On Tue, Jan 11, 2022 at 09:56:36AM +0800, Ming Lei wrote:
> 2) block throttle is block layer internal stuff, and we shouldn't expose
> blk_throtl_charge_bio_split() to driver.
>
> Maybe rename the new API as submit_split_bio_noacct(), but we can't
> reuse submit_bio_noacct() simply, otherwise blk_throtl_charge_bio_split()
> needs to be exported.
submit_bio_noacct should only ever be used for resubmitting bios that
came up from the upper layer, although they might not always be "split".
blk_throtl_charge_bio_split
> Another ides is to clearing BIO_THROTTLED before calling submit_bio_noacct(),
> meantime blk-throttle code needs to change to avoid double accounting of bio
> bytes, so the caller of submit_bio_noacct() still needs some change.
> This way can get smooth IOPS throttle, but needs to call __blk_throtl_bio
> for split bio one more time.
>
> Or other idea for this bio split vs. iops limit issue?
Well, if you want a helper specificly for splits, add one that actally
specifically handles splits and makes the callers life easier, something
like:
void bio_submit_splice(struct bio *split, struct bio *orig)
{
split->bi_opf |= REQ_NOMERGE;
trace_block_split(split, orig->bi_iter.bi_sector);
submit_bio_noacct(orig);
blk_throtl_charge_bio_split(orig);
}
including a proper kerneldoc comment.
But I still fail to grasp how a split is so different from just
resubmitting a not split bio.
