From 4177af3a8a6f119484c7903845c6693d7381c13e Mon Sep 17 00:00:00 2001 From: Chandra Seetharaman Date: Mon, 23 Jan 2012 17:31:43 +0000 Subject: Define new macro XFS_ALL_QUOTA_ACTIVE and simply some usage Define new macro XFS_ALL_QUOTA_ACTIVE and simply some usage of quota macros. Signed-off-by: Chandra Seetharaman Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers --- fs/xfs/xfs_super.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'fs/xfs/xfs_super.c') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index ee5b695c99a..5e0d43f231a 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -324,10 +324,9 @@ xfs_parseargs( } else if (!strcmp(this_char, MNTOPT_FILESTREAM)) { mp->m_flags |= XFS_MOUNT_FILESTREAMS; } else if (!strcmp(this_char, MNTOPT_NOQUOTA)) { - mp->m_qflags &= ~(XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE | - XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE | - XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE | - XFS_UQUOTA_ENFD | XFS_OQUOTA_ENFD); + mp->m_qflags &= ~XFS_ALL_QUOTA_ACCT; + mp->m_qflags &= ~XFS_ALL_QUOTA_ENFD; + mp->m_qflags &= ~XFS_ALL_QUOTA_ACTIVE; } else if (!strcmp(this_char, MNTOPT_QUOTA) || !strcmp(this_char, MNTOPT_UQUOTA) || !strcmp(this_char, MNTOPT_USRQUOTA)) { -- cgit v1.2.3-70-g09d2 From aa6bf01d391935a8929333bc2e243084ea0c58db Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 29 Feb 2012 09:53:48 +0000 Subject: xfs: use per-filesystem I/O completion workqueues The new concurrency managed workqueues are cheap enough that we can create per-filesystem instead of global workqueues. This allows us to remove the trylock or defer scheme on the ilock, which is not helpful once we have outstanding log reservations until finishing a size update. Also allow the default concurrency on this workqueues so that I/O completions blocking on the ilock for one inode do not block process for another inode. Reviewed-by: Dave Chinner Reviewed-by: Mark Tinguely Signed-off-by: Christoph Hellwig Signed-off-by: Ben Myers --- fs/xfs/xfs_aops.c | 39 ++++++++++----------------------------- fs/xfs/xfs_aops.h | 2 -- fs/xfs/xfs_buf.c | 17 ----------------- fs/xfs/xfs_mount.h | 3 +++ fs/xfs/xfs_super.c | 39 ++++++++++++++++++++++++++++++++++++++- 5 files changed, 51 insertions(+), 49 deletions(-) (limited to 'fs/xfs/xfs_super.c') diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 74b9baf36ac..540a01742c6 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -126,21 +126,15 @@ static inline bool xfs_ioend_is_append(struct xfs_ioend *ioend) /* * Update on-disk file size now that data has been written to disk. - * - * This function does not block as blocking on the inode lock in IO completion - * can lead to IO completion order dependency deadlocks.. If it can't get the - * inode ilock it will return EAGAIN. Callers must handle this. */ -STATIC int +STATIC void xfs_setfilesize( - xfs_ioend_t *ioend) + struct xfs_ioend *ioend) { - xfs_inode_t *ip = XFS_I(ioend->io_inode); + struct xfs_inode *ip = XFS_I(ioend->io_inode); xfs_fsize_t isize; - if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) - return EAGAIN; - + xfs_ilock(ip, XFS_ILOCK_EXCL); isize = xfs_ioend_new_eof(ioend); if (isize) { trace_xfs_setfilesize(ip, ioend->io_offset, ioend->io_size); @@ -149,7 +143,6 @@ xfs_setfilesize( } xfs_iunlock(ip, XFS_ILOCK_EXCL); - return 0; } /* @@ -163,10 +156,12 @@ xfs_finish_ioend( struct xfs_ioend *ioend) { if (atomic_dec_and_test(&ioend->io_remaining)) { + struct xfs_mount *mp = XFS_I(ioend->io_inode)->i_mount; + if (ioend->io_type == IO_UNWRITTEN) - queue_work(xfsconvertd_workqueue, &ioend->io_work); + queue_work(mp->m_unwritten_workqueue, &ioend->io_work); else if (xfs_ioend_is_append(ioend)) - queue_work(xfsdatad_workqueue, &ioend->io_work); + queue_work(mp->m_data_workqueue, &ioend->io_work); else xfs_destroy_ioend(ioend); } @@ -207,23 +202,9 @@ xfs_end_io( * We might have to update the on-disk file size after extending * writes. */ - error = xfs_setfilesize(ioend); - ASSERT(!error || error == EAGAIN); - + xfs_setfilesize(ioend); done: - /* - * If we didn't complete processing of the ioend, requeue it to the - * tail of the workqueue for another attempt later. Otherwise destroy - * it. - */ - if (error == EAGAIN) { - atomic_inc(&ioend->io_remaining); - xfs_finish_ioend(ioend); - /* ensure we don't spin on blocked ioends */ - delay(1); - } else { - xfs_destroy_ioend(ioend); - } + xfs_destroy_ioend(ioend); } /* diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index 116dd5c3703..06e4caf3820 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -18,8 +18,6 @@ #ifndef __XFS_AOPS_H__ #define __XFS_AOPS_H__ -extern struct workqueue_struct *xfsdatad_workqueue; -extern struct workqueue_struct *xfsconvertd_workqueue; extern mempool_t *xfs_ioend_pool; /* diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 4dff85c7d7e..6819b5163e3 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -45,8 +45,6 @@ static kmem_zone_t *xfs_buf_zone; STATIC int xfsbufd(void *); static struct workqueue_struct *xfslogd_workqueue; -struct workqueue_struct *xfsdatad_workqueue; -struct workqueue_struct *xfsconvertd_workqueue; #ifdef XFS_BUF_LOCK_TRACKING # define XB_SET_OWNER(bp) ((bp)->b_last_holder = current->pid) @@ -1793,21 +1791,8 @@ xfs_buf_init(void) if (!xfslogd_workqueue) goto out_free_buf_zone; - xfsdatad_workqueue = alloc_workqueue("xfsdatad", WQ_MEM_RECLAIM, 1); - if (!xfsdatad_workqueue) - goto out_destroy_xfslogd_workqueue; - - xfsconvertd_workqueue = alloc_workqueue("xfsconvertd", - WQ_MEM_RECLAIM, 1); - if (!xfsconvertd_workqueue) - goto out_destroy_xfsdatad_workqueue; - return 0; - out_destroy_xfsdatad_workqueue: - destroy_workqueue(xfsdatad_workqueue); - out_destroy_xfslogd_workqueue: - destroy_workqueue(xfslogd_workqueue); out_free_buf_zone: kmem_zone_destroy(xfs_buf_zone); out: @@ -1817,8 +1802,6 @@ xfs_buf_init(void) void xfs_buf_terminate(void) { - destroy_workqueue(xfsconvertd_workqueue); - destroy_workqueue(xfsdatad_workqueue); destroy_workqueue(xfslogd_workqueue); kmem_zone_destroy(xfs_buf_zone); } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index c082e44dad2..9eba7388782 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -211,6 +211,9 @@ typedef struct xfs_mount { struct shrinker m_inode_shrink; /* inode reclaim shrinker */ int64_t m_low_space[XFS_LOWSP_MAX]; /* low free space thresholds */ + + struct workqueue_struct *m_data_workqueue; + struct workqueue_struct *m_unwritten_workqueue; } xfs_mount_t; /* diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 5e0d43f231a..c7f7bc2855a 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -759,6 +759,36 @@ xfs_setup_devices( return 0; } +STATIC int +xfs_init_mount_workqueues( + struct xfs_mount *mp) +{ + mp->m_data_workqueue = alloc_workqueue("xfs-data/%s", + WQ_MEM_RECLAIM, 0, mp->m_fsname); + if (!mp->m_data_workqueue) + goto out; + + mp->m_unwritten_workqueue = alloc_workqueue("xfs-conv/%s", + WQ_MEM_RECLAIM, 0, mp->m_fsname); + if (!mp->m_unwritten_workqueue) + goto out_destroy_data_iodone_queue; + + return 0; + +out_destroy_data_iodone_queue: + destroy_workqueue(mp->m_data_workqueue); +out: + return -ENOMEM; +} + +STATIC void +xfs_destroy_mount_workqueues( + struct xfs_mount *mp) +{ + destroy_workqueue(mp->m_data_workqueue); + destroy_workqueue(mp->m_unwritten_workqueue); +} + /* Catch misguided souls that try to use this interface on XFS */ STATIC struct inode * xfs_fs_alloc_inode( @@ -982,6 +1012,7 @@ xfs_fs_put_super( xfs_unmountfs(mp); xfs_freesb(mp); xfs_icsb_destroy_counters(mp); + xfs_destroy_mount_workqueues(mp); xfs_close_devices(mp); xfs_free_fsname(mp); kfree(mp); @@ -1308,10 +1339,14 @@ xfs_fs_fill_super( if (error) goto out_free_fsname; - error = xfs_icsb_init_counters(mp); + error = xfs_init_mount_workqueues(mp); if (error) goto out_close_devices; + error = xfs_icsb_init_counters(mp); + if (error) + goto out_destroy_workqueues; + error = xfs_readsb(mp, flags); if (error) goto out_destroy_counters; @@ -1374,6 +1409,8 @@ xfs_fs_fill_super( xfs_freesb(mp); out_destroy_counters: xfs_icsb_destroy_counters(mp); +out_destroy_workqueues: + xfs_destroy_mount_workqueues(mp); out_close_devices: xfs_close_devices(mp); out_free_fsname: -- cgit v1.2.3-70-g09d2 From 8a9c9980f24f6d86e0ec0150ed35fba45d0c9f88 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 29 Feb 2012 09:53:52 +0000 Subject: xfs: log timestamp updates Timestamps on regular files are the last metadata that XFS does not update transactionally. Now that we use the delaylog mode exclusively and made the log scode scale extremly well there is no need to bypass that code for timestamp updates. Logging all updates allows to drop a lot of code, and will allow for further performance improvements later on. Note that this patch drops optimized handling of fdatasync - it will be added back in a separate commit. Reviewed-by: Dave Chinner Signed-off-by: Christoph Hellwig Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_file.c | 83 ++++++------------------------------ fs/xfs/xfs_iget.c | 1 - fs/xfs/xfs_inode.c | 25 +---------- fs/xfs/xfs_inode.h | 5 --- fs/xfs/xfs_inode_item.c | 36 ---------------- fs/xfs/xfs_inode_item.h | 5 +-- fs/xfs/xfs_iops.c | 58 ------------------------- fs/xfs/xfs_itable.c | 21 +++------ fs/xfs/xfs_super.c | 108 ++++++++++++++++------------------------------- fs/xfs/xfs_sync.c | 36 ---------------- fs/xfs/xfs_sync.h | 2 - fs/xfs/xfs_trace.h | 2 +- fs/xfs/xfs_trans_inode.c | 4 ++ 13 files changed, 65 insertions(+), 321 deletions(-) (limited to 'fs/xfs/xfs_super.c') diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 7e5bc872f2b..78d8b029959 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -163,7 +163,6 @@ xfs_file_fsync( struct inode *inode = file->f_mapping->host; struct xfs_inode *ip = XFS_I(inode); struct xfs_mount *mp = ip->i_mount; - struct xfs_trans *tp; int error = 0; int log_flushed = 0; xfs_lsn_t lsn = 0; @@ -194,75 +193,15 @@ xfs_file_fsync( } /* - * We always need to make sure that the required inode state is safe on - * disk. The inode might be clean but we still might need to force the - * log because of committed transactions that haven't hit the disk yet. - * Likewise, there could be unflushed non-transactional changes to the - * inode core that have to go to disk and this requires us to issue - * a synchronous transaction to capture these changes correctly. - * - * This code relies on the assumption that if the i_update_core field - * of the inode is clear and the inode is unpinned then it is clean - * and no action is required. + * All metadata updates are logged, which means that we just have + * to flush the log up to the latest LSN that touched the inode. */ xfs_ilock(ip, XFS_ILOCK_SHARED); - - /* - * First check if the VFS inode is marked dirty. All the dirtying - * of non-transactional updates do not go through mark_inode_dirty*, - * which allows us to distinguish between pure timestamp updates - * and i_size updates which need to be caught for fdatasync. - * After that also check for the dirty state in the XFS inode, which - * might gets cleared when the inode gets written out via the AIL - * or xfs_iflush_cluster. - */ - if (((inode->i_state & I_DIRTY_DATASYNC) || - ((inode->i_state & I_DIRTY_SYNC) && !datasync)) && - ip->i_update_core) { - /* - * Kick off a transaction to log the inode core to get the - * updates. The sync transaction will also force the log. - */ - xfs_iunlock(ip, XFS_ILOCK_SHARED); - tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); - error = xfs_trans_reserve(tp, 0, - XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0); - if (error) { - xfs_trans_cancel(tp, 0); - return -error; - } - xfs_ilock(ip, XFS_ILOCK_EXCL); - - /* - * Note - it's possible that we might have pushed ourselves out - * of the way during trans_reserve which would flush the inode. - * But there's no guarantee that the inode buffer has actually - * gone out yet (it's delwri). Plus the buffer could be pinned - * anyway if it's part of an inode in another recent - * transaction. So we play it safe and fire off the - * transaction anyway. - */ - xfs_trans_ijoin(tp, ip, 0); - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - error = xfs_trans_commit(tp, 0); - + if (xfs_ipincount(ip)) lsn = ip->i_itemp->ili_last_lsn; - xfs_iunlock(ip, XFS_ILOCK_EXCL); - } else { - /* - * Timestamps/size haven't changed since last inode flush or - * inode transaction commit. That means either nothing got - * written or a transaction committed which caught the updates. - * If the latter happened and the transaction hasn't hit the - * disk yet, the inode will be still be pinned. If it is, - * force the log. - */ - if (xfs_ipincount(ip)) - lsn = ip->i_itemp->ili_last_lsn; - xfs_iunlock(ip, XFS_ILOCK_SHARED); - } + xfs_iunlock(ip, XFS_ILOCK_SHARED); - if (!error && lsn) + if (lsn) error = _xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, &log_flushed); /* @@ -659,9 +598,6 @@ restart: return error; } - if (likely(!(file->f_mode & FMODE_NOCMTIME))) - file_update_time(file); - /* * If the offset is beyond the size of the file, we need to zero any * blocks that fall between the existing EOF and the start of this @@ -684,6 +620,15 @@ restart: if (error) return error; + /* + * Updating the timestamps will grab the ilock again from + * xfs_fs_dirty_inode, so we have to call it after dropping the + * lock above. Eventually we should look into a way to avoid + * the pointless lock roundtrip. + */ + if (likely(!(file->f_mode & FMODE_NOCMTIME))) + file_update_time(file); + /* * If we're writing the file then make sure to clear the setuid and * setgid bits if the process is not being run by root. This keeps diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c index 37f22dad5f5..af3f30a3d9c 100644 --- a/fs/xfs/xfs_iget.c +++ b/fs/xfs/xfs_iget.c @@ -91,7 +91,6 @@ xfs_inode_alloc( ip->i_afp = NULL; memset(&ip->i_df, 0, sizeof(xfs_ifork_t)); ip->i_flags = 0; - ip->i_update_core = 0; ip->i_delayed_blks = 0; memset(&ip->i_d, 0, sizeof(xfs_icdinode_t)); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index b21022499c2..7ce9ccbf17c 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1656,7 +1656,6 @@ retry: iip = ip->i_itemp; if (!iip || xfs_inode_clean(ip)) { ASSERT(ip != free_ip); - ip->i_update_core = 0; xfs_ifunlock(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); continue; @@ -2451,7 +2450,6 @@ xfs_iflush( * to disk, because the log record didn't make it to disk! */ if (XFS_FORCED_SHUTDOWN(mp)) { - ip->i_update_core = 0; if (iip) iip->ili_format.ilf_fields = 0; xfs_ifunlock(ip); @@ -2533,26 +2531,6 @@ xfs_iflush_int( /* set *dip = inode's place in the buffer */ dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset); - /* - * Clear i_update_core before copying out the data. - * This is for coordination with our timestamp updates - * that don't hold the inode lock. They will always - * update the timestamps BEFORE setting i_update_core, - * so if we clear i_update_core after they set it we - * are guaranteed to see their updates to the timestamps. - * I believe that this depends on strongly ordered memory - * semantics, but we have that. We use the SYNCHRONIZE - * macro to make sure that the compiler does not reorder - * the i_update_core access below the data copy below. - */ - ip->i_update_core = 0; - SYNCHRONIZE(); - - /* - * Make sure to get the latest timestamps from the Linux inode. - */ - xfs_synchronize_times(ip); - if (XFS_TEST_ERROR(dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC), mp, XFS_ERRTAG_IFLUSH_1, XFS_RANDOM_IFLUSH_1)) { xfs_alert_tag(mp, XFS_PTAG_IFLUSH, @@ -2711,8 +2689,7 @@ xfs_iflush_int( } else { /* * We're flushing an inode which is not in the AIL and has - * not been logged but has i_update_core set. For this - * case we can use a B_DELWRI flush and immediately drop + * not been logged. For this case we can immediately drop * the inode flush lock because we can avoid the whole * AIL state thing. It's OK to drop the flush lock now, * because we've already locked the buffer and to do anything diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 7f90469141d..f123dbe6d42 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -241,7 +241,6 @@ typedef struct xfs_inode { spinlock_t i_flags_lock; /* inode i_flags lock */ /* Miscellaneous state. */ unsigned long i_flags; /* see defined flags below */ - unsigned char i_update_core; /* timestamps/size is dirty */ unsigned int i_delayed_blks; /* count of delay alloc blks */ xfs_icdinode_t i_d; /* most of ondisk inode */ @@ -534,10 +533,6 @@ void xfs_promote_inode(struct xfs_inode *); void xfs_lock_inodes(xfs_inode_t **, int, uint); void xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint); -void xfs_synchronize_times(xfs_inode_t *); -void xfs_mark_inode_dirty(xfs_inode_t *); -void xfs_mark_inode_dirty_sync(xfs_inode_t *); - #define IHOLD(ip) \ do { \ ASSERT(atomic_read(&VFS_I(ip)->i_count) > 0) ; \ diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index adc8a261b5d..7a60da64f31 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -254,42 +254,6 @@ xfs_inode_item_format( vecp++; nvecs = 1; - /* - * Clear i_update_core if the timestamps (or any other - * non-transactional modification) need flushing/logging - * and we're about to log them with the rest of the core. - * - * This is the same logic as xfs_iflush() but this code can't - * run at the same time as xfs_iflush because we're in commit - * processing here and so we have the inode lock held in - * exclusive mode. Although it doesn't really matter - * for the timestamps if both routines were to grab the - * timestamps or not. That would be ok. - * - * We clear i_update_core before copying out the data. - * This is for coordination with our timestamp updates - * that don't hold the inode lock. They will always - * update the timestamps BEFORE setting i_update_core, - * so if we clear i_update_core after they set it we - * are guaranteed to see their updates to the timestamps - * either here. Likewise, if they set it after we clear it - * here, we'll see it either on the next commit of this - * inode or the next time the inode gets flushed via - * xfs_iflush(). This depends on strongly ordered memory - * semantics, but we have that. We use the SYNCHRONIZE - * macro to make sure that the compiler does not reorder - * the i_update_core access below the data copy below. - */ - if (ip->i_update_core) { - ip->i_update_core = 0; - SYNCHRONIZE(); - } - - /* - * Make sure to get the latest timestamps from the Linux inode. - */ - xfs_synchronize_times(ip); - vecp->i_addr = &ip->i_d; vecp->i_len = sizeof(struct xfs_icdinode); vecp->i_type = XLOG_REG_TYPE_ICORE; diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h index d3dee61e6d9..25784b06656 100644 --- a/fs/xfs/xfs_inode_item.h +++ b/fs/xfs/xfs_inode_item.h @@ -148,9 +148,8 @@ typedef struct xfs_inode_log_item { static inline int xfs_inode_clean(xfs_inode_t *ip) { - return (!ip->i_itemp || - !(ip->i_itemp->ili_format.ilf_fields & XFS_ILOG_ALL)) && - !ip->i_update_core; + return !ip->i_itemp || + !(ip->i_itemp->ili_format.ilf_fields & XFS_ILOG_ALL); } extern void xfs_inode_item_init(struct xfs_inode *, struct xfs_mount *); diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index ab302539e5b..7c01cda1672 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -50,59 +50,6 @@ #include #include -/* - * Bring the timestamps in the XFS inode uptodate. - * - * Used before writing the inode to disk. - */ -void -xfs_synchronize_times( - xfs_inode_t *ip) -{ - struct inode *inode = VFS_I(ip); - - ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec; - ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec; - ip->i_d.di_ctime.t_sec = (__int32_t)inode->i_ctime.tv_sec; - ip->i_d.di_ctime.t_nsec = (__int32_t)inode->i_ctime.tv_nsec; - ip->i_d.di_mtime.t_sec = (__int32_t)inode->i_mtime.tv_sec; - ip->i_d.di_mtime.t_nsec = (__int32_t)inode->i_mtime.tv_nsec; -} - -/* - * If the linux inode is valid, mark it dirty, else mark the dirty state - * in the XFS inode to make sure we pick it up when reclaiming the inode. - */ -void -xfs_mark_inode_dirty_sync( - xfs_inode_t *ip) -{ - struct inode *inode = VFS_I(ip); - - if (!(inode->i_state & (I_WILL_FREE|I_FREEING))) - mark_inode_dirty_sync(inode); - else { - barrier(); - ip->i_update_core = 1; - } -} - -void -xfs_mark_inode_dirty( - xfs_inode_t *ip) -{ - struct inode *inode = VFS_I(ip); - - if (!(inode->i_state & (I_WILL_FREE|I_FREEING))) - mark_inode_dirty(inode); - else { - barrier(); - ip->i_update_core = 1; - } - -} - - int xfs_initxattrs(struct inode *inode, const struct xattr *xattr_array, void *fs_info) { @@ -678,19 +625,16 @@ xfs_setattr_nonsize( inode->i_atime = iattr->ia_atime; ip->i_d.di_atime.t_sec = iattr->ia_atime.tv_sec; ip->i_d.di_atime.t_nsec = iattr->ia_atime.tv_nsec; - ip->i_update_core = 1; } if (mask & ATTR_CTIME) { inode->i_ctime = iattr->ia_ctime; ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec; ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec; - ip->i_update_core = 1; } if (mask & ATTR_MTIME) { inode->i_mtime = iattr->ia_mtime; ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec; ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec; - ip->i_update_core = 1; } xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); @@ -918,13 +862,11 @@ xfs_setattr_size( inode->i_ctime = iattr->ia_ctime; ip->i_d.di_ctime.t_sec = iattr->ia_ctime.tv_sec; ip->i_d.di_ctime.t_nsec = iattr->ia_ctime.tv_nsec; - ip->i_update_core = 1; } if (mask & ATTR_MTIME) { inode->i_mtime = iattr->ia_mtime; ip->i_d.di_mtime.t_sec = iattr->ia_mtime.tv_sec; ip->i_d.di_mtime.t_nsec = iattr->ia_mtime.tv_nsec; - ip->i_update_core = 1; } xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 751e94fe1f7..9720c54bbed 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -62,7 +62,6 @@ xfs_bulkstat_one_int( { struct xfs_icdinode *dic; /* dinode core info pointer */ struct xfs_inode *ip; /* incore inode pointer */ - struct inode *inode; struct xfs_bstat *buf; /* return buffer */ int error = 0; /* error value */ @@ -86,7 +85,6 @@ xfs_bulkstat_one_int( ASSERT(ip->i_imap.im_blkno != 0); dic = &ip->i_d; - inode = VFS_I(ip); /* xfs_iget returns the following without needing * further change. @@ -99,19 +97,12 @@ xfs_bulkstat_one_int( buf->bs_uid = dic->di_uid; buf->bs_gid = dic->di_gid; buf->bs_size = dic->di_size; - - /* - * We need to read the timestamps from the Linux inode because - * the VFS keeps writing directly into the inode structure instead - * of telling us about the updates. - */ - buf->bs_atime.tv_sec = inode->i_atime.tv_sec; - buf->bs_atime.tv_nsec = inode->i_atime.tv_nsec; - buf->bs_mtime.tv_sec = inode->i_mtime.tv_sec; - buf->bs_mtime.tv_nsec = inode->i_mtime.tv_nsec; - buf->bs_ctime.tv_sec = inode->i_ctime.tv_sec; - buf->bs_ctime.tv_nsec = inode->i_ctime.tv_nsec; - + buf->bs_atime.tv_sec = dic->di_atime.t_sec; + buf->bs_atime.tv_nsec = dic->di_atime.t_nsec; + buf->bs_mtime.tv_sec = dic->di_mtime.t_sec; + buf->bs_mtime.tv_nsec = dic->di_mtime.t_nsec; + buf->bs_ctime.tv_sec = dic->di_ctime.t_sec; + buf->bs_ctime.tv_nsec = dic->di_ctime.t_nsec; buf->bs_xflags = xfs_ip2xflags(ip); buf->bs_extsize = dic->di_extsize << mp->m_sb.sb_blocklog; buf->bs_extents = dic->di_nextents; diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index c7f7bc2855a..e602c8c67c5 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -863,91 +863,58 @@ xfs_fs_inode_init_once( } /* - * Dirty the XFS inode when mark_inode_dirty_sync() is called so that - * we catch unlogged VFS level updates to the inode. + * This is called by the VFS when dirtying inode metadata. This can happen + * for a few reasons, but we only care about timestamp updates, given that + * we handled the rest ourselves. In theory no other calls should happen, + * but for example generic_write_end() keeps dirtying the inode after + * updating i_size. Thus we check that the flags are exactly I_DIRTY_SYNC, + * and skip this call otherwise. * - * We need the barrier() to maintain correct ordering between unlogged - * updates and the transaction commit code that clears the i_update_core - * field. This requires all updates to be completed before marking the - * inode dirty. + * We'll hopefull get a different method just for updating timestamps soon, + * at which point this hack can go away, and maybe we'll also get real + * error handling here. */ STATIC void xfs_fs_dirty_inode( - struct inode *inode, - int flags) -{ - barrier(); - XFS_I(inode)->i_update_core = 1; -} - -STATIC int -xfs_fs_write_inode( struct inode *inode, - struct writeback_control *wbc) + int flags) { struct xfs_inode *ip = XFS_I(inode); struct xfs_mount *mp = ip->i_mount; - int error = EAGAIN; - - trace_xfs_write_inode(ip); - - if (XFS_FORCED_SHUTDOWN(mp)) - return -XFS_ERROR(EIO); - - if (wbc->sync_mode == WB_SYNC_ALL || wbc->for_kupdate) { - /* - * Make sure the inode has made it it into the log. Instead - * of forcing it all the way to stable storage using a - * synchronous transaction we let the log force inside the - * ->sync_fs call do that for thus, which reduces the number - * of synchronous log forces dramatically. - */ - error = xfs_log_dirty_inode(ip, NULL, 0); - if (error) - goto out; - return 0; - } else { - if (!ip->i_update_core) - return 0; + struct xfs_trans *tp; + int error; - /* - * We make this non-blocking if the inode is contended, return - * EAGAIN to indicate to the caller that they did not succeed. - * This prevents the flush path from blocking on inodes inside - * another operation right now, they get caught later by - * xfs_sync. - */ - if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) - goto out; + if (flags != I_DIRTY_SYNC) + return; - if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) - goto out_unlock; + trace_xfs_dirty_inode(ip); - /* - * Now we have the flush lock and the inode is not pinned, we - * can check if the inode is really clean as we know that - * there are no pending transaction completions, it is not - * waiting on the delayed write queue and there is no IO in - * progress. - */ - if (xfs_inode_clean(ip)) { - xfs_ifunlock(ip); - error = 0; - goto out_unlock; - } - error = xfs_iflush(ip, SYNC_TRYLOCK); + tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); + error = xfs_trans_reserve(tp, 0, XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0); + if (error) { + xfs_trans_cancel(tp, 0); + goto trouble; } - - out_unlock: - xfs_iunlock(ip, XFS_ILOCK_SHARED); - out: + xfs_ilock(ip, XFS_ILOCK_EXCL); /* - * if we failed to write out the inode then mark - * it dirty again so we'll try again later. + * Grab all the latest timestamps from the Linux inode. */ + ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec; + ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec; + ip->i_d.di_ctime.t_sec = (__int32_t)inode->i_ctime.tv_sec; + ip->i_d.di_ctime.t_nsec = (__int32_t)inode->i_ctime.tv_nsec; + ip->i_d.di_mtime.t_sec = (__int32_t)inode->i_mtime.tv_sec; + ip->i_d.di_mtime.t_nsec = (__int32_t)inode->i_mtime.tv_nsec; + + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + error = xfs_trans_commit(tp, 0); if (error) - xfs_mark_inode_dirty_sync(ip); - return -error; + goto trouble; + return; + +trouble: + xfs_warn(mp, "failed to update timestamps for inode 0x%llx", ip->i_ino); } STATIC void @@ -1466,7 +1433,6 @@ static const struct super_operations xfs_super_operations = { .alloc_inode = xfs_fs_alloc_inode, .destroy_inode = xfs_fs_destroy_inode, .dirty_inode = xfs_fs_dirty_inode, - .write_inode = xfs_fs_write_inode, .evict_inode = xfs_fs_evict_inode, .put_super = xfs_fs_put_super, .sync_fs = xfs_fs_sync_fs, diff --git a/fs/xfs/xfs_sync.c b/fs/xfs/xfs_sync.c index 71bf846b728..205ebcb34d9 100644 --- a/fs/xfs/xfs_sync.c +++ b/fs/xfs/xfs_sync.c @@ -336,32 +336,6 @@ xfs_sync_fsdata( return error; } -int -xfs_log_dirty_inode( - struct xfs_inode *ip, - struct xfs_perag *pag, - int flags) -{ - struct xfs_mount *mp = ip->i_mount; - struct xfs_trans *tp; - int error; - - if (!ip->i_update_core) - return 0; - - tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); - error = xfs_trans_reserve(tp, 0, XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0); - if (error) { - xfs_trans_cancel(tp, 0); - return error; - } - - xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - return xfs_trans_commit(tp, 0); -} - /* * When remounting a filesystem read-only or freezing the filesystem, we have * two phases to execute. This first phase is syncing the data before we @@ -385,16 +359,6 @@ xfs_quiesce_data( { int error, error2 = 0; - /* - * Log all pending size and timestamp updates. The vfs writeback - * code is supposed to do this, but due to its overagressive - * livelock detection it will skip inodes where appending writes - * were written out in the first non-blocking sync phase if their - * completion took long enough that it happened after taking the - * timestamp for the cut-off in the blocking phase. - */ - xfs_inode_ag_iterator(mp, xfs_log_dirty_inode, 0); - /* force out the log */ xfs_log_force(mp, XFS_LOG_SYNC); diff --git a/fs/xfs/xfs_sync.h b/fs/xfs/xfs_sync.h index fa965479d78..941202e7ac6 100644 --- a/fs/xfs/xfs_sync.h +++ b/fs/xfs/xfs_sync.h @@ -34,8 +34,6 @@ void xfs_quiesce_attr(struct xfs_mount *mp); void xfs_flush_inodes(struct xfs_inode *ip); -int xfs_log_dirty_inode(struct xfs_inode *ip, struct xfs_perag *pag, int flags); - int xfs_reclaim_inodes(struct xfs_mount *mp, int mode); int xfs_reclaim_inodes_count(struct xfs_mount *mp); void xfs_reclaim_inodes_nr(struct xfs_mount *mp, int nr_to_scan); diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 3b369c1277f..ceaf6fe67e4 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -580,7 +580,7 @@ DEFINE_INODE_EVENT(xfs_ioctl_setattr); DEFINE_INODE_EVENT(xfs_dir_fsync); DEFINE_INODE_EVENT(xfs_file_fsync); DEFINE_INODE_EVENT(xfs_destroy_inode); -DEFINE_INODE_EVENT(xfs_write_inode); +DEFINE_INODE_EVENT(xfs_dirty_inode); DEFINE_INODE_EVENT(xfs_evict_inode); DEFINE_INODE_EVENT(xfs_dquot_dqalloc); diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c index 32f0288ae10..892763effdf 100644 --- a/fs/xfs/xfs_trans_inode.c +++ b/fs/xfs/xfs_trans_inode.c @@ -95,10 +95,14 @@ xfs_trans_ichgtime( if ((flags & XFS_ICHGTIME_MOD) && !timespec_equal(&inode->i_mtime, &tv)) { inode->i_mtime = tv; + ip->i_d.di_mtime.t_sec = tv.tv_sec; + ip->i_d.di_mtime.t_nsec = tv.tv_nsec; } if ((flags & XFS_ICHGTIME_CHG) && !timespec_equal(&inode->i_ctime, &tv)) { inode->i_ctime = tv; + ip->i_d.di_ctime.t_sec = tv.tv_sec; + ip->i_d.di_ctime.t_nsec = tv.tv_nsec; } } -- cgit v1.2.3-70-g09d2 From 8f639ddea0c4978ae9b4e46ea041c9e5afe0ee8d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 29 Feb 2012 09:53:55 +0000 Subject: xfs: reimplement fdatasync support Add an in-memory only flag to say we logged timestamps only, and use it to check if fdatasync can optimize away the log force. Reviewed-by: Dave Chinner Signed-off-by: Christoph Hellwig Reviewed-by: Mark Tinguely Signed-off-by: Ben Myers --- fs/xfs/xfs_file.c | 7 +++++-- fs/xfs/xfs_inode_item.c | 3 ++- fs/xfs/xfs_inode_item.h | 11 ++++++++++- fs/xfs/xfs_super.c | 2 +- 4 files changed, 18 insertions(+), 5 deletions(-) (limited to 'fs/xfs/xfs_super.c') diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 78d8b029959..54a67dd9ac0 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -197,8 +197,11 @@ xfs_file_fsync( * to flush the log up to the latest LSN that touched the inode. */ xfs_ilock(ip, XFS_ILOCK_SHARED); - if (xfs_ipincount(ip)) - lsn = ip->i_itemp->ili_last_lsn; + if (xfs_ipincount(ip)) { + if (!datasync || + (ip->i_itemp->ili_fields & ~XFS_ILOG_TIMESTAMP)) + lsn = ip->i_itemp->ili_last_lsn; + } xfs_iunlock(ip, XFS_ILOCK_SHARED); if (lsn) diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 8becef4f9e6..05d924efcea 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -438,7 +438,8 @@ out: * games in recovery easier, which isn't a big deal as just about any * transaction would dirty it anyway. */ - iip->ili_format.ilf_fields = XFS_ILOG_CORE | iip->ili_fields; + iip->ili_format.ilf_fields = XFS_ILOG_CORE | + (iip->ili_fields & ~XFS_ILOG_TIMESTAMP); iip->ili_format.ilf_size = nvecs; } diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h index bc183d81ad3..41d61c3b7a3 100644 --- a/fs/xfs/xfs_inode_item.h +++ b/fs/xfs/xfs_inode_item.h @@ -86,6 +86,15 @@ typedef struct xfs_inode_log_format_64 { #define XFS_ILOG_AEXT 0x080 /* log i_af.if_extents */ #define XFS_ILOG_ABROOT 0x100 /* log i_af.i_broot */ + +/* + * The timestamps are dirty, but not necessarily anything else in the inode + * core. Unlike the other fields above this one must never make it to disk + * in the ilf_fields of the inode_log_format, but is purely store in-memory in + * ili_fields in the inode_log_item. + */ +#define XFS_ILOG_TIMESTAMP 0x4000 + #define XFS_ILOG_NONCORE (XFS_ILOG_DDATA | XFS_ILOG_DEXT | \ XFS_ILOG_DBROOT | XFS_ILOG_DEV | \ XFS_ILOG_UUID | XFS_ILOG_ADATA | \ @@ -101,7 +110,7 @@ typedef struct xfs_inode_log_format_64 { XFS_ILOG_DEXT | XFS_ILOG_DBROOT | \ XFS_ILOG_DEV | XFS_ILOG_UUID | \ XFS_ILOG_ADATA | XFS_ILOG_AEXT | \ - XFS_ILOG_ABROOT) + XFS_ILOG_ABROOT | XFS_ILOG_TIMESTAMP) static inline int xfs_ilog_fbroot(int w) { diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index e602c8c67c5..e9ad7894648 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -907,7 +907,7 @@ xfs_fs_dirty_inode( ip->i_d.di_mtime.t_nsec = (__int32_t)inode->i_mtime.tv_nsec; xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + xfs_trans_log_inode(tp, ip, XFS_ILOG_TIMESTAMP); error = xfs_trans_commit(tp, 0); if (error) goto trouble; -- cgit v1.2.3-70-g09d2 From a05931ceb0160deadbd7798d60d01b17f2d81b09 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 13 Mar 2012 08:52:37 +0000 Subject: xfs: remove the global xfs_Gqm structure If we initialize the slab caches for the quota code when XFS is loaded there is no need for a global and reference counted quota manager structure. Drop all this overhead and also fix the error handling during quota initialization. Reviewed-by: Dave Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Ben Myers --- fs/xfs/xfs_dquot.c | 35 ++++++++++++- fs/xfs/xfs_qm.c | 132 ----------------------------------------------- fs/xfs/xfs_qm.h | 15 +----- fs/xfs/xfs_qm_bhv.c | 16 ------ fs/xfs/xfs_super.c | 10 ++-- fs/xfs/xfs_super.h | 8 ++- fs/xfs/xfs_trans_dquot.c | 4 +- 7 files changed, 46 insertions(+), 174 deletions(-) (limited to 'fs/xfs/xfs_super.c') diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 2896ac953ed..4be16a0cbe5 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -59,6 +59,9 @@ int xfs_dqreq_num; int xfs_dqerror_mod = 33; #endif +struct kmem_zone *xfs_qm_dqtrxzone; +static struct kmem_zone *xfs_qm_dqzone; + static struct lock_class_key xfs_dquot_other_class; /* @@ -71,7 +74,7 @@ xfs_qm_dqdestroy( ASSERT(list_empty(&dqp->q_lru)); mutex_destroy(&dqp->q_qlock); - kmem_zone_free(xfs_Gqm->qm_dqzone, dqp); + kmem_zone_free(xfs_qm_dqzone, dqp); XFS_STATS_DEC(xs_qm_dquot); } @@ -491,7 +494,7 @@ xfs_qm_dqread( int cancelflags = 0; - dqp = kmem_zone_zalloc(xfs_Gqm->qm_dqzone, KM_SLEEP); + dqp = kmem_zone_zalloc(xfs_qm_dqzone, KM_SLEEP); dqp->dq_flags = type; dqp->q_core.d_id = cpu_to_be32(id); @@ -1040,3 +1043,31 @@ xfs_dqflock_pushbuf_wait( out_lock: xfs_dqflock(dqp); } + +int __init +xfs_qm_init(void) +{ + xfs_qm_dqzone = + kmem_zone_init(sizeof(struct xfs_dquot), "xfs_dquot"); + if (!xfs_qm_dqzone) + goto out; + + xfs_qm_dqtrxzone = + kmem_zone_init(sizeof(struct xfs_dquot_acct), "xfs_dqtrx"); + if (!xfs_qm_dqtrxzone) + goto out_free_dqzone; + + return 0; + +out_free_dqzone: + kmem_zone_destroy(xfs_qm_dqzone); +out: + return -ENOMEM; +} + +void __exit +xfs_qm_exit(void) +{ + kmem_zone_destroy(xfs_qm_dqtrxzone); + kmem_zone_destroy(xfs_qm_dqzone); +} diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 2f92d3b0d8a..55c6afedc87 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -48,126 +48,10 @@ * quota functionality, including maintaining the freelist and hash * tables of dquots. */ -struct mutex xfs_Gqm_lock; -struct xfs_qm *xfs_Gqm; - -kmem_zone_t *qm_dqzone; -kmem_zone_t *qm_dqtrxzone; - STATIC int xfs_qm_init_quotainos(xfs_mount_t *); STATIC int xfs_qm_init_quotainfo(xfs_mount_t *); STATIC int xfs_qm_shake(struct shrinker *, struct shrink_control *); -/* - * Initialize the XQM structure. - * Note that there is not one quota manager per file system. - */ -STATIC struct xfs_qm * -xfs_Gqm_init(void) -{ - xfs_qm_t *xqm; - - xqm = kmem_zalloc(sizeof(xfs_qm_t), KM_SLEEP); - - /* - * dquot zone. we register our own low-memory callback. - */ - if (!qm_dqzone) { - xqm->qm_dqzone = kmem_zone_init(sizeof(xfs_dquot_t), - "xfs_dquots"); - qm_dqzone = xqm->qm_dqzone; - } else - xqm->qm_dqzone = qm_dqzone; - - /* - * The t_dqinfo portion of transactions. - */ - if (!qm_dqtrxzone) { - xqm->qm_dqtrxzone = kmem_zone_init(sizeof(xfs_dquot_acct_t), - "xfs_dqtrx"); - qm_dqtrxzone = xqm->qm_dqtrxzone; - } else - xqm->qm_dqtrxzone = qm_dqtrxzone; - - xqm->qm_nrefs = 0; - return xqm; -} - -/* - * Destroy the global quota manager when its reference count goes to zero. - */ -STATIC void -xfs_qm_destroy( - struct xfs_qm *xqm) -{ - ASSERT(xqm != NULL); - ASSERT(xqm->qm_nrefs == 0); - - kmem_free(xqm); -} - -/* - * Called at mount time to let XQM know that another file system is - * starting quotas. This isn't crucial information as the individual mount - * structures are pretty independent, but it helps the XQM keep a - * global view of what's going on. - */ -/* ARGSUSED */ -STATIC int -xfs_qm_hold_quotafs_ref( - struct xfs_mount *mp) -{ - /* - * Need to lock the xfs_Gqm structure for things like this. For example, - * the structure could disappear between the entry to this routine and - * a HOLD operation if not locked. - */ - mutex_lock(&xfs_Gqm_lock); - - if (!xfs_Gqm) { - xfs_Gqm = xfs_Gqm_init(); - if (!xfs_Gqm) { - mutex_unlock(&xfs_Gqm_lock); - return ENOMEM; - } - } - - /* - * We can keep a list of all filesystems with quotas mounted for - * debugging and statistical purposes, but ... - * Just take a reference and get out. - */ - xfs_Gqm->qm_nrefs++; - mutex_unlock(&xfs_Gqm_lock); - - return 0; -} - - -/* - * Release the reference that a filesystem took at mount time, - * so that we know when we need to destroy the entire quota manager. - */ -/* ARGSUSED */ -STATIC void -xfs_qm_rele_quotafs_ref( - struct xfs_mount *mp) -{ - ASSERT(xfs_Gqm); - ASSERT(xfs_Gqm->qm_nrefs > 0); - - /* - * Destroy the entire XQM. If somebody mounts with quotaon, this'll - * be restarted. - */ - mutex_lock(&xfs_Gqm_lock); - if (--xfs_Gqm->qm_nrefs == 0) { - xfs_qm_destroy(xfs_Gqm); - xfs_Gqm = NULL; - } - mutex_unlock(&xfs_Gqm_lock); -} - /* * We use the batch lookup interface to iterate over the dquots as it * currently is the only interface into the radix tree code that allows @@ -738,13 +622,6 @@ xfs_qm_init_quotainfo( ASSERT(XFS_IS_QUOTA_RUNNING(mp)); - /* - * Tell XQM that we exist as soon as possible. - */ - if ((error = xfs_qm_hold_quotafs_ref(mp))) { - return error; - } - qinf = mp->m_quotainfo = kmem_zalloc(sizeof(xfs_quotainfo_t), KM_SLEEP); /* @@ -850,17 +727,9 @@ xfs_qm_destroy_quotainfo( qi = mp->m_quotainfo; ASSERT(qi != NULL); - ASSERT(xfs_Gqm != NULL); unregister_shrinker(&qi->qi_shrinker); - /* - * Release the reference that XQM kept, so that we know - * when the XQM structure should be freed. We cannot assume - * that xfs_Gqm is non-null after this point. - */ - xfs_qm_rele_quotafs_ref(mp); - if (qi->qi_uquotaip) { IRELE(qi->qi_uquotaip); qi->qi_uquotaip = NULL; /* paranoia */ @@ -1447,7 +1316,6 @@ xfs_qm_quotacheck( * We must turn off quotas. */ ASSERT(mp->m_quotainfo != NULL); - ASSERT(xfs_Gqm != NULL); xfs_qm_destroy_quotainfo(mp); if (xfs_mount_reset_sbqflags(mp)) { xfs_warn(mp, diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h index 76447060cd4..44b858b79d7 100644 --- a/fs/xfs/xfs_qm.h +++ b/fs/xfs/xfs_qm.h @@ -22,13 +22,9 @@ #include "xfs_dquot.h" #include "xfs_quota_priv.h" -struct xfs_qm; struct xfs_inode; -extern struct mutex xfs_Gqm_lock; -extern struct xfs_qm *xfs_Gqm; -extern kmem_zone_t *qm_dqzone; -extern kmem_zone_t *qm_dqtrxzone; +extern struct kmem_zone *xfs_qm_dqtrxzone; /* * This defines the unit of allocation of dquots. @@ -41,15 +37,6 @@ extern kmem_zone_t *qm_dqtrxzone; */ #define XFS_DQUOT_CLUSTER_SIZE_FSB (xfs_filblks_t)1 -/* - * Quota Manager (global) structure. Lives only in core. - */ -typedef struct xfs_qm { - uint qm_nrefs; /* file systems with quota on */ - kmem_zone_t *qm_dqzone; /* dquot mem-alloc zone */ - kmem_zone_t *qm_dqtrxzone; /* t_dqinfo of transactions */ -} xfs_qm_t; - /* * Various quota information for individual filesystems. * The mount structure keeps a pointer to this. diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c index 809f86857c6..e6986b5d80d 100644 --- a/fs/xfs/xfs_qm_bhv.c +++ b/fs/xfs/xfs_qm_bhv.c @@ -156,19 +156,3 @@ xfs_qm_newmount( return 0; } - -void __init -xfs_qm_init(void) -{ - printk(KERN_INFO "SGI XFS Quota Management subsystem\n"); - mutex_init(&xfs_Gqm_lock); -} - -void __exit -xfs_qm_exit(void) -{ - if (qm_dqzone) - kmem_zone_destroy(qm_dqzone); - if (qm_dqtrxzone) - kmem_zone_destroy(qm_dqtrxzone); -} diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index e9ad7894648..06d23b976f4 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1654,13 +1654,17 @@ init_xfs_fs(void) if (error) goto out_cleanup_procfs; - vfs_initquota(); + error = xfs_qm_init(); + if (error) + goto out_sysctl_unregister; error = register_filesystem(&xfs_fs_type); if (error) - goto out_sysctl_unregister; + goto out_qm_exit; return 0; + out_qm_exit: + xfs_qm_exit(); out_sysctl_unregister: xfs_sysctl_unregister(); out_cleanup_procfs: @@ -1682,7 +1686,7 @@ init_xfs_fs(void) STATIC void __exit exit_xfs_fs(void) { - vfs_exitquota(); + xfs_qm_exit(); unregister_filesystem(&xfs_fs_type); xfs_sysctl_unregister(); xfs_cleanup_procfs(); diff --git a/fs/xfs/xfs_super.h b/fs/xfs/xfs_super.h index 50a3266c999..09b0c26b224 100644 --- a/fs/xfs/xfs_super.h +++ b/fs/xfs/xfs_super.h @@ -21,13 +21,11 @@ #include #ifdef CONFIG_XFS_QUOTA -extern void xfs_qm_init(void); +extern int xfs_qm_init(void); extern void xfs_qm_exit(void); -# define vfs_initquota() xfs_qm_init() -# define vfs_exitquota() xfs_qm_exit() #else -# define vfs_initquota() do { } while (0) -# define vfs_exitquota() do { } while (0) +# define xfs_qm_init() (0) +# define xfs_qm_exit() do { } while (0) #endif #ifdef CONFIG_XFS_POSIX_ACL diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c index 877fe6367d2..279099717ed 100644 --- a/fs/xfs/xfs_trans_dquot.c +++ b/fs/xfs/xfs_trans_dquot.c @@ -875,7 +875,7 @@ STATIC void xfs_trans_alloc_dqinfo( xfs_trans_t *tp) { - tp->t_dqinfo = kmem_zone_zalloc(xfs_Gqm->qm_dqtrxzone, KM_SLEEP); + tp->t_dqinfo = kmem_zone_zalloc(xfs_qm_dqtrxzone, KM_SLEEP); } void @@ -884,6 +884,6 @@ xfs_trans_free_dqinfo( { if (!tp->t_dqinfo) return; - kmem_zone_free(xfs_Gqm->qm_dqtrxzone, tp->t_dqinfo); + kmem_zone_free(xfs_qm_dqtrxzone, tp->t_dqinfo); tp->t_dqinfo = NULL; } -- cgit v1.2.3-70-g09d2