Re: [PATCH v2 5/5] cpufreq: Catch double invocations of cpufreq_freq_transition_begin/end

["Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>]

On 04/29/2014 01:34 PM, Viresh Kumar wrote:
> On 29 April 2014 13:05, Srivatsa S. Bhat
> <srivatsa.bhat@linux.vnet.ibm.com> wrote:
>> On 04/29/2014 12:19 PM, Viresh Kumar wrote:
>>> +       WARN_ON(!(cpufreq_driver->flags & CPUFREQ_ASYNC_NOTIFICATION)
>>>                              && (current == policy->transition_task));
>>>
>>> which you already mentioned.
>>
>> Yeah, I think we should just go with this. I thought we needed lots of
>> if-conditions to do exclude these drivers (which would have made it ugly),
>> but as you pointed above, just this one would suffice.
> 
> Okay, I think we can do one more modification here:
> 
>>> +       WARN_ON(unlikely(!(cpufreq_driver->flags & CPUFREQ_ASYNC_NOTIFICATION)
>>>                              && (current == policy->transition_task)));
>

WARN_ON and friends already wrap their arguments within unlikely().
So we don't need to add it explicitly.
 
> 
>> Besides, the cpufreq core doesn't automatically invoke _begin() and
>> _end() for ASYNC_NOTIFICATION drivers. So that means the probability
>> that such drivers will hit this problem is extremely low, since the
>> driver alone is responsible for invoking _begin/_end and hence there
>> shouldn't be much of a conflict. So I think we should really just
>> skip ASYNC_NOTIFICATION drivers in this debug infrastructure.
> 
> The only way it can happen (I don't hope somebody would be so
> stupid to call begin twice from target() :)), is via transition notifiers,
> which in some case starting a new transition..

Hmm..

Regards,
Srivatsa S. Bhat