linux-kernel - Re: [PATCH v2 0/4] sched/fair: Active balancer RT/DL preemption fix

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] [day] [month] [year] [list]

Message-ID: <c4f7940e-8a8a-378e-ca02-034e3b7348ef@arm.com>
Date:   Tue, 1 Oct 2019 15:15:28 +0100
From:   Valentin Schneider <valentin.schneider@....com>
To:     Juri Lelli <juri.lelli@...hat.com>
Cc:     linux-kernel@...r.kernel.org, mingo@...nel.org,
        peterz@...radead.org, vincent.guittot@...aro.org,
        tglx@...utronix.de, qais.yousef@....com,
        Steven Rostedt <rostedt@...dmis.org>
Subject: Re: [PATCH v2 0/4] sched/fair: Active balancer RT/DL preemption fix

Hi Juri,

On 01/10/2019 14:31, Juri Lelli wrote:
> Hi Valentin,
> 
> On 01/10/19 11:29, Valentin Schneider wrote:
>> (expanded the Cc list)
>> RT/DL folks, any thought on the thing?
> 
> Even if I like your idea and it looks theoretically the right thing to
> do, I'm not sure we want it in practice if it adds complexity to CFS.
> 
> I personally never noticed this kind of interference from CFS, but, at
> the same time, for RT we usually like more to be safe than sorry.
> However, since this doesn't seem to be bullet-proof (as you also say), I
> guess it all boils down again to complexity vs. practical benefits.
> 

Thanks for having a look.

IMO worst part is the local detach_one_task() thing, I added that after v1
following Qais' comments but perhaps it doesn't gain us much.

I'll try to cook something up with rt-app and see if I can get sensible
numbers.

> Best,
> 
> Juri
>