public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Nick Piggin <nickpiggin@yahoo.com.au>
To: Peter Williams <pwil3058@bigpond.net.au>
Cc: "Chen, Kenneth W" <kenneth.w.chen@intel.com>,
	linux-kernel@vger.kernel.org
Subject: Re: bug in sched.c:task_hot()
Date: Tue, 05 Oct 2004 17:44:48 +1000	[thread overview]
Message-ID: <416250F0.5010008@yahoo.com.au> (raw)
In-Reply-To: <41624E42.8030105@bigpond.net.au>

Peter Williams wrote:
> Chen, Kenneth W wrote:
> 
>> Current implementation of task_hot() has a performance bug in it
>> that it will cause integer underflow.
>>
>> Variable "now" (typically passed in as rq->timestamp_last_tick)
>> and p->timestamp are all defined as unsigned long long.  However,
>> If former is smaller than the latter, integer under flow occurs
>> which make the result of subtraction a huge positive number. Then
>> it is compared to sd->cache_hot_time and it will wrongly identify
>> a cache hot task as cache cold.
>>
>> This bug causes large amount of incorrect process migration across
>> cpus (at stunning 10,000 per second) and we lost cache affinity very
>> quickly and almost took double digit performance regression on a db
>> transaction processing workload.  Patch to fix the bug.  Diff'ed against
>> 2.6.9-rc3.
>>
>> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
>>
>>
>> --- linux-2.6.9-rc3/kernel/sched.c.orig    2004-10-04 
>> 19:11:21.000000000 -0700
>> +++ linux-2.6.9-rc3/kernel/sched.c    2004-10-04 19:19:27.000000000 -0700
>> @@ -180,7 +180,8 @@ static unsigned int task_timeslice(task_
>>      else
>>          return SCALE_PRIO(DEF_TIMESLICE, p->static_prio);
>>  }
>> -#define task_hot(p, now, sd) ((now) - (p)->timestamp < 
>> (sd)->cache_hot_time)
>> +#define task_hot(p, now, sd) ((long long) ((now) - (p)->timestamp)    \
>> +                < (long long) (sd)->cache_hot_time)
>>
>>  enum idle_type
>>  {
> 
> 
> The interesting question is: How does now get to be less than timestamp? 
>  This probably means that timestamp_last_tick is not a good way of 
> getting a value for "now".

It is the best we can do.

>  By the way, neither is sched_clock() when 
> measuring small time differences as it is not monotonic (something that 
> I had to allow for in my scheduling code).

I'm pretty sure it is monotonic, actually. I know some CPUs can execute
rdtsc speculatively, but I don't think it would ever be sane to execute
two rdtsc's in the wrong order.

>  I applied no such safeguards 
> to the timing used by the load balancing code as I assumed that it 
> already worked.

It should (modulo this bug).

  reply	other threads:[~2004-10-05  7:44 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-10-05  2:38 bug in sched.c:task_hot() Chen, Kenneth W
2004-10-05  3:17 ` Nick Piggin
2004-10-05 17:15   ` Chen, Kenneth W
2004-10-05  7:33 ` Peter Williams
2004-10-05  7:44   ` Nick Piggin [this message]
2004-10-05  8:07     ` Peter Williams
2004-10-05  8:42       ` Nick Piggin
2004-10-05 10:03         ` Peter Williams
2004-10-05 17:39   ` Chen, Kenneth W
2004-10-05 22:09     ` Peter Williams
2004-10-05  8:03 ` Ingo Molnar

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=416250F0.5010008@yahoo.com.au \
    --to=nickpiggin@yahoo.com.au \
    --cc=kenneth.w.chen@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pwil3058@bigpond.net.au \
    /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