From: Paul Jackson <pj@sgi.com>
To: KUROSAWA Takahiro <kurosawa@valinux.co.jp>,
Dinakar Guniguntala <dino@in.ibm.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH 0/5] SUBCPUSETS: a resource control functionality using CPUSETS
Date: Thu, 8 Sep 2005 00:23:23 -0700 [thread overview]
Message-ID: <20050908002323.181fd7d5.pj@sgi.com> (raw)
In-Reply-To: <20050908053912.1352770031@sv1.valinux.co.jp>
[ Adding Dinakar to explicit cc list, since I mention his work.
Hopefully he can correct any mispresentations of his work I
might have presented. - pj ]
I've just started reading this - it seems well presented and I think
you have put much effort into it. Thank-you for posting it.
I have not yet taken the time to understand it properly, but a couple
of questions come to mind offhand. I hope these questions will not be
too silly.
1) What is the relation of this patch set to CKRM?
2) Would a structure similar to Dinakar's patches to connect
cpusets and dynamic sched domains (posted to linux-mm)
work here as well?
Let me describe what I see so far, and hope you will be patient with my
confusions. Then perhaps I can better explain my question (2) above.
My initial understanding is that subcpusets provide a way to partial
out the proportion of cpu and memory used by various tasks. A leaf
node cpuset can partition the tasks attached to it into subsets, called
subcpusets, where each subcpuset gets a proportion of cpu and memory
available to the original leaf node cpuset.
These original leaf node cpusets (the parents of the 'subcpusets') form
a non-overlapping and partial covering of the cpus and memory nodes in
a system. It is a _partial_ covering because not every leaf node
cpuset need be subdivided into 'subcpusets', nor does every cpu or
memory node need to be in such a cpuset.
I'm guessing you do not want such cpusets (the parents of subcpusets)
to overlap, because if they did, it would seem to confuse the meaning
of getting a fixed proportion of available cpu and memory resources. I
was a little surprised not to see any additional checks that
cpu_exclusive and mem_exclusive must be set true in these cpusets, to
insure non- overlapping cpusets.
Dinakar's patches used cpu_exclusive cpusets to form a complete
non-overlapping covering of the cpus in a system (the cpus not
in cpu_exclusive cpusets got their own separate covers). Then each
such cover defined a different dynamic sched domain (a separately
scheduled set of tasks).
Your cover elements (your subcpusets) seem to be a somewhat new kind of
cpuset, that can only occur as children of original leaf node cpusets.
Instead of that, I might prefer doing as Dinakar did, and use a single
boolean flag per existing cpuset, to mark a subset of the cpusets that
form the covering. Dinakar got lucky, and was able to use an existing
cpuset boolean flag - cpu_exclusive, which had not been very useful
until then. You will have to add a new flag.
On the other hand, Dinakar had more work to do than you might, because
he needed a complete covering (so had to round up cpus in non exclusive
cpusets to form more covering elements). From what I can tell, you
don't need a complete covering - it seems fine if some cpus are not
managed by this resource control function.
I hope my above remarks are not to far off the mark and do not misrepresent
your work too painfully. And I hope they make at least a little sense.
Sometimes I have a strange way of describing my thoughts.
Thank-you again for this good work. I look forward to your further
considerations. I will make the time soon to read this patch more
closely.
--
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <pj@sgi.com> 1.925.600.0401
next prev parent reply other threads:[~2005-09-08 7:23 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-09-08 5:39 [PATCH 0/5] SUBCPUSETS: a resource control functionality using CPUSETS KUROSAWA Takahiro
2005-09-08 7:23 ` Paul Jackson [this message]
2005-09-08 8:18 ` KUROSAWA Takahiro
2005-09-08 12:02 ` Paul Jackson
2005-09-09 1:38 ` KUROSAWA Takahiro
2005-09-09 4:12 ` Magnus Damm
2005-09-09 5:55 ` Paul Jackson
2005-09-09 7:52 ` Magnus Damm
2005-09-09 8:39 ` Paul Jackson
2005-09-09 11:38 ` Hirokazu Takahashi
2005-09-09 13:31 ` Paul Jackson
2005-09-10 7:11 ` Hirokazu Takahashi
2005-09-10 8:52 ` Paul Jackson
2005-09-11 16:02 ` Hirokazu Takahashi
2005-09-26 9:33 ` [PATCH 0/3] CPUMETER (Re: [PATCH 0/5] SUBCPUSETS: a resource control functionality using CPUSETS) KUROSAWA Takahiro
2005-10-02 4:20 ` Paul Jackson
2005-10-04 2:49 ` KUROSAWA Takahiro
2005-09-26 9:34 ` [PATCH 1/3] CPUMETER: add cpumeter framework to the CPUSETS KUROSAWA Takahiro
2005-09-27 8:37 ` Paul Jackson
2005-09-27 9:22 ` Nick Piggin
2005-09-27 15:53 ` [ckrm-tech] " Paul Jackson
2005-09-27 21:45 ` Chandra Seetharaman
2005-09-28 6:35 ` KUROSAWA Takahiro
2005-09-28 10:08 ` Hirokazu Takahashi
2005-09-28 10:32 ` KUROSAWA Takahiro
2005-09-27 11:39 ` KUROSAWA Takahiro
2005-09-27 15:49 ` [ckrm-tech] " Paul Jackson
2005-09-28 6:21 ` KUROSAWA Takahiro
2005-09-28 6:43 ` Paul Jackson
2005-09-28 7:08 ` Paul Jackson
2005-09-28 7:53 ` KUROSAWA Takahiro
2005-09-28 16:49 ` Paul Jackson
2005-09-29 2:53 ` KUROSAWA Takahiro
2005-09-29 2:58 ` Paul Jackson
2005-09-30 9:39 ` Simon Derr
2005-09-30 14:21 ` Paul Jackson
2005-10-02 7:01 ` Ok to change cpuset flags for sched domains? (was [PATCH 1/3] CPUMETER ...) Paul Jackson
2005-10-03 14:00 ` Dinakar Guniguntala
2005-10-03 23:36 ` [ckrm-tech] " Paul Jackson
2005-09-28 9:25 ` [PATCH][BUG] fix memory leak on reading cpuset files after seeking beyond eof KUROSAWA Takahiro
2005-09-28 13:42 ` Paul Jackson
2005-09-28 13:42 ` [PATCH] cpuset read past eof memory leak fix Paul Jackson
2005-09-28 15:01 ` Linus Torvalds
2005-09-28 17:53 ` Paul Jackson
2005-09-28 18:03 ` Linus Torvalds
2005-09-28 18:03 ` Randy.Dunlap
2005-09-28 19:04 ` [ckrm-tech] " Paul Jackson
2005-09-28 14:29 ` [PATCH 1/3] CPUMETER: add cpumeter framework to the CPUSETS Paul Jackson
2005-09-26 9:34 ` [PATCH 2/3] CPUMETER: CPU resource controller KUROSAWA Takahiro
2005-09-26 9:34 ` [PATCH 3/3] CPUMETER: connect the CPU resource controller to CPUMETER KUROSAWA Takahiro
2005-09-09 22:26 ` [PATCH 0/5] SUBCPUSETS: a resource control functionality using CPUSETS Matthew Helsley
2005-09-08 13:14 ` Dinakar Guniguntala
2005-09-08 14:11 ` Dipankar Sarma
2005-09-08 14:55 ` Paul Jackson
2005-09-08 14:59 ` Paul Jackson
2005-09-08 22:51 ` [ckrm-tech] " Chandra Seetharaman
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=20050908002323.181fd7d5.pj@sgi.com \
--to=pj@sgi.com \
--cc=dino@in.ibm.com \
--cc=kurosawa@valinux.co.jp \
--cc=linux-kernel@vger.kernel.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).