linux-kernel - Re: [PATCH] sched/cputime: make scale

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: <20190723140043.GB8994@redhat.com>
Date:   Tue, 23 Jul 2019 16:00:43 +0200
From:   Oleg Nesterov <oleg@...hat.com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Ingo Molnar <mingo@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Andrew Fox <afox@...hat.com>,
        Stephen Johnston <sjohnsto@...hat.com>,
        linux-kernel@...r.kernel.org,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Stanislaw Gruszka <sgruszka@...hat.com>
Subject: Re: [PATCH] sched/cputime: make scale_stime() more precise

On 07/22, Peter Zijlstra wrote:
>
> On Fri, Jul 19, 2019 at 04:37:42PM +0200, Oleg Nesterov wrote:
> > On 07/19, Peter Zijlstra wrote:
>
> > > But I'm still confused, since in the long run, it should still end up
> > > with a proportionally divided user/system, irrespective of some short
> > > term wobblies.
> >
> > Why?
> >
> > Yes, statistically the numbers are proportionally divided.
>
> This; due to the loss in precision the distribution is like a step
> function around the actual s:u ratio line, but on average it still is
> s:u.

You know, I am no longer sure... perhaps it is even worse, I need to recheck.

> Even if it were a perfect function, we'd still see increments in stime even
> if the current program state never does syscalls, simply because it
> needs to stay on that s:u line.
>
> > but you will (probably) never see the real stime == 1000 && utime == 10000
> > numbers if you watch incrementally.
>
> See, there are no 'real' stime and utime numbers. What we have are user
> and system samples -- tick based.

Yes, yes, I know.

> Sure, we take a shortcut, it wobbles a bit, but seriously, the samples
> are inaccurate anyway, so who bloody cares :-)
...
> People always complain, just tell em to go pound sand :-)

I tried ;) this was my initial reaction to this bug report.

However,

> You can construct a program that runs 99% in userspace but has all
> system samples.

Yes, but with the current implementation you do not need to construct
such a program, this is what you can easily get "in practice". And this
confuses people.

They can watch /proc/pid/stat incrementally and (when the numbers are big)
find that a program that runs 100% in userspace somehow spends 10 minutes
almost entirely in kernel. Or at least more in kernel than in userspace.
Even if task->stime doesn't grow at all.

Oleg.