linux-kernel - Re: [GIT PULL] scheduler updates for v4.15

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: <1510565015.15042.38.camel@gmx.de>
Date:   Mon, 13 Nov 2017 10:23:35 +0100
From:   Mike Galbraith <efault@....de>
To:     Ingo Molnar <mingo@...nel.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     linux-kernel@...r.kernel.org,
        Peter Zijlstra <a.p.zijlstra@...llo.nl>,
        Thomas Gleixner <tglx@...utronix.de>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Steven Rostedt <rostedt@...dmis.org>
Subject: Re: [GIT PULL] scheduler updates for v4.15

On Mon, 2017-11-13 at 09:00 +0100, Ingo Molnar wrote:
> 
>  - Improve the RT IPI based balancing logic (Steven Rostedt)

Well how 'bout that, serendipity strikes again:

I just got a report that a proggy Gregory Haskins wrote back in 2007
(preempt-test) was claiming RT was broken, but no, it's every kernel
containing b6366f048e0c.  What the proggy does is to start a gaggle of
RT tasks of descending priority, giving each an identical quantum of
work to do, expecting higher priority tasks to get it done before
lower.  That currently does not always happen, we occasionally do not
get a task pushed for long enough that a task that should have been
preempted by a push shortly after hitting the CPU, manages to remain
long enough to finish well before it should have been able to.

This patch fixed it up before I could even get my troubleshooting
wheels properly smoking (they spin a bit before getting traction).
 Given that, not to mention b6366f048e0c being capable of log jamming
even fairly modest boxen, perhaps this patch wants a stable tag?

	-Mike