[PATCH 0/3] bpf: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[PATCH 0/3] bpf: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[Marco Crivellari]

Hi!

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 (that this patch is introducing)
    -   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.

2) [P 3] add WQ_PERCPU to remaining alloc_workqueue() users

        Every alloc_workqueue() caller should use one among WQ_PERCPU or
        WQ_UNBOUND. This is actually enforced warning if both or none of them
        are present at the same time.

        WQ_UNBOUND will be removed in a next release cycle.

=== 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!

Marco Crivellari (3):
  bpf: replace use of system_wq with system_percpu_wq
  bpf: replace use of system_unbound_wq with system_dfl_wq
  bpf: WQ_PERCPU added to alloc_workqueue users

 kernel/bpf/cgroup.c   | 5 +++--
 kernel/bpf/cpumap.c   | 2 +-
 kernel/bpf/helpers.c  | 4 ++--
 kernel/bpf/memalloc.c | 2 +-
 kernel/bpf/syscall.c  | 2 +-
 5 files changed, 8 insertions(+), 7 deletions(-)

-- 
2.51.0

[PATCH 1/3] bpf: 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.

queue_work() / queue_delayed_work() mod_delayed_work() will now use the
new per-cpu wq: whether the user still stick on the old name a warn will
be printed along a wq redirect to the new one.

This patch add the new system_percpu_wq except for mm, fs and net
subsystem, whom are handled in separated patches.

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>
---
 kernel/bpf/cgroup.c | 2 +-
 kernel/bpf/cpumap.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index 84f58f3d028a..b8699ec4d766 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -27,7 +27,7 @@ EXPORT_SYMBOL(cgroup_bpf_enabled_key);
 /*
  * cgroup bpf destruction makes heavy use of work items and there can be a lot
  * of concurrent destructions.  Use a separate workqueue so that cgroup bpf
- * destruction work items don't end up filling up max_active of system_wq
+ * destruction work items don't end up filling up max_active of system_percpu_wq
  * which may lead to deadlock.
  */
 static struct workqueue_struct *cgroup_bpf_destroy_wq;
diff --git a/kernel/bpf/cpumap.c b/kernel/bpf/cpumap.c
index 67e8a2fc1a99..1ab8e6876618 100644
--- a/kernel/bpf/cpumap.c
+++ b/kernel/bpf/cpumap.c
@@ -551,7 +551,7 @@ static void __cpu_map_entry_replace(struct bpf_cpu_map *cmap,
 	old_rcpu = unrcu_pointer(xchg(&cmap->cpu_map[key_cpu], RCU_INITIALIZER(rcpu)));
 	if (old_rcpu) {
 		INIT_RCU_WORK(&old_rcpu->free_work, __cpu_map_entry_free);
-		queue_rcu_work(system_wq, &old_rcpu->free_work);
+		queue_rcu_work(system_percpu_wq, &old_rcpu->free_work);
 	}
 }
 
-- 
2.51.0

[PATCH 2/3] bpf: 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.

queue_work() / queue_delayed_work() / mod_delayed_work() will now use the
new unbound wq: whether the user still use the old wq a warn will be
printed along with a wq redirect to the new one.

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>
---
 kernel/bpf/helpers.c  | 4 ++--
 kernel/bpf/memalloc.c | 2 +-
 kernel/bpf/syscall.c  | 2 +-
 3 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index e3a2662f4e33..b969ca4d7af0 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -1593,7 +1593,7 @@ void bpf_timer_cancel_and_free(void *val)
 	 * timer callback.
 	 */
 	if (this_cpu_read(hrtimer_running)) {
-		queue_work(system_unbound_wq, &t->cb.delete_work);
+		queue_work(system_dfl_wq, &t->cb.delete_work);
 		return;
 	}
 
@@ -1606,7 +1606,7 @@ void bpf_timer_cancel_and_free(void *val)
 		if (hrtimer_try_to_cancel(&t->timer) >= 0)
 			kfree_rcu(t, cb.rcu);
 		else
-			queue_work(system_unbound_wq, &t->cb.delete_work);
+			queue_work(system_dfl_wq, &t->cb.delete_work);
 	} else {
 		bpf_timer_delete_work(&t->cb.delete_work);
 	}
diff --git a/kernel/bpf/memalloc.c b/kernel/bpf/memalloc.c
index 889374722d0a..bd45dda9dc35 100644
--- a/kernel/bpf/memalloc.c
+++ b/kernel/bpf/memalloc.c
@@ -736,7 +736,7 @@ static void destroy_mem_alloc(struct bpf_mem_alloc *ma, int rcu_in_progress)
 	/* Defer barriers into worker to let the rest of map memory to be freed */
 	memset(ma, 0, sizeof(*ma));
 	INIT_WORK(&copy->work, free_mem_alloc_deferred);
-	queue_work(system_unbound_wq, &copy->work);
+	queue_work(system_dfl_wq, &copy->work);
 }
 
 void bpf_mem_alloc_destroy(struct bpf_mem_alloc *ma)
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 9794446bc8c6..bb6f85fda240 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -901,7 +901,7 @@ static void bpf_map_free_in_work(struct bpf_map *map)
 	/* Avoid spawning kworkers, since they all might contend
 	 * for the same mutex like slab_mutex.
 	 */
-	queue_work(system_unbound_wq, &map->work);
+	queue_work(system_dfl_wq, &map->work);
 }
 
 static void bpf_map_free_rcu_gp(struct rcu_head *rcu)
-- 
2.51.0

[PATCH 3/3] bpf: 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 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 patch adds a new WQ_PERCPU flag to explicitly request the use of
the per-CPU behavior. Both flags coexist for one release cycle to allow
callers to transition their calls.

Once migration is complete, WQ_UNBOUND can be removed and unbound will
become the implicit default.

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.

All existing users have been updated accordingly.

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

diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index b8699ec4d766..f3da9400c178 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -34,7 +34,8 @@ static struct workqueue_struct *cgroup_bpf_destroy_wq;
 
 static int __init cgroup_bpf_wq_init(void)
 {
-	cgroup_bpf_destroy_wq = alloc_workqueue("cgroup_bpf_destroy", 0, 1);
+	cgroup_bpf_destroy_wq = alloc_workqueue("cgroup_bpf_destroy",
+						WQ_PERCPU, 1);
 	if (!cgroup_bpf_destroy_wq)
 		panic("Failed to alloc workqueue for cgroup bpf destroy.\n");
 	return 0;
-- 
2.51.0

Re: [PATCH 0/3] bpf: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[Alexei Starovoitov]

On Fri, Sep 5, 2025 at 1:53 AM Marco Crivellari
<marco.crivellari@suse.com> wrote:
>
> === 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 (that this patch is introducing)
>     -   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.
>
> 2) [P 3] add WQ_PERCPU to remaining alloc_workqueue() users
>
>         Every alloc_workqueue() caller should use one among WQ_PERCPU or
>         WQ_UNBOUND. This is actually enforced warning if both or none of them
>         are present at the same time.
>
>         WQ_UNBOUND will be removed in a next release cycle.
>
> === For Maintainers ===
>
> There are prerequisites for this series, already merged in the master branch.

Everything makes sense.

Tejun,
please ack this set just to make sure it's all going as planned.

Re: [PATCH 0/3] bpf: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[Tejun Heo]

On Fri, Sep 05, 2025 at 10:53:06AM +0200, Marco Crivellari wrote:
...
> Marco Crivellari (3):
>   bpf: replace use of system_wq with system_percpu_wq
>   bpf: replace use of system_unbound_wq with system_dfl_wq
>   bpf: WQ_PERCPU added to alloc_workqueue users

Acked-by: Tejun Heo <tj@kernel.org>

Thanks.

-- 
tejun

Re: [PATCH 0/3] bpf: replace wq users and add WQ_PERCPU to alloc_workqueue() users

[patchwork-bot+netdevbpf]

Hello:

This series was applied to bpf/bpf-next.git (master)
by Alexei Starovoitov <ast@kernel.org>:

On Fri,  5 Sep 2025 10:53:06 +0200 you wrote:
> Hi!
> 
> 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/
> 
> [...]

Here is the summary with links:
  - [1/3] bpf: replace use of system_wq with system_percpu_wq
    https://git.kernel.org/bpf/bpf-next/c/34f86083a4e1
  - [2/3] bpf: replace use of system_unbound_wq with system_dfl_wq
    https://git.kernel.org/bpf/bpf-next/c/0409819a0021
  - [3/3] bpf: WQ_PERCPU added to alloc_workqueue users
    https://git.kernel.org/bpf/bpf-next/c/a857210b104f

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html