Skip to contentSkip to conversation

thermal: core: Do not fail cdev registration because of invalid initial state

PATCH Superseded Rafael J. Wysocki 2 participants 3 messages Jun 6, 2024

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 thermal: core: Do not fail cdev registration because of invalid initial state

From: Rafael J. Wysocki <[email protected]>

It is reported that commit 31a0fa0019b0 ("thermal/debugfs: Pass cooling device state to thermal_debug_cdev_add()") causes the ACPI fan driver to fail probing on some systems which turns out to be due to the _FST control method returning an invalid value until _FSL is first evaluated for the given fan. If this happens, the .get_cur_state() cooling device callback returns an error and __thermal_cooling_device_register() fails as uses that callback after commit 31a0fa0019b0.

Arguably, _FST should not return an invalid value even if it is evaluated before _FSL, so this may be regarded as a platform firmware issue, but at the same time it is not a good enough reason for failing the cooling device registration where the initial cooling device state is only needed to initialize a thermal debug facility.

Accordingly, modify __thermal_cooling_device_register() to avoid calling thermal_debug_cdev_add() instead of returning an error if the initial .get_cur_state() callback invocation fails.

Fixes 31a0fa0019b0 ("thermal/debugfs: Pass cooling device state to thermal_debug_cdev_add()")

---

v1 -> v2:
* Instead of making the thermal debug code effectively ignore the invalid initial cooling device state, simply don't register thermal debugfs for a cooling device if its initial state returned by the driver's .get_cur_state() is invalid (Daniel).

Laura, please test this one even though I don't see why it wouldn't work for you if the v1 did.

---
 drivers/thermal/thermal_core.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/thermal/thermal_core.c +3 -2
Index: linux-pm/drivers/thermal/thermal_core.c
===================================================================
@@ -1001,7 +1001,7 @@ __thermal_cooling_device_register(struct
 
 	ret = cdev->ops->get_cur_state(cdev, &current_state);
 	if (ret)
-		goto out_cdev_type;
+		current_state = ULONG_MAX;
 
 	thermal_cooling_device_setup_sysfs(cdev);
 
@@ -1016,7 +1016,8 @@ __thermal_cooling_device_register(struct
 		return ERR_PTR(ret);
 	}
 
-	thermal_debug_cdev_add(cdev, current_state);
+	if (current_state <= cdev->max_state)
+		thermal_debug_cdev_add(cdev, current_state);
 
 	/* Add 'this' new cdev to the global cdev list */
 	mutex_lock(&thermal_list_lock);
#1 lore raw
Quoted 44 lines · Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

It is reported that commit 31a0fa0019b0 ("thermal/debugfs: Pass cooling device state to thermal_debug_cdev_add()") causes the ACPI fan driver to fail probing on some systems which turns out to be due to the _FST control method returning an invalid value until _FSL is first evaluated for the given fan. If this happens, the .get_cur_state() cooling device callback returns an error and __thermal_cooling_device_register() fails as uses that callback after commit 31a0fa0019b0.

Arguably, _FST should not return an invalid value even if it is evaluated before _FSL, so this may be regarded as a platform firmware issue, but at the same time it is not a good enough reason for failing the cooling device registration where the initial cooling device state is only needed to initialize a thermal debug facility.

Accordingly, modify __thermal_cooling_device_register() to avoid calling thermal_debug_cdev_add() instead of returning an error if the initial .get_cur_state() callback invocation fails.

Fixes 31a0fa0019b0 ("thermal/debugfs: Pass cooling device state to thermal_debug_cdev_add()")

---

v1 -> v2:
* Instead of making the thermal debug code effectively ignore the invalid initial cooling device state, simply don't register thermal debugfs for a cooling device if its initial state returned by the driver's .get_cur_state() is invalid (Daniel).

Laura, please test this one even though I don't see why it wouldn't work for you if the v1 did.

---
  drivers/thermal/thermal_core.c |    5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)

Index: linux-pm/drivers/thermal/thermal_core.c
===================================================================

drivers/thermal/thermal_core.c __thermal_cooling_device_register(struct · line 1004

Quoted 5 lines
drivers/thermal/thermal_core.c +1 -1
@@ -1001,7 +1001,7 @@ __thermal_cooling_device_register(struct
 
  	ret = cdev->ops->get_cur_state(cdev, &current_state);
  	if (ret)
-		goto out_cdev_type;
+		current_state = ULONG_MAX;

Why not move the section ? So we end up below.

drivers/thermal/thermal_core.c __thermal_cooling_device_register(struct · line 1020

Quoted 10 lines
drivers/thermal/thermal_core.c +2 -1
  	thermal_cooling_device_setup_sysfs(cdev);
 
@@ -1016,7 +1016,8 @@ __thermal_cooling_device_register(struct
  		return ERR_PTR(ret);
  	}
 
-	thermal_debug_cdev_add(cdev, current_state);
+	if (current_state <= cdev->max_state)
+		thermal_debug_cdev_add(cdev, current_state);
	ret = cdev->ops->get_cur_state(cdev, &current_state);
	if (!ret)
		thermal_debug_cdev_add(cdev, current_state);

Additionally a comment here to explain why get_cur_state can fail and telling it is up to the driver to fix its routine?

Quoted 5 lines
drivers/thermal/thermal_core.c +0 -0
  	/* Add 'this' new cdev to the global cdev list */
  	mutex_lock(&thermal_list_lock);
 
Signature · 5 lines
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
#2 lore raw
Quoted 46 lines · CEST Daniel Lezcano
Quoted 45 lines · Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

It is reported that commit 31a0fa0019b0 ("thermal/debugfs: Pass cooling device state to thermal_debug_cdev_add()") causes the ACPI fan driver to fail probing on some systems which turns out to be due to the _FST control method returning an invalid value until _FSL is first evaluated for the given fan. If this happens, the .get_cur_state() cooling device callback returns an error and __thermal_cooling_device_register() fails as uses that callback after commit 31a0fa0019b0.

Arguably, _FST should not return an invalid value even if it is evaluated before _FSL, so this may be regarded as a platform firmware issue, but at the same time it is not a good enough reason for failing the cooling device registration where the initial cooling device state is only needed to initialize a thermal debug facility.

Accordingly, modify __thermal_cooling_device_register() to avoid calling thermal_debug_cdev_add() instead of returning an error if the initial .get_cur_state() callback invocation fails.

Fixes 31a0fa0019b0 ("thermal/debugfs: Pass cooling device state to thermal_debug_cdev_add()")

---

v1 -> v2:
* Instead of making the thermal debug code effectively ignore the invalid initial cooling device state, simply don't register thermal debugfs for a cooling device if its initial state returned by the driver's .get_cur_state() is invalid (Daniel).

Laura, please test this one even though I don't see why it wouldn't work for you if the v1 did.

---
  drivers/thermal/thermal_core.c |    5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)

Index: linux-pm/drivers/thermal/thermal_core.c
===================================================================

drivers/thermal/thermal_core.c +0 -0
@@ -1001,7 +1001,7 @@ __thermal_cooling_device_register(struct
 
Quoted 6 lines

drivers/thermal/thermal_core.c __thermal_cooling_device_register(struct · line 1004

Quoted 4 lines
drivers/thermal/thermal_core.c +1 -1
  	ret = cdev->ops->get_cur_state(cdev, &current_state);
  	if (ret)
-		goto out_cdev_type;
+		current_state = ULONG_MAX;

Why not move the section ? So we end up below.

Because it's fewer changes in this patch and it doesn't matter.

Quoted 7 lines
Quoted 7 lines
drivers/thermal/thermal_core.c +0 -0
  	thermal_cooling_device_setup_sysfs(cdev);
 
@@ -1016,7 +1016,8 @@ __thermal_cooling_device_register(struct
  		return ERR_PTR(ret);
  	}
 
Quoted 7 lines

drivers/thermal/thermal_core.c __thermal_cooling_device_register(struct · line 1020

Quoted 3 lines
drivers/thermal/thermal_core.c +2 -1
-	thermal_debug_cdev_add(cdev, current_state);
+	if (current_state <= cdev->max_state)
+		thermal_debug_cdev_add(cdev, current_state);
	ret = cdev->ops->get_cur_state(cdev, &current_state);
	if (!ret)
		thermal_debug_cdev_add(cdev, current_state);

So the current_state <= cdev->max_state is actually useful regardless and consistent with the analogous check in thermal_cooling_device_update().

Quoted 2 lines

Additionally a comment here to explain why get_cur_state can fail and telling it is up to the driver to fix its routine?

I've added a comment in v3.

Quoted 8 lines
Quoted 6 lines
drivers/thermal/thermal_core.c +0 -0
  	/* Add 'this' new cdev to the global cdev list */
  	mutex_lock(&thermal_list_lock);