From: Salman Qazi <sqazi@google.com>
To: Ming Lei <ming.lei@redhat.com>
Cc: Bart Van Assche <bvanassche@acm.org>,
Jens Axboe <axboe@kernel.dk>,
linux-block@vger.kernel.org,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Jesse Barnes <jsbarnes@google.com>,
Gwendal Grignou <gwendal@google.com>
Subject: Re: [PATCH] block: Limit number of items taken from the I/O scheduler in one go
Date: Tue, 4 Feb 2020 10:26:29 -0800 [thread overview]
Message-ID: <CAKUOC8Xvxa8nixstFOdjuf7_sCZNV6EqSDxTAQZjLhvf86LESA@mail.gmail.com> (raw)
In-Reply-To: <20200204092004.GB19922@ming.t460p>
On Tue, Feb 4, 2020 at 1:20 AM Ming Lei <ming.lei@redhat.com> wrote:
>
> On Mon, Feb 03, 2020 at 12:59:50PM -0800, Salman Qazi wrote:
> > We observed that it is possible for a flush to bypass the I/O
> > scheduler and get added to hctx->dispatch in blk_mq_sched_bypass_insert.
>
> We always bypass io scheduler for flush request.
>
> > This can happen while a kworker is running blk_mq_do_dispatch_sched call
> > in blk_mq_sched_dispatch_requests.
> >
> > However, the blk_mq_do_dispatch_sched call doesn't end in bounded time.
> > As a result, the flush can sit there indefinitely, as the I/O scheduler
> > feeds an arbitrary number of requests to the hardware.
>
> blk-mq supposes to handle requests in hctx->dispatch with higher
> priority, see blk_mq_sched_dispatch_requests().
>
> However, there is single hctx->run_work, so async run queue for dispatching
> flush request can only be run until another async run queue from scheduler
> is done.
>
> >
> > The solution is to periodically poll hctx->dispatch in
> > blk_mq_do_dispatch_sched, to put a bound on the latency of the commands
> > sitting there.
> >
> > Signed-off-by: Salman Qazi <sqazi@google.com>
> > ---
> > block/blk-mq-sched.c | 6 ++++++
> > block/blk-mq.c | 4 ++++
> > block/blk-sysfs.c | 33 +++++++++++++++++++++++++++++++++
> > include/linux/blkdev.h | 2 ++
> > 4 files changed, 45 insertions(+)
> >
> > diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c
> > index ca22afd47b3d..75cdec64b9c7 100644
> > --- a/block/blk-mq-sched.c
> > +++ b/block/blk-mq-sched.c
> > @@ -90,6 +90,7 @@ static void blk_mq_do_dispatch_sched(struct blk_mq_hw_ctx *hctx)
> > struct request_queue *q = hctx->queue;
> > struct elevator_queue *e = q->elevator;
> > LIST_HEAD(rq_list);
> > + int count = 0;
> >
> > do {
> > struct request *rq;
> > @@ -97,6 +98,10 @@ static void blk_mq_do_dispatch_sched(struct blk_mq_hw_ctx *hctx)
> > if (e->type->ops.has_work && !e->type->ops.has_work(hctx))
> > break;
> >
> > + if (count > 0 && count % q->max_sched_batch == 0 &&
> > + !list_empty_careful(&hctx->dispatch))
> > + break;
>
> q->max_sched_batch may not be needed, and it is reasonable to always
> disaptch requests in hctx->dispatch first.
>
> Also such trick is missed in dispatch from sw queue.
I will update the commit message, drop max_sched_batch and just turn
it into a simple check and add the same
thing to the dispatch from the sw queue.
>
>
> Thanks,
> Ming
>
next prev parent reply other threads:[~2020-02-04 18:26 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-30 19:34 Hung tasks with multiple partitions Salman Qazi
2020-01-30 20:49 ` Bart Van Assche
2020-01-30 21:02 ` Salman Qazi
[not found] ` <20200203204554.119849-1-sqazi@google.com>
2020-02-03 20:59 ` [PATCH] block: Limit number of items taken from the I/O scheduler in one go Salman Qazi
2020-02-04 3:47 ` Bart Van Assche
2020-02-04 9:20 ` Ming Lei
2020-02-04 18:26 ` Salman Qazi [this message]
2020-02-04 19:37 ` Salman Qazi
2020-02-05 4:55 ` Ming Lei
2020-02-05 19:57 ` Salman Qazi
2020-02-06 10:18 ` Ming Lei
2020-02-06 21:12 ` Salman Qazi
2020-02-07 2:07 ` Ming Lei
2020-02-07 15:26 ` Bart Van Assche
2020-02-07 18:45 ` Salman Qazi
2020-02-07 19:04 ` Salman Qazi
2020-02-07 20:19 ` Bart Van Assche
2020-02-07 20:37 ` Salman Qazi
2020-04-20 16:42 ` Doug Anderson
2020-04-23 20:13 ` Jesse Barnes
2020-04-23 20:34 ` Jens Axboe
2020-04-23 20:40 ` Salman Qazi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAKUOC8Xvxa8nixstFOdjuf7_sCZNV6EqSDxTAQZjLhvf86LESA@mail.gmail.com \
--to=sqazi@google.com \
--cc=axboe@kernel.dk \
--cc=bvanassche@acm.org \
--cc=gwendal@google.com \
--cc=jsbarnes@google.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ming.lei@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).