linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Valente <paolo.valente@unimore.it>
To: Vivek Goyal <vgoyal@redhat.com>
Cc: Tejun Heo <tj@kernel.org>, Shaohua Li <shli@fb.com>,
	linux-block@vger.kernel.org, linux-kernel@vger.kernel.org,
	Jens Axboe <axboe@fb.com>,
	Kernel-team@fb.com, jmoyer@redhat.com,
	Mark Brown <broonie@kernel.org>,
	Linus Walleij <linus.walleij@linaro.org>,
	Ulf Hansson <ulf.hansson@linaro.org>
Subject: Re: [PATCH V3 00/11] block-throttle: add .high limit
Date: Wed, 5 Oct 2016 16:04:55 +0200	[thread overview]
Message-ID: <872670CC-AE1E-4AD9-8019-67EEC0EEE276@unimore.it> (raw)
In-Reply-To: <20161005131218.GB17339@redhat.com>


> Il giorno 05 ott 2016, alle ore 15:12, Vivek Goyal <vgoyal@redhat.com> ha scritto:
> 
> On Wed, Oct 05, 2016 at 02:37:00PM +0200, Paolo Valente wrote:
> 
> [..]
>> Anyway, to avoid going on with trying speculations and arguments, let
>> me retry with a practical proposal.  BFQ is out there, free.  Let's
>> just test, measure and check whether we have already a solution to
>> the problems you/we are still trying to solve in Linux.
> 
> Hi Paolo,
> 
> Does BFQ implementaiton scale for fast storage devices using blk-mq
> interface. We will want to make sure that locking and other overhead of
> BFQ is very minimal so that overall throughput does not suffer.
> 

Of course BFQ needs to be modified to work in blk-mq.  I'm rather sure
its overhead will then be small enough, just because I have already
collaborated to a basically equivalent port from single to multi-queue for
packet scheduling (with Luigi Rizzo and others), and our prototype can
make over 15 million scheduling decisions per second, and keep latency
low, even with tens of concurrent clients running on a multi-core,
multi-socket system.

For details, here is the paper [1], plus some slides [2].

Actually, the solution in [1] is a global scheduler, which is more
complex than the first blk-mq version of BFQ that I have in mind,
namely, partitioned scheduling, in which there should be one
independent scheduler instance per core.  But this is still investigation
territory. BTW, I would really appreciate help/feedback on this task [3].

Thanks,
Paolo

[1] http://info.iet.unipi.it/~luigi/papers/20160921-pspat.pdf
[2] http://info.iet.unipi.it/~luigi/pspat/
[3] https://marc.info/?l=linux-kernel&m=147066540916339&w=2

> Vivek
> 


--
Paolo Valente
Algogroup
Dipartimento di Scienze Fisiche, Informatiche e Matematiche
Via Campi 213/B
41125 Modena - Italy
http://algogroup.unimore.it/people/paolo/

  reply	other threads:[~2016-10-05 14:05 UTC|newest]

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-10-03 21:20 [PATCH V3 00/11] block-throttle: add .high limit Shaohua Li
2016-10-03 21:20 ` [PATCH v3 01/11] block-throttle: prepare support multiple limits Shaohua Li
2016-10-03 21:20 ` [PATCH v3 02/11] block-throttle: add .high interface Shaohua Li
2016-10-03 21:20 ` [PATCH v3 03/11] block-throttle: configure bps/iops limit for cgroup in high limit Shaohua Li
2016-10-03 21:20 ` [PATCH v3 04/11] block-throttle: add upgrade logic for LIMIT_HIGH state Shaohua Li
2016-10-03 21:20 ` [PATCH v3 05/11] block-throttle: add downgrade logic Shaohua Li
2016-10-03 21:20 ` [PATCH v3 06/11] blk-throttle: make sure expire time isn't too big Shaohua Li
2016-10-03 21:20 ` [PATCH v3 07/11] blk-throttle: make throtl_slice tunable Shaohua Li
2016-10-03 21:20 ` [PATCH v3 08/11] blk-throttle: detect completed idle cgroup Shaohua Li
2016-10-03 21:20 ` [PATCH v3 09/11] block-throttle: make bandwidth change smooth Shaohua Li
2016-10-03 21:20 ` [PATCH v3 10/11] block-throttle: add a simple idle detection Shaohua Li
2016-10-03 21:20 ` [PATCH v3 11/11] blk-throttle: ignore idle cgroup limit Shaohua Li
2016-10-04 13:28 ` [PATCH V3 00/11] block-throttle: add .high limit Vivek Goyal
2016-10-04 15:56   ` Tejun Heo
2016-10-04 16:22     ` Paolo Valente
2016-10-04 16:27       ` Tejun Heo
2016-10-04 17:01         ` Paolo Valente
2016-10-04 17:28           ` Shaohua Li
2016-10-04 17:43             ` Paolo Valente
2016-10-04 18:28               ` Shaohua Li
2016-10-04 19:49                 ` Paolo Valente
2016-10-04 18:54               ` Tejun Heo
2016-10-04 19:02                 ` Paolo Valente
2016-10-04 19:14                   ` Tejun Heo
2016-10-04 19:29                     ` Paolo Valente
2016-10-04 20:27                       ` Tejun Heo
2016-10-05 12:37                         ` Paolo Valente
2016-10-05 13:12                           ` Vivek Goyal
2016-10-05 14:04                             ` Paolo Valente [this message]
2016-10-05 14:49                           ` Tejun Heo
2016-10-05 18:30                             ` Shaohua Li
2016-10-05 19:08                               ` Shaohua Li
2016-10-05 19:57                                 ` Paolo Valente
2016-10-05 20:36                                   ` Shaohua Li
2016-10-06  7:22                                     ` Paolo Valente
2016-10-05 19:47                               ` Paolo Valente
2016-10-05 20:07                                 ` Paolo Valente
2016-10-05 20:46                                 ` Shaohua Li
2016-10-06  7:58                                   ` Paolo Valente
2016-10-06 13:15                                     ` Paolo Valente
2016-10-06 17:49                                       ` Vivek Goyal
2016-10-06 18:01                                         ` Paolo Valente
2016-10-06 18:32                                           ` Vivek Goyal
2016-10-06 20:51                                             ` Paolo Valente
2016-10-06 19:44                                         ` Mark Brown
2016-10-06 19:57                                     ` Shaohua Li
2016-10-06 22:24                                       ` Paolo Valente
     [not found]                         ` <CACsaVZ+AqSXHTRdpdrQQp6PuynEPeB-5YOyweWsenjvuKsD12w@mail.gmail.com>
2016-10-09  1:15                           ` Fwd: " Kyle Sanderson
2016-10-14 16:40                             ` Tejun Heo
2016-10-14 17:13                               ` Paolo Valente
2016-10-14 18:35                                 ` Tejun Heo
2016-10-16 19:02                                   ` Paolo Valente
2016-10-18  5:15                                     ` Kyle Sanderson
2016-10-06  8:04                     ` Linus Walleij
2016-10-06 11:03                       ` Mark Brown
2016-10-06 11:57                         ` Austin S. Hemmelgarn
2016-10-06 12:50                           ` Paolo Valente
2016-10-06 13:52                             ` Austin S. Hemmelgarn
2016-10-06 15:05                               ` Paolo Valente
2016-10-06 15:10                                 ` Austin S. Hemmelgarn
2016-10-08 10:46                       ` Heinz Diehl
2016-10-04 18:12     ` Vivek Goyal
2016-10-04 18:50       ` Tejun Heo
2016-10-04 18:56         ` Paolo Valente
2016-10-04 17:08   ` Shaohua Li

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=872670CC-AE1E-4AD9-8019-67EEC0EEE276@unimore.it \
    --to=paolo.valente@unimore.it \
    --cc=Kernel-team@fb.com \
    --cc=axboe@fb.com \
    --cc=broonie@kernel.org \
    --cc=jmoyer@redhat.com \
    --cc=linus.walleij@linaro.org \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=shli@fb.com \
    --cc=tj@kernel.org \
    --cc=ulf.hansson@linaro.org \
    --cc=vgoyal@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).