Patchwork [v2,6/7] PCI/PM: warn about incomplete actions in ->runtime_suspend() callback

login
register
mail settings
Submitter Konstantin Khlebnikov
Date Feb. 4, 2013, 11:56 a.m.
Message ID <20130204115612.5569.49895.stgit@zurg>
Download mbox | patch
Permalink /patch/217883/
State Rejected
Headers show

Comments

Konstantin Khlebnikov - Feb. 4, 2013, 11:56 a.m.
Documentation/power/pci.txt says:
| It is expected that the device driver's pm->runtime_suspend() callback will
| not attempt to prepare the device for signaling wakeup or to put it into a
| low-power state.  The driver ought to leave these tasks to the PCI subsystem
| that has all of the information necessary to perform them.

After commit 42eca2302146fed51335b95128e949ee6f54478f
("PCI: Don't touch card regs after runtime suspend D3")
| If the driver takes care of state saving, don't touch any registers on it.
pci_pm_runtime_suspend() thinks if state has been saved by ->runtime_suspend()
that means device alredy prepared for wakeup and probably no longer accessible.

Thus driver must either do all actions or leave all these tasks to PCI subsystem.

Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
---
 drivers/pci/pci-driver.c |   15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)


--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rafael J. Wysocki - Feb. 4, 2013, 8:22 p.m.
On Monday, February 04, 2013 03:56:12 PM Konstantin Khlebnikov wrote:
> Documentation/power/pci.txt says:
> | It is expected that the device driver's pm->runtime_suspend() callback will
> | not attempt to prepare the device for signaling wakeup or to put it into a
> | low-power state.  The driver ought to leave these tasks to the PCI subsystem
> | that has all of the information necessary to perform them.
> 
> After commit 42eca2302146fed51335b95128e949ee6f54478f
> ("PCI: Don't touch card regs after runtime suspend D3")
> | If the driver takes care of state saving, don't touch any registers on it.
> pci_pm_runtime_suspend() thinks if state has been saved by ->runtime_suspend()
> that means device alredy prepared for wakeup and probably no longer accessible.
> 
> Thus driver must either do all actions or leave all these tasks to PCI subsystem.

I don't like this one, because it will generate noise for drivers that
legitimately use pci_save_state() in their runtime suspend callbacks
and know what they are doing.

Thanks,
Rafael


> Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
> Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> Cc: Bjorn Helgaas <bhelgaas@google.com>
> Cc: linux-pci@vger.kernel.org
> ---
>  drivers/pci/pci-driver.c |   15 +++++++++++----
>  1 file changed, 11 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
> index f9aa311..2b0ff9a 100644
> --- a/drivers/pci/pci-driver.c
> +++ b/drivers/pci/pci-driver.c
> @@ -1009,12 +1009,19 @@ static int pci_pm_runtime_suspend(struct device *dev)
>  		return 0;
>  	}
>  
> -	if (!pci_dev->state_saved) {
> -		pci_save_state(pci_dev);
> -		pci_finish_runtime_suspend(pci_dev);
> +	if (pci_dev->state_saved) {
> +		WARN_ONCE(pci_dev->current_state == prev,
> +			"PCI PM: Power state of device not changed by %pF\n",
> +			pm->runtime_suspend);
> +		WARN_ONCE(pci_dev_run_wake(pci_dev) &&
> +				!pci_dev->wakeup_prepared,
> +			"PCI PM: Waking of device not configured by %pF\n",
> +			pm->runtime_suspend);
> +		return 0;
>  	}
>  
> -	return 0;
> +	pci_save_state(pci_dev);
> +	return pci_finish_runtime_suspend(pci_dev);
>  }
>  
>  static int pci_pm_runtime_resume(struct device *dev)
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
Konstantin Khlebnikov - Feb. 4, 2013, 8:57 p.m.
Rafael J. Wysocki wrote:
> On Monday, February 04, 2013 03:56:12 PM Konstantin Khlebnikov wrote:
>> Documentation/power/pci.txt says:
>> | It is expected that the device driver's pm->runtime_suspend() callback will
>> | not attempt to prepare the device for signaling wakeup or to put it into a
>> | low-power state.  The driver ought to leave these tasks to the PCI subsystem
>> | that has all of the information necessary to perform them.
>>
>> After commit 42eca2302146fed51335b95128e949ee6f54478f
>> ("PCI: Don't touch card regs after runtime suspend D3")
>> | If the driver takes care of state saving, don't touch any registers on it.
>> pci_pm_runtime_suspend() thinks if state has been saved by ->runtime_suspend()
>> that means device alredy prepared for wakeup and probably no longer accessible.
>>
>> Thus driver must either do all actions or leave all these tasks to PCI subsystem.
>
> I don't like this one, because it will generate noise for drivers that
> legitimately use pci_save_state() in their runtime suspend callbacks
> and know what they are doing.

Ok, you're right. Both these warnings can be false-positive.
For example some of wifi cards easily might be weird enough for this.

>
> Thanks,
> Rafael
>
>
>> Signed-off-by: Konstantin Khlebnikov<khlebnikov@openvz.org>
>> Cc: Rafael J. Wysocki<rafael.j.wysocki@intel.com>
>> Cc: Bjorn Helgaas<bhelgaas@google.com>
>> Cc: linux-pci@vger.kernel.org
>> ---
>>   drivers/pci/pci-driver.c |   15 +++++++++++----
>>   1 file changed, 11 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
>> index f9aa311..2b0ff9a 100644
>> --- a/drivers/pci/pci-driver.c
>> +++ b/drivers/pci/pci-driver.c
>> @@ -1009,12 +1009,19 @@ static int pci_pm_runtime_suspend(struct device *dev)
>>   		return 0;
>>   	}
>>
>> -	if (!pci_dev->state_saved) {
>> -		pci_save_state(pci_dev);
>> -		pci_finish_runtime_suspend(pci_dev);
>> +	if (pci_dev->state_saved) {
>> +		WARN_ONCE(pci_dev->current_state == prev,
>> +			"PCI PM: Power state of device not changed by %pF\n",
>> +			pm->runtime_suspend);
>> +		WARN_ONCE(pci_dev_run_wake(pci_dev)&&
>> +				!pci_dev->wakeup_prepared,
>> +			"PCI PM: Waking of device not configured by %pF\n",
>> +			pm->runtime_suspend);
>> +		return 0;
>>   	}
>>
>> -	return 0;
>> +	pci_save_state(pci_dev);
>> +	return pci_finish_runtime_suspend(pci_dev);
>>   }
>>
>>   static int pci_pm_runtime_resume(struct device *dev)
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
index f9aa311..2b0ff9a 100644
--- a/drivers/pci/pci-driver.c
+++ b/drivers/pci/pci-driver.c
@@ -1009,12 +1009,19 @@  static int pci_pm_runtime_suspend(struct device *dev)
 		return 0;
 	}
 
-	if (!pci_dev->state_saved) {
-		pci_save_state(pci_dev);
-		pci_finish_runtime_suspend(pci_dev);
+	if (pci_dev->state_saved) {
+		WARN_ONCE(pci_dev->current_state == prev,
+			"PCI PM: Power state of device not changed by %pF\n",
+			pm->runtime_suspend);
+		WARN_ONCE(pci_dev_run_wake(pci_dev) &&
+				!pci_dev->wakeup_prepared,
+			"PCI PM: Waking of device not configured by %pF\n",
+			pm->runtime_suspend);
+		return 0;
 	}
 
-	return 0;
+	pci_save_state(pci_dev);
+	return pci_finish_runtime_suspend(pci_dev);
 }
 
 static int pci_pm_runtime_resume(struct device *dev)