linux-kernel - Re: [PATCH v4 1/2] rust: add system_dfl() around the new system_dfl

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: <CAAofZF6CrTkkffCZ1q_cjPZzUDBeHoJJOEwhMOd90KXu4VZSEQ@mail.gmail.com>
Date: Mon, 9 Feb 2026 18:14:22 +0100
From: Marco Crivellari <marco.crivellari@...e.com>
To: Gary Guo <gary@...yguo.net>
Cc: linux-kernel@...r.kernel.org, rust-for-linux@...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>, Miguel Ojeda <ojeda@...nel.org>, Alex Gaynor <alex.gaynor@...il.com>, 
	Alice Ryhl <aliceryhl@...gle.com>
Subject: Re: [PATCH v4 1/2] rust: add system_dfl() around the new system_dfl_wq

On Wed, Feb 4, 2026 at 5:50 PM Gary Guo <gary@...yguo.net> wrote:
> [...]
> Yeah, I appreciate the system -> system_percpu renaming which does make it clear
> that this is percpu. However I think `system_unbound_wq` to `system_dfl_wq` is
> somewhat less clear?
>
> What is a problem to leave this still being `system_unbound`? User would need to
> make a choice, and unless they know they want `percpu` it feels like natural
> that they're going to use the unbounded option.
>

Hi Gary,

Yes, it feels natural, as you said. But there are cases where
system_wq has been used,
maybe because the name felt like a "default system workqueue", or for
other reasons.

There are currently some places where we have already converted from
system_wq to
system_dfl_wq, because they don't need to be per-cpu. So I think that
suggesting the
use of an unbound workqueue as default it could work better. :-)
Then whoever really needs to be per-cpu, would use the system_percpu_wq.

BTW, few examples about the mentioned conversions:

"ASoC: SDCA: Replace use of system_wq with system_dfl_wq"
https://lore.kernel.org/all/20251230143429.179643-1-marco.crivellari@suse.com/

"Input: gpio_keys - replace use of system_wq with system_dfl_wq"
https://lore.kernel.org/all/20251106141955.218911-2-marco.crivellari@suse.com/

"regulator: irq_helper: replace use of system_wq with system_dfl_wq"
https://lore.kernel.org/all/20250929155053.400342-2-marco.crivellari@suse.com/

"regulator: qcomm-labibb: replace use of system_wq with system_dfl_wq"
https://lore.kernel.org/all/20251106145003.245866-1-marco.crivellari@suse.com/

We're going to have more changes like the above.

Also noticed that system_unbound_wq has been already renamed to system_dfl_wq
in about ~120 places in the code. :-)

Thanks!


--

Marco Crivellari

L3 Support Engineer