From aab3a8c7a3a6a001dd439ed00d4db17a1059803e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= Date: Tue, 19 Aug 2008 14:23:37 +0000 Subject: [CIFS] reindent misindented statement MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Ilpo Järvinen Signed-off-by: Steve French --- fs/cifs/inode.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c index 848286861c3..9c548f11010 100644 --- a/fs/cifs/inode.c +++ b/fs/cifs/inode.c @@ -546,7 +546,8 @@ int cifs_get_inode_info(struct inode **pinode, if ((inode->i_mode & S_IWUGO) == 0 && (attr & ATTR_READONLY) == 0) inode->i_mode |= (S_IWUGO & default_mode); - inode->i_mode &= ~S_IFMT; + + inode->i_mode &= ~S_IFMT; } /* clear write bits if ATTR_READONLY is set */ if (attr & ATTR_READONLY) -- cgit v1.2.3 From cb7691b648bddbfaf6dd8d8068273dbb18d2484c Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Mon, 18 Aug 2008 15:41:05 -0400 Subject: cifs: add local server pointer to cifs_setup_session cifs_setup_session references pSesInfo->server several times. That pointer shouldn't change during the life of the function so grab it once and store it in a local var. This makes the code look a little cleaner too. Signed-off-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/connect.c | 33 +++++++++++++++++---------------- 1 file changed, 17 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 0711db65afe..4c13bcdb92a 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -3598,19 +3598,21 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *pSesInfo, char ntlm_session_key[CIFS_SESS_KEY_SIZE]; bool ntlmv2_flag = false; int first_time = 0; + struct TCP_Server_Info *server = pSesInfo->server; /* what if server changes its buffer size after dropping the session? */ - if (pSesInfo->server->maxBuf == 0) /* no need to send on reconnect */ { + if (server->maxBuf == 0) /* no need to send on reconnect */ { rc = CIFSSMBNegotiate(xid, pSesInfo); - if (rc == -EAGAIN) /* retry only once on 1st time connection */ { + if (rc == -EAGAIN) { + /* retry only once on 1st time connection */ rc = CIFSSMBNegotiate(xid, pSesInfo); if (rc == -EAGAIN) rc = -EHOSTDOWN; } if (rc == 0) { spin_lock(&GlobalMid_Lock); - if (pSesInfo->server->tcpStatus != CifsExiting) - pSesInfo->server->tcpStatus = CifsGood; + if (server->tcpStatus != CifsExiting) + server->tcpStatus = CifsGood; else rc = -EHOSTDOWN; spin_unlock(&GlobalMid_Lock); @@ -3623,23 +3625,22 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *pSesInfo, goto ss_err_exit; pSesInfo->flags = 0; - pSesInfo->capabilities = pSesInfo->server->capabilities; + pSesInfo->capabilities = server->capabilities; if (linuxExtEnabled == 0) pSesInfo->capabilities &= (~CAP_UNIX); /* pSesInfo->sequence_number = 0;*/ cFYI(1, ("Security Mode: 0x%x Capabilities: 0x%x TimeAdjust: %d", - pSesInfo->server->secMode, - pSesInfo->server->capabilities, - pSesInfo->server->timeAdj)); + server->secMode, server->capabilities, server->timeAdj)); + if (experimEnabled < 2) rc = CIFS_SessSetup(xid, pSesInfo, first_time, nls_info); else if (extended_security && (pSesInfo->capabilities & CAP_EXTENDED_SECURITY) - && (pSesInfo->server->secType == NTLMSSP)) { + && (server->secType == NTLMSSP)) { rc = -EOPNOTSUPP; } else if (extended_security && (pSesInfo->capabilities & CAP_EXTENDED_SECURITY) - && (pSesInfo->server->secType == RawNTLMSSP)) { + && (server->secType == RawNTLMSSP)) { cFYI(1, ("NTLMSSP sesssetup")); rc = CIFSNTLMSSPNegotiateSessSetup(xid, pSesInfo, &ntlmv2_flag, nls_info); @@ -3668,12 +3669,12 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *pSesInfo, } else { SMBNTencrypt(pSesInfo->password, - pSesInfo->server->cryptKey, + server->cryptKey, ntlm_session_key); if (first_time) cifs_calculate_mac_key( - &pSesInfo->server->mac_signing_key, + &server->mac_signing_key, ntlm_session_key, pSesInfo->password); } @@ -3686,13 +3687,13 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *pSesInfo, nls_info); } } else { /* old style NTLM 0.12 session setup */ - SMBNTencrypt(pSesInfo->password, pSesInfo->server->cryptKey, + SMBNTencrypt(pSesInfo->password, server->cryptKey, ntlm_session_key); if (first_time) - cifs_calculate_mac_key( - &pSesInfo->server->mac_signing_key, - ntlm_session_key, pSesInfo->password); + cifs_calculate_mac_key(&server->mac_signing_key, + ntlm_session_key, + pSesInfo->password); rc = CIFSSessSetup(xid, pSesInfo, ntlm_session_key, nls_info); } -- cgit v1.2.3 From c16fefa56334e8d0197492607e473fdbb813073f Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 19 Aug 2008 19:35:33 +0000 Subject: [CIFS] distinguish between Kerberos and MSKerberos in upcall Properly handle MSKRB5 by passing sec=mskrb5 to the upcall so that the spengo blob can be generated appropriately. Also, make decode_negTokenInit prefer whichever mechanism is first in the list. Needed for some NetApp servers, and possibly some older versions of Windows which treat the two KRB5 mechanisms differently. Signed-off-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/asn1.c | 11 ++++++++--- fs/cifs/cifs_spnego.c | 4 +++- fs/cifs/cifsglob.h | 3 ++- fs/cifs/sess.c | 2 +- 4 files changed, 14 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/cifs/asn1.c b/fs/cifs/asn1.c index 5fabd2caf93..1b09f167006 100644 --- a/fs/cifs/asn1.c +++ b/fs/cifs/asn1.c @@ -476,6 +476,7 @@ decode_negTokenInit(unsigned char *security_blob, int length, unsigned int cls, con, tag, oidlen, rc; bool use_ntlmssp = false; bool use_kerberos = false; + bool use_mskerberos = false; *secType = NTLM; /* BB eventually make Kerberos or NLTMSSP the default*/ @@ -574,10 +575,12 @@ decode_negTokenInit(unsigned char *security_blob, int length, *(oid + 1), *(oid + 2), *(oid + 3))); if (compare_oid(oid, oidlen, MSKRB5_OID, - MSKRB5_OID_LEN)) - use_kerberos = true; + MSKRB5_OID_LEN) && + !use_kerberos) + use_mskerberos = true; else if (compare_oid(oid, oidlen, KRB5_OID, - KRB5_OID_LEN)) + KRB5_OID_LEN) && + !use_mskerberos) use_kerberos = true; else if (compare_oid(oid, oidlen, NTLMSSP_OID, NTLMSSP_OID_LEN)) @@ -630,6 +633,8 @@ decode_negTokenInit(unsigned char *security_blob, int length, if (use_kerberos) *secType = Kerberos; + else if (use_mskerberos) + *secType = MSKerberos; else if (use_ntlmssp) *secType = NTLMSSP; diff --git a/fs/cifs/cifs_spnego.c b/fs/cifs/cifs_spnego.c index 2434ab0e879..117ef4bba68 100644 --- a/fs/cifs/cifs_spnego.c +++ b/fs/cifs/cifs_spnego.c @@ -114,9 +114,11 @@ cifs_get_spnego_key(struct cifsSesInfo *sesInfo) dp = description + strlen(description); - /* for now, only sec=krb5 is valid */ + /* for now, only sec=krb5 and sec=mskrb5 are valid */ if (server->secType == Kerberos) sprintf(dp, ";sec=krb5"); + else if (server->secType == MSKerberos) + sprintf(dp, ";sec=mskrb5"); else goto out; diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index 7e1cf262eff..8dfd6f24d48 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -80,7 +80,8 @@ enum securityEnum { NTLMv2, /* Legacy NTLM auth with NTLMv2 hash */ RawNTLMSSP, /* NTLMSSP without SPNEGO */ NTLMSSP, /* NTLMSSP via SPNEGO */ - Kerberos /* Kerberos via SPNEGO */ + Kerberos, /* Kerberos via SPNEGO */ + MSKerberos, /* MS Kerberos via SPNEGO */ }; enum protocolEnum { diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index ed150efbe27..3188e4d9cdd 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -505,7 +505,7 @@ CIFS_SessSetup(unsigned int xid, struct cifsSesInfo *ses, int first_time, unicode_ssetup_strings(&bcc_ptr, ses, nls_cp); } else ascii_ssetup_strings(&bcc_ptr, ses, nls_cp); - } else if (type == Kerberos) { + } else if (type == Kerberos || type == MSKerberos) { #ifdef CONFIG_CIFS_UPCALL struct cifs_spnego_msg *msg; spnego_key = cifs_get_spnego_key(ses); -- cgit v1.2.3 From 3d2af3465e91335bd1dbf36b19e92079d901409f Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 19 Aug 2008 20:51:09 +0000 Subject: [CIFS] Kerberos support not considered experimental anymore Acked-by: Jeff Layton Signed-off-by: Steve French --- fs/Kconfig | 1 - fs/cifs/README | 30 ++++++++++++++++++++++++++---- 2 files changed, 26 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index d3873583360..f0427105a61 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1984,7 +1984,6 @@ config CIFS_EXPERIMENTAL config CIFS_UPCALL bool "Kerberos/SPNEGO advanced session setup (EXPERIMENTAL)" - depends on CIFS_EXPERIMENTAL depends on KEYS help Enables an upcall mechanism for CIFS which accesses diff --git a/fs/cifs/README b/fs/cifs/README index 2bd6fe556f8..68b5c1169d9 100644 --- a/fs/cifs/README +++ b/fs/cifs/README @@ -642,8 +642,30 @@ The statistics for the number of total SMBs and oplock breaks are different in that they represent all for that share, not just those for which the server returned success. -Also note that "cat /proc/fs/cifs/DebugData" will display information about +Also note that "cat /proc/fs/cifs/DebugData" will display information about the active sessions and the shares that are mounted. -Enabling Kerberos (extended security) works when CONFIG_CIFS_EXPERIMENTAL is -on but requires a user space helper (from the Samba project). NTLM and NTLMv2 and -LANMAN support do not require this helper. + +Enabling Kerberos (extended security) works but requires version 1.2 or later +of the helper program cifs.upcall to be present and to be configured in the +/etc/request-key.conf file. The cifs.upcall helper program is from the Samba +project(http://www.samba.org). NTLM and NTLMv2 and LANMAN support do not +require this helper. Note that NTLMv2 security (which does not require the +cifs.upcall helper program), instead of using Kerberos, is sufficient for +some use cases. + +Enabling DFS support (used to access shares transparently in an MS-DFS +global name space) requires that CONFIG_CIFS_EXPERIMENTAL be enabled. In +addition, DFS support for target shares which are specified as UNC +names which begin with host names (rather than IP addresses) requires +a user space helper (such as cifs.upcall) to be present in order to +translate host names to ip address, and the user space helper must also +be configured in the file /etc/request-key.conf + +To use cifs Kerberos and DFS support, the Linux keyutils package should be +installed and something like the following lines should be added to the +/etc/request-key.conf file: + +create cifs.spnego * * /usr/local/sbin/cifs.upcall %k +create dns_resolver * * /usr/local/sbin/cifs.upcall %k + + -- cgit v1.2.3 From 04da11bfcf511544ae19e0a7e5f994b3237752ac Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Wed, 20 Aug 2008 17:16:34 +0300 Subject: UBIFS: fix zero-length truncations Always allow truncations to zero, even if budgeting thinks there is no space. UBIFS reserves some space for deletions anyway. Otherwise, the following happans: 1. create a file, and write as much as possible there, until ENOSPC 2. truncate the file, which fails with ENOSPC, which is not good. Signed-off-by: Artem Bityutskiy --- fs/ubifs/dir.c | 1 - fs/ubifs/file.c | 20 ++++++++++++++++---- 2 files changed, 16 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c index 5c96f1fb701..2b267c9a180 100644 --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -587,7 +587,6 @@ static int ubifs_unlink(struct inode *dir, struct dentry *dentry) if (err) { if (err != -ENOSPC) return err; - err = 0; budgeted = 0; } diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 4071d1cae29..3d698e2022b 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -793,7 +793,7 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode, int err; struct ubifs_budget_req req; loff_t old_size = inode->i_size, new_size = attr->ia_size; - int offset = new_size & (UBIFS_BLOCK_SIZE - 1); + int offset = new_size & (UBIFS_BLOCK_SIZE - 1), budgeted = 1; struct ubifs_inode *ui = ubifs_inode(inode); dbg_gen("ino %lu, size %lld -> %lld", inode->i_ino, old_size, new_size); @@ -811,8 +811,15 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode, /* A funny way to budget for truncation node */ req.dirtied_ino_d = UBIFS_TRUN_NODE_SZ; err = ubifs_budget_space(c, &req); - if (err) - return err; + if (err) { + /* + * Treat truncations to zero as deletion and always allow them, + * just like we do for '->unlink()'. + */ + if (new_size || err != -ENOSPC) + return err; + budgeted = 0; + } err = vmtruncate(inode, new_size); if (err) @@ -869,7 +876,12 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode, err = ubifs_jnl_truncate(c, inode, old_size, new_size); mutex_unlock(&ui->ui_mutex); out_budg: - ubifs_release_budget(c, &req); + if (budgeted) + ubifs_release_budget(c, &req); + else { + c->nospace = c->nospace_rp = 0; + smp_wmb(); + } return err; } -- cgit v1.2.3 From 18496e80f729be5f536d0315751b3bbb95ca913e Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Thu, 7 Aug 2008 00:11:12 +0300 Subject: [PATCH] ocfs2/cluster/tcp.c: make some functions static Commit 0f475b2abed6cbccee1da20a0bef2895eb2a0edd (ocfs2/net: Silence build warnings) made sense as far as it fixed compile warnings, but it was not required that it made the functions global. Signed-off-by: Adrian Bunk Signed-off-by: Mark Fasheh --- fs/ocfs2/cluster/tcp.c | 44 ++++++++++++++++++++++++++++++++++------- fs/ocfs2/cluster/tcp_internal.h | 32 ------------------------------ 2 files changed, 37 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index a27d61581bd..2bcf706d9dd 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -143,8 +143,8 @@ static void o2net_sc_postpone_idle(struct o2net_sock_container *sc); static void o2net_sc_reset_idle_timer(struct o2net_sock_container *sc); #ifdef CONFIG_DEBUG_FS -void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, - u32 msgkey, struct task_struct *task, u8 node) +static void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, + u32 msgkey, struct task_struct *task, u8 node) { INIT_LIST_HEAD(&nst->st_net_debug_item); nst->st_task = task; @@ -153,31 +153,61 @@ void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, nst->st_node = node; } -void o2net_set_nst_sock_time(struct o2net_send_tracking *nst) +static void o2net_set_nst_sock_time(struct o2net_send_tracking *nst) { do_gettimeofday(&nst->st_sock_time); } -void o2net_set_nst_send_time(struct o2net_send_tracking *nst) +static void o2net_set_nst_send_time(struct o2net_send_tracking *nst) { do_gettimeofday(&nst->st_send_time); } -void o2net_set_nst_status_time(struct o2net_send_tracking *nst) +static void o2net_set_nst_status_time(struct o2net_send_tracking *nst) { do_gettimeofday(&nst->st_status_time); } -void o2net_set_nst_sock_container(struct o2net_send_tracking *nst, +static void o2net_set_nst_sock_container(struct o2net_send_tracking *nst, struct o2net_sock_container *sc) { nst->st_sc = sc; } -void o2net_set_nst_msg_id(struct o2net_send_tracking *nst, u32 msg_id) +static void o2net_set_nst_msg_id(struct o2net_send_tracking *nst, u32 msg_id) { nst->st_id = msg_id; } + +#else /* CONFIG_DEBUG_FS */ + +static inline void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, + u32 msgkey, struct task_struct *task, u8 node) +{ +} + +static inline void o2net_set_nst_sock_time(struct o2net_send_tracking *nst) +{ +} + +static inline void o2net_set_nst_send_time(struct o2net_send_tracking *nst) +{ +} + +static inline void o2net_set_nst_status_time(struct o2net_send_tracking *nst) +{ +} + +static inline void o2net_set_nst_sock_container(struct o2net_send_tracking *nst, + struct o2net_sock_container *sc) +{ +} + +static inline void o2net_set_nst_msg_id(struct o2net_send_tracking *nst, + u32 msg_id) +{ +} + #endif /* CONFIG_DEBUG_FS */ static inline int o2net_reconnect_delay(void) diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h index 18307ff81b7..8d58cfe410b 100644 --- a/fs/ocfs2/cluster/tcp_internal.h +++ b/fs/ocfs2/cluster/tcp_internal.h @@ -224,42 +224,10 @@ struct o2net_send_tracking { struct timeval st_send_time; struct timeval st_status_time; }; - -void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, - u32 msgkey, struct task_struct *task, u8 node); -void o2net_set_nst_sock_time(struct o2net_send_tracking *nst); -void o2net_set_nst_send_time(struct o2net_send_tracking *nst); -void o2net_set_nst_status_time(struct o2net_send_tracking *nst); -void o2net_set_nst_sock_container(struct o2net_send_tracking *nst, - struct o2net_sock_container *sc); -void o2net_set_nst_msg_id(struct o2net_send_tracking *nst, u32 msg_id); - #else struct o2net_send_tracking { u32 dummy; }; - -static inline void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, - u32 msgkey, struct task_struct *task, u8 node) -{ -} -static inline void o2net_set_nst_sock_time(struct o2net_send_tracking *nst) -{ -} -static inline void o2net_set_nst_send_time(struct o2net_send_tracking *nst) -{ -} -static inline void o2net_set_nst_status_time(struct o2net_send_tracking *nst) -{ -} -static inline void o2net_set_nst_sock_container(struct o2net_send_tracking *nst, - struct o2net_sock_container *sc) -{ -} -static inline void o2net_set_nst_msg_id(struct o2net_send_tracking *nst, - u32 msg_id) -{ -} #endif /* CONFIG_DEBUG_FS */ #endif /* O2CLUSTER_TCP_INTERNAL_H */ -- cgit v1.2.3 From a57a874b04e27cb530a0e18c244387452e73ccce Mon Sep 17 00:00:00 2001 From: Alexander Beregalov Date: Wed, 6 Aug 2008 00:50:41 +0400 Subject: [PATCH] ocfs2/cluster/netdebug.c: fix warning ocfs2/cluster/netdebug.c: fix warning fs/ocfs2/cluster/netdebug.c:154: warning: format '%lu' expects type 'long unsigned int', but argument 17 has type 'suseconds_t' Signed-off-by: Alexander Beregalov Signed-off-by: Mark Fasheh --- fs/ocfs2/cluster/netdebug.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c index d8bfa0eb41b..52276c02f71 100644 --- a/fs/ocfs2/cluster/netdebug.c +++ b/fs/ocfs2/cluster/netdebug.c @@ -138,20 +138,20 @@ static int nst_seq_show(struct seq_file *seq, void *v) " message id: %d\n" " message type: %u\n" " message key: 0x%08x\n" - " sock acquiry: %lu.%lu\n" - " send start: %lu.%lu\n" - " wait start: %lu.%lu\n", + " sock acquiry: %lu.%ld\n" + " send start: %lu.%ld\n" + " wait start: %lu.%ld\n", nst, (unsigned long)nst->st_task->pid, (unsigned long)nst->st_task->tgid, nst->st_task->comm, nst->st_node, nst->st_sc, nst->st_id, nst->st_msg_type, nst->st_msg_key, nst->st_sock_time.tv_sec, - (unsigned long)nst->st_sock_time.tv_usec, + (long)nst->st_sock_time.tv_usec, nst->st_send_time.tv_sec, - (unsigned long)nst->st_send_time.tv_usec, + (long)nst->st_send_time.tv_usec, nst->st_status_time.tv_sec, - nst->st_status_time.tv_usec); + (long)nst->st_status_time.tv_usec); } spin_unlock(&o2net_debug_lock); @@ -276,7 +276,7 @@ static void *sc_seq_next(struct seq_file *seq, void *v, loff_t *pos) return sc; /* unused, just needs to be null when done */ } -#define TV_SEC_USEC(TV) TV.tv_sec, (unsigned long)TV.tv_usec +#define TV_SEC_USEC(TV) TV.tv_sec, (long)TV.tv_usec static int sc_seq_show(struct seq_file *seq, void *v) { @@ -309,12 +309,12 @@ static int sc_seq_show(struct seq_file *seq, void *v) " remote node: %s\n" " page off: %zu\n" " handshake ok: %u\n" - " timer: %lu.%lu\n" - " data ready: %lu.%lu\n" - " advance start: %lu.%lu\n" - " advance stop: %lu.%lu\n" - " func start: %lu.%lu\n" - " func stop: %lu.%lu\n" + " timer: %lu.%ld\n" + " data ready: %lu.%ld\n" + " advance start: %lu.%ld\n" + " advance stop: %lu.%ld\n" + " func start: %lu.%ld\n" + " func stop: %lu.%ld\n" " func key: %u\n" " func type: %u\n", sc, -- cgit v1.2.3 From a1af7d15a18d1e375b0a6fee93789a0bbfe088b4 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 19 Aug 2008 17:20:28 -0700 Subject: ocfs2: Fix sleep-with-spinlock recovery regression This fixes a bug introduced with 539d8264093560b917ee3afe4c7f74e5da09d6a5: [PATCH 2/2] ocfs2: Fix race between mount and recovery ocfs2_mark_dead_nodes() was reading journal inodes while holding the spinlock protecting our in-memory recovery state. The fix is very simple - the disk state is protected by a cluster lock that's already held, so we just move the spinlock down past the read. Reviewed-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/journal.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 7a37240f7a3..c47bc2a809c 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1418,13 +1418,13 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb) { unsigned int node_num; int status, i; + u32 gen; struct buffer_head *bh = NULL; struct ocfs2_dinode *di; /* This is called with the super block cluster lock, so we * know that the slot map can't change underneath us. */ - spin_lock(&osb->osb_lock); for (i = 0; i < osb->max_slots; i++) { /* Read journal inode to get the recovery generation */ status = ocfs2_read_journal_inode(osb, i, &bh, NULL); @@ -1433,23 +1433,31 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb) goto bail; } di = (struct ocfs2_dinode *)bh->b_data; - osb->slot_recovery_generations[i] = - ocfs2_get_recovery_generation(di); + gen = ocfs2_get_recovery_generation(di); brelse(bh); bh = NULL; + spin_lock(&osb->osb_lock); + osb->slot_recovery_generations[i] = gen; + mlog(0, "Slot %u recovery generation is %u\n", i, osb->slot_recovery_generations[i]); - if (i == osb->slot_num) + if (i == osb->slot_num) { + spin_unlock(&osb->osb_lock); continue; + } status = ocfs2_slot_to_node_num_locked(osb, i, &node_num); - if (status == -ENOENT) + if (status == -ENOENT) { + spin_unlock(&osb->osb_lock); continue; + } - if (__ocfs2_recovery_map_test(osb, node_num)) + if (__ocfs2_recovery_map_test(osb, node_num)) { + spin_unlock(&osb->osb_lock); continue; + } spin_unlock(&osb->osb_lock); /* Ok, we have a slot occupied by another node which @@ -1465,10 +1473,7 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb) mlog_errno(status); goto bail; } - - spin_lock(&osb->osb_lock); } - spin_unlock(&osb->osb_lock); status = 0; bail: -- cgit v1.2.3 From 83cab5338fa8c74f979223698c8d4cc88f2ab68e Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Thu, 21 Aug 2008 14:14:27 +0800 Subject: ocfs2: Jump to correct label in ocfs2_expand_inline_dir() When we fail to insert extent in ocfs2_expand_inline_dir(), we should go to out_commit, not out. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/dir.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c index 8a187584808..8e9c4a47d81 100644 --- a/fs/ocfs2/dir.c +++ b/fs/ocfs2/dir.c @@ -1310,7 +1310,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, NULL); if (ret) { mlog_errno(ret); - goto out; + goto out_commit; } ret = ocfs2_journal_dirty(handle, di_bh); @@ -1336,7 +1336,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, len, 0, NULL); if (ret) { mlog_errno(ret); - goto out; + goto out_commit; } } -- cgit v1.2.3 From 9780eb6cfaf7d2d5ccc061eaf94e7aec6a17791e Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 5 Aug 2008 11:32:46 -0700 Subject: ocfs2: correctly set i_blocks after inline dir gets expanded We were setting i_blocks based on allocation before the extent insert, which is wrong as the value is a calculation based on ip_clusters which gets updated as a result of the insert. This patch moves the line in question to just after the call to ocfs2_insert_extent(). Without this fix, inline directories were temporarily having an i_blocks value of zero immediately after expansion to extents. Reported-and-tested-by: Tristan Ye Signed-off-by: Mark Fasheh --- fs/ocfs2/dir.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c index 8e9c4a47d81..9cce563fd62 100644 --- a/fs/ocfs2/dir.c +++ b/fs/ocfs2/dir.c @@ -1300,7 +1300,6 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, di->i_size = cpu_to_le64(sb->s_blocksize); di->i_ctime = di->i_mtime = cpu_to_le64(dir->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(dir->i_ctime.tv_nsec); - dir->i_blocks = ocfs2_inode_sector_count(dir); /* * This should never fail as our extent list is empty and all @@ -1313,6 +1312,12 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, goto out_commit; } + /* + * Set i_blocks after the extent insert for the most up to + * date ip_clusters value. + */ + dir->i_blocks = ocfs2_inode_sector_count(dir); + ret = ocfs2_journal_dirty(handle, di_bh); if (ret) { mlog_errno(ret); -- cgit v1.2.3 From de6bf18e9ce0df807dab08cff08751cac383429d Mon Sep 17 00:00:00 2001 From: Louis Rilling Date: Fri, 15 Aug 2008 12:37:23 -0700 Subject: [PATCH] configfs: Consolidate locking around configfs_detach_prep() in configfs_rmdir() It appears that configfs_rmdir() can protect configfs_detach_prep() retries with less calls to {spin,mutex}_{lock,unlock}, and a cleaner code. This patch does not change any behavior, except that it removes two useless lock/unlock pairs having nothing inside to protect and providing a useless barrier. Signed-off-by: Louis Rilling Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/configfs/dir.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/configfs/dir.c b/fs/configfs/dir.c index 7a8db78a91d..8e93341f3e8 100644 --- a/fs/configfs/dir.c +++ b/fs/configfs/dir.c @@ -1311,16 +1311,18 @@ static int configfs_rmdir(struct inode *dir, struct dentry *dentry) * Ensure that no racing symlink() will make detach_prep() fail while * the new link is temporarily attached */ - mutex_lock(&configfs_symlink_mutex); - spin_lock(&configfs_dirent_lock); do { struct mutex *wait_mutex; + mutex_lock(&configfs_symlink_mutex); + spin_lock(&configfs_dirent_lock); ret = configfs_detach_prep(dentry, &wait_mutex); - if (ret) { + if (ret) configfs_detach_rollback(dentry); - spin_unlock(&configfs_dirent_lock); - mutex_unlock(&configfs_symlink_mutex); + spin_unlock(&configfs_dirent_lock); + mutex_unlock(&configfs_symlink_mutex); + + if (ret) { if (ret != -EAGAIN) { config_item_put(parent_item); return ret; @@ -1329,13 +1331,8 @@ static int configfs_rmdir(struct inode *dir, struct dentry *dentry) /* Wait until the racing operation terminates */ mutex_lock(wait_mutex); mutex_unlock(wait_mutex); - - mutex_lock(&configfs_symlink_mutex); - spin_lock(&configfs_dirent_lock); } } while (ret == -EAGAIN); - spin_unlock(&configfs_dirent_lock); - mutex_unlock(&configfs_symlink_mutex); /* Get a working ref for the duration of this function */ item = configfs_get_config_item(dentry); -- cgit v1.2.3 From cc996099174dc05b35b7a29301026987990e7f8c Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Sat, 2 Aug 2008 07:30:48 +0400 Subject: [PATCH] proc: inode number fixlet Ouch, if number taken from IDA is too big, the intent was to signal an error, not check for overflow and still do overflowing addition. One still needs 2^28 proc entries to notice this. Signed-off-by: Alexey Dobriyan Signed-off-by: Al Viro --- fs/proc/generic.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/proc/generic.c b/fs/proc/generic.c index 4fb81e9c94e..bca0f81eb68 100644 --- a/fs/proc/generic.c +++ b/fs/proc/generic.c @@ -330,6 +330,7 @@ retry: spin_lock(&proc_inum_lock); ida_remove(&proc_inum_ida, i); spin_unlock(&proc_inum_lock); + return 0; } return PROC_DYNAMIC_FIRST + i; } -- cgit v1.2.3 From 2d8a10cd1760e7ecc07a21e409485947c68a3291 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 11 Aug 2008 11:33:57 -0400 Subject: [PATCH] fix efs_lookup() it needs to use d_splice_alias(), not d_add() Signed-off-by: Al Viro --- fs/efs/namei.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/efs/namei.c b/fs/efs/namei.c index 3a404e7fad5..291abb11e20 100644 --- a/fs/efs/namei.c +++ b/fs/efs/namei.c @@ -74,8 +74,7 @@ struct dentry *efs_lookup(struct inode *dir, struct dentry *dentry, struct namei } unlock_kernel(); - d_add(dentry, inode); - return NULL; + return d_splice_alias(inode, dentry); } static struct inode *efs_nfs_get_inode(struct super_block *sb, u64 ino, -- cgit v1.2.3 From e45b590b976465c258f3e2a6cc84573fc19e16d3 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 7 Aug 2008 23:49:07 +0200 Subject: [PATCH] change d_add_ci argument ordering As pointed out during review d_add_ci argument order should match d_add, so switch the dentry and inode arguments. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/dcache.c | 2 +- fs/xfs/linux-2.6/xfs_iops.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 101663d15e9..80e93956ace 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1236,7 +1236,7 @@ struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry) * If no entry exists with the exact case name, allocate new dentry with * the exact case, and return the spliced entry. */ -struct dentry *d_add_ci(struct inode *inode, struct dentry *dentry, +struct dentry *d_add_ci(struct dentry *dentry, struct inode *inode, struct qstr *name) { int error; diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 91bcd979242..095d271f343 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -355,7 +355,7 @@ xfs_vn_ci_lookup( /* else case-insensitive match... */ dname.name = ci_name.name; dname.len = ci_name.len; - dentry = d_add_ci(VFS_I(ip), dentry, &dname); + dentry = d_add_ci(dentry, VFS_I(ip), &dname); kmem_free(ci_name.name); return dentry; } -- cgit v1.2.3 From 2690421743b03c9be05d8e44c3b827986d1329a7 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 7 Aug 2008 23:50:21 +0200 Subject: [PATCH] ntfs: use d_add_ci d_add_ci was lifted 1:1 from ntfs. Change ntfs to use the common version. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/ntfs/namei.c | 89 ++------------------------------------------------------- 1 file changed, 2 insertions(+), 87 deletions(-) (limited to 'fs') diff --git a/fs/ntfs/namei.c b/fs/ntfs/namei.c index e1781c8b165..9e8a95be7a1 100644 --- a/fs/ntfs/namei.c +++ b/fs/ntfs/namei.c @@ -174,7 +174,6 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent, // TODO: Consider moving this lot to a separate function! (AIA) handle_name: { - struct dentry *real_dent, *new_dent; MFT_RECORD *m; ntfs_attr_search_ctx *ctx; ntfs_inode *ni = NTFS_I(dent_inode); @@ -255,93 +254,9 @@ handle_name: } nls_name.hash = full_name_hash(nls_name.name, nls_name.len); - /* - * Note: No need for dent->d_lock lock as i_mutex is held on the - * parent inode. - */ - - /* Does a dentry matching the nls_name exist already? */ - real_dent = d_lookup(dent->d_parent, &nls_name); - /* If not, create it now. */ - if (!real_dent) { - real_dent = d_alloc(dent->d_parent, &nls_name); - kfree(nls_name.name); - if (!real_dent) { - err = -ENOMEM; - goto err_out; - } - new_dent = d_splice_alias(dent_inode, real_dent); - if (new_dent) - dput(real_dent); - else - new_dent = real_dent; - ntfs_debug("Done. (Created new dentry.)"); - return new_dent; - } + dent = d_add_ci(dent, dent_inode, &nls_name); kfree(nls_name.name); - /* Matching dentry exists, check if it is negative. */ - if (real_dent->d_inode) { - if (unlikely(real_dent->d_inode != dent_inode)) { - /* This can happen because bad inodes are unhashed. */ - BUG_ON(!is_bad_inode(dent_inode)); - BUG_ON(!is_bad_inode(real_dent->d_inode)); - } - /* - * Already have the inode and the dentry attached, decrement - * the reference count to balance the ntfs_iget() we did - * earlier on. We found the dentry using d_lookup() so it - * cannot be disconnected and thus we do not need to worry - * about any NFS/disconnectedness issues here. - */ - iput(dent_inode); - ntfs_debug("Done. (Already had inode and dentry.)"); - return real_dent; - } - /* - * Negative dentry: instantiate it unless the inode is a directory and - * has a 'disconnected' dentry (i.e. IS_ROOT and DCACHE_DISCONNECTED), - * in which case d_move() that in place of the found dentry. - */ - if (!S_ISDIR(dent_inode->i_mode)) { - /* Not a directory; everything is easy. */ - d_instantiate(real_dent, dent_inode); - ntfs_debug("Done. (Already had negative file dentry.)"); - return real_dent; - } - spin_lock(&dcache_lock); - if (list_empty(&dent_inode->i_dentry)) { - /* - * Directory without a 'disconnected' dentry; we need to do - * d_instantiate() by hand because it takes dcache_lock which - * we already hold. - */ - list_add(&real_dent->d_alias, &dent_inode->i_dentry); - real_dent->d_inode = dent_inode; - spin_unlock(&dcache_lock); - security_d_instantiate(real_dent, dent_inode); - ntfs_debug("Done. (Already had negative directory dentry.)"); - return real_dent; - } - /* - * Directory with a 'disconnected' dentry; get a reference to the - * 'disconnected' dentry. - */ - new_dent = list_entry(dent_inode->i_dentry.next, struct dentry, - d_alias); - dget_locked(new_dent); - spin_unlock(&dcache_lock); - /* Do security vodoo. */ - security_d_instantiate(real_dent, dent_inode); - /* Move new_dent in place of real_dent. */ - d_move(new_dent, real_dent); - /* Balance the ntfs_iget() we did above. */ - iput(dent_inode); - /* Throw away real_dent. */ - dput(real_dent); - /* Use new_dent as the actual dentry. */ - ntfs_debug("Done. (Already had negative, disconnected directory " - "dentry.)"); - return new_dent; + return dent; eio_err_out: ntfs_error(vol->sb, "Illegal file name attribute. Run chkdsk."); -- cgit v1.2.3 From 8f3f655da7288504c1013621090ecc940173ae1c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 12 Aug 2008 00:28:24 -0400 Subject: [PATCH] fix regular readdir() and friends Handling of -EOVERFLOW. Signed-off-by: Al Viro --- fs/compat.c | 8 ++++++-- fs/readdir.c | 8 ++++++-- 2 files changed, 12 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/compat.c b/fs/compat.c index c9d1472e65c..075d0509970 100644 --- a/fs/compat.c +++ b/fs/compat.c @@ -792,8 +792,10 @@ static int compat_fillonedir(void *__buf, const char *name, int namlen, if (buf->result) return -EINVAL; d_ino = ino; - if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) + if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) { + buf->result = -EOVERFLOW; return -EOVERFLOW; + } buf->result++; dirent = buf->dirent; if (!access_ok(VERIFY_WRITE, dirent, @@ -862,8 +864,10 @@ static int compat_filldir(void *__buf, const char *name, int namlen, if (reclen > buf->count) return -EINVAL; d_ino = ino; - if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) + if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) { + buf->error = -EOVERFLOW; return -EOVERFLOW; + } dirent = buf->previous; if (dirent) { if (__put_user(offset, &dirent->d_off)) diff --git a/fs/readdir.c b/fs/readdir.c index 4e026e5407f..93a7559bbfd 100644 --- a/fs/readdir.c +++ b/fs/readdir.c @@ -80,8 +80,10 @@ static int fillonedir(void * __buf, const char * name, int namlen, loff_t offset if (buf->result) return -EINVAL; d_ino = ino; - if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) + if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) { + buf->result = -EOVERFLOW; return -EOVERFLOW; + } buf->result++; dirent = buf->dirent; if (!access_ok(VERIFY_WRITE, dirent, @@ -155,8 +157,10 @@ static int filldir(void * __buf, const char * name, int namlen, loff_t offset, if (reclen > buf->count) return -EINVAL; d_ino = ino; - if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) + if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) { + buf->error = -EOVERFLOW; return -EOVERFLOW; + } dirent = buf->previous; if (dirent) { if (__put_user(offset, &dirent->d_off)) -- cgit v1.2.3 From 59af1584bf33810639cb98d79856021253e2177c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 24 Aug 2008 07:24:41 -0400 Subject: [PATCH] fix ->llseek() for a bunch of directories Signed-off-by: Al Viro --- fs/9p/vfs_dir.c | 1 + fs/adfs/dir.c | 1 + fs/affs/dir.c | 1 + fs/autofs4/root.c | 2 ++ fs/befs/linuxvfs.c | 1 + fs/xfs/linux-2.6/xfs_file.c | 1 + 6 files changed, 7 insertions(+) (limited to 'fs') diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c index 88e3787c6ea..e298fe19409 100644 --- a/fs/9p/vfs_dir.c +++ b/fs/9p/vfs_dir.c @@ -119,6 +119,7 @@ int v9fs_dir_release(struct inode *inode, struct file *filp) const struct file_operations v9fs_dir_operations = { .read = generic_read_dir, + .llseek = generic_file_llseek, .readdir = v9fs_dir_readdir, .open = v9fs_file_open, .release = v9fs_dir_release, diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c index fc1a8dc64d7..85a30e92980 100644 --- a/fs/adfs/dir.c +++ b/fs/adfs/dir.c @@ -197,6 +197,7 @@ out: const struct file_operations adfs_dir_operations = { .read = generic_read_dir, + .llseek = generic_file_llseek, .readdir = adfs_readdir, .fsync = file_fsync, }; diff --git a/fs/affs/dir.c b/fs/affs/dir.c index 6e3f282424b..7b36904dbea 100644 --- a/fs/affs/dir.c +++ b/fs/affs/dir.c @@ -19,6 +19,7 @@ static int affs_readdir(struct file *, void *, filldir_t); const struct file_operations affs_dir_operations = { .read = generic_read_dir, + .llseek = generic_file_llseek, .readdir = affs_readdir, .fsync = file_fsync, }; diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index bcfb2dc0a61..2a41c2a7fc5 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -36,6 +36,7 @@ const struct file_operations autofs4_root_operations = { .release = dcache_dir_close, .read = generic_read_dir, .readdir = dcache_readdir, + .llseek = dcache_dir_lseek, .ioctl = autofs4_root_ioctl, }; @@ -44,6 +45,7 @@ const struct file_operations autofs4_dir_operations = { .release = dcache_dir_close, .read = generic_read_dir, .readdir = dcache_readdir, + .llseek = dcache_dir_lseek, }; const struct inode_operations autofs4_indirect_root_inode_operations = { diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c index 02c6e62b72f..740f53672a8 100644 --- a/fs/befs/linuxvfs.c +++ b/fs/befs/linuxvfs.c @@ -66,6 +66,7 @@ static struct kmem_cache *befs_inode_cachep; static const struct file_operations befs_dir_operations = { .read = generic_read_dir, .readdir = befs_readdir, + .llseek = generic_file_llseek, }; static const struct inode_operations befs_dir_inode_operations = { diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c index 5f60363b934..5311c1acdd4 100644 --- a/fs/xfs/linux-2.6/xfs_file.c +++ b/fs/xfs/linux-2.6/xfs_file.c @@ -475,6 +475,7 @@ const struct file_operations xfs_invis_file_operations = { const struct file_operations xfs_dir_file_operations = { .read = generic_read_dir, .readdir = xfs_file_readdir, + .llseek = generic_file_llseek, .unlocked_ioctl = xfs_file_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = xfs_file_compat_ioctl, -- cgit v1.2.3 From 4cdfe84b51420c9ac95c7133da2d4c8a191094af Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 24 Aug 2008 07:45:33 -0400 Subject: [PATCH] deal with the first call of ->show() generating no output seq_read() has a subtle bug - we want the first loop there to go until at least one *non-empty* record had fit entirely into buffer. Signed-off-by: Al Viro --- fs/seq_file.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/seq_file.c b/fs/seq_file.c index 5d54205e486..bd20f7f5a93 100644 --- a/fs/seq_file.c +++ b/fs/seq_file.c @@ -108,9 +108,9 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos) goto Done; } /* we need at least one record in buffer */ + pos = m->index; + p = m->op->start(m, &pos); while (1) { - pos = m->index; - p = m->op->start(m, &pos); err = PTR_ERR(p); if (!p || IS_ERR(p)) break; @@ -119,6 +119,11 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos) break; if (unlikely(err)) m->count = 0; + if (unlikely(!m->count)) { + p = m->op->next(m, p, &pos); + m->index = pos; + continue; + } if (m->count < m->size) goto Fill; m->op->stop(m, p); @@ -128,6 +133,8 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos) goto Enomem; m->count = 0; m->version = 0; + pos = m->index; + p = m->op->start(m, &pos); } m->op->stop(m, p); m->count = 0; -- cgit v1.2.3 From 761e29f3bb19b05bea55285dfdf2d28e001a63b8 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Wed, 20 Aug 2008 16:32:40 +0300 Subject: UBIFS: always read hashed-key nodes under TNC mutex Leaf-nodes that have a hashed key are stored in the leaf-node-cache (LNC) which is protected by the TNC mutex. Consequently, when reading a leaf node with a hashed key (i.e. directory entries, xattr entries) the TNC mutex is always required. Signed-off-by: Adrian Hunter --- fs/ubifs/tnc.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c index e909f4a9644..4fbc5921688 100644 --- a/fs/ubifs/tnc.c +++ b/fs/ubifs/tnc.c @@ -1498,7 +1498,6 @@ static int do_lookup_nm(struct ubifs_info *c, const union ubifs_key *key, { int found, n, err; struct ubifs_znode *znode; - struct ubifs_zbranch zbr; dbg_tnc("name '%.*s' key %s", nm->len, nm->name, DBGKEY(key)); mutex_lock(&c->tnc_mutex); @@ -1522,11 +1521,7 @@ static int do_lookup_nm(struct ubifs_info *c, const union ubifs_key *key, goto out_unlock; } - zbr = znode->zbranch[n]; - mutex_unlock(&c->tnc_mutex); - - err = tnc_read_node_nm(c, &zbr, node); - return err; + err = tnc_read_node_nm(c, &znode->zbranch[n], node); out_unlock: mutex_unlock(&c->tnc_mutex); -- cgit v1.2.3 From 601c0bc46753007be011b513ba4fc50ed8e30aef Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Fri, 22 Aug 2008 14:23:35 +0300 Subject: UBIFS: allow for racing between GC and TNC The TNC mutex is unlocked prematurely when reading leaf nodes with non-hashed keys. This is unsafe because the node may be moved by garbage collection and the eraseblock unmapped, although that has never actually happened during stress testing. This patch fixes the flaw by detecting the race and retrying with the TNC mutex locked. Signed-off-by: Adrian Hunter --- fs/ubifs/gc.c | 6 +++ fs/ubifs/misc.h | 17 +++++++++ fs/ubifs/tnc.c | 109 ++++++++++++++++++++++++++++++------------------------- fs/ubifs/ubifs.h | 6 ++- 4 files changed, 87 insertions(+), 51 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/gc.c b/fs/ubifs/gc.c index d0f3dac2908..13f1019c859 100644 --- a/fs/ubifs/gc.c +++ b/fs/ubifs/gc.c @@ -344,6 +344,12 @@ int ubifs_garbage_collect_leb(struct ubifs_info *c, struct ubifs_lprops *lp) if (err) goto out; + /* Allow for races with TNC */ + c->gced_lnum = lnum; + smp_wmb(); + c->gc_seq += 1; + smp_wmb(); + if (c->gc_lnum == -1) { c->gc_lnum = lnum; err = LEB_RETAINED; diff --git a/fs/ubifs/misc.h b/fs/ubifs/misc.h index 87dabf9fe74..87ced4c74a6 100644 --- a/fs/ubifs/misc.h +++ b/fs/ubifs/misc.h @@ -325,4 +325,21 @@ static inline struct timespec ubifs_current_time(struct inode *inode) current_fs_time(inode->i_sb) : CURRENT_TIME_SEC; } +/** + * ubifs_tnc_lookup - look up a file-system node. + * @c: UBIFS file-system description object + * @key: node key to lookup + * @node: the node is returned here + * + * This function look 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. + */ +static inline int ubifs_tnc_lookup(struct ubifs_info *c, + const union ubifs_key *key, void *node) +{ + return ubifs_tnc_locate(c, key, node, NULL, NULL); +} + #endif /* __UBIFS_MISC_H__ */ diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c index 4fbc5921688..7da209ab937 100644 --- a/fs/ubifs/tnc.c +++ b/fs/ubifs/tnc.c @@ -506,7 +506,7 @@ static int fallible_read_node(struct ubifs_info *c, const union ubifs_key *key, if (keys_cmp(c, key, &node_key) != 0) ret = 0; } - if (ret == 0) + if (ret == 0 && c->replaying) dbg_mnt("dangling branch LEB %d:%d len %d, key %s", zbr->lnum, zbr->offs, zbr->len, DBGKEY(key)); return ret; @@ -1382,50 +1382,39 @@ static int lookup_level0_dirty(struct ubifs_info *c, const union ubifs_key *key, } /** - * ubifs_tnc_lookup - look up a file-system node. + * maybe_leb_gced - determine if a LEB may have been garbage collected. * @c: UBIFS file-system description object - * @key: node key to lookup - * @node: the node is returned here + * @lnum: LEB number + * @gc_seq1: garbage collection sequence number * - * This function look 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. + * This function determines if @lnum may have been garbage collected since + * sequence number @gc_seq1. If it may have been then %1 is returned, otherwise + * %0 is returned. */ -int ubifs_tnc_lookup(struct ubifs_info *c, const union ubifs_key *key, - void *node) +static int maybe_leb_gced(struct ubifs_info *c, int lnum, int gc_seq1) { - int found, n, err; - struct ubifs_znode *znode; - struct ubifs_zbranch zbr, *zt; - - mutex_lock(&c->tnc_mutex); - found = ubifs_lookup_level0(c, key, &znode, &n); - if (!found) { - err = -ENOENT; - goto out; - } else if (found < 0) { - err = found; - goto out; - } - zt = &znode->zbranch[n]; - if (is_hash_key(c, key)) { - /* - * In this case the leaf node cache gets used, so we pass the - * address of the zbranch and keep the mutex locked - */ - err = tnc_read_node_nm(c, zt, node); - goto out; - } - zbr = znode->zbranch[n]; - mutex_unlock(&c->tnc_mutex); - - err = ubifs_tnc_read_node(c, &zbr, node); - return err; + int gc_seq2, gced_lnum; -out: - mutex_unlock(&c->tnc_mutex); - return err; + gced_lnum = c->gced_lnum; + smp_rmb(); + gc_seq2 = c->gc_seq; + /* Same seq means no GC */ + if (gc_seq1 == gc_seq2) + return 0; + /* Different by more than 1 means we don't know */ + if (gc_seq1 + 1 != gc_seq2) + return 1; + /* + * We have seen the sequence number has increased by 1. Now we need to + * be sure we read the right LEB number, so read it again. + */ + smp_rmb(); + if (gced_lnum != c->gced_lnum) + return 1; + /* Finally we can check lnum */ + if (gced_lnum == lnum) + return 1; + return 0; } /** @@ -1436,16 +1425,19 @@ out: * @lnum: LEB number is returned here * @offs: offset is returned here * - * This function is the same as 'ubifs_tnc_lookup()' but it returns the node - * location also. See 'ubifs_tnc_lookup()'. + * This function look 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. */ int ubifs_tnc_locate(struct ubifs_info *c, const union ubifs_key *key, void *node, int *lnum, int *offs) { - int found, n, err; + int found, n, err, safely = 0, gc_seq1; struct ubifs_znode *znode; struct ubifs_zbranch zbr, *zt; +again: mutex_lock(&c->tnc_mutex); found = ubifs_lookup_level0(c, key, &znode, &n); if (!found) { @@ -1456,24 +1448,43 @@ int ubifs_tnc_locate(struct ubifs_info *c, const union ubifs_key *key, goto out; } zt = &znode->zbranch[n]; + if (lnum) { + *lnum = zt->lnum; + *offs = zt->offs; + } if (is_hash_key(c, key)) { /* * In this case the leaf node cache gets used, so we pass the * address of the zbranch and keep the mutex locked */ - *lnum = zt->lnum; - *offs = zt->offs; err = tnc_read_node_nm(c, zt, node); goto out; } + if (safely) { + err = ubifs_tnc_read_node(c, zt, node); + goto out; + } + /* Drop the TNC mutex prematurely and race with garbage collection */ zbr = znode->zbranch[n]; + gc_seq1 = c->gc_seq; mutex_unlock(&c->tnc_mutex); - *lnum = zbr.lnum; - *offs = zbr.offs; + if (ubifs_get_wbuf(c, zbr.lnum)) { + /* We do not GC journal heads */ + err = ubifs_tnc_read_node(c, &zbr, node); + return err; + } - err = ubifs_tnc_read_node(c, &zbr, node); - return err; + err = fallible_read_node(c, key, &zbr, node); + if (maybe_leb_gced(c, zbr.lnum, gc_seq1)) { + /* + * The node may have been GC'ed out from under us so try again + * while keeping the TNC mutex locked. + */ + safely = 1; + goto again; + } + return 0; out: mutex_unlock(&c->tnc_mutex); diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index d7f706f7a30..7828d69ca4f 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1028,6 +1028,8 @@ struct ubifs_mount_opts { * @sbuf: a buffer of LEB size used by GC and replay for scanning * @idx_gc: list of index LEBs that have been garbage collected * @idx_gc_cnt: number of elements on the idx_gc list + * @gc_seq: incremented for every non-index LEB garbage collected + * @gced_lnum: last non-index LEB that was garbage collected * * @infos_list: links all 'ubifs_info' objects * @umount_mutex: serializes shrinker and un-mount @@ -1257,6 +1259,8 @@ struct ubifs_info { void *sbuf; struct list_head idx_gc; int idx_gc_cnt; + volatile int gc_seq; + volatile int gced_lnum; struct list_head infos_list; struct mutex umount_mutex; @@ -1451,8 +1455,6 @@ int ubifs_save_dirty_idx_lnums(struct ubifs_info *c); /* tnc.c */ int ubifs_lookup_level0(struct ubifs_info *c, const union ubifs_key *key, struct ubifs_znode **zn, int *n); -int ubifs_tnc_lookup(struct ubifs_info *c, const union ubifs_key *key, - void *node); int ubifs_tnc_lookup_nm(struct ubifs_info *c, const union ubifs_key *key, void *node, const struct qstr *nm); int ubifs_tnc_locate(struct ubifs_info *c, const union ubifs_key *key, -- cgit v1.2.3 From d6817cdbd143f87f9d7c59a4c3194091190eeb84 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 22 Aug 2008 14:30:10 -0700 Subject: ocfs2: Increment the reference count of an already-active stack. The ocfs2_stack_driver_request() function failed to increment the refcount of an already-active stack. It only did the increment on the first reference. Whoops. Signed-off-by: Joel Becker Tested-by: Marcos Matsunaga Signed-off-by: Mark Fasheh --- fs/ocfs2/stackglue.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 10e149ae5e3..07f348b8d72 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -97,13 +97,14 @@ static int ocfs2_stack_driver_request(const char *stack_name, goto out; } - /* Ok, the stack is pinned */ - p->sp_count++; active_stack = p; - rc = 0; out: + /* If we found it, pin it */ + if (!rc) + active_stack->sp_count++; + spin_unlock(&ocfs2_stack_lock); return rc; } -- cgit v1.2.3 From 6ce5eecb9cd3ac97b952c50309b87c31488a45e9 Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 26 Aug 2008 00:37:14 +0000 Subject: [CIFS] check version in spnego upcall response Currently, we don't check the version in the SPNEGO upcall response even though one is provided. Jeff and Q have made the corresponding change to the Samba client (cifs.upcall). Acked-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/CHANGES | 6 +++++- fs/cifs/cifs_spnego.h | 2 +- fs/cifs/sess.c | 9 +++++++++ 3 files changed, 15 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES index f5d0083e09f..526041a52d3 100644 --- a/fs/cifs/CHANGES +++ b/fs/cifs/CHANGES @@ -4,7 +4,11 @@ Fix premature write failure on congested networks (we would give up on EAGAIN from the socket too quickly on large writes). Cifs_mkdir and cifs_create now respect the setgid bit on parent dir. Fix endian problems in acl (mode from/to cifs acl) on bigendian -architectures. +architectures. Fix problems with preserving timestamps on copying open +files (e.g. "cp -a") to Windows servers. For mkdir and create honor setgid bit +on parent directory when server supports Unix Extensions but not POSIX +create. Update cifs.upcall version to handle new Kerberos sec flags +(this requires update of cifs.upcall program from Samba). Version 1.53 ------------ diff --git a/fs/cifs/cifs_spnego.h b/fs/cifs/cifs_spnego.h index 05a34b17a1a..e4041ec4d71 100644 --- a/fs/cifs/cifs_spnego.h +++ b/fs/cifs/cifs_spnego.h @@ -23,7 +23,7 @@ #ifndef _CIFS_SPNEGO_H #define _CIFS_SPNEGO_H -#define CIFS_SPNEGO_UPCALL_VERSION 1 +#define CIFS_SPNEGO_UPCALL_VERSION 2 /* * The version field should always be set to CIFS_SPNEGO_UPCALL_VERSION. diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index 3188e4d9cdd..b537fad3bf5 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -516,6 +516,15 @@ CIFS_SessSetup(unsigned int xid, struct cifsSesInfo *ses, int first_time, } msg = spnego_key->payload.data; + /* check version field to make sure that cifs.upcall is + sending us a response in an expected form */ + if (msg->version != CIFS_SPNEGO_UPCALL_VERSION) { + cERROR(1, ("incorrect version of cifs.upcall (expected" + " %d but got %d)", + CIFS_SPNEGO_UPCALL_VERSION, msg->version)); + rc = -EKEYREJECTED; + goto ssetup_exit; + } /* bail out if key is too long */ if (msg->sesskey_len > sizeof(ses->server->mac_signing_key.data.krb5)) { -- cgit v1.2.3 From e9775843ecb039318dbc9ded6da9c762bff28a0b Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 26 Aug 2008 18:22:50 +0000 Subject: [CIFS] Correct keys dependency for cifs kerberos support Must also depend on CIFS ... Signed-off-by: Steve French --- fs/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index f0427105a61..3fab3901e0e 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1984,7 +1984,7 @@ config CIFS_EXPERIMENTAL config CIFS_UPCALL bool "Kerberos/SPNEGO advanced session setup (EXPERIMENTAL)" - depends on KEYS + depends on CIFS && KEYS help Enables an upcall mechanism for CIFS which accesses userspace helper utilities to provide SPNEGO packaged (RFC 4178) -- cgit v1.2.3 From 96c2a1137b9e00bcdbe3a95113ea8f42ca994f76 Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 26 Aug 2008 18:32:28 +0000 Subject: [CIFS] Reorder cifs config item for better clarity Signed-off-by: Steve French --- fs/Kconfig | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 3fab3901e0e..abccb5dab9a 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1930,6 +1930,16 @@ config CIFS_WEAK_PW_HASH If unsure, say N. +config CIFS_UPCALL + bool "Kerberos/SPNEGO advanced session setup" + depends on CIFS && KEYS + help + Enables an upcall mechanism for CIFS which accesses + userspace helper utilities to provide SPNEGO packaged (RFC 4178) + Kerberos tickets which are needed to mount to certain secure servers + (for which more secure Kerberos authentication is required). If + unsure, say N. + config CIFS_XATTR bool "CIFS extended attributes" depends on CIFS @@ -1982,16 +1992,6 @@ config CIFS_EXPERIMENTAL (which is disabled by default). See the file fs/cifs/README for more details. If unsure, say N. -config CIFS_UPCALL - bool "Kerberos/SPNEGO advanced session setup (EXPERIMENTAL)" - depends on CIFS && KEYS - help - Enables an upcall mechanism for CIFS which accesses - userspace helper utilities to provide SPNEGO packaged (RFC 4178) - Kerberos tickets which are needed to mount to certain secure servers - (for which more secure Kerberos authentication is required). If - unsure, say N. - config CIFS_DFS_UPCALL bool "DFS feature support (EXPERIMENTAL)" depends on CIFS_EXPERIMENTAL -- cgit v1.2.3 From 48fd4f93a00eac844678629f2f00518e146ed30d Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Fri, 22 Aug 2008 10:00:36 +0200 Subject: block: submit_bh() inadvertently discards barrier flag on a sync write Reported by Milan Broz , commit 18ce3751 inadvertently made submit_bh() discard the barrier bit for a WRITE_SYNC request. Fix that up. Signed-off-by: Jens Axboe --- fs/buffer.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 38653e36e22..ac78d4c19b3 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -2926,14 +2926,17 @@ int submit_bh(int rw, struct buffer_head * bh) BUG_ON(!buffer_mapped(bh)); BUG_ON(!bh->b_end_io); - if (buffer_ordered(bh) && (rw == WRITE)) - rw = WRITE_BARRIER; + /* + * Mask in barrier bit for a write (could be either a WRITE or a + * WRITE_SYNC + */ + if (buffer_ordered(bh) && (rw & WRITE)) + rw |= WRITE_BARRIER; /* - * Only clear out a write error when rewriting, should this - * include WRITE_SYNC as well? + * Only clear out a write error when rewriting */ - if (test_set_buffer_req(bh) && (rw == WRITE || rw == WRITE_BARRIER)) + if (test_set_buffer_req(bh) && (rw & WRITE)) clear_buffer_write_io_error(bh); /* -- cgit v1.2.3 From 76029ff37f31dad64641489c610d98955217bb68 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Mon, 25 Aug 2008 20:36:08 +0200 Subject: bio: fix bio_copy_kern() handling of bio->bv_len The commit 68154e90c9d1492d570671ae181d9a8f8530da55 introduced bio_copy_kern() to add bounce support to blk_rq_map_kern. bio_copy_kern() uses bio->bv_len to copy data for READ commands after the completion but it doesn't work with a request that partially completed. SCSI always completes a PC request as a whole but seems some don't. This patch fixes bio_copy_kern to handle the above case. As bio_copy_user does, bio_copy_kern uses struct bio_map_data to store struct bio_vec. Signed-off-by: FUJITA Tomonori Reported-by: Nix Tested-by: Nix Cc: stable@kernel.org Signed-off-by: Jens Axboe --- fs/bio.c | 38 ++++++++++++++++++++++++++++---------- 1 file changed, 28 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/bio.c b/fs/bio.c index 8000e2fa16c..8b1f5ee6f83 100644 --- a/fs/bio.c +++ b/fs/bio.c @@ -469,20 +469,21 @@ static void bio_free_map_data(struct bio_map_data *bmd) kfree(bmd); } -static struct bio_map_data *bio_alloc_map_data(int nr_segs, int iov_count) +static struct bio_map_data *bio_alloc_map_data(int nr_segs, int iov_count, + gfp_t gfp_mask) { - struct bio_map_data *bmd = kmalloc(sizeof(*bmd), GFP_KERNEL); + struct bio_map_data *bmd = kmalloc(sizeof(*bmd), gfp_mask); if (!bmd) return NULL; - bmd->iovecs = kmalloc(sizeof(struct bio_vec) * nr_segs, GFP_KERNEL); + bmd->iovecs = kmalloc(sizeof(struct bio_vec) * nr_segs, gfp_mask); if (!bmd->iovecs) { kfree(bmd); return NULL; } - bmd->sgvecs = kmalloc(sizeof(struct sg_iovec) * iov_count, GFP_KERNEL); + bmd->sgvecs = kmalloc(sizeof(struct sg_iovec) * iov_count, gfp_mask); if (bmd->sgvecs) return bmd; @@ -596,7 +597,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q, struct sg_iovec *iov, len += iov[i].iov_len; } - bmd = bio_alloc_map_data(nr_pages, iov_count); + bmd = bio_alloc_map_data(nr_pages, iov_count, GFP_KERNEL); if (!bmd) return ERR_PTR(-ENOMEM); @@ -942,19 +943,22 @@ static void bio_copy_kern_endio(struct bio *bio, int err) { struct bio_vec *bvec; const int read = bio_data_dir(bio) == READ; - char *p = bio->bi_private; + struct bio_map_data *bmd = bio->bi_private; int i; + char *p = bmd->sgvecs[0].iov_base; __bio_for_each_segment(bvec, bio, i, 0) { char *addr = page_address(bvec->bv_page); + int len = bmd->iovecs[i].bv_len; if (read && !err) - memcpy(p, addr, bvec->bv_len); + memcpy(p, addr, len); __free_page(bvec->bv_page); - p += bvec->bv_len; + p += len; } + bio_free_map_data(bmd); bio_put(bio); } @@ -978,11 +982,21 @@ struct bio *bio_copy_kern(struct request_queue *q, void *data, unsigned int len, const int nr_pages = end - start; struct bio *bio; struct bio_vec *bvec; + struct bio_map_data *bmd; int i, ret; + struct sg_iovec iov; + + iov.iov_base = data; + iov.iov_len = len; + + bmd = bio_alloc_map_data(nr_pages, 1, gfp_mask); + if (!bmd) + return ERR_PTR(-ENOMEM); + ret = -ENOMEM; bio = bio_alloc(gfp_mask, nr_pages); if (!bio) - return ERR_PTR(-ENOMEM); + goto out_bmd; while (len) { struct page *page; @@ -1016,14 +1030,18 @@ struct bio *bio_copy_kern(struct request_queue *q, void *data, unsigned int len, } } - bio->bi_private = data; + bio->bi_private = bmd; bio->bi_end_io = bio_copy_kern_endio; + + bio_set_map_data(bmd, bio, &iov, 1); return bio; cleanup: bio_for_each_segment(bvec, bio, i) __free_page(bvec->bv_page); bio_put(bio); +out_bmd: + bio_free_map_data(bmd); return ERR_PTR(ret); } -- cgit v1.2.3 From aefcc28a3a63ac33a298777aa50ba43641c75241 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Mon, 25 Aug 2008 20:36:08 +0200 Subject: bio: fix __bio_copy_iov() handling of bio->bv_len The commit c5dec1c3034f1ae3503efbf641ff3b0273b64797 introduced __bio_copy_iov() to add bounce support to blk_rq_map_user_iov. __bio_copy_iov() uses bio->bv_len to copy data for READ commands after the completion but it doesn't work with a request that partially completed. SCSI always completes a PC request as a whole but seems some don't. Signed-off-by: FUJITA Tomonori Cc: stable@kernel.org Signed-off-by: Jens Axboe --- fs/bio.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/bio.c b/fs/bio.c index 8b1f5ee6f83..3cba7ae34d7 100644 --- a/fs/bio.c +++ b/fs/bio.c @@ -492,8 +492,8 @@ static struct bio_map_data *bio_alloc_map_data(int nr_segs, int iov_count, return NULL; } -static int __bio_copy_iov(struct bio *bio, struct sg_iovec *iov, int iov_count, - int uncopy) +static int __bio_copy_iov(struct bio *bio, struct bio_vec *iovecs, + struct sg_iovec *iov, int iov_count, int uncopy) { int ret = 0, i; struct bio_vec *bvec; @@ -503,7 +503,7 @@ static int __bio_copy_iov(struct bio *bio, struct sg_iovec *iov, int iov_count, __bio_for_each_segment(bvec, bio, i, 0) { char *bv_addr = page_address(bvec->bv_page); - unsigned int bv_len = bvec->bv_len; + unsigned int bv_len = iovecs[i].bv_len; while (bv_len && iov_idx < iov_count) { unsigned int bytes; @@ -555,7 +555,7 @@ int bio_uncopy_user(struct bio *bio) struct bio_map_data *bmd = bio->bi_private; int ret; - ret = __bio_copy_iov(bio, bmd->sgvecs, bmd->nr_sgvecs, 1); + ret = __bio_copy_iov(bio, bmd->iovecs, bmd->sgvecs, bmd->nr_sgvecs, 1); bio_free_map_data(bmd); bio_put(bio); @@ -634,7 +634,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q, struct sg_iovec *iov, * success */ if (!write_to_vm) { - ret = __bio_copy_iov(bio, iov, iov_count, 0); + ret = __bio_copy_iov(bio, bio->bi_io_vec, iov, iov_count, 0); if (ret) goto cleanup; } -- cgit v1.2.3 From 87ed1d65fb536a0cd4e84874c0b038f953e448aa Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 27 Aug 2008 17:53:30 +0000 Subject: [CIFS] Add destroy routine for dns_resolver Otherwise, we're leaking the payload memory. CC: Stable Kernel Acked-by: David Howells Signed-off-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/CHANGES | 3 ++- fs/cifs/dns_resolve.c | 7 +++++++ 2 files changed, 9 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES index 526041a52d3..f9e4ad97a79 100644 --- a/fs/cifs/CHANGES +++ b/fs/cifs/CHANGES @@ -8,7 +8,8 @@ architectures. Fix problems with preserving timestamps on copying open files (e.g. "cp -a") to Windows servers. For mkdir and create honor setgid bit on parent directory when server supports Unix Extensions but not POSIX create. Update cifs.upcall version to handle new Kerberos sec flags -(this requires update of cifs.upcall program from Samba). +(this requires update of cifs.upcall program from Samba). Fix memory leak +on dns_upcall (resolving DFS referralls). Version 1.53 ------------ diff --git a/fs/cifs/dns_resolve.c b/fs/cifs/dns_resolve.c index f730ef35499..a2e0673e1b0 100644 --- a/fs/cifs/dns_resolve.c +++ b/fs/cifs/dns_resolve.c @@ -47,11 +47,18 @@ static int dns_resolver_instantiate(struct key *key, const void *data, return rc; } +static void +dns_resolver_destroy(struct key *key) +{ + kfree(key->payload.data); +} + struct key_type key_type_dns_resolver = { .name = "dns_resolver", .def_datalen = sizeof(struct in_addr), .describe = user_describe, .instantiate = dns_resolver_instantiate, + .destroy = dns_resolver_destroy, .match = user_match, }; -- cgit v1.2.3 From bcc55c6664a90146149ba0fd93052adc94287b9f Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 27 Aug 2008 21:30:22 +0000 Subject: [CIFS] Fix plaintext authentication The last eight bytes of the password field were not cleared when doing lanman plaintext password authentication. This patch fixes that. I tested it with Samba by setting password encryption to no in the server's smb.conf. Other servers also can be configured to force plaintext authentication. Note that plaintexti authentication requires setting /proc/fs/cifs/SecurityFlags to 0x30030 on the client (enabling both LANMAN and also plaintext password support). Also note that LANMAN support (and thus plaintext password support) requires CONFIG_CIFS_WEAK_PW_HASH to be enabled in menuconfig. CC: Jeff Layton CC: Stable Kernel Signed-off-by: Steve French --- fs/cifs/cifsencrypt.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c index 83fd40dc1ef..bd5f13d3845 100644 --- a/fs/cifs/cifsencrypt.c +++ b/fs/cifs/cifsencrypt.c @@ -294,6 +294,7 @@ void calc_lanman_hash(struct cifsSesInfo *ses, char *lnm_session_key) if ((ses->server->secMode & SECMODE_PW_ENCRYPT) == 0) if (extended_security & CIFSSEC_MAY_PLNTXT) { + memset(lnm_session_key, 0, CIFS_SESS_KEY_SIZE); memcpy(lnm_session_key, password_with_pad, CIFS_ENCPWD_SIZE); return; -- cgit v1.2.3 From 838726c4756813576078203eb7e1e219db0da870 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 28 Aug 2008 07:54:59 -0400 Subject: cifs: fix O_APPEND on directio mounts The direct I/O write codepath for CIFS is done through cifs_user_write(). That function does not currently call generic_write_checks() so the file position isn't being properly set when the file is opened with O_APPEND. It's also not doing the other "normal" checks that should be done for a write call. The problem is currently that when you open a file with O_APPEND on a mount with the directio mount option, the file position is set to the beginning of the file. This makes any subsequent writes clobber the data in the file starting at the beginning. This seems to fix the problem in cursory testing. It is, however important to note that NFS disallows the combination of (O_DIRECT|O_APPEND). If my understanding is correct, the concern is races with multiple clients appending to a file clobbering each others' data. Since the write model for CIFS and NFS is pretty similar in this regard, CIFS is probably subject to the same sort of races. What's unclear to me is why this is a particular problem with O_DIRECT and not with buffered writes... Regardless, disallowing O_APPEND on an entire mount is probably not reasonable, so we'll probably just have to deal with it and reevaluate this flag combination when we get proper support for O_DIRECT. In the meantime this patch at least fixes the existing problem. Signed-off-by: Jeff Layton Cc: Stable Tree Signed-off-by: Steve French --- fs/cifs/file.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/cifs/file.c b/fs/cifs/file.c index ff14d14903a..cbefe1f1f9f 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -833,6 +833,10 @@ ssize_t cifs_user_write(struct file *file, const char __user *write_data, return -EBADF; open_file = (struct cifsFileInfo *) file->private_data; + rc = generic_write_checks(file, poffset, &write_size, 0); + if (rc) + return rc; + xid = GetXid(); if (*poffset > file->f_path.dentry->d_inode->i_size) -- cgit v1.2.3 From 2e655021b8d50b5d90ce442f3de6bf3667729910 Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 28 Aug 2008 15:30:06 +0000 Subject: [CIFS] update cifs change log Signed-off-by: Steve French --- fs/cifs/CHANGES | 5 ++++- fs/cifs/README | 14 ++++++++++++-- 2 files changed, 16 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES index f9e4ad97a79..06e521a945c 100644 --- a/fs/cifs/CHANGES +++ b/fs/cifs/CHANGES @@ -9,7 +9,10 @@ files (e.g. "cp -a") to Windows servers. For mkdir and create honor setgid bit on parent directory when server supports Unix Extensions but not POSIX create. Update cifs.upcall version to handle new Kerberos sec flags (this requires update of cifs.upcall program from Samba). Fix memory leak -on dns_upcall (resolving DFS referralls). +on dns_upcall (resolving DFS referralls). Fix plain text password +authentication (requires setting SecurityFlags to 0x30030 to enable +lanman and plain text though). Fix writes to be at correct offset when +file is open with O_APPEND and file is on a directio (forcediretio) mount. Version 1.53 ------------ diff --git a/fs/cifs/README b/fs/cifs/README index 68b5c1169d9..bd2343d4c6a 100644 --- a/fs/cifs/README +++ b/fs/cifs/README @@ -542,10 +542,20 @@ SecurityFlags Flags which control security negotiation and hashing mechanisms (as "must use") on the other hand does not make much sense. Default flags are 0x07007 - (NTLM, NTLMv2 and packet signing allowed). Maximum + (NTLM, NTLMv2 and packet signing allowed). The maximum allowable flags if you want to allow mounts to servers using weaker password hashes is 0x37037 (lanman, - plaintext, ntlm, ntlmv2, signing allowed): + plaintext, ntlm, ntlmv2, signing allowed). Some + SecurityFlags require the corresponding menuconfig + options to be enabled (lanman and plaintext require + CONFIG_CIFS_WEAK_PW_HASH for example). Enabling + plaintext authentication currently requires also + enabling lanman authentication in the security flags + because the cifs module only supports sending + laintext passwords using the older lanman dialect + form of the session setup SMB. (e.g. for authentication + using plain text passwords, set the SecurityFlags + to 0x30030): may use packet signing 0x00001 must use packet signing 0x01001 -- cgit v1.2.3 From c76da9da1fffa6de263486df54950eb328d58f71 Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 28 Aug 2008 15:32:22 +0000 Subject: [CIFS] Turn off Unicode during session establishment for plaintext authentication LANMAN session setup did not support Unicode (after session setup, unicode can still be used though). Fixes samba bug# 5319 CC: Jeff Layton CC: Stable Kernel Signed-off-by: Steve French --- fs/cifs/sess.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index b537fad3bf5..252fdc0567f 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -409,6 +409,8 @@ CIFS_SessSetup(unsigned int xid, struct cifsSesInfo *ses, int first_time, #ifdef CONFIG_CIFS_WEAK_PW_HASH char lnm_session_key[CIFS_SESS_KEY_SIZE]; + pSMB->req.hdr.Flags2 &= ~SMBFLG2_UNICODE; + /* no capabilities flags in old lanman negotiation */ pSMB->old_req.PasswordLength = cpu_to_le16(CIFS_SESS_KEY_SIZE); -- cgit v1.2.3 From 8191e1fa8131a422f4bf7b0f2dc1f8543fd17783 Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Fri, 22 Aug 2008 12:26:40 +0300 Subject: UBIFS: do not update min_idx_lebs in stafs This is bad because the rest of the code should not depend on it, and this may hide bugss, instead of revealing them. Signed-off-by: Artem Bityutskiy --- fs/ubifs/budget.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index 15409815747..ac0d2e1e73b 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -741,7 +741,6 @@ long long ubifs_budg_get_free_space(struct ubifs_info *c) available = ubifs_calc_available(c, min_idx_lebs); outstanding = c->budg_data_growth + c->budg_dd_growth; - c->min_idx_lebs = min_idx_lebs; spin_unlock(&c->space_lock); if (available > outstanding) -- cgit v1.2.3 From 9e5de3549615818cae9c20a0ee1fd3ad4a747758 Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Mon, 25 Aug 2008 17:29:43 +0300 Subject: UBIFS: push empty flash hack down We have a hack which forces the amount of flash space to be equivalent to 'c->blocks_cnt' in case of empty FS. This is to make users happy and see '%0' used in 'df' when they mount an empty FS. This hack is not needed in 'ubifs_calc_available()', but it is only needed the caller, in 'ubifs_budg_get_free_space()'. So push it down there. Signed-off-by: Artem Bityutskiy --- fs/ubifs/budget.c | 24 +++++++++++------------- fs/ubifs/super.c | 2 -- 2 files changed, 11 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index ac0d2e1e73b..f6d2eaa7a06 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -302,18 +302,6 @@ long long ubifs_calc_available(const struct ubifs_info *c, int min_idx_lebs) int subtract_lebs; long long available; - /* - * Force the amount available to the total size reported if the used - * space is zero. - */ - if (c->lst.total_used <= UBIFS_INO_NODE_SZ && - c->budg_data_growth + c->budg_dd_growth == 0) { - /* Do the same calculation as for c->block_cnt */ - available = c->main_lebs - 2; - available *= c->leb_size - c->dark_wm; - return available; - } - available = c->main_bytes - c->lst.total_used; /* @@ -739,8 +727,18 @@ long long ubifs_budg_get_free_space(struct ubifs_info *c) return 0; } - available = ubifs_calc_available(c, min_idx_lebs); outstanding = c->budg_data_growth + c->budg_dd_growth; + + /* + * Force the amount available to the total size reported if the used + * space is zero. + */ + if (c->lst.total_used <= UBIFS_INO_NODE_SZ && !outstanding) { + spin_unlock(&c->space_lock); + return (long long)c->block_cnt << UBIFS_BLOCK_SHIFT; + } + + available = ubifs_calc_available(c, min_idx_lebs); spin_unlock(&c->space_lock); if (available > outstanding) diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index f71e6b8822c..1018053519e 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -649,8 +649,6 @@ static int init_constants_late(struct ubifs_info *c) * * Subtract the LEB reserved for GC and the LEB which is reserved for * deletions. - * - * Review 'ubifs_calc_available()' if changing this calculation. */ tmp64 = c->main_lebs - 2; tmp64 *= (uint64_t)c->leb_size - c->dark_wm; -- cgit v1.2.3 From 8aabb75017291ba68c09ff5fdb998ef0a1fdaaf9 Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Mon, 25 Aug 2008 16:02:31 +0300 Subject: UBIFS: remove incorrect index space check When we report free space to user-space, we should not report 0 if the amount of empty LEBs is too low, because they would be produced by GC when needed. Thus, just call 'ubifs_calc_available()' straight away which would take 'min_idx_lebs' into account anyway. Signed-off-by: Artem Bityutskiy --- fs/ubifs/budget.c | 15 +-------------- 1 file changed, 1 insertion(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index f6d2eaa7a06..9ef630a594c 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -709,24 +709,11 @@ void ubifs_release_dirty_inode_budget(struct ubifs_info *c, */ long long ubifs_budg_get_free_space(struct ubifs_info *c) { - int min_idx_lebs, rsvd_idx_lebs; + int min_idx_lebs; long long available, outstanding, free; - /* Do exactly the same calculations as in 'do_budget_space()' */ spin_lock(&c->space_lock); min_idx_lebs = ubifs_calc_min_idx_lebs(c); - - if (min_idx_lebs > c->lst.idx_lebs) - rsvd_idx_lebs = min_idx_lebs - c->lst.idx_lebs; - else - rsvd_idx_lebs = 0; - - if (rsvd_idx_lebs > c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt - - c->lst.taken_empty_lebs) { - spin_unlock(&c->space_lock); - return 0; - } - outstanding = c->budg_data_growth + c->budg_dd_growth; /* -- cgit v1.2.3 From 4b5f2762ec914c9dfd0e9d2377c0574f2ee9a8f9 Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Mon, 25 Aug 2008 16:15:56 +0300 Subject: UBIFS: improve statfs reporting Make free space calculation less pessimistic and more realistic, which in turn improves 'statfs()' reports. Now it lies by 10%-20%, instead of 20%-30% (10% more honest). Results of "freespace" test (120MiB volume, 16KiB LEB size, 512 bytes page size). Before the change: freespace: Test 1: fill the space we have 3 times freespace: was free: 78274560 bytes 74.6 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 18214912 bytes 17.4 MiB, wrote 23.3% more than predicted freespace: was free: 76754944 bytes 73.2 MiB, wrote: 96493568 bytes 92.0 MiB, delta: 19738624 bytes 18.8 MiB, wrote 25.7% more than predicted freespace: was free: 76759040 bytes 73.2 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 19730432 bytes 18.8 MiB, wrote 25.7% more than predicted freespace: Test 1 finished freespace: Test 2: gradually lessen amount of free space and fill the FS freespace: do 10 steps, lessen free space by 6977722 bytes 6.7 MiB each time freespace: was free: 72273920 bytes 68.9 MiB, wrote: 88891392 bytes 84.8 MiB, delta: 16617472 bytes 15.8 MiB, wrote 23.0% more than predicted freespace: was free: 66154496 bytes 63.1 MiB, wrote: 81506304 bytes 77.7 MiB, delta: 15351808 bytes 14.6 MiB, wrote 23.2% more than predicted freespace: was free: 58732544 bytes 56.0 MiB, wrote: 72572928 bytes 69.2 MiB, delta: 13840384 bytes 13.2 MiB, wrote 23.6% more than predicted freespace: was free: 51552256 bytes 49.2 MiB, wrote: 63754240 bytes 60.8 MiB, delta: 12201984 bytes 11.6 MiB, wrote 23.7% more than predicted freespace: was free: 44404736 bytes 42.3 MiB, wrote: 54943744 bytes 52.4 MiB, delta: 10539008 bytes 10.1 MiB, wrote 23.7% more than predicted freespace: was free: 37285888 bytes 35.6 MiB, wrote: 46161920 bytes 44.0 MiB, delta: 8876032 bytes 8.5 MiB, wrote 23.8% more than predicted freespace: was free: 30171136 bytes 28.8 MiB, wrote: 37384192 bytes 35.7 MiB, delta: 7213056 bytes 6.9 MiB, wrote 23.9% more than predicted freespace: was free: 23048192 bytes 22.0 MiB, wrote: 28606464 bytes 27.3 MiB, delta: 5558272 bytes 5.3 MiB, wrote 24.1% more than predicted freespace: was free: 15941632 bytes 15.2 MiB, wrote: 19828736 bytes 18.9 MiB, delta: 3887104 bytes 3.7 MiB, wrote 24.4% more than predicted freespace: was free: 8830976 bytes 8.4 MiB, wrote: 11063296 bytes 10.6 MiB, delta: 2232320 bytes 2.1 MiB, wrote 25.3% more than predicted freespace: Test 2 finished freespace: Test 3: gradually lessen amount of free space by trashing and fill the FS freespace: do 10 steps, lessen free space by 6985541 bytes 6.7 MiB each time freespace: trashing: was free: 76840960 bytes 73.3 MiB, need free: 6985550 bytes 6.7 MiB, files created: 248311, delete 225737 (90.9% of them) freespace: was free: 65228800 bytes 62.2 MiB, wrote: 82530304 bytes 78.7 MiB, delta: 17301504 bytes 16.5 MiB, wrote 26.5% more than predicted freespace: trashing: was free: 74485760 bytes 71.0 MiB, need free: 13971091 bytes 13.3 MiB, files created: 248712, delete 202061 (81.2% of them) freespace: was free: 55025664 bytes 52.5 MiB, wrote: 71925760 bytes 68.6 MiB, delta: 16900096 bytes 16.1 MiB, wrote 30.7% more than predicted freespace: trashing: was free: 75550720 bytes 72.1 MiB, need free: 20956632 bytes 20.0 MiB, files created: 248849, delete 179822 (72.3% of them) freespace: was free: 46669824 bytes 44.5 MiB, wrote: 63197184 bytes 60.3 MiB, delta: 16527360 bytes 15.8 MiB, wrote 35.4% more than predicted freespace: trashing: was free: 76214272 bytes 72.7 MiB, need free: 27942173 bytes 26.6 MiB, files created: 248789, delete 157576 (63.3% of them) freespace: was free: 39129088 bytes 37.3 MiB, wrote: 55164928 bytes 52.6 MiB, delta: 16035840 bytes 15.3 MiB, wrote 41.0% more than predicted freespace: trashing: was free: 77398016 bytes 73.8 MiB, need free: 34927714 bytes 33.3 MiB, files created: 248711, delete 136474 (54.9% of them) freespace: was free: 32325632 bytes 30.8 MiB, wrote: 48234496 bytes 46.0 MiB, delta: 15908864 bytes 15.2 MiB, wrote 49.2% more than predicted freespace: trashing: was free: 75796480 bytes 72.3 MiB, need free: 41913255 bytes 40.0 MiB, files created: 248674, delete 111164 (44.7% of them) freespace: was free: 25079808 bytes 23.9 MiB, wrote: 40775680 bytes 38.9 MiB, delta: 15695872 bytes 15.0 MiB, wrote 62.6% more than predicted freespace: trashing: was free: 78209024 bytes 74.6 MiB, need free: 48898796 bytes 46.6 MiB, files created: 248708, delete 93207 (37.5% of them) freespace: was free: 20582400 bytes 19.6 MiB, wrote: 34844672 bytes 33.2 MiB, delta: 14262272 bytes 13.6 MiB, wrote 69.3% more than predicted freespace: trashing: was free: 77328384 bytes 73.7 MiB, need free: 55884337 bytes 53.3 MiB, files created: 248644, delete 68951 (27.7% of them) freespace: was free: 14368768 bytes 13.7 MiB, wrote: 28278784 bytes 27.0 MiB, delta: 13910016 bytes 13.3 MiB, wrote 96.8% more than predicted freespace: trashing: was free: 77434880 bytes 73.8 MiB, need free: 62869878 bytes 60.0 MiB, files created: 248640, delete 46767 (18.8% of them) freespace: was free: 8286208 bytes 7.9 MiB, wrote: 21811200 bytes 20.8 MiB, delta: 13524992 bytes 12.9 MiB, wrote 163.2% more than predicted freespace: trashing: was free: 77856768 bytes 74.2 MiB, need free: 69855419 bytes 66.6 MiB, files created: 248576, delete 25546 (10.3% of them) freespace: was free: 5570560 bytes 5.3 MiB, wrote: 8187904 bytes 7.8 MiB, delta: 2617344 bytes 2.5 MiB, wrote 47.0% more than predicted freespace: Test 3 finished freespace: finished successfully After the change: freespace: Test 1: fill the space we have 3 times freespace: was free: 85204992 bytes 81.3 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 11284480 bytes 10.8 MiB, wrote 13.2% more than predicted freespace: was free: 83554304 bytes 79.7 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 12935168 bytes 12.3 MiB, wrote 15.5% more than predicted freespace: was free: 83554304 bytes 79.7 MiB, wrote: 96493568 bytes 92.0 MiB, delta: 12939264 bytes 12.3 MiB, wrote 15.5% more than predicted freespace: Test 1 finished freespace: Test 2: gradually lessen amount of free space and fill the FS freespace: do 10 steps, lessen free space by 7596218 bytes 7.2 MiB each time freespace: was free: 78675968 bytes 75.0 MiB, wrote: 88903680 bytes 84.8 MiB, delta: 10227712 bytes 9.8 MiB, wrote 13.0% more than predicted freespace: was free: 72015872 bytes 68.7 MiB, wrote: 81514496 bytes 77.7 MiB, delta: 9498624 bytes 9.1 MiB, wrote 13.2% more than predicted freespace: was free: 63938560 bytes 61.0 MiB, wrote: 72589312 bytes 69.2 MiB, delta: 8650752 bytes 8.2 MiB, wrote 13.5% more than predicted freespace: was free: 56127488 bytes 53.5 MiB, wrote: 63762432 bytes 60.8 MiB, delta: 7634944 bytes 7.3 MiB, wrote 13.6% more than predicted freespace: was free: 48336896 bytes 46.1 MiB, wrote: 54935552 bytes 52.4 MiB, delta: 6598656 bytes 6.3 MiB, wrote 13.7% more than predicted freespace: was free: 40587264 bytes 38.7 MiB, wrote: 46157824 bytes 44.0 MiB, delta: 5570560 bytes 5.3 MiB, wrote 13.7% more than predicted freespace: was free: 32841728 bytes 31.3 MiB, wrote: 37384192 bytes 35.7 MiB, delta: 4542464 bytes 4.3 MiB, wrote 13.8% more than predicted freespace: was free: 25100288 bytes 23.9 MiB, wrote: 28618752 bytes 27.3 MiB, delta: 3518464 bytes 3.4 MiB, wrote 14.0% more than predicted freespace: was free: 17342464 bytes 16.5 MiB, wrote: 19841024 bytes 18.9 MiB, delta: 2498560 bytes 2.4 MiB, wrote 14.4% more than predicted freespace: was free: 9605120 bytes 9.2 MiB, wrote: 11063296 bytes 10.6 MiB, delta: 1458176 bytes 1.4 MiB, wrote 15.2% more than predicted freespace: Test 2 finished freespace: Test 3: gradually lessen amount of free space by trashing and fill the FS freespace: do 10 steps, lessen free space by 7606272 bytes 7.3 MiB each time freespace: trashing: was free: 83668992 bytes 79.8 MiB, need free: 7606272 bytes 7.3 MiB, files created: 248297, delete 225724 (90.9% of them) freespace: was free: 70803456 bytes 67.5 MiB, wrote: 82485248 bytes 78.7 MiB, delta: 11681792 bytes 11.1 MiB, wrote 16.5% more than predicted freespace: trashing: was free: 81080320 bytes 77.3 MiB, need free: 15212544 bytes 14.5 MiB, files created: 248711, delete 202047 (81.2% of them) freespace: was free: 59867136 bytes 57.1 MiB, wrote: 71897088 bytes 68.6 MiB, delta: 12029952 bytes 11.5 MiB, wrote 20.1% more than predicted freespace: trashing: was free: 82243584 bytes 78.4 MiB, need free: 22818816 bytes 21.8 MiB, files created: 248866, delete 179817 (72.3% of them) freespace: was free: 50905088 bytes 48.5 MiB, wrote: 63168512 bytes 60.2 MiB, delta: 12263424 bytes 11.7 MiB, wrote 24.1% more than predicted freespace: trashing: was free: 83402752 bytes 79.5 MiB, need free: 30425088 bytes 29.0 MiB, files created: 248920, delete 158114 (63.5% of them) freespace: was free: 42651648 bytes 40.7 MiB, wrote: 55406592 bytes 52.8 MiB, delta: 12754944 bytes 12.2 MiB, wrote 29.9% more than predicted freespace: trashing: was free: 84402176 bytes 80.5 MiB, need free: 38031360 bytes 36.3 MiB, files created: 248709, delete 136641 (54.9% of them) freespace: was free: 35233792 bytes 33.6 MiB, wrote: 48250880 bytes 46.0 MiB, delta: 13017088 bytes 12.4 MiB, wrote 36.9% more than predicted freespace: trashing: was free: 82530304 bytes 78.7 MiB, need free: 45637632 bytes 43.5 MiB, files created: 248778, delete 111208 (44.7% of them) freespace: was free: 27287552 bytes 26.0 MiB, wrote: 40267776 bytes 38.4 MiB, delta: 12980224 bytes 12.4 MiB, wrote 47.6% more than predicted freespace: trashing: was free: 85114880 bytes 81.2 MiB, need free: 53243904 bytes 50.8 MiB, files created: 248508, delete 93052 (37.4% of them) freespace: was free: 22437888 bytes 21.4 MiB, wrote: 35328000 bytes 33.7 MiB, delta: 12890112 bytes 12.3 MiB, wrote 57.4% more than predicted freespace: trashing: was free: 84103168 bytes 80.2 MiB, need free: 60850176 bytes 58.0 MiB, files created: 248637, delete 68743 (27.6% of them) freespace: was free: 15536128 bytes 14.8 MiB, wrote: 28319744 bytes 27.0 MiB, delta: 12783616 bytes 12.2 MiB, wrote 82.3% more than predicted freespace: trashing: was free: 84357120 bytes 80.4 MiB, need free: 68456448 bytes 65.3 MiB, files created: 248567, delete 46852 (18.8% of them) freespace: was free: 9015296 bytes 8.6 MiB, wrote: 22044672 bytes 21.0 MiB, delta: 13029376 bytes 12.4 MiB, wrote 144.5% more than predicted freespace: trashing: was free: 84942848 bytes 81.0 MiB, need free: 76062720 bytes 72.5 MiB, files created: 248636, delete 25993 (10.5% of them) freespace: was free: 6086656 bytes 5.8 MiB, wrote: 8331264 bytes 7.9 MiB, delta: 2244608 bytes 2.1 MiB, wrote 36.9% more than predicted freespace: Test 3 finished freespace: finished successfully Signed-off-by: Artem Bityutskiy --- fs/ubifs/budget.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ fs/ubifs/misc.h | 32 -------------------------------- fs/ubifs/ubifs.h | 1 + 3 files changed, 46 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index 9ef630a594c..7851480a6ce 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -701,6 +701,51 @@ void ubifs_release_dirty_inode_budget(struct ubifs_info *c, ubifs_release_budget(c, &req); } +/** + * ubifs_reported_space - calculate reported free space. + * @c: the UBIFS file-system description object + * @free: amount of free space + * + * This function calculates amount of free space which will be reported to + * user-space. User-space application tend to expect that if the file-system + * (e.g., via the 'statfs()' call) reports that it has N bytes available, they + * are able to write a file of size N. UBIFS attaches node headers to each data + * node and it has to write indexind nodes as well. This introduces additional + * overhead, and UBIFS it has to report sligtly less free space to meet the + * above expectetion. + * + * This function assumes free space is made up of uncompressed data nodes and + * full index nodes (one per data node, tripled because we always allow enough + * space to write the index thrice). + * + * Note, the calculation is pessimistic, which means that most of the time + * UBIFS reports less space than it actually has. + */ +long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free) +{ + int divisor, factor; + + /* + * Reported space size is @free * X, where X is UBIFS block size + * divided by UBIFS block size + all overhead one data block + * introduces. The overhead is the node header + indexing overhead. + * + * Indexing overhead is calculations are based on the following + * formula: I = N/(f - 1) + 1, where I - number of indexing nodes, N - + * number of data nodes, f - fanout. Because effective UBIFS fanout is + * twice as less than maximum fanout, we assume that each data node + * introduces 3 * @c->max_idx_node_sz / (@c->fanout/2 - 1) bytes. + * Note, the multiplier 3 is because UBIFS reseves thrice as more space + * for the index. + */ + factor = UBIFS_BLOCK_SIZE; + divisor = UBIFS_MAX_DATA_NODE_SZ; + divisor += (c->max_idx_node_sz * 3) / ((c->fanout >> 1) - 1); + free *= factor; + do_div(free, divisor); + return free; +} + /** * ubifs_budg_get_free_space - return amount of free space. * @c: UBIFS file-system description object diff --git a/fs/ubifs/misc.h b/fs/ubifs/misc.h index 87ced4c74a6..4c12a9215d7 100644 --- a/fs/ubifs/misc.h +++ b/fs/ubifs/misc.h @@ -283,38 +283,6 @@ static inline void *ubifs_idx_key(const struct ubifs_info *c, return (void *)((struct ubifs_branch *)idx->branches)->key; } -/** - * ubifs_reported_space - calculate reported free space. - * @c: the UBIFS file-system description object - * @free: amount of free space - * - * This function calculates amount of free space which will be reported to - * user-space. User-space application tend to expect that if the file-system - * (e.g., via the 'statfs()' call) reports that it has N bytes available, they - * are able to write a file of size N. UBIFS attaches node headers to each data - * node and it has to write indexind nodes as well. This introduces additional - * overhead, and UBIFS it has to report sligtly less free space to meet the - * above expectetion. - * - * This function assumes free space is made up of uncompressed data nodes and - * full index nodes (one per data node, doubled because we always allow enough - * space to write the index twice). - * - * Note, the calculation is pessimistic, which means that most of the time - * UBIFS reports less space than it actually has. - */ -static inline long long ubifs_reported_space(const struct ubifs_info *c, - uint64_t free) -{ - int divisor, factor; - - divisor = UBIFS_MAX_DATA_NODE_SZ + (c->max_idx_node_sz * 3); - factor = UBIFS_MAX_DATA_NODE_SZ - UBIFS_DATA_NODE_SZ; - do_div(free, divisor); - - return free * factor; -} - /** * ubifs_current_time - round current time to time granularity. * @inode: inode diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 7828d69ca4f..681d46e1628 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1441,6 +1441,7 @@ void ubifs_cancel_ino_op(struct ubifs_info *c, struct inode *inode, long long ubifs_budg_get_free_space(struct ubifs_info *c); int ubifs_calc_min_idx_lebs(struct ubifs_info *c); void ubifs_convert_page_budget(struct ubifs_info *c); +long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free); long long ubifs_calc_available(const struct ubifs_info *c, int min_idx_lebs); /* find.c */ -- cgit v1.2.3 From ad507653a39e0d27404291e5d813683265388a20 Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Mon, 25 Aug 2008 18:32:57 +0300 Subject: UBIFS: fix assertion The assertion was incorrect, because it did not take into account free space. This patch also amends the comments correspondingly, and cleans them up a little. Signed-off-by: Artem Bityutskiy --- fs/ubifs/find.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/find.c b/fs/ubifs/find.c index adee7b5ddea..9fc55ae7b03 100644 --- a/fs/ubifs/find.c +++ b/fs/ubifs/find.c @@ -211,14 +211,8 @@ static const struct ubifs_lprops *scan_for_dirty(struct ubifs_info *c, * dirty index heap, and it falls-back to LPT scanning if the heaps are empty * or do not have an LEB which satisfies the @min_space criteria. * - * Note: - * o LEBs which have less than dead watermark of dirty space are never picked - * by this function; - * - * Returns zero and the LEB properties of - * found dirty LEB in case of success, %-ENOSPC if no dirty LEB was found and a - * negative error code in case of other failures. The returned LEB is marked as - * "taken". + * Note, LEBs which have less than dead watermark of free + dirty space are + * never picked by this function. * * The additional @pick_free argument controls if this function has to return a * free or freeable LEB if one is present. For example, GC must to set it to %1, @@ -231,6 +225,10 @@ static const struct ubifs_lprops *scan_for_dirty(struct ubifs_info *c, * * In addition @pick_free is set to %2 by the recovery process in order to * recover gc_lnum in which case an index LEB must not be returned. + * + * This function returns zero and the LEB properties of found dirty LEB in case + * of success, %-ENOSPC if no dirty LEB was found and a negative error code in + * case of other failures. The returned LEB is marked as "taken". */ int ubifs_find_dirty_leb(struct ubifs_info *c, struct ubifs_lprops *ret_lp, int min_space, int pick_free) @@ -317,7 +315,7 @@ int ubifs_find_dirty_leb(struct ubifs_info *c, struct ubifs_lprops *ret_lp, lp = idx_lp; if (lp) { - ubifs_assert(lp->dirty >= c->dead_wm); + ubifs_assert(lp->free + lp->dirty >= c->dead_wm); goto found; } -- cgit v1.2.3 From 131130b9a1e6e523c64b34137b14f88ae1382a6a Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Mon, 25 Aug 2008 18:34:45 +0300 Subject: UBIFS: add forgotten gc_idx_lebs component We add this component at other similar places, but not in this one. Signed-off-by: Artem Bityutskiy --- fs/ubifs/find.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ubifs/find.c b/fs/ubifs/find.c index 9fc55ae7b03..e045c8b5542 100644 --- a/fs/ubifs/find.c +++ b/fs/ubifs/find.c @@ -243,7 +243,7 @@ int ubifs_find_dirty_leb(struct ubifs_info *c, struct ubifs_lprops *ret_lp, int lebs, rsvd_idx_lebs = 0; spin_lock(&c->space_lock); - lebs = c->lst.empty_lebs; + lebs = c->lst.empty_lebs + c->idx_gc_cnt; lebs += c->freeable_cnt - c->lst.taken_empty_lebs; /* -- cgit v1.2.3 From 9bbb5726efb64e2cfed42f6eec07db80cd87e63b Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Fri, 22 Aug 2008 18:23:22 +0300 Subject: UBIFS: introduce LEB overhead This is a preparational patch for the following statfs() report fix. Signed-off-by: Artem Bityutskiy --- fs/ubifs/super.c | 6 ++++++ fs/ubifs/ubifs.h | 5 +++++ 2 files changed, 11 insertions(+) (limited to 'fs') diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 1018053519e..be23fd3cfd8 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -530,6 +530,12 @@ static int init_constants_early(struct ubifs_info *c) c->dead_wm = ALIGN(MIN_WRITE_SZ, c->min_io_size); c->dark_wm = ALIGN(UBIFS_MAX_NODE_SZ, c->min_io_size); + /* + * Calculate how many bytes would be wasted at the end of LEB if it was + * fully filled with data nodes of maximum size. This is used in + * calculations when reporting free space. + */ + c->leb_overhead = c->leb_size % UBIFS_MAX_DATA_NODE_SZ; return 0; } diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 681d46e1628..57e58541de2 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -995,6 +995,9 @@ struct ubifs_mount_opts { * @max_idx_node_sz: maximum indexing node aligned on 8-bytes boundary * @max_inode_sz: maximum possible inode size in bytes * @max_znode_sz: size of znode in bytes + * + * @leb_overhead: how many bytes are wasted in an LEB when it is filled with + * data nodes of maximum size - used in free space reporting * @dead_wm: LEB dead space watermark * @dark_wm: LEB dark space watermark * @block_cnt: count of 4KiB blocks on the FS @@ -1226,6 +1229,8 @@ struct ubifs_info { int max_idx_node_sz; long long max_inode_sz; int max_znode_sz; + + int leb_overhead; int dead_wm; int dark_wm; int block_cnt; -- cgit v1.2.3 From 7dad181bbe58b8fe9e170da28bcd5f6ec9addd6d Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Mon, 25 Aug 2008 18:58:19 +0300 Subject: UBIFS: improve statfs reporting even more Since free space we report in statfs is file size which should fit to the FS - change the way we calculate free space and use leb_overhead instead of dark_wm in calculations. Results of "freespace" test (120MiB volume, 16KiB LEB size, 512 bytes page size). Before the change: freespace: Test 1: fill the space we have 3 times freespace: was free: 85204992 bytes 81.3 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 11284480 bytes 10.8 MiB, wrote 13.2% more than predicted freespace: was free: 83554304 bytes 79.7 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 12935168 bytes 12.3 MiB, wrote 15.5% more than predicted freespace: was free: 83554304 bytes 79.7 MiB, wrote: 96493568 bytes 92.0 MiB, delta: 12939264 bytes 12.3 MiB, wrote 15.5% more than predicted freespace: Test 1 finished freespace: Test 2: gradually lessen amount of free space and fill the FS freespace: do 10 steps, lessen free space by 7596218 bytes 7.2 MiB each time freespace: was free: 78675968 bytes 75.0 MiB, wrote: 88903680 bytes 84.8 MiB, delta: 10227712 bytes 9.8 MiB, wrote 13.0% more than predicted freespace: was free: 72015872 bytes 68.7 MiB, wrote: 81514496 bytes 77.7 MiB, delta: 9498624 bytes 9.1 MiB, wrote 13.2% more than predicted freespace: was free: 63938560 bytes 61.0 MiB, wrote: 72589312 bytes 69.2 MiB, delta: 8650752 bytes 8.2 MiB, wrote 13.5% more than predicted freespace: was free: 56127488 bytes 53.5 MiB, wrote: 63762432 bytes 60.8 MiB, delta: 7634944 bytes 7.3 MiB, wrote 13.6% more than predicted freespace: was free: 48336896 bytes 46.1 MiB, wrote: 54935552 bytes 52.4 MiB, delta: 6598656 bytes 6.3 MiB, wrote 13.7% more than predicted freespace: was free: 40587264 bytes 38.7 MiB, wrote: 46157824 bytes 44.0 MiB, delta: 5570560 bytes 5.3 MiB, wrote 13.7% more than predicted freespace: was free: 32841728 bytes 31.3 MiB, wrote: 37384192 bytes 35.7 MiB, delta: 4542464 bytes 4.3 MiB, wrote 13.8% more than predicted freespace: was free: 25100288 bytes 23.9 MiB, wrote: 28618752 bytes 27.3 MiB, delta: 3518464 bytes 3.4 MiB, wrote 14.0% more than predicted freespace: was free: 17342464 bytes 16.5 MiB, wrote: 19841024 bytes 18.9 MiB, delta: 2498560 bytes 2.4 MiB, wrote 14.4% more than predicted freespace: was free: 9605120 bytes 9.2 MiB, wrote: 11063296 bytes 10.6 MiB, delta: 1458176 bytes 1.4 MiB, wrote 15.2% more than predicted freespace: Test 2 finished freespace: Test 3: gradually lessen amount of free space by trashing and fill the FS freespace: do 10 steps, lessen free space by 7606272 bytes 7.3 MiB each time freespace: trashing: was free: 83668992 bytes 79.8 MiB, need free: 7606272 bytes 7.3 MiB, files created: 248297, delete 225724 (90.9% of them) freespace: was free: 70803456 bytes 67.5 MiB, wrote: 82485248 bytes 78.7 MiB, delta: 11681792 bytes 11.1 MiB, wrote 16.5% more than predicted freespace: trashing: was free: 81080320 bytes 77.3 MiB, need free: 15212544 bytes 14.5 MiB, files created: 248711, delete 202047 (81.2% of them) freespace: was free: 59867136 bytes 57.1 MiB, wrote: 71897088 bytes 68.6 MiB, delta: 12029952 bytes 11.5 MiB, wrote 20.1% more than predicted freespace: trashing: was free: 82243584 bytes 78.4 MiB, need free: 22818816 bytes 21.8 MiB, files created: 248866, delete 179817 (72.3% of them) freespace: was free: 50905088 bytes 48.5 MiB, wrote: 63168512 bytes 60.2 MiB, delta: 12263424 bytes 11.7 MiB, wrote 24.1% more than predicted freespace: trashing: was free: 83402752 bytes 79.5 MiB, need free: 30425088 bytes 29.0 MiB, files created: 248920, delete 158114 (63.5% of them) freespace: was free: 42651648 bytes 40.7 MiB, wrote: 55406592 bytes 52.8 MiB, delta: 12754944 bytes 12.2 MiB, wrote 29.9% more than predicted freespace: trashing: was free: 84402176 bytes 80.5 MiB, need free: 38031360 bytes 36.3 MiB, files created: 248709, delete 136641 (54.9% of them) freespace: was free: 35233792 bytes 33.6 MiB, wrote: 48250880 bytes 46.0 MiB, delta: 13017088 bytes 12.4 MiB, wrote 36.9% more than predicted freespace: trashing: was free: 82530304 bytes 78.7 MiB, need free: 45637632 bytes 43.5 MiB, files created: 248778, delete 111208 (44.7% of them) freespace: was free: 27287552 bytes 26.0 MiB, wrote: 40267776 bytes 38.4 MiB, delta: 12980224 bytes 12.4 MiB, wrote 47.6% more than predicted freespace: trashing: was free: 85114880 bytes 81.2 MiB, need free: 53243904 bytes 50.8 MiB, files created: 248508, delete 93052 (37.4% of them) freespace: was free: 22437888 bytes 21.4 MiB, wrote: 35328000 bytes 33.7 MiB, delta: 12890112 bytes 12.3 MiB, wrote 57.4% more than predicted freespace: trashing: was free: 84103168 bytes 80.2 MiB, need free: 60850176 bytes 58.0 MiB, files created: 248637, delete 68743 (27.6% of them) freespace: was free: 15536128 bytes 14.8 MiB, wrote: 28319744 bytes 27.0 MiB, delta: 12783616 bytes 12.2 MiB, wrote 82.3% more than predicted freespace: trashing: was free: 84357120 bytes 80.4 MiB, need free: 68456448 bytes 65.3 MiB, files created: 248567, delete 46852 (18.8% of them) freespace: was free: 9015296 bytes 8.6 MiB, wrote: 22044672 bytes 21.0 MiB, delta: 13029376 bytes 12.4 MiB, wrote 144.5% more than predicted freespace: trashing: was free: 84942848 bytes 81.0 MiB, need free: 76062720 bytes 72.5 MiB, files created: 248636, delete 25993 (10.5% of them) freespace: was free: 6086656 bytes 5.8 MiB, wrote: 8331264 bytes 7.9 MiB, delta: 2244608 bytes 2.1 MiB, wrote 36.9% more than predicted freespace: Test 3 finished freespace: finished successfully After the change: freespace: Test 1: fill the space we have 3 times freespace: was free: 94048256 bytes 89.7 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 2441216 bytes 2.3 MiB, wrote 2.6% more than predicted freespace: was free: 92246016 bytes 88.0 MiB, wrote: 96493568 bytes 92.0 MiB, delta: 4247552 bytes 4.1 MiB, wrote 4.6% more than predicted freespace: was free: 92254208 bytes 88.0 MiB, wrote: 96489472 bytes 92.0 MiB, delta: 4235264 bytes 4.0 MiB, wrote 4.6% more than predicted freespace: Test 1 finished freespace: Test 2: gradually lessen amount of free space and fill the FS freespace: do 10 steps, lessen free space by 8386001 bytes 8.0 MiB each time freespace: was free: 86605824 bytes 82.6 MiB, wrote: 88252416 bytes 84.2 MiB, delta: 1646592 bytes 1.6 MiB, wrote 1.9% more than predicted freespace: was free: 78667776 bytes 75.0 MiB, wrote: 80715776 bytes 77.0 MiB, delta: 2048000 bytes 2.0 MiB, wrote 2.6% more than predicted freespace: was free: 69615616 bytes 66.4 MiB, wrote: 71630848 bytes 68.3 MiB, delta: 2015232 bytes 1.9 MiB, wrote 2.9% more than predicted freespace: was free: 61018112 bytes 58.2 MiB, wrote: 62783488 bytes 59.9 MiB, delta: 1765376 bytes 1.7 MiB, wrote 2.9% more than predicted freespace: was free: 52424704 bytes 50.0 MiB, wrote: 53968896 bytes 51.5 MiB, delta: 1544192 bytes 1.5 MiB, wrote 2.9% more than predicted freespace: was free: 43880448 bytes 41.8 MiB, wrote: 45199360 bytes 43.1 MiB, delta: 1318912 bytes 1.3 MiB, wrote 3.0% more than predicted freespace: was free: 35332096 bytes 33.7 MiB, wrote: 36425728 bytes 34.7 MiB, delta: 1093632 bytes 1.0 MiB, wrote 3.1% more than predicted freespace: was free: 26771456 bytes 25.5 MiB, wrote: 27643904 bytes 26.4 MiB, delta: 872448 bytes 852.0 KiB, wrote 3.3% more than predicted freespace: was free: 18231296 bytes 17.4 MiB, wrote: 18878464 bytes 18.0 MiB, delta: 647168 bytes 632.0 KiB, wrote 3.5% more than predicted freespace: was free: 9674752 bytes 9.2 MiB, wrote: 10088448 bytes 9.6 MiB, delta: 413696 bytes 404.0 KiB, wrote 4.3% more than predicted freespace: Test 2 finished freespace: Test 3: gradually lessen amount of free space by trashing and fill the FS freespace: do 10 steps, lessen free space by 8397544 bytes 8.0 MiB each time freespace: trashing: was free: 92372992 bytes 88.1 MiB, need free: 8397552 bytes 8.0 MiB, files created: 248296, delete 225723 (90.9% of them) freespace: was free: 71909376 bytes 68.6 MiB, wrote: 82472960 bytes 78.7 MiB, delta: 10563584 bytes 10.1 MiB, wrote 14.7% more than predicted freespace: trashing: was free: 88989696 bytes 84.9 MiB, need free: 16795096 bytes 16.0 MiB, files created: 248794, delete 201838 (81.1% of them) freespace: was free: 60354560 bytes 57.6 MiB, wrote: 71782400 bytes 68.5 MiB, delta: 11427840 bytes 10.9 MiB, wrote 18.9% more than predicted freespace: trashing: was free: 90304512 bytes 86.1 MiB, need free: 25192640 bytes 24.0 MiB, files created: 248733, delete 179342 (72.1% of them) freespace: was free: 51187712 bytes 48.8 MiB, wrote: 62943232 bytes 60.0 MiB, delta: 11755520 bytes 11.2 MiB, wrote 23.0% more than predicted freespace: trashing: was free: 91209728 bytes 87.0 MiB, need free: 33590184 bytes 32.0 MiB, files created: 248779, delete 157160 (63.2% of them) freespace: was free: 42704896 bytes 40.7 MiB, wrote: 55050240 bytes 52.5 MiB, delta: 12345344 bytes 11.8 MiB, wrote 28.9% more than predicted freespace: trashing: was free: 92700672 bytes 88.4 MiB, need free: 41987728 bytes 40.0 MiB, files created: 248848, delete 136135 (54.7% of them) freespace: was free: 35250176 bytes 33.6 MiB, wrote: 48115712 bytes 45.9 MiB, delta: 12865536 bytes 12.3 MiB, wrote 36.5% more than predicted freespace: trashing: was free: 93986816 bytes 89.6 MiB, need free: 50385272 bytes 48.1 MiB, files created: 248723, delete 115385 (46.4% of them) freespace: was free: 29995008 bytes 28.6 MiB, wrote: 41582592 bytes 39.7 MiB, delta: 11587584 bytes 11.1 MiB, wrote 38.6% more than predicted freespace: trashing: was free: 91881472 bytes 87.6 MiB, need free: 58782816 bytes 56.1 MiB, files created: 248645, delete 89569 (36.0% of them) freespace: was free: 22511616 bytes 21.5 MiB, wrote: 34705408 bytes 33.1 MiB, delta: 12193792 bytes 11.6 MiB, wrote 54.2% more than predicted freespace: trashing: was free: 91774976 bytes 87.5 MiB, need free: 67180360 bytes 64.1 MiB, files created: 248580, delete 66616 (26.8% of them) freespace: was free: 16908288 bytes 16.1 MiB, wrote: 26898432 bytes 25.7 MiB, delta: 9990144 bytes 9.5 MiB, wrote 59.1% more than predicted freespace: trashing: was free: 92450816 bytes 88.2 MiB, need free: 75577904 bytes 72.1 MiB, files created: 248654, delete 45381 (18.3% of them) freespace: was free: 10170368 bytes 9.7 MiB, wrote: 19111936 bytes 18.2 MiB, delta: 8941568 bytes 8.5 MiB, wrote 87.9% more than predicted freespace: trashing: was free: 93282304 bytes 89.0 MiB, need free: 83975448 bytes 80.1 MiB, files created: 248513, delete 24794 (10.0% of them) freespace: was free: 3911680 bytes 3.7 MiB, wrote: 7872512 bytes 7.5 MiB, delta: 3960832 bytes 3.8 MiB, wrote 101.3% more than predicted freespace: Test 3 finished freespace: finished successfully Signed-off-by: Artem Bityutskiy --- fs/ubifs/budget.c | 38 ++++++++++++++++++++++++++++++++++---- fs/ubifs/super.c | 10 +++++----- fs/ubifs/ubifs.h | 2 +- 3 files changed, 40 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index 7851480a6ce..101d278c591 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -747,14 +747,24 @@ long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free) } /** - * ubifs_budg_get_free_space - return amount of free space. + * ubifs_get_free_space - return amount of free space. * @c: UBIFS file-system description object * - * This function returns amount of free space on the file-system. + * This function calculates amount of free space to report to user-space. + * + * Because UBIFS may introduce substantial overhead (the index, node headers, + * alighment, wastage at the end of eraseblocks, etc), it cannot report real + * amount of free flash space it has (well, because not all dirty space is + * reclamable, UBIFS does not actually know the real amount). If UBIFS did so, + * it would bread user expectetion about what free space is. Users seem to + * accustomed to assume that if the file-system reports N bytes of free space, + * they would be able to fit a file of N bytes to the FS. This almost works for + * traditional file-systems, because they have way less overhead than UBIFS. + * So, to keep users happy, UBIFS tries to take the overhead into account. */ -long long ubifs_budg_get_free_space(struct ubifs_info *c) +long long ubifs_get_free_space(struct ubifs_info *c) { - int min_idx_lebs; + int min_idx_lebs, rsvd_idx_lebs, lebs; long long available, outstanding, free; spin_lock(&c->space_lock); @@ -771,6 +781,26 @@ long long ubifs_budg_get_free_space(struct ubifs_info *c) } available = ubifs_calc_available(c, min_idx_lebs); + + /* + * When reporting free space to user-space, UBIFS guarantees that it is + * possible to write a file of free space size. This means that for + * empty LEBs we may use more precise calculations than + * 'ubifs_calc_available()' is using. Namely, we know that in empty + * LEBs we would waste only @c->leb_overhead bytes, not @c->dark_wm. + * Thus, amend the available space. + * + * Note, the calculations below are similar to what we have in + * 'do_budget_space()', so refer there for comments. + */ + if (min_idx_lebs > c->lst.idx_lebs) + rsvd_idx_lebs = min_idx_lebs - c->lst.idx_lebs; + else + rsvd_idx_lebs = 0; + lebs = c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt - + c->lst.taken_empty_lebs; + lebs -= rsvd_idx_lebs; + available += lebs * (c->dark_wm - c->leb_overhead); spin_unlock(&c->space_lock); if (available > outstanding) diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index be23fd3cfd8..1207bd51ead 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -371,7 +371,7 @@ static int ubifs_statfs(struct dentry *dentry, struct kstatfs *buf) struct ubifs_info *c = dentry->d_sb->s_fs_info; unsigned long long free; - free = ubifs_budg_get_free_space(c); + free = ubifs_get_free_space(c); dbg_gen("free space %lld bytes (%lld blocks)", free, free >> UBIFS_BLOCK_SHIFT); @@ -653,11 +653,11 @@ static int init_constants_late(struct ubifs_info *c) * internally because it does not make much sense for UBIFS, but it is * necessary to report something for the 'statfs()' call. * - * Subtract the LEB reserved for GC and the LEB which is reserved for - * deletions. + * Subtract the LEB reserved for GC, the LEB which is reserved for + * deletions, and assume only one journal head is available. */ - tmp64 = c->main_lebs - 2; - tmp64 *= (uint64_t)c->leb_size - c->dark_wm; + tmp64 = c->main_lebs - 2 - c->jhead_cnt + 1; + tmp64 *= (uint64_t)c->leb_size - c->leb_overhead; tmp64 = ubifs_reported_space(c, tmp64); c->block_cnt = tmp64 >> UBIFS_BLOCK_SHIFT; diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 57e58541de2..17c620b93ee 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1443,7 +1443,7 @@ void ubifs_release_ino_dirty(struct ubifs_info *c, struct inode *inode, struct ubifs_budget_req *req); void ubifs_cancel_ino_op(struct ubifs_info *c, struct inode *inode, struct ubifs_budget_req *req); -long long ubifs_budg_get_free_space(struct ubifs_info *c); +long long ubifs_get_free_space(struct ubifs_info *c); int ubifs_calc_min_idx_lebs(struct ubifs_info *c); void ubifs_convert_page_budget(struct ubifs_info *c); long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free); -- cgit v1.2.3 From b3385c278d3c32aec68d4900b35bc07df1b2240c Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Sun, 31 Aug 2008 17:13:18 +0300 Subject: UBIFS: fill f_fsid UBIFS stores 16-bit UUID in the superblock, and it is a good idea to return part of it in 'f_fsid' filed of kstatfs structure. Signed-off-by: Artem Bityutskiy --- fs/ubifs/super.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 1207bd51ead..0dee4042c6c 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -386,6 +386,7 @@ static int ubifs_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_files = 0; buf->f_ffree = 0; buf->f_namelen = UBIFS_MAX_NLEN; + memcpy(&buf->f_fsid, c->uuid, sizeof(__kernel_fsid_t)); return 0; } -- cgit v1.2.3 From c228c24bf1138d4757dbe20615df655815446da3 Mon Sep 17 00:00:00 2001 From: Andy Adamson Date: Thu, 21 Aug 2008 08:42:16 -0400 Subject: nfsd: fix compound state allocation error handling Move the cstate_alloc call so that if it fails, the response is setup to encode the NFS error. The out label now means that the nfsd4_compound_state has not been allocated. Signed-off-by: Andy Adamson Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4proc.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index 2e51adac65d..e5b51ffafc6 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -867,11 +867,6 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, int slack_bytes; __be32 status; - status = nfserr_resource; - cstate = cstate_alloc(); - if (cstate == NULL) - goto out; - resp->xbuf = &rqstp->rq_res; resp->p = rqstp->rq_res.head[0].iov_base + rqstp->rq_res.head[0].iov_len; resp->tagp = resp->p; @@ -890,6 +885,11 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, if (args->minorversion > NFSD_SUPPORTED_MINOR_VERSION) goto out; + status = nfserr_resource; + cstate = cstate_alloc(); + if (cstate == NULL) + goto out; + status = nfs_ok; while (!status && resp->opcnt < args->opcnt) { op = &args->ops[resp->opcnt++]; @@ -957,9 +957,9 @@ encode_op: nfsd4_increment_op_stats(op->opnum); } + cstate_free(cstate); out: nfsd4_release_compoundargs(args); - cstate_free(cstate); dprintk("nfsv4 compound returned %d\n", ntohl(status)); return status; } -- cgit v1.2.3 From 91b80969ba466ba4b915a4a1d03add8c297add3f Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 29 Aug 2008 19:18:45 -0400 Subject: nfsd: fix buffer overrun decoding NFSv4 acl The array we kmalloc() here is not large enough. Thanks to Johann Dahm and David Richter for bug report and testing. Signed-off-by: J. Bruce Fields Cc: David Richter Tested-by: Johann Dahm --- fs/nfsd/nfs4acl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4acl.c b/fs/nfsd/nfs4acl.c index b6ed38380ab..54b8b4140c8 100644 --- a/fs/nfsd/nfs4acl.c +++ b/fs/nfsd/nfs4acl.c @@ -443,7 +443,7 @@ init_state(struct posix_acl_state *state, int cnt) * enough space for either: */ alloc = sizeof(struct posix_ace_state_array) - + cnt*sizeof(struct posix_ace_state); + + cnt*sizeof(struct posix_user_ace_state); state->users = kzalloc(alloc, GFP_KERNEL); if (!state->users) return -ENOMEM; -- cgit v1.2.3 From 169ccbd44eb20f5bb7e4352451eba25397e29749 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Tue, 2 Sep 2008 14:35:37 -0700 Subject: NTFS: update homepage Update the location of the NTFS homepage in several files. Signed-off-by: Adrian Bunk Cc: Jeff Garzik Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ntfs/usnjrnl.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ntfs/usnjrnl.h b/fs/ntfs/usnjrnl.h index 3a8af75351e..4087fbdac32 100644 --- a/fs/ntfs/usnjrnl.h +++ b/fs/ntfs/usnjrnl.h @@ -113,7 +113,7 @@ typedef struct { * Reason flags (32-bit). Cumulative flags describing the change(s) to the * file since it was last opened. I think the names speak for themselves but * if you disagree check out the descriptions in the Linux NTFS project NTFS - * documentation: http://linux-ntfs.sourceforge.net/ntfs/files/usnjrnl.html + * documentation: http://www.linux-ntfs.org/ */ enum { USN_REASON_DATA_OVERWRITE = const_cpu_to_le32(0x00000001), @@ -145,7 +145,7 @@ typedef le32 USN_REASON_FLAGS; * Source info flags (32-bit). Information about the source of the change(s) * to the file. For detailed descriptions of what these mean, see the Linux * NTFS project NTFS documentation: - * http://linux-ntfs.sourceforge.net/ntfs/files/usnjrnl.html + * http://www.linux-ntfs.org/ */ enum { USN_SOURCE_DATA_MANAGEMENT = const_cpu_to_le32(0x00000001), -- cgit v1.2.3 From 4b8561521dbaa3d766b198496b220e984e3bf756 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Tue, 2 Sep 2008 14:35:53 -0700 Subject: mm: show quicklist usage in /proc/meminfo Quicklists can consume several GB of memory. We should provide a means of monitoring this. After this patch is applied, /proc/meminfo will output the following: % cat /proc/meminfo MemTotal: 7715392 kB MemFree: 5401600 kB Buffers: 80384 kB Cached: 300800 kB SwapCached: 0 kB Active: 235584 kB Inactive: 262656 kB SwapTotal: 2031488 kB SwapFree: 2031488 kB Dirty: 3520 kB Writeback: 0 kB AnonPages: 117696 kB Mapped: 38528 kB Slab: 1589952 kB SReclaimable: 23104 kB SUnreclaim: 1566848 kB PageTables: 14656 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 5889152 kB Committed_AS: 393152 kB VmallocTotal: 17592177655808 kB VmallocUsed: 29056 kB VmallocChunk: 17592177626432 kB Quicklists: 130944 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 262144 kB Signed-off-by: KOSAKI Motohiro Cc: Christoph Lameter Cc: Keiichiro Tokunaga Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/proc_misc.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c index ded96986296..00f10a2dcf1 100644 --- a/fs/proc/proc_misc.c +++ b/fs/proc/proc_misc.c @@ -24,6 +24,7 @@ #include #include #include +#include #include #include #include @@ -189,7 +190,8 @@ static int meminfo_read_proc(char *page, char **start, off_t off, "Committed_AS: %8lu kB\n" "VmallocTotal: %8lu kB\n" "VmallocUsed: %8lu kB\n" - "VmallocChunk: %8lu kB\n", + "VmallocChunk: %8lu kB\n" + "Quicklists: %8lu kB\n", K(i.totalram), K(i.freeram), K(i.bufferram), @@ -221,7 +223,8 @@ static int meminfo_read_proc(char *page, char **start, off_t off, K(committed), (unsigned long)VMALLOC_TOTAL >> 10, vmi.used >> 10, - vmi.largest_chunk >> 10 + vmi.largest_chunk >> 10, + K(quicklist_total_size()) ); len += hugetlb_report_meminfo(page + len); -- cgit v1.2.3 From 7c7cbadf7341a0792879c67d6e3020f040d6cd7f Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Wed, 3 Sep 2008 14:16:42 +0300 Subject: UBIFS: amend f_fsid David Woodhouse suggested to be consistent with other FSes and xor the beginning and the end of the UUID. Signed-off-by: Artem Bityutskiy --- fs/ubifs/super.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 0dee4042c6c..7562464ac83 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -370,6 +370,7 @@ static int ubifs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct ubifs_info *c = dentry->d_sb->s_fs_info; unsigned long long free; + __le32 *uuid = (__le32 *)c->uuid; free = ubifs_get_free_space(c); dbg_gen("free space %lld bytes (%lld blocks)", @@ -386,8 +387,8 @@ static int ubifs_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_files = 0; buf->f_ffree = 0; buf->f_namelen = UBIFS_MAX_NLEN; - memcpy(&buf->f_fsid, c->uuid, sizeof(__kernel_fsid_t)); - + buf->f_fsid.val[0] = le32_to_cpu(uuid[0]) ^ le32_to_cpu(uuid[2]); + buf->f_fsid.val[1] = le32_to_cpu(uuid[1]) ^ le32_to_cpu(uuid[3]); return 0; } -- cgit v1.2.3 From 49048622eae698e5c4ae61f7e71200f265ccc529 Mon Sep 17 00:00:00 2001 From: Balbir Singh Date: Fri, 5 Sep 2008 18:12:23 +0200 Subject: sched: fix process time monotonicity Spencer reported a problem where utime and stime were going negative despite the fixes in commit b27f03d4bdc145a09fb7b0c0e004b29f1ee555fa. The suspected reason for the problem is that signal_struct maintains it's own utime and stime (of exited tasks), these are not updated using the new task_utime() routine, hence sig->utime can go backwards and cause the same problem to occur (sig->utime, adds tsk->utime and not task_utime()). This patch fixes the problem TODO: using max(task->prev_utime, derived utime) works for now, but a more generic solution is to implement cputime_max() and use the cputime_gt() function for comparison. Reported-by: spencer@bluehost.com Signed-off-by: Balbir Singh Signed-off-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- fs/proc/array.c | 59 --------------------------------------------------------- 1 file changed, 59 deletions(-) (limited to 'fs') diff --git a/fs/proc/array.c b/fs/proc/array.c index 0d6eb33597c..71c9be59c9c 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -337,65 +337,6 @@ int proc_pid_status(struct seq_file *m, struct pid_namespace *ns, return 0; } -/* - * Use precise platform statistics if available: - */ -#ifdef CONFIG_VIRT_CPU_ACCOUNTING -static cputime_t task_utime(struct task_struct *p) -{ - return p->utime; -} - -static cputime_t task_stime(struct task_struct *p) -{ - return p->stime; -} -#else -static cputime_t task_utime(struct task_struct *p) -{ - clock_t utime = cputime_to_clock_t(p->utime), - total = utime + cputime_to_clock_t(p->stime); - u64 temp; - - /* - * Use CFS's precise accounting: - */ - temp = (u64)nsec_to_clock_t(p->se.sum_exec_runtime); - - if (total) { - temp *= utime; - do_div(temp, total); - } - utime = (clock_t)temp; - - p->prev_utime = max(p->prev_utime, clock_t_to_cputime(utime)); - return p->prev_utime; -} - -static cputime_t task_stime(struct task_struct *p) -{ - clock_t stime; - - /* - * Use CFS's precise accounting. (we subtract utime from - * the total, to make sure the total observed by userspace - * grows monotonically - apps rely on that): - */ - stime = nsec_to_clock_t(p->se.sum_exec_runtime) - - cputime_to_clock_t(task_utime(p)); - - if (stime >= 0) - p->prev_stime = max(p->prev_stime, clock_t_to_cputime(stime)); - - return p->prev_stime; -} -#endif - -static cputime_t task_gtime(struct task_struct *p) -{ - return p->gtime; -} - static int do_task_stat(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *task, int whole) { -- cgit v1.2.3 From f171d4d769c8ccac6675892960e37f6485837fae Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Wed, 3 Sep 2008 16:17:14 +0300 Subject: UBIFS: fix division by zero If fanout is 3, we have division by zero in 'ubifs_read_superblock()': divide error: 0000 [#1] PREEMPT SMP Pid: 28744, comm: mount Not tainted (2.6.27-rc4-ubifs-2.6 #23) EIP: 0060:[] EFLAGS: 00010202 CPU: 0 EIP is at ubifs_reported_space+0x2d/0x69 [ubifs] EAX: 00000000 EBX: 00000000 ECX: 00000000 EDX: 00000000 ESI: 00000000 EDI: f0ae64b0 EBP: f1f9fcf4 ESP: f1f9fce0 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 Signed-off-by: Artem Bityutskiy --- fs/ubifs/budget.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c index 101d278c591..73db464cd08 100644 --- a/fs/ubifs/budget.c +++ b/fs/ubifs/budget.c @@ -723,24 +723,25 @@ void ubifs_release_dirty_inode_budget(struct ubifs_info *c, */ long long ubifs_reported_space(const struct ubifs_info *c, uint64_t free) { - int divisor, factor; + int divisor, factor, f; /* * Reported space size is @free * X, where X is UBIFS block size * divided by UBIFS block size + all overhead one data block * introduces. The overhead is the node header + indexing overhead. * - * Indexing overhead is calculations are based on the following - * formula: I = N/(f - 1) + 1, where I - number of indexing nodes, N - - * number of data nodes, f - fanout. Because effective UBIFS fanout is - * twice as less than maximum fanout, we assume that each data node + * Indexing overhead calculations are based on the following formula: + * I = N/(f - 1) + 1, where I - number of indexing nodes, N - number + * of data nodes, f - fanout. Because effective UBIFS fanout is twice + * as less than maximum fanout, we assume that each data node * introduces 3 * @c->max_idx_node_sz / (@c->fanout/2 - 1) bytes. * Note, the multiplier 3 is because UBIFS reseves thrice as more space * for the index. */ + f = c->fanout > 3 ? c->fanout >> 1 : 2; factor = UBIFS_BLOCK_SIZE; divisor = UBIFS_MAX_DATA_NODE_SZ; - divisor += (c->max_idx_node_sz * 3) / ((c->fanout >> 1) - 1); + divisor += (c->max_idx_node_sz * 3) / (f - 1); free *= factor; do_div(free, divisor); return free; -- cgit v1.2.3 From a5cb562d6977d9d7989c346b7b153cef31ec0228 Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy Date: Wed, 3 Sep 2008 18:26:47 +0300 Subject: UBIFS: make minimum fanout 3 UBIFS does not really work correctly when fanout is 2, because of the way we manage the indexing tree. It may just become a list and UBIFS screws up. Signed-off-by: Artem Bityutskiy --- fs/ubifs/ubifs-media.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ubifs/ubifs-media.h b/fs/ubifs/ubifs-media.h index bd2121f3426..a9ecbd9af20 100644 --- a/fs/ubifs/ubifs-media.h +++ b/fs/ubifs/ubifs-media.h @@ -87,7 +87,7 @@ #define UBIFS_SK_LEN 8 /* Minimum index tree fanout */ -#define UBIFS_MIN_FANOUT 2 +#define UBIFS_MIN_FANOUT 3 /* Maximum number of levels in UBIFS indexing B-tree */ #define UBIFS_MAX_LEVELS 512 -- cgit v1.2.3 From af904deaf6da3f3285eb0a06a3dc6a1af0251030 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 8 Sep 2008 11:58:13 -0400 Subject: NFS: Restore missing hunk in NFS mount option parser Automounter maps can contain mount options valid for other NFS implementations but not for Linux. The Linux automounter uses the mount command's "-s" command line option ("s" for "sloppy") so that mount requests containing such options are not rejected. Commit f45663ce5fb30f76a3414ab3ac69f4dd320e760a attempted to address a known regression with text-based NFS mount option parsing. Unrecognized mount options would cause mount requests to fail, even if the "-s" option was used on the mount command line. Unfortunately, this commit was not complete as submitted. It adds a new mount option, "sloppy". But it is missing a hunk, so it now allows NFS mounts with unrecognized mount options, even if the "sloppy" option is not present. This could be a problem if a required critical mount option such as "sync" is misspelled, for example, and is considered a regression from 2.6.26. This patch restores the missing hunk. Now, the default behavior of text-based NFS mount options is as before: any unrecognized mount option will cause the mount to fail. Please include this in 2.6.27-rc. Thanks to Neil Brown for reporting this. Signed-off-by: Chuck Lever Acked-by: J. Bruce Fields Signed-off-by: Linus Torvalds --- fs/nfs/super.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 9abcd2b329f..e9b20173fef 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1279,6 +1279,12 @@ static int nfs_parse_mount_options(char *raw, } } + if (errors > 0) { + dfprintk(MOUNT, "NFS: parsing encountered %d error%s\n", + errors, (errors == 1 ? "" : "s")); + if (!sloppy) + return 0; + } return 1; out_nomem: -- cgit v1.2.3