diff options
| author | Darrick J. Wong <djwong@kernel.org> | 2021-08-06 11:05:40 -0700 |
|---|---|---|
| committer | Darrick J. Wong <djwong@kernel.org> | 2021-08-09 10:52:18 -0700 |
| commit | 108523b8de676a45cef1f6c8566c444222b85de0 (patch) | |
| tree | f9a89cf8a2114065ecb7f17846dbb20008f5d615 /fs/xfs/xfs_icache.c | |
| parent | 7d6f07d2c5ad9fce298889eeed317d512a2df8cd (diff) | |
xfs: queue inactivation immediately when quota is nearing enforcement
Now that we have made the inactivation of unlinked inodes a background
task to increase the throughput of file deletions, we need to be a
little more careful about how long of a delay we can tolerate.
Specifically, if the dquots attached to the inode being inactivated are
nearing any kind of enforcement boundary, we want to queue that
inactivation work immediately so that users don't get EDQUOT/ENOSPC
errors even after they deleted a bunch of files to stay within quota.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Diffstat (limited to 'fs/xfs/xfs_icache.c')
| -rw-r--r-- | fs/xfs/xfs_icache.c | 10 |
1 files changed, 10 insertions, 0 deletions
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 8b2ac969c60c..0bea604f320a 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -1921,6 +1921,7 @@ xfs_inodegc_start( * * - We've accumulated more than one inode cluster buffer's worth of inodes. * - There is less than 5% free space left. + * - Any of the quotas for this inode are near an enforcement limit. */ static inline bool xfs_inodegc_want_queue_work( @@ -1937,6 +1938,15 @@ xfs_inodegc_want_queue_work( XFS_FDBLOCKS_BATCH) < 0) return true; + if (xfs_inode_near_dquot_enforcement(ip, XFS_DQTYPE_USER)) + return true; + + if (xfs_inode_near_dquot_enforcement(ip, XFS_DQTYPE_GROUP)) + return true; + + if (xfs_inode_near_dquot_enforcement(ip, XFS_DQTYPE_PROJ)) + return true; + return false; } |