From 859d718279b6e1d6bc27a701db47c1be720b5907 Mon Sep 17 00:00:00 2001 From: Vlad Apostolov Date: Thu, 11 Oct 2007 17:44:18 +1000 Subject: [PATCH] [XFS] get_bulkall() could return incorrect inode state In the following scenario xfs_bulkstat() returns incorrect stale inode state: 1. File_A is created and its inode synced to disk. 2. File_A is unlinked and doesn't exist anymore. 3. Filesystem sync is invoked. 4. File_B is created. File_B happens to reclaim File_A's inode. 5. xfs_bulkstat() is called and detects File_B but reports the incorrect File_A inode state. Explanation for the incorrect inode state is that inodes are not immediately synced on file create for performance reasons. This leaves the on-disk inode buffer uninitialized (or with old state from a previous generation inode) and this is what xfs_bulkstat() would report. The patch marks the on-disk inode buffer "dirty" on unlink. When the inode is reclaimed (by a new file create), xfs_bulkstat() would filter this inode by the "dirty" mark. Once the inode is flushed to disk, the on-disk buffer "dirty" mark is automatically removed and a following xfs_bulkstat() would return the correct inode state. Marking the on-disk inode buffer "dirty" on unlink is achieved by setting the on-disk di_nlink field to 0. Note that the in-core di_nlink has already been set to 0 and a corresponding transaction logged by xfs_droplink(). This is an exception from the rule that any on-disk inode buffer changes has to be followed by a disk write (inode flush). Synchronizing the in-core to on-disk di_nlink values in advance (before the actual inode flush to disk) should be fine in this case because the inode is already unlinked and it would never change its di_nlink again for this inode generation. SGI-PV: 970842 SGI-Modid: xfs-linux-melb:xfs-kern:29757a Signed-off-by: Vlad Apostolov Signed-off-by: Alex Elder Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Mark Goodwin Signed-off-by: Tim Shimmin --- fs/xfs/xfs_inode.c | 26 ++++++++++++++++++++------ fs/xfs/xfs_itable.c | 10 +++++++++- 2 files changed, 29 insertions(+), 7 deletions(-) diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 3d8ba8fec191..abf509a88915 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1931,9 +1931,9 @@ xfs_iunlink( */ error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, agdaddr, XFS_FSS_TO_BB(mp, 1), 0, &agibp); - if (error) { + if (error) return error; - } + /* * Validate the magic number of the agi block. */ @@ -1957,6 +1957,24 @@ xfs_iunlink( ASSERT(agi->agi_unlinked[bucket_index]); ASSERT(be32_to_cpu(agi->agi_unlinked[bucket_index]) != agino); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0); + if (error) + return error; + + /* + * Clear the on-disk di_nlink. This is to prevent xfs_bulkstat + * from picking up this inode when it is reclaimed (its incore state + * initialzed but not flushed to disk yet). The in-core di_nlink is + * already cleared in xfs_droplink() and a corresponding transaction + * logged. The hack here just synchronizes the in-core to on-disk + * di_nlink value in advance before the actual inode sync to disk. + * This is OK because the inode is already unlinked and would never + * change its di_nlink again for this inode generation. + * This is a temporary hack that would require a proper fix + * in the future. + */ + dip->di_core.di_nlink = 0; + if (be32_to_cpu(agi->agi_unlinked[bucket_index]) != NULLAGINO) { /* * There is already another inode in the bucket we need @@ -1964,10 +1982,6 @@ xfs_iunlink( * Here we put the head pointer into our next pointer, * and then we fall through to point the head at us. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0); - if (error) { - return error; - } ASSERT(be32_to_cpu(dip->di_next_unlinked) == NULLAGINO); /* both on-disk, don't endian flip twice */ dip->di_next_unlinked = agi->agi_unlinked[bucket_index]; diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index efeeafe275b9..1edd9afb664b 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -290,8 +290,16 @@ xfs_bulkstat_use_dinode( return 1; dip = (xfs_dinode_t *) xfs_buf_offset(bp, clustidx << mp->m_sb.sb_inodelog); + /* + * Check the buffer containing the on-disk inode for di_nlink == 0. + * This is to prevent xfs_bulkstat from picking up just reclaimed + * inodes that have their in-core state initialized but not flushed + * to disk yet. This is a temporary hack that would require a proper + * fix in the future. + */ if (be16_to_cpu(dip->di_core.di_magic) != XFS_DINODE_MAGIC || - !XFS_DINODE_GOOD_VERSION(dip->di_core.di_version)) + !XFS_DINODE_GOOD_VERSION(dip->di_core.di_version) || + !dip->di_core.di_nlink) return 0; if (flags & BULKSTAT_FG_QUICK) { *dipp = dip;