+	context->cow_start = cow_start;
+	context->cow_len = cow_len;
+	context->ref_tree = ref_tree;
+	context->ref_root_bh = ref_root_bh;;
+	context->cow_object = xv;
+
+	context->cow_duplicate_clusters = ocfs2_duplicate_clusters_by_jbd;
+	/* We need the extra credits for duplicate_clusters by jbd. */
+	context->extra_credits =
+		ocfs2_clusters_to_blocks(inode->i_sb, 1) * cow_len;
+	context->get_clusters = ocfs2_xattr_value_get_clusters;
+	context->post_refcount = post;
+
+	ocfs2_init_xattr_value_extent_tree(&context->data_et,
+					   INODE_CACHE(inode), vb);
+
+	ret = ocfs2_replace_cow(context);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	kfree(context);
+	return ret;
+}
+
+/*
+ * Insert a new extent into refcount tree and mark a extent rec
+ * as refcounted in the dinode tree.
+ */
+int ocfs2_add_refcount_flag(struct inode *inode,
+			    struct ocfs2_extent_tree *data_et,
+			    struct ocfs2_caching_info *ref_ci,
+			    struct buffer_head *ref_root_bh,
+			    u32 cpos, u32 p_cluster, u32 num_clusters,
+			    struct ocfs2_cached_dealloc_ctxt *dealloc,
+			    struct ocfs2_post_refcount *post)
+{
+	int ret;
+	handle_t *handle;
+	int credits = 1, ref_blocks = 0;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_alloc_context *meta_ac = NULL;
+
+	ret = ocfs2_calc_refcount_meta_credits(inode->i_sb,
+					       ref_ci, ref_root_bh,
+					       p_cluster, num_clusters,
+					       &ref_blocks, &credits);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mlog(0, "reserve new metadata %d, credits = %d\n",
+	     ref_blocks, credits);
+
+	if (ref_blocks) {
+		ret = ocfs2_reserve_new_metadata_blocks(OCFS2_SB(inode->i_sb),
+							ref_blocks, &meta_ac);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+	}
+
+	if (post)
+		credits += post->credits;
+
+	handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_mark_extent_refcounted(inode, data_et, handle,
+					   cpos, num_clusters, p_cluster,
+					   meta_ac, dealloc);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ret = __ocfs2_increase_refcount(handle, ref_ci, ref_root_bh,
+					p_cluster, num_clusters, 0,
+					meta_ac, dealloc);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	if (post && post->func) {
+		ret = post->func(inode, handle, post->para);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+out:
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+	return ret;
+}
+
+static int ocfs2_change_ctime(struct inode *inode,
+			      struct buffer_head *di_bh)
+{
+	int ret;
+	handle_t *handle;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+
+	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_di(handle, INODE_CACHE(inode), di_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	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(handle, di_bh);
+
+out_commit:
+	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
+out:
+	return ret;
+}
+
+static int ocfs2_attach_refcount_tree(struct inode *inode,
+				      struct buffer_head *di_bh)
+{
+	int ret, data_changed = 0;
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_refcount_tree *ref_tree;
+	unsigned int ext_flags;
+	loff_t size;
+	u32 cpos, num_clusters, clusters, p_cluster;
+	struct ocfs2_cached_dealloc_ctxt dealloc;
+	struct ocfs2_extent_tree di_et;
+
+	ocfs2_init_dealloc_ctxt(&dealloc);
+
+	if (!(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL)) {
+		ret = ocfs2_create_refcount_tree(inode, di_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+	}
+
+	BUG_ON(!di->i_refcount_loc);
+	ret = ocfs2_lock_refcount_tree(osb,
+				       le64_to_cpu(di->i_refcount_loc), 1,
+				       &ref_tree, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ocfs2_init_dinode_extent_tree(&di_et, INODE_CACHE(inode), di_bh);
+
+	size = i_size_read(inode);
+	clusters = ocfs2_clusters_for_bytes(inode->i_sb, size);
+
+	cpos = 0;
+	while (cpos < clusters) {
+		ret = ocfs2_get_clusters(inode, cpos, &p_cluster,
+					 &num_clusters, &ext_flags);
+
+		if (p_cluster && !(ext_flags & OCFS2_EXT_REFCOUNTED)) {
+			ret = ocfs2_add_refcount_flag(inode, &di_et,
+						      &ref_tree->rf_ci,
+						      ref_root_bh, cpos,
+						      p_cluster, num_clusters,
+						      &dealloc, NULL);
+			if (ret) {
+				mlog_errno(ret);
+				goto unlock;
+			}
+
+			data_changed = 1;
+		}
+		cpos += num_clusters;
+	}
+
+	if (oi->ip_dyn_features & OCFS2_HAS_XATTR_FL) {
+		ret = ocfs2_xattr_attach_refcount_tree(inode, di_bh,
+						       &ref_tree->rf_ci,
+						       ref_root_bh,
+						       &dealloc);
+		if (ret) {
+			mlog_errno(ret);
+			goto unlock;
+		}
+	}
+
+	if (data_changed) {
+		ret = ocfs2_change_ctime(inode, di_bh);
+		if (ret)
+			mlog_errno(ret);
+	}
+
+unlock:
+	ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+	brelse(ref_root_bh);
+
+	if (!ret && ocfs2_dealloc_has_cluster(&dealloc)) {
+		ocfs2_schedule_truncate_log_flush(osb, 1);
+		ocfs2_run_deallocs(osb, &dealloc);
+	}
+out:
+	/*
+	 * Empty the extent map so that we may get the right extent
+	 * record from the disk.
+	 */
+	ocfs2_extent_map_trunc(inode, 0);
+
+	return ret;
+}
+
+static int ocfs2_add_refcounted_extent(struct inode *inode,
+				   struct ocfs2_extent_tree *et,
+				   struct ocfs2_caching_info *ref_ci,
+				   struct buffer_head *ref_root_bh,
+				   u32 cpos, u32 p_cluster, u32 num_clusters,
+				   unsigned int ext_flags,
+				   struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret;
+	handle_t *handle;
+	int credits = 0;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_alloc_context *meta_ac = NULL;
+
+	ret = ocfs2_lock_refcount_allocators(inode->i_sb,
+					     p_cluster, num_clusters,
+					     et, ref_ci,
+					     ref_root_bh, &meta_ac,
+					     NULL, &credits);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_insert_extent(handle, et, cpos,
+			cpu_to_le64(ocfs2_clusters_to_blocks(inode->i_sb,
+							     p_cluster)),
+			num_clusters, ext_flags, meta_ac);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ret = ocfs2_increase_refcount(handle, ref_ci, ref_root_bh,
+				      p_cluster, num_clusters,
+				      meta_ac, dealloc);
+	if (ret)
+		mlog_errno(ret);
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+out:
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+	return ret;
+}
+
+static int ocfs2_duplicate_extent_list(struct inode *s_inode,
+				struct inode *t_inode,
+				struct buffer_head *t_bh,
+				struct ocfs2_caching_info *ref_ci,
+				struct buffer_head *ref_root_bh,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret = 0;
+	u32 p_cluster, num_clusters, clusters, cpos;
+	loff_t size;
+	unsigned int ext_flags;
+	struct ocfs2_extent_tree et;
+
+	ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(t_inode), t_bh);
+
+	size = i_size_read(s_inode);
+	clusters = ocfs2_clusters_for_bytes(s_inode->i_sb, size);
+
+	cpos = 0;
+	while (cpos < clusters) {
+		ret = ocfs2_get_clusters(s_inode, cpos, &p_cluster,
+					 &num_clusters, &ext_flags);
+
+		if (p_cluster) {
+			ret = ocfs2_add_refcounted_extent(t_inode, &et,
+							  ref_ci, ref_root_bh,
+							  cpos, p_cluster,
+							  num_clusters,
+							  ext_flags,
+							  dealloc);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+		}
+
+		cpos += num_clusters;
+	}
+
+out:
+	return ret;
+}
+
+/*
+ * change the new file's attributes to the src.
+ *
+ * reflink creates a snapshot of a file, that means the attributes
+ * must be identical except for three exceptions - nlink, ino, and ctime.
+ */
+static int ocfs2_complete_reflink(struct inode *s_inode,
+				  struct buffer_head *s_bh,
+				  struct inode *t_inode,
+				  struct buffer_head *t_bh,
+				  bool preserve)
+{
+	int ret;
+	handle_t *handle;
+	struct ocfs2_dinode *s_di = (struct ocfs2_dinode *)s_bh->b_data;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)t_bh->b_data;
+	loff_t size = i_size_read(s_inode);
+
+	handle = ocfs2_start_trans(OCFS2_SB(t_inode->i_sb),
+				   OCFS2_INODE_UPDATE_CREDITS);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		return ret;
+	}
+
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(t_inode), t_bh,
+				      OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	spin_lock(&OCFS2_I(t_inode)->ip_lock);
+	OCFS2_I(t_inode)->ip_clusters = OCFS2_I(s_inode)->ip_clusters;
+	OCFS2_I(t_inode)->ip_attr = OCFS2_I(s_inode)->ip_attr;
+	OCFS2_I(t_inode)->ip_dyn_features = OCFS2_I(s_inode)->ip_dyn_features;
+	spin_unlock(&OCFS2_I(t_inode)->ip_lock);
+	i_size_write(t_inode, size);
+
+	di->i_xattr_inline_size = s_di->i_xattr_inline_size;
+	di->i_clusters = s_di->i_clusters;
+	di->i_size = s_di->i_size;
+	di->i_dyn_features = s_di->i_dyn_features;
+	di->i_attr = s_di->i_attr;
+
+	if (preserve) {
+		di->i_uid = s_di->i_uid;
+		di->i_gid = s_di->i_gid;
+		di->i_mode = s_di->i_mode;
+
+		/*
+		 * update time.
+		 * we want mtime to appear identical to the source and
+		 * update ctime.
+		 */
+		t_inode->i_ctime = CURRENT_TIME;
+
+		di->i_ctime = cpu_to_le64(t_inode->i_ctime.tv_sec);
+		di->i_ctime_nsec = cpu_to_le32(t_inode->i_ctime.tv_nsec);
+
+		t_inode->i_mtime = s_inode->i_mtime;
+		di->i_mtime = s_di->i_mtime;
+		di->i_mtime_nsec = s_di->i_mtime_nsec;
+	}
+
+	ocfs2_journal_dirty(handle, t_bh);
+
+out_commit:
+	ocfs2_commit_trans(OCFS2_SB(t_inode->i_sb), handle);
+	return ret;
+}
+
+static int ocfs2_create_reflink_node(struct inode *s_inode,
+				     struct buffer_head *s_bh,
+				     struct inode *t_inode,
+				     struct buffer_head *t_bh,
+				     bool preserve)
+{
+	int ret;
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_cached_dealloc_ctxt dealloc;
+	struct ocfs2_super *osb = OCFS2_SB(s_inode->i_sb);
+	struct ocfs2_refcount_block *rb;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)s_bh->b_data;
+	struct ocfs2_refcount_tree *ref_tree;
+
+	ocfs2_init_dealloc_ctxt(&dealloc);
+
+	ret = ocfs2_set_refcount_tree(t_inode, t_bh,
+				      le64_to_cpu(di->i_refcount_loc));
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_lock_refcount_tree(osb, le64_to_cpu(di->i_refcount_loc),
+				       1, &ref_tree, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+	rb = (struct ocfs2_refcount_block *)ref_root_bh->b_data;
+
+	ret = ocfs2_duplicate_extent_list(s_inode, t_inode, t_bh,
+					  &ref_tree->rf_ci, ref_root_bh,
+					  &dealloc);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_unlock_refcount;
+	}
+
+	ret = ocfs2_complete_reflink(s_inode, s_bh, t_inode, t_bh, preserve);
+	if (ret)
+		mlog_errno(ret);
+
+out_unlock_refcount:
+	ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
+	brelse(ref_root_bh);
+out:
+	if (ocfs2_dealloc_has_cluster(&dealloc)) {
+		ocfs2_schedule_truncate_log_flush(osb, 1);
+		ocfs2_run_deallocs(osb, &dealloc);
+	}
+
+	return ret;
+}
+
+static int __ocfs2_reflink(struct dentry *old_dentry,
+			   struct buffer_head *old_bh,
+			   struct inode *new_inode,
+			   bool preserve)
+{
+	int ret;
+	struct inode *inode = old_dentry->d_inode;
+	struct buffer_head *new_bh = NULL;
+
+	ret = filemap_fdatawrite(inode->i_mapping);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_attach_refcount_tree(inode, old_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mutex_lock(&new_inode->i_mutex);
+	ret = ocfs2_inode_lock(new_inode, &new_bh, 1);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_unlock;
+	}
+
+	ret = ocfs2_create_reflink_node(inode, old_bh,
+					new_inode, new_bh, preserve);
+	if (ret) {
+		mlog_errno(ret);
+		goto inode_unlock;
+	}
+
+	if (OCFS2_I(inode)->ip_dyn_features & OCFS2_HAS_XATTR_FL) {
+		ret = ocfs2_reflink_xattrs(inode, old_bh,
+					   new_inode, new_bh,
+					   preserve);
+		if (ret)
+			mlog_errno(ret);
+	}
+inode_unlock:
+	ocfs2_inode_unlock(new_inode, 1);
+	brelse(new_bh);
+out_unlock:
+	mutex_unlock(&new_inode->i_mutex);
+out:
+	if (!ret) {
+		ret = filemap_fdatawait(inode->i_mapping);
+		if (ret)
+			mlog_errno(ret);
+	}
+	return ret;
+}
+
+static int ocfs2_reflink(struct dentry *old_dentry, struct inode *dir,
+			 struct dentry *new_dentry, bool preserve)
+{
+	int error;
+	struct inode *inode = old_dentry->d_inode;
+	struct buffer_head *old_bh = NULL;
+	struct inode *new_orphan_inode = NULL;
+
+	if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb)))
+		return -EOPNOTSUPP;
+
+	error = ocfs2_create_inode_in_orphan(dir, inode->i_mode,
+					     &new_orphan_inode);
+	if (error) {
+		mlog_errno(error);
+		goto out;
+	}
+
+	error = ocfs2_inode_lock(inode, &old_bh, 1);
+	if (error) {
+		mlog_errno(error);
+		goto out;
+	}
+
+	down_write(&OCFS2_I(inode)->ip_xattr_sem);
+	down_write(&OCFS2_I(inode)->ip_alloc_sem);
+	error = __ocfs2_reflink(old_dentry, old_bh,
+				new_orphan_inode, preserve);
+	up_write(&OCFS2_I(inode)->ip_alloc_sem);
+	up_write(&OCFS2_I(inode)->ip_xattr_sem);
+
+	ocfs2_inode_unlock(inode, 1);
+	brelse(old_bh);
+
+	if (error) {
+		mlog_errno(error);
+		goto out;
+	}
+
+	/* If the security isn't preserved, we need to re-initialize them. */
+	if (!preserve) {
+		error = ocfs2_init_security_and_acl(dir, new_orphan_inode);
+		if (error)
+			mlog_errno(error);
+	}
+out:
+	if (!error) {
+		error = ocfs2_mv_orphaned_inode_to_new(dir, new_orphan_inode,
+						       new_dentry);
+		if (error)
+			mlog_errno(error);
+	}
+
+	if (new_orphan_inode) {
+		/*
+		 * We need to open_unlock the inode no matter whether we
+		 * succeed or not, so that other nodes can delete it later.
+		 */
+		ocfs2_open_unlock(new_orphan_inode);
+		if (error)
+			iput(new_orphan_inode);
+	}
+
+	return error;
+}
+
+/*
+ * Below here are the bits used by OCFS2_IOC_REFLINK() to fake
+ * sys_reflink().  This will go away when vfs_reflink() exists in
+ * fs/namei.c.
+ */
+
+/* copied from may_create in VFS. */
+static inline int ocfs2_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);
+}
+
+/* copied from user_path_parent. */
+static int ocfs2_user_path_parent(const char __user *path,
+				  struct nameidata *nd, char **name)
+{
+	char *s = getname(path);
+	int error;
+
+	if (IS_ERR(s))
+		return PTR_ERR(s);
+
+	error = path_lookup(s, LOOKUP_PARENT, nd);
+	if (error)
+		putname(s);
+	else
+		*name = s;
+
+	return error;
+}
+
+/**
+ * ocfs2_vfs_reflink - Create a reference-counted link
+ *
+ * @old_dentry:        source dentry + inode
+ * @dir:       directory to create the target
+ * @new_dentry:        target dentry
+ * @preserve:  if true, preserve all file attributes
+ */
+int ocfs2_vfs_reflink(struct dentry *old_dentry, struct inode *dir,
+		      struct dentry *new_dentry, bool preserve)
+{
+	struct inode *inode = old_dentry->d_inode;
+	int error;
+
+	if (!inode)
+		return -ENOENT;
+
+	error = ocfs2_may_create(dir, new_dentry);
+	if (error)
+		return error;
+
+	if (dir->i_sb != inode->i_sb)
+		return -EXDEV;
+
+	/*
+	 * A reflink to an append-only or immutable file cannot be created.
+	 */
+	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
+		return -EPERM;
+
+	/* Only regular files can be reflinked. */
+	if (!S_ISREG(inode->i_mode))
+		return -EPERM;
+
+	/*
+	 * If the caller wants to preserve ownership, they require the
+	 * rights to do so.
+	 */
+	if (preserve) {
+		if ((current_fsuid() != inode->i_uid) && !capable(CAP_CHOWN))
+			return -EPERM;
+		if (!in_group_p(inode->i_gid) && !capable(CAP_CHOWN))
+			return -EPERM;
+	}
+
+	/*
+	 * If the caller is modifying any aspect of the attributes, they
+	 * are not creating a snapshot.  They need read permission on the
+	 * file.
+	 */
+	if (!preserve) {
+		error = inode_permission(inode, MAY_READ);
+		if (error)
+			return error;
+	}
+
+	mutex_lock(&inode->i_mutex);
+	vfs_dq_init(dir);
+	error = ocfs2_reflink(old_dentry, dir, new_dentry, preserve);
+	mutex_unlock(&inode->i_mutex);
+	if (!error)
+		fsnotify_create(dir, new_dentry);
+	return error;
+}
+/*
+ * Most codes are copied from sys_linkat.
+ */
+int ocfs2_reflink_ioctl(struct inode *inode,
+			const char __user *oldname,
+			const char __user *newname,
+			bool preserve)
+{
+	struct dentry *new_dentry;
+	struct nameidata nd;
+	struct path old_path;
+	int error;
+	char *to = NULL;
+
+	if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb)))
+		return -EOPNOTSUPP;
+
+	error = user_path_at(AT_FDCWD, oldname, 0, &old_path);
+	if (error) {
+		mlog_errno(error);
+		return error;
+	}
+
+	error = ocfs2_user_path_parent(newname, &nd, &to);
+	if (error) {
+		mlog_errno(error);
+		goto out;
+	}
+
+	error = -EXDEV;
+	if (old_path.mnt != nd.path.mnt)
+		goto out_release;
+	new_dentry = lookup_create(&nd, 0);
+	error = PTR_ERR(new_dentry);
+	if (IS_ERR(new_dentry)) {
+		mlog_errno(error);
+		goto out_unlock;
+	}
+
+	error = mnt_want_write(nd.path.mnt);
+	if (error) {
+		mlog_errno(error);
+		goto out_dput;
+	}
+
+	error = ocfs2_vfs_reflink(old_path.dentry,
+				  nd.path.dentry->d_inode,
+				  new_dentry, preserve);
+	mnt_drop_write(nd.path.mnt);
+out_dput:
+	dput(new_dentry);
+out_unlock:
+	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
+out_release:
+	path_put(&nd.path);
+	putname(to);
+out:
+	path_put(&old_path);
+
+	return error;
+}
diff --git a/fs/ocfs2/refcounttree.h b/fs/ocfs2/refcounttree.h
new file mode 100644
index 00000000000..c1d19b1d3ec
--- /dev/null
+++ b/fs/ocfs2/refcounttree.h
@@ -0,0 +1,106 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * refcounttree.h
+ *
+ * Copyright (C) 2009 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_REFCOUNTTREE_H
+#define OCFS2_REFCOUNTTREE_H
+
+struct ocfs2_refcount_tree {
+	struct rb_node rf_node;
+	u64 rf_blkno;
+	u32 rf_generation;
+	struct rw_semaphore rf_sem;
+	struct ocfs2_lock_res rf_lockres;
+	struct kref rf_getcnt;
+	int rf_removed;
+
+	/* the following 4 fields are used by caching_info. */
+	struct ocfs2_caching_info rf_ci;
+	spinlock_t rf_lock;
+	struct mutex rf_io_mutex;
+	struct super_block *rf_sb;
+};
+
+void ocfs2_purge_refcount_trees(struct ocfs2_super *osb);
+int ocfs2_lock_refcount_tree(struct ocfs2_super *osb, u64 ref_blkno, int rw,
+			     struct ocfs2_refcount_tree **tree,
+			     struct buffer_head **ref_bh);
+void ocfs2_unlock_refcount_tree(struct ocfs2_super *osb,
+				struct ocfs2_refcount_tree *tree,
+				int rw);
+
+int ocfs2_decrease_refcount(struct inode *inode,
+			    handle_t *handle, u32 cpos, u32 len,
+			    struct ocfs2_alloc_context *meta_ac,
+			    struct ocfs2_cached_dealloc_ctxt *dealloc,
+			    int delete);
+int ocfs2_prepare_refcount_change_for_del(struct inode *inode,
+					  struct buffer_head *di_bh,
+					  u64 phys_blkno,
+					  u32 clusters,
+					  int *credits,
+					  struct ocfs2_alloc_context **meta_ac);
+int ocfs2_refcount_cow(struct inode *inode, struct buffer_head *di_bh,
+		       u32 cpos, u32 write_len, u32 max_cpos);
+
+typedef int (ocfs2_post_refcount_func)(struct inode *inode,
+				       handle_t *handle,
+				       void *para);
+/*
+ * Some refcount caller need to do more work after we modify the data b-tree
+ * during refcount operation(including CoW and add refcount flag), and make the
+ * transaction complete. So it must give us this structure so that we can do it
+ * within our transaction.
+ *
+ */
+struct ocfs2_post_refcount {
+	int credits;			/* credits it need for journal. */
+	ocfs2_post_refcount_func *func;	/* real function. */
+	void *para;
+};
+
+int ocfs2_refcounted_xattr_delete_need(struct inode *inode,
+				       struct ocfs2_caching_info *ref_ci,
+				       struct buffer_head *ref_root_bh,
+				       struct ocfs2_xattr_value_root *xv,
+				       int *meta_add, int *credits);
+int ocfs2_refcount_cow_xattr(struct inode *inode,
+			     struct ocfs2_dinode *di,
+			     struct ocfs2_xattr_value_buf *vb,
+			     struct ocfs2_refcount_tree *ref_tree,
+			     struct buffer_head *ref_root_bh,
+			     u32 cpos, u32 write_len,
+			     struct ocfs2_post_refcount *post);
+int ocfs2_add_refcount_flag(struct inode *inode,
+			    struct ocfs2_extent_tree *data_et,
+			    struct ocfs2_caching_info *ref_ci,
+			    struct buffer_head *ref_root_bh,
+			    u32 cpos, u32 p_cluster, u32 num_clusters,
+			    struct ocfs2_cached_dealloc_ctxt *dealloc,
+			    struct ocfs2_post_refcount *post);
+int ocfs2_remove_refcount_tree(struct inode *inode, struct buffer_head *di_bh);
+int ocfs2_try_remove_refcount_tree(struct inode *inode,
+				   struct buffer_head *di_bh);
+int ocfs2_increase_refcount(handle_t *handle,
+			    struct ocfs2_caching_info *ci,
+			    struct buffer_head *ref_root_bh,
+			    u64 cpos, u32 len,
+			    struct ocfs2_alloc_context *meta_ac,
+			    struct ocfs2_cached_dealloc_ctxt *dealloc);
+int ocfs2_reflink_ioctl(struct inode *inode,
+			const char __user *oldname,
+			const char __user *newname,
+			bool preserve);
+#endif /* OCFS2_REFCOUNTTREE_H */
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index 424adaa5f90..3c3d673a4d2 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_gd(handle, bm_inode, group_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_gd(handle, INODE_CACHE(bm_inode),
+				      group_bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -141,7 +141,7 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
 	}
 
 	/* update the inode accordingly. */
-	ret = ocfs2_journal_access_di(handle, bm_inode, bm_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(bm_inode), bm_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -514,7 +514,7 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
 		goto out_unlock;
 	}
 
-	ocfs2_set_new_buffer_uptodate(inode, group_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), group_bh);
 
 	ret = ocfs2_verify_group_and_input(main_bm_inode, fe, input, group_bh);
 	if (ret) {
@@ -536,8 +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_gd(handle, main_bm_inode, group_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_gd(handle, INODE_CACHE(main_bm_inode),
+				      group_bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out_commit;
@@ -552,8 +552,8 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
 		goto out_commit;
 	}
 
-	ret = ocfs2_journal_access_di(handle, main_bm_inode, main_bm_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(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 40661e7824e..bfbd7e9e949 100644
--- a/fs/ocfs2/slot_map.c
+++ b/fs/ocfs2/slot_map.c
@@ -150,8 +150,8 @@ int ocfs2_refresh_slot_info(struct ocfs2_super *osb)
 	 * be !NULL.  Thus, ocfs2_read_blocks() will ignore blocknr.  If
 	 * 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, NULL);
+	ret = ocfs2_read_blocks(INODE_CACHE(si->si_inode), -1, si->si_blocks,
+				si->si_bh, OCFS2_BH_IGNORE_CACHE, NULL);
 	if (ret == 0) {
 		spin_lock(&osb->osb_lock);
 		ocfs2_update_slot_info(si);
@@ -213,7 +213,7 @@ static int ocfs2_update_disk_slot(struct ocfs2_super *osb,
 		ocfs2_update_disk_slot_old(si, slot_num, &bh);
 	spin_unlock(&osb->osb_lock);
 
-	status = ocfs2_write_block(osb, bh, si->si_inode);
+	status = ocfs2_write_block(osb, bh, INODE_CACHE(si->si_inode));
 	if (status < 0)
 		mlog_errno(status);
 
@@ -404,8 +404,8 @@ static int ocfs2_map_slot_buffers(struct ocfs2_super *osb,
 		     (unsigned long long)blkno);
 
 		bh = NULL;  /* Acquire a fresh bh */
-		status = ocfs2_read_blocks(si->si_inode, blkno, 1, &bh,
-					   OCFS2_BH_IGNORE_CACHE, NULL);
+		status = ocfs2_read_blocks(INODE_CACHE(si->si_inode), blkno,
+					   1, &bh, 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 73a16d4666d..c30b644d957 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -310,7 +310,7 @@ int ocfs2_read_group_descriptor(struct inode *inode, struct ocfs2_dinode *di,
 	int rc;
 	struct buffer_head *tmp = *bh;
 
-	rc = ocfs2_read_block(inode, gd_blkno, &tmp,
+	rc = ocfs2_read_block(INODE_CACHE(inode), gd_blkno, &tmp,
 			      ocfs2_validate_group_descriptor);
 	if (rc)
 		goto out;
@@ -352,7 +352,7 @@ static int ocfs2_block_group_fill(handle_t *handle,
 	}
 
 	status = ocfs2_journal_access_gd(handle,
-					 alloc_inode,
+					 INODE_CACHE(alloc_inode),
 					 bg_bh,
 					 OCFS2_JOURNAL_ACCESS_CREATE);
 	if (status < 0) {
@@ -476,7 +476,7 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
 		mlog_errno(status);
 		goto bail;
 	}
-	ocfs2_set_new_buffer_uptodate(alloc_inode, bg_bh);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(alloc_inode), bg_bh);
 
 	status = ocfs2_block_group_fill(handle,
 					alloc_inode,
@@ -491,7 +491,7 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
 
 	bg = (struct ocfs2_group_desc *) bg_bh->b_data;
 
-	status = ocfs2_journal_access_di(handle, alloc_inode,
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(alloc_inode),
 					 bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1033,7 +1033,7 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
 		journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
 
 	status = ocfs2_journal_access_gd(handle,
-					 alloc_inode,
+					 INODE_CACHE(alloc_inode),
 					 group_bh,
 					 journal_type);
 	if (status < 0) {
@@ -1106,7 +1106,8 @@ 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_gd(handle, alloc_inode, prev_bg_bh,
+	status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode),
+					 prev_bg_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -1121,8 +1122,8 @@ static int ocfs2_relink_block_group(handle_t *handle,
 		goto out_rollback;
 	}
 
-	status = ocfs2_journal_access_gd(handle, alloc_inode, bg_bh,
-					 OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode),
+					 bg_bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_rollback;
@@ -1136,8 +1137,8 @@ static int ocfs2_relink_block_group(handle_t *handle,
 		goto out_rollback;
 	}
 
-	status = ocfs2_journal_access_di(handle, alloc_inode, fe_bh,
-					 OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(alloc_inode),
+					 fe_bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto out_rollback;
@@ -1288,7 +1289,7 @@ 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_di(handle, inode, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -1461,7 +1462,7 @@ 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_di(handle,
-					 alloc_inode,
+					 INODE_CACHE(alloc_inode),
 					 ac->ac_bh,
 					 OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
@@ -1907,8 +1908,8 @@ static inline int ocfs2_block_group_clear_bits(handle_t *handle,
 	if (ocfs2_is_cluster_bitmap(alloc_inode))
 		journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
 
-	status = ocfs2_journal_access_gd(handle, alloc_inode, group_bh,
-					 journal_type);
+	status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode),
+					 group_bh, journal_type);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -1993,8 +1994,8 @@ int ocfs2_free_suballoc_bits(handle_t *handle,
 		goto bail;
 	}
 
-	status = ocfs2_journal_access_di(handle, alloc_inode, alloc_bh,
-					 OCFS2_JOURNAL_ACCESS_WRITE);
+	status = ocfs2_journal_access_di(handle, INODE_CACHE(alloc_inode),
+					 alloc_bh, OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
@@ -2151,7 +2152,7 @@ int ocfs2_lock_allocators(struct inode *inode,
 
 	BUG_ON(clusters_to_add != 0 && data_ac == NULL);
 
-	num_free_extents = ocfs2_num_free_extents(osb, inode, et);
+	num_free_extents = ocfs2_num_free_extents(osb, et);
 	if (num_free_extents < 0) {
 		ret = num_free_extents;
 		mlog_errno(ret);
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index a3f8871d21f..24feb449a1d 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -69,6 +69,7 @@
 #include "ver.h"
 #include "xattr.h"
 #include "quota.h"
+#include "refcounttree.h"
 
 #include "buffer_head_io.h"
 
@@ -965,7 +966,7 @@ static int ocfs2_quota_off(struct super_block *sb, int type, int remount)
 	return vfs_quota_disable(sb, type, DQUOT_LIMITS_ENABLED);
 }
 
-static struct quotactl_ops ocfs2_quotactl_ops = {
+static const struct quotactl_ops ocfs2_quotactl_ops = {
 	.quota_on	= ocfs2_quota_on,
 	.quota_off	= ocfs2_quota_off,
 	.quota_sync	= vfs_quota_sync,
@@ -1668,8 +1669,6 @@ static void ocfs2_inode_init_once(void *data)
 	spin_lock_init(&oi->ip_lock);
 	ocfs2_extent_map_init(&oi->vfs_inode);
 	INIT_LIST_HEAD(&oi->ip_io_markers);
-	oi->ip_created_trans = 0;
-	oi->ip_last_trans = 0;
 	oi->ip_dir_start_lookup = 0;
 
 	init_rwsem(&oi->ip_alloc_sem);
@@ -1683,7 +1682,8 @@ static void ocfs2_inode_init_once(void *data)
 	ocfs2_lock_res_init_once(&oi->ip_inode_lockres);
 	ocfs2_lock_res_init_once(&oi->ip_open_lockres);
 
-	ocfs2_metadata_cache_init(&oi->vfs_inode);
+	ocfs2_metadata_cache_init(INODE_CACHE(&oi->vfs_inode),
+				  &ocfs2_inode_caching_ops);
 
 	inode_init_once(&oi->vfs_inode);
 }
@@ -1859,6 +1859,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	ocfs2_sync_blockdev(sb);
 
+	ocfs2_purge_refcount_trees(osb);
+
 	/* No cluster connection means we've failed during mount, so skip
 	 * all the steps which depended on that to complete. */
 	if (osb->cconn) {
@@ -2065,6 +2067,8 @@ static int ocfs2_initialize_super(struct super_block *sb,
 		goto bail;
 	}
 
+	osb->osb_rf_lock_tree = RB_ROOT;
+
 	osb->s_feature_compat =
 		le32_to_cpu(OCFS2_RAW_SB(di)->s_feature_compat);
 	osb->s_feature_ro_compat =
@@ -2490,7 +2494,8 @@ void __ocfs2_abort(struct super_block* sb,
 	/* Force a panic(). This stinks, but it's better than letting
 	 * things continue without having a proper hard readonly
 	 * here. */
-	OCFS2_SB(sb)->s_mount_opt |= OCFS2_MOUNT_ERRORS_PANIC;
+	if (!ocfs2_mount_local(OCFS2_SB(sb)))
+		OCFS2_SB(sb)->s_mount_opt |= OCFS2_MOUNT_ERRORS_PANIC;
 	ocfs2_handle_error(sb);
 }
 
diff --git a/fs/ocfs2/uptodate.c b/fs/ocfs2/uptodate.c
index 187b99ff036..b6284f235d2 100644
--- a/fs/ocfs2/uptodate.c
+++ b/fs/ocfs2/uptodate.c
@@ -75,15 +75,77 @@ struct ocfs2_meta_cache_item {
 
 static struct kmem_cache *ocfs2_uptodate_cachep = NULL;
 
-void ocfs2_metadata_cache_init(struct inode *inode)
+u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci)
 {
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
-	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
+	BUG_ON(!ci || !ci->ci_ops);
 
-	oi->ip_flags |= OCFS2_INODE_CACHE_INLINE;
+	return ci->ci_ops->co_owner(ci);
+}
+
+struct super_block *ocfs2_metadata_cache_get_super(struct ocfs2_caching_info *ci)
+{
+	BUG_ON(!ci || !ci->ci_ops);
+
+	return ci->ci_ops->co_get_super(ci);
+}
+
+static void ocfs2_metadata_cache_lock(struct ocfs2_caching_info *ci)
+{
+	BUG_ON(!ci || !ci->ci_ops);
+
+	ci->ci_ops->co_cache_lock(ci);
+}
+
+static void ocfs2_metadata_cache_unlock(struct ocfs2_caching_info *ci)
+{
+	BUG_ON(!ci || !ci->ci_ops);
+
+	ci->ci_ops->co_cache_unlock(ci);
+}
+
+void ocfs2_metadata_cache_io_lock(struct ocfs2_caching_info *ci)
+{
+	BUG_ON(!ci || !ci->ci_ops);
+
+	ci->ci_ops->co_io_lock(ci);
+}
+
+void ocfs2_metadata_cache_io_unlock(struct ocfs2_caching_info *ci)
+{
+	BUG_ON(!ci || !ci->ci_ops);
+
+	ci->ci_ops->co_io_unlock(ci);
+}
+
+
+static void ocfs2_metadata_cache_reset(struct ocfs2_caching_info *ci,
+				       int clear)
+{
+	ci->ci_flags |= OCFS2_CACHE_FL_INLINE;
 	ci->ci_num_cached = 0;
+
+	if (clear) {
+		ci->ci_created_trans = 0;
+		ci->ci_last_trans = 0;
+	}
+}
+
+void ocfs2_metadata_cache_init(struct ocfs2_caching_info *ci,
+			       const struct ocfs2_caching_operations *ops)
+{
+	BUG_ON(!ops);
+
+	ci->ci_ops = ops;
+	ocfs2_metadata_cache_reset(ci, 1);
 }
 
+void ocfs2_metadata_cache_exit(struct ocfs2_caching_info *ci)
+{
+	ocfs2_metadata_cache_purge(ci);
+	ocfs2_metadata_cache_reset(ci, 1);
+}
+
+
 /* No lock taken here as 'root' is not expected to be visible to other
  * processes. */
 static unsigned int ocfs2_purge_copied_metadata_tree(struct rb_root *root)
@@ -112,19 +174,20 @@ static unsigned int ocfs2_purge_copied_metadata_tree(struct rb_root *root)
  * This function is a few more lines longer than necessary due to some
  * accounting done here, but I think it's worth tracking down those
  * bugs sooner -- Mark */
-void ocfs2_metadata_cache_purge(struct inode *inode)
+void ocfs2_metadata_cache_purge(struct ocfs2_caching_info *ci)
 {
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 	unsigned int tree, to_purge, purged;
-	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 	struct rb_root root = RB_ROOT;
 
-	spin_lock(&oi->ip_lock);
-	tree = !(oi->ip_flags & OCFS2_INODE_CACHE_INLINE);
+	BUG_ON(!ci || !ci->ci_ops);
+
+	ocfs2_metadata_cache_lock(ci);
+	tree = !(ci->ci_flags & OCFS2_CACHE_FL_INLINE);
 	to_purge = ci->ci_num_cached;
 
-	mlog(0, "Purge %u %s items from Inode %llu\n", to_purge,
-	     tree ? "array" : "tree", (unsigned long long)oi->ip_blkno);
+	mlog(0, "Purge %u %s items from Owner %llu\n", to_purge,
+	     tree ? "array" : "tree",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci));
 
 	/* If we're a tree, save off the root so that we can safely
 	 * initialize the cache. We do the work to free tree members
@@ -132,16 +195,17 @@ void ocfs2_metadata_cache_purge(struct inode *inode)
 	if (tree)
 		root = ci->ci_cache.ci_tree;
 
-	ocfs2_metadata_cache_init(inode);
-	spin_unlock(&oi->ip_lock);
+	ocfs2_metadata_cache_reset(ci, 0);
+	ocfs2_metadata_cache_unlock(ci);
 
 	purged = ocfs2_purge_copied_metadata_tree(&root);
 	/* If possible, track the number wiped so that we can more
 	 * easily detect counting errors. Unfortunately, this is only
 	 * meaningful for trees. */
 	if (tree && purged != to_purge)
-		mlog(ML_ERROR, "Inode %llu, count = %u, purged = %u\n",
-		     (unsigned long long)oi->ip_blkno, to_purge, purged);
+		mlog(ML_ERROR, "Owner %llu, count = %u, purged = %u\n",
+		     (unsigned long long)ocfs2_metadata_cache_owner(ci),
+		     to_purge, purged);
 }
 
 /* Returns the index in the cache array, -1 if not found.
@@ -182,27 +246,25 @@ ocfs2_search_cache_tree(struct ocfs2_caching_info *ci,
 	return NULL;
 }
 
-static int ocfs2_buffer_cached(struct ocfs2_inode_info *oi,
+static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
 			       struct buffer_head *bh)
 {
 	int index = -1;
 	struct ocfs2_meta_cache_item *item = NULL;
 
-	spin_lock(&oi->ip_lock);
+	ocfs2_metadata_cache_lock(ci);
 
-	mlog(0, "Inode %llu, query block %llu (inline = %u)\n",
-	     (unsigned long long)oi->ip_blkno,
+	mlog(0, "Owner %llu, query block %llu (inline = %u)\n",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
 	     (unsigned long long) bh->b_blocknr,
-	     !!(oi->ip_flags & OCFS2_INODE_CACHE_INLINE));
+	     !!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
 
-	if (oi->ip_flags & OCFS2_INODE_CACHE_INLINE)
-		index = ocfs2_search_cache_array(&oi->ip_metadata_cache,
-						 bh->b_blocknr);
+	if (ci->ci_flags & OCFS2_CACHE_FL_INLINE)
+		index = ocfs2_search_cache_array(ci, bh->b_blocknr);
 	else
-		item = ocfs2_search_cache_tree(&oi->ip_metadata_cache,
-					       bh->b_blocknr);
+		item = ocfs2_search_cache_tree(ci, bh->b_blocknr);
 
-	spin_unlock(&oi->ip_lock);
+	ocfs2_metadata_cache_unlock(ci);
 
 	mlog(0, "index = %d, item = %p\n", index, item);
 
@@ -214,7 +276,7 @@ static int ocfs2_buffer_cached(struct ocfs2_inode_info *oi,
  * 
  * This can be called under lock_buffer()
  */
-int ocfs2_buffer_uptodate(struct inode *inode,
+int ocfs2_buffer_uptodate(struct ocfs2_caching_info *ci,
 			  struct buffer_head *bh)
 {
 	/* Doesn't matter if the bh is in our cache or not -- if it's
@@ -230,24 +292,24 @@ int ocfs2_buffer_uptodate(struct inode *inode,
 
 	/* Ok, locally the buffer is marked as up to date, now search
 	 * our cache to see if we can trust that. */
-	return ocfs2_buffer_cached(OCFS2_I(inode), bh);
+	return ocfs2_buffer_cached(ci, bh);
 }
 
-/* 
+/*
  * Determine whether a buffer is currently out on a read-ahead request.
- * ip_io_sem should be held to serialize submitters with the logic here.
+ * ci_io_sem should be held to serialize submitters with the logic here.
  */
-int ocfs2_buffer_read_ahead(struct inode *inode,
+int ocfs2_buffer_read_ahead(struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh)
 {
-	return buffer_locked(bh) && ocfs2_buffer_cached(OCFS2_I(inode), bh);
+	return buffer_locked(bh) && ocfs2_buffer_cached(ci, bh);
 }
 
 /* Requires ip_lock */
 static void ocfs2_append_cache_array(struct ocfs2_caching_info *ci,
 				     sector_t block)
 {
-	BUG_ON(ci->ci_num_cached >= OCFS2_INODE_MAX_CACHE_ARRAY);
+	BUG_ON(ci->ci_num_cached >= OCFS2_CACHE_INFO_MAX_ARRAY);
 
 	mlog(0, "block %llu takes position %u\n", (unsigned long long) block,
 	     ci->ci_num_cached);
@@ -292,66 +354,64 @@ static void __ocfs2_insert_cache_tree(struct ocfs2_caching_info *ci,
 	ci->ci_num_cached++;
 }
 
-static inline int ocfs2_insert_can_use_array(struct ocfs2_inode_info *oi,
-					     struct ocfs2_caching_info *ci)
+/* co_cache_lock() must be held */
+static inline int ocfs2_insert_can_use_array(struct ocfs2_caching_info *ci)
 {
-	assert_spin_locked(&oi->ip_lock);
-
-	return (oi->ip_flags & OCFS2_INODE_CACHE_INLINE) &&
-		(ci->ci_num_cached < OCFS2_INODE_MAX_CACHE_ARRAY);
+	return (ci->ci_flags & OCFS2_CACHE_FL_INLINE) &&
+		(ci->ci_num_cached < OCFS2_CACHE_INFO_MAX_ARRAY);
 }
 
-/* tree should be exactly OCFS2_INODE_MAX_CACHE_ARRAY wide. NULL the
+/* tree should be exactly OCFS2_CACHE_INFO_MAX_ARRAY wide. NULL the
  * pointers in tree after we use them - this allows caller to detect
- * when to free in case of error. */
-static void ocfs2_expand_cache(struct ocfs2_inode_info *oi,
+ * when to free in case of error.
+ *
+ * The co_cache_lock() must be held. */
+static void ocfs2_expand_cache(struct ocfs2_caching_info *ci,
 			       struct ocfs2_meta_cache_item **tree)
 {
 	int i;
-	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 
-	mlog_bug_on_msg(ci->ci_num_cached != OCFS2_INODE_MAX_CACHE_ARRAY,
-			"Inode %llu, num cached = %u, should be %u\n",
-			(unsigned long long)oi->ip_blkno, ci->ci_num_cached,
-			OCFS2_INODE_MAX_CACHE_ARRAY);
-	mlog_bug_on_msg(!(oi->ip_flags & OCFS2_INODE_CACHE_INLINE),
-			"Inode %llu not marked as inline anymore!\n",
-			(unsigned long long)oi->ip_blkno);
-	assert_spin_locked(&oi->ip_lock);
+	mlog_bug_on_msg(ci->ci_num_cached != OCFS2_CACHE_INFO_MAX_ARRAY,
+			"Owner %llu, num cached = %u, should be %u\n",
+			(unsigned long long)ocfs2_metadata_cache_owner(ci),
+			ci->ci_num_cached, OCFS2_CACHE_INFO_MAX_ARRAY);
+	mlog_bug_on_msg(!(ci->ci_flags & OCFS2_CACHE_FL_INLINE),
+			"Owner %llu not marked as inline anymore!\n",
+			(unsigned long long)ocfs2_metadata_cache_owner(ci));
 
 	/* Be careful to initialize the tree members *first* because
 	 * once the ci_tree is used, the array is junk... */
-	for(i = 0; i < OCFS2_INODE_MAX_CACHE_ARRAY; i++)
+	for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
 		tree[i]->c_block = ci->ci_cache.ci_array[i];
 
-	oi->ip_flags &= ~OCFS2_INODE_CACHE_INLINE;
+	ci->ci_flags &= ~OCFS2_CACHE_FL_INLINE;
 	ci->ci_cache.ci_tree = RB_ROOT;
 	/* this will be set again by __ocfs2_insert_cache_tree */
 	ci->ci_num_cached = 0;
 
-	for(i = 0; i < OCFS2_INODE_MAX_CACHE_ARRAY; i++) {
+	for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
 		__ocfs2_insert_cache_tree(ci, tree[i]);
 		tree[i] = NULL;
 	}
 
 	mlog(0, "Expanded %llu to a tree cache: flags 0x%x, num = %u\n",
-	     (unsigned long long)oi->ip_blkno, oi->ip_flags, ci->ci_num_cached);
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
+	     ci->ci_flags, ci->ci_num_cached);
 }
 
 /* Slow path function - memory allocation is necessary. See the
  * comment above ocfs2_set_buffer_uptodate for more information. */
-static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi,
+static void __ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
 					sector_t block,
 					int expand_tree)
 {
 	int i;
-	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 	struct ocfs2_meta_cache_item *new = NULL;
-	struct ocfs2_meta_cache_item *tree[OCFS2_INODE_MAX_CACHE_ARRAY] =
+	struct ocfs2_meta_cache_item *tree[OCFS2_CACHE_INFO_MAX_ARRAY] =
 		{ NULL, };
 
-	mlog(0, "Inode %llu, block %llu, expand = %d\n",
-	     (unsigned long long)oi->ip_blkno,
+	mlog(0, "Owner %llu, block %llu, expand = %d\n",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
 	     (unsigned long long)block, expand_tree);
 
 	new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_NOFS);
@@ -364,7 +424,7 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi,
 	if (expand_tree) {
 		/* Do *not* allocate an array here - the removal code
 		 * has no way of tracking that. */
-		for(i = 0; i < OCFS2_INODE_MAX_CACHE_ARRAY; i++) {
+		for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++) {
 			tree[i] = kmem_cache_alloc(ocfs2_uptodate_cachep,
 						   GFP_NOFS);
 			if (!tree[i]) {
@@ -376,21 +436,21 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi,
 		}
 	}
 
-	spin_lock(&oi->ip_lock);
-	if (ocfs2_insert_can_use_array(oi, ci)) {
+	ocfs2_metadata_cache_lock(ci);
+	if (ocfs2_insert_can_use_array(ci)) {
 		mlog(0, "Someone cleared the tree underneath us\n");
 		/* Ok, items were removed from the cache in between
 		 * locks. Detect this and revert back to the fast path */
 		ocfs2_append_cache_array(ci, block);
-		spin_unlock(&oi->ip_lock);
+		ocfs2_metadata_cache_unlock(ci);
 		goto out_free;
 	}
 
 	if (expand_tree)
-		ocfs2_expand_cache(oi, tree);
+		ocfs2_expand_cache(ci, tree);
 
 	__ocfs2_insert_cache_tree(ci, new);
-	spin_unlock(&oi->ip_lock);
+	ocfs2_metadata_cache_unlock(ci);
 
 	new = NULL;
 out_free:
@@ -400,14 +460,14 @@ out_free:
 	/* If these were used, then ocfs2_expand_cache re-set them to
 	 * NULL for us. */
 	if (tree[0]) {
-		for(i = 0; i < OCFS2_INODE_MAX_CACHE_ARRAY; i++)
+		for (i = 0; i < OCFS2_CACHE_INFO_MAX_ARRAY; i++)
 			if (tree[i])
 				kmem_cache_free(ocfs2_uptodate_cachep,
 						tree[i]);
 	}
 }
 
-/* Item insertion is guarded by ip_io_mutex, so the insertion path takes
+/* Item insertion is guarded by co_io_lock(), so the insertion path takes
  * advantage of this by not rechecking for a duplicate insert during
  * the slow case. Additionally, if the cache needs to be bumped up to
  * a tree, the code will not recheck after acquiring the lock --
@@ -425,59 +485,55 @@ out_free:
  * Readahead buffers can be passed in here before the I/O request is
  * completed.
  */
-void ocfs2_set_buffer_uptodate(struct inode *inode,
+void ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
 			       struct buffer_head *bh)
 {
 	int expand;
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
-	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 
 	/* The block may very well exist in our cache already, so avoid
 	 * doing any more work in that case. */
-	if (ocfs2_buffer_cached(oi, bh))
+	if (ocfs2_buffer_cached(ci, bh))
 		return;
 
-	mlog(0, "Inode %llu, inserting block %llu\n",
-	     (unsigned long long)oi->ip_blkno,
+	mlog(0, "Owner %llu, inserting block %llu\n",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
 	     (unsigned long long)bh->b_blocknr);
 
 	/* No need to recheck under spinlock - insertion is guarded by
-	 * ip_io_mutex */
-	spin_lock(&oi->ip_lock);
-	if (ocfs2_insert_can_use_array(oi, ci)) {
+	 * co_io_lock() */
+	ocfs2_metadata_cache_lock(ci);
+	if (ocfs2_insert_can_use_array(ci)) {
 		/* Fast case - it's an array and there's a free
 		 * spot. */
 		ocfs2_append_cache_array(ci, bh->b_blocknr);
-		spin_unlock(&oi->ip_lock);
+		ocfs2_metadata_cache_unlock(ci);
 		return;
 	}
 
 	expand = 0;
-	if (oi->ip_flags & OCFS2_INODE_CACHE_INLINE) {
+	if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
 		/* We need to bump things up to a tree. */
 		expand = 1;
 	}
-	spin_unlock(&oi->ip_lock);
+	ocfs2_metadata_cache_unlock(ci);
 
-	__ocfs2_set_buffer_uptodate(oi, bh->b_blocknr, expand);
+	__ocfs2_set_buffer_uptodate(ci, bh->b_blocknr, expand);
 }
 
 /* Called against a newly allocated buffer. Most likely nobody should
  * be able to read this sort of metadata while it's still being
- * allocated, but this is careful to take ip_io_mutex anyway. */
-void ocfs2_set_new_buffer_uptodate(struct inode *inode,
+ * allocated, but this is careful to take co_io_lock() anyway. */
+void ocfs2_set_new_buffer_uptodate(struct ocfs2_caching_info *ci,
 				   struct buffer_head *bh)
 {
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
-
 	/* This should definitely *not* exist in our cache */
-	BUG_ON(ocfs2_buffer_cached(oi, bh));
+	BUG_ON(ocfs2_buffer_cached(ci, bh));
 
 	set_buffer_uptodate(bh);
 
-	mutex_lock(&oi->ip_io_mutex);
-	ocfs2_set_buffer_uptodate(inode, bh);
-	mutex_unlock(&oi->ip_io_mutex);
+	ocfs2_metadata_cache_io_lock(ci);
+	ocfs2_set_buffer_uptodate(ci, bh);
+	ocfs2_metadata_cache_io_unlock(ci);
 }
 
 /* Requires ip_lock. */
@@ -487,7 +543,7 @@ static void ocfs2_remove_metadata_array(struct ocfs2_caching_info *ci,
 	sector_t *array = ci->ci_cache.ci_array;
 	int bytes;
 
-	BUG_ON(index < 0 || index >= OCFS2_INODE_MAX_CACHE_ARRAY);
+	BUG_ON(index < 0 || index >= OCFS2_CACHE_INFO_MAX_ARRAY);
 	BUG_ON(index >= ci->ci_num_cached);
 	BUG_ON(!ci->ci_num_cached);
 
@@ -515,21 +571,19 @@ static void ocfs2_remove_metadata_tree(struct ocfs2_caching_info *ci,
 	ci->ci_num_cached--;
 }
 
-static void ocfs2_remove_block_from_cache(struct inode *inode,
+static void ocfs2_remove_block_from_cache(struct ocfs2_caching_info *ci,
 					  sector_t block)
 {
 	int index;
 	struct ocfs2_meta_cache_item *item = NULL;
-	struct ocfs2_inode_info *oi = OCFS2_I(inode);
-	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 
-	spin_lock(&oi->ip_lock);
-	mlog(0, "Inode %llu, remove %llu, items = %u, array = %u\n",
-	     (unsigned long long)oi->ip_blkno,
+	ocfs2_metadata_cache_lock(ci);
+	mlog(0, "Owner %llu, remove %llu, items = %u, array = %u\n",
+	     (unsigned long long)ocfs2_metadata_cache_owner(ci),
 	     (unsigned long long) block, ci->ci_num_cached,
-	     oi->ip_flags & OCFS2_INODE_CACHE_INLINE);
+	     ci->ci_flags & OCFS2_CACHE_FL_INLINE);
 
-	if (oi->ip_flags & OCFS2_INODE_CACHE_INLINE) {
+	if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
 		index = ocfs2_search_cache_array(ci, block);
 		if (index != -1)
 			ocfs2_remove_metadata_array(ci, index);
@@ -538,7 +592,7 @@ static void ocfs2_remove_block_from_cache(struct inode *inode,
 		if (item)
 			ocfs2_remove_metadata_tree(ci, item);
 	}
-	spin_unlock(&oi->ip_lock);
+	ocfs2_metadata_cache_unlock(ci);
 
 	if (item)
 		kmem_cache_free(ocfs2_uptodate_cachep, item);
@@ -549,23 +603,24 @@ static void ocfs2_remove_block_from_cache(struct inode *inode,
  * bother reverting things to an inlined array in the case of a remove
  * which moves us back under the limit.
  */
-void ocfs2_remove_from_cache(struct inode *inode,
+void ocfs2_remove_from_cache(struct ocfs2_caching_info *ci,
 			     struct buffer_head *bh)
 {
 	sector_t block = bh->b_blocknr;
 
-	ocfs2_remove_block_from_cache(inode, block);
+	ocfs2_remove_block_from_cache(ci, block);
 }
 
 /* Called when we remove xattr clusters from an inode. */
-void ocfs2_remove_xattr_clusters_from_cache(struct inode *inode,
+void ocfs2_remove_xattr_clusters_from_cache(struct ocfs2_caching_info *ci,
 					    sector_t block,
 					    u32 c_len)
 {
-	unsigned int i, b_len = ocfs2_clusters_to_blocks(inode->i_sb, 1) * c_len;
+	struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
+	unsigned int i, b_len = ocfs2_clusters_to_blocks(sb, 1) * c_len;
 
 	for (i = 0; i < b_len; i++, block++)
-		ocfs2_remove_block_from_cache(inode, block);
+		ocfs2_remove_block_from_cache(ci, block);
 }
 
 int __init init_ocfs2_uptodate_cache(void)
@@ -577,7 +632,7 @@ int __init init_ocfs2_uptodate_cache(void)
 		return -ENOMEM;
 
 	mlog(0, "%u inlined cache items per inode.\n",
-	     OCFS2_INODE_MAX_CACHE_ARRAY);
+	     OCFS2_CACHE_INFO_MAX_ARRAY);
 
 	return 0;
 }
diff --git a/fs/ocfs2/uptodate.h b/fs/ocfs2/uptodate.h
index 531b4b3a0c4..0d826fe2da0 100644
--- a/fs/ocfs2/uptodate.h
+++ b/fs/ocfs2/uptodate.h
@@ -26,24 +26,59 @@
 #ifndef OCFS2_UPTODATE_H
 #define OCFS2_UPTODATE_H
 
+/*
+ * The caching code relies on locking provided by the user of
+ * struct ocfs2_caching_info.  These operations connect that up.
+ */
+struct ocfs2_caching_operations {
+	/*
+	 * A u64 representing the owning structure.  Usually this
+	 * is the block number (i_blkno or whatnot).  This is used so
+	 * that caching log messages can identify the owning structure.
+	 */
+	u64	(*co_owner)(struct ocfs2_caching_info *ci);
+
+	/* The superblock is needed during I/O. */
+	struct super_block *(*co_get_super)(struct ocfs2_caching_info *ci);
+	/*
+	 * Lock and unlock the caching data.  These will not sleep, and
+	 * should probably be spinlocks.
+	 */
+	void	(*co_cache_lock)(struct ocfs2_caching_info *ci);
+	void	(*co_cache_unlock)(struct ocfs2_caching_info *ci);
+
+	/*
+	 * Lock and unlock for disk I/O.  These will sleep, and should
+	 * be mutexes.
+	 */
+	void	(*co_io_lock)(struct ocfs2_caching_info *ci);
+	void	(*co_io_unlock)(struct ocfs2_caching_info *ci);
+};
+
 int __init init_ocfs2_uptodate_cache(void);
 void exit_ocfs2_uptodate_cache(void);
 
-void ocfs2_metadata_cache_init(struct inode *inode);
-void ocfs2_metadata_cache_purge(struct inode *inode);
+void ocfs2_metadata_cache_init(struct ocfs2_caching_info *ci,
+			       const struct ocfs2_caching_operations *ops);
+void ocfs2_metadata_cache_purge(struct ocfs2_caching_info *ci);
+void ocfs2_metadata_cache_exit(struct ocfs2_caching_info *ci);
+
+u64 ocfs2_metadata_cache_owner(struct ocfs2_caching_info *ci);
+void ocfs2_metadata_cache_io_lock(struct ocfs2_caching_info *ci);
+void ocfs2_metadata_cache_io_unlock(struct ocfs2_caching_info *ci);
 
-int ocfs2_buffer_uptodate(struct inode *inode,
+int ocfs2_buffer_uptodate(struct ocfs2_caching_info *ci,
 			  struct buffer_head *bh);
-void ocfs2_set_buffer_uptodate(struct inode *inode,
+void ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
 			       struct buffer_head *bh);
-void ocfs2_set_new_buffer_uptodate(struct inode *inode,
+void ocfs2_set_new_buffer_uptodate(struct ocfs2_caching_info *ci,
 				   struct buffer_head *bh);
-void ocfs2_remove_from_cache(struct inode *inode,
+void ocfs2_remove_from_cache(struct ocfs2_caching_info *ci,
 			     struct buffer_head *bh);
-void ocfs2_remove_xattr_clusters_from_cache(struct inode *inode,
+void ocfs2_remove_xattr_clusters_from_cache(struct ocfs2_caching_info *ci,
 					    sector_t block,
 					    u32 c_len);
-int ocfs2_buffer_read_ahead(struct inode *inode,
+int ocfs2_buffer_read_ahead(struct ocfs2_caching_info *ci,
 			    struct buffer_head *bh);
 
 #endif /* OCFS2_UPTODATE_H */
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index d1a27cda984..fe3419068df 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -55,7 +55,8 @@
 #include "buffer_head_io.h"
 #include "super.h"
 #include "xattr.h"
-
+#include "refcounttree.h"
+#include "acl.h"
 
 struct ocfs2_xattr_def_value_root {
 	struct ocfs2_xattr_value_root	xv;
@@ -140,7 +141,7 @@ struct ocfs2_xattr_search {
 	int not_found;
 };
 
-static int ocfs2_xattr_bucket_get_name_value(struct inode *inode,
+static int ocfs2_xattr_bucket_get_name_value(struct super_block *sb,
 					     struct ocfs2_xattr_header *xh,
 					     int index,
 					     int *block_off,
@@ -157,7 +158,7 @@ static int ocfs2_xattr_index_block_find(struct inode *inode,
 					struct ocfs2_xattr_search *xs);
 
 static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
-					struct ocfs2_xattr_tree_root *xt,
+					struct buffer_head *blk_bh,
 					char *buffer,
 					size_t buffer_size);
 
@@ -170,12 +171,42 @@ static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
 					     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);
+typedef int (xattr_tree_rec_func)(struct inode *inode,
+				  struct buffer_head *root_bh,
+				  u64 blkno, u32 cpos, u32 len, void *para);
+static int ocfs2_iterate_xattr_index_block(struct inode *inode,
+					   struct buffer_head *root_bh,
+					   xattr_tree_rec_func *rec_func,
+					   void *para);
+static int ocfs2_delete_xattr_in_bucket(struct inode *inode,
+					struct ocfs2_xattr_bucket *bucket,
+					void *para);
+static int ocfs2_rm_xattr_cluster(struct inode *inode,
+				  struct buffer_head *root_bh,
+				  u64 blkno,
+				  u32 cpos,
+				  u32 len,
+				  void *para);
+
 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 int ocfs2_prepare_refcount_xattr(struct inode *inode,
+					struct ocfs2_dinode *di,
+					struct ocfs2_xattr_info *xi,
+					struct ocfs2_xattr_search *xis,
+					struct ocfs2_xattr_search *xbs,
+					struct ocfs2_refcount_tree **ref_tree,
+					int *meta_need,
+					int *credits);
+static int ocfs2_get_xattr_tree_value_root(struct super_block *sb,
+					   struct ocfs2_xattr_bucket *bucket,
+					   int offset,
+					   struct ocfs2_xattr_value_root **xv,
+					   struct buffer_head **bh);
+static int ocfs2_xattr_security_set(struct inode *inode, const char *name,
+				    const void *value, size_t size, int flags);
 
 static inline u16 ocfs2_xattr_buckets_per_cluster(struct ocfs2_super *osb)
 {
@@ -254,9 +285,9 @@ static int ocfs2_init_xattr_bucket(struct ocfs2_xattr_bucket *bucket,
 			break;
 		}
 
-		if (!ocfs2_buffer_uptodate(bucket->bu_inode,
+		if (!ocfs2_buffer_uptodate(INODE_CACHE(bucket->bu_inode),
 					   bucket->bu_bhs[i]))
-			ocfs2_set_new_buffer_uptodate(bucket->bu_inode,
+			ocfs2_set_new_buffer_uptodate(INODE_CACHE(bucket->bu_inode),
 						      bucket->bu_bhs[i]);
 	}
 
@@ -271,7 +302,7 @@ static int ocfs2_read_xattr_bucket(struct ocfs2_xattr_bucket *bucket,
 {
 	int rc;
 
-	rc = ocfs2_read_blocks(bucket->bu_inode, xb_blkno,
+	rc = ocfs2_read_blocks(INODE_CACHE(bucket->bu_inode), xb_blkno,
 			       bucket->bu_blocks, bucket->bu_bhs, 0,
 			       NULL);
 	if (!rc) {
@@ -297,7 +328,8 @@ static int ocfs2_xattr_bucket_journal_access(handle_t *handle,
 	int i, rc = 0;
 
 	for (i = 0; i < bucket->bu_blocks; i++) {
-		rc = ocfs2_journal_access(handle, bucket->bu_inode,
+		rc = ocfs2_journal_access(handle,
+					  INODE_CACHE(bucket->bu_inode),
 					  bucket->bu_bhs[i], type);
 		if (rc) {
 			mlog_errno(rc);
@@ -399,7 +431,7 @@ static int ocfs2_read_xattr_block(struct inode *inode, u64 xb_blkno,
 	int rc;
 	struct buffer_head *tmp = *bh;
 
-	rc = ocfs2_read_block(inode, xb_blkno, &tmp,
+	rc = ocfs2_read_block(INODE_CACHE(inode), xb_blkno, &tmp,
 			      ocfs2_validate_xattr_block);
 
 	/* If ocfs2_read_block() got us a new bh, pass it up. */
@@ -596,15 +628,14 @@ static int ocfs2_xattr_extend_allocation(struct inode *inode,
 	int status = 0;
 	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(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, vb);
+	ocfs2_init_xattr_value_extent_tree(&et, INODE_CACHE(inode), vb);
 
-	status = vb->vb_access(handle, inode, vb->vb_bh,
+	status = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
 			      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (status < 0) {
 		mlog_errno(status);
@@ -612,13 +643,11 @@ static int ocfs2_xattr_extend_allocation(struct inode *inode,
 	}
 
 	prev_clusters = le32_to_cpu(vb->vb_xv->xr_clusters);
-	status = ocfs2_add_clusters_in_btree(osb,
-					     inode,
+	status = ocfs2_add_clusters_in_btree(handle,
+					     &et,
 					     &logical_start,
 					     clusters_to_add,
 					     0,
-					     &et,
-					     handle,
 					     ctxt->data_ac,
 					     ctxt->meta_ac,
 					     &why);
@@ -649,6 +678,7 @@ leave:
 static int __ocfs2_remove_xattr_range(struct inode *inode,
 				      struct ocfs2_xattr_value_buf *vb,
 				      u32 cpos, u32 phys_cpos, u32 len,
+				      unsigned int ext_flags,
 				      struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret;
@@ -656,16 +686,16 @@ static int __ocfs2_remove_xattr_range(struct inode *inode,
 	handle_t *handle = ctxt->handle;
 	struct ocfs2_extent_tree et;
 
-	ocfs2_init_xattr_value_extent_tree(&et, inode, vb);
+	ocfs2_init_xattr_value_extent_tree(&et, INODE_CACHE(inode), vb);
 
-	ret = vb->vb_access(handle, inode, vb->vb_bh,
+	ret = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
 			    OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
 	}
 
-	ret = ocfs2_remove_extent(inode, &et, cpos, len, handle, ctxt->meta_ac,
+	ret = ocfs2_remove_extent(handle, &et, cpos, len, ctxt->meta_ac,
 				  &ctxt->dealloc);
 	if (ret) {
 		mlog_errno(ret);
@@ -680,7 +710,14 @@ static int __ocfs2_remove_xattr_range(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_cache_cluster_dealloc(&ctxt->dealloc, phys_blkno, len);
+	if (ext_flags & OCFS2_EXT_REFCOUNTED)
+		ret = ocfs2_decrease_refcount(inode, handle,
+					ocfs2_blocks_to_clusters(inode->i_sb,
+								 phys_blkno),
+					len, ctxt->meta_ac, &ctxt->dealloc, 1);
+	else
+		ret = ocfs2_cache_cluster_dealloc(&ctxt->dealloc,
+						  phys_blkno, len);
 	if (ret)
 		mlog_errno(ret);
 
@@ -695,6 +732,7 @@ static int ocfs2_xattr_shrink_size(struct inode *inode,
 				   struct ocfs2_xattr_set_ctxt *ctxt)
 {
 	int ret = 0;
+	unsigned int ext_flags;
 	u32 trunc_len, cpos, phys_cpos, alloc_size;
 	u64 block;
 
@@ -706,7 +744,7 @@ static int ocfs2_xattr_shrink_size(struct inode *inode,
 	while (trunc_len) {
 		ret = ocfs2_xattr_get_clusters(inode, cpos, &phys_cpos,
 					       &alloc_size,
-					       &vb->vb_xv->xr_list);
+					       &vb->vb_xv->xr_list, &ext_flags);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -717,15 +755,15 @@ static int ocfs2_xattr_shrink_size(struct inode *inode,
 
 		ret = __ocfs2_remove_xattr_range(inode, vb, cpos,
 						 phys_cpos, alloc_size,
-						 ctxt);
+						 ext_flags, ctxt);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
 		block = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
-		ocfs2_remove_xattr_clusters_from_cache(inode, block,
-						       alloc_size);
+		ocfs2_remove_xattr_clusters_from_cache(INODE_CACHE(inode),
+						       block, alloc_size);
 		cpos += alloc_size;
 		trunc_len -= alloc_size;
 	}
@@ -810,6 +848,23 @@ static int ocfs2_xattr_list_entries(struct inode *inode,
 	return result;
 }
 
+int ocfs2_has_inline_xattr_value_outside(struct inode *inode,
+					 struct ocfs2_dinode *di)
+{
+	struct ocfs2_xattr_header *xh;
+	int i;
+
+	xh = (struct ocfs2_xattr_header *)
+		 ((void *)di + inode->i_sb->s_blocksize -
+		 le16_to_cpu(di->i_xattr_inline_size));
+
+	for (i = 0; i < le16_to_cpu(xh->xh_count); i++)
+		if (!ocfs2_xattr_is_local(&xh->xh_entries[i]))
+			return 1;
+
+	return 0;
+}
+
 static int ocfs2_xattr_ibody_list(struct inode *inode,
 				  struct ocfs2_dinode *di,
 				  char *buffer,
@@ -855,11 +910,9 @@ static int ocfs2_xattr_block_list(struct inode *inode,
 		struct ocfs2_xattr_header *header = &xb->xb_attrs.xb_header;
 		ret = ocfs2_xattr_list_entries(inode, header,
 					       buffer, buffer_size);
-	} else {
-		struct ocfs2_xattr_tree_root *xt = &xb->xb_attrs.xb_root;
-		ret = ocfs2_xattr_tree_list_index_block(inode, xt,
+	} else
+		ret = ocfs2_xattr_tree_list_index_block(inode, blk_bh,
 						   buffer, buffer_size);
-	}
 
 	brelse(blk_bh);
 
@@ -961,7 +1014,7 @@ static int ocfs2_xattr_get_value_outside(struct inode *inode,
 	cpos = 0;
 	while (cpos < clusters) {
 		ret = ocfs2_xattr_get_clusters(inode, cpos, &p_cluster,
-					       &num_clusters, el);
+					       &num_clusters, el, NULL);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -970,7 +1023,8 @@ 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, NULL);
+			ret = ocfs2_read_block(INODE_CACHE(inode), blkno,
+					       &bh, NULL);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
@@ -1085,7 +1139,7 @@ static int ocfs2_xattr_block_get(struct inode *inode,
 		i = xs->here - xs->header->xh_entries;
 
 		if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED) {
-			ret = ocfs2_xattr_bucket_get_name_value(inode,
+			ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb,
 								bucket_xh(xs->bucket),
 								i,
 								&block_off,
@@ -1183,7 +1237,7 @@ 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,
+					   struct ocfs2_xattr_value_buf *vb,
 					   const void *value,
 					   int value_len)
 {
@@ -1194,28 +1248,34 @@ static int __ocfs2_xattr_set_value_outside(struct inode *inode,
 	u32 clusters = ocfs2_clusters_for_bytes(inode->i_sb, value_len);
 	u64 blkno;
 	struct buffer_head *bh = NULL;
+	unsigned int ext_flags;
+	struct ocfs2_xattr_value_root *xv = vb->vb_xv;
 
 	BUG_ON(clusters > le32_to_cpu(xv->xr_clusters));
 
 	while (cpos < clusters) {
 		ret = ocfs2_xattr_get_clusters(inode, cpos, &p_cluster,
-					       &num_clusters, &xv->xr_list);
+					       &num_clusters, &xv->xr_list,
+					       &ext_flags);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
 
+		BUG_ON(ext_flags & OCFS2_EXT_REFCOUNTED);
+
 		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, NULL);
+			ret = ocfs2_read_block(INODE_CACHE(inode), blkno,
+					       &bh, NULL);
 			if (ret) {
 				mlog_errno(ret);
 				goto out;
 			}
 
 			ret = ocfs2_journal_access(handle,
-						   inode,
+						   INODE_CACHE(inode),
 						   bh,
 						   OCFS2_JOURNAL_ACCESS_WRITE);
 			if (ret < 0) {
@@ -1266,7 +1326,7 @@ static int ocfs2_xattr_cleanup(struct inode *inode,
 	void *val = xs->base + offs;
 	size_t size = OCFS2_XATTR_SIZE(name_len) + OCFS2_XATTR_ROOT_SIZE;
 
-	ret = vb->vb_access(handle, inode, vb->vb_bh,
+	ret = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
 			    OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1294,7 +1354,7 @@ static int ocfs2_xattr_update_entry(struct inode *inode,
 {
 	int ret;
 
-	ret = vb->vb_access(handle, inode, vb->vb_bh,
+	ret = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
 			    OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1355,7 +1415,7 @@ static int ocfs2_xattr_set_value_outside(struct inode *inode,
 		mlog_errno(ret);
 		return ret;
 	}
-	ret = __ocfs2_xattr_set_value_outside(inode, ctxt->handle, vb->vb_xv,
+	ret = __ocfs2_xattr_set_value_outside(inode, ctxt->handle, vb,
 					      xi->value, xi->value_len);
 	if (ret < 0)
 		mlog_errno(ret);
@@ -1594,7 +1654,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 
 				ret = __ocfs2_xattr_set_value_outside(inode,
 								handle,
-								vb.vb_xv,
+								&vb,
 								xi->value,
 								xi->value_len);
 				if (ret < 0)
@@ -1615,7 +1675,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 		}
 	}
 
-	ret = ocfs2_journal_access_di(handle, inode, xs->inode_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), xs->inode_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1623,7 +1683,7 @@ static int ocfs2_xattr_set_entry(struct inode *inode,
 	}
 
 	if (!(flag & OCFS2_INLINE_XATTR_FL)) {
-		ret = vb.vb_access(handle, inode, vb.vb_bh,
+		ret = vb.vb_access(handle, INODE_CACHE(inode), vb.vb_bh,
 				   OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
 			mlog_errno(ret);
@@ -1700,51 +1760,112 @@ out:
 	return ret;
 }
 
+/*
+ * In xattr remove, if it is stored outside and refcounted, we may have
+ * the chance to split the refcount tree. So need the allocators.
+ */
+static int ocfs2_lock_xattr_remove_allocators(struct inode *inode,
+					struct ocfs2_xattr_value_root *xv,
+					struct ocfs2_caching_info *ref_ci,
+					struct buffer_head *ref_root_bh,
+					struct ocfs2_alloc_context **meta_ac,
+					int *ref_credits)
+{
+	int ret, meta_add = 0;
+	u32 p_cluster, num_clusters;
+	unsigned int ext_flags;
+
+	*ref_credits = 0;
+	ret = ocfs2_xattr_get_clusters(inode, 0, &p_cluster,
+				       &num_clusters,
+				       &xv->xr_list,
+				       &ext_flags);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (!(ext_flags & OCFS2_EXT_REFCOUNTED))
+		goto out;
+
+	ret = ocfs2_refcounted_xattr_delete_need(inode, ref_ci,
+						 ref_root_bh, xv,
+						 &meta_add, ref_credits);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_reserve_new_metadata_blocks(OCFS2_SB(inode->i_sb),
+						meta_add, meta_ac);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	return ret;
+}
+
 static int ocfs2_remove_value_outside(struct inode*inode,
 				      struct ocfs2_xattr_value_buf *vb,
-				      struct ocfs2_xattr_header *header)
+				      struct ocfs2_xattr_header *header,
+				      struct ocfs2_caching_info *ref_ci,
+				      struct buffer_head *ref_root_bh)
 {
-	int ret = 0, i;
+	int ret = 0, i, ref_credits;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_xattr_set_ctxt ctxt = { NULL, NULL, };
+	void *val;
 
 	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)) {
-			void *val;
+		if (ocfs2_xattr_is_local(entry))
+			continue;
 
-			val = (void *)header +
-				le16_to_cpu(entry->xe_name_offset);
-			vb->vb_xv = (struct ocfs2_xattr_value_root *)
-				(val + OCFS2_XATTR_SIZE(entry->xe_name_len));
-			ret = ocfs2_xattr_value_truncate(inode, vb, 0, &ctxt);
-			if (ret < 0) {
-				mlog_errno(ret);
-				break;
-			}
+		val = (void *)header +
+			le16_to_cpu(entry->xe_name_offset);
+		vb->vb_xv = (struct ocfs2_xattr_value_root *)
+			(val + OCFS2_XATTR_SIZE(entry->xe_name_len));
+
+		ret = ocfs2_lock_xattr_remove_allocators(inode, vb->vb_xv,
+							 ref_ci, ref_root_bh,
+							 &ctxt.meta_ac,
+							 &ref_credits);
+
+		ctxt.handle = ocfs2_start_trans(osb, ref_credits +
+					ocfs2_remove_extent_credits(osb->sb));
+		if (IS_ERR(ctxt.handle)) {
+			ret = PTR_ERR(ctxt.handle);
+			mlog_errno(ret);
+			break;
+		}
+
+		ret = ocfs2_xattr_value_truncate(inode, vb, 0, &ctxt);
+		if (ret < 0) {
+			mlog_errno(ret);
+			break;
+		}
+
+		ocfs2_commit_trans(osb, ctxt.handle);
+		if (ctxt.meta_ac) {
+			ocfs2_free_alloc_context(ctxt.meta_ac);
+			ctxt.meta_ac = NULL;
 		}
 	}
 
-	ocfs2_commit_trans(osb, ctxt.handle);
+	if (ctxt.meta_ac)
+		ocfs2_free_alloc_context(ctxt.meta_ac);
 	ocfs2_schedule_truncate_log_flush(osb, 1);
 	ocfs2_run_deallocs(osb, &ctxt.dealloc);
-out:
 	return ret;
 }
 
 static int ocfs2_xattr_ibody_remove(struct inode *inode,
-				    struct buffer_head *di_bh)
+				    struct buffer_head *di_bh,
+				    struct ocfs2_caching_info *ref_ci,
+				    struct buffer_head *ref_root_bh)
 {
 
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
@@ -1759,13 +1880,21 @@ static int ocfs2_xattr_ibody_remove(struct inode *inode,
 		 ((void *)di + inode->i_sb->s_blocksize -
 		 le16_to_cpu(di->i_xattr_inline_size));
 
-	ret = ocfs2_remove_value_outside(inode, &vb, header);
+	ret = ocfs2_remove_value_outside(inode, &vb, header,
+					 ref_ci, ref_root_bh);
 
 	return ret;
 }
 
+struct ocfs2_rm_xattr_bucket_para {
+	struct ocfs2_caching_info *ref_ci;
+	struct buffer_head *ref_root_bh;
+};
+
 static int ocfs2_xattr_block_remove(struct inode *inode,
-				    struct buffer_head *blk_bh)
+				    struct buffer_head *blk_bh,
+				    struct ocfs2_caching_info *ref_ci,
+				    struct buffer_head *ref_root_bh)
 {
 	struct ocfs2_xattr_block *xb;
 	int ret = 0;
@@ -1773,19 +1902,29 @@ static int ocfs2_xattr_block_remove(struct inode *inode,
 		.vb_bh = blk_bh,
 		.vb_access = ocfs2_journal_access_xb,
 	};
+	struct ocfs2_rm_xattr_bucket_para args = {
+		.ref_ci = ref_ci,
+		.ref_root_bh = ref_root_bh,
+	};
 
 	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, &vb, header);
+		ret = ocfs2_remove_value_outside(inode, &vb, header,
+						 ref_ci, ref_root_bh);
 	} else
-		ret = ocfs2_delete_xattr_index_block(inode, blk_bh);
+		ret = ocfs2_iterate_xattr_index_block(inode,
+						blk_bh,
+						ocfs2_rm_xattr_cluster,
+						&args);
 
 	return ret;
 }
 
 static int ocfs2_xattr_free_block(struct inode *inode,
-				  u64 block)
+				  u64 block,
+				  struct ocfs2_caching_info *ref_ci,
+				  struct buffer_head *ref_root_bh)
 {
 	struct inode *xb_alloc_inode;
 	struct buffer_head *xb_alloc_bh = NULL;
@@ -1803,7 +1942,7 @@ static int ocfs2_xattr_free_block(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_xattr_block_remove(inode, blk_bh);
+	ret = ocfs2_xattr_block_remove(inode, blk_bh, ref_ci, ref_root_bh);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -1863,6 +2002,9 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh)
 {
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+	struct ocfs2_refcount_tree *ref_tree = NULL;
+	struct buffer_head *ref_root_bh = NULL;
+	struct ocfs2_caching_info *ref_ci = NULL;
 	handle_t *handle;
 	int ret;
 
@@ -1872,8 +2014,21 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh)
 	if (!(oi->ip_dyn_features & OCFS2_HAS_XATTR_FL))
 		return 0;
 
+	if (OCFS2_I(inode)->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL) {
+		ret = ocfs2_lock_refcount_tree(OCFS2_SB(inode->i_sb),
+					       le64_to_cpu(di->i_refcount_loc),
+					       1, &ref_tree, &ref_root_bh);
+		if (ret) {
+			mlog_errno(ret);
+			goto out;
+		}
+		ref_ci = &ref_tree->rf_ci;
+
+	}
+
 	if (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL) {
-		ret = ocfs2_xattr_ibody_remove(inode, di_bh);
+		ret = ocfs2_xattr_ibody_remove(inode, di_bh,
+					       ref_ci, ref_root_bh);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -1882,7 +2037,8 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh)
 
 	if (di->i_xattr_loc) {
 		ret = ocfs2_xattr_free_block(inode,
-					     le64_to_cpu(di->i_xattr_loc));
+					     le64_to_cpu(di->i_xattr_loc),
+					     ref_ci, ref_root_bh);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
@@ -1896,7 +2052,7 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh)
 		mlog_errno(ret);
 		goto out;
 	}
-	ret = ocfs2_journal_access_di(handle, inode, di_bh,
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -1916,6 +2072,9 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh)
 out_commit:
 	ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
 out:
+	if (ref_tree)
+		ocfs2_unlock_refcount_tree(OCFS2_SB(inode->i_sb), ref_tree, 1);
+	brelse(ref_root_bh);
 	return ret;
 }
 
@@ -2083,6 +2242,84 @@ cleanup:
 	return ret;
 }
 
+static int ocfs2_create_xattr_block(handle_t *handle,
+				    struct inode *inode,
+				    struct buffer_head *inode_bh,
+				    struct ocfs2_alloc_context *meta_ac,
+				    struct buffer_head **ret_bh,
+				    int indexed)
+{
+	int ret;
+	u16 suballoc_bit_start;
+	u32 num_got;
+	u64 first_blkno;
+	struct ocfs2_dinode *di =  (struct ocfs2_dinode *)inode_bh->b_data;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct buffer_head *new_bh = NULL;
+	struct ocfs2_xattr_block *xblk;
+
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), inode_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto end;
+	}
+
+	ret = ocfs2_claim_metadata(osb, handle, meta_ac, 1,
+				   &suballoc_bit_start, &num_got,
+				   &first_blkno);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto end;
+	}
+
+	new_bh = sb_getblk(inode->i_sb, first_blkno);
+	ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), new_bh);
+
+	ret = ocfs2_journal_access_xb(handle, INODE_CACHE(inode),
+				      new_bh,
+				      OCFS2_JOURNAL_ACCESS_CREATE);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto end;
+	}
+
+	/* Initialize ocfs2_xattr_block */
+	xblk = (struct ocfs2_xattr_block *)new_bh->b_data;
+	memset(xblk, 0, inode->i_sb->s_blocksize);
+	strcpy((void *)xblk, OCFS2_XATTR_BLOCK_SIGNATURE);
+	xblk->xb_suballoc_slot = cpu_to_le16(osb->slot_num);
+	xblk->xb_suballoc_bit = cpu_to_le16(suballoc_bit_start);
+	xblk->xb_fs_generation = cpu_to_le32(osb->fs_generation);
+	xblk->xb_blkno = cpu_to_le64(first_blkno);
+
+	if (indexed) {
+		struct ocfs2_xattr_tree_root *xr = &xblk->xb_attrs.xb_root;
+		xr->xt_clusters = cpu_to_le32(1);
+		xr->xt_last_eb_blk = 0;
+		xr->xt_list.l_tree_depth = 0;
+		xr->xt_list.l_count = cpu_to_le16(
+					ocfs2_xattr_recs_per_xb(inode->i_sb));
+		xr->xt_list.l_next_free_rec = cpu_to_le16(1);
+		xblk->xb_flags = cpu_to_le16(OCFS2_XATTR_INDEXED);
+	}
+
+	ret = ocfs2_journal_dirty(handle, new_bh);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto end;
+	}
+	di->i_xattr_loc = cpu_to_le64(first_blkno);
+	ocfs2_journal_dirty(handle, inode_bh);
+
+	*ret_bh = new_bh;
+	new_bh = NULL;
+
+end:
+	brelse(new_bh);
+	return ret;
+}
+
 /*
  * ocfs2_xattr_block_set()
  *
@@ -2095,63 +2332,24 @@ static int ocfs2_xattr_block_set(struct inode *inode,
 				 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;
 	handle_t *handle = ctxt->handle;
 	struct ocfs2_xattr_block *xblk = NULL;
-	u16 suballoc_bit_start;
-	u32 num_got;
-	u64 first_blkno;
 	int ret;
 
 	if (!xs->xattr_bh) {
-		ret = ocfs2_journal_access_di(handle, inode, xs->inode_bh,
-					      OCFS2_JOURNAL_ACCESS_CREATE);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto end;
-		}
-
-		ret = ocfs2_claim_metadata(osb, handle, ctxt->meta_ac, 1,
-					   &suballoc_bit_start, &num_got,
-					   &first_blkno);
-		if (ret < 0) {
-			mlog_errno(ret);
-			goto end;
-		}
-
-		new_bh = sb_getblk(inode->i_sb, first_blkno);
-		ocfs2_set_new_buffer_uptodate(inode, new_bh);
-
-		ret = ocfs2_journal_access_xb(handle, inode, new_bh,
-					      OCFS2_JOURNAL_ACCESS_CREATE);
-		if (ret < 0) {
+		ret = ocfs2_create_xattr_block(handle, inode, xs->inode_bh,
+					       ctxt->meta_ac, &new_bh, 0);
+		if (ret) {
 			mlog_errno(ret);
 			goto end;
 		}
 
-		/* Initialize ocfs2_xattr_block */
 		xs->xattr_bh = new_bh;
-		xblk = (struct ocfs2_xattr_block *)new_bh->b_data;
-		memset(xblk, 0, inode->i_sb->s_blocksize);
-		strcpy((void *)xblk, OCFS2_XATTR_BLOCK_SIGNATURE);
-		xblk->xb_suballoc_slot = cpu_to_le16(osb->slot_num);
-		xblk->xb_suballoc_bit = cpu_to_le16(suballoc_bit_start);
-		xblk->xb_fs_generation = cpu_to_le32(osb->fs_generation);
-		xblk->xb_blkno = cpu_to_le64(first_blkno);
-
+		xblk = (struct ocfs2_xattr_block *)xs->xattr_bh->b_data;
 		xs->header = &xblk->xb_attrs.xb_header;
 		xs->base = (void *)xs->header;
 		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 end;
-		}
-		di->i_xattr_loc = cpu_to_le64(first_blkno);
-		ocfs2_journal_dirty(handle, xs->inode_bh);
 	} else
 		xblk = (struct ocfs2_xattr_block *)xs->xattr_bh->b_data;
 
@@ -2273,7 +2471,7 @@ static int ocfs2_calc_xattr_set_need(struct inode *inode,
 		old_in_xb = 1;
 
 		if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED) {
-			ret = ocfs2_xattr_bucket_get_name_value(inode,
+			ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb,
 							bucket_xh(xbs->bucket),
 							i, &block_off,
 							&name_offset);
@@ -2428,6 +2626,7 @@ static int ocfs2_init_xattr_set_ctxt(struct inode *inode,
 				     struct ocfs2_xattr_search *xis,
 				     struct ocfs2_xattr_search *xbs,
 				     struct ocfs2_xattr_set_ctxt *ctxt,
+				     int extra_meta,
 				     int *credits)
 {
 	int clusters_add, meta_add, ret;
@@ -2444,6 +2643,7 @@ static int ocfs2_init_xattr_set_ctxt(struct inode *inode,
 		return ret;
 	}
 
+	meta_add += extra_meta;
 	mlog(0, "Set xattr %s, reserve meta blocks = %d, clusters = %d, "
 	     "credits = %d\n", xi->name, meta_add, clusters_add, *credits);
 
@@ -2598,7 +2798,7 @@ static int __ocfs2_xattr_set_handle(struct inode *inode,
 
 	if (!ret) {
 		/* Update inode ctime. */
-		ret = ocfs2_journal_access_di(ctxt->handle, inode,
+		ret = ocfs2_journal_access_di(ctxt->handle, INODE_CACHE(inode),
 					      xis->inode_bh,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 		if (ret) {
@@ -2711,10 +2911,11 @@ int ocfs2_xattr_set(struct inode *inode,
 {
 	struct buffer_head *di_bh = NULL;
 	struct ocfs2_dinode *di;
-	int ret, credits;
+	int ret, credits, ref_meta = 0, ref_credits = 0;
 	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_refcount_tree *ref_tree = NULL;
 
 	struct ocfs2_xattr_info xi = {
 		.name_index = name_index,
@@ -2779,6 +2980,17 @@ int ocfs2_xattr_set(struct inode *inode,
 			goto cleanup;
 	}
 
+	/* Check whether the value is refcounted and do some prepartion. */
+	if (OCFS2_I(inode)->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL &&
+	    (!xis.not_found || !xbs.not_found)) {
+		ret = ocfs2_prepare_refcount_xattr(inode, di, &xi,
+						   &xis, &xbs, &ref_tree,
+						   &ref_meta, &ref_credits);
+		if (ret) {
+			mlog_errno(ret);
+			goto cleanup;
+		}
+	}
 
 	mutex_lock(&tl_inode->i_mutex);
 
@@ -2793,7 +3005,7 @@ int ocfs2_xattr_set(struct inode *inode,
 	mutex_unlock(&tl_inode->i_mutex);
 
 	ret = ocfs2_init_xattr_set_ctxt(inode, di, &xi, &xis,
-					&xbs, &ctxt, &credits);
+					&xbs, &ctxt, ref_meta, &credits);
 	if (ret) {
 		mlog_errno(ret);
 		goto cleanup;
@@ -2801,7 +3013,7 @@ int ocfs2_xattr_set(struct inode *inode,
 
 	/* 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);
+	ctxt.handle = ocfs2_start_trans(osb, credits + ref_credits);
 	if (IS_ERR(ctxt.handle)) {
 		ret = PTR_ERR(ctxt.handle);
 		mlog_errno(ret);
@@ -2819,8 +3031,16 @@ int ocfs2_xattr_set(struct inode *inode,
 	if (ocfs2_dealloc_has_cluster(&ctxt.dealloc))
 		ocfs2_schedule_truncate_log_flush(osb, 1);
 	ocfs2_run_deallocs(osb, &ctxt.dealloc);
+
 cleanup:
+	if (ref_tree)
+		ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
 	up_write(&OCFS2_I(inode)->ip_xattr_sem);
+	if (!value && !ret) {
+		ret = ocfs2_try_remove_refcount_tree(inode, di_bh);
+		if (ret)
+			mlog_errno(ret);
+	}
 	ocfs2_inode_unlock(inode, 1);
 cleanup_nolock:
 	brelse(di_bh);
@@ -2849,7 +3069,8 @@ static int ocfs2_xattr_get_rec(struct inode *inode,
 	u64 e_blkno = 0;
 
 	if (el->l_tree_depth) {
-		ret = ocfs2_find_leaf(inode, el, name_hash, &eb_bh);
+		ret = ocfs2_find_leaf(INODE_CACHE(inode), el, name_hash,
+				      &eb_bh);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
@@ -2931,7 +3152,7 @@ static int ocfs2_find_xe_in_bucket(struct inode *inode,
 		if (cmp)
 			continue;
 
-		ret = ocfs2_xattr_bucket_get_name_value(inode,
+		ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb,
 							xh,
 							i,
 							&block_off,
@@ -3175,7 +3396,7 @@ struct ocfs2_xattr_tree_list {
 	size_t result;
 };
 
-static int ocfs2_xattr_bucket_get_name_value(struct inode *inode,
+static int ocfs2_xattr_bucket_get_name_value(struct super_block *sb,
 					     struct ocfs2_xattr_header *xh,
 					     int index,
 					     int *block_off,
@@ -3188,8 +3409,8 @@ static int ocfs2_xattr_bucket_get_name_value(struct inode *inode,
 
 	name_offset = le16_to_cpu(xh->xh_entries[index].xe_name_offset);
 
-	*block_off = name_offset >> inode->i_sb->s_blocksize_bits;
-	*new_offset = name_offset % inode->i_sb->s_blocksize;
+	*block_off = name_offset >> sb->s_blocksize_bits;
+	*new_offset = name_offset % sb->s_blocksize;
 
 	return 0;
 }
@@ -3209,7 +3430,7 @@ static int ocfs2_list_xattr_bucket(struct inode *inode,
 		prefix = ocfs2_xattr_prefix(type);
 
 		if (prefix) {
-			ret = ocfs2_xattr_bucket_get_name_value(inode,
+			ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb,
 								bucket_xh(bucket),
 								i,
 								&block_off,
@@ -3232,22 +3453,19 @@ static int ocfs2_list_xattr_bucket(struct inode *inode,
 	return ret;
 }
 
-static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
-					     struct ocfs2_xattr_tree_root *xt,
-					     char *buffer,
-					     size_t buffer_size)
+static int ocfs2_iterate_xattr_index_block(struct inode *inode,
+					   struct buffer_head *blk_bh,
+					   xattr_tree_rec_func *rec_func,
+					   void *para)
 {
-	struct ocfs2_extent_list *el = &xt->xt_list;
+	struct ocfs2_xattr_block *xb =
+			(struct ocfs2_xattr_block *)blk_bh->b_data;
+	struct ocfs2_extent_list *el = &xb->xb_attrs.xb_root.xt_list;
 	int ret = 0;
 	u32 name_hash = UINT_MAX, e_cpos = 0, num_clusters = 0;
 	u64 p_blkno = 0;
-	struct ocfs2_xattr_tree_list xl = {
-		.buffer = buffer,
-		.buffer_size = buffer_size,
-		.result = 0,
-	};
 
-	if (le16_to_cpu(el->l_next_free_rec) == 0)
+	if (!el->l_next_free_rec || !rec_func)
 		return 0;
 
 	while (name_hash > 0) {
@@ -3255,16 +3473,15 @@ static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
 					  &e_cpos, &num_clusters, el);
 		if (ret) {
 			mlog_errno(ret);
-			goto out;
+			break;
 		}
 
-		ret = ocfs2_iterate_xattr_buckets(inode, p_blkno, num_clusters,
-						  ocfs2_list_xattr_bucket,
-						  &xl);
+		ret = rec_func(inode, blk_bh, p_blkno, e_cpos,
+			       num_clusters, para);
 		if (ret) {
 			if (ret != -ERANGE)
 				mlog_errno(ret);
-			goto out;
+			break;
 		}
 
 		if (e_cpos == 0)
@@ -3273,6 +3490,37 @@ static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
 		name_hash = e_cpos - 1;
 	}
 
+	return ret;
+
+}
+
+static int ocfs2_list_xattr_tree_rec(struct inode *inode,
+				     struct buffer_head *root_bh,
+				     u64 blkno, u32 cpos, u32 len, void *para)
+{
+	return ocfs2_iterate_xattr_buckets(inode, blkno, len,
+					   ocfs2_list_xattr_bucket, para);
+}
+
+static int ocfs2_xattr_tree_list_index_block(struct inode *inode,
+					     struct buffer_head *blk_bh,
+					     char *buffer,
+					     size_t buffer_size)
+{
+	int ret;
+	struct ocfs2_xattr_tree_list xl = {
+		.buffer = buffer,
+		.buffer_size = buffer_size,
+		.result = 0,
+	};
+
+	ret = ocfs2_iterate_xattr_index_block(inode, blk_bh,
+					      ocfs2_list_xattr_tree_rec, &xl);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
 	ret = xl.result;
 out:
 	return ret;
@@ -3426,7 +3674,7 @@ static int ocfs2_xattr_create_index_block(struct inode *inode,
 	 */
 	down_write(&oi->ip_alloc_sem);
 
-	ret = ocfs2_journal_access_xb(handle, inode, xb_bh,
+	ret = ocfs2_journal_access_xb(handle, INODE_CACHE(inode), xb_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret) {
 		mlog_errno(ret);
@@ -4263,9 +4511,9 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
 	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     prev_cpos, (unsigned long long)bucket_blkno(first));
 
-	ocfs2_init_xattr_tree_extent_tree(&et, inode, root_bh);
+	ocfs2_init_xattr_tree_extent_tree(&et, INODE_CACHE(inode), root_bh);
 
-	ret = ocfs2_journal_access_xb(handle, inode, root_bh,
+	ret = ocfs2_journal_access_xb(handle, INODE_CACHE(inode), root_bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -4319,7 +4567,7 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
 
 	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,
+	ret = ocfs2_insert_extent(handle, &et, v_start, block,
 				  num_bits, 0, ctxt->meta_ac);
 	if (ret < 0) {
 		mlog_errno(ret);
@@ -4798,10 +5046,13 @@ static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode,
 	struct ocfs2_xattr_entry *xe = xs->here;
 	struct ocfs2_xattr_header *xh = bucket_xh(xs->bucket);
 	void *base;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_access = ocfs2_journal_access,
+	};
 
 	BUG_ON(!xs->base || !xe || ocfs2_xattr_is_local(xe));
 
-	ret = ocfs2_xattr_bucket_get_name_value(inode, xh,
+	ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb, xh,
 						xe - xh->xh_entries,
 						&block_off,
 						&offset);
@@ -4814,8 +5065,10 @@ static int ocfs2_xattr_bucket_set_value_outside(struct inode *inode,
 	xv = (struct ocfs2_xattr_value_root *)(base + offset +
 		 OCFS2_XATTR_SIZE(xe->xe_name_len));
 
+	vb.vb_xv = xv;
+	vb.vb_bh = xs->bucket->bu_bhs[block_off];
 	ret = __ocfs2_xattr_set_value_outside(inode, handle,
-					      xv, val, value_len);
+					      &vb, val, value_len);
 	if (ret)
 		mlog_errno(ret);
 out:
@@ -4826,7 +5079,8 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode,
 				  struct buffer_head *root_bh,
 				  u64 blkno,
 				  u32 cpos,
-				  u32 len)
+				  u32 len,
+				  void *para)
 {
 	int ret;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
@@ -4838,14 +5092,22 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode,
 	struct ocfs2_cached_dealloc_ctxt dealloc;
 	struct ocfs2_extent_tree et;
 
-	ocfs2_init_xattr_tree_extent_tree(&et, inode, root_bh);
+	ret = ocfs2_iterate_xattr_buckets(inode, blkno, len,
+					  ocfs2_delete_xattr_in_bucket, para);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	ocfs2_init_xattr_tree_extent_tree(&et, INODE_CACHE(inode), root_bh);
 
 	ocfs2_init_dealloc_ctxt(&dealloc);
 
 	mlog(0, "rm xattr extent rec at %u len = %u, start from %llu\n",
 	     cpos, len, (unsigned long long)blkno);
 
-	ocfs2_remove_xattr_clusters_from_cache(inode, blkno, len);
+	ocfs2_remove_xattr_clusters_from_cache(INODE_CACHE(inode), blkno,
+					       len);
 
 	ret = ocfs2_lock_allocators(inode, &et, 0, 1, NULL, &meta_ac);
 	if (ret) {
@@ -4870,14 +5132,14 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_journal_access_xb(handle, inode, root_bh,
+	ret = ocfs2_journal_access_xb(handle, INODE_CACHE(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,
+	ret = ocfs2_remove_extent(handle, &et, cpos, len, meta_ac,
 				  &dealloc);
 	if (ret) {
 		mlog_errno(ret);
@@ -5220,7 +5482,7 @@ static int ocfs2_delete_xattr_in_bucket(struct inode *inode,
 					struct ocfs2_xattr_bucket *bucket,
 					void *para)
 {
-	int ret = 0;
+	int ret = 0, ref_credits;
 	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
 	u16 i;
 	struct ocfs2_xattr_entry *xe;
@@ -5228,7 +5490,9 @@ static int ocfs2_delete_xattr_in_bucket(struct inode *inode,
 	struct ocfs2_xattr_set_ctxt ctxt = {NULL, NULL,};
 	int credits = ocfs2_remove_extent_credits(osb->sb) +
 		ocfs2_blocks_per_xattr_bucket(inode->i_sb);
-
+	struct ocfs2_xattr_value_root *xv;
+	struct ocfs2_rm_xattr_bucket_para *args =
+			(struct ocfs2_rm_xattr_bucket_para *)para;
 
 	ocfs2_init_dealloc_ctxt(&ctxt.dealloc);
 
@@ -5237,7 +5501,16 @@ static int ocfs2_delete_xattr_in_bucket(struct inode *inode,
 		if (ocfs2_xattr_is_local(xe))
 			continue;
 
-		ctxt.handle = ocfs2_start_trans(osb, credits);
+		ret = ocfs2_get_xattr_tree_value_root(inode->i_sb, bucket,
+						      i, &xv, NULL);
+
+		ret = ocfs2_lock_xattr_remove_allocators(inode, xv,
+							 args->ref_ci,
+							 args->ref_root_bh,
+							 &ctxt.meta_ac,
+							 &ref_credits);
+
+		ctxt.handle = ocfs2_start_trans(osb, credits + ref_credits);
 		if (IS_ERR(ctxt.handle)) {
 			ret = PTR_ERR(ctxt.handle);
 			mlog_errno(ret);
@@ -5248,57 +5521,1439 @@ static int ocfs2_delete_xattr_in_bucket(struct inode *inode,
 							i, 0, &ctxt);
 
 		ocfs2_commit_trans(osb, ctxt.handle);
+		if (ctxt.meta_ac) {
+			ocfs2_free_alloc_context(ctxt.meta_ac);
+			ctxt.meta_ac = NULL;
+		}
 		if (ret) {
 			mlog_errno(ret);
 			break;
 		}
 	}
 
+	if (ctxt.meta_ac)
+		ocfs2_free_alloc_context(ctxt.meta_ac);
 	ocfs2_schedule_truncate_log_flush(osb, 1);
 	ocfs2_run_deallocs(osb, &ctxt.dealloc);
 	return ret;
 }
 
-static int ocfs2_delete_xattr_index_block(struct inode *inode,
-					  struct buffer_head *xb_bh)
+/*
+ * Whenever we modify a xattr value root in the bucket(e.g, CoW
+ * or change the extent record flag), we need to recalculate
+ * the metaecc for the whole bucket. So it is done here.
+ *
+ * Note:
+ * We have to give the extra credits for the caller.
+ */
+static int ocfs2_xattr_bucket_post_refcount(struct inode *inode,
+					    handle_t *handle,
+					    void *para)
+{
+	int ret;
+	struct ocfs2_xattr_bucket *bucket =
+			(struct ocfs2_xattr_bucket *)para;
+
+	ret = ocfs2_xattr_bucket_journal_access(handle, bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	ocfs2_xattr_bucket_journal_dirty(handle, bucket);
+
+	return 0;
+}
+
+/*
+ * Special action we need if the xattr value is refcounted.
+ *
+ * 1. If the xattr is refcounted, lock the tree.
+ * 2. CoW the xattr if we are setting the new value and the value
+ *    will be stored outside.
+ * 3. In other case, decrease_refcount will work for us, so just
+ *    lock the refcount tree, calculate the meta and credits is OK.
+ *
+ * We have to do CoW before ocfs2_init_xattr_set_ctxt since
+ * currently CoW is a completed transaction, while this function
+ * will also lock the allocators and let us deadlock. So we will
+ * CoW the whole xattr value.
+ */
+static int ocfs2_prepare_refcount_xattr(struct inode *inode,
+					struct ocfs2_dinode *di,
+					struct ocfs2_xattr_info *xi,
+					struct ocfs2_xattr_search *xis,
+					struct ocfs2_xattr_search *xbs,
+					struct ocfs2_refcount_tree **ref_tree,
+					int *meta_add,
+					int *credits)
 {
-	struct ocfs2_xattr_block *xb =
-			(struct ocfs2_xattr_block *)xb_bh->b_data;
-	struct ocfs2_extent_list *el = &xb->xb_attrs.xb_root.xt_list;
 	int ret = 0;
-	u32 name_hash = UINT_MAX, e_cpos, num_clusters;
-	u64 p_blkno;
+	struct ocfs2_xattr_block *xb;
+	struct ocfs2_xattr_entry *xe;
+	char *base;
+	u32 p_cluster, num_clusters;
+	unsigned int ext_flags;
+	int name_offset, name_len;
+	struct ocfs2_xattr_value_buf vb;
+	struct ocfs2_xattr_bucket *bucket = NULL;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_post_refcount refcount;
+	struct ocfs2_post_refcount *p = NULL;
+	struct buffer_head *ref_root_bh = NULL;
 
-	if (le16_to_cpu(el->l_next_free_rec) == 0)
-		return 0;
+	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;
+		vb.vb_bh = xis->inode_bh;
+		vb.vb_access = ocfs2_journal_access_di;
+	} 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;
 
-	while (name_hash > 0) {
-		ret = ocfs2_xattr_get_rec(inode, name_hash, &p_blkno,
-					  &e_cpos, &num_clusters, el);
+		if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED) {
+			ret = ocfs2_xattr_bucket_get_name_value(inode->i_sb,
+							bucket_xh(xbs->bucket),
+							i, &block_off,
+							&name_offset);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+			base = bucket_block(xbs->bucket, block_off);
+			vb.vb_bh = xbs->bucket->bu_bhs[block_off];
+			vb.vb_access = ocfs2_journal_access;
+
+			if (ocfs2_meta_ecc(osb)) {
+				/*create parameters for ocfs2_post_refcount. */
+				bucket = xbs->bucket;
+				refcount.credits = bucket->bu_blocks;
+				refcount.para = bucket;
+				refcount.func =
+					ocfs2_xattr_bucket_post_refcount;
+				p = &refcount;
+			}
+		} else {
+			base = xbs->base;
+			vb.vb_bh = xbs->xattr_bh;
+			vb.vb_access = ocfs2_journal_access_xb;
+		}
+	}
+
+	if (ocfs2_xattr_is_local(xe))
+		goto out;
+
+	vb.vb_xv = (struct ocfs2_xattr_value_root *)
+				(base + name_offset + name_len);
+
+	ret = ocfs2_xattr_get_clusters(inode, 0, &p_cluster,
+				       &num_clusters, &vb.vb_xv->xr_list,
+				       &ext_flags);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/*
+	 * We just need to check the 1st extent record, since we always
+	 * CoW the whole xattr. So there shouldn't be a xattr with
+	 * some REFCOUNT extent recs after the 1st one.
+	 */
+	if (!(ext_flags & OCFS2_EXT_REFCOUNTED))
+		goto out;
+
+	ret = ocfs2_lock_refcount_tree(osb, le64_to_cpu(di->i_refcount_loc),
+				       1, ref_tree, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/*
+	 * If we are deleting the xattr or the new size will be stored inside,
+	 * cool, leave it there, the xattr truncate process will remove them
+	 * for us(it still needs the refcount tree lock and the meta, credits).
+	 * And the worse case is that every cluster truncate will split the
+	 * refcount tree, and make the original extent become 3. So we will need
+	 * 2 * cluster more extent recs at most.
+	 */
+	if (!xi->value || xi->value_len <= OCFS2_XATTR_INLINE_SIZE) {
+
+		ret = ocfs2_refcounted_xattr_delete_need(inode,
+							 &(*ref_tree)->rf_ci,
+							 ref_root_bh, vb.vb_xv,
+							 meta_add, credits);
+		if (ret)
+			mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_refcount_cow_xattr(inode, di, &vb,
+				       *ref_tree, ref_root_bh, 0,
+				       le32_to_cpu(vb.vb_xv->xr_clusters), p);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	brelse(ref_root_bh);
+	return ret;
+}
+
+/*
+ * Add the REFCOUNTED flags for all the extent rec in ocfs2_xattr_value_root.
+ * The physical clusters will be added to refcount tree.
+ */
+static int ocfs2_xattr_value_attach_refcount(struct inode *inode,
+				struct ocfs2_xattr_value_root *xv,
+				struct ocfs2_extent_tree *value_et,
+				struct ocfs2_caching_info *ref_ci,
+				struct buffer_head *ref_root_bh,
+				struct ocfs2_cached_dealloc_ctxt *dealloc,
+				struct ocfs2_post_refcount *refcount)
+{
+	int ret = 0;
+	u32 clusters = le32_to_cpu(xv->xr_clusters);
+	u32 cpos, p_cluster, num_clusters;
+	struct ocfs2_extent_list *el = &xv->xr_list;
+	unsigned int ext_flags;
+
+	cpos = 0;
+	while (cpos < clusters) {
+		ret = ocfs2_xattr_get_clusters(inode, cpos, &p_cluster,
+					       &num_clusters, el, &ext_flags);
+
+		cpos += num_clusters;
+		if ((ext_flags & OCFS2_EXT_REFCOUNTED))
+			continue;
+
+		BUG_ON(!p_cluster);
+
+		ret = ocfs2_add_refcount_flag(inode, value_et,
+					      ref_ci, ref_root_bh,
+					      cpos - num_clusters,
+					      p_cluster, num_clusters,
+					      dealloc, refcount);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+	}
+
+	return ret;
+}
+
+/*
+ * Given a normal ocfs2_xattr_header, refcount all the entries which
+ * have value stored outside.
+ * Used for xattrs stored in inode and ocfs2_xattr_block.
+ */
+static int ocfs2_xattr_attach_refcount_normal(struct inode *inode,
+				struct ocfs2_xattr_value_buf *vb,
+				struct ocfs2_xattr_header *header,
+				struct ocfs2_caching_info *ref_ci,
+				struct buffer_head *ref_root_bh,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+
+	struct ocfs2_xattr_entry *xe;
+	struct ocfs2_xattr_value_root *xv;
+	struct ocfs2_extent_tree et;
+	int i, ret = 0;
+
+	for (i = 0; i < le16_to_cpu(header->xh_count); i++) {
+		xe = &header->xh_entries[i];
+
+		if (ocfs2_xattr_is_local(xe))
+			continue;
+
+		xv = (struct ocfs2_xattr_value_root *)((void *)header +
+			le16_to_cpu(xe->xe_name_offset) +
+			OCFS2_XATTR_SIZE(xe->xe_name_len));
+
+		vb->vb_xv = xv;
+		ocfs2_init_xattr_value_extent_tree(&et, INODE_CACHE(inode), vb);
+
+		ret = ocfs2_xattr_value_attach_refcount(inode, xv, &et,
+							ref_ci, ref_root_bh,
+							dealloc, NULL);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+	}
+
+	return ret;
+}
+
+static int ocfs2_xattr_inline_attach_refcount(struct inode *inode,
+				struct buffer_head *fe_bh,
+				struct ocfs2_caching_info *ref_ci,
+				struct buffer_head *ref_root_bh,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)fe_bh->b_data;
+	struct ocfs2_xattr_header *header = (struct ocfs2_xattr_header *)
+				(fe_bh->b_data + inode->i_sb->s_blocksize -
+				le16_to_cpu(di->i_xattr_inline_size));
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_bh = fe_bh,
+		.vb_access = ocfs2_journal_access_di,
+	};
+
+	return ocfs2_xattr_attach_refcount_normal(inode, &vb, header,
+						  ref_ci, ref_root_bh, dealloc);
+}
+
+struct ocfs2_xattr_tree_value_refcount_para {
+	struct ocfs2_caching_info *ref_ci;
+	struct buffer_head *ref_root_bh;
+	struct ocfs2_cached_dealloc_ctxt *dealloc;
+};
+
+static int ocfs2_get_xattr_tree_value_root(struct super_block *sb,
+					   struct ocfs2_xattr_bucket *bucket,
+					   int offset,
+					   struct ocfs2_xattr_value_root **xv,
+					   struct buffer_head **bh)
+{
+	int ret, block_off, name_offset;
+	struct ocfs2_xattr_header *xh = bucket_xh(bucket);
+	struct ocfs2_xattr_entry *xe = &xh->xh_entries[offset];
+	void *base;
+
+	ret = ocfs2_xattr_bucket_get_name_value(sb,
+						bucket_xh(bucket),
+						offset,
+						&block_off,
+						&name_offset);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	base = bucket_block(bucket, block_off);
+
+	*xv = (struct ocfs2_xattr_value_root *)(base + name_offset +
+			 OCFS2_XATTR_SIZE(xe->xe_name_len));
+
+	if (bh)
+		*bh = bucket->bu_bhs[block_off];
+out:
+	return ret;
+}
+
+/*
+ * For a given xattr bucket, refcount all the entries which
+ * have value stored outside.
+ */
+static int ocfs2_xattr_bucket_value_refcount(struct inode *inode,
+					     struct ocfs2_xattr_bucket *bucket,
+					     void *para)
+{
+	int i, ret = 0;
+	struct ocfs2_extent_tree et;
+	struct ocfs2_xattr_tree_value_refcount_para *ref =
+			(struct ocfs2_xattr_tree_value_refcount_para *)para;
+	struct ocfs2_xattr_header *xh =
+			(struct ocfs2_xattr_header *)bucket->bu_bhs[0]->b_data;
+	struct ocfs2_xattr_entry *xe;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_access = ocfs2_journal_access,
+	};
+	struct ocfs2_post_refcount refcount = {
+		.credits = bucket->bu_blocks,
+		.para = bucket,
+		.func = ocfs2_xattr_bucket_post_refcount,
+	};
+	struct ocfs2_post_refcount *p = NULL;
+
+	/* We only need post_refcount if we support metaecc. */
+	if (ocfs2_meta_ecc(OCFS2_SB(inode->i_sb)))
+		p = &refcount;
+
+	mlog(0, "refcount bucket %llu, count = %u\n",
+	     (unsigned long long)bucket_blkno(bucket),
+	     le16_to_cpu(xh->xh_count));
+	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_get_xattr_tree_value_root(inode->i_sb, bucket, i,
+						      &vb.vb_xv, &vb.vb_bh);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		ocfs2_init_xattr_value_extent_tree(&et,
+						   INODE_CACHE(inode), &vb);
+
+		ret = ocfs2_xattr_value_attach_refcount(inode, vb.vb_xv,
+							&et, ref->ref_ci,
+							ref->ref_root_bh,
+							ref->dealloc, p);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+	}
+
+	return ret;
+
+}
+
+static int ocfs2_refcount_xattr_tree_rec(struct inode *inode,
+				     struct buffer_head *root_bh,
+				     u64 blkno, u32 cpos, u32 len, void *para)
+{
+	return ocfs2_iterate_xattr_buckets(inode, blkno, len,
+					   ocfs2_xattr_bucket_value_refcount,
+					   para);
+}
+
+static int ocfs2_xattr_block_attach_refcount(struct inode *inode,
+				struct buffer_head *blk_bh,
+				struct ocfs2_caching_info *ref_ci,
+				struct buffer_head *ref_root_bh,
+				struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret = 0;
+	struct ocfs2_xattr_block *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;
+		struct ocfs2_xattr_value_buf vb = {
+			.vb_bh = blk_bh,
+			.vb_access = ocfs2_journal_access_xb,
+		};
+
+		ret = ocfs2_xattr_attach_refcount_normal(inode, &vb, header,
+							 ref_ci, ref_root_bh,
+							 dealloc);
+	} else {
+		struct ocfs2_xattr_tree_value_refcount_para para = {
+			.ref_ci = ref_ci,
+			.ref_root_bh = ref_root_bh,
+			.dealloc = dealloc,
+		};
+
+		ret = ocfs2_iterate_xattr_index_block(inode, blk_bh,
+						ocfs2_refcount_xattr_tree_rec,
+						&para);
+	}
+
+	return ret;
+}
+
+int ocfs2_xattr_attach_refcount_tree(struct inode *inode,
+				     struct buffer_head *fe_bh,
+				     struct ocfs2_caching_info *ref_ci,
+				     struct buffer_head *ref_root_bh,
+				     struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+	int ret = 0;
+	struct ocfs2_inode_info *oi = OCFS2_I(inode);
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)fe_bh->b_data;
+	struct buffer_head *blk_bh = NULL;
+
+	if (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL) {
+		ret = ocfs2_xattr_inline_attach_refcount(inode, fe_bh,
+							 ref_ci, ref_root_bh,
+							 dealloc);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
+	}
+
+	if (!di->i_xattr_loc)
+		goto out;
+
+	ret = ocfs2_read_xattr_block(inode, le64_to_cpu(di->i_xattr_loc),
+				     &blk_bh);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_xattr_block_attach_refcount(inode, blk_bh, ref_ci,
+						ref_root_bh, dealloc);
+	if (ret)
+		mlog_errno(ret);
+
+	brelse(blk_bh);
+out:
+
+	return ret;
+}
+
+typedef int (should_xattr_reflinked)(struct ocfs2_xattr_entry *xe);
+/*
+ * Store the information we need in xattr reflink.
+ * old_bh and new_bh are inode bh for the old and new inode.
+ */
+struct ocfs2_xattr_reflink {
+	struct inode *old_inode;
+	struct inode *new_inode;
+	struct buffer_head *old_bh;
+	struct buffer_head *new_bh;
+	struct ocfs2_caching_info *ref_ci;
+	struct buffer_head *ref_root_bh;
+	struct ocfs2_cached_dealloc_ctxt *dealloc;
+	should_xattr_reflinked *xattr_reflinked;
+};
+
+/*
+ * Given a xattr header and xe offset,
+ * return the proper xv and the corresponding bh.
+ * xattr in inode, block and xattr tree have different implementaions.
+ */
+typedef int (get_xattr_value_root)(struct super_block *sb,
+				   struct buffer_head *bh,
+				   struct ocfs2_xattr_header *xh,
+				   int offset,
+				   struct ocfs2_xattr_value_root **xv,
+				   struct buffer_head **ret_bh,
+				   void *para);
+
+/*
+ * Calculate all the xattr value root metadata stored in this xattr header and
+ * credits we need if we create them from the scratch.
+ * We use get_xattr_value_root so that all types of xattr container can use it.
+ */
+static int ocfs2_value_metas_in_xattr_header(struct super_block *sb,
+					     struct buffer_head *bh,
+					     struct ocfs2_xattr_header *xh,
+					     int *metas, int *credits,
+					     int *num_recs,
+					     get_xattr_value_root *func,
+					     void *para)
+{
+	int i, ret = 0;
+	struct ocfs2_xattr_value_root *xv;
+	struct ocfs2_xattr_entry *xe;
+
+	for (i = 0; i < le16_to_cpu(xh->xh_count); i++) {
+		xe = &xh->xh_entries[i];
+		if (ocfs2_xattr_is_local(xe))
+			continue;
+
+		ret = func(sb, bh, xh, i, &xv, NULL, para);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		*metas += le16_to_cpu(xv->xr_list.l_tree_depth) *
+			  le16_to_cpu(xv->xr_list.l_next_free_rec);
+
+		*credits += ocfs2_calc_extend_credits(sb,
+						&def_xv.xv.xr_list,
+						le32_to_cpu(xv->xr_clusters));
+
+		/*
+		 * If the value is a tree with depth > 1, We don't go deep
+		 * to the extent block, so just calculate a maximum record num.
+		 */
+		if (!xv->xr_list.l_tree_depth)
+			*num_recs += xv->xr_list.l_next_free_rec;
+		else
+			*num_recs += ocfs2_clusters_for_bytes(sb,
+							      XATTR_SIZE_MAX);
+	}
+
+	return ret;
+}
+
+/* Used by xattr inode and block to return the right xv and buffer_head. */
+static int ocfs2_get_xattr_value_root(struct super_block *sb,
+				      struct buffer_head *bh,
+				      struct ocfs2_xattr_header *xh,
+				      int offset,
+				      struct ocfs2_xattr_value_root **xv,
+				      struct buffer_head **ret_bh,
+				      void *para)
+{
+	struct ocfs2_xattr_entry *xe = &xh->xh_entries[offset];
+
+	*xv = (struct ocfs2_xattr_value_root *)((void *)xh +
+		le16_to_cpu(xe->xe_name_offset) +
+		OCFS2_XATTR_SIZE(xe->xe_name_len));
+
+	if (ret_bh)
+		*ret_bh = bh;
+
+	return 0;
+}
+
+/*
+ * Lock the meta_ac and caculate how much credits we need for reflink xattrs.
+ * It is only used for inline xattr and xattr block.
+ */
+static int ocfs2_reflink_lock_xattr_allocators(struct ocfs2_super *osb,
+					struct ocfs2_xattr_header *xh,
+					struct buffer_head *ref_root_bh,
+					int *credits,
+					struct ocfs2_alloc_context **meta_ac)
+{
+	int ret, meta_add = 0, num_recs = 0;
+	struct ocfs2_refcount_block *rb =
+			(struct ocfs2_refcount_block *)ref_root_bh->b_data;
+
+	*credits = 0;
+
+	ret = ocfs2_value_metas_in_xattr_header(osb->sb, NULL, xh,
+						&meta_add, credits, &num_recs,
+						ocfs2_get_xattr_value_root,
+						NULL);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	/*
+	 * We need to add/modify num_recs in refcount tree, so just calculate
+	 * an approximate number we need for refcount tree change.
+	 * Sometimes we need to split the tree, and after split,  half recs
+	 * will be moved to the new block, and a new block can only provide
+	 * half number of recs. So we multiple new blocks by 2.
+	 */
+	num_recs = num_recs / ocfs2_refcount_recs_per_rb(osb->sb) * 2;
+	meta_add += num_recs;
+	*credits += num_recs + num_recs * OCFS2_EXPAND_REFCOUNT_TREE_CREDITS;
+	if (le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL)
+		*credits += le16_to_cpu(rb->rf_list.l_tree_depth) *
+			    le16_to_cpu(rb->rf_list.l_next_free_rec) + 1;
+	else
+		*credits += 1;
+
+	ret = ocfs2_reserve_new_metadata_blocks(osb, meta_add, meta_ac);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	return ret;
+}
 
-		ret = ocfs2_iterate_xattr_buckets(inode, p_blkno, num_clusters,
-						  ocfs2_delete_xattr_in_bucket,
-						  NULL);
+/*
+ * Given a xattr header, reflink all the xattrs in this container.
+ * It can be used for inode, block and bucket.
+ *
+ * NOTE:
+ * Before we call this function, the caller has memcpy the xattr in
+ * old_xh to the new_xh.
+ *
+ * If args.xattr_reflinked is set, call it to decide whether the xe should
+ * be reflinked or not. If not, remove it from the new xattr header.
+ */
+static int ocfs2_reflink_xattr_header(handle_t *handle,
+				      struct ocfs2_xattr_reflink *args,
+				      struct buffer_head *old_bh,
+				      struct ocfs2_xattr_header *xh,
+				      struct buffer_head *new_bh,
+				      struct ocfs2_xattr_header *new_xh,
+				      struct ocfs2_xattr_value_buf *vb,
+				      struct ocfs2_alloc_context *meta_ac,
+				      get_xattr_value_root *func,
+				      void *para)
+{
+	int ret = 0, i, j;
+	struct super_block *sb = args->old_inode->i_sb;
+	struct buffer_head *value_bh;
+	struct ocfs2_xattr_entry *xe, *last;
+	struct ocfs2_xattr_value_root *xv, *new_xv;
+	struct ocfs2_extent_tree data_et;
+	u32 clusters, cpos, p_cluster, num_clusters;
+	unsigned int ext_flags = 0;
+
+	mlog(0, "reflink xattr in container %llu, count = %u\n",
+	     (unsigned long long)old_bh->b_blocknr, le16_to_cpu(xh->xh_count));
+
+	last = &new_xh->xh_entries[le16_to_cpu(new_xh->xh_count)];
+	for (i = 0, j = 0; i < le16_to_cpu(xh->xh_count); i++, j++) {
+		xe = &xh->xh_entries[i];
+
+		if (args->xattr_reflinked && !args->xattr_reflinked(xe)) {
+			xe = &new_xh->xh_entries[j];
+
+			le16_add_cpu(&new_xh->xh_count, -1);
+			if (new_xh->xh_count) {
+				memmove(xe, xe + 1,
+					(void *)last - (void *)xe);
+				memset(last, 0,
+				       sizeof(struct ocfs2_xattr_entry));
+			}
+
+			/*
+			 * We don't want j to increase in the next round since
+			 * it is already moved ahead.
+			 */
+			j--;
+			continue;
+		}
+
+		if (ocfs2_xattr_is_local(xe))
+			continue;
+
+		ret = func(sb, old_bh, xh, i, &xv, NULL, para);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		ret = func(sb, new_bh, new_xh, j, &new_xv, &value_bh, para);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		/*
+		 * For the xattr which has l_tree_depth = 0, all the extent
+		 * recs have already be copied to the new xh with the
+		 * propriate OCFS2_EXT_REFCOUNTED flag we just need to
+		 * increase the refount count int the refcount tree.
+		 *
+		 * For the xattr which has l_tree_depth > 0, we need
+		 * to initialize it to the empty default value root,
+		 * and then insert the extents one by one.
+		 */
+		if (xv->xr_list.l_tree_depth) {
+			memcpy(new_xv, &def_xv, sizeof(def_xv));
+			vb->vb_xv = new_xv;
+			vb->vb_bh = value_bh;
+			ocfs2_init_xattr_value_extent_tree(&data_et,
+					INODE_CACHE(args->new_inode), vb);
+		}
+
+		clusters = le32_to_cpu(xv->xr_clusters);
+		cpos = 0;
+		while (cpos < clusters) {
+			ret = ocfs2_xattr_get_clusters(args->old_inode,
+						       cpos,
+						       &p_cluster,
+						       &num_clusters,
+						       &xv->xr_list,
+						       &ext_flags);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+
+			BUG_ON(!p_cluster);
+
+			if (xv->xr_list.l_tree_depth) {
+				ret = ocfs2_insert_extent(handle,
+						&data_et, cpos,
+						ocfs2_clusters_to_blocks(
+							args->old_inode->i_sb,
+							p_cluster),
+						num_clusters, ext_flags,
+						meta_ac);
+				if (ret) {
+					mlog_errno(ret);
+					goto out;
+				}
+			}
+
+			ret = ocfs2_increase_refcount(handle, args->ref_ci,
+						      args->ref_root_bh,
+						      p_cluster, num_clusters,
+						      meta_ac, args->dealloc);
+			if (ret) {
+				mlog_errno(ret);
+				goto out;
+			}
+
+			cpos += num_clusters;
+		}
+	}
+
+out:
+	return ret;
+}
+
+static int ocfs2_reflink_xattr_inline(struct ocfs2_xattr_reflink *args)
+{
+	int ret = 0, credits = 0;
+	handle_t *handle;
+	struct ocfs2_super *osb = OCFS2_SB(args->old_inode->i_sb);
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)args->old_bh->b_data;
+	int inline_size = le16_to_cpu(di->i_xattr_inline_size);
+	int header_off = osb->sb->s_blocksize - inline_size;
+	struct ocfs2_xattr_header *xh = (struct ocfs2_xattr_header *)
+					(args->old_bh->b_data + header_off);
+	struct ocfs2_xattr_header *new_xh = (struct ocfs2_xattr_header *)
+					(args->new_bh->b_data + header_off);
+	struct ocfs2_alloc_context *meta_ac = NULL;
+	struct ocfs2_inode_info *new_oi;
+	struct ocfs2_dinode *new_di;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_bh = args->new_bh,
+		.vb_access = ocfs2_journal_access_di,
+	};
+
+	ret = ocfs2_reflink_lock_xattr_allocators(osb, xh, args->ref_root_bh,
+						  &credits, &meta_ac);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_journal_access_di(handle, INODE_CACHE(args->new_inode),
+				      args->new_bh, OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	memcpy(args->new_bh->b_data + header_off,
+	       args->old_bh->b_data + header_off, inline_size);
+
+	new_di = (struct ocfs2_dinode *)args->new_bh->b_data;
+	new_di->i_xattr_inline_size = cpu_to_le16(inline_size);
+
+	ret = ocfs2_reflink_xattr_header(handle, args, args->old_bh, xh,
+					 args->new_bh, new_xh, &vb, meta_ac,
+					 ocfs2_get_xattr_value_root, NULL);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	new_oi = OCFS2_I(args->new_inode);
+	spin_lock(&new_oi->ip_lock);
+	new_oi->ip_dyn_features |= OCFS2_HAS_XATTR_FL | OCFS2_INLINE_XATTR_FL;
+	new_di->i_dyn_features = cpu_to_le16(new_oi->ip_dyn_features);
+	spin_unlock(&new_oi->ip_lock);
+
+	ocfs2_journal_dirty(handle, args->new_bh);
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+
+out:
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+	return ret;
+}
+
+static int ocfs2_create_empty_xattr_block(struct inode *inode,
+					  struct buffer_head *fe_bh,
+					  struct buffer_head **ret_bh,
+					  int indexed)
+{
+	int ret;
+	handle_t *handle;
+	struct ocfs2_alloc_context *meta_ac;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+
+	ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
+	if (ret < 0) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	handle = ocfs2_start_trans(osb, OCFS2_XATTR_BLOCK_CREATE_CREDITS);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	mlog(0, "create new xattr block for inode %llu, index = %d\n",
+	     (unsigned long long)fe_bh->b_blocknr, indexed);
+	ret = ocfs2_create_xattr_block(handle, inode, fe_bh,
+				       meta_ac, ret_bh, indexed);
+	if (ret)
+		mlog_errno(ret);
+
+	ocfs2_commit_trans(osb, handle);
+out:
+	ocfs2_free_alloc_context(meta_ac);
+	return ret;
+}
+
+static int ocfs2_reflink_xattr_block(struct ocfs2_xattr_reflink *args,
+				     struct buffer_head *blk_bh,
+				     struct buffer_head *new_blk_bh)
+{
+	int ret = 0, credits = 0;
+	handle_t *handle;
+	struct ocfs2_inode_info *new_oi = OCFS2_I(args->new_inode);
+	struct ocfs2_dinode *new_di;
+	struct ocfs2_super *osb = OCFS2_SB(args->new_inode->i_sb);
+	int header_off = offsetof(struct ocfs2_xattr_block, xb_attrs.xb_header);
+	struct ocfs2_xattr_block *xb =
+			(struct ocfs2_xattr_block *)blk_bh->b_data;
+	struct ocfs2_xattr_header *xh = &xb->xb_attrs.xb_header;
+	struct ocfs2_xattr_block *new_xb =
+			(struct ocfs2_xattr_block *)new_blk_bh->b_data;
+	struct ocfs2_xattr_header *new_xh = &new_xb->xb_attrs.xb_header;
+	struct ocfs2_alloc_context *meta_ac;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_bh = new_blk_bh,
+		.vb_access = ocfs2_journal_access_xb,
+	};
+
+	ret = ocfs2_reflink_lock_xattr_allocators(osb, xh, args->ref_root_bh,
+						  &credits, &meta_ac);
+	if (ret) {
+		mlog_errno(ret);
+		return ret;
+	}
+
+	/* One more credits in case we need to add xattr flags in new inode. */
+	handle = ocfs2_start_trans(osb, credits + 1);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (!(new_oi->ip_dyn_features & OCFS2_HAS_XATTR_FL)) {
+		ret = ocfs2_journal_access_di(handle,
+					      INODE_CACHE(args->new_inode),
+					      args->new_bh,
+					      OCFS2_JOURNAL_ACCESS_WRITE);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_commit;
+		}
+	}
+
+	ret = ocfs2_journal_access_xb(handle, INODE_CACHE(args->new_inode),
+				      new_blk_bh, OCFS2_JOURNAL_ACCESS_WRITE);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	memcpy(new_blk_bh->b_data + header_off, blk_bh->b_data + header_off,
+	       osb->sb->s_blocksize - header_off);
+
+	ret = ocfs2_reflink_xattr_header(handle, args, blk_bh, xh,
+					 new_blk_bh, new_xh, &vb, meta_ac,
+					 ocfs2_get_xattr_value_root, NULL);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	ocfs2_journal_dirty(handle, new_blk_bh);
+
+	if (!(new_oi->ip_dyn_features & OCFS2_HAS_XATTR_FL)) {
+		new_di = (struct ocfs2_dinode *)args->new_bh->b_data;
+		spin_lock(&new_oi->ip_lock);
+		new_oi->ip_dyn_features |= OCFS2_HAS_XATTR_FL;
+		new_di->i_dyn_features = cpu_to_le16(new_oi->ip_dyn_features);
+		spin_unlock(&new_oi->ip_lock);
+
+		ocfs2_journal_dirty(handle, args->new_bh);
+	}
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+
+out:
+	ocfs2_free_alloc_context(meta_ac);
+	return ret;
+}
+
+struct ocfs2_reflink_xattr_tree_args {
+	struct ocfs2_xattr_reflink *reflink;
+	struct buffer_head *old_blk_bh;
+	struct buffer_head *new_blk_bh;
+	struct ocfs2_xattr_bucket *old_bucket;
+	struct ocfs2_xattr_bucket *new_bucket;
+};
+
+/*
+ * NOTE:
+ * We have to handle the case that both old bucket and new bucket
+ * will call this function to get the right ret_bh.
+ * So The caller must give us the right bh.
+ */
+static int ocfs2_get_reflink_xattr_value_root(struct super_block *sb,
+					struct buffer_head *bh,
+					struct ocfs2_xattr_header *xh,
+					int offset,
+					struct ocfs2_xattr_value_root **xv,
+					struct buffer_head **ret_bh,
+					void *para)
+{
+	struct ocfs2_reflink_xattr_tree_args *args =
+			(struct ocfs2_reflink_xattr_tree_args *)para;
+	struct ocfs2_xattr_bucket *bucket;
+
+	if (bh == args->old_bucket->bu_bhs[0])
+		bucket = args->old_bucket;
+	else
+		bucket = args->new_bucket;
+
+	return ocfs2_get_xattr_tree_value_root(sb, bucket, offset,
+					       xv, ret_bh);
+}
+
+struct ocfs2_value_tree_metas {
+	int num_metas;
+	int credits;
+	int num_recs;
+};
+
+static int ocfs2_value_tree_metas_in_bucket(struct super_block *sb,
+					struct buffer_head *bh,
+					struct ocfs2_xattr_header *xh,
+					int offset,
+					struct ocfs2_xattr_value_root **xv,
+					struct buffer_head **ret_bh,
+					void *para)
+{
+	struct ocfs2_xattr_bucket *bucket =
+				(struct ocfs2_xattr_bucket *)para;
+
+	return ocfs2_get_xattr_tree_value_root(sb, bucket, offset,
+					       xv, ret_bh);
+}
+
+static int ocfs2_calc_value_tree_metas(struct inode *inode,
+				      struct ocfs2_xattr_bucket *bucket,
+				      void *para)
+{
+	struct ocfs2_value_tree_metas *metas =
+			(struct ocfs2_value_tree_metas *)para;
+	struct ocfs2_xattr_header *xh =
+			(struct ocfs2_xattr_header *)bucket->bu_bhs[0]->b_data;
+
+	/* Add the credits for this bucket first. */
+	metas->credits += bucket->bu_blocks;
+	return ocfs2_value_metas_in_xattr_header(inode->i_sb, bucket->bu_bhs[0],
+					xh, &metas->num_metas,
+					&metas->credits, &metas->num_recs,
+					ocfs2_value_tree_metas_in_bucket,
+					bucket);
+}
+
+/*
+ * Given a xattr extent rec starting from blkno and having len clusters,
+ * iterate all the buckets calculate how much metadata we need for reflinking
+ * all the ocfs2_xattr_value_root and lock the allocators accordingly.
+ */
+static int ocfs2_lock_reflink_xattr_rec_allocators(
+				struct ocfs2_reflink_xattr_tree_args *args,
+				struct ocfs2_extent_tree *xt_et,
+				u64 blkno, u32 len, int *credits,
+				struct ocfs2_alloc_context **meta_ac,
+				struct ocfs2_alloc_context **data_ac)
+{
+	int ret, num_free_extents;
+	struct ocfs2_value_tree_metas metas;
+	struct ocfs2_super *osb = OCFS2_SB(args->reflink->old_inode->i_sb);
+	struct ocfs2_refcount_block *rb;
+
+	memset(&metas, 0, sizeof(metas));
+
+	ret = ocfs2_iterate_xattr_buckets(args->reflink->old_inode, blkno, len,
+					  ocfs2_calc_value_tree_metas, &metas);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	*credits = metas.credits;
+
+	/*
+	 * Calculate we need for refcount tree change.
+	 *
+	 * We need to add/modify num_recs in refcount tree, so just calculate
+	 * an approximate number we need for refcount tree change.
+	 * Sometimes we need to split the tree, and after split,  half recs
+	 * will be moved to the new block, and a new block can only provide
+	 * half number of recs. So we multiple new blocks by 2.
+	 * In the end, we have to add credits for modifying the already
+	 * existed refcount block.
+	 */
+	rb = (struct ocfs2_refcount_block *)args->reflink->ref_root_bh->b_data;
+	metas.num_recs =
+		(metas.num_recs + ocfs2_refcount_recs_per_rb(osb->sb) - 1) /
+		 ocfs2_refcount_recs_per_rb(osb->sb) * 2;
+	metas.num_metas += metas.num_recs;
+	*credits += metas.num_recs +
+		    metas.num_recs * OCFS2_EXPAND_REFCOUNT_TREE_CREDITS;
+	if (le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL)
+		*credits += le16_to_cpu(rb->rf_list.l_tree_depth) *
+			    le16_to_cpu(rb->rf_list.l_next_free_rec) + 1;
+	else
+		*credits += 1;
+
+	/* count in the xattr tree change. */
+	num_free_extents = ocfs2_num_free_extents(osb, xt_et);
+	if (num_free_extents < 0) {
+		ret = num_free_extents;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (num_free_extents < len)
+		metas.num_metas += ocfs2_extend_meta_needed(xt_et->et_root_el);
+
+	*credits += ocfs2_calc_extend_credits(osb->sb,
+					      xt_et->et_root_el, len);
+
+	if (metas.num_metas) {
+		ret = ocfs2_reserve_new_metadata_blocks(osb, metas.num_metas,
+							meta_ac);
 		if (ret) {
 			mlog_errno(ret);
 			goto out;
 		}
+	}
 
-		ret = ocfs2_rm_xattr_cluster(inode, xb_bh,
-					     p_blkno, e_cpos, num_clusters);
+	if (len) {
+		ret = ocfs2_reserve_clusters(osb, len, data_ac);
+		if (ret)
+			mlog_errno(ret);
+	}
+out:
+	if (ret) {
+		if (*meta_ac) {
+			ocfs2_free_alloc_context(*meta_ac);
+			meta_ac = NULL;
+		}
+	}
+
+	return ret;
+}
+
+static int ocfs2_reflink_xattr_buckets(handle_t *handle,
+				u64 blkno, u64 new_blkno, u32 clusters,
+				struct ocfs2_alloc_context *meta_ac,
+				struct ocfs2_alloc_context *data_ac,
+				struct ocfs2_reflink_xattr_tree_args *args)
+{
+	int i, j, ret = 0;
+	struct super_block *sb = args->reflink->old_inode->i_sb;
+	u32 bpc = ocfs2_xattr_buckets_per_cluster(OCFS2_SB(sb));
+	u32 num_buckets = clusters * bpc;
+	int bpb = args->old_bucket->bu_blocks;
+	struct ocfs2_xattr_value_buf vb = {
+		.vb_access = ocfs2_journal_access,
+	};
+
+	for (i = 0; i < num_buckets; i++, blkno += bpb, new_blkno += bpb) {
+		ret = ocfs2_read_xattr_bucket(args->old_bucket, blkno);
 		if (ret) {
 			mlog_errno(ret);
 			break;
 		}
 
-		if (e_cpos == 0)
+		ret = ocfs2_init_xattr_bucket(args->new_bucket, new_blkno);
+		if (ret) {
+			mlog_errno(ret);
 			break;
+		}
 
-		name_hash = e_cpos - 1;
+		/*
+		 * 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(args->old_bucket)->xh_num_buckets);
+
+		ret = ocfs2_xattr_bucket_journal_access(handle,
+						args->new_bucket,
+						OCFS2_JOURNAL_ACCESS_CREATE);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		for (j = 0; j < bpb; j++)
+			memcpy(bucket_block(args->new_bucket, j),
+			       bucket_block(args->old_bucket, j),
+			       sb->s_blocksize);
+
+		ocfs2_xattr_bucket_journal_dirty(handle, args->new_bucket);
+
+		ret = ocfs2_reflink_xattr_header(handle, args->reflink,
+					args->old_bucket->bu_bhs[0],
+					bucket_xh(args->old_bucket),
+					args->new_bucket->bu_bhs[0],
+					bucket_xh(args->new_bucket),
+					&vb, meta_ac,
+					ocfs2_get_reflink_xattr_value_root,
+					args);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		/*
+		 * Re-access and dirty the bucket to calculate metaecc.
+		 * Because we may extend the transaction in reflink_xattr_header
+		 * which will let the already accessed block gone.
+		 */
+		ret = ocfs2_xattr_bucket_journal_access(handle,
+						args->new_bucket,
+						OCFS2_JOURNAL_ACCESS_WRITE);
+		if (ret) {
+			mlog_errno(ret);
+			break;
+		}
+
+		ocfs2_xattr_bucket_journal_dirty(handle, args->new_bucket);
+		ocfs2_xattr_bucket_relse(args->old_bucket);
+		ocfs2_xattr_bucket_relse(args->new_bucket);
+	}
+
+	ocfs2_xattr_bucket_relse(args->old_bucket);
+	ocfs2_xattr_bucket_relse(args->new_bucket);
+	return ret;
+}
+/*
+ * Create the same xattr extent record in the new inode's xattr tree.
+ */
+static int ocfs2_reflink_xattr_rec(struct inode *inode,
+				   struct buffer_head *root_bh,
+				   u64 blkno,
+				   u32 cpos,
+				   u32 len,
+				   void *para)
+{
+	int ret, credits = 0;
+	u32 p_cluster, num_clusters;
+	u64 new_blkno;
+	handle_t *handle;
+	struct ocfs2_reflink_xattr_tree_args *args =
+			(struct ocfs2_reflink_xattr_tree_args *)para;
+	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+	struct ocfs2_alloc_context *meta_ac = NULL;
+	struct ocfs2_alloc_context *data_ac = NULL;
+	struct ocfs2_extent_tree et;
+
+	ocfs2_init_xattr_tree_extent_tree(&et,
+					  INODE_CACHE(args->reflink->new_inode),
+					  args->new_blk_bh);
+
+	ret = ocfs2_lock_reflink_xattr_rec_allocators(args, &et, blkno,
+						      len, &credits,
+						      &meta_ac, &data_ac);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	handle = ocfs2_start_trans(osb, credits);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_claim_clusters(osb, handle, data_ac,
+				   len, &p_cluster, &num_clusters);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	new_blkno = ocfs2_clusters_to_blocks(osb->sb, p_cluster);
+
+	mlog(0, "reflink xattr buckets %llu to %llu, len %u\n",
+	     (unsigned long long)blkno, (unsigned long long)new_blkno, len);
+	ret = ocfs2_reflink_xattr_buckets(handle, blkno, new_blkno, len,
+					  meta_ac, data_ac, args);
+	if (ret) {
+		mlog_errno(ret);
+		goto out_commit;
+	}
+
+	mlog(0, "insert new xattr extent rec start %llu len %u to %u\n",
+	     (unsigned long long)new_blkno, len, cpos);
+	ret = ocfs2_insert_extent(handle, &et, cpos, new_blkno,
+				  len, 0, meta_ac);
+	if (ret)
+		mlog_errno(ret);
+
+out_commit:
+	ocfs2_commit_trans(osb, handle);
+
+out:
+	if (meta_ac)
+		ocfs2_free_alloc_context(meta_ac);
+	if (data_ac)
+		ocfs2_free_alloc_context(data_ac);
+	return ret;
+}
+
+/*
+ * Create reflinked xattr buckets.
+ * We will add bucket one by one, and refcount all the xattrs in the bucket
+ * if they are stored outside.
+ */
+static int ocfs2_reflink_xattr_tree(struct ocfs2_xattr_reflink *args,
+				    struct buffer_head *blk_bh,
+				    struct buffer_head *new_blk_bh)
+{
+	int ret;
+	struct ocfs2_reflink_xattr_tree_args para;
+
+	memset(&para, 0, sizeof(para));
+	para.reflink = args;
+	para.old_blk_bh = blk_bh;
+	para.new_blk_bh = new_blk_bh;
+
+	para.old_bucket = ocfs2_xattr_bucket_new(args->old_inode);
+	if (!para.old_bucket) {
+		mlog_errno(-ENOMEM);
+		return -ENOMEM;
+	}
+
+	para.new_bucket = ocfs2_xattr_bucket_new(args->new_inode);
+	if (!para.new_bucket) {
+		ret = -ENOMEM;
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ret = ocfs2_iterate_xattr_index_block(args->old_inode, blk_bh,
+					      ocfs2_reflink_xattr_rec,
+					      &para);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	ocfs2_xattr_bucket_free(para.old_bucket);
+	ocfs2_xattr_bucket_free(para.new_bucket);
+	return ret;
+}
+
+static int ocfs2_reflink_xattr_in_block(struct ocfs2_xattr_reflink *args,
+					struct buffer_head *blk_bh)
+{
+	int ret, indexed = 0;
+	struct buffer_head *new_blk_bh = NULL;
+	struct ocfs2_xattr_block *xb =
+			(struct ocfs2_xattr_block *)blk_bh->b_data;
+
+
+	if (le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED)
+		indexed = 1;
+
+	ret = ocfs2_create_empty_xattr_block(args->new_inode, args->new_bh,
+					     &new_blk_bh, indexed);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	if (!(le16_to_cpu(xb->xb_flags) & OCFS2_XATTR_INDEXED))
+		ret = ocfs2_reflink_xattr_block(args, blk_bh, new_blk_bh);
+	else
+		ret = ocfs2_reflink_xattr_tree(args, blk_bh, new_blk_bh);
+	if (ret)
+		mlog_errno(ret);
+
+out:
+	brelse(new_blk_bh);
+	return ret;
+}
+
+static int ocfs2_reflink_xattr_no_security(struct ocfs2_xattr_entry *xe)
+{
+	int type = ocfs2_xattr_get_type(xe);
+
+	return type != OCFS2_XATTR_INDEX_SECURITY &&
+	       type != OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS &&
+	       type != OCFS2_XATTR_INDEX_POSIX_ACL_DEFAULT;
+}
+
+int ocfs2_reflink_xattrs(struct inode *old_inode,
+			 struct buffer_head *old_bh,
+			 struct inode *new_inode,
+			 struct buffer_head *new_bh,
+			 bool preserve_security)
+{
+	int ret;
+	struct ocfs2_xattr_reflink args;
+	struct ocfs2_inode_info *oi = OCFS2_I(old_inode);
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)old_bh->b_data;
+	struct buffer_head *blk_bh = NULL;
+	struct ocfs2_cached_dealloc_ctxt dealloc;
+	struct ocfs2_refcount_tree *ref_tree;
+	struct buffer_head *ref_root_bh = NULL;
+
+	ret = ocfs2_lock_refcount_tree(OCFS2_SB(old_inode->i_sb),
+				       le64_to_cpu(di->i_refcount_loc),
+				       1, &ref_tree, &ref_root_bh);
+	if (ret) {
+		mlog_errno(ret);
+		goto out;
+	}
+
+	ocfs2_init_dealloc_ctxt(&dealloc);
+
+	args.old_inode = old_inode;
+	args.new_inode = new_inode;
+	args.old_bh = old_bh;
+	args.new_bh = new_bh;
+	args.ref_ci = &ref_tree->rf_ci;
+	args.ref_root_bh = ref_root_bh;
+	args.dealloc = &dealloc;
+	if (preserve_security)
+		args.xattr_reflinked = NULL;
+	else
+		args.xattr_reflinked = ocfs2_reflink_xattr_no_security;
+
+	if (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL) {
+		ret = ocfs2_reflink_xattr_inline(&args);
+		if (ret) {
+			mlog_errno(ret);
+			goto out_unlock;
+		}
+	}
+
+	if (!di->i_xattr_loc)
+		goto out_unlock;
+
+	ret = ocfs2_read_xattr_block(old_inode, le64_to_cpu(di->i_xattr_loc),
+				     &blk_bh);
+	if (ret < 0) {
+		mlog_errno(ret);
+		goto out_unlock;
+	}
+
+	ret = ocfs2_reflink_xattr_in_block(&args, blk_bh);
+	if (ret)
+		mlog_errno(ret);
+
+	brelse(blk_bh);
+
+out_unlock:
+	ocfs2_unlock_refcount_tree(OCFS2_SB(old_inode->i_sb),
+				   ref_tree, 1);
+	brelse(ref_root_bh);
+
+	if (ocfs2_dealloc_has_cluster(&dealloc)) {
+		ocfs2_schedule_truncate_log_flush(OCFS2_SB(old_inode->i_sb), 1);
+		ocfs2_run_deallocs(OCFS2_SB(old_inode->i_sb), &dealloc);
 	}
 
 out:
@@ -5306,6 +6961,51 @@ out:
 }
 
 /*
+ * Initialize security and acl for a already created inode.
+ * Used for reflink a non-preserve-security file.
+ *
+ * It uses common api like ocfs2_xattr_set, so the caller
+ * must not hold any lock expect i_mutex.
+ */
+int ocfs2_init_security_and_acl(struct inode *dir,
+				struct inode *inode)
+{
+	int ret = 0;
+	struct buffer_head *dir_bh = NULL;
+	struct ocfs2_security_xattr_info si = {
+		.enable = 1,
+	};
+
+	ret = ocfs2_init_security_get(inode, dir, &si);
+	if (!ret) {
+		ret = ocfs2_xattr_security_set(inode, si.name,
+					       si.value, si.value_len,
+					       XATTR_CREATE);
+		if (ret) {
+			mlog_errno(ret);
+			goto leave;
+		}
+	} else if (ret != -EOPNOTSUPP) {
+		mlog_errno(ret);
+		goto leave;
+	}
+
+	ret = ocfs2_inode_lock(dir, &dir_bh, 0);
+	if (ret) {
+		mlog_errno(ret);
+		goto leave;
+	}
+
+	ret = ocfs2_init_acl(NULL, inode, dir, NULL, dir_bh, NULL, NULL);
+	if (ret)
+		mlog_errno(ret);
+
+	ocfs2_inode_unlock(dir, 0);
+	brelse(dir_bh);
+leave:
+	return ret;
+}
+/*
  * 'security' attributes support
  */
 static size_t ocfs2_xattr_security_list(struct inode *inode, char *list,
diff --git a/fs/ocfs2/xattr.h b/fs/ocfs2/xattr.h
index 1ca7e9a1b7b..08e36389f56 100644
--- a/fs/ocfs2/xattr.h
+++ b/fs/ocfs2/xattr.h
@@ -55,6 +55,8 @@ 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_has_inline_xattr_value_outside(struct inode *inode,
+					 struct ocfs2_dinode *di);
 int ocfs2_xattr_remove(struct inode *, struct buffer_head *);
 int ocfs2_init_security_get(struct inode *, struct inode *,
 			    struct ocfs2_security_xattr_info *);
@@ -83,5 +85,16 @@ struct ocfs2_xattr_value_buf {
 	struct ocfs2_xattr_value_root	*vb_xv;
 };
 
-
+int ocfs2_xattr_attach_refcount_tree(struct inode *inode,
+				     struct buffer_head *fe_bh,
+				     struct ocfs2_caching_info *ref_ci,
+				     struct buffer_head *ref_root_bh,
+				     struct ocfs2_cached_dealloc_ctxt *dealloc);
+int ocfs2_reflink_xattrs(struct inode *old_inode,
+			 struct buffer_head *old_bh,
+			 struct inode *new_inode,
+			 struct buffer_head *new_bh,
+			 bool preserve_security);
+int ocfs2_init_security_and_acl(struct inode *dir,
+				struct inode *inode);
 #endif /* OCFS2_XATTR_H */
diff --git a/fs/omfs/dir.c b/fs/omfs/dir.c
index c7275cfbdcf..3680bae335b 100644
--- a/fs/omfs/dir.c
+++ b/fs/omfs/dir.c
@@ -489,7 +489,7 @@ out:
 	return ret;
 }
 
-struct inode_operations omfs_dir_inops = {
+const struct inode_operations omfs_dir_inops = {
 	.lookup = omfs_lookup,
 	.mkdir = omfs_mkdir,
 	.rename = omfs_rename,
diff --git a/fs/omfs/file.c b/fs/omfs/file.c
index d17e774eaf4..4845fbb18e6 100644
--- a/fs/omfs/file.c
+++ b/fs/omfs/file.c
@@ -333,11 +333,11 @@ struct file_operations omfs_file_operations = {
 	.splice_read = generic_file_splice_read,
 };
 
-struct inode_operations omfs_file_inops = {
+const struct inode_operations omfs_file_inops = {
 	.truncate = omfs_truncate
 };
 
-struct address_space_operations omfs_aops = {
+const struct address_space_operations omfs_aops = {
 	.readpage = omfs_readpage,
 	.readpages = omfs_readpages,
 	.writepage = omfs_writepage,
diff --git a/fs/omfs/inode.c b/fs/omfs/inode.c
index 379ae5fb441..f3b7c1541f3 100644
--- a/fs/omfs/inode.c
+++ b/fs/omfs/inode.c
@@ -278,7 +278,7 @@ static int omfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	return 0;
 }
 
-static struct super_operations omfs_sops = {
+static const struct super_operations omfs_sops = {
 	.write_inode	= omfs_write_inode,
 	.delete_inode	= omfs_delete_inode,
 	.put_super	= omfs_put_super,
diff --git a/fs/omfs/omfs.h b/fs/omfs/omfs.h
index 2bc0f067040..df71039945a 100644
--- a/fs/omfs/omfs.h
+++ b/fs/omfs/omfs.h
@@ -45,15 +45,15 @@ extern int omfs_clear_range(struct super_block *sb, u64 block, int count);
 
 /* dir.c */
 extern struct file_operations omfs_dir_operations;
-extern struct inode_operations omfs_dir_inops;
+extern const struct inode_operations omfs_dir_inops;
 extern int omfs_make_empty(struct inode *inode, struct super_block *sb);
 extern int omfs_is_bad(struct omfs_sb_info *sbi, struct omfs_header *header,
 			u64 fsblock);
 
 /* file.c */
 extern struct file_operations omfs_file_operations;
-extern struct inode_operations omfs_file_inops;
-extern struct address_space_operations omfs_aops;
+extern const struct inode_operations omfs_file_inops;
+extern const struct address_space_operations omfs_aops;
 extern void omfs_make_empty_table(struct buffer_head *bh, int offset);
 extern int omfs_shrink_inode(struct inode *inode);
 
diff --git a/fs/open.c b/fs/open.c
index dd98e807602..4f01e06227c 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -199,7 +199,7 @@ out:
 int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
 	struct file *filp)
 {
-	int err;
+	int ret;
 	struct iattr newattrs;
 
 	/* Not pretty: "inode->i_size" shouldn't really be signed. But it is. */
@@ -214,12 +214,14 @@ int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
 	}
 
 	/* Remove suid/sgid on truncate too */
-	newattrs.ia_valid |= should_remove_suid(dentry);
+	ret = should_remove_suid(dentry);
+	if (ret)
+		newattrs.ia_valid |= ret | ATTR_FORCE;
 
 	mutex_lock(&dentry->d_inode->i_mutex);
-	err = notify_change(dentry, &newattrs);
+	ret = notify_change(dentry, &newattrs);
 	mutex_unlock(&dentry->d_inode->i_mutex);
-	return err;
+	return ret;
 }
 
 static long do_sys_truncate(const char __user *pathname, loff_t length)
@@ -288,10 +290,9 @@ out:
 	return error;
 }
 
-SYSCALL_DEFINE2(truncate, const char __user *, path, unsigned long, length)
+SYSCALL_DEFINE2(truncate, const char __user *, path, long, length)
 {
-	/* on 32-bit boxen it will cut the range 2^31--2^32-1 off */
-	return do_sys_truncate(path, (long)length);
+	return do_sys_truncate(path, length);
 }
 
 static long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
@@ -957,6 +958,8 @@ struct file *dentry_open(struct dentry *dentry, struct vfsmount *mnt, int flags,
 	int error;
 	struct file *f;
 
+	validate_creds(cred);
+
 	/*
 	 * We must always pass in a valid mount pointer.   Historically
 	 * callers got away with not passing it, but we must enforce this at
diff --git a/fs/partitions/check.c b/fs/partitions/check.c
index ea4e6cb29e1..7b685e10cba 100644
--- a/fs/partitions/check.c
+++ b/fs/partitions/check.c
@@ -248,11 +248,19 @@ ssize_t part_stat_show(struct device *dev,
 		part_stat_read(p, merges[WRITE]),
 		(unsigned long long)part_stat_read(p, sectors[WRITE]),
 		jiffies_to_msecs(part_stat_read(p, ticks[WRITE])),
-		p->in_flight,
+		part_in_flight(p),
 		jiffies_to_msecs(part_stat_read(p, io_ticks)),
 		jiffies_to_msecs(part_stat_read(p, time_in_queue)));
 }
 
+ssize_t part_inflight_show(struct device *dev,
+			struct device_attribute *attr, char *buf)
+{
+	struct hd_struct *p = dev_to_part(dev);
+
+	return sprintf(buf, "%8u %8u\n", p->in_flight[0], p->in_flight[1]);
+}
+
 #ifdef CONFIG_FAIL_MAKE_REQUEST
 ssize_t part_fail_show(struct device *dev,
 		       struct device_attribute *attr, char *buf)
@@ -281,6 +289,7 @@ static DEVICE_ATTR(start, S_IRUGO, part_start_show, NULL);
 static DEVICE_ATTR(size, S_IRUGO, part_size_show, NULL);
 static DEVICE_ATTR(alignment_offset, S_IRUGO, part_alignment_offset_show, NULL);
 static DEVICE_ATTR(stat, S_IRUGO, part_stat_show, NULL);
+static DEVICE_ATTR(inflight, S_IRUGO, part_inflight_show, NULL);
 #ifdef CONFIG_FAIL_MAKE_REQUEST
 static struct device_attribute dev_attr_fail =
 	__ATTR(make-it-fail, S_IRUGO|S_IWUSR, part_fail_show, part_fail_store);
@@ -292,6 +301,7 @@ static struct attribute *part_attrs[] = {
 	&dev_attr_size.attr,
 	&dev_attr_alignment_offset.attr,
 	&dev_attr_stat.attr,
+	&dev_attr_inflight.attr,
 #ifdef CONFIG_FAIL_MAKE_REQUEST
 	&dev_attr_fail.attr,
 #endif
@@ -302,7 +312,7 @@ static struct attribute_group part_attr_group = {
 	.attrs = part_attrs,
 };
 
-static struct attribute_group *part_attr_groups[] = {
+static const struct attribute_group *part_attr_groups[] = {
 	&part_attr_group,
 #ifdef CONFIG_BLK_DEV_IO_TRACE
 	&blk_trace_attr_group,
@@ -571,7 +581,7 @@ try_scan:
 		}
 
 		if (from + size > get_capacity(disk)) {
-			struct block_device_operations *bdops = disk->fops;
+			const struct block_device_operations *bdops = disk->fops;
 			unsigned long long capacity;
 
 			printk(KERN_WARNING
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 725a650bbbb..0c6bc602e6c 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -82,6 +82,7 @@
 #include <linux/pid_namespace.h>
 #include <linux/ptrace.h>
 #include <linux/tracehook.h>
+#include <linux/swapops.h>
 
 #include <asm/pgtable.h>
 #include <asm/processor.h>
@@ -321,6 +322,87 @@ static inline void task_context_switch_counts(struct seq_file *m,
 			p->nivcsw);
 }
 
+struct stack_stats {
+	struct vm_area_struct *vma;
+	unsigned long	startpage;
+	unsigned long	usage;
+};
+
+static int stack_usage_pte_range(pmd_t *pmd, unsigned long addr,
+				unsigned long end, struct mm_walk *walk)
+{
+	struct stack_stats *ss = walk->private;
+	struct vm_area_struct *vma = ss->vma;
+	pte_t *pte, ptent;
+	spinlock_t *ptl;
+	int ret = 0;
+
+	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
+	for (; addr != end; pte++, addr += PAGE_SIZE) {
+		ptent = *pte;
+
+#ifdef CONFIG_STACK_GROWSUP
+		if (pte_present(ptent) || is_swap_pte(ptent))
+			ss->usage = addr - ss->startpage + PAGE_SIZE;
+#else
+		if (pte_present(ptent) || is_swap_pte(ptent)) {
+			ss->usage = ss->startpage - addr + PAGE_SIZE;
+			pte++;
+			ret = 1;
+			break;
+		}
+#endif
+	}
+	pte_unmap_unlock(pte - 1, ptl);
+	cond_resched();
+	return ret;
+}
+
+static inline unsigned long get_stack_usage_in_bytes(struct vm_area_struct *vma,
+				struct task_struct *task)
+{
+	struct stack_stats ss;
+	struct mm_walk stack_walk = {
+		.pmd_entry = stack_usage_pte_range,
+		.mm = vma->vm_mm,
+		.private = &ss,
+	};
+
+	if (!vma->vm_mm || is_vm_hugetlb_page(vma))
+		return 0;
+
+	ss.vma = vma;
+	ss.startpage = task->stack_start & PAGE_MASK;
+	ss.usage = 0;
+
+#ifdef CONFIG_STACK_GROWSUP
+	walk_page_range(KSTK_ESP(task) & PAGE_MASK, vma->vm_end,
+		&stack_walk);
+#else
+	walk_page_range(vma->vm_start, (KSTK_ESP(task) & PAGE_MASK) + PAGE_SIZE,
+		&stack_walk);
+#endif
+	return ss.usage;
+}
+
+static inline void task_show_stack_usage(struct seq_file *m,
+						struct task_struct *task)
+{
+	struct vm_area_struct	*vma;
+	struct mm_struct	*mm = get_task_mm(task);
+
+	if (mm) {
+		down_read(&mm->mmap_sem);
+		vma = find_vma(mm, task->stack_start);
+		if (vma)
+			seq_printf(m, "Stack usage:\t%lu kB\n",
+				get_stack_usage_in_bytes(vma, task) >> 10);
+
+		up_read(&mm->mmap_sem);
+		mmput(mm);
+	}
+}
+
 int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
 			struct pid *pid, struct task_struct *task)
 {
@@ -340,6 +422,7 @@ int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
 	task_show_regs(m, task);
 #endif
 	task_context_switch_counts(m, task);
+	task_show_stack_usage(m, task);
 	return 0;
 }
 
@@ -481,7 +564,7 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns,
 		rsslim,
 		mm ? mm->start_code : 0,
 		mm ? mm->end_code : 0,
-		(permitted && mm) ? mm->start_stack : 0,
+		(permitted) ? task->stack_start : 0,
 		esp,
 		eip,
 		/* The signal information here is obsolete.
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 6f742f6658a..837469a9659 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -447,7 +447,7 @@ static int proc_oom_score(struct task_struct *task, char *buffer)
 
 	do_posix_clock_monotonic_gettime(&uptime);
 	read_lock(&tasklist_lock);
-	points = badness(task, uptime.tv_sec);
+	points = badness(task->group_leader, uptime.tv_sec);
 	read_unlock(&tasklist_lock);
 	return sprintf(buffer, "%lu\n", points);
 }
@@ -458,7 +458,7 @@ struct limit_names {
 };
 
 static const struct limit_names lnames[RLIM_NLIMITS] = {
-	[RLIMIT_CPU] = {"Max cpu time", "ms"},
+	[RLIMIT_CPU] = {"Max cpu time", "seconds"},
 	[RLIMIT_FSIZE] = {"Max file size", "bytes"},
 	[RLIMIT_DATA] = {"Max data size", "bytes"},
 	[RLIMIT_STACK] = {"Max stack size", "bytes"},
@@ -999,11 +999,17 @@ static ssize_t oom_adjust_read(struct file *file, char __user *buf,
 	struct task_struct *task = get_proc_task(file->f_path.dentry->d_inode);
 	char buffer[PROC_NUMBUF];
 	size_t len;
-	int oom_adjust;
+	int oom_adjust = OOM_DISABLE;
+	unsigned long flags;
 
 	if (!task)
 		return -ESRCH;
-	oom_adjust = task->oomkilladj;
+
+	if (lock_task_sighand(task, &flags)) {
+		oom_adjust = task->signal->oom_adj;
+		unlock_task_sighand(task, &flags);
+	}
+
 	put_task_struct(task);
 
 	len = snprintf(buffer, sizeof(buffer), "%i\n", oom_adjust);
@@ -1015,32 +1021,44 @@ static ssize_t oom_adjust_write(struct file *file, const char __user *buf,
 				size_t count, loff_t *ppos)
 {
 	struct task_struct *task;
-	char buffer[PROC_NUMBUF], *end;
-	int oom_adjust;
+	char buffer[PROC_NUMBUF];
+	long oom_adjust;
+	unsigned long flags;
+	int err;
 
 	memset(buffer, 0, sizeof(buffer));
 	if (count > sizeof(buffer) - 1)
 		count = sizeof(buffer) - 1;
 	if (copy_from_user(buffer, buf, count))
 		return -EFAULT;
-	oom_adjust = simple_strtol(buffer, &end, 0);
+
+	err = strict_strtol(strstrip(buffer), 0, &oom_adjust);
+	if (err)
+		return -EINVAL;
 	if ((oom_adjust < OOM_ADJUST_MIN || oom_adjust > OOM_ADJUST_MAX) &&
 	     oom_adjust != OOM_DISABLE)
 		return -EINVAL;
-	if (*end == '\n')
-		end++;
+
 	task = get_proc_task(file->f_path.dentry->d_inode);
 	if (!task)
 		return -ESRCH;
-	if (oom_adjust < task->oomkilladj && !capable(CAP_SYS_RESOURCE)) {
+	if (!lock_task_sighand(task, &flags)) {
+		put_task_struct(task);
+		return -ESRCH;
+	}
+
+	if (oom_adjust < task->signal->oom_adj && !capable(CAP_SYS_RESOURCE)) {
+		unlock_task_sighand(task, &flags);
 		put_task_struct(task);
 		return -EACCES;
 	}
-	task->oomkilladj = oom_adjust;
+
+	task->signal->oom_adj = oom_adjust;
+
+	unlock_task_sighand(task, &flags);
 	put_task_struct(task);
-	if (end - buffer == 0)
-		return -EIO;
-	return end - buffer;
+
+	return count;
 }
 
 static const struct file_operations proc_oom_adjust_operations = {
@@ -1169,17 +1187,16 @@ static ssize_t proc_fault_inject_write(struct file * file,
 		count = sizeof(buffer) - 1;
 	if (copy_from_user(buffer, buf, count))
 		return -EFAULT;
-	make_it_fail = simple_strtol(buffer, &end, 0);
-	if (*end == '\n')
-		end++;
+	make_it_fail = simple_strtol(strstrip(buffer), &end, 0);
+	if (*end)
+		return -EINVAL;
 	task = get_proc_task(file->f_dentry->d_inode);
 	if (!task)
 		return -ESRCH;
 	task->make_it_fail = make_it_fail;
 	put_task_struct(task);
-	if (end - buffer == 0)
-		return -EIO;
-	return end - buffer;
+
+	return count;
 }
 
 static const struct file_operations proc_fault_inject_operations = {
@@ -2586,9 +2603,6 @@ static void proc_flush_task_mnt(struct vfsmount *mnt, pid_t pid, pid_t tgid)
 		dput(dentry);
 	}
 
-	if (tgid == 0)
-		goto out;
-
 	name.name = buf;
 	name.len = snprintf(buf, sizeof(buf), "%d", tgid);
 	leader = d_hash_and_lookup(mnt->mnt_root, &name);
@@ -2645,17 +2659,16 @@ out:
 void proc_flush_task(struct task_struct *task)
 {
 	int i;
-	struct pid *pid, *tgid = NULL;
+	struct pid *pid, *tgid;
 	struct upid *upid;
 
 	pid = task_pid(task);
-	if (thread_group_leader(task))
-		tgid = task_tgid(task);
+	tgid = task_tgid(task);
 
 	for (i = 0; i <= pid->level; i++) {
 		upid = &pid->numbers[i];
 		proc_flush_task_mnt(upid->ns->proc_mnt, upid->nr,
-			tgid ? tgid->numbers[i].nr : 0);
+					tgid->numbers[i].nr);
 	}
 
 	upid = &pid->numbers[pid->level];
diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c
index 59b43a06887..56013371f9f 100644
--- a/fs/proc/kcore.c
+++ b/fs/proc/kcore.c
@@ -17,9 +17,15 @@
 #include <linux/elfcore.h>
 #include <linux/vmalloc.h>
 #include <linux/highmem.h>
+#include <linux/bootmem.h>
 #include <linux/init.h>
 #include <asm/uaccess.h>
 #include <asm/io.h>
+#include <linux/list.h>
+#include <linux/ioport.h>
+#include <linux/mm.h>
+#include <linux/memory.h>
+#include <asm/sections.h>
 
 #define CORE_STR "CORE"
 
@@ -29,17 +35,6 @@
 
 static struct proc_dir_entry *proc_root_kcore;
 
-static int open_kcore(struct inode * inode, struct file * filp)
-{
-	return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
-}
-
-static ssize_t read_kcore(struct file *, char __user *, size_t, loff_t *);
-
-static const struct file_operations proc_kcore_operations = {
-	.read		= read_kcore,
-	.open		= open_kcore,
-};
 
 #ifndef kc_vaddr_to_offset
 #define	kc_vaddr_to_offset(v) ((v) - PAGE_OFFSET)
@@ -57,18 +52,19 @@ struct memelfnote
 	void *data;
 };
 
-static struct kcore_list *kclist;
+static LIST_HEAD(kclist_head);
 static DEFINE_RWLOCK(kclist_lock);
+static int kcore_need_update = 1;
 
 void
-kclist_add(struct kcore_list *new, void *addr, size_t size)
+kclist_add(struct kcore_list *new, void *addr, size_t size, int type)
 {
 	new->addr = (unsigned long)addr;
 	new->size = size;
+	new->type = type;
 
 	write_lock(&kclist_lock);
-	new->next = kclist;
-	kclist = new;
+	list_add_tail(&new->list, &kclist_head);
 	write_unlock(&kclist_lock);
 }
 
@@ -80,7 +76,7 @@ static size_t get_kcore_size(int *nphdr, size_t *elf_buflen)
 	*nphdr = 1; /* PT_NOTE */
 	size = 0;
 
-	for (m=kclist; m; m=m->next) {
+	list_for_each_entry(m, &kclist_head, list) {
 		try = kc_vaddr_to_offset((size_t)m->addr + m->size);
 		if (try > size)
 			size = try;
@@ -97,6 +93,177 @@ static size_t get_kcore_size(int *nphdr, size_t *elf_buflen)
 	return size + *elf_buflen;
 }
 
+static void free_kclist_ents(struct list_head *head)
+{
+	struct kcore_list *tmp, *pos;
+
+	list_for_each_entry_safe(pos, tmp, head, list) {
+		list_del(&pos->list);
+		kfree(pos);
+	}
+}
+/*
+ * Replace all KCORE_RAM/KCORE_VMEMMAP information with passed list.
+ */
+static void __kcore_update_ram(struct list_head *list)
+{
+	int nphdr;
+	size_t size;
+	struct kcore_list *tmp, *pos;
+	LIST_HEAD(garbage);
+
+	write_lock(&kclist_lock);
+	if (kcore_need_update) {
+		list_for_each_entry_safe(pos, tmp, &kclist_head, list) {
+			if (pos->type == KCORE_RAM
+				|| pos->type == KCORE_VMEMMAP)
+				list_move(&pos->list, &garbage);
+		}
+		list_splice_tail(list, &kclist_head);
+	} else
+		list_splice(list, &garbage);
+	kcore_need_update = 0;
+	proc_root_kcore->size = get_kcore_size(&nphdr, &size);
+	write_unlock(&kclist_lock);
+
+	free_kclist_ents(&garbage);
+}
+
+
+#ifdef CONFIG_HIGHMEM
+/*
+ * If no highmem, we can assume [0...max_low_pfn) continuous range of memory
+ * because memory hole is not as big as !HIGHMEM case.
+ * (HIGHMEM is special because part of memory is _invisible_ from the kernel.)
+ */
+static int kcore_update_ram(void)
+{
+	LIST_HEAD(head);
+	struct kcore_list *ent;
+	int ret = 0;
+
+	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
+	if (!ent)
+		return -ENOMEM;
+	ent->addr = (unsigned long)__va(0);
+	ent->size = max_low_pfn << PAGE_SHIFT;
+	ent->type = KCORE_RAM;
+	list_add(&ent->list, &head);
+	__kcore_update_ram(&head);
+	return ret;
+}
+
+#else /* !CONFIG_HIGHMEM */
+
+#ifdef CONFIG_SPARSEMEM_VMEMMAP
+/* calculate vmemmap's address from given system ram pfn and register it */
+int get_sparsemem_vmemmap_info(struct kcore_list *ent, struct list_head *head)
+{
+	unsigned long pfn = __pa(ent->addr) >> PAGE_SHIFT;
+	unsigned long nr_pages = ent->size >> PAGE_SHIFT;
+	unsigned long start, end;
+	struct kcore_list *vmm, *tmp;
+
+
+	start = ((unsigned long)pfn_to_page(pfn)) & PAGE_MASK;
+	end = ((unsigned long)pfn_to_page(pfn + nr_pages)) - 1;
+	end = ALIGN(end, PAGE_SIZE);
+	/* overlap check (because we have to align page */
+	list_for_each_entry(tmp, head, list) {
+		if (tmp->type != KCORE_VMEMMAP)
+			continue;
+		if (start < tmp->addr + tmp->size)
+			if (end > tmp->addr)
+				end = tmp->addr;
+	}
+	if (start < end) {
+		vmm = kmalloc(sizeof(*vmm), GFP_KERNEL);
+		if (!vmm)
+			return 0;
+		vmm->addr = start;
+		vmm->size = end - start;
+		vmm->type = KCORE_VMEMMAP;
+		list_add_tail(&vmm->list, head);
+	}
+	return 1;
+
+}
+#else
+int get_sparsemem_vmemmap_info(struct kcore_list *ent, struct list_head *head)
+{
+	return 1;
+}
+
+#endif
+
+static int
+kclist_add_private(unsigned long pfn, unsigned long nr_pages, void *arg)
+{
+	struct list_head *head = (struct list_head *)arg;
+	struct kcore_list *ent;
+
+	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
+	if (!ent)
+		return -ENOMEM;
+	ent->addr = (unsigned long)__va((pfn << PAGE_SHIFT));
+	ent->size = nr_pages << PAGE_SHIFT;
+
+	/* Sanity check: Can happen in 32bit arch...maybe */
+	if (ent->addr < (unsigned long) __va(0))
+		goto free_out;
+
+	/* cut not-mapped area. ....from ppc-32 code. */
+	if (ULONG_MAX - ent->addr < ent->size)
+		ent->size = ULONG_MAX - ent->addr;
+
+	/* cut when vmalloc() area is higher than direct-map area */
+	if (VMALLOC_START > (unsigned long)__va(0)) {
+		if (ent->addr > VMALLOC_START)
+			goto free_out;
+		if (VMALLOC_START - ent->addr < ent->size)
+			ent->size = VMALLOC_START - ent->addr;
+	}
+
+	ent->type = KCORE_RAM;
+	list_add_tail(&ent->list, head);
+
+	if (!get_sparsemem_vmemmap_info(ent, head)) {
+		list_del(&ent->list);
+		goto free_out;
+	}
+
+	return 0;
+free_out:
+	kfree(ent);
+	return 1;
+}
+
+static int kcore_update_ram(void)
+{
+	int nid, ret;
+	unsigned long end_pfn;
+	LIST_HEAD(head);
+
+	/* Not inialized....update now */
+	/* find out "max pfn" */
+	end_pfn = 0;
+	for_each_node_state(nid, N_HIGH_MEMORY) {
+		unsigned long node_end;
+		node_end  = NODE_DATA(nid)->node_start_pfn +
+			NODE_DATA(nid)->node_spanned_pages;
+		if (end_pfn < node_end)
+			end_pfn = node_end;
+	}
+	/* scan 0 to max_pfn */
+	ret = walk_system_ram_range(0, end_pfn, &head, kclist_add_private);
+	if (ret) {
+		free_kclist_ents(&head);
+		return -ENOMEM;
+	}
+	__kcore_update_ram(&head);
+	return ret;
+}
+#endif /* CONFIG_HIGHMEM */
 
 /*****************************************************************************/
 /*
@@ -192,7 +359,7 @@ static void elf_kcore_store_hdr(char *bufp, int nphdr, int dataoff)
 	nhdr->p_align	= 0;
 
 	/* setup ELF PT_LOAD program header for every area */
-	for (m=kclist; m; m=m->next) {
+	list_for_each_entry(m, &kclist_head, list) {
 		phdr = (struct elf_phdr *) bufp;
 		bufp += sizeof(struct elf_phdr);
 		offset += sizeof(struct elf_phdr);
@@ -265,7 +432,8 @@ read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
 	unsigned long start;
 
 	read_lock(&kclist_lock);
-	proc_root_kcore->size = size = get_kcore_size(&nphdr, &elf_buflen);
+	size = get_kcore_size(&nphdr, &elf_buflen);
+
 	if (buflen == 0 || *fpos >= size) {
 		read_unlock(&kclist_lock);
 		return 0;
@@ -317,7 +485,7 @@ read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
 		struct kcore_list *m;
 
 		read_lock(&kclist_lock);
-		for (m=kclist; m; m=m->next) {
+		list_for_each_entry(m, &kclist_head, list) {
 			if (start >= m->addr && start < (m->addr+m->size))
 				break;
 		}
@@ -326,45 +494,14 @@ read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
 		if (m == NULL) {
 			if (clear_user(buffer, tsz))
 				return -EFAULT;
-		} else if (is_vmalloc_addr((void *)start)) {
+		} else if (is_vmalloc_or_module_addr((void *)start)) {
 			char * elf_buf;
-			struct vm_struct *m;
-			unsigned long curstart = start;
-			unsigned long cursize = tsz;
 
 			elf_buf = kzalloc(tsz, GFP_KERNEL);
 			if (!elf_buf)
 				return -ENOMEM;
-
-			read_lock(&vmlist_lock);
-			for (m=vmlist; m && cursize; m=m->next) {
-				unsigned long vmstart;
-				unsigned long vmsize;
-				unsigned long msize = m->size - PAGE_SIZE;
-
-				if (((unsigned long)m->addr + msize) < 
-								curstart)
-					continue;
-				if ((unsigned long)m->addr > (curstart + 
-								cursize))
-					break;
-				vmstart = (curstart < (unsigned long)m->addr ? 
-					(unsigned long)m->addr : curstart);
-				if (((unsigned long)m->addr + msize) > 
-							(curstart + cursize))
-					vmsize = curstart + cursize - vmstart;
-				else
-					vmsize = (unsigned long)m->addr + 
-							msize - vmstart;
-				curstart = vmstart + vmsize;
-				cursize -= vmsize;
-				/* don't dump ioremap'd stuff! (TA) */
-				if (m->flags & VM_IOREMAP)
-					continue;
-				memcpy(elf_buf + (vmstart - start),
-					(char *)vmstart, vmsize);
-			}
-			read_unlock(&vmlist_lock);
+			vread(elf_buf, (char *)start, tsz);
+			/* we have to zero-fill user buffer even if no read */
 			if (copy_to_user(buffer, elf_buf, tsz)) {
 				kfree(elf_buf);
 				return -EFAULT;
@@ -402,12 +539,96 @@ read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
 	return acc;
 }
 
+
+static int open_kcore(struct inode *inode, struct file *filp)
+{
+	if (!capable(CAP_SYS_RAWIO))
+		return -EPERM;
+	if (kcore_need_update)
+		kcore_update_ram();
+	if (i_size_read(inode) != proc_root_kcore->size) {
+		mutex_lock(&inode->i_mutex);
+		i_size_write(inode, proc_root_kcore->size);
+		mutex_unlock(&inode->i_mutex);
+	}
+	return 0;
+}
+
+
+static const struct file_operations proc_kcore_operations = {
+	.read		= read_kcore,
+	.open		= open_kcore,
+};
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+/* just remember that we have to update kcore */
+static int __meminit kcore_callback(struct notifier_block *self,
+				    unsigned long action, void *arg)
+{
+	switch (action) {
+	case MEM_ONLINE:
+	case MEM_OFFLINE:
+		write_lock(&kclist_lock);
+		kcore_need_update = 1;
+		write_unlock(&kclist_lock);
+	}
+	return NOTIFY_OK;
+}
+#endif
+
+
+static struct kcore_list kcore_vmalloc;
+
+#ifdef CONFIG_ARCH_PROC_KCORE_TEXT
+static struct kcore_list kcore_text;
+/*
+ * If defined, special segment is used for mapping kernel text instead of
+ * direct-map area. We need to create special TEXT section.
+ */
+static void __init proc_kcore_text_init(void)
+{
+	kclist_add(&kcore_text, _stext, _end - _stext, KCORE_TEXT);
+}
+#else
+static void __init proc_kcore_text_init(void)
+{
+}
+#endif
+
+#if defined(CONFIG_MODULES) && defined(MODULES_VADDR)
+/*
+ * MODULES_VADDR has no intersection with VMALLOC_ADDR.
+ */
+struct kcore_list kcore_modules;
+static void __init add_modules_range(void)
+{
+	kclist_add(&kcore_modules, (void *)MODULES_VADDR,
+			MODULES_END - MODULES_VADDR, KCORE_VMALLOC);
+}
+#else
+static void __init add_modules_range(void)
+{
+}
+#endif
+
 static int __init proc_kcore_init(void)
 {
-	proc_root_kcore = proc_create("kcore", S_IRUSR, NULL, &proc_kcore_operations);
-	if (proc_root_kcore)
-		proc_root_kcore->size =
-				(size_t)high_memory - PAGE_OFFSET + PAGE_SIZE;
+	proc_root_kcore = proc_create("kcore", S_IRUSR, NULL,
+				      &proc_kcore_operations);
+	if (!proc_root_kcore) {
+		printk(KERN_ERR "couldn't create /proc/kcore\n");
+		return 0; /* Always returns 0. */
+	}
+	/* Store text area if it's special */
+	proc_kcore_text_init();
+	/* Store vmalloc area */
+	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START,
+		VMALLOC_END - VMALLOC_START, KCORE_VMALLOC);
+	add_modules_range();
+	/* Store direct-map area from physical memory map */
+	kcore_update_ram();
+	hotplug_memory_notifier(kcore_callback, 0);
+
 	return 0;
 }
 module_init(proc_kcore_init);
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index d5c410d47fa..171e052c07b 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -81,9 +81,11 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		"Writeback:      %8lu kB\n"
 		"AnonPages:      %8lu kB\n"
 		"Mapped:         %8lu kB\n"
+		"Shmem:          %8lu kB\n"
 		"Slab:           %8lu kB\n"
 		"SReclaimable:   %8lu kB\n"
 		"SUnreclaim:     %8lu kB\n"
+		"KernelStack:    %8lu kB\n"
 		"PageTables:     %8lu kB\n"
 #ifdef CONFIG_QUICKLIST
 		"Quicklists:     %8lu kB\n"
@@ -124,10 +126,12 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
 		K(global_page_state(NR_WRITEBACK)),
 		K(global_page_state(NR_ANON_PAGES)),
 		K(global_page_state(NR_FILE_MAPPED)),
+		K(global_page_state(NR_SHMEM)),
 		K(global_page_state(NR_SLAB_RECLAIMABLE) +
 				global_page_state(NR_SLAB_UNRECLAIMABLE)),
 		K(global_page_state(NR_SLAB_RECLAIMABLE)),
 		K(global_page_state(NR_SLAB_UNRECLAIMABLE)),
+		global_page_state(NR_KERNEL_STACK) * THREAD_SIZE / 1024,
 		K(global_page_state(NR_PAGETABLE)),
 #ifdef CONFIG_QUICKLIST
 		K(quicklist_total_size()),
diff --git a/fs/proc/nommu.c b/fs/proc/nommu.c
index 7e14d1a0400..9fe7d7ebe11 100644
--- a/fs/proc/nommu.c
+++ b/fs/proc/nommu.c
@@ -109,7 +109,7 @@ static void *nommu_region_list_next(struct seq_file *m, void *v, loff_t *pos)
 	return rb_next((struct rb_node *) v);
 }
 
-static struct seq_operations proc_nommu_region_list_seqop = {
+static const struct seq_operations proc_nommu_region_list_seqop = {
 	.start	= nommu_region_list_start,
 	.next	= nommu_region_list_next,
 	.stop	= nommu_region_list_stop,
diff --git a/fs/proc/page.c b/fs/proc/page.c
index 2707c6c7a20..2281c2cbfe2 100644
--- a/fs/proc/page.c
+++ b/fs/proc/page.c
@@ -2,6 +2,7 @@
 #include <linux/compiler.h>
 #include <linux/fs.h>
 #include <linux/init.h>
+#include <linux/ksm.h>
 #include <linux/mm.h>
 #include <linux/mmzone.h>
 #include <linux/proc_fs.h>
@@ -95,6 +96,8 @@ static const struct file_operations proc_kpagecount_operations = {
 #define KPF_UNEVICTABLE		18
 #define KPF_NOPAGE		20
 
+#define KPF_KSM			21
+
 /* kernel hacking assistances
  * WARNING: subject to change, never rely on them!
  */
@@ -137,6 +140,8 @@ static u64 get_uflags(struct page *page)
 		u |= 1 << KPF_MMAP;
 	if (PageAnon(page))
 		u |= 1 << KPF_ANON;
+	if (PageKsm(page))
+		u |= 1 << KPF_KSM;
 
 	/*
 	 * compound pages: export both head/tail info
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 9bd8be1d235..2a1bef9203c 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -243,6 +243,25 @@ static void show_map_vma(struct seq_file *m, struct vm_area_struct *vma)
 				} else if (vma->vm_start <= mm->start_stack &&
 					   vma->vm_end >= mm->start_stack) {
 					name = "[stack]";
+				} else {
+					unsigned long stack_start;
+					struct proc_maps_private *pmp;
+
+					pmp = m->private;
+					stack_start = pmp->task->stack_start;
+
+					if (vma->vm_start <= stack_start &&
+					    vma->vm_end >= stack_start) {
+						pad_len_spaces(m, len);
+						seq_printf(m,
+						 "[threadstack:%08lx]",
+#ifdef CONFIG_STACK_GROWSUP
+						 vma->vm_end - stack_start
+#else
+						 stack_start - vma->vm_start
+#endif
+						);
+					}
 				}
 			} else {
 				name = "[vdso]";
@@ -465,23 +484,28 @@ static int clear_refs_pte_range(pmd_t *pmd, unsigned long addr,
 	return 0;
 }
 
+#define CLEAR_REFS_ALL 1
+#define CLEAR_REFS_ANON 2
+#define CLEAR_REFS_MAPPED 3
+
 static ssize_t clear_refs_write(struct file *file, const char __user *buf,
 				size_t count, loff_t *ppos)
 {
 	struct task_struct *task;
-	char buffer[PROC_NUMBUF], *end;
+	char buffer[PROC_NUMBUF];
 	struct mm_struct *mm;
 	struct vm_area_struct *vma;
+	long type;
 
 	memset(buffer, 0, sizeof(buffer));
 	if (count > sizeof(buffer) - 1)
 		count = sizeof(buffer) - 1;
 	if (copy_from_user(buffer, buf, count))
 		return -EFAULT;
-	if (!simple_strtol(buffer, &end, 0))
+	if (strict_strtol(strstrip(buffer), 10, &type))
+		return -EINVAL;
+	if (type < CLEAR_REFS_ALL || type > CLEAR_REFS_MAPPED)
 		return -EINVAL;
-	if (*end == '\n')
-		end++;
 	task = get_proc_task(file->f_path.dentry->d_inode);
 	if (!task)
 		return -ESRCH;
@@ -494,18 +518,31 @@ static ssize_t clear_refs_write(struct file *file, const char __user *buf,
 		down_read(&mm->mmap_sem);
 		for (vma = mm->mmap; vma; vma = vma->vm_next) {
 			clear_refs_walk.private = vma;
-			if (!is_vm_hugetlb_page(vma))
-				walk_page_range(vma->vm_start, vma->vm_end,
-						&clear_refs_walk);
+			if (is_vm_hugetlb_page(vma))
+				continue;
+			/*
+			 * Writing 1 to /proc/pid/clear_refs affects all pages.
+			 *
+			 * Writing 2 to /proc/pid/clear_refs only affects
+			 * Anonymous pages.
+			 *
+			 * Writing 3 to /proc/pid/clear_refs only affects file
+			 * mapped pages.
+			 */
+			if (type == CLEAR_REFS_ANON && vma->vm_file)
+				continue;
+			if (type == CLEAR_REFS_MAPPED && !vma->vm_file)
+				continue;
+			walk_page_range(vma->vm_start, vma->vm_end,
+					&clear_refs_walk);
 		}
 		flush_tlb_mm(mm);
 		up_read(&mm->mmap_sem);
 		mmput(mm);
 	}
 	put_task_struct(task);
-	if (end - buffer == 0)
-		return -EIO;
-	return end - buffer;
+
+	return count;
 }
 
 const struct file_operations proc_clear_refs_operations = {
diff --git a/fs/qnx4/Kconfig b/fs/qnx4/Kconfig
index be8e0e1445b..5f608999404 100644
--- a/fs/qnx4/Kconfig
+++ b/fs/qnx4/Kconfig
@@ -6,20 +6,9 @@ config QNX4FS_FS
 	  QNX 4 and QNX 6 (the latter is also called QNX RTP).
 	  Further information is available at <http://www.qnx.com/>.
 	  Say Y if you intend to mount QNX hard disks or floppies.
-	  Unless you say Y to "QNX4FS read-write support" below, you will
-	  only be able to read these file systems.
 
 	  To compile this file system support as a module, choose M here: the
 	  module will be called qnx4.
 
 	  If you don't know whether you need it, then you don't need it:
 	  answer N.
-
-config QNX4FS_RW
-	bool "QNX4FS write support (DANGEROUS)"
-	depends on QNX4FS_FS && EXPERIMENTAL && BROKEN
-	help
-	  Say Y if you want to test write support for QNX4 file systems.
-
-	  It's currently broken, so for now:
-	  answer N.
diff --git a/fs/qnx4/Makefile b/fs/qnx4/Makefile
index e4d408cc547..4a283b3f87f 100644
--- a/fs/qnx4/Makefile
+++ b/fs/qnx4/Makefile
@@ -4,4 +4,4 @@
 
 obj-$(CONFIG_QNX4FS_FS) += qnx4.o
 
-qnx4-objs := inode.o dir.o namei.o file.o bitmap.o truncate.o
+qnx4-objs := inode.o dir.o namei.o bitmap.o
diff --git a/fs/qnx4/bitmap.c b/fs/qnx4/bitmap.c
index e1cd061a25f..0afba069d56 100644
--- a/fs/qnx4/bitmap.c
+++ b/fs/qnx4/bitmap.c
@@ -78,84 +78,3 @@ unsigned long qnx4_count_free_blocks(struct super_block *sb)
 
 	return total_free;
 }
-
-#ifdef CONFIG_QNX4FS_RW
-
-int qnx4_is_free(struct super_block *sb, long block)
-{
-	int start = le32_to_cpu(qnx4_sb(sb)->BitMap->di_first_xtnt.xtnt_blk) - 1;
-	int size = le32_to_cpu(qnx4_sb(sb)->BitMap->di_size);
-	struct buffer_head *bh;
-	const char *g;
-	int ret = -EIO;
-
-	start += block / (QNX4_BLOCK_SIZE * 8);
-	QNX4DEBUG(("qnx4: is_free requesting block [%lu], bitmap in block [%lu]\n",
-		   (unsigned long) block, (unsigned long) start));
-	(void) size;		/* CHECKME */
-	bh = sb_bread(sb, start);
-	if (bh == NULL) {
-		return -EIO;
-	}
-	g = bh->b_data + (block % QNX4_BLOCK_SIZE);
-	if (((*g) & (1 << (block % 8))) == 0) {
-		QNX4DEBUG(("qnx4: is_free -> block is free\n"));
-		ret = 1;
-	} else {
-		QNX4DEBUG(("qnx4: is_free -> block is busy\n"));
-		ret = 0;
-	}
-	brelse(bh);
-
-	return ret;
-}
-
-int qnx4_set_bitmap(struct super_block *sb, long block, int busy)
-{
-	int start = le32_to_cpu(qnx4_sb(sb)->BitMap->di_first_xtnt.xtnt_blk) - 1;
-	int size = le32_to_cpu(qnx4_sb(sb)->BitMap->di_size);
-	struct buffer_head *bh;
-	char *g;
-
-	start += block / (QNX4_BLOCK_SIZE * 8);
-	QNX4DEBUG(("qnx4: set_bitmap requesting block [%lu], bitmap in block [%lu]\n",
-		   (unsigned long) block, (unsigned long) start));
-	(void) size;		/* CHECKME */
-	bh = sb_bread(sb, start);
-	if (bh == NULL) {
-		return -EIO;
-	}
-	g = bh->b_data + (block % QNX4_BLOCK_SIZE);
-	if (busy == 0) {
-		(*g) &= ~(1 << (block % 8));
-	} else {
-		(*g) |= (1 << (block % 8));
-	}
-	mark_buffer_dirty(bh);
-	brelse(bh);
-
-	return 0;
-}
-
-static void qnx4_clear_inode(struct inode *inode)
-{
-	struct qnx4_inode_entry *qnx4_ino = qnx4_raw_inode(inode);
-	/* What for? */
-	memset(qnx4_ino->di_fname, 0, sizeof qnx4_ino->di_fname);
-	qnx4_ino->di_size = 0;
-	qnx4_ino->di_num_xtnts = 0;
-	qnx4_ino->di_mode = 0;
-	qnx4_ino->di_status = 0;
-}
-
-void qnx4_free_inode(struct inode *inode)
-{
-	if (inode->i_ino < 1) {
-		printk("free_inode: inode 0 or nonexistent inode\n");
-		return;
-	}
-	qnx4_clear_inode(inode);
-	clear_inode(inode);
-}
-
-#endif
diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c
index 003c68f3238..86cc39cb139 100644
--- a/fs/qnx4/dir.c
+++ b/fs/qnx4/dir.c
@@ -85,9 +85,4 @@ const struct file_operations qnx4_dir_operations =
 const struct inode_operations qnx4_dir_inode_operations =
 {
 	.lookup		= qnx4_lookup,
-#ifdef CONFIG_QNX4FS_RW
-	.create		= qnx4_create,
-	.unlink		= qnx4_unlink,
-	.rmdir		= qnx4_rmdir,
-#endif
 };
diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c
deleted file mode 100644
index 09b170ac936..00000000000
--- a/fs/qnx4/file.c
+++ /dev/null
@@ -1,40 +0,0 @@
-/*
- * QNX4 file system, Linux implementation.
- *
- * Version : 0.2.1
- *
- * Using parts of the xiafs filesystem.
- *
- * History :
- *
- * 25-05-1998 by Richard Frowijn : first release.
- * 21-06-1998 by Frank Denis : wrote qnx4_readpage to use generic_file_read.
- * 27-06-1998 by Frank Denis : file overwriting.
- */
-
-#include "qnx4.h"
-
-/*
- * We have mostly NULL's here: the current defaults are ok for
- * the qnx4 filesystem.
- */
-const struct file_operations qnx4_file_operations =
-{
-	.llseek		= generic_file_llseek,
-	.read		= do_sync_read,
-	.aio_read	= generic_file_aio_read,
-	.mmap		= generic_file_mmap,
-	.splice_read	= generic_file_splice_read,
-#ifdef CONFIG_QNX4FS_RW
-	.write		= do_sync_write,
-	.aio_write	= generic_file_aio_write,
-	.fsync		= simple_fsync,
-#endif
-};
-
-const struct inode_operations qnx4_file_inode_operations =
-{
-#ifdef CONFIG_QNX4FS_RW
-	.truncate	= qnx4_truncate,
-#endif
-};
diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c
index 681df5fcd16..d2cd1798d8c 100644
--- a/fs/qnx4/inode.c
+++ b/fs/qnx4/inode.c
@@ -28,73 +28,6 @@
 
 static const struct super_operations qnx4_sops;
 
-#ifdef CONFIG_QNX4FS_RW
-
-static void qnx4_delete_inode(struct inode *inode)
-{
-	QNX4DEBUG(("qnx4: deleting inode [%lu]\n", (unsigned long) inode->i_ino));
-	truncate_inode_pages(&inode->i_data, 0);
-	inode->i_size = 0;
-	qnx4_truncate(inode);
-	lock_kernel();
-	qnx4_free_inode(inode);
-	unlock_kernel();
-}
-
-static int qnx4_write_inode(struct inode *inode, int do_sync)
-{
-	struct qnx4_inode_entry *raw_inode;
-	int block, ino;
-	struct buffer_head *bh;
-	ino = inode->i_ino;
-
-	QNX4DEBUG(("qnx4: write inode 1.\n"));
-	if (inode->i_nlink == 0) {
-		return 0;
-	}
-	if (!ino) {
-		printk("qnx4: bad inode number on dev %s: %d is out of range\n",
-		       inode->i_sb->s_id, ino);
-		return -EIO;
-	}
-	QNX4DEBUG(("qnx4: write inode 2.\n"));
-	block = ino / QNX4_INODES_PER_BLOCK;
-	lock_kernel();
-	if (!(bh = sb_bread(inode->i_sb, block))) {
-		printk("qnx4: major problem: unable to read inode from dev "
-		       "%s\n", inode->i_sb->s_id);
-		unlock_kernel();
-		return -EIO;
-	}
-	raw_inode = ((struct qnx4_inode_entry *) bh->b_data) +
-	    (ino % QNX4_INODES_PER_BLOCK);
-	raw_inode->di_mode  = cpu_to_le16(inode->i_mode);
-	raw_inode->di_uid   = cpu_to_le16(fs_high2lowuid(inode->i_uid));
-	raw_inode->di_gid   = cpu_to_le16(fs_high2lowgid(inode->i_gid));
-	raw_inode->di_nlink = cpu_to_le16(inode->i_nlink);
-	raw_inode->di_size  = cpu_to_le32(inode->i_size);
-	raw_inode->di_mtime = cpu_to_le32(inode->i_mtime.tv_sec);
-	raw_inode->di_atime = cpu_to_le32(inode->i_atime.tv_sec);
-	raw_inode->di_ctime = cpu_to_le32(inode->i_ctime.tv_sec);
-	raw_inode->di_first_xtnt.xtnt_size = cpu_to_le32(inode->i_blocks);
-	mark_buffer_dirty(bh);
-	if (do_sync) {
-		sync_dirty_buffer(bh);
-		if (buffer_req(bh) && !buffer_uptodate(bh)) {
-			printk("qnx4: IO error syncing inode [%s:%08x]\n",
-					inode->i_sb->s_id, ino);
-			brelse(bh);
-			unlock_kernel();
-			return -EIO;
-		}
-	}
-	brelse(bh);
-	unlock_kernel();
-	return 0;
-}
-
-#endif
-
 static void qnx4_put_super(struct super_block *sb);
 static struct inode *qnx4_alloc_inode(struct super_block *sb);
 static void qnx4_destroy_inode(struct inode *inode);
@@ -108,10 +41,6 @@ static const struct super_operations qnx4_sops =
 	.put_super	= qnx4_put_super,
 	.statfs		= qnx4_statfs,
 	.remount_fs	= qnx4_remount,
-#ifdef CONFIG_QNX4FS_RW
-	.write_inode	= qnx4_write_inode,
-	.delete_inode	= qnx4_delete_inode,
-#endif
 };
 
 static int qnx4_remount(struct super_block *sb, int *flags, char *data)
@@ -120,15 +49,7 @@ static int qnx4_remount(struct super_block *sb, int *flags, char *data)
 
 	qs = qnx4_sb(sb);
 	qs->Version = QNX4_VERSION;
-#ifndef CONFIG_QNX4FS_RW
 	*flags |= MS_RDONLY;
-#endif
-	if (*flags & MS_RDONLY) {
-		return 0;
-	}
-
-	mark_buffer_dirty(qs->sb_buf);
-
 	return 0;
 }
 
@@ -354,9 +275,7 @@ static int qnx4_fill_super(struct super_block *s, void *data, int silent)
 	}
 	s->s_op = &qnx4_sops;
 	s->s_magic = QNX4_SUPER_MAGIC;
-#ifndef CONFIG_QNX4FS_RW
 	s->s_flags |= MS_RDONLY;	/* Yup, read-only yet */
-#endif
 	qnx4_sb(s)->sb_buf = bh;
 	qnx4_sb(s)->sb = (struct qnx4_super_block *) bh->b_data;
 
@@ -489,8 +408,7 @@ struct inode *qnx4_iget(struct super_block *sb, unsigned long ino)
 
 	memcpy(qnx4_inode, raw_inode, QNX4_DIR_ENTRY_SIZE);
 	if (S_ISREG(inode->i_mode)) {
-		inode->i_op = &qnx4_file_inode_operations;
-		inode->i_fop = &qnx4_file_operations;
+		inode->i_fop = &generic_ro_fops;
 		inode->i_mapping->a_ops = &qnx4_aops;
 		qnx4_i(inode)->mmu_private = inode->i_size;
 	} else if (S_ISDIR(inode->i_mode)) {
diff --git a/fs/qnx4/namei.c b/fs/qnx4/namei.c
index 5972ed21493..ae1e7edbacd 100644
--- a/fs/qnx4/namei.c
+++ b/fs/qnx4/namei.c
@@ -134,108 +134,3 @@ out:
 
 	return NULL;
 }
-
-#ifdef CONFIG_QNX4FS_RW
-int qnx4_create(struct inode *dir, struct dentry *dentry, int mode,
-		struct nameidata *nd)
-{
-	QNX4DEBUG(("qnx4: qnx4_create\n"));
-	if (dir == NULL) {
-		return -ENOENT;
-	}
-	return -ENOSPC;
-}
-
-int qnx4_rmdir(struct inode *dir, struct dentry *dentry)
-{
-	struct buffer_head *bh;
-	struct qnx4_inode_entry *de;
-	struct inode *inode;
-	int retval;
-	int ino;
-
-	QNX4DEBUG(("qnx4: qnx4_rmdir [%s]\n", dentry->d_name.name));
-	lock_kernel();
-	bh = qnx4_find_entry(dentry->d_name.len, dir, dentry->d_name.name,
-			     &de, &ino);
-	if (bh == NULL) {
-		unlock_kernel();
-		return -ENOENT;
-	}
-	inode = dentry->d_inode;
-	if (inode->i_ino != ino) {
-		retval = -EIO;
-		goto end_rmdir;
-	}
-#if 0
-	if (!empty_dir(inode)) {
-		retval = -ENOTEMPTY;
-		goto end_rmdir;
-	}
-#endif
-	if (inode->i_nlink != 2) {
-		QNX4DEBUG(("empty directory has nlink!=2 (%d)\n", inode->i_nlink));
-	}
-	QNX4DEBUG(("qnx4: deleting directory\n"));
-	de->di_status = 0;
-	memset(de->di_fname, 0, sizeof de->di_fname);
-	de->di_mode = 0;
-	mark_buffer_dirty_inode(bh, dir);
-	clear_nlink(inode);
-	mark_inode_dirty(inode);
-	inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC;
-	inode_dec_link_count(dir);
-	retval = 0;
-
-      end_rmdir:
-	brelse(bh);
-
-	unlock_kernel();
-	return retval;
-}
-
-int qnx4_unlink(struct inode *dir, struct dentry *dentry)
-{
-	struct buffer_head *bh;
-	struct qnx4_inode_entry *de;
-	struct inode *inode;
-	int retval;
-	int ino;
-
-	QNX4DEBUG(("qnx4: qnx4_unlink [%s]\n", dentry->d_name.name));
-	lock_kernel();
-	bh = qnx4_find_entry(dentry->d_name.len, dir, dentry->d_name.name,
-			     &de, &ino);
-	if (bh == NULL) {
-		unlock_kernel();
-		return -ENOENT;
-	}
-	inode = dentry->d_inode;
-	if (inode->i_ino != ino) {
-		retval = -EIO;
-		goto end_unlink;
-	}
-	retval = -EPERM;
-	if (!inode->i_nlink) {
-		QNX4DEBUG(("Deleting nonexistent file (%s:%lu), %d\n",
-			   inode->i_sb->s_id,
-			   inode->i_ino, inode->i_nlink));
-		inode->i_nlink = 1;
-	}
-	de->di_status = 0;
-	memset(de->di_fname, 0, sizeof de->di_fname);
-	de->di_mode = 0;
-	mark_buffer_dirty_inode(bh, dir);
-	dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC;
-	mark_inode_dirty(dir);
-	inode->i_ctime = dir->i_ctime;
-	inode_dec_link_count(inode);
-	retval = 0;
-
-end_unlink:
-	unlock_kernel();
-	brelse(bh);
-
-	return retval;
-}
-#endif
diff --git a/fs/qnx4/qnx4.h b/fs/qnx4/qnx4.h
index 9efc089454f..33a60858203 100644
--- a/fs/qnx4/qnx4.h
+++ b/fs/qnx4/qnx4.h
@@ -29,17 +29,9 @@ extern unsigned long qnx4_block_map(struct inode *inode, long iblock);
 
 extern struct buffer_head *qnx4_bread(struct inode *, int, int);
 
-extern const struct inode_operations qnx4_file_inode_operations;
 extern const struct inode_operations qnx4_dir_inode_operations;
-extern const struct file_operations qnx4_file_operations;
 extern const struct file_operations qnx4_dir_operations;
 extern int qnx4_is_free(struct super_block *sb, long block);
-extern int qnx4_set_bitmap(struct super_block *sb, long block, int busy);
-extern int qnx4_create(struct inode *inode, struct dentry *dentry, int mode, struct nameidata *nd);
-extern void qnx4_truncate(struct inode *inode);
-extern void qnx4_free_inode(struct inode *inode);
-extern int qnx4_unlink(struct inode *dir, struct dentry *dentry);
-extern int qnx4_rmdir(struct inode *dir, struct dentry *dentry);
 
 static inline struct qnx4_sb_info *qnx4_sb(struct super_block *sb)
 {
diff --git a/fs/qnx4/truncate.c b/fs/qnx4/truncate.c
deleted file mode 100644
index d94d9ee241f..00000000000
--- a/fs/qnx4/truncate.c
+++ /dev/null
@@ -1,34 +0,0 @@
-/* 
- * QNX4 file system, Linux implementation.
- * 
- * Version : 0.1
- * 
- * Using parts of the xiafs filesystem.
- * 
- * History :
- * 
- * 30-06-1998 by Frank DENIS : ugly filler.
- */
-
-#include <linux/smp_lock.h>
-#include "qnx4.h"
-
-#ifdef CONFIG_QNX4FS_RW
-
-void qnx4_truncate(struct inode *inode)
-{
-	if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
-	      S_ISLNK(inode->i_mode))) {
-		return;
-	}
-	lock_kernel();
-	if (!(S_ISDIR(inode->i_mode))) {
-		/* TODO */
-	}
-	QNX4DEBUG(("qnx4: qnx4_truncate called\n"));
-	inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC;
-	mark_inode_dirty(inode);
-	unlock_kernel();
-}
-
-#endif
diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c
index 38f7bd559f3..39b49c42a7e 100644
--- a/fs/quota/dquot.c
+++ b/fs/quota/dquot.c
@@ -1839,7 +1839,7 @@ EXPORT_SYMBOL(dquot_commit_info);
 /*
  * Definitions of diskquota operations.
  */
-struct dquot_operations dquot_operations = {
+const struct dquot_operations dquot_operations = {
 	.initialize	= dquot_initialize,
 	.drop		= dquot_drop,
 	.alloc_space	= dquot_alloc_space,
@@ -2461,7 +2461,7 @@ out:
 }
 EXPORT_SYMBOL(vfs_set_dqinfo);
 
-struct quotactl_ops vfs_quotactl_ops = {
+const struct quotactl_ops vfs_quotactl_ops = {
 	.quota_on	= vfs_quota_on,
 	.quota_off	= vfs_quota_off,
 	.quota_sync	= vfs_quota_sync,
diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c
index 0ff7566c767..a6090aa1a7c 100644
--- a/fs/ramfs/inode.c
+++ b/fs/ramfs/inode.c
@@ -34,18 +34,17 @@
 #include <linux/ramfs.h>
 #include <linux/sched.h>
 #include <linux/parser.h>
+#include <linux/magic.h>
 #include <asm/uaccess.h>
 #include "internal.h"
 
-/* some random number */
-#define RAMFS_MAGIC	0x858458f6
-
 #define RAMFS_DEFAULT_MODE	0755
 
 static const struct super_operations ramfs_ops;
 static const struct inode_operations ramfs_dir_inode_operations;
 
 static struct backing_dev_info ramfs_backing_dev_info = {
+	.name		= "ramfs",
 	.ra_pages	= 0,	/* No readahead */
 	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK |
 			  BDI_CAP_MAP_DIRECT | BDI_CAP_MAP_COPY |
diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c
index 7adea74d6a8..f0ad05f3802 100644
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -612,7 +612,7 @@ static int reiserfs_mark_dquot_dirty(struct dquot *);
 static int reiserfs_write_info(struct super_block *, int);
 static int reiserfs_quota_on(struct super_block *, int, int, char *, int);
 
-static struct dquot_operations reiserfs_quota_operations = {
+static const struct dquot_operations reiserfs_quota_operations = {
 	.initialize = dquot_initialize,
 	.drop = dquot_drop,
 	.alloc_space = dquot_alloc_space,
@@ -629,7 +629,7 @@ static struct dquot_operations reiserfs_quota_operations = {
 	.destroy_dquot	= dquot_destroy,
 };
 
-static struct quotactl_ops reiserfs_qctl_operations = {
+static const struct quotactl_ops reiserfs_qctl_operations = {
 	.quota_on = reiserfs_quota_on,
 	.quota_off = vfs_quota_off,
 	.quota_sync = vfs_quota_sync,
diff --git a/fs/romfs/super.c b/fs/romfs/super.c
index 4ab3c03d8f9..47f132df0c3 100644
--- a/fs/romfs/super.c
+++ b/fs/romfs/super.c
@@ -284,7 +284,7 @@ static const struct file_operations romfs_dir_operations = {
 	.readdir	= romfs_readdir,
 };
 
-static struct inode_operations romfs_dir_inode_operations = {
+static const struct inode_operations romfs_dir_inode_operations = {
 	.lookup		= romfs_lookup,
 };
 
diff --git a/fs/select.c b/fs/select.c
index 8084834e123..a201fc37022 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -41,22 +41,28 @@
  * better solutions..
  */
 
+#define MAX_SLACK	(100 * NSEC_PER_MSEC)
+
 static long __estimate_accuracy(struct timespec *tv)
 {
 	long slack;
 	int divfactor = 1000;
 
+	if (tv->tv_sec < 0)
+		return 0;
+
 	if (task_nice(current) > 0)
 		divfactor = divfactor / 5;
 
+	if (tv->tv_sec > MAX_SLACK / (NSEC_PER_SEC/divfactor))
+		return MAX_SLACK;
+
 	slack = tv->tv_nsec / divfactor;
 	slack += tv->tv_sec * (NSEC_PER_SEC/divfactor);
 
-	if (slack > 100 * NSEC_PER_MSEC)
-		slack =  100 * NSEC_PER_MSEC;
+	if (slack > MAX_SLACK)
+		return MAX_SLACK;
 
-	if (slack < 0)
-		slack = 0;
 	return slack;
 }
 
diff --git a/fs/smbfs/proc.c b/fs/smbfs/proc.c
index 9468168b9af..71c29b6670b 100644
--- a/fs/smbfs/proc.c
+++ b/fs/smbfs/proc.c
@@ -509,7 +509,7 @@ date_unix2dos(struct smb_sb_info *server,
 		month = 2;
 	} else {
 		nl_day = (year & 3) || day <= 59 ? day : day - 1;
-		for (month = 0; month < 12; month++)
+		for (month = 1; month < 12; month++)
 			if (day_n[month] > nl_day)
 				break;
 	}
diff --git a/fs/splice.c b/fs/splice.c
index 73766d24f97..7394e9e1753 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -502,8 +502,10 @@ ssize_t generic_file_splice_read(struct file *in, loff_t *ppos,
 		len = left;
 
 	ret = __generic_file_splice_read(in, ppos, pipe, len, flags);
-	if (ret > 0)
+	if (ret > 0) {
 		*ppos += ret;
+		file_accessed(in);
+	}
 
 	return ret;
 }
@@ -963,8 +965,10 @@ generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out,
 
 		mutex_lock_nested(&inode->i_mutex, I_MUTEX_CHILD);
 		ret = file_remove_suid(out);
-		if (!ret)
+		if (!ret) {
+			file_update_time(out);
 			ret = splice_from_pipe_feed(pipe, &sd, pipe_to_file);
+		}
 		mutex_unlock(&inode->i_mutex);
 	} while (ret > 0);
 	splice_from_pipe_end(pipe, &sd);
@@ -976,25 +980,15 @@ generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out,
 
 	if (ret > 0) {
 		unsigned long nr_pages;
+		int err;
 
-		*ppos += ret;
 		nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
 
-		/*
-		 * If file or inode is SYNC and we actually wrote some data,
-		 * sync it.
-		 */
-		if (unlikely((out->f_flags & O_SYNC) || IS_SYNC(inode))) {
-			int err;
-
-			mutex_lock(&inode->i_mutex);
-			err = generic_osync_inode(inode, mapping,
-						  OSYNC_METADATA|OSYNC_DATA);
-			mutex_unlock(&inode->i_mutex);
-
-			if (err)
-				ret = err;
-		}
+		err = generic_write_sync(out, *ppos, ret);
+		if (err)
+			ret = err;
+		else
+			*ppos += ret;
 		balance_dirty_pages_ratelimited_nr(mapping, nr_pages);
 	}
 
diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c
index cb5fc57e370..6c197ef53ad 100644
--- a/fs/squashfs/super.c
+++ b/fs/squashfs/super.c
@@ -44,7 +44,7 @@
 #include "squashfs.h"
 
 static struct file_system_type squashfs_fs_type;
-static struct super_operations squashfs_super_ops;
+static const struct super_operations squashfs_super_ops;
 
 static int supported_squashfs_filesystem(short major, short minor, short comp)
 {
@@ -444,7 +444,7 @@ static struct file_system_type squashfs_fs_type = {
 	.fs_flags = FS_REQUIRES_DEV
 };
 
-static struct super_operations squashfs_super_ops = {
+static const struct super_operations squashfs_super_ops = {
 	.alloc_inode = squashfs_alloc_inode,
 	.destroy_inode = squashfs_destroy_inode,
 	.statfs = squashfs_statfs,
diff --git a/fs/super.c b/fs/super.c
index 2761d3e22ed..0e7207b9815 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -54,7 +54,7 @@ DEFINE_SPINLOCK(sb_lock);
 static struct super_block *alloc_super(struct file_system_type *type)
 {
 	struct super_block *s = kzalloc(sizeof(struct super_block),  GFP_USER);
-	static struct super_operations default_op;
+	static const struct super_operations default_op;
 
 	if (s) {
 		if (security_sb_alloc(s)) {
@@ -62,9 +62,6 @@ static struct super_block *alloc_super(struct file_system_type *type)
 			s = NULL;
 			goto out;
 		}
-		INIT_LIST_HEAD(&s->s_dirty);
-		INIT_LIST_HEAD(&s->s_io);
-		INIT_LIST_HEAD(&s->s_more_io);
 		INIT_LIST_HEAD(&s->s_files);
 		INIT_LIST_HEAD(&s->s_instances);
 		INIT_HLIST_HEAD(&s->s_anon);
@@ -171,7 +168,7 @@ int __put_super_and_need_restart(struct super_block *sb)
  *	Drops a temporary reference, frees superblock if there's no
  *	references left.
  */
-static void put_super(struct super_block *sb)
+void put_super(struct super_block *sb)
 {
 	spin_lock(&sb_lock);
 	__put_super(sb);
@@ -710,6 +707,12 @@ static int set_bdev_super(struct super_block *s, void *data)
 {
 	s->s_bdev = data;
 	s->s_dev = s->s_bdev->bd_dev;
+
+	/*
+	 * We set the bdi here to the queue backing, file systems can
+	 * overwrite this in ->fill_super()
+	 */
+	s->s_bdi = &bdev_get_queue(s->s_bdev)->backing_dev_info;
 	return 0;
 }
 
diff --git a/fs/sync.c b/fs/sync.c
index 3422ba61d86..d104591b066 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -19,20 +19,29 @@
 			SYNC_FILE_RANGE_WAIT_AFTER)
 
 /*
- * Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0)
- * just dirties buffers with inodes so we have to submit IO for these buffers
- * via __sync_blockdev(). This also speeds up the wait == 1 case since in that
- * case write_inode() functions do sync_dirty_buffer() and thus effectively
- * write one block at a time.
+ * Do the filesystem syncing work. For simple filesystems
+ * writeback_inodes_sb(sb) just dirties buffers with inodes so we have to
+ * submit IO for these buffers via __sync_blockdev(). This also speeds up the
+ * wait == 1 case since in that case write_inode() functions do
+ * sync_dirty_buffer() and thus effectively write one block at a time.
  */
 static int __sync_filesystem(struct super_block *sb, int wait)
 {
+	/*
+	 * This should be safe, as we require bdi backing to actually
+	 * write out data in the first place
+	 */
+	if (!sb->s_bdi)
+		return 0;
+
 	/* Avoid doing twice syncing and cache pruning for quota sync */
-	if (!wait)
+	if (!wait) {
 		writeout_quota_sb(sb, -1);
-	else
+		writeback_inodes_sb(sb);
+	} else {
 		sync_quota_sb(sb, -1);
-	sync_inodes_sb(sb, wait);
+		sync_inodes_sb(sb);
+	}
 	if (sb->s_op->sync_fs)
 		sb->s_op->sync_fs(sb, wait);
 	return __sync_blockdev(sb->s_bdev, wait);
@@ -99,7 +108,7 @@ restart:
 		spin_unlock(&sb_lock);
 
 		down_read(&sb->s_umount);
-		if (!(sb->s_flags & MS_RDONLY) && sb->s_root)
+		if (!(sb->s_flags & MS_RDONLY) && sb->s_root && sb->s_bdi)
 			__sync_filesystem(sb, wait);
 		up_read(&sb->s_umount);
 
@@ -118,7 +127,7 @@ restart:
  */
 SYSCALL_DEFINE0(sync)
 {
-	wakeup_pdflush(0);
+	wakeup_flusher_threads(0);
 	sync_filesystems(0);
 	sync_filesystems(1);
 	if (unlikely(laptop_mode))
@@ -174,21 +183,26 @@ int file_fsync(struct file *filp, struct dentry *dentry, int datasync)
 		ret = err;
 	return ret;
 }
+EXPORT_SYMBOL(file_fsync);
 
 /**
- * vfs_fsync - perform a fsync or fdatasync on a file
+ * vfs_fsync_range - helper to sync a range of data & metadata to disk
  * @file:		file to sync
  * @dentry:		dentry of @file
- * @data:		only perform a fdatasync operation
+ * @start:		offset in bytes of the beginning of data range to sync
+ * @end:		offset in bytes of the end of data range (inclusive)
+ * @datasync:		perform only datasync
  *
- * Write back data and metadata for @file to disk.  If @datasync is
- * set only metadata needed to access modified file data is written.
+ * Write back data in range @start..@end 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 vfs_fsync_range(struct file *file, struct dentry *dentry, loff_t start,
+		    loff_t end, int datasync)
 {
 	const struct file_operations *fop;
 	struct address_space *mapping;
@@ -212,7 +226,7 @@ int vfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 		goto out;
 	}
 
-	ret = filemap_fdatawrite(mapping);
+	ret = filemap_write_and_wait_range(mapping, start, end);
 
 	/*
 	 * We need to protect against concurrent writers, which could cause
@@ -223,12 +237,29 @@ int vfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 	if (!ret)
 		ret = err;
 	mutex_unlock(&mapping->host->i_mutex);
-	err = filemap_fdatawait(mapping);
-	if (!ret)
-		ret = err;
+
 out:
 	return ret;
 }
+EXPORT_SYMBOL(vfs_fsync_range);
+
+/**
+ * vfs_fsync - perform a fsync or fdatasync on a file
+ * @file:		file to sync
+ * @dentry:		dentry of @file
+ * @datasync:		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)
+{
+	return vfs_fsync_range(file, dentry, 0, LLONG_MAX, datasync);
+}
 EXPORT_SYMBOL(vfs_fsync);
 
 static int do_fsync(unsigned int fd, int datasync)
@@ -254,6 +285,23 @@ SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
 	return do_fsync(fd, 1);
 }
 
+/**
+ * generic_write_sync - perform syncing after a write if file / inode is sync
+ * @file:	file to which the write happened
+ * @pos:	offset where the write started
+ * @count:	length of the write
+ *
+ * This is just a simple wrapper about our general syncing function.
+ */
+int generic_write_sync(struct file *file, loff_t pos, loff_t count)
+{
+	if (!(file->f_flags & O_SYNC) && !IS_SYNC(file->f_mapping->host))
+		return 0;
+	return vfs_fsync_range(file, file->f_path.dentry, pos,
+			       pos + count - 1, 1);
+}
+EXPORT_SYMBOL(generic_write_sync);
+
 /*
  * sys_sync_file_range() permits finely controlled syncing over a segment of
  * a file in the range offset .. (offset+nbytes-1) inclusive.  If nbytes is
diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c
index 14f2d71ea3c..0050fc40e8c 100644
--- a/fs/sysfs/dir.c
+++ b/fs/sysfs/dir.c
@@ -760,6 +760,7 @@ static struct dentry * sysfs_lookup(struct inode *dir, struct dentry *dentry,
 const struct inode_operations sysfs_dir_inode_operations = {
 	.lookup		= sysfs_lookup,
 	.setattr	= sysfs_setattr,
+	.setxattr	= sysfs_setxattr,
 };
 
 static void remove_dir(struct sysfs_dirent *sd)
diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c
index 555f0ff988d..e28cecf179f 100644
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -18,6 +18,8 @@
 #include <linux/capability.h>
 #include <linux/errno.h>
 #include <linux/sched.h>
+#include <linux/xattr.h>
+#include <linux/security.h>
 #include "sysfs.h"
 
 extern struct super_block * sysfs_sb;
@@ -29,12 +31,14 @@ static const struct address_space_operations sysfs_aops = {
 };
 
 static struct backing_dev_info sysfs_backing_dev_info = {
+	.name		= "sysfs",
 	.ra_pages	= 0,	/* No readahead */
 	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK,
 };
 
 static const struct inode_operations sysfs_inode_operations ={
 	.setattr	= sysfs_setattr,
+	.setxattr	= sysfs_setxattr,
 };
 
 int __init sysfs_inode_init(void)
@@ -42,18 +46,37 @@ int __init sysfs_inode_init(void)
 	return bdi_init(&sysfs_backing_dev_info);
 }
 
+struct sysfs_inode_attrs *sysfs_init_inode_attrs(struct sysfs_dirent *sd)
+{
+	struct sysfs_inode_attrs *attrs;
+	struct iattr *iattrs;
+
+	attrs = kzalloc(sizeof(struct sysfs_inode_attrs), GFP_KERNEL);
+	if (!attrs)
+		return NULL;
+	iattrs = &attrs->ia_iattr;
+
+	/* assign default attributes */
+	iattrs->ia_mode = sd->s_mode;
+	iattrs->ia_uid = 0;
+	iattrs->ia_gid = 0;
+	iattrs->ia_atime = iattrs->ia_mtime = iattrs->ia_ctime = CURRENT_TIME;
+
+	return attrs;
+}
 int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 {
 	struct inode * inode = dentry->d_inode;
 	struct sysfs_dirent * sd = dentry->d_fsdata;
-	struct iattr * sd_iattr;
+	struct sysfs_inode_attrs *sd_attrs;
+	struct iattr *iattrs;
 	unsigned int ia_valid = iattr->ia_valid;
 	int error;
 
 	if (!sd)
 		return -EINVAL;
 
-	sd_iattr = sd->s_iattr;
+	sd_attrs = sd->s_iattr;
 
 	error = inode_change_ok(inode, iattr);
 	if (error)
@@ -65,42 +88,77 @@ int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 	if (error)
 		return error;
 
-	if (!sd_iattr) {
+	if (!sd_attrs) {
 		/* setting attributes for the first time, allocate now */
-		sd_iattr = kzalloc(sizeof(struct iattr), GFP_KERNEL);
-		if (!sd_iattr)
+		sd_attrs = sysfs_init_inode_attrs(sd);
+		if (!sd_attrs)
 			return -ENOMEM;
-		/* assign default attributes */
-		sd_iattr->ia_mode = sd->s_mode;
-		sd_iattr->ia_uid = 0;
-		sd_iattr->ia_gid = 0;
-		sd_iattr->ia_atime = sd_iattr->ia_mtime = sd_iattr->ia_ctime = CURRENT_TIME;
-		sd->s_iattr = sd_iattr;
+		sd->s_iattr = sd_attrs;
+	} else {
+		/* attributes were changed at least once in past */
+		iattrs = &sd_attrs->ia_iattr;
+
+		if (ia_valid & ATTR_UID)
+			iattrs->ia_uid = iattr->ia_uid;
+		if (ia_valid & ATTR_GID)
+			iattrs->ia_gid = iattr->ia_gid;
+		if (ia_valid & ATTR_ATIME)
+			iattrs->ia_atime = timespec_trunc(iattr->ia_atime,
+					inode->i_sb->s_time_gran);
+		if (ia_valid & ATTR_MTIME)
+			iattrs->ia_mtime = timespec_trunc(iattr->ia_mtime,
+					inode->i_sb->s_time_gran);
+		if (ia_valid & ATTR_CTIME)
+			iattrs->ia_ctime = timespec_trunc(iattr->ia_ctime,
+					inode->i_sb->s_time_gran);
+		if (ia_valid & ATTR_MODE) {
+			umode_t mode = iattr->ia_mode;
+
+			if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
+				mode &= ~S_ISGID;
+			iattrs->ia_mode = sd->s_mode = mode;
+		}
 	}
+	return error;
+}
 
-	/* attributes were changed atleast once in past */
-
-	if (ia_valid & ATTR_UID)
-		sd_iattr->ia_uid = iattr->ia_uid;
-	if (ia_valid & ATTR_GID)
-		sd_iattr->ia_gid = iattr->ia_gid;
-	if (ia_valid & ATTR_ATIME)
-		sd_iattr->ia_atime = timespec_trunc(iattr->ia_atime,
-						inode->i_sb->s_time_gran);
-	if (ia_valid & ATTR_MTIME)
-		sd_iattr->ia_mtime = timespec_trunc(iattr->ia_mtime,
-						inode->i_sb->s_time_gran);
-	if (ia_valid & ATTR_CTIME)
-		sd_iattr->ia_ctime = timespec_trunc(iattr->ia_ctime,
-						inode->i_sb->s_time_gran);
-	if (ia_valid & ATTR_MODE) {
-		umode_t mode = iattr->ia_mode;
-
-		if (!in_group_p(inode->i_gid) && !capable(CAP_FSETID))
-			mode &= ~S_ISGID;
-		sd_iattr->ia_mode = sd->s_mode = mode;
-	}
+int sysfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags)
+{
+	struct sysfs_dirent *sd = dentry->d_fsdata;
+	struct sysfs_inode_attrs *iattrs;
+	void *secdata;
+	int error;
+	u32 secdata_len = 0;
+
+	if (!sd)
+		return -EINVAL;
+	if (!sd->s_iattr)
+		sd->s_iattr = sysfs_init_inode_attrs(sd);
+	if (!sd->s_iattr)
+		return -ENOMEM;
+
+	iattrs = sd->s_iattr;
+
+	if (!strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN)) {
+		const char *suffix = name + XATTR_SECURITY_PREFIX_LEN;
+		error = security_inode_setsecurity(dentry->d_inode, suffix,
+						value, size, flags);
+		if (error)
+			goto out;
+		error = security_inode_getsecctx(dentry->d_inode,
+						&secdata, &secdata_len);
+		if (error)
+			goto out;
+		if (iattrs->ia_secdata)
+			security_release_secctx(iattrs->ia_secdata,
+						iattrs->ia_secdata_len);
+		iattrs->ia_secdata = secdata;
+		iattrs->ia_secdata_len = secdata_len;
 
+	} else
+		return -EINVAL;
+out:
 	return error;
 }
 
@@ -146,6 +204,7 @@ static int sysfs_count_nlink(struct sysfs_dirent *sd)
 static void sysfs_init_inode(struct sysfs_dirent *sd, struct inode *inode)
 {
 	struct bin_attribute *bin_attr;
+	struct sysfs_inode_attrs *iattrs;
 
 	inode->i_private = sysfs_get(sd);
 	inode->i_mapping->a_ops = &sysfs_aops;
@@ -154,16 +213,20 @@ static void sysfs_init_inode(struct sysfs_dirent *sd, struct inode *inode)
 	inode->i_ino = sd->s_ino;
 	lockdep_set_class(&inode->i_mutex, &sysfs_inode_imutex_key);
 
-	if (sd->s_iattr) {
+	iattrs = sd->s_iattr;
+	if (iattrs) {
 		/* sysfs_dirent has non-default attributes
 		 * get them for the new inode from persistent copy
 		 * in sysfs_dirent
 		 */
-		set_inode_attr(inode, sd->s_iattr);
+		set_inode_attr(inode, &iattrs->ia_iattr);
+		if (iattrs->ia_secdata)
+			security_inode_notifysecctx(inode,
+						iattrs->ia_secdata,
+						iattrs->ia_secdata_len);
 	} else
 		set_default_inode_attr(inode, sd->s_mode);
 
-
 	/* initialize inode according to type */
 	switch (sysfs_type(sd)) {
 	case SYSFS_DIR:
diff --git a/fs/sysfs/symlink.c b/fs/sysfs/symlink.c
index 1d897ad808e..c5081ad7702 100644
--- a/fs/sysfs/symlink.c
+++ b/fs/sysfs/symlink.c
@@ -16,6 +16,7 @@
 #include <linux/kobject.h>
 #include <linux/namei.h>
 #include <linux/mutex.h>
+#include <linux/security.h>
 
 #include "sysfs.h"
 
@@ -209,6 +210,7 @@ static void sysfs_put_link(struct dentry *dentry, struct nameidata *nd, void *co
 }
 
 const struct inode_operations sysfs_symlink_inode_operations = {
+	.setxattr = sysfs_setxattr,
 	.readlink = generic_readlink,
 	.follow_link = sysfs_follow_link,
 	.put_link = sysfs_put_link,
diff --git a/fs/sysfs/sysfs.h b/fs/sysfs/sysfs.h
index 3fa0d98481e..af4c4e7482a 100644
--- a/fs/sysfs/sysfs.h
+++ b/fs/sysfs/sysfs.h
@@ -8,6 +8,8 @@
  * This file is released under the GPLv2.
  */
 
+#include <linux/fs.h>
+
 struct sysfs_open_dirent;
 
 /* type-specific structures for sysfs_dirent->s_* union members */
@@ -31,6 +33,12 @@ struct sysfs_elem_bin_attr {
 	struct hlist_head	buffers;
 };
 
+struct sysfs_inode_attrs {
+	struct iattr	ia_iattr;
+	void		*ia_secdata;
+	u32		ia_secdata_len;
+};
+
 /*
  * sysfs_dirent - the building block of sysfs hierarchy.  Each and
  * every sysfs node is represented by single sysfs_dirent.
@@ -56,7 +64,7 @@ struct sysfs_dirent {
 	unsigned int		s_flags;
 	ino_t			s_ino;
 	umode_t			s_mode;
-	struct iattr		*s_iattr;
+	struct sysfs_inode_attrs *s_iattr;
 };
 
 #define SD_DEACTIVATED_BIAS		INT_MIN
@@ -148,6 +156,8 @@ static inline void __sysfs_put(struct sysfs_dirent *sd)
 struct inode *sysfs_get_inode(struct sysfs_dirent *sd);
 void sysfs_delete_inode(struct inode *inode);
 int sysfs_setattr(struct dentry *dentry, struct iattr *iattr);
+int sysfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags);
 int sysfs_hash_and_remove(struct sysfs_dirent *dir_sd, const char *name);
 int sysfs_inode_init(void);
 
diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c
index eaf6d891d46..076ca50e993 100644
--- a/fs/ubifs/budget.c
+++ b/fs/ubifs/budget.c
@@ -54,41 +54,15 @@
  * @nr_to_write: how many dirty pages to write-back
  *
  * This function shrinks UBIFS liability by means of writing back some amount
- * of dirty inodes and their pages. Returns the amount of pages which were
- * written back. The returned value does not include dirty inodes which were
- * synchronized.
+ * of dirty inodes and their pages.
  *
  * Note, this function synchronizes even VFS inodes which are locked
  * (@i_mutex) by the caller of the budgeting function, because write-back does
  * not touch @i_mutex.
  */
-static int shrink_liability(struct ubifs_info *c, int nr_to_write)
+static void shrink_liability(struct ubifs_info *c, int nr_to_write)
 {
-	int nr_written;
-	struct writeback_control wbc = {
-		.sync_mode   = WB_SYNC_NONE,
-		.range_end   = LLONG_MAX,
-		.nr_to_write = nr_to_write,
-	};
-
-	generic_sync_sb_inodes(c->vfs_sb, &wbc);
-	nr_written = nr_to_write - wbc.nr_to_write;
-
-	if (!nr_written) {
-		/*
-		 * Re-try again but wait on pages/inodes which are being
-		 * written-back concurrently (e.g., by pdflush).
-		 */
-		memset(&wbc, 0, sizeof(struct writeback_control));
-		wbc.sync_mode   = WB_SYNC_ALL;
-		wbc.range_end   = LLONG_MAX;
-		wbc.nr_to_write = nr_to_write;
-		generic_sync_sb_inodes(c->vfs_sb, &wbc);
-		nr_written = nr_to_write - wbc.nr_to_write;
-	}
-
-	dbg_budg("%d pages were written back", nr_written);
-	return nr_written;
+	writeback_inodes_sb(c->vfs_sb);
 }
 
 /**
@@ -741,7 +715,7 @@ long long ubifs_get_free_space_nolock(struct ubifs_info *c)
  * ubifs_get_free_space - return amount of free space.
  * @c: UBIFS file-system description object
  *
- * This function calculates and retuns amount of free space to report to
+ * This function calculates and returns amount of free space to report to
  * user-space.
  */
 long long ubifs_get_free_space(struct ubifs_info *c)
diff --git a/fs/ubifs/commit.c b/fs/ubifs/commit.c
index f3a7945527f..4775af40116 100644
--- a/fs/ubifs/commit.c
+++ b/fs/ubifs/commit.c
@@ -510,7 +510,7 @@ 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;
+	union ubifs_key uninitialized_var(lower_key), upper_key, l_key, u_key;
 	unsigned long long uninitialized_var(last_sqnum);
 	struct ubifs_idx_node *idx;
 	struct list_head list;
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index ce2cd834361..dbc093afd94 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -210,6 +210,20 @@ const char *dbg_cstate(int cmt_state)
 	}
 }
 
+const char *dbg_jhead(int jhead)
+{
+	switch (jhead) {
+	case GCHD:
+		return "0 (GC)";
+	case BASEHD:
+		return "1 (base)";
+	case DATAHD:
+		return "2 (data)";
+	default:
+		return "unknown journal head";
+	}
+}
+
 static void dump_ch(const struct ubifs_ch *ch)
 {
 	printk(KERN_DEBUG "\tmagic          %#x\n", le32_to_cpu(ch->magic));
@@ -623,8 +637,9 @@ void dbg_dump_budg(struct ubifs_info *c)
 	/* If we are in R/O mode, journal heads do not exist */
 	if (c->jheads)
 		for (i = 0; i < c->jhead_cnt; i++)
-			printk(KERN_DEBUG "\tjhead %d\t LEB %d\n",
-			       c->jheads[i].wbuf.jhead, c->jheads[i].wbuf.lnum);
+			printk(KERN_DEBUG "\tjhead %s\t LEB %d\n",
+			       dbg_jhead(c->jheads[i].wbuf.jhead),
+			       c->jheads[i].wbuf.lnum);
 	for (rb = rb_first(&c->buds); rb; rb = rb_next(rb)) {
 		bud = rb_entry(rb, struct ubifs_bud, rb);
 		printk(KERN_DEBUG "\tbud LEB %d\n", bud->lnum);
@@ -648,9 +663,90 @@ void dbg_dump_budg(struct ubifs_info *c)
 
 void dbg_dump_lprop(const struct ubifs_info *c, const struct ubifs_lprops *lp)
 {
-	printk(KERN_DEBUG "LEB %d lprops: free %d, dirty %d (used %d), "
-	       "flags %#x\n", lp->lnum, lp->free, lp->dirty,
-	       c->leb_size - lp->free - lp->dirty, lp->flags);
+	int i, spc, dark = 0, dead = 0;
+	struct rb_node *rb;
+	struct ubifs_bud *bud;
+
+	spc = lp->free + lp->dirty;
+	if (spc < c->dead_wm)
+		dead = spc;
+	else
+		dark = ubifs_calc_dark(c, spc);
+
+	if (lp->flags & LPROPS_INDEX)
+		printk(KERN_DEBUG "LEB %-7d free %-8d dirty %-8d used %-8d "
+		       "free + dirty %-8d flags %#x (", lp->lnum, lp->free,
+		       lp->dirty, c->leb_size - spc, spc, lp->flags);
+	else
+		printk(KERN_DEBUG "LEB %-7d free %-8d dirty %-8d used %-8d "
+		       "free + dirty %-8d dark %-4d dead %-4d nodes fit %-3d "
+		       "flags %#-4x (", lp->lnum, lp->free, lp->dirty,
+		       c->leb_size - spc, spc, dark, dead,
+		       (int)(spc / UBIFS_MAX_NODE_SZ), lp->flags);
+
+	if (lp->flags & LPROPS_TAKEN) {
+		if (lp->flags & LPROPS_INDEX)
+			printk(KERN_CONT "index, taken");
+		else
+			printk(KERN_CONT "taken");
+	} else {
+		const char *s;
+
+		if (lp->flags & LPROPS_INDEX) {
+			switch (lp->flags & LPROPS_CAT_MASK) {
+			case LPROPS_DIRTY_IDX:
+				s = "dirty index";
+				break;
+			case LPROPS_FRDI_IDX:
+				s = "freeable index";
+				break;
+			default:
+				s = "index";
+			}
+		} else {
+			switch (lp->flags & LPROPS_CAT_MASK) {
+			case LPROPS_UNCAT:
+				s = "not categorized";
+				break;
+			case LPROPS_DIRTY:
+				s = "dirty";
+				break;
+			case LPROPS_FREE:
+				s = "free";
+				break;
+			case LPROPS_EMPTY:
+				s = "empty";
+				break;
+			case LPROPS_FREEABLE:
+				s = "freeable";
+				break;
+			default:
+				s = NULL;
+				break;
+			}
+		}
+		printk(KERN_CONT "%s", s);
+	}
+
+	for (rb = rb_first((struct rb_root *)&c->buds); rb; rb = rb_next(rb)) {
+		bud = rb_entry(rb, struct ubifs_bud, rb);
+		if (bud->lnum == lp->lnum) {
+			int head = 0;
+			for (i = 0; i < c->jhead_cnt; i++) {
+				if (lp->lnum == c->jheads[i].wbuf.lnum) {
+					printk(KERN_CONT ", jhead %s",
+					       dbg_jhead(i));
+					head = 1;
+				}
+			}
+			if (!head)
+				printk(KERN_CONT ", bud of jhead %s",
+				       dbg_jhead(bud->jhead));
+		}
+	}
+	if (lp->lnum == c->gc_lnum)
+		printk(KERN_CONT ", GC LEB");
+	printk(KERN_CONT ")\n");
 }
 
 void dbg_dump_lprops(struct ubifs_info *c)
@@ -724,7 +820,7 @@ void dbg_dump_leb(const struct ubifs_info *c, int lnum)
 
 	printk(KERN_DEBUG "(pid %d) start dumping LEB %d\n",
 	       current->pid, lnum);
-	sleb = ubifs_scan(c, lnum, 0, c->dbg->buf);
+	sleb = ubifs_scan(c, lnum, 0, c->dbg->buf, 0);
 	if (IS_ERR(sleb)) {
 		ubifs_err("scan error %d", (int)PTR_ERR(sleb));
 		return;
@@ -909,8 +1005,10 @@ out:
 	ubifs_msg("saved lprops statistics dump");
 	dbg_dump_lstats(&d->saved_lst);
 	ubifs_get_lp_stats(c, &lst);
+
 	ubifs_msg("current lprops statistics dump");
-	dbg_dump_lstats(&d->saved_lst);
+	dbg_dump_lstats(&lst);
+
 	spin_lock(&c->space_lock);
 	dbg_dump_budg(c);
 	spin_unlock(&c->space_lock);
diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h
index c1cd73b2e06..29d960101ea 100644
--- a/fs/ubifs/debug.h
+++ b/fs/ubifs/debug.h
@@ -271,6 +271,7 @@ 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_jhead(int jhead);
 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);
@@ -321,6 +322,8 @@ void dbg_check_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat,
 int dbg_check_lprops(struct ubifs_info *c);
 int dbg_check_lpt_nodes(struct ubifs_info *c, struct ubifs_cnode *cnode,
 			int row, int col);
+int dbg_check_inode_size(struct ubifs_info *c, const struct inode *inode,
+			 loff_t size);
 
 /* Force the use of in-the-gaps method for testing */
 
@@ -425,6 +428,7 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c);
 
 #define dbg_ntype(type)                        ""
 #define dbg_cstate(cmt_state)                  ""
+#define dbg_jhead(jhead)                       ""
 #define dbg_get_key_dump(c, key)               ({})
 #define dbg_dump_inode(c, inode)               ({})
 #define dbg_dump_node(c, node)                 ({})
@@ -460,6 +464,7 @@ void dbg_debugfs_exit_fs(struct ubifs_info *c);
 #define dbg_check_heap(c, heap, cat, add_pos)      ({})
 #define dbg_check_lprops(c)                        0
 #define dbg_check_lpt_nodes(c, cnode, row, col)    0
+#define dbg_check_inode_size(c, inode, size)       0
 #define dbg_force_in_the_gaps_enabled              0
 #define dbg_force_in_the_gaps()                    0
 #define dbg_failure_mode                           0
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index 6d34dc7e33e..2e6481a7701 100644
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -21,34 +21,32 @@
  */
 
 /*
- * This file implements VFS file and inode operations of regular files, device
+ * This file implements VFS file and inode operations for regular files, device
  * nodes and symlinks as well as address space operations.
  *
- * UBIFS uses 2 page flags: PG_private and PG_checked. PG_private is set if the
- * page is dirty and is used for budgeting purposes - dirty pages should not be
- * budgeted. The PG_checked flag is set if full budgeting is required for the
- * page e.g., when it corresponds to a file hole or it is just beyond the file
- * size. The budgeting is done in 'ubifs_write_begin()', because it is OK to
- * fail in this function, and the budget is released in 'ubifs_write_end()'. So
- * the PG_private and PG_checked flags carry the information about how the page
- * was budgeted, to make it possible to release the budget properly.
+ * UBIFS uses 2 page flags: @PG_private and @PG_checked. @PG_private is set if
+ * the page is dirty and is used for optimization purposes - dirty pages are
+ * not budgeted so the flag shows that 'ubifs_write_end()' should not release
+ * the budget for this page. The @PG_checked flag is set if full budgeting is
+ * required for the page e.g., when it corresponds to a file hole or it is
+ * beyond the file size. The budgeting is done in 'ubifs_write_begin()', because
+ * it is OK to fail in this function, and the budget is released in
+ * 'ubifs_write_end()'. So the @PG_private and @PG_checked flags carry
+ * information about how the page was budgeted, to make it possible to release
+ * the budget properly.
  *
- * A thing to keep in mind: inode's 'i_mutex' is locked in most VFS operations
- * we implement. However, this is not true for '->writepage()', which might be
- * called with 'i_mutex' unlocked. For example, when pdflush is performing
- * write-back, it calls 'writepage()' with unlocked 'i_mutex', although the
- * inode has 'I_LOCK' flag in this case. At "normal" work-paths 'i_mutex' is
- * locked in '->writepage', e.g. in "sys_write -> alloc_pages -> direct reclaim
- * path'. So, in '->writepage()' we are only guaranteed that the page is
- * locked.
+ * A thing to keep in mind: inode @i_mutex is locked in most VFS operations we
+ * implement. However, this is not true for 'ubifs_writepage()', which may be
+ * called with @i_mutex unlocked. For example, when pdflush is doing background
+ * write-back, it calls 'ubifs_writepage()' with unlocked @i_mutex. At "normal"
+ * work-paths the @i_mutex is locked in 'ubifs_writepage()', e.g. in the
+ * "sys_write -> alloc_pages -> direct reclaim path". So, in 'ubifs_writepage()'
+ * we are only guaranteed that the page is locked.
  *
- * Similarly, 'i_mutex' does not have to be locked in readpage(), e.g.,
- * readahead path does not have it locked ("sys_read -> generic_file_aio_read
- * -> ondemand_readahead -> readpage"). In case of readahead, 'I_LOCK' flag is
- * not set as well. However, UBIFS disables readahead.
- *
- * This, for example means that there might be 2 concurrent '->writepage()'
- * calls for the same inode, but different inode dirty pages.
+ * Similarly, @i_mutex is not always locked in 'ubifs_readpage()', e.g., the
+ * read-ahead path does not lock it ("sys_read -> generic_file_aio_read ->
+ * ondemand_readahead -> readpage"). In case of readahead, @I_LOCK flag is not
+ * set as well. However, UBIFS disables readahead.
  */
 
 #include "ubifs.h"
@@ -449,9 +447,9 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 			/*
 			 * We change whole page so no need to load it. But we
 			 * have to set the @PG_checked flag to make the further
-			 * code the page is new. This might be not true, but it
-			 * is better to budget more that to read the page from
-			 * the media.
+			 * code know that the page is new. This might be not
+			 * true, but it is better to budget more than to read
+			 * the page from the media.
 			 */
 			SetPageChecked(page);
 			skipped_read = 1;
@@ -497,8 +495,8 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping,
 	}
 
 	/*
-	 * Whee, we aquired budgeting quickly - without involving
-	 * garbage-collection, committing or forceing write-back. We return
+	 * Whee, we acquired budgeting quickly - without involving
+	 * garbage-collection, committing or forcing write-back. We return
 	 * with @ui->ui_mutex locked if we are appending pages, and unlocked
 	 * otherwise. This is an optimization (slightly hacky though).
 	 */
@@ -562,7 +560,7 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping,
 
 		/*
 		 * Return 0 to force VFS to repeat the whole operation, or the
-		 * error code if 'do_readpage()' failes.
+		 * error code if 'do_readpage()' fails.
 		 */
 		copied = do_readpage(page);
 		goto out;
@@ -1175,11 +1173,11 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode,
 	ui->ui_size = inode->i_size;
 	/* Truncation changes inode [mc]time */
 	inode->i_mtime = inode->i_ctime = ubifs_current_time(inode);
-	/* The other attributes may be changed at the same time as well */
+	/* Other attributes may be changed at the same time as well */
 	do_attr_changes(inode, attr);
-
 	err = ubifs_jnl_truncate(c, inode, old_size, new_size);
 	mutex_unlock(&ui->ui_mutex);
+
 out_budg:
 	if (budgeted)
 		ubifs_release_budget(c, &req);
diff --git a/fs/ubifs/gc.c b/fs/ubifs/gc.c
index f0f5f15d384..618c2701d3a 100644
--- a/fs/ubifs/gc.c
+++ b/fs/ubifs/gc.c
@@ -529,7 +529,7 @@ int ubifs_garbage_collect_leb(struct ubifs_info *c, struct ubifs_lprops *lp)
 	 * We scan the entire LEB even though we only really need to scan up to
 	 * (c->leb_size - lp->free).
 	 */
-	sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+	sleb = ubifs_scan(c, lnum, 0, c->sbuf, 0);
 	if (IS_ERR(sleb))
 		return PTR_ERR(sleb);
 
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
index 762a7d6cec7..e589fedaf1e 100644
--- a/fs/ubifs/io.c
+++ b/fs/ubifs/io.c
@@ -297,7 +297,7 @@ static enum hrtimer_restart wbuf_timer_callback_nolock(struct hrtimer *timer)
 {
 	struct ubifs_wbuf *wbuf = container_of(timer, struct ubifs_wbuf, timer);
 
-	dbg_io("jhead %d", wbuf->jhead);
+	dbg_io("jhead %s", dbg_jhead(wbuf->jhead));
 	wbuf->need_sync = 1;
 	wbuf->c->need_wbuf_sync = 1;
 	ubifs_wake_up_bgt(wbuf->c);
@@ -314,7 +314,8 @@ static void new_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
 
 	if (wbuf->no_timer)
 		return;
-	dbg_io("set timer for jhead %d, %llu-%llu millisecs", wbuf->jhead,
+	dbg_io("set timer for jhead %s, %llu-%llu millisecs",
+	       dbg_jhead(wbuf->jhead),
 	       div_u64(ktime_to_ns(wbuf->softlimit), USEC_PER_SEC),
 	       div_u64(ktime_to_ns(wbuf->softlimit) + wbuf->delta,
 		       USEC_PER_SEC));
@@ -351,8 +352,8 @@ int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf)
 		/* Write-buffer is empty or not seeked */
 		return 0;
 
-	dbg_io("LEB %d:%d, %d bytes, jhead %d",
-	       wbuf->lnum, wbuf->offs, wbuf->used, wbuf->jhead);
+	dbg_io("LEB %d:%d, %d bytes, jhead %s",
+	       wbuf->lnum, wbuf->offs, wbuf->used, dbg_jhead(wbuf->jhead));
 	ubifs_assert(!(c->vfs_sb->s_flags & MS_RDONLY));
 	ubifs_assert(!(wbuf->avail & 7));
 	ubifs_assert(wbuf->offs + c->min_io_size <= c->leb_size);
@@ -401,7 +402,7 @@ int ubifs_wbuf_seek_nolock(struct ubifs_wbuf *wbuf, int lnum, int offs,
 {
 	const struct ubifs_info *c = wbuf->c;
 
-	dbg_io("LEB %d:%d, jhead %d", lnum, offs, wbuf->jhead);
+	dbg_io("LEB %d:%d, jhead %s", lnum, offs, dbg_jhead(wbuf->jhead));
 	ubifs_assert(lnum >= 0 && lnum < c->leb_cnt);
 	ubifs_assert(offs >= 0 && offs <= c->leb_size);
 	ubifs_assert(offs % c->min_io_size == 0 && !(offs & 7));
@@ -508,9 +509,9 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 	struct ubifs_info *c = wbuf->c;
 	int err, written, n, aligned_len = ALIGN(len, 8), offs;
 
-	dbg_io("%d bytes (%s) to jhead %d wbuf at LEB %d:%d", len,
-	       dbg_ntype(((struct ubifs_ch *)buf)->node_type), wbuf->jhead,
-	       wbuf->lnum, wbuf->offs + wbuf->used);
+	dbg_io("%d bytes (%s) to jhead %s wbuf at LEB %d:%d", len,
+	       dbg_ntype(((struct ubifs_ch *)buf)->node_type),
+	       dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs + wbuf->used);
 	ubifs_assert(len > 0 && wbuf->lnum >= 0 && wbuf->lnum < c->leb_cnt);
 	ubifs_assert(wbuf->offs >= 0 && wbuf->offs % c->min_io_size == 0);
 	ubifs_assert(!(wbuf->offs & 7) && wbuf->offs <= c->leb_size);
@@ -535,8 +536,8 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 		memcpy(wbuf->buf + wbuf->used, buf, len);
 
 		if (aligned_len == wbuf->avail) {
-			dbg_io("flush jhead %d wbuf to LEB %d:%d",
-			       wbuf->jhead, wbuf->lnum, wbuf->offs);
+			dbg_io("flush jhead %s wbuf to LEB %d:%d",
+			       dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
 			err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf,
 					    wbuf->offs, c->min_io_size,
 					    wbuf->dtype);
@@ -564,8 +565,8 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
 	 * minimal I/O unit. We have to fill and flush write-buffer and switch
 	 * to the next min. I/O unit.
 	 */
-	dbg_io("flush jhead %d wbuf to LEB %d:%d",
-	       wbuf->jhead, wbuf->lnum, wbuf->offs);
+	dbg_io("flush jhead %s wbuf to LEB %d:%d",
+	       dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
 	memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
 	err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs,
 			    c->min_io_size, wbuf->dtype);
@@ -698,8 +699,8 @@ int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
 	int err, rlen, overlap;
 	struct ubifs_ch *ch = buf;
 
-	dbg_io("LEB %d:%d, %s, length %d, jhead %d", lnum, offs,
-	       dbg_ntype(type), len, wbuf->jhead);
+	dbg_io("LEB %d:%d, %s, length %d, jhead %s", lnum, offs,
+	       dbg_ntype(type), len, dbg_jhead(wbuf->jhead));
 	ubifs_assert(wbuf && lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
 	ubifs_assert(!(offs & 7) && offs < c->leb_size);
 	ubifs_assert(type >= 0 && type < UBIFS_NODE_TYPES_CNT);
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index 64b5f3a309f..d321baeca68 100644
--- a/fs/ubifs/journal.c
+++ b/fs/ubifs/journal.c
@@ -158,7 +158,7 @@ again:
 	 * some. But the write-buffer mutex has to be unlocked because
 	 * GC also takes it.
 	 */
-	dbg_jnl("no free space  jhead %d, run GC", jhead);
+	dbg_jnl("no free space in jhead %s, run GC", dbg_jhead(jhead));
 	mutex_unlock(&wbuf->io_mutex);
 
 	lnum = ubifs_garbage_collect(c, 0);
@@ -173,7 +173,8 @@ again:
 		 * because we dropped @wbuf->io_mutex, so try once
 		 * again.
 		 */
-		dbg_jnl("GC couldn't make a free LEB for jhead %d", jhead);
+		dbg_jnl("GC couldn't make a free LEB for jhead %s",
+			dbg_jhead(jhead));
 		if (retries++ < 2) {
 			dbg_jnl("retry (%d)", retries);
 			goto again;
@@ -184,7 +185,7 @@ again:
 	}
 
 	mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
-	dbg_jnl("got LEB %d for jhead %d", lnum, jhead);
+	dbg_jnl("got LEB %d for jhead %s", lnum, dbg_jhead(jhead));
 	avail = c->leb_size - wbuf->offs - wbuf->used;
 
 	if (wbuf->lnum != -1 && avail >= len) {
@@ -255,7 +256,8 @@ static int write_node(struct ubifs_info *c, int jhead, void *node, int len,
 	*lnum = c->jheads[jhead].wbuf.lnum;
 	*offs = c->jheads[jhead].wbuf.offs + c->jheads[jhead].wbuf.used;
 
-	dbg_jnl("jhead %d, LEB %d:%d, len %d", jhead, *lnum, *offs, len);
+	dbg_jnl("jhead %s, LEB %d:%d, len %d",
+		dbg_jhead(jhead), *lnum, *offs, len);
 	ubifs_prepare_node(c, node, len, 0);
 
 	return ubifs_wbuf_write_nolock(wbuf, node, len);
@@ -285,7 +287,8 @@ static int write_head(struct ubifs_info *c, int jhead, void *buf, int len,
 
 	*lnum = c->jheads[jhead].wbuf.lnum;
 	*offs = c->jheads[jhead].wbuf.offs + c->jheads[jhead].wbuf.used;
-	dbg_jnl("jhead %d, LEB %d:%d, len %d", jhead, *lnum, *offs, len);
+	dbg_jnl("jhead %s, LEB %d:%d, len %d",
+		dbg_jhead(jhead), *lnum, *offs, len);
 
 	err = ubifs_wbuf_write_nolock(wbuf, buf, len);
 	if (err)
diff --git a/fs/ubifs/key.h b/fs/ubifs/key.h
index 5fa27ea031b..0f530c684f0 100644
--- a/fs/ubifs/key.h
+++ b/fs/ubifs/key.h
@@ -229,23 +229,6 @@ static inline void xent_key_init(const struct ubifs_info *c,
 }
 
 /**
- * xent_key_init_hash - initialize extended attribute entry key without
- *                      re-calculating hash function.
- * @c: UBIFS file-system description object
- * @key: key to initialize
- * @inum: host inode number
- * @hash: extended attribute entry name hash
- */
-static inline void xent_key_init_hash(const struct ubifs_info *c,
-				      union ubifs_key *key, ino_t inum,
-				      uint32_t hash)
-{
-	ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
-	key->u32[0] = inum;
-	key->u32[1] = hash | (UBIFS_XENT_KEY << UBIFS_S_KEY_HASH_BITS);
-}
-
-/**
  * xent_key_init_flash - initialize on-flash extended attribute entry key.
  * @c: UBIFS file-system description object
  * @k: key to initialize
@@ -295,22 +278,15 @@ static inline void data_key_init(const struct ubifs_info *c,
 }
 
 /**
- * data_key_init_flash - initialize on-flash data key.
+ * highest_data_key - get the highest possible data key for an inode.
  * @c: UBIFS file-system description object
- * @k: key to initialize
+ * @key: key to initialize
  * @inum: inode number
- * @block: block number
  */
-static inline void data_key_init_flash(const struct ubifs_info *c, void *k,
-				       ino_t inum, unsigned int block)
+static inline void highest_data_key(const struct ubifs_info *c,
+				   union ubifs_key *key, ino_t inum)
 {
-	union ubifs_key *key = k;
-
-	ubifs_assert(!(block & ~UBIFS_S_KEY_BLOCK_MASK));
-	key->j32[0] = cpu_to_le32(inum);
-	key->j32[1] = cpu_to_le32(block |
-				  (UBIFS_DATA_KEY << UBIFS_S_KEY_BLOCK_BITS));
-	memset(k + 8, 0, UBIFS_MAX_KEY_LEN - 8);
+	data_key_init(c, key, inum, UBIFS_S_KEY_BLOCK_MASK);
 }
 
 /**
@@ -554,4 +530,5 @@ static inline unsigned long long key_max_inode_size(const struct ubifs_info *c)
 		return 0;
 	}
 }
+
 #endif /* !__UBIFS_KEY_H__ */
diff --git a/fs/ubifs/log.c b/fs/ubifs/log.c
index 56e33772a1e..c345e125f42 100644
--- a/fs/ubifs/log.c
+++ b/fs/ubifs/log.c
@@ -169,8 +169,8 @@ void ubifs_add_bud(struct ubifs_info *c, struct ubifs_bud *bud)
 	 */
 	c->bud_bytes += c->leb_size - bud->start;
 
-	dbg_log("LEB %d:%d, jhead %d, bud_bytes %lld", bud->lnum,
-		bud->start, bud->jhead, c->bud_bytes);
+	dbg_log("LEB %d:%d, jhead %s, bud_bytes %lld", bud->lnum,
+		bud->start, dbg_jhead(bud->jhead), c->bud_bytes);
 	spin_unlock(&c->buds_lock);
 }
 
@@ -355,16 +355,16 @@ static void remove_buds(struct ubifs_info *c)
 			 * heads (non-closed buds).
 			 */
 			c->cmt_bud_bytes += wbuf->offs - bud->start;
-			dbg_log("preserve %d:%d, jhead %d, bud bytes %d, "
+			dbg_log("preserve %d:%d, jhead %s, bud bytes %d, "
 				"cmt_bud_bytes %lld", bud->lnum, bud->start,
-				bud->jhead, wbuf->offs - bud->start,
+				dbg_jhead(bud->jhead), wbuf->offs - bud->start,
 				c->cmt_bud_bytes);
 			bud->start = wbuf->offs;
 		} else {
 			c->cmt_bud_bytes += c->leb_size - bud->start;
-			dbg_log("remove %d:%d, jhead %d, bud bytes %d, "
+			dbg_log("remove %d:%d, jhead %s, bud bytes %d, "
 				"cmt_bud_bytes %lld", bud->lnum, bud->start,
-				bud->jhead, c->leb_size - bud->start,
+				dbg_jhead(bud->jhead), c->leb_size - bud->start,
 				c->cmt_bud_bytes);
 			rb_erase(p1, &c->buds);
 			/*
@@ -429,7 +429,8 @@ int ubifs_log_start_commit(struct ubifs_info *c, int *ltail_lnum)
 		if (lnum == -1 || offs == c->leb_size)
 			continue;
 
-		dbg_log("add ref to LEB %d:%d for jhead %d", lnum, offs, i);
+		dbg_log("add ref to LEB %d:%d for jhead %s",
+			lnum, offs, dbg_jhead(i));
 		ref = buf + len;
 		ref->ch.node_type = UBIFS_REF_NODE;
 		ref->lnum = cpu_to_le32(lnum);
@@ -695,7 +696,7 @@ int ubifs_consolidate_log(struct ubifs_info *c)
 	lnum = c->ltail_lnum;
 	write_lnum = lnum;
 	while (1) {
-		sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 0);
 		if (IS_ERR(sleb)) {
 			err = PTR_ERR(sleb);
 			goto out_free;
diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c
index 4cdd284dea5..4d4ca388889 100644
--- a/fs/ubifs/lprops.c
+++ b/fs/ubifs/lprops.c
@@ -281,7 +281,7 @@ void ubifs_add_to_cat(struct ubifs_info *c, struct ubifs_lprops *lprops,
 	case LPROPS_FREE:
 		if (add_to_lpt_heap(c, lprops, cat))
 			break;
-		/* No more room on heap so make it uncategorized */
+		/* No more room on heap so make it un-categorized */
 		cat = LPROPS_UNCAT;
 		/* Fall through */
 	case LPROPS_UNCAT:
@@ -375,8 +375,8 @@ void ubifs_replace_cat(struct ubifs_info *c, struct ubifs_lprops *old_lprops,
  * @lprops: LEB properties
  *
  * A LEB may have fallen off of the bottom of a heap, and ended up as
- * uncategorized even though it has enough space for us now. If that is the case
- * this function will put the LEB back onto a heap.
+ * un-categorized even though it has enough space for us now. If that is the
+ * case this function will put the LEB back onto a heap.
  */
 void ubifs_ensure_cat(struct ubifs_info *c, struct ubifs_lprops *lprops)
 {
@@ -436,10 +436,10 @@ int ubifs_categorize_lprops(const struct ubifs_info *c,
 /**
  * change_category - change LEB properties category.
  * @c: UBIFS file-system description object
- * @lprops: LEB properties to recategorize
+ * @lprops: LEB properties to re-categorize
  *
  * LEB properties are categorized to enable fast find operations. When the LEB
- * properties change they must be recategorized.
+ * properties change they must be re-categorized.
  */
 static void change_category(struct ubifs_info *c, struct ubifs_lprops *lprops)
 {
@@ -461,21 +461,18 @@ static void change_category(struct ubifs_info *c, struct ubifs_lprops *lprops)
 }
 
 /**
- * calc_dark - calculate LEB dark space size.
+ * ubifs_calc_dark - calculate LEB dark space size.
  * @c: the UBIFS file-system description object
  * @spc: amount of free and dirty space in the LEB
  *
- * This function calculates amount of dark space in an LEB which has @spc bytes
- * of free and dirty space. Returns the calculations result.
+ * This function calculates and returns amount of dark space in an LEB which
+ * has @spc bytes of free and dirty space.
  *
- * Dark space is the space which is not always usable - it depends on which
- * nodes are written in which order. E.g., if an LEB has only 512 free bytes,
- * it is dark space, because it cannot fit a large data node. So UBIFS cannot
- * count on this LEB and treat these 512 bytes as usable because it is not true
- * if, for example, only big chunks of uncompressible data will be written to
- * the FS.
+ * UBIFS is trying to account the space which might not be usable, and this
+ * space is called "dark space". For example, if an LEB has only %512 free
+ * bytes, it is dark space, because it cannot fit a large data node.
  */
-static int calc_dark(struct ubifs_info *c, int spc)
+int ubifs_calc_dark(const struct ubifs_info *c, int spc)
 {
 	ubifs_assert(!(spc & 7));
 
@@ -518,7 +515,7 @@ static int is_lprops_dirty(struct ubifs_info *c, struct ubifs_lprops *lprops)
  * @free: new free space amount
  * @dirty: new dirty space amount
  * @flags: new flags
- * @idx_gc_cnt: change to the count of idx_gc list
+ * @idx_gc_cnt: change to the count of @idx_gc list
  *
  * 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
@@ -535,7 +532,7 @@ const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
 {
 	/*
 	 * This is the only function that is allowed to change lprops, so we
-	 * discard the const qualifier.
+	 * discard the "const" qualifier.
 	 */
 	struct ubifs_lprops *lprops = (struct ubifs_lprops *)lp;
 
@@ -575,7 +572,7 @@ const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
 		if (old_spc < c->dead_wm)
 			c->lst.total_dead -= old_spc;
 		else
-			c->lst.total_dark -= calc_dark(c, old_spc);
+			c->lst.total_dark -= ubifs_calc_dark(c, old_spc);
 
 		c->lst.total_used -= c->leb_size - old_spc;
 	}
@@ -616,7 +613,7 @@ const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
 		if (new_spc < c->dead_wm)
 			c->lst.total_dead += new_spc;
 		else
-			c->lst.total_dark += calc_dark(c, new_spc);
+			c->lst.total_dark += ubifs_calc_dark(c, new_spc);
 
 		c->lst.total_used += c->leb_size - new_spc;
 	}
@@ -1096,7 +1093,7 @@ 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, 0);
 	if (IS_ERR(sleb)) {
 		/*
 		 * After an unclean unmount, empty and freeable LEBs
@@ -1107,7 +1104,7 @@ static int scan_check_cb(struct ubifs_info *c,
 				  "- continuing checking");
 			lst->empty_lebs += 1;
 			lst->total_free += c->leb_size;
-			lst->total_dark += calc_dark(c, c->leb_size);
+			lst->total_dark += ubifs_calc_dark(c, c->leb_size);
 			return LPT_SCAN_CONTINUE;
 		}
 
@@ -1117,7 +1114,7 @@ static int scan_check_cb(struct ubifs_info *c,
 				  "- continuing checking");
 			lst->total_free  += lp->free;
 			lst->total_dirty += lp->dirty;
-			lst->total_dark  +=  calc_dark(c, c->leb_size);
+			lst->total_dark  +=  ubifs_calc_dark(c, c->leb_size);
 			return LPT_SCAN_CONTINUE;
 		}
 		data->err = PTR_ERR(sleb);
@@ -1235,7 +1232,7 @@ static int scan_check_cb(struct ubifs_info *c,
 		if (spc < c->dead_wm)
 			lst->total_dead += spc;
 		else
-			lst->total_dark += calc_dark(c, spc);
+			lst->total_dark += ubifs_calc_dark(c, spc);
 	}
 
 	ubifs_scan_destroy(sleb);
diff --git a/fs/ubifs/master.c b/fs/ubifs/master.c
index a88f33801b9..28beaeedadc 100644
--- a/fs/ubifs/master.c
+++ b/fs/ubifs/master.c
@@ -29,7 +29,8 @@
  * @c: UBIFS file-system description object
  *
  * This function scans the master node LEBs and search for the latest master
- * node. Returns zero in case of success and a negative error code in case of
+ * node. Returns zero in case of success, %-EUCLEAN if there master area is
+ * corrupted and requires recovery, and a negative error code in case of
  * failure.
  */
 static int scan_for_master(struct ubifs_info *c)
@@ -40,7 +41,7 @@ static int scan_for_master(struct ubifs_info *c)
 
 	lnum = UBIFS_MST_LNUM;
 
-	sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+	sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
 	if (IS_ERR(sleb))
 		return PTR_ERR(sleb);
 	nodes_cnt = sleb->nodes_cnt;
@@ -48,7 +49,7 @@ static int scan_for_master(struct ubifs_info *c)
 		snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node,
 				  list);
 		if (snod->type != UBIFS_MST_NODE)
-			goto out;
+			goto out_dump;
 		memcpy(c->mst_node, snod->node, snod->len);
 		offs = snod->offs;
 	}
@@ -56,7 +57,7 @@ static int scan_for_master(struct ubifs_info *c)
 
 	lnum += 1;
 
-	sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+	sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
 	if (IS_ERR(sleb))
 		return PTR_ERR(sleb);
 	if (sleb->nodes_cnt != nodes_cnt)
@@ -65,7 +66,7 @@ static int scan_for_master(struct ubifs_info *c)
 		goto out;
 	snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node, list);
 	if (snod->type != UBIFS_MST_NODE)
-		goto out;
+		goto out_dump;
 	if (snod->offs != offs)
 		goto out;
 	if (memcmp((void *)c->mst_node + UBIFS_CH_SZ,
@@ -78,6 +79,12 @@ static int scan_for_master(struct ubifs_info *c)
 
 out:
 	ubifs_scan_destroy(sleb);
+	return -EUCLEAN;
+
+out_dump:
+	ubifs_err("unexpected node type %d master LEB %d:%d",
+		  snod->type, lnum, snod->offs);
+	ubifs_scan_destroy(sleb);
 	return -EINVAL;
 }
 
@@ -256,7 +263,8 @@ int ubifs_read_master(struct ubifs_info *c)
 
 	err = scan_for_master(c);
 	if (err) {
-		err = ubifs_recover_master_node(c);
+		if (err == -EUCLEAN)
+			err = ubifs_recover_master_node(c);
 		if (err)
 			/*
 			 * Note, we do not free 'c->mst_node' here because the
diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
index 152a7b34a14..82009c74b6a 100644
--- a/fs/ubifs/orphan.c
+++ b/fs/ubifs/orphan.c
@@ -670,9 +670,10 @@ static int kill_orphans(struct ubifs_info *c)
 		struct ubifs_scan_leb *sleb;
 
 		dbg_rcvry("LEB %d", lnum);
-		sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
 		if (IS_ERR(sleb)) {
-			sleb = ubifs_recover_leb(c, lnum, 0, c->sbuf, 0);
+			if (PTR_ERR(sleb) == -EUCLEAN)
+				sleb = ubifs_recover_leb(c, lnum, 0, c->sbuf, 0);
 			if (IS_ERR(sleb)) {
 				err = PTR_ERR(sleb);
 				break;
@@ -899,7 +900,7 @@ 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, 0);
 		if (IS_ERR(sleb)) {
 			err = PTR_ERR(sleb);
 			break;
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
index e5f6cf8a115..f94ddf7efba 100644
--- a/fs/ubifs/recovery.c
+++ b/fs/ubifs/recovery.c
@@ -286,7 +286,7 @@ int ubifs_recover_master_node(struct ubifs_info *c)
 		mst = mst2;
 	}
 
-	dbg_rcvry("recovered master node from LEB %d",
+	ubifs_msg("recovered master node from LEB %d",
 		  (mst == mst1 ? UBIFS_MST_LNUM : UBIFS_MST_LNUM + 1));
 
 	memcpy(c->mst_node, mst, UBIFS_MST_NODE_SZ);
@@ -790,7 +790,7 @@ struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
 		 * We can only recover at the end of the log, so check that the
 		 * next log LEB is empty or out of date.
 		 */
-		sleb = ubifs_scan(c, next_lnum, 0, sbuf);
+		sleb = ubifs_scan(c, next_lnum, 0, sbuf, 0);
 		if (IS_ERR(sleb))
 			return sleb;
 		if (sleb->nodes_cnt) {
diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
index 2970500f32d..5c2d6d759a3 100644
--- a/fs/ubifs/replay.c
+++ b/fs/ubifs/replay.c
@@ -506,7 +506,7 @@ static int replay_bud(struct ubifs_info *c, int lnum, int offs, int jhead,
 	if (c->need_recovery)
 		sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf, jhead != GCHD);
 	else
-		sleb = ubifs_scan(c, lnum, offs, c->sbuf);
+		sleb = ubifs_scan(c, lnum, offs, c->sbuf, 0);
 	if (IS_ERR(sleb))
 		return PTR_ERR(sleb);
 
@@ -836,8 +836,8 @@ static int replay_log_leb(struct ubifs_info *c, int lnum, int offs, void *sbuf)
 	const struct ubifs_cs_node *node;
 
 	dbg_mnt("replay log LEB %d:%d", lnum, offs);
-	sleb = ubifs_scan(c, lnum, offs, sbuf);
-	if (IS_ERR(sleb) ) {
+	sleb = ubifs_scan(c, lnum, offs, sbuf, c->need_recovery);
+	if (IS_ERR(sleb)) {
 		if (PTR_ERR(sleb) != -EUCLEAN || !c->need_recovery)
 			return PTR_ERR(sleb);
 		sleb = ubifs_recover_log_leb(c, lnum, offs, sbuf);
diff --git a/fs/ubifs/scan.c b/fs/ubifs/scan.c
index 892ebfee4fe..96c52538419 100644
--- a/fs/ubifs/scan.c
+++ b/fs/ubifs/scan.c
@@ -108,10 +108,9 @@ int ubifs_scan_a_node(const struct ubifs_info *c, void *buf, int len, int lnum,
 
 		/* Make the node pads to 8-byte boundary */
 		if ((node_len + pad_len) & 7) {
-			if (!quiet) {
+			if (!quiet)
 				dbg_err("bad padding length %d - %d",
 					offs, offs + node_len + pad_len);
-			}
 			return SCANNED_A_BAD_PAD_NODE;
 		}
 
@@ -253,15 +252,19 @@ void ubifs_scanned_corruption(const struct ubifs_info *c, int lnum, int offs,
  * @c: UBIFS file-system description object
  * @lnum: logical eraseblock number
  * @offs: offset to start at (usually zero)
- * @sbuf: scan buffer (must be c->leb_size)
+ * @sbuf: scan buffer (must be of @c->leb_size bytes in size)
+ * @quiet: print no messages
  *
  * This function scans LEB number @lnum and returns complete information about
  * its contents. Returns the scaned information in case of success and,
  * %-EUCLEAN if the LEB neads recovery, and other negative error codes in case
  * of failure.
+ *
+ * If @quiet is non-zero, this function does not print large and scary
+ * error messages and flash dumps in case of errors.
  */
 struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
-				  int offs, void *sbuf)
+				  int offs, void *sbuf, int quiet)
 {
 	void *buf = sbuf + offs;
 	int err, len = c->leb_size - offs;
@@ -280,7 +283,7 @@ struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
 
 		cond_resched();
 
-		ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 0);
+		ret = ubifs_scan_a_node(c, buf, len, lnum, offs, quiet);
 		if (ret > 0) {
 			/* Padding bytes or a valid padding node */
 			offs += ret;
@@ -320,7 +323,9 @@ struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
 	}
 
 	if (offs % c->min_io_size) {
-		ubifs_err("empty space starts at non-aligned offset %d", offs);
+		if (!quiet)
+			ubifs_err("empty space starts at non-aligned offset %d",
+				  offs);
 		goto corrupted;;
 	}
 
@@ -331,18 +336,25 @@ struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
 			break;
 	for (; len; offs++, buf++, len--)
 		if (*(uint8_t *)buf != 0xff) {
-			ubifs_err("corrupt empty space at LEB %d:%d",
-				  lnum, offs);
+			if (!quiet)
+				ubifs_err("corrupt empty space at LEB %d:%d",
+					  lnum, offs);
 			goto corrupted;
 		}
 
 	return sleb;
 
 corrupted:
-	ubifs_scanned_corruption(c, lnum, offs, buf);
+	if (!quiet) {
+		ubifs_scanned_corruption(c, lnum, offs, buf);
+		ubifs_err("LEB %d scanning failed", lnum);
+	}
 	err = -EUCLEAN;
+	ubifs_scan_destroy(sleb);
+	return ERR_PTR(err);
+
 error:
-	ubifs_err("LEB %d scanning failed", lnum);
+	ubifs_err("LEB %d scanning failed, error %d", lnum, err);
 	ubifs_scan_destroy(sleb);
 	return ERR_PTR(err);
 }
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 26d2e0d8046..333e181ee98 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -36,7 +36,6 @@
 #include <linux/mount.h>
 #include <linux/math64.h>
 #include <linux/writeback.h>
-#include <linux/smp_lock.h>
 #include "ubifs.h"
 
 /*
@@ -318,6 +317,8 @@ static int ubifs_write_inode(struct inode *inode, int wait)
 		if (err)
 			ubifs_err("can't write inode %lu, error %d",
 				  inode->i_ino, err);
+		else
+			err = dbg_check_inode_size(c, inode, ui->ui_size);
 	}
 
 	ui->dirty = 0;
@@ -438,12 +439,6 @@ static int ubifs_sync_fs(struct super_block *sb, int wait)
 {
 	int i, err;
 	struct ubifs_info *c = sb->s_fs_info;
-	struct writeback_control wbc = {
-		.sync_mode   = WB_SYNC_ALL,
-		.range_start = 0,
-		.range_end   = LLONG_MAX,
-		.nr_to_write = LONG_MAX,
-	};
 
 	/*
 	 * Zero @wait is just an advisory thing to help the file system shove
@@ -454,17 +449,6 @@ static int ubifs_sync_fs(struct super_block *sb, int wait)
 		return 0;
 
 	/*
-	 * 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.
-	 */
-	generic_sync_sb_inodes(sb, &wbc);
-
-	/*
 	 * Synchronize write buffers, because 'ubifs_run_commit()' does not
 	 * do this if it waits for an already running commit.
 	 */
@@ -474,6 +458,13 @@ static int ubifs_sync_fs(struct super_block *sb, int wait)
 			return err;
 	}
 
+	/*
+	 * 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.
+	 */
 	err = ubifs_run_commit(c);
 	if (err)
 		return err;
@@ -1726,8 +1717,6 @@ static void ubifs_put_super(struct super_block *sb)
 	ubifs_msg("un-mount UBI device %d, volume %d", c->vi.ubi_num,
 		  c->vi.vol_id);
 
-	lock_kernel();
-
 	/*
 	 * The following asserts are only valid if there has not been a failure
 	 * of the media. For example, there will be dirty inodes if we failed
@@ -1792,8 +1781,6 @@ static void ubifs_put_super(struct super_block *sb)
 	ubi_close_volume(c->ubi);
 	mutex_unlock(&c->umount_mutex);
 	kfree(c);
-
-	unlock_kernel();
 }
 
 static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data)
@@ -1809,22 +1796,17 @@ static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data)
 		return err;
 	}
 
-	lock_kernel();
 	if ((sb->s_flags & MS_RDONLY) && !(*flags & MS_RDONLY)) {
 		if (c->ro_media) {
 			ubifs_msg("cannot re-mount due to prior errors");
-			unlock_kernel();
 			return -EROFS;
 		}
 		err = ubifs_remount_rw(c);
-		if (err) {
-			unlock_kernel();
+		if (err)
 			return err;
-		}
 	} else if (!(sb->s_flags & MS_RDONLY) && (*flags & MS_RDONLY)) {
 		if (c->ro_media) {
 			ubifs_msg("cannot re-mount due to prior errors");
-			unlock_kernel();
 			return -EROFS;
 		}
 		ubifs_remount_ro(c);
@@ -1839,7 +1821,6 @@ static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data)
 	}
 
 	ubifs_assert(c->lst.taken_empty_lebs > 0);
-	unlock_kernel();
 	return 0;
 }
 
@@ -1971,6 +1952,7 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 	 *
 	 * Read-ahead will be disabled because @c->bdi.ra_pages is 0.
 	 */
+	c->bdi.name = "ubifs",
 	c->bdi.capabilities = BDI_CAP_MAP_COPY;
 	c->bdi.unplug_io_fn = default_unplug_io_fn;
 	err  = bdi_init(&c->bdi);
@@ -1985,6 +1967,7 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
 	if (err)
 		goto out_bdi;
 
+	sb->s_bdi = &c->bdi;
 	sb->s_fs_info = c;
 	sb->s_magic = UBIFS_SUPER_MAGIC;
 	sb->s_blocksize = UBIFS_BLOCK_SIZE;
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index f249f7b0d65..e5b1a7d00fa 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -1159,8 +1159,8 @@ static struct ubifs_znode *dirty_cow_bottom_up(struct ubifs_info *c,
  *   o exact match, i.e. the found zero-level znode contains key @key, then %1
  *     is returned and slot number of the matched branch is stored in @n;
  *   o not exact match, which means that zero-level znode does not contain
- *     @key, then %0 is returned and slot number of the closed branch is stored
- *     in  @n;
+ *     @key, then %0 is returned and slot number of the closest branch is stored
+ *     in @n;
  *   o @key is so small that it is even less than the lowest key of the
  *     leftmost zero-level node, then %0 is returned and %0 is stored in @n.
  *
@@ -1433,7 +1433,7 @@ static int maybe_leb_gced(struct ubifs_info *c, int lnum, int gc_seq1)
  * @lnum: LEB number is returned here
  * @offs: offset is returned here
  *
- * This function look up and reads node with key @key. The caller has to make
+ * This function looks up and reads node with key @key. The caller has to make
  * sure the @node buffer is large enough to fit the node. Returns zero in case
  * of success, %-ENOENT if the node was not found, and a negative error code in
  * case of failure. The node location can be returned in @lnum and @offs.
@@ -3268,3 +3268,73 @@ out_unlock:
 	mutex_unlock(&c->tnc_mutex);
 	return err;
 }
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+
+/**
+ * dbg_check_inode_size - check if inode size is correct.
+ * @c: UBIFS file-system description object
+ * @inum: inode number
+ * @size: inode size
+ *
+ * This function makes sure that the inode size (@size) is correct and it does
+ * not have any pages beyond @size. Returns zero if the inode is OK, %-EINVAL
+ * if it has a data page beyond @size, and other negative error code in case of
+ * other errors.
+ */
+int dbg_check_inode_size(struct ubifs_info *c, const struct inode *inode,
+			 loff_t size)
+{
+	int err, n;
+	union ubifs_key from_key, to_key, *key;
+	struct ubifs_znode *znode;
+	unsigned int block;
+
+	if (!S_ISREG(inode->i_mode))
+		return 0;
+	if (!(ubifs_chk_flags & UBIFS_CHK_GEN))
+		return 0;
+
+	block = (size + UBIFS_BLOCK_SIZE - 1) >> UBIFS_BLOCK_SHIFT;
+	data_key_init(c, &from_key, inode->i_ino, block);
+	highest_data_key(c, &to_key, inode->i_ino);
+
+	mutex_lock(&c->tnc_mutex);
+	err = ubifs_lookup_level0(c, &from_key, &znode, &n);
+	if (err < 0)
+		goto out_unlock;
+
+	if (err) {
+		err = -EINVAL;
+		key = &from_key;
+		goto out_dump;
+	}
+
+	err = tnc_next(c, &znode, &n);
+	if (err == -ENOENT) {
+		err = 0;
+		goto out_unlock;
+	}
+	if (err < 0)
+		goto out_unlock;
+
+	ubifs_assert(err == 0);
+	key = &znode->zbranch[n].key;
+	if (!key_in_range(c, key, &from_key, &to_key))
+		goto out_unlock;
+
+out_dump:
+	block = key_block(c, key);
+	ubifs_err("inode %lu has size %lld, but there are data at offset %lld "
+		  "(data key %s)", (unsigned long)inode->i_ino, size,
+		  ((loff_t)block) << UBIFS_BLOCK_SHIFT, DBGKEY(key));
+	dbg_dump_inode(c, inode);
+	dbg_dump_stack();
+	err = -EINVAL;
+
+out_unlock:
+	mutex_unlock(&c->tnc_mutex);
+	return err;
+}
+
+#endif /* CONFIG_UBIFS_FS_DEBUG */
diff --git a/fs/ubifs/tnc_commit.c b/fs/ubifs/tnc_commit.c
index fde8d127c76..53288e5d604 100644
--- a/fs/ubifs/tnc_commit.c
+++ b/fs/ubifs/tnc_commit.c
@@ -245,7 +245,7 @@ static int layout_leb_in_gaps(struct ubifs_info *c, int *p)
 	 * it is more comprehensive and less efficient than is needed for this
 	 * purpose.
 	 */
-	sleb = ubifs_scan(c, lnum, 0, c->ileb_buf);
+	sleb = ubifs_scan(c, lnum, 0, c->ileb_buf, 0);
 	c->ileb_len = 0;
 	if (IS_ERR(sleb))
 		return PTR_ERR(sleb);
diff --git a/fs/ubifs/ubifs-media.h b/fs/ubifs/ubifs-media.h
index 3eee07e0c49..191ca7863fe 100644
--- a/fs/ubifs/ubifs-media.h
+++ b/fs/ubifs/ubifs-media.h
@@ -135,6 +135,13 @@
 /* The key is always at the same position in all keyed nodes */
 #define UBIFS_KEY_OFFSET offsetof(struct ubifs_ino_node, key)
 
+/* Garbage collector journal head number */
+#define UBIFS_GC_HEAD   0
+/* Base journal head number */
+#define UBIFS_BASE_HEAD 1
+/* Data journal head number */
+#define UBIFS_DATA_HEAD 2
+
 /*
  * LEB Properties Tree node types.
  *
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index a2934909442..b2d976366a4 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -105,12 +105,10 @@
 /* Number of non-data journal heads */
 #define NONDATA_JHEADS_CNT 2
 
-/* Garbage collector head */
-#define GCHD   0
-/* Base journal head number */
-#define BASEHD 1
-/* First "general purpose" journal head */
-#define DATAHD 2
+/* Shorter names for journal head numbers for internal usage */
+#define GCHD   UBIFS_GC_HEAD
+#define BASEHD UBIFS_BASE_HEAD
+#define DATAHD UBIFS_DATA_HEAD
 
 /* 'No change' value for 'ubifs_change_lp()' */
 #define LPROPS_NC 0x80000001
@@ -1451,7 +1449,7 @@ int ubifs_sync_wbufs_by_inode(struct ubifs_info *c, struct inode *inode);
 
 /* scan.c */
 struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
-				  int offs, void *sbuf);
+				  int offs, void *sbuf, int quiet);
 void ubifs_scan_destroy(struct ubifs_scan_leb *sleb);
 int ubifs_scan_a_node(const struct ubifs_info *c, void *buf, int len, int lnum,
 		      int offs, int quiet);
@@ -1676,6 +1674,7 @@ const struct ubifs_lprops *ubifs_fast_find_free(struct ubifs_info *c);
 const struct ubifs_lprops *ubifs_fast_find_empty(struct ubifs_info *c);
 const struct ubifs_lprops *ubifs_fast_find_freeable(struct ubifs_info *c);
 const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c);
+int ubifs_calc_dark(const struct ubifs_info *c, int spc);
 
 /* file.c */
 int ubifs_fsync(struct file *file, struct dentry *dentry, int datasync);
diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c
index adafcf55653..195830f4756 100644
--- a/fs/ubifs/xattr.c
+++ b/fs/ubifs/xattr.c
@@ -78,9 +78,9 @@ enum {
 	SECURITY_XATTR,
 };
 
-static struct inode_operations none_inode_operations;
-static struct address_space_operations none_address_operations;
-static struct file_operations none_file_operations;
+static const struct inode_operations none_inode_operations;
+static const struct address_space_operations none_address_operations;
+static const struct file_operations none_file_operations;
 
 /**
  * create_xattr - create an extended attribute.
diff --git a/fs/udf/directory.c b/fs/udf/directory.c
index 1d2c570704c..2ffdb6733af 100644
--- a/fs/udf/directory.c
+++ b/fs/udf/directory.c
@@ -18,59 +18,6 @@
 #include <linux/string.h>
 #include <linux/buffer_head.h>
 
-#if 0
-static uint8_t *udf_filead_read(struct inode *dir, uint8_t *tmpad,
-				uint8_t ad_size, struct kernel_lb_addr fe_loc,
-				int *pos, int *offset, struct buffer_head **bh,
-				int *error)
-{
-	int loffset = *offset;
-	int block;
-	uint8_t *ad;
-	int remainder;
-
-	*error = 0;
-
-	ad = (uint8_t *)(*bh)->b_data + *offset;
-	*offset += ad_size;
-
-	if (!ad) {
-		brelse(*bh);
-		*error = 1;
-		return NULL;
-	}
-
-	if (*offset == dir->i_sb->s_blocksize) {
-		brelse(*bh);
-		block = udf_get_lb_pblock(dir->i_sb, fe_loc, ++*pos);
-		if (!block)
-			return NULL;
-		*bh = udf_tread(dir->i_sb, block);
-		if (!*bh)
-			return NULL;
-	} else if (*offset > dir->i_sb->s_blocksize) {
-		ad = tmpad;
-
-		remainder = dir->i_sb->s_blocksize - loffset;
-		memcpy((uint8_t *)ad, (*bh)->b_data + loffset, remainder);
-
-		brelse(*bh);
-		block = udf_get_lb_pblock(dir->i_sb, fe_loc, ++*pos);
-		if (!block)
-			return NULL;
-		(*bh) = udf_tread(dir->i_sb, block);
-		if (!*bh)
-			return NULL;
-
-		memcpy((uint8_t *)ad + remainder, (*bh)->b_data,
-			ad_size - remainder);
-		*offset = ad_size - remainder;
-	}
-
-	return ad;
-}
-#endif
-
 struct fileIdentDesc *udf_fileident_read(struct inode *dir, loff_t *nf_pos,
 					 struct udf_fileident_bh *fibh,
 					 struct fileIdentDesc *cfi,
@@ -248,39 +195,6 @@ struct fileIdentDesc *udf_get_fileident(void *buffer, int bufsize, int *offset)
 	return fi;
 }
 
-#if 0
-static struct extent_ad *udf_get_fileextent(void *buffer, int bufsize, int *offset)
-{
-	struct extent_ad *ext;
-	struct fileEntry *fe;
-	uint8_t *ptr;
-
-	if ((!buffer) || (!offset)) {
-		printk(KERN_ERR "udf: udf_get_fileextent() invalidparms\n");
-		return NULL;
-	}
-
-	fe = (struct fileEntry *)buffer;
-
-	if (fe->descTag.tagIdent != cpu_to_le16(TAG_IDENT_FE)) {
-		udf_debug("0x%x != TAG_IDENT_FE\n",
-			  le16_to_cpu(fe->descTag.tagIdent));
-		return NULL;
-	}
-
-	ptr = (uint8_t *)(fe->extendedAttr) +
-		le32_to_cpu(fe->lengthExtendedAttr);
-
-	if ((*offset > 0) && (*offset < le32_to_cpu(fe->lengthAllocDescs)))
-		ptr += *offset;
-
-	ext = (struct extent_ad *)ptr;
-
-	*offset = *offset + sizeof(struct extent_ad);
-	return ext;
-}
-#endif
-
 struct short_ad *udf_get_fileshortad(uint8_t *ptr, int maxoffset, uint32_t *offset,
 			      int inc)
 {
diff --git a/fs/udf/file.c b/fs/udf/file.c
index 7464305382b..b80cbd78833 100644
--- a/fs/udf/file.c
+++ b/fs/udf/file.c
@@ -193,9 +193,11 @@ int udf_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 static int udf_release_file(struct inode *inode, struct file *filp)
 {
 	if (filp->f_mode & FMODE_WRITE) {
+		mutex_lock(&inode->i_mutex);
 		lock_kernel();
 		udf_discard_prealloc(inode);
 		unlock_kernel();
+		mutex_unlock(&inode->i_mutex);
 	}
 	return 0;
 }
diff --git a/fs/udf/inode.c b/fs/udf/inode.c
index e7533f78563..6d24c2c63f9 100644
--- a/fs/udf/inode.c
+++ b/fs/udf/inode.c
@@ -90,19 +90,16 @@ no_delete:
 }
 
 /*
- * If we are going to release inode from memory, we discard preallocation and
- * truncate last inode extent to proper length. We could use drop_inode() but
- * it's called under inode_lock and thus we cannot mark inode dirty there.  We
- * use clear_inode() but we have to make sure to write inode as it's not written
- * automatically.
+ * If we are going to release inode from memory, we truncate last inode extent
+ * to proper length. We could use drop_inode() but it's called under inode_lock
+ * and thus we cannot mark inode dirty there.  We use clear_inode() but we have
+ * to make sure to write inode as it's not written automatically.
  */
 void udf_clear_inode(struct inode *inode)
 {
 	struct udf_inode_info *iinfo;
 	if (!(inode->i_sb->s_flags & MS_RDONLY)) {
 		lock_kernel();
-		/* Discard preallocation for directories, symlinks, etc. */
-		udf_discard_prealloc(inode);
 		udf_truncate_tail_extent(inode);
 		unlock_kernel();
 		write_inode_now(inode, 0);
@@ -664,8 +661,12 @@ static struct buffer_head *inode_getblk(struct inode *inode, sector_t block,
 	udf_split_extents(inode, &c, offset, newblocknum, laarr, &endnum);
 
 #ifdef UDF_PREALLOCATE
-	/* preallocate blocks */
-	udf_prealloc_extents(inode, c, lastblock, laarr, &endnum);
+	/* We preallocate blocks only for regular files. It also makes sense
+	 * for directories but there's a problem when to drop the
+	 * preallocation. We might use some delayed work for that but I feel
+	 * it's overengineering for a filesystem like UDF. */
+	if (S_ISREG(inode->i_mode))
+		udf_prealloc_extents(inode, c, lastblock, laarr, &endnum);
 #endif
 
 	/* merge any continuous blocks in laarr */
diff --git a/fs/udf/lowlevel.c b/fs/udf/lowlevel.c
index 1b88fd5df05..43e24a3b8e1 100644
--- a/fs/udf/lowlevel.c
+++ b/fs/udf/lowlevel.c
@@ -36,14 +36,10 @@ unsigned int udf_get_last_session(struct super_block *sb)
 	ms_info.addr_format = CDROM_LBA;
 	i = ioctl_by_bdev(bdev, CDROMMULTISESSION, (unsigned long)&ms_info);
 
-#define WE_OBEY_THE_WRITTEN_STANDARDS 1
-
 	if (i == 0) {
 		udf_debug("XA disk: %s, vol_desc_start=%d\n",
 			  (ms_info.xa_flag ? "yes" : "no"), ms_info.addr.lba);
-#if WE_OBEY_THE_WRITTEN_STANDARDS
 		if (ms_info.xa_flag) /* necessary for a valid ms_info.addr */
-#endif
 			vol_desc_start = ms_info.addr.lba;
 	} else {
 		udf_debug("CDROMMULTISESSION not supported: rc=%d\n", i);
diff --git a/fs/udf/namei.c b/fs/udf/namei.c
index 6a29fa34c47..21dad8c608f 100644
--- a/fs/udf/namei.c
+++ b/fs/udf/namei.c
@@ -943,7 +943,6 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry,
 		pc->componentType = 1;
 		pc->lengthComponentIdent = 0;
 		pc->componentFileVersionNum = 0;
-		pc += sizeof(struct pathComponent);
 		elen += sizeof(struct pathComponent);
 	}
 
diff --git a/fs/xattr.c b/fs/xattr.c
index 1c3d0af59dd..6d4f6d3449f 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
@@ -66,22 +66,28 @@ xattr_permission(struct inode *inode, const char *name, int mask)
 	return inode_permission(inode, mask);
 }
 
-int
-vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
-		size_t size, int flags)
+/**
+ *  __vfs_setxattr_noperm - perform setxattr operation without performing
+ *  permission checks.
+ *
+ *  @dentry - object to perform setxattr on
+ *  @name - xattr name to set
+ *  @value - value to set @name to
+ *  @size - size of @value
+ *  @flags - flags to pass into filesystem operations
+ *
+ *  returns the result of the internal setxattr or setsecurity operations.
+ *
+ *  This function requires the caller to lock the inode's i_mutex before it
+ *  is executed. It also assumes that the caller will make the appropriate
+ *  permission checks.
+ */
+int __vfs_setxattr_noperm(struct dentry *dentry, const char *name,
+		const void *value, size_t size, int flags)
 {
 	struct inode *inode = dentry->d_inode;
-	int error;
-
-	error = xattr_permission(inode, name, MAY_WRITE);
-	if (error)
-		return error;
+	int error = -EOPNOTSUPP;
 
-	mutex_lock(&inode->i_mutex);
-	error = security_inode_setxattr(dentry, name, value, size, flags);
-	if (error)
-		goto out;
-	error = -EOPNOTSUPP;
 	if (inode->i_op->setxattr) {
 		error = inode->i_op->setxattr(dentry, name, value, size, flags);
 		if (!error) {
@@ -97,6 +103,29 @@ vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
 		if (!error)
 			fsnotify_xattr(dentry);
 	}
+
+	return error;
+}
+
+
+int
+vfs_setxattr(struct dentry *dentry, const char *name, const void *value,
+		size_t size, int flags)
+{
+	struct inode *inode = dentry->d_inode;
+	int error;
+
+	error = xattr_permission(inode, name, MAY_WRITE);
+	if (error)
+		return error;
+
+	mutex_lock(&inode->i_mutex);
+	error = security_inode_setxattr(dentry, name, value, size, flags);
+	if (error)
+		goto out;
+
+	error = __vfs_setxattr_noperm(dentry, name, value, size, flags);
+
 out:
 	mutex_unlock(&inode->i_mutex);
 	return error;
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index aecf2519db7..d5e5559e31d 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -216,7 +216,6 @@ xfs_setfilesize(
 	if (ip->i_d.di_size < isize) {
 		ip->i_d.di_size = isize;
 		ip->i_update_core = 1;
-		ip->i_update_size = 1;
 		xfs_mark_inode_dirty_sync(ip);
 	}
 
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index 0542fd50764..988d8f87bc0 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -172,12 +172,21 @@ xfs_file_release(
  */
 STATIC int
 xfs_file_fsync(
-	struct file	*filp,
-	struct dentry	*dentry,
-	int		datasync)
+	struct file		*file,
+	struct dentry		*dentry,
+	int			datasync)
 {
-	xfs_iflags_clear(XFS_I(dentry->d_inode), XFS_ITRUNCATED);
-	return -xfs_fsync(XFS_I(dentry->d_inode));
+	struct inode		*inode = dentry->d_inode;
+	struct xfs_inode	*ip = XFS_I(inode);
+	int			error;
+
+	/* capture size updates in I/O completion before writing the inode. */
+	error = filemap_fdatawait(inode->i_mapping);
+	if (error)
+		return error;
+
+	xfs_iflags_clear(ip, XFS_ITRUNCATED);
+	return -xfs_fsync(ip);
 }
 
 STATIC int
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index 8070b34cc28..da0159d99f8 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -43,7 +43,6 @@
 #include "xfs_error.h"
 #include "xfs_itable.h"
 #include "xfs_rw.h"
-#include "xfs_acl.h"
 #include "xfs_attr.h"
 #include "xfs_buf_item.h"
 #include "xfs_utils.h"
@@ -485,14 +484,6 @@ xfs_vn_put_link(
 }
 
 STATIC int
-xfs_vn_permission(
-	struct inode		*inode,
-	int			mask)
-{
-	return generic_permission(inode, mask, xfs_check_acl);
-}
-
-STATIC int
 xfs_vn_getattr(
 	struct vfsmount		*mnt,
 	struct dentry		*dentry,
@@ -696,7 +687,7 @@ xfs_vn_fiemap(
 }
 
 static const struct inode_operations xfs_inode_operations = {
-	.permission		= xfs_vn_permission,
+	.check_acl		= xfs_check_acl,
 	.truncate		= xfs_vn_truncate,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
@@ -724,7 +715,7 @@ static const struct inode_operations xfs_dir_inode_operations = {
 	.rmdir			= xfs_vn_unlink,
 	.mknod			= xfs_vn_mknod,
 	.rename			= xfs_vn_rename,
-	.permission		= xfs_vn_permission,
+	.check_acl		= xfs_check_acl,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
@@ -749,7 +740,7 @@ static const struct inode_operations xfs_dir_ci_inode_operations = {
 	.rmdir			= xfs_vn_unlink,
 	.mknod			= xfs_vn_mknod,
 	.rename			= xfs_vn_rename,
-	.permission		= xfs_vn_permission,
+	.check_acl		= xfs_check_acl,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
@@ -762,7 +753,7 @@ static const struct inode_operations xfs_symlink_inode_operations = {
 	.readlink		= generic_readlink,
 	.follow_link		= xfs_vn_follow_link,
 	.put_link		= xfs_vn_put_link,
-	.permission		= xfs_vn_permission,
+	.check_acl		= xfs_check_acl,
 	.getattr		= xfs_vn_getattr,
 	.setattr		= xfs_vn_setattr,
 	.setxattr		= generic_setxattr,
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
index 7078974a6ee..49e4a6aea73 100644
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ b/fs/xfs/linux-2.6/xfs_lrw.c
@@ -812,18 +812,21 @@ write_retry:
 
 	/* Handle various SYNC-type writes */
 	if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
+		loff_t end = pos + ret - 1;
 		int error2;
 
 		xfs_iunlock(xip, iolock);
 		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
-		error2 = sync_page_range(inode, mapping, pos, ret);
+
+		error2 = filemap_write_and_wait_range(mapping, pos, end);
 		if (!error)
 			error = error2;
 		if (need_i_mutex)
 			mutex_lock(&inode->i_mutex);
 		xfs_ilock(xip, iolock);
-		error2 = xfs_write_sync_logforce(mp, xip);
+
+		error2 = xfs_fsync(xip);
 		if (!error)
 			error = error2;
 	}
diff --git a/fs/xfs/linux-2.6/xfs_quotaops.c b/fs/xfs/linux-2.6/xfs_quotaops.c
index cb6e2cca214..9e41f91aa26 100644
--- a/fs/xfs/linux-2.6/xfs_quotaops.c
+++ b/fs/xfs/linux-2.6/xfs_quotaops.c
@@ -150,7 +150,7 @@ xfs_fs_set_xquota(
 	return -xfs_qm_scall_setqlim(mp, id, xfs_quota_type(type), fdq);
 }
 
-struct quotactl_ops xfs_quotactl_operations = {
+const struct quotactl_ops xfs_quotactl_operations = {
 	.quota_sync		= xfs_fs_quota_sync,
 	.get_xstate		= xfs_fs_get_xstate,
 	.set_xstate		= xfs_fs_set_xstate,
diff --git a/fs/xfs/linux-2.6/xfs_stats.c b/fs/xfs/linux-2.6/xfs_stats.c
index c3526d445f6..76fdc586193 100644
--- a/fs/xfs/linux-2.6/xfs_stats.c
+++ b/fs/xfs/linux-2.6/xfs_stats.c
@@ -20,16 +20,9 @@
 
 DEFINE_PER_CPU(struct xfsstats, xfsstats);
 
-STATIC int
-xfs_read_xfsstats(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
+static int xfs_stat_proc_show(struct seq_file *m, void *v)
 {
-	int		c, i, j, len, val;
+	int		c, i, j, val;
 	__uint64_t	xs_xstrat_bytes = 0;
 	__uint64_t	xs_write_bytes = 0;
 	__uint64_t	xs_read_bytes = 0;
@@ -60,18 +53,18 @@ xfs_read_xfsstats(
 	};
 
 	/* Loop over all stats groups */
-	for (i=j=len = 0; i < ARRAY_SIZE(xstats); i++) {
-		len += sprintf(buffer + len, "%s", xstats[i].desc);
+	for (i=j = 0; i < ARRAY_SIZE(xstats); i++) {
+		seq_printf(m, "%s", xstats[i].desc);
 		/* inner loop does each group */
 		while (j < xstats[i].endpoint) {
 			val = 0;
 			/* sum over all cpus */
 			for_each_possible_cpu(c)
 				val += *(((__u32*)&per_cpu(xfsstats, c) + j));
-			len += sprintf(buffer + len, " %u", val);
+			seq_printf(m, " %u", val);
 			j++;
 		}
-		buffer[len++] = '\n';
+		seq_putc(m, '\n');
 	}
 	/* extra precision counters */
 	for_each_possible_cpu(i) {
@@ -80,36 +73,38 @@ xfs_read_xfsstats(
 		xs_read_bytes += per_cpu(xfsstats, i).xs_read_bytes;
 	}
 
-	len += sprintf(buffer + len, "xpc %Lu %Lu %Lu\n",
+	seq_printf(m, "xpc %Lu %Lu %Lu\n",
 			xs_xstrat_bytes, xs_write_bytes, xs_read_bytes);
-	len += sprintf(buffer + len, "debug %u\n",
+	seq_printf(m, "debug %u\n",
 #if defined(DEBUG)
 		1);
 #else
 		0);
 #endif
+	return 0;
+}
 
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
+static int xfs_stat_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xfs_stat_proc_show, NULL);
 }
 
+static const struct file_operations xfs_stat_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xfs_stat_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
 int
 xfs_init_procfs(void)
 {
 	if (!proc_mkdir("fs/xfs", NULL))
 		goto out;
 
-	if (!create_proc_read_entry("fs/xfs/stat", 0, NULL,
-			xfs_read_xfsstats, NULL))
+	if (!proc_create("fs/xfs/stat", 0, NULL,
+			 &xfs_stat_proc_fops))
 		goto out_remove_entry;
 	return 0;
 
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index a220d36f789..bdd41c8c342 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -67,7 +67,7 @@
 #include <linux/freezer.h>
 #include <linux/parser.h>
 
-static struct super_operations xfs_super_operations;
+static const struct super_operations xfs_super_operations;
 static kmem_zone_t *xfs_ioend_zone;
 mempool_t *xfs_ioend_pool;
 
@@ -579,15 +579,19 @@ xfs_showargs(
 	else if (mp->m_qflags & XFS_UQUOTA_ACCT)
 		seq_puts(m, "," MNTOPT_UQUOTANOENF);
 
-	if (mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_PRJQUOTA);
-	else if (mp->m_qflags & XFS_PQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_PQUOTANOENF);
-
-	if (mp->m_qflags & (XFS_GQUOTA_ACCT|XFS_OQUOTA_ENFD))
-		seq_puts(m, "," MNTOPT_GRPQUOTA);
-	else if (mp->m_qflags & XFS_GQUOTA_ACCT)
-		seq_puts(m, "," MNTOPT_GQUOTANOENF);
+	/* Either project or group quotas can be active, not both */
+
+	if (mp->m_qflags & XFS_PQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_PRJQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_PQUOTANOENF);
+	} else if (mp->m_qflags & XFS_GQUOTA_ACCT) {
+		if (mp->m_qflags & XFS_OQUOTA_ENFD)
+			seq_puts(m, "," MNTOPT_GRPQUOTA);
+		else
+			seq_puts(m, "," MNTOPT_GQUOTANOENF);
+	}
 
 	if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
 		seq_puts(m, "," MNTOPT_NOQUOTA);
@@ -687,7 +691,7 @@ xfs_barrier_test(
 	return error;
 }
 
-void
+STATIC void
 xfs_mountfs_check_barriers(xfs_mount_t *mp)
 {
 	int error;
@@ -1532,7 +1536,7 @@ xfs_fs_get_sb(
 			   mnt);
 }
 
-static struct super_operations xfs_super_operations = {
+static const struct super_operations xfs_super_operations = {
 	.alloc_inode		= xfs_fs_alloc_inode,
 	.destroy_inode		= xfs_fs_destroy_inode,
 	.write_inode		= xfs_fs_write_inode,
diff --git a/fs/xfs/linux-2.6/xfs_super.h b/fs/xfs/linux-2.6/xfs_super.h
index 5a2ea3a2178..18175ebd58e 100644
--- a/fs/xfs/linux-2.6/xfs_super.h
+++ b/fs/xfs/linux-2.6/xfs_super.h
@@ -93,7 +93,7 @@ extern void xfs_blkdev_issue_flush(struct xfs_buftarg *);
 
 extern const struct export_operations xfs_export_operations;
 extern struct xattr_handler *xfs_xattr_handlers[];
-extern struct quotactl_ops xfs_quotactl_operations;
+extern const struct quotactl_ops xfs_quotactl_operations;
 
 #define XFS_M(sb)		((struct xfs_mount *)((sb)->s_fs_info))
 
diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c
index 98ef624d9ba..320be6aea49 100644
--- a/fs/xfs/linux-2.6/xfs_sync.c
+++ b/fs/xfs/linux-2.6/xfs_sync.c
@@ -749,21 +749,6 @@ __xfs_inode_clear_reclaim_tag(
 			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 int
 xfs_reclaim_inode_now(
 	struct xfs_inode	*ip,
diff --git a/fs/xfs/linux-2.6/xfs_sync.h b/fs/xfs/linux-2.6/xfs_sync.h
index 59120602588..27920eb7a82 100644
--- a/fs/xfs/linux-2.6/xfs_sync.h
+++ b/fs/xfs/linux-2.6/xfs_sync.h
@@ -49,7 +49,6 @@ int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
 
 void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
 void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, 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);
 
diff --git a/fs/xfs/quota/xfs_qm_stats.c b/fs/xfs/quota/xfs_qm_stats.c
index 21b08c0396a..83e7ea3e25f 100644
--- a/fs/xfs/quota/xfs_qm_stats.c
+++ b/fs/xfs/quota/xfs_qm_stats.c
@@ -48,50 +48,34 @@
 
 struct xqmstats xqmstats;
 
-STATIC int
-xfs_qm_read_xfsquota(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
+static int xqm_proc_show(struct seq_file *m, void *v)
 {
-	int		len;
-
 	/* maximum; incore; ratio free to inuse; freelist */
-	len = sprintf(buffer, "%d\t%d\t%d\t%u\n",
+	seq_printf(m, "%d\t%d\t%d\t%u\n",
 			ndquot,
 			xfs_Gqm? atomic_read(&xfs_Gqm->qm_totaldquots) : 0,
 			xfs_Gqm? xfs_Gqm->qm_dqfree_ratio : 0,
 			xfs_Gqm? xfs_Gqm->qm_dqfreelist.qh_nelems : 0);
-
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
+	return 0;
 }
 
-STATIC int
-xfs_qm_read_stats(
-	char		*buffer,
-	char		**start,
-	off_t		offset,
-	int		count,
-	int		*eof,
-	void		*data)
+static int xqm_proc_open(struct inode *inode, struct file *file)
 {
-	int		len;
+	return single_open(file, xqm_proc_show, NULL);
+}
+
+static const struct file_operations xqm_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xqm_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
 
+static int xqmstat_proc_show(struct seq_file *m, void *v)
+{
 	/* quota performance statistics */
-	len = sprintf(buffer, "qm %u %u %u %u %u %u %u %u\n",
+	seq_printf(m, "qm %u %u %u %u %u %u %u %u\n",
 			xqmstats.xs_qm_dqreclaims,
 			xqmstats.xs_qm_dqreclaim_misses,
 			xqmstats.xs_qm_dquot_dups,
@@ -100,25 +84,27 @@ xfs_qm_read_stats(
 			xqmstats.xs_qm_dqwants,
 			xqmstats.xs_qm_dqshake_reclaims,
 			xqmstats.xs_qm_dqinact_reclaims);
+	return 0;
+}
 
-	if (offset >= len) {
-		*start = buffer;
-		*eof = 1;
-		return 0;
-	}
-	*start = buffer + offset;
-	if ((len -= offset) > count)
-		return count;
-	*eof = 1;
-
-	return len;
+static int xqmstat_proc_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, xqmstat_proc_show, NULL);
 }
 
+static const struct file_operations xqmstat_proc_fops = {
+	.owner		= THIS_MODULE,
+	.open		= xqmstat_proc_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
 void
 xfs_qm_init_procfs(void)
 {
-	create_proc_read_entry("fs/xfs/xqmstat", 0, NULL, xfs_qm_read_stats, NULL);
-	create_proc_read_entry("fs/xfs/xqm", 0, NULL, xfs_qm_read_xfsquota, NULL);
+	proc_create("fs/xfs/xqmstat", 0, NULL, &xqmstat_proc_fops);
+	proc_create("fs/xfs/xqm", 0, NULL, &xqm_proc_fops);
 }
 
 void
diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h
index f24b50b68d0..a5d54bf4931 100644
--- a/fs/xfs/xfs_ag.h
+++ b/fs/xfs/xfs_ag.h
@@ -198,6 +198,15 @@ typedef struct xfs_perag
 	xfs_agino_t	pagi_count;	/* number of allocated inodes */
 	int		pagb_count;	/* pagb slots in use */
 	xfs_perag_busy_t *pagb_list;	/* unstable blocks */
+
+	/*
+	 * Inode allocation search lookup optimisation.
+	 * If the pagino matches, the search for new inodes
+	 * doesn't need to search the near ones again straight away
+	 */
+	xfs_agino_t	pagl_pagino;
+	xfs_agino_t	pagl_leftrec;
+	xfs_agino_t	pagl_rightrec;
 #ifdef __KERNEL__
 	spinlock_t	pagb_lock;	/* lock for pagb_list */
 
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 8ee5b5a76a2..8971fb09d38 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -3713,7 +3713,7 @@ done:
  * entry (null if none).  Else, *lastxp will be set to the index
  * of the found entry; *gotp will contain the entry.
  */
-xfs_bmbt_rec_host_t *			/* pointer to found extent entry */
+STATIC xfs_bmbt_rec_host_t *		/* pointer to found extent entry */
 xfs_bmap_search_multi_extents(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fileoff_t	bno,		/* block number searched for */
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index 1b8ff9256bd..56f62d2edc3 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -392,17 +392,6 @@ xfs_bmap_count_blocks(
 	int			whichfork,
 	int			*count);
 
-/*
- * 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
- * entry (null if none).  Else, *lastxp will be set to the index
- * of the found entry; *gotp will contain the entry.
- */
-xfs_bmbt_rec_host_t *
-xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
-			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
-
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index 5c1ade06578..eb7b702d069 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -202,16 +202,6 @@ xfs_bmbt_get_state(
 				ext_flag);
 }
 
-/* Endian flipping versions of the bmbt extraction functions */
-void
-xfs_bmbt_disk_get_all(
-	xfs_bmbt_rec_t	*r,
-	xfs_bmbt_irec_t *s)
-{
-	__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
-				get_unaligned_be64(&r->l1), s);
-}
-
 /*
  * Extract the blockcount field from an on disk bmap extent record.
  */
@@ -816,6 +806,16 @@ xfs_bmbt_trace_key(
 	*l1 = 0;
 }
 
+/* Endian flipping versions of the bmbt extraction functions */
+STATIC void
+xfs_bmbt_disk_get_all(
+	xfs_bmbt_rec_t	*r,
+	xfs_bmbt_irec_t *s)
+{
+	__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
+				get_unaligned_be64(&r->l1), s);
+}
+
 STATIC void
 xfs_bmbt_trace_record(
 	struct xfs_btree_cur	*cur,
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index 0e8df007615..5549d495947 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -220,7 +220,6 @@ 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);
 extern xfs_exntst_t xfs_bmbt_get_state(xfs_bmbt_rec_host_t *r);
 
-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);
 
diff --git a/fs/xfs/xfs_btree.c b/fs/xfs/xfs_btree.c
index 26717388acf..52b5f14d0c3 100644
--- a/fs/xfs/xfs_btree.c
+++ b/fs/xfs/xfs_btree.c
@@ -646,46 +646,6 @@ xfs_btree_read_bufl(
 }
 
 /*
- * Get a buffer for the block, return it read in.
- * Short-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufs(
-	xfs_mount_t	*mp,		/* file system mount point */
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_agnumber_t	agno,		/* allocation group number */
-	xfs_agblock_t	agbno,		/* allocation group block number */
-	uint		lock,		/* lock flags for read_buf */
-	xfs_buf_t	**bpp,		/* buffer for agno/agbno */
-	int		refval)		/* ref count value for buffer */
-{
-	xfs_buf_t	*bp;		/* return value */
-	xfs_daddr_t	d;		/* real disk block address */
-	int		error;
-
-	ASSERT(agno != NULLAGNUMBER);
-	ASSERT(agbno != NULLAGBLOCK);
-	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
-					mp->m_bsize, lock, &bp))) {
-		return error;
-	}
-	ASSERT(!bp || !XFS_BUF_GETERROR(bp));
-	if (bp != NULL) {
-		switch (refval) {
-		case XFS_ALLOC_BTREE_REF:
-			XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, refval);
-			break;
-		case XFS_INO_BTREE_REF:
-			XFS_BUF_SET_VTYPE_REF(bp, B_FS_INOMAP, refval);
-			break;
-		}
-	}
-	*bpp = bp;
-	return 0;
-}
-
-/*
  * Read-ahead the block, don't wait for it, don't return a buffer.
  * Long-form addressing.
  */
@@ -2951,7 +2911,7 @@ error0:
  * inode we have to copy the single block it was pointing to into the
  * inode.
  */
-int
+STATIC int
 xfs_btree_kill_iroot(
 	struct xfs_btree_cur	*cur)
 {
diff --git a/fs/xfs/xfs_btree.h b/fs/xfs/xfs_btree.h
index 4f852b735b9..7fa07062bdd 100644
--- a/fs/xfs/xfs_btree.h
+++ b/fs/xfs/xfs_btree.h
@@ -379,20 +379,6 @@ xfs_btree_read_bufl(
 	int			refval);/* ref count value for buffer */
 
 /*
- * Get a buffer for the block, return it read in.
- * Short-form addressing.
- */
-int					/* error */
-xfs_btree_read_bufs(
-	struct xfs_mount	*mp,	/* file system mount point */
-	struct xfs_trans	*tp,	/* transaction pointer */
-	xfs_agnumber_t		agno,	/* allocation group number */
-	xfs_agblock_t		agbno,	/* allocation group block number */
-	uint			lock,	/* lock flags for read_buf */
-	struct xfs_buf		**bpp,	/* buffer for agno/agbno */
-	int			refval);/* ref count value for buffer */
-
-/*
  * Read-ahead the block, don't wait for it, don't return a buffer.
  * Long-form addressing.
  */
@@ -432,7 +418,6 @@ 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 *);
diff --git a/fs/xfs/xfs_fs.h b/fs/xfs/xfs_fs.h
index c4ea51b55dc..f52ac276277 100644
--- a/fs/xfs/xfs_fs.h
+++ b/fs/xfs/xfs_fs.h
@@ -117,7 +117,7 @@ struct getbmapx {
 #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 */
+/*	bmv_oflags values - returned for each non-header segment */
 #define BMV_OF_PREALLOC		0x1	/* segment = unwritten pre-allocation */
 #define BMV_OF_DELALLOC		0x2	/* segment = delayed allocation */
 #define BMV_OF_LAST		0x4	/* segment is the last in the file */
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 3120a3a5e20..ab64f3efb43 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -57,75 +57,35 @@ 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.
+ * Lookup a record by ino in the btree given by cur.
  */
 int					/* error */
-xfs_inobt_lookup_ge(
+xfs_inobt_lookup(
 	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 */
+	xfs_lookup_t		dir,	/* <=, >=, == */
 	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);
+	cur->bc_rec.i.ir_freecount = 0;
+	cur->bc_rec.i.ir_free = 0;
+	return xfs_btree_lookup(cur, dir, 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].
+ * Update the record referred to by cur to the value given.
  * 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 */
+	xfs_inobt_rec_incore_t	*irec)	/* btree record */
 {
 	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);
+	rec.inobt.ir_startino = cpu_to_be32(irec->ir_startino);
+	rec.inobt.ir_freecount = cpu_to_be32(irec->ir_freecount);
+	rec.inobt.ir_free = cpu_to_be64(irec->ir_free);
 	return xfs_btree_update(cur, &rec);
 }
 
@@ -135,9 +95,7 @@ xfs_inobt_update(
 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 */
+	xfs_inobt_rec_incore_t	*irec,	/* btree record */
 	int			*stat)	/* output: success/failure */
 {
 	union xfs_btree_rec	*rec;
@@ -145,14 +103,136 @@ xfs_inobt_get_rec(
 
 	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);
+		irec->ir_startino = be32_to_cpu(rec->inobt.ir_startino);
+		irec->ir_freecount = be32_to_cpu(rec->inobt.ir_freecount);
+		irec->ir_free = be64_to_cpu(rec->inobt.ir_free);
 	}
 	return error;
 }
 
 /*
+ * Verify that the number of free inodes in the AGI is correct.
+ */
+#ifdef DEBUG
+STATIC int
+xfs_check_agi_freecount(
+	struct xfs_btree_cur	*cur,
+	struct xfs_agi		*agi)
+{
+	if (cur->bc_nlevels == 1) {
+		xfs_inobt_rec_incore_t rec;
+		int		freecount = 0;
+		int		error;
+		int		i;
+
+		error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
+		if (error)
+			return error;
+
+		do {
+			error = xfs_inobt_get_rec(cur, &rec, &i);
+			if (error)
+				return error;
+
+			if (i) {
+				freecount += rec.ir_freecount;
+				error = xfs_btree_increment(cur, 0, &i);
+				if (error)
+					return error;
+			}
+		} while (i == 1);
+
+		if (!XFS_FORCED_SHUTDOWN(cur->bc_mp))
+			ASSERT(freecount == be32_to_cpu(agi->agi_freecount));
+	}
+	return 0;
+}
+#else
+#define xfs_check_agi_freecount(cur, agi)	0
+#endif
+
+/*
+ * Initialise a new set of inodes.
+ */
+STATIC void
+xfs_ialloc_inode_init(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	xfs_agblock_t		agbno,
+	xfs_agblock_t		length,
+	unsigned int		gen)
+{
+	struct xfs_buf		*fbuf;
+	struct xfs_dinode	*free;
+	int			blks_per_cluster, nbufs, ninodes;
+	int			version;
+	int			i, j;
+	xfs_daddr_t		d;
+
+	/*
+	 * Loop over the new block(s), filling in the inodes.
+	 * For small block sizes, manipulate the inodes in buffers
+	 * which are multiples of the blocks size.
+	 */
+	if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
+		blks_per_cluster = 1;
+		nbufs = length;
+		ninodes = mp->m_sb.sb_inopblock;
+	} else {
+		blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) /
+				   mp->m_sb.sb_blocksize;
+		nbufs = length / blks_per_cluster;
+		ninodes = blks_per_cluster * mp->m_sb.sb_inopblock;
+	}
+
+	/*
+	 * Figure out what version number to use in the inodes we create.
+	 * If the superblock version has caught up to the one that supports
+	 * the new inode format, then use the new inode version.  Otherwise
+	 * use the old version so that old kernels will continue to be
+	 * able to use the file system.
+	 */
+	if (xfs_sb_version_hasnlink(&mp->m_sb))
+		version = 2;
+	else
+		version = 1;
+
+	for (j = 0; j < nbufs; j++) {
+		/*
+		 * Get the block.
+		 */
+		d = XFS_AGB_TO_DADDR(mp, agno, agbno + (j * blks_per_cluster));
+		fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
+					 mp->m_bsize * blks_per_cluster,
+					 XFS_BUF_LOCK);
+		ASSERT(fbuf);
+		ASSERT(!XFS_BUF_GETERROR(fbuf));
+
+		/*
+		 * 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
+		 *	individual transactions causing a lot of log traffic.
+		 */
+		xfs_biozero(fbuf, 0, ninodes << mp->m_sb.sb_inodelog);
+		for (i = 0; i < ninodes; i++) {
+			int	ioffset = i << mp->m_sb.sb_inodelog;
+			uint	isize = sizeof(struct xfs_dinode);
+
+			free = xfs_make_iptr(mp, fbuf, i);
+			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_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
+		}
+		xfs_trans_inode_alloc_buf(tp, fbuf);
+	}
+}
+
+/*
  * Allocate new inodes in the allocation group specified by agbp.
  * Return 0 for success, else error code.
  */
@@ -164,24 +244,15 @@ xfs_ialloc_ag_alloc(
 {
 	xfs_agi_t	*agi;		/* allocation group header */
 	xfs_alloc_arg_t	args;		/* allocation argument structure */
-	int		blks_per_cluster;  /* fs blocks per inode cluster */
 	xfs_btree_cur_t	*cur;		/* inode btree cursor */
-	xfs_daddr_t	d;		/* disk addr of buffer */
 	xfs_agnumber_t	agno;
 	int		error;
-	xfs_buf_t	*fbuf;		/* new free inodes' buffer */
-	xfs_dinode_t	*free;		/* new free inode structure */
-	int		i;		/* inode counter */
-	int		j;		/* block counter */
-	int		nbufs;		/* num bufs of new inodes */
+	int		i;
 	xfs_agino_t	newino;		/* new first inode's number */
 	xfs_agino_t	newlen;		/* new number of inodes */
-	int		ninodes;	/* num inodes per buf */
 	xfs_agino_t	thisino;	/* current inode number, for loop */
-	int		version;	/* inode version number to use */
 	int		isaligned = 0;	/* inode allocation at stripe unit */
 					/* boundary */
-	unsigned int	gen;
 
 	args.tp = tp;
 	args.mp = tp->t_mountp;
@@ -202,12 +273,12 @@ xfs_ialloc_ag_alloc(
  	 */
 	agi = XFS_BUF_TO_AGI(agbp);
 	newino = be32_to_cpu(agi->agi_newino);
+	agno = be32_to_cpu(agi->agi_seqno);
 	args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
 			XFS_IALLOC_BLOCKS(args.mp);
 	if (likely(newino != NULLAGINO &&
 		  (args.agbno < be32_to_cpu(agi->agi_length)))) {
-		args.fsbno = XFS_AGB_TO_FSB(args.mp,
-				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
 		args.type = XFS_ALLOCTYPE_THIS_BNO;
 		args.mod = args.total = args.wasdel = args.isfl =
 			args.userdata = args.minalignslop = 0;
@@ -258,8 +329,7 @@ xfs_ialloc_ag_alloc(
 		 * For now, just allocate blocks up front.
 		 */
 		args.agbno = be32_to_cpu(agi->agi_root);
-		args.fsbno = XFS_AGB_TO_FSB(args.mp,
-				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
 		/*
 		 * Allocate a fixed-size extent of inodes.
 		 */
@@ -282,8 +352,7 @@ xfs_ialloc_ag_alloc(
 	if (isaligned && args.fsbno == NULLFSBLOCK) {
 		args.type = XFS_ALLOCTYPE_NEAR_BNO;
 		args.agbno = be32_to_cpu(agi->agi_root);
-		args.fsbno = XFS_AGB_TO_FSB(args.mp,
-				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
 		args.alignment = xfs_ialloc_cluster_alignment(&args);
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
@@ -294,85 +363,30 @@ xfs_ialloc_ag_alloc(
 		return 0;
 	}
 	ASSERT(args.len == args.minlen);
-	/*
-	 * Convert the results.
-	 */
-	newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0);
-	/*
-	 * Loop over the new block(s), filling in the inodes.
-	 * For small block sizes, manipulate the inodes in buffers
-	 * which are multiples of the blocks size.
-	 */
-	if (args.mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(args.mp)) {
-		blks_per_cluster = 1;
-		nbufs = (int)args.len;
-		ninodes = args.mp->m_sb.sb_inopblock;
-	} else {
-		blks_per_cluster = XFS_INODE_CLUSTER_SIZE(args.mp) /
-				   args.mp->m_sb.sb_blocksize;
-		nbufs = (int)args.len / blks_per_cluster;
-		ninodes = blks_per_cluster * args.mp->m_sb.sb_inopblock;
-	}
-	/*
-	 * Figure out what version number to use in the inodes we create.
-	 * If the superblock version has caught up to the one that supports
-	 * the new inode format, then use the new inode version.  Otherwise
-	 * use the old version so that old kernels will continue to be
-	 * able to use the file system.
-	 */
-	if (xfs_sb_version_hasnlink(&args.mp->m_sb))
-		version = 2;
-	else
-		version = 1;
 
 	/*
+	 * Stamp and write the inode buffers.
+	 *
 	 * Seed the new inode cluster with a random generation number. This
 	 * prevents short-term reuse of generation numbers if a chunk is
 	 * freed and then immediately reallocated. We use random numbers
 	 * rather than a linear progression to prevent the next generation
 	 * number from being easily guessable.
 	 */
-	gen = random32();
-	for (j = 0; j < nbufs; j++) {
-		/*
-		 * Get the block.
-		 */
-		d = XFS_AGB_TO_DADDR(args.mp, be32_to_cpu(agi->agi_seqno),
-				     args.agbno + (j * blks_per_cluster));
-		fbuf = xfs_trans_get_buf(tp, args.mp->m_ddev_targp, d,
-					 args.mp->m_bsize * blks_per_cluster,
-					 XFS_BUF_LOCK);
-		ASSERT(fbuf);
-		ASSERT(!XFS_BUF_GETERROR(fbuf));
+	xfs_ialloc_inode_init(args.mp, tp, agno, args.agbno, args.len,
+			      random32());
 
-		/*
-		 * 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 individual
-		 *      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_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_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
-		}
-		xfs_trans_inode_alloc_buf(tp, fbuf);
-	}
+	/*
+	 * Convert the results.
+	 */
+	newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0);
 	be32_add_cpu(&agi->agi_count, newlen);
 	be32_add_cpu(&agi->agi_freecount, newlen);
-	agno = be32_to_cpu(agi->agi_seqno);
 	down_read(&args.mp->m_peraglock);
 	args.mp->m_perag[agno].pagi_freecount += newlen;
 	up_read(&args.mp->m_peraglock);
 	agi->agi_newino = cpu_to_be32(newino);
+
 	/*
 	 * Insert records describing the new inode chunk into the btree.
 	 */
@@ -380,13 +394,17 @@ xfs_ialloc_ag_alloc(
 	for (thisino = newino;
 	     thisino < newino + newlen;
 	     thisino += XFS_INODES_PER_CHUNK) {
-		if ((error = xfs_inobt_lookup_eq(cur, thisino,
-				XFS_INODES_PER_CHUNK, XFS_INOBT_ALL_FREE, &i))) {
+		cur->bc_rec.i.ir_startino = thisino;
+		cur->bc_rec.i.ir_freecount = XFS_INODES_PER_CHUNK;
+		cur->bc_rec.i.ir_free = XFS_INOBT_ALL_FREE;
+		error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, &i);
+		if (error) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 			return error;
 		}
 		ASSERT(i == 0);
-		if ((error = xfs_btree_insert(cur, &i))) {
+		error = xfs_btree_insert(cur, &i);
+		if (error) {
 			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 			return error;
 		}
@@ -539,6 +557,62 @@ nextag:
 }
 
 /*
+ * Try to retrieve the next record to the left/right from the current one.
+ */
+STATIC int
+xfs_ialloc_next_rec(
+	struct xfs_btree_cur	*cur,
+	xfs_inobt_rec_incore_t	*rec,
+	int			*done,
+	int			left)
+{
+	int                     error;
+	int			i;
+
+	if (left)
+		error = xfs_btree_decrement(cur, 0, &i);
+	else
+		error = xfs_btree_increment(cur, 0, &i);
+
+	if (error)
+		return error;
+	*done = !i;
+	if (i) {
+		error = xfs_inobt_get_rec(cur, rec, &i);
+		if (error)
+			return error;
+		XFS_WANT_CORRUPTED_RETURN(i == 1);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_ialloc_get_rec(
+	struct xfs_btree_cur	*cur,
+	xfs_agino_t		agino,
+	xfs_inobt_rec_incore_t	*rec,
+	int			*done,
+	int			left)
+{
+	int                     error;
+	int			i;
+
+	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_EQ, &i);
+	if (error)
+		return error;
+	*done = !i;
+	if (i) {
+		error = xfs_inobt_get_rec(cur, rec, &i);
+		if (error)
+			return error;
+		XFS_WANT_CORRUPTED_RETURN(i == 1);
+	}
+
+	return 0;
+}
+
+/*
  * Visible inode allocation functions.
  */
 
@@ -592,8 +666,8 @@ xfs_dialloc(
 	int		j;		/* result code */
 	xfs_mount_t	*mp;		/* file system mount structure */
 	int		offset;		/* index of inode in chunk */
-	xfs_agino_t	pagino;		/* parent's a.g. relative inode # */
-	xfs_agnumber_t	pagno;		/* parent's allocation group number */
+	xfs_agino_t	pagino;		/* parent's AG relative inode # */
+	xfs_agnumber_t	pagno;		/* parent's AG number */
 	xfs_inobt_rec_incore_t rec;	/* inode allocation record */
 	xfs_agnumber_t	tagno;		/* testing allocation group number */
 	xfs_btree_cur_t	*tcur;		/* temp cursor */
@@ -716,6 +790,8 @@ nextag:
 	 */
 	agno = tagno;
 	*IO_agbp = NULL;
+
+ restart_pagno:
 	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.
@@ -723,220 +799,199 @@ nextag:
 	 */
 	if (!pagino)
 		pagino = be32_to_cpu(agi->agi_newino);
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int	freecount = 0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-		do {
-			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-					&rec.ir_freecount, &rec.ir_free, &i)))
-				goto error0;
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			freecount += rec.ir_freecount;
-			if ((error = xfs_btree_increment(cur, 0, &i)))
-				goto error0;
-		} while (i == 1);
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
 
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
 	/*
-	 * If in the same a.g. as the parent, try to get near the parent.
+	 * If in the same AG as the parent, try to get near the parent.
 	 */
 	if (pagno == agno) {
-		if ((error = xfs_inobt_lookup_le(cur, pagino, 0, 0, &i)))
+		xfs_perag_t	*pag = &mp->m_perag[agno];
+		int		doneleft;	/* done, to the left */
+		int		doneright;	/* done, to the right */
+		int		searchdistance = 10;
+
+		error = xfs_inobt_lookup(cur, pagino, XFS_LOOKUP_LE, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		error = xfs_inobt_get_rec(cur, &rec, &j);
+		if (error)
 			goto error0;
-		if (i != 0 &&
-		    (error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-			    &rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
-		    j == 1 &&
-		    rec.ir_freecount > 0) {
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+		if (rec.ir_freecount > 0) {
 			/*
 			 * Found a free inode in the same chunk
-			 * as parent, done.
+			 * as the parent, done.
 			 */
+			goto alloc_inode;
 		}
+
+
+		/*
+		 * In the same AG as parent, but parent's chunk is full.
+		 */
+
+		/* duplicate the cursor, search left & right simultaneously */
+		error = xfs_btree_dup_cursor(cur, &tcur);
+		if (error)
+			goto error0;
+
 		/*
-		 * In the same a.g. as parent, but parent's chunk is full.
+		 * Skip to last blocks looked up if same parent inode.
 		 */
-		else {
-			int	doneleft;	/* done, to the left */
-			int	doneright;	/* done, to the right */
+		if (pagino != NULLAGINO &&
+		    pag->pagl_pagino == pagino &&
+		    pag->pagl_leftrec != NULLAGINO &&
+		    pag->pagl_rightrec != NULLAGINO) {
+			error = xfs_ialloc_get_rec(tcur, pag->pagl_leftrec,
+						   &trec, &doneleft, 1);
+			if (error)
+				goto error1;
 
+			error = xfs_ialloc_get_rec(cur, pag->pagl_rightrec,
+						   &rec, &doneright, 0);
 			if (error)
-				goto error0;
-			ASSERT(i == 1);
-			ASSERT(j == 1);
-			/*
-			 * Duplicate the cursor, search left & right
-			 * simultaneously.
-			 */
-			if ((error = xfs_btree_dup_cursor(cur, &tcur)))
-				goto error0;
-			/*
-			 * Search left with tcur, back up 1 record.
-			 */
-			if ((error = xfs_btree_decrement(tcur, 0, &i)))
 				goto error1;
-			doneleft = !i;
-			if (!doneleft) {
-				if ((error = xfs_inobt_get_rec(tcur,
-						&trec.ir_startino,
-						&trec.ir_freecount,
-						&trec.ir_free, &i)))
-					goto error1;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
-			}
-			/*
-			 * Search right with cur, go forward 1 record.
-			 */
-			if ((error = xfs_btree_increment(cur, 0, &i)))
+		} else {
+			/* search left with tcur, back up 1 record */
+			error = xfs_ialloc_next_rec(tcur, &trec, &doneleft, 1);
+			if (error)
 				goto error1;
-			doneright = !i;
-			if (!doneright) {
-				if ((error = xfs_inobt_get_rec(cur,
-						&rec.ir_startino,
-						&rec.ir_freecount,
-						&rec.ir_free, &i)))
-					goto error1;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
-			}
-			/*
-			 * Loop until we find the closest inode chunk
-			 * with a free one.
-			 */
-			while (!doneleft || !doneright) {
-				int	useleft;  /* using left inode
-						     chunk this time */
 
+			/* search right with cur, go forward 1 record. */
+			error = xfs_ialloc_next_rec(cur, &rec, &doneright, 0);
+			if (error)
+				goto error1;
+		}
+
+		/*
+		 * Loop until we find an inode chunk with a free inode.
+		 */
+		while (!doneleft || !doneright) {
+			int	useleft;  /* using left inode chunk this time */
+
+			if (!--searchdistance) {
 				/*
-				 * Figure out which block is closer,
-				 * if both are valid.
-				 */
-				if (!doneleft && !doneright)
-					useleft =
-						pagino -
-						(trec.ir_startino +
-						 XFS_INODES_PER_CHUNK - 1) <
-						 rec.ir_startino - pagino;
-				else
-					useleft = !doneleft;
-				/*
-				 * If checking the left, does it have
-				 * free inodes?
-				 */
-				if (useleft && trec.ir_freecount) {
-					/*
-					 * Yes, set it up as the chunk to use.
-					 */
-					rec = trec;
-					xfs_btree_del_cursor(cur,
-						XFS_BTREE_NOERROR);
-					cur = tcur;
-					break;
-				}
-				/*
-				 * If checking the right, does it have
-				 * free inodes?
-				 */
-				if (!useleft && rec.ir_freecount) {
-					/*
-					 * Yes, it's already set up.
-					 */
-					xfs_btree_del_cursor(tcur,
-						XFS_BTREE_NOERROR);
-					break;
-				}
-				/*
-				 * If used the left, get another one
-				 * further left.
-				 */
-				if (useleft) {
-					if ((error = xfs_btree_decrement(tcur, 0,
-							&i)))
-						goto error1;
-					doneleft = !i;
-					if (!doneleft) {
-						if ((error = xfs_inobt_get_rec(
-							    tcur,
-							    &trec.ir_startino,
-							    &trec.ir_freecount,
-							    &trec.ir_free, &i)))
-							goto error1;
-						XFS_WANT_CORRUPTED_GOTO(i == 1,
-							error1);
-					}
-				}
-				/*
-				 * If used the right, get another one
-				 * further right.
+				 * Not in range - save last search
+				 * location and allocate a new inode
 				 */
-				else {
-					if ((error = xfs_btree_increment(cur, 0,
-							&i)))
-						goto error1;
-					doneright = !i;
-					if (!doneright) {
-						if ((error = xfs_inobt_get_rec(
-							    cur,
-							    &rec.ir_startino,
-							    &rec.ir_freecount,
-							    &rec.ir_free, &i)))
-							goto error1;
-						XFS_WANT_CORRUPTED_GOTO(i == 1,
-							error1);
-					}
-				}
+				pag->pagl_leftrec = trec.ir_startino;
+				pag->pagl_rightrec = rec.ir_startino;
+				pag->pagl_pagino = pagino;
+				goto newino;
+			}
+
+			/* figure out the closer block if both are valid. */
+			if (!doneleft && !doneright) {
+				useleft = pagino -
+				 (trec.ir_startino + XFS_INODES_PER_CHUNK - 1) <
+				  rec.ir_startino - pagino;
+			} else {
+				useleft = !doneleft;
 			}
-			ASSERT(!doneleft || !doneright);
+
+			/* free inodes to the left? */
+			if (useleft && trec.ir_freecount) {
+				rec = trec;
+				xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+				cur = tcur;
+
+				pag->pagl_leftrec = trec.ir_startino;
+				pag->pagl_rightrec = rec.ir_startino;
+				pag->pagl_pagino = pagino;
+				goto alloc_inode;
+			}
+
+			/* free inodes to the right? */
+			if (!useleft && rec.ir_freecount) {
+				xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+
+				pag->pagl_leftrec = trec.ir_startino;
+				pag->pagl_rightrec = rec.ir_startino;
+				pag->pagl_pagino = pagino;
+				goto alloc_inode;
+			}
+
+			/* get next record to check */
+			if (useleft) {
+				error = xfs_ialloc_next_rec(tcur, &trec,
+								 &doneleft, 1);
+			} else {
+				error = xfs_ialloc_next_rec(cur, &rec,
+								 &doneright, 0);
+			}
+			if (error)
+				goto error1;
 		}
+
+		/*
+		 * We've reached the end of the btree. because
+		 * we are only searching a small chunk of the
+		 * btree each search, there is obviously free
+		 * inodes closer to the parent inode than we
+		 * are now. restart the search again.
+		 */
+		pag->pagl_pagino = NULLAGINO;
+		pag->pagl_leftrec = NULLAGINO;
+		pag->pagl_rightrec = NULLAGINO;
+		xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
+		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+		goto restart_pagno;
 	}
+
 	/*
-	 * In a different a.g. from the parent.
+	 * In a different AG from the parent.
 	 * See if the most recently allocated block has any free.
 	 */
-	else if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
-		if ((error = xfs_inobt_lookup_eq(cur,
-				be32_to_cpu(agi->agi_newino), 0, 0, &i)))
+newino:
+	if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
+		error = xfs_inobt_lookup(cur, be32_to_cpu(agi->agi_newino),
+					 XFS_LOOKUP_EQ, &i);
+		if (error)
 			goto error0;
-		if (i == 1 &&
-		    (error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-			    &rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
-		    j == 1 &&
-		    rec.ir_freecount > 0) {
-			/*
-			 * The last chunk allocated in the group still has
-			 * a free inode.
-			 */
-		}
-		/*
-		 * None left in the last group, search the whole a.g.
-		 */
-		else {
+
+		if (i == 1) {
+			error = xfs_inobt_get_rec(cur, &rec, &j);
 			if (error)
 				goto error0;
-			if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-				goto error0;
-			ASSERT(i == 1);
-			for (;;) {
-				if ((error = xfs_inobt_get_rec(cur,
-						&rec.ir_startino,
-						&rec.ir_freecount, &rec.ir_free,
-						&i)))
-					goto error0;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-				if (rec.ir_freecount > 0)
-					break;
-				if ((error = xfs_btree_increment(cur, 0, &i)))
-					goto error0;
-				XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+			if (j == 1 && rec.ir_freecount > 0) {
+				/*
+				 * The last chunk allocated in the group
+				 * still has a free inode.
+				 */
+				goto alloc_inode;
 			}
 		}
 	}
+
+	/*
+	 * None left in the last group, search the whole AG
+	 */
+	error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
+	if (error)
+		goto error0;
+	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+
+	for (;;) {
+		error = xfs_inobt_get_rec(cur, &rec, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+		if (rec.ir_freecount > 0)
+			break;
+		error = xfs_btree_increment(cur, 0, &i);
+		if (error)
+			goto error0;
+		XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
+	}
+
+alloc_inode:
 	offset = xfs_ialloc_find_free(&rec.ir_free);
 	ASSERT(offset >= 0);
 	ASSERT(offset < XFS_INODES_PER_CHUNK);
@@ -945,33 +1000,19 @@ nextag:
 	ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
 	rec.ir_free &= ~XFS_INOBT_MASK(offset);
 	rec.ir_freecount--;
-	if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount,
-			rec.ir_free)))
+	error = xfs_inobt_update(cur, &rec);
+	if (error)
 		goto error0;
 	be32_add_cpu(&agi->agi_freecount, -1);
 	xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
 	down_read(&mp->m_peraglock);
 	mp->m_perag[tagno].pagi_freecount--;
 	up_read(&mp->m_peraglock);
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int	freecount = 0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		do {
-			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-					&rec.ir_freecount, &rec.ir_free, &i)))
-				goto error0;
-			XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-			freecount += rec.ir_freecount;
-			if ((error = xfs_btree_increment(cur, 0, &i)))
-				goto error0;
-		} while (i == 1);
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
+
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
 	*inop = ino;
@@ -1062,38 +1103,23 @@ xfs_difree(
 	 * Initialize the cursor.
 	 */
 	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int freecount = 0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		do {
-			if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
-					&rec.ir_freecount, &rec.ir_free, &i)))
-				goto error0;
-			if (i) {
-				freecount += rec.ir_freecount;
-				if ((error = xfs_btree_increment(cur, 0, &i)))
-					goto error0;
-			}
-		} while (i == 1);
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
+
 	/*
 	 * Look for the entry describing this inode.
 	 */
-	if ((error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i))) {
+	if ((error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i))) {
 		cmn_err(CE_WARN,
-			"xfs_difree: xfs_inobt_lookup_le returned()  an error %d on %s.  Returning error.",
+			"xfs_difree: xfs_inobt_lookup returned()  an error %d on %s.  Returning error.",
 			error, mp->m_fsname);
 		goto error0;
 	}
 	XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
-	if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino, &rec.ir_freecount,
-			&rec.ir_free, &i))) {
+	error = xfs_inobt_get_rec(cur, &rec, &i);
+	if (error) {
 		cmn_err(CE_WARN,
 			"xfs_difree: xfs_inobt_get_rec()  returned an error %d on %s.  Returning error.",
 			error, mp->m_fsname);
@@ -1148,12 +1174,14 @@ xfs_difree(
 	} else {
 		*delete = 0;
 
-		if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount, rec.ir_free))) {
+		error = xfs_inobt_update(cur, &rec);
+		if (error) {
 			cmn_err(CE_WARN,
-				"xfs_difree: xfs_inobt_update()  returned an error %d on %s.  Returning error.",
+	"xfs_difree: xfs_inobt_update returned an error %d on %s.",
 				error, mp->m_fsname);
 			goto error0;
 		}
+
 		/* 
 		 * Change the inode free counts and log the ag/sb changes.
 		 */
@@ -1165,28 +1193,10 @@ xfs_difree(
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1);
 	}
 
-#ifdef DEBUG
-	if (cur->bc_nlevels == 1) {
-		int freecount = 0;
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error0;
 
-		if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
-			goto error0;
-		do {
-			if ((error = xfs_inobt_get_rec(cur,
-					&rec.ir_startino,
-					&rec.ir_freecount,
-					&rec.ir_free, &i)))
-				goto error0;
-			if (i) {
-				freecount += rec.ir_freecount;
-				if ((error = xfs_btree_increment(cur, 0, &i)))
-					goto error0;
-			}
-		} while (i == 1);
-		ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
-		       XFS_FORCED_SHUTDOWN(mp));
-	}
-#endif
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	return 0;
 
@@ -1297,9 +1307,7 @@ xfs_imap(
 		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_inobt_rec_incore_t chunk_rec;
 		xfs_buf_t	*agbp;	/* agi buffer */
 		int		i;	/* temp state */
 
@@ -1315,15 +1323,14 @@ xfs_imap(
 		}
 
 		cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
-		error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i);
+		error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i);
 		if (error) {
 			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
-					"xfs_inobt_lookup_le() failed");
+					"xfs_inobt_lookup() failed");
 			goto error0;
 		}
 
-		error = xfs_inobt_get_rec(cur, &chunk_agino, &chunk_cnt,
-				&chunk_free, &i);
+		error = xfs_inobt_get_rec(cur, &chunk_rec, &i);
 		if (error) {
 			xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
 					"xfs_inobt_get_rec() failed");
@@ -1341,7 +1348,7 @@ xfs_imap(
 		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 		if (error)
 			return error;
-		chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_agino);
+		chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_rec.ir_startino);
 		offset_agbno = agbno - chunk_agbno;
 	}
 
diff --git a/fs/xfs/xfs_ialloc.h b/fs/xfs/xfs_ialloc.h
index aeee8278f92..bb5385475e1 100644
--- a/fs/xfs/xfs_ialloc.h
+++ b/fs/xfs/xfs_ialloc.h
@@ -150,23 +150,15 @@ xfs_ialloc_pagi_init(
         xfs_agnumber_t  agno);		/* allocation group number */
 
 /*
- * Lookup the first record greater than or equal to ino
- * in the btree given by cur.
+ * Lookup a record by 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);
+int xfs_inobt_lookup(struct xfs_btree_cur *cur, xfs_agino_t ino,
+		xfs_lookup_t dir, 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);
+extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur,
+		xfs_inobt_rec_incore_t *rec, int *stat);
 
 #endif	/* __XFS_IALLOC_H__ */
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index ecbf8b4d2e2..80e526489be 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -82,7 +82,6 @@ xfs_inode_alloc(
 	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;
@@ -456,32 +455,6 @@ out_error_or_again:
 	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.
- * Otherwise, return NULL.
- */
-xfs_inode_t *
-xfs_inode_incore(xfs_mount_t	*mp,
-		 xfs_ino_t	ino,
-		 xfs_trans_t	*tp)
-{
-	xfs_inode_t	*ip;
-	xfs_perag_t	*pag;
-
-	pag = xfs_get_perag(mp, ino);
-	read_lock(&pag->pag_ici_lock);
-	ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ino));
-	read_unlock(&pag->pag_ici_lock);
-	xfs_put_perag(mp, pag);
-
-	/* the returned inode must match the transaction */
-	if (ip && (ip->i_transp != tp))
-		return NULL;
-	return ip;
-}
-
 /*
  * Decrement reference count of an inode structure and unlock it.
  *
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index da428b3fe0f..c1dc7ef5a1d 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -651,7 +651,7 @@ xfs_iformat_btree(
 	return 0;
 }
 
-void
+STATIC void
 xfs_dinode_from_disk(
 	xfs_icdinode_t		*to,
 	xfs_dinode_t		*from)
@@ -1247,7 +1247,7 @@ xfs_isize_check(
  * In that case the pages will still be in memory, but the inode size
  * will never have been updated.
  */
-xfs_fsize_t
+STATIC xfs_fsize_t
 xfs_file_last_byte(
 	xfs_inode_t	*ip)
 {
@@ -3837,7 +3837,7 @@ xfs_iext_inline_to_direct(
 /*
  * Resize an extent indirection array to new_size bytes.
  */
-void
+STATIC void
 xfs_iext_realloc_indirect(
 	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	int		new_size)	/* new indirection array size */
@@ -3862,7 +3862,7 @@ xfs_iext_realloc_indirect(
 /*
  * Switch from indirection array to linear (direct) extent allocations.
  */
-void
+STATIC void
 xfs_iext_indirect_to_direct(
 	 xfs_ifork_t	*ifp)		/* inode fork pointer */
 {
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 65f24a3cc99..0b38b9a869e 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -261,7 +261,6 @@ typedef struct xfs_inode {
 	/* Miscellaneous state. */
 	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_delayed_blks;	/* count of delay alloc blks */
 
 	xfs_icdinode_t		i_d;		/* most of ondisk inode */
@@ -468,8 +467,6 @@ static inline void xfs_ifunlock(xfs_inode_t *ip)
 /*
  * xfs_iget.c prototypes.
  */
-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,
 			 uint, uint, xfs_inode_t **, xfs_daddr_t);
 void		xfs_iput(xfs_inode_t *, uint);
@@ -504,7 +501,6 @@ void		xfs_ipin(xfs_inode_t *);
 void		xfs_iunpin(xfs_inode_t *);
 int		xfs_iflush(xfs_inode_t *, uint);
 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);
 void		xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint);
 
@@ -572,8 +568,6 @@ int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
 			  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);
@@ -592,8 +586,6 @@ void		xfs_iext_remove_inline(xfs_ifork_t *, xfs_extnum_t, int);
 void		xfs_iext_remove_direct(xfs_ifork_t *, xfs_extnum_t, int);
 void		xfs_iext_remove_indirect(xfs_ifork_t *, xfs_extnum_t, int);
 void		xfs_iext_realloc_direct(xfs_ifork_t *, int);
-void		xfs_iext_realloc_indirect(xfs_ifork_t *, int);
-void		xfs_iext_indirect_to_direct(xfs_ifork_t *);
 void		xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
 void		xfs_iext_inline_to_direct(xfs_ifork_t *, int);
 void		xfs_iext_destroy(xfs_ifork_t *);
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 977c4aec587..47d5b663c37 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -263,14 +263,6 @@ xfs_inode_item_format(
 	}
 
 	/*
-	 * We don't have to worry about re-ordering here because
-	 * the update_size field is protected by the inode lock
-	 * and we have that held in exclusive mode.
-	 */
-	if (ip->i_update_size)
-		ip->i_update_size = 0;
-
-	/*
 	 * Make sure to get the latest atime from the Linux inode.
 	 */
 	xfs_synchronize_atime(ip);
@@ -712,8 +704,6 @@ xfs_inode_item_unlock(
 	 * Clear out the fields of the inode log item particular
 	 * to the current transaction.
 	 */
-	iip->ili_ilock_recur = 0;
-	iip->ili_iolock_recur = 0;
 	iip->ili_flags = 0;
 
 	/*
diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h
index a52ac125f05..65bae4c9b8b 100644
--- a/fs/xfs/xfs_inode_item.h
+++ b/fs/xfs/xfs_inode_item.h
@@ -137,8 +137,6 @@ typedef struct xfs_inode_log_item {
 	struct xfs_inode	*ili_inode;	   /* inode ptr */
 	xfs_lsn_t		ili_flush_lsn;	   /* lsn at last flush */
 	xfs_lsn_t		ili_last_lsn;	   /* lsn at last transaction */
-	unsigned short		ili_ilock_recur;   /* lock recursion count */
-	unsigned short		ili_iolock_recur;  /* lock recursion count */
 	unsigned short		ili_flags;	   /* misc flags */
 	unsigned short		ili_logged;	   /* flushed logged data */
 	unsigned int		ili_last_fields;   /* fields when flushed */
diff --git a/fs/xfs/xfs_inum.h b/fs/xfs/xfs_inum.h
index 7a28191cb0d..b8e4ee4e89a 100644
--- a/fs/xfs/xfs_inum.h
+++ b/fs/xfs/xfs_inum.h
@@ -72,7 +72,6 @@ struct xfs_mount;
 
 #if XFS_BIG_INUMS
 #define	XFS_MAXINUMBER		((xfs_ino_t)((1ULL << 56) - 1ULL))
-#define	XFS_INO64_OFFSET	((xfs_ino_t)(1ULL << 32))
 #else
 #define	XFS_MAXINUMBER		((xfs_ino_t)((1ULL << 32) - 1ULL))
 #endif
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index aeb2d2221c7..b68f9107e26 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -39,7 +39,7 @@
 #include "xfs_error.h"
 #include "xfs_btree.h"
 
-int
+STATIC int
 xfs_internal_inum(
 	xfs_mount_t	*mp,
 	xfs_ino_t	ino)
@@ -353,9 +353,6 @@ xfs_bulkstat(
 	int			end_of_ag; /* set if we've seen the ag end */
 	int			error;	/* error code */
 	int                     fmterror;/* bulkstat formatter result */
-	__int32_t		gcnt;	/* current btree rec's count */
-	xfs_inofree_t		gfree;	/* current btree rec's free mask */
-	xfs_agino_t		gino;	/* current btree rec's start inode */
 	int			i;	/* loop index */
 	int			icount;	/* count of inodes good in irbuf */
 	size_t			irbsize; /* size of irec buffer in bytes */
@@ -442,40 +439,43 @@ xfs_bulkstat(
 		 * we need to get the remainder of the chunk we're in.
 		 */
 		if (agino > 0) {
+			xfs_inobt_rec_incore_t r;
+
 			/*
 			 * Lookup the inode chunk that this inode lives in.
 			 */
-			error = xfs_inobt_lookup_le(cur, agino, 0, 0, &tmp);
+			error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE,
+						 &tmp);
 			if (!error &&	/* no I/O error */
 			    tmp &&	/* lookup succeeded */
 					/* got the record, should always work */
-			    !(error = xfs_inobt_get_rec(cur, &gino, &gcnt,
-				    &gfree, &i)) &&
+			    !(error = xfs_inobt_get_rec(cur, &r, &i)) &&
 			    i == 1 &&
 					/* this is the right chunk */
-			    agino < gino + XFS_INODES_PER_CHUNK &&
+			    agino < r.ir_startino + XFS_INODES_PER_CHUNK &&
 					/* lastino was not last in chunk */
-			    (chunkidx = agino - gino + 1) <
+			    (chunkidx = agino - r.ir_startino + 1) <
 				    XFS_INODES_PER_CHUNK &&
 					/* there are some left allocated */
 			    xfs_inobt_maskn(chunkidx,
-				    XFS_INODES_PER_CHUNK - chunkidx) & ~gfree) {
+				    XFS_INODES_PER_CHUNK - chunkidx) &
+				    ~r.ir_free) {
 				/*
 				 * Grab the chunk record.  Mark all the
 				 * uninteresting inodes (because they're
 				 * before our start point) free.
 				 */
 				for (i = 0; i < chunkidx; i++) {
-					if (XFS_INOBT_MASK(i) & ~gfree)
-						gcnt++;
+					if (XFS_INOBT_MASK(i) & ~r.ir_free)
+						r.ir_freecount++;
 				}
-				gfree |= xfs_inobt_maskn(0, chunkidx);
-				irbp->ir_startino = gino;
-				irbp->ir_freecount = gcnt;
-				irbp->ir_free = gfree;
+				r.ir_free |= xfs_inobt_maskn(0, chunkidx);
+				irbp->ir_startino = r.ir_startino;
+				irbp->ir_freecount = r.ir_freecount;
+				irbp->ir_free = r.ir_free;
 				irbp++;
-				agino = gino + XFS_INODES_PER_CHUNK;
-				icount = XFS_INODES_PER_CHUNK - gcnt;
+				agino = r.ir_startino + XFS_INODES_PER_CHUNK;
+				icount = XFS_INODES_PER_CHUNK - r.ir_freecount;
 			} else {
 				/*
 				 * If any of those tests failed, bump the
@@ -493,7 +493,7 @@ xfs_bulkstat(
 			/*
 			 * Start of ag.  Lookup the first inode chunk.
 			 */
-			error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &tmp);
+			error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &tmp);
 			icount = 0;
 		}
 		/*
@@ -501,6 +501,8 @@ xfs_bulkstat(
 		 * until we run out of inodes or space in the buffer.
 		 */
 		while (irbp < irbufend && icount < ubcount) {
+			xfs_inobt_rec_incore_t r;
+
 			/*
 			 * Loop as long as we're unable to read the
 			 * inode btree.
@@ -510,51 +512,55 @@ xfs_bulkstat(
 				if (XFS_AGINO_TO_AGBNO(mp, agino) >=
 						be32_to_cpu(agi->agi_length))
 					break;
-				error = xfs_inobt_lookup_ge(cur, agino, 0, 0,
-							    &tmp);
+				error = xfs_inobt_lookup(cur, agino,
+							 XFS_LOOKUP_GE, &tmp);
 				cond_resched();
 			}
 			/*
 			 * If ran off the end of the ag either with an error,
 			 * or the normal way, set end and stop collecting.
 			 */
-			if (error ||
-			    (error = xfs_inobt_get_rec(cur, &gino, &gcnt,
-				    &gfree, &i)) ||
-			    i == 0) {
+			if (error) {
 				end_of_ag = 1;
 				break;
 			}
+
+			error = xfs_inobt_get_rec(cur, &r, &i);
+			if (error || i == 0) {
+				end_of_ag = 1;
+				break;
+			}
+
 			/*
 			 * If this chunk has any allocated inodes, save it.
 			 * Also start read-ahead now for this chunk.
 			 */
-			if (gcnt < XFS_INODES_PER_CHUNK) {
+			if (r.ir_freecount < XFS_INODES_PER_CHUNK) {
 				/*
 				 * Loop over all clusters in the next chunk.
 				 * Do a readahead if there are any allocated
 				 * inodes in that cluster.
 				 */
-				for (agbno = XFS_AGINO_TO_AGBNO(mp, gino),
-				     chunkidx = 0;
+				agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino);
+				for (chunkidx = 0;
 				     chunkidx < XFS_INODES_PER_CHUNK;
 				     chunkidx += nicluster,
 				     agbno += nbcluster) {
-					if (xfs_inobt_maskn(chunkidx,
-							    nicluster) & ~gfree)
+					if (xfs_inobt_maskn(chunkidx, nicluster)
+							& ~r.ir_free)
 						xfs_btree_reada_bufs(mp, agno,
 							agbno, nbcluster);
 				}
-				irbp->ir_startino = gino;
-				irbp->ir_freecount = gcnt;
-				irbp->ir_free = gfree;
+				irbp->ir_startino = r.ir_startino;
+				irbp->ir_freecount = r.ir_freecount;
+				irbp->ir_free = r.ir_free;
 				irbp++;
-				icount += XFS_INODES_PER_CHUNK - gcnt;
+				icount += XFS_INODES_PER_CHUNK - r.ir_freecount;
 			}
 			/*
 			 * Set agino to after this chunk and bump the cursor.
 			 */
-			agino = gino + XFS_INODES_PER_CHUNK;
+			agino = r.ir_startino + XFS_INODES_PER_CHUNK;
 			error = xfs_btree_increment(cur, 0, &tmp);
 			cond_resched();
 		}
@@ -820,9 +826,7 @@ xfs_inumbers(
 	int		bufidx;
 	xfs_btree_cur_t	*cur;
 	int		error;
-	__int32_t	gcnt;
-	xfs_inofree_t	gfree;
-	xfs_agino_t	gino;
+	xfs_inobt_rec_incore_t r;
 	int		i;
 	xfs_ino_t	ino;
 	int		left;
@@ -855,7 +859,8 @@ xfs_inumbers(
 				continue;
 			}
 			cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
-			error = xfs_inobt_lookup_ge(cur, agino, 0, 0, &tmp);
+			error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_GE,
+						 &tmp);
 			if (error) {
 				xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 				cur = NULL;
@@ -870,9 +875,8 @@ xfs_inumbers(
 				continue;
 			}
 		}
-		if ((error = xfs_inobt_get_rec(cur, &gino, &gcnt, &gfree,
-			&i)) ||
-		    i == 0) {
+		error = xfs_inobt_get_rec(cur, &r, &i);
+		if (error || i == 0) {
 			xfs_buf_relse(agbp);
 			agbp = NULL;
 			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
@@ -881,10 +885,12 @@ xfs_inumbers(
 			agino = 0;
 			continue;
 		}
-		agino = gino + XFS_INODES_PER_CHUNK - 1;
-		buffer[bufidx].xi_startino = XFS_AGINO_TO_INO(mp, agno, gino);
-		buffer[bufidx].xi_alloccount = XFS_INODES_PER_CHUNK - gcnt;
-		buffer[bufidx].xi_allocmask = ~gfree;
+		agino = r.ir_startino + XFS_INODES_PER_CHUNK - 1;
+		buffer[bufidx].xi_startino =
+			XFS_AGINO_TO_INO(mp, agno, r.ir_startino);
+		buffer[bufidx].xi_alloccount =
+			XFS_INODES_PER_CHUNK - r.ir_freecount;
+		buffer[bufidx].xi_allocmask = ~r.ir_free;
 		bufidx++;
 		left--;
 		if (bufidx == bcount) {
diff --git a/fs/xfs/xfs_itable.h b/fs/xfs/xfs_itable.h
index 1fb04e7deb6..20792bf4594 100644
--- a/fs/xfs/xfs_itable.h
+++ b/fs/xfs/xfs_itable.h
@@ -99,11 +99,6 @@ xfs_bulkstat_one(
 	void			*dibuff,
 	int			*stat);
 
-int
-xfs_internal_inum(
-	xfs_mount_t		*mp,
-	xfs_ino_t		ino);
-
 typedef int (*inumbers_fmt_pf)(
 	void			__user *ubuffer, /* buffer to write to */
 	const xfs_inogrp_t	*buffer,	/* buffer to read from */
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index bcad5f4c1fd..679c7c4926a 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -451,8 +451,6 @@ extern int	 xlog_find_tail(xlog_t	*log,
 extern int	 xlog_recover(xlog_t *log);
 extern int	 xlog_recover_finish(xlog_t *log);
 extern void	 xlog_pack_data(xlog_t *log, xlog_in_core_t *iclog, int);
-extern void	 xlog_recover_process_iunlinks(xlog_t *log);
-
 extern struct xfs_buf *xlog_get_bp(xlog_t *, int);
 extern void	 xlog_put_bp(struct xfs_buf *);
 
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 47da2fb4537..1099395d7d6 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -3263,7 +3263,7 @@ xlog_recover_process_one_iunlink(
  * freeing of the inode and its removal from the list must be
  * atomic.
  */
-void
+STATIC void
 xlog_recover_process_iunlinks(
 	xlog_t		*log)
 {
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 5c6f092659c..8b6c9e807ef 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -1568,7 +1568,7 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
  *
  * The m_sb_lock must be held when this routine is called.
  */
-int
+STATIC int
 xfs_mod_incore_sb_unlocked(
 	xfs_mount_t	*mp,
 	xfs_sb_field_t	field,
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index a5122382afd..a6c023bc0fb 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -414,13 +414,10 @@ typedef struct xfs_mod_sb {
 
 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_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);
 extern int	xfs_mod_incore_sb_batch(xfs_mount_t *, xfs_mod_sb_t *,
 			uint, int);
 extern int	xfs_mount_log_sb(xfs_mount_t *, __int64_t);
diff --git a/fs/xfs/xfs_mru_cache.c b/fs/xfs/xfs_mru_cache.c
index afee7eb2432..4b0613d99fa 100644
--- a/fs/xfs/xfs_mru_cache.c
+++ b/fs/xfs/xfs_mru_cache.c
@@ -564,35 +564,6 @@ xfs_mru_cache_lookup(
 }
 
 /*
- * To look up an element using its key, but leave its location in the internal
- * lists alone, call xfs_mru_cache_peek().  If the element isn't found, this
- * function returns NULL.
- *
- * See the comments above the declaration of the xfs_mru_cache_lookup() function
- * for important locking information pertaining to this call.
- */
-void *
-xfs_mru_cache_peek(
-	xfs_mru_cache_t	*mru,
-	unsigned long	key)
-{
-	xfs_mru_cache_elem_t *elem;
-
-	ASSERT(mru && mru->lists);
-	if (!mru || !mru->lists)
-		return NULL;
-
-	spin_lock(&mru->lock);
-	elem = radix_tree_lookup(&mru->store, key);
-	if (!elem)
-		spin_unlock(&mru->lock);
-	else
-		__release(mru_lock); /* help sparse not be stupid */
-
-	return elem ? elem->value : NULL;
-}
-
-/*
  * To release the internal data structure spinlock after having performed an
  * xfs_mru_cache_lookup() or an xfs_mru_cache_peek(), call xfs_mru_cache_done()
  * with the data store pointer.
diff --git a/fs/xfs/xfs_mru_cache.h b/fs/xfs/xfs_mru_cache.h
index dd58ea1bbeb..5d439f34b0c 100644
--- a/fs/xfs/xfs_mru_cache.h
+++ b/fs/xfs/xfs_mru_cache.h
@@ -49,7 +49,6 @@ int xfs_mru_cache_insert(struct xfs_mru_cache *mru, unsigned long key,
 void * xfs_mru_cache_remove(struct xfs_mru_cache *mru, unsigned long key);
 void xfs_mru_cache_delete(struct xfs_mru_cache *mru, unsigned long key);
 void *xfs_mru_cache_lookup(struct xfs_mru_cache *mru, unsigned long key);
-void *xfs_mru_cache_peek(struct xfs_mru_cache *mru, unsigned long key);
 void xfs_mru_cache_done(struct xfs_mru_cache *mru);
 
 #endif /* __XFS_MRU_CACHE_H__ */
diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c
index fea68615ed2..3f816ad7ff1 100644
--- a/fs/xfs/xfs_rw.c
+++ b/fs/xfs/xfs_rw.c
@@ -88,90 +88,6 @@ xfs_write_clear_setuid(
 }
 
 /*
- * Handle logging requirements of various synchronous types of write.
- */
-int
-xfs_write_sync_logforce(
-	xfs_mount_t	*mp,
-	xfs_inode_t	*ip)
-{
-	int		error = 0;
-
-	/*
-	 * If we're treating this as O_DSYNC and we have not updated the
-	 * size, force the log.
-	 */
-	if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
-	    !(ip->i_update_size)) {
-		xfs_inode_log_item_t	*iip = ip->i_itemp;
-
-		/*
-		 * If an allocation transaction occurred
-		 * without extending the size, then we have to force
-		 * the log up the proper point to ensure that the
-		 * allocation is permanent.  We can't count on
-		 * the fact that buffered writes lock out direct I/O
-		 * writes - the direct I/O write could have extended
-		 * the size nontransactionally, then finished before
-		 * we started.  xfs_write_file will think that the file
-		 * didn't grow but the update isn't safe unless the
-		 * size change is logged.
-		 *
-		 * Force the log if we've committed a transaction
-		 * against the inode or if someone else has and
-		 * the commit record hasn't gone to disk (e.g.
-		 * the inode is pinned).  This guarantees that
-		 * all changes affecting the inode are permanent
-		 * when we return.
-		 */
-		if (iip && iip->ili_last_lsn) {
-			error = _xfs_log_force(mp, iip->ili_last_lsn,
-					XFS_LOG_FORCE | XFS_LOG_SYNC, NULL);
-		} else if (xfs_ipincount(ip) > 0) {
-			error = _xfs_log_force(mp, (xfs_lsn_t)0,
-					XFS_LOG_FORCE | XFS_LOG_SYNC, NULL);
-		}
-
-	} else {
-		xfs_trans_t	*tp;
-
-		/*
-		 * O_SYNC or O_DSYNC _with_ a size update are handled
-		 * the same way.
-		 *
-		 * If the write was synchronous then we need to make
-		 * sure that the inode modification time is permanent.
-		 * We'll have updated the timestamp above, so here
-		 * we use a synchronous transaction to log the inode.
-		 * It's not fast, but it's necessary.
-		 *
-		 * If this a dsync write and the size got changed
-		 * non-transactionally, then we need to ensure that
-		 * the size change gets logged in a synchronous
-		 * transaction.
-		 */
-		tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
-		if ((error = xfs_trans_reserve(tp, 0,
-						XFS_SWRITE_LOG_RES(mp),
-						0, 0, 0))) {
-			/* Transaction reserve failed */
-			xfs_trans_cancel(tp, 0);
-		} else {
-			/* Transaction reserve successful */
-			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);
-			xfs_trans_set_sync(tp);
-			error = xfs_trans_commit(tp, 0);
-			xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		}
-	}
-
-	return error;
-}
-
-/*
  * Force a shutdown of the filesystem instantly while keeping
  * the filesystem consistent. We don't do an unmount here; just shutdown
  * the shop, make sure that absolutely nothing persistent happens to
diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h
index f76c003ec55..f5e4874c37d 100644
--- a/fs/xfs/xfs_rw.h
+++ b/fs/xfs/xfs_rw.h
@@ -68,7 +68,6 @@ xfs_get_extsz_hint(
  * Prototypes for functions in xfs_rw.c.
  */
 extern int xfs_write_clear_setuid(struct xfs_inode *ip);
-extern int xfs_write_sync_logforce(struct xfs_mount *mp, struct xfs_inode *ip);
 extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp);
 extern int xfs_bioerror(struct xfs_buf *bp);
 extern int xfs_bioerror_relse(struct xfs_buf *bp);
@@ -78,10 +77,4 @@ extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp,
 extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
 				xfs_buf_t *bp, xfs_daddr_t blkno);
 
-/*
- * Prototypes for functions in xfs_vnodeops.c.
- */
-extern int xfs_free_eofblocks(struct xfs_mount *mp, struct xfs_inode *ip,
-			int flags);
-
 #endif /* __XFS_RW_H__ */
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 775249a54f6..ed47fc77759 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -68,7 +68,7 @@ typedef struct xfs_trans_header {
 #define XFS_TRANS_GROWFS		14
 #define XFS_TRANS_STRAT_WRITE		15
 #define XFS_TRANS_DIOSTRAT		16
-#define	XFS_TRANS_WRITE_SYNC		17
+/* 17 was XFS_TRANS_WRITE_SYNC */
 #define	XFS_TRANS_WRITEID		18
 #define	XFS_TRANS_ADDAFORK		19
 #define	XFS_TRANS_ATTRINVAL		20
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 8ee2f8c8b0a..218829e6a15 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -307,7 +307,7 @@ xfs_trans_read_buf(
 			return (flags & XFS_BUF_TRYLOCK) ?
 					EAGAIN : XFS_ERROR(ENOMEM);
 
-		if ((bp != NULL) && (XFS_BUF_GETERROR(bp) != 0)) {
+		if (XFS_BUF_GETERROR(bp) != 0) {
 			xfs_ioerror_alert("xfs_trans_read_buf", mp,
 					  bp, blkno);
 			error = XFS_BUF_GETERROR(bp);
@@ -315,7 +315,7 @@ xfs_trans_read_buf(
 			return error;
 		}
 #ifdef DEBUG
-		if (xfs_do_error && (bp != NULL)) {
+		if (xfs_do_error) {
 			if (xfs_error_target == target) {
 				if (((xfs_req_num++) % xfs_error_mod) == 0) {
 					xfs_buf_relse(bp);
diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c
index 23d276af2e0..785ff101da0 100644
--- a/fs/xfs/xfs_trans_inode.c
+++ b/fs/xfs/xfs_trans_inode.c
@@ -49,30 +49,7 @@ xfs_trans_inode_broot_debug(
 
 
 /*
- * Get and lock the inode for the caller if it is not already
- * locked within the given transaction.  If it is already locked
- * within the transaction, just increment its lock recursion count
- * and return a pointer to it.
- *
- * For an inode to be locked in a transaction, the inode lock, as
- * opposed to the io lock, must be taken exclusively.  This ensures
- * that the inode can be involved in only 1 transaction at a time.
- * Lock recursion is handled on the io lock, but only for lock modes
- * of equal or lesser strength.  That is, you can recur on the io lock
- * held EXCL with a SHARED request but not vice versa.  Also, if
- * the inode is already a part of the transaction then you cannot
- * go from not holding the io lock to having it EXCL or SHARED.
- *
- * Use the inode cache routine xfs_inode_incore() to find the inode
- * if it is already owned by this transaction.
- *
- * If we don't already own the inode, use xfs_iget() to get it.
- * Since the inode log item structure is embedded in the incore
- * inode structure and is initialized when the inode is brought
- * into memory, there is nothing to do with it here.
- *
- * If the given transaction pointer is NULL, just call xfs_iget().
- * This simplifies code which must handle both cases.
+ * Get an inode and join it to the transaction.
  */
 int
 xfs_trans_iget(
@@ -84,62 +61,11 @@ xfs_trans_iget(
 	xfs_inode_t	**ipp)
 {
 	int			error;
-	xfs_inode_t		*ip;
-
-	/*
-	 * If the transaction pointer is NULL, just call the normal
-	 * xfs_iget().
-	 */
-	if (tp == NULL)
-		return xfs_iget(mp, NULL, ino, flags, lock_flags, ipp, 0);
-
-	/*
-	 * If we find the inode in core with this transaction
-	 * pointer in its i_transp field, then we know we already
-	 * have it locked.  In this case we just increment the lock
-	 * recursion count and return the inode to the caller.
-	 * Assert that the inode is already locked in the mode requested
-	 * by the caller.  We cannot do lock promotions yet, so
-	 * die if someone gets this wrong.
-	 */
-	if ((ip = xfs_inode_incore(tp->t_mountp, ino, tp)) != NULL) {
-		/*
-		 * Make sure that the inode lock is held EXCL and
-		 * that the io lock is never upgraded when the inode
-		 * is already a part of the transaction.
-		 */
-		ASSERT(ip->i_itemp != NULL);
-		ASSERT(lock_flags & XFS_ILOCK_EXCL);
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-		ASSERT((!(lock_flags & XFS_IOLOCK_EXCL)) ||
-		       xfs_isilocked(ip, XFS_IOLOCK_EXCL));
-		ASSERT((!(lock_flags & XFS_IOLOCK_EXCL)) ||
-		       (ip->i_itemp->ili_flags & XFS_ILI_IOLOCKED_EXCL));
-		ASSERT((!(lock_flags & XFS_IOLOCK_SHARED)) ||
-		       xfs_isilocked(ip, XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED));
-		ASSERT((!(lock_flags & XFS_IOLOCK_SHARED)) ||
-		       (ip->i_itemp->ili_flags & XFS_ILI_IOLOCKED_ANY));
-
-		if (lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) {
-			ip->i_itemp->ili_iolock_recur++;
-		}
-		if (lock_flags & XFS_ILOCK_EXCL) {
-			ip->i_itemp->ili_ilock_recur++;
-		}
-		*ipp = ip;
-		return 0;
-	}
-
-	ASSERT(lock_flags & XFS_ILOCK_EXCL);
-	error = xfs_iget(tp->t_mountp, tp, ino, flags, lock_flags, &ip, 0);
-	if (error) {
-		return error;
-	}
-	ASSERT(ip != NULL);
 
-	xfs_trans_ijoin(tp, ip, lock_flags);
-	*ipp = ip;
-	return 0;
+	error = xfs_iget(mp, tp, ino, flags, lock_flags, ipp, 0);
+	if (!error && tp)
+		xfs_trans_ijoin(tp, *ipp, lock_flags);
+	return error;
 }
 
 /*
@@ -163,8 +89,6 @@ xfs_trans_ijoin(
 		xfs_inode_item_init(ip, ip->i_mount);
 	iip = ip->i_itemp;
 	ASSERT(iip->ili_flags == 0);
-	ASSERT(iip->ili_ilock_recur == 0);
-	ASSERT(iip->ili_iolock_recur == 0);
 
 	/*
 	 * Get a log_item_desc to point at the new item.
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 492d75bae2b..a434f287962 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -611,7 +611,7 @@ xfs_fsync(
 	xfs_inode_t	*ip)
 {
 	xfs_trans_t	*tp;
-	int		error;
+	int		error = 0;
 	int		log_flushed = 0, changed = 1;
 
 	xfs_itrace_entry(ip);
@@ -619,14 +619,9 @@ xfs_fsync(
 	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
 		return XFS_ERROR(EIO);
 
-	/* capture size updates in I/O completion before writing the inode. */
-	error = xfs_wait_on_pages(ip, 0, -1);
-	if (error)
-		return XFS_ERROR(error);
-
 	/*
 	 * We always need to make sure that the required inode state is safe on
-	 * disk.  The vnode might be clean but we still might need to force the
+	 * disk.  The inode might be clean but we still might need to force the
 	 * log because of committed transactions that haven't hit the disk yet.
 	 * Likewise, there could be unflushed non-transactional changes to the
 	 * inode core that have to go to disk and this requires us to issue
@@ -638,7 +633,7 @@ xfs_fsync(
 	 */
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
 
-	if (!(ip->i_update_size || ip->i_update_core)) {
+	if (!ip->i_update_core) {
 		/*
 		 * Timestamps/size haven't changed since last inode flush or
 		 * inode transaction commit.  That means either nothing got
@@ -718,7 +713,7 @@ xfs_fsync(
  * when the link count isn't zero and by xfs_dm_punch_hole() when
  * punching a hole to EOF.
  */
-int
+STATIC int
 xfs_free_eofblocks(
 	xfs_mount_t	*mp,
 	xfs_inode_t	*ip,
@@ -1476,8 +1471,8 @@ xfs_create(
 	if (error == ENOSPC) {
 		/* flush outstanding delalloc blocks and retry */
 		xfs_flush_inodes(dp);
-		error = xfs_trans_reserve(tp, resblks, XFS_CREATE_LOG_RES(mp), 0,
-			XFS_TRANS_PERM_LOG_RES, XFS_CREATE_LOG_COUNT);
+		error = xfs_trans_reserve(tp, resblks, log_res, 0,
+				XFS_TRANS_PERM_LOG_RES, log_count);
 	}
 	if (error == ENOSPC) {
 		/* No space at all so try a "no-allocation" reservation */
author	Frederic Weisbecker <fweisbec@gmail.com>	2009-09-23 23:08:43 +0200
committer	Frederic Weisbecker <fweisbec@gmail.com>	2009-09-23 23:08:43 +0200
commit	d7a4b414eed51f1653bb05ebe84122bf9a7ae18b (patch)
tree	bd6603a0c27de4c138a1767871897e9cd3e1a1d2 /fs
parent	1f0ab40976460bc4673fa204ce917a725185d8f2 (diff)
parent	a724eada8c2a7b62463b73ccf73fd0bb6e928aeb (diff)