From: Peter Williams <pwil3058@bigpond.net.au>
To: Mike Galbraith <efault@gmx.de>
Cc: Helge Hafting <helgehaf@aitel.hist.no>,
Trond Myklebust <trond.myklebust@fys.uio.no>,
Ingo Molnar <mingo@elte.hu>, Con Kolivas <kernel@kolivas.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] sched: Fix adverse effects of NFS client on interactive response
Date: Sat, 07 Jan 2006 12:11:06 +1100 [thread overview]
Message-ID: <43BF152A.80509@bigpond.net.au> (raw)
In-Reply-To: <5.2.1.1.2.20060106074738.00bbaeb8@pop.gmx.net>
Mike Galbraith wrote:
> At 10:13 AM 1/6/2006 +1100, Peter Williams wrote:
>
>> Mike Galbraith wrote:
>>
>>> At 10:31 PM 1/5/2006 +1100, Peter Williams wrote:
>>>
>>>> Mike Galbraith wrote:
>>>>
>>>>> At 08:51 AM 1/5/2006 +1100, Peter Williams wrote:
>>>>>
>>>>>> I think that some of the harder to understand parts of the
>>>>>> scheduler code are actually attempts to overcome the undesirable
>>>>>> effects (such as those I've described) of inappropriately
>>>>>> identifying tasks as interactive. I think that it would have been
>>>>>> better to attempt to fix the inappropriate identifications rather
>>>>>> than their effects and I think the prudent use of
>>>>>> TASK_NONINTERACTIVE is an important tool for achieving this.
>>>>>
>>>>>
>>>>>
>>>>> IMHO, that's nothing but a cover for the weaknesses induced by
>>>>> using exclusively sleep time as an information source for the
>>>>> priority calculation. While this heuristic does work pretty darn
>>>>> well, it's easily fooled (intentionally or otherwise). The
>>>>> challenge is to find the right low cost informational component,
>>>>> and to stir it in at O(1).
>>>>
>>>>
>>>>
>>>> TASK_NONINTERACTIVE helps in this regard, is no cost in the code
>>>> where it's used and probably decreases the costs in the scheduler
>>>> code by enabling some processing to be skipped. If by its judicious
>>>> use the heuristic is only fed interactive sleep data the heuristics
>>>> accuracy in identifying interactive tasks should be improved. It
>>>> may also allow the heuristic to be simplified.
>>>
>>>
>>> I disagree. You can nip and tuck all the bits of sleep time you
>>> want, and it'll just shift the lumpy spots around (btdt).
>>
>>
>> Yes, but there's a lot of (understandable) reluctance to do any major
>> rework of this part of the scheduler so we're stuck with nips and
>> tucks for the time being. This patch is a zero cost nip and tuck.
>
>
> Color me skeptical, but nonetheless, it looks to me like the mechanism
> might need the attached.
Is that patch complete? (This is all I got.)
--- linux-2.6.15/kernel/sched.c.org Fri Jan 6 08:44:09 2006
+++ linux-2.6.15/kernel/sched.c Fri Jan 6 08:51:03 2006
@@ -1353,7 +1353,7 @@
out_activate:
#endif /* CONFIG_SMP */
- if (old_state == TASK_UNINTERRUPTIBLE) {
+ if (old_state & TASK_UNINTERRUPTIBLE) {
rq->nr_uninterruptible--;
/*
* Tasks on involuntary sleep don't earn
@@ -3010,7 +3010,7 @@
unlikely(signal_pending(prev))))
prev->state = TASK_RUNNING;
else {
- if (prev->state == TASK_UNINTERRUPTIBLE)
+ if (prev->state & TASK_UNINTERRUPTIBLE)
rq->nr_uninterruptible++;
deactivate_task(prev, rq);
}
In the absence of any use of TASK_NONINTERACTIVE in conjunction with
TASK_UNINTERRUPTIBLE it will have no effect. Personally, I think that
all TASK_UNINTERRUPTIBLE sleeps should be treated as non interactive
rather than just be heavily discounted (and that TASK_NONINTERACTIVE
shouldn't be needed in conjunction with it) BUT I may be wrong
especially w.r.t. media streamers such as audio and video players and
the mechanisms they use to do sleeps between cpu bursts.
>
> On the subject of nip and tuck, take a look at the little proggy posted
> in thread [SCHED] wrong priority calc - SIMPLE test case. That testcase
> was the result of Paolo Ornati looking into a real problem on his
> system. I just 'fixed' that nanosleep() problem by judicious
> application of TASK_NONINTERACTIVE to the schedule_timeout(). Sure, it
> works, but it doesn't look like anything but a bandaid (tourniquet in
> this case:) to me.
>
> -Mike
Peter
--
Peter Williams pwil3058@bigpond.net.au
"Learning, n. The kind of ignorance distinguishing the studious."
-- Ambrose Bierce
next prev parent reply other threads:[~2006-01-07 1:11 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-12-21 6:00 [PATCH] sched: Fix adverse effects of NFS client on interactive response Peter Williams
2005-12-21 6:09 ` Trond Myklebust
2005-12-21 6:32 ` Peter Williams
2005-12-21 13:21 ` Trond Myklebust
2005-12-21 13:36 ` Kyle Moffett
2005-12-21 13:40 ` Trond Myklebust
2005-12-22 2:26 ` Peter Williams
2005-12-22 22:08 ` Trond Myklebust
2005-12-22 22:33 ` Peter Williams
2005-12-22 22:59 ` Trond Myklebust
2005-12-23 0:02 ` Kyle Moffett
2005-12-23 0:25 ` Trond Myklebust
2005-12-23 3:06 ` Peter Williams
2005-12-23 9:39 ` Trond Myklebust
2005-12-23 10:49 ` Peter Williams
2005-12-23 12:51 ` Trond Myklebust
2005-12-23 13:36 ` Peter Williams
2006-01-02 12:09 ` Pekka Enberg
2005-12-23 19:07 ` Lee Revell
2005-12-23 21:08 ` Trond Myklebust
2005-12-23 21:17 ` Lee Revell
2005-12-23 21:23 ` Trond Myklebust
2005-12-23 22:04 ` Lee Revell
2005-12-23 22:10 ` Trond Myklebust
2005-12-21 16:10 ` Horst von Brand
2005-12-21 20:36 ` Kyle Moffett
2005-12-21 22:59 ` Peter Williams
2005-12-21 16:11 ` Ingo Molnar
2005-12-21 22:49 ` Peter Williams
2006-01-02 11:01 ` Helge Hafting
2006-01-02 23:54 ` Peter Williams
2006-01-04 1:25 ` Peter Williams
2006-01-04 9:40 ` Marcelo Tosatti
2006-01-04 12:18 ` Con Kolivas
2006-01-04 10:31 ` Marcelo Tosatti
2006-01-04 21:51 ` Peter Williams
2006-01-05 6:31 ` Mike Galbraith
2006-01-05 11:31 ` Peter Williams
2006-01-05 14:31 ` Mike Galbraith
2006-01-05 23:13 ` Peter Williams
2006-01-05 23:33 ` Con Kolivas
2006-01-06 0:02 ` Peter Williams
2006-01-06 0:08 ` Con Kolivas
2006-01-06 0:40 ` Peter Williams
2006-01-06 7:39 ` Mike Galbraith
2006-01-07 1:11 ` Peter Williams [this message]
2006-01-07 5:27 ` Mike Galbraith
2006-01-07 6:34 ` Peter Williams
2006-01-07 8:54 ` Mike Galbraith
2006-01-07 23:40 ` Peter Williams
2006-01-08 5:51 ` Mike Galbraith
2006-01-07 9:30 ` Con Kolivas
2006-01-07 10:23 ` Mike Galbraith
2006-01-07 23:31 ` Peter Williams
2006-01-08 0:38 ` Con Kolivas
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=43BF152A.80509@bigpond.net.au \
--to=pwil3058@bigpond.net.au \
--cc=efault@gmx.de \
--cc=helgehaf@aitel.hist.no \
--cc=kernel@kolivas.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=trond.myklebust@fys.uio.no \
/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).