Merge tag 'for-6.0/dm-changes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

Pull more device mapper updates from Mike Snitzer:

 - Add flags argument to dm_bufio_client_create and introduce
   DM_BUFIO_CLIENT_NO_SLEEP flag to have dm-bufio use spinlock rather
   than mutex for its locking.

 - Add optional "try_verify_in_tasklet" feature to DM verity target.
   This feature gives users the option to improve IO latency by using a
   tasklet to verify, using hashes in bufio's cache, rather than wait to
   schedule a work item via workqueue. But if there is a bufio cache
   miss, or an error, then the tasklet will fallback to using workqueue.

 - Incremental changes to both dm-bufio and the DM verity target to use
   jump_label to minimize cost of branching associated with the niche
   "try_verify_in_tasklet" feature. DM-bufio in particular is used by
   quite a few other DM targets so it doesn't make sense to incur
   additional bufio cost in those targets purely for the benefit of this
   niche verity feature if the feature isn't ever used.

 - Optimize verity_verify_io, which is used by both workqueue and
   tasklet based verification, if FEC is not configured or tasklet based
   verification isn't used.

 - Remove DM verity target's verify_wq's use of the WQ_CPU_INTENSIVE
   flag since it uses WQ_UNBOUND. Also, use the WQ_HIGHPRI flag if
   "try_verify_in_tasklet" is specified.

* tag 'for-6.0/dm-changes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm verity: have verify_wq use WQ_HIGHPRI if "try_verify_in_tasklet"
  dm verity: remove WQ_CPU_INTENSIVE flag since using WQ_UNBOUND
  dm verity: only copy bvec_iter in verity_verify_io if in_tasklet
  dm verity: optimize verity_verify_io if FEC not configured
  dm verity: conditionally enable branching for "try_verify_in_tasklet"
  dm bufio: conditionally enable branching for DM_BUFIO_CLIENT_NO_SLEEP
  dm verity: allow optional args to alter primary args handling
  dm verity: Add optional "try_verify_in_tasklet" feature
  dm bufio: Add DM_BUFIO_CLIENT_NO_SLEEP flag
  dm bufio: Add flags argument to dm_bufio_client_create

9 files changed, 187 insertions, 33 deletions

diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c
index 514a802e499b..acd6d6b47434 100644
--- a/drivers/md/dm-bufio.c
+++ b/drivers/md/dm-bufio.c
@@ -18,6 +18,7 @@
 #include <linux/module.h>
 #include <linux/rbtree.h>
 #include <linux/stacktrace.h>
+#include <linux/jump_label.h>
 
 #define DM_MSG_PREFIX "bufio"
 
@@ -81,6 +82,8 @@
  */
 struct dm_bufio_client {
 	struct mutex lock;
+	spinlock_t spinlock;
+	unsigned long spinlock_flags;
 
 	struct list_head lru[LIST_SIZE];
 	unsigned long n_buffers[LIST_SIZE];
@@ -90,6 +93,7 @@ struct dm_bufio_client {
 	s8 sectors_per_block_bits;
 	void (*alloc_callback)(struct dm_buffer *);
 	void (*write_callback)(struct dm_buffer *);
+	bool no_sleep;
 
 	struct kmem_cache *slab_buffer;
 	struct kmem_cache *slab_cache;
@@ -161,23 +165,34 @@ struct dm_buffer {
 #endif
 };
 
+static DEFINE_STATIC_KEY_FALSE(no_sleep_enabled);
+
 /*----------------------------------------------------------------*/
 
 #define dm_bufio_in_request()	(!!current->bio_list)
 
 static void dm_bufio_lock(struct dm_bufio_client *c)
 {
-	mutex_lock_nested(&c->lock, dm_bufio_in_request());
+	if (static_branch_unlikely(&no_sleep_enabled) && c->no_sleep)
+		spin_lock_irqsave_nested(&c->spinlock, c->spinlock_flags, dm_bufio_in_request());
+	else
+		mutex_lock_nested(&c->lock, dm_bufio_in_request());
 }
 
 static int dm_bufio_trylock(struct dm_bufio_client *c)
 {
-	return mutex_trylock(&c->lock);
+	if (static_branch_unlikely(&no_sleep_enabled) && c->no_sleep)
+		return spin_trylock_irqsave(&c->spinlock, c->spinlock_flags);
+	else
+		return mutex_trylock(&c->lock);
 }
 
 static void dm_bufio_unlock(struct dm_bufio_client *c)
 {
-	mutex_unlock(&c->lock);
+	if (static_branch_unlikely(&no_sleep_enabled) && c->no_sleep)
+		spin_unlock_irqrestore(&c->spinlock, c->spinlock_flags);
+	else
+		mutex_unlock(&c->lock);
 }
 
 /*----------------------------------------------------------------*/
@@ -1715,7 +1730,8 @@ static unsigned long dm_bufio_shrink_count(struct shrinker *shrink, struct shrin
 struct dm_bufio_client *dm_bufio_client_create(struct block_device *bdev, unsigned block_size,
 					       unsigned reserved_buffers, unsigned aux_size,
 					       void (*alloc_callback)(struct dm_buffer *),
-					       void (*write_callback)(struct dm_buffer *))
+					       void (*write_callback)(struct dm_buffer *),
+					       unsigned int flags)
 {
 	int r;
 	struct dm_bufio_client *c;
@@ -1745,12 +1761,18 @@ struct dm_bufio_client *dm_bufio_client_create(struct block_device *bdev, unsign
 	c->alloc_callback = alloc_callback;
 	c->write_callback = write_callback;
 
+	if (flags & DM_BUFIO_CLIENT_NO_SLEEP) {
+		c->no_sleep = true;
+		static_branch_inc(&no_sleep_enabled);
+	}
+
 	for (i = 0; i < LIST_SIZE; i++) {
 		INIT_LIST_HEAD(&c->lru[i]);
 		c->n_buffers[i] = 0;
 	}
 
 	mutex_init(&c->lock);
+	spin_lock_init(&c->spinlock);
 	INIT_LIST_HEAD(&c->reserved_buffers);
 	c->need_reserved_buffers = reserved_buffers;
 
@@ -1877,6 +1899,8 @@ void dm_bufio_client_destroy(struct dm_bufio_client *c)
 	kmem_cache_destroy(c->slab_buffer);
 	dm_io_client_destroy(c->dm_io);
 	mutex_destroy(&c->lock);
+	if (c->no_sleep)
+		static_branch_dec(&no_sleep_enabled);
 	kfree(c);
 }
 EXPORT_SYMBOL_GPL(dm_bufio_client_destroy);
diff --git a/drivers/md/dm-ebs-target.c b/drivers/md/dm-ebs-target.c
index 223e8e1a7a13..512cc6cea095 100644
--- a/drivers/md/dm-ebs-target.c
+++ b/drivers/md/dm-ebs-target.c
@@ -313,7 +313,8 @@ static int ebs_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 		goto bad;
 	}
 
-	ec->bufio = dm_bufio_client_create(ec->dev->bdev, to_bytes(ec->u_bs), 1, 0, NULL, NULL);
+	ec->bufio = dm_bufio_client_create(ec->dev->bdev, to_bytes(ec->u_bs), 1,
+					   0, NULL, NULL, 0);
 	if (IS_ERR(ec->bufio)) {
 		ti->error = "Cannot create dm bufio client";
 		r = PTR_ERR(ec->bufio);
diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c
index c60f9b2ece2d..aaf2472df6e5 100644
--- a/drivers/md/dm-integrity.c
+++ b/drivers/md/dm-integrity.c
@@ -4439,7 +4439,7 @@ try_smaller_buffer:
 	}
 
 	ic->bufio = dm_bufio_client_create(ic->meta_dev ? ic->meta_dev->bdev : ic->dev->bdev,
-			1U << (SECTOR_SHIFT + ic->log2_buffer_sectors), 1, 0, NULL, NULL);
+			1U << (SECTOR_SHIFT + ic->log2_buffer_sectors), 1, 0, NULL, NULL, 0);
 	if (IS_ERR(ic->bufio)) {
 		r = PTR_ERR(ic->bufio);
 		ti->error = "Cannot initialize dm-bufio";
diff --git a/drivers/md/dm-snap-persistent.c b/drivers/md/dm-snap-persistent.c
index f46f930eedf9..680cc05ec654 100644
--- a/drivers/md/dm-snap-persistent.c
+++ b/drivers/md/dm-snap-persistent.c
@@ -493,7 +493,7 @@ static int read_exceptions(struct pstore *ps,
 
 	client = dm_bufio_client_create(dm_snap_cow(ps->store->snap)->bdev,
 					ps->store->chunk_size << SECTOR_SHIFT,
-					1, 0, NULL, NULL);
+					1, 0, NULL, NULL, 0);
 
 	if (IS_ERR(client))
 		return PTR_ERR(client);
diff --git a/drivers/md/dm-verity-fec.c b/drivers/md/dm-verity-fec.c
index cea2b3789736..23cffce56403 100644
--- a/drivers/md/dm-verity-fec.c
+++ b/drivers/md/dm-verity-fec.c
@@ -749,7 +749,7 @@ int verity_fec_ctr(struct dm_verity *v)
 
 	f->bufio = dm_bufio_client_create(f->dev->bdev,
 					  f->io_size,
-					  1, 0, NULL, NULL);
+					  1, 0, NULL, NULL, 0);
 	if (IS_ERR(f->bufio)) {
 		ti->error = "Cannot initialize FEC bufio client";
 		return PTR_ERR(f->bufio);
@@ -765,7 +765,7 @@ int verity_fec_ctr(struct dm_verity *v)
 
 	f->data_bufio = dm_bufio_client_create(v->data_dev->bdev,
 					       1 << v->data_dev_block_bits,
-					       1, 0, NULL, NULL);
+					       1, 0, NULL, NULL, 0);
 	if (IS_ERR(f->data_bufio)) {
 		ti->error = "Cannot initialize FEC data bufio client";
 		return PTR_ERR(f->data_bufio);
diff --git a/drivers/md/dm-verity-target.c b/drivers/md/dm-verity-target.c
index 4fd853a56b1a..2347e83902f1 100644
--- a/drivers/md/dm-verity-target.c
+++ b/drivers/md/dm-verity-target.c
@@ -20,6 +20,7 @@
 #include <linux/reboot.h>
 #include <linux/scatterlist.h>
 #include <linux/string.h>
+#include <linux/jump_label.h>
 
 #define DM_MSG_PREFIX			"verity"
 
@@ -35,6 +36,7 @@
 #define DM_VERITY_OPT_PANIC		"panic_on_corruption"
 #define DM_VERITY_OPT_IGN_ZEROES	"ignore_zero_blocks"
 #define DM_VERITY_OPT_AT_MOST_ONCE	"check_at_most_once"
+#define DM_VERITY_OPT_TASKLET_VERIFY	"try_verify_in_tasklet"
 
 #define DM_VERITY_OPTS_MAX		(3 + DM_VERITY_OPTS_FEC + \
 					 DM_VERITY_ROOT_HASH_VERIFICATION_OPTS)
@@ -43,6 +45,8 @@ static unsigned dm_verity_prefetch_cluster = DM_VERITY_DEFAULT_PREFETCH_SIZE;
 
 module_param_named(prefetch_cluster, dm_verity_prefetch_cluster, uint, S_IRUGO | S_IWUSR);
 
+static DEFINE_STATIC_KEY_FALSE(use_tasklet_enabled);
+
 struct dm_verity_prefetch_work {
 	struct work_struct work;
 	struct dm_verity *v;
@@ -221,7 +225,7 @@ static int verity_handle_err(struct dm_verity *v, enum verity_block_type type,
 	struct mapped_device *md = dm_table_get_md(v->ti->table);
 
 	/* Corruption should be visible in device status in all modes */
-	v->hash_failed = 1;
+	v->hash_failed = true;
 
 	if (v->corrupted_errs >= DM_VERITY_MAX_CORRUPTED_ERRS)
 		goto out;
@@ -287,7 +291,19 @@ static int verity_verify_level(struct dm_verity *v, struct dm_verity_io *io,
 
 	verity_hash_at_level(v, block, level, &hash_block, &offset);
 
-	data = dm_bufio_read(v->bufio, hash_block, &buf);
+	if (static_branch_unlikely(&use_tasklet_enabled) && io->in_tasklet) {
+		data = dm_bufio_get(v->bufio, hash_block, &buf);
+		if (data == NULL) {
+			/*
+			 * In tasklet and the hash was not in the bufio cache.
+			 * Return early and resume execution from a work-queue
+			 * to read the hash from disk.
+			 */
+			return -EAGAIN;
+		}
+	} else
+		data = dm_bufio_read(v->bufio, hash_block, &buf);
+
 	if (IS_ERR(data))
 		return PTR_ERR(data);
 
@@ -308,6 +324,15 @@ static int verity_verify_level(struct dm_verity *v, struct dm_verity_io *io,
 		if (likely(memcmp(verity_io_real_digest(v, io), want_digest,
 				  v->digest_size) == 0))
 			aux->hash_verified = 1;
+		else if (static_branch_unlikely(&use_tasklet_enabled) &&
+			 io->in_tasklet) {
+			/*
+			 * Error handling code (FEC included) cannot be run in a
+			 * tasklet since it may sleep, so fallback to work-queue.
+			 */
+			r = -EAGAIN;
+			goto release_ret_r;
+		}
 		else if (verity_fec_decode(v, io,
 					   DM_VERITY_BLOCK_TYPE_METADATA,
 					   hash_block, data, NULL) == 0)
@@ -474,10 +499,24 @@ static int verity_verify_io(struct dm_verity_io *io)
 {
 	bool is_zero;
 	struct dm_verity *v = io->v;
+#if defined(CONFIG_DM_VERITY_FEC)
 	struct bvec_iter start;
-	unsigned b;
+#endif
+	struct bvec_iter iter_copy;
+	struct bvec_iter *iter;
 	struct crypto_wait wait;
 	struct bio *bio = dm_bio_from_per_bio_data(io, v->ti->per_io_data_size);
+	unsigned int b;
+
+	if (static_branch_unlikely(&use_tasklet_enabled) && io->in_tasklet) {
+		/*
+		 * Copy the iterator in case we need to restart
+		 * verification in a work-queue.
+		 */
+		iter_copy = io->iter;
+		iter = &iter_copy;
+	} else
+		iter = &io->iter;
 
 	for (b = 0; b < io->n_blocks; b++) {
 		int r;
@@ -486,7 +525,7 @@ static int verity_verify_io(struct dm_verity_io *io)
 
 		if (v->validated_blocks &&
 		    likely(test_bit(cur_block, v->validated_blocks))) {
-			verity_bv_skip_block(v, io, &io->iter);
+			verity_bv_skip_block(v, io, iter);
 			continue;
 		}
 
@@ -501,7 +540,7 @@ static int verity_verify_io(struct dm_verity_io *io)
 			 * If we expect a zero block, don't validate, just
 			 * return zeros.
 			 */
-			r = verity_for_bv_block(v, io, &io->iter,
+			r = verity_for_bv_block(v, io, iter,
 						verity_bv_zero);
 			if (unlikely(r < 0))
 				return r;
@@ -513,8 +552,11 @@ static int verity_verify_io(struct dm_verity_io *io)
 		if (unlikely(r < 0))
 			return r;
 
-		start = io->iter;
-		r = verity_for_io_block(v, io, &io->iter, &wait);
+#if defined(CONFIG_DM_VERITY_FEC)
+		if (verity_fec_is_enabled(v))
+			start = *iter;
+#endif
+		r = verity_for_io_block(v, io, iter, &wait);
 		if (unlikely(r < 0))
 			return r;
 
@@ -528,9 +570,18 @@ static int verity_verify_io(struct dm_verity_io *io)
 			if (v->validated_blocks)
 				set_bit(cur_block, v->validated_blocks);
 			continue;
+		} else if (static_branch_unlikely(&use_tasklet_enabled) &&
+			   io->in_tasklet) {
+			/*
+			 * Error handling code (FEC included) cannot be run in a
+			 * tasklet since it may sleep, so fallback to work-queue.
+			 */
+			return -EAGAIN;
+#if defined(CONFIG_DM_VERITY_FEC)
 		} else if (verity_fec_decode(v, io, DM_VERITY_BLOCK_TYPE_DATA,
-					   cur_block, NULL, &start) == 0) {
+					     cur_block, NULL, &start) == 0) {
 			continue;
+#endif
 		} else {
 			if (bio->bi_status) {
 				/*
@@ -567,7 +618,8 @@ static void verity_finish_io(struct dm_verity_io *io, blk_status_t status)
 	bio->bi_end_io = io->orig_bi_end_io;
 	bio->bi_status = status;
 
-	verity_fec_finish_io(io);
+	if (!static_branch_unlikely(&use_tasklet_enabled) || !io->in_tasklet)
+		verity_fec_finish_io(io);
 
 	bio_endio(bio);
 }
@@ -576,9 +628,29 @@ static void verity_work(struct work_struct *w)
 {
 	struct dm_verity_io *io = container_of(w, struct dm_verity_io, work);
 
+	io->in_tasklet = false;
+
+	verity_fec_init_io(io);
 	verity_finish_io(io, errno_to_blk_status(verity_verify_io(io)));
 }
 
+static void verity_tasklet(unsigned long data)
+{
+	struct dm_verity_io *io = (struct dm_verity_io *)data;
+	int err;
+
+	io->in_tasklet = true;
+	err = verity_verify_io(io);
+	if (err == -EAGAIN) {
+		/* fallback to retrying with work-queue */
+		INIT_WORK(&io->work, verity_work);
+		queue_work(io->v->verify_wq, &io->work);
+		return;
+	}
+
+	verity_finish_io(io, errno_to_blk_status(err));
+}
+
 static void verity_end_io(struct bio *bio)
 {
 	struct dm_verity_io *io = bio->bi_private;
@@ -589,8 +661,13 @@ static void verity_end_io(struct bio *bio)
 		return;
 	}
 
-	INIT_WORK(&io->work, verity_work);
-	queue_work(io->v->verify_wq, &io->work);
+	if (static_branch_unlikely(&use_tasklet_enabled) && io->v->use_tasklet) {
+		tasklet_init(&io->tasklet, verity_tasklet, (unsigned long)io);
+		tasklet_schedule(&io->tasklet);
+	} else {
+		INIT_WORK(&io->work, verity_work);
+		queue_work(io->v->verify_wq, &io->work);
+	}
 }
 
 /*
@@ -701,8 +778,6 @@ static int verity_map(struct dm_target *ti, struct bio *bio)
 	bio->bi_private = io;
 	io->iter = bio->bi_iter;
 
-	verity_fec_init_io(io);
-
 	verity_submit_prefetch(v, io);
 
 	submit_bio_noacct(bio);
@@ -752,6 +827,8 @@ static void verity_status(struct dm_target *ti, status_type_t type,
 			args++;
 		if (v->validated_blocks)
 			args++;
+		if (v->use_tasklet)
+			args++;
 		if (v->signature_key_desc)
 			args += DM_VERITY_ROOT_HASH_VERIFICATION_OPTS;
 		if (!args)
@@ -777,6 +854,8 @@ static void verity_status(struct dm_target *ti, status_type_t type,
 			DMEMIT(" " DM_VERITY_OPT_IGN_ZEROES);
 		if (v->validated_blocks)
 			DMEMIT(" " DM_VERITY_OPT_AT_MOST_ONCE);
+		if (v->use_tasklet)
+			DMEMIT(" " DM_VERITY_OPT_TASKLET_VERIFY);
 		sz = verity_fec_status_table(v, sz, result, maxlen);
 		if (v->signature_key_desc)
 			DMEMIT(" " DM_VERITY_ROOT_HASH_VERIFICATION_OPT_SIG_KEY
@@ -890,6 +969,9 @@ static void verity_dtr(struct dm_target *ti)
 
 	kfree(v->signature_key_desc);
 
+	if (v->use_tasklet)
+		static_branch_dec(&use_tasklet_enabled);
+
 	kfree(v);
 }
 
@@ -968,7 +1050,8 @@ static int verity_parse_verity_mode(struct dm_verity *v, const char *arg_name)
 }
 
 static int verity_parse_opt_args(struct dm_arg_set *as, struct dm_verity *v,
-				 struct dm_verity_sig_opts *verify_args)
+				 struct dm_verity_sig_opts *verify_args,
+				 bool only_modifier_opts)
 {
 	int r;
 	unsigned argc;
@@ -991,6 +1074,8 @@ static int verity_parse_opt_args(struct dm_arg_set *as, struct dm_verity *v,
 		argc--;
 
 		if (verity_is_verity_mode(arg_name)) {
+			if (only_modifier_opts)
+				continue;
 			r = verity_parse_verity_mode(v, arg_name);
 			if (r) {
 				ti->error = "Conflicting error handling parameters";
@@ -999,6 +1084,8 @@ static int verity_parse_opt_args(struct dm_arg_set *as, struct dm_verity *v,
 			continue;
 
 		} else if (!strcasecmp(arg_name, DM_VERITY_OPT_IGN_ZEROES)) {
+			if (only_modifier_opts)
+				continue;
 			r = verity_alloc_zero_digest(v);
 			if (r) {
 				ti->error = "Cannot allocate zero digest";
@@ -1007,24 +1094,35 @@ static int verity_parse_opt_args(struct dm_arg_set *as, struct dm_verity *v,
 			continue;
 
 		} else if (!strcasecmp(arg_name, DM_VERITY_OPT_AT_MOST_ONCE)) {
+			if (only_modifier_opts)
+				continue;
 			r = verity_alloc_most_once(v);
 			if (r)
 				return r;
 			continue;
 
+		} else if (!strcasecmp(arg_name, DM_VERITY_OPT_TASKLET_VERIFY)) {
+			v->use_tasklet = true;
+			static_branch_inc(&use_tasklet_enabled);
+			continue;
+
 		} else if (verity_is_fec_opt_arg(arg_name)) {
+			if (only_modifier_opts)
+				continue;
 			r = verity_fec_parse_opt_args(as, v, &argc, arg_name);
 			if (r)
 				return r;
 			continue;
+
 		} else if (verity_verify_is_sig_opt_arg(arg_name)) {
+			if (only_modifier_opts)
+				continue;
 			r = verity_verify_sig_parse_opt_args(as, v,
 							     verify_args,
 							     &argc, arg_name);
 			if (r)
 				return r;
 			continue;
-
 		}
 
 		ti->error = "Unrecognized verity feature request";
@@ -1054,6 +1152,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 	struct dm_verity_sig_opts verify_args = {0};
 	struct dm_arg_set as;
 	unsigned int num;
+	unsigned int wq_flags;
 	unsigned long long num_ll;
 	int r;
 	int i;
@@ -1085,6 +1184,15 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 		goto bad;
 	}
 
+	/* Parse optional parameters that modify primary args */
+	if (argc > 10) {
+		as.argc = argc - 10;
+		as.argv = argv + 10;
+		r = verity_parse_opt_args(&as, v, &verify_args, true);
+		if (r < 0)
+			goto bad;
+	}
+
 	if (sscanf(argv[0], "%u%c", &num, &dummy) != 1 ||
 	    num > 1) {
 		ti->error = "Invalid version";
@@ -1156,7 +1264,8 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 		goto bad;
 	}
 
-	v->tfm = crypto_alloc_ahash(v->alg_name, 0, 0);
+	v->tfm = crypto_alloc_ahash(v->alg_name, 0,
+				    v->use_tasklet ? CRYPTO_ALG_ASYNC : 0);
 	if (IS_ERR(v->tfm)) {
 		ti->error = "Cannot initialize hash function";
 		r = PTR_ERR(v->tfm);
@@ -1218,8 +1327,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 	if (argc) {
 		as.argc = argc;
 		as.argv = argv;
-
-		r = verity_parse_opt_args(&as, v, &verify_args);
+		r = verity_parse_opt_args(&as, v, &verify_args, false);
 		if (r < 0)
 			goto bad;
 	}
@@ -1266,7 +1374,8 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
 	v->bufio = dm_bufio_client_create(v->hash_dev->bdev,
 		1 << v->hash_dev_block_bits, 1, sizeof(struct buffer_aux),
-		dm_bufio_alloc_callback, NULL);
+		dm_bufio_alloc_callback, NULL,
+		v->use_tasklet ? DM_BUFIO_CLIENT_NO_SLEEP : 0);
 	if (IS_ERR(v->bufio)) {
 		ti->error = "Cannot initialize dm-bufio";
 		r = PTR_ERR(v->bufio);
@@ -1281,7 +1390,16 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 	}
 
 	/* WQ_UNBOUND greatly improves performance when running on ramdisk */
-	v->verify_wq = alloc_workqueue("kverityd", WQ_CPU_INTENSIVE | WQ_MEM_RECLAIM | WQ_UNBOUND, num_online_cpus());
+	wq_flags = WQ_MEM_RECLAIM | WQ_UNBOUND;
+	if (v->use_tasklet) {
+		/*
+		 * Allow verify_wq to preempt softirq since verification in
+		 * tasklet will fall-back to using it for error handling
+		 * (or if the bufio cache doesn't have required hashes).
+		 */
+		wq_flags |= WQ_HIGHPRI;
+	}
+	v->verify_wq = alloc_workqueue("kverityd", wq_flags, num_online_cpus());
 	if (!v->verify_wq) {
 		ti->error = "Cannot allocate workqueue";
 		r = -ENOMEM;
@@ -1343,7 +1461,7 @@ int dm_verity_get_root_digest(struct dm_target *ti, u8 **root_digest, unsigned i
 static struct target_type verity_target = {
 	.name		= "verity",
 	.features	= DM_TARGET_IMMUTABLE,
-	.version	= {1, 8, 1},
+	.version	= {1, 9, 0},
 	.module		= THIS_MODULE,
 	.ctr		= verity_ctr,
 	.dtr		= verity_dtr,
diff --git a/drivers/md/dm-verity.h b/drivers/md/dm-verity.h
index c832cc3e3d24..45455de1b4bc 100644
--- a/drivers/md/dm-verity.h
+++ b/drivers/md/dm-verity.h
@@ -13,6 +13,7 @@
 
 #include <linux/dm-bufio.h>
 #include <linux/device-mapper.h>
+#include <linux/interrupt.h>
 #include <crypto/hash.h>
 
 #define DM_VERITY_MAX_LEVELS		63
@@ -51,9 +52,10 @@ struct dm_verity {
 	unsigned char hash_per_block_bits;	/* log2(hashes in hash block) */
 	unsigned char levels;	/* the number of tree levels */
 	unsigned char version;
+	bool hash_failed:1;	/* set if hash of any block failed */
+	bool use_tasklet:1;	/* try to verify in tasklet before work-queue */
 	unsigned digest_size;	/* digest size for the current hash algorithm */
 	unsigned int ahash_reqsize;/* the size of temporary space for crypto */
-	int hash_failed;	/* set to 1 if hash of any block failed */
 	enum verity_mode mode;	/* mode for handling verification errors */
 	unsigned corrupted_errs;/* Number of errors for corrupted blocks */
 
@@ -76,10 +78,12 @@ struct dm_verity_io {
 
 	sector_t block;
 	unsigned n_blocks;
+	bool in_tasklet;
 
 	struct bvec_iter iter;
 
 	struct work_struct work;
+	struct tasklet_struct tasklet;
 
 	/*
 	 * Three variably-size fields follow this struct:
diff --git a/drivers/md/persistent-data/dm-block-manager.c b/drivers/md/persistent-data/dm-block-manager.c
index 54c089a50b15..11935864f50f 100644
--- a/drivers/md/persistent-data/dm-block-manager.c
+++ b/drivers/md/persistent-data/dm-block-manager.c
@@ -391,7 +391,8 @@ struct dm_block_manager *dm_block_manager_create(struct block_device *bdev,
 	bm->bufio = dm_bufio_client_create(bdev, block_size, max_held_per_thread,
 					   sizeof(struct buffer_aux),
 					   dm_block_manager_alloc_callback,
-					   dm_block_manager_write_callback);
+					   dm_block_manager_write_callback,
+					   0);
 	if (IS_ERR(bm->bufio)) {
 		r = PTR_ERR(bm->bufio);
 		kfree(bm);
diff --git a/include/linux/dm-bufio.h b/include/linux/dm-bufio.h
index 90bd558a17f5..15d9e15ca830 100644
--- a/include/linux/dm-bufio.h
+++ b/include/linux/dm-bufio.h
@@ -18,13 +18,19 @@ struct dm_bufio_client;
 struct dm_buffer;
 
 /*
+ * Flags for dm_bufio_client_create
+ */
+#define DM_BUFIO_CLIENT_NO_SLEEP 0x1
+
+/*
  * Create a buffered IO cache on a given device
  */
 struct dm_bufio_client *
 dm_bufio_client_create(struct block_device *bdev, unsigned block_size,
 		       unsigned reserved_buffers, unsigned aux_size,
 		       void (*alloc_callback)(struct dm_buffer *),
-		       void (*write_callback)(struct dm_buffer *));
+		       void (*write_callback)(struct dm_buffer *),
+		       unsigned int flags);
 
 /*
  * Release a buffered IO cache.