Re: [PATCH RFC v2 4/5] blk-mq-sched: refactor __blk_mq_do_dispatch_sched()

[Damien Le Moal <dlemoal@kernel.org>]

On 6/14/25 18:25, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Introduce struct sched_dispatch_ctx, and split the helper into
> elevator_dispatch_one_request() and elevator_finish_dispatch(). Make
> code cleaner and prepare to support request batch dispatching.

It is not clear how this patch prepares for supporting batch dispatching. Since
this is only a refactor without any semantic change, I would either drop this
comment or explain more clearly what you mean. This patch can also probably come
earlier in the series.

> 
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>


> +static int elevator_finish_dispatch(struct sched_dispatch_ctx *ctx)

Please add a comment here to document the return values.

> +/*
> + * Only SCSI implements .get_budget and .put_budget, and SCSI restarts
> + * its queue by itself in its completion handler, so we don't need to
> + * restart queue if .get_budget() fails to get the budget.
> + *
> + * Returns -EAGAIN if hctx->dispatch was found non-empty and run_work has to
> + * be run again.  This is necessary to avoid starving flushes.

And what is returned for the non error case ? (e.e. document the meaning of 0
and 1 return values).

> + */
> +static int __blk_mq_do_dispatch_sched(struct blk_mq_hw_ctx *hctx)



-- 
Damien Le Moal
Western Digital Research