linux-kernel - Re: timers/migration: add 'notmigr' kernel parameter to disable timer migration

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: <87v7lzy2c7.ffs@tglx>
Date: Wed, 03 Sep 2025 17:37:12 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Sehee Jeong <sehee1.jeong@...sung.com>, anna-maria@...utronix.de,
 frederic@...nel.org, corbet@....net
Cc: linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
 sehee1.jeong@...sung.com
Subject: Re: timers/migration: add 'notmigr' kernel parameter to disable
 timer migration

On Thu, Aug 07 2025 at 15:48, Sehee Jeong wrote:
> On heterogeneous systems with big.LITTLE architectures, timer migration
> may cause timers from little cores to run on big cores, or vice versa,
> because core type differences are not taken into account in the current
> timer migration logic.

And what's the reason why this should be done?

> This can be undesirable in systems that require strict power
> management, predictable latency, or core isolation.

Undesirable is not really a technical argument. Can you please describe
the actual problems instead of handwaving?

> This patch does not attempt to solve the structural limitation,

# git grep "This patch" Documentation/process/

> but provides a workaround by introducing an optional early boot parameter:
>
>     notmigr
>
> When specified, timer migration initialization is skipped entirely.

I agree with Frederic that this naming is suboptimal and should be
tmigr=[on/off].

But aside of that, disabling timer migration causes other power related
issues as it keeps timers always CPU local and therefore fails to let
idle CPUs stay truly idle.

So instead of just having a lazy work around and disabling everything,
this should really be addressed properly. Though without a proper
technical problem description, that's pretty much impossible.

Thanks,

        tglx