1 files changed, 53 insertions, 113 deletions

diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c
index 69cdb29ef6be..280873b13e74 100644
--- a/drivers/md/dm-cache-target.c
+++ b/drivers/md/dm-cache-target.c
@@ -410,7 +410,6 @@ struct cache {
 	spinlock_t lock;
 	struct list_head deferred_cells;
 	struct bio_list deferred_bios;
-	struct bio_list deferred_writethrough_bios;
 	sector_t migration_threshold;
 	wait_queue_head_t migration_wait;
 	atomic_t nr_allocated_migrations;
@@ -446,10 +445,10 @@ struct cache {
 	struct dm_kcopyd_client *copier;
 	struct workqueue_struct *wq;
 	struct work_struct deferred_bio_worker;
-	struct work_struct deferred_writethrough_worker;
 	struct work_struct migration_worker;
 	struct delayed_work waker;
 	struct dm_bio_prison_v2 *prison;
+	struct bio_set *bs;
 
 	mempool_t *migration_pool;
 
@@ -490,15 +489,6 @@ struct per_bio_data {
 	struct dm_bio_prison_cell_v2 *cell;
 	struct dm_hook_info hook_info;
 	sector_t len;
-
-	/*
-	 * writethrough fields.  These MUST remain at the end of this
-	 * structure and the 'cache' member must be the first as it
-	 * is used to determine the offset of the writethrough fields.
-	 */
-	struct cache *cache;
-	dm_cblock_t cblock;
-	struct dm_bio_details bio_details;
 };
 
 struct dm_cache_migration {
@@ -515,19 +505,19 @@ struct dm_cache_migration {
 
 /*----------------------------------------------------------------*/
 
-static bool writethrough_mode(struct cache_features *f)
+static bool writethrough_mode(struct cache *cache)
 {
-	return f->io_mode == CM_IO_WRITETHROUGH;
+	return cache->features.io_mode == CM_IO_WRITETHROUGH;
 }
 
-static bool writeback_mode(struct cache_features *f)
+static bool writeback_mode(struct cache *cache)
 {
-	return f->io_mode == CM_IO_WRITEBACK;
+	return cache->features.io_mode == CM_IO_WRITEBACK;
 }
 
-static inline bool passthrough_mode(struct cache_features *f)
+static inline bool passthrough_mode(struct cache *cache)
 {
-	return unlikely(f->io_mode == CM_IO_PASSTHROUGH);
+	return unlikely(cache->features.io_mode == CM_IO_PASSTHROUGH);
 }
 
 /*----------------------------------------------------------------*/
@@ -537,14 +527,9 @@ static void wake_deferred_bio_worker(struct cache *cache)
 	queue_work(cache->wq, &cache->deferred_bio_worker);
 }
 
-static void wake_deferred_writethrough_worker(struct cache *cache)
-{
-	queue_work(cache->wq, &cache->deferred_writethrough_worker);
-}
-
 static void wake_migration_worker(struct cache *cache)
 {
-	if (passthrough_mode(&cache->features))
+	if (passthrough_mode(cache))
 		return;
 
 	queue_work(cache->wq, &cache->migration_worker);
@@ -618,15 +603,9 @@ static unsigned lock_level(struct bio *bio)
  * Per bio data
  *--------------------------------------------------------------*/
 
-/*
- * If using writeback, leave out struct per_bio_data's writethrough fields.
- */
-#define PB_DATA_SIZE_WB (offsetof(struct per_bio_data, cache))
-#define PB_DATA_SIZE_WT (sizeof(struct per_bio_data))
-
 static size_t get_per_bio_data_size(struct cache *cache)
 {
-	return writethrough_mode(&cache->features) ? PB_DATA_SIZE_WT : PB_DATA_SIZE_WB;
+	return sizeof(struct per_bio_data);
 }
 
 static struct per_bio_data *get_per_bio_data(struct bio *bio, size_t data_size)
@@ -868,16 +847,23 @@ static void check_if_tick_bio_needed(struct cache *cache, struct bio *bio)
 	spin_unlock_irqrestore(&cache->lock, flags);
 }
 
-static void remap_to_origin_clear_discard(struct cache *cache, struct bio *bio,
-					  dm_oblock_t oblock)
+static void __remap_to_origin_clear_discard(struct cache *cache, struct bio *bio,
+					    dm_oblock_t oblock, bool bio_has_pbd)
 {
-	// FIXME: this is called way too much.
-	check_if_tick_bio_needed(cache, bio);
+	if (bio_has_pbd)
+		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));
 }
 
+static void remap_to_origin_clear_discard(struct cache *cache, struct bio *bio,
+					  dm_oblock_t oblock)
+{
+	// FIXME: check_if_tick_bio_needed() is called way too much through this interface
+	__remap_to_origin_clear_discard(cache, bio, oblock, true);
+}
+
 static void remap_to_cache_dirty(struct cache *cache, struct bio *bio,
 				 dm_oblock_t oblock, dm_cblock_t cblock)
 {
@@ -937,57 +923,26 @@ static void issue_op(struct bio *bio, void *context)
 	accounted_request(cache, bio);
 }
 
-static void defer_writethrough_bio(struct cache *cache, struct bio *bio)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&cache->lock, flags);
-	bio_list_add(&cache->deferred_writethrough_bios, bio);
-	spin_unlock_irqrestore(&cache->lock, flags);
-
-	wake_deferred_writethrough_worker(cache);
-}
-
-static void writethrough_endio(struct bio *bio)
-{
-	struct per_bio_data *pb = get_per_bio_data(bio, PB_DATA_SIZE_WT);
-
-	dm_unhook_bio(&pb->hook_info, bio);
-
-	if (bio->bi_status) {
-		bio_endio(bio);
-		return;
-	}
-
-	dm_bio_restore(&pb->bio_details, bio);
-	remap_to_cache(pb->cache, bio, pb->cblock);
-
-	/*
-	 * We can't issue this bio directly, since we're in interrupt
-	 * context.  So it gets put on a bio list for processing by the
-	 * worker thread.
-	 */
-	defer_writethrough_bio(pb->cache, bio);
-}
-
 /*
- * FIXME: send in parallel, huge latency as is.
  * When running in writethrough mode we need to send writes to clean blocks
- * to both the cache and origin devices.  In future we'd like to clone the
- * bio and send them in parallel, but for now we're doing them in
- * series as this is easier.
+ * to both the cache and origin devices.  Clone the bio and send them in parallel.
  */
-static void remap_to_origin_then_cache(struct cache *cache, struct bio *bio,
-				       dm_oblock_t oblock, dm_cblock_t cblock)
+static void remap_to_origin_and_cache(struct cache *cache, struct bio *bio,
+				      dm_oblock_t oblock, dm_cblock_t cblock)
 {
-	struct per_bio_data *pb = get_per_bio_data(bio, PB_DATA_SIZE_WT);
+	struct bio *origin_bio = bio_clone_fast(bio, GFP_NOIO, cache->bs);
 
-	pb->cache = cache;
-	pb->cblock = cblock;
-	dm_hook_bio(&pb->hook_info, bio, writethrough_endio, NULL);
-	dm_bio_record(&pb->bio_details, bio);
+	BUG_ON(!origin_bio);
 
-	remap_to_origin_clear_discard(pb->cache, bio, oblock);
+	bio_chain(origin_bio, bio);
+	/*
+	 * Passing false to __remap_to_origin_clear_discard() skips
+	 * all code that might use per_bio_data (since clone doesn't have it)
+	 */
+	__remap_to_origin_clear_discard(cache, origin_bio, oblock, false);
+	submit_bio(origin_bio);
+
+	remap_to_cache(cache, bio, cblock);
 }
 
 /*----------------------------------------------------------------
@@ -1209,7 +1164,7 @@ static bool bio_writes_complete_block(struct cache *cache, struct bio *bio)
 
 static bool optimisable_bio(struct cache *cache, struct bio *bio, dm_oblock_t block)
 {
-	return writeback_mode(&cache->features) &&
+	return writeback_mode(cache) &&
 		(is_discarded_oblock(cache, block) || bio_writes_complete_block(cache, bio));
 }
 
@@ -1862,7 +1817,7 @@ static int map_bio(struct cache *cache, struct bio *bio, dm_oblock_t block,
 		 * Passthrough always maps to the origin, invalidating any
 		 * cache blocks that are written to.
 		 */
-		if (passthrough_mode(&cache->features)) {
+		if (passthrough_mode(cache)) {
 			if (bio_data_dir(bio) == WRITE) {
 				bio_drop_shared_lock(cache, bio);
 				atomic_inc(&cache->stats.demotion);
@@ -1871,9 +1826,9 @@ static int map_bio(struct cache *cache, struct bio *bio, dm_oblock_t block,
 				remap_to_origin_clear_discard(cache, bio, block);
 
 		} else {
-			if (bio_data_dir(bio) == WRITE && writethrough_mode(&cache->features) &&
+			if (bio_data_dir(bio) == WRITE && writethrough_mode(cache) &&
 			    !is_dirty(cache, cblock)) {
-				remap_to_origin_then_cache(cache, bio, block, cblock);
+				remap_to_origin_and_cache(cache, bio, block, cblock);
 				accounted_begin(cache, bio);
 			} else
 				remap_to_cache_dirty(cache, bio, block, cblock);
@@ -2003,28 +1958,6 @@ static void process_deferred_bios(struct work_struct *ws)
 		schedule_commit(&cache->committer);
 }
 
-static void process_deferred_writethrough_bios(struct work_struct *ws)
-{
-	struct cache *cache = container_of(ws, struct cache, deferred_writethrough_worker);
-
-	unsigned long flags;
-	struct bio_list bios;
-	struct bio *bio;
-
-	bio_list_init(&bios);
-
-	spin_lock_irqsave(&cache->lock, flags);
-	bio_list_merge(&bios, &cache->deferred_writethrough_bios);
-	bio_list_init(&cache->deferred_writethrough_bios);
-	spin_unlock_irqrestore(&cache->lock, flags);
-
-	/*
-	 * These bios have already been through accounted_begin()
-	 */
-	while ((bio = bio_list_pop(&bios)))
-		generic_make_request(bio);
-}
-
 /*----------------------------------------------------------------
  * Main worker loop
  *--------------------------------------------------------------*/
@@ -2132,6 +2065,9 @@ static void destroy(struct cache *cache)
 		kfree(cache->ctr_args[i]);
 	kfree(cache->ctr_args);
 
+	if (cache->bs)
+		bioset_free(cache->bs);
+
 	kfree(cache);
 }
 
@@ -2589,6 +2525,13 @@ static int cache_create(struct cache_args *ca, struct cache **result)
 	cache->features = ca->features;
 	ti->per_io_data_size = get_per_bio_data_size(cache);
 
+	if (writethrough_mode(cache)) {
+		/* Create bioset for writethrough bios issued to origin */
+		cache->bs = bioset_create(BIO_POOL_SIZE, 0, 0);
+		if (!cache->bs)
+			goto bad;
+	}
+
 	cache->callbacks.congested_fn = cache_is_congested;
 	dm_table_add_target_callbacks(ti->table, &cache->callbacks);
 
@@ -2649,7 +2592,7 @@ static int cache_create(struct cache_args *ca, struct cache **result)
 		goto bad;
 	}
 
-	if (passthrough_mode(&cache->features)) {
+	if (passthrough_mode(cache)) {
 		bool all_clean;
 
 		r = dm_cache_metadata_all_clean(cache->cmd, &all_clean);
@@ -2670,7 +2613,6 @@ static int cache_create(struct cache_args *ca, struct cache **result)
 	spin_lock_init(&cache->lock);
 	INIT_LIST_HEAD(&cache->deferred_cells);
 	bio_list_init(&cache->deferred_bios);
-	bio_list_init(&cache->deferred_writethrough_bios);
 	atomic_set(&cache->nr_allocated_migrations, 0);
 	atomic_set(&cache->nr_io_migrations, 0);
 	init_waitqueue_head(&cache->migration_wait);
@@ -2709,8 +2651,6 @@ static int cache_create(struct cache_args *ca, struct cache **result)
 		goto bad;
 	}
 	INIT_WORK(&cache->deferred_bio_worker, process_deferred_bios);
-	INIT_WORK(&cache->deferred_writethrough_worker,
-		  process_deferred_writethrough_bios);
 	INIT_WORK(&cache->migration_worker, check_migrations);
 	INIT_DELAYED_WORK(&cache->waker, do_waker);
 
@@ -3279,13 +3219,13 @@ static void cache_status(struct dm_target *ti, status_type_t type,
 		else
 			DMEMIT("1 ");
 
-		if (writethrough_mode(&cache->features))
+		if (writethrough_mode(cache))
 			DMEMIT("writethrough ");
 
-		else if (passthrough_mode(&cache->features))
+		else if (passthrough_mode(cache))
 			DMEMIT("passthrough ");
 
-		else if (writeback_mode(&cache->features))
+		else if (writeback_mode(cache))
 			DMEMIT("writeback ");
 
 		else {
@@ -3451,7 +3391,7 @@ static int process_invalidate_cblocks_message(struct cache *cache, unsigned coun
 	unsigned i;
 	struct cblock_range range;
 
-	if (!passthrough_mode(&cache->features)) {
+	if (!passthrough_mode(cache)) {
 		DMERR("%s: cache has to be in passthrough mode for invalidation",
 		      cache_device_name(cache));
 		return -EPERM;