linux-kernel - Re: [PATCH] platform/surface: acpi-notify: add WQ_PERCPU to alloc

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: <88c0ecd8-0d61-f4bc-ae13-cce971b9c69c@linux.intel.com>
Date: Tue, 18 Nov 2025 13:29:54 +0200 (EET)
From: Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>
To: Marco Crivellari <marco.crivellari@...e.com>
cc: LKML <linux-kernel@...r.kernel.org>, platform-driver-x86@...r.kernel.org, 
    Tejun Heo <tj@...nel.org>, Lai Jiangshan <jiangshanlai@...il.com>, 
    Frederic Weisbecker <frederic@...nel.org>, 
    Sebastian Andrzej Siewior <bigeasy@...utronix.de>, 
    Michal Hocko <mhocko@...e.com>, Maximilian Luz <luzmaximilian@...il.com>, 
    Hans de Goede <hansg@...nel.org>
Subject: Re: [PATCH] platform/surface: acpi-notify: add WQ_PERCPU to
 alloc_workqueue users

On Fri, 7 Nov 2025, Marco Crivellari wrote:

> Currently if a user enqueues a work item using schedule_delayed_work() the
> used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
> WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
> schedule_work() that is using system_wq and queue_work(), that makes use
> again of WORK_CPU_UNBOUND.
> This lack of consistency cannot be addressed without refactoring the API.
> 
> alloc_workqueue() treats all queues as per-CPU by default, while unbound
> workqueues must opt-in via WQ_UNBOUND.
> 
> This default is suboptimal: most workloads benefit from unbound queues,
> allowing the scheduler to place worker threads where they’re needed and
> reducing noise when CPUs are isolated.
> 
> This continues the effort to refactor workqueue APIs, which began with
> the introduction of new workqueues and a new alloc_workqueue flag in:
> 
> commit 128ea9f6ccfb ("workqueue: Add system_percpu_wq and system_dfl_wq")
> commit 930c2ea566af ("workqueue: Add new WQ_PERCPU flag")
> 
> This change adds a new WQ_PERCPU flag to explicitly request
> alloc_workqueue() to be per-cpu when WQ_UNBOUND has not been specified.
> 
> With the introduction of the WQ_PERCPU flag (equivalent to !WQ_UNBOUND),
> any alloc_workqueue() caller that doesn’t explicitly specify WQ_UNBOUND
> must now use WQ_PERCPU.
> 
> Once migration is complete, WQ_UNBOUND can be removed and unbound will
> become the implicit default.

Hi,

There's lots of words in this but it's extremely confusingly written.

On one hand, you're saying "most workloads benefit from unbound queues" 
but then end up using percpu anyway in the actual diff!?!

If this is to support refactoring efforts, start the 
description/justification in the changelog with that, with only a concise 
explanation of the refactoring goal (you don't need to explain everything 
that is already in the changelogs of the commits you refer above).

And I'd expect explanation to the forementioned inconsistency, it might be 
because of the refactoring in steps. But if that's the case, you should 
instead mention that either a) there will be a follow-up to migrate to 
unbound where applicable (assuming it cannot be done now) or b) state 
that this change is just to keep behavior the same and each case needs to 
be evaluated once the refactoring is complete whether unbound could be 
used or not.

Also, there's one copy-pasted typo "worqueue" in the other patches
relating to pdx86 (and likely all your other copy-pasted
descriptions to other subsystems than pdx86 ;-)).

-- 
 i.

> Suggested-by: Tejun Heo <tj@...nel.org>
> Signed-off-by: Marco Crivellari <marco.crivellari@...e.com>
> ---
>  drivers/platform/surface/surface_acpi_notify.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/platform/surface/surface_acpi_notify.c b/drivers/platform/surface/surface_acpi_notify.c
> index 3b30cfe3466b..a9dcb0bbe90e 100644
> --- a/drivers/platform/surface/surface_acpi_notify.c
> +++ b/drivers/platform/surface/surface_acpi_notify.c
> @@ -862,7 +862,7 @@ static int __init san_init(void)
>  {
>  	int ret;
>  
> -	san_wq = alloc_workqueue("san_wq", 0, 0);
> +	san_wq = alloc_workqueue("san_wq", WQ_PERCPU, 0);
>  	if (!san_wq)
>  		return -ENOMEM;
>  	ret = platform_driver_register(&surface_acpi_notify);
>