Merge tag 'dm-3.9-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm

Pull device-mapper fixes from Alasdair G Kergon:
 "Fix reported data loss with discards and thin snapshots; avoid a
  deadlock observed in dm verity; fix a race in the new dm cache code
  along with some other minor bugs; store the cache policy version on
  disk to make the stored hints format future-proof."

* tag 'dm-3.9-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm:
  dm cache: policy ignore hints if generated by different version
  dm cache: policy change version from string to integer set
  dm cache: fix race in writethrough implementation
  dm cache: metadata clear dirty bits on clean shutdown
  dm cache: avoid calling policy destructor twice on error
  dm cache: detect cache_create failure
  dm cache: avoid 64 bit division on 32 bit
  dm verity: avoid deadlock
  dm thin: fix non power of two discard granularity calc
  dm thin: fix discard corruption

drivers/md/dm-bufio.c

@@ -1025,6 +1025,8 @@ void dm_bufio_prefetch(struct dm_bufio_client *c,
 {
 	struct blk_plug plug;
 
+	BUG_ON(dm_bufio_in_request());
+
 	blk_start_plug(&plug);
 	dm_bufio_lock(c);
 

drivers/md/dm-cache-metadata.c

@@ -83,6 +83,8 @@ struct cache_disk_superblock {
 	__le32 read_misses;
 	__le32 write_hits;
 	__le32 write_misses;
+
+	__le32 policy_version[CACHE_POLICY_VERSION_SIZE];
 } __packed;
 
 struct dm_cache_metadata {
@@ -109,6 +111,7 @@ struct dm_cache_metadata {
 	bool clean_when_opened:1;
 
 	char policy_name[CACHE_POLICY_NAME_SIZE];
+	unsigned policy_version[CACHE_POLICY_VERSION_SIZE];
 	size_t policy_hint_size;
 	struct dm_cache_statistics stats;
 };
@@ -268,7 +271,8 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 	memset(disk_super->uuid, 0, sizeof(disk_super->uuid));
 	disk_super->magic = cpu_to_le64(CACHE_SUPERBLOCK_MAGIC);
 	disk_super->version = cpu_to_le32(CACHE_VERSION);
-	memset(disk_super->policy_name, 0, CACHE_POLICY_NAME_SIZE);
+	memset(disk_super->policy_name, 0, sizeof(disk_super->policy_name));
+	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,
@@ -284,7 +288,6 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 	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);
-	memset(disk_super->policy_name, 0, sizeof(disk_super->policy_name));
 
 	disk_super->read_hits = cpu_to_le32(0);
 	disk_super->read_misses = cpu_to_le32(0);
@@ -478,6 +481,9 @@ static void read_superblock_fields(struct dm_cache_metadata *cmd,
 	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));
+	cmd->policy_version[0] = le32_to_cpu(disk_super->policy_version[0]);
+	cmd->policy_version[1] = le32_to_cpu(disk_super->policy_version[1]);
+	cmd->policy_version[2] = le32_to_cpu(disk_super->policy_version[2]);
 	cmd->policy_hint_size = le32_to_cpu(disk_super->policy_hint_size);
 
 	cmd->stats.read_hits = le32_to_cpu(disk_super->read_hits);
@@ -572,6 +578,9 @@ static int __commit_transaction(struct dm_cache_metadata *cmd,
 	disk_super->discard_nr_blocks = cpu_to_le64(from_dblock(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]);
+	disk_super->policy_version[1] = cpu_to_le32(cmd->policy_version[1]);
+	disk_super->policy_version[2] = cpu_to_le32(cmd->policy_version[2]);
 
 	disk_super->read_hits = cpu_to_le32(cmd->stats.read_hits);
 	disk_super->read_misses = cpu_to_le32(cmd->stats.read_misses);
@@ -854,18 +863,43 @@ struct thunk {
 	bool hints_valid;
 };
 
+static bool policy_unchanged(struct dm_cache_metadata *cmd,
+			     struct dm_cache_policy *policy)
+{
+	const char *policy_name = dm_cache_policy_get_name(policy);
+	const unsigned *policy_version = dm_cache_policy_get_version(policy);
+	size_t policy_hint_size = dm_cache_policy_get_hint_size(policy);
+
+	/*
+	 * Ensure policy names match.
+	 */
+	if (strncmp(cmd->policy_name, policy_name, sizeof(cmd->policy_name)))
+		return false;
+
+	/*
+	 * Ensure policy major versions match.
+	 */
+	if (cmd->policy_version[0] != policy_version[0])
+		return false;
+
+	/*
+	 * Ensure policy hint sizes match.
+	 */
+	if (cmd->policy_hint_size != policy_hint_size)
+		return false;
+
+	return true;
+}
+
 static bool hints_array_initialized(struct dm_cache_metadata *cmd)
 {
 	return cmd->hint_root && cmd->policy_hint_size;
 }
 
 static bool hints_array_available(struct dm_cache_metadata *cmd,
-				  const char *policy_name)
+				  struct dm_cache_policy *policy)
 {
-	bool policy_names_match = !strncmp(cmd->policy_name, policy_name,
-					   sizeof(cmd->policy_name));
-
-	return cmd->clean_when_opened && policy_names_match &&
+	return cmd->clean_when_opened && policy_unchanged(cmd, policy) &&
 		hints_array_initialized(cmd);
 }
 
@@ -899,7 +933,8 @@ static int __load_mapping(void *context, uint64_t cblock, void *leaf)
 	return r;
 }
 
-static int __load_mappings(struct dm_cache_metadata *cmd, const char *policy_name,
+static int __load_mappings(struct dm_cache_metadata *cmd,
+			   struct dm_cache_policy *policy,
 			   load_mapping_fn fn, void *context)
 {
 	struct thunk thunk;
@@ -909,18 +944,19 @@ static int __load_mappings(struct dm_cache_metadata *cmd, const char *policy_nam
 
 	thunk.cmd = cmd;
 	thunk.respect_dirty_flags = cmd->clean_when_opened;
-	thunk.hints_valid = hints_array_available(cmd, policy_name);
+	thunk.hints_valid = hints_array_available(cmd, policy);
 
 	return dm_array_walk(&cmd->info, cmd->root, __load_mapping, &thunk);
 }
 
-int dm_cache_load_mappings(struct dm_cache_metadata *cmd, const char *policy_name,
+int dm_cache_load_mappings(struct dm_cache_metadata *cmd,
+			   struct dm_cache_policy *policy,
 			   load_mapping_fn fn, void *context)
 {
 	int r;
 
 	down_read(&cmd->root_lock);
-	r = __load_mappings(cmd, policy_name, fn, context);
+	r = __load_mappings(cmd, policy, fn, context);
 	up_read(&cmd->root_lock);
 
 	return r;
@@ -979,7 +1015,7 @@ static int __dirty(struct dm_cache_metadata *cmd, dm_cblock_t cblock, bool dirty
 		/* nothing to be done */
 		return 0;
 
-	value = pack_value(oblock, flags | (dirty ? M_DIRTY : 0));
+	value = pack_value(oblock, (flags & ~M_DIRTY) | (dirty ? M_DIRTY : 0));
 	__dm_bless_for_disk(&value);
 
 	r = dm_array_set_value(&cmd->info, cmd->root, from_cblock(cblock),
@@ -1070,13 +1106,15 @@ static int begin_hints(struct dm_cache_metadata *cmd, struct dm_cache_policy *po
 	__le32 value;
 	size_t hint_size;
 	const char *policy_name = dm_cache_policy_get_name(policy);
+	const unsigned *policy_version = dm_cache_policy_get_version(policy);
 
 	if (!policy_name[0] ||
 	    (strlen(policy_name) > sizeof(cmd->policy_name) - 1))
 		return -EINVAL;
 
-	if (strcmp(cmd->policy_name, policy_name)) {
+	if (!policy_unchanged(cmd, policy)) {
 		strncpy(cmd->policy_name, policy_name, sizeof(cmd->policy_name));
+		memcpy(cmd->policy_version, policy_version, sizeof(cmd->policy_version));
 
 		hint_size = dm_cache_policy_get_hint_size(policy);
 		if (!hint_size)

drivers/md/dm-cache-metadata.h

@@ -89,7 +89,7 @@ typedef int (*load_mapping_fn)(void *context, dm_oblock_t oblock,
 			       dm_cblock_t cblock, bool dirty,
 			       uint32_t hint, bool hint_valid);
 int dm_cache_load_mappings(struct dm_cache_metadata *cmd,
-			   const char *policy_name,
+			   struct dm_cache_policy *policy,
 			   load_mapping_fn fn,
 			   void *context);
 

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

@@ -17,7 +17,6 @@
 /*----------------------------------------------------------------*/
 
 #define DM_MSG_PREFIX "cache cleaner"
-#define CLEANER_VERSION "1.0.0"
 
 /* Cache entry struct. */
 struct wb_cache_entry {
@@ -434,6 +433,7 @@ static struct dm_cache_policy *wb_create(dm_cblock_t cache_size,
 
 static struct dm_cache_policy_type wb_policy_type = {
 	.name = "cleaner",
+	.version = {1, 0, 0},
 	.hint_size = 0,
 	.owner = THIS_MODULE,
 	.create = wb_create
@@ -446,7 +446,10 @@ static int __init wb_init(void)
 	if (r < 0)
 		DMERR("register failed %d", r);
 	else
-		DMINFO("version " CLEANER_VERSION " loaded");
+		DMINFO("version %u.%u.%u loaded",
+		       wb_policy_type.version[0],
+		       wb_policy_type.version[1],
+		       wb_policy_type.version[2]);
 
 	return r;
 }

drivers/md/dm-cache-policy-internal.h

@@ -117,6 +117,8 @@ void dm_cache_policy_destroy(struct dm_cache_policy *p);
  */
 const char *dm_cache_policy_get_name(struct dm_cache_policy *p);
 
+const unsigned *dm_cache_policy_get_version(struct dm_cache_policy *p);
+
 size_t dm_cache_policy_get_hint_size(struct dm_cache_policy *p);
 
 /*----------------------------------------------------------------*/

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

@@ -14,7 +14,6 @@
 #include <linux/vmalloc.h>
 
 #define DM_MSG_PREFIX "cache-policy-mq"
-#define MQ_VERSION	"1.0.0"
 
 static struct kmem_cache *mq_entry_cache;
 
@@ -1133,6 +1132,7 @@ static struct dm_cache_policy *mq_create(dm_cblock_t cache_size,
 
 static struct dm_cache_policy_type mq_policy_type = {
 	.name = "mq",
+	.version = {1, 0, 0},
 	.hint_size = 4,
 	.owner = THIS_MODULE,
 	.create = mq_create
@@ -1140,6 +1140,7 @@ static struct dm_cache_policy_type mq_policy_type = {
 
 static struct dm_cache_policy_type default_policy_type = {
 	.name = "default",
+	.version = {1, 0, 0},
 	.hint_size = 4,
 	.owner = THIS_MODULE,
 	.create = mq_create
@@ -1164,7 +1165,10 @@ static int __init mq_init(void)
 
 	r = dm_cache_policy_register(&default_policy_type);
 	if (!r) {
-		DMINFO("version " MQ_VERSION " loaded");
+		DMINFO("version %u.%u.%u loaded",
+		       mq_policy_type.version[0],
+		       mq_policy_type.version[1],
+		       mq_policy_type.version[2]);
 		return 0;
 	}
 

drivers/md/dm-cache-policy.c

@@ -150,6 +150,14 @@ const char *dm_cache_policy_get_name(struct dm_cache_policy *p)
 }
 EXPORT_SYMBOL_GPL(dm_cache_policy_get_name);
 
+const unsigned *dm_cache_policy_get_version(struct dm_cache_policy *p)
+{
+	struct dm_cache_policy_type *t = p->private;
+
+	return t->version;
+}
+EXPORT_SYMBOL_GPL(dm_cache_policy_get_version);
+
 size_t dm_cache_policy_get_hint_size(struct dm_cache_policy *p)
 {
 	struct dm_cache_policy_type *t = p->private;

drivers/md/dm-cache-policy.h

@@ -196,6 +196,7 @@ struct dm_cache_policy {
  * We maintain a little register of the different policy types.
  */
 #define CACHE_POLICY_NAME_SIZE 16
+#define CACHE_POLICY_VERSION_SIZE 3
 
 struct dm_cache_policy_type {
 	/* For use by the register code only. */
@@ -206,6 +207,7 @@ struct dm_cache_policy_type {
 	 * what gets passed on the target line to select your policy.
 	 */
 	char name[CACHE_POLICY_NAME_SIZE];
+	unsigned version[CACHE_POLICY_VERSION_SIZE];
 
 	/*
 	 * Policies may store a hint for each each cache block.

drivers/md/dm-thin.c

@@ -1577,6 +1577,11 @@ static bool data_dev_supports_discard(struct pool_c *pt)
 	return q && blk_queue_discard(q);
 }
 
+static bool is_factor(sector_t block_size, uint32_t n)
+{
+	return !sector_div(block_size, n);
+}
+
 /*
  * If discard_passdown was enabled verify that the data device
  * supports discards.  Disable discard_passdown if not.
@@ -1602,7 +1607,7 @@ static void disable_passdown_if_not_supported(struct pool_c *pt)
 	else if (data_limits->discard_granularity > block_size)
 		reason = "discard granularity larger than a block";
 
-	else if (block_size & (data_limits->discard_granularity - 1))
+	else if (!is_factor(block_size, data_limits->discard_granularity))
 		reason = "discard granularity not a factor of block size";
 
 	if (reason) {
@@ -2544,7 +2549,7 @@ static struct target_type pool_target = {
 	.name = "thin-pool",
 	.features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE |
 		    DM_TARGET_IMMUTABLE,
-	.version = {1, 6, 1},
+	.version = {1, 7, 0},
 	.module = THIS_MODULE,
 	.ctr = pool_ctr,
 	.dtr = pool_dtr,
@@ -2831,7 +2836,7 @@ static int thin_iterate_devices(struct dm_target *ti,
 
 static struct target_type thin_target = {
 	.name = "thin",
-	.version = {1, 7, 1},
+	.version = {1, 8, 0},
 	.module	= THIS_MODULE,
 	.ctr = thin_ctr,
 	.dtr = thin_dtr,

drivers/md/dm-verity.c

@@ -93,6 +93,13 @@ struct dm_verity_io {
 	 */
 };
 
+struct dm_verity_prefetch_work {
+	struct work_struct work;
+	struct dm_verity *v;
+	sector_t block;
+	unsigned n_blocks;
+};
+
 static struct shash_desc *io_hash_desc(struct dm_verity *v, struct dm_verity_io *io)
 {
 	return (struct shash_desc *)(io + 1);
@@ -424,15 +431,18 @@ static void verity_end_io(struct bio *bio, int error)
  * The root buffer is not prefetched, it is assumed that it will be cached
  * all the time.
  */
-static void verity_prefetch_io(struct dm_verity *v, struct dm_verity_io *io)
+static void verity_prefetch_io(struct work_struct *work)
 {
+	struct dm_verity_prefetch_work *pw =
+		container_of(work, struct dm_verity_prefetch_work, work);
+	struct dm_verity *v = pw->v;
 	int i;
 
 	for (i = v->levels - 2; i >= 0; i--) {
 		sector_t hash_block_start;
 		sector_t hash_block_end;
-		verity_hash_at_level(v, io->block, i, &hash_block_start, NULL);
-		verity_hash_at_level(v, io->block + io->n_blocks - 1, i, &hash_block_end, NULL);
+		verity_hash_at_level(v, pw->block, i, &hash_block_start, NULL);
+		verity_hash_at_level(v, pw->block + pw->n_blocks - 1, i, &hash_block_end, NULL);
 		if (!i) {
 			unsigned cluster = ACCESS_ONCE(dm_verity_prefetch_cluster);
 
@@ -452,6 +462,25 @@ static void verity_prefetch_io(struct dm_verity *v, struct dm_verity_io *io)
 		dm_bufio_prefetch(v->bufio, hash_block_start,
 				  hash_block_end - hash_block_start + 1);
 	}
+
+	kfree(pw);
+}
+
+static void verity_submit_prefetch(struct dm_verity *v, struct dm_verity_io *io)
+{
+	struct dm_verity_prefetch_work *pw;
+
+	pw = kmalloc(sizeof(struct dm_verity_prefetch_work),
+		GFP_NOIO | __GFP_NORETRY | __GFP_NOMEMALLOC | __GFP_NOWARN);
+
+	if (!pw)
+		return;
+
+	INIT_WORK(&pw->work, verity_prefetch_io);
+	pw->v = v;
+	pw->block = io->block;
+	pw->n_blocks = io->n_blocks;
+	queue_work(v->verify_wq, &pw->work);
 }
 
 /*
@@ -498,7 +527,7 @@ static int verity_map(struct dm_target *ti, struct bio *bio)
 	memcpy(io->io_vec, bio_iovec(bio),
 	       io->io_vec_size * sizeof(struct bio_vec));
 
-	verity_prefetch_io(v, io);
+	verity_submit_prefetch(v, io);
 
 	generic_make_request(bio);
 
@@ -858,7 +887,7 @@ static int verity_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
 static struct target_type verity_target = {
 	.name		= "verity",
-	.version	= {1, 1, 1},
+	.version	= {1, 2, 0},
 	.module		= THIS_MODULE,
 	.ctr		= verity_ctr,
 	.dtr		= verity_dtr,

drivers/md/persistent-data/dm-btree-remove.c

@@ -139,15 +139,8 @@ struct child {
 	struct btree_node *n;
 };
 
-static struct dm_btree_value_type le64_type = {
-	.context = NULL,
-	.size = sizeof(__le64),
-	.inc = NULL,
-	.dec = NULL,
-	.equal = NULL
-};
-
-static int init_child(struct dm_btree_info *info, struct btree_node *parent,
+static int init_child(struct dm_btree_info *info, struct dm_btree_value_type *vt,
+		      struct btree_node *parent,
 		      unsigned index, struct child *result)
 {
 	int r, inc;
@@ -164,7 +157,7 @@ static int init_child(struct dm_btree_info *info, struct btree_node *parent,
 	result->n = dm_block_data(result->block);
 
 	if (inc)
-		inc_children(info->tm, result->n, &le64_type);
+		inc_children(info->tm, result->n, vt);
 
 	*((__le64 *) value_ptr(parent, index)) =
 		cpu_to_le64(dm_block_location(result->block));
@@ -236,7 +229,7 @@ static void __rebalance2(struct dm_btree_info *info, struct btree_node *parent,
 }
 
 static int rebalance2(struct shadow_spine *s, struct dm_btree_info *info,
-		      unsigned left_index)
+		      struct dm_btree_value_type *vt, unsigned left_index)
 {
 	int r;
 	struct btree_node *parent;
@@ -244,11 +237,11 @@ static int rebalance2(struct shadow_spine *s, struct dm_btree_info *info,
 
 	parent = dm_block_data(shadow_current(s));
 
-	r = init_child(info, parent, left_index, &left);
+	r = init_child(info, vt, parent, left_index, &left);
 	if (r)
 		return r;
 
-	r = init_child(info, parent, left_index + 1, &right);
+	r = init_child(info, vt, parent, left_index + 1, &right);
 	if (r) {
 		exit_child(info, &left);
 		return r;
@@ -368,7 +361,7 @@ static void __rebalance3(struct dm_btree_info *info, struct btree_node *parent,
 }
 
 static int rebalance3(struct shadow_spine *s, struct dm_btree_info *info,
-		      unsigned left_index)
+		      struct dm_btree_value_type *vt, unsigned left_index)
 {
 	int r;
 	struct btree_node *parent = dm_block_data(shadow_current(s));
@@ -377,17 +370,17 @@ static int rebalance3(struct shadow_spine *s, struct dm_btree_info *info,
 	/*
 	 * FIXME: fill out an array?
 	 */
-	r = init_child(info, parent, left_index, &left);
+	r = init_child(info, vt, parent, left_index, &left);
 	if (r)
 		return r;
 
-	r = init_child(info, parent, left_index + 1, &center);
+	r = init_child(info, vt, parent, left_index + 1, &center);
 	if (r) {
 		exit_child(info, &left);
 		return r;
 	}
 
-	r = init_child(info, parent, left_index + 2, &right);
+	r = init_child(info, vt, parent, left_index + 2, &right);
 	if (r) {
 		exit_child(info, &left);
 		exit_child(info, &center);
@@ -434,7 +427,8 @@ static int get_nr_entries(struct dm_transaction_manager *tm,
 }
 
 static int rebalance_children(struct shadow_spine *s,
-			      struct dm_btree_info *info, uint64_t key)
+			      struct dm_btree_info *info,
+			      struct dm_btree_value_type *vt, uint64_t key)
 {
 	int i, r, has_left_sibling, has_right_sibling;
 	uint32_t child_entries;
@@ -472,13 +466,13 @@ static int rebalance_children(struct shadow_spine *s,
 	has_right_sibling = i < (le32_to_cpu(n->header.nr_entries) - 1);
 
 	if (!has_left_sibling)
-		r = rebalance2(s, info, i);
+		r = rebalance2(s, info, vt, i);
 
 	else if (!has_right_sibling)
-		r = rebalance2(s, info, i - 1);
+		r = rebalance2(s, info, vt, i - 1);
 
 	else
-		r = rebalance3(s, info, i - 1);
+		r = rebalance3(s, info, vt, i - 1);
 
 	return r;
 }
@@ -529,7 +523,7 @@ static int remove_raw(struct shadow_spine *s, struct dm_btree_info *info,
 		if (le32_to_cpu(n->header.flags) & LEAF_NODE)
 			return do_leaf(n, key, index);
 
-		r = rebalance_children(s, info, key);
+		r = rebalance_children(s, info, vt, key);
 		if (r)
 			break;
 
@@ -550,6 +544,14 @@ static int remove_raw(struct shadow_spine *s, struct dm_btree_info *info,
 	return r;
 }
 
+static struct dm_btree_value_type le64_type = {
+	.context = NULL,
+	.size = sizeof(__le64),
+	.inc = NULL,
+	.dec = NULL,
+	.equal = NULL
+};
+
 int dm_btree_remove(struct dm_btree_info *info, dm_block_t root,
 		    uint64_t *keys, dm_block_t *new_root)
 {