Merge tag 'dm-3.14-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

Pull device mapper fixes from Mike Snitzer:
 "A few dm-cache fixes, an invalid ioctl handling fix for dm multipath,
  a couple immutable biovec fixups for dm mirror, and a few dm-thin
  fixes.

  There will likely be additional dm-thin metadata and data resize fixes
  to include in 3.14-rc6 next week.

  Note to stable-minded folks: Immutable biovecs were introduced in
  3.14, so the related fixups for dm mirror are not needed in stable@
  kernels"

* tag 'dm-3.14-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm cache: fix truncation bug when mapping I/O to >2TB fast device
  dm thin: allow metadata space larger than supported to go unused
  dm mpath: fix stalls when handling invalid ioctls
  dm thin: fix the error path for the thin device constructor
  dm raid1: fix immutable biovec related BUG when retrying read bio
  dm io: fix I/O to multiple destinations
  dm thin: avoid metadata commit if a pool's thin devices haven't changed
  dm cache: do not add migration to completed list before unhooking bio
  dm cache: move hook_info into common portion of per_bio_data structure

drivers/md/dm-cache-target.c

@@ -289,6 +289,7 @@ struct per_bio_data {
 	bool tick:1;
 	unsigned req_nr:2;
 	struct dm_deferred_entry *all_io_entry;
+	struct dm_hook_info hook_info;
 
 	/*
 	 * writethrough fields.  These MUST remain at the end of this
@@ -297,7 +298,6 @@ struct per_bio_data {
 	 */
 	struct cache *cache;
 	dm_cblock_t cblock;
-	struct dm_hook_info hook_info;
 	struct dm_bio_details bio_details;
 };
 
@@ -671,15 +671,16 @@ static void remap_to_cache(struct cache *cache, struct bio *bio,
 			   dm_cblock_t cblock)
 {
 	sector_t bi_sector = bio->bi_iter.bi_sector;
+	sector_t block = from_cblock(cblock);
 
 	bio->bi_bdev = cache->cache_dev->bdev;
 	if (!block_size_is_power_of_two(cache))
 		bio->bi_iter.bi_sector =
-			(from_cblock(cblock) * cache->sectors_per_block) +
+			(block * cache->sectors_per_block) +
 			sector_div(bi_sector, cache->sectors_per_block);
 	else
 		bio->bi_iter.bi_sector =
-			(from_cblock(cblock) << cache->sectors_per_block_shift) |
+			(block << cache->sectors_per_block_shift) |
 			(bi_sector & (cache->sectors_per_block - 1));
 }
 
@@ -1010,13 +1011,15 @@ static void overwrite_endio(struct bio *bio, int err)
 	struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size);
 	unsigned long flags;
 
+	dm_unhook_bio(&pb->hook_info, bio);
+
 	if (err)
 		mg->err = true;
 
+	mg->requeue_holder = false;
+
 	spin_lock_irqsave(&cache->lock, flags);
 	list_add_tail(&mg->list, &cache->completed_migrations);
-	dm_unhook_bio(&pb->hook_info, bio);
-	mg->requeue_holder = false;
 	spin_unlock_irqrestore(&cache->lock, flags);
 
 	wake_worker(cache);

drivers/md/dm-io.c

@@ -201,29 +201,28 @@ static void list_dp_init(struct dpages *dp, struct page_list *pl, unsigned offse
 /*
  * Functions for getting the pages from a bvec.
  */
-static void bio_get_page(struct dpages *dp,
-		  struct page **p, unsigned long *len, unsigned *offset)
+static void bio_get_page(struct dpages *dp, struct page **p,
+			 unsigned long *len, unsigned *offset)
 {
-	struct bio *bio = dp->context_ptr;
-	struct bio_vec bvec = bio_iovec(bio);
-	*p = bvec.bv_page;
-	*len = bvec.bv_len;
-	*offset = bvec.bv_offset;
+	struct bio_vec *bvec = dp->context_ptr;
+	*p = bvec->bv_page;
+	*len = bvec->bv_len - dp->context_u;
+	*offset = bvec->bv_offset + dp->context_u;
 }
 
 static void bio_next_page(struct dpages *dp)
 {
-	struct bio *bio = dp->context_ptr;
-	struct bio_vec bvec = bio_iovec(bio);
-
-	bio_advance(bio, bvec.bv_len);
+	struct bio_vec *bvec = dp->context_ptr;
+	dp->context_ptr = bvec + 1;
+	dp->context_u = 0;
 }
 
 static void bio_dp_init(struct dpages *dp, struct bio *bio)
 {
 	dp->get_page = bio_get_page;
 	dp->next_page = bio_next_page;
-	dp->context_ptr = bio;
+	dp->context_ptr = __bvec_iter_bvec(bio->bi_io_vec, bio->bi_iter);
+	dp->context_u = bio->bi_iter.bi_bvec_done;
 }
 
 /*

drivers/md/dm-mpath.c

@@ -1626,8 +1626,11 @@ static int multipath_ioctl(struct dm_target *ti, unsigned int cmd,
 	/*
 	 * Only pass ioctls through if the device sizes match exactly.
 	 */
-	if (!r && ti->len != i_size_read(bdev->bd_inode) >> SECTOR_SHIFT)
-		r = scsi_verify_blk_ioctl(NULL, cmd);
+	if (!bdev || ti->len != i_size_read(bdev->bd_inode) >> SECTOR_SHIFT) {
+		int err = scsi_verify_blk_ioctl(NULL, cmd);
+		if (err)
+			r = err;
+	}
 
 	if (r == -ENOTCONN && !fatal_signal_pending(current))
 		queue_work(kmultipathd, &m->process_queued_ios);

drivers/md/dm-raid1.c

@@ -1244,6 +1244,9 @@ static int mirror_end_io(struct dm_target *ti, struct bio *bio, int error)
 
 			dm_bio_restore(bd, bio);
 			bio_record->details.bi_bdev = NULL;
+
+			atomic_inc(&bio->bi_remaining);
+
 			queue_bio(ms, bio, rw);
 			return DM_ENDIO_INCOMPLETE;
 		}

drivers/md/dm-thin-metadata.c

@@ -483,7 +483,7 @@ static int __write_initial_superblock(struct dm_pool_metadata *pmd)
 
 	disk_super->data_mapping_root = cpu_to_le64(pmd->root);
 	disk_super->device_details_root = cpu_to_le64(pmd->details_root);
-	disk_super->metadata_block_size = cpu_to_le32(THIN_METADATA_BLOCK_SIZE >> SECTOR_SHIFT);
+	disk_super->metadata_block_size = cpu_to_le32(THIN_METADATA_BLOCK_SIZE);
 	disk_super->metadata_nr_blocks = cpu_to_le64(bdev_size >> SECTOR_TO_BLOCK_SHIFT);
 	disk_super->data_block_size = cpu_to_le32(pmd->data_block_size);
 
@@ -651,7 +651,7 @@ static int __create_persistent_data_objects(struct dm_pool_metadata *pmd, bool f
 {
 	int r;
 
-	pmd->bm = dm_block_manager_create(pmd->bdev, THIN_METADATA_BLOCK_SIZE,
+	pmd->bm = dm_block_manager_create(pmd->bdev, THIN_METADATA_BLOCK_SIZE << SECTOR_SHIFT,
 					  THIN_METADATA_CACHE_SIZE,
 					  THIN_MAX_CONCURRENT_LOCKS);
 	if (IS_ERR(pmd->bm)) {
@@ -1489,6 +1489,23 @@ bool dm_thin_changed_this_transaction(struct dm_thin_device *td)
 	return r;
 }
 
+bool dm_pool_changed_this_transaction(struct dm_pool_metadata *pmd)
+{
+	bool r = false;
+	struct dm_thin_device *td, *tmp;
+
+	down_read(&pmd->root_lock);
+	list_for_each_entry_safe(td, tmp, &pmd->thin_devices, list) {
+		if (td->changed) {
+			r = td->changed;
+			break;
+		}
+	}
+	up_read(&pmd->root_lock);
+
+	return r;
+}
+
 bool dm_thin_aborted_changes(struct dm_thin_device *td)
 {
 	bool r;

drivers/md/dm-thin-metadata.h

@@ -9,16 +9,14 @@
 
 #include "persistent-data/dm-block-manager.h"
 #include "persistent-data/dm-space-map.h"
+#include "persistent-data/dm-space-map-metadata.h"
 
-#define THIN_METADATA_BLOCK_SIZE 4096
+#define THIN_METADATA_BLOCK_SIZE DM_SM_METADATA_BLOCK_SIZE
 
 /*
  * The metadata device is currently limited in size.
- *
- * We have one block of index, which can hold 255 index entries.  Each
- * index entry contains allocation info about 16k metadata blocks.
  */
-#define THIN_METADATA_MAX_SECTORS (255 * (1 << 14) * (THIN_METADATA_BLOCK_SIZE / (1 << SECTOR_SHIFT)))
+#define THIN_METADATA_MAX_SECTORS DM_SM_METADATA_MAX_SECTORS
 
 /*
  * A metadata device larger than 16GB triggers a warning.
@@ -161,6 +159,8 @@ int dm_thin_remove_block(struct dm_thin_device *td, dm_block_t block);
  */
 bool dm_thin_changed_this_transaction(struct dm_thin_device *td);
 
+bool dm_pool_changed_this_transaction(struct dm_pool_metadata *pmd);
+
 bool dm_thin_aborted_changes(struct dm_thin_device *td);
 
 int dm_thin_get_highest_mapped_block(struct dm_thin_device *td,

drivers/md/dm-thin.c

@@ -1357,7 +1357,8 @@ static void process_deferred_bios(struct pool *pool)
 	bio_list_init(&pool->deferred_flush_bios);
 	spin_unlock_irqrestore(&pool->lock, flags);
 
-	if (bio_list_empty(&bios) && !need_commit_due_to_time(pool))
+	if (bio_list_empty(&bios) &&
+	    !(dm_pool_changed_this_transaction(pool->pmd) && need_commit_due_to_time(pool)))
 		return;
 
 	if (commit(pool)) {
@@ -1999,16 +2000,27 @@ static void metadata_low_callback(void *context)
 	dm_table_event(pool->ti->table);
 }
 
-static sector_t get_metadata_dev_size(struct block_device *bdev)
+static sector_t get_dev_size(struct block_device *bdev)
+{
+	return i_size_read(bdev->bd_inode) >> SECTOR_SHIFT;
+}
+
+static void warn_if_metadata_device_too_big(struct block_device *bdev)
 {
-	sector_t metadata_dev_size = i_size_read(bdev->bd_inode) >> SECTOR_SHIFT;
+	sector_t metadata_dev_size = get_dev_size(bdev);
 	char buffer[BDEVNAME_SIZE];
 
-	if (metadata_dev_size > THIN_METADATA_MAX_SECTORS_WARNING) {
+	if (metadata_dev_size > THIN_METADATA_MAX_SECTORS_WARNING)
 		DMWARN("Metadata device %s is larger than %u sectors: excess space will not be used.",
 		       bdevname(bdev, buffer), THIN_METADATA_MAX_SECTORS);
-		metadata_dev_size = THIN_METADATA_MAX_SECTORS_WARNING;
-	}
+}
+
+static sector_t get_metadata_dev_size(struct block_device *bdev)
+{
+	sector_t metadata_dev_size = get_dev_size(bdev);
+
+	if (metadata_dev_size > THIN_METADATA_MAX_SECTORS)
+		metadata_dev_size = THIN_METADATA_MAX_SECTORS;
 
 	return metadata_dev_size;
 }
@@ -2017,7 +2029,7 @@ static dm_block_t get_metadata_dev_size_in_blocks(struct block_device *bdev)
 {
 	sector_t metadata_dev_size = get_metadata_dev_size(bdev);
 
-	sector_div(metadata_dev_size, THIN_METADATA_BLOCK_SIZE >> SECTOR_SHIFT);
+	sector_div(metadata_dev_size, THIN_METADATA_BLOCK_SIZE);
 
 	return metadata_dev_size;
 }
@@ -2095,12 +2107,7 @@ static int pool_ctr(struct dm_target *ti, unsigned argc, char **argv)
 		ti->error = "Error opening metadata block device";
 		goto out_unlock;
 	}
-
-	/*
-	 * Run for the side-effect of possibly issuing a warning if the
-	 * device is too big.
-	 */
-	(void) get_metadata_dev_size(metadata_dev->bdev);
+	warn_if_metadata_device_too_big(metadata_dev->bdev);
 
 	r = dm_get_device(ti, argv[1], FMODE_READ | FMODE_WRITE, &data_dev);
 	if (r) {
@@ -2287,6 +2294,7 @@ static int maybe_resize_metadata_dev(struct dm_target *ti, bool *need_commit)
 		return -EINVAL;
 
 	} else if (metadata_dev_size > sb_metadata_dev_size) {
+		warn_if_metadata_device_too_big(pool->md_dev);
 		DMINFO("%s: growing the metadata device from %llu to %llu blocks",
 		       dm_device_name(pool->pool_md),
 		       sb_metadata_dev_size, metadata_dev_size);
@@ -2894,6 +2902,7 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
 	if (get_pool_mode(tc->pool) == PM_FAIL) {
 		ti->error = "Couldn't open thin device, Pool is in fail mode";
+		r = -EINVAL;
 		goto bad_thin_open;
 	}
 
@@ -2905,7 +2914,7 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
 	r = dm_set_target_max_io_len(ti, tc->pool->sectors_per_block);
 	if (r)
-		goto bad_thin_open;
+		goto bad_target_max_io_len;
 
 	ti->num_flush_bios = 1;
 	ti->flush_supported = true;
@@ -2926,6 +2935,8 @@ static int thin_ctr(struct dm_target *ti, unsigned argc, char **argv)
 
 	return 0;
 
+bad_target_max_io_len:
+	dm_pool_close_thin_device(tc->td);
 bad_thin_open:
 	__pool_dec(tc->pool);
 bad_pool_lookup:

drivers/md/persistent-data/dm-space-map-metadata.c

@@ -680,6 +680,8 @@ int dm_sm_metadata_create(struct dm_space_map *sm,
 	if (r)
 		return r;
 
+	if (nr_blocks > DM_SM_METADATA_MAX_BLOCKS)
+		nr_blocks = DM_SM_METADATA_MAX_BLOCKS;
 	r = sm_ll_extend(&smm->ll, nr_blocks);
 	if (r)
 		return r;

drivers/md/persistent-data/dm-space-map-metadata.h

@@ -9,6 +9,17 @@
 
 #include "dm-transaction-manager.h"
 
+#define DM_SM_METADATA_BLOCK_SIZE (4096 >> SECTOR_SHIFT)
+
+/*
+ * The metadata device is currently limited in size.
+ *
+ * We have one block of index, which can hold 255 index entries.  Each
+ * index entry contains allocation info about ~16k metadata blocks.
+ */
+#define DM_SM_METADATA_MAX_BLOCKS (255 * ((1 << 14) - 64))
+#define DM_SM_METADATA_MAX_SECTORS (DM_SM_METADATA_MAX_BLOCKS * DM_SM_METADATA_BLOCK_SIZE)
+
 /*
  * Unfortunately we have to use two-phase construction due to the cycle
  * between the tm and sm.