Merge tag 'dm-3.15-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

Pull device mapper changes from Mike Snitzer:

 - Fix dm-cache corruption caused by discard_block_size > cache_block_size

 - Fix a lock-inversion detected by LOCKDEP in dm-cache

 - Fix a dangling bio bug in the dm-thinp target's process_deferred_bios
   error path

 - Fix corruption due to non-atomic transaction commit which allowed a
   metadata superblock to be written before all other metadata was
   successfully written -- this is common to all targets that use the
   persistent-data library's transaction manager (dm-thinp, dm-cache and
   dm-era).

 - Various small cleanups in the DM core

 - Add the dm-era target which is useful for keeping track of which
   blocks were written within a user defined period of time called an
   'era'.  Use cases include tracking changed blocks for backup
   software, and partially invalidating the contents of a cache to
   restore cache coherency after rolling back a vendor snapshot.

 - Improve the on-disk layout of multithreaded writes to the
   dm-thin-pool by splitting the pool's deferred bio list to be a
   per-thin device list and then sorting that list using an rb_tree.
   The subsequent read throughput of the data written via multiple
   threads improved by ~70%.

 - Simplify the multipath target's handling of queuing IO by pushing
   requests back to the request queue rather than queueing the IO
   internally.

* tag 'dm-3.15-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm: (24 commits)
  dm cache: fix a lock-inversion
  dm thin: sort the per thin deferred bios using an rb_tree
  dm thin: use per thin device deferred bio lists
  dm thin: simplify pool_is_congested
  dm thin: fix dangling bio in process_deferred_bios error path
  dm mpath: print more useful warnings in multipath_message()
  dm-mpath: do not activate failed paths
  dm mpath: remove extra nesting in map function
  dm mpath: remove map_io()
  dm mpath: reduce memory pressure when requeuing
  dm mpath: remove process_queued_ios()
  dm mpath: push back requests instead of queueing
  dm table: add dm_table_run_md_queue_async
  dm mpath: do not call pg_init when it is already running
  dm: use RCU_INIT_POINTER instead of rcu_assign_pointer in __unbind
  dm: stop using bi_private
  dm: remove dm_get_mapinfo
  dm: make dm_table_alloc_md_mempools static
  dm: take care to copy the space map roots before locking the superblock
  dm transaction manager: fix corruption due to non-atomic transaction commit
  ...

Documentation/device-mapper/era.txt

+Introduction
+============
+
+dm-era is a target that behaves similar to the linear target.  In
+addition it keeps track of which blocks were written within a user
+defined period of time called an 'era'.  Each era target instance
+maintains the current era as a monotonically increasing 32-bit
+counter.
+
+Use cases include tracking changed blocks for backup software, and
+partially invalidating the contents of a cache to restore cache
+coherency after rolling back a vendor snapshot.
+
+Constructor
+===========
+
+ era <metadata dev> <origin dev> <block size>
+
+ metadata dev    : fast device holding the persistent metadata
+ origin dev	 : device holding data blocks that may change
+ block size      : block size of origin data device, granularity that is
+		     tracked by the target
+
+Messages
+========
+
+None of the dm messages take any arguments.
+
+checkpoint
+----------
+
+Possibly move to a new era.  You shouldn't assume the era has
+incremented.  After sending this message, you should check the
+current era via the status line.
+
+take_metadata_snap
+------------------
+
+Create a clone of the metadata, to allow a userland process to read it.
+
+drop_metadata_snap
+------------------
+
+Drop the metadata snapshot.
+
+Status
+======
+
+<metadata block size> <#used metadata blocks>/<#total metadata blocks>
+<current era> <held metadata root | '-'>
+
+metadata block size	 : Fixed block size for each metadata block in
+			     sectors
+#used metadata blocks	 : Number of metadata blocks used
+#total metadata blocks	 : Total number of metadata blocks
+current era		 : The current era
+held metadata root	 : The location, in blocks, of the metadata root
+			     that has been 'held' for userspace read
+			     access. '-' indicates there is no held root
+
+Detailed use case
+=================
+
+The scenario of invalidating a cache when rolling back a vendor
+snapshot was the primary use case when developing this target:
+
+Taking a vendor snapshot
+------------------------
+
+- Send a checkpoint message to the era target
+- Make a note of the current era in its status line
+- Take vendor snapshot (the era and snapshot should be forever
+  associated now).
+
+Rolling back to an vendor snapshot
+----------------------------------
+
+- Cache enters passthrough mode (see: dm-cache's docs in cache.txt)
+- Rollback vendor storage
+- Take metadata snapshot
+- Ascertain which blocks have been written since the snapshot was taken
+  by checking each block's era
+- Invalidate those blocks in the caching software
+- Cache returns to writeback/writethrough mode
+
+Memory usage
+============
+
+The target uses a bitset to record writes in the current era.  It also
+has a spare bitset ready for switching over to a new era.  Other than
+that it uses a few 4k blocks for updating metadata.
+
+   (4 * nr_blocks) bytes + buffers
+
+Resilience
+==========
+
+Metadata is updated on disk before a write to a previously unwritten
+block is performed.  As such dm-era should not be effected by a hard
+crash such as power failure.
+
+Userland tools
+==============
+
+Userland tools are found in the increasingly poorly named
+thin-provisioning-tools project:
+
+    https://github.com/jthornber/thin-provisioning-tools

drivers/md/Kconfig

@@ -285,6 +285,17 @@ config DM_CACHE_CLEANER
          A simple cache policy that writes back all data to the
          origin.  Used when decommissioning a dm-cache.
 
+config DM_ERA
+       tristate "Era target (EXPERIMENTAL)"
+       depends on BLK_DEV_DM
+       default n
+       select DM_PERSISTENT_DATA
+       select DM_BIO_PRISON
+       ---help---
+         dm-era tracks which parts of a block device are written to
+         over time.  Useful for maintaining cache coherency when using
+         vendor snapshots.
+
 config DM_MIRROR
        tristate "Mirror target"
        depends on BLK_DEV_DM

drivers/md/Makefile

@@ -14,6 +14,7 @@ dm-thin-pool-y	+= dm-thin.o dm-thin-metadata.o
 dm-cache-y	+= dm-cache-target.o dm-cache-metadata.o dm-cache-policy.o
 dm-cache-mq-y   += dm-cache-policy-mq.o
 dm-cache-cleaner-y += dm-cache-policy-cleaner.o
+dm-era-y	+= dm-era-target.o
 md-mod-y	+= md.o bitmap.o
 raid456-y	+= raid5.o
 
@@ -53,6 +54,7 @@ obj-$(CONFIG_DM_VERITY)		+= dm-verity.o
 obj-$(CONFIG_DM_CACHE)		+= dm-cache.o
 obj-$(CONFIG_DM_CACHE_MQ)	+= dm-cache-mq.o
 obj-$(CONFIG_DM_CACHE_CLEANER)	+= dm-cache-cleaner.o
+obj-$(CONFIG_DM_ERA)		+= dm-era.o
 
 ifeq ($(CONFIG_DM_UEVENT),y)
 dm-mod-objs			+= dm-uevent.o

drivers/md/dm-cache-block-types.h

@@ -19,7 +19,6 @@
 
 typedef dm_block_t __bitwise__ dm_oblock_t;
 typedef uint32_t __bitwise__ dm_cblock_t;
-typedef dm_block_t __bitwise__ dm_dblock_t;
 
 static inline dm_oblock_t to_oblock(dm_block_t b)
 {
@@ -41,14 +40,4 @@ static inline uint32_t from_cblock(dm_cblock_t b)
 	return (__force uint32_t) b;
 }
 
-static inline dm_dblock_t to_dblock(dm_block_t b)
-{
-	return (__force dm_dblock_t) b;
-}
-
-static inline dm_block_t from_dblock(dm_dblock_t b)
-{
-	return (__force dm_block_t) b;
-}
-
 #endif /* DM_CACHE_BLOCK_TYPES_H */

drivers/md/dm-cache-metadata.c

@@ -109,7 +109,7 @@ struct dm_cache_metadata {
 	dm_block_t discard_root;
 
 	sector_t discard_block_size;
-	dm_dblock_t discard_nr_blocks;
+	dm_oblock_t discard_nr_blocks;
 
 	sector_t data_block_size;
 	dm_cblock_t cache_blocks;
@@ -120,6 +120,12 @@ struct dm_cache_metadata {
 	unsigned policy_version[CACHE_POLICY_VERSION_SIZE];
 	size_t policy_hint_size;
 	struct dm_cache_statistics stats;
+
+	/*
+	 * Reading the space map root can fail, so we read it into this
+	 * buffer before the superblock is locked and updated.
+	 */
+	__u8 metadata_space_map_root[SPACE_MAP_ROOT_SIZE];
 };
 
 /*-------------------------------------------------------------------
@@ -260,11 +266,31 @@ static void __setup_mapping_info(struct dm_cache_metadata *cmd)
 	}
 }
 
+static int __save_sm_root(struct dm_cache_metadata *cmd)
+{
+	int r;
+	size_t metadata_len;
+
+	r = dm_sm_root_size(cmd->metadata_sm, &metadata_len);
+	if (r < 0)
+		return r;
+
+	return dm_sm_copy_root(cmd->metadata_sm, &cmd->metadata_space_map_root,
+			       metadata_len);
+}
+
+static void __copy_sm_root(struct dm_cache_metadata *cmd,
+			   struct cache_disk_superblock *disk_super)
+{
+	memcpy(&disk_super->metadata_space_map_root,
+	       &cmd->metadata_space_map_root,
+	       sizeof(cmd->metadata_space_map_root));
+}
+
 static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 {
 	int r;
 	struct dm_block *sblock;
-	size_t metadata_len;
 	struct cache_disk_superblock *disk_super;
 	sector_t bdev_size = i_size_read(cmd->bdev->bd_inode) >> SECTOR_SHIFT;
 
@@ -272,12 +298,16 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 	if (bdev_size > DM_CACHE_METADATA_MAX_SECTORS)
 		bdev_size = DM_CACHE_METADATA_MAX_SECTORS;
 
-	r = dm_sm_root_size(cmd->metadata_sm, &metadata_len);
+	r = dm_tm_pre_commit(cmd->tm);
 	if (r < 0)
 		return r;
 
-	r = dm_tm_pre_commit(cmd->tm);
-	if (r < 0)
+	/*
+	 * dm_sm_copy_root() can fail.  So we need to do it before we start
+	 * updating the superblock.
+	 */
+	r = __save_sm_root(cmd);
+	if (r)
 		return r;
 
 	r = superblock_lock_zero(cmd, &sblock);
@@ -293,16 +323,13 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 	memset(disk_super->policy_version, 0, sizeof(disk_super->policy_version));
 	disk_super->policy_hint_size = 0;
 
-	r = dm_sm_copy_root(cmd->metadata_sm, &disk_super->metadata_space_map_root,
-			    metadata_len);
-	if (r < 0)
-		goto bad_locked;
+	__copy_sm_root(cmd, disk_super);
 
 	disk_super->mapping_root = cpu_to_le64(cmd->root);
 	disk_super->hint_root = cpu_to_le64(cmd->hint_root);
 	disk_super->discard_root = cpu_to_le64(cmd->discard_root);
 	disk_super->discard_block_size = cpu_to_le64(cmd->discard_block_size);
-	disk_super->discard_nr_blocks = cpu_to_le64(from_dblock(cmd->discard_nr_blocks));
+	disk_super->discard_nr_blocks = cpu_to_le64(from_oblock(cmd->discard_nr_blocks));
 	disk_super->metadata_block_size = cpu_to_le32(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT);
 	disk_super->data_block_size = cpu_to_le32(cmd->data_block_size);
 	disk_super->cache_blocks = cpu_to_le32(0);
@@ -313,10 +340,6 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 	disk_super->write_misses = cpu_to_le32(0);
 
 	return dm_tm_commit(cmd->tm, sblock);
-
-bad_locked:
-	dm_bm_unlock(sblock);
-	return r;
 }
 
 static int __format_metadata(struct dm_cache_metadata *cmd)
@@ -496,7 +519,7 @@ static void read_superblock_fields(struct dm_cache_metadata *cmd,
 	cmd->hint_root = le64_to_cpu(disk_super->hint_root);
 	cmd->discard_root = le64_to_cpu(disk_super->discard_root);
 	cmd->discard_block_size = le64_to_cpu(disk_super->discard_block_size);
-	cmd->discard_nr_blocks = to_dblock(le64_to_cpu(disk_super->discard_nr_blocks));
+	cmd->discard_nr_blocks = to_oblock(le64_to_cpu(disk_super->discard_nr_blocks));
 	cmd->data_block_size = le32_to_cpu(disk_super->data_block_size);
 	cmd->cache_blocks = to_cblock(le32_to_cpu(disk_super->cache_blocks));
 	strncpy(cmd->policy_name, disk_super->policy_name, sizeof(cmd->policy_name));
@@ -530,8 +553,9 @@ static int __begin_transaction_flags(struct dm_cache_metadata *cmd,
 	disk_super = dm_block_data(sblock);
 	update_flags(disk_super, mutator);
 	read_superblock_fields(cmd, disk_super);
+	dm_bm_unlock(sblock);
 
-	return dm_bm_flush_and_unlock(cmd->bm, sblock);
+	return dm_bm_flush(cmd->bm);
 }
 
 static int __begin_transaction(struct dm_cache_metadata *cmd)
@@ -559,7 +583,6 @@ static int __commit_transaction(struct dm_cache_metadata *cmd,
 				flags_mutator mutator)
 {
 	int r;
-	size_t metadata_len;
 	struct cache_disk_superblock *disk_super;
 	struct dm_block *sblock;
 
@@ -577,8 +600,8 @@ static int __commit_transaction(struct dm_cache_metadata *cmd,
 	if (r < 0)
 		return r;
 
-	r = dm_sm_root_size(cmd->metadata_sm, &metadata_len);
-	if (r < 0)
+	r = __save_sm_root(cmd);
+	if (r)
 		return r;
 
 	r = superblock_lock(cmd, &sblock);
@@ -594,7 +617,7 @@ static int __commit_transaction(struct dm_cache_metadata *cmd,
 	disk_super->hint_root = cpu_to_le64(cmd->hint_root);
 	disk_super->discard_root = cpu_to_le64(cmd->discard_root);
 	disk_super->discard_block_size = cpu_to_le64(cmd->discard_block_size);
-	disk_super->discard_nr_blocks = cpu_to_le64(from_dblock(cmd->discard_nr_blocks));
+	disk_super->discard_nr_blocks = cpu_to_le64(from_oblock(cmd->discard_nr_blocks));
 	disk_super->cache_blocks = cpu_to_le32(from_cblock(cmd->cache_blocks));
 	strncpy(disk_super->policy_name, cmd->policy_name, sizeof(disk_super->policy_name));
 	disk_super->policy_version[0] = cpu_to_le32(cmd->policy_version[0]);
@@ -605,13 +628,7 @@ static int __commit_transaction(struct dm_cache_metadata *cmd,
 	disk_super->read_misses = cpu_to_le32(cmd->stats.read_misses);
 	disk_super->write_hits = cpu_to_le32(cmd->stats.write_hits);
 	disk_super->write_misses = cpu_to_le32(cmd->stats.write_misses);
-
-	r = dm_sm_copy_root(cmd->metadata_sm, &disk_super->metadata_space_map_root,
-			    metadata_len);
-	if (r < 0) {
-		dm_bm_unlock(sblock);
-		return r;
-	}
+	__copy_sm_root(cmd, disk_super);
 
 	return dm_tm_commit(cmd->tm, sblock);
 }
@@ -771,15 +788,15 @@ int dm_cache_resize(struct dm_cache_metadata *cmd, dm_cblock_t new_cache_size)
 
 int dm_cache_discard_bitset_resize(struct dm_cache_metadata *cmd,
 				   sector_t discard_block_size,
-				   dm_dblock_t new_nr_entries)
+				   dm_oblock_t new_nr_entries)
 {
 	int r;
 
 	down_write(&cmd->root_lock);
 	r = dm_bitset_resize(&cmd->discard_info,
 			     cmd->discard_root,
-			     from_dblock(cmd->discard_nr_blocks),
-			     from_dblock(new_nr_entries),
+			     from_oblock(cmd->discard_nr_blocks),
+			     from_oblock(new_nr_entries),
 			     false, &cmd->discard_root);
 	if (!r) {
 		cmd->discard_block_size = discard_block_size;
@@ -792,28 +809,28 @@ int dm_cache_discard_bitset_resize(struct dm_cache_metadata *cmd,
 	return r;
 }
 
-static int __set_discard(struct dm_cache_metadata *cmd, dm_dblock_t b)
+static int __set_discard(struct dm_cache_metadata *cmd, dm_oblock_t b)
 {
 	return dm_bitset_set_bit(&cmd->discard_info, cmd->discard_root,
-				 from_dblock(b), &cmd->discard_root);
+				 from_oblock(b), &cmd->discard_root);
 }
 
-static int __clear_discard(struct dm_cache_metadata *cmd, dm_dblock_t b)
+static int __clear_discard(struct dm_cache_metadata *cmd, dm_oblock_t b)
 {
 	return dm_bitset_clear_bit(&cmd->discard_info, cmd->discard_root,
-				   from_dblock(b), &cmd->discard_root);
+				   from_oblock(b), &cmd->discard_root);
 }
 
-static int __is_discarded(struct dm_cache_metadata *cmd, dm_dblock_t b,
+static int __is_discarded(struct dm_cache_metadata *cmd, dm_oblock_t b,
 			  bool *is_discarded)
 {
 	return dm_bitset_test_bit(&cmd->discard_info, cmd->discard_root,
-				  from_dblock(b), &cmd->discard_root,
+				  from_oblock(b), &cmd->discard_root,
 				  is_discarded);
 }
 
 static int __discard(struct dm_cache_metadata *cmd,
-		     dm_dblock_t dblock, bool discard)
+		     dm_oblock_t dblock, bool discard)
 {
 	int r;
 
@@ -826,7 +843,7 @@ static int __discard(struct dm_cache_metadata *cmd,
 }
 
 int dm_cache_set_discard(struct dm_cache_metadata *cmd,
-			 dm_dblock_t dblock, bool discard)
+			 dm_oblock_t dblock, bool discard)
 {
 	int r;
 
@@ -844,8 +861,8 @@ static int __load_discards(struct dm_cache_metadata *cmd,
 	dm_block_t b;
 	bool discard;
 
-	for (b = 0; b < from_dblock(cmd->discard_nr_blocks); b++) {
-		dm_dblock_t dblock = to_dblock(b);
+	for (b = 0; b < from_oblock(cmd->discard_nr_blocks); b++) {
+		dm_oblock_t dblock = to_oblock(b);
 
 		if (cmd->clean_when_opened) {
 			r = __is_discarded(cmd, dblock, &discard);
@@ -1228,22 +1245,12 @@ static int begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *po
 	return 0;
 }
 
-int dm_cache_begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *policy)
+static int save_hint(void *context, dm_cblock_t cblock, dm_oblock_t oblock, uint32_t hint)
 {
+	struct dm_cache_metadata *cmd = context;
+	__le32 value = cpu_to_le32(hint);
 	int r;
 
-	down_write(&cmd->root_lock);
-	r = begin_hints(cmd, policy);
-	up_write(&cmd->root_lock);
-
-	return r;
-}
-
-static int save_hint(struct dm_cache_metadata *cmd, dm_cblock_t cblock,
-		     uint32_t hint)
-{
-	int r;
-	__le32 value = cpu_to_le32(hint);
 	__dm_bless_for_disk(&value);
 
 	r = dm_array_set_value(&cmd->hint_info, cmd->hint_root,
@@ -1253,16 +1260,25 @@ static int save_hint(struct dm_cache_metadata *cmd, dm_cblock_t cblock,
 	return r;
 }
 
-int dm_cache_save_hint(struct dm_cache_metadata *cmd, dm_cblock_t cblock,
-		       uint32_t hint)
+static int write_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *policy)
 {
 	int r;
 
-	if (!hints_array_initialized(cmd))
-		return 0;
+	r = begin_hints(cmd, policy);
+	if (r) {
+		DMERR("begin_hints failed");
+		return r;
+	}
+
+	return policy_walk_mappings(policy, save_hint, cmd);
+}
+
+int dm_cache_write_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *policy)
+{
+	int r;
 
 	down_write(&cmd->root_lock);
-	r = save_hint(cmd, cblock, hint);
+	r = write_hints(cmd, policy);
 	up_write(&cmd->root_lock);
 
 	return r;

drivers/md/dm-cache-metadata.h

@@ -72,14 +72,14 @@ dm_cblock_t dm_cache_size(struct dm_cache_metadata *cmd);
 
 int dm_cache_discard_bitset_resize(struct dm_cache_metadata *cmd,
 				   sector_t discard_block_size,
-				   dm_dblock_t new_nr_entries);
+				   dm_oblock_t new_nr_entries);
 
 typedef int (*load_discard_fn)(void *context, sector_t discard_block_size,
-			       dm_dblock_t dblock, bool discarded);
+			       dm_oblock_t dblock, bool discarded);
 int dm_cache_load_discards(struct dm_cache_metadata *cmd,
 			   load_discard_fn fn, void *context);
 
-int dm_cache_set_discard(struct dm_cache_metadata *cmd, dm_dblock_t dblock, bool discard);
+int dm_cache_set_discard(struct dm_cache_metadata *cmd, dm_oblock_t dblock, bool discard);
 
 int dm_cache_remove_mapping(struct dm_cache_metadata *cmd, dm_cblock_t cblock);
 int dm_cache_insert_mapping(struct dm_cache_metadata *cmd, dm_cblock_t cblock, dm_oblock_t oblock);
@@ -128,14 +128,7 @@ void dm_cache_dump(struct dm_cache_metadata *cmd);
  * rather than querying the policy for each cblock, we let it walk its data
  * structures and fill in the hints in whatever order it wishes.
  */
-
-int dm_cache_begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *p);
-
-/*
- * requests hints for every cblock and stores in the metadata device.
- */
-int dm_cache_save_hint(struct dm_cache_metadata *cmd,
-		       dm_cblock_t cblock, uint32_t hint);
+int dm_cache_write_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *p);
 
 /*
  * Query method.  Are all the blocks in the cache clean?

drivers/md/dm-cache-target.c

@@ -237,9 +237,8 @@ struct cache {
 	/*
 	 * origin_blocks entries, discarded if set.
 	 */
-	dm_dblock_t discard_nr_blocks;
+	dm_oblock_t discard_nr_blocks;
 	unsigned long *discard_bitset;
-	uint32_t discard_block_size; /* a power of 2 times sectors per block */
 
 	/*
 	 * Rather than reconstructing the table line for the status we just
@@ -526,48 +525,33 @@ static dm_block_t block_div(dm_block_t b, uint32_t n)
 	return b;
 }
 
-static dm_dblock_t oblock_to_dblock(struct cache *cache, dm_oblock_t oblock)
-{
-	uint32_t discard_blocks = cache->discard_block_size;
-	dm_block_t b = from_oblock(oblock);
-
-	if (!block_size_is_power_of_two(cache))
-		discard_blocks = discard_blocks / cache->sectors_per_block;
-	else
-		discard_blocks >>= cache->sectors_per_block_shift;
-
-	b = block_div(b, discard_blocks);
-
-	return to_dblock(b);
-}
-
-static void set_discard(struct cache *cache, dm_dblock_t b)
+static void set_discard(struct cache *cache, dm_oblock_t b)
 {
 	unsigned long flags;
 
 	atomic_inc(&cache->stats.discard_count);
 
 	spin_lock_irqsave(&cache->lock, flags);
-	set_bit(from_dblock(b), cache->discard_bitset);
+	set_bit(from_oblock(b), cache->discard_bitset);
 	spin_unlock_irqrestore(&cache->lock, flags);
 }
 
-static void clear_discard(struct cache *cache, dm_dblock_t b)
+static void clear_discard(struct cache *cache, dm_oblock_t b)
 {
 	unsigned long flags;
 
 	spin_lock_irqsave(&cache->lock, flags);
-	clear_bit(from_dblock(b), cache->discard_bitset);
+	clear_bit(from_oblock(b), cache->discard_bitset);
 	spin_unlock_irqrestore(&cache->lock, flags);
 }
 
-static bool is_discarded(struct cache *cache, dm_dblock_t b)
+static bool is_discarded(struct cache *cache, dm_oblock_t b)
 {
 	int r;
 	unsigned long flags;
 
 	spin_lock_irqsave(&cache->lock, flags);
-	r = test_bit(from_dblock(b), cache->discard_bitset);
+	r = test_bit(from_oblock(b), cache->discard_bitset);
 	spin_unlock_irqrestore(&cache->lock, flags);
 
 	return r;
@@ -579,8 +563,7 @@ static bool is_discarded_oblock(struct cache *cache, dm_oblock_t b)
 	unsigned long flags;
 
 	spin_lock_irqsave(&cache->lock, flags);
-	r = test_bit(from_dblock(oblock_to_dblock(cache, b)),
-		     cache->discard_bitset);
+	r = test_bit(from_oblock(b), cache->discard_bitset);
 	spin_unlock_irqrestore(&cache->lock, flags);
 
 	return r;
@@ -705,7 +688,7 @@ static void remap_to_origin_clear_discard(struct cache *cache, struct bio *bio,
 	check_if_tick_bio_needed(cache, bio);
 	remap_to_origin(cache, bio);
 	if (bio_data_dir(bio) == WRITE)
-		clear_discard(cache, oblock_to_dblock(cache, oblock));
+		clear_discard(cache, oblock);
 }
 
 static void remap_to_cache_dirty(struct cache *cache, struct bio *bio,
@@ -715,7 +698,7 @@ static void remap_to_cache_dirty(struct cache *cache, struct bio *bio,
 	remap_to_cache(cache, bio, cblock);
 	if (bio_data_dir(bio) == WRITE) {
 		set_dirty(cache, oblock, cblock);
-		clear_discard(cache, oblock_to_dblock(cache, oblock));
+		clear_discard(cache, oblock);
 	}
 }
 
@@ -1288,14 +1271,14 @@ static void process_flush_bio(struct cache *cache, struct bio *bio)
 static void process_discard_bio(struct cache *cache, struct bio *bio)
 {
 	dm_block_t start_block = dm_sector_div_up(bio->bi_iter.bi_sector,
-						  cache->discard_block_size);
+						  cache->sectors_per_block);
 	dm_block_t end_block = bio_end_sector(bio);
 	dm_block_t b;
 
-	end_block = block_div(end_block, cache->discard_block_size);
+	end_block = block_div(end_block, cache->sectors_per_block);
 
 	for (b = start_block; b < end_block; b++)
-		set_discard(cache, to_dblock(b));
+		set_discard(cache, to_oblock(b));
 
 	bio_endio(bio, 0);
 }
@@ -2171,35 +2154,6 @@ static int create_cache_policy(struct cache *cache, struct cache_args *ca,
 	return 0;
 }
 
-/*
- * We want the discard block size to be a power of two, at least the size
- * of the cache block size, and have no more than 2^14 discard blocks
- * across the origin.
- */
-#define MAX_DISCARD_BLOCKS (1 << 14)
-
-static bool too_many_discard_blocks(sector_t discard_block_size,
-				    sector_t origin_size)
-{
-	(void) sector_div(origin_size, discard_block_size);
-
-	return origin_size > MAX_DISCARD_BLOCKS;
-}
-
-static sector_t calculate_discard_block_size(sector_t cache_block_size,
-					     sector_t origin_size)
-{
-	sector_t discard_block_size;
-
-	discard_block_size = roundup_pow_of_two(cache_block_size);
-
-	if (origin_size)
-		while (too_many_discard_blocks(discard_block_size, origin_size))
-			discard_block_size *= 2;
-
-	return discard_block_size;
-}
-
 #define DEFAULT_MIGRATION_THRESHOLD 2048
 
 static int cache_create(struct cache_args *ca, struct cache **result)
@@ -2321,16 +2275,13 @@ static int cache_create(struct cache_args *ca, struct cache **result)
 	}
 	clear_bitset(cache->dirty_bitset, from_cblock(cache->cache_size));
 
-	cache->discard_block_size =
-		calculate_discard_block_size(cache->sectors_per_block,
-					     cache->origin_sectors);
-	cache->discard_nr_blocks = oblock_to_dblock(cache, cache->origin_blocks);
-	cache->discard_bitset = alloc_bitset(from_dblock(cache->discard_nr_blocks));
+	cache->discard_nr_blocks = cache->origin_blocks;
+	cache->discard_bitset = alloc_bitset(from_oblock(cache->discard_nr_blocks));
 	if (!cache->discard_bitset) {
 		*error = "could not allocate discard bitset";
 		goto bad;
 	}
-	clear_bitset(cache->discard_bitset, from_dblock(cache->discard_nr_blocks));
+	clear_bitset(cache->discard_bitset, from_oblock(cache->discard_nr_blocks));
 
 	cache->copier = dm_kcopyd_client_create(&dm_kcopyd_throttle);
 	if (IS_ERR(cache->copier)) {
@@ -2614,16 +2565,16 @@ static int write_discard_bitset(struct cache *cache)
 {
 	unsigned i, r;
 
-	r = dm_cache_discard_bitset_resize(cache->cmd, cache->discard_block_size,
-					   cache->discard_nr_blocks);
+	r = dm_cache_discard_bitset_resize(cache->cmd, cache->sectors_per_block,
+					   cache->origin_blocks);
 	if (r) {
 		DMERR("could not resize on-disk discard bitset");
 		return r;
 	}
 
-	for (i = 0; i < from_dblock(cache->discard_nr_blocks); i++) {
-		r = dm_cache_set_discard(cache->cmd, to_dblock(i),
-					 is_discarded(cache, to_dblock(i)));
+	for (i = 0; i < from_oblock(cache->discard_nr_blocks); i++) {
+		r = dm_cache_set_discard(cache->cmd, to_oblock(i),
+					 is_discarded(cache, to_oblock(i)));
 		if (r)
 			return r;
 	}
@@ -2631,30 +2582,6 @@ static int write_discard_bitset(struct cache *cache)
 	return 0;
 }
 
-static int save_hint(void *context, dm_cblock_t cblock, dm_oblock_t oblock,
-		     uint32_t hint)
-{
-	struct cache *cache = context;
-	return dm_cache_save_hint(cache->cmd, cblock, hint);
-}
-
-static int write_hints(struct cache *cache)
-{
-	int r;
-
-	r = dm_cache_begin_hints(cache->cmd, cache->policy);
-	if (r) {
-		DMERR("dm_cache_begin_hints failed");
-		return r;
-	}
-
-	r = policy_walk_mappings(cache->policy, save_hint, cache);
-	if (r)
-		DMERR("policy_walk_mappings failed");
-
-	return r;
-}
-
 /*
  * returns true on success
  */
@@ -2672,7 +2599,7 @@ static bool sync_metadata(struct cache *cache)
 
 	save_stats(cache);
 
-	r3 = write_hints(cache);
+	r3 = dm_cache_write_hints(cache->cmd, cache->policy);
 	if (r3)
 		DMERR("could not write hints");
 
@@ -2720,16 +2647,14 @@ static int load_mapping(void *context, dm_oblock_t oblock, dm_cblock_t cblock,
 }
 
 static int load_discard(void *context, sector_t discard_block_size,
-			dm_dblock_t dblock, bool discard)
+			dm_oblock_t oblock, bool discard)
 {
 	struct cache *cache = context;
 
-	/* FIXME: handle mis-matched block size */
-
 	if (discard)
-		set_discard(cache, dblock);
+		set_discard(cache, oblock);
 	else
-		clear_discard(cache, dblock);
+		clear_discard(cache, oblock);
 
 	return 0;
 }
@@ -3120,8 +3045,8 @@ static void set_discard_limits(struct cache *cache, struct queue_limits *limits)
 	/*
 	 * FIXME: these limits may be incompatible with the cache device
 	 */
-	limits->max_discard_sectors = cache->discard_block_size * 1024;
-	limits->discard_granularity = cache->discard_block_size << SECTOR_SHIFT;
+	limits->max_discard_sectors = cache->sectors_per_block;
+	limits->discard_granularity = cache->sectors_per_block << SECTOR_SHIFT;
 }
 
 static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits)
@@ -3145,7 +3070,7 @@ static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits)
 
 static struct target_type cache_target = {
 	.name = "cache",
-	.version = {1, 3, 0},
+	.version = {1, 4, 0},
 	.module = THIS_MODULE,
 	.ctr = cache_ctr,
 	.dtr = cache_dtr,

drivers/md/dm-mpath.c

@@ -93,10 +93,6 @@ struct multipath {
 	unsigned pg_init_count;		/* Number of times pg_init called */
 	unsigned pg_init_delay_msecs;	/* Number of msecs before pg_init retry */
 
-	unsigned queue_size;
-	struct work_struct process_queued_ios;
-	struct list_head queued_ios;
-
 	struct work_struct trigger_event;
 
 	/*
@@ -121,9 +117,9 @@ typedef int (*action_fn) (struct pgpath *pgpath);
 static struct kmem_cache *_mpio_cache;
 
 static struct workqueue_struct *kmultipathd, *kmpath_handlerd;
-static void process_queued_ios(struct work_struct *work);
 static void trigger_event(struct work_struct *work);
 static void activate_path(struct work_struct *work);
+static int __pgpath_busy(struct pgpath *pgpath);
 
 
 /*-----------------------------------------------
@@ -195,11 +191,9 @@ static struct multipath *alloc_multipath(struct dm_target *ti)
 	m = kzalloc(sizeof(*m), GFP_KERNEL);
 	if (m) {
 		INIT_LIST_HEAD(&m->priority_groups);
-		INIT_LIST_HEAD(&m->queued_ios);
 		spin_lock_init(&m->lock);
 		m->queue_io = 1;
 		m->pg_init_delay_msecs = DM_PG_INIT_DELAY_DEFAULT;
-		INIT_WORK(&m->process_queued_ios, process_queued_ios);
 		INIT_WORK(&m->trigger_event, trigger_event);
 		init_waitqueue_head(&m->pg_init_wait);
 		mutex_init(&m->work_mutex);
@@ -256,13 +250,21 @@ static void clear_mapinfo(struct multipath *m, union map_info *info)
  * Path selection
  *-----------------------------------------------*/
 
-static void __pg_init_all_paths(struct multipath *m)
+static int __pg_init_all_paths(struct multipath *m)
 {
 	struct pgpath *pgpath;
 	unsigned long pg_init_delay = 0;
 
+	if (m->pg_init_in_progress || m->pg_init_disabled)
+		return 0;
+
 	m->pg_init_count++;
 	m->pg_init_required = 0;
+
+	/* Check here to reset pg_init_required */
+	if (!m->current_pg)
+		return 0;
+
 	if (m->pg_init_delay_retry)
 		pg_init_delay = msecs_to_jiffies(m->pg_init_delay_msecs != DM_PG_INIT_DELAY_DEFAULT ?
 						 m->pg_init_delay_msecs : DM_PG_INIT_DELAY_MSECS);
@@ -274,6 +276,7 @@ static void __pg_init_all_paths(struct multipath *m)
 				       pg_init_delay))
 			m->pg_init_in_progress++;
 	}
+	return m->pg_init_in_progress;
 }
 
 static void __switch_pg(struct multipath *m, struct pgpath *pgpath)
@@ -365,19 +368,26 @@ static void __choose_pgpath(struct multipath *m, size_t nr_bytes)
  */
 static int __must_push_back(struct multipath *m)
 {
-	return (m->queue_if_no_path != m->saved_queue_if_no_path &&
-		dm_noflush_suspending(m->ti));
+	return (m->queue_if_no_path ||
+		(m->queue_if_no_path != m->saved_queue_if_no_path &&
+		 dm_noflush_suspending(m->ti)));
 }
 
-static int map_io(struct multipath *m, struct request *clone,
-		  union map_info *map_context, unsigned was_queued)
+#define pg_ready(m) (!(m)->queue_io && !(m)->pg_init_required)
+
+/*
+ * Map cloned requests
+ */
+static int multipath_map(struct dm_target *ti, struct request *clone,
+			 union map_info *map_context)
 {
-	int r = DM_MAPIO_REMAPPED;
+	struct multipath *m = (struct multipath *) ti->private;
+	int r = DM_MAPIO_REQUEUE;
 	size_t nr_bytes = blk_rq_bytes(clone);
 	unsigned long flags;
 	struct pgpath *pgpath;
 	struct block_device *bdev;
-	struct dm_mpath_io *mpio = map_context->ptr;
+	struct dm_mpath_io *mpio;
 
 	spin_lock_irqsave(&m->lock, flags);
 
@@ -388,38 +398,33 @@ static int map_io(struct multipath *m, struct request *clone,
 
 	pgpath = m->current_pgpath;
 
-	if (was_queued)
-		m->queue_size--;
-
-	if (m->pg_init_required) {
-		if (!m->pg_init_in_progress)
-			queue_work(kmultipathd, &m->process_queued_ios);
-		r = DM_MAPIO_REQUEUE;
-	} else if ((pgpath && m->queue_io) ||
-		   (!pgpath && m->queue_if_no_path)) {
-		/* Queue for the daemon to resubmit */
-		list_add_tail(&clone->queuelist, &m->queued_ios);
-		m->queue_size++;
-		if (!m->queue_io)
-			queue_work(kmultipathd, &m->process_queued_ios);
-		pgpath = NULL;
-		r = DM_MAPIO_SUBMITTED;
-	} else if (pgpath) {
-		bdev = pgpath->path.dev->bdev;
-		clone->q = bdev_get_queue(bdev);
-		clone->rq_disk = bdev->bd_disk;
-	} else if (__must_push_back(m))
-		r = DM_MAPIO_REQUEUE;
-	else
-		r = -EIO;	/* Failed */
+	if (!pgpath) {
+		if (!__must_push_back(m))
+			r = -EIO;	/* Failed */
+		goto out_unlock;
+	}
+	if (!pg_ready(m)) {
+		__pg_init_all_paths(m);
+		goto out_unlock;
+	}
+	if (set_mapinfo(m, map_context) < 0)
+		/* ENOMEM, requeue */
+		goto out_unlock;
 
+	bdev = pgpath->path.dev->bdev;
+	clone->q = bdev_get_queue(bdev);
+	clone->rq_disk = bdev->bd_disk;
+	clone->cmd_flags |= REQ_FAILFAST_TRANSPORT;
+	mpio = map_context->ptr;
 	mpio->pgpath = pgpath;
 	mpio->nr_bytes = nr_bytes;
-
-	if (r == DM_MAPIO_REMAPPED && pgpath->pg->ps.type->start_io)
-		pgpath->pg->ps.type->start_io(&pgpath->pg->ps, &pgpath->path,
+	if (pgpath->pg->ps.type->start_io)
+		pgpath->pg->ps.type->start_io(&pgpath->pg->ps,
+					      &pgpath->path,
 					      nr_bytes);
+	r = DM_MAPIO_REMAPPED;
 
+out_unlock:
 	spin_unlock_irqrestore(&m->lock, flags);
 
 	return r;
@@ -440,76 +445,14 @@ static int queue_if_no_path(struct multipath *m, unsigned queue_if_no_path,
 	else
 		m->saved_queue_if_no_path = queue_if_no_path;
 	m->queue_if_no_path = queue_if_no_path;
-	if (!m->queue_if_no_path && m->queue_size)
-		queue_work(kmultipathd, &m->process_queued_ios);
+	if (!m->queue_if_no_path)
+		dm_table_run_md_queue_async(m->ti->table);
 
 	spin_unlock_irqrestore(&m->lock, flags);
 
 	return 0;
 }
 
-/*-----------------------------------------------------------------
- * The multipath daemon is responsible for resubmitting queued ios.
- *---------------------------------------------------------------*/
-
-static void dispatch_queued_ios(struct multipath *m)
-{
-	int r;
-	unsigned long flags;
-	union map_info *info;
-	struct request *clone, *n;
-	LIST_HEAD(cl);
-
-	spin_lock_irqsave(&m->lock, flags);
-	list_splice_init(&m->queued_ios, &cl);
-	spin_unlock_irqrestore(&m->lock, flags);
-
-	list_for_each_entry_safe(clone, n, &cl, queuelist) {
-		list_del_init(&clone->queuelist);
-
-		info = dm_get_rq_mapinfo(clone);
-
-		r = map_io(m, clone, info, 1);
-		if (r < 0) {
-			clear_mapinfo(m, info);
-			dm_kill_unmapped_request(clone, r);
-		} else if (r == DM_MAPIO_REMAPPED)
-			dm_dispatch_request(clone);
-		else if (r == DM_MAPIO_REQUEUE) {
-			clear_mapinfo(m, info);
-			dm_requeue_unmapped_request(clone);
-		}
-	}
-}
-
-static void process_queued_ios(struct work_struct *work)
-{
-	struct multipath *m =
-		container_of(work, struct multipath, process_queued_ios);
-	struct pgpath *pgpath = NULL;
-	unsigned must_queue = 1;
-	unsigned long flags;
-
-	spin_lock_irqsave(&m->lock, flags);
-
-	if (!m->current_pgpath)
-		__choose_pgpath(m, 0);
-
-	pgpath = m->current_pgpath;
-
-	if ((pgpath && !m->queue_io) ||
-	    (!pgpath && !m->queue_if_no_path))
-		must_queue = 0;
-
-	if (m->pg_init_required && !m->pg_init_in_progress && pgpath &&
-	    !m->pg_init_disabled)
-		__pg_init_all_paths(m);
-
-	spin_unlock_irqrestore(&m->lock, flags);
-	if (!must_queue)
-		dispatch_queued_ios(m);
-}
-
 /*
  * An event is triggered whenever a path is taken out of use.
  * Includes path failure and PG bypass.
@@ -971,27 +914,6 @@ static void multipath_dtr(struct dm_target *ti)
 	free_multipath(m);
 }
 
-/*
- * Map cloned requests
- */
-static int multipath_map(struct dm_target *ti, struct request *clone,
-			 union map_info *map_context)
-{
-	int r;
-	struct multipath *m = (struct multipath *) ti->private;
-
-	if (set_mapinfo(m, map_context) < 0)
-		/* ENOMEM, requeue */
-		return DM_MAPIO_REQUEUE;
-
-	clone->cmd_flags |= REQ_FAILFAST_TRANSPORT;
-	r = map_io(m, clone, map_context, 0);
-	if (r < 0 || r == DM_MAPIO_REQUEUE)
-		clear_mapinfo(m, map_context);
-
-	return r;
-}
-
 /*
  * Take a path out of use.
  */
@@ -1054,9 +976,9 @@ static int reinstate_path(struct pgpath *pgpath)
 
 	pgpath->is_active = 1;
 
-	if (!m->nr_valid_paths++ && m->queue_size) {
+	if (!m->nr_valid_paths++) {
 		m->current_pgpath = NULL;
-		queue_work(kmultipathd, &m->process_queued_ios);
+		dm_table_run_md_queue_async(m->ti->table);
 	} else if (m->hw_handler_name && (m->current_pg == pgpath->pg)) {
 		if (queue_work(kmpath_handlerd, &pgpath->activate_path.work))
 			m->pg_init_in_progress++;
@@ -1252,11 +1174,12 @@ static void pg_init_done(void *data, int errors)
 		/* Activations of other paths are still on going */
 		goto out;
 
-	if (!m->pg_init_required)
-		m->queue_io = 0;
-
-	m->pg_init_delay_retry = delay_retry;
-	queue_work(kmultipathd, &m->process_queued_ios);
+	if (m->pg_init_required) {
+		m->pg_init_delay_retry = delay_retry;
+		if (__pg_init_all_paths(m))
+			goto out;
+	}
+	m->queue_io = 0;
 
 	/*
 	 * Wake up any thread waiting to suspend.
@@ -1272,8 +1195,11 @@ static void activate_path(struct work_struct *work)
 	struct pgpath *pgpath =
 		container_of(work, struct pgpath, activate_path.work);
 
-	scsi_dh_activate(bdev_get_queue(pgpath->path.dev->bdev),
-				pg_init_done, pgpath);
+	if (pgpath->is_active)
+		scsi_dh_activate(bdev_get_queue(pgpath->path.dev->bdev),
+				 pg_init_done, pgpath);
+	else
+		pg_init_done(pgpath, SCSI_DH_DEV_OFFLINED);
 }
 
 static int noretry_error(int error)
@@ -1433,7 +1359,7 @@ static void multipath_status(struct dm_target *ti, status_type_t type,
 
 	/* Features */
 	if (type == STATUSTYPE_INFO)
-		DMEMIT("2 %u %u ", m->queue_size, m->pg_init_count);
+		DMEMIT("2 %u %u ", m->queue_io, m->pg_init_count);
 	else {
 		DMEMIT("%u ", m->queue_if_no_path +
 			      (m->pg_init_retries > 0) * 2 +
@@ -1552,7 +1478,7 @@ static int multipath_message(struct dm_target *ti, unsigned argc, char **argv)
 	}
 
 	if (argc != 2) {
-		DMWARN("Unrecognised multipath message received.");
+		DMWARN("Invalid multipath message arguments. Expected 2 arguments, got %d.", argc);
 		goto out;
 	}
 
@@ -1570,7 +1496,7 @@ static int multipath_message(struct dm_target *ti, unsigned argc, char **argv)
 	else if (!strcasecmp(argv[0], "fail_path"))
 		action = fail_path;
 	else {
-		DMWARN("Unrecognised multipath message received.");
+		DMWARN("Unrecognised multipath message received: %s", argv[0]);
 		goto out;
 	}
 
@@ -1632,8 +1558,17 @@ static int multipath_ioctl(struct dm_target *ti, unsigned int cmd,
 			r = err;
 	}
 
-	if (r == -ENOTCONN && !fatal_signal_pending(current))
-		queue_work(kmultipathd, &m->process_queued_ios);
+	if (r == -ENOTCONN && !fatal_signal_pending(current)) {
+		spin_lock_irqsave(&m->lock, flags);
+		if (!m->current_pg) {
+			/* Path status changed, redo selection */
+			__choose_pgpath(m, 0);
+		}
+		if (m->pg_init_required)
+			__pg_init_all_paths(m);
+		spin_unlock_irqrestore(&m->lock, flags);
+		dm_table_run_md_queue_async(m->ti->table);
+	}
 
 	return r ? : __blkdev_driver_ioctl(bdev, mode, cmd, arg);
 }
@@ -1684,7 +1619,7 @@ static int multipath_busy(struct dm_target *ti)
 	spin_lock_irqsave(&m->lock, flags);
 
 	/* pg_init in progress, requeue until done */
-	if (m->pg_init_in_progress) {
+	if (!pg_ready(m)) {
 		busy = 1;
 		goto out;
 	}
@@ -1737,7 +1672,7 @@ static int multipath_busy(struct dm_target *ti)
  *---------------------------------------------------------------*/
 static struct target_type multipath_target = {
 	.name = "multipath",
-	.version = {1, 6, 0},
+	.version = {1, 7, 0},
 	.module = THIS_MODULE,
 	.ctr = multipath_ctr,
 	.dtr = multipath_dtr,

drivers/md/dm-table.c

@@ -945,7 +945,7 @@ bool dm_table_request_based(struct dm_table *t)
 	return dm_table_get_type(t) == DM_TYPE_REQUEST_BASED;
 }
 
-int dm_table_alloc_md_mempools(struct dm_table *t)
+static int dm_table_alloc_md_mempools(struct dm_table *t)
 {
 	unsigned type = dm_table_get_type(t);
 	unsigned per_bio_data_size = 0;
@@ -1618,6 +1618,25 @@ struct mapped_device *dm_table_get_md(struct dm_table *t)
 }
 EXPORT_SYMBOL(dm_table_get_md);
 
+void dm_table_run_md_queue_async(struct dm_table *t)
+{
+	struct mapped_device *md;
+	struct request_queue *queue;
+	unsigned long flags;
+
+	if (!dm_table_request_based(t))
+		return;
+
+	md = dm_table_get_md(t);
+	queue = dm_get_md_queue(md);
+	if (queue) {
+		spin_lock_irqsave(queue->queue_lock, flags);
+		blk_run_queue_async(queue);
+		spin_unlock_irqrestore(queue->queue_lock, flags);
+	}
+}
+EXPORT_SYMBOL(dm_table_run_md_queue_async);
+
 static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev,
 				  sector_t start, sector_t len, void *data)
 {

drivers/md/dm-thin-metadata.c

@@ -192,6 +192,13 @@ struct dm_pool_metadata {
 	 * operation possible in this state is the closing of the device.
 	 */
 	bool fail_io:1;
+
+	/*
+	 * Reading the space map roots can fail, so we read it into these
+	 * buffers before the superblock is locked and updated.
+	 */
+	__u8 data_space_map_root[SPACE_MAP_ROOT_SIZE];
+	__u8 metadata_space_map_root[SPACE_MAP_ROOT_SIZE];
 };
 
 struct dm_thin_device {
@@ -431,26 +438,53 @@ static void __setup_btree_details(struct dm_pool_metadata *pmd)
 	pmd->details_info.value_type.equal = NULL;
 }
 
+static int save_sm_roots(struct dm_pool_metadata *pmd)
+{
+	int r;
+	size_t len;
+
+	r = dm_sm_root_size(pmd->metadata_sm, &len);
+	if (r < 0)
+		return r;
+
+	r = dm_sm_copy_root(pmd->metadata_sm, &pmd->metadata_space_map_root, len);
+	if (r < 0)
+		return r;
+
+	r = dm_sm_root_size(pmd->data_sm, &len);
+	if (r < 0)
+		return r;
+
+	return dm_sm_copy_root(pmd->data_sm, &pmd->data_space_map_root, len);
+}
+
+static void copy_sm_roots(struct dm_pool_metadata *pmd,
+			  struct thin_disk_superblock *disk)
+{
+	memcpy(&disk->metadata_space_map_root,
+	       &pmd->metadata_space_map_root,
+	       sizeof(pmd->metadata_space_map_root));
+
+	memcpy(&disk->data_space_map_root,
+	       &pmd->data_space_map_root,
+	       sizeof(pmd->data_space_map_root));
+}
+
 static int __write_initial_superblock(struct dm_pool_metadata *pmd)
 {
 	int r;
 	struct dm_block *sblock;
-	size_t metadata_len, data_len;
 	struct thin_disk_superblock *disk_super;
 	sector_t bdev_size = i_size_read(pmd->bdev->bd_inode) >> SECTOR_SHIFT;
 
 	if (bdev_size > THIN_METADATA_MAX_SECTORS)
 		bdev_size = THIN_METADATA_MAX_SECTORS;
 
-	r = dm_sm_root_size(pmd->metadata_sm, &metadata_len);
-	if (r < 0)
-		return r;
-
-	r = dm_sm_root_size(pmd->data_sm, &data_len);
+	r = dm_sm_commit(pmd->data_sm);
 	if (r < 0)
 		return r;
 
-	r = dm_sm_commit(pmd->data_sm);
+	r = save_sm_roots(pmd);
 	if (r < 0)
 		return r;
 
@@ -471,15 +505,7 @@ static int __write_initial_superblock(struct dm_pool_metadata *pmd)
 	disk_super->trans_id = 0;
 	disk_super->held_root = 0;
 
-	r = dm_sm_copy_root(pmd->metadata_sm, &disk_super->metadata_space_map_root,
-			    metadata_len);
-	if (r < 0)
-		goto bad_locked;
-
-	r = dm_sm_copy_root(pmd->data_sm, &disk_super->data_space_map_root,
-			    data_len);
-	if (r < 0)
-		goto bad_locked;
+	copy_sm_roots(pmd, disk_super);
 
 	disk_super->data_mapping_root = cpu_to_le64(pmd->root);
 	disk_super->device_details_root = cpu_to_le64(pmd->details_root);
@@ -488,10 +514,6 @@ static int __write_initial_superblock(struct dm_pool_metadata *pmd)
 	disk_super->data_block_size = cpu_to_le32(pmd->data_block_size);
 
 	return dm_tm_commit(pmd->tm, sblock);
-
-bad_locked:
-	dm_bm_unlock(sblock);
-	return r;
 }
 
 static int __format_metadata(struct dm_pool_metadata *pmd)
@@ -769,6 +791,10 @@ static int __commit_transaction(struct dm_pool_metadata *pmd)
 	if (r < 0)
 		return r;
 
+	r = save_sm_roots(pmd);
+	if (r < 0)
+		return r;
+
 	r = superblock_lock(pmd, &sblock);
 	if (r)
 		return r;
@@ -780,21 +806,9 @@ static int __commit_transaction(struct dm_pool_metadata *pmd)
 	disk_super->trans_id = cpu_to_le64(pmd->trans_id);
 	disk_super->flags = cpu_to_le32(pmd->flags);
 
-	r = dm_sm_copy_root(pmd->metadata_sm, &disk_super->metadata_space_map_root,
-			    metadata_len);
-	if (r < 0)
-		goto out_locked;
-
-	r = dm_sm_copy_root(pmd->data_sm, &disk_super->data_space_map_root,
-			    data_len);
-	if (r < 0)
-		goto out_locked;
+	copy_sm_roots(pmd, disk_super);
 
 	return dm_tm_commit(pmd->tm, sblock);
-
-out_locked:
-	dm_bm_unlock(sblock);
-	return r;
 }
 
 struct dm_pool_metadata *dm_pool_metadata_open(struct block_device *bdev,

drivers/md/dm-thin.c

@@ -12,9 +12,11 @@
 #include <linux/dm-io.h>
 #include <linux/dm-kcopyd.h>
 #include <linux/list.h>
+#include <linux/rculist.h>
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/slab.h>
+#include <linux/rbtree.h>
 
 #define	DM_MSG_PREFIX	"thin"
 
@@ -178,12 +180,10 @@ struct pool {
 	unsigned ref_count;
 
 	spinlock_t lock;
-	struct bio_list deferred_bios;
 	struct bio_list deferred_flush_bios;
 	struct list_head prepared_mappings;
 	struct list_head prepared_discards;
-
-	struct bio_list retry_on_resume_list;
+	struct list_head active_thins;
 
 	struct dm_deferred_set *shared_read_ds;
 	struct dm_deferred_set *all_io_ds;
@@ -220,6 +220,7 @@ struct pool_c {
  * Target context for a thin.
  */
 struct thin_c {
+	struct list_head list;
 	struct dm_dev *pool_dev;
 	struct dm_dev *origin_dev;
 	dm_thin_id dev_id;
@@ -227,6 +228,10 @@ struct thin_c {
 	struct pool *pool;
 	struct dm_thin_device *td;
 	bool requeue_mode:1;
+	spinlock_t lock;
+	struct bio_list deferred_bio_list;
+	struct bio_list retry_on_resume_list;
+	struct rb_root sort_bio_list; /* sorted list of deferred bios */
 };
 
 /*----------------------------------------------------------------*/
@@ -287,9 +292,9 @@ static void cell_defer_no_holder_no_free(struct thin_c *tc,
 	struct pool *pool = tc->pool;
 	unsigned long flags;
 
-	spin_lock_irqsave(&pool->lock, flags);
-	dm_cell_release_no_holder(pool->prison, cell, &pool->deferred_bios);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
+	dm_cell_release_no_holder(pool->prison, cell, &tc->deferred_bio_list);
+	spin_unlock_irqrestore(&tc->lock, flags);
 
 	wake_worker(pool);
 }
@@ -368,6 +373,7 @@ struct dm_thin_endio_hook {
 	struct dm_deferred_entry *shared_read_entry;
 	struct dm_deferred_entry *all_io_entry;
 	struct dm_thin_new_mapping *overwrite_mapping;
+	struct rb_node rb_node;
 };
 
 static void requeue_bio_list(struct thin_c *tc, struct bio_list *master)
@@ -378,30 +384,22 @@ static void requeue_bio_list(struct thin_c *tc, struct bio_list *master)
 
 	bio_list_init(&bios);
 
-	spin_lock_irqsave(&tc->pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
 	bio_list_merge(&bios, master);
 	bio_list_init(master);
-	spin_unlock_irqrestore(&tc->pool->lock, flags);
+	spin_unlock_irqrestore(&tc->lock, flags);
 
-	while ((bio = bio_list_pop(&bios))) {
-		struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
-
-		if (h->tc == tc)
-			bio_endio(bio, DM_ENDIO_REQUEUE);
-		else
-			bio_list_add(master, bio);
-	}
+	while ((bio = bio_list_pop(&bios)))
+		bio_endio(bio, DM_ENDIO_REQUEUE);
 }
 
 static void requeue_io(struct thin_c *tc)
 {
-	struct pool *pool = tc->pool;
-
-	requeue_bio_list(tc, &pool->deferred_bios);
-	requeue_bio_list(tc, &pool->retry_on_resume_list);
+	requeue_bio_list(tc, &tc->deferred_bio_list);
+	requeue_bio_list(tc, &tc->retry_on_resume_list);
 }
 
-static void error_retry_list(struct pool *pool)
+static void error_thin_retry_list(struct thin_c *tc)
 {
 	struct bio *bio;
 	unsigned long flags;
@@ -409,15 +407,25 @@ static void error_retry_list(struct pool *pool)
 
 	bio_list_init(&bios);
 
-	spin_lock_irqsave(&pool->lock, flags);
-	bio_list_merge(&bios, &pool->retry_on_resume_list);
-	bio_list_init(&pool->retry_on_resume_list);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
+	bio_list_merge(&bios, &tc->retry_on_resume_list);
+	bio_list_init(&tc->retry_on_resume_list);
+	spin_unlock_irqrestore(&tc->lock, flags);
 
 	while ((bio = bio_list_pop(&bios)))
 		bio_io_error(bio);
 }
 
+static void error_retry_list(struct pool *pool)
+{
+	struct thin_c *tc;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(tc, &pool->active_thins, list)
+		error_thin_retry_list(tc);
+	rcu_read_unlock();
+}
+
 /*
  * This section of code contains the logic for processing a thin device's IO.
  * Much of the code depends on pool object resources (lists, workqueues, etc)
@@ -608,9 +616,9 @@ static void cell_defer(struct thin_c *tc, struct dm_bio_prison_cell *cell)
 	struct pool *pool = tc->pool;
 	unsigned long flags;
 
-	spin_lock_irqsave(&pool->lock, flags);
-	cell_release(pool, cell, &pool->deferred_bios);
-	spin_unlock_irqrestore(&tc->pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
+	cell_release(pool, cell, &tc->deferred_bio_list);
+	spin_unlock_irqrestore(&tc->lock, flags);
 
 	wake_worker(pool);
 }
@@ -623,9 +631,9 @@ static void cell_defer_no_holder(struct thin_c *tc, struct dm_bio_prison_cell *c
 	struct pool *pool = tc->pool;
 	unsigned long flags;
 
-	spin_lock_irqsave(&pool->lock, flags);
-	cell_release_no_holder(pool, cell, &pool->deferred_bios);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
+	cell_release_no_holder(pool, cell, &tc->deferred_bio_list);
+	spin_unlock_irqrestore(&tc->lock, flags);
 
 	wake_worker(pool);
 }
@@ -1001,12 +1009,11 @@ static void retry_on_resume(struct bio *bio)
 {
 	struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
 	struct thin_c *tc = h->tc;
-	struct pool *pool = tc->pool;
 	unsigned long flags;
 
-	spin_lock_irqsave(&pool->lock, flags);
-	bio_list_add(&pool->retry_on_resume_list, bio);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
+	bio_list_add(&tc->retry_on_resume_list, bio);
+	spin_unlock_irqrestore(&tc->lock, flags);
 }
 
 static bool should_error_unserviceable_bio(struct pool *pool)
@@ -1363,38 +1370,111 @@ static int need_commit_due_to_time(struct pool *pool)
 	       jiffies > pool->last_commit_jiffies + COMMIT_PERIOD;
 }
 
-static void process_deferred_bios(struct pool *pool)
+#define thin_pbd(node) rb_entry((node), struct dm_thin_endio_hook, rb_node)
+#define thin_bio(pbd) dm_bio_from_per_bio_data((pbd), sizeof(struct dm_thin_endio_hook))
+
+static void __thin_bio_rb_add(struct thin_c *tc, struct bio *bio)
+{
+	struct rb_node **rbp, *parent;
+	struct dm_thin_endio_hook *pbd;
+	sector_t bi_sector = bio->bi_iter.bi_sector;
+
+	rbp = &tc->sort_bio_list.rb_node;
+	parent = NULL;
+	while (*rbp) {
+		parent = *rbp;
+		pbd = thin_pbd(parent);
+
+		if (bi_sector < thin_bio(pbd)->bi_iter.bi_sector)
+			rbp = &(*rbp)->rb_left;
+		else
+			rbp = &(*rbp)->rb_right;
+	}
+
+	pbd = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
+	rb_link_node(&pbd->rb_node, parent, rbp);
+	rb_insert_color(&pbd->rb_node, &tc->sort_bio_list);
+}
+
+static void __extract_sorted_bios(struct thin_c *tc)
+{
+	struct rb_node *node;
+	struct dm_thin_endio_hook *pbd;
+	struct bio *bio;
+
+	for (node = rb_first(&tc->sort_bio_list); node; node = rb_next(node)) {
+		pbd = thin_pbd(node);
+		bio = thin_bio(pbd);
+
+		bio_list_add(&tc->deferred_bio_list, bio);
+		rb_erase(&pbd->rb_node, &tc->sort_bio_list);
+	}
+
+	WARN_ON(!RB_EMPTY_ROOT(&tc->sort_bio_list));
+}
+
+static void __sort_thin_deferred_bios(struct thin_c *tc)
+{
+	struct bio *bio;
+	struct bio_list bios;
+
+	bio_list_init(&bios);
+	bio_list_merge(&bios, &tc->deferred_bio_list);
+	bio_list_init(&tc->deferred_bio_list);
+
+	/* Sort deferred_bio_list using rb-tree */
+	while ((bio = bio_list_pop(&bios)))
+		__thin_bio_rb_add(tc, bio);
+
+	/*
+	 * Transfer the sorted bios in sort_bio_list back to
+	 * deferred_bio_list to allow lockless submission of
+	 * all bios.
+	 */
+	__extract_sorted_bios(tc);
+}
+
+static void process_thin_deferred_bios(struct thin_c *tc)
 {
+	struct pool *pool = tc->pool;
 	unsigned long flags;
 	struct bio *bio;
 	struct bio_list bios;
+	struct blk_plug plug;
+
+	if (tc->requeue_mode) {
+		requeue_bio_list(tc, &tc->deferred_bio_list);
+		return;
+	}
 
 	bio_list_init(&bios);
 
-	spin_lock_irqsave(&pool->lock, flags);
-	bio_list_merge(&bios, &pool->deferred_bios);
-	bio_list_init(&pool->deferred_bios);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
 
-	while ((bio = bio_list_pop(&bios))) {
-		struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
-		struct thin_c *tc = h->tc;
+	if (bio_list_empty(&tc->deferred_bio_list)) {
+		spin_unlock_irqrestore(&tc->lock, flags);
+		return;
+	}
 
-		if (tc->requeue_mode) {
-			bio_endio(bio, DM_ENDIO_REQUEUE);
-			continue;
-		}
+	__sort_thin_deferred_bios(tc);
+
+	bio_list_merge(&bios, &tc->deferred_bio_list);
+	bio_list_init(&tc->deferred_bio_list);
 
+	spin_unlock_irqrestore(&tc->lock, flags);
+
+	blk_start_plug(&plug);
+	while ((bio = bio_list_pop(&bios))) {
 		/*
 		 * If we've got no free new_mapping structs, and processing
 		 * this bio might require one, we pause until there are some
 		 * prepared mappings to process.
 		 */
 		if (ensure_next_mapping(pool)) {
-			spin_lock_irqsave(&pool->lock, flags);
-			bio_list_merge(&pool->deferred_bios, &bios);
-			spin_unlock_irqrestore(&pool->lock, flags);
-
+			spin_lock_irqsave(&tc->lock, flags);
+			bio_list_add(&tc->deferred_bio_list, bio);
+			bio_list_merge(&tc->deferred_bio_list, &bios);
+			spin_unlock_irqrestore(&tc->lock, flags);
 			break;
 		}
 
@@ -1403,6 +1483,20 @@ static void process_deferred_bios(struct pool *pool)
 		else
 			pool->process_bio(tc, bio);
 	}
+	blk_finish_plug(&plug);
+}
+
+static void process_deferred_bios(struct pool *pool)
+{
+	unsigned long flags;
+	struct bio *bio;
+	struct bio_list bios;
+	struct thin_c *tc;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(tc, &pool->active_thins, list)
+		process_thin_deferred_bios(tc);
+	rcu_read_unlock();
 
 	/*
 	 * If there are any deferred flush bios, we must commit
@@ -1634,9 +1728,9 @@ static void thin_defer_bio(struct thin_c *tc, struct bio *bio)
 	unsigned long flags;
 	struct pool *pool = tc->pool;
 
-	spin_lock_irqsave(&pool->lock, flags);
-	bio_list_add(&pool->deferred_bios, bio);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	spin_lock_irqsave(&tc->lock, flags);
+	bio_list_add(&tc->deferred_bio_list, bio);
+	spin_unlock_irqrestore(&tc->lock, flags);
 
 	wake_worker(pool);
 }
@@ -1757,26 +1851,29 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio)
 
 static int pool_is_congested(struct dm_target_callbacks *cb, int bdi_bits)
 {
-	int r;
-	unsigned long flags;
 	struct pool_c *pt = container_of(cb, struct pool_c, callbacks);
+	struct request_queue *q;
 
-	spin_lock_irqsave(&pt->pool->lock, flags);
-	r = !bio_list_empty(&pt->pool->retry_on_resume_list);
-	spin_unlock_irqrestore(&pt->pool->lock, flags);
+	if (get_pool_mode(pt->pool) == PM_OUT_OF_DATA_SPACE)
+		return 1;
 
-	if (!r) {
-		struct request_queue *q = bdev_get_queue(pt->data_dev->bdev);
-		r = bdi_congested(&q->backing_dev_info, bdi_bits);
-	}
-
-	return r;
+	q = bdev_get_queue(pt->data_dev->bdev);
+	return bdi_congested(&q->backing_dev_info, bdi_bits);
 }
 
-static void __requeue_bios(struct pool *pool)
+static void requeue_bios(struct pool *pool)
 {
-	bio_list_merge(&pool->deferred_bios, &pool->retry_on_resume_list);
-	bio_list_init(&pool->retry_on_resume_list);
+	unsigned long flags;
+	struct thin_c *tc;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(tc, &pool->active_thins, list) {
+		spin_lock_irqsave(&tc->lock, flags);
+		bio_list_merge(&tc->deferred_bio_list, &tc->retry_on_resume_list);
+		bio_list_init(&tc->retry_on_resume_list);
+		spin_unlock_irqrestore(&tc->lock, flags);
+	}
+	rcu_read_unlock();
 }
 
 /*----------------------------------------------------------------
@@ -1957,12 +2054,11 @@ static struct pool *pool_create(struct mapped_device *pool_md,
 	INIT_WORK(&pool->worker, do_worker);
 	INIT_DELAYED_WORK(&pool->waker, do_waker);
 	spin_lock_init(&pool->lock);
-	bio_list_init(&pool->deferred_bios);
 	bio_list_init(&pool->deferred_flush_bios);
 	INIT_LIST_HEAD(&pool->prepared_mappings);
 	INIT_LIST_HEAD(&pool->prepared_discards);
+	INIT_LIST_HEAD(&pool->active_thins);
 	pool->low_water_triggered = false;
-	bio_list_init(&pool->retry_on_resume_list);
 
 	pool->shared_read_ds = dm_deferred_set_create();
 	if (!pool->shared_read_ds) {
@@ -2507,8 +2603,8 @@ static void pool_resume(struct dm_target *ti)
 
 	spin_lock_irqsave(&pool->lock, flags);
 	pool->low_water_triggered = false;
-	__requeue_bios(pool);
 	spin_unlock_irqrestore(&pool->lock, flags);
+	requeue_bios(pool);
 
 	do_waker(&pool->waker.work);
 }
@@ -2947,7 +3043,7 @@ static struct target_type pool_target = {
 	.name = "thin-pool",
 	.features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE |
 		    DM_TARGET_IMMUTABLE,
-	.version = {1, 11, 0},
+	.version = {1, 12, 0},
 	.module = THIS_MODULE,
 	.ctr = pool_ctr,
 	.dtr = pool_dtr,
@@ -2968,6 +3064,12 @@ static struct target_type pool_target = {
 static void thin_dtr(struct dm_target *ti)
 {
 	struct thin_c *tc = ti->private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&tc->pool->lock, flags);
+	list_del_rcu(&tc->list);
+	spin_unlock_irqrestore(&tc->pool->lock, flags);
+	synchronize_rcu();
 
 	mutex_lock(&dm_thin_pool_table.mutex);
 
@@ -3014,6 +3116,10 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
 		r = -ENOMEM;
 		goto out_unlock;
 	}
+	spin_lock_init(&tc->lock);
+	bio_list_init(&tc->deferred_bio_list);
+	bio_list_init(&tc->retry_on_resume_list);
+	tc->sort_bio_list = RB_ROOT;
 
 	if (argc == 3) {
 		r = dm_get_device(ti, argv[2], FMODE_READ, &origin_dev);
@@ -3085,6 +3191,17 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
 	mutex_unlock(&dm_thin_pool_table.mutex);
 
+	spin_lock(&tc->pool->lock);
+	list_add_tail_rcu(&tc->list, &tc->pool->active_thins);
+	spin_unlock(&tc->pool->lock);
+	/*
+	 * This synchronize_rcu() call is needed here otherwise we risk a
+	 * wake_worker() call finding no bios to process (because the newly
+	 * added tc isn't yet visible).  So this reduces latency since we
+	 * aren't then dependent on the periodic commit to wake_worker().
+	 */
+	synchronize_rcu();
+
 	return 0;
 
 bad_target_max_io_len:
@@ -3250,7 +3367,7 @@ static int thin_iterate_devices(struct dm_target *ti,
 
 static struct target_type thin_target = {
 	.name = "thin",
-	.version = {1, 11, 0},
+	.version = {1, 12, 0},
 	.module	= THIS_MODULE,
 	.ctr = thin_ctr,
 	.dtr = thin_dtr,

drivers/md/dm.c

@@ -94,13 +94,6 @@ struct dm_rq_clone_bio_info {
 	struct bio clone;
 };
 
-union map_info *dm_get_mapinfo(struct bio *bio)
-{
-	if (bio && bio->bi_private)
-		return &((struct dm_target_io *)bio->bi_private)->info;
-	return NULL;
-}
-
 union map_info *dm_get_rq_mapinfo(struct request *rq)
 {
 	if (rq && rq->end_io_data)
@@ -475,6 +468,11 @@ sector_t dm_get_size(struct mapped_device *md)
 	return get_capacity(md->disk);
 }
 
+struct request_queue *dm_get_md_queue(struct mapped_device *md)
+{
+	return md->queue;
+}
+
 struct dm_stats *dm_get_stats(struct mapped_device *md)
 {
 	return &md->stats;
@@ -760,7 +758,7 @@ static void dec_pending(struct dm_io *io, int error)
 static void clone_endio(struct bio *bio, int error)
 {
 	int r = 0;
-	struct dm_target_io *tio = bio->bi_private;
+	struct dm_target_io *tio = container_of(bio, struct dm_target_io, clone);
 	struct dm_io *io = tio->io;
 	struct mapped_device *md = tio->io->md;
 	dm_endio_fn endio = tio->ti->type->end_io;
@@ -794,7 +792,8 @@ static void clone_endio(struct bio *bio, int error)
  */
 static void end_clone_bio(struct bio *clone, int error)
 {
-	struct dm_rq_clone_bio_info *info = clone->bi_private;
+	struct dm_rq_clone_bio_info *info =
+		container_of(clone, struct dm_rq_clone_bio_info, clone);
 	struct dm_rq_target_io *tio = info->tio;
 	struct bio *bio = info->orig;
 	unsigned int nr_bytes = info->orig->bi_iter.bi_size;
@@ -1120,7 +1119,6 @@ static void __map_bio(struct dm_target_io *tio)
 	struct dm_target *ti = tio->ti;
 
 	clone->bi_end_io = clone_endio;
-	clone->bi_private = tio;
 
 	/*
 	 * Map the clone.  If r == 0 we don't need to do
@@ -1195,7 +1193,6 @@ static struct dm_target_io *alloc_tio(struct clone_info *ci,
 
 	tio->io = ci->io;
 	tio->ti = ti;
-	memset(&tio->info, 0, sizeof(tio->info));
 	tio->target_bio_nr = target_bio_nr;
 
 	return tio;
@@ -1530,7 +1527,6 @@ static int dm_rq_bio_constructor(struct bio *bio, struct bio *bio_orig,
 	info->orig = bio_orig;
 	info->tio = tio;
 	bio->bi_end_io = end_clone_bio;
-	bio->bi_private = info;
 
 	return 0;
 }
@@ -2172,7 +2168,7 @@ static struct dm_table *__unbind(struct mapped_device *md)
 		return NULL;
 
 	dm_table_event_callback(map, NULL, NULL);
-	rcu_assign_pointer(md->map, NULL);
+	RCU_INIT_POINTER(md->map, NULL);
 	dm_sync_table(md);
 
 	return map;
@@ -2873,8 +2869,6 @@ static const struct block_device_operations dm_blk_dops = {
 	.owner = THIS_MODULE
 };
 
-EXPORT_SYMBOL(dm_get_mapinfo);
-
 /*
  * module hooks
  */

drivers/md/dm.h

@@ -73,7 +73,6 @@ unsigned dm_table_get_type(struct dm_table *t);
 struct target_type *dm_table_get_immutable_target_type(struct dm_table *t);
 bool dm_table_request_based(struct dm_table *t);
 bool dm_table_supports_discards(struct dm_table *t);
-int dm_table_alloc_md_mempools(struct dm_table *t);
 void dm_table_free_md_mempools(struct dm_table *t);
 struct dm_md_mempools *dm_table_get_md_mempools(struct dm_table *t);
 
@@ -189,6 +188,7 @@ int dm_lock_for_deletion(struct mapped_device *md, bool mark_deferred, bool only
 int dm_cancel_deferred_remove(struct mapped_device *md);
 int dm_request_based(struct mapped_device *md);
 sector_t dm_get_size(struct mapped_device *md);
+struct request_queue *dm_get_md_queue(struct mapped_device *md);
 struct dm_stats *dm_get_stats(struct mapped_device *md);
 
 int dm_kobject_uevent(struct mapped_device *md, enum kobject_action action,

drivers/md/persistent-data/dm-bitset.c

@@ -65,7 +65,7 @@ int dm_bitset_flush(struct dm_disk_bitset *info, dm_block_t root,
 	int r;
 	__le64 value;
 
-	if (!info->current_index_set)
+	if (!info->current_index_set || !info->dirty)
 		return 0;
 
 	value = cpu_to_le64(info->current_bits);
@@ -77,6 +77,8 @@ int dm_bitset_flush(struct dm_disk_bitset *info, dm_block_t root,
 		return r;
 
 	info->current_index_set = false;
+	info->dirty = false;
+
 	return 0;
 }
 EXPORT_SYMBOL_GPL(dm_bitset_flush);
@@ -94,6 +96,8 @@ static int read_bits(struct dm_disk_bitset *info, dm_block_t root,
 	info->current_bits = le64_to_cpu(value);
 	info->current_index_set = true;
 	info->current_index = array_index;
+	info->dirty = false;
+
 	return 0;
 }
 
@@ -126,6 +130,8 @@ int dm_bitset_set_bit(struct dm_disk_bitset *info, dm_block_t root,
 		return r;
 
 	set_bit(b, (unsigned long *) &info->current_bits);
+	info->dirty = true;
+
 	return 0;
 }
 EXPORT_SYMBOL_GPL(dm_bitset_set_bit);
@@ -141,6 +147,8 @@ int dm_bitset_clear_bit(struct dm_disk_bitset *info, dm_block_t root,
 		return r;
 
 	clear_bit(b, (unsigned long *) &info->current_bits);
+	info->dirty = true;
+
 	return 0;
 }
 EXPORT_SYMBOL_GPL(dm_bitset_clear_bit);

drivers/md/persistent-data/dm-bitset.h

@@ -71,6 +71,7 @@ struct dm_disk_bitset {
 	uint64_t current_bits;
 
 	bool current_index_set:1;
+	bool dirty:1;
 };
 
 /*

drivers/md/persistent-data/dm-block-manager.c

@@ -595,25 +595,14 @@ int dm_bm_unlock(struct dm_block *b)
 }
 EXPORT_SYMBOL_GPL(dm_bm_unlock);
 
-int dm_bm_flush_and_unlock(struct dm_block_manager *bm,
-			   struct dm_block *superblock)
+int dm_bm_flush(struct dm_block_manager *bm)
 {
-	int r;
-
 	if (bm->read_only)
 		return -EPERM;
 
-	r = dm_bufio_write_dirty_buffers(bm->bufio);
-	if (unlikely(r)) {
-		dm_bm_unlock(superblock);
-		return r;
-	}
-
-	dm_bm_unlock(superblock);
-
 	return dm_bufio_write_dirty_buffers(bm->bufio);
 }
-EXPORT_SYMBOL_GPL(dm_bm_flush_and_unlock);
+EXPORT_SYMBOL_GPL(dm_bm_flush);
 
 void dm_bm_prefetch(struct dm_block_manager *bm, dm_block_t b)
 {

drivers/md/persistent-data/dm-block-manager.h

@@ -105,8 +105,7 @@ int dm_bm_unlock(struct dm_block *b);
  *
  * This method always blocks.
  */
-int dm_bm_flush_and_unlock(struct dm_block_manager *bm,
-			   struct dm_block *superblock);
+int dm_bm_flush(struct dm_block_manager *bm);
 
 /*
  * Request data is prefetched into the cache.

drivers/md/persistent-data/dm-transaction-manager.c

@@ -154,7 +154,7 @@ int dm_tm_pre_commit(struct dm_transaction_manager *tm)
 	if (r < 0)
 		return r;
 
-	return 0;
+	return dm_bm_flush(tm->bm);
 }
 EXPORT_SYMBOL_GPL(dm_tm_pre_commit);
 
@@ -164,8 +164,9 @@ int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *root)
 		return -EWOULDBLOCK;
 
 	wipe_shadow_table(tm);
+	dm_bm_unlock(root);
 
-	return dm_bm_flush_and_unlock(tm->bm, root);
+	return dm_bm_flush(tm->bm);
 }
 EXPORT_SYMBOL_GPL(dm_tm_commit);
 

drivers/md/persistent-data/dm-transaction-manager.h

@@ -38,18 +38,17 @@ struct dm_transaction_manager *dm_tm_create_non_blocking_clone(struct dm_transac
 /*
  * We use a 2-phase commit here.
  *
- * i) In the first phase the block manager is told to start flushing, and
- * the changes to the space map are written to disk.  You should interrogate
- * your particular space map to get detail of its root node etc. to be
- * included in your superblock.
+ * i) Make all changes for the transaction *except* for the superblock.
+ * Then call dm_tm_pre_commit() to flush them to disk.
  *
- * ii) @root will be committed last.  You shouldn't use more than the
- * first 512 bytes of @root if you wish the transaction to survive a power
- * failure.  You *must* have a write lock held on @root for both stage (i)
- * and (ii).  The commit will drop the write lock.
+ * ii) Lock your superblock.  Update.  Then call dm_tm_commit() which will
+ * unlock the superblock and flush it.  No other blocks should be updated
+ * during this period.  Care should be taken to never unlock a partially
+ * updated superblock; perform any operations that could fail *before* you
+ * take the superblock lock.
  */
 int dm_tm_pre_commit(struct dm_transaction_manager *tm);
-int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *root);
+int dm_tm_commit(struct dm_transaction_manager *tm, struct dm_block *superblock);
 
 /*
  * These methods are the only way to get hold of a writeable block.

include/linux/device-mapper.h

@@ -23,7 +23,6 @@ typedef enum { STATUSTYPE_INFO, STATUSTYPE_TABLE } status_type_t;
 
 union map_info {
 	void *ptr;
-	unsigned long long ll;
 };
 
 /*
@@ -291,7 +290,6 @@ struct dm_target_callbacks {
 struct dm_target_io {
 	struct dm_io *io;
 	struct dm_target *ti;
-	union map_info info;
 	unsigned target_bio_nr;
 	struct bio clone;
 };
@@ -403,7 +401,6 @@ int dm_copy_name_and_uuid(struct mapped_device *md, char *name, char *uuid);
 struct gendisk *dm_disk(struct mapped_device *md);
 int dm_suspended(struct dm_target *ti);
 int dm_noflush_suspending(struct dm_target *ti);
-union map_info *dm_get_mapinfo(struct bio *bio);
 union map_info *dm_get_rq_mapinfo(struct request *rq);
 
 struct queue_limits *dm_get_queue_limits(struct mapped_device *md);
@@ -465,6 +462,11 @@ struct mapped_device *dm_table_get_md(struct dm_table *t);
  */
 void dm_table_event(struct dm_table *t);
 
+/*
+ * Run the queue for request-based targets.
+ */
+void dm_table_run_md_queue_async(struct dm_table *t);
+
 /*
  * The device must be suspended before calling this method.
  * Returns the previous table, which the caller must destroy.