[PATCH v2 0/2] rust: replace use of system_unbound_wq and system_wq

[PATCH v2 0/2] rust: replace use of system_unbound_wq and system_wq

[Marco Crivellari]

Below is a summary of a discussion about the Workqueue API and cpu isolation
considerations. Details and more information are available here:

        "workqueue: Always use wq_select_unbound_cpu() for WORK_CPU_UNBOUND."
        https://lore.kernel.org/all/20250221112003.1dSuoGyc@linutronix.de/

=== Current situation: problems ===

Let's consider a nohz_full system with isolated CPUs: wq_unbound_cpumask is
set to the housekeeping CPUs, for !WQ_UNBOUND the local CPU is selected.

This leads to different scenarios if a work item is scheduled on an isolated
CPU where "delay" value is 0 or greater then 0:
        schedule_delayed_work(, 0);

This will be handled by __queue_work() that will queue the work item on the
current local (isolated) CPU, while:

        schedule_delayed_work(, 1);

Will move the timer on an housekeeping CPU, and schedule the work there.

Currently if a user enqueue 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 consistentcy cannot be addressed without refactoring the API.

=== Plan and future plans ===

This patchset is the first stone on a refactoring needed in order to
address the points aforementioned; it will have a positive impact also
on the cpu isolation, in the long term, moving away percpu workqueue in
favor to an unbound model.

These are the main steps:
1)  API refactoring (changes introduced by this series)
    -   Make more clear and uniform the system wq names, both per-cpu and
        unbound. This to avoid any possible confusion on what should be
        used.

    -   Introduction of WQ_PERCPU: this flag is the complement of WQ_UNBOUND,
        introduced in this patchset and used on all the callers that are not
        currently using WQ_UNBOUND.

        WQ_UNBOUND will be removed in a future release cycle.

        Most users don't need to be per-cpu, because they don't have
        locality requirements, because of that, a next future step will be
        make "unbound" the default behavior.

2)  Check who really needs to be per-cpu
    -   Remove the WQ_PERCPU flag when is not strictly required.

3)  Add a new API (prefer local cpu)
    -   There are users that don't require a local execution, like mentioned
        above; despite that, local execution yeld to performance gain.

        This new API will prefer the local execution, without requiring it.

=== Introduced Changes by this series ===

1) [P 1-2] Replace use of system_wq and system_unbound_wq

        system_wq is a per-CPU workqueue, but his name is not clear.
        system_unbound_wq is to be used when locality is not required.

        Because of that, system_wq has been renamed in system_percpu_wq, and
        system_unbound_wq has been renamed in system_dfl_wq.


=== For Maintainers ===

There are prerequisites for this series, already merged in the master branch.
The commits are:

128ea9f6ccfb6960293ae4212f4f97165e42222d ("workqueue: Add system_percpu_wq and
system_dfl_wq")

930c2ea566aff59e962c50b2421d5fcc3b98b8be ("workqueue: Add new WQ_PERCPU flag")


Thanks!

---
Changes in v2:
- added system_percpu() and system_dfl() in order to use the new
  wq defined in the C code.
- fixed misleading paragraph in the commit log (no warnings are currently
  present).

Marco Crivellari (2):
  rust: replace use of system_unbound_wq with system_dfl_wq
  rust: replace use of system_wq with system_percpu_wq

 rust/kernel/workqueue.rs | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

-- 
2.51.0

[PATCH v2 1/2] rust: replace use of system_unbound_wq with system_dfl_wq

[Marco Crivellari]

Currently if a user enqueue 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 consistentcy cannot be addressed without refactoring the API.

system_unbound_wq should be the default workqueue so as not to enforce
locality constraints for random work whenever it's not required.

Adding system_dfl_wq to encourage its use when unbound work should be used.

The old system_unbound_wq will be kept for a few release cycles.

Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 rust/kernel/workqueue.rs | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/rust/kernel/workqueue.rs b/rust/kernel/workqueue.rs
index f98bd02b838f..01e331a1f11b 100644
--- a/rust/kernel/workqueue.rs
+++ b/rust/kernel/workqueue.rs
@@ -667,11 +667,18 @@ pub fn system_long() -> &'static Queue {
 /// Workers are not bound to any specific CPU, not concurrency managed, and all queued work items
 /// are executed immediately as long as `max_active` limit is not reached and resources are
 /// available.
+///
+/// Note: `system_unbound_wq` will be removed in a future release cycle. Use [`system_dfl_wq`] instead.
 pub fn system_unbound() -> &'static Queue {
     // SAFETY: `system_unbound_wq` is a C global, always available.
     unsafe { Queue::from_raw(bindings::system_unbound_wq) }
 }
 
+pub fn system_dfl() -> &'static Queue {
+    // SAFETY: `system_dfl_wq` is a C global, always available.
+    unsafe { Queue::from_raw(bindings::system_dfl_wq) }
+}
+
 /// Returns the system freezable work queue (`system_freezable_wq`).
 ///
 /// It is equivalent to the one returned by [`system`] except that it's freezable.
-- 
2.51.0

[PATCH v2 2/2] rust: replace use of system_wq with system_percpu_wq

[Marco Crivellari]

Currently if a user enqueue 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 consistentcy cannot be addressed without refactoring the API.

system_wq is a per-CPU worqueue, yet nothing in its name tells about that
CPU affinity constraint, which is very often not required by users. Make
it clear by adding a system_percpu_wq.

The old wq will be kept for a few release cylces.

Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 rust/kernel/workqueue.rs | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/rust/kernel/workqueue.rs b/rust/kernel/workqueue.rs
index 01e331a1f11b..8fb16d1015c9 100644
--- a/rust/kernel/workqueue.rs
+++ b/rust/kernel/workqueue.rs
@@ -639,11 +639,18 @@ unsafe fn __enqueue<F>(self, queue_work_on: F) -> Self::EnqueueOutput
 /// users which expect relatively short queue flush time.
 ///
 /// Callers shouldn't queue work items which can run for too long.
+///
+/// Note: `system_wq` will be removed in a future release cycle. Use [`system_percpu_wq`] instead.
 pub fn system() -> &'static Queue {
     // SAFETY: `system_wq` is a C global, always available.
     unsafe { Queue::from_raw(bindings::system_wq) }
 }
 
+pub fn system_percpu() -> &'static Queue {
+    // SAFETY: `system_percpu_wq` is a C global, always available.
+    unsafe { Queue::from_raw(bindings::system_percpu_wq) }
+}
+
 /// Returns the system high-priority work queue (`system_highpri_wq`).
 ///
 /// It is similar to the one returned by [`system`] but for work items which require higher
-- 
2.51.0

Re: [PATCH v2 0/2] rust: replace use of system_unbound_wq and system_wq

[Miguel Ojeda]

On Mon, Sep 8, 2025 at 6:02 PM Marco Crivellari
<marco.crivellari@suse.com> wrote:
>
> Below is a summary of a discussion about the Workqueue API and cpu isolation
> considerations. Details and more information are available here:

If I understand correctly, these two patches are "just" following the
C side, right?

i.e. the summary below comes from that original patch series, but this
one is just adding the `system_{dfl,percpu}_wq`s that were added on
the C side.

In other words, I would have just said that, plus a link to the patch
series or commit that actually added the new wqs.

I would do so similarly in the commit messages.

> === Introduced Changes by this series ===
>
> 1) [P 1-2] Replace use of system_wq and system_unbound_wq
>
>         system_wq is a per-CPU workqueue, but his name is not clear.
>         system_unbound_wq is to be used when locality is not required.
>
>         Because of that, system_wq has been renamed in system_percpu_wq, and
>         system_unbound_wq has been renamed in system_dfl_wq.

This is not what the patches are doing? There is no replacement nor rename.

> === For Maintainers ===
>
> There are prerequisites for this series, already merged in the master branch.
> The commits are:
>
> 128ea9f6ccfb6960293ae4212f4f97165e42222d ("workqueue: Add system_percpu_wq and
> system_dfl_wq")
>
> 930c2ea566aff59e962c50b2421d5fcc3b98b8be ("workqueue: Add new WQ_PERCPU flag")

Why are these prerequisites if they are already merged?

Thanks!

Cheers,
Miguel

Re: [PATCH v2 0/2] rust: replace use of system_unbound_wq and system_wq

[Marco Crivellari]

On Mon, Sep 8, 2025 at 6:32 PM Miguel Ojeda
<miguel.ojeda.sandonis@gmail.com> wrote:
> If I understand correctly, these two patches are "just" following the
> C side, right?
>
> i.e. the summary below comes from that original patch series, but this
> one is just adding the `system_{dfl,percpu}_wq`s that were added on
> the C side.
>
> In other words, I would have just said that, plus a link to the patch
> series or commit that actually added the new wqs.
>
> I would do so similarly in the commit messages.

Hello Miguel,

Exactly.
I kept the same cover letter for every subsystem involved to give all
the information.

> This is not what the patches are doing? There is no replacement nor rename.

Yes you're right. This is done on the C side, the rename of all the
uses; but the old wq(s) are not removed.

> > === For Maintainers ===
> >
> > There are prerequisites for this series, already merged in the master branch.
> > The commits are:
> >
> > 128ea9f6ccfb6960293ae4212f4f97165e42222d ("workqueue: Add system_percpu_wq and
> > system_dfl_wq")
> >
> > 930c2ea566aff59e962c50b2421d5fcc3b98b8be ("workqueue: Add new WQ_PERCPU flag")
>
> Why are these prerequisites if they are already merged?

I kept the old cover letter paragraph name.

Unfortunately I noticed I made some mistakes not updating the cover
letter accordingly... sorry!

Thanks!
-- 

Marco Crivellari

L3 Support Engineer, Technology & Product

marco.crivellari@suse.com

Re: [PATCH v2 0/2] rust: replace use of system_unbound_wq and system_wq

[Miguel Ojeda]

On Tue, Sep 9, 2025 at 9:25 AM Marco Crivellari
<marco.crivellari@suse.com> wrote:
>
> Unfortunately I noticed I made some mistakes not updating the cover
> letter accordingly... sorry!

No worries! It was just a bit confusing :)

Thanks!

Cheers,
Miguel

Re: [PATCH v2 0/2] rust: replace use of system_unbound_wq and system_wq

[Marco Crivellari]

On Mon, Sep 8, 2025 at 6:02 PM Marco Crivellari
<marco.crivellari@suse.com> wrote:
> [...]
> Marco Crivellari (2):
>   rust: replace use of system_unbound_wq with system_dfl_wq
>   rust: replace use of system_wq with system_percpu_wq
>

Gentle ping :-)

Thanks!
-- 

Marco Crivellari

L3 Support Engineer, Technology & Product

Re: [PATCH v2 1/2] rust: replace use of system_unbound_wq with system_dfl_wq

[Alice Ryhl]

On Mon, Sep 08, 2025 at 06:02:23PM +0200, Marco Crivellari wrote:
> Currently if a user enqueue 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 consistentcy cannot be addressed without refactoring the API.
> 
> system_unbound_wq should be the default workqueue so as not to enforce
> locality constraints for random work whenever it's not required.
> 
> Adding system_dfl_wq to encourage its use when unbound work should be used.
> 
> The old system_unbound_wq will be kept for a few release cycles.
> 
> Suggested-by: Tejun Heo <tj@kernel.org>
> Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
> ---
>  rust/kernel/workqueue.rs | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/rust/kernel/workqueue.rs b/rust/kernel/workqueue.rs
> index f98bd02b838f..01e331a1f11b 100644
> --- a/rust/kernel/workqueue.rs
> +++ b/rust/kernel/workqueue.rs
> @@ -667,11 +667,18 @@ pub fn system_long() -> &'static Queue {
>  /// Workers are not bound to any specific CPU, not concurrency managed, and all queued work items
>  /// are executed immediately as long as `max_active` limit is not reached and resources are
>  /// available.
> +///
> +/// Note: `system_unbound_wq` will be removed in a future release cycle. Use [`system_dfl_wq`] instead.
>  pub fn system_unbound() -> &'static Queue {
>      // SAFETY: `system_unbound_wq` is a C global, always available.
>      unsafe { Queue::from_raw(bindings::system_unbound_wq) }
>  }
>  
> +pub fn system_dfl() -> &'static Queue {
> +    // SAFETY: `system_dfl_wq` is a C global, always available.
> +    unsafe { Queue::from_raw(bindings::system_dfl_wq) }
> +}

Please add documentation to the new methods, similar to the
documentation that existing methods already have.

Alice

Re: [PATCH v2 1/2] rust: replace use of system_unbound_wq with system_dfl_wq

[Marco Crivellari]

On Wed, Oct 8, 2025 at 2:59 PM Alice Ryhl <aliceryhl@google.com> wrote:
> Please add documentation to the new methods, similar to the
> documentation that existing methods already have.

Sure, thanks for the feedback.

-- 

Marco Crivellari

L3 Support Engineer, Technology & Product