On Tue, Oct 29, 2019 at 01:30:24PM +0900, Keith Busch wrote:
> On Tue, Oct 29, 2019 at 12:19:04PM +0800, Ming Lei wrote:
> > @@ -309,6 +309,10 @@ void __blk_queue_split(struct request_queue *q, struct bio **bio,
> > nr_segs);
> > break;
> > default:
> > + if (!bio_flagged(*bio, BIO_MULTI_PAGE)) {
> > + *nr_segs = 1;
> > + return;
> > + }
> > split = blk_bio_segment_split(q, *bio, &q->bio_split, nr_segs);
> > break;
> > }
>
> Is there anything to gain by clearing this new flag if the result of
> blk_bio_segment_split() creates single page bio's?
That may save nothing except for one bio split for stacking devices.
thanks,
Ming
