Merge tag 'for-5.13/dm-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

Pull device mapper updates from Mike Snitzer:

 - Improve scalability of DM's device hash by switching to rbtree

 - Extend DM ioctl's DM_LIST_DEVICES_CMD handling to include UUID and
   allow filtering based on name or UUID prefix.

 - Various small fixes for typos, warnings, unused function, or
   needlessly exported interfaces.

 - Remove needless request_queue NULL pointer checks in DM thin and
   cache targets.

 - Remove unnecessary loop in DM core's __split_and_process_bio().

 - Remove DM core's dm_vcalloc() and just use kvcalloc or kvmalloc_array
   instead (depending whether zeroing is useful).

 - Fix request-based DM's double free of blk_mq_tag_set in device remove
   after table load fails.

 - Improve DM persistent data performance on non-x86 by fixing packed
   structs to have a stated alignment. Also remove needless extra work
   from redundant calls to sm_disk_get_nr_free() and a paranoid BUG_ON()
   that caused duplicate checksum calculation.

 - Fix missing goto in DM integrity's bitmap_flush_interval error
   handling.

 - Add "reset_recalculate" feature flag to DM integrity.

 - Improve DM integrity by leveraging discard support to avoid needless
   re-writing of metadata and also use discard support to improve hash
   recalculation.

 - Fix race with DM raid target's reshape and MD raid4/5/6 resync that
   resulted in inconsistant reshape state during table reloads.

 - Update DM raid target to temove unnecessary discard limits for raid0
   and raid10 now that MD has optimized discard handling for both raid
   levels.

* tag 'for-5.13/dm-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: (26 commits)
  dm raid: remove unnecessary discard limits for raid0 and raid10
  dm rq: fix double free of blk_mq_tag_set in dev remove after table load fails
  dm integrity: use discard support when recalculating
  dm integrity: increase RECALC_SECTORS to improve recalculate speed
  dm integrity: don't re-write metadata if discarding same blocks
  dm raid: fix inconclusive reshape layout on fast raid4/5/6 table reload sequences
  dm raid: fix fall-through warning in rs_check_takeover() for Clang
  dm clone metadata: remove unused function
  dm integrity: fix missing goto in bitmap_flush_interval error handling
  dm: replace dm_vcalloc()
  dm space map common: fix division bug in sm_ll_find_free_block()
  dm persistent data: packed struct should have an aligned() attribute too
  dm btree spine: remove paranoid node_check call in node_prep_for_write()
  dm space map disk: remove redundant calls to sm_disk_get_nr_free()
  dm integrity: add the "reset_recalculate" feature flag
  dm persistent data: remove unused return from exit_shadow_spine()
  dm cache: remove needless request_queue NULL pointer checks
  dm thin: remove needless request_queue NULL pointer check
  dm: unexport dm_{get,put}_table_device
  dm ebs: fix a few typos
  ...

1 files changed, 29 insertions, 15 deletions

diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index cab12b2251ba..bf4a467fc73a 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -1853,6 +1853,7 @@ static int rs_check_takeover(struct raid_set *rs)
 		    ((mddev->layout == ALGORITHM_PARITY_N && mddev->new_layout == ALGORITHM_PARITY_N) ||
 		     __within_range(mddev->new_layout, ALGORITHM_LEFT_ASYMMETRIC, ALGORITHM_RIGHT_SYMMETRIC)))
 			return 0;
+		break;
 
 	default:
 		break;
@@ -1868,6 +1869,14 @@ static bool rs_takeover_requested(struct raid_set *rs)
 	return rs->md.new_level != rs->md.level;
 }
 
+/* True if layout is set to reshape. */
+static bool rs_is_layout_change(struct raid_set *rs, bool use_mddev)
+{
+	return (use_mddev ? rs->md.delta_disks : rs->delta_disks) ||
+	       rs->md.new_layout != rs->md.layout ||
+	       rs->md.new_chunk_sectors != rs->md.chunk_sectors;
+}
+
 /* True if @rs is requested to reshape by ctr */
 static bool rs_reshape_requested(struct raid_set *rs)
 {
@@ -1880,9 +1889,7 @@ static bool rs_reshape_requested(struct raid_set *rs)
 	if (rs_is_raid0(rs))
 		return false;
 
-	change = mddev->new_layout != mddev->layout ||
-		 mddev->new_chunk_sectors != mddev->chunk_sectors ||
-		 rs->delta_disks;
+	change = rs_is_layout_change(rs, false);
 
 	/* Historical case to support raid1 reshape without delta disks */
 	if (rs_is_raid1(rs)) {
@@ -2817,7 +2824,7 @@ static sector_t _get_reshape_sectors(struct raid_set *rs)
 }
 
 /*
- *
+ * Reshape:
  * - change raid layout
  * - change chunk size
  * - add disks
@@ -2927,6 +2934,20 @@ static int rs_setup_reshape(struct raid_set *rs)
 }
 
 /*
+ * If the md resync thread has updated superblock with max reshape position
+ * at the end of a reshape but not (yet) reset the layout configuration
+ * changes -> reset the latter.
+ */
+static void rs_reset_inconclusive_reshape(struct raid_set *rs)
+{
+	if (!rs_is_reshaping(rs) && rs_is_layout_change(rs, true)) {
+		rs_set_cur(rs);
+		rs->md.delta_disks = 0;
+		rs->md.reshape_backwards = 0;
+	}
+}
+
+/*
  * Enable/disable discard support on RAID set depending on
  * RAID level and discard properties of underlying RAID members.
  */
@@ -3212,11 +3233,14 @@ size_check:
 	if (r)
 		goto bad;
 
+	/* Catch any inconclusive reshape superblock content. */
+	rs_reset_inconclusive_reshape(rs);
+
 	/* Start raid set read-only and assumed clean to change in raid_resume() */
 	rs->md.ro = 1;
 	rs->md.in_sync = 1;
 
-	/* Keep array frozen */
+	/* Keep array frozen until resume. */
 	set_bit(MD_RECOVERY_FROZEN, &rs->md.recovery);
 
 	/* Has to be held on running the array */
@@ -3230,7 +3254,6 @@ size_check:
 	}
 
 	r = md_start(&rs->md);
-
 	if (r) {
 		ti->error = "Failed to start raid array";
 		mddev_unlock(&rs->md);
@@ -3727,15 +3750,6 @@ static void raid_io_hints(struct dm_target *ti, struct queue_limits *limits)
 
 	blk_limits_io_min(limits, chunk_size_bytes);
 	blk_limits_io_opt(limits, chunk_size_bytes * mddev_data_stripes(rs));
-
-	/*
-	 * RAID0 and RAID10 personalities require bio splitting,
-	 * RAID1/4/5/6 don't and process large discard bios properly.
-	 */
-	if (rs_is_raid0(rs) || rs_is_raid10(rs)) {
-		limits->discard_granularity = chunk_size_bytes;
-		limits->max_discard_sectors = rs->md.chunk_sectors;
-	}
 }
 
 static void raid_postsuspend(struct dm_target *ti)