linux-kernel - Re: [PATCHSET v1 wq/for-6.5] workqueue: Improve unbound workqueue execution locality

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 for Android: free password hash cracker in your pocket

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

Message-ID: <05fe8027e0f447d3249d423d843f943c3974489b.camel@fb.com>
Date:   Tue, 23 May 2023 20:08:42 +0000
From:   Rik van Riel <riel@...a.com>
To:     "torvalds@...ux-foundation.org" <torvalds@...ux-foundation.org>,
        "tj@...nel.org" <tj@...nel.org>,
        "dhavale@...gle.com" <dhavale@...gle.com>
CC:     "nhuck@...gle.com" <nhuck@...gle.com>,
        "brho@...gle.com" <brho@...gle.com>,
        Kernel Team <kernel-team@...a.com>,
        "agk@...hat.com" <agk@...hat.com>,
        "peterz@...radead.org" <peterz@...radead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "joshdon@...gle.com" <joshdon@...gle.com>,
        "briannorris@...omium.org" <briannorris@...omium.org>,
        "snitzer@...nel.org" <snitzer@...nel.org>,
        "jiangshanlai@...il.com" <jiangshanlai@...il.com>,
        "void@...ifault.com" <void@...ifault.com>
Subject: Re: [PATCHSET v1 wq/for-6.5] workqueue: Improve unbound workqueue
 execution locality

On Tue, 2023-05-23 at 10:59 -0700, Linus Torvalds wrote:
> 
> I really hate how we have random drivers and filesystems doing random
> workarounds for "kthread workers don't work well enough, so add
> random
> tweaks".

Part of this seems to be due to the way CFS works.

CFS policy seems to make sense for a lot of workloads, but
there are some cases with kworkers where the CFS policies
just don't work quite right. Unfortunately the scheduler
problem space is not all that well explored, and it isn't
clear what the desired behavior of a scheduler should be
in every case.