gfs2: skip if we cannot defer delete

In gfs2_evict_inode(), in the unlikely case that we cannot defer
deleting the inode, it is not safe to fall back to deleting the inode;
the only valid choice we have is to skip the delete.

In addition, in evict_should_delete(), if we cannot lock the inode glock
exclusively, we are in a bad enough state that skipping the delete is
likely a better choice than trying to recover from the failure later.

Fixes: c5b7a24 ("gfs2: Only defer deletes when we have an iopen glock")
Signed-off-by: Andreas Gruenbacher <[email protected]>

Author: Andreas Gruenbacher

fs/gfs2/super.c

@@ -1340,7 +1340,7 @@ static enum evict_behavior evict_should_delete(struct inode *inode,
 	/* Must not read inode block until block type has been verified */
 	ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, GL_SKIP, gh);
 	if (unlikely(ret))
-		return EVICT_SHOULD_DEFER_DELETE;
+		return EVICT_SHOULD_SKIP_DELETE;
 
 	if (gfs2_inode_already_deleted(ip->i_gl, ip->i_no_formal_ino))
 		return EVICT_SHOULD_SKIP_DELETE;
@@ -1499,7 +1499,7 @@ static void gfs2_evict_inode(struct inode *inode)
 				gfs2_glock_put(io_gl);
 			goto out;
 		}
-		behavior = EVICT_SHOULD_DELETE;
+		behavior = EVICT_SHOULD_SKIP_DELETE;
 	}
 	if (behavior == EVICT_SHOULD_DELETE)
 		ret = evict_unlinked_inode(inode);