linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jens Axboe <axboe@fb.com>
To: Paolo Valente <paolo.valente@linaro.org>
Cc: <linux-block@vger.kernel.org>,
	Linux-Kernal <linux-kernel@vger.kernel.org>, <osandov@fb.com>
Subject: Re: [PATCH 7/8] mq-deadline: add blk-mq adaptation of the deadline IO scheduler
Date: Fri, 20 Jan 2017 07:26:18 -0700	[thread overview]
Message-ID: <20170120142618.GA18162@kernel.dk> (raw)
In-Reply-To: <8191D993-76A7-47F9-AF5D-585B0EA1718B@linaro.org>

On Fri, Jan 20 2017, Paolo Valente wrote:
> 
> > Il giorno 17 gen 2017, alle ore 03:47, Jens Axboe <axboe@fb.com> ha scritto:
> > 
> > On 12/22/2016 09:49 AM, Paolo Valente wrote:
> >> 
> >>> Il giorno 17 dic 2016, alle ore 01:12, Jens Axboe <axboe@fb.com> ha scritto:
> >>> 
> >>> This is basically identical to deadline-iosched, except it registers
> >>> as a MQ capable scheduler. This is still a single queue design.
> >>> 
> >> 
> >> One last question (for today ...):in mq-deadline there are no
> >> "schedule dispatch" or "unplug work" functions.  In blk, CFQ and BFQ
> >> do these schedules/unplugs in a lot of cases.  What's the right
> >> replacement?  Just doing nothing?
> > 
> > You just use blk_mq_run_hw_queue() or variants thereof to kick off queue
> > runs.
> > 
> 
> Hi Jens,
> I'm working on this right now.  I have a pair of quick questions about
> performance.
> 
> In the blk version of bfq, if the in-service bfq_queue happen to have
> no more budget when the bfq dispatch function is invoked, then bfq:
> returns no request (NULL), immediately expires the in-service
> bfq_queue, and schedules a new dispatch.  The third step is taken so
> that, if other bfq_queues have requests, then a new in-service
> bfq_queue will be selected on the upcoming new dispatch, and a new
> request will be provided right away.
> 
> My questions are: is this dispatch-schedule step still needed with
> blk-mq, to avoid a stall?  If it is not needed to avoid a stall, would
> it still be needed to boost throughput, because it would force an
> immediate, next dispatch?

Generally that step is only needed if you don't dispatch a request for
that invocation, yet you have requests to dispatch. For that case, you
must ensure that the queues are run at some point in the future. So I'm
inclined to answer yes to your question, though it depends on exactly
how it happens. If you have the queue run in the code, comment it like
that, and we can always revisit.

> BTW, bfq-mq survived its first request completion.  I will provide you
> with a link to a github branch as soon as bfq-mq seems able to stand
> up with a minimal workload.

Congratulations, that's a nice milestone!

-- 
Jens Axboe

  reply	other threads:[~2017-01-20 14:26 UTC|newest]

Thread overview: 69+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-12-17  0:12 [PATCHSET v4] blk-mq-scheduling framework Jens Axboe
2016-12-17  0:12 ` [PATCH 1/8] block: move existing elevator ops to union Jens Axboe
2016-12-17  0:12 ` [PATCH 2/8] blk-mq: make mq_ops a const pointer Jens Axboe
2016-12-17  0:12 ` [PATCH 3/8] block: move rq_ioc() to blk.h Jens Axboe
2016-12-20 10:12   ` Paolo Valente
2016-12-20 15:46     ` Jens Axboe
2016-12-20 22:14       ` Jens Axboe
2016-12-17  0:12 ` [PATCH 4/8] blk-mq: un-export blk_mq_free_hctx_request() Jens Axboe
2016-12-17  0:12 ` [PATCH 5/8] blk-mq: export some helpers we need to the scheduling framework Jens Axboe
2016-12-17  0:12 ` [PATCH 6/8] blk-mq-sched: add framework for MQ capable IO schedulers Jens Axboe
2016-12-20 11:55   ` Paolo Valente
2016-12-20 15:45     ` Jens Axboe
2016-12-21  2:22     ` Jens Axboe
2016-12-22 15:20       ` Paolo Valente
2016-12-22  9:59   ` Paolo Valente
2016-12-22 11:13     ` Paolo Valente
2017-01-17  2:47       ` Jens Axboe
2017-01-17 10:13         ` Paolo Valente
2017-01-17 12:38           ` Jens Axboe
2016-12-23 10:12     ` Paolo Valente
2017-01-17  2:47     ` Jens Axboe
2017-01-17  9:17       ` Paolo Valente
2016-12-17  0:12 ` [PATCH 7/8] mq-deadline: add blk-mq adaptation of the deadline IO scheduler Jens Axboe
2016-12-20  9:34   ` Paolo Valente
2016-12-20 15:46     ` Jens Axboe
2016-12-21 11:59   ` Bart Van Assche
2016-12-21 14:22     ` Jens Axboe
2016-12-22 16:07   ` Paolo Valente
2017-01-17  2:47     ` Jens Axboe
2016-12-22 16:49   ` Paolo Valente
2017-01-17  2:47     ` Jens Axboe
2017-01-20 11:07       ` Paolo Valente
2017-01-20 14:26         ` Jens Axboe [this message]
2017-01-20 13:14   ` Paolo Valente
2017-01-20 13:18     ` Paolo Valente
2017-01-20 14:28       ` Jens Axboe
2017-01-20 14:28     ` Jens Axboe
2017-02-01 11:11   ` Paolo Valente
2017-02-02  5:19     ` Jens Axboe
2017-02-02  9:19       ` Paolo Valente
2017-02-02 15:30         ` Jens Axboe
2017-02-02 21:15           ` Paolo Valente
2017-02-02 21:32             ` Jens Axboe
2017-02-07 17:27               ` Paolo Valente
2017-02-01 11:56   ` Paolo Valente
2017-02-02  5:20     ` Jens Axboe
2017-02-16 10:46   ` Paolo Valente
2017-02-16 15:35     ` Jens Axboe
2016-12-17  0:12 ` [PATCH 8/8] blk-mq-sched: allow setting of default " Jens Axboe
2016-12-19 11:32 ` [PATCHSET v4] blk-mq-scheduling framework Paolo Valente
2016-12-19 15:20   ` Jens Axboe
2016-12-19 15:33     ` Jens Axboe
2016-12-19 18:21     ` Paolo Valente
2016-12-19 21:05       ` Jens Axboe
2016-12-22 15:28         ` Paolo Valente
2017-01-17  2:47           ` Jens Axboe
2017-01-17 10:49             ` Paolo Valente
2017-01-18 16:14               ` Paolo Valente
2017-01-18 16:21                 ` Jens Axboe
2017-01-23 17:04                   ` Paolo Valente
2017-01-23 17:42                     ` Jens Axboe
2017-01-25  8:46                       ` Paolo Valente
2017-01-25 16:13                         ` Jens Axboe
2017-01-26 14:23                           ` Paolo Valente
2016-12-22 16:23 ` Bart Van Assche
2016-12-22 16:52   ` Omar Sandoval
2016-12-22 16:57     ` Bart Van Assche
2016-12-22 17:12       ` Omar Sandoval
2016-12-22 17:39         ` Bart Van Assche

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=20170120142618.GA18162@kernel.dk \
    --to=axboe@fb.com \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=osandov@fb.com \
    --cc=paolo.valente@linaro.org \
    /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).