Merge git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm

* git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-2.6-dm: (33 commits)
dm mpath: support discard
dm stripe: support discards
dm: split discard requests on target boundaries
dm stripe: optimize sector division
dm stripe: move sector translation to a function
dm: error return error for discards
dm delay: support discard
dm: zero silently drop discards
dm: use dm_target_offset macro
dm: factor out max_io_len_target_boundary
dm: use common __issue_target_request for flush and discard support
dm: linear support discard
dm crypt: simplify crypt_ctr
dm crypt: simplify crypt_config destruction logic
dm: allow autoloading of dm mod
dm: rename map_info flush_request to target_request_nr
dm ioctl: refactor dm_table_complete
dm snapshot: implement merge
dm: do not initialise full request queue when bio based
dm ioctl: make bio or request based device type immutable
...

Documentation/devices.txt

···
 		233 = /dev/kmview	View-OS A process with a view
 		234 = /dev/btrfs-control	Btrfs control device
 		235 = /dev/autofs	Autofs control device
+		236 = /dev/mapper/control	Device-Mapper control device
 		240-254			Reserved for local use
 		255			Reserved for MISC_DYNAMIC_MINOR
 

drivers/md/dm-crypt.c

···
 	struct workqueue_struct *io_queue;
 	struct workqueue_struct *crypt_queue;
 
-	/*
-	 * crypto related data
-	 */
+	char *cipher;
+	char *cipher_mode;
+
 	struct crypt_iv_operations *iv_gen_ops;
-	char *iv_mode;
 	union {
 		struct iv_essiv_private essiv;
 		struct iv_benbi_private benbi;
···
 	unsigned int dmreq_start;
 	struct ablkcipher_request *req;
 
-	char cipher[CRYPTO_MAX_ALG_NAME];
-	char chainmode[CRYPTO_MAX_ALG_NAME];
 	struct crypto_ablkcipher *tfm;
 	unsigned long flags;
 	unsigned int key_size;
···
 	return crypto_ablkcipher_setkey(cc->tfm, cc->key, cc->key_size);
 }
 
-/*
- * Construct an encryption mapping:
- * <cipher> <key> <iv_offset> <dev_path> <start>
- */
-static int crypt_ctr(struct dm_target *ti, unsigned int argc, char **argv)
+static void crypt_dtr(struct dm_target *ti)
 {
-	struct crypt_config *cc;
-	struct crypto_ablkcipher *tfm;
-	char *tmp;
-	char *cipher;
-	char *chainmode;
-	char *ivmode;
-	char *ivopts;
-	unsigned int key_size;
-	unsigned long long tmpll;
+	struct crypt_config *cc = ti->private;
 
-	if (argc != 5) {
-		ti->error = "Not enough arguments";
+	ti->private = NULL;
+
+	if (!cc)
+		return;
+
+	if (cc->io_queue)
+		destroy_workqueue(cc->io_queue);
+	if (cc->crypt_queue)
+		destroy_workqueue(cc->crypt_queue);
+
+	if (cc->bs)
+		bioset_free(cc->bs);
+
+	if (cc->page_pool)
+		mempool_destroy(cc->page_pool);
+	if (cc->req_pool)
+		mempool_destroy(cc->req_pool);
+	if (cc->io_pool)
+		mempool_destroy(cc->io_pool);
+
+	if (cc->iv_gen_ops && cc->iv_gen_ops->dtr)
+		cc->iv_gen_ops->dtr(cc);
+
+	if (cc->tfm && !IS_ERR(cc->tfm))
+		crypto_free_ablkcipher(cc->tfm);
+
+	if (cc->dev)
+		dm_put_device(ti, cc->dev);
+
+	kzfree(cc->cipher);
+	kzfree(cc->cipher_mode);
+
+	/* Must zero key material before freeing */
+	kzfree(cc);
+}
+
+static int crypt_ctr_cipher(struct dm_target *ti,
+			    char *cipher_in, char *key)
+{
+	struct crypt_config *cc = ti->private;
+	char *tmp, *cipher, *chainmode, *ivmode, *ivopts;
+	char *cipher_api = NULL;
+	int ret = -EINVAL;
+
+	/* Convert to crypto api definition? */
+	if (strchr(cipher_in, '(')) {
+		ti->error = "Bad cipher specification";
 		return -EINVAL;
 	}
 
-	tmp = argv[0];
+	/*
+	 * Legacy dm-crypt cipher specification
+	 * cipher-mode-iv:ivopts
+	 */
+	tmp = cipher_in;
 	cipher = strsep(&tmp, "-");
+
+	cc->cipher = kstrdup(cipher, GFP_KERNEL);
+	if (!cc->cipher)
+		goto bad_mem;
+
+	if (tmp) {
+		cc->cipher_mode = kstrdup(tmp, GFP_KERNEL);
+		if (!cc->cipher_mode)
+			goto bad_mem;
+	}
+
 	chainmode = strsep(&tmp, "-");
 	ivopts = strsep(&tmp, "-");
 	ivmode = strsep(&ivopts, ":");
 
 	if (tmp)
-		DMWARN("Unexpected additional cipher options");
+		DMWARN("Ignoring unexpected additional cipher options");
 
-	key_size = strlen(argv[1]) >> 1;
-
- 	cc = kzalloc(sizeof(*cc) + key_size * sizeof(u8), GFP_KERNEL);
-	if (cc == NULL) {
-		ti->error =
-			"Cannot allocate transparent encryption context";
-		return -ENOMEM;
-	}
-
-	/* Compatibility mode for old dm-crypt cipher strings */
-	if (!chainmode || (strcmp(chainmode, "plain") == 0 && !ivmode)) {
+	/* Compatibility mode for old dm-crypt mappings */
+	if (!chainmode || (!strcmp(chainmode, "plain") && !ivmode)) {
+		kfree(cc->cipher_mode);
+		cc->cipher_mode = kstrdup("cbc-plain", GFP_KERNEL);
 		chainmode = "cbc";
 		ivmode = "plain";
 	}
 
 	if (strcmp(chainmode, "ecb") && !ivmode) {
-		ti->error = "This chaining mode requires an IV mechanism";
-		goto bad_cipher;
+		ti->error = "IV mechanism required";
+		return -EINVAL;
 	}
 
-	if (snprintf(cc->cipher, CRYPTO_MAX_ALG_NAME, "%s(%s)",
-		     chainmode, cipher) >= CRYPTO_MAX_ALG_NAME) {
-		ti->error = "Chain mode + cipher name is too long";
-		goto bad_cipher;
+	cipher_api = kmalloc(CRYPTO_MAX_ALG_NAME, GFP_KERNEL);
+	if (!cipher_api)
+		goto bad_mem;
+
+	ret = snprintf(cipher_api, CRYPTO_MAX_ALG_NAME,
+		       "%s(%s)", chainmode, cipher);
+	if (ret < 0) {
+		kfree(cipher_api);
+		goto bad_mem;
 	}
 
-	tfm = crypto_alloc_ablkcipher(cc->cipher, 0, 0);
-	if (IS_ERR(tfm)) {
+	/* Allocate cipher */
+	cc->tfm = crypto_alloc_ablkcipher(cipher_api, 0, 0);
+	if (IS_ERR(cc->tfm)) {
+		ret = PTR_ERR(cc->tfm);
 		ti->error = "Error allocating crypto tfm";
-		goto bad_cipher;
+		goto bad;
 	}
 
-	strcpy(cc->cipher, cipher);
-	strcpy(cc->chainmode, chainmode);
-	cc->tfm = tfm;
-
-	if (crypt_set_key(cc, argv[1]) < 0) {
+	/* Initialize and set key */
+	ret = crypt_set_key(cc, key);
+	if (ret < 0) {
 		ti->error = "Error decoding and setting key";
-		goto bad_ivmode;
+		goto bad;
 	}
 
-	/*
-	 * Choose ivmode. Valid modes: "plain", "essiv:<esshash>", "benbi".
-	 * See comments at iv code
-	 */
+	/* Initialize IV */
+	cc->iv_size = crypto_ablkcipher_ivsize(cc->tfm);
+	if (cc->iv_size)
+		/* at least a 64 bit sector number should fit in our buffer */
+		cc->iv_size = max(cc->iv_size,
+				  (unsigned int)(sizeof(u64) / sizeof(u8)));
+	else if (ivmode) {
+		DMWARN("Selected cipher does not support IVs");
+		ivmode = NULL;
+	}
 
+	/* Choose ivmode, see comments at iv code. */
 	if (ivmode == NULL)
 		cc->iv_gen_ops = NULL;
 	else if (strcmp(ivmode, "plain") == 0)
···
 	else if (strcmp(ivmode, "null") == 0)
 		cc->iv_gen_ops = &crypt_iv_null_ops;
 	else {
+		ret = -EINVAL;
 		ti->error = "Invalid IV mode";
-		goto bad_ivmode;
+		goto bad;
 	}
 
-	if (cc->iv_gen_ops && cc->iv_gen_ops->ctr &&
-	    cc->iv_gen_ops->ctr(cc, ti, ivopts) < 0)
-		goto bad_ivmode;
-
-	if (cc->iv_gen_ops && cc->iv_gen_ops->init &&
-	    cc->iv_gen_ops->init(cc) < 0) {
-		ti->error = "Error initialising IV";
-		goto bad_slab_pool;
-	}
-
-	cc->iv_size = crypto_ablkcipher_ivsize(tfm);
-	if (cc->iv_size)
-		/* at least a 64 bit sector number should fit in our buffer */
-		cc->iv_size = max(cc->iv_size,
-				  (unsigned int)(sizeof(u64) / sizeof(u8)));
-	else {
-		if (cc->iv_gen_ops) {
-			DMWARN("Selected cipher does not support IVs");
-			if (cc->iv_gen_ops->dtr)
-				cc->iv_gen_ops->dtr(cc);
-			cc->iv_gen_ops = NULL;
+	/* Allocate IV */
+	if (cc->iv_gen_ops && cc->iv_gen_ops->ctr) {
+		ret = cc->iv_gen_ops->ctr(cc, ti, ivopts);
+		if (ret < 0) {
+			ti->error = "Error creating IV";
+			goto bad;
 		}
 	}
 
+	/* Initialize IV (set keys for ESSIV etc) */
+	if (cc->iv_gen_ops && cc->iv_gen_ops->init) {
+		ret = cc->iv_gen_ops->init(cc);
+		if (ret < 0) {
+			ti->error = "Error initialising IV";
+			goto bad;
+		}
+	}
+
+	ret = 0;
+bad:
+	kfree(cipher_api);
+	return ret;
+
+bad_mem:
+	ti->error = "Cannot allocate cipher strings";
+	return -ENOMEM;
+}
+
+/*
+ * Construct an encryption mapping:
+ * <cipher> <key> <iv_offset> <dev_path> <start>
+ */
+static int crypt_ctr(struct dm_target *ti, unsigned int argc, char **argv)
+{
+	struct crypt_config *cc;
+	unsigned int key_size;
+	unsigned long long tmpll;
+	int ret;
+
+	if (argc != 5) {
+		ti->error = "Not enough arguments";
+		return -EINVAL;
+	}
+
+	key_size = strlen(argv[1]) >> 1;
+
+	cc = kzalloc(sizeof(*cc) + key_size * sizeof(u8), GFP_KERNEL);
+	if (!cc) {
+		ti->error = "Cannot allocate encryption context";
+		return -ENOMEM;
+	}
+
+	ti->private = cc;
+	ret = crypt_ctr_cipher(ti, argv[0], argv[1]);
+	if (ret < 0)
+		goto bad;
+
+	ret = -ENOMEM;
 	cc->io_pool = mempool_create_slab_pool(MIN_IOS, _crypt_io_pool);
 	if (!cc->io_pool) {
 		ti->error = "Cannot allocate crypt io mempool";
-		goto bad_slab_pool;
+		goto bad;
 	}
 
 	cc->dmreq_start = sizeof(struct ablkcipher_request);
-	cc->dmreq_start += crypto_ablkcipher_reqsize(tfm);
+	cc->dmreq_start += crypto_ablkcipher_reqsize(cc->tfm);
 	cc->dmreq_start = ALIGN(cc->dmreq_start, crypto_tfm_ctx_alignment());
-	cc->dmreq_start += crypto_ablkcipher_alignmask(tfm) &
+	cc->dmreq_start += crypto_ablkcipher_alignmask(cc->tfm) &
 			   ~(crypto_tfm_ctx_alignment() - 1);
 
 	cc->req_pool = mempool_create_kmalloc_pool(MIN_IOS, cc->dmreq_start +
 			sizeof(struct dm_crypt_request) + cc->iv_size);
 	if (!cc->req_pool) {
 		ti->error = "Cannot allocate crypt request mempool";
-		goto bad_req_pool;
+		goto bad;
 	}
 	cc->req = NULL;
 
 	cc->page_pool = mempool_create_page_pool(MIN_POOL_PAGES, 0);
 	if (!cc->page_pool) {
 		ti->error = "Cannot allocate page mempool";
-		goto bad_page_pool;
+		goto bad;
 	}
 
 	cc->bs = bioset_create(MIN_IOS, 0);
 	if (!cc->bs) {
 		ti->error = "Cannot allocate crypt bioset";
-		goto bad_bs;
+		goto bad;
 	}
 
+	ret = -EINVAL;
 	if (sscanf(argv[2], "%llu", &tmpll) != 1) {
 		ti->error = "Invalid iv_offset sector";
-		goto bad_device;
+		goto bad;
 	}
 	cc->iv_offset = tmpll;
 
+	if (dm_get_device(ti, argv[3], dm_table_get_mode(ti->table), &cc->dev)) {
+		ti->error = "Device lookup failed";
+		goto bad;
+	}
+
 	if (sscanf(argv[4], "%llu", &tmpll) != 1) {
 		ti->error = "Invalid device sector";
-		goto bad_device;
+		goto bad;
 	}
 	cc->start = tmpll;
 
-	if (dm_get_device(ti, argv[3], dm_table_get_mode(ti->table), &cc->dev)) {
-		ti->error = "Device lookup failed";
-		goto bad_device;
-	}
-
-	if (ivmode && cc->iv_gen_ops) {
-		if (ivopts)
-			*(ivopts - 1) = ':';
-		cc->iv_mode = kmalloc(strlen(ivmode) + 1, GFP_KERNEL);
-		if (!cc->iv_mode) {
-			ti->error = "Error kmallocing iv_mode string";
-			goto bad_ivmode_string;
-		}
-		strcpy(cc->iv_mode, ivmode);
-	} else
-		cc->iv_mode = NULL;
-
+	ret = -ENOMEM;
 	cc->io_queue = create_singlethread_workqueue("kcryptd_io");
 	if (!cc->io_queue) {
 		ti->error = "Couldn't create kcryptd io queue";
-		goto bad_io_queue;
+		goto bad;
 	}
 
 	cc->crypt_queue = create_singlethread_workqueue("kcryptd");
 	if (!cc->crypt_queue) {
 		ti->error = "Couldn't create kcryptd queue";
-		goto bad_crypt_queue;
+		goto bad;
 	}
 
 	ti->num_flush_requests = 1;
-	ti->private = cc;
 	return 0;
 
-bad_crypt_queue:
-	destroy_workqueue(cc->io_queue);
-bad_io_queue:
-	kfree(cc->iv_mode);
-bad_ivmode_string:
-	dm_put_device(ti, cc->dev);
-bad_device:
-	bioset_free(cc->bs);
-bad_bs:
-	mempool_destroy(cc->page_pool);
-bad_page_pool:
-	mempool_destroy(cc->req_pool);
-bad_req_pool:
-	mempool_destroy(cc->io_pool);
-bad_slab_pool:
-	if (cc->iv_gen_ops && cc->iv_gen_ops->dtr)
-		cc->iv_gen_ops->dtr(cc);
-bad_ivmode:
-	crypto_free_ablkcipher(tfm);
-bad_cipher:
-	/* Must zero key material before freeing */
-	kzfree(cc);
-	return -EINVAL;
-}
-
-static void crypt_dtr(struct dm_target *ti)
-{
-	struct crypt_config *cc = (struct crypt_config *) ti->private;
-
-	destroy_workqueue(cc->io_queue);
-	destroy_workqueue(cc->crypt_queue);
-
-	if (cc->req)
-		mempool_free(cc->req, cc->req_pool);
-
-	bioset_free(cc->bs);
-	mempool_destroy(cc->page_pool);
-	mempool_destroy(cc->req_pool);
-	mempool_destroy(cc->io_pool);
-
-	kfree(cc->iv_mode);
-	if (cc->iv_gen_ops && cc->iv_gen_ops->dtr)
-		cc->iv_gen_ops->dtr(cc);
-	crypto_free_ablkcipher(cc->tfm);
-	dm_put_device(ti, cc->dev);
-
-	/* Must zero key material before freeing */
-	kzfree(cc);
+bad:
+	crypt_dtr(ti);
+	return ret;
 }
 
 static int crypt_map(struct dm_target *ti, struct bio *bio,
···
 		return DM_MAPIO_REMAPPED;
 	}
 
-	io = crypt_io_alloc(ti, bio, bio->bi_sector - ti->begin);
+	io = crypt_io_alloc(ti, bio, dm_target_offset(ti, bio->bi_sector));
 
 	if (bio_data_dir(io->base_bio) == READ)
 		kcryptd_queue_io(io);
···
 static int crypt_status(struct dm_target *ti, status_type_t type,
 			char *result, unsigned int maxlen)
 {
-	struct crypt_config *cc = (struct crypt_config *) ti->private;
+	struct crypt_config *cc = ti->private;
 	unsigned int sz = 0;
 
 	switch (type) {
···
 		break;
 
 	case STATUSTYPE_TABLE:
-		if (cc->iv_mode)
-			DMEMIT("%s-%s-%s ", cc->cipher, cc->chainmode,
-			       cc->iv_mode);
+		if (cc->cipher_mode)
+			DMEMIT("%s-%s ", cc->cipher, cc->cipher_mode);
 		else
-			DMEMIT("%s-%s ", cc->cipher, cc->chainmode);
+			DMEMIT("%s ", cc->cipher);
 
 		if (cc->key_size > 0) {
 			if ((maxlen - sz) < ((cc->key_size << 1) + 1))
···
 		return max_size;
 
 	bvm->bi_bdev = cc->dev->bdev;
-	bvm->bi_sector = cc->start + bvm->bi_sector - ti->begin;
+	bvm->bi_sector = cc->start + dm_target_offset(ti, bvm->bi_sector);
 
 	return min(max_size, q->merge_bvec_fn(q, bvm, biovec));
 }

drivers/md/dm-delay.c

···
 	atomic_set(&dc->may_delay, 1);
 
 	ti->num_flush_requests = 1;
+	ti->num_discard_requests = 1;
 	ti->private = dc;
 	return 0;
 
···
 		bio->bi_bdev = dc->dev_write->bdev;
 		if (bio_sectors(bio))
 			bio->bi_sector = dc->start_write +
-					 (bio->bi_sector - ti->begin);
+					 dm_target_offset(ti, bio->bi_sector);
 
 		return delay_bio(dc, dc->write_delay, bio);
 	}
 
 	bio->bi_bdev = dc->dev_read->bdev;
-	bio->bi_sector = dc->start_read +
-			 (bio->bi_sector - ti->begin);
+	bio->bi_sector = dc->start_read + dm_target_offset(ti, bio->bi_sector);
 
 	return delay_bio(dc, dc->read_delay, bio);
 }

drivers/md/dm-exception-store.c

···
 
 	/* Validate the chunk size against the device block size */
 	if (chunk_size %
-	    (bdev_logical_block_size(dm_snap_cow(store->snap)->bdev) >> 9)) {
+	    (bdev_logical_block_size(dm_snap_cow(store->snap)->bdev) >> 9) ||
+	    chunk_size %
+	    (bdev_logical_block_size(dm_snap_origin(store->snap)->bdev) >> 9)) {
 		*error = "Chunk size is not a multiple of device blocksize";
 		return -EINVAL;
 	}

drivers/md/dm-exception-store.h

···
 };
 
 /*
- * Obtain the cow device used by a given snapshot.
+ * Obtain the origin or cow device used by a given snapshot.
  */
+struct dm_dev *dm_snap_origin(struct dm_snapshot *snap);
 struct dm_dev *dm_snap_cow(struct dm_snapshot *snap);
 
 /*

drivers/md/dm-ioctl.c

···
 
 static void dm_hash_remove_all(int keep_open_devices)
 {
-	int i, dev_skipped, dev_removed;
+	int i, dev_skipped;
 	struct hash_cell *hc;
-	struct list_head *tmp, *n;
+	struct mapped_device *md;
+
+retry:
+	dev_skipped = 0;
 
 	down_write(&_hash_lock);
 
-retry:
-	dev_skipped = dev_removed = 0;
 	for (i = 0; i < NUM_BUCKETS; i++) {
-		list_for_each_safe (tmp, n, _name_buckets + i) {
-			hc = list_entry(tmp, struct hash_cell, name_list);
+		list_for_each_entry(hc, _name_buckets + i, name_list) {
+			md = hc->md;
+			dm_get(md);
 
-			if (keep_open_devices &&
-			    dm_lock_for_deletion(hc->md)) {
+			if (keep_open_devices && dm_lock_for_deletion(md)) {
+				dm_put(md);
 				dev_skipped++;
 				continue;
 			}
+
 			__hash_remove(hc);
-			dev_removed = 1;
+
+			up_write(&_hash_lock);
+
+			dm_put(md);
+			if (likely(keep_open_devices))
+				dm_destroy(md);
+			else
+				dm_destroy_immediate(md);
+
+			/*
+			 * Some mapped devices may be using other mapped
+			 * devices, so repeat until we make no further
+			 * progress.  If a new mapped device is created
+			 * here it will also get removed.
+			 */
+			goto retry;
 		}
 	}
 
-	/*
-	 * Some mapped devices may be using other mapped devices, so if any
-	 * still exist, repeat until we make no further progress.
-	 */
-	if (dev_skipped) {
-		if (dev_removed)
-			goto retry;
-
-		DMWARN("remove_all left %d open device(s)", dev_skipped);
-	}
-
 	up_write(&_hash_lock);
+
+	if (dev_skipped)
+		DMWARN("remove_all left %d open device(s)", dev_skipped);
 }
 
-static int dm_hash_rename(uint32_t cookie, uint32_t *flags, const char *old,
-			  const char *new)
+static struct mapped_device *dm_hash_rename(struct dm_ioctl *param,
+					    const char *new)
 {
 	char *new_name, *old_name;
 	struct hash_cell *hc;
 	struct dm_table *table;
+	struct mapped_device *md;
 
 	/*
 	 * duplicate new.
 	 */
 	new_name = kstrdup(new, GFP_KERNEL);
 	if (!new_name)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	down_write(&_hash_lock);
 
···
 	 */
 	hc = __get_name_cell(new);
 	if (hc) {
-		DMWARN("asked to rename to an already existing name %s -> %s",
-		       old, new);
+		DMWARN("asked to rename to an already-existing name %s -> %s",
+		       param->name, new);
 		dm_put(hc->md);
 		up_write(&_hash_lock);
 		kfree(new_name);
-		return -EBUSY;
+		return ERR_PTR(-EBUSY);
 	}
 
 	/*
 	 * Is there such a device as 'old' ?
 	 */
-	hc = __get_name_cell(old);
+	hc = __get_name_cell(param->name);
 	if (!hc) {
-		DMWARN("asked to rename a non existent device %s -> %s",
-		       old, new);
+		DMWARN("asked to rename a non-existent device %s -> %s",
+		       param->name, new);
 		up_write(&_hash_lock);
 		kfree(new_name);
-		return -ENXIO;
+		return ERR_PTR(-ENXIO);
 	}
 
 	/*
···
 		dm_table_put(table);
 	}
 
-	if (!dm_kobject_uevent(hc->md, KOBJ_CHANGE, cookie))
-		*flags |= DM_UEVENT_GENERATED_FLAG;
+	if (!dm_kobject_uevent(hc->md, KOBJ_CHANGE, param->event_nr))
+		param->flags |= DM_UEVENT_GENERATED_FLAG;
 
-	dm_put(hc->md);
+	md = hc->md;
 	up_write(&_hash_lock);
 	kfree(old_name);
-	return 0;
+
+	return md;
 }
 
 /*-----------------------------------------------------------------
···
  * Fills in a dm_ioctl structure, ready for sending back to
  * userland.
  */
-static int __dev_status(struct mapped_device *md, struct dm_ioctl *param)
+static void __dev_status(struct mapped_device *md, struct dm_ioctl *param)
 {
 	struct gendisk *disk = dm_disk(md);
 	struct dm_table *table;
···
 			dm_table_put(table);
 		}
 	}
-
-	return 0;
 }
 
 static int dev_create(struct dm_ioctl *param, size_t param_size)
···
 	r = dm_hash_insert(param->name, *param->uuid ? param->uuid : NULL, md);
 	if (r) {
 		dm_put(md);
+		dm_destroy(md);
 		return r;
 	}
 
 	param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
 
-	r = __dev_status(md, param);
+	__dev_status(md, param);
+
 	dm_put(md);
 
-	return r;
+	return 0;
 }
 
 /*
···
 		param->flags |= DM_UEVENT_GENERATED_FLAG;
 
 	dm_put(md);
+	dm_destroy(md);
 	return 0;
 }
 
···
 {
 	int r;
 	char *new_name = (char *) param + param->data_start;
+	struct mapped_device *md;
 
 	if (new_name < param->data ||
 	    invalid_str(new_name, (void *) param + param_size) ||
···
 	if (r)
 		return r;
 
-	param->data_size = 0;
+	md = dm_hash_rename(param, new_name);
+	if (IS_ERR(md))
+		return PTR_ERR(md);
 
-	return dm_hash_rename(param->event_nr, &param->flags, param->name,
-			      new_name);
+	__dev_status(md, param);
+	dm_put(md);
+
+	return 0;
 }
 
 static int dev_set_geometry(struct dm_ioctl *param, size_t param_size)
···
 	geometry.start = indata[3];
 
 	r = dm_set_geometry(md, &geometry);
-	if (!r)
-		r = __dev_status(md, param);
 
 	param->data_size = 0;
 
···
 	if (param->flags & DM_NOFLUSH_FLAG)
 		suspend_flags |= DM_SUSPEND_NOFLUSH_FLAG;
 
-	if (!dm_suspended_md(md))
+	if (!dm_suspended_md(md)) {
 		r = dm_suspend(md, suspend_flags);
+		if (r)
+			goto out;
+	}
 
-	if (!r)
-		r = __dev_status(md, param);
+	__dev_status(md, param);
 
+out:
 	dm_put(md);
+
 	return r;
 }
 
···
 		dm_table_destroy(old_map);
 
 	if (!r)
-		r = __dev_status(md, param);
+		__dev_status(md, param);
 
 	dm_put(md);
 	return r;
···
  */
 static int dev_status(struct dm_ioctl *param, size_t param_size)
 {
-	int r;
 	struct mapped_device *md;
 
 	md = find_device(param);
 	if (!md)
 		return -ENXIO;
 
-	r = __dev_status(md, param);
+	__dev_status(md, param);
 	dm_put(md);
-	return r;
+
+	return 0;
 }
 
 /*
···
  */
 static int dev_wait(struct dm_ioctl *param, size_t param_size)
 {
-	int r;
+	int r = 0;
 	struct mapped_device *md;
 	struct dm_table *table;
 
···
 	 * changed to trigger the event, so we may as well tell
 	 * him and save an ioctl.
 	 */
-	r = __dev_status(md, param);
-	if (r)
-		goto out;
+	__dev_status(md, param);
 
 	table = dm_get_live_or_inactive_table(md, param);
 	if (table) {
···
 		dm_table_put(table);
 	}
 
- out:
+out:
 	dm_put(md);
+
 	return r;
 }
 
···
 		next = spec->next;
 	}
 
-	r = dm_table_set_type(table);
-	if (r) {
-		DMWARN("unable to set table type");
-		return r;
-	}
-
 	return dm_table_complete(table);
-}
-
-static int table_prealloc_integrity(struct dm_table *t,
-				    struct mapped_device *md)
-{
-	struct list_head *devices = dm_table_get_devices(t);
-	struct dm_dev_internal *dd;
-
-	list_for_each_entry(dd, devices, list)
-		if (bdev_get_integrity(dd->dm_dev.bdev))
-			return blk_integrity_register(dm_disk(md), NULL);
-
-	return 0;
 }
 
 static int table_load(struct dm_ioctl *param, size_t param_size)
···
 		goto out;
 	}
 
-	r = table_prealloc_integrity(t, md);
-	if (r) {
-		DMERR("%s: could not register integrity profile.",
-		      dm_device_name(md));
+	/* Protect md->type and md->queue against concurrent table loads. */
+	dm_lock_md_type(md);
+	if (dm_get_md_type(md) == DM_TYPE_NONE)
+		/* Initial table load: acquire type of table. */
+		dm_set_md_type(md, dm_table_get_type(t));
+	else if (dm_get_md_type(md) != dm_table_get_type(t)) {
+		DMWARN("can't change device type after initial table load.");
 		dm_table_destroy(t);
+		dm_unlock_md_type(md);
+		r = -EINVAL;
 		goto out;
 	}
 
-	r = dm_table_alloc_md_mempools(t);
+	/* setup md->queue to reflect md's type (may block) */
+	r = dm_setup_md_queue(md);
 	if (r) {
-		DMWARN("unable to allocate mempools for this table");
+		DMWARN("unable to set up device queue for new table.");
 		dm_table_destroy(t);
+		dm_unlock_md_type(md);
 		goto out;
 	}
+	dm_unlock_md_type(md);
 
+	/* stage inactive table */
 	down_write(&_hash_lock);
 	hc = dm_get_mdptr(md);
 	if (!hc || hc->md != md) {
···
 	up_write(&_hash_lock);
 
 	param->flags |= DM_INACTIVE_PRESENT_FLAG;
-	r = __dev_status(md, param);
+	__dev_status(md, param);
 
 out:
 	dm_put(md);
···
 
 static int table_clear(struct dm_ioctl *param, size_t param_size)
 {
-	int r;
 	struct hash_cell *hc;
 	struct mapped_device *md;
 
···
 
 	param->flags &= ~DM_INACTIVE_PRESENT_FLAG;
 
-	r = __dev_status(hc->md, param);
+	__dev_status(hc->md, param);
 	md = hc->md;
 	up_write(&_hash_lock);
 	dm_put(md);
-	return r;
+
+	return 0;
 }
 
 /*
···
 
 static int table_deps(struct dm_ioctl *param, size_t param_size)
 {
-	int r = 0;
 	struct mapped_device *md;
 	struct dm_table *table;
 
···
 	if (!md)
 		return -ENXIO;
 
-	r = __dev_status(md, param);
-	if (r)
-		goto out;
+	__dev_status(md, param);
 
 	table = dm_get_live_or_inactive_table(md, param);
 	if (table) {
···
 		dm_table_put(table);
 	}
 
- out:
 	dm_put(md);
-	return r;
+
+	return 0;
 }
 
 /*
···
  */
 static int table_status(struct dm_ioctl *param, size_t param_size)
 {
-	int r;
 	struct mapped_device *md;
 	struct dm_table *table;
 
···
 	if (!md)
 		return -ENXIO;
 
-	r = __dev_status(md, param);
-	if (r)
-		goto out;
+	__dev_status(md, param);
 
 	table = dm_get_live_or_inactive_table(md, param);
 	if (table) {
···
 		dm_table_put(table);
 	}
 
-out:
 	dm_put(md);
-	return r;
+
+	return 0;
 }
 
 /*
···
 	md = find_device(param);
 	if (!md)
 		return -ENXIO;
-
-	r = __dev_status(md, param);
-	if (r)
-		goto out;
 
 	if (tmsg < (struct dm_target_msg *) param->data ||
 	    invalid_str(tmsg->message, (void *) param + param_size)) {
···
 #endif
 
 static const struct file_operations _ctl_fops = {
+	.open = nonseekable_open,
 	.unlocked_ioctl	 = dm_ctl_ioctl,
 	.compat_ioctl = dm_compat_ctl_ioctl,
 	.owner	 = THIS_MODULE,
 };
 
 static struct miscdevice _dm_misc = {
-	.minor 		= MISC_DYNAMIC_MINOR,
+	.minor		= MAPPER_CTRL_MINOR,
 	.name  		= DM_NAME,
-	.nodename	= "mapper/control",
+	.nodename	= DM_DIR "/" DM_CONTROL_NODE,
 	.fops  		= &_ctl_fops
 };
+
+MODULE_ALIAS_MISCDEV(MAPPER_CTRL_MINOR);
+MODULE_ALIAS("devname:" DM_DIR "/" DM_CONTROL_NODE);
 
 /*
  * Create misc character device and link to DM_DIR/control.

drivers/md/dm-linear.c

···
 	}
 
 	ti->num_flush_requests = 1;
+	ti->num_discard_requests = 1;
 	ti->private = lc;
 	return 0;
 
···
 {
 	struct linear_c *lc = ti->private;
 
-	return lc->start + (bi_sector - ti->begin);
+	return lc->start + dm_target_offset(ti, bi_sector);
 }
 
 static void linear_map_bio(struct dm_target *ti, struct bio *bio)

drivers/md/dm-mpath.c

···
 
 		if (as->argc < nr_params) {
 			ti->error = "not enough path parameters";
+			r = -EINVAL;
 			goto bad;
 		}
 
···
 	}
 
 	ti->num_flush_requests = 1;
+	ti->num_discard_requests = 1;
 
 	return 0;
 
···
 		return 0;	/* I/O complete */
 
 	if (error == -EOPNOTSUPP)
+		return error;
+
+	if (clone->cmd_flags & REQ_DISCARD)
+		/*
+		 * Pass all discard request failures up.
+		 * FIXME: only fail_path if the discard failed due to a
+		 * transport problem.  This requires precise understanding
+		 * of the underlying failure (e.g. the SCSI sense).
+		 */
 		return error;
 
 	if (mpio->pgpath)

drivers/md/dm-raid1.c

···
 {
 	if (unlikely(!bio->bi_size))
 		return 0;
-	return m->offset + (bio->bi_sector - m->ms->ti->begin);
+	return m->offset + dm_target_offset(m->ms->ti, bio->bi_sector);
 }
 
 static void map_bio(struct mirror *m, struct bio *bio)

drivers/md/dm-snap-persistent.c

···
  */
 static chunk_t area_location(struct pstore *ps, chunk_t area)
 {
-	return 1 + ((ps->exceptions_per_area + 1) * area);
+	return NUM_SNAPSHOT_HDR_CHUNKS + ((ps->exceptions_per_area + 1) * area);
 }
 
 /*
···
 	 * ps->current_area does not get reduced by prepare_merge() until
 	 * after commit_merge() has removed the nr_merged previous exceptions.
 	 */
-	ps->next_free = (area_location(ps, ps->current_area) - 1) +
-			(ps->current_committed + 1) + NUM_SNAPSHOT_HDR_CHUNKS;
+	ps->next_free = area_location(ps, ps->current_area) +
+			ps->current_committed + 1;
 
 	return 0;
 }

drivers/md/dm-snap.c

···
 #define RUNNING_MERGE          0
 #define SHUTDOWN_MERGE         1
 
+struct dm_dev *dm_snap_origin(struct dm_snapshot *s)
+{
+	return s->origin;
+}
+EXPORT_SYMBOL(dm_snap_origin);
+
 struct dm_dev *dm_snap_cow(struct dm_snapshot *s)
 {
 	return s->cow;
···
 		origin_mode = FMODE_WRITE;
 	}
 
-	origin_path = argv[0];
-	argv++;
-	argc--;
-
 	s = kmalloc(sizeof(*s), GFP_KERNEL);
 	if (!s) {
 		ti->error = "Cannot allocate snapshot context private "
 		    "structure";
 		r = -ENOMEM;
 		goto bad;
+	}
+
+	origin_path = argv[0];
+	argv++;
+	argc--;
+
+	r = dm_get_device(ti, origin_path, origin_mode, &s->origin);
+	if (r) {
+		ti->error = "Cannot get origin device";
+		goto bad_origin;
 	}
 
 	cow_path = argv[0];
···
 
 	argv += args_used;
 	argc -= args_used;
-
-	r = dm_get_device(ti, origin_path, origin_mode, &s->origin);
-	if (r) {
-		ti->error = "Cannot get origin device";
-		goto bad_origin;
-	}
 
 	s->ti = ti;
 	s->valid = 1;
···
 	dm_exception_table_exit(&s->complete, exception_cache);
 
 bad_hash_tables:
-	dm_put_device(ti, s->origin);
-
-bad_origin:
 	dm_exception_store_destroy(s->store);
 
 bad_store:
 	dm_put_device(ti, s->cow);
 
 bad_cow:
+	dm_put_device(ti, s->origin);
+
+bad_origin:
 	kfree(s);
 
 bad:
···
 
 	mempool_destroy(s->pending_pool);
 
-	dm_put_device(ti, s->origin);
-
 	dm_exception_store_destroy(s->store);
 
 	dm_put_device(ti, s->cow);
+
+	dm_put_device(ti, s->origin);
 
 	kfree(s);
 }
···
 	chunk_t chunk;
 
 	if (unlikely(bio_empty_barrier(bio))) {
-		if (!map_context->flush_request)
+		if (!map_context->target_request_nr)
 			bio->bi_bdev = s->origin->bdev;
 		else
 			bio->bi_bdev = s->cow->bdev;
···
 				    iterate_devices_callout_fn fn, void *data)
 {
 	struct dm_snapshot *snap = ti->private;
+	int r;
 
-	return fn(ti, snap->origin, 0, ti->len, data);
+	r = fn(ti, snap->origin, 0, ti->len, data);
+
+	if (!r)
+		r = fn(ti, snap->cow, 0, get_dev_size(snap->cow->bdev), data);
+
+	return r;
 }
 
 
···
 	return 0;
 }
 
+static int origin_merge(struct dm_target *ti, struct bvec_merge_data *bvm,
+			struct bio_vec *biovec, int max_size)
+{
+	struct dm_dev *dev = ti->private;
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	if (!q->merge_bvec_fn)
+		return max_size;
+
+	bvm->bi_bdev = dev->bdev;
+	bvm->bi_sector = bvm->bi_sector;
+
+	return min(max_size, q->merge_bvec_fn(q, bvm, biovec));
+}
+
 static int origin_iterate_devices(struct dm_target *ti,
 				  iterate_devices_callout_fn fn, void *data)
 {
···
 	.map     = origin_map,
 	.resume  = origin_resume,
 	.status  = origin_status,
+	.merge	 = origin_merge,
 	.iterate_devices = origin_iterate_devices,
 };
 

drivers/md/dm-stripe.c

···
 
 struct stripe_c {
 	uint32_t stripes;
+	int stripes_shift;
+	sector_t stripes_mask;
 
 	/* The size of this target / num. stripes */
 	sector_t stripe_width;
···
 
 	/* Set pointer to dm target; used in trigger_event */
 	sc->ti = ti;
-
 	sc->stripes = stripes;
 	sc->stripe_width = width;
+
+	if (stripes & (stripes - 1))
+		sc->stripes_shift = -1;
+	else {
+		sc->stripes_shift = ffs(stripes) - 1;
+		sc->stripes_mask = ((sector_t) stripes) - 1;
+	}
+
 	ti->split_io = chunk_size;
 	ti->num_flush_requests = stripes;
+	ti->num_discard_requests = stripes;
 
+	sc->chunk_shift = ffs(chunk_size) - 1;
 	sc->chunk_mask = ((sector_t) chunk_size) - 1;
-	for (sc->chunk_shift = 0; chunk_size; sc->chunk_shift++)
-		chunk_size >>= 1;
-	sc->chunk_shift--;
 
 	/*
 	 * Get the stripe destinations.
···
 	kfree(sc);
 }
 
+static void stripe_map_sector(struct stripe_c *sc, sector_t sector,
+			      uint32_t *stripe, sector_t *result)
+{
+	sector_t offset = dm_target_offset(sc->ti, sector);
+	sector_t chunk = offset >> sc->chunk_shift;
+
+	if (sc->stripes_shift < 0)
+		*stripe = sector_div(chunk, sc->stripes);
+	else {
+		*stripe = chunk & sc->stripes_mask;
+		chunk >>= sc->stripes_shift;
+	}
+
+	*result = (chunk << sc->chunk_shift) | (offset & sc->chunk_mask);
+}
+
+static void stripe_map_range_sector(struct stripe_c *sc, sector_t sector,
+				    uint32_t target_stripe, sector_t *result)
+{
+	uint32_t stripe;
+
+	stripe_map_sector(sc, sector, &stripe, result);
+	if (stripe == target_stripe)
+		return;
+	*result &= ~sc->chunk_mask;			/* round down */
+	if (target_stripe < stripe)
+		*result += sc->chunk_mask + 1;		/* next chunk */
+}
+
+static int stripe_map_discard(struct stripe_c *sc, struct bio *bio,
+			      uint32_t target_stripe)
+{
+	sector_t begin, end;
+
+	stripe_map_range_sector(sc, bio->bi_sector, target_stripe, &begin);
+	stripe_map_range_sector(sc, bio->bi_sector + bio_sectors(bio),
+				target_stripe, &end);
+	if (begin < end) {
+		bio->bi_bdev = sc->stripe[target_stripe].dev->bdev;
+		bio->bi_sector = begin + sc->stripe[target_stripe].physical_start;
+		bio->bi_size = to_bytes(end - begin);
+		return DM_MAPIO_REMAPPED;
+	} else {
+		/* The range doesn't map to the target stripe */
+		bio_endio(bio, 0);
+		return DM_MAPIO_SUBMITTED;
+	}
+}
+
 static int stripe_map(struct dm_target *ti, struct bio *bio,
 		      union map_info *map_context)
 {
-	struct stripe_c *sc = (struct stripe_c *) ti->private;
-	sector_t offset, chunk;
+	struct stripe_c *sc = ti->private;
 	uint32_t stripe;
+	unsigned target_request_nr;
 
 	if (unlikely(bio_empty_barrier(bio))) {
-		BUG_ON(map_context->flush_request >= sc->stripes);
-		bio->bi_bdev = sc->stripe[map_context->flush_request].dev->bdev;
+		target_request_nr = map_context->target_request_nr;
+		BUG_ON(target_request_nr >= sc->stripes);
+		bio->bi_bdev = sc->stripe[target_request_nr].dev->bdev;
 		return DM_MAPIO_REMAPPED;
 	}
+	if (unlikely(bio->bi_rw & REQ_DISCARD)) {
+		target_request_nr = map_context->target_request_nr;
+		BUG_ON(target_request_nr >= sc->stripes);
+		return stripe_map_discard(sc, bio, target_request_nr);
+	}
 
-	offset = bio->bi_sector - ti->begin;
-	chunk = offset >> sc->chunk_shift;
-	stripe = sector_div(chunk, sc->stripes);
+	stripe_map_sector(sc, bio->bi_sector, &stripe, &bio->bi_sector);
 
+	bio->bi_sector += sc->stripe[stripe].physical_start;
 	bio->bi_bdev = sc->stripe[stripe].dev->bdev;
-	bio->bi_sector = sc->stripe[stripe].physical_start +
-	    (chunk << sc->chunk_shift) + (offset & sc->chunk_mask);
+
 	return DM_MAPIO_REMAPPED;
 }
 

drivers/md/dm-table.c

···
 	sector_t *highs;
 	struct dm_target *targets;
 
+	unsigned discards_supported:1;
+
 	/*
 	 * Indicates the rw permissions for the new logical
 	 * device.  This should be a combination of FMODE_READ
···
 
 	INIT_LIST_HEAD(&t->devices);
 	atomic_set(&t->holders, 0);
+	t->discards_supported = 1;
 
 	if (!num_targets)
 		num_targets = KEYS_PER_NODE;
···
 		msleep(1);
 	smp_mb();
 
-	/* free the indexes (see dm_table_complete) */
+	/* free the indexes */
 	if (t->depth >= 2)
 		vfree(t->index[t->depth - 2]);
 
···
 
 	t->highs[t->num_targets++] = tgt->begin + tgt->len - 1;
 
+	if (!tgt->num_discard_requests)
+		t->discards_supported = 0;
+
 	return 0;
 
  bad:
···
 	return r;
 }
 
-int dm_table_set_type(struct dm_table *t)
+static int dm_table_set_type(struct dm_table *t)
 {
 	unsigned i;
 	unsigned bio_based = 0, request_based = 0;
···
 /*
  * Builds the btree to index the map.
  */
-int dm_table_complete(struct dm_table *t)
+static int dm_table_build_index(struct dm_table *t)
 {
 	int r = 0;
 	unsigned int leaf_nodes;
···
 
 	if (t->depth >= 2)
 		r = setup_indexes(t);
+
+	return r;
+}
+
+/*
+ * Register the mapped device for blk_integrity support if
+ * the underlying devices support it.
+ */
+static int dm_table_prealloc_integrity(struct dm_table *t, struct mapped_device *md)
+{
+	struct list_head *devices = dm_table_get_devices(t);
+	struct dm_dev_internal *dd;
+
+	list_for_each_entry(dd, devices, list)
+		if (bdev_get_integrity(dd->dm_dev.bdev))
+			return blk_integrity_register(dm_disk(md), NULL);
+
+	return 0;
+}
+
+/*
+ * Prepares the table for use by building the indices,
+ * setting the type, and allocating mempools.
+ */
+int dm_table_complete(struct dm_table *t)
+{
+	int r;
+
+	r = dm_table_set_type(t);
+	if (r) {
+		DMERR("unable to set table type");
+		return r;
+	}
+
+	r = dm_table_build_index(t);
+	if (r) {
+		DMERR("unable to build btrees");
+		return r;
+	}
+
+	r = dm_table_prealloc_integrity(t, t->md);
+	if (r) {
+		DMERR("could not register integrity profile.");
+		return r;
+	}
+
+	r = dm_table_alloc_md_mempools(t);
+	if (r)
+		DMERR("unable to allocate mempools");
 
 	return r;
 }
···
 	else
 		queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, q);
 
+	if (!dm_table_supports_discards(t))
+		queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, q);
+	else
+		queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
+
 	dm_table_set_integrity(t);
 
 	/*
···
 struct mapped_device *dm_table_get_md(struct dm_table *t)
 {
 	return t->md;
+}
+
+static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev,
+				  sector_t start, sector_t len, void *data)
+{
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	return q && blk_queue_discard(q);
+}
+
+bool dm_table_supports_discards(struct dm_table *t)
+{
+	struct dm_target *ti;
+	unsigned i = 0;
+
+	if (!t->discards_supported)
+		return 0;
+
+	/*
+	 * Ensure that at least one underlying device supports discards.
+	 * t->devices includes internal dm devices such as mirror logs
+	 * so we need to use iterate_devices here, which targets
+	 * supporting discard must provide.
+	 */
+	while (i < dm_table_get_num_targets(t)) {
+		ti = dm_table_get_target(t, i++);
+
+		if (ti->type->iterate_devices &&
+		    ti->type->iterate_devices(ti, device_discard_capable, NULL))
+			return 1;
+	}
+
+	return 0;
 }
 
 EXPORT_SYMBOL(dm_vcalloc);

drivers/md/dm-target.c

···
  */
 static int io_err_ctr(struct dm_target *tt, unsigned int argc, char **args)
 {
+	/*
+	 * Return error for discards instead of -EOPNOTSUPP
+	 */
+	tt->num_discard_requests = 1;
+
 	return 0;
 }
 

drivers/md/dm-zero.c

···
 		return -EINVAL;
 	}
 
+	/*
+	 * Silently drop discards, avoiding -EOPNOTSUPP.
+	 */
+	ti->num_discard_requests = 1;
+
 	return 0;
 }
 

drivers/md/dm.c

···
 #include <linux/slab.h>
 #include <linux/idr.h>
 #include <linux/hdreg.h>
+#include <linux/delay.h>
 
 #include <trace/events/block.h>
 
···
 	unsigned long flags;
 
 	struct request_queue *queue;
+	unsigned type;
+	/* Protect queue and type against concurrent access. */
+	struct mutex type_lock;
+
 	struct gendisk *disk;
 	char name[16];
 
···
 			 * There can be just one barrier request so we use
 			 * a per-device variable for error reporting.
 			 * Note that you can't touch the bio after end_io_acct
+			 *
+			 * We ignore -EOPNOTSUPP for empty flush reported by
+			 * underlying devices. We assume that if the device
+			 * doesn't support empty barriers, it doesn't need
+			 * cache flushing commands.
 			 */
-			if (!md->barrier_error && io_error != -EOPNOTSUPP)
+			if (!md->barrier_error &&
+			    !(bio_empty_barrier(bio) && io_error == -EOPNOTSUPP))
 				md->barrier_error = io_error;
 			end_io_acct(io);
 			free_io(md, io);
···
 	dm_complete_request(clone, error);
 }
 
-static sector_t max_io_len(struct mapped_device *md,
-			   sector_t sector, struct dm_target *ti)
+/*
+ * Return maximum size of I/O possible at the supplied sector up to the current
+ * target boundary.
+ */
+static sector_t max_io_len_target_boundary(sector_t sector, struct dm_target *ti)
 {
-	sector_t offset = sector - ti->begin;
-	sector_t len = ti->len - offset;
+	sector_t target_offset = dm_target_offset(ti, sector);
+
+	return ti->len - target_offset;
+}
+
+static sector_t max_io_len(sector_t sector, struct dm_target *ti)
+{
+	sector_t len = max_io_len_target_boundary(sector, ti);
 
 	/*
 	 * Does the target need to split even further ?
 	 */
 	if (ti->split_io) {
 		sector_t boundary;
+		sector_t offset = dm_target_offset(ti, sector);
 		boundary = ((offset + ti->split_io) & ~(ti->split_io - 1))
 			   - offset;
 		if (len > boundary)
···
 	return tio;
 }
 
-static void __flush_target(struct clone_info *ci, struct dm_target *ti,
-			  unsigned flush_nr)
+static void __issue_target_request(struct clone_info *ci, struct dm_target *ti,
+				   unsigned request_nr, sector_t len)
 {
 	struct dm_target_io *tio = alloc_tio(ci, ti);
 	struct bio *clone;
 
-	tio->info.flush_request = flush_nr;
+	tio->info.target_request_nr = request_nr;
 
-	clone = bio_alloc_bioset(GFP_NOIO, 0, ci->md->bs);
+	/*
+	 * Discard requests require the bio's inline iovecs be initialized.
+	 * ci->bio->bi_max_vecs is BIO_INLINE_VECS anyway, for both flush
+	 * and discard, so no need for concern about wasted bvec allocations.
+	 */
+	clone = bio_alloc_bioset(GFP_NOIO, ci->bio->bi_max_vecs, ci->md->bs);
 	__bio_clone(clone, ci->bio);
 	clone->bi_destructor = dm_bio_destructor;
+	if (len) {
+		clone->bi_sector = ci->sector;
+		clone->bi_size = to_bytes(len);
+	}
 
 	__map_bio(ti, clone, tio);
 }
 
+static void __issue_target_requests(struct clone_info *ci, struct dm_target *ti,
+				    unsigned num_requests, sector_t len)
+{
+	unsigned request_nr;
+
+	for (request_nr = 0; request_nr < num_requests; request_nr++)
+		__issue_target_request(ci, ti, request_nr, len);
+}
+
 static int __clone_and_map_empty_barrier(struct clone_info *ci)
 {
-	unsigned target_nr = 0, flush_nr;
+	unsigned target_nr = 0;
 	struct dm_target *ti;
 
 	while ((ti = dm_table_get_target(ci->map, target_nr++)))
-		for (flush_nr = 0; flush_nr < ti->num_flush_requests;
-		     flush_nr++)
-			__flush_target(ci, ti, flush_nr);
+		__issue_target_requests(ci, ti, ti->num_flush_requests, 0);
 
 	ci->sector_count = 0;
+
+	return 0;
+}
+
+/*
+ * Perform all io with a single clone.
+ */
+static void __clone_and_map_simple(struct clone_info *ci, struct dm_target *ti)
+{
+	struct bio *clone, *bio = ci->bio;
+	struct dm_target_io *tio;
+
+	tio = alloc_tio(ci, ti);
+	clone = clone_bio(bio, ci->sector, ci->idx,
+			  bio->bi_vcnt - ci->idx, ci->sector_count,
+			  ci->md->bs);
+	__map_bio(ti, clone, tio);
+	ci->sector_count = 0;
+}
+
+static int __clone_and_map_discard(struct clone_info *ci)
+{
+	struct dm_target *ti;
+	sector_t len;
+
+	do {
+		ti = dm_table_find_target(ci->map, ci->sector);
+		if (!dm_target_is_valid(ti))
+			return -EIO;
+
+		/*
+		 * Even though the device advertised discard support,
+		 * reconfiguration might have changed that since the
+		 * check was performed.
+		 */
+		if (!ti->num_discard_requests)
+			return -EOPNOTSUPP;
+
+		len = min(ci->sector_count, max_io_len_target_boundary(ci->sector, ti));
+
+		__issue_target_requests(ci, ti, ti->num_discard_requests, len);
+
+		ci->sector += len;
+	} while (ci->sector_count -= len);
 
 	return 0;
 }
···
 	if (unlikely(bio_empty_barrier(bio)))
 		return __clone_and_map_empty_barrier(ci);
 
+	if (unlikely(bio->bi_rw & REQ_DISCARD))
+		return __clone_and_map_discard(ci);
+
 	ti = dm_table_find_target(ci->map, ci->sector);
 	if (!dm_target_is_valid(ti))
 		return -EIO;
 
-	max = max_io_len(ci->md, ci->sector, ti);
-
-	/*
-	 * Allocate a target io object.
-	 */
-	tio = alloc_tio(ci, ti);
+	max = max_io_len(ci->sector, ti);
 
 	if (ci->sector_count <= max) {
 		/*
 		 * Optimise for the simple case where we can do all of
 		 * the remaining io with a single clone.
 		 */
-		clone = clone_bio(bio, ci->sector, ci->idx,
-				  bio->bi_vcnt - ci->idx, ci->sector_count,
-				  ci->md->bs);
-		__map_bio(ti, clone, tio);
-		ci->sector_count = 0;
+		__clone_and_map_simple(ci, ti);
 
 	} else if (to_sector(bio->bi_io_vec[ci->idx].bv_len) <= max) {
 		/*
···
 			len += bv_len;
 		}
 
+		tio = alloc_tio(ci, ti);
 		clone = clone_bio(bio, ci->sector, ci->idx, i - ci->idx, len,
 				  ci->md->bs);
 		__map_bio(ti, clone, tio);
···
 				if (!dm_target_is_valid(ti))
 					return -EIO;
 
-				max = max_io_len(ci->md, ci->sector, ti);
-
-				tio = alloc_tio(ci, ti);
+				max = max_io_len(ci->sector, ti);
 			}
 
 			len = min(remaining, max);
 
+			tio = alloc_tio(ci, ti);
 			clone = split_bvec(bio, ci->sector, ci->idx,
 					   bv->bv_offset + offset, len,
 					   ci->md->bs);
···
 	/*
 	 * Find maximum amount of I/O that won't need splitting
 	 */
-	max_sectors = min(max_io_len(md, bvm->bi_sector, ti),
+	max_sectors = min(max_io_len(bvm->bi_sector, ti),
 			  (sector_t) BIO_MAX_SECTORS);
 	max_size = (max_sectors << SECTOR_SHIFT) - bvm->bi_size;
 	if (max_size < 0)
···
 static void dm_wq_work(struct work_struct *work);
 static void dm_rq_barrier_work(struct work_struct *work);
 
+static void dm_init_md_queue(struct mapped_device *md)
+{
+	/*
+	 * Request-based dm devices cannot be stacked on top of bio-based dm
+	 * devices.  The type of this dm device has not been decided yet.
+	 * The type is decided at the first table loading time.
+	 * To prevent problematic device stacking, clear the queue flag
+	 * for request stacking support until then.
+	 *
+	 * This queue is new, so no concurrency on the queue_flags.
+	 */
+	queue_flag_clear_unlocked(QUEUE_FLAG_STACKABLE, md->queue);
+
+	md->queue->queuedata = md;
+	md->queue->backing_dev_info.congested_fn = dm_any_congested;
+	md->queue->backing_dev_info.congested_data = md;
+	blk_queue_make_request(md->queue, dm_request);
+	blk_queue_bounce_limit(md->queue, BLK_BOUNCE_ANY);
+	md->queue->unplug_fn = dm_unplug_all;
+	blk_queue_merge_bvec(md->queue, dm_merge_bvec);
+}
+
 /*
  * Allocate and initialise a blank device with a given minor.
  */
···
 	if (r < 0)
 		goto bad_minor;
 
+	md->type = DM_TYPE_NONE;
 	init_rwsem(&md->io_lock);
 	mutex_init(&md->suspend_lock);
+	mutex_init(&md->type_lock);
 	spin_lock_init(&md->deferred_lock);
 	spin_lock_init(&md->barrier_error_lock);
 	rwlock_init(&md->map_lock);
···
 	INIT_LIST_HEAD(&md->uevent_list);
 	spin_lock_init(&md->uevent_lock);
 
-	md->queue = blk_init_queue(dm_request_fn, NULL);
+	md->queue = blk_alloc_queue(GFP_KERNEL);
 	if (!md->queue)
 		goto bad_queue;
 
-	/*
-	 * Request-based dm devices cannot be stacked on top of bio-based dm
-	 * devices.  The type of this dm device has not been decided yet,
-	 * although we initialized the queue using blk_init_queue().
-	 * The type is decided at the first table loading time.
-	 * To prevent problematic device stacking, clear the queue flag
-	 * for request stacking support until then.
-	 *
-	 * This queue is new, so no concurrency on the queue_flags.
-	 */
-	queue_flag_clear_unlocked(QUEUE_FLAG_STACKABLE, md->queue);
-	md->saved_make_request_fn = md->queue->make_request_fn;
-	md->queue->queuedata = md;
-	md->queue->backing_dev_info.congested_fn = dm_any_congested;
-	md->queue->backing_dev_info.congested_data = md;
-	blk_queue_make_request(md->queue, dm_request);
-	blk_queue_bounce_limit(md->queue, BLK_BOUNCE_ANY);
-	md->queue->unplug_fn = dm_unplug_all;
-	blk_queue_merge_bvec(md->queue, dm_merge_bvec);
-	blk_queue_softirq_done(md->queue, dm_softirq_done);
-	blk_queue_prep_rq(md->queue, dm_prep_fn);
-	blk_queue_lld_busy(md->queue, dm_lld_busy);
-	blk_queue_ordered(md->queue, QUEUE_ORDERED_DRAIN_FLUSH);
+	dm_init_md_queue(md);
 
 	md->disk = alloc_disk(1);
 	if (!md->disk)
···
 	return 0;
 }
 
+/*
+ * Functions to manage md->type.
+ * All are required to hold md->type_lock.
+ */
+void dm_lock_md_type(struct mapped_device *md)
+{
+	mutex_lock(&md->type_lock);
+}
+
+void dm_unlock_md_type(struct mapped_device *md)
+{
+	mutex_unlock(&md->type_lock);
+}
+
+void dm_set_md_type(struct mapped_device *md, unsigned type)
+{
+	md->type = type;
+}
+
+unsigned dm_get_md_type(struct mapped_device *md)
+{
+	return md->type;
+}
+
+/*
+ * Fully initialize a request-based queue (->elevator, ->request_fn, etc).
+ */
+static int dm_init_request_based_queue(struct mapped_device *md)
+{
+	struct request_queue *q = NULL;
+
+	if (md->queue->elevator)
+		return 1;
+
+	/* Fully initialize the queue */
+	q = blk_init_allocated_queue(md->queue, dm_request_fn, NULL);
+	if (!q)
+		return 0;
+
+	md->queue = q;
+	md->saved_make_request_fn = md->queue->make_request_fn;
+	dm_init_md_queue(md);
+	blk_queue_softirq_done(md->queue, dm_softirq_done);
+	blk_queue_prep_rq(md->queue, dm_prep_fn);
+	blk_queue_lld_busy(md->queue, dm_lld_busy);
+	blk_queue_ordered(md->queue, QUEUE_ORDERED_DRAIN_FLUSH);
+
+	elv_register_queue(md->queue);
+
+	return 1;
+}
+
+/*
+ * Setup the DM device's queue based on md's type
+ */
+int dm_setup_md_queue(struct mapped_device *md)
+{
+	if ((dm_get_md_type(md) == DM_TYPE_REQUEST_BASED) &&
+	    !dm_init_request_based_queue(md)) {
+		DMWARN("Cannot initialize queue for request-based mapped device");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static struct mapped_device *dm_find_md(dev_t dev)
 {
 	struct mapped_device *md;
···
 	md = idr_find(&_minor_idr, minor);
 	if (md && (md == MINOR_ALLOCED ||
 		   (MINOR(disk_devt(dm_disk(md))) != minor) ||
+		   dm_deleting_md(md) ||
 		   test_bit(DMF_FREEING, &md->flags))) {
 		md = NULL;
 		goto out;
···
 void dm_get(struct mapped_device *md)
 {
 	atomic_inc(&md->holders);
+	BUG_ON(test_bit(DMF_FREEING, &md->flags));
 }
 
 const char *dm_device_name(struct mapped_device *md)
···
 }
 EXPORT_SYMBOL_GPL(dm_device_name);
 
-void dm_put(struct mapped_device *md)
+static void __dm_destroy(struct mapped_device *md, bool wait)
 {
 	struct dm_table *map;
 
-	BUG_ON(test_bit(DMF_FREEING, &md->flags));
+	might_sleep();
 
-	if (atomic_dec_and_lock(&md->holders, &_minor_lock)) {
-		map = dm_get_live_table(md);
-		idr_replace(&_minor_idr, MINOR_ALLOCED,
-			    MINOR(disk_devt(dm_disk(md))));
-		set_bit(DMF_FREEING, &md->flags);
-		spin_unlock(&_minor_lock);
-		if (!dm_suspended_md(md)) {
-			dm_table_presuspend_targets(map);
-			dm_table_postsuspend_targets(map);
-		}
-		dm_sysfs_exit(md);
-		dm_table_put(map);
-		dm_table_destroy(__unbind(md));
-		free_dev(md);
+	spin_lock(&_minor_lock);
+	map = dm_get_live_table(md);
+	idr_replace(&_minor_idr, MINOR_ALLOCED, MINOR(disk_devt(dm_disk(md))));
+	set_bit(DMF_FREEING, &md->flags);
+	spin_unlock(&_minor_lock);
+
+	if (!dm_suspended_md(md)) {
+		dm_table_presuspend_targets(map);
+		dm_table_postsuspend_targets(map);
 	}
+
+	/*
+	 * Rare, but there may be I/O requests still going to complete,
+	 * for example.  Wait for all references to disappear.
+	 * No one should increment the reference count of the mapped_device,
+	 * after the mapped_device state becomes DMF_FREEING.
+	 */
+	if (wait)
+		while (atomic_read(&md->holders))
+			msleep(1);
+	else if (atomic_read(&md->holders))
+		DMWARN("%s: Forcibly removing mapped_device still in use! (%d users)",
+		       dm_device_name(md), atomic_read(&md->holders));
+
+	dm_sysfs_exit(md);
+	dm_table_put(map);
+	dm_table_destroy(__unbind(md));
+	free_dev(md);
+}
+
+void dm_destroy(struct mapped_device *md)
+{
+	__dm_destroy(md, true);
+}
+
+void dm_destroy_immediate(struct mapped_device *md)
+{
+	__dm_destroy(md, false);
+}
+
+void dm_put(struct mapped_device *md)
+{
+	atomic_dec(&md->holders);
 }
 EXPORT_SYMBOL_GPL(dm_put);
 
···
 
 	if (!bio_empty_barrier(bio)) {
 		__split_and_process_bio(md, bio);
-		dm_flush(md);
+		/*
+		 * If the request isn't supported, don't waste time with
+		 * the second flush.
+		 */
+		if (md->barrier_error != -EOPNOTSUPP)
+			dm_flush(md);
 	}
 
 	if (md->barrier_error != DM_ENDIO_REQUEUE)
···
 	queue_work(md->wq, &md->work);
 }
 
-static void dm_rq_set_flush_nr(struct request *clone, unsigned flush_nr)
+static void dm_rq_set_target_request_nr(struct request *clone, unsigned request_nr)
 {
 	struct dm_rq_target_io *tio = clone->end_io_data;
 
-	tio->info.flush_request = flush_nr;
+	tio->info.target_request_nr = request_nr;
 }
 
 /* Issue barrier requests to targets and wait for their completion. */
···
 		ti = dm_table_get_target(map, i);
 		for (j = 0; j < ti->num_flush_requests; j++) {
 			clone = clone_rq(md->flush_request, md, GFP_NOIO);
-			dm_rq_set_flush_nr(clone, j);
+			dm_rq_set_target_request_nr(clone, j);
 			atomic_inc(&md->pending[rq_data_dir(clone)]);
 			map_request(ti, clone, md);
 		}
···
 	r = dm_calculate_queue_limits(table, &limits);
 	if (r) {
 		map = ERR_PTR(r);
-		goto out;
-	}
-
-	/* cannot change the device type, once a table is bound */
-	if (md->map &&
-	    (dm_table_get_type(md->map) != dm_table_get_type(table))) {
-		DMWARN("can't change the device type after a table is bound");
 		goto out;
 	}
 

drivers/md/dm.h

···
 int dm_table_resume_targets(struct dm_table *t);
 int dm_table_any_congested(struct dm_table *t, int bdi_bits);
 int dm_table_any_busy_target(struct dm_table *t);
-int dm_table_set_type(struct dm_table *t);
 unsigned dm_table_get_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);
+
+void dm_lock_md_type(struct mapped_device *md);
+void dm_unlock_md_type(struct mapped_device *md);
+void dm_set_md_type(struct mapped_device *md, unsigned type);
+unsigned dm_get_md_type(struct mapped_device *md);
+
+int dm_setup_md_queue(struct mapped_device *md);
 
 /*
  * To check the return value from dm_table_find_target().
···
 int dm_stripe_init(void);
 void dm_stripe_exit(void);
 
+/*
+ * mapped_device operations
+ */
+void dm_destroy(struct mapped_device *md);
+void dm_destroy_immediate(struct mapped_device *md);
 int dm_open_count(struct mapped_device *md);
 int dm_lock_for_deletion(struct mapped_device *md);
 

include/linux/device-mapper.h

···
 union map_info {
 	void *ptr;
 	unsigned long long ll;
-	unsigned flush_request;
+	unsigned target_request_nr;
 };
 
 /*
···
 	 * A number of zero-length barrier requests that will be submitted
 	 * to the target for the purpose of flushing cache.
 	 *
-	 * The request number will be placed in union map_info->flush_request.
+	 * The request number will be placed in union map_info->target_request_nr.
 	 * It is a responsibility of the target driver to remap these requests
 	 * to the real underlying devices.
 	 */
 	unsigned num_flush_requests;
+
+	/*
+	 * The number of discard requests that will be submitted to the
+	 * target.  map_info->request_nr is used just like num_flush_requests.
+	 */
+	unsigned num_discard_requests;
 
 	/* target specific data */
 	void *private;
···
 
 #define dm_array_too_big(fixed, obj, num) \
 	((num) > (UINT_MAX - (fixed)) / (obj))
+
+/*
+ * Sector offset taken relative to the start of the target instead of
+ * relative to the start of the device.
+ */
+#define dm_target_offset(ti, sector) ((sector) - (ti)->begin)
 
 static inline sector_t to_sector(unsigned long n)
 {

include/linux/dm-ioctl.h

···
 #include <linux/types.h>
 
 #define DM_DIR "mapper"		/* Slashes not supported */
+#define DM_CONTROL_NODE "control"
 #define DM_MAX_TYPE_NAME 16
 #define DM_NAME_LEN 128
 #define DM_UUID_LEN 129
···
 #define DM_DEV_SET_GEOMETRY	_IOWR(DM_IOCTL, DM_DEV_SET_GEOMETRY_CMD, struct dm_ioctl)
 
 #define DM_VERSION_MAJOR	4
-#define DM_VERSION_MINOR	17
+#define DM_VERSION_MINOR	18
 #define DM_VERSION_PATCHLEVEL	0
-#define DM_VERSION_EXTRA	"-ioctl (2010-03-05)"
+#define DM_VERSION_EXTRA	"-ioctl (2010-06-29)"
 
 /* Status bits */
 #define DM_READONLY_FLAG	(1 << 0) /* In/Out */

include/linux/miscdevice.h

···
 #define KVM_MINOR		232
 #define BTRFS_MINOR		234
 #define AUTOFS_MINOR		235
+#define MAPPER_CTRL_MINOR	236
 #define MISC_DYNAMIC_MINOR	255
 
 struct device;