When unregistering a thermal zone device, update the cooling device in
case the cooling device is activated by the current thermal zone.
This fixes a problem that the frequency of ACPI processors are still
limited after unloading ACPI thermal driver while ACPI passive cooling
is activated.
Signed-off-by: Zhang Rui <[email protected]>
---
drivers/thermal/thermal_core.c | 26 +++++++++++++++++++++++++-
1 file changed, 25 insertions(+), 1 deletion(-)
diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
index cfd4c1afeae7..9f120e3c9bf0 100644
--- a/drivers/thermal/thermal_core.c
+++ b/drivers/thermal/thermal_core.c
@@ -1477,6 +1477,7 @@ void thermal_zone_device_unregister(struct thermal_zone_device *tz)
const struct thermal_zone_params *tzp;
struct thermal_cooling_device *cdev;
struct thermal_zone_device *pos = NULL;
+ struct thermal_instance *ti;
if (!tz)
return;
@@ -1497,9 +1498,22 @@ void thermal_zone_device_unregister(struct thermal_zone_device *tz)
/* Unbind all cdevs associated with 'this' thermal zone */
list_for_each_entry(cdev, &thermal_cdev_list, node) {
+ mutex_lock(&tz->lock);
+ list_for_each_entry(ti, &tz->thermal_instances, tz_node) {
+ if (ti->cdev == cdev)
+ break;
+ }
+
+ /* The cooling device is not used by current thermal zone */
+ if (ti->cdev != cdev) {
+ mutex_unlock(&tz->lock);
+ continue;
+ }
+ mutex_unlock(&tz->lock);
+
if (tz->ops->unbind) {
tz->ops->unbind(tz, cdev);
- continue;
+ goto deactivate_cdev;
}
if (!tzp || !tzp->tbp)
@@ -1511,6 +1525,16 @@ void thermal_zone_device_unregister(struct thermal_zone_device *tz)
tzp->tbp[i].cdev = NULL;
}
}
+
+deactivate_cdev:
+ /*
+ * The thermal instances for current thermal zone has been
+ * removed. Update the cooling device in case it is activated
+ * by current thermal zone device.
+ */
+ mutex_lock(&cdev->lock);
+ __thermal_cdev_update(cdev);
+ mutex_unlock(&cdev->lock);
}
mutex_unlock(&thermal_list_lock);
--
2.25.1
On Tue, Mar 21, 2023 at 4:46 PM Zhang Rui <[email protected]> wrote:
>
> When unregistering a thermal zone device, update the cooling device in
> case the cooling device is activated by the current thermal zone.
I think that all cooling devices bound to the thermal zone being
removed need to be updated at this point? Which is what the patch
really does IIUC.
It also avoids unbinding cooling devices that have not been bound to that zone.
> This fixes a problem that the frequency of ACPI processors are still
> limited after unloading ACPI thermal driver while ACPI passive cooling
> is activated.
>
Cc: stable@vger ?
> Signed-off-by: Zhang Rui <[email protected]>
> ---
> drivers/thermal/thermal_core.c | 26 +++++++++++++++++++++++++-
> 1 file changed, 25 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
> index cfd4c1afeae7..9f120e3c9bf0 100644
> --- a/drivers/thermal/thermal_core.c
> +++ b/drivers/thermal/thermal_core.c
> @@ -1477,6 +1477,7 @@ void thermal_zone_device_unregister(struct thermal_zone_device *tz)
> const struct thermal_zone_params *tzp;
> struct thermal_cooling_device *cdev;
> struct thermal_zone_device *pos = NULL;
> + struct thermal_instance *ti;
>
> if (!tz)
> return;
> @@ -1497,9 +1498,22 @@ void thermal_zone_device_unregister(struct thermal_zone_device *tz)
>
I would rearrange the code as follows.
> /* Unbind all cdevs associated with 'this' thermal zone */
> list_for_each_entry(cdev, &thermal_cdev_list, node) {
struct thermal_instance *ti;
> + mutex_lock(&tz->lock);
> + list_for_each_entry(ti, &tz->thermal_instances, tz_node) {
if (ti->cdev == cdev) {
mutex_unlock(&tz->lock);
goto unbind;
}
}
/* The cooling device is not used by this thermal zone. */
mutex_unlock(&tz->lock);
continue;
unbind:
> if (tz->ops->unbind) {
> tz->ops->unbind(tz, cdev);
/*
* The thermal instance for current
thermal zone has been
* removed, so update the cooling device
in case it has been
* activated by the thermal zone device going away.
*/
mutex_lock(&cdev->lock);
__thermal_cdev_update(cdev);
mutex_unlock(&cdev->lock);
continue;
> }
>
> if (!tzp || !tzp->tbp)
On Tue, 2023-03-21 at 20:43 +0100, Rafael J. Wysocki wrote:
> On Tue, Mar 21, 2023 at 4:46 PM Zhang Rui <[email protected]>
> wrote:
> > When unregistering a thermal zone device, update the cooling device
> > in
> > case the cooling device is activated by the current thermal zone.
>
> I think that all cooling devices bound to the thermal zone being
> removed need to be updated at this point? Which is what the patch
> really does IIUC.
yes, that is what I want to say.
>
> It also avoids unbinding cooling devices that have not been bound to
> that zone.
>
The thermal zone device driver' .unbind() callback should be able to
handle this. But still, this is a valid improvement.
> > This fixes a problem that the frequency of ACPI processors are
> > still
> > limited after unloading ACPI thermal driver while ACPI passive
> > cooling
> > is activated.
> >
>
> Cc: stable@vger ?
yeah, will add it.
>
> > Signed-off-by: Zhang Rui <[email protected]>
> > ---
> > drivers/thermal/thermal_core.c | 26 +++++++++++++++++++++++++-
> > 1 file changed, 25 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/thermal/thermal_core.c
> > b/drivers/thermal/thermal_core.c
> > index cfd4c1afeae7..9f120e3c9bf0 100644
> > --- a/drivers/thermal/thermal_core.c
> > +++ b/drivers/thermal/thermal_core.c
> > @@ -1477,6 +1477,7 @@ void thermal_zone_device_unregister(struct
> > thermal_zone_device *tz)
> > const struct thermal_zone_params *tzp;
> > struct thermal_cooling_device *cdev;
> > struct thermal_zone_device *pos = NULL;
> > + struct thermal_instance *ti;
> >
> > if (!tz)
> > return;
> > @@ -1497,9 +1498,22 @@ void thermal_zone_device_unregister(struct
> > thermal_zone_device *tz)
> >
>
> I would rearrange the code as follows.
>
> > /* Unbind all cdevs associated with 'this' thermal zone */
> > list_for_each_entry(cdev, &thermal_cdev_list, node) {
> struct thermal_instance *ti;
>
> > + mutex_lock(&tz->lock);
> > + list_for_each_entry(ti, &tz->thermal_instances,
> > tz_node) {
>
> if (ti->cdev == cdev) {
> mutex_unlock(&tz->lock);
> goto unbind;
> }
> }
> /* The cooling device is not used by this thermal
> zone. */
> mutex_unlock(&tz->lock);
> continue;
>
> unbind:
>
> > if (tz->ops->unbind) {
> > tz->ops->unbind(tz, cdev);
>
> /*
> * The thermal instance for current
> thermal zone has been
> * removed, so update the cooling device
> in case it has been
> * activated by the thermal zone device
> going away.
> */
> mutex_lock(&cdev->lock);
> __thermal_cdev_update(cdev);
> mutex_unlock(&cdev->lock);
>
> continue;
> > }
IMO, updating the cooling device is required, no matter the cooling
device is bound by .bind() callback or by static thermal_bind_params
structure.
Actually, I think struct thermal_bind_params is obsoleted and nobody is
using it now. I will write a cleanup patch to remove it after this one.
thanks,
rui
> >
> > if (!tzp || !tzp->tbp)
Hi Zhang,
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Zhang-Rui/thermal-core-update-cooling-device-during-thermal-zone-unregistration/20230321-234754
base: https://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git thermal
patch link: https://lore.kernel.org/r/20230321154627.17787-1-rui.zhang%40intel.com
patch subject: [PATCH] thermal/core: update cooling device during thermal zone unregistration
config: i386-randconfig-m021 (https://download.01.org/0day-ci/archive/20230322/[email protected]/config)
compiler: gcc-11 (Debian 11.3.0-8) 11.3.0
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>
| Reported-by: Dan Carpenter <[email protected]>
| Link: https://lore.kernel.org/r/[email protected]/
smatch warnings:
drivers/thermal/thermal_core.c:1436 thermal_zone_device_unregister() warn: iterator used outside loop: 'ti'
vim +/ti +1436 drivers/thermal/thermal_core.c
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1402 void thermal_zone_device_unregister(struct thermal_zone_device *tz)
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1403 {
a5f785ce608caf drivers/thermal/thermal_core.c Dmitry Osipenko 2020-08-18 1404 int i, tz_id;
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1405 const struct thermal_zone_params *tzp;
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1406 struct thermal_cooling_device *cdev;
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1407 struct thermal_zone_device *pos = NULL;
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1408 struct thermal_instance *ti;
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1409
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1410 if (!tz)
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1411 return;
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1412
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1413 tzp = tz->tzp;
a5f785ce608caf drivers/thermal/thermal_core.c Dmitry Osipenko 2020-08-18 1414 tz_id = tz->id;
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1415
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1416 mutex_lock(&thermal_list_lock);
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1417 list_for_each_entry(pos, &thermal_tz_list, node)
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1418 if (pos == tz)
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1419 break;
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1420 if (pos != tz) {
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1421 /* thermal zone device not found */
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1422 mutex_unlock(&thermal_list_lock);
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1423 return;
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1424 }
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1425 list_del(&tz->node);
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1426
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1427 /* Unbind all cdevs associated with 'this' thermal zone */
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1428 list_for_each_entry(cdev, &thermal_cdev_list, node) {
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1429 mutex_lock(&tz->lock);
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1430 list_for_each_entry(ti, &tz->thermal_instances, tz_node) {
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1431 if (ti->cdev == cdev)
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1432 break;
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1433 }
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1434
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1435 /* The cooling device is not used by current thermal zone */
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 @1436 if (ti->cdev != cdev) {
Here we are testing to see if the loop exited by hitting the break. If
the condition is != then "ti" is not a valid pointer and it's kind of an
out of bounds access. I used to fix these with:
- if (ti->cdev != cdev) {
+ if (list_entry_is_head(ti, &tz->thermal_instances, tz_node)) {
But these days we just prefer using a found variable:
found = false;
list_for_each_entry(ti, &tz->thermal_instances, tz_node) {
if (ti->cdev == cdev) {
found = true;
break;
}
}
if (!found) {
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1437 mutex_unlock(&tz->lock);
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1438 continue;
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1439 }
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1440 mutex_unlock(&tz->lock);
55a1117eebc62b drivers/thermal/thermal_core.c Zhang Rui 2023-03-21 1441
7e8ee1e9d7561f drivers/thermal/thermal_sys.c Durgadoss R 2012-09-18 1442 if (tz->ops->unbind) {
203d3d4aa48233 drivers/thermal/thermal.c Zhang Rui 2008-01-17 1443 tz->ops->unbind(tz, cdev);
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests