dm thin: resume even if in FAIL mode

Bugzilla: https://bugzilla.redhat.com/2162536
Upstream Status: kernel/git/torvalds/linux.git

commit 19eb165
Author: Luo Meng <luomeng12@huawei.com>
Date:   Wed Nov 30 10:09:45 2022 +0800

    dm thin: resume even if in FAIL mode

    If a thinpool set fail_io while suspending, resume will fail with:
     device-mapper: resume ioctl on vg-thinpool  failed: Invalid argument

    The thin-pool also can't be removed if an in-flight bio is in the
    deferred list.

    This can be easily reproduced using:

      echo "offline" > /sys/block/sda/device/state
      dd if=/dev/zero of=/dev/mapper/thin bs=4K count=1
      dmsetup suspend /dev/mapper/pool
      mkfs.ext4 /dev/mapper/thin
      dmsetup resume /dev/mapper/pool

    The root cause is maybe_resize_data_dev() will check fail_io and return
    error before called dm_resume.

    Fix this by adding FAIL mode check at the end of pool_preresume().

    Cc: stable@vger.kernel.org
    Fixes: da105ed ("dm thin metadata: introduce dm_pool_abort_metadata")
    Signed-off-by: Luo Meng <luomeng12@huawei.com>
    Signed-off-by: Mike Snitzer <snitzer@kernel.org>

Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>

Author: bmarzins

drivers/md/dm-thin.c

@@ -3542,20 +3542,28 @@ static int pool_preresume(struct dm_target *ti)
 	 */
 	r = bind_control_target(pool, ti);
 	if (r)
-		return r;
+		goto out;
 
 	r = maybe_resize_data_dev(ti, &need_commit1);
 	if (r)
-		return r;
+		goto out;
 
 	r = maybe_resize_metadata_dev(ti, &need_commit2);
 	if (r)
-		return r;
+		goto out;
 
 	if (need_commit1 || need_commit2)
 		(void) commit(pool);
+out:
+	/*
+	 * When a thin-pool is PM_FAIL, it cannot be rebuilt if
+	 * bio is in deferred list. Therefore need to return 0
+	 * to allow pool_resume() to flush IO.
+	 */
+	if (r && get_pool_mode(pool) == PM_FAIL)
+		r = 0;
 
-	return 0;
+	return r;
 }
 
 static void pool_suspend_active_thins(struct pool *pool)