[PATCH 0/2] replace system_wq with system_percpu_wq

[PATCH 0/2] replace system_wq with system_percpu_wq

[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.
    Because of that, system_wq has been replaced with system_percpu_wq,
    keeping the old behavior.

Thanks!

Marco Crivellari (2):
  platform: surface: replace use of system_wq with system_percpu_wq
  platform: x86: replace use of system_wq with system_percpu_wq

 drivers/platform/surface/aggregator/ssh_packet_layer.c     | 2 +-
 drivers/platform/surface/aggregator/ssh_request_layer.c    | 2 +-
 drivers/platform/x86/gpd-pocket-fan.c                      | 4 ++--
 drivers/platform/x86/x86-android-tablets/vexia_atla10_ec.c | 2 +-
 4 files changed, 5 insertions(+), 5 deletions(-)

-- 
2.51.1

[PATCH 1/2] platform: surface: replace use of system_wq with system_percpu_wq

[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.

This patch continues the effort to refactor worqueue APIs, which has begun
with the change introducing new workqueues and a new alloc_workqueue flag:

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

Replace system_wq with system_percpu_wq, keeping the same behavior.
The old wq (system_wq) will be kept for a few release cycles.

CC: Maximilian Luz <luzmaximilian@gmail.com>
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
---
 drivers/platform/surface/aggregator/ssh_packet_layer.c  | 2 +-
 drivers/platform/surface/aggregator/ssh_request_layer.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/platform/surface/aggregator/ssh_packet_layer.c b/drivers/platform/surface/aggregator/ssh_packet_layer.c
index 6081b0146d5f..3dd22856570f 100644
--- a/drivers/platform/surface/aggregator/ssh_packet_layer.c
+++ b/drivers/platform/surface/aggregator/ssh_packet_layer.c
@@ -671,7 +671,7 @@ static void ssh_ptl_timeout_reaper_mod(struct ssh_ptl *ptl, ktime_t now,
 	/* Re-adjust / schedule reaper only if it is above resolution delta. */
 	if (ktime_before(aexp, ptl->rtx_timeout.expires)) {
 		ptl->rtx_timeout.expires = expires;
-		mod_delayed_work(system_wq, &ptl->rtx_timeout.reaper, delta);
+		mod_delayed_work(system_percpu_wq, &ptl->rtx_timeout.reaper, delta);
 	}
 
 	spin_unlock(&ptl->rtx_timeout.lock);
diff --git a/drivers/platform/surface/aggregator/ssh_request_layer.c b/drivers/platform/surface/aggregator/ssh_request_layer.c
index 879ca9ee7ff6..a356e4956562 100644
--- a/drivers/platform/surface/aggregator/ssh_request_layer.c
+++ b/drivers/platform/surface/aggregator/ssh_request_layer.c
@@ -434,7 +434,7 @@ static void ssh_rtl_timeout_reaper_mod(struct ssh_rtl *rtl, ktime_t now,
 	/* Re-adjust / schedule reaper only if it is above resolution delta. */
 	if (ktime_before(aexp, rtl->rtx_timeout.expires)) {
 		rtl->rtx_timeout.expires = expires;
-		mod_delayed_work(system_wq, &rtl->rtx_timeout.reaper, delta);
+		mod_delayed_work(system_percpu_wq, &rtl->rtx_timeout.reaper, delta);
 	}
 
 	spin_unlock(&rtl->rtx_timeout.lock);
-- 
2.51.1

[PATCH 2/2] platform: x86: replace use of system_wq with system_percpu_wq

[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.

This patch continues the effort to refactor worqueue APIs, which has begun
with the change introducing new workqueues and a new alloc_workqueue flag:

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

Replace system_wq with system_percpu_wq, keeping the same behavior.
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/platform/x86/gpd-pocket-fan.c                      | 4 ++--
 drivers/platform/x86/x86-android-tablets/vexia_atla10_ec.c | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/platform/x86/gpd-pocket-fan.c b/drivers/platform/x86/gpd-pocket-fan.c
index 7a20f68ae206..c9236738f896 100644
--- a/drivers/platform/x86/gpd-pocket-fan.c
+++ b/drivers/platform/x86/gpd-pocket-fan.c
@@ -112,14 +112,14 @@ static void gpd_pocket_fan_worker(struct work_struct *work)
 	gpd_pocket_fan_set_speed(fan, speed);
 
 	/* When mostly idle (low temp/speed), slow down the poll interval. */
-	queue_delayed_work(system_wq, &fan->work,
+	queue_delayed_work(system_percpu_wq, &fan->work,
 			   msecs_to_jiffies(4000 / (speed + 1)));
 }
 
 static void gpd_pocket_fan_force_update(struct gpd_pocket_fan_data *fan)
 {
 	fan->last_speed = -1;
-	mod_delayed_work(system_wq, &fan->work, 0);
+	mod_delayed_work(system_percpu_wq, &fan->work, 0);
 }
 
 static int gpd_pocket_fan_probe(struct platform_device *pdev)
diff --git a/drivers/platform/x86/x86-android-tablets/vexia_atla10_ec.c b/drivers/platform/x86/x86-android-tablets/vexia_atla10_ec.c
index 2f8cd8d9e0ab..ebbedfe5f4e8 100644
--- a/drivers/platform/x86/x86-android-tablets/vexia_atla10_ec.c
+++ b/drivers/platform/x86/x86-android-tablets/vexia_atla10_ec.c
@@ -183,7 +183,7 @@ static void atla10_ec_external_power_changed(struct power_supply *psy)
 	struct atla10_ec_data *data = power_supply_get_drvdata(psy);
 
 	/* After charger plug in/out wait 0.5s for things to stabilize */
-	mod_delayed_work(system_wq, &data->work, HZ / 2);
+	mod_delayed_work(system_percpu_wq, &data->work, HZ / 2);
 }
 
 static const enum power_supply_property atla10_ec_psy_props[] = {
-- 
2.51.1