[patch 15/16] cpufreq: intel-pstate: Use del_timer_sync in intel_pstate_cpu_exit()

[patch 15/16] cpufreq: intel-pstate: Use del_timer_sync in intel_pstate_cpu_exit()

[Thomas Gleixner]

[-- Attachment #1: cpufreq-intel-pstate-use-del_timer_sync.patch --]
[-- Type: text/plain, Size: 1152 bytes --]

We are about to free the data structure. Make sure no timer callback
is running. I might be paranoid, but the ->exit callback can be
invoked from so many places, that it is not entirely clear whether
del_timer is always called on the cpu on which it is enqueued.

While looking through the call sites I noticed, that
cpufreq_init_policy() can fail and invoke cpufreq_driver->exit() but
it does not return the failure and the callsite happily proceeds.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: cpufreq <cpufreq@vger.kernel.org>
Cc: pm <linux-pm@vger.kernel.org>
---

 drivers/cpufreq/intel_pstate.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: tip/drivers/cpufreq/intel_pstate.c
===================================================================
--- tip.orig/drivers/cpufreq/intel_pstate.c
+++ tip/drivers/cpufreq/intel_pstate.c
@@ -777,7 +777,7 @@ static int intel_pstate_cpu_exit(struct
 {
 	int cpu = policy->cpu;
 
-	del_timer(&all_cpu_data[cpu]->timer);
+	del_timer_sync(&all_cpu_data[cpu]->timer);
 	kfree(all_cpu_data[cpu]);
 	all_cpu_data[cpu] = NULL;
 	return 0;

Re: [patch 15/16] cpufreq: intel-pstate: Use del_timer_sync in intel_pstate_cpu_exit()

[Rafael J. Wysocki]

On Sunday, March 23, 2014 03:09:32 PM Thomas Gleixner wrote:
> We are about to free the data structure. Make sure no timer callback
> is running. I might be paranoid, but the ->exit callback can be
> invoked from so many places, that it is not entirely clear whether
> del_timer is always called on the cpu on which it is enqueued.
> 
> While looking through the call sites I noticed, that
> cpufreq_init_policy() can fail and invoke cpufreq_driver->exit() but
> it does not return the failure and the callsite happily proceeds.
> 
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: cpufreq <cpufreq@vger.kernel.org>
> Cc: pm <linux-pm@vger.kernel.org>

Dirk?

> ---
> 
>  drivers/cpufreq/intel_pstate.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: tip/drivers/cpufreq/intel_pstate.c
> ===================================================================
> --- tip.orig/drivers/cpufreq/intel_pstate.c
> +++ tip/drivers/cpufreq/intel_pstate.c
> @@ -777,7 +777,7 @@ static int intel_pstate_cpu_exit(struct
>  {
>  	int cpu = policy->cpu;
>  
> -	del_timer(&all_cpu_data[cpu]->timer);
> +	del_timer_sync(&all_cpu_data[cpu]->timer);
>  	kfree(all_cpu_data[cpu]);
>  	all_cpu_data[cpu] = NULL;
>  	return 0;
> 
> 

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [patch 15/16] cpufreq: intel-pstate: Use del_timer_sync in intel_pstate_cpu_exit()

[Dirk Brandewie]

Hi Thomas,
On 03/23/2014 06:56 PM, Rafael J. Wysocki wrote:
> On Sunday, March 23, 2014 03:09:32 PM Thomas Gleixner wrote:
>> We are about to free the data structure. Make sure no timer callback
>> is running. I might be paranoid, but the ->exit callback can be
>> invoked from so many places, that it is not entirely clear whether
>> del_timer is always called on the cpu on which it is enqueued.
>>
>> While looking through the call sites I noticed, that
>> cpufreq_init_policy() can fail and invoke cpufreq_driver->exit() but
>> it does not return the failure and the callsite happily proceeds.
>>

The call to del_timer() has been moved to a new callback in material
in Rafaels pull request for v3.15.

I will send a patch adding this change to the v3.15 material.

--Dirk
>> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
>> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
>> Cc: cpufreq <cpufreq@vger.kernel.org>
>> Cc: pm <linux-pm@vger.kernel.org>
>
> Dirk?
>
>> ---
>>
>>   drivers/cpufreq/intel_pstate.c |    2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> Index: tip/drivers/cpufreq/intel_pstate.c
>> ===================================================================
>> --- tip.orig/drivers/cpufreq/intel_pstate.c
>> +++ tip/drivers/cpufreq/intel_pstate.c
>> @@ -777,7 +777,7 @@ static int intel_pstate_cpu_exit(struct
>>   {
>>   	int cpu = policy->cpu;
>>
>> -	del_timer(&all_cpu_data[cpu]->timer);
>> +	del_timer_sync(&all_cpu_data[cpu]->timer);
>>   	kfree(all_cpu_data[cpu]);
>>   	all_cpu_data[cpu] = NULL;
>>   	return 0;
>>
>>
>

Re: [patch 15/16] cpufreq: intel-pstate: Use del_timer_sync in intel_pstate_cpu_exit()

[Viresh Kumar]

On Sun, Mar 23, 2014 at 8:39 PM, Thomas Gleixner <tglx@linutronix.de> wrote:
> We are about to free the data structure. Make sure no timer callback
> is running. I might be paranoid, but the ->exit callback can be
> invoked from so many places, that it is not entirely clear whether
> del_timer is always called on the cpu on which it is enqueued.

In normal cases it will be called only when CPU goes offline and I
hope that isn't called on the dying CPU.

> While looking through the call sites I noticed, that
> cpufreq_init_policy() can fail and invoke cpufreq_driver->exit() but
> it does not return the failure and the callsite happily proceeds.

Thanks for reporting this, will get it fixed.

> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: cpufreq <cpufreq@vger.kernel.org>
> Cc: pm <linux-pm@vger.kernel.org>
> ---
>
>  drivers/cpufreq/intel_pstate.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> Index: tip/drivers/cpufreq/intel_pstate.c
> ===================================================================
> --- tip.orig/drivers/cpufreq/intel_pstate.c
> +++ tip/drivers/cpufreq/intel_pstate.c
> @@ -777,7 +777,7 @@ static int intel_pstate_cpu_exit(struct
>  {
>         int cpu = policy->cpu;
>
> -       del_timer(&all_cpu_data[cpu]->timer);
> +       del_timer_sync(&all_cpu_data[cpu]->timer);

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>