Hi, This patchset introduces flag of BLK_MQ_F_SCHED_USE_HW_TAG and allows to use hardware tag directly for IO scheduling if the queue's depth is big enough. In this way, we can avoid to allocate extra tags and request pool for IO schedule, and the schedule tag allocation/release can be saved in I/O submit path. V3: - respect q->nr_requests by resizing hw tags, as suggested by Omar V2: - fix oops when kyber is used - move dumping the new flag into patch 1 - support to use hw tag for shared tags - update hctx->sched_tag when BLK_MQ_F_SCHED_USE_HW_TAG is changed - clear the flag in patch of blk_mq_exit_sched() - don't update q->nr_requests when updating hw queue's depth - fix blk_mq_get_queue_depth() Ming Lei (4): blk-mq: introduce BLK_MQ_F_SCHED_USE_HW_TAG blk-mq: introduce blk_mq_get_queue_depth() blk-mq: use hw tag for scheduling if hw tag space is big enough blk-mq: allow to use hw tag for shared tags block/blk-mq-debugfs.c | 1 + block/blk-mq-sched.c | 90 ++++++++++++++++++++++++++++++++++++++++---- block/blk-mq-sched.h | 9 +++++ block/blk-mq.c | 100 +++++++++++++++++++++++++++++++++++++++++++++---- block/blk-mq.h | 24 ++++++++++++ block/kyber-iosched.c | 7 +++- include/linux/blk-mq.h | 1 + include/linux/blkdev.h | 8 ++++ 8 files changed, 223 insertions(+), 17 deletions(-) -- 2.9.3
