drivers/cxl/pci.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
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/cxl/pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
index bd100ac31672..0be4e508affe 100644
--- a/drivers/cxl/pci.c
+++ b/drivers/cxl/pci.c
@@ -136,7 +136,7 @@ static irqreturn_t cxl_pci_mbox_irq(int irq, void *id)
if (opcode == CXL_MBOX_OP_SANITIZE) {
mutex_lock(&cxl_mbox->mbox_mutex);
if (mds->security.sanitize_node)
- mod_delayed_work(system_wq, &mds->security.poll_dwork, 0);
+ mod_delayed_work(system_percpu_wq, &mds->security.poll_dwork, 0);
mutex_unlock(&cxl_mbox->mbox_mutex);
} else {
/* short-circuit the wait in __cxl_pci_mbox_send_cmd() */
--
2.51.0
On 10/30/25 9:38 AM, 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 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>
Applied to cxl/next
952e9057e66c17a9718232664368ffdaca468f93
> ---
> drivers/cxl/pci.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
> index bd100ac31672..0be4e508affe 100644
> --- a/drivers/cxl/pci.c
> +++ b/drivers/cxl/pci.c
> @@ -136,7 +136,7 @@ static irqreturn_t cxl_pci_mbox_irq(int irq, void *id)
> if (opcode == CXL_MBOX_OP_SANITIZE) {
> mutex_lock(&cxl_mbox->mbox_mutex);
> if (mds->security.sanitize_node)
> - mod_delayed_work(system_wq, &mds->security.poll_dwork, 0);
> + mod_delayed_work(system_percpu_wq, &mds->security.poll_dwork, 0);
> mutex_unlock(&cxl_mbox->mbox_mutex);
> } else {
> /* short-circuit the wait in __cxl_pci_mbox_send_cmd() */
On Tue, Nov 4, 2025 at 12:44 AM Dave Jiang <dave.jiang@intel.com> wrote: >[...] > Applied to cxl/next > 952e9057e66c17a9718232664368ffdaca468f93 Many thanks Dave! -- Marco Crivellari L3 Support Engineer, Technology & Product
On Thu, 30 Oct 2025, 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 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> Acked-by: Davidlohr Bueso <dave@stgolabs.net>
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 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.
A reference to:
commit 128ea9f6ccfb6960293ae4212f4f97165e42222d
Author: Marco Crivellari <marco.crivellari@suse.com>
Date: Sat Jun 14 15:35:29 2025 +0200
workqueue: Add system_percpu_wq and system_dfl_wq
...
Would have been nice just to save reviewers time. Regardless.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
[snip]
On Fri, Oct 31, 2025 at 3:47 PM Ira Weiny <ira.weiny@intel.com> wrote: >[...] > A reference to: > > commit 128ea9f6ccfb6960293ae4212f4f97165e42222d > Author: Marco Crivellari <marco.crivellari@suse.com> > Date: Sat Jun 14 15:35:29 2025 +0200 > > workqueue: Add system_percpu_wq and system_dfl_wq > > ... > > Would have been nice just to save reviewers time. Regardless. > > Reviewed-by: Ira Weiny <ira.weiny@intel.com> Many thanks, and sorry for the trouble! -- Marco Crivellari L3 Support Engineer, Technology & Product
On 10/30/25 9:38 AM, 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 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>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>> ---
> drivers/cxl/pci.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c
> index bd100ac31672..0be4e508affe 100644
> --- a/drivers/cxl/pci.c
> +++ b/drivers/cxl/pci.c
> @@ -136,7 +136,7 @@ static irqreturn_t cxl_pci_mbox_irq(int irq, void *id)
> if (opcode == CXL_MBOX_OP_SANITIZE) {
> mutex_lock(&cxl_mbox->mbox_mutex);
> if (mds->security.sanitize_node)
> - mod_delayed_work(system_wq, &mds->security.poll_dwork, 0);
> + mod_delayed_work(system_percpu_wq, &mds->security.poll_dwork, 0);
> mutex_unlock(&cxl_mbox->mbox_mutex);
> } else {
> /* short-circuit the wait in __cxl_pci_mbox_send_cmd() */
© 2016 - 2026 Red Hat, Inc.