Merge tag 'for-5.18/dm-changes' of git://git.kernel.org/pub/scm/linux…

…/kernel/git/device-mapper/linux-dm

Pull device mapper updates from Mike Snitzer:

 - Significant refactoring and fixing of how DM core does bio-based IO
   accounting with focus on fixing wildly inaccurate IO stats for
   dm-crypt (and other DM targets that defer bio submission in their own
   workqueues). End result is proper IO accounting, made possible by
   targets being updated to use the new dm_submit_bio_remap() interface.

 - Add hipri bio polling support (REQ_POLLED) to bio-based DM.

 - Reduce dm_io and dm_target_io structs so that a single dm_io (which
   contains dm_target_io and first clone bio) weighs in at 256 bytes.
   For reference the bio struct is 128 bytes.

 - Various other small cleanups, fixes or improvements in DM core and
   targets.

 - Update MAINTAINERS with my kernel.org email address to allow
   distinction between my "upstream" and "Red" Hats.

* tag 'for-5.18/dm-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: (46 commits)
  dm: consolidate spinlocks in dm_io struct
  dm: reduce size of dm_io and dm_target_io structs
  dm: switch dm_target_io booleans over to proper flags
  dm: switch dm_io booleans over to proper flags
  dm: update email address in MAINTAINERS
  dm: return void from __send_empty_flush
  dm: factor out dm_io_complete
  dm cache: use dm_submit_bio_remap
  dm: simplify dm_sumbit_bio_remap interface
  dm thin: use dm_submit_bio_remap
  dm: add WARN_ON_ONCE to dm_submit_bio_remap
  dm: support bio polling
  block: add ->poll_bio to block_device_operations
  dm mpath: use DMINFO instead of printk with KERN_INFO
  dm: stop using bdevname
  dm-zoned: remove the ->name field in struct dmz_dev
  dm: remove unnecessary local variables in __bind
  dm: requeue IO if mapping table not yet available
  dm io: remove stale comment block for dm_io()
  dm thin metadata: remove unused dm_thin_remove_block and __remove
  ...

MAINTAINERS

@@ -5605,7 +5605,7 @@ F:	include/linux/devm-helpers.h
 
 DEVICE-MAPPER  (LVM)
 M:	Alasdair Kergon <agk@redhat.com>
-M:	Mike Snitzer <snitzer@redhat.com>
+M:	Mike Snitzer <snitzer@kernel.org>
 M:	dm-devel@redhat.com
 L:	dm-devel@redhat.com
 S:	Maintained

block/blk-core.c

@@ -688,7 +688,7 @@ static void __submit_bio(struct bio *bio)
  *
  * bio_list_on_stack[0] contains bios submitted by the current ->submit_bio.
  * bio_list_on_stack[1] contains bios that were submitted before the current
- *	->submit_bio_bio, but that haven't been processed yet.
+ *	->submit_bio, but that haven't been processed yet.
  */
 static void __submit_bio_noacct(struct bio *bio)
 {
@@ -955,7 +955,7 @@ int bio_poll(struct bio *bio, struct io_comp_batch *iob, unsigned int flags)
 {
 	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
 	blk_qc_t cookie = READ_ONCE(bio->bi_cookie);
-	int ret;
+	int ret = 0;
 
 	if (cookie == BLK_QC_T_NONE ||
 	    !test_bit(QUEUE_FLAG_POLL, &q->queue_flags))
@@ -965,10 +965,14 @@ int bio_poll(struct bio *bio, struct io_comp_batch *iob, unsigned int flags)
 
 	if (blk_queue_enter(q, BLK_MQ_REQ_NOWAIT))
 		return 0;
-	if (WARN_ON_ONCE(!queue_is_mq(q)))
-		ret = 0;	/* not yet implemented, should not happen */
-	else
+	if (queue_is_mq(q)) {
 		ret = blk_mq_poll(q, cookie, iob, flags);
+	} else {
+		struct gendisk *disk = q->disk;
+
+		if (disk && disk->fops->poll_bio)
+			ret = disk->fops->poll_bio(bio, iob, flags);
+	}
 	blk_queue_exit(q);
 	return ret;
 }

block/genhd.c

@@ -412,6 +412,10 @@ int __must_check device_add_disk(struct device *parent, struct gendisk *disk,
 	struct device *ddev = disk_to_dev(disk);
 	int ret;
 
+	/* Only makes sense for bio-based to set ->poll_bio */
+	if (queue_is_mq(disk->queue) && disk->fops->poll_bio)
+		return -EINVAL;
+
 	/*
 	 * The disk queue should now be all set with enough information about
 	 * the device for the elevator code to pick an adequate default

drivers/md/dm-cache-policy-smq.c

@@ -1026,7 +1026,9 @@ static unsigned default_promote_level(struct smq_policy *mq)
 	 * This scheme reminds me of a graph of entropy vs probability of a
 	 * binary variable.
 	 */
-	static unsigned table[] = {1, 1, 1, 2, 4, 6, 7, 8, 7, 6, 4, 4, 3, 3, 2, 2, 1};
+	static const unsigned int table[] = {
+		1, 1, 1, 2, 4, 6, 7, 8, 7, 6, 4, 4, 3, 3, 2, 2, 1
+	};
 
 	unsigned hits = mq->cache_stats.hits;
 	unsigned misses = mq->cache_stats.misses;

drivers/md/dm-cache-target.c

@@ -803,7 +803,7 @@ static void accounted_complete(struct cache *cache, struct bio *bio)
 static void accounted_request(struct cache *cache, struct bio *bio)
 {
 	accounted_begin(cache, bio);
-	submit_bio_noacct(bio);
+	dm_submit_bio_remap(bio, NULL);
 }
 
 static void issue_op(struct bio *bio, void *context)
@@ -1708,7 +1708,7 @@ static bool process_bio(struct cache *cache, struct bio *bio)
 	bool commit_needed;
 
 	if (map_bio(cache, bio, get_bio_block(cache, bio), &commit_needed) == DM_MAPIO_REMAPPED)
-		submit_bio_noacct(bio);
+		dm_submit_bio_remap(bio, NULL);
 
 	return commit_needed;
 }
@@ -1774,7 +1774,7 @@ static bool process_discard_bio(struct cache *cache, struct bio *bio)
 
 	if (cache->features.discard_passdown) {
 		remap_to_origin(cache, bio);
-		submit_bio_noacct(bio);
+		dm_submit_bio_remap(bio, NULL);
 	} else
 		bio_endio(bio);
 
@@ -2015,7 +2015,6 @@ static int parse_metadata_dev(struct cache_args *ca, struct dm_arg_set *as,
 {
 	int r;
 	sector_t metadata_dev_size;
-	char b[BDEVNAME_SIZE];
 
 	if (!at_least_one_arg(as, error))
 		return -EINVAL;
@@ -2029,8 +2028,8 @@ static int parse_metadata_dev(struct cache_args *ca, struct dm_arg_set *as,
 
 	metadata_dev_size = get_dev_size(ca->metadata_dev);
 	if (metadata_dev_size > DM_CACHE_METADATA_MAX_SECTORS_WARNING)
-		DMWARN("Metadata device %s is larger than %u sectors: excess space will not be used.",
-		       bdevname(ca->metadata_dev->bdev, b), THIN_METADATA_MAX_SECTORS);
+		DMWARN("Metadata device %pg is larger than %u sectors: excess space will not be used.",
+		       ca->metadata_dev->bdev, THIN_METADATA_MAX_SECTORS);
 
 	return 0;
 }
@@ -2357,6 +2356,7 @@ static int cache_create(struct cache_args *ca, struct cache **result)
 
 	cache->ti = ca->ti;
 	ti->private = cache;
+	ti->accounts_remapped_io = true;
 	ti->num_flush_bios = 2;
 	ti->flush_supported = true;
 
@@ -3345,7 +3345,6 @@ static void disable_passdown_if_not_supported(struct cache *cache)
 	struct block_device *origin_bdev = cache->origin_dev->bdev;
 	struct queue_limits *origin_limits = &bdev_get_queue(origin_bdev)->limits;
 	const char *reason = NULL;
-	char buf[BDEVNAME_SIZE];
 
 	if (!cache->features.discard_passdown)
 		return;
@@ -3357,8 +3356,8 @@ static void disable_passdown_if_not_supported(struct cache *cache)
 		reason = "max discard sectors smaller than a block";
 
 	if (reason) {
-		DMWARN("Origin device (%s) %s: Disabling discard passdown.",
-		       bdevname(origin_bdev, buf), reason);
+		DMWARN("Origin device (%pg) %s: Disabling discard passdown.",
+		       origin_bdev, reason);
 		cache->features.discard_passdown = false;
 	}
 }

drivers/md/dm-clone-target.c

@@ -1682,7 +1682,6 @@ static int parse_metadata_dev(struct clone *clone, struct dm_arg_set *as, char *
 {
 	int r;
 	sector_t metadata_dev_size;
-	char b[BDEVNAME_SIZE];
 
 	r = dm_get_device(clone->ti, dm_shift_arg(as), FMODE_READ | FMODE_WRITE,
 			  &clone->metadata_dev);
@@ -1693,8 +1692,8 @@ static int parse_metadata_dev(struct clone *clone, struct dm_arg_set *as, char *
 
 	metadata_dev_size = get_dev_size(clone->metadata_dev);
 	if (metadata_dev_size > DM_CLONE_METADATA_MAX_SECTORS_WARNING)
-		DMWARN("Metadata device %s is larger than %u sectors: excess space will not be used.",
-		       bdevname(clone->metadata_dev->bdev, b), DM_CLONE_METADATA_MAX_SECTORS);
+		DMWARN("Metadata device %pg is larger than %u sectors: excess space will not be used.",
+		       clone->metadata_dev->bdev, DM_CLONE_METADATA_MAX_SECTORS);
 
 	return 0;
 }
@@ -2033,7 +2032,6 @@ static void disable_passdown_if_not_supported(struct clone *clone)
 	struct block_device *dest_dev = clone->dest_dev->bdev;
 	struct queue_limits *dest_limits = &bdev_get_queue(dest_dev)->limits;
 	const char *reason = NULL;
-	char buf[BDEVNAME_SIZE];
 
 	if (!test_bit(DM_CLONE_DISCARD_PASSDOWN, &clone->flags))
 		return;
@@ -2044,8 +2042,8 @@ static void disable_passdown_if_not_supported(struct clone *clone)
 		reason = "max discard sectors smaller than a region";
 
 	if (reason) {
-		DMWARN("Destination device (%s) %s: Disabling discard passdown.",
-		       bdevname(dest_dev, buf), reason);
+		DMWARN("Destination device (%pd) %s: Disabling discard passdown.",
+		       dest_dev, reason);
 		clear_bit(DM_CLONE_DISCARD_PASSDOWN, &clone->flags);
 	}
 }

drivers/md/dm-core.h

@@ -64,11 +64,21 @@ struct mapped_device {
 	struct gendisk *disk;
 	struct dax_device *dax_dev;
 
+	wait_queue_head_t wait;
+	unsigned long __percpu *pending_io;
+
+	/* forced geometry settings */
+	struct hd_geometry geometry;
+
+	/*
+	 * Processing queue (flush)
+	 */
+	struct workqueue_struct *wq;
+
 	/*
 	 * A list of ios that arrived while we were suspended.
 	 */
 	struct work_struct work;
-	wait_queue_head_t wait;
 	spinlock_t deferred_lock;
 	struct bio_list deferred;
 
@@ -83,36 +93,28 @@ struct mapped_device {
 	struct list_head uevent_list;
 	spinlock_t uevent_lock; /* Protect access to uevent_list */
 
+	/* for blk-mq request-based DM support */
+	bool init_tio_pdu:1;
+	struct blk_mq_tag_set *tag_set;
+
+	struct dm_stats stats;
+
 	/* the number of internal suspends */
 	unsigned internal_suspend_count;
 
+	int swap_bios;
+	struct semaphore swap_bios_semaphore;
+	struct mutex swap_bios_lock;
+
 	/*
 	 * io objects are allocated from here.
 	 */
 	struct bio_set io_bs;
 	struct bio_set bs;
 
-	/*
-	 * Processing queue (flush)
-	 */
-	struct workqueue_struct *wq;
-
-	/* forced geometry settings */
-	struct hd_geometry geometry;
-
 	/* kobject and completion */
 	struct dm_kobject_holder kobj_holder;
 
-	int swap_bios;
-	struct semaphore swap_bios_semaphore;
-	struct mutex swap_bios_lock;
-
-	struct dm_stats stats;
-
-	/* for blk-mq request-based DM support */
-	struct blk_mq_tag_set *tag_set;
-	bool init_tio_pdu:1;
-
 	struct srcu_struct io_barrier;
 
 #ifdef CONFIG_BLK_DEV_ZONED
@@ -206,35 +208,76 @@ struct dm_table {
 /*
  * One of these is allocated per clone bio.
  */
-#define DM_TIO_MAGIC 7282014
+#define DM_TIO_MAGIC 28714
 struct dm_target_io {
-	unsigned int magic;
+	unsigned short magic;
+	unsigned short flags;
+	unsigned int target_bio_nr;
 	struct dm_io *io;
 	struct dm_target *ti;
-	unsigned int target_bio_nr;
 	unsigned int *len_ptr;
-	bool inside_dm_io;
+	sector_t old_sector;
 	struct bio clone;
 };
 
+/*
+ * dm_target_io flags
+ */
+enum {
+	DM_TIO_INSIDE_DM_IO,
+	DM_TIO_IS_DUPLICATE_BIO
+};
+
+static inline bool dm_tio_flagged(struct dm_target_io *tio, unsigned int bit)
+{
+	return (tio->flags & (1U << bit)) != 0;
+}
+
+static inline void dm_tio_set_flag(struct dm_target_io *tio, unsigned int bit)
+{
+	tio->flags |= (1U << bit);
+}
+
 /*
  * One of these is allocated per original bio.
  * It contains the first clone used for that original.
  */
-#define DM_IO_MAGIC 5191977
+#define DM_IO_MAGIC 19577
 struct dm_io {
-	unsigned int magic;
-	struct mapped_device *md;
-	blk_status_t status;
+	unsigned short magic;
+	unsigned short flags;
 	atomic_t io_count;
+	struct mapped_device *md;
 	struct bio *orig_bio;
+	blk_status_t status;
+	spinlock_t lock;
 	unsigned long start_time;
-	spinlock_t endio_lock;
+	void *data;
+	struct hlist_node node;
+	struct task_struct *map_task;
 	struct dm_stats_aux stats_aux;
 	/* last member of dm_target_io is 'struct bio' */
 	struct dm_target_io tio;
 };
 
+/*
+ * dm_io flags
+ */
+enum {
+	DM_IO_START_ACCT,
+	DM_IO_ACCOUNTED
+};
+
+static inline bool dm_io_flagged(struct dm_io *io, unsigned int bit)
+{
+	return (io->flags & (1U << bit)) != 0;
+}
+
+static inline void dm_io_set_flag(struct dm_io *io, unsigned int bit)
+{
+	io->flags |= (1U << bit);
+}
+
 static inline void dm_io_inc_pending(struct dm_io *io)
 {
 	atomic_inc(&io->io_count);