On 6/9/21 1:07 AM, Bart Van Assche wrote:
> For interactive workloads it is important that synchronous requests are
> not delayed. Hence reserve 25% of scheduler tags for synchronous requests.
> This patch still allows asynchronous requests to fill the hardware queues
> since blk_mq_init_sched() makes sure that the number of scheduler requests
> is the double of the hardware queue depth. From blk_mq_init_sched():
>
> q->nr_requests = 2 * min_t(unsigned int, q->tag_set->queue_depth,
> BLKDEV_MAX_RQ);
>
> Cc: Damien Le Moal <damien.lemoal@xxxxxxx>
> Cc: Hannes Reinecke <hare@xxxxxxx>
> Cc: Christoph Hellwig <hch@xxxxxx>
> Cc: Ming Lei <ming.lei@xxxxxxxxxx>
> Cc: Johannes Thumshirn <johannes.thumshirn@xxxxxxx>
> Cc: Himanshu Madhani <himanshu.madhani@xxxxxxxxxx>
> Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
> ---
> block/mq-deadline.c | 52 +++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 52 insertions(+)
>
> diff --git a/block/mq-deadline.c b/block/mq-deadline.c
> index 1d1bb7a41d2a..a7d0584437d1 100644
> --- a/block/mq-deadline.c
> +++ b/block/mq-deadline.c
> @@ -67,6 +67,7 @@ struct deadline_data {
> int fifo_batch;
> int writes_starved;
> int front_merges;
> + u32 async_depth;
>
> spinlock_t lock;
> spinlock_t zone_lock;
> @@ -397,6 +398,44 @@ static struct request *dd_dispatch_request(struct blk_mq_hw_ctx *hctx)
> return rq;
> }
>
> +/*
> + * Called by __blk_mq_alloc_request(). The shallow_depth value set by this
> + * function is used by __blk_mq_get_tag().
> + */
> +static void dd_limit_depth(unsigned int op, struct blk_mq_alloc_data *data)
> +{
> + struct deadline_data *dd = data->q->elevator->elevator_data;
> +
> + /* Do not throttle synchronous reads. */
> + if (op_is_sync(op) && !op_is_write(op))
> + return;
> +
> + /*
> + * Throttle asynchronous requests and writes such that these requests
> + * do not block the allocation of synchronous requests.
> + */
> + data->shallow_depth = dd->async_depth;
> +}
> +
> +/* Called by blk_mq_update_nr_requests(). */
> +static void dd_depth_updated(struct blk_mq_hw_ctx *hctx)
> +{
> + struct request_queue *q = hctx->queue;
> + struct deadline_data *dd = q->elevator->elevator_data;
> + struct blk_mq_tags *tags = hctx->sched_tags;
> +
> + dd->async_depth = max(1UL, 3 * q->nr_requests / 4);
> +
> + sbitmap_queue_min_shallow_depth(tags->bitmap_tags, dd->async_depth);
> +}
> +
> +/* Called by blk_mq_init_hctx() and blk_mq_init_sched(). */
> +static int dd_init_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
> +{
> + dd_depth_updated(hctx);
> + return 0;
> +}
> +
> static void dd_exit_sched(struct elevator_queue *e)
> {
> struct deadline_data *dd = e->elevator_data;
> @@ -733,6 +772,15 @@ static int deadline_starved_show(void *data, struct seq_file *m)
> return 0;
> }
>
> +static int dd_async_depth_show(void *data, struct seq_file *m)
> +{
> + struct request_queue *q = data;
> + struct deadline_data *dd = q->elevator->elevator_data;
> +
> + seq_printf(m, "%u\n", dd->async_depth);
> + return 0;
> +}
> +
> static void *deadline_dispatch_start(struct seq_file *m, loff_t *pos)
> __acquires(&dd->lock)
> {
> @@ -775,6 +823,7 @@ static const struct blk_mq_debugfs_attr deadline_queue_debugfs_attrs[] = {
> DEADLINE_QUEUE_DDIR_ATTRS(write),
> {"batching", 0400, deadline_batching_show},
> {"starved", 0400, deadline_starved_show},
> + {"async_depth", 0400, dd_async_depth_show},
> {"dispatch", 0400, .seq_ops = &deadline_dispatch_seq_ops},
> {},
> };
> @@ -783,6 +832,8 @@ static const struct blk_mq_debugfs_attr deadline_queue_debugfs_attrs[] = {
>
> static struct elevator_type mq_deadline = {
> .ops = {
> + .depth_updated = dd_depth_updated,
> + .limit_depth = dd_limit_depth,
> .insert_requests = dd_insert_requests,
> .dispatch_request = dd_dispatch_request,
> .prepare_request = dd_prepare_request,
> @@ -796,6 +847,7 @@ static struct elevator_type mq_deadline = {
> .has_work = dd_has_work,
> .init_sched = dd_init_sched,
> .exit_sched = dd_exit_sched,
> + .init_hctx = dd_init_hctx,
> },
>
> #ifdef CONFIG_BLK_DEBUG_FS
>
Can't say I like heuristics. Might be needed, but there will always be
use-cases where the heuristics fail. Can't we make this value
configurable via sysfs?
Cheers,
Hannes
--
Dr. Hannes Reinecke Kernel Storage Architect
hare@xxxxxxx +49 911 74053 688
SUSE Software Solutions Germany GmbH, 90409 Nürnberg
GF: F. Imendörffer, HRB 36809 (AG Nürnberg)
