This is a note to let you know that I've just added the patch titled dm cache: submit writethrough writes in parallel to origin and cache to the 4.14-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: dm-cache-submit-writethrough-writes-in-parallel-to-origin-and-cache.patch and it can be found in the queue-4.14 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From 2df3bae9a6543e90042291707b8db0cbfbae9ee9 Mon Sep 17 00:00:00 2001 From: Mike Snitzer <snitzer@xxxxxxxxxx> Date: Thu, 19 Oct 2017 17:16:54 -0400 Subject: dm cache: submit writethrough writes in parallel to origin and cache From: Mike Snitzer <snitzer@xxxxxxxxxx> commit 2df3bae9a6543e90042291707b8db0cbfbae9ee9 upstream. Discontinue issuing writethrough write IO in series to the origin and then cache. Use bio_clone_fast() to create a new origin clone bio that will be mapped to the origin device and then bio_chain() it to the bio that gets remapped to the cache device. The origin clone bio does _not_ have a copy of the per_bio_data -- as such check_if_tick_bio_needed() will not be called. The cache bio (parent bio) will not complete until the origin bio has completed -- this fulfills bio_clone_fast()'s requirements as well as the requirement to not complete the original IO until the write IO has completed to both the origin and cache device. Signed-off-by: Mike Snitzer <snitzer@xxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/md/dm-cache-target.c | 54 +++++++++++++++++++++++++++++-------------- 1 file changed, 37 insertions(+), 17 deletions(-) --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -450,6 +450,7 @@ struct cache { struct work_struct migration_worker; struct delayed_work waker; struct dm_bio_prison_v2 *prison; + struct bio_set *bs; mempool_t *migration_pool; @@ -868,16 +869,23 @@ static void check_if_tick_bio_needed(str 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) { @@ -971,23 +979,25 @@ static void writethrough_endio(struct bi } /* - * 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); + + BUG_ON(!origin_bio); - pb->cache = cache; - pb->cblock = cblock; - dm_hook_bio(&pb->hook_info, bio, writethrough_endio, NULL); - dm_bio_record(&pb->bio_details, bio); + 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_origin_clear_discard(pb->cache, bio, oblock); + remap_to_cache(cache, bio, cblock); } /*---------------------------------------------------------------- @@ -1873,7 +1883,7 @@ static int map_bio(struct cache *cache, } else { 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); @@ -2132,6 +2142,9 @@ 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 +2602,13 @@ static int cache_create(struct cache_arg 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); Patches currently in stable-queue which might be from snitzer@xxxxxxxxxx are queue-4.14/dm-cache-submit-writethrough-writes-in-parallel-to-origin-and-cache.patch queue-4.14/dm-cache-pass-cache-structure-to-mode-functions.patch queue-4.14/dm-rq-don-t-call-blk_mq_queue_stopped-in-dm_stop_que.patch queue-4.14/dm-cache-remove-all-obsolete-writethrough-specific-code.patch