Hi Ming,

>At 2016-07-06 09:10:00, "Ming Lei" <tom.leiming@gmail.com> wrote: >Cc block list > >On Mon, Jul 4, 2016 at 10:46 PM, Jiale Li <aaronlee0817@163.com> wrote: >> Hi Tejun, >> >> These days, we have tested the cgroup blkio throttle function use fio, >> and we found a problem that when we use buffered IO or set the big block >> size like 1M, then the IO performance cannot reach the value we set. >> For example we set blkio.throttle.read_bps_device as 10M, in kernel >> version 4.3 IO performance can only reach 6M, and in kernel version >> 4.4 the actual IO bps is only 3.1M. >> >> Then we did some research and find that in kernel version 4.3 brought in >> blk_queue_split() function to split the big size bio into several parts, >> and some of them are calling the generic_make_request() again, this result >> the bio been throttled more than once. so the actual bio sent to device is >> less than we expected. > >Except for blk_queue_split(), there are other(stacked) drivers which call
>generic_make_request() too, such as drbd, dm, md and bcache.

Yes, but only when using cgroup blkio throttle function will cause problem.

> >> >> We have checked the newest kernel of 4.7-rc5, this problem is still exist. >> >> Based on this kind of situation, we propose a fix solution to add a flag bit >> in bio to let the splited bio bypass the blk_queue_split(). Below is the patch >> we used to fix this problem. > >The splitted bio is just a fast-cloned bio(except for discard bio) and not very >special compared with other fast-cloned bio, which is quite common used. > >So I guess what you need is to bypass BIO_CLONED bio for this purpose
>since all fast-cloned bio shares the same bvec table of the source bio.

Sorry about my mistake, not bypass the blk_queue_split(), should be let
the remainder bio to bypass the blk_throtl_bio() in generic_make_request().
So, bypass BIO_CLONED bio might not fix this, the problem I introduced is
in blk_throtl_bio(), it counts the dispatched bytes of the same bio more than
once, once for a bio and remainder bio of this bio after split. we need to count
one bio for only once.

> >Thanks, >Ming > >> >> Thanks >> >> From b5ea98c9fc5612f9390b65bd9cf4ff344b6cfe92 Mon Sep 17 00:00:00 2001 >> From: Jiale Li <aaronlee0817@163.com> >> Date: Mon, 4 Jul 2016 09:23:32 -0400 >> Subject: [PATCH] cgroup: Fix split bio been throttled more than once >> >> From kernel version 4.3, blk_queue_split() been added into the kernel, >> it calls the generic_make_request() function, witch means a part of >> the bio will be counted more than once in blk_throtl_bio(). This result >> the throttle function of cgroup cannot work as we expected. >> >> This patch add a new flag bit in bio, to let the split bio bypass the >> blk_throtl_bio(). >> >> Signed-off-by: Jiale Li <aaronlee0817@163.com> >> --- >> block/blk-merge.c | 1 + >> block/blk-throttle.c | 4 ++++ >> include/linux/blk_types.h | 1 + >> 3 files changed, 6 insertions(+) >> >> diff --git a/block/blk-merge.c b/block/blk-merge.c >> index 2613531..7b17a65 100644 >> --- a/block/blk-merge.c >> +++ b/block/blk-merge.c >> @@ -190,6 +190,7 @@ void blk_queue_split(struct request_queue *q, struct bio **bio, >> >> bio_chain(split, *bio); >> trace_block_split(q, split, (*bio)->bi_iter.bi_sector); >> + bio_set_flag(*bio, BIO_SPLITED); >> generic_make_request(*bio); >> *bio = split; >> } >> diff --git a/block/blk-throttle.c b/block/blk-throttle.c >> index 47a3e54..4ffde95 100644 >> --- a/block/blk-throttle.c >> +++ b/block/blk-throttle.c >> @@ -1403,6 +1403,10 @@ bool blk_throtl_bio(struct request_queue *q, struct blkcg_gq *blkg, >> >> WARN_ON_ONCE(!rcu_read_lock_held()); >> >> + /* if the bio has been splited, should not throttle again */ >> + if (bio_flagged(bio, BIO_SPLITED)) >> + goto out; >> + >> /* see throtl_charge_bio() */ >> if ((bio->bi_rw & REQ_THROTTLED) || !tg->has_rules[rw]) >> goto out; >> diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h >> index 77e5d81..b294780 100644 >> --- a/include/linux/blk_types.h >> +++ b/include/linux/blk_types.h >> @@ -120,6 +120,7 @@ struct bio { >> #define BIO_QUIET 6 /* Make BIO Quiet */ >> #define BIO_CHAIN 7 /* chained bio, ->bi_remaining in effect */ >> #define BIO_REFFED 8 /* bio has elevated ->bi_cnt */ >> +#define BIO_SPLITED 9 /* bio has been splited */ >> >> /* >> * Flags starting here get preserved by bio_reset() - this includes >> -- >> 1.9.1 >>
Thanks,

Jiale