gfs2: gfs2_evict_inode: Put glocks asynchronously

gfs2_evict_inode is called to free inodes under memory pressure.  The
function calls into DLM when an inode's last cluster-wide reference goes
away (remote unlink) and to release the glock and associated DLM lock
before finally destroying the inode.  However, if DLM is blocked on
memory to become available, calling into DLM again will deadlock.

Avoid that by decoupling releasing glocks from destroying inodes in that
case: with gfs2_glock_queue_put, glocks will be dequeued asynchronously
in work queue context, when the associated inodes have likely already
been destroyed.

With this change, inodes can end up being unlinked, remote-unlink can be
triggered, and then the inode can be reallocated before all
remote-unlink callbacks are processed.  To detect that, revalidate the
link count in gfs2_evict_inode to make sure we're not deleting an
allocated, referenced inode.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>

1 files changed, 28 insertions, 2 deletions

diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c
index 6c39bb1ec100..4089dbe617a6 100644
--- a/fs/gfs2/super.c
+++ b/fs/gfs2/super.c
@@ -1502,6 +1502,22 @@ out_qs:
 }
 
 /**
+ * gfs2_glock_put_eventually
+ * @gl:	The glock to put
+ *
+ * When under memory pressure, trigger a deferred glock put to make sure we
+ * won't call into DLM and deadlock.  Otherwise, put the glock directly.
+ */
+
+static void gfs2_glock_put_eventually(struct gfs2_glock *gl)
+{
+	if (current->flags & PF_MEMALLOC)
+		gfs2_glock_queue_put(gl);
+	else
+		gfs2_glock_put(gl);
+}
+
+/**
  * gfs2_evict_inode - Remove an inode from cache
  * @inode: The inode to evict
  *
@@ -1564,6 +1580,12 @@ static void gfs2_evict_inode(struct inode *inode)
 			goto out_truncate;
 	}
 
+	/*
+	 * The inode may have been recreated in the meantime.
+	 */
+	if (inode->i_nlink)
+		goto out_truncate;
+
 alloc_failed:
 	if (gfs2_holder_initialized(&ip->i_iopen_gh) &&
 	    test_bit(HIF_HOLDER, &ip->i_iopen_gh.gh_iflags)) {
@@ -1653,12 +1675,16 @@ out:
 	glock_clear_object(ip->i_gl, ip);
 	wait_on_bit_io(&ip->i_flags, GIF_GLOP_PENDING, TASK_UNINTERRUPTIBLE);
 	gfs2_glock_add_to_lru(ip->i_gl);
-	gfs2_glock_put(ip->i_gl);
+	gfs2_glock_put_eventually(ip->i_gl);
 	ip->i_gl = NULL;
 	if (gfs2_holder_initialized(&ip->i_iopen_gh)) {
-		glock_clear_object(ip->i_iopen_gh.gh_gl, ip);
+		struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
+
+		glock_clear_object(gl, ip);
 		ip->i_iopen_gh.gh_flags |= GL_NOCACHE;
+		gfs2_glock_hold(gl);
 		gfs2_glock_dq_uninit(&ip->i_iopen_gh);
+		gfs2_glock_put_eventually(gl);
 	}
 }