On 3/13/20 10:15 AM, Quentin Perret wrote:
> On Monday 09 Mar 2020 at 13:41:15 (+0000), Lukasz Luba wrote:
>> diff --git a/drivers/cpufreq/cpufreq-dt.c b/drivers/cpufreq/cpufreq-dt.c
>> index d2b5f062a07b..676b56424886 100644
>> --- a/drivers/cpufreq/cpufreq-dt.c
>> +++ b/drivers/cpufreq/cpufreq-dt.c
>> @@ -275,7 +275,9 @@ static int cpufreq_init(struct cpufreq_policy *policy)
>>   	policy->cpuinfo.transition_latency = transition_latency;
>>   	policy->dvfs_possible_from_any_cpu = true;
>> 
>> -	dev_pm_opp_of_register_em(policy->cpus);
>> +	ret = dev_pm_opp_of_register_em(cpu_dev, policy->cpus);
>> +	if (ret)
>> +		dev_dbg(cpu_dev, "Couldn't register Energy Model %d\n", ret);
>> 
>>   	return 0;
> 
> Ah, that answers my comment on patch 01. You're adding the error
> messages here.
> 
> Isn't this more boilerplate for the drivers ? All they do is print the
> same debug message. Maybe just move it inside dev_pm_opp_of_register_em
> directly ?
> 

Agree. I could add a 'fail' label in dev_pm_opp_of_register_em
something like:

--------------------------------->8----------------
dev_pm_opp_of_register_em
...
	nr_opp = dev_pm_opp_get_opp_count(dev);
	if (nr_opp <= 0) {
		ret = -EINVAL;
		goto fail;
	}

...
	ret = em_register_perf_domain()
	if (ret)
		goto fail;

	return 0;

fail:
	dev_dbg(cpu_dev, "Couldn't register Energy Model %d\n", ret);
	return ret
------------------------8<-----------------

Makes more sense, agree, thank you for your suggestion.

I will remove this
   if (ret)
	dev_dbg(cpu_dev, "Couldn't register Energy Model %d\n", ret);
from the cpufreq drivers.


Regards,
Lukasz