xfs: don't vmap inode cluster buffers during free
authorDave Chinner <dchinner@redhat.com>
Fri, 2 Nov 2012 00:38:42 +0000 (11:38 +1100)
committerBen Myers <bpm@sgi.com>
Thu, 8 Nov 2012 17:10:18 +0000 (11:10 -0600)
Inode buffers do not need to be mapped as inodes are read or written
directly from/to the pages underlying the buffer. This fixes a
regression introduced by commit 611c994 ("xfs: make XBF_MAPPED the
default behaviour").

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>
fs/xfs/xfs_inode.c

index 2778258..1938b41 100644 (file)
@@ -1509,7 +1509,8 @@ xfs_ifree_cluster(
                 * to mark all the active inodes on the buffer stale.
                 */
                bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno,
-                                       mp->m_bsize * blks_per_cluster, 0);
+                                       mp->m_bsize * blks_per_cluster,
+                                       XBF_UNMAPPED);
 
                if (!bp)
                        return ENOMEM;