linux-kernel - Re: Utime and stime are less when getrusage (RUSAGE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <OSBPR01MB21839A0A9B2464953E1A316DEB0F9@OSBPR01MB2183.jpnprd01.prod.outlook.com>
Date:   Wed, 16 Jun 2021 02:27:18 +0000
From:   "hasegawa-hitomi@...itsu.com" <hasegawa-hitomi@...itsu.com>
To:     Mel Gorman <mgorman@...e.de>,
        Peter Zijlstra <peterz@...radead.org>,
        "'fweisbec@...il.com'" <fweisbec@...il.com>,
        "'mingo@...nel.org'" <mingo@...nel.org>,
        "'tglx@...utronix.de'" <tglx@...utronix.de>,
        "'juri.lelli@...hat.com'" <juri.lelli@...hat.com>,
        "'vincent.guittot@...aro.org'" <vincent.guittot@...aro.org>
CC:     "'dietmar.eggemann@....com'" <dietmar.eggemann@....com>,
        "'rostedt@...dmis.org'" <rostedt@...dmis.org>,
        "'bsegall@...gle.com'" <bsegall@...gle.com>,
        "'bristot@...hat.com'" <bristot@...hat.com>,
        "'linux-kernel@...r.kernel.org'" <linux-kernel@...r.kernel.org>
Subject: Re: Utime and stime are less when getrusage (RUSAGE_THREAD) is
 executed on a tickless CPU.

Hi Mel, Peter, and Frederic

I'm sorry for the late reply.

> I don't think it's a straight-forward issue. I know we've had to deal with
> bugs in the past where the overhead of getting CPU usage statistics was
> high enough to dominate workloads that had self-monitoring capabilities to
> the extent the self-monitoring was counter-productive. It was particularly
> problematic when self-monitoring was being activated to find the source
> of a slowdown. I tend to agree with Peter here that the fix may be worse
> than the problem ultimately where workloads are not necessarily willing
> to pay the cost of accuracy and as he pointed out already, it's expected
> nohz_full tasks are avoiding syscalls as much as possible.

I understand that the use of syscall should be avoided as much as possible.
However, I think it is necessary to get accurate utime and stime
for performance tuning in nohz_full.
For example, how about introducing a new procfs entry or kernel configuration
so that the RUSAGE_THREAD process can switch the execution of
task_sched_runtime()? That way, It can leave it to the user to decide
whether to take performance or correctness.

Thanks.
Hitomi Hasegawa