Merge tag 'dm-4.8-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

Pull device mapper fixes from Mike Snitzer:

 - a stable dm-flakey fix to error read IO during the 'down_interval'

 - a DM core suspend fix to establish the SUSPENDED flag before dropping
   the SUSPENDING flag

 - a blk-mq request-based DM (dm-mq) dm_stop_queue() fix to properly
   stop the blk-mq hw_queues (and cancel pending requeue work); also
   set/clear QUEUE_FLAG_STOPPED when stopping/starting the dm-mq
   request_queue.

 - a DM multipath fix to harden locking of in-core state flags in the
   face of concurrent access while handling path failures under heavy
   IO.

 - a few small DM raid fixes to edge cases caught with further testing.

* tag 'dm-4.8-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm raid: fix use of wrong status char during resynchronization
  dm raid: constructor fails on non-zero incompat_features
  dm raid: fix processing of max_recovery_rate constructor flag
  dm: set DMF_SUSPENDED* _before_ clearing DMF_NOFLUSH_SUSPENDING
  dm rq: fix the starting and stopping of blk-mq queues
  dm mpath: add locking to multipath_resume and must_push_back
  dm flakey: error READ bios during the down_interval

drivers/md/dm-flakey.c

@@ -289,10 +289,16 @@ static int flakey_map(struct dm_target *ti, struct bio *bio)
 		pb->bio_submitted = true;
 
 		/*
-		 * Map reads as normal.
+		 * Map reads as normal only if corrupt_bio_byte set.
 		 */
-		if (bio_data_dir(bio) == READ)
+		if (bio_data_dir(bio) == READ) {
+			/* If flags were specified, only corrupt those that match. */
+			if (fc->corrupt_bio_byte && (fc->corrupt_bio_rw == READ) &&
+			    all_corrupt_bio_flags_match(bio, fc))
 				goto map_bio;
+			else
+				return -EIO;
+		}
 
 		/*
 		 * Drop writes?
@@ -330,12 +336,13 @@ static int flakey_end_io(struct dm_target *ti, struct bio *bio, int error)
 
 	/*
 	 * Corrupt successful READs while in down state.
-	 * If flags were specified, only corrupt those that match.
 	 */
-	if (fc->corrupt_bio_byte && !error && pb->bio_submitted &&
-	    (bio_data_dir(bio) == READ) && (fc->corrupt_bio_rw == READ) &&
-	    all_corrupt_bio_flags_match(bio, fc))
+	if (!error && pb->bio_submitted && (bio_data_dir(bio) == READ)) {
+		if (fc->corrupt_bio_byte)
 			corrupt_bio_data(bio, fc);
+		else
+			return -EIO;
+	}
 
 	return error;
 }

drivers/md/dm-mpath.c

@@ -507,13 +507,27 @@ static bool __must_push_back(struct multipath *m)
 
 static bool must_push_back_rq(struct multipath *m)
 {
-	return (test_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags) ||
+	bool r;
+	unsigned long flags;
+
+	spin_lock_irqsave(&m->lock, flags);
+	r = (test_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags) ||
 	     __must_push_back(m));
+	spin_unlock_irqrestore(&m->lock, flags);
+
+	return r;
 }
 
 static bool must_push_back_bio(struct multipath *m)
 {
-	return __must_push_back(m);
+	bool r;
+	unsigned long flags;
+
+	spin_lock_irqsave(&m->lock, flags);
+	r = __must_push_back(m);
+	spin_unlock_irqrestore(&m->lock, flags);
+
+	return r;
 }
 
 /*
@@ -1680,12 +1694,14 @@ static void multipath_postsuspend(struct dm_target *ti)
 static void multipath_resume(struct dm_target *ti)
 {
 	struct multipath *m = ti->private;
+	unsigned long flags;
 
+	spin_lock_irqsave(&m->lock, flags);
 	if (test_bit(MPATHF_SAVED_QUEUE_IF_NO_PATH, &m->flags))
 		set_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags);
 	else
 		clear_bit(MPATHF_QUEUE_IF_NO_PATH, &m->flags);
-	smp_mb__after_atomic();
+	spin_unlock_irqrestore(&m->lock, flags);
 }
 
 /*

drivers/md/dm-raid.c

@@ -1270,7 +1270,7 @@ static int parse_raid_params(struct raid_set *rs, struct dm_arg_set *as,
 			}
 			rs->md.sync_speed_min = (int)value;
 		} else if (!strcasecmp(key, dm_raid_arg_name_by_flag(CTR_FLAG_MAX_RECOVERY_RATE))) {
-			if (test_and_set_bit(__CTR_FLAG_MIN_RECOVERY_RATE, &rs->ctr_flags)) {
+			if (test_and_set_bit(__CTR_FLAG_MAX_RECOVERY_RATE, &rs->ctr_flags)) {
 				rs->ti->error = "Only one max_recovery_rate argument pair allowed";
 				return -EINVAL;
 			}
@@ -1960,6 +1960,7 @@ static void super_sync(struct mddev *mddev, struct md_rdev *rdev)
 	sb->data_offset = cpu_to_le64(rdev->data_offset);
 	sb->new_data_offset = cpu_to_le64(rdev->new_data_offset);
 	sb->sectors = cpu_to_le64(rdev->sectors);
+	sb->incompat_features = cpu_to_le32(0);
 
 	/* Zero out the rest of the payload after the size of the superblock */
 	memset(sb + 1, 0, rdev->sb_size - sizeof(*sb));
@@ -3577,7 +3578,6 @@ static int raid_preresume(struct dm_target *ti)
 	/* Be prepared for mddev_resume() in raid_resume() */
 	set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
 	if (mddev->recovery_cp && mddev->recovery_cp < MaxSector) {
-		set_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
 		set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
 		mddev->resync_min = mddev->recovery_cp;
 	}

drivers/md/dm-rq.c

@@ -78,6 +78,7 @@ void dm_start_queue(struct request_queue *q)
 	if (!q->mq_ops)
 		dm_old_start_queue(q);
 	else {
+		queue_flag_clear_unlocked(QUEUE_FLAG_STOPPED, q);
 		blk_mq_start_stopped_hw_queues(q, true);
 		blk_mq_kick_requeue_list(q);
 	}
@@ -101,8 +102,14 @@ void dm_stop_queue(struct request_queue *q)
 {
 	if (!q->mq_ops)
 		dm_old_stop_queue(q);
-	else
+	else {
+		spin_lock_irq(q->queue_lock);
+		queue_flag_set(QUEUE_FLAG_STOPPED, q);
+		spin_unlock_irq(q->queue_lock);
+
+		blk_mq_cancel_requeue_work(q);
 		blk_mq_stop_hw_queues(q);
+	}
 }
 
 static struct dm_rq_target_io *alloc_old_rq_tio(struct mapped_device *md,
@@ -864,6 +871,17 @@ static int dm_mq_queue_rq(struct blk_mq_hw_ctx *hctx,
 		dm_put_live_table(md, srcu_idx);
 	}
 
+	/*
+	 * On suspend dm_stop_queue() handles stopping the blk-mq
+	 * request_queue BUT: even though the hw_queues are marked
+	 * BLK_MQ_S_STOPPED at that point there is still a race that
+	 * is allowing block/blk-mq.c to call ->queue_rq against a
+	 * hctx that it really shouldn't.  The following check guards
+	 * against this rarity (albeit _not_ race-free).
+	 */
+	if (unlikely(test_bit(BLK_MQ_S_STOPPED, &hctx->state)))
+		return BLK_MQ_RQ_QUEUE_BUSY;
+
 	if (ti->type->busy && ti->type->busy(ti))
 		return BLK_MQ_RQ_QUEUE_BUSY;
 

drivers/md/dm.c

@@ -2082,7 +2082,8 @@ static void unlock_fs(struct mapped_device *md)
  * Caller must hold md->suspend_lock
  */
 static int __dm_suspend(struct mapped_device *md, struct dm_table *map,
-			unsigned suspend_flags, int interruptible)
+			unsigned suspend_flags, int interruptible,
+			int dmf_suspended_flag)
 {
 	bool do_lockfs = suspend_flags & DM_SUSPEND_LOCKFS_FLAG;
 	bool noflush = suspend_flags & DM_SUSPEND_NOFLUSH_FLAG;
@@ -2149,6 +2150,8 @@ static int __dm_suspend(struct mapped_device *md, struct dm_table *map,
 	 * to finish.
 	 */
 	r = dm_wait_for_completion(md, interruptible);
+	if (!r)
+		set_bit(dmf_suspended_flag, &md->flags);
 
 	if (noflush)
 		clear_bit(DMF_NOFLUSH_SUSPENDING, &md->flags);
@@ -2210,12 +2213,10 @@ int dm_suspend(struct mapped_device *md, unsigned suspend_flags)
 
 	map = rcu_dereference_protected(md->map, lockdep_is_held(&md->suspend_lock));
 
-	r = __dm_suspend(md, map, suspend_flags, TASK_INTERRUPTIBLE);
+	r = __dm_suspend(md, map, suspend_flags, TASK_INTERRUPTIBLE, DMF_SUSPENDED);
 	if (r)
 		goto out_unlock;
 
-	set_bit(DMF_SUSPENDED, &md->flags);
-
 	dm_table_postsuspend_targets(map);
 
 out_unlock:
@@ -2309,9 +2310,8 @@ static void __dm_internal_suspend(struct mapped_device *md, unsigned suspend_fla
 	 * would require changing .presuspend to return an error -- avoid this
 	 * until there is a need for more elaborate variants of internal suspend.
 	 */
-	(void) __dm_suspend(md, map, suspend_flags, TASK_UNINTERRUPTIBLE);
-
-	set_bit(DMF_SUSPENDED_INTERNALLY, &md->flags);
+	(void) __dm_suspend(md, map, suspend_flags, TASK_UNINTERRUPTIBLE,
+			    DMF_SUSPENDED_INTERNALLY);
 
 	dm_table_postsuspend_targets(map);
 }