[PATCH 0/2] add WQ_PERCPU to alloc_workqueue() in virtio_balloon and vduse

[PATCH 0/2] add WQ_PERCPU to alloc_workqueue() in virtio_balloon and vduse

[Marco Crivellari]

Hi,

=== 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 consistency cannot be addressed without refactoring the API.

=== Recent changes to the WQ API ===

The following, address the recent changes in the Workqueue API:

- commit 128ea9f6ccfb ("workqueue: Add system_percpu_wq and system_dfl_wq")
- commit 930c2ea566af ("workqueue: Add new WQ_PERCPU flag")

The old workqueues will be removed in a future release cycle.

=== Introduced Changes by this series ===

1) [P 1-2] WQ_PERCPU added to alloc_workqueue()

    This adds a new WQ_PERCPU flag to explicitly request alloc_workqueue()
    to be per-cpu when WQ_UNBOUND has not been specified.

Thanks!

Marco Crivellari (2):
  virtio_balloon: add WQ_PERCPU to alloc_workqueue users
  vduse: add WQ_PERCPU to alloc_workqueue users

 drivers/vdpa/vdpa_user/vduse_dev.c | 3 ++-
 drivers/virtio/virtio_balloon.c    | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)

-- 
2.51.1

[PATCH 1/2] virtio_balloon: add WQ_PERCPU to alloc_workqueue users

[Marco Crivellari]

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.

Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 drivers/virtio/virtio_balloon.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
index 1b93d8c64361..74fe59f5a78c 100644
--- a/drivers/virtio/virtio_balloon.c
+++ b/drivers/virtio/virtio_balloon.c
@@ -983,7 +983,8 @@ static int virtballoon_probe(struct virtio_device *vdev)
 			goto out_del_vqs;
 		}
 		vb->balloon_wq = alloc_workqueue("balloon-wq",
-					WQ_FREEZABLE | WQ_CPU_INTENSIVE, 0);
+					WQ_FREEZABLE | WQ_CPU_INTENSIVE | WQ_PERCPU,
+					0);
 		if (!vb->balloon_wq) {
 			err = -ENOMEM;
 			goto out_del_vqs;
-- 
2.51.1

[PATCH 2/2] vduse: add WQ_PERCPU to alloc_workqueue users

[Marco Crivellari]

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.

Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 drivers/vdpa/vdpa_user/vduse_dev.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/vdpa/vdpa_user/vduse_dev.c b/drivers/vdpa/vdpa_user/vduse_dev.c
index e7bced0b5542..ae357d014564 100644
--- a/drivers/vdpa/vdpa_user/vduse_dev.c
+++ b/drivers/vdpa/vdpa_user/vduse_dev.c
@@ -2173,7 +2173,8 @@ static int vduse_init(void)
 	if (!vduse_irq_wq)
 		goto err_wq;
 
-	vduse_irq_bound_wq = alloc_workqueue("vduse-irq-bound", WQ_HIGHPRI, 0);
+	vduse_irq_bound_wq = alloc_workqueue("vduse-irq-bound",
+					     WQ_HIGHPRI | WQ_PERCPU, 0);
 	if (!vduse_irq_bound_wq)
 		goto err_bound_wq;
 
-- 
2.51.1

Re: [PATCH 0/2] add WQ_PERCPU to alloc_workqueue() in virtio_balloon and vduse

[Michael S. Tsirkin]

On Fri, Nov 07, 2025 at 04:49:15PM +0100, Marco Crivellari wrote:
> Hi,
> 
> === 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 consistency cannot be addressed without refactoring the API.
> 
> === Recent changes to the WQ API ===
> 
> The following, address the recent changes in the Workqueue API:
> 
> - commit 128ea9f6ccfb ("workqueue: Add system_percpu_wq and system_dfl_wq")
> - commit 930c2ea566af ("workqueue: Add new WQ_PERCPU flag")
> 
> The old workqueues will be removed in a future release cycle.
> 
> === Introduced Changes by this series ===
> 
> 1) [P 1-2] WQ_PERCPU added to alloc_workqueue()
> 
>     This adds a new WQ_PERCPU flag to explicitly request alloc_workqueue()
>     to be per-cpu when WQ_UNBOUND has not been specified.
> 
> Thanks!
> 
> Marco Crivellari (2):
>   virtio_balloon: add WQ_PERCPU to alloc_workqueue users
>   vduse: add WQ_PERCPU to alloc_workqueue users

To make sure, this does not seem to introduce any
functional change - you want me to queue this now?


>  drivers/vdpa/vdpa_user/vduse_dev.c | 3 ++-
>  drivers/virtio/virtio_balloon.c    | 3 ++-
>  2 files changed, 4 insertions(+), 2 deletions(-)
> 
> -- 
> 2.51.1

Re: [PATCH 0/2] add WQ_PERCPU to alloc_workqueue() in virtio_balloon and vduse

[Marco Crivellari]

On Mon, Nov 17, 2025 at 11:17 AM Michael S. Tsirkin <mst@redhat.com> wrote:
> [...]
> To make sure, this does not seem to introduce any
> functional change - you want me to queue this now?

Hi,

Yes please, there are no functional changes as you said.
We are just marking explicitly this workqueue as per-cpu.

Thanks!

-- 

Marco Crivellari

L3 Support Engineer, Technology & Product