Merge commit 'c039c332f23e794deb6d6f37b9f07ff3b27fb2cf' into md

Pull in pre-requisites for adding raid10 support to dm-raid.

18 files changed, 280 insertions, 687 deletions

diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig
index 10f122a3a85..1eee45b69b7 100644
--- a/drivers/md/Kconfig
+++ b/drivers/md/Kconfig
@@ -260,15 +260,6 @@ config DM_DEBUG_BLOCK_STACK_TRACING
 
 	  If unsure, say N.
 
-config DM_DEBUG_SPACE_MAPS
-	boolean "Extra validation for thin provisioning space maps"
-	depends on DM_THIN_PROVISIONING
-	---help---
-	  Enable this for messages that may help debug problems with the
-	  space maps used by thin provisioning.
-
-          If unsure, say N.
-
 config DM_MIRROR
        tristate "Mirror target"
        depends on BLK_DEV_DM
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 3f06df59fd8..e2b32401ecc 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -1241,7 +1241,6 @@ static void kcryptd_queue_crypt(struct dm_crypt_io *io)
 static int crypt_decode_key(u8 *key, char *hex, unsigned int size)
 {
 	char buffer[3];
-	char *endp;
 	unsigned int i;
 
 	buffer[2] = '\0';
@@ -1250,9 +1249,7 @@ static int crypt_decode_key(u8 *key, char *hex, unsigned int size)
 		buffer[0] = *hex++;
 		buffer[1] = *hex++;
 
-		key[i] = (u8)simple_strtoul(buffer, &endp, 16);
-
-		if (endp != &buffer[2])
+		if (kstrtou8(buffer, 16, &key[i]))
 			return -EINVAL;
 	}
 
diff --git a/drivers/md/dm-exception-store.c b/drivers/md/dm-exception-store.c
index aa70f7d43a1..ebaa4f803ee 100644
--- a/drivers/md/dm-exception-store.c
+++ b/drivers/md/dm-exception-store.c
@@ -142,24 +142,19 @@ EXPORT_SYMBOL(dm_exception_store_type_unregister);
 static int set_chunk_size(struct dm_exception_store *store,
 			  const char *chunk_size_arg, char **error)
 {
-	unsigned long chunk_size_ulong;
-	char *value;
+	unsigned chunk_size;
 
-	chunk_size_ulong = simple_strtoul(chunk_size_arg, &value, 10);
-	if (*chunk_size_arg == '\0' || *value != '\0' ||
-	    chunk_size_ulong > UINT_MAX) {
+	if (kstrtouint(chunk_size_arg, 10, &chunk_size)) {
 		*error = "Invalid chunk size";
 		return -EINVAL;
 	}
 
-	if (!chunk_size_ulong) {
+	if (!chunk_size) {
 		store->chunk_size = store->chunk_mask = store->chunk_shift = 0;
 		return 0;
 	}
 
-	return dm_exception_store_set_chunk_size(store,
-						 (unsigned) chunk_size_ulong,
-						 error);
+	return dm_exception_store_set_chunk_size(store, chunk_size, error);
 }
 
 int dm_exception_store_set_chunk_size(struct dm_exception_store *store,
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index 638dae048b4..8a3b2d53f81 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -85,6 +85,7 @@ struct multipath {
 	unsigned queue_io:1;		/* Must we queue all I/O? */
 	unsigned queue_if_no_path:1;	/* Queue I/O if last path fails? */
 	unsigned saved_queue_if_no_path:1; /* Saved state during suspension */
+	unsigned retain_attached_hw_handler:1; /* If there's already a hw_handler present, don't change it. */
 
 	unsigned pg_init_retries;	/* Number of times to retry pg_init */
 	unsigned pg_init_count;		/* Number of times pg_init called */
@@ -568,6 +569,8 @@ static struct pgpath *parse_path(struct dm_arg_set *as, struct path_selector *ps
 	int r;
 	struct pgpath *p;
 	struct multipath *m = ti->private;
+	struct request_queue *q = NULL;
+	const char *attached_handler_name;
 
 	/* we need at least a path arg */
 	if (as->argc < 1) {
@@ -586,13 +589,37 @@ static struct pgpath *parse_path(struct dm_arg_set *as, struct path_selector *ps
 		goto bad;
 	}
 
-	if (m->hw_handler_name) {
-		struct request_queue *q = bdev_get_queue(p->path.dev->bdev);
+	if (m->retain_attached_hw_handler || m->hw_handler_name)
+		q = bdev_get_queue(p->path.dev->bdev);
+
+	if (m->retain_attached_hw_handler) {
+		attached_handler_name = scsi_dh_attached_handler_name(q, GFP_KERNEL);
+		if (attached_handler_name) {
+			/*
+			 * Reset hw_handler_name to match the attached handler
+			 * and clear any hw_handler_params associated with the
+			 * ignored handler.
+			 *
+			 * NB. This modifies the table line to show the actual
+			 * handler instead of the original table passed in.
+			 */
+			kfree(m->hw_handler_name);
+			m->hw_handler_name = attached_handler_name;
+
+			kfree(m->hw_handler_params);
+			m->hw_handler_params = NULL;
+		}
+	}
 
+	if (m->hw_handler_name) {
+		/*
+		 * Increments scsi_dh reference, even when using an
+		 * already-attached handler.
+		 */
 		r = scsi_dh_attach(q, m->hw_handler_name);
 		if (r == -EBUSY) {
 			/*
-			 * Already attached to different hw_handler,
+			 * Already attached to different hw_handler:
 			 * try to reattach with correct one.
 			 */
 			scsi_dh_detach(q);
@@ -760,7 +787,7 @@ static int parse_features(struct dm_arg_set *as, struct multipath *m)
 	const char *arg_name;
 
 	static struct dm_arg _args[] = {
-		{0, 5, "invalid number of feature args"},
+		{0, 6, "invalid number of feature args"},
 		{1, 50, "pg_init_retries must be between 1 and 50"},
 		{0, 60000, "pg_init_delay_msecs must be between 0 and 60000"},
 	};
@@ -781,6 +808,11 @@ static int parse_features(struct dm_arg_set *as, struct multipath *m)
 			continue;
 		}
 
+		if (!strcasecmp(arg_name, "retain_attached_hw_handler")) {
+			m->retain_attached_hw_handler = 1;
+			continue;
+		}
+
 		if (!strcasecmp(arg_name, "pg_init_retries") &&
 		    (argc >= 1)) {
 			r = dm_read_arg(_args + 1, as, &m->pg_init_retries, &ti->error);
@@ -1364,13 +1396,16 @@ static int multipath_status(struct dm_target *ti, status_type_t type,
 	else {
 		DMEMIT("%u ", m->queue_if_no_path +
 			      (m->pg_init_retries > 0) * 2 +
-			      (m->pg_init_delay_msecs != DM_PG_INIT_DELAY_DEFAULT) * 2);
+			      (m->pg_init_delay_msecs != DM_PG_INIT_DELAY_DEFAULT) * 2 +
+			      m->retain_attached_hw_handler);
 		if (m->queue_if_no_path)
 			DMEMIT("queue_if_no_path ");
 		if (m->pg_init_retries)
 			DMEMIT("pg_init_retries %u ", m->pg_init_retries);
 		if (m->pg_init_delay_msecs != DM_PG_INIT_DELAY_DEFAULT)
 			DMEMIT("pg_init_delay_msecs %u ", m->pg_init_delay_msecs);
+		if (m->retain_attached_hw_handler)
+			DMEMIT("retain_attached_hw_handler ");
 	}
 
 	if (!m->hw_handler_name || type == STATUSTYPE_INFO)
@@ -1656,7 +1691,7 @@ out:
  *---------------------------------------------------------------*/
 static struct target_type multipath_target = {
 	.name = "multipath",
-	.version = {1, 4, 0},
+	.version = {1, 5, 0},
 	.module = THIS_MODULE,
 	.ctr = multipath_ctr,
 	.dtr = multipath_dtr,
diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index 017c34d78d6..f4275a8e860 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -101,20 +101,12 @@ static struct raid_set *context_alloc(struct dm_target *ti, struct raid_type *ra
 {
 	unsigned i;
 	struct raid_set *rs;
-	sector_t sectors_per_dev;
 
 	if (raid_devs <= raid_type->parity_devs) {
 		ti->error = "Insufficient number of devices";
 		return ERR_PTR(-EINVAL);
 	}
 
-	sectors_per_dev = ti->len;
-	if ((raid_type->level > 1) &&
-	    sector_div(sectors_per_dev, (raid_devs - raid_type->parity_devs))) {
-		ti->error = "Target length not divisible by number of data devices";
-		return ERR_PTR(-EINVAL);
-	}
-
 	rs = kzalloc(sizeof(*rs) + raid_devs * sizeof(rs->dev[0]), GFP_KERNEL);
 	if (!rs) {
 		ti->error = "Cannot allocate raid context";
@@ -128,7 +120,6 @@ static struct raid_set *context_alloc(struct dm_target *ti, struct raid_type *ra
 	rs->md.raid_disks = raid_devs;
 	rs->md.level = raid_type->level;
 	rs->md.new_level = rs->md.level;
-	rs->md.dev_sectors = sectors_per_dev;
 	rs->md.layout = raid_type->algorithm;
 	rs->md.new_layout = rs->md.layout;
 	rs->md.delta_disks = 0;
@@ -143,6 +134,7 @@ static struct raid_set *context_alloc(struct dm_target *ti, struct raid_type *ra
 	 *  rs->md.external
 	 *  rs->md.chunk_sectors
 	 *  rs->md.new_chunk_sectors
+	 *  rs->md.dev_sectors
 	 */
 
 	return rs;
@@ -353,6 +345,8 @@ static int parse_raid_params(struct raid_set *rs, char **argv,
 {
 	unsigned i, rebuild_cnt = 0;
 	unsigned long value, region_size = 0;
+	sector_t sectors_per_dev = rs->ti->len;
+	sector_t max_io_len;
 	char *key;
 
 	/*
@@ -429,13 +423,28 @@ static int parse_raid_params(struct raid_set *rs, char **argv,
 
 		if (!strcasecmp(key, "rebuild")) {
 			rebuild_cnt++;
-			if (((rs->raid_type->level != 1) &&
-			     (rebuild_cnt > rs->raid_type->parity_devs)) ||
-			    ((rs->raid_type->level == 1) &&
-			     (rebuild_cnt > (rs->md.raid_disks - 1)))) {
-				rs->ti->error = "Too many rebuild devices specified for given RAID type";
+
+			switch (rs->raid_type->level) {
+			case 1:
+				if (rebuild_cnt >= rs->md.raid_disks) {
+					rs->ti->error = "Too many rebuild devices specified";
+					return -EINVAL;
+				}
+				break;
+			case 4:
+			case 5:
+			case 6:
+				if (rebuild_cnt > rs->raid_type->parity_devs) {
+					rs->ti->error = "Too many rebuild devices specified for given RAID type";
+					return -EINVAL;
+				}
+				break;
+			default:
+				DMERR("The rebuild parameter is not supported for %s", rs->raid_type->name);
+				rs->ti->error = "Rebuild not supported for this RAID type";
 				return -EINVAL;
 			}
+
 			if (value > rs->md.raid_disks) {
 				rs->ti->error = "Invalid rebuild index given";
 				return -EINVAL;
@@ -522,14 +531,19 @@ static int parse_raid_params(struct raid_set *rs, char **argv,
 		return -EINVAL;
 
 	if (rs->md.chunk_sectors)
-		rs->ti->split_io = rs->md.chunk_sectors;
+		max_io_len = rs->md.chunk_sectors;
 	else
-		rs->ti->split_io = region_size;
+		max_io_len = region_size;
 
-	if (rs->md.chunk_sectors)
-		rs->ti->split_io = rs->md.chunk_sectors;
-	else
-		rs->ti->split_io = region_size;
+	if (dm_set_target_max_io_len(rs->ti, max_io_len))
+		return -EINVAL;
+
+	if ((rs->raid_type->level > 1) &&
+	    sector_div(sectors_per_dev, (rs->md.raid_disks - rs->raid_type->parity_devs))) {
+		rs->ti->error = "Target length not divisible by number of data devices";
+		return -EINVAL;
+	}
+	rs->md.dev_sectors = sectors_per_dev;
 
 	/* Assume there are no metadata devices until the drives are parsed */
 	rs->md.persistent = 0;
diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c
index d039de8322f..819ccba6591 100644
--- a/drivers/md/dm-raid1.c
+++ b/drivers/md/dm-raid1.c
@@ -1081,9 +1081,14 @@ static int mirror_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 	}
 
 	ti->private = ms;
-	ti->split_io = dm_rh_get_region_size(ms->rh);
+
+	r = dm_set_target_max_io_len(ti, dm_rh_get_region_size(ms->rh));
+	if (r)
+		goto err_free_context;
+
 	ti->num_flush_requests = 1;
 	ti->num_discard_requests = 1;
+	ti->discard_zeroes_data_unsupported = 1;
 
 	ms->kmirrord_wq = alloc_workqueue("kmirrord",
 					  WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0);
@@ -1214,7 +1219,7 @@ static int mirror_end_io(struct dm_target *ti, struct bio *bio,
 	 * We need to dec pending if this was a write.
 	 */
 	if (rw == WRITE) {
-		if (!(bio->bi_rw & REQ_FLUSH))
+		if (!(bio->bi_rw & (REQ_FLUSH | REQ_DISCARD)))
 			dm_rh_dec(ms->rh, map_context->ll);
 		return error;
 	}
diff --git a/drivers/md/dm-region-hash.c b/drivers/md/dm-region-hash.c
index 7771ed21218..69732e03eb3 100644
--- a/drivers/md/dm-region-hash.c
+++ b/drivers/md/dm-region-hash.c
@@ -404,6 +404,9 @@ void dm_rh_mark_nosync(struct dm_region_hash *rh, struct bio *bio)
 		return;
 	}
 
+	if (bio->bi_rw & REQ_DISCARD)
+		return;
+
 	/* We must inform the log that the sync count has changed. */
 	log->type->set_region_sync(log, region, 0);
 
@@ -524,7 +527,7 @@ void dm_rh_inc_pending(struct dm_region_hash *rh, struct bio_list *bios)
 	struct bio *bio;
 
 	for (bio = bios->head; bio; bio = bio->bi_next) {
-		if (bio->bi_rw & REQ_FLUSH)
+		if (bio->bi_rw & (REQ_FLUSH | REQ_DISCARD))
 			continue;
 		rh_inc(rh, dm_rh_bio_to_region(rh, bio));
 	}
diff --git a/drivers/md/dm-snap.c b/drivers/md/dm-snap.c
index 6f758870fc1..6c0f3e33923 100644
--- a/drivers/md/dm-snap.c
+++ b/drivers/md/dm-snap.c
@@ -691,7 +691,7 @@ static int dm_add_exception(void *context, chunk_t old, chunk_t new)
  * Return a minimum chunk size of all snapshots that have the specified origin.
  * Return zero if the origin has no snapshots.
  */
-static sector_t __minimum_chunk_size(struct origin *o)
+static uint32_t __minimum_chunk_size(struct origin *o)
 {
 	struct dm_snapshot *snap;
 	unsigned chunk_size = 0;
@@ -701,7 +701,7 @@ static sector_t __minimum_chunk_size(struct origin *o)
 			chunk_size = min_not_zero(chunk_size,
 						  snap->store->chunk_size);
 
-	return chunk_size;
+	return (uint32_t) chunk_size;
 }
 
 /*
@@ -1172,7 +1172,10 @@ static int snapshot_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 		ti->error = "Chunk size not set";
 		goto bad_read_metadata;
 	}
-	ti->split_io = s->store->chunk_size;
+
+	r = dm_set_target_max_io_len(ti, s->store->chunk_size);
+	if (r)
+		goto bad_read_metadata;
 
 	return 0;
 
@@ -1239,7 +1242,7 @@ static void __handover_exceptions(struct dm_snapshot *snap_src,
 	snap_dest->store->snap = snap_dest;
 	snap_src->store->snap = snap_src;
 
-	snap_dest->ti->split_io = snap_dest->store->chunk_size;
+	snap_dest->ti->max_io_len = snap_dest->store->chunk_size;
 	snap_dest->valid = snap_src->valid;
 
 	/*
@@ -1817,9 +1820,9 @@ static void snapshot_resume(struct dm_target *ti)
 	up_write(&s->lock);
 }
 
-static sector_t get_origin_minimum_chunksize(struct block_device *bdev)
+static uint32_t get_origin_minimum_chunksize(struct block_device *bdev)
 {
-	sector_t min_chunksize;
+	uint32_t min_chunksize;
 
 	down_read(&_origins_lock);
 	min_chunksize = __minimum_chunk_size(__lookup_origin(bdev));
@@ -1838,9 +1841,9 @@ static void snapshot_merge_resume(struct dm_target *ti)
 	snapshot_resume(ti);
 
 	/*
-	 * snapshot-merge acts as an origin, so set ti->split_io
+	 * snapshot-merge acts as an origin, so set ti->max_io_len
 	 */
-	ti->split_io = get_origin_minimum_chunksize(s->origin->bdev);
+	ti->max_io_len = get_origin_minimum_chunksize(s->origin->bdev);
 
 	start_merge(s);
 }
@@ -2073,12 +2076,12 @@ static int origin_write_extent(struct dm_snapshot *merging_snap,
 	struct origin *o;
 
 	/*
-	 * The origin's __minimum_chunk_size() got stored in split_io
+	 * The origin's __minimum_chunk_size() got stored in max_io_len
 	 * by snapshot_merge_resume().
 	 */
 	down_read(&_origins_lock);
 	o = __lookup_origin(merging_snap->origin->bdev);
-	for (n = 0; n < size; n += merging_snap->ti->split_io)
+	for (n = 0; n < size; n += merging_snap->ti->max_io_len)
 		if (__origin_write(&o->snapshots, sector + n, NULL) ==
 		    DM_MAPIO_SUBMITTED)
 			must_wait = 1;
@@ -2138,14 +2141,14 @@ static int origin_map(struct dm_target *ti, struct bio *bio,
 }
 
 /*
- * Set the target "split_io" field to the minimum of all the snapshots'
+ * Set the target "max_io_len" field to the minimum of all the snapshots'
  * chunk sizes.
  */
 static void origin_resume(struct dm_target *ti)
 {
 	struct dm_dev *dev = ti->private;
 
-	ti->split_io = get_origin_minimum_chunksize(dev->bdev);
+	ti->max_io_len = get_origin_minimum_chunksize(dev->bdev);
 }
 
 static int origin_status(struct dm_target *ti, status_type_t type, char *result,
@@ -2176,7 +2179,6 @@ static int origin_merge(struct dm_target *ti, struct bvec_merge_data *bvm,
 		return max_size;
 
 	bvm->bi_bdev = dev->bdev;
-	bvm->bi_sector = bvm->bi_sector;
 
 	return min(max_size, q->merge_bvec_fn(q, bvm, biovec));
 }
diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c
index 35c94ff24ad..9e8f4cc63d6 100644
--- a/drivers/md/dm-stripe.c
+++ b/drivers/md/dm-stripe.c
@@ -26,14 +26,12 @@ struct stripe {
 struct stripe_c {
 	uint32_t stripes;
 	int stripes_shift;
-	sector_t stripes_mask;
 
 	/* The size of this target / num. stripes */
 	sector_t stripe_width;
 
-	/* stripe chunk size */
-	uint32_t chunk_shift;
-	sector_t chunk_mask;
+	uint32_t chunk_size;
+	int chunk_size_shift;
 
 	/* Needed for handling events */
 	struct dm_target *ti;
@@ -91,7 +89,7 @@ static int get_stripe(struct dm_target *ti, struct stripe_c *sc,
 
 /*
  * Construct a striped mapping.
- * <number of stripes> <chunk size (2^^n)> [<dev_path> <offset>]+
+ * <number of stripes> <chunk size> [<dev_path> <offset>]+
  */
 static int stripe_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 {
@@ -99,7 +97,6 @@ static int stripe_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 	sector_t width;
 	uint32_t stripes;
 	uint32_t chunk_size;
-	char *end;
 	int r;
 	unsigned int i;
 
@@ -108,34 +105,23 @@ static int stripe_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 		return -EINVAL;
 	}
 
-	stripes = simple_strtoul(argv[0], &end, 10);
-	if (!stripes || *end) {
+	if (kstrtouint(argv[0], 10, &stripes) || !stripes) {
 		ti->error = "Invalid stripe count";
 		return -EINVAL;
 	}
 
-	chunk_size = simple_strtoul(argv[1], &end, 10);
-	if (*end) {
+	if (kstrtouint(argv[1], 10, &chunk_size) || !chunk_size) {
 		ti->error = "Invalid chunk_size";
 		return -EINVAL;
 	}
 
-	/*
-	 * chunk_size is a power of two
-	 */
-	if (!is_power_of_2(chunk_size) ||
-	    (chunk_size < (PAGE_SIZE >> SECTOR_SHIFT))) {
-		ti->error = "Invalid chunk size";
-		return -EINVAL;
-	}
-
-	if (ti->len & (chunk_size - 1)) {
+	width = ti->len;
+	if (sector_div(width, chunk_size)) {
 		ti->error = "Target length not divisible by "
 		    "chunk size";
 		return -EINVAL;
 	}
 
-	width = ti->len;
 	if (sector_div(width, stripes)) {
 		ti->error = "Target length not divisible by "
 		    "number of stripes";
@@ -167,17 +153,21 @@ static int stripe_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 
 	if (stripes & (stripes - 1))
 		sc->stripes_shift = -1;
-	else {
-		sc->stripes_shift = ffs(stripes) - 1;
-		sc->stripes_mask = ((sector_t) stripes) - 1;
-	}
+	else
+		sc->stripes_shift = __ffs(stripes);
+
+	r = dm_set_target_max_io_len(ti, chunk_size);
+	if (r)
+		return r;
 
-	ti->split_io = chunk_size;
 	ti->num_flush_requests = stripes;
 	ti->num_discard_requests = stripes;
 
-	sc->chunk_shift = ffs(chunk_size) - 1;
-	sc->chunk_mask = ((sector_t) chunk_size) - 1;
+	sc->chunk_size = chunk_size;
+	if (chunk_size & (chunk_size - 1))
+		sc->chunk_size_shift = -1;
+	else
+		sc->chunk_size_shift = __ffs(chunk_size);
 
 	/*
 	 * Get the stripe destinations.
@@ -216,17 +206,29 @@ static void stripe_dtr(struct dm_target *ti)
 static void stripe_map_sector(struct stripe_c *sc, sector_t sector,
 			      uint32_t *stripe, sector_t *result)
 {
-	sector_t offset = dm_target_offset(sc->ti, sector);
-	sector_t chunk = offset >> sc->chunk_shift;
+	sector_t chunk = dm_target_offset(sc->ti, sector);
+	sector_t chunk_offset;
+
+	if (sc->chunk_size_shift < 0)
+		chunk_offset = sector_div(chunk, sc->chunk_size);
+	else {
+		chunk_offset = chunk & (sc->chunk_size - 1);
+		chunk >>= sc->chunk_size_shift;
+	}
 
 	if (sc->stripes_shift < 0)
 		*stripe = sector_div(chunk, sc->stripes);
 	else {
-		*stripe = chunk & sc->stripes_mask;
+		*stripe = chunk & (sc->stripes - 1);
 		chunk >>= sc->stripes_shift;
 	}
 
-	*result = (chunk << sc->chunk_shift) | (offset & sc->chunk_mask);
+	if (sc->chunk_size_shift < 0)
+		chunk *= sc->chunk_size;
+	else
+		chunk <<= sc->chunk_size_shift;
+
+	*result = chunk + chunk_offset;
 }
 
 static void stripe_map_range_sector(struct stripe_c *sc, sector_t sector,
@@ -237,9 +239,16 @@ static void stripe_map_range_sector(struct stripe_c *sc, sector_t sector,
 	stripe_map_sector(sc, sector, &stripe, result);
 	if (stripe == target_stripe)
 		return;
-	*result &= ~sc->chunk_mask;			/* round down */
+
+	/* round down */
+	sector = *result;
+	if (sc->chunk_size_shift < 0)
+		*result -= sector_div(sector, sc->chunk_size);
+	else
+		*result = sector & ~(sector_t)(sc->chunk_size - 1);
+
 	if (target_stripe < stripe)
-		*result += sc->chunk_mask + 1;		/* next chunk */
+		*result += sc->chunk_size;		/* next chunk */
 }
 
 static int stripe_map_discard(struct stripe_c *sc, struct bio *bio,
@@ -324,7 +333,7 @@ static int stripe_status(struct dm_target *ti,
 
 	case STATUSTYPE_TABLE:
 		DMEMIT("%d %llu", sc->stripes,
-			(unsigned long long)sc->chunk_mask + 1);
+			(unsigned long long)sc->chunk_size);
 		for (i = 0; i < sc->stripes; i++)
 			DMEMIT(" %s %llu", sc->stripe[i].dev->name,
 			    (unsigned long long)sc->stripe[i].physical_start);
@@ -391,7 +400,7 @@ static void stripe_io_hints(struct dm_target *ti,
 			    struct queue_limits *limits)
 {
 	struct stripe_c *sc = ti->private;
-	unsigned chunk_size = (sc->chunk_mask + 1) << 9;
+	unsigned chunk_size = sc->chunk_size << SECTOR_SHIFT;
 
 	blk_limits_io_min(limits, chunk_size);
 	blk_limits_io_opt(limits, chunk_size * sc->stripes);
@@ -419,7 +428,7 @@ static int stripe_merge(struct dm_target *ti, struct bvec_merge_data *bvm,
 
 static struct target_type stripe_target = {
 	.name   = "striped",
-	.version = {1, 4, 0},
+	.version = {1, 5, 0},
 	.module = THIS_MODULE,
 	.ctr    = stripe_ctr,
 	.dtr    = stripe_dtr,
diff --git a/drivers/md/dm-thin-metadata.c b/drivers/md/dm-thin-metadata.c
index 3e2907f0bc4..31f9827dfb5 100644
--- a/drivers/md/dm-thin-metadata.c
+++ b/drivers/md/dm-thin-metadata.c
@@ -80,6 +80,12 @@
 #define THIN_METADATA_CACHE_SIZE 64
 #define SECTOR_TO_BLOCK_SHIFT 3
 
+/*
+ *  3 for btree insert +
+ *  2 for btree lookup used within space map
+ */
+#define THIN_MAX_CONCURRENT_LOCKS 5
+
 /* This should be plenty */
 #define SPACE_MAP_ROOT_SIZE 128
 
@@ -597,31 +603,31 @@ static int __commit_transaction(struct dm_pool_metadata *pmd)
 
 	r = __write_changed_details(pmd);
 	if (r < 0)
-		goto out;
+		return r;
 
 	if (!pmd->need_commit)
-		goto out;
+		return r;
 
 	r = dm_sm_commit(pmd->data_sm);
 	if (r < 0)
-		goto out;
+		return r;
 
 	r = dm_tm_pre_commit(pmd->tm);
 	if (r < 0)
-		goto out;
+		return r;
 
 	r = dm_sm_root_size(pmd->metadata_sm, &metadata_len);
 	if (r < 0)
-		goto out;
+		return r;
 
 	r = dm_sm_root_size(pmd->data_sm, &data_len);
 	if (r < 0)
-		goto out;
+		return r;
 
 	r = dm_bm_write_lock(pmd->bm, THIN_SUPERBLOCK_LOCATION,
 			     &sb_validator, &sblock);
 	if (r)
-		goto out;
+		return r;
 
 	disk_super = dm_block_data(sblock);
 	disk_super->time = cpu_to_le32(pmd->time);
@@ -644,7 +650,6 @@ static int __commit_transaction(struct dm_pool_metadata *pmd)
 	if (!r)
 		pmd->need_commit = 0;
 
-out:
 	return r;
 
 out_locked:
@@ -669,13 +674,9 @@ struct dm_pool_metadata *dm_pool_metadata_open(struct block_device *bdev,
 		return ERR_PTR(-ENOMEM);
 	}
 
-	/*
-	 * Max hex locks:
-	 *  3 for btree insert +
-	 *  2 for btree lookup used within space map
-	 */
 	bm = dm_block_manager_create(bdev, THIN_METADATA_BLOCK_SIZE,
-				     THIN_METADATA_CACHE_SIZE, 5);
+				     THIN_METADATA_CACHE_SIZE,
+				     THIN_MAX_CONCURRENT_LOCKS);
 	if (!bm) {
 		DMERR("could not create block manager");
 		kfree(pmd);
@@ -1262,7 +1263,7 @@ dm_thin_id dm_thin_dev_id(struct dm_thin_device *td)
 	return td->id;
 }
 
-static int __snapshotted_since(struct dm_thin_device *td, uint32_t time)
+static bool __snapshotted_since(struct dm_thin_device *td, uint32_t time)
 {
 	return td->snapshotted_time > time;
 }
diff --git a/drivers/md/dm-thin-metadata.h b/drivers/md/dm-thin-metadata.h
index b88918ccdaf..7b47c0a9a8e 100644
--- a/drivers/md/dm-thin-metadata.h
+++ b/drivers/md/dm-thin-metadata.h
@@ -119,7 +119,7 @@ dm_thin_id dm_thin_dev_id(struct dm_thin_device *td);
 
 struct dm_thin_lookup_result {
 	dm_block_t block;
-	int shared;
+	unsigned shared:1;
 };
 
 /*
diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c
index ce59824fb41..93e3e542cff 100644
--- a/drivers/md/dm-thin.c
+++ b/drivers/md/dm-thin.c
@@ -19,7 +19,7 @@
 /*
  * Tunable constants
  */
-#define ENDIO_HOOK_POOL_SIZE 10240
+#define ENDIO_HOOK_POOL_SIZE 1024
 #define DEFERRED_SET_SIZE 64
 #define MAPPING_POOL_SIZE 1024
 #define PRISON_CELLS 1024
@@ -510,10 +510,9 @@ struct pool {
 	struct block_device *md_dev;
 	struct dm_pool_metadata *pmd;
 
-	uint32_t sectors_per_block;
-	unsigned block_shift;
-	dm_block_t offset_mask;
 	dm_block_t low_water_blocks;
+	uint32_t sectors_per_block;
+	int sectors_per_block_shift;
 
 	struct pool_features pf;
 	unsigned low_water_triggered:1;	/* A dm event has been sent */
@@ -526,8 +525,8 @@ struct pool {
 	struct work_struct worker;
 	struct delayed_work waker;
 
-	unsigned ref_count;
 	unsigned long last_commit_jiffies;
+	unsigned ref_count;
 
 	spinlock_t lock;
 	struct bio_list deferred_bios;
@@ -679,16 +678,28 @@ static void requeue_io(struct thin_c *tc)
 
 static dm_block_t get_bio_block(struct thin_c *tc, struct bio *bio)
 {
-	return bio->bi_sector >> tc->pool->block_shift;
+	sector_t block_nr = bio->bi_sector;
+
+	if (tc->pool->sectors_per_block_shift < 0)
+		(void) sector_div(block_nr, tc->pool->sectors_per_block);
+	else
+		block_nr >>= tc->pool->sectors_per_block_shift;
+
+	return block_nr;
 }
 
 static void remap(struct thin_c *tc, struct bio *bio, dm_block_t block)
 {
 	struct pool *pool = tc->pool;
+	sector_t bi_sector = bio->bi_sector;
 
 	bio->bi_bdev = tc->pool_dev->bdev;
-	bio->bi_sector = (block << pool->block_shift) +
-		(bio->bi_sector & pool->offset_mask);
+	if (tc->pool->sectors_per_block_shift < 0)
+		bio->bi_sector = (block * pool->sectors_per_block) +
+				 sector_div(bi_sector, pool->sectors_per_block);
+	else
+		bio->bi_sector = (block << pool->sectors_per_block_shift) |
+				(bi_sector & (pool->sectors_per_block - 1));
 }
 
 static void remap_to_origin(struct thin_c *tc, struct bio *bio)
@@ -933,9 +944,7 @@ static void process_prepared(struct pool *pool, struct list_head *head,
  */
 static int io_overlaps_block(struct pool *pool, struct bio *bio)
 {
-	return !(bio->bi_sector & pool->offset_mask) &&
-		(bio->bi_size == (pool->sectors_per_block << SECTOR_SHIFT));
-
+	return bio->bi_size == (pool->sectors_per_block << SECTOR_SHIFT);
 }
 
 static int io_overwrites_block(struct pool *pool, struct bio *bio)
@@ -1218,7 +1227,7 @@ static void process_discard(struct thin_c *tc, struct bio *bio)
 			 */
 			m = get_next_mapping(pool);
 			m->tc = tc;
-			m->pass_discard = (!lookup_result.shared) & pool->pf.discard_passdown;
+			m->pass_discard = (!lookup_result.shared) && pool->pf.discard_passdown;
 			m->virt_block = block;
 			m->data_block = lookup_result.block;
 			m->cell = cell;
@@ -1234,18 +1243,16 @@ static void process_discard(struct thin_c *tc, struct bio *bio)
 			}
 		} else {
 			/*
-			 * This path is hit if people are ignoring
-			 * limits->discard_granularity.  It ignores any
-			 * part of the discard that is in a subsequent
-			 * block.
+			 * The DM core makes sure that the discard doesn't span
+			 * a block boundary.  So we submit the discard of a
+			 * partial block appropriately.
 			 */
-			sector_t offset = bio->bi_sector - (block << pool->block_shift);
-			unsigned remaining = (pool->sectors_per_block - offset) << 9;
-			bio->bi_size = min(bio->bi_size, remaining);
-
 			cell_release_singleton(cell, bio);
 			cell_release_singleton(cell2, bio);
-			remap_and_issue(tc, bio, lookup_result.block);
+			if ((!lookup_result.shared) && pool->pf.discard_passdown)
+				remap_and_issue(tc, bio, lookup_result.block);
+			else
+				bio_endio(bio, 0);
 		}
 		break;
 
@@ -1719,8 +1726,10 @@ static struct pool *pool_create(struct mapped_device *pool_md,
 
 	pool->pmd = pmd;
 	pool->sectors_per_block = block_size;
-	pool->block_shift = ffs(block_size) - 1;
-	pool->offset_mask = block_size - 1;
+	if (block_size & (block_size - 1))
+		pool->sectors_per_block_shift = -1;
+	else
+		pool->sectors_per_block_shift = __ffs(block_size);
 	pool->low_water_blocks = 0;
 	pool_features_init(&pool->pf);
 	pool->prison = prison_create(PRISON_CELLS);
@@ -1825,15 +1834,19 @@ static struct pool *__pool_find(struct mapped_device *pool_md,
 	struct pool *pool = __pool_table_lookup_metadata_dev(metadata_dev);
 
 	if (pool) {
-		if (pool->pool_md != pool_md)
+		if (pool->pool_md != pool_md) {
+			*error = "metadata device already in use by a pool";
 			return ERR_PTR(-EBUSY);
+		}
 		__pool_inc(pool);
 
 	} else {
 		pool = __pool_table_lookup(pool_md);
 		if (pool) {
-			if (pool->md_dev != metadata_dev)
+			if (pool->md_dev != metadata_dev) {
+				*error = "different pool cannot replace a pool";
 				return ERR_PTR(-EINVAL);
+			}
 			__pool_inc(pool);
 
 		} else {
@@ -1964,7 +1977,7 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv)
 	if (kstrtoul(argv[2], 10, &block_size) || !block_size ||
 	    block_size < DATA_DEV_BLOCK_SIZE_MIN_SECTORS ||
 	    block_size > DATA_DEV_BLOCK_SIZE_MAX_SECTORS ||
-	    !is_power_of_2(block_size)) {
+	    block_size & (DATA_DEV_BLOCK_SIZE_MIN_SEC