summaryrefslogtreecommitdiffstats
path: root/fs/xfs/xfs_sync.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/xfs/xfs_sync.c')
-rw-r--r--fs/xfs/xfs_sync.c4
1 files changed, 2 insertions, 2 deletions
diff --git a/fs/xfs/xfs_sync.c b/fs/xfs/xfs_sync.c
index 1e9ee064dbb..e61fc151907 100644
--- a/fs/xfs/xfs_sync.c
+++ b/fs/xfs/xfs_sync.c
@@ -712,8 +712,8 @@ restart:
* Note that xfs_iflush will never block on the inode buffer lock, as
* xfs_ifree_cluster() can lock the inode buffer before it locks the
* ip->i_lock, and we are doing the exact opposite here. As a result,
- * doing a blocking xfs_itobp() to get the cluster buffer would result
- * in an ABBA deadlock with xfs_ifree_cluster().
+ * doing a blocking xfs_imap_to_bp() to get the cluster buffer would
+ * result in an ABBA deadlock with xfs_ifree_cluster().
*
* As xfs_ifree_cluser() must gather all inodes that are active in the
* cache to mark them stale, if we hit this case we don't actually want