linux-kernel - Re: [PATCH v6 6/6] timers: Exclude isolated cpus from timer migation

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

Message-ID: <aFwKzZiIRnJM4TIE@localhost.localdomain>
Date: Wed, 25 Jun 2025 16:42:21 +0200
From: Frederic Weisbecker <frederic@...nel.org>
To: Gabriele Monaco <gmonaco@...hat.com>
Cc: Thomas Gleixner <tglx@...utronix.de>, linux-kernel@...r.kernel.org,
	Anna-Maria Behnsen <anna-maria@...utronix.de>,
	Waiman Long <longman@...hat.com>
Subject: Re: [PATCH v6 6/6] timers: Exclude isolated cpus from timer migation

Le Wed, Jun 25, 2025 at 03:46:02PM +0200, Gabriele Monaco a écrit :
> Thanks both for the reviews.
> I'm a bit puzzled by what is expected now, though.
> 
> The late initcall would work just fine to replace the call to
> tick_nohz_cpu_hotpluggable(), indeed superfluous for hotplug calls,
> however the checks for housekeeping CPUs is required to prevent
> isolated CPUs getting online from becoming available and so will run on
> early boot too (without any practical reason, only because the hotplug
> handlers run there).
> 
> I might avoid it by playing with cpuhp_setup_state_nocalls perhaps, but
> that feels even more hacky.
> 
> Otherwise, I can refactor the code to maintain a separate field
> (isolated), restore the 'online' field and keep the functions for
> online/offline and isolation as separate as possible, while considering
> available = !isolated && online
> 
> This would make reading housekeeping masks superfluous on hotplug (and
> boot) code, but again, it doesn't look simpler to me.
> 
> Am I missing some obviously elegant solution here?

I keep being confused as well but yes, I think you're right, we need to
keep the checks anyway on CPU up.

> 
> Thanks,
> Gabriele
> 

-- 
Frederic Weisbecker
SUSE Labs