Re: [RFC PATCH v3 3/3] blk-mq: Lockout tagset iterator when exiting elevator

[Bart Van Assche]

On 3/5/21 7:14 AM, John Garry wrote:
> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
> index 7ff1b20d58e7..5950fee490e8 100644
> --- a/block/blk-mq-tag.c
> +++ b/block/blk-mq-tag.c
> @@ -358,11 +358,16 @@ void blk_mq_tagset_busy_iter(struct blk_mq_tag_set *tagset,
>  {
>  	int i;
>  
> +	if (!atomic_inc_not_zero(&tagset->iter_usage_counter))
> +		return;
> +
>  	for (i = 0; i < tagset->nr_hw_queues; i++) {
>  		if (tagset->tags && tagset->tags[i])
>  			__blk_mq_all_tag_iter(tagset->tags[i], fn, priv,
>  					      BT_TAG_ITER_STARTED);
>  	}
> +
> +	atomic_dec(&tagset->iter_usage_counter);
>  }
>  EXPORT_SYMBOL(blk_mq_tagset_busy_iter);

This changes the behavior of blk_mq_tagset_busy_iter(). What will e.g.
happen if the mtip driver calls blk_mq_tagset_busy_iter(&dd->tags,
mtip_abort_cmd, dd) concurrently with another blk_mq_tagset_busy_iter()
call and if that causes all mtip_abort_cmd() calls to be skipped?

> +	while (atomic_cmpxchg(&set->iter_usage_counter, 1, 0) != 1);

Isn't it recommended to call cpu_relax() inside busy-waiting loops?

>  	blk_mq_sched_free_requests(q);
>  	__elevator_exit(q, e);
>  
> +	atomic_set(&set->iter_usage_counter, 1);

Can it happen that the above atomic_set() call happens while a
blk_mq_tagset_busy_iter() call is in progress? Should that atomic_set()
call perhaps be changed into an atomic_inc() call?

Thanks,

Bart.