[PATCH 0/5] replace old wq(s), added WQ_PERCPU to alloc_workqueue

[PATCH 0/5] replace old wq(s), added WQ_PERCPU to alloc_workqueue

[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]  Replace uses 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 replaced with system_percpu_wq, and
	system_unbound_wq has been replaced with system_dfl_wq.

2) [P 3-4-5] WQ_PERCPU added to alloc_workqueue()

	This change 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 (5):
  ACPI: scan: replace use of system_unbound_wq with system_dfl_wq
  ACPI: OSL: replace use of system_wq with system_percpu_wq
  ACPI: EC: WQ_PERCPU added to alloc_workqueue users
  ACPI: OSL: WQ_PERCPU added to alloc_workqueue users
  ACPI: thermal: WQ_PERCPU added to alloc_workqueue users

 drivers/acpi/ec.c      | 3 ++-
 drivers/acpi/osl.c     | 6 +++---
 drivers/acpi/scan.c    | 2 +-
 drivers/acpi/thermal.c | 3 ++-
 4 files changed, 8 insertions(+), 6 deletions(-)

-- 
2.51.0

[PATCH 1/5] ACPI: scan: 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 consistency 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>
---
 drivers/acpi/scan.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/acpi/scan.c b/drivers/acpi/scan.c
index ef16d58b2949..e0c0b5a50d26 100644
--- a/drivers/acpi/scan.c
+++ b/drivers/acpi/scan.c
@@ -2397,7 +2397,7 @@ static bool acpi_scan_clear_dep_queue(struct acpi_device *adev)
 	 * initial enumeration of devices is complete, put it into the unbound
 	 * workqueue.
 	 */
-	queue_work(system_unbound_wq, &cdw->work);
+	queue_work(system_dfl_wq, &cdw->work);
 
 	return true;
 }
-- 
2.51.0

[PATCH 2/5] ACPI: OSL: 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 consistency cannot be addressed without refactoring the API.

system_wq should be the per-cpu workqueue, yet in this name nothing makes
that clear, so replace system_wq with system_percpu_wq.

The old wq (system_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>
---
 drivers/acpi/osl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
index 5ff343096ece..a79a5d47bdb8 100644
--- a/drivers/acpi/osl.c
+++ b/drivers/acpi/osl.c
@@ -398,7 +398,7 @@ static void acpi_os_drop_map_ref(struct acpi_ioremap *map)
 	list_del_rcu(&map->list);
 
 	INIT_RCU_WORK(&map->track.rwork, acpi_os_map_remove);
-	queue_rcu_work(system_wq, &map->track.rwork);
+	queue_rcu_work(system_percpu_wq, &map->track.rwork);
 }
 
 /**
-- 
2.51.0

[PATCH 3/5] ACPI: EC: WQ_PERCPU added to alloc_workqueue users

[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.

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 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/acpi/ec.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/acpi/ec.c b/drivers/acpi/ec.c
index 7855bbf752b1..59b3d50ff01e 100644
--- a/drivers/acpi/ec.c
+++ b/drivers/acpi/ec.c
@@ -2294,7 +2294,8 @@ static int acpi_ec_init_workqueues(void)
 		ec_wq = alloc_ordered_workqueue("kec", 0);
 
 	if (!ec_query_wq)
-		ec_query_wq = alloc_workqueue("kec_query", 0, ec_max_queries);
+		ec_query_wq = alloc_workqueue("kec_query", WQ_PERCPU,
+					      ec_max_queries);
 
 	if (!ec_wq || !ec_query_wq) {
 		acpi_ec_destroy_workqueues();
-- 
2.51.0

[PATCH 4/5] ACPI: OSL: WQ_PERCPU added to alloc_workqueue users

[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.

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 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/acpi/osl.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
index a79a5d47bdb8..05393a7315fe 100644
--- a/drivers/acpi/osl.c
+++ b/drivers/acpi/osl.c
@@ -1694,8 +1694,8 @@ acpi_status __init acpi_os_initialize(void)
 
 acpi_status __init acpi_os_initialize1(void)
 {
-	kacpid_wq = alloc_workqueue("kacpid", 0, 1);
-	kacpi_notify_wq = alloc_workqueue("kacpi_notify", 0, 0);
+	kacpid_wq = alloc_workqueue("kacpid", WQ_PERCPU, 1);
+	kacpi_notify_wq = alloc_workqueue("kacpi_notify", WQ_PERCPU, 0);
 	kacpi_hotplug_wq = alloc_ordered_workqueue("kacpi_hotplug", 0);
 	BUG_ON(!kacpid_wq);
 	BUG_ON(!kacpi_notify_wq);
-- 
2.51.0

[PATCH 5/5] ACPI: thermal: WQ_PERCPU added to alloc_workqueue users

[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.

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 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>
CC: Zhang Rui <rui.zhang@intel.com>
---
 drivers/acpi/thermal.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/acpi/thermal.c b/drivers/acpi/thermal.c
index 8537395b417b..a511f9ea0267 100644
--- a/drivers/acpi/thermal.c
+++ b/drivers/acpi/thermal.c
@@ -1060,7 +1060,8 @@ static int __init acpi_thermal_init(void)
 	}
 
 	acpi_thermal_pm_queue = alloc_workqueue("acpi_thermal_pm",
-						WQ_HIGHPRI | WQ_MEM_RECLAIM, 0);
+						WQ_HIGHPRI | WQ_MEM_RECLAIM | WQ_PERCPU,
+						0);
 	if (!acpi_thermal_pm_queue)
 		return -ENODEV;
 
-- 
2.51.0

Re: [PATCH 0/5] replace old wq(s), added WQ_PERCPU to alloc_workqueue

[Rafael J. Wysocki]

On Thu, Oct 30, 2025 at 4:47 PM Marco Crivellari
<marco.crivellari@suse.com> 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]  Replace uses 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 replaced with system_percpu_wq, and
>         system_unbound_wq has been replaced with system_dfl_wq.
>
> 2) [P 3-4-5] WQ_PERCPU added to alloc_workqueue()
>
>         This change 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 (5):
>   ACPI: scan: replace use of system_unbound_wq with system_dfl_wq
>   ACPI: OSL: replace use of system_wq with system_percpu_wq
>   ACPI: EC: WQ_PERCPU added to alloc_workqueue users
>   ACPI: OSL: WQ_PERCPU added to alloc_workqueue users
>   ACPI: thermal: WQ_PERCPU added to alloc_workqueue users
>
>  drivers/acpi/ec.c      | 3 ++-
>  drivers/acpi/osl.c     | 6 +++---
>  drivers/acpi/scan.c    | 2 +-
>  drivers/acpi/thermal.c | 3 ++-
>  4 files changed, 8 insertions(+), 6 deletions(-)
>
> --

All applied as 6.19 material with minor tweaks in some subjects.

Thanks!

Re: [PATCH 0/5] replace old wq(s), added WQ_PERCPU to alloc_workqueue

[Marco Crivellari]

On Mon, Nov 3, 2025 at 6:47 PM Rafael J. Wysocki <rafael@kernel.org> wrote:
>[...]
> All applied as 6.19 material with minor tweaks in some subjects.
>
> Thanks!

Many thanks Rafael!


-- 

Marco Crivellari

L3 Support Engineer, Technology & Product