Merge tag 'ceph-for-5.14-rc3' of git://github.com/ceph/ceph-client

Pull ceph fixes from Ilya Dryomov:
 "A subtle deadlock on lock_rwsem (marked for stable) and rbd fixes for
  a -rc1 regression.

  Also included a rare WARN condition tweak"

* tag 'ceph-for-5.14-rc3' of git://github.com/ceph/ceph-client:
  rbd: resurrect setting of disk->private_data in rbd_init_disk()
  ceph: don't WARN if we're still opening a session to an MDS
  rbd: don't hold lock_rwsem while running_list is being drained
  rbd: always kick acquire on "acquired" and "released" notifications

Author: torvalds

drivers/block/rbd.c

@@ -4100,8 +4100,6 @@ static void rbd_acquire_lock(struct work_struct *work)
 
 static bool rbd_quiesce_lock(struct rbd_device *rbd_dev)
 {
-	bool need_wait;
-
 	dout("%s rbd_dev %p\n", __func__, rbd_dev);
 	lockdep_assert_held_write(&rbd_dev->lock_rwsem);
 
@@ -4113,11 +4111,11 @@ static bool rbd_quiesce_lock(struct rbd_device *rbd_dev)
 	 */
 	rbd_dev->lock_state = RBD_LOCK_STATE_RELEASING;
 	rbd_assert(!completion_done(&rbd_dev->releasing_wait));
-	need_wait = !list_empty(&rbd_dev->running_list);
-	downgrade_write(&rbd_dev->lock_rwsem);
-	if (need_wait)
-		wait_for_completion(&rbd_dev->releasing_wait);
-	up_read(&rbd_dev->lock_rwsem);
+	if (list_empty(&rbd_dev->running_list))
+		return true;
+
+	up_write(&rbd_dev->lock_rwsem);
+	wait_for_completion(&rbd_dev->releasing_wait);
 
 	down_write(&rbd_dev->lock_rwsem);
 	if (rbd_dev->lock_state != RBD_LOCK_STATE_RELEASING)
@@ -4203,15 +4201,11 @@ static void rbd_handle_acquired_lock(struct rbd_device *rbd_dev, u8 struct_v,
 	if (!rbd_cid_equal(&cid, &rbd_empty_cid)) {
 		down_write(&rbd_dev->lock_rwsem);
 		if (rbd_cid_equal(&cid, &rbd_dev->owner_cid)) {
-			/*
-			 * we already know that the remote client is
-			 * the owner
-			 */
-			up_write(&rbd_dev->lock_rwsem);
-			return;
+			dout("%s rbd_dev %p cid %llu-%llu == owner_cid\n",
+			     __func__, rbd_dev, cid.gid, cid.handle);
+		} else {
+			rbd_set_owner_cid(rbd_dev, &cid);
 		}
-
-		rbd_set_owner_cid(rbd_dev, &cid);
 		downgrade_write(&rbd_dev->lock_rwsem);
 	} else {
 		down_read(&rbd_dev->lock_rwsem);
@@ -4236,14 +4230,12 @@ static void rbd_handle_released_lock(struct rbd_device *rbd_dev, u8 struct_v,
 	if (!rbd_cid_equal(&cid, &rbd_empty_cid)) {
 		down_write(&rbd_dev->lock_rwsem);
 		if (!rbd_cid_equal(&cid, &rbd_dev->owner_cid)) {
-			dout("%s rbd_dev %p unexpected owner, cid %llu-%llu != owner_cid %llu-%llu\n",
+			dout("%s rbd_dev %p cid %llu-%llu != owner_cid %llu-%llu\n",
 			     __func__, rbd_dev, cid.gid, cid.handle,
 			     rbd_dev->owner_cid.gid, rbd_dev->owner_cid.handle);
-			up_write(&rbd_dev->lock_rwsem);
-			return;
+		} else {
+			rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
 		}
-
-		rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
 		downgrade_write(&rbd_dev->lock_rwsem);
 	} else {
 		down_read(&rbd_dev->lock_rwsem);
@@ -4951,6 +4943,7 @@ static int rbd_init_disk(struct rbd_device *rbd_dev)
 		disk->minors = RBD_MINORS_PER_MAJOR;
 	}
 	disk->fops = &rbd_bd_ops;
+	disk->private_data = rbd_dev;
 
 	blk_queue_flag_set(QUEUE_FLAG_NONROT, q);
 	/* QUEUE_FLAG_ADD_RANDOM is off by default for blk-mq */

fs/ceph/mds_client.c

@@ -4456,7 +4456,7 @@ bool check_session_state(struct ceph_mds_session *s)
 		break;
 	case CEPH_MDS_SESSION_CLOSING:
 		/* Should never reach this when we're unmounting */
-		WARN_ON_ONCE(true);
+		WARN_ON_ONCE(s->s_ttl);
 		fallthrough;
 	case CEPH_MDS_SESSION_NEW:
 	case CEPH_MDS_SESSION_RESTARTING: