Re: [PATCH 6/6] PM / devfreq: Set the min_freq and max_freq of devfreq device

From: Chanwoo Choi
Date: Thu Nov 19 2015 - 05:37:34 EST


On 2015ë 11ì 19ì 19:10, MyungJoo Ham wrote:
>> After probing the devfreq device driver, the value of both min_freq and
>> max_freq are zero(0). So, this patch initializes the 'min_freq' and 'max_freq'
>> field of devfreq device by using the freq_table array.
>>
>> Signed-off-by: Chanwoo Choi <cw00.choi@xxxxxxxxxxx>
>> ---
>> drivers/devfreq/devfreq.c | 5 +++++
>> 1 file changed, 5 insertions(+)
>>
>> diff --git a/drivers/devfreq/devfreq.c b/drivers/devfreq/devfreq.c
>> index c292ceb7ff19..0b24ae7b7a48 100644
>> --- a/drivers/devfreq/devfreq.c
>> +++ b/drivers/devfreq/devfreq.c
>> @@ -121,6 +121,11 @@ static void devfreq_set_freq_table(struct devfreq *devfreq)
>> profile->freq_table[i] = freq;
>> }
>> rcu_read_unlock();
>> +
>> + mutex_lock(&devfreq->lock);
>> + devfreq->min_freq = profile->freq_table[0];
>> + devfreq->max_freq = profile->freq_table[profile->max_state - 1];
>> + mutex_unlock(&devfreq->lock);
>> }
>
> No, you should not do this.
>
> It is allowed to use devfreq without both OPP and freq_table
> assuming that the devfreq device may operate with very many
> frequencies so that practically, we can virtually give it
> any frequency numbers in a given range.
> (cases where profile->max_state is 0 and it is not an error)
>
> The value 0 is used for min/max_freq to declare
> that min/max_freq is deactivated. Therefore, it is not
> required to do so; they are not intended to show the hardware
> configuration as well.

This case consider the devfreq device using OPP because devfreq_set_freq_table()
get the number of OPP entry in OPP list before setting the min_freq/max_freq.
If the devfreq device don't use the OPP entry, devfreq_set_freq_table()
will return without any operation.

IMHO, when devfreq device uses the OPP table including the frequency,
min_freq/max_freq should show the correct value as CPUFREQ framework.

Regards,
Chanwoo CHoi



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/