From: "David Schwartz" <davids@webmaster.com>
To: <jeremy@goop.org>
Cc: "Linux-Kernel@Vger. Kernel. Org" <linux-kernel@vger.kernel.org>
Subject: RE: [PATCH][RSDL-mm 0/7] RSDL cpu scheduler for 2.6.21-rc3-mm2
Date: Tue, 13 Mar 2007 12:58:01 -0700 [thread overview]
Message-ID: <MDEHLPKNGKAHNMBLJOLKIELACDAC.davids@webmaster.com> (raw)
In-Reply-To: <45F6E69B.9000602@goop.org>
> There's a distinction between giving it more cpu and giving it higher
> priority: the important part about having high priority is getting low
> latency access to the cpu when its needed.
I agree. Tasks that voluntarily relinquish their timeslices should get lower
latency compared to other processes at the same static priority.
> This really seems like the wrong approach to me. The implication here
> and in other mails is that fairness is an inherently good thing which
> should obviously take preference over any other property.
Yes, that is the implication. The alternative to fairness is arbitrary
unfairness. "Rational unfairness" is a form of fairness.
> The old unix-style dynamic priority scheme was designed to give
> interactive processes high priorities, by using the observation that
> "interactive" means "spends a lot of time blocked waiting for input".
> That model of interactive is way too simple now, and the attempts to try
> an find an equivalent heuristic have been flawed and lead to - in some
> cases - wildly bad behaviours. I'm guessing the emphasis on "fairness"
> is in reaction to this, which is fair enough.
I don't think it makes sense for the scheduler to look for some hint that
the user would prefer a task to get more CPU and try to give it more. That's
what 'nice' is for.
> But saying that the user needs to explicitly hold the schedulers hand
> and nice everything to tell it how to schedule seems to be an abdication
> of duty, an admission of failure. We can't expect users to finesse all
> their processes with nice, and it would be a bad user interface to ask
> them to do so.
Then you will always get cases where the scheduler does not do what the user
wants because the scheduler does not *know* what the user wants. You always
have to tell a computer what you want it to do, and the best it can do is
faithfully follow your request.
I think it's completely irrational to ask for a scheduler that automatically
gives more CPU time to CPU hogs.
> And if someone/distro *does* go to all the effort of managing how to get
> all the processes at the right nice levels, you have this big legacy
> problem where you're now stuck keeping all those nice values meaningful
> as you continue to develop the scheduler. Its bad enough to make them
> do the work in the first place, but its worse if they need to make it a
> kernel version dependent function.
I agree. I'm not claiming to have the perfect solution. Let's not let the
perfect be the enemy of the good though.
DS
next prev parent reply other threads:[~2007-03-13 19:58 UTC|newest]
Thread overview: 91+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-03-11 3:57 [PATCH][RSDL-mm 0/7] RSDL cpu scheduler for 2.6.21-rc3-mm2 Con Kolivas
2007-03-11 11:39 ` Mike Galbraith
2007-03-11 11:48 ` Con Kolivas
2007-03-11 12:08 ` Mike Galbraith
2007-03-11 12:10 ` Ingo Molnar
2007-03-11 12:20 ` Mike Galbraith
2007-03-11 21:18 ` Mike Galbraith
2007-03-12 7:22 ` Mike Galbraith
2007-03-12 7:48 ` Con Kolivas
2007-03-12 8:29 ` Con Kolivas
2007-03-12 8:55 ` Mike Galbraith
2007-03-12 9:22 ` Con Kolivas
2007-03-12 9:38 ` Mike Galbraith
2007-03-12 10:27 ` Con Kolivas
2007-03-12 10:57 ` Mike Galbraith
2007-03-12 11:08 ` Ingo Molnar
2007-03-12 11:23 ` Con Kolivas
2007-03-12 13:48 ` Theodore Tso
2007-03-12 18:09 ` Con Kolivas
2007-03-12 14:34 ` Mike Galbraith
2007-03-12 15:26 ` Linus Torvalds
2007-03-12 18:10 ` Con Kolivas
2007-03-12 19:36 ` Peter Zijlstra
2007-03-12 20:36 ` Mike Galbraith
2007-03-13 4:17 ` Kyle Moffett
2007-03-13 8:09 ` Ingo Molnar
2007-03-12 18:49 ` Con Kolivas
2007-03-12 19:06 ` Xavier Bestel
2007-03-13 17:21 ` Valdis.Kletnieks
2007-03-12 20:11 ` Mike Galbraith
2007-03-12 20:38 ` Con Kolivas
2007-03-12 20:45 ` Mike Galbraith
2007-03-12 22:51 ` Con Kolivas
2007-03-13 5:10 ` Mike Galbraith
2007-03-13 5:53 ` Con Kolivas
2007-03-13 6:08 ` [ck] " Rodney Gordon II
2007-03-13 6:17 ` Mike Galbraith
2007-03-13 7:53 ` Mike Galbraith
2007-03-13 8:22 ` Ingo Molnar
2007-03-13 8:18 ` Ingo Molnar
2007-03-13 8:22 ` Mike Galbraith
2007-03-13 9:21 ` Con Kolivas
2007-03-13 9:29 ` Ingo Molnar
2007-03-13 9:41 ` Con Kolivas
2007-03-13 10:50 ` Bill Huey
2007-03-13 9:31 ` [ck] " Con Kolivas
2007-03-13 10:24 ` Xavier Bestel
2007-03-13 23:19 ` Sanjoy Mahajan
2007-03-13 9:33 ` Mike Galbraith
2007-03-13 9:39 ` Ingo Molnar
2007-03-13 10:06 ` Con Kolivas
2007-03-13 11:23 ` Mike Galbraith
2007-03-13 11:41 ` Serge Belyshev
2007-03-13 11:46 ` Mike Galbraith
2007-03-13 15:36 ` John Stoffel
2007-03-13 14:17 ` Matt Mackall
2007-03-13 15:15 ` David Schwartz
2007-03-13 17:59 ` Jeremy Fitzhardinge
2007-03-13 19:58 ` David Schwartz [this message]
2007-03-13 20:10 ` Jeremy Fitzhardinge
2007-03-13 20:35 ` Bill Huey
2007-03-13 20:27 ` Bill Huey
2007-03-16 16:42 ` Pavel Machek
2007-03-12 23:43 ` David Lang
2007-03-13 2:23 ` Lee Revell
2007-03-13 6:00 ` David Lang
2007-03-12 21:34 ` [ck] " jos poortvliet
2007-03-12 21:38 ` michael chang
2007-03-13 0:09 ` Thibaut VARENE
2007-03-13 6:08 ` Mike Galbraith
2007-03-13 6:16 ` Con Kolivas
2007-03-13 6:30 ` Mike Galbraith
2007-03-12 20:42 ` Peter Zijlstra
2007-03-12 21:05 ` Serge Belyshev
2007-03-12 21:41 ` Mike Galbraith
2007-03-12 11:25 ` Mike Galbraith
2007-03-12 9:38 ` Xavier Bestel
2007-03-12 10:34 ` Con Kolivas
2007-03-12 16:38 ` Kasper Sandberg
2007-03-14 2:25 ` Valdis.Kletnieks
2007-03-14 3:25 ` Gabriel C
2007-03-14 9:44 ` Xavier Bestel
2007-03-12 8:44 ` Mike Galbraith
2007-03-11 14:32 ` Gene Heskett
2007-03-12 6:58 ` Radoslaw Szkodzinski
2007-03-12 11:16 ` Gene Heskett
2007-03-12 11:49 ` Gene Heskett
2007-03-12 11:58 ` Con Kolivas
2007-03-12 16:38 ` Gene Heskett
2007-03-12 18:34 ` Gene Heskett
2007-03-12 19:53 Al Boldi
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=MDEHLPKNGKAHNMBLJOLKIELACDAC.davids@webmaster.com \
--to=davids@webmaster.com \
--cc=jeremy@goop.org \
--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).