Re: [PATCH 01/16] blk-mq-sched: add new parameter nr_requests in blk_mq_alloc_sched_tags()

[Ming Lei <ming.lei@redhat.com>]

On Thu, Aug 14, 2025 at 11:35:07AM +0800, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> This helper only support to iallocate the default number of requests,
> add a new parameter to support specific number of requests.
> 
> Prepare to fix tags double free problem if nr_requests is grown by
> queue sysfs attribute nr_requests.
> 
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>  block/blk-mq-sched.c | 11 +++++++----
>  block/blk-mq-sched.h |  2 +-
>  block/elevator.c     |  2 +-
>  3 files changed, 9 insertions(+), 6 deletions(-)
> 
> diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c
> index e2ce4a28e6c9..9a8a0b5e04a9 100644
> --- a/block/blk-mq-sched.c
> +++ b/block/blk-mq-sched.c
> @@ -454,7 +454,7 @@ void blk_mq_free_sched_tags_batch(struct xarray *et_table,
>  }
>  
>  struct elevator_tags *blk_mq_alloc_sched_tags(struct blk_mq_tag_set *set,
> -		unsigned int nr_hw_queues)
> +		unsigned int nr_hw_queues, unsigned int nr_requests)
>  {
>  	unsigned int nr_tags;
>  	int i;
> @@ -475,8 +475,11 @@ struct elevator_tags *blk_mq_alloc_sched_tags(struct blk_mq_tag_set *set,
>  	 * 128, since we don't split into sync/async like the old code
>  	 * did. Additionally, this is a per-hw queue depth.
>  	 */
> -	et->nr_requests = 2 * min_t(unsigned int, set->queue_depth,
> -			BLKDEV_DEFAULT_RQ);
> +	if (nr_requests)
> +		et->nr_requests = nr_requests;
> +	else
> +		et->nr_requests = 2 * min_t(unsigned int, set->queue_depth,
> +				BLKDEV_DEFAULT_RQ);

It looks more readable to add helper blk_mq_default_nr_requests(),
and pass it from call sites directly, then people won't be confused
with the passed zero `nr_requests`.


Thanks, 
Ming