From 786228ab3095fe18dba3bc0d62055a123991d9d9 Mon Sep 17 00:00:00 2001 From: Marcos Paulo de Souza Date: Wed, 23 Nov 2011 20:56:45 +0800 Subject: writeback: Fix issue on make htmldocs Document the @reason parameter to make "make htmldocs" happy. Acked-by: Randy Dunlap Signed-off-by: Marcos Paulo de Souza Signed-off-by: Wu Fengguang --- fs/fs-writeback.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'fs') diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 73c3992b2bb..ac86f8b3e3c 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -156,6 +156,7 @@ __bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages, * bdi_start_writeback - start writeback * @bdi: the backing device to write from * @nr_pages: the number of pages to write + * @reason: reason why some writeback work was initiated * * Description: * This does WB_SYNC_NONE opportunistic writeback. The IO is only @@ -1223,6 +1224,7 @@ static void wait_sb_inodes(struct super_block *sb) * writeback_inodes_sb_nr - writeback dirty inodes from given super_block * @sb: the superblock * @nr: the number of pages to write + * @reason: reason why some writeback work initiated * * Start writeback on some inodes on this super_block. No guarantees are made * on how many (if any) will be written, and this function does not wait @@ -1251,6 +1253,7 @@ EXPORT_SYMBOL(writeback_inodes_sb_nr); /** * writeback_inodes_sb - writeback dirty inodes from given super_block * @sb: the superblock + * @reason: reason why some writeback work was initiated * * Start writeback on some inodes on this super_block. No guarantees are made * on how many (if any) will be written, and this function does not wait @@ -1265,6 +1268,7 @@ EXPORT_SYMBOL(writeback_inodes_sb); /** * writeback_inodes_sb_if_idle - start writeback if none underway * @sb: the superblock + * @reason: reason why some writeback work was initiated * * Invoke writeback_inodes_sb if no writeback is currently underway. * Returns 1 if writeback was started, 0 if not. @@ -1285,6 +1289,7 @@ EXPORT_SYMBOL(writeback_inodes_sb_if_idle); * writeback_inodes_sb_if_idle - start writeback if none underway * @sb: the superblock * @nr: the number of pages to write + * @reason: reason why some writeback work was initiated * * Invoke writeback_inodes_sb if no writeback is currently underway. * Returns 1 if writeback was started, 0 if not. -- cgit v1.2.3-70-g09d2 From 4f38e4aadcca319650c0de31edf7193d7c384de2 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 1 Dec 2011 16:31:34 -0500 Subject: NFSv4: Don't error if we handled it in nfs4_recovery_handle_error If we handled an error condition, then nfs4_recovery_handle_error should return '0' so that the state recovery thread can continue. Also ensure that nfs4_check_lease() continues to abort if we haven't got any credentials by having it return ENOKEY (which is not handled). Signed-off-by: Trond Myklebust --- fs/nfs/nfs4state.c | 21 +++++++++++++-------- 1 file changed, 13 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 39914be40b0..1da95e3947b 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1350,12 +1350,14 @@ static void nfs4_warn_keyexpired(const char *s) static int nfs4_recovery_handle_error(struct nfs_client *clp, int error) { switch (error) { + case 0: + break; case -NFS4ERR_CB_PATH_DOWN: nfs_handle_cb_pathdown(clp); - return 0; + break; case -NFS4ERR_NO_GRACE: nfs4_state_end_reclaim_reboot(clp); - return 0; + break; case -NFS4ERR_STALE_CLIENTID: case -NFS4ERR_LEASE_MOVED: set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state); @@ -1375,13 +1377,15 @@ static int nfs4_recovery_handle_error(struct nfs_client *clp, int error) case -NFS4ERR_SEQ_MISORDERED: set_bit(NFS4CLNT_SESSION_RESET, &clp->cl_state); /* Zero session reset errors */ - return 0; + break; case -EKEYEXPIRED: /* Nothing we can do */ nfs4_warn_keyexpired(clp->cl_hostname); - return 0; + break; + default: + return error; } - return error; + return 0; } static int nfs4_do_reclaim(struct nfs_client *clp, const struct nfs4_state_recovery_ops *ops) @@ -1428,7 +1432,7 @@ static int nfs4_check_lease(struct nfs_client *clp) struct rpc_cred *cred; const struct nfs4_state_maintenance_ops *ops = clp->cl_mvops->state_renewal_ops; - int status = -NFS4ERR_EXPIRED; + int status; /* Is the client already known to have an expired lease? */ if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) @@ -1438,6 +1442,7 @@ static int nfs4_check_lease(struct nfs_client *clp) spin_unlock(&clp->cl_lock); if (cred == NULL) { cred = nfs4_get_setclientid_cred(clp); + status = -ENOKEY; if (cred == NULL) goto out; } @@ -1662,10 +1667,10 @@ static void nfs4_state_manager(struct nfs_client *clp) if (test_and_clear_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state)) { status = nfs4_check_lease(clp); + if (status < 0) + goto out_error; if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) continue; - if (status < 0 && status != -NFS4ERR_CB_PATH_DOWN) - goto out_error; } /* Initialize or reset the session */ -- cgit v1.2.3-70-g09d2 From 111d489f0fb431f4ae85d96851fbf8d3248c09d8 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 1 Dec 2011 16:37:42 -0500 Subject: NFSv4.1: Ensure that we handle _all_ SEQUENCE status bits. Currently, the code assumes that the SEQUENCE status bits are mutually exclusive. They are not... Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org [>= 2.6.34] --- fs/nfs/nfs4state.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 1da95e3947b..da8d73ed9e0 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1530,16 +1530,16 @@ void nfs41_handle_sequence_flag_errors(struct nfs_client *clp, u32 flags) { if (!flags) return; - else if (flags & SEQ4_STATUS_RESTART_RECLAIM_NEEDED) + if (flags & SEQ4_STATUS_RESTART_RECLAIM_NEEDED) nfs41_handle_server_reboot(clp); - else if (flags & (SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED | + if (flags & (SEQ4_STATUS_EXPIRED_ALL_STATE_REVOKED | SEQ4_STATUS_EXPIRED_SOME_STATE_REVOKED | SEQ4_STATUS_ADMIN_STATE_REVOKED | SEQ4_STATUS_LEASE_MOVED)) nfs41_handle_state_revoked(clp); - else if (flags & SEQ4_STATUS_RECALLABLE_STATE_REVOKED) + if (flags & SEQ4_STATUS_RECALLABLE_STATE_REVOKED) nfs41_handle_recallable_state_revoked(clp); - else if (flags & (SEQ4_STATUS_CB_PATH_DOWN | + if (flags & (SEQ4_STATUS_CB_PATH_DOWN | SEQ4_STATUS_BACKCHANNEL_FAULT | SEQ4_STATUS_CB_PATH_DOWN_SESSION)) nfs41_handle_cb_path_down(clp); -- cgit v1.2.3-70-g09d2 From 2151937d7ce491bfbe269a1ae742c6686904474c Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Thu, 1 Dec 2011 08:06:52 -0800 Subject: ceph: fix rasize reporting by ceph_show_options Fix typo. Reported-by: mowang da Signed-off-by: Sage Weil --- fs/ceph/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 8dc73a594a9..b48f15f101a 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -383,7 +383,7 @@ static int ceph_show_options(struct seq_file *m, struct vfsmount *mnt) if (fsopt->rsize != CEPH_RSIZE_DEFAULT) seq_printf(m, ",rsize=%d", fsopt->rsize); if (fsopt->rasize != CEPH_RASIZE_DEFAULT) - seq_printf(m, ",rasize=%d", fsopt->rsize); + seq_printf(m, ",rasize=%d", fsopt->rasize); if (fsopt->congestion_kb != default_congestion_kb()) seq_printf(m, ",write_congestion_kb=%d", fsopt->congestion_kb); if (fsopt->caps_wanted_delay_min != CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT) -- cgit v1.2.3-70-g09d2 From a99ebf43f49f4499ab0e2a8a9132ad6ed6ba2409 Mon Sep 17 00:00:00 2001 From: Dave Chinner Date: Thu, 1 Dec 2011 11:24:20 +0000 Subject: xfs: fix allocation length overflow in xfs_bmapi_write() When testing the new xfstests --large-fs option that does very large file preallocations, this assert was tripped deep in xfs_alloc_vextent(): XFS: Assertion failed: args->minlen <= args->maxlen, file: fs/xfs/xfs_alloc.c, line: 2239 The allocation was trying to allocate a zero length extent because the lower 32 bits of the allocation length was zero. The remaining length of the allocation to be done was an exact multiple of 2^32 - the first case I saw was at 496TB remaining to be allocated. This turns out to be an overflow when converting the allocation length (a 64 bit quantity) into the extent length to allocate (a 32 bit quantity), and it requires the length to be allocated an exact multiple of 2^32 blocks to trip the assert. Fix it by limiting the extent lenth to allocate to MAXEXTLEN. Signed-off-by: Dave Chinner Signed-off-by: Ben Myers Reviewed-by: Christoph Hellwig --- fs/xfs/xfs_bmap.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index c68baeb0974..d0ab7883705 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2383,6 +2383,8 @@ xfs_bmap_btalloc( int tryagain; int error; + ASSERT(ap->length); + mp = ap->ip->i_mount; align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0; if (unlikely(align)) { @@ -4629,6 +4631,8 @@ xfs_bmapi_allocate( int error; int rt; + ASSERT(bma->length > 0); + rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(bma->ip); /* @@ -4849,6 +4853,7 @@ xfs_bmapi_write( ASSERT(*nmap <= XFS_BMAP_MAX_NMAP); ASSERT(!(flags & XFS_BMAPI_IGSTATE)); ASSERT(tp != NULL); + ASSERT(len > 0); whichfork = (flags & XFS_BMAPI_ATTRFORK) ? XFS_ATTR_FORK : XFS_DATA_FORK; @@ -4918,9 +4923,22 @@ xfs_bmapi_write( bma.eof = eof; bma.conv = !!(flags & XFS_BMAPI_CONVERT); bma.wasdel = wasdelay; - bma.length = len; bma.offset = bno; + /* + * There's a 32/64 bit type mismatch between the + * allocation length request (which can be 64 bits in + * length) and the bma length request, which is + * xfs_extlen_t and therefore 32 bits. Hence we have to + * check for 32-bit overflows and handle them here. + */ + if (len > (xfs_filblks_t)MAXEXTLEN) + bma.length = MAXEXTLEN; + else + bma.length = len; + + ASSERT(len > 0); + ASSERT(bma.length > 0); error = xfs_bmapi_allocate(&bma, flags); if (error) goto error0; -- cgit v1.2.3-70-g09d2 From c29f7d457ac63311feb11928a866efd2fe153d74 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 30 Nov 2011 08:58:18 +0000 Subject: xfs: fix nfs export of 64-bit inodes numbers on 32-bit kernels The i_ino field in the VFS inode is of type unsigned long and thus can't hold the full 64-bit inode number on 32-bit kernels. We have the full inode number in the XFS inode, so use that one for nfs exports. Note that I've also switched the 32-bit file handles types to it, just to make the code more consistent and copy & paste errors less likely to happen. Reported-by: Guoquan Yang Reported-by: Hank Peng Signed-off-by: Christoph Hellwig Signed-off-by: Ben Myers --- fs/xfs/xfs_export.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_export.c b/fs/xfs/xfs_export.c index da108977b21..558910f5e3c 100644 --- a/fs/xfs/xfs_export.c +++ b/fs/xfs/xfs_export.c @@ -98,22 +98,22 @@ xfs_fs_encode_fh( switch (fileid_type) { case FILEID_INO32_GEN_PARENT: spin_lock(&dentry->d_lock); - fid->i32.parent_ino = dentry->d_parent->d_inode->i_ino; + fid->i32.parent_ino = XFS_I(dentry->d_parent->d_inode)->i_ino; fid->i32.parent_gen = dentry->d_parent->d_inode->i_generation; spin_unlock(&dentry->d_lock); /*FALLTHRU*/ case FILEID_INO32_GEN: - fid->i32.ino = inode->i_ino; + fid->i32.ino = XFS_I(inode)->i_ino; fid->i32.gen = inode->i_generation; break; case FILEID_INO32_GEN_PARENT | XFS_FILEID_TYPE_64FLAG: spin_lock(&dentry->d_lock); - fid64->parent_ino = dentry->d_parent->d_inode->i_ino; + fid64->parent_ino = XFS_I(dentry->d_parent->d_inode)->i_ino; fid64->parent_gen = dentry->d_parent->d_inode->i_generation; spin_unlock(&dentry->d_lock); /*FALLTHRU*/ case FILEID_INO32_GEN | XFS_FILEID_TYPE_64FLAG: - fid64->ino = inode->i_ino; + fid64->ino = XFS_I(inode)->i_ino; fid64->gen = inode->i_generation; break; } -- cgit v1.2.3-70-g09d2 From 9f9c19ec1a59422c7687b11847ed3408128aa0d6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 28 Nov 2011 08:17:36 +0000 Subject: xfs: fix the logspace waiting algorithm Apply the scheme used in log_regrant_write_log_space to wake up any other threads waiting for log space before the newly added one to log_regrant_write_log_space as well, and factor the code into readable helpers. For each of the queues we have add two helpers: - one to try to wake up all waiting threads. This helper will also be usable by xfs_log_move_tail once we remove the current opportunistic wakeups in it. - one to sleep on t_wait until enough log space is available, loosely modelled after Linux waitqueues. And use them to reimplement the guts of log_regrant_write_log_space and log_regrant_write_log_space. These two function now use one and the same algorithm for waiting on log space instead of subtly different ones before, with an option to completely unify them in the near future. Also move the filesystem shutdown handling to the common caller given that we had to touch it anyway. Based on hard debugging and an earlier patch from Chandra Seetharaman . Signed-off-by: Christoph Hellwig Reviewed-by: Chandra Seetharaman Tested-by: Chandra Seetharaman Signed-off-by: Ben Myers --- fs/xfs/xfs_log.c | 348 ++++++++++++++++++++++++++--------------------------- fs/xfs/xfs_trace.h | 12 +- 2 files changed, 177 insertions(+), 183 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index a14cd89fe46..34817adf4b9 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -150,6 +150,117 @@ xlog_grant_add_space( } while (head_val != old); } +STATIC bool +xlog_reserveq_wake( + struct log *log, + int *free_bytes) +{ + struct xlog_ticket *tic; + int need_bytes; + + list_for_each_entry(tic, &log->l_reserveq, t_queue) { + if (tic->t_flags & XLOG_TIC_PERM_RESERV) + need_bytes = tic->t_unit_res * tic->t_cnt; + else + need_bytes = tic->t_unit_res; + + if (*free_bytes < need_bytes) + return false; + *free_bytes -= need_bytes; + + trace_xfs_log_grant_wake_up(log, tic); + wake_up(&tic->t_wait); + } + + return true; +} + +STATIC bool +xlog_writeq_wake( + struct log *log, + int *free_bytes) +{ + struct xlog_ticket *tic; + int need_bytes; + + list_for_each_entry(tic, &log->l_writeq, t_queue) { + ASSERT(tic->t_flags & XLOG_TIC_PERM_RESERV); + + need_bytes = tic->t_unit_res; + + if (*free_bytes < need_bytes) + return false; + *free_bytes -= need_bytes; + + trace_xfs_log_regrant_write_wake_up(log, tic); + wake_up(&tic->t_wait); + } + + return true; +} + +STATIC int +xlog_reserveq_wait( + struct log *log, + struct xlog_ticket *tic, + int need_bytes) +{ + list_add_tail(&tic->t_queue, &log->l_reserveq); + + do { + if (XLOG_FORCED_SHUTDOWN(log)) + goto shutdown; + xlog_grant_push_ail(log, need_bytes); + + XFS_STATS_INC(xs_sleep_logspace); + trace_xfs_log_grant_sleep(log, tic); + + xlog_wait(&tic->t_wait, &log->l_grant_reserve_lock); + trace_xfs_log_grant_wake(log, tic); + + spin_lock(&log->l_grant_reserve_lock); + if (XLOG_FORCED_SHUTDOWN(log)) + goto shutdown; + } while (xlog_space_left(log, &log->l_grant_reserve_head) < need_bytes); + + list_del_init(&tic->t_queue); + return 0; +shutdown: + list_del_init(&tic->t_queue); + return XFS_ERROR(EIO); +} + +STATIC int +xlog_writeq_wait( + struct log *log, + struct xlog_ticket *tic, + int need_bytes) +{ + list_add_tail(&tic->t_queue, &log->l_writeq); + + do { + if (XLOG_FORCED_SHUTDOWN(log)) + goto shutdown; + xlog_grant_push_ail(log, need_bytes); + + XFS_STATS_INC(xs_sleep_logspace); + trace_xfs_log_regrant_write_sleep(log, tic); + + xlog_wait(&tic->t_wait, &log->l_grant_write_lock); + trace_xfs_log_regrant_write_wake(log, tic); + + spin_lock(&log->l_grant_write_lock); + if (XLOG_FORCED_SHUTDOWN(log)) + goto shutdown; + } while (xlog_space_left(log, &log->l_grant_write_head) < need_bytes); + + list_del_init(&tic->t_queue); + return 0; +shutdown: + list_del_init(&tic->t_queue); + return XFS_ERROR(EIO); +} + static void xlog_tic_reset_res(xlog_ticket_t *tic) { @@ -350,8 +461,19 @@ xfs_log_reserve( retval = xlog_grant_log_space(log, internal_ticket); } + if (unlikely(retval)) { + /* + * If we are failing, make sure the ticket doesn't have any + * current reservations. We don't want to add this back + * when the ticket/ transaction gets cancelled. + */ + internal_ticket->t_curr_res = 0; + /* ungrant will give back unit_res * t_cnt. */ + internal_ticket->t_cnt = 0; + } + return retval; -} /* xfs_log_reserve */ +} /* @@ -2481,8 +2603,8 @@ restart: /* * Atomically get the log space required for a log ticket. * - * Once a ticket gets put onto the reserveq, it will only return after - * the needed reservation is satisfied. + * Once a ticket gets put onto the reserveq, it will only return after the + * needed reservation is satisfied. * * This function is structured so that it has a lock free fast path. This is * necessary because every new transaction reservation will come through this @@ -2490,113 +2612,53 @@ restart: * every pass. * * As tickets are only ever moved on and off the reserveq under the - * l_grant_reserve_lock, we only need to take that lock if we are going - * to add the ticket to the queue and sleep. We can avoid taking the lock if the - * ticket was never added to the reserveq because the t_queue list head will be - * empty and we hold the only reference to it so it can safely be checked - * unlocked. + * l_grant_reserve_lock, we only need to take that lock if we are going to add + * the ticket to the queue and sleep. We can avoid taking the lock if the ticket + * was never added to the reserveq because the t_queue list head will be empty + * and we hold the only reference to it so it can safely be checked unlocked. */ STATIC int -xlog_grant_log_space(xlog_t *log, - xlog_ticket_t *tic) +xlog_grant_log_space( + struct log *log, + struct xlog_ticket *tic) { - int free_bytes; - int need_bytes; + int free_bytes, need_bytes; + int error = 0; -#ifdef DEBUG - if (log->l_flags & XLOG_ACTIVE_RECOVERY) - panic("grant Recovery problem"); -#endif + ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); trace_xfs_log_grant_enter(log, tic); + /* + * If there are other waiters on the queue then give them a chance at + * logspace before us. Wake up the first waiters, if we do not wake + * up all the waiters then go to sleep waiting for more free space, + * otherwise try to get some space for this transaction. + */ need_bytes = tic->t_unit_res; if (tic->t_flags & XFS_LOG_PERM_RESERV) need_bytes *= tic->t_ocnt; - - /* something is already sleeping; insert new transaction at end */ - if (!list_empty_careful(&log->l_reserveq)) { - spin_lock(&log->l_grant_reserve_lock); - /* recheck the queue now we are locked */ - if (list_empty(&log->l_reserveq)) { - spin_unlock(&log->l_grant_reserve_lock); - goto redo; - } - list_add_tail(&tic->t_queue, &log->l_reserveq); - - trace_xfs_log_grant_sleep1(log, tic); - - /* - * Gotta check this before going to sleep, while we're - * holding the grant lock. - */ - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - XFS_STATS_INC(xs_sleep_logspace); - xlog_wait(&tic->t_wait, &log->l_grant_reserve_lock); - - /* - * If we got an error, and the filesystem is shutting down, - * we'll catch it down below. So just continue... - */ - trace_xfs_log_grant_wake1(log, tic); - } - -redo: - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return_unlocked; - free_bytes = xlog_space_left(log, &log->l_grant_reserve_head); - if (free_bytes < need_bytes) { + if (!list_empty_careful(&log->l_reserveq)) { spin_lock(&log->l_grant_reserve_lock); - if (list_empty(&tic->t_queue)) - list_add_tail(&tic->t_queue, &log->l_reserveq); - - trace_xfs_log_grant_sleep2(log, tic); - - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - xlog_grant_push_ail(log, need_bytes); - - XFS_STATS_INC(xs_sleep_logspace); - xlog_wait(&tic->t_wait, &log->l_grant_reserve_lock); - - trace_xfs_log_grant_wake2(log, tic); - goto redo; - } - - if (!list_empty(&tic->t_queue)) { + if (!xlog_reserveq_wake(log, &free_bytes) || + free_bytes < need_bytes) + error = xlog_reserveq_wait(log, tic, need_bytes); + spin_unlock(&log->l_grant_reserve_lock); + } else if (free_bytes < need_bytes) { spin_lock(&log->l_grant_reserve_lock); - list_del_init(&tic->t_queue); + error = xlog_reserveq_wait(log, tic, need_bytes); spin_unlock(&log->l_grant_reserve_lock); } + if (error) + return error; - /* we've got enough space */ xlog_grant_add_space(log, &log->l_grant_reserve_head, need_bytes); xlog_grant_add_space(log, &log->l_grant_write_head, need_bytes); trace_xfs_log_grant_exit(log, tic); xlog_verify_grant_tail(log); return 0; - -error_return_unlocked: - spin_lock(&log->l_grant_reserve_lock); -error_return: - list_del_init(&tic->t_queue); - spin_unlock(&log->l_grant_reserve_lock); - trace_xfs_log_grant_error(log, tic); - - /* - * If we are failing, make sure the ticket doesn't have any - * current reservations. We don't want to add this back when - * the ticket/transaction gets cancelled. - */ - tic->t_curr_res = 0; - tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ - return XFS_ERROR(EIO); -} /* xlog_grant_log_space */ - +} /* * Replenish the byte reservation required by moving the grant write head. @@ -2605,10 +2667,12 @@ error_return: * free fast path. */ STATIC int -xlog_regrant_write_log_space(xlog_t *log, - xlog_ticket_t *tic) +xlog_regrant_write_log_space( + struct log *log, + struct xlog_ticket *tic) { - int free_bytes, need_bytes; + int free_bytes, need_bytes; + int error = 0; tic->t_curr_res = tic->t_unit_res; xlog_tic_reset_res(tic); @@ -2616,104 +2680,38 @@ xlog_regrant_write_log_space(xlog_t *log, if (tic->t_cnt > 0) return 0; -#ifdef DEBUG - if (log->l_flags & XLOG_ACTIVE_RECOVERY) - panic("regrant Recovery problem"); -#endif + ASSERT(!(log->l_flags & XLOG_ACTIVE_RECOVERY)); trace_xfs_log_regrant_write_enter(log, tic); - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return_unlocked; - /* If there are other waiters on the queue then give them a - * chance at logspace before us. Wake up the first waiters, - * if we do not wake up all the waiters then go to sleep waiting - * for more free space, otherwise try to get some space for - * this transaction. + /* + * If there are other waiters on the queue then give them a chance at + * logspace before us. Wake up the first waiters, if we do not wake + * up all the waiters then go to sleep waiting for more free space, + * otherwise try to get some space for this transaction. */ need_bytes = tic->t_unit_res; - if (!list_empty_careful(&log->l_writeq)) { - struct xlog_ticket *ntic; - - spin_lock(&log->l_grant_write_lock); - free_bytes = xlog_space_left(log, &log->l_grant_write_head); - list_for_each_entry(ntic, &log->l_writeq, t_queue) { - ASSERT(ntic->t_flags & XLOG_TIC_PERM_RESERV); - - if (free_bytes < ntic->t_unit_res) - break; - free_bytes -= ntic->t_unit_res; - wake_up(&ntic->t_wait); - } - - if (ntic != list_first_entry(&log->l_writeq, - struct xlog_ticket, t_queue)) { - if (list_empty(&tic->t_queue)) - list_add_tail(&tic->t_queue, &log->l_writeq); - trace_xfs_log_regrant_write_sleep1(log, tic); - - xlog_grant_push_ail(log, need_bytes); - - XFS_STATS_INC(xs_sleep_logspace); - xlog_wait(&tic->t_wait, &log->l_grant_write_lock); - trace_xfs_log_regrant_write_wake1(log, tic); - } else - spin_unlock(&log->l_grant_write_lock); - } - -redo: - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return_unlocked; - free_bytes = xlog_space_left(log, &log->l_grant_write_head); - if (free_bytes < need_bytes) { + if (!list_empty_careful(&log->l_writeq)) { spin_lock(&log->l_grant_write_lock); - if (list_empty(&tic->t_queue)) - list_add_tail(&tic->t_queue, &log->l_writeq); - - if (XLOG_FORCED_SHUTDOWN(log)) - goto error_return; - - xlog_grant_push_ail(log, need_bytes); - - XFS_STATS_INC(xs_sleep_logspace); - trace_xfs_log_regrant_write_sleep2(log, tic); - xlog_wait(&tic->t_wait, &log->l_grant_write_lock); - - trace_xfs_log_regrant_write_wake2(log, tic); - goto redo; - } - - if (!list_empty(&tic->t_queue)) { + if (!xlog_writeq_wake(log, &free_bytes) || + free_bytes < need_bytes) + error = xlog_writeq_wait(log, tic, need_bytes); + spin_unlock(&log->l_grant_write_lock); + } else if (free_bytes < need_bytes) { spin_lock(&log->l_grant_write_lock); - list_del_init(&tic->t_queue); + error = xlog_writeq_wait(log, tic, need_bytes); spin_unlock(&log->l_grant_write_lock); } - /* we've got enough space */ + if (error) + return error; + xlog_grant_add_space(log, &log->l_grant_write_head, need_bytes); trace_xfs_log_regrant_write_exit(log, tic); xlog_verify_grant_tail(log); return 0; - - - error_return_unlocked: - spin_lock(&log->l_grant_write_lock); - error_return: - list_del_init(&tic->t_queue); - spin_unlock(&log->l_grant_write_lock); - trace_xfs_log_regrant_write_error(log, tic); - - /* - * If we are failing, make sure the ticket doesn't have any - * current reservations. We don't want to add this back when - * the ticket/transaction gets cancelled. - */ - tic->t_curr_res = 0; - tic->t_cnt = 0; /* ungrant will give back unit_res * t_cnt. */ - return XFS_ERROR(EIO); -} /* xlog_regrant_write_log_space */ - +} /* The first cnt-1 times through here we don't need to * move the grant write head because the permanent diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index f1d2802b2f0..49403579887 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -834,18 +834,14 @@ DEFINE_LOGGRANT_EVENT(xfs_log_umount_write); DEFINE_LOGGRANT_EVENT(xfs_log_grant_enter); DEFINE_LOGGRANT_EVENT(xfs_log_grant_exit); DEFINE_LOGGRANT_EVENT(xfs_log_grant_error); -DEFINE_LOGGRANT_EVENT(xfs_log_grant_sleep1); -DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake1); -DEFINE_LOGGRANT_EVENT(xfs_log_grant_sleep2); -DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake2); +DEFINE_LOGGRANT_EVENT(xfs_log_grant_sleep); +DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake); DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake_up); DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_enter); DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_exit); DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_error); -DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_sleep1); -DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_wake1); -DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_sleep2); -DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_wake2); +DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_sleep); +DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_wake); DEFINE_LOGGRANT_EVENT(xfs_log_regrant_write_wake_up); DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_enter); DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_exit); -- cgit v1.2.3-70-g09d2 From 02125a826459a6ad142f8d91c5b6357562f96615 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 5 Dec 2011 08:43:34 -0500 Subject: fix apparmor dereferencing potentially freed dentry, sanitize __d_path() API __d_path() API is asking for trouble and in case of apparmor d_namespace_path() getting just that. The root cause is that when __d_path() misses the root it had been told to look for, it stores the location of the most remote ancestor in *root. Without grabbing references. Sure, at the moment of call it had been pinned down by what we have in *path. And if we raced with umount -l, we could have very well stopped at vfsmount/dentry that got freed as soon as prepend_path() dropped vfsmount_lock. It is safe to compare these pointers with pre-existing (and known to be still alive) vfsmount and dentry, as long as all we are asking is "is it the same address?". Dereferencing is not safe and apparmor ended up stepping into that. d_namespace_path() really wants to examine the place where we stopped, even if it's not connected to our namespace. As the result, it looked at ->d_sb->s_magic of a dentry that might've been already freed by that point. All other callers had been careful enough to avoid that, but it's really a bad interface - it invites that kind of trouble. The fix is fairly straightforward, even though it's bigger than I'd like: * prepend_path() root argument becomes const. * __d_path() is never called with NULL/NULL root. It was a kludge to start with. Instead, we have an explicit function - d_absolute_root(). Same as __d_path(), except that it doesn't get root passed and stops where it stops. apparmor and tomoyo are using it. * __d_path() returns NULL on path outside of root. The main caller is show_mountinfo() and that's precisely what we pass root for - to skip those outside chroot jail. Those who don't want that can (and do) use d_path(). * __d_path() root argument becomes const. Everyone agrees, I hope. * apparmor does *NOT* try to use __d_path() or any of its variants when it sees that path->mnt is an internal vfsmount. In that case it's definitely not mounted anywhere and dentry_path() is exactly what we want there. Handling of sysctl()-triggered weirdness is moved to that place. * if apparmor is asked to do pathname relative to chroot jail and __d_path() tells it we it's not in that jail, the sucker just calls d_absolute_path() instead. That's the other remaining caller of __d_path(), BTW. * seq_path_root() does _NOT_ return -ENAMETOOLONG (it's stupid anyway - the normal seq_file logics will take care of growing the buffer and redoing the call of ->show() just fine). However, if it gets path not reachable from root, it returns SEQ_SKIP. The only caller adjusted (i.e. stopped ignoring the return value as it used to do). Reviewed-by: John Johansen ACKed-by: John Johansen Signed-off-by: Al Viro Cc: stable@vger.kernel.org --- fs/dcache.c | 71 ++++++++++++++++++++++++++++------------------ fs/namespace.c | 20 +++++++------ fs/seq_file.c | 6 ++-- include/linux/dcache.h | 3 +- include/linux/fs.h | 1 + security/apparmor/path.c | 65 ++++++++++++++++++++++++------------------ security/tomoyo/realpath.c | 3 +- 7 files changed, 100 insertions(+), 69 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 10ba92def3f..89509b5a090 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -2439,16 +2439,14 @@ static int prepend_name(char **buffer, int *buflen, struct qstr *name) /** * prepend_path - Prepend path string to a buffer * @path: the dentry/vfsmount to report - * @root: root vfsmnt/dentry (may be modified by this function) + * @root: root vfsmnt/dentry * @buffer: pointer to the end of the buffer * @buflen: pointer to buffer length * * Caller holds the rename_lock. - * - * If path is not reachable from the supplied root, then the value of - * root is changed (without modifying refcounts). */ -static int prepend_path(const struct path *path, struct path *root, +static int prepend_path(const struct path *path, + const struct path *root, char **buffer, int *buflen) { struct dentry *dentry = path->dentry; @@ -2483,10 +2481,10 @@ static int prepend_path(const struct path *path, struct path *root, dentry = parent; } -out: if (!error && !slash) error = prepend(buffer, buflen, "/", 1); +out: br_read_unlock(vfsmount_lock); return error; @@ -2500,15 +2498,17 @@ global_root: WARN(1, "Root dentry has weird name <%.*s>\n", (int) dentry->d_name.len, dentry->d_name.name); } - root->mnt = vfsmnt; - root->dentry = dentry; + if (!slash) + error = prepend(buffer, buflen, "/", 1); + if (!error) + error = vfsmnt->mnt_ns ? 1 : 2; goto out; } /** * __d_path - return the path of a dentry * @path: the dentry/vfsmount to report - * @root: root vfsmnt/dentry (may be modified by this function) + * @root: root vfsmnt/dentry * @buf: buffer to return value in * @buflen: buffer length * @@ -2519,10 +2519,10 @@ global_root: * * "buflen" should be positive. * - * If path is not reachable from the supplied root, then the value of - * root is changed (without modifying refcounts). + * If the path is not reachable from the supplied root, return %NULL. */ -char *__d_path(const struct path *path, struct path *root, +char *__d_path(const struct path *path, + const struct path *root, char *buf, int buflen) { char *res = buf + buflen; @@ -2533,7 +2533,28 @@ char *__d_path(const struct path *path, struct path *root, error = prepend_path(path, root, &res, &buflen); write_sequnlock(&rename_lock); - if (error) + if (error < 0) + return ERR_PTR(error); + if (error > 0) + return NULL; + return res; +} + +char *d_absolute_path(const struct path *path, + char *buf, int buflen) +{ + struct path root = {}; + char *res = buf + buflen; + int error; + + prepend(&res, &buflen, "\0", 1); + write_seqlock(&rename_lock); + error = prepend_path(path, &root, &res, &buflen); + write_sequnlock(&rename_lock); + + if (error > 1) + error = -EINVAL; + if (error < 0) return ERR_PTR(error); return res; } @@ -2541,8 +2562,9 @@ char *__d_path(const struct path *path, struct path *root, /* * same as __d_path but appends "(deleted)" for unlinked files. */ -static int path_with_deleted(const struct path *path, struct path *root, - char **buf, int *buflen) +static int path_with_deleted(const struct path *path, + const struct path *root, + char **buf, int *buflen) { prepend(buf, buflen, "\0", 1); if (d_unlinked(path->dentry)) { @@ -2579,7 +2601,6 @@ char *d_path(const struct path *path, char *buf, int buflen) { char *res = buf + buflen; struct path root; - struct path tmp; int error; /* @@ -2594,9 +2615,8 @@ char *d_path(const struct path *path, char *buf, int buflen) get_fs_root(current->fs, &root); write_seqlock(&rename_lock); - tmp = root; - error = path_with_deleted(path, &tmp, &res, &buflen); - if (error) + error = path_with_deleted(path, &root, &res, &buflen); + if (error < 0) res = ERR_PTR(error); write_sequnlock(&rename_lock); path_put(&root); @@ -2617,7 +2637,6 @@ char *d_path_with_unreachable(const struct path *path, char *buf, int buflen) { char *res = buf + buflen; struct path root; - struct path tmp; int error; if (path->dentry->d_op && path->dentry->d_op->d_dname) @@ -2625,9 +2644,8 @@ char *d_path_with_unreachable(const struct path *path, char *buf, int buflen) get_fs_root(current->fs, &root); write_seqlock(&rename_lock); - tmp = root; - error = path_with_deleted(path, &tmp, &res, &buflen); - if (!error && !path_equal(&tmp, &root)) + error = path_with_deleted(path, &root, &res, &buflen); + if (error > 0) error = prepend_unreachable(&res, &buflen); write_sequnlock(&rename_lock); path_put(&root); @@ -2758,19 +2776,18 @@ SYSCALL_DEFINE2(getcwd, char __user *, buf, unsigned long, size) write_seqlock(&rename_lock); if (!d_unlinked(pwd.dentry)) { unsigned long len; - struct path tmp = root; char *cwd = page + PAGE_SIZE; int buflen = PAGE_SIZE; prepend(&cwd, &buflen, "\0", 1); - error = prepend_path(&pwd, &tmp, &cwd, &buflen); + error = prepend_path(&pwd, &root, &cwd, &buflen); write_sequnlock(&rename_lock); - if (error) + if (error < 0) goto out; /* Unreachable from current root */ - if (!path_equal(&tmp, &root)) { + if (error > 0) { error = prepend_unreachable(&cwd, &buflen); if (error) goto out; diff --git a/fs/namespace.c b/fs/namespace.c index 6d3a1963879..cfc6d4448aa 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1048,15 +1048,12 @@ static int show_mountinfo(struct seq_file *m, void *v) if (err) goto out; seq_putc(m, ' '); - seq_path_root(m, &mnt_path, &root, " \t\n\\"); - if (root.mnt != p->root.mnt || root.dentry != p->root.dentry) { - /* - * Mountpoint is outside root, discard that one. Ugly, - * but less so than trying to do that in iterator in a - * race-free way (due to renames). - */ - return SEQ_SKIP; - } + + /* mountpoints outside of chroot jail will give SEQ_SKIP on this */ + err = seq_path_root(m, &mnt_path, &root, " \t\n\\"); + if (err) + goto out; + seq_puts(m, mnt->mnt_flags & MNT_READONLY ? " ro" : " rw"); show_mnt_opts(m, mnt); @@ -2776,3 +2773,8 @@ void kern_unmount(struct vfsmount *mnt) } } EXPORT_SYMBOL(kern_unmount); + +bool our_mnt(struct vfsmount *mnt) +{ + return check_mnt(mnt); +} diff --git a/fs/seq_file.c b/fs/seq_file.c index 05d6b0e78c9..dba43c3ea3a 100644 --- a/fs/seq_file.c +++ b/fs/seq_file.c @@ -449,8 +449,6 @@ EXPORT_SYMBOL(seq_path); /* * Same as seq_path, but relative to supplied root. - * - * root may be changed, see __d_path(). */ int seq_path_root(struct seq_file *m, struct path *path, struct path *root, char *esc) @@ -463,6 +461,8 @@ int seq_path_root(struct seq_file *m, struct path *path, struct path *root, char *p; p = __d_path(path, root, buf, size); + if (!p) + return SEQ_SKIP; res = PTR_ERR(p); if (!IS_ERR(p)) { char *end = mangle_path(buf, p, esc); @@ -474,7 +474,7 @@ int seq_path_root(struct seq_file *m, struct path *path, struct path *root, } seq_commit(m, res); - return res < 0 ? res : 0; + return res < 0 && res != -ENAMETOOLONG ? res : 0; } /* diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 4df92619936..ed9f74f6c51 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -339,7 +339,8 @@ extern int d_validate(struct dentry *, struct dentry *); */ extern char *dynamic_dname(struct dentry *, char *, int, const char *, ...); -extern char *__d_path(const struct path *path, struct path *root, char *, int); +extern char *__d_path(const struct path *, const struct path *, char *, int); +extern char *d_absolute_path(const struct path *, char *, int); extern char *d_path(const struct path *, char *, int); extern char *d_path_with_unreachable(const struct path *, char *, int); extern char *dentry_path_raw(struct dentry *, char *, int); diff --git a/include/linux/fs.h b/include/linux/fs.h index e3130220ce3..019dc558df1 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1942,6 +1942,7 @@ extern int fd_statfs(int, struct kstatfs *); extern int statfs_by_dentry(struct dentry *, struct kstatfs *); extern int freeze_super(struct super_block *super); extern int thaw_super(struct super_block *super); +extern bool our_mnt(struct vfsmount *mnt); extern int current_umask(void); diff --git a/security/apparmor/path.c b/security/apparmor/path.c index 36cc0cc39e7..b566eba4a65 100644 --- a/security/apparmor/path.c +++ b/security/apparmor/path.c @@ -57,23 +57,44 @@ static int prepend(char **buffer, int buflen, const char *str, int namelen) static int d_namespace_path(struct path *path, char *buf, int buflen, char **name, int flags) { - struct path root, tmp; char *res; - int connected, error = 0; + int error = 0; + int connected = 1; + + if (path->mnt->mnt_flags & MNT_INTERNAL) { + /* it's not mounted anywhere */ + res = dentry_path(path->dentry, buf, buflen); + *name = res; + if (IS_ERR(res)) { + *name = buf; + return PTR_ERR(res); + } + if (path->dentry->d_sb->s_magic == PROC_SUPER_MAGIC && + strncmp(*name, "/sys/", 5) == 0) { + /* TODO: convert over to using a per namespace + * control instead of hard coded /proc + */ + return prepend(name, *name - buf, "/proc", 5); + } + return 0; + } - /* Get the root we want to resolve too, released below */ + /* resolve paths relative to chroot?*/ if (flags & PATH_CHROOT_REL) { - /* resolve paths relative to chroot */ + struct path root; get_fs_root(current->fs, &root); - } else { - /* resolve paths relative to namespace */ - root.mnt = current->nsproxy->mnt_ns->root; - root.dentry = root.mnt->mnt_root; - path_get(&root); + res = __d_path(path, &root, buf, buflen); + if (res && !IS_ERR(res)) { + /* everything's fine */ + *name = res; + path_put(&root); + goto ok; + } + path_put(&root); + connected = 0; } - tmp = root; - res = __d_path(path, &tmp, buf, buflen); + res = d_absolute_path(path, buf, buflen); *name = res; /* handle error conditions - and still allow a partial path to @@ -84,7 +105,10 @@ static int d_namespace_path(struct path *path, char *buf, int buflen, *name = buf; goto out; } + if (!our_mnt(path->mnt)) + connected = 0; +ok: /* Handle two cases: * 1. A deleted dentry && profile is not allowing mediation of deleted * 2. On some filesystems, newly allocated dentries appear to the @@ -97,10 +121,7 @@ static int d_namespace_path(struct path *path, char *buf, int buflen, goto out; } - /* Determine if the path is connected to the expected root */ - connected = tmp.dentry == root.dentry && tmp.mnt == root.mnt; - - /* If the path is not connected, + /* If the path is not connected to the expected root, * check if it is a sysctl and handle specially else remove any * leading / that __d_path may have returned. * Unless @@ -112,17 +133,9 @@ static int d_namespace_path(struct path *path, char *buf, int buflen, * namespace root. */ if (!connected) { - /* is the disconnect path a sysctl? */ - if (tmp.dentry->d_sb->s_magic == PROC_SUPER_MAGIC && - strncmp(*name, "/sys/", 5) == 0) { - /* TODO: convert over to using a per namespace - * control instead of hard coded /proc - */ - error = prepend(name, *name - buf, "/proc", 5); - } else if (!(flags & PATH_CONNECT_PATH) && + if (!(flags & PATH_CONNECT_PATH) && !(((flags & CHROOT_NSCONNECT) == CHROOT_NSCONNECT) && - (tmp.mnt == current->nsproxy->mnt_ns->root && - tmp.dentry == tmp.mnt->mnt_root))) { + our_mnt(path->mnt))) { /* disconnected path, don't return pathname starting * with '/' */ @@ -133,8 +146,6 @@ static int d_namespace_path(struct path *path, char *buf, int buflen, } out: - path_put(&root); - return error; } diff --git a/security/tomoyo/realpath.c b/security/tomoyo/realpath.c index 738bbdf8d4c..36fa7c9bedc 100644 --- a/security/tomoyo/realpath.c +++ b/security/tomoyo/realpath.c @@ -101,9 +101,8 @@ static char *tomoyo_get_absolute_path(struct path *path, char * const buffer, { char *pos = ERR_PTR(-ENOMEM); if (buflen >= 256) { - struct path ns_root = { }; /* go to whatever namespace root we are under */ - pos = __d_path(path, &ns_root, buffer, buflen - 1); + pos = d_absolute_path(path, buffer, buflen - 1); if (!IS_ERR(pos) && *pos == '/' && pos[1]) { struct inode *inode = path->dentry->d_inode; if (inode && S_ISDIR(inode->i_mode)) { -- cgit v1.2.3-70-g09d2 From be655596b3de5873f994ddbe205751a5ffb4de39 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Wed, 30 Nov 2011 09:47:09 -0800 Subject: ceph: use i_ceph_lock instead of i_lock We have been using i_lock to protect all kinds of data structures in the ceph_inode_info struct, including lists of inodes that we need to iterate over while avoiding races with inode destruction. That requires grabbing a reference to the inode with the list lock protected, but igrab() now takes i_lock to check the inode flags. Changing the list lock ordering would be a painful process. However, using a ceph-specific i_ceph_lock in the ceph inode instead of i_lock is a simple mechanical change and avoids the ordering constraints imposed by igrab(). Reported-by: Amon Ott Signed-off-by: Sage Weil --- fs/ceph/addr.c | 8 +-- fs/ceph/caps.c | 187 ++++++++++++++++++++++++++------------------------- fs/ceph/dir.c | 24 +++---- fs/ceph/file.c | 20 +++--- fs/ceph/inode.c | 53 ++++++++------- fs/ceph/ioctl.c | 4 +- fs/ceph/mds_client.c | 32 ++++----- fs/ceph/mds_client.h | 2 +- fs/ceph/snap.c | 16 ++--- fs/ceph/super.h | 31 ++++----- fs/ceph/xattr.c | 42 ++++++------ 11 files changed, 212 insertions(+), 207 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 4144caf2f9d..173b1d22e59 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -87,7 +87,7 @@ static int ceph_set_page_dirty(struct page *page) snapc = ceph_get_snap_context(ci->i_snap_realm->cached_context); /* dirty the head */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_head_snapc == NULL) ci->i_head_snapc = ceph_get_snap_context(snapc); ++ci->i_wrbuffer_ref_head; @@ -100,7 +100,7 @@ static int ceph_set_page_dirty(struct page *page) ci->i_wrbuffer_ref-1, ci->i_wrbuffer_ref_head-1, ci->i_wrbuffer_ref, ci->i_wrbuffer_ref_head, snapc, snapc->seq, snapc->num_snaps); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); /* now adjust page */ spin_lock_irq(&mapping->tree_lock); @@ -391,7 +391,7 @@ static struct ceph_snap_context *get_oldest_context(struct inode *inode, struct ceph_snap_context *snapc = NULL; struct ceph_cap_snap *capsnap = NULL; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { dout(" cap_snap %p snapc %p has %d dirty pages\n", capsnap, capsnap->context, capsnap->dirty_pages); @@ -407,7 +407,7 @@ static struct ceph_snap_context *get_oldest_context(struct inode *inode, dout(" head snapc %p has %d dirty pages\n", snapc, ci->i_wrbuffer_ref_head); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return snapc; } diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 0f327c6c967..8b53193e4f7 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -309,7 +309,7 @@ void ceph_reservation_status(struct ceph_fs_client *fsc, /* * Find ceph_cap for given mds, if any. * - * Called with i_lock held. + * Called with i_ceph_lock held. */ static struct ceph_cap *__get_cap_for_mds(struct ceph_inode_info *ci, int mds) { @@ -332,9 +332,9 @@ struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci, int mds) { struct ceph_cap *cap; - spin_lock(&ci->vfs_inode.i_lock); + spin_lock(&ci->i_ceph_lock); cap = __get_cap_for_mds(ci, mds); - spin_unlock(&ci->vfs_inode.i_lock); + spin_unlock(&ci->i_ceph_lock); return cap; } @@ -361,15 +361,16 @@ static int __ceph_get_cap_mds(struct ceph_inode_info *ci) int ceph_get_cap_mds(struct inode *inode) { + struct ceph_inode_info *ci = ceph_inode(inode); int mds; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); mds = __ceph_get_cap_mds(ceph_inode(inode)); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return mds; } /* - * Called under i_lock. + * Called under i_ceph_lock. */ static void __insert_cap_node(struct ceph_inode_info *ci, struct ceph_cap *new) @@ -415,7 +416,7 @@ static void __cap_set_timeouts(struct ceph_mds_client *mdsc, * * If I_FLUSH is set, leave the inode at the front of the list. * - * Caller holds i_lock + * Caller holds i_ceph_lock * -> we take mdsc->cap_delay_lock */ static void __cap_delay_requeue(struct ceph_mds_client *mdsc, @@ -457,7 +458,7 @@ static void __cap_delay_requeue_front(struct ceph_mds_client *mdsc, /* * Cancel delayed work on cap. * - * Caller must hold i_lock. + * Caller must hold i_ceph_lock. */ static void __cap_delay_cancel(struct ceph_mds_client *mdsc, struct ceph_inode_info *ci) @@ -532,14 +533,14 @@ int ceph_add_cap(struct inode *inode, wanted |= ceph_caps_for_mode(fmode); retry: - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap = __get_cap_for_mds(ci, mds); if (!cap) { if (new_cap) { cap = new_cap; new_cap = NULL; } else { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); new_cap = get_cap(mdsc, caps_reservation); if (new_cap == NULL) return -ENOMEM; @@ -625,7 +626,7 @@ retry: if (fmode >= 0) __ceph_get_fmode(ci, fmode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); wake_up_all(&ci->i_cap_wq); return 0; } @@ -792,7 +793,7 @@ int ceph_caps_revoking(struct ceph_inode_info *ci, int mask) struct rb_node *p; int ret = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { cap = rb_entry(p, struct ceph_cap, ci_node); if (__cap_is_valid(cap) && @@ -801,7 +802,7 @@ int ceph_caps_revoking(struct ceph_inode_info *ci, int mask) break; } } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("ceph_caps_revoking %p %s = %d\n", inode, ceph_cap_string(mask), ret); return ret; @@ -855,7 +856,7 @@ int __ceph_caps_mds_wanted(struct ceph_inode_info *ci) } /* - * called under i_lock + * called under i_ceph_lock */ static int __ceph_is_any_caps(struct ceph_inode_info *ci) { @@ -865,7 +866,7 @@ static int __ceph_is_any_caps(struct ceph_inode_info *ci) /* * Remove a cap. Take steps to deal with a racing iterate_session_caps. * - * caller should hold i_lock. + * caller should hold i_ceph_lock. * caller will not hold session s_mutex if called from destroy_inode. */ void __ceph_remove_cap(struct ceph_cap *cap) @@ -1028,7 +1029,7 @@ static void __queue_cap_release(struct ceph_mds_session *session, /* * Queue cap releases when an inode is dropped from our cache. Since - * inode is about to be destroyed, there is no need for i_lock. + * inode is about to be destroyed, there is no need for i_ceph_lock. */ void ceph_queue_caps_release(struct inode *inode) { @@ -1049,7 +1050,7 @@ void ceph_queue_caps_release(struct inode *inode) /* * Send a cap msg on the given inode. Update our caps state, then - * drop i_lock and send the message. + * drop i_ceph_lock and send the message. * * Make note of max_size reported/requested from mds, revoked caps * that have now been implemented. @@ -1061,13 +1062,13 @@ void ceph_queue_caps_release(struct inode *inode) * Return non-zero if delayed release, or we experienced an error * such that the caller should requeue + retry later. * - * called with i_lock, then drops it. + * called with i_ceph_lock, then drops it. * caller should hold snap_rwsem (read), s_mutex. */ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, int op, int used, int want, int retain, int flushing, unsigned *pflush_tid) - __releases(cap->ci->vfs_inode->i_lock) + __releases(cap->ci->i_ceph_lock) { struct ceph_inode_info *ci = cap->ci; struct inode *inode = &ci->vfs_inode; @@ -1170,7 +1171,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, xattr_version = ci->i_xattrs.version; } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); ret = send_cap_msg(session, ceph_vino(inode).ino, cap_id, op, keep, want, flushing, seq, flush_tid, issue_seq, mseq, @@ -1198,13 +1199,13 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, * Unless @again is true, skip cap_snaps that were already sent to * the MDS (i.e., during this session). * - * Called under i_lock. Takes s_mutex as needed. + * Called under i_ceph_lock. Takes s_mutex as needed. */ void __ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_session **psession, int again) - __releases(ci->vfs_inode->i_lock) - __acquires(ci->vfs_inode->i_lock) + __releases(ci->i_ceph_lock) + __acquires(ci->i_ceph_lock) { struct inode *inode = &ci->vfs_inode; int mds; @@ -1261,7 +1262,7 @@ retry: session = NULL; } if (!session) { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); mutex_lock(&mdsc->mutex); session = __ceph_lookup_mds_session(mdsc, mds); mutex_unlock(&mdsc->mutex); @@ -1275,7 +1276,7 @@ retry: * deletion or migration. retry, and we'll * get a better @mds value next time. */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); goto retry; } @@ -1285,7 +1286,7 @@ retry: list_del_init(&capsnap->flushing_item); list_add_tail(&capsnap->flushing_item, &session->s_cap_snaps_flushing); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("flush_snaps %p cap_snap %p follows %lld tid %llu\n", inode, capsnap, capsnap->follows, capsnap->flush_tid); @@ -1302,7 +1303,7 @@ retry: next_follows = capsnap->follows + 1; ceph_put_cap_snap(capsnap); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); goto retry; } @@ -1322,11 +1323,9 @@ out: static void ceph_flush_snaps(struct ceph_inode_info *ci) { - struct inode *inode = &ci->vfs_inode; - - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __ceph_flush_snaps(ci, NULL, 0); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } /* @@ -1373,7 +1372,7 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask) * Add dirty inode to the flushing list. Assigned a seq number so we * can wait for caps to flush without starving. * - * Called under i_lock. + * Called under i_ceph_lock. */ static int __mark_caps_flushing(struct inode *inode, struct ceph_mds_session *session) @@ -1421,9 +1420,9 @@ static int try_nonblocking_invalidate(struct inode *inode) struct ceph_inode_info *ci = ceph_inode(inode); u32 invalidating_gen = ci->i_rdcache_gen; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); invalidate_mapping_pages(&inode->i_data, 0, -1); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (inode->i_data.nrpages == 0 && invalidating_gen == ci->i_rdcache_gen) { @@ -1470,7 +1469,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, if (mdsc->stopping) is_delayed = 1; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_ceph_flags & CEPH_I_FLUSH) flags |= CHECK_CAPS_FLUSH; @@ -1480,7 +1479,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, __ceph_flush_snaps(ci, &session, 0); goto retry_locked; retry: - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); retry_locked: file_wanted = __ceph_caps_file_wanted(ci); used = __ceph_caps_used(ci); @@ -1634,7 +1633,7 @@ ack: if (mutex_trylock(&session->s_mutex) == 0) { dout("inverting session/ino locks on %p\n", session); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (took_snap_rwsem) { up_read(&mdsc->snap_rwsem); took_snap_rwsem = 0; @@ -1648,7 +1647,7 @@ ack: if (down_read_trylock(&mdsc->snap_rwsem) == 0) { dout("inverting snap/in locks on %p\n", inode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); down_read(&mdsc->snap_rwsem); took_snap_rwsem = 1; goto retry; @@ -1664,10 +1663,10 @@ ack: mds = cap->mds; /* remember mds, so we don't repeat */ sent++; - /* __send_cap drops i_lock */ + /* __send_cap drops i_ceph_lock */ delayed += __send_cap(mdsc, cap, CEPH_CAP_OP_UPDATE, used, want, retain, flushing, NULL); - goto retry; /* retake i_lock and restart our cap scan. */ + goto retry; /* retake i_ceph_lock and restart our cap scan. */ } /* @@ -1681,7 +1680,7 @@ ack: else if (!is_delayed || force_requeue) __cap_delay_requeue(mdsc, ci); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (queue_invalidate) ceph_queue_invalidate(inode); @@ -1704,7 +1703,7 @@ static int try_flush_caps(struct inode *inode, struct ceph_mds_session *session, int flushing = 0; retry: - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_ceph_flags & CEPH_I_NOFLUSH) { dout("try_flush_caps skipping %p I_NOFLUSH set\n", inode); goto out; @@ -1716,7 +1715,7 @@ retry: int delayed; if (!session) { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); session = cap->session; mutex_lock(&session->s_mutex); goto retry; @@ -1727,18 +1726,18 @@ retry: flushing = __mark_caps_flushing(inode, session); - /* __send_cap drops i_lock */ + /* __send_cap drops i_ceph_lock */ delayed = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, used, want, cap->issued | cap->implemented, flushing, flush_tid); if (!delayed) goto out_unlocked; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __cap_delay_requeue(mdsc, ci); } out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); out_unlocked: if (session && unlock_session) mutex_unlock(&session->s_mutex); @@ -1753,7 +1752,7 @@ static int caps_are_flushed(struct inode *inode, unsigned tid) struct ceph_inode_info *ci = ceph_inode(inode); int i, ret = 1; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); for (i = 0; i < CEPH_CAP_BITS; i++) if ((ci->i_flushing_caps & (1 << i)) && ci->i_cap_flush_tid[i] <= tid) { @@ -1761,7 +1760,7 @@ static int caps_are_flushed(struct inode *inode, unsigned tid) ret = 0; break; } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return ret; } @@ -1868,10 +1867,10 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc) struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (__ceph_caps_dirty(ci)) __cap_delay_requeue_front(mdsc, ci); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } return err; } @@ -1894,7 +1893,7 @@ static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc, struct inode *inode = &ci->vfs_inode; struct ceph_cap *cap; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; if (cap && cap->session == session) { dout("kick_flushing_caps %p cap %p capsnap %p\n", inode, @@ -1904,7 +1903,7 @@ static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc, pr_err("%p auth cap %p not mds%d ???\n", inode, cap, session->s_mds); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } } @@ -1921,7 +1920,7 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_cap *cap; int delayed = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; if (cap && cap->session == session) { dout("kick_flushing_caps %p cap %p %s\n", inode, @@ -1932,14 +1931,14 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, cap->issued | cap->implemented, ci->i_flushing_caps, NULL); if (delayed) { - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __cap_delay_requeue(mdsc, ci); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } } else { pr_err("%p auth cap %p not mds%d ???\n", inode, cap, session->s_mds); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } } } @@ -1952,7 +1951,7 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, struct ceph_cap *cap; int delayed = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; dout("kick_flushing_inode_caps %p flushing %s flush_seq %lld\n", inode, ceph_cap_string(ci->i_flushing_caps), ci->i_cap_flush_seq); @@ -1964,12 +1963,12 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, cap->issued | cap->implemented, ci->i_flushing_caps, NULL); if (delayed) { - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __cap_delay_requeue(mdsc, ci); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } } else { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } } @@ -1978,7 +1977,7 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, * Take references to capabilities we hold, so that we don't release * them to the MDS prematurely. * - * Protected by i_lock. + * Protected by i_ceph_lock. */ static void __take_cap_refs(struct ceph_inode_info *ci, int got) { @@ -2016,7 +2015,7 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want, dout("get_cap_refs %p need %s want %s\n", inode, ceph_cap_string(need), ceph_cap_string(want)); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); /* make sure file is actually open */ file_wanted = __ceph_caps_file_wanted(ci); @@ -2077,7 +2076,7 @@ static int try_get_cap_refs(struct ceph_inode_info *ci, int need, int want, ceph_cap_string(have), ceph_cap_string(need)); } out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("get_cap_refs %p ret %d got %s\n", inode, ret, ceph_cap_string(*got)); return ret; @@ -2094,7 +2093,7 @@ static void check_max_size(struct inode *inode, loff_t endoff) int check = 0; /* do we need to explicitly request a larger max_size? */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if ((endoff >= ci->i_max_size || endoff > (inode->i_size << 1)) && endoff > ci->i_wanted_max_size) { @@ -2103,7 +2102,7 @@ static void check_max_size(struct inode *inode, loff_t endoff) ci->i_wanted_max_size = endoff; check = 1; } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (check) ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); } @@ -2140,9 +2139,9 @@ retry: */ void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps) { - spin_lock(&ci->vfs_inode.i_lock); + spin_lock(&ci->i_ceph_lock); __take_cap_refs(ci, caps); - spin_unlock(&ci->vfs_inode.i_lock); + spin_unlock(&ci->i_ceph_lock); } /* @@ -2160,7 +2159,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) int last = 0, put = 0, flushsnaps = 0, wake = 0; struct ceph_cap_snap *capsnap; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (had & CEPH_CAP_PIN) --ci->i_pin_ref; if (had & CEPH_CAP_FILE_RD) @@ -2193,7 +2192,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) } } } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("put_cap_refs %p had %s%s%s\n", inode, ceph_cap_string(had), last ? " last" : "", put ? " put" : ""); @@ -2225,7 +2224,7 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, int found = 0; struct ceph_cap_snap *capsnap = NULL; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_wrbuffer_ref -= nr; last = !ci->i_wrbuffer_ref; @@ -2274,7 +2273,7 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, } } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (last) { ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); @@ -2291,7 +2290,7 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, * Handle a cap GRANT message from the MDS. (Note that a GRANT may * actually be a revocation if it specifies a smaller cap set.) * - * caller holds s_mutex and i_lock, we drop both. + * caller holds s_mutex and i_ceph_lock, we drop both. * * return value: * 0 - ok @@ -2302,7 +2301,7 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant, struct ceph_mds_session *session, struct ceph_cap *cap, struct ceph_buffer *xattr_buf) - __releases(inode->i_lock) + __releases(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); int mds = session->s_mds; @@ -2453,7 +2452,7 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant, } BUG_ON(cap->issued & ~cap->implemented); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (writeback) /* * queue inode for writeback: we can't actually call @@ -2483,7 +2482,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, struct ceph_mds_caps *m, struct ceph_mds_session *session, struct ceph_cap *cap) - __releases(inode->i_lock) + __releases(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; @@ -2539,7 +2538,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, wake_up_all(&ci->i_cap_wq); out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (drop) iput(inode); } @@ -2562,7 +2561,7 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, dout("handle_cap_flushsnap_ack inode %p ci %p mds%d follows %lld\n", inode, ci, session->s_mds, follows); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->follows == follows) { if (capsnap->flush_tid != flush_tid) { @@ -2585,7 +2584,7 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, capsnap, capsnap->follows); } } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (drop) iput(inode); } @@ -2598,7 +2597,7 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, static void handle_cap_trunc(struct inode *inode, struct ceph_mds_caps *trunc, struct ceph_mds_session *session) - __releases(inode->i_lock) + __releases(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); int mds = session->s_mds; @@ -2617,7 +2616,7 @@ static void handle_cap_trunc(struct inode *inode, inode, mds, seq, truncate_size, truncate_seq); queue_trunc = ceph_fill_file_size(inode, issued, truncate_seq, truncate_size, size); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (queue_trunc) ceph_queue_vmtruncate(inode); @@ -2646,7 +2645,7 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, dout("handle_cap_export inode %p ci %p mds%d mseq %d\n", inode, ci, mds, mseq); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); /* make sure we haven't seen a higher mseq */ for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) { @@ -2690,7 +2689,7 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex, } /* else, we already released it */ - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } /* @@ -2745,9 +2744,9 @@ static void handle_cap_import(struct ceph_mds_client *mdsc, up_read(&mdsc->snap_rwsem); /* make sure we re-request max_size, if necessary */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_requested_max_size = 0; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } /* @@ -2762,6 +2761,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, struct ceph_mds_client *mdsc = session->s_mdsc; struct super_block *sb = mdsc->fsc->sb; struct inode *inode; + struct ceph_inode_info *ci; struct ceph_cap *cap; struct ceph_mds_caps *h; int mds = session->s_mds; @@ -2815,6 +2815,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, /* lookup ino */ inode = ceph_find_inode(sb, vino); + ci = ceph_inode(inode); dout(" op %s ino %llx.%llx inode %p\n", ceph_cap_op_name(op), vino.ino, vino.snap, inode); if (!inode) { @@ -2844,16 +2845,16 @@ void ceph_handle_caps(struct ceph_mds_session *session, } /* the rest require a cap */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap = __get_cap_for_mds(ceph_inode(inode), mds); if (!cap) { dout(" no cap on %p ino %llx.%llx from mds%d\n", inode, ceph_ino(inode), ceph_snap(inode), mds); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); goto flush_cap_releases; } - /* note that each of these drops i_lock for us */ + /* note that each of these drops i_ceph_lock for us */ switch (op) { case CEPH_CAP_OP_REVOKE: case CEPH_CAP_OP_GRANT: @@ -2869,7 +2870,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, break; default: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); pr_err("ceph_handle_caps: unknown cap op %d %s\n", op, ceph_cap_op_name(op)); } @@ -2962,13 +2963,13 @@ void ceph_put_fmode(struct ceph_inode_info *ci, int fmode) struct inode *inode = &ci->vfs_inode; int last = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dout("put_fmode %p fmode %d %d -> %d\n", inode, fmode, ci->i_nr_by_mode[fmode], ci->i_nr_by_mode[fmode]-1); BUG_ON(ci->i_nr_by_mode[fmode] == 0); if (--ci->i_nr_by_mode[fmode] == 0) last++; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (last && ci->i_vino.snap == CEPH_NOSNAP) ceph_check_caps(ci, 0, NULL); @@ -2991,7 +2992,7 @@ int ceph_encode_inode_release(void **p, struct inode *inode, int used, dirty; int ret = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); used = __ceph_caps_used(ci); dirty = __ceph_caps_dirty(ci); @@ -3046,7 +3047,7 @@ int ceph_encode_inode_release(void **p, struct inode *inode, inode, cap, ceph_cap_string(cap->issued)); } } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return ret; } @@ -3061,7 +3062,7 @@ int ceph_encode_dentry_release(void **p, struct dentry *dentry, /* * force an record for the directory caps if we have a dentry lease. - * this is racy (can't take i_lock and d_lock together), but it + * this is racy (can't take i_ceph_lock and d_lock together), but it * doesn't have to be perfect; the mds will revoke anything we don't * release. */ diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index bca3948e9db..3eeb9766126 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -281,18 +281,18 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir) } /* can we use the dcache? */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if ((filp->f_pos == 2 || fi->dentry) && !ceph_test_mount_opt(fsc, NOASYNCREADDIR) && ceph_snap(inode) != CEPH_SNAPDIR && ceph_dir_test_complete(inode) && __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); err = __dcache_readdir(filp, dirent, filldir); if (err != -EAGAIN) return err; } else { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } if (fi->dentry) { err = note_last_dentry(fi, fi->dentry->d_name.name, @@ -428,12 +428,12 @@ more: * were released during the whole readdir, and we should have * the complete dir contents in our cache. */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_release_count == fi->dir_release_count) { ceph_dir_set_complete(inode); ci->i_max_offset = filp->f_pos; } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("readdir %p filp %p done.\n", inode, filp); return 0; @@ -607,7 +607,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, struct ceph_inode_info *ci = ceph_inode(dir); struct ceph_dentry_info *di = ceph_dentry(dentry); - spin_lock(&dir->i_lock); + spin_lock(&ci->i_ceph_lock); dout(" dir %p flags are %d\n", dir, ci->i_ceph_flags); if (strncmp(dentry->d_name.name, fsc->mount_options->snapdir_name, @@ -615,13 +615,13 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry, !is_root_ceph_dentry(dir, dentry) && ceph_dir_test_complete(dir) && (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) { - spin_unlock(&dir->i_lock); + spin_unlock(&ci->i_ceph_lock); dout(" dir %p complete, -ENOENT\n", dir); d_add(dentry, NULL); di->lease_shared_gen = ci->i_shared_gen; return NULL; } - spin_unlock(&dir->i_lock); + spin_unlock(&ci->i_ceph_lock); } op = ceph_snap(dir) == CEPH_SNAPDIR ? @@ -841,12 +841,12 @@ static int drop_caps_for_unlink(struct inode *inode) struct ceph_inode_info *ci = ceph_inode(inode); int drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (inode->i_nlink == 1) { drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN); ci->i_ceph_flags |= CEPH_I_NODELAY; } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return drop; } @@ -1015,10 +1015,10 @@ static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry) struct ceph_dentry_info *di = ceph_dentry(dentry); int valid = 0; - spin_lock(&dir->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_shared_gen == di->lease_shared_gen) valid = __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1); - spin_unlock(&dir->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("dir_lease_is_valid dir %p v%u dentry %p v%u = %d\n", dir, (unsigned)ci->i_shared_gen, dentry, (unsigned)di->lease_shared_gen, valid); diff --git a/fs/ceph/file.c b/fs/ceph/file.c index ce549d31eeb..af187330cc8 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -147,9 +147,9 @@ int ceph_open(struct inode *inode, struct file *file) /* trivially open snapdir */ if (ceph_snap(inode) == CEPH_SNAPDIR) { - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __ceph_get_fmode(ci, fmode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return ceph_init_file(inode, file, fmode); } @@ -158,7 +158,7 @@ int ceph_open(struct inode *inode, struct file *file) * write) or any MDS (for read). Update wanted set * asynchronously. */ - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (__ceph_is_any_real_caps(ci) && (((fmode & CEPH_FILE_MODE_WR) == 0) || ci->i_auth_cap)) { int mds_wanted = __ceph_caps_mds_wanted(ci); @@ -168,7 +168,7 @@ int ceph_open(struct inode *inode, struct file *file) inode, fmode, ceph_cap_string(wanted), ceph_cap_string(issued)); __ceph_get_fmode(ci, fmode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); /* adjust wanted? */ if ((issued & wanted) != wanted && @@ -180,10 +180,10 @@ int ceph_open(struct inode *inode, struct file *file) } else if (ceph_snap(inode) != CEPH_NOSNAP && (ci->i_snap_caps & wanted) == wanted) { __ceph_get_fmode(ci, fmode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return ceph_init_file(inode, file, fmode); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("open fmode %d wants %s\n", fmode, ceph_cap_string(wanted)); req = prepare_open_request(inode->i_sb, flags, 0); @@ -743,9 +743,9 @@ retry_snap: */ int dirty; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); ceph_put_cap_refs(ci, got); ret = generic_file_aio_write(iocb, iov, nr_segs, pos); @@ -764,9 +764,9 @@ retry_snap: if (ret >= 0) { int dirty; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_FILE_WR); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (dirty) __mark_inode_dirty(inode, dirty); } diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 116f36502f1..87fb132fb33 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -297,6 +297,8 @@ struct inode *ceph_alloc_inode(struct super_block *sb) dout("alloc_inode %p\n", &ci->vfs_inode); + spin_lock_init(&ci->i_ceph_lock); + ci->i_version = 0; ci->i_time_warp_seq = 0; ci->i_ceph_flags = 0; @@ -583,7 +585,7 @@ static int fill_inode(struct inode *inode, iinfo->xattr_len); } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); /* * provided version will be odd if inode value is projected, @@ -680,7 +682,7 @@ static int fill_inode(struct inode *inode, char *sym; BUG_ON(symlen != inode->i_size); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); err = -ENOMEM; sym = kmalloc(symlen+1, GFP_NOFS); @@ -689,7 +691,7 @@ static int fill_inode(struct inode *inode, memcpy(sym, iinfo->symlink, symlen); sym[symlen] = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (!ci->i_symlink) ci->i_symlink = sym; else @@ -715,7 +717,7 @@ static int fill_inode(struct inode *inode, } no_change: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); /* queue truncate if we saw i_size decrease */ if (queue_trunc) @@ -750,13 +752,13 @@ no_change: info->cap.flags, caps_reservation); } else { - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dout(" %p got snap_caps %s\n", inode, ceph_cap_string(le32_to_cpu(info->cap.caps))); ci->i_snap_caps |= le32_to_cpu(info->cap.caps); if (cap_fmode >= 0) __ceph_get_fmode(ci, cap_fmode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } } else if (cap_fmode >= 0) { pr_warning("mds issued no caps on %llx.%llx\n", @@ -849,19 +851,20 @@ static void ceph_set_dentry_offset(struct dentry *dn) { struct dentry *dir = dn->d_parent; struct inode *inode = dir->d_inode; + struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_dentry_info *di; BUG_ON(!inode); di = ceph_dentry(dn); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (!ceph_dir_test_complete(inode)) { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return; } di->offset = ceph_inode(inode)->i_max_offset++; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); spin_lock(&dir->d_lock); spin_lock_nested(&dn->d_lock, DENTRY_D_LOCK_NESTED); @@ -1308,7 +1311,7 @@ int ceph_inode_set_size(struct inode *inode, loff_t size) struct ceph_inode_info *ci = ceph_inode(inode); int ret = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dout("set_size %p %llu -> %llu\n", inode, inode->i_size, size); inode->i_size = size; inode->i_blocks = (size + (1 << 9) - 1) >> 9; @@ -1318,7 +1321,7 @@ int ceph_inode_set_size(struct inode *inode, loff_t size) (ci->i_reported_size << 1) < ci->i_max_size) ret = 1; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return ret; } @@ -1376,20 +1379,20 @@ static void ceph_invalidate_work(struct work_struct *work) u32 orig_gen; int check = 0; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dout("invalidate_pages %p gen %d revoking %d\n", inode, ci->i_rdcache_gen, ci->i_rdcache_revoking); if (ci->i_rdcache_revoking != ci->i_rdcache_gen) { /* nevermind! */ - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); goto out; } orig_gen = ci->i_rdcache_gen; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); truncate_inode_pages(&inode->i_data, 0); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (orig_gen == ci->i_rdcache_gen && orig_gen == ci->i_rdcache_revoking) { dout("invalidate_pages %p gen %d successful\n", inode, @@ -1401,7 +1404,7 @@ static void ceph_invalidate_work(struct work_struct *work) inode, orig_gen, ci->i_rdcache_gen, ci->i_rdcache_revoking); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (check) ceph_check_caps(ci, 0, NULL); @@ -1460,10 +1463,10 @@ void __ceph_do_pending_vmtruncate(struct inode *inode) int wrbuffer_refs, wake = 0; retry: - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_truncate_pending == 0) { dout("__do_pending_vmtruncate %p none pending\n", inode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return; } @@ -1474,7 +1477,7 @@ retry: if (ci->i_wrbuffer_ref_head < ci->i_wrbuffer_ref) { dout("__do_pending_vmtruncate %p flushing snaps first\n", inode); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); filemap_write_and_wait_range(&inode->i_data, 0, inode->i_sb->s_maxbytes); goto retry; @@ -1484,15 +1487,15 @@ retry: wrbuffer_refs = ci->i_wrbuffer_ref; dout("__do_pending_vmtruncate %p (%d) to %lld\n", inode, ci->i_truncate_pending, to); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); truncate_inode_pages(inode->i_mapping, to); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_truncate_pending--; if (ci->i_truncate_pending == 0) wake = 1; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (wrbuffer_refs == 0) ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); @@ -1547,7 +1550,7 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) if (IS_ERR(req)) return PTR_ERR(req); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); issued = __ceph_caps_issued(ci, NULL); dout("setattr %p issued %s\n", inode, ceph_cap_string(issued)); @@ -1695,7 +1698,7 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) } release &= issued; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (inode_dirty_flags) __mark_inode_dirty(inode, inode_dirty_flags); @@ -1717,7 +1720,7 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) __ceph_do_pending_vmtruncate(inode); return err; out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); ceph_mdsc_put_request(req); return err; } diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 5a14c29cbba..790914a598d 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -241,11 +241,11 @@ static long ceph_ioctl_lazyio(struct file *file) struct ceph_inode_info *ci = ceph_inode(inode); if ((fi->fmode & CEPH_FILE_MODE_LAZY) == 0) { - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_nr_by_mode[fi->fmode]--; fi->fmode |= CEPH_FILE_MODE_LAZY; ci->i_nr_by_mode[fi->fmode]++; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout("ioctl_layzio: file %p marked lazy\n", file); ceph_check_caps(ci, 0, NULL); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 264ab701154..34bc35084ae 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -732,21 +732,21 @@ static int __choose_mds(struct ceph_mds_client *mdsc, } } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap = NULL; if (mode == USE_AUTH_MDS) cap = ci->i_auth_cap; if (!cap && !RB_EMPTY_ROOT(&ci->i_caps)) cap = rb_entry(rb_first(&ci->i_caps), struct ceph_cap, ci_node); if (!cap) { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); goto random; } mds = cap->session->s_mds; dout("choose_mds %p %llx.%llx mds%d (%scap %p)\n", inode, ceph_vinop(inode), mds, cap == ci->i_auth_cap ? "auth " : "", cap); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return mds; random: @@ -951,7 +951,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, dout("removing cap %p, ci is %p, inode is %p\n", cap, ci, &ci->vfs_inode); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __ceph_remove_cap(cap); if (!__ceph_is_any_real_caps(ci)) { struct ceph_mds_client *mdsc = @@ -984,7 +984,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, } spin_unlock(&mdsc->cap_dirty_lock); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); while (drop--) iput(inode); return 0; @@ -1015,10 +1015,10 @@ static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap, wake_up_all(&ci->i_cap_wq); if (arg) { - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_wanted_max_size = 0; ci->i_requested_max_size = 0; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } return 0; } @@ -1151,7 +1151,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) if (session->s_trim_caps <= 0) return -1; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); mine = cap->issued | cap->implemented; used = __ceph_caps_used(ci); oissued = __ceph_caps_issued_other(ci, cap); @@ -1170,7 +1170,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) __ceph_remove_cap(cap); } else { /* try to drop referring dentries */ - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); d_prune_aliases(inode); dout("trim_caps_cb %p cap %p pruned, count now %d\n", inode, cap, atomic_read(&inode->i_count)); @@ -1178,7 +1178,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg) } out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return 0; } @@ -1296,7 +1296,7 @@ static int check_cap_flush(struct ceph_mds_client *mdsc, u64 want_flush_seq) i_flushing_item); struct inode *inode = &ci->vfs_inode; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_cap_flush_seq <= want_flush_seq) { dout("check_cap_flush still flushing %p " "seq %lld <= %lld to mds%d\n", inode, @@ -1304,7 +1304,7 @@ static int check_cap_flush(struct ceph_mds_client *mdsc, u64 want_flush_seq) session->s_mds); ret = 0; } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); @@ -2011,10 +2011,10 @@ void ceph_invalidate_dir_request(struct ceph_mds_request *req) struct ceph_inode_info *ci = ceph_inode(inode); dout("invalidate_dir_request %p (D_COMPLETE, lease(s))\n", inode); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ceph_dir_clear_complete(inode); ci->i_release_count++; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (req->r_dentry) ceph_invalidate_dentry_lease(req->r_dentry); @@ -2422,7 +2422,7 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, if (err) goto out_free; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); cap->seq = 0; /* reset cap seq */ cap->issue_seq = 0; /* and issue_seq */ @@ -2445,7 +2445,7 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, rec.v1.pathbase = cpu_to_le64(pathbase); reclen = sizeof(rec.v1); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (recon_state->flock) { int num_fcntl_locks, num_flock_locks; diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 4bb239921db..a50ca0e3947 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -20,7 +20,7 @@ * * mdsc->snap_rwsem * - * inode->i_lock + * ci->i_ceph_lock * mdsc->snap_flush_lock * mdsc->cap_delay_lock * diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index e2643719133..a559c80f127 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -446,7 +446,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) return; } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); used = __ceph_caps_used(ci); dirty = __ceph_caps_dirty(ci); @@ -528,7 +528,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) kfree(capsnap); } - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } /* @@ -537,7 +537,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) * * If capsnap can now be flushed, add to snap_flush list, and return 1. * - * Caller must hold i_lock. + * Caller must hold i_ceph_lock. */ int __ceph_finish_cap_snap(struct ceph_inode_info *ci, struct ceph_cap_snap *capsnap) @@ -739,9 +739,9 @@ static void flush_snaps(struct ceph_mds_client *mdsc) inode = &ci->vfs_inode; ihold(inode); spin_unlock(&mdsc->snap_flush_lock); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __ceph_flush_snaps(ci, &session, 0); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); iput(inode); spin_lock(&mdsc->snap_flush_lock); } @@ -847,7 +847,7 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, continue; ci = ceph_inode(inode); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (!ci->i_snap_realm) goto skip_inode; /* @@ -876,7 +876,7 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, oldrealm = ci->i_snap_realm; ci->i_snap_realm = realm; spin_unlock(&realm->inodes_with_caps_lock); - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); ceph_get_snap_realm(mdsc, realm); ceph_put_snap_realm(mdsc, oldrealm); @@ -885,7 +885,7 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc, continue; skip_inode: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); iput(inode); } diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 01bf189e08a..edcbf3774a5 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -220,7 +220,7 @@ struct ceph_dentry_info { * The locking for D_COMPLETE is a bit odd: * - we can clear it at almost any time (see ceph_d_prune) * - it is only meaningful if: - * - we hold dir inode i_lock + * - we hold dir inode i_ceph_lock * - we hold dir FILE_SHARED caps * - the dentry D_COMPLETE is set */ @@ -250,6 +250,8 @@ struct ceph_inode_xattrs_info { struct ceph_inode_info { struct ceph_vino i_vino; /* ceph ino + snap */ + spinlock_t i_ceph_lock; + u64 i_version; u32 i_time_warp_seq; @@ -271,7 +273,7 @@ struct ceph_inode_info { struct ceph_inode_xattrs_info i_xattrs; - /* capabilities. protected _both_ by i_lock and cap->session's + /* capabilities. protected _both_ by i_ceph_lock and cap->session's * s_mutex. */ struct rb_root i_caps; /* cap list */ struct ceph_cap *i_auth_cap; /* authoritative cap, if any */ @@ -437,18 +439,18 @@ static inline void ceph_i_clear(struct inode *inode, unsigned mask) { struct ceph_inode_info *ci = ceph_inode(inode); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_ceph_flags &= ~mask; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } static inline void ceph_i_set(struct inode *inode, unsigned mask) { struct ceph_inode_info *ci = ceph_inode(inode); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); ci->i_ceph_flags |= mask; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); } static inline bool ceph_i_test(struct inode *inode, unsigned mask) @@ -456,9 +458,9 @@ static inline bool ceph_i_test(struct inode *inode, unsigned mask) struct ceph_inode_info *ci = ceph_inode(inode); bool r; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); r = (ci->i_ceph_flags & mask) == mask; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return r; } @@ -508,9 +510,9 @@ extern int __ceph_caps_issued_other(struct ceph_inode_info *ci, static inline int ceph_caps_issued(struct ceph_inode_info *ci) { int issued; - spin_lock(&ci->vfs_inode.i_lock); + spin_lock(&ci->i_ceph_lock); issued = __ceph_caps_issued(ci, NULL); - spin_unlock(&ci->vfs_inode.i_lock); + spin_unlock(&ci->i_ceph_lock); return issued; } @@ -518,9 +520,9 @@ static inline int ceph_caps_issued_mask(struct ceph_inode_info *ci, int mask, int touch) { int r; - spin_lock(&ci->vfs_inode.i_lock); + spin_lock(&ci->i_ceph_lock); r = __ceph_caps_issued_mask(ci, mask, touch); - spin_unlock(&ci->vfs_inode.i_lock); + spin_unlock(&ci->i_ceph_lock); return r; } @@ -743,10 +745,9 @@ extern int ceph_add_cap(struct inode *inode, extern void __ceph_remove_cap(struct ceph_cap *cap); static inline void ceph_remove_cap(struct ceph_cap *cap) { - struct inode *inode = &cap->ci->vfs_inode; - spin_lock(&inode->i_lock); + spin_lock(&cap->ci->i_ceph_lock); __ceph_remove_cap(cap); - spin_unlock(&inode->i_lock); + spin_unlock(&cap->ci->i_ceph_lock); } extern void ceph_put_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap); diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 96c6739a028..a5e36e4488a 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -343,8 +343,8 @@ void __ceph_destroy_xattrs(struct ceph_inode_info *ci) } static int __build_xattrs(struct inode *inode) - __releases(inode->i_lock) - __acquires(inode->i_lock) + __releases(ci->i_ceph_lock) + __acquires(ci->i_ceph_lock) { u32 namelen; u32 numattr = 0; @@ -372,7 +372,7 @@ start: end = p + ci->i_xattrs.blob->vec.iov_len; ceph_decode_32_safe(&p, end, numattr, bad); xattr_version = ci->i_xattrs.version; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); xattrs = kcalloc(numattr, sizeof(struct ceph_xattr *), GFP_NOFS); @@ -387,7 +387,7 @@ start: goto bad_lock; } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_xattrs.version != xattr_version) { /* lost a race, retry */ for (i = 0; i < numattr; i++) @@ -418,7 +418,7 @@ start: return err; bad_lock: - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); bad: if (xattrs) { for (i = 0; i < numattr; i++) @@ -512,7 +512,7 @@ ssize_t ceph_getxattr(struct dentry *dentry, const char *name, void *value, if (vxattrs) vxattr = ceph_match_vxattr(vxattrs, name); - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dout("getxattr %p ver=%lld index_ver=%lld\n", inode, ci->i_xattrs.version, ci->i_xattrs.index_version); @@ -520,14 +520,14 @@ ssize_t ceph_getxattr(struct dentry *dentry, const char *name, void *value, (ci->i_xattrs.index_version >= ci->i_xattrs.version)) { goto get_xattr; } else { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); /* get xattrs from mds (if we don't already have them) */ err = ceph_do_getattr(inode, CEPH_STAT_CAP_XATTR); if (err) return err; } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (vxattr && vxattr->readonly) { err = vxattr->getxattr_cb(ci, value, size); @@ -558,7 +558,7 @@ get_xattr: memcpy(value, xattr->val, xattr->val_len); out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return err; } @@ -573,7 +573,7 @@ ssize_t ceph_listxattr(struct dentry *dentry, char *names, size_t size) u32 len; int i; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); dout("listxattr %p ver=%lld index_ver=%lld\n", inode, ci->i_xattrs.version, ci->i_xattrs.index_version); @@ -581,13 +581,13 @@ ssize_t ceph_listxattr(struct dentry *dentry, char *names, size_t size) (ci->i_xattrs.index_version >= ci->i_xattrs.version)) { goto list_xattr; } else { - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); err = ceph_do_getattr(inode, CEPH_STAT_CAP_XATTR); if (err) return err; } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); err = __build_xattrs(inode); if (err < 0) @@ -619,7 +619,7 @@ list_xattr: } out: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); return err; } @@ -739,7 +739,7 @@ int ceph_setxattr(struct dentry *dentry, const char *name, if (!xattr) goto out; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); retry: issued = __ceph_caps_issued(ci, NULL); if (!(issued & CEPH_CAP_XATTR_EXCL)) @@ -752,12 +752,12 @@ retry: required_blob_size > ci->i_xattrs.prealloc_blob->alloc_len) { struct ceph_buffer *blob = NULL; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); dout(" preaallocating new blob size=%d\n", required_blob_size); blob = ceph_buffer_new(required_blob_size, GFP_NOFS); if (!blob) goto out; - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); if (ci->i_xattrs.prealloc_blob) ceph_buffer_put(ci->i_xattrs.prealloc_blob); ci->i_xattrs.prealloc_blob = blob; @@ -770,13 +770,13 @@ retry: dirty = __ceph_mark_dirty_caps(ci, CEPH_CAP_XATTR_EXCL); ci->i_xattrs.dirty = true; inode->i_ctime = CURRENT_TIME; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (dirty) __mark_inode_dirty(inode, dirty); return err; do_sync: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); err = ceph_sync_setxattr(dentry, name, value, size, flags); out: kfree(newname); @@ -833,7 +833,7 @@ int ceph_removexattr(struct dentry *dentry, const char *name) return -EOPNOTSUPP; } - spin_lock(&inode->i_lock); + spin_lock(&ci->i_ceph_lock); __build_xattrs(inode); issued = __ceph_caps_issued(ci, NULL); dout("removexattr %p issued %s\n", inode, ceph_cap_string(issued)); @@ -846,12 +846,12 @@ int ceph_removexattr(struct dentry *dentry, const char *name) ci->i_xattrs.dirty = true; inode->i_ctime = CURRENT_TIME; - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); if (dirty) __mark_inode_dirty(inode, dirty); return err; do_sync: - spin_unlock(&inode->i_lock); + spin_unlock(&ci->i_ceph_lock); err = ceph_send_removexattr(dentry, name); return err; } -- cgit v1.2.3-70-g09d2 From 062c05c46bd4358aad7a0e0cb5ffeb98ab935286 Mon Sep 17 00:00:00 2001 From: Alexandre Oliva Date: Wed, 7 Dec 2011 19:50:42 -0500 Subject: Btrfs: try to allocate from cluster even at LOOP_NO_EMPTY_SIZE If we reach LOOP_NO_EMPTY_SIZE, we won't even try to use a cluster that others might have set up. Odds are that there won't be one, but if someone else succeeded in setting it up, we might as well use it, even if we don't try to set up a cluster again. Signed-off-by: Alexandre Oliva Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 26 ++++++++++++++++++-------- 1 file changed, 18 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 813c6bb96c9..db0b23b14f2 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5285,15 +5285,10 @@ alloc: spin_unlock(&block_group->free_space_ctl->tree_lock); /* - * Ok we want to try and use the cluster allocator, so lets look - * there, unless we are on LOOP_NO_EMPTY_SIZE, since we will - * have tried the cluster allocator plenty of times at this - * point and not have found anything, so we are likely way too - * fragmented for the clustering stuff to find anything, so lets - * just skip it and let the allocator find whatever block it can - * find + * Ok we want to try and use the cluster allocator, so + * lets look there */ - if (last_ptr && loop < LOOP_NO_EMPTY_SIZE) { + if (last_ptr) { /* * the refill lock keeps out other * people trying to start a new cluster @@ -5342,6 +5337,20 @@ alloc: } spin_unlock(&last_ptr->lock); refill_cluster: + /* If we are on LOOP_NO_EMPTY_SIZE, we can't + * set up a new clusters, so lets just skip it + * and let the allocator find whatever block + * it can find. If we reach this point, we + * will have tried the cluster allocator + * plenty of times and not have found + * anything, so we are likely way too + * fragmented for the clustering stuff to find + * anything. */ + if (loop >= LOOP_NO_EMPTY_SIZE) { + spin_unlock(&last_ptr->refill_lock); + goto unclustered_alloc; + } + /* * this cluster didn't work out, free it and * start over @@ -5389,6 +5398,7 @@ refill_cluster: goto loop; } +unclustered_alloc: offset = btrfs_find_space_for_alloc(block_group, search_start, num_bytes, empty_size); /* -- cgit v1.2.3-70-g09d2 From 274bd4fb3ed6b72c1d77ef8850511f09fc6b8e4d Mon Sep 17 00:00:00 2001 From: Alexandre Oliva Date: Wed, 7 Dec 2011 20:08:40 -0500 Subject: Btrfs: try cluster but don't advance in search list When we find an existing cluster, we switch to its block group as the current block group, possibly skipping multiple blocks in the process. Furthermore, under heavy contention, multiple threads may fail to allocate from a cluster and then release just-created clusters just to proceed to create new ones in a different block group. This patch tries to allocate from an existing cluster regardless of its block group, and doesn't switch to that group, instead proceeding to try to allocate a cluster from the group it was iterating before the attempt. Signed-off-by: Alexandre Oliva Signed-off-by: Chris Mason --- fs/btrfs/extent-tree.c | 74 +++++++++++++++++++++----------------------------- 1 file changed, 31 insertions(+), 43 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index db0b23b14f2..05e1386b8be 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5106,11 +5106,11 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root = orig_root->fs_info->extent_root; struct btrfs_free_cluster *last_ptr = NULL; struct btrfs_block_group_cache *block_group = NULL; + struct btrfs_block_group_cache *used_block_group; int empty_cluster = 2 * 1024 * 1024; int allowed_chunk_alloc = 0; int done_chunk_alloc = 0; struct btrfs_space_info *space_info; - int last_ptr_loop = 0; int loop = 0; int index = 0; int alloc_type = (data & BTRFS_BLOCK_GROUP_DATA) ? @@ -5172,6 +5172,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, ideal_cache: block_group = btrfs_lookup_block_group(root->fs_info, search_start); + used_block_group = block_group; /* * we don't want to use the block group if it doesn't match our * allocation bits, or if its not cached. @@ -5209,6 +5210,7 @@ search: u64 offset; int cached; + used_block_group = block_group; btrfs_get_block_group(block_group); search_start = block_group->key.objectid; @@ -5294,49 +5296,33 @@ alloc: * people trying to start a new cluster */ spin_lock(&last_ptr->refill_lock); - if (!last_ptr->block_group || - last_ptr->block_group->ro || - !block_group_bits(last_ptr->block_group, data)) + used_block_group = last_ptr->block_group; + if (used_block_group != block_group && + (!used_block_group || + used_block_group->ro || + !block_group_bits(used_block_group, data))) { + used_block_group = block_group; goto refill_cluster; + } + + if (used_block_group != block_group) + btrfs_get_block_group(used_block_group); - offset = btrfs_alloc_from_cluster(block_group, last_ptr, - num_bytes, search_start); + offset = btrfs_alloc_from_cluster(used_block_group, + last_ptr, num_bytes, used_block_group->key.objectid); if (offset) { /* we have a block, we're done */ spin_unlock(&last_ptr->refill_lock); goto checks; } - spin_lock(&last_ptr->lock); - /* - * whoops, this cluster doesn't actually point to - * this block group. Get a ref on the block - * group is does point to and try again - */ - if (!last_ptr_loop && last_ptr->block_group && - last_ptr->block_group != block_group && - index <= - get_block_group_index(last_ptr->block_group)) { - - btrfs_put_block_group(block_group); - block_group = last_ptr->block_group; - btrfs_get_block_group(block_group); - spin_unlock(&last_ptr->lock); - spin_unlock(&last_ptr->refill_lock); - - last_ptr_loop = 1; - search_start = block_group->key.objectid; - /* - * we know this block group is properly - * in the list because - * btrfs_remove_block_group, drops the - * cluster before it removes the block - * group from the list - */ - goto have_block_group; + WARN_ON(last_ptr->block_group != used_block_group); + if (used_block_group != block_group) { + btrfs_put_block_group(used_block_group); + used_block_group = block_group; } - spin_unlock(&last_ptr->lock); refill_cluster: + BUG_ON(used_block_group != block_group); /* If we are on LOOP_NO_EMPTY_SIZE, we can't * set up a new clusters, so lets just skip it * and let the allocator find whatever block @@ -5357,8 +5343,6 @@ refill_cluster: */ btrfs_return_cluster_to_free_space(NULL, last_ptr); - last_ptr_loop = 0; - /* allocate a cluster in this block group */ ret = btrfs_find_space_cluster(trans, root, block_group, last_ptr, @@ -5425,14 +5409,14 @@ checks: search_start = stripe_align(root, offset); /* move on to the next group */ if (search_start + num_bytes >= search_end) { - btrfs_add_free_space(block_group, offset, num_bytes); + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } /* move on to the next group */ if (search_start + num_bytes > - block_group->key.objectid + block_group->key.offset) { - btrfs_add_free_space(block_group, offset, num_bytes); + used_block_group->key.objectid + used_block_group->key.offset) { + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } @@ -5440,14 +5424,14 @@ checks: ins->offset = num_bytes; if (offset < search_start) - btrfs_add_free_space(block_group, offset, + btrfs_add_free_space(used_block_group, offset, search_start - offset); BUG_ON(offset > search_start); - ret = btrfs_update_reserved_bytes(block_group, num_bytes, + ret = btrfs_update_reserved_bytes(used_block_group, num_bytes, alloc_type); if (ret == -EAGAIN) { - btrfs_add_free_space(block_group, offset, num_bytes); + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } @@ -5456,15 +5440,19 @@ checks: ins->offset = num_bytes; if (offset < search_start) - btrfs_add_free_space(block_group, offset, + btrfs_add_free_space(used_block_group, offset, search_start - offset); BUG_ON(offset > search_start); + if (used_block_group != block_group) + btrfs_put_block_group(used_block_group); btrfs_put_block_group(block_group); break; loop: failed_cluster_refill = false; failed_alloc = false; BUG_ON(index != get_block_group_index(block_group)); + if (used_block_group != block_group) + btrfs_put_block_group(used_block_group); btrfs_put_block_group(block_group); } up_read(&space_info->groups_sem); -- cgit v1.2.3-70-g09d2 From a5d16333612718569ffd26064270e535cb9c3928 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 7 Dec 2011 20:08:40 -0500 Subject: Btrfs: check if the to-be-added device is writable If we call ioctl(BTRFS_IOC_ADD_DEV) directly, we'll succeed in adding a readonly device to a btrfs filesystem, and btrfs will write to that device, emitting kernel errors: [ 3109.833692] lost page write due to I/O error on loop2 [ 3109.833720] lost page write due to I/O error on loop2 ... Signed-off-by: Li Zefan Signed-off-by: Chris Mason --- fs/btrfs/volumes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index c37433d3cd8..0a8c8f8304b 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1611,7 +1611,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding) return -EINVAL; - bdev = blkdev_get_by_path(device_path, FMODE_EXCL, + bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL, root->fs_info->bdev_holder); if (IS_ERR(bdev)) return PTR_ERR(bdev); -- cgit v1.2.3-70-g09d2 From 1cf4ffdb3289624a6462c94f2ce05545b32ef736 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 7 Dec 2011 20:08:40 -0500 Subject: Btrfs: drop spin lock when memory alloc fails Drop spin lock in convert_extent_bit() when memory alloc fails, otherwise, it will be a deadlock. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/extent_io.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index be1bf627a14..49f3c9dc09f 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -935,8 +935,10 @@ again: node = tree_search(tree, start); if (!node) { prealloc = alloc_extent_state_atomic(prealloc); - if (!prealloc) - return -ENOMEM; + if (!prealloc) { + err = -ENOMEM; + goto out; + } err = insert_state(tree, prealloc, start, end, &bits); prealloc = NULL; BUG_ON(err == -EEXIST); @@ -992,8 +994,10 @@ hit_next: */ if (state->start < start) { prealloc = alloc_extent_state_atomic(prealloc); - if (!prealloc) - return -ENOMEM; + if (!prealloc) { + err = -ENOMEM; + goto out; + } err = split_state(tree, state, prealloc, start); BUG_ON(err == -EEXIST); prealloc = NULL; @@ -1024,8 +1028,10 @@ hit_next: this_end = last_start - 1; prealloc = alloc_extent_state_atomic(prealloc); - if (!prealloc) - return -ENOMEM; + if (!prealloc) { + err = -ENOMEM; + goto out; + } /* * Avoid to free 'prealloc' if it can be merged with @@ -1051,8 +1057,10 @@ hit_next: */ if (state->start <= end && state->end > end) { prealloc = alloc_extent_state_atomic(prealloc); - if (!prealloc) - return -ENOMEM; + if (!prealloc) { + err = -ENOMEM; + goto out; + } err = split_state(tree, state, prealloc, end + 1); BUG_ON(err == -EEXIST); -- cgit v1.2.3-70-g09d2 From 9a5101c8968c7b3e7de5564af6358d7332cf9a6b Mon Sep 17 00:00:00 2001 From: Pavel Shilovsky Date: Mon, 7 Nov 2011 16:11:24 +0300 Subject: CIFS: Add descriptions to the brlock cache functions Signed-off-by: Pavel Shilovsky Signed-off-by: Steve French --- fs/cifs/file.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) (limited to 'fs') diff --git a/fs/cifs/file.c b/fs/cifs/file.c index cf0b1539b32..4dd9283885e 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -702,6 +702,13 @@ cifs_find_lock_conflict(struct cifsInodeInfo *cinode, struct cifsLockInfo *lock, lock->type, lock->netfid, conf_lock); } +/* + * Check if there is another lock that prevents us to set the lock (mandatory + * style). If such a lock exists, update the flock structure with its + * properties. Otherwise, set the flock type to F_UNLCK if we can cache brlocks + * or leave it the same if we can't. Returns 0 if we don't need to request to + * the server or 1 otherwise. + */ static int cifs_lock_test(struct cifsInodeInfo *cinode, __u64 offset, __u64 length, __u8 type, __u16 netfid, struct file_lock *flock) @@ -739,6 +746,12 @@ cifs_lock_add(struct cifsInodeInfo *cinode, struct cifsLockInfo *lock) mutex_unlock(&cinode->lock_mutex); } +/* + * Set the byte-range lock (mandatory style). Returns: + * 1) 0, if we set the lock and don't need to request to the server; + * 2) 1, if no locks prevent us but we need to request to the server; + * 3) -EACCESS, if there is a lock that prevents us and wait is false. + */ static int cifs_lock_add_if(struct cifsInodeInfo *cinode, struct cifsLockInfo *lock, bool wait) @@ -778,6 +791,13 @@ try_again: return rc; } +/* + * Check if there is another lock that prevents us to set the lock (posix + * style). If such a lock exists, update the flock structure with its + * properties. Otherwise, set the flock type to F_UNLCK if we can cache brlocks + * or leave it the same if we can't. Returns 0 if we don't need to request to + * the server or 1 otherwise. + */ static int cifs_posix_lock_test(struct file *file, struct file_lock *flock) { @@ -800,6 +820,12 @@ cifs_posix_lock_test(struct file *file, struct file_lock *flock) return rc; } +/* + * Set the byte-range lock (posix style). Returns: + * 1) 0, if we set the lock and don't need to request to the server; + * 2) 1, if we need to request to the server; + * 3) <0, if the error occurs while setting the lock. + */ static int cifs_posix_lock_set(struct file *file, struct file_lock *flock) { -- cgit v1.2.3-70-g09d2 From 59edb63ad08e4f3a9477223c9aaf3fba753085bf Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 10 Nov 2011 12:48:20 -0600 Subject: cifs: Fix sparse warning when calling cifs_strtoUCS Fix sparse endian check warning while calling cifs_strtoUCS CHECK fs/cifs/smbencrypt.c fs/cifs/smbencrypt.c:216:37: warning: incorrect type in argument 1 (different base types) fs/cifs/smbencrypt.c:216:37: expected restricted __le16 [usertype] * fs/cifs/smbencrypt.c:216:37: got unsigned short * Signed-off-by: Steve French Acked-by: Shirish Pargaonkar Date: Thu, 1 Dec 2011 20:22:41 -0500 Subject: cifs: attempt to freeze while looping on a receive attempt In the recent overhaul of the demultiplex thread receive path, I neglected to ensure that we attempt to freeze on each pass through the receive loop. Reported-and-Tested-by: Woody Suwalski Reported-and-Tested-by: Adam Williamson Signed-off-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/connect.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index d6a972df033..8cd4b52d421 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -441,6 +441,8 @@ cifs_readv_from_socket(struct TCP_Server_Info *server, struct kvec *iov_orig, smb_msg.msg_controllen = 0; for (total_read = 0; to_read; total_read += length, to_read -= length) { + try_to_freeze(); + if (server_unresponsive(server)) { total_read = -EAGAIN; break; -- cgit v1.2.3-70-g09d2 From 7023676f9ee851d94f0942e879243fc1f9081c47 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 1 Dec 2011 20:23:34 -0500 Subject: cifs: check for NULL last_entry before calling cifs_save_resume_key Prior to commit eaf35b1, cifs_save_resume_key had some NULL pointer checks at the top. It turns out that at least one of those NULL pointer checks is needed after all. When the LastNameOffset in a FIND reply appears to be beyond the end of the buffer, CIFSFindFirst and CIFSFindNext will set srch_inf.last_entry to NULL. Since eaf35b1, the code will now oops in this situation. Fix this by having the callers check for a NULL last entry pointer before calling cifs_save_resume_key. No change is needed for the call site in cifs_readdir as it's not reachable with a NULL current_entry pointer. This should fix: https://bugzilla.redhat.com/show_bug.cgi?id=750247 Cc: stable@vger.kernel.org Cc: Christoph Hellwig Reported-by: Adam G. Metzler Signed-off-by: Jeff Layton Signed-off-by: Steve French --- fs/cifs/readdir.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c index 5de03ec2014..a090bbe6ee2 100644 --- a/fs/cifs/readdir.c +++ b/fs/cifs/readdir.c @@ -554,7 +554,10 @@ static int find_cifs_entry(const int xid, struct cifs_tcon *pTcon, rc); return rc; } - cifs_save_resume_key(cifsFile->srch_inf.last_entry, cifsFile); + /* FindFirst/Next set last_entry to NULL on malformed reply */ + if (cifsFile->srch_inf.last_entry) + cifs_save_resume_key(cifsFile->srch_inf.last_entry, + cifsFile); } while ((index_to_find >= cifsFile->srch_inf.index_of_last_entry) && @@ -562,7 +565,10 @@ static int find_cifs_entry(const int xid, struct cifs_tcon *pTcon, cFYI(1, "calling findnext2"); rc = CIFSFindNext(xid, pTcon, cifsFile->netfid, &cifsFile->srch_inf); - cifs_save_resume_key(cifsFile->srch_inf.last_entry, cifsFile); + /* FindFirst/Next set last_entry to NULL on malformed reply */ + if (cifsFile->srch_inf.last_entry) + cifs_save_resume_key(cifsFile->srch_inf.last_entry, + cifsFile); if (rc) return -ENOENT; } -- cgit v1.2.3-70-g09d2 From 905ad269c55fc62bee3da29f7b1d1efeba8aa1e1 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 8 Dec 2011 23:20:45 -0500 Subject: procfs: fix a vfsmount longterm reference leak kern_mount() doesn't pair with plain mntput()... Signed-off-by: Al Viro --- fs/proc/root.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/proc/root.c b/fs/proc/root.c index 9a8a2b77b87..03102d97818 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -91,20 +91,18 @@ static struct file_system_type proc_fs_type = { void __init proc_root_init(void) { - struct vfsmount *mnt; int err; proc_init_inodecache(); err = register_filesystem(&proc_fs_type); if (err) return; - mnt = kern_mount_data(&proc_fs_type, &init_pid_ns); - if (IS_ERR(mnt)) { + err = pid_ns_prepare_proc(&init_pid_ns); + if (err) { unregister_filesystem(&proc_fs_type); return; } - init_pid_ns.proc_mnt = mnt; proc_symlink("mounts", NULL, "self/mounts"); proc_net_init(); @@ -209,5 +207,5 @@ int pid_ns_prepare_proc(struct pid_namespace *ns) void pid_ns_release_proc(struct pid_namespace *ns) { - mntput(ns->proc_mnt); + kern_unmount(ns->proc_mnt); } -- cgit v1.2.3-70-g09d2 From b53fc7c2974a50913f49e1d800fe904a28c338e3 Mon Sep 17 00:00:00 2001 From: Claudio Scordino Date: Thu, 8 Dec 2011 14:33:56 -0800 Subject: fs/proc/meminfo.c: fix compilation error Fix the error message "directives may not be used inside a macro argument" which appears when the kernel is compiled for the cris architecture. Signed-off-by: Claudio Scordino Cc: Andrea Arcangeli Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/meminfo.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c index 586174168e2..80e4645f799 100644 --- a/fs/proc/meminfo.c +++ b/fs/proc/meminfo.c @@ -131,12 +131,13 @@ static int meminfo_proc_show(struct seq_file *m, void *v) K(i.freeswap), K(global_page_state(NR_FILE_DIRTY)), K(global_page_state(NR_WRITEBACK)), - K(global_page_state(NR_ANON_PAGES) #ifdef CONFIG_TRANSPARENT_HUGEPAGE + K(global_page_state(NR_ANON_PAGES) + global_page_state(NR_ANON_TRANSPARENT_HUGEPAGES) * - HPAGE_PMD_NR + HPAGE_PMD_NR), +#else + K(global_page_state(NR_ANON_PAGES)), #endif - ), K(global_page_state(NR_FILE_MAPPED)), K(global_page_state(NR_SHMEM)), K(global_page_state(NR_SLAB_RECLAIMABLE) + -- cgit v1.2.3-70-g09d2 From 2a95ea6c0d129b4568fb64e1deda16ceb20e6636 Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Thu, 8 Dec 2011 14:34:32 -0800 Subject: procfs: do not overflow get_{idle,iowait}_time for nohz Since commit a25cac5198d4 ("proc: Consider NO_HZ when printing idle and iowait times") we are reporting idle/io_wait time also while a CPU is tickless. We rely on get_{idle,iowait}_time functions to retrieve proper data. These functions, however, use usecs_to_cputime to translate micro seconds time to cputime64_t. This is just an alias to usecs_to_jiffies which reduces the data type from u64 to unsigned int and also checks whether the given parameter overflows jiffies_to_usecs(MAX_JIFFY_OFFSET) and returns MAX_JIFFY_OFFSET in that case. When we overflow depends on CONFIG_HZ but especially for CONFIG_HZ_300 it is quite low (1431649781) so we are getting MAX_JIFFY_OFFSET for >3000s! until we overflow unsigned int. Just for reference CONFIG_HZ_100 has an overflow window around 20s, CONFIG_HZ_250 ~8s and CONFIG_HZ_1000 ~2s. This results in a bug when people saw [h]top going mad reporting 100% CPU usage even though there was basically no CPU load. The reason was simply that /proc/stat stopped reporting idle/io_wait changes (and reported MAX_JIFFY_OFFSET) and so the only change happening was for user system time. Let's use nsecs_to_jiffies64 instead which doesn't reduce the precision to 32b type and it is much more appropriate for cumulative time values (unlike usecs_to_jiffies which intended for timeout calculations). Signed-off-by: Michal Hocko Tested-by: Artem S. Tashkinov Cc: Dave Jones Cc: Arnd Bergmann Cc: Alexey Dobriyan Cc: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/stat.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/proc/stat.c b/fs/proc/stat.c index 42b274da92c..2a30d67dd6b 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -32,7 +32,7 @@ static cputime64_t get_idle_time(int cpu) idle = kstat_cpu(cpu).cpustat.idle; idle = cputime64_add(idle, arch_idle_time(cpu)); } else - idle = usecs_to_cputime(idle_time); + idle = nsecs_to_jiffies64(1000 * idle_time); return idle; } @@ -46,7 +46,7 @@ static cputime64_t get_iowait_time(int cpu) /* !NO_HZ so we can rely on cpustat.iowait */ iowait = kstat_cpu(cpu).cpustat.iowait; else - iowait = usecs_to_cputime(iowait_time); + iowait = nsecs_to_jiffies64(1000 * iowait_time); return iowait; } -- cgit v1.2.3-70-g09d2 From 5dbc8fca8ef5d719014f22345d990e957dcfc692 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Fri, 9 Dec 2011 11:07:37 -0500 Subject: Btrfs: fix btrfs_end_bio to deal with write errors to a single mirror btrfs_end_bio checks the number of errors on a bio against the max number of errors allowed before sending any EIOs up to the higher levels. If we got enough copies of the bio done for a given raid level, it is supposed to clear the bio error flag and return success. We have pointers to the original bio sent down by the higher layers and pointers to any cloned bios we made for raid purposes. If the original bio happens to be the one that got an io error, but not the last one to finish, it might not have the BIO_UPTODATE bit set. Then, when the last bio does finish, we'll call bio_end_io on the original bio. It won't have the uptodate bit set and we'll end up sending EIO to the higher layers. We already had a check for this, it just was conditional on getting the IO error on the very last bio. Make the check unconditional so we eat the EIOs properly. Signed-off-by: Chris Mason --- fs/btrfs/volumes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 0a8c8f8304b..91ea57a1474 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3258,7 +3258,7 @@ static void btrfs_end_bio(struct bio *bio, int err) */ if (atomic_read(&bbio->error) > bbio->max_errors) { err = -EIO; - } else if (err) { + } else { /* * this bio is actually up to date, we didn't * go over the max number of errors -- cgit v1.2.3-70-g09d2 From 4b44b40e04a758e2242ff4a3f7c15982801ec8bc Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 9 Dec 2011 16:31:52 -0500 Subject: NFSv4: Ensure correct locking when accessing the 'lock_states' list There are currently 2 places in the state recovery code, where we do not take sufficient precautions before accessing the state->lock_states. In both cases, we should be holding the state->state_lock. Reported-by: Pascal Bouchareine Signed-off-by: Trond Myklebust --- fs/nfs/nfs4state.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index da8d73ed9e0..6a7107ae6b7 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1156,11 +1156,13 @@ restart: if (status >= 0) { status = nfs4_reclaim_locks(state, ops); if (status >= 0) { + spin_lock(&state->state_lock); list_for_each_entry(lock, &state->lock_states, ls_locks) { if (!(lock->ls_flags & NFS_LOCK_INITIALIZED)) printk("%s: Lock reclaim failed!\n", __func__); } + spin_unlock(&state->state_lock); nfs4_put_open_state(state); goto restart; } @@ -1224,10 +1226,12 @@ static void nfs4_clear_open_state(struct nfs4_state *state) clear_bit(NFS_O_RDONLY_STATE, &state->flags); clear_bit(NFS_O_WRONLY_STATE, &state->flags); clear_bit(NFS_O_RDWR_STATE, &state->flags); + spin_lock(&state->state_lock); list_for_each_entry(lock, &state->lock_states, ls_locks) { lock->ls_seqid.flags = 0; lock->ls_flags &= ~NFS_LOCK_INITIALIZED; } + spin_unlock(&state->state_lock); } static void nfs4_reset_seqids(struct nfs_server *server, -- cgit v1.2.3-70-g09d2 From 652f89f64fabcdae9143ee2b4253cfa838fb0279 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 9 Dec 2011 19:05:58 -0500 Subject: NFSv4: Do not accept delegated opens when a delegation recall is in effect ...and report the servers that try to return a delegation when the client is using the CLAIM_DELEG_CUR open mode. That behaviour is explicitly forbidden in RFC3530. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index be2bbac1381..d9f4d78c341 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -39,6 +39,8 @@ #include #include #include +#include +#include #include #include #include @@ -894,6 +896,8 @@ out: static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode) { + if (delegation == NULL) + return 0; if ((delegation->type & fmode) != fmode) return 0; if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags)) @@ -1036,8 +1040,7 @@ static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata) } rcu_read_lock(); delegation = rcu_dereference(nfsi->delegation); - if (delegation == NULL || - !can_open_delegated(delegation, fmode)) { + if (!can_open_delegated(delegation, fmode)) { rcu_read_unlock(); break; } @@ -1091,7 +1094,12 @@ static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data if (delegation) delegation_flags = delegation->flags; rcu_read_unlock(); - if ((delegation_flags & 1UL<o_arg.claim == NFS4_OPEN_CLAIM_DELEGATE_CUR) { + pr_err_ratelimited("NFS: Broken NFSv4 server %s is " + "returning a delegation for " + "OPEN(CLAIM_DELEGATE_CUR)\n", + NFS_CLIENT(inode)->cl_server); + } else if ((delegation_flags & 1UL<inode, data->owner->so_cred, &data->o_res); @@ -1423,11 +1431,9 @@ static void nfs4_open_prepare(struct rpc_task *task, void *calldata) goto out_no_action; rcu_read_lock(); delegation = rcu_dereference(NFS_I(data->state->inode)->delegation); - if (delegation != NULL && - test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) == 0) { - rcu_read_unlock(); - goto out_no_action; - } + if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR && + can_open_delegated(delegation, data->o_arg.fmode)) + goto unlock_no_action; rcu_read_unlock(); } /* Update sequence id. */ @@ -1444,6 +1450,8 @@ static void nfs4_open_prepare(struct rpc_task *task, void *calldata) return; rpc_call_start(task); return; +unlock_no_action: + rcu_read_unlock(); out_no_action: task->tk_action = NULL; -- cgit v1.2.3-70-g09d2 From b5a7e97039a80fae673ccc115ce595d5b88fb4ee Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 12 Dec 2011 10:53:02 -0500 Subject: ext4: fix ext4_end_io_dio() racing against fsync() We need to make sure iocb->private is cleared *before* we put the io_end structure on i_completed_io_list. Otherwise fsync() could potentially run on another CPU and free the iocb structure out from under us. Reported-by: Kent Overstreet Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 848f436df29..d2c86633811 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2781,10 +2781,11 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, iocb->private, io_end->inode->i_ino, iocb, offset, size); + iocb->private = NULL; + /* if not aio dio with unwritten extents, just free io and return */ if (!(io_end->flag & EXT4_IO_END_UNWRITTEN)) { ext4_free_io_end(io_end); - iocb->private = NULL; out: if (is_async) aio_complete(iocb, ret, 0); @@ -2807,7 +2808,6 @@ out: spin_unlock_irqrestore(&ei->i_completed_io_lock, flags); /* queue the work to convert unwritten extents to written */ - iocb->private = NULL; queue_work(wq, &io_end->work); /* XXX: probably should move into the real I/O completion handler */ -- cgit v1.2.3-70-g09d2 From b4611abfa98af8351bf32e8b9ecc9d3384931c37 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Mon, 12 Dec 2011 11:00:56 -0500 Subject: ext4: Fix crash due to getting bogus eh_depth value on big-endian systems Commit 1939dd84b3 ("ext4: cleanup ext4_ext_grow_indepth code") added a reference to ext4_extent_header.eh_depth, but forget to pass the value read through le16_to_cpu. The result is a crash on big-endian machines, such as this crash on a POWER7 server: attempt to access beyond end of device sda8: rw=0, want=776392648163376, limit=168558560 Unable to handle kernel paging request for data at address 0x6b6b6b6b6b6b6bcb Faulting instruction address: 0xc0000000001f5f38 cpu 0x14: Vector: 300 (Data Access) at [c000001bd1aaecf0] pc: c0000000001f5f38: .__brelse+0x18/0x60 lr: c0000000002e07a4: .ext4_ext_drop_refs+0x44/0x80 sp: c000001bd1aaef70 msr: 9000000000009032 dar: 6b6b6b6b6b6b6bcb dsisr: 40000000 current = 0xc000001bd15b8010 paca = 0xc00000000ffe4600 pid = 19911, comm = flush-8:0 enter ? for help [c000001bd1aaeff0] c0000000002e07a4 .ext4_ext_drop_refs+0x44/0x80 [c000001bd1aaf090] c0000000002e0c58 .ext4_ext_find_extent+0x408/0x4c0 [c000001bd1aaf180] c0000000002e145c .ext4_ext_insert_extent+0x2bc/0x14c0 [c000001bd1aaf2c0] c0000000002e3fb8 .ext4_ext_map_blocks+0x628/0x1710 [c000001bd1aaf420] c0000000002b2974 .ext4_map_blocks+0x224/0x310 [c000001bd1aaf4d0] c0000000002b7f2c .mpage_da_map_and_submit+0xbc/0x490 [c000001bd1aaf5a0] c0000000002b8688 .write_cache_pages_da+0x2c8/0x430 [c000001bd1aaf720] c0000000002b8b28 .ext4_da_writepages+0x338/0x670 [c000001bd1aaf8d0] c000000000157280 .do_writepages+0x40/0x90 [c000001bd1aaf940] c0000000001ea830 .writeback_single_inode+0xe0/0x530 [c000001bd1aafa00] c0000000001eb680 .writeback_sb_inodes+0x210/0x300 [c000001bd1aafb20] c0000000001ebc84 .__writeback_inodes_wb+0xd4/0x140 [c000001bd1aafbe0] c0000000001ebfec .wb_writeback+0x2fc/0x3e0 [c000001bd1aafce0] c0000000001ed770 .wb_do_writeback+0x2f0/0x300 [c000001bd1aafdf0] c0000000001ed848 .bdi_writeback_thread+0xc8/0x340 [c000001bd1aafed0] c0000000000c5494 .kthread+0xb4/0xc0 [c000001bd1aaff90] c000000000021f48 .kernel_thread+0x54/0x70 This is due to getting ext_depth(inode) == 0x101 and therefore running off the end of the path array in ext4_ext_drop_refs into following unallocated structures. This fixes it by adding the necessary le16_to_cpu. Signed-off-by: Paul Mackerras Signed-off-by: "Theodore Ts'o" --- fs/ext4/extents.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 61fa9e1614a..61cdb97f338 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -1095,7 +1095,7 @@ static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode, le32_to_cpu(EXT_FIRST_INDEX(neh)->ei_block), ext4_idx_pblock(EXT_FIRST_INDEX(neh))); - neh->eh_depth = cpu_to_le16(neh->eh_depth + 1); + neh->eh_depth = cpu_to_le16(le16_to_cpu(neh->eh_depth) + 1); ext4_mark_inode_dirty(handle, inode); out: brelse(bh); -- cgit v1.2.3-70-g09d2 From fc6cb1cda5db7b2d24bf32890826214b857c728e Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 12 Dec 2011 22:06:18 -0500 Subject: ext4: display the correct mount option in /proc/mounts for [no]init_itable /proc/mounts was showing the mount option [no]init_inode_table when the correct mount option that will be accepted by parse_options() is [no]init_itable. Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/super.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 3858767ec67..3e1329e2f82 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1155,9 +1155,9 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",block_validity"); if (!test_opt(sb, INIT_INODE_TABLE)) - seq_puts(seq, ",noinit_inode_table"); + seq_puts(seq, ",noinit_itable"); else if (sbi->s_li_wait_mult != EXT4_DEF_LI_WAIT_MULT) - seq_printf(seq, ",init_inode_table=%u", + seq_printf(seq, ",init_itable=%u", (unsigned) sbi->s_li_wait_mult); ext4_show_quota_options(seq, sb); @@ -1333,8 +1333,7 @@ enum { Opt_nomblk_io_submit, Opt_block_validity, Opt_noblock_validity, Opt_inode_readahead_blks, Opt_journal_ioprio, Opt_dioread_nolock, Opt_dioread_lock, - Opt_discard, Opt_nodiscard, - Opt_init_inode_table, Opt_noinit_inode_table, + Opt_discard, Opt_nodiscard, Opt_init_itable, Opt_noinit_itable, }; static const match_table_t tokens = { @@ -1407,9 +1406,9 @@ static const match_table_t tokens = { {Opt_dioread_lock, "dioread_lock"}, {Opt_discard, "discard"}, {Opt_nodiscard, "nodiscard"}, - {Opt_init_inode_table, "init_itable=%u"}, - {Opt_init_inode_table, "init_itable"}, - {Opt_noinit_inode_table, "noinit_itable"}, + {Opt_init_itable, "init_itable=%u"}, + {Opt_init_itable, "init_itable"}, + {Opt_noinit_itable, "noinit_itable"}, {Opt_err, NULL}, }; @@ -1892,7 +1891,7 @@ set_qf_format: case Opt_dioread_lock: clear_opt(sb, DIOREAD_NOLOCK); break; - case Opt_init_inode_table: + case Opt_init_itable: set_opt(sb, INIT_INODE_TABLE); if (args[0].from) { if (match_int(&args[0], &option)) @@ -1903,7 +1902,7 @@ set_qf_format: return 0; sbi->s_li_wait_mult = option; break; - case Opt_noinit_inode_table: + case Opt_noinit_itable: clear_opt(sb, INIT_INODE_TABLE); break; default: -- cgit v1.2.3-70-g09d2 From 48706d0a91583d08c56e7ef2a7602d99c8d4133f Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 13 Dec 2011 10:36:59 +0100 Subject: fuse: fix fuse_retrieve Fix two bugs in fuse_retrieve(): - retrieving more than one page would yield repeated instances of the first page - if more than FUSE_MAX_PAGES_PER_REQ pages were requested than the request page array would overflow fuse_retrieve() was added in 2.6.36 and these bugs had been there since the beginning. Signed-off-by: Miklos Szeredi CC: stable@vger.kernel.org --- fs/fuse/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 5cb8614508c..2aaf3eaaf13 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -1512,7 +1512,7 @@ static int fuse_retrieve(struct fuse_conn *fc, struct inode *inode, else if (outarg->offset + num > file_size) num = file_size - outarg->offset; - while (num) { + while (num && req->num_pages < FUSE_MAX_PAGES_PER_REQ) { struct page *page; unsigned int this_num; @@ -1526,6 +1526,7 @@ static int fuse_retrieve(struct fuse_conn *fc, struct inode *inode, num -= this_num; total_len += this_num; + index++; } req->misc.retrieve_in.offset = outarg->offset; req->misc.retrieve_in.size = total_len; -- cgit v1.2.3-70-g09d2 From b48c6af2086ab2ba8a9c9b6ce9ecb34592ce500c Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Tue, 13 Dec 2011 10:37:00 +0100 Subject: fuse: fix llseek bug The test in fuse_file_llseek() "not SEEK_CUR or not SEEK_SET" always evaluates to true. This was introduced in 3.1 by commit 06222e49 (fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek) and changed the behavior of SEEK_CUR and SEEK_SET to always retrieve the file attributes. This is a performance regression. Fix the test so that it makes sense. Signed-off-by: Miklos Szeredi CC: stable@vger.kernel.org CC: Josef Bacik CC: Al Viro --- fs/fuse/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 594f07a81c2..19029e97ca2 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1556,7 +1556,7 @@ static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin) struct inode *inode = file->f_path.dentry->d_inode; mutex_lock(&inode->i_mutex); - if (origin != SEEK_CUR || origin != SEEK_SET) { + if (origin != SEEK_CUR && origin != SEEK_SET) { retval = fuse_update_attributes(inode, NULL, file, NULL); if (retval) goto exit; -- cgit v1.2.3-70-g09d2 From 73104b6e37f75f822e3dcf016147c322d1f6dad9 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 13 Dec 2011 11:40:59 +0100 Subject: fuse: llseek fix race Fix race between lseek(fd, 0, SEEK_CUR) and read/write. This was fixed in generic code by commit 5b6f1eb97d (vfs: lseek(fd, 0, SEEK_CUR) race condition). Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 19029e97ca2..0c84100acd4 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1567,6 +1567,10 @@ static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin) offset += i_size_read(inode); break; case SEEK_CUR: + if (offset == 0) { + retval = file->f_pos; + goto exit; + } offset += file->f_pos; break; case SEEK_DATA: -- cgit v1.2.3-70-g09d2 From 6a82c47aa84ab22cb5969a44105cca5358879d21 Mon Sep 17 00:00:00 2001 From: Sage Weil Date: Tue, 13 Dec 2011 09:19:26 -0800 Subject: ceph: fix SEEK_CUR, SEEK_SET regression Commit 06222e491e663dac939f04b125c9dc52126a75c4 got the if wrong so that it always evaluates as true. This is semantically harmless, but makes SEEK_CUR and SEEK_SET needlessly query the server. Rewrite the if to explicitly enumerate the cases we DO need a valid i_size to make this code less fragile. Reported-by: Roel Kluin Signed-off-by: Sage Weil --- fs/ceph/file.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/file.c b/fs/ceph/file.c index af187330cc8..ed72428d9c7 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -797,7 +797,8 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int origin) mutex_lock(&inode->i_mutex); __ceph_do_pending_vmtruncate(inode); - if (origin != SEEK_CUR || origin != SEEK_SET) { + + if (origin == SEEK_END || origin == SEEK_DATA || origin == SEEK_HOLE) { ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE); if (ret < 0) { offset = ret; -- cgit v1.2.3-70-g09d2 From 5cc361e3b8241989d8240014dbbe6214c9292541 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 12 Dec 2011 23:58:21 -0500 Subject: ubifs: too early register_filesystem() doing that before you are ready to handle mount() is a Bad Idea(tm)... Signed-off-by: Al Viro --- fs/ubifs/super.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 20403dc5d43..ae0e76bb6eb 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2264,19 +2264,12 @@ static int __init ubifs_init(void) return -EINVAL; } - err = register_filesystem(&ubifs_fs_type); - if (err) { - ubifs_err("cannot register file system, error %d", err); - return err; - } - - err = -ENOMEM; ubifs_inode_slab = kmem_cache_create("ubifs_inode_slab", sizeof(struct ubifs_inode), 0, SLAB_MEM_SPREAD | SLAB_RECLAIM_ACCOUNT, &inode_slab_ctor); if (!ubifs_inode_slab) - goto out_reg; + return -ENOMEM; register_shrinker(&ubifs_shrinker_info); @@ -2288,15 +2281,20 @@ static int __init ubifs_init(void) if (err) goto out_compr; + err = register_filesystem(&ubifs_fs_type); + if (err) { + ubifs_err("cannot register file system, error %d", err); + goto out_dbg; + } return 0; +out_dbg: + dbg_debugfs_exit(); out_compr: ubifs_compressors_exit(); out_shrinker: unregister_shrinker(&ubifs_shrinker_info); kmem_cache_destroy(ubifs_inode_slab); -out_reg: - unregister_filesystem(&ubifs_fs_type); return err; } /* late_initcall to let compressors initialize first */ -- cgit v1.2.3-70-g09d2 From 988f032567eaba3c5896c5de05270b9ff71b2a9d Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 13 Dec 2011 12:25:27 -0500 Subject: fuse: register_filesystem() called too early same story as with ubifs Signed-off-by: Al Viro --- fs/fuse/inode.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 3e6d7275647..aa83109b943 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -1138,28 +1138,28 @@ static int __init fuse_fs_init(void) { int err; - err = register_filesystem(&fuse_fs_type); - if (err) - goto out; - - err = register_fuseblk(); - if (err) - goto out_unreg; - fuse_inode_cachep = kmem_cache_create("fuse_inode", sizeof(struct fuse_inode), 0, SLAB_HWCACHE_ALIGN, fuse_inode_init_once); err = -ENOMEM; if (!fuse_inode_cachep) - goto out_unreg2; + goto out; + + err = register_fuseblk(); + if (err) + goto out2; + + err = register_filesystem(&fuse_fs_type); + if (err) + goto out3; return 0; - out_unreg2: + out3: unregister_fuseblk(); - out_unreg: - unregister_filesystem(&fuse_fs_type); + out2: + kmem_cache_destroy(fuse_inode_cachep); out: return err; } -- cgit v1.2.3-70-g09d2 From 7c6455e368bc87ef38df7b6ddba2339a67107bdf Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 13 Dec 2011 12:32:42 -0500 Subject: configfs: register_filesystem() called too early Signed-off-by: Al Viro --- fs/configfs/inode.c | 2 +- fs/configfs/mount.c | 36 +++++++++++++++++------------------- 2 files changed, 18 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/configfs/inode.c b/fs/configfs/inode.c index ca418aaf635..9d8715c45f2 100644 --- a/fs/configfs/inode.c +++ b/fs/configfs/inode.c @@ -292,7 +292,7 @@ int __init configfs_inode_init(void) return bdi_init(&configfs_backing_dev_info); } -void __exit configfs_inode_exit(void) +void configfs_inode_exit(void) { bdi_destroy(&configfs_backing_dev_info); } diff --git a/fs/configfs/mount.c b/fs/configfs/mount.c index ecc62178bed..276e15cafd5 100644 --- a/fs/configfs/mount.c +++ b/fs/configfs/mount.c @@ -143,28 +143,26 @@ static int __init configfs_init(void) goto out; config_kobj = kobject_create_and_add("config", kernel_kobj); - if (!config_kobj) { - kmem_cache_destroy(configfs_dir_cachep); - configfs_dir_cachep = NULL; - goto out; - } + if (!config_kobj) + goto out2; + + err = configfs_inode_init(); + if (err) + goto out3; err = register_filesystem(&configfs_fs_type); - if (err) { - printk(KERN_ERR "configfs: Unable to register filesystem!\n"); - kobject_put(config_kobj); - kmem_cache_destroy(configfs_dir_cachep); - configfs_dir_cachep = NULL; - goto out; - } + if (err) + goto out4; - err = configfs_inode_init(); - if (err) { - unregister_filesystem(&configfs_fs_type); - kobject_put(config_kobj); - kmem_cache_destroy(configfs_dir_cachep); - configfs_dir_cachep = NULL; - } + return 0; +out4: + printk(KERN_ERR "configfs: Unable to register filesystem!\n"); + configfs_inode_exit(); +out3: + kobject_put(config_kobj); +out2: + kmem_cache_destroy(configfs_dir_cachep); + configfs_dir_cachep = NULL; out: return err; } -- cgit v1.2.3-70-g09d2 From 9d5a09e659f8414dd3713e2acbfaf8a9e9794aa1 Mon Sep 17 00:00:00 2001 From: Yehuda Sadeh Date: Tue, 13 Dec 2011 09:57:44 -0800 Subject: ceph: add missing spin_unlock at ceph_mdsc_build_path() one of the paths was missing spin_unlock Signed-off-by: Yehuda Sadeh --- fs/ceph/mds_client.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 34bc35084ae..6203d805eb4 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1495,6 +1495,7 @@ retry: pos, temp); } else if (stop_on_nosnap && inode && ceph_snap(inode) == CEPH_NOSNAP) { + spin_unlock(&temp->d_lock); break; } else { pos -= temp->d_name.len; -- cgit v1.2.3-70-g09d2 From ea51d132dbf9b00063169c1159bee253d9649224 Mon Sep 17 00:00:00 2001 From: Andrea Arcangeli Date: Tue, 13 Dec 2011 21:41:15 -0500 Subject: ext4: avoid hangs in ext4_da_should_update_i_disksize() If the pte mapping in generic_perform_write() is unmapped between iov_iter_fault_in_readable() and iov_iter_copy_from_user_atomic(), the "copied" parameter to ->end_write can be zero. ext4 couldn't cope with it with delayed allocations enabled. This skips the i_disksize enlargement logic if copied is zero and no new data was appeneded to the inode. gdb> bt #0 0xffffffff811afe80 in ext4_da_should_update_i_disksize (file=0xffff88003f606a80, mapping=0xffff88001d3824e0, pos=0x1\ 08000, len=0x1000, copied=0x0, page=0xffffea0000d792e8, fsdata=0x0) at fs/ext4/inode.c:2467 #1 ext4_da_write_end (file=0xffff88003f606a80, mapping=0xffff88001d3824e0, pos=0x108000, len=0x1000, copied=0x0, page=0\ xffffea0000d792e8, fsdata=0x0) at fs/ext4/inode.c:2512 #2 0xffffffff810d97f1 in generic_perform_write (iocb=, iov=, nr_segs=, pos=0x108000, ppos=0xffff88001e26be40, count=, written=0x0) at mm/filemap.c:2440 #3 generic_file_buffered_write (iocb=, iov=, nr_segs=, p\ os=0x108000, ppos=0xffff88001e26be40, count=, written=0x0) at mm/filemap.c:2482 #4 0xffffffff810db5d1 in __generic_file_aio_write (iocb=0xffff88001e26bde8, iov=0xffff88001e26bec8, nr_segs=0x1, ppos=0\ xffff88001e26be40) at mm/filemap.c:2600 #5 0xffffffff810db853 in generic_file_aio_write (iocb=0xffff88001e26bde8, iov=0xffff88001e26bec8, nr_segs=, pos=) at mm/filemap.c:2632 #6 0xffffffff811a71aa in ext4_file_write (iocb=0xffff88001e26bde8, iov=0xffff88001e26bec8, nr_segs=0x1, pos=0x108000) a\ t fs/ext4/file.c:136 #7 0xffffffff811375aa in do_sync_write (filp=0xffff88003f606a80, buf=, len=, \ ppos=0xffff88001e26bf48) at fs/read_write.c:406 #8 0xffffffff81137e56 in vfs_write (file=0xffff88003f606a80, buf=0x1ec2960
, count=0x4\ 000, pos=0xffff88001e26bf48) at fs/read_write.c:435 #9 0xffffffff8113816c in sys_write (fd=, buf=0x1ec2960
, count=0x\ 4000) at fs/read_write.c:487 #10 #11 0x00007f120077a390 in __brk_reservation_fn_dmi_alloc__ () #12 0x0000000000000000 in ?? () gdb> print offset $22 = 0xffffffffffffffff gdb> print idx $23 = 0xffffffff gdb> print inode->i_blkbits $24 = 0xc gdb> up #1 ext4_da_write_end (file=0xffff88003f606a80, mapping=0xffff88001d3824e0, pos=0x108000, len=0x1000, copied=0x0, page=0\ xffffea0000d792e8, fsdata=0x0) at fs/ext4/inode.c:2512 2512 if (ext4_da_should_update_i_disksize(page, end)) { gdb> print start $25 = 0x0 gdb> print end $26 = 0xffffffffffffffff gdb> print pos $27 = 0x108000 gdb> print new_i_size $28 = 0x108000 gdb> print ((struct ext4_inode_info *)((char *)inode-((int)(&((struct ext4_inode_info *)0)->vfs_inode))))->i_disksize $29 = 0xd9000 gdb> down 2467 for (i = 0; i < idx; i++) gdb> print i $30 = 0xd44acbee This is 100% reproducible with some autonuma development code tuned in a very aggressive manner (not normal way even for knumad) which does "exotic" changes to the ptes. It wouldn't normally trigger but I don't see why it can't happen normally if the page is added to swap cache in between the two faults leading to "copied" being zero (which then hangs in ext4). So it should be fixed. Especially possible with lumpy reclaim (albeit disabled if compaction is enabled) as that would ignore the young bits in the ptes. Signed-off-by: Andrea Arcangeli Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index d2c86633811..b5edc6e4bab 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2508,7 +2508,7 @@ static int ext4_da_write_end(struct file *file, */ new_i_size = pos + copied; - if (new_i_size > EXT4_I(inode)->i_disksize) { + if (copied && new_i_size > EXT4_I(inode)->i_disksize) { if (ext4_da_should_update_i_disksize(page, end)) { down_write(&EXT4_I(inode)->i_data_sem); if (new_i_size > EXT4_I(inode)->i_disksize) { -- cgit v1.2.3-70-g09d2 From 13a79a4741d37fda2fbafb953f0f301dc007928f Mon Sep 17 00:00:00 2001 From: Yongqiang Yang Date: Tue, 13 Dec 2011 21:51:55 -0500 Subject: ext4: avoid potential hang in mpage_submit_io() when blocksize < pagesize If there is an unwritten but clean buffer in a page and there is a dirty buffer after the buffer, then mpage_submit_io does not write the dirty buffer out. As a result, da_writepages loops forever. This patch fixes the problem by checking dirty flag. Signed-off-by: Yongqiang Yang Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/inode.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index b5edc6e4bab..cf20e1fe778 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1339,8 +1339,11 @@ static int mpage_da_submit_io(struct mpage_da_data *mpd, clear_buffer_unwritten(bh); } - /* skip page if block allocation undone */ - if (buffer_delay(bh) || buffer_unwritten(bh)) + /* + * skip page if block allocation undone and + * block is dirty + */ + if (ext4_bh_delay_or_unwritten(NULL, bh)) skip_page = 1; bh = bh->b_this_page; block_start += bh->b_size; -- cgit v1.2.3-70-g09d2 From 093e6e3666f47d29763a235b404c84ee47ba8bb0 Mon Sep 17 00:00:00 2001 From: Yongqiang Yang Date: Tue, 13 Dec 2011 22:05:05 -0500 Subject: ext4: correctly handle pages w/o buffers in ext4_discard_partial_buffers() If a page has been read into memory and never been written, it has no buffers, but we should handle the page in truncate or punch hole. VFS code of writing operations has handled holes correctly, so this patch removes the code handling holes in writing operations. Signed-off-by: Yongqiang Yang Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/inode.c | 41 ++--------------------------------------- 1 file changed, 2 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index cf20e1fe778..92655fd8965 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2390,7 +2390,6 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping, pgoff_t index; struct inode *inode = mapping->host; handle_t *handle; - loff_t page_len; index = pos >> PAGE_CACHE_SHIFT; @@ -2437,13 +2436,6 @@ retry: */ if (pos + len > inode->i_size) ext4_truncate_failed_write(inode); - } else { - page_len = pos & (PAGE_CACHE_SIZE - 1); - if (page_len > 0) { - ret = ext4_discard_partial_page_buffers_no_lock(handle, - inode, page, pos - page_len, page_len, - EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED); - } } if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries)) @@ -2486,7 +2478,6 @@ static int ext4_da_write_end(struct file *file, loff_t new_i_size; unsigned long start, end; int write_mode = (int)(unsigned long)fsdata; - loff_t page_len; if (write_mode == FALL_BACK_TO_NONDELALLOC) { if (ext4_should_order_data(inode)) { @@ -2535,16 +2526,6 @@ static int ext4_da_write_end(struct file *file, } ret2 = generic_write_end(file, mapping, pos, len, copied, page, fsdata); - - page_len = PAGE_CACHE_SIZE - - ((pos + copied - 1) & (PAGE_CACHE_SIZE - 1)); - - if (page_len > 0) { - ret = ext4_discard_partial_page_buffers_no_lock(handle, - inode, page, pos + copied - 1, page_len, - EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED); - } - copied = ret2; if (ret2 < 0) ret = ret2; @@ -3206,26 +3187,8 @@ int ext4_discard_partial_page_buffers_no_lock(handle_t *handle, iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits); - if (!page_has_buffers(page)) { - /* - * If the range to be discarded covers a partial block - * we need to get the page buffers. This is because - * partial blocks cannot be released and the page needs - * to be updated with the contents of the block before - * we write the zeros on top of it. - */ - if ((from & (blocksize - 1)) || - ((from + length) & (blocksize - 1))) { - create_empty_buffers(page, blocksize, 0); - } else { - /* - * If there are no partial blocks, - * there is nothing to update, - * so we can return now - */ - return 0; - } - } + if (!page_has_buffers(page)) + create_empty_buffers(page, blocksize, 0); /* Find the buffer that contains "offset" */ bh = page_buffers(page); -- cgit v1.2.3-70-g09d2 From 5b5ffa49d4bbb8ca4c41c094261660264f16bd20 Mon Sep 17 00:00:00 2001 From: Yongqiang Yang Date: Tue, 13 Dec 2011 22:13:42 -0500 Subject: ext4: remove a wrong BUG_ON in ext4_ext_convert_to_initialized If a file is fallocated on a hole, map->m_lblk + map->m_len may be greater than ee_block + ee_len. Signed-off-by: Yongqiang Yang Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/extents.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 61cdb97f338..607b1557d29 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -2955,7 +2955,6 @@ static int ext4_ext_convert_to_initialized(handle_t *handle, /* Pre-conditions */ BUG_ON(!ext4_ext_is_uninitialized(ex)); BUG_ON(!in_range(map->m_lblk, ee_block, ee_len)); - BUG_ON(map->m_lblk + map->m_len > ee_block + ee_len); /* * Attempt to transfer newly initialized blocks from the currently -- cgit v1.2.3-70-g09d2 From 5a0dc7365c240795bf190766eba7a27600be3b3e Mon Sep 17 00:00:00 2001 From: Yongqiang Yang Date: Tue, 13 Dec 2011 22:29:12 -0500 Subject: ext4: handle EOF correctly in ext4_bio_write_page() We need to zero out part of a page which beyond EOF before setting uptodate, otherwise, mapread or write will see non-zero data beyond EOF. Signed-off-by: Yongqiang Yang Signed-off-by: "Theodore Ts'o" Cc: stable@kernel.org --- fs/ext4/page-io.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs') diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c index 7ce1d0b19c9..7e106c810c6 100644 --- a/fs/ext4/page-io.c +++ b/fs/ext4/page-io.c @@ -385,6 +385,18 @@ int ext4_bio_write_page(struct ext4_io_submit *io, block_end = block_start + blocksize; if (block_start >= len) { + /* + * Comments copied from block_write_full_page_endio: + * + * The page straddles i_size. It must be zeroed out on + * each and every writepage invocation because it may + * be mmapped. "A file is mapped in multiples of the + * page size. For a file that is not a multiple of + * the page size, the remaining memory is zeroed when + * mapped, and writes to that region are not written + * out to the file." + */ + zero_user_segment(page, block_start, block_end); clear_buffer_dirty(bh); set_buffer_uptodate(bh); continue; -- cgit v1.2.3-70-g09d2 From 759c361eb95964d0d786f8962224dd0d9e780e6e Mon Sep 17 00:00:00 2001 From: Djalal Harouni Date: Tue, 13 Dec 2011 02:47:29 +0100 Subject: fs/ncpfs: fix error paths and goto statements in ncp_fill_super() The label 'out_bdi' should be followed by bdi_destroy() instead of fput() which should be after the 'out_fput' label. If bdi_setup_and_register() fails then jump to the 'out_fput' label instead of the 'out_bdi' one. If fget(data.info_fd) fails then jump to the previously fixed 'out_bdi' label to call bdi_destroy() otherwise the bdi object will not be destroyed. Compile tested only. Signed-off-by: Djalal Harouni Signed-off-by: Al Viro --- fs/ncpfs/inode.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ncpfs/inode.c b/fs/ncpfs/inode.c index 5b5fa33b6b9..cbd1a61c110 100644 --- a/fs/ncpfs/inode.c +++ b/fs/ncpfs/inode.c @@ -548,7 +548,7 @@ static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent) error = bdi_setup_and_register(&server->bdi, "ncpfs", BDI_CAP_MAP_COPY); if (error) - goto out_bdi; + goto out_fput; server->ncp_filp = ncp_filp; server->ncp_sock = sock; @@ -559,7 +559,7 @@ static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent) error = -EBADF; server->info_filp = fget(data.info_fd); if (!server->info_filp) - goto out_fput; + goto out_bdi; error = -ENOTSOCK; sock_inode = server->info_filp->f_path.dentry->d_inode; if (!S_ISSOCK(sock_inode->i_mode)) @@ -746,9 +746,9 @@ out_nls: out_fput2: if (server->info_filp) fput(server->info_filp); -out_fput: - bdi_destroy(&server->bdi); out_bdi: + bdi_destroy(&server->bdi); +out_fput: /* 23/12/1998 Marcin Dalecki : * * The previously used put_filp(ncp_filp); was bogus, since -- cgit v1.2.3-70-g09d2 From f4a2f4c548296168832ad4ab7e7f7b0cd0bf1214 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Wed, 14 Dec 2011 20:12:01 -0500 Subject: Btrfs: fix wrong i_size when truncating a file to a larger size Btrfsck report error 100 after the 83th case of xfstests was run, it means the i_size of the file is wrong. The reason of this bug is that: Btrfs increased i_size of the file at the beginning, but it failed to expand the file, and failed to update the i_size to the old size because there is no enough space in the file system, so we found a wrong i_size. This patch fixes this bug by updating the i_size just when we pass the file expanding and get enough space to update i-node. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index c5ccec23984..4bbceb928af 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3370,6 +3370,8 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) static int btrfs_setsize(struct inode *inode, loff_t newsize) { + struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_trans_handle *trans; loff_t oldsize = i_size_read(inode); int ret; @@ -3377,16 +3379,20 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return 0; if (newsize > oldsize) { - i_size_write(inode, newsize); - btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); truncate_pagecache(inode, oldsize, newsize); ret = btrfs_cont_expand(inode, oldsize, newsize); - if (ret) { - btrfs_setsize(inode, oldsize); + if (ret) return ret; - } - mark_inode_dirty(inode); + trans = btrfs_start_transaction(root, 1); + if (IS_ERR(trans)) + return PTR_ERR(trans); + + i_size_write(inode, newsize); + btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); + ret = btrfs_update_inode(trans, root, inode); + + btrfs_end_transaction_throttle(trans, root); } else { /* -- cgit v1.2.3-70-g09d2 From 3642320e07444cc46327b24977d752f99706dac2 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Wed, 14 Dec 2011 20:12:02 -0500 Subject: Btrfs: fix wrong disk space information of the files Btrfsck report errors after the 83th case of xfstests was run, The error number is 400, it means the used disk space of the file is wrong. The reason of this bug is that: The file truncation may fail when the space of the file system is not enough, and leave some file extents, whose offset are beyond the end of the files. When we want to expand those files, we will drop those file extents, and put in dummy file extents, and then we should update the i-node. But btrfs forgets to do it. This patch adds the forgotten i-node update. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4bbceb928af..f1c4bceed07 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3327,7 +3327,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) u64 hint_byte = 0; hole_size = last_byte - cur_offset; - trans = btrfs_start_transaction(root, 2); + trans = btrfs_start_transaction(root, 3); if (IS_ERR(trans)) { err = PTR_ERR(trans); break; @@ -3337,6 +3337,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) cur_offset + hole_size, &hint_byte, 1); if (err) { + btrfs_update_inode(trans, root, inode); btrfs_end_transaction(trans, root); break; } @@ -3346,6 +3347,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) 0, hole_size, 0, hole_size, 0, 0, 0); if (err) { + btrfs_update_inode(trans, root, inode); btrfs_end_transaction(trans, root); break; } @@ -3353,6 +3355,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); + btrfs_update_inode(trans, root, inode); btrfs_end_transaction(trans, root); } free_extent_map(em); -- cgit v1.2.3-70-g09d2 From 39fb26c398ddf8d7794a85e896cfe1a42e55524b Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Wed, 14 Dec 2011 20:12:02 -0500 Subject: Btrfs: fix inaccurate available space on raid0 profile When we use raid0 as the data profile, df command may show us a very inaccurate value of the available space, which may be much less than the real one. It may make the users puzzled. Fix it by changing the calculation of the available space, and making it be more similar to a fake chunk allocation. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/super.c | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 1a3ce9e0b49..41495dce4db 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1079,7 +1079,7 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) u64 avail_space; u64 used_space; u64 min_stripe_size; - int min_stripes = 1; + int min_stripes = 1, num_stripes = 1; int i = 0, nr_devices; int ret; @@ -1093,12 +1093,16 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) /* calc min stripe number for data space alloction */ type = btrfs_get_alloc_profile(root, 1); - if (type & BTRFS_BLOCK_GROUP_RAID0) + if (type & BTRFS_BLOCK_GROUP_RAID0) { min_stripes = 2; - else if (type & BTRFS_BLOCK_GROUP_RAID1) + num_stripes = nr_devices; + } else if (type & BTRFS_BLOCK_GROUP_RAID1) { min_stripes = 2; - else if (type & BTRFS_BLOCK_GROUP_RAID10) + num_stripes = 2; + } else if (type & BTRFS_BLOCK_GROUP_RAID10) { min_stripes = 4; + num_stripes = 4; + } if (type & BTRFS_BLOCK_GROUP_DUP) min_stripe_size = 2 * BTRFS_STRIPE_LEN; @@ -1167,13 +1171,16 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) i = nr_devices - 1; avail_space = 0; while (nr_devices >= min_stripes) { + if (num_stripes > nr_devices) + num_stripes = nr_devices; + if (devices_info[i].max_avail >= min_stripe_size) { int j; u64 alloc_size; - avail_space += devices_info[i].max_avail * min_stripes; + avail_space += devices_info[i].max_avail * num_stripes; alloc_size = devices_info[i].max_avail; - for (j = i + 1 - min_stripes; j <= i; j++) + for (j = i + 1 - num_stripes; j <= i; j++) devices_info[j].max_avail -= alloc_size; } i--; -- cgit v1.2.3-70-g09d2 From f8e9e0b07be0464e12366631da3da73a1a62449c Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Wed, 14 Dec 2011 20:12:02 -0500 Subject: btrfs: keep orphans for subvolume deletion Since we have the free space caches, btrfs_orphan_cleanup also runs for the tree_root. Unfortunately this also cleans up the orphans used to mark subvol deletions in progress. Currently if a subvol deletion gets interrupted twice by umount/mount, the deletion will not be continued and the space permanently lost, though it would be possible to write a tool to recover those lost subvol deletions. This patch checks if the orphan belongs to a subvol (dead root) and skips the deletion. Signed-off-by: Arne Jansen Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index f1c4bceed07..4a31493d97a 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2158,6 +2158,38 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) if (ret && ret != -ESTALE) goto out; + if (ret == -ESTALE && root == root->fs_info->tree_root) { + struct btrfs_root *dead_root; + struct btrfs_fs_info *fs_info = root->fs_info; + int is_dead_root = 0; + + /* + * this is an orphan in the tree root. Currently these + * could come from 2 sources: + * a) a snapshot deletion in progress + * b) a free space cache inode + * We need to distinguish those two, as the snapshot + * orphan must not get deleted. + * find_dead_roots already ran before us, so if this + * is a snapshot deletion, we should find the root + * in the dead_roots list + */ + spin_lock(&fs_info->trans_lock); + list_for_each_entry(dead_root, &fs_info->dead_roots, + root_list) { + if (dead_root->root_key.objectid == + found_key.objectid) { + is_dead_root = 1; + break; + } + } + spin_unlock(&fs_info->trans_lock); + if (is_dead_root) { + /* prevent this orphan from being found again */ + key.offset = found_key.objectid - 1; + continue; + } + } /* * Inode is already gone but the orphan item is still there, * kill the orphan item. -- cgit v1.2.3-70-g09d2 From 306424cc880a0fbbdc99eee1f43d056a301a180f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 14 Dec 2011 20:12:02 -0500 Subject: Btrfs: fix ctime update of on-disk inode To reproduce the bug: # touch /mnt/tmp # stat /mnt/tmp | grep Change Change: 2011-12-09 09:32:23.412105981 +0800 # chattr +i /mnt/tmp # stat /mnt/tmp | grep Change Change: 2011-12-09 09:32:43.198105295 +0800 # umount /mnt # mount /dev/loop1 /mnt # stat /mnt/tmp | grep Change Change: 2011-12-09 09:32:23.412105981 +0800 We should update ctime of in-memory inode before calling btrfs_update_inode(). Signed-off-by: Li Zefan Signed-off-by: Chris Mason --- fs/btrfs/ioctl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 72d461656f6..40eaa9fdf07 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -252,11 +252,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); + btrfs_update_iflags(inode); + inode->i_ctime = CURRENT_TIME; ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); - btrfs_update_iflags(inode); - inode->i_ctime = CURRENT_TIME; btrfs_end_transaction(trans, root); mnt_drop_write(file->f_path.mnt); -- cgit v1.2.3-70-g09d2 From 8f3b65a3d66bcc086e1eb040b7545e70681f2ed1 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Thu, 15 Dec 2011 09:29:43 -0500 Subject: Btrfs: add a cond_resched() into the worker loop If we have a constant stream of end_io completions or crc work, we can hit softlockup messages from the async helper threads. This adds a cond_resched() into the loop to avoid them. Signed-off-by: Chris Mason --- fs/btrfs/async-thread.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c index 7ec14097fef..203e849dc9c 100644 --- a/fs/btrfs/async-thread.c +++ b/fs/btrfs/async-thread.c @@ -331,7 +331,7 @@ again: run_ordered_completions(worker->workers, work); check_pending_worker_creates(worker); - + cond_resched(); } spin_lock_irq(&worker->lock); -- cgit v1.2.3-70-g09d2 From ad19db71f498fd858dd84ce603efcf97e321f184 Mon Sep 17 00:00:00 2001 From: Casey Schaufler Date: Thu, 15 Dec 2011 10:09:07 -0500 Subject: BTRFS: Establish i_ops before calling d_instantiate The Smack LSM hook for security_d_instantiate checks the inode's i_op->getxattr value to determine if the containing filesystem supports extended attributes. The BTRFS filesystem sets the inode's i_op value only after it has instantiated the inode. This results in Smack incorrectly giving new BTRFS inodes attributes from the filesystem defaults on the assumption that values can't be stored on the filesystem. This patch moves the assignment of inode operation vectors ahead of the calls to d_instantiate, letting Smack know that the filesystem supports extended attributes. There should be no impact on the performance or behavior of BTRFS. Signed-off-by: Casey Schaufler Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 31 ++++++++++++++++++++++++++----- 1 file changed, 26 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4a31493d97a..d4a9195c7f0 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4596,11 +4596,18 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + + inode->i_op = &btrfs_special_inode_operations; err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { - inode->i_op = &btrfs_special_inode_operations; init_special_inode(inode, inode->i_mode, rdev); btrfs_update_inode(trans, root, inode); } @@ -4654,14 +4661,21 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + inode->i_fop = &btrfs_file_operations; + inode->i_op = &btrfs_file_inode_operations; + err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { inode->i_mapping->a_ops = &btrfs_aops; inode->i_mapping->backing_dev_info = &root->fs_info->bdi; - inode->i_fop = &btrfs_file_operations; - inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } out_unlock: @@ -7117,14 +7131,21 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + inode->i_fop = &btrfs_file_operations; + inode->i_op = &btrfs_file_inode_operations; + err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { inode->i_mapping->a_ops = &btrfs_aops; inode->i_mapping->backing_dev_info = &root->fs_info->bdi; - inode->i_fop = &btrfs_file_operations; - inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } if (drop_inode) -- cgit v1.2.3-70-g09d2 From 0dc3b84a73267f47a75468f924f5d58a840e3152 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 18 Nov 2011 14:37:27 -0500 Subject: Btrfs: fix num_workers_starting bug and other bugs in async thread Al pointed out we have some random problems with the way we account for num_workers_starting in the async thread stuff. First of all we need to make sure to decrement num_workers_starting if we fail to start the worker, so make __btrfs_start_workers do this. Also fix __btrfs_start_workers so that it doesn't call btrfs_stop_workers(), there is no point in stopping everybody if we failed to create a worker. Also check_pending_worker_creates needs to call __btrfs_start_work in it's work function since it already increments num_workers_starting. People only start one worker at a time, so get rid of the num_workers argument everywhere, and make btrfs_queue_worker a void since it will always succeed. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/async-thread.c | 115 +++++++++++++++++++++++------------------------- fs/btrfs/async-thread.h | 4 +- fs/btrfs/disk-io.c | 34 ++++++++------ fs/btrfs/scrub.c | 8 +++- 4 files changed, 83 insertions(+), 78 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c index 7ec14097fef..af8e117c897 100644 --- a/fs/btrfs/async-thread.c +++ b/fs/btrfs/async-thread.c @@ -64,6 +64,8 @@ struct btrfs_worker_thread { int idle; }; +static int __btrfs_start_workers(struct btrfs_workers *workers); + /* * btrfs_start_workers uses kthread_run, which can block waiting for memory * for a very long time. It will actually throttle on page writeback, @@ -88,27 +90,10 @@ static void start_new_worker_func(struct btrfs_work *work) { struct worker_start *start; start = container_of(work, struct worker_start, work); - btrfs_start_workers(start->queue, 1); + __btrfs_start_workers(start->queue); kfree(start); } -static int start_new_worker(struct btrfs_workers *queue) -{ - struct worker_start *start; - int ret; - - start = kzalloc(sizeof(*start), GFP_NOFS); - if (!start) - return -ENOMEM; - - start->work.func = start_new_worker_func; - start->queue = queue; - ret = btrfs_queue_worker(queue->atomic_worker_start, &start->work); - if (ret) - kfree(start); - return ret; -} - /* * helper function to move a thread onto the idle list after it * has finished some requests. @@ -153,12 +138,20 @@ static void check_busy_worker(struct btrfs_worker_thread *worker) static void check_pending_worker_creates(struct btrfs_worker_thread *worker) { struct btrfs_workers *workers = worker->workers; + struct worker_start *start; unsigned long flags; rmb(); if (!workers->atomic_start_pending) return; + start = kzalloc(sizeof(*start), GFP_NOFS); + if (!start) + return; + + start->work.func = start_new_worker_func; + start->queue = workers; + spin_lock_irqsave(&workers->lock, flags); if (!workers->atomic_start_pending) goto out; @@ -170,10 +163,11 @@ static void check_pending_worker_creates(struct btrfs_worker_thread *worker) workers->num_workers_starting += 1; spin_unlock_irqrestore(&workers->lock, flags); - start_new_worker(workers); + btrfs_queue_worker(workers->atomic_worker_start, &start->work); return; out: + kfree(start); spin_unlock_irqrestore(&workers->lock, flags); } @@ -462,56 +456,55 @@ void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max, * starts new worker threads. This does not enforce the max worker * count in case you need to temporarily go past it. */ -static int __btrfs_start_workers(struct btrfs_workers *workers, - int num_workers) +static int __btrfs_start_workers(struct btrfs_workers *workers) { struct btrfs_worker_thread *worker; int ret = 0; - int i; - for (i = 0; i < num_workers; i++) { - worker = kzalloc(sizeof(*worker), GFP_NOFS); - if (!worker) { - ret = -ENOMEM; - goto fail; - } + worker = kzalloc(sizeof(*worker), GFP_NOFS); + if (!worker) { + ret = -ENOMEM; + goto fail; + } - INIT_LIST_HEAD(&worker->pending); - INIT_LIST_HEAD(&worker->prio_pending); - INIT_LIST_HEAD(&worker->worker_list); - spin_lock_init(&worker->lock); - - atomic_set(&worker->num_pending, 0); - atomic_set(&worker->refs, 1); - worker->workers = workers; - worker->task = kthread_run(worker_loop, worker, - "btrfs-%s-%d", workers->name, - workers->num_workers + i); - if (IS_ERR(worker->task)) { - ret = PTR_ERR(worker->task); - kfree(worker); - goto fail; - } - spin_lock_irq(&workers->lock); - list_add_tail(&worker->worker_list, &workers->idle_list); - worker->idle = 1; - workers->num_workers++; - workers->num_workers_starting--; - WARN_ON(workers->num_workers_starting < 0); - spin_unlock_irq(&workers->lock); + INIT_LIST_HEAD(&worker->pending); + INIT_LIST_HEAD(&worker->prio_pending); + INIT_LIST_HEAD(&worker->worker_list); + spin_lock_init(&worker->lock); + + atomic_set(&worker->num_pending, 0); + atomic_set(&worker->refs, 1); + worker->workers = workers; + worker->task = kthread_run(worker_loop, worker, + "btrfs-%s-%d", workers->name, + workers->num_workers + 1); + if (IS_ERR(worker->task)) { + ret = PTR_ERR(worker->task); + kfree(worker); + goto fail; } + spin_lock_irq(&workers->lock); + list_add_tail(&worker->worker_list, &workers->idle_list); + worker->idle = 1; + workers->num_workers++; + workers->num_workers_starting--; + WARN_ON(workers->num_workers_starting < 0); + spin_unlock_irq(&workers->lock); + return 0; fail: - btrfs_stop_workers(workers); + spin_lock_irq(&workers->lock); + workers->num_workers_starting--; + spin_unlock_irq(&workers->lock); return ret; } -int btrfs_start_workers(struct btrfs_workers *workers, int num_workers) +int btrfs_start_workers(struct btrfs_workers *workers) { spin_lock_irq(&workers->lock); - workers->num_workers_starting += num_workers; + workers->num_workers_starting++; spin_unlock_irq(&workers->lock); - return __btrfs_start_workers(workers, num_workers); + return __btrfs_start_workers(workers); } /* @@ -568,6 +561,7 @@ static struct btrfs_worker_thread *find_worker(struct btrfs_workers *workers) struct btrfs_worker_thread *worker; unsigned long flags; struct list_head *fallback; + int ret; again: spin_lock_irqsave(&workers->lock, flags); @@ -584,7 +578,9 @@ again: workers->num_workers_starting++; spin_unlock_irqrestore(&workers->lock, flags); /* we're below the limit, start another worker */ - __btrfs_start_workers(workers, 1); + ret = __btrfs_start_workers(workers); + if (ret) + goto fallback; goto again; } } @@ -665,7 +661,7 @@ void btrfs_set_work_high_prio(struct btrfs_work *work) /* * places a struct btrfs_work into the pending queue of one of the kthreads */ -int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) +void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) { struct btrfs_worker_thread *worker; unsigned long flags; @@ -673,7 +669,7 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) /* don't requeue something already on a list */ if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags)) - goto out; + return; worker = find_worker(workers); if (workers->ordered) { @@ -712,7 +708,4 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) if (wake) wake_up_process(worker->task); spin_unlock_irqrestore(&worker->lock, flags); - -out: - return 0; } diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h index 5077746cf85..f34cc31fa3c 100644 --- a/fs/btrfs/async-thread.h +++ b/fs/btrfs/async-thread.h @@ -109,8 +109,8 @@ struct btrfs_workers { char *name; }; -int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work); -int btrfs_start_workers(struct btrfs_workers *workers, int num_workers); +void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work); +int btrfs_start_workers(struct btrfs_workers *workers); int btrfs_stop_workers(struct btrfs_workers *workers); void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max, struct btrfs_workers *async_starter); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 94abc25392f..3f9d5551e58 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2194,19 +2194,27 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->endio_meta_write_workers.idle_thresh = 2; fs_info->readahead_workers.idle_thresh = 2; - btrfs_start_workers(&fs_info->workers, 1); - btrfs_start_workers(&fs_info->generic_worker, 1); - btrfs_start_workers(&fs_info->submit_workers, 1); - btrfs_start_workers(&fs_info->delalloc_workers, 1); - btrfs_start_workers(&fs_info->fixup_workers, 1); - btrfs_start_workers(&fs_info->endio_workers, 1); - btrfs_start_workers(&fs_info->endio_meta_workers, 1); - btrfs_start_workers(&fs_info->endio_meta_write_workers, 1); - btrfs_start_workers(&fs_info->endio_write_workers, 1); - btrfs_start_workers(&fs_info->endio_freespace_worker, 1); - btrfs_start_workers(&fs_info->delayed_workers, 1); - btrfs_start_workers(&fs_info->caching_workers, 1); - btrfs_start_workers(&fs_info->readahead_workers, 1); + /* + * btrfs_start_workers can really only fail because of ENOMEM so just + * return -ENOMEM if any of these fail. + */ + ret = btrfs_start_workers(&fs_info->workers); + ret |= btrfs_start_workers(&fs_info->generic_worker); + ret |= btrfs_start_workers(&fs_info->submit_workers); + ret |= btrfs_start_workers(&fs_info->delalloc_workers); + ret |= btrfs_start_workers(&fs_info->fixup_workers); + ret |= btrfs_start_workers(&fs_info->endio_workers); + ret |= btrfs_start_workers(&fs_info->endio_meta_workers); + ret |= btrfs_start_workers(&fs_info->endio_meta_write_workers); + ret |= btrfs_start_workers(&fs_info->endio_write_workers); + ret |= btrfs_start_workers(&fs_info->endio_freespace_worker); + ret |= btrfs_start_workers(&fs_info->delayed_workers); + ret |= btrfs_start_workers(&fs_info->caching_workers); + ret |= btrfs_start_workers(&fs_info->readahead_workers); + if (ret) { + ret = -ENOMEM; + goto fail_sb_buffer; + } fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super); fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages, diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index c27bcb67f33..ddf2c90d3fc 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -1535,18 +1535,22 @@ static noinline_for_stack int scrub_supers(struct scrub_dev *sdev) static noinline_for_stack int scrub_workers_get(struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; + int ret = 0; mutex_lock(&fs_info->scrub_lock); if (fs_info->scrub_workers_refcnt == 0) { btrfs_init_workers(&fs_info->scrub_workers, "scrub", fs_info->thread_pool_size, &fs_info->generic_worker); fs_info->scrub_workers.idle_thresh = 4; - btrfs_start_workers(&fs_info->scrub_workers, 1); + ret = btrfs_start_workers(&fs_info->scrub_workers); + if (ret) + goto out; } ++fs_info->scrub_workers_refcnt; +out: mutex_unlock(&fs_info->scrub_lock); - return 0; + return ret; } static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) -- cgit v1.2.3-70-g09d2 From 22c44fe65adacd20a174f3f54686509ee94ef7be Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 30 Nov 2011 10:45:38 -0500 Subject: Btrfs: deal with enospc from dirtying inodes properly Now that we're properly keeping track of delayed inode space we've been getting a lot of warnings out of btrfs_dirty_inode() when running xfstest 83. This is because a bunch of people call mark_inode_dirty, which is void so we can't return ENOSPC. This needs to be fixed in a few areas 1) file_update_time - this updates the mtime and such when writing to a file, which will call mark_inode_dirty. So copy file_update_time into btrfs so we can call btrfs_dirty_inode directly and return an error if we get one appropriately. 2) fix symlinks to use btrfs_setattr for ->setattr. For some reason we weren't setting ->setattr for symlinks, even though we should have been. This catches one of the cases where we were getting errors in mark_inode_dirty. 3) Fix btrfs_setattr and btrfs_setsize to call btrfs_dirty_inode directly instead of mark_inode_dirty. This lets us return errors properly for truncate and chown/anything related to setattr. 4) Add a new btrfs_fs_dirty_inode which will just call btrfs_dirty_inode and print an error if we have one. The only remaining user we can't control for this is touch_atime(), but we don't really want to keep people from walking down the tree if we don't have space to save the atime update, so just complain but don't worry about it. With this patch xfstests 83 complains a handful of times instead of hundreds of times. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/ctree.h | 3 ++- fs/btrfs/file.c | 6 ++++- fs/btrfs/inode.c | 80 ++++++++++++++++++++++++++++++++++++++++++-------------- fs/btrfs/super.c | 13 ++++++++- 4 files changed, 80 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 50634abef9b..67385033323 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2692,7 +2692,8 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf); int btrfs_readpage(struct file *file, struct page *page); void btrfs_evict_inode(struct inode *inode); int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc); -void btrfs_dirty_inode(struct inode *inode, int flags); +int btrfs_dirty_inode(struct inode *inode); +int btrfs_update_time(struct file *file); struct inode *btrfs_alloc_inode(struct super_block *sb); void btrfs_destroy_inode(struct inode *inode); int btrfs_drop_inode(struct inode *inode); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index f2e92828960..cc7492c823f 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1387,7 +1387,11 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, goto out; } - file_update_time(file); + err = btrfs_update_time(file); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } BTRFS_I(inode)->sequence++; start_pos = round_down(pos, root->sectorsize); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index c5ccec23984..eec33b9c953 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -38,6 +38,7 @@ #include #include #include +#include #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -3386,7 +3387,7 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return ret; } - mark_inode_dirty(inode); + ret = btrfs_dirty_inode(inode); } else { /* @@ -3426,9 +3427,9 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) if (attr->ia_valid) { setattr_copy(inode, attr); - mark_inode_dirty(inode); + err = btrfs_dirty_inode(inode); - if (attr->ia_valid & ATTR_MODE) + if (!err && attr->ia_valid & ATTR_MODE) err = btrfs_acl_chmod(inode); } @@ -4204,42 +4205,80 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) * FIXME, needs more benchmarking...there are no reasons other than performance * to keep or drop this code. */ -void btrfs_dirty_inode(struct inode *inode, int flags) +int btrfs_dirty_inode(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; int ret; if (BTRFS_I(inode)->dummy_inode) - return; + return 0; trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); ret = btrfs_update_inode(trans, root, inode); if (ret && ret == -ENOSPC) { /* whoops, lets try again with the full transaction */ btrfs_end_transaction(trans, root); trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) { - printk_ratelimited(KERN_ERR "btrfs: fail to " - "dirty inode %llu error %ld\n", - (unsigned long long)btrfs_ino(inode), - PTR_ERR(trans)); - return; - } + if (IS_ERR(trans)) + return PTR_ERR(trans); ret = btrfs_update_inode(trans, root, inode); - if (ret) { - printk_ratelimited(KERN_ERR "btrfs: fail to " - "dirty inode %llu error %d\n", - (unsigned long long)btrfs_ino(inode), - ret); - } } btrfs_end_transaction(trans, root); if (BTRFS_I(inode)->delayed_node) btrfs_balance_delayed_items(root); + + return ret; +} + +/* + * This is a copy of file_update_time. We need this so we can return error on + * ENOSPC for updating the inode in the case of file write and mmap writes. + */ +int btrfs_update_time(struct file *file) +{ + struct inode *inode = file->f_path.dentry->d_inode; + struct timespec now; + int ret; + enum { S_MTIME = 1, S_CTIME = 2, S_VERSION = 4 } sync_it = 0; + + /* First try to exhaust all avenues to not sync */ + if (IS_NOCMTIME(inode)) + return 0; + + now = current_fs_time(inode->i_sb); + if (!timespec_equal(&inode->i_mtime, &now)) + sync_it = S_MTIME; + + if (!timespec_equal(&inode->i_ctime, &now)) + sync_it |= S_CTIME; + + if (IS_I_VERSION(inode)) + sync_it |= S_VERSION; + + if (!sync_it) + return 0; + + /* Finally allowed to write? Takes lock. */ + if (mnt_want_write_file(file)) + return 0; + + /* Only change inode inside the lock region */ + if (sync_it & S_VERSION) + inode_inc_iversion(inode); + if (sync_it & S_CTIME) + inode->i_ctime = now; + if (sync_it & S_MTIME) + inode->i_mtime = now; + ret = btrfs_dirty_inode(inode); + if (!ret) + mark_inode_dirty_sync(inode); + mnt_drop_write(file->f_path.mnt); + return ret; } /* @@ -6304,6 +6343,8 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) u64 page_end; ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); + if (!ret) + ret = btrfs_update_time(vma->vm_file); if (ret) { if (ret == -ENOMEM) ret = VM_FAULT_OOM; @@ -7353,6 +7394,7 @@ static const struct inode_operations btrfs_symlink_inode_operations = { .follow_link = page_follow_link_light, .put_link = page_put_link, .getattr = btrfs_getattr, + .setattr = btrfs_setattr, .permission = btrfs_permission, .setxattr = btrfs_setxattr, .getxattr = btrfs_getxattr, diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 1a3ce9e0b49..05b40457e26 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -41,6 +41,7 @@ #include #include #include +#include #include "compat.h" #include "delayed-inode.h" #include "ctree.h" @@ -1290,6 +1291,16 @@ static int btrfs_unfreeze(struct super_block *sb) return 0; } +static void btrfs_fs_dirty_inode(struct inode *inode, int flags) +{ + int ret; + + ret = btrfs_dirty_inode(inode); + if (ret) + printk_ratelimited(KERN_ERR "btrfs: fail to dirty inode %Lu " + "error %d\n", btrfs_ino(inode), ret); +} + static const struct super_operations btrfs_super_ops = { .drop_inode = btrfs_drop_inode, .evict_inode = btrfs_evict_inode, @@ -1297,7 +1308,7 @@ static const struct super_operations btrfs_super_ops = { .sync_fs = btrfs_sync_fs, .show_options = btrfs_show_options, .write_inode = btrfs_write_inode, - .dirty_inode = btrfs_dirty_inode, + .dirty_inode = btrfs_fs_dirty_inode, .alloc_inode = btrfs_alloc_inode, .destroy_inode = btrfs_destroy_inode, .statfs = btrfs_statfs, -- cgit v1.2.3-70-g09d2 From 660d3f6cde552323578b85fc5a09a6742f1fe804 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 9 Dec 2011 11:18:51 -0500 Subject: Btrfs: fix how we do delalloc reservations and how we free reservations on error Running xfstests 269 with some tracing my scripts kept spitting out errors about releasing bytes that we didn't actually have reserved. This took me down a huge rabbit hole and it turns out the way we deal with reserved_extents is wrong, we need to only be setting it if the reservation succeeds, otherwise the free() method will come in and unreserve space that isn't actually reserved yet, which can lead to other warnings and such. The math was all working out right in the end, but it caused all sorts of other issues in addition to making my scripts yell and scream and generally make it impossible for me to track down the original issue I was looking for. The other problem is with our error handling in the reservation code. There are two cases that we need to deal with 1) We raced with free. In this case free won't free anything because csum_bytes is modified before we dro the lock in our reservation path, so free rightly doesn't release any space because the reservation code may be depending on that reservation. However if we fail, we need the reservation side to do the free at that point since that space is no longer in use. So as it stands the code was doing this fine and it worked out, except in case #2 2) We don't race with free. Nobody comes in and changes anything, and our reservation fails. In this case we didn't reserve anything anyway and we just need to clean up csum_bytes but not free anything. So we keep track of csum_bytes before we drop the lock and if it hasn't changed we know we can just decrement csum_bytes and carry on. Because of the case where we can race with free()'s since we have to drop our spin_lock to do the reservation, I'm going to serialize all reservations with the i_mutex. We already get this for free in the heavy use paths, truncate and file write all hold the i_mutex, just needed to add it to page_mkwrite and various ioctl/balance things. With this patch my space leak scripts no longer scream bloody murder. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/extent-tree.c | 43 ++++++++++++++++++++++++++++++------------- fs/btrfs/inode.c | 10 ++++++++++ fs/btrfs/ioctl.c | 2 ++ fs/btrfs/relocation.c | 2 ++ 4 files changed, 44 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 05e1386b8be..4eb4d2748be 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4203,12 +4203,17 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; u64 to_reserve = 0; + u64 csum_bytes; unsigned nr_extents = 0; + int extra_reserve = 0; int flush = 1; int ret; + /* Need to be holding the i_mutex here if we aren't free space cache */ if (btrfs_is_free_space_inode(root, inode)) flush = 0; + else + WARN_ON(!mutex_is_locked(&inode->i_mutex)); if (flush && btrfs_transaction_in_commit(root->fs_info)) schedule_timeout(1); @@ -4219,11 +4224,9 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) BTRFS_I(inode)->outstanding_extents++; if (BTRFS_I(inode)->outstanding_extents > - BTRFS_I(inode)->reserved_extents) { + BTRFS_I(inode)->reserved_extents) nr_extents = BTRFS_I(inode)->outstanding_extents - BTRFS_I(inode)->reserved_extents; - BTRFS_I(inode)->reserved_extents += nr_extents; - } /* * Add an item to reserve for updating the inode when we complete the @@ -4231,11 +4234,12 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) */ if (!BTRFS_I(inode)->delalloc_meta_reserved) { nr_extents++; - BTRFS_I(inode)->delalloc_meta_reserved = 1; + extra_reserve = 1; } to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); + csum_bytes = BTRFS_I(inode)->csum_bytes; spin_unlock(&BTRFS_I(inode)->lock); ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); @@ -4245,22 +4249,35 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) spin_lock(&BTRFS_I(inode)->lock); dropped = drop_outstanding_extent(inode); - to_free = calc_csum_metadata_size(inode, num_bytes, 0); - spin_unlock(&BTRFS_I(inode)->lock); - to_free += btrfs_calc_trans_metadata_size(root, dropped); - /* - * Somebody could have come in and twiddled with the - * reservation, so if we have to free more than we would have - * reserved from this reservation go ahead and release those - * bytes. + * If the inodes csum_bytes is the same as the original + * csum_bytes then we know we haven't raced with any free()ers + * so we can just reduce our inodes csum bytes and carry on. + * Otherwise we have to do the normal free thing to account for + * the case that the free side didn't free up its reserve + * because of this outstanding reservation. */ - to_free -= to_reserve; + if (BTRFS_I(inode)->csum_bytes == csum_bytes) + calc_csum_metadata_size(inode, num_bytes, 0); + else + to_free = calc_csum_metadata_size(inode, num_bytes, 0); + spin_unlock(&BTRFS_I(inode)->lock); + if (dropped) + to_free += btrfs_calc_trans_metadata_size(root, dropped); + if (to_free) btrfs_block_rsv_release(root, block_rsv, to_free); return ret; } + spin_lock(&BTRFS_I(inode)->lock); + if (extra_reserve) { + BTRFS_I(inode)->delalloc_meta_reserved = 1; + nr_extents--; + } + BTRFS_I(inode)->reserved_extents += nr_extents; + spin_unlock(&BTRFS_I(inode)->lock); + block_rsv_add_bytes(block_rsv, to_reserve, 1); return 0; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index eec33b9c953..8938174e6bc 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2192,7 +2192,14 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) continue; } nr_truncate++; + /* + * Need to hold the imutex for reservation purposes, not + * a huge deal here but I have a WARN_ON in + * btrfs_delalloc_reserve_space to catch offenders. + */ + mutex_lock(&inode->i_mutex); ret = btrfs_truncate(inode); + mutex_unlock(&inode->i_mutex); } else { nr_unlink++; } @@ -6342,7 +6349,10 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) u64 page_start; u64 page_end; + /* Need this to keep space reservations serialized */ + mutex_lock(&inode->i_mutex); ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); + mutex_unlock(&inode->i_mutex); if (!ret) ret = btrfs_update_time(vma->vm_file); if (ret) { diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 72d461656f6..dd889166235 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -858,8 +858,10 @@ static int cluster_pages_for_defrag(struct inode *inode, return 0; file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + mutex_lock(&inode->i_mutex); ret = btrfs_delalloc_reserve_space(inode, num_pages << PAGE_CACHE_SHIFT); + mutex_unlock(&inode->i_mutex); if (ret) return ret; again: diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index dff29d5e151..cfb55434a46 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2947,7 +2947,9 @@ static int relocate_file_extent_cluster(struct inode *inode, index = (cluster->start - offset) >> PAGE_CACHE_SHIFT; last_index = (cluster->end - offset) >> PAGE_CACHE_SHIFT; while (index <= last_index) { + mutex_lock(&inode->i_mutex); ret = btrfs_delalloc_reserve_metadata(inode, PAGE_CACHE_SIZE); + mutex_unlock(&inode->i_mutex); if (ret) goto out; -- cgit v1.2.3-70-g09d2 From 7041ee97281c30a78658904140c7bd9373a36142 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 9 Dec 2011 13:26:22 -0500 Subject: Btrfs: fix leaked space in truncate We were occasionaly leaking space when running xfstest 269. This is because if we failed to start the transaction in the truncate loop we'd just goto out, but we need to break so that the inode is removed from the orphan list and the space is properly freed. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 8938174e6bc..6349c63a4b3 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6566,8 +6566,9 @@ static int btrfs_truncate(struct inode *inode) /* Just need the 1 for updating the inode */ trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { - err = PTR_ERR(trans); - goto out; + ret = err = PTR_ERR(trans); + trans = NULL; + break; } } -- cgit v1.2.3-70-g09d2 From ee4d89f0c4967c624c92516fcc37b41069bfdc23 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 13 Dec 2011 12:55:58 -0500 Subject: Btrfs: don't panic if orphan item already exists I've been hitting this BUG_ON() in btrfs_orphan_add when running xfstest 269 in a loop. This is because we will add an orphan item, do the truncate, the truncate will fail for whatever reason (*cough*ENOSPC*cough*) and then we're left with an orphan item still in the fs. Then we come back later to do another truncate and it blows up because we already have an orphan item. This is ok so just fix the BUG_ON() to only BUG() if ret is not EEXIST. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 6349c63a4b3..b212f391cea 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2032,7 +2032,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) /* insert an orphan item to track this unlinked/truncated file */ if (insert >= 1) { ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); - BUG_ON(ret); + BUG_ON(ret && ret != -EEXIST); } /* insert an orphan item to track subvolume contains orphan files */ -- cgit v1.2.3-70-g09d2 From e65cbb94e036058128a5dec6398be2fd64cf88ba Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 13 Dec 2011 16:04:54 -0500 Subject: Btrfs: only set cache_generation if we setup the block group A user reported a problem booting into a new kernel with the old format inodes. He was panicing in cow_file_range while writing out the inode cache. This is because if the block group is not cached we'll just skip writing out the cache, however if it gets dirtied again in the same transaction and it finished caching we'd go ahead and write it out, but since we set cache_generation to the transid we think we've already truncated it and will just carry on, running into cow_file_range and blowing up. We need to make sure we only set cache_generation if we've done the truncate. The user tested this patch and verified that the panic no longer occured. Thanks, Reported-and-Tested-by: Klaus Bitto Signed-off-by: Josef Bacik --- fs/btrfs/extent-tree.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 4eb4d2748be..8603ee4e3df 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2822,7 +2822,7 @@ out_free: btrfs_release_path(path); out: spin_lock(&block_group->lock); - if (!ret) + if (!ret && dcs == BTRFS_DC_SETUP) block_group->cache_generation = trans->transid; block_group->disk_cache_state = dcs; spin_unlock(&block_group->lock); -- cgit v1.2.3-70-g09d2 From e755d9ab387c4c2401dce5b6f435432a829f41cb Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Thu, 15 Dec 2011 13:36:29 -0500 Subject: Btrfs: deal with NULL srv_rsv in the delalloc inode reservation code btrfs_update_inode is sometimes called with a null reservation. Signed-off-by: Chris Mason --- fs/btrfs/delayed-inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 6a1a6800776..c7ddf8a01c5 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -640,8 +640,8 @@ static int btrfs_delayed_inode_reserve_metadata( * Now if src_rsv == delalloc_block_rsv we'll let it just steal since * we're accounted for. */ - if (!trans->bytes_reserved && - src_rsv != &root->fs_info->delalloc_block_rsv) { + if (!src_rsv || (!trans->bytes_reserved && + src_rsv != &root->fs_info->delalloc_block_rsv)) { ret = btrfs_block_rsv_add_noflush(root, dst_rsv, num_bytes); /* * Since we're under a transaction reserve_metadata_bytes could -- cgit v1.2.3-70-g09d2 From d85c8a6f1bc083279215ff6e79b7c292bf3ec905 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Thu, 15 Dec 2011 15:38:41 -0500 Subject: Btrfs: unplug every once and a while The btrfs io submission threads can build up massive plug lists. This keeps things more reasonable so we don't hand over huge dumps of IO at once. Signed-off-by: Chris Mason --- fs/btrfs/volumes.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 91ea57a1474..f4b839fd3c9 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -295,6 +295,12 @@ loop_lock: btrfs_requeue_work(&device->work); goto done; } + /* unplug every 64 requests just for good measure */ + if (batch_run % 64 == 0) { + blk_finish_plug(&plug); + blk_start_plug(&plug); + sync_pending = 0; + } } cond_resched(); -- cgit v1.2.3-70-g09d2 From 6c52961743f38747401b47127b82159ab6d8a7a4 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 15 Dec 2011 18:38:10 -0500 Subject: NFS: Fix a regression in nfs_file_llseek() After commit 06222e491e663dac939f04b125c9dc52126a75c4 (fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek) the behaviour of llseek() was changed so that it always revalidates the file size. The bug appears to be due to a logic error in the afore-mentioned commit, which always evaluates to 'true'. Reported-by: Roel Kluin Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org [>=3.1] --- fs/nfs/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index eca56d4b39c..606ef0f20ae 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -147,7 +147,7 @@ static loff_t nfs_file_llseek(struct file *filp, loff_t offset, int origin) * origin == SEEK_END || SEEK_DATA || SEEK_HOLE => we must revalidate * the cached file length */ - if (origin != SEEK_SET || origin != SEEK_CUR) { + if (origin != SEEK_SET && origin != SEEK_CUR) { struct inode *inode = filp->f_mapping->host; int retval = nfs_revalidate_file_size(inode, filp); -- cgit v1.2.3-70-g09d2 From 142349f541d0bb6bc3e0d4563268105aada42b0b Mon Sep 17 00:00:00 2001 From: Wu Fengguang Date: Fri, 16 Dec 2011 12:32:57 -0500 Subject: btrfs: lower the dirty balance poll interval Tests show that the original large intervals can easily make the dirty limit exceeded on 100 concurrent dd's. So adapt to as large as the next check point selected by the dirty throttling algorithm. Signed-off-by: Wu Fengguang Signed-off-by: Chris Mason --- fs/btrfs/file.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index dafdfa059bf..52305a885c3 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1167,6 +1167,8 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE, PAGE_CACHE_SIZE / (sizeof(struct page *))); + nrptrs = min(nrptrs, current->nr_dirtied_pause - current->nr_dirtied); + nrptrs = max(nrptrs, 8); pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL); if (!pages) return -ENOMEM; -- cgit v1.2.3-70-g09d2 From 695c60f21c69e525a89279a5f35bae4ff237afbc Mon Sep 17 00:00:00 2001 From: Thomas Meyer Date: Mon, 19 Dec 2011 17:11:55 -0800 Subject: nilfs2: unbreak compat ioctl commit 828b1c50ae ("nilfs2: add compat ioctl") incidentally broke all other NILFS compat ioctls. Make them work again. Signed-off-by: Thomas Meyer Signed-off-by: Ryusuke Konishi Tested-by: Ryusuke Konishi Cc: [3.0+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/ioctl.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'fs') diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index 41d6743d303..3e654273cfc 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -842,6 +842,19 @@ long nilfs_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) case FS_IOC32_GETVERSION: cmd = FS_IOC_GETVERSION; break; + case NILFS_IOCTL_CHANGE_CPMODE: + case NILFS_IOCTL_DELETE_CHECKPOINT: + case NILFS_IOCTL_GET_CPINFO: + case NILFS_IOCTL_GET_CPSTAT: + case NILFS_IOCTL_GET_SUINFO: + case NILFS_IOCTL_GET_SUSTAT: + case NILFS_IOCTL_GET_VINFO: + case NILFS_IOCTL_GET_BDESCS: + case NILFS_IOCTL_CLEAN_SEGMENTS: + case NILFS_IOCTL_SYNC: + case NILFS_IOCTL_RESIZE: + case NILFS_IOCTL_SET_ALLOC_RANGE: + break; default: return -ENOIOCTLCMD; } -- cgit v1.2.3-70-g09d2 From 481fe17e973fb97aa3edf17c69557afe88d8334f Mon Sep 17 00:00:00 2001 From: Haogang Chen Date: Mon, 19 Dec 2011 17:11:56 -0800 Subject: nilfs2: potential integer overflow in nilfs_ioctl_clean_segments() There is a potential integer overflow in nilfs_ioctl_clean_segments(). When a large argv[n].v_nmembs is passed from the userspace, the subsequent call to vmalloc() will allocate a buffer smaller than expected, which leads to out-of-bound access in nilfs_ioctl_move_blocks() and lfs_clean_segments(). The following check does not prevent the overflow because nsegs is also controlled by the userspace and could be very large. if (argv[n].v_nmembs > nsegs * nilfs->ns_blocks_per_segment) goto out_free; This patch clamps argv[n].v_nmembs to UINT_MAX / argv[n].v_size, and returns -EINVAL when overflow. Signed-off-by: Haogang Chen Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/ioctl.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index 3e654273cfc..ac258beeda3 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -625,6 +625,9 @@ static int nilfs_ioctl_clean_segments(struct inode *inode, struct file *filp, if (argv[n].v_nmembs > nsegs * nilfs->ns_blocks_per_segment) goto out_free; + if (argv[n].v_nmembs >= UINT_MAX / argv[n].v_size) + goto out_free; + len = argv[n].v_size * argv[n].v_nmembs; base = (void __user *)(unsigned long)argv[n].v_base; if (len == 0) { -- cgit v1.2.3-70-g09d2