From: Dave Chinner Date: Wed, 28 Nov 2012 02:01:02 +0000 (+1100) Subject: xfs: fix stray dquot unlock when reclaiming dquots X-Git-Tag: v3.7.2~30 X-Git-Url: https://git.karo-electronics.de/?a=commitdiff_plain;h=f8b723cc9f1a9eec83c8527d67eedc6249b7fb66;p=karo-tx-linux.git xfs: fix stray dquot unlock when reclaiming dquots commit b870553cdecb26d5291af09602352b763e323df2 upstream. When we fail to get a dquot lock during reclaim, we jump to an error handler that unlocks the dquot. This is wrong as we didn't lock the dquot, and unlocking it means who-ever is holding the lock has had it silently taken away, and hence it results in a lock imbalance. Found by inspection while modifying the code for the numa-lru patchset. This fixes a random hang I've been seeing on xfstest 232 for the past several months. Signed-off-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Ben Myers Signed-off-by: Greg Kroah-Hartman --- diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 2e86fa0cfc0d..335a34041ea1 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1453,7 +1453,7 @@ xfs_qm_dqreclaim_one( int error; if (!xfs_dqlock_nowait(dqp)) - goto out_busy; + goto out_move_tail; /* * This dquot has acquired a reference in the meantime remove it from @@ -1476,7 +1476,7 @@ xfs_qm_dqreclaim_one( * getting flushed to disk, we don't want to reclaim it. */ if (!xfs_dqflock_nowait(dqp)) - goto out_busy; + goto out_unlock_move_tail; if (XFS_DQ_IS_DIRTY(dqp)) { struct xfs_buf *bp = NULL; @@ -1487,7 +1487,7 @@ xfs_qm_dqreclaim_one( if (error) { xfs_warn(mp, "%s: dquot %p flush failed", __func__, dqp); - goto out_busy; + goto out_unlock_move_tail; } xfs_buf_delwri_queue(bp, buffer_list); @@ -1496,7 +1496,7 @@ xfs_qm_dqreclaim_one( * Give the dquot another try on the freelist, as the * flushing will take some time. */ - goto out_busy; + goto out_unlock_move_tail; } xfs_dqfunlock(dqp); @@ -1515,14 +1515,13 @@ xfs_qm_dqreclaim_one( XFS_STATS_INC(xs_qm_dqreclaims); return; -out_busy: - xfs_dqunlock(dqp); - /* * Move the dquot to the tail of the list so that we don't spin on it. */ +out_unlock_move_tail: + xfs_dqunlock(dqp); +out_move_tail: list_move_tail(&dqp->q_lru, &qi->qi_lru_list); - trace_xfs_dqreclaim_busy(dqp); XFS_STATS_INC(xs_qm_dqreclaim_misses); }