During reclaim under clear_inode, one thread holds the iprune_mutex while
trying to get an inode buffer lock. The other thread has the inode buffer
lock while trying to get the iprune_mutex lock.
Avoid reclaim recursing into the file system by using KM_NOFS in
xfs_trans_alloc.
Signed-off-by: Peter Watkins <treestem@xxxxxxxxx>
---
fs/xfs/xfs_trans.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 7692279..f26f6d8 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -578,7 +578,7 @@ xfs_trans_alloc(
uint type)
{
xfs_wait_for_freeze(mp, SB_FREEZE_TRANS);
- return _xfs_trans_alloc(mp, type, KM_SLEEP);
+ return _xfs_trans_alloc(mp, type, KM_SLEEP|KM_NOFS);
}
xfs_trans_t *
--
1.7.0.4
_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs
