From: Jan Kara jack@suse.cz

mainline inclusion from mainline-v5.12-rc2 commit b6e68ee82585f2ee890b0a897a6aacbf49a467bb category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I7QFJD?from=project-issue CVE: NA

Upstream commit 5d05426e2d5fd7df8af fixes a bug introduced by this patch, but it does not need to be backported.

This patch was merged in mainline-v5.12-rc1-dontuse at first, thus I denote it comes from mainline-v5.12-rc2.

--------------------------------

Currently when non-mq aware IO scheduler (BFQ, mq-deadline) is used for a queue with multiple HW queues, the performance it rather bad. The problem is that these IO schedulers use queue-wide locking and their dispatch function does not respect the hctx it is passed in and returns any request it finds appropriate. Thus locality of request access is broken and dispatch from multiple CPUs just contends on IO scheduler locks. For these IO schedulers there's little point in dispatching from multiple CPUs. Instead dispatch always only from a single CPU to limit contention.

Below is a comparison of dbench runs on XFS filesystem where the storage is a raid card with 64 HW queues and to it attached a single rotating disk. BFQ is used as IO scheduler:

clients MQ SQ MQ-Patched Amean 1 39.12 (0.00%) 43.29 * -10.67%* 36.09 * 7.74%* Amean 2 128.58 (0.00%) 101.30 * 21.22%* 96.14 * 25.23%* Amean 4 577.42 (0.00%) 494.47 * 14.37%* 508.49 * 11.94%* Amean 8 610.95 (0.00%) 363.86 * 40.44%* 362.12 * 40.73%* Amean 16 391.78 (0.00%) 261.49 * 33.25%* 282.94 * 27.78%* Amean 32 324.64 (0.00%) 267.71 * 17.54%* 233.00 * 28.23%* Amean 64 295.04 (0.00%) 253.02 * 14.24%* 242.37 * 17.85%* Amean 512 10281.61 (0.00%) 10211.16 * 0.69%* 10447.53 * -1.61%*

Numbers are times so lower is better. MQ is stock 5.10-rc6 kernel. SQ is the same kernel with megaraid_sas.host_tagset_enable=0 so that the card advertises just a single HW queue. MQ-Patched is a kernel with this patch applied.

You can see multiple hardware queues heavily hurt performance in combination with BFQ. The patch restores the performance.

Signed-off-by: Jan Kara jack@suse.cz Reviewed-by: Ming Lei ming.lei@redhat.com Signed-off-by: Jens Axboe axboe@kernel.dk Signed-off-by: Lei Chen lei.chen@smartx.com --- block/blk-mq.c | 63 +++++++++++++++++++++++++++++++++++++--- block/kyber-iosched.c | 1 + include/linux/elevator.h | 4 +++ 3 files changed, 64 insertions(+), 4 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c index 96debbe63ad2..2b0f0a9f40d9 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1521,17 +1521,62 @@ bool blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async) return false; } EXPORT_SYMBOL(blk_mq_run_hw_queue); +/* + * Is the request queue handled by an IO scheduler that does not respect + * hardware queues when dispatching? + */ +static bool blk_mq_has_sqsched(struct request_queue *q) +{ + struct elevator_queue *e = q->elevator;

-void blk_mq_run_hw_queues(struct request_queue *q, bool async) + if (e && e->type->ops.mq.dispatch_request && + !(e->type->elevator_features & ELEVATOR_F_MQ_AWARE)) + return true; + return false; +} + +/* + * Return prefered queue to dispatch from (if any) for non-mq aware IO + * scheduler. + */ +static struct blk_mq_hw_ctx *blk_mq_get_sq_hctx(struct request_queue *q) { struct blk_mq_hw_ctx *hctx; + + /* + * If the IO scheduler does not respect hardware queues when + * dispatching, we just don't bother with multiple HW queues and + * dispatch from hctx for the current CPU since running multiple queues + * just causes lock contention inside the scheduler and pointless cache + * bouncing. + */ + hctx = blk_mq_map_queue(q, raw_smp_processor_id()); + if (!blk_mq_hctx_stopped(hctx)) + return hctx; + return NULL; +} + + +void blk_mq_run_hw_queues(struct request_queue *q, bool async) +{ + struct blk_mq_hw_ctx *hctx, *sq_hctx; int i;

+ sq_hctx = NULL; + if (blk_mq_has_sqsched(q)) + sq_hctx = blk_mq_get_sq_hctx(q); queue_for_each_hw_ctx(q, hctx, i) { if (blk_mq_hctx_stopped(hctx)) continue;

- blk_mq_run_hw_queue(hctx, async); + /* + * Dispatch from this hctx either if there's no hctx preferred + * by IO scheduler or if it has requests that bypass the + * scheduler. + */ + if (!sq_hctx || sq_hctx == hctx || + !list_empty_careful(&hctx->dispatch)) + blk_mq_run_hw_queue(hctx, async); } } EXPORT_SYMBOL(blk_mq_run_hw_queues); @@ -1543,9 +1588,12 @@ EXPORT_SYMBOL(blk_mq_run_hw_queues); */ void blk_mq_delay_run_hw_queues(struct request_queue *q, unsigned long msecs) { - struct blk_mq_hw_ctx *hctx; + struct blk_mq_hw_ctx *hctx, *sq_hctx; int i;

+ sq_hctx = NULL; + if (blk_mq_has_sqsched(q)) + sq_hctx = blk_mq_get_sq_hctx(q); queue_for_each_hw_ctx(q, hctx, i) { if (blk_mq_hctx_stopped(hctx)) continue; @@ -1559,7 +1607,14 @@ void blk_mq_delay_run_hw_queues(struct request_queue *q, unsigned long msecs) if (delayed_work_pending(&hctx->run_work)) continue;

- blk_mq_delay_run_hw_queue(hctx, msecs); + /* + * Dispatch from this hctx either if there's no hctx preferred + * by IO scheduler or if it has requests that bypass the + * scheduler. + */ + if (!sq_hctx || sq_hctx == hctx || + !list_empty_careful(&hctx->dispatch)) + blk_mq_delay_run_hw_queue(hctx, msecs); } } EXPORT_SYMBOL(blk_mq_delay_run_hw_queues); diff --git a/block/kyber-iosched.c b/block/kyber-iosched.c index d8f3cb1bffa6..bc0376c338ff 100644 --- a/block/kyber-iosched.c +++ b/block/kyber-iosched.c @@ -972,6 +972,7 @@ static struct elevator_type kyber_sched = { #endif .elevator_attrs = kyber_sched_attrs, .elevator_name = "kyber", + .elevator_features = ELEVATOR_F_MQ_AWARE, .elevator_owner = THIS_MODULE, };

diff --git a/include/linux/elevator.h b/include/linux/elevator.h index 663ce1780c5d..70c3c1fd5cf0 100644 --- a/include/linux/elevator.h +++ b/include/linux/elevator.h @@ -155,6 +155,7 @@ struct elevator_type size_t icq_align; /* ditto */ struct elv_fs_entry *elevator_attrs; char elevator_name[ELV_NAME_MAX]; + unsigned int elevator_features; const char *elevator_alias; struct module *elevator_owner; bool uses_mq; @@ -276,6 +277,9 @@ enum { #define rq_entry_fifo(ptr) list_entry((ptr), struct request, queuelist) #define rq_fifo_clear(rq) list_del_init(&(rq)->queuelist)

+/* Supports scheduling on multiple hardware queues */ +#define ELEVATOR_F_MQ_AWARE (1U << 1) + #else /* CONFIG_BLOCK */

static inline void load_default_elevator_module(void) { }