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

[PATCH 0/3] block: 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):
  drivers/block: replace use of system_wq with system_percpu_wq
  drivers/block: replace use of system_unbound_wq with system_dfl_wq
  drivers/block: WQ_PERCPU added to alloc_workqueue users

 drivers/block/aoe/aoemain.c   | 2 +-
 drivers/block/nbd.c           | 2 +-
 drivers/block/rbd.c           | 2 +-
 drivers/block/rnbd/rnbd-clt.c | 2 +-
 drivers/block/sunvdc.c        | 4 ++--
 drivers/block/virtio_blk.c    | 2 +-
 drivers/block/zram/zram_drv.c | 2 +-
 7 files changed, 8 insertions(+), 8 deletions(-)

-- 
2.51.0

[PATCH 1/3] drivers/block: 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_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>
---
 drivers/block/nbd.c    | 2 +-
 drivers/block/sunvdc.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 7bdc7eb808ea..7738fce177fa 100644
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -311,7 +311,7 @@ static void nbd_mark_nsock_dead(struct nbd_device *nbd, struct nbd_sock *nsock,
 		if (args) {
 			INIT_WORK(&args->work, nbd_dead_link_work);
 			args->index = nbd->index;
-			queue_work(system_wq, &args->work);
+			queue_work(system_percpu_wq, &args->work);
 		}
 	}
 	if (!nsock->dead) {
diff --git a/drivers/block/sunvdc.c b/drivers/block/sunvdc.c
index b5727dea15bd..442546b05df8 100644
--- a/drivers/block/sunvdc.c
+++ b/drivers/block/sunvdc.c
@@ -1187,7 +1187,7 @@ static void vdc_ldc_reset(struct vdc_port *port)
 	}
 
 	if (port->ldc_timeout)
-		mod_delayed_work(system_wq, &port->ldc_reset_timer_work,
+		mod_delayed_work(system_percpu_wq, &port->ldc_reset_timer_work,
 			  round_jiffies(jiffies + HZ * port->ldc_timeout));
 	mod_timer(&port->vio.timer, round_jiffies(jiffies + HZ));
 	return;
-- 
2.51.0

[PATCH 2/3] drivers/block: 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>
---
 drivers/block/zram/zram_drv.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
index fda7d8624889..c7e0fa29a572 100644
--- a/drivers/block/zram/zram_drv.c
+++ b/drivers/block/zram/zram_drv.c
@@ -975,7 +975,7 @@ static int read_from_bdev_sync(struct zram *zram, struct page *page,
 	work.entry = entry;
 
 	INIT_WORK_ONSTACK(&work.work, zram_sync_read);
-	queue_work(system_unbound_wq, &work.work);
+	queue_work(system_dfl_wq, &work.work);
 	flush_work(&work.work);
 	destroy_work_on_stack(&work.work);
 
-- 
2.51.0

[PATCH 3/3] drivers/block: 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>
---
 drivers/block/aoe/aoemain.c   | 2 +-
 drivers/block/rbd.c           | 2 +-
 drivers/block/rnbd/rnbd-clt.c | 2 +-
 drivers/block/sunvdc.c        | 2 +-
 drivers/block/virtio_blk.c    | 2 +-
 5 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/block/aoe/aoemain.c b/drivers/block/aoe/aoemain.c
index cdf6e4041bb9..3b21750038ee 100644
--- a/drivers/block/aoe/aoemain.c
+++ b/drivers/block/aoe/aoemain.c
@@ -44,7 +44,7 @@ aoe_init(void)
 {
 	int ret;
 
-	aoe_wq = alloc_workqueue("aoe_wq", 0, 0);
+	aoe_wq = alloc_workqueue("aoe_wq", WQ_PERCPU, 0);
 	if (!aoe_wq)
 		return -ENOMEM;
 
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index faafd7ff43d6..af0e21149dbc 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -7389,7 +7389,7 @@ static int __init rbd_init(void)
 	 * The number of active work items is limited by the number of
 	 * rbd devices * queue depth, so leave @max_active at default.
 	 */
-	rbd_wq = alloc_workqueue(RBD_DRV_NAME, WQ_MEM_RECLAIM, 0);
+	rbd_wq = alloc_workqueue(RBD_DRV_NAME, WQ_MEM_RECLAIM | WQ_PERCPU, 0);
 	if (!rbd_wq) {
 		rc = -ENOMEM;
 		goto err_out_slab;
diff --git a/drivers/block/rnbd/rnbd-clt.c b/drivers/block/rnbd/rnbd-clt.c
index 15627417f12e..b3a0470f9e80 100644
--- a/drivers/block/rnbd/rnbd-clt.c
+++ b/drivers/block/rnbd/rnbd-clt.c
@@ -1809,7 +1809,7 @@ static int __init rnbd_client_init(void)
 		unregister_blkdev(rnbd_client_major, "rnbd");
 		return err;
 	}
-	rnbd_clt_wq = alloc_workqueue("rnbd_clt_wq", 0, 0);
+	rnbd_clt_wq = alloc_workqueue("rnbd_clt_wq", WQ_PERCPU, 0);
 	if (!rnbd_clt_wq) {
 		pr_err("Failed to load module, alloc_workqueue failed.\n");
 		rnbd_clt_destroy_sysfs_files();
diff --git a/drivers/block/sunvdc.c b/drivers/block/sunvdc.c
index 442546b05df8..851763e5dd18 100644
--- a/drivers/block/sunvdc.c
+++ b/drivers/block/sunvdc.c
@@ -1215,7 +1215,7 @@ static int __init vdc_init(void)
 {
 	int err;
 
-	sunvdc_wq = alloc_workqueue("sunvdc", 0, 0);
+	sunvdc_wq = alloc_workqueue("sunvdc", WQ_PERCPU, 0);
 	if (!sunvdc_wq)
 		return -ENOMEM;
 
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index 7cffea01d868..a5a48f976a20 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -1683,7 +1683,7 @@ static int __init virtio_blk_init(void)
 {
 	int error;
 
-	virtblk_wq = alloc_workqueue("virtio-blk", 0, 0);
+	virtblk_wq = alloc_workqueue("virtio-blk", WQ_PERCPU, 0);
 	if (!virtblk_wq)
 		return -ENOMEM;
 
-- 
2.51.0

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

[Jens Axboe]

On Fri, 05 Sep 2025 10:51:38 +0200, Marco Crivellari wrote:
> 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 ===
> 
> [...]

Applied, thanks!

[1/3] drivers/block: replace use of system_wq with system_percpu_wq
      commit: 51723bf92679427bba09a76fc13e1b0a93b680bc
[2/3] drivers/block: replace use of system_unbound_wq with system_dfl_wq
      commit: 456cefcb312d90d12dbcf7eaf6b3f7cfae6f622b
[3/3] drivers/block: WQ_PERCPU added to alloc_workqueue users
      commit: d7b1cdc9108f46f47a0899597d6fa270f64dd98c

Best regards,
-- 
Jens Axboe

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

[Marco Crivellari]

On Tue, Sep 9, 2025 at 5:12 PM Jens Axboe <axboe@kernel.dk> wrote:
> Applied, thanks!
>
> [1/3] drivers/block: replace use of system_wq with system_percpu_wq
>       commit: 51723bf92679427bba09a76fc13e1b0a93b680bc
> [2/3] drivers/block: replace use of system_unbound_wq with system_dfl_wq
>       commit: 456cefcb312d90d12dbcf7eaf6b3f7cfae6f622b
> [3/3] drivers/block: WQ_PERCPU added to alloc_workqueue users
>       commit: d7b1cdc9108f46f47a0899597d6fa270f64dd98c

Many thanks, Jens!

-- 

Marco Crivellari

L3 Support Engineer, Technology & Product