mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Glauber Costa <glommer@parallels.com>
Cc: linux-kernel@vger.kernel.org, paul@paulmenage.org,
	lizf@cn.fujitsu.com, daniel.lezcano@free.fr,
	jbottomley@parallels.com
Subject: Re: [RFD 0/9] per-cgroup /proc/stat statistics
Date: Wed, 28 Sep 2011 00:11:53 +0200	[thread overview]
Message-ID: <1317161513.21836.28.camel@twins> (raw)
In-Reply-To: <1316816432-9237-1-git-send-email-glommer@parallels.com>

On Fri, 2011-09-23 at 19:20 -0300, Glauber Costa wrote:
> Hi,
> 
> Since I've sent already a RFC about it, I am sending now a RFD.
> If you eager for meaning, this can then be a "Request for Doctors",
> since Peter is likely to have a heart attack now.

:-)

All we need is to ensure the case of cgroups enabled but not used isn't
actually more expensive that what we have now, after that, if people
create a 100 deep cgroup hierarchy they get what they asked.

>From a conceptual pov this patch-set is a lot saner than the previous
one, doesn't duplicate nearly as much and actually tries to improve the
code (although I suspect simply killing off cputime64_t as a whole will
get us even more).

> So here's the deal:
> 
> * My main goal here was to demonstrate that we can avoid double accounting
>   in a lot of places. So what I did was getting rid of the original and first
>   kstat mechanism, and use only cgroups accounting for that. Since the parent
>   is always updated, the original stats are the one for the root cgroup.

Right, current patch-set won't compile for those who have CGROUP=n
kernels though, need to find something for that. Shouldn't be too hard
though. It looks like you only need to provide static per-cpu storage
and a custom version of task_cgroup_account_field().

> * I believe that all those cpu cgroups are confusing and should be unified. Not
>   that we can simply get rid of it, but my goal here is to provide all the
>   information they do, in cpu cgroup. If the set of tasks needed for accounting
>   is not independent of the ones in cpu cgroup, we can avoid double accounting
>   for that. I default cpuacct to n, but leave it to people that wants to use it
>   alone. 

Amen! Ideally we place cpuacct on the deprecated list or somesuch..

> * Well, I'm also doing what I was doing originally: Providing a per-cgroup version
>   of the /proc/stat file.

Right, so how much sense does it make to keep calling it proc.stat?

  parent reply	other threads:[~2011-09-27 22:12 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-09-23 22:20 Glauber Costa
2011-09-23 22:20 ` [RFD 1/9] Change cpustat fields to an array Glauber Costa
2011-09-27 21:00   ` Peter Zijlstra
2011-09-28 15:13     ` Glauber Costa
2011-09-28 15:23       ` Peter Zijlstra
2011-09-28 18:19     ` Glauber Costa
2011-09-28 19:09       ` Peter Zijlstra
2011-09-28 20:04         ` Glauber Costa
2011-10-01 17:47         ` Glauber Costa
2011-09-27 21:03   ` Peter Zijlstra
2011-09-28 15:14     ` Glauber Costa
2011-09-23 22:20 ` [RFD 2/9] Move /proc/stat logic inside sched.c Glauber Costa
2011-09-23 22:20 ` [RFD 3/9] Display /proc/stat information per cgroup Glauber Costa
2011-09-27 17:01   ` Balbir Singh
2011-09-27 18:42     ` Glauber Costa
2011-09-27 22:21       ` Peter Zijlstra
2011-09-28 15:22         ` Glauber Costa
2011-09-28 15:23         ` Glauber Costa
2011-09-27 21:48   ` Peter Zijlstra
2011-09-28 15:14     ` Glauber Costa
2011-09-27 21:52   ` Peter Zijlstra
2011-09-28 15:15     ` Glauber Costa
2011-09-23 22:20 ` [RFD 4/9] Make total_forks per-cgroup Glauber Costa
2011-09-27 22:00   ` Peter Zijlstra
2011-09-28  8:13     ` Martin Schwidefsky
2011-09-28 10:35       ` Peter Zijlstra
2011-09-28 12:42         ` Martin Schwidefsky
2011-09-28 12:53           ` Peter Zijlstra
2011-09-28 15:29             ` Glauber Costa
2011-09-28 15:33               ` Peter Zijlstra
2011-09-28 15:35                 ` Glauber Costa
2011-09-28 15:37                   ` Peter Zijlstra
2011-09-28 15:39                     ` Glauber Costa
2011-09-28 15:28           ` Glauber Costa
2011-09-28 15:27         ` Glauber Costa
2011-09-28 15:26       ` Glauber Costa
2011-09-23 22:20 ` [RFD 5/9] per-cgroup boot time Glauber Costa
2011-09-23 22:20 ` [RFD 6/9] Report steal time for cgroup Glauber Costa
2011-09-23 22:20 ` [RFD 7/9] provide a version of cpuacct statistics inside cpu cgroup Glauber Costa
2011-09-23 22:20 ` [RFD 8/9] provide a version of cpuusage " Glauber Costa
2011-09-23 22:20 ` [RFD 9/9] Change CPUACCT to default n Glauber Costa
2011-09-27 22:11 ` Peter Zijlstra [this message]
2011-09-28 15:21   ` [RFD 0/9] per-cgroup /proc/stat statistics Glauber Costa
2011-09-28 15:27     ` Peter Zijlstra

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=1317161513.21836.28.camel@twins \
    --to=a.p.zijlstra@chello.nl \
    --cc=daniel.lezcano@free.fr \
    --cc=glommer@parallels.com \
    --cc=jbottomley@parallels.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lizf@cn.fujitsu.com \
    --cc=paul@paulmenage.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

Powered by JetHome