[spruce-project] [PATCH] xfs: fix deadlock on failure path in xfs_setattr_nonsize()

[Alexey Khoroshilov]

If xfs_trans_reserve() fails, xfs_setattr_nonsize() does not deallocate
the transaction and does not release "freeze" lock. That leads to the
following warning from lockdep:
  [ BUG: lock held when returning to user space! ]
  ------------------------------------------------
  fs-driver-tests/7127 is leaving the kernel with locks still held!
  1 lock held by fs-driver-tests/7127:
   #0:  (sb_internal){.+.+.+}, at: [<ffffffffa0389a44>] xfs_trans_alloc+0x24/0x40 [xfs]

The patch adds xfs_trans_cancel() on the failure path.

Found by Linux File System Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov at ispras.ru>
---
 fs/xfs/xfs_iops.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index 72129493e9d3..e2e785eec831 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -599,8 +599,10 @@ xfs_setattr_nonsize(
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_NOT_SIZE);
 	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ichange, 0, 0);
-	if (error)
+	if (error) {
+		xfs_trans_cancel(tp, 0);
 		goto out_dqrele;
+	}
 
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 
-- 
1.9.1