[GFS2] Mark nlink cleared so VFS sees it happen
authorSteven Whitehouse <swhiteho@redhat.com>
Mon, 2 Oct 2006 20:01:53 +0000 (16:01 -0400)
committerSteven Whitehouse <swhiteho@redhat.com>
Mon, 2 Oct 2006 20:01:53 +0000 (16:01 -0400)
This does nothing atm, but will be required for later support of
r/o bind mounts.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
fs/gfs2/inode.c

index 660dc81..81a9211 100644 (file)
@@ -353,6 +353,7 @@ int gfs2_change_nlink(struct gfs2_inode *ip, int diff)
                if (error)
                        goto out_norgrp;
 
+               clear_nlink(&ip->i_inode);
                gfs2_unlink_di(&ip->i_inode); /* mark inode unlinked */
                gfs2_glock_dq_uninit(&rg_gh);
 out_norgrp: