Re: [PATCH v6] block: disable iopoll for split bio

[Christoph Hellwig <hch@infradead.org>]

On Wed, Nov 25, 2020 at 03:19:44PM +0800, Ming Lei wrote:
> Not sure the new bio flag is really required for this case, just wondering
> why not take the following simple way? BTW we are really going to run
> out of bio flag.
> 
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 55bcee5dc032..1139b1efd712 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -2157,6 +2157,7 @@ blk_qc_t blk_mq_submit_bio(struct bio *bio)
>  	unsigned int nr_segs;
>  	blk_qc_t cookie;
>  	blk_status_t ret;
> +	struct bio *orig_bio = bio;
>  
>  	blk_queue_bounce(q, &bio);
>  	__blk_queue_split(&bio, &nr_segs);
> @@ -2265,6 +2266,10 @@ blk_qc_t blk_mq_submit_bio(struct bio *bio)
>  		blk_mq_sched_insert_request(rq, false, true, true);
>  	}
>  
> +	/* don't poll splitted bio */
> +	if (orig_bio != bio)
> +		return BLK_QC_T_NONE;
> +
>  	return cookie;
>  queue_exit:
>  	blk_queue_exit(q);

Yeah, this looks simpler and more obvious, and it also catches the
bounce buffering case (not that polling + bounce buffering should ever
happen).