Re: [PATCH v7 04/23] PM: EM: Refactor em_pd_get_efficient_state() to be more flexible

[Lukasz Luba <lukasz.luba@arm.com>]

On 1/17/24 12:45, Hongyan Xia wrote:
> On 17/01/2024 09:56, Lukasz Luba wrote:
>> The Energy Model (EM) is going to support runtime modification. There
>> are going to be 2 EM tables which store information. This patch aims
>> to prepare the code to be generic and use one of the tables. The function
>> will no longer get a pointer to 'struct em_perf_domain' (the EM) but
>> instead a pointer to 'struct em_perf_state' (which is one of the EM's
>> tables).
>>
>> Prepare em_pd_get_efficient_state() for the upcoming changes and
>> make it possible to be re-used. Return an index for the best performance
>> state for a given EM table. The function arguments that are introduced
>> should allow to work on different performance state arrays. The caller of
>> em_pd_get_efficient_state() should be able to use the index either
>> on the default or the modifiable EM table.
>>
>> Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>
>> Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
>> ---
>>   include/linux/energy_model.h | 30 +++++++++++++++++-------------
>>   1 file changed, 17 insertions(+), 13 deletions(-)
>>
>> diff --git a/include/linux/energy_model.h b/include/linux/energy_model.h
>> index c19e7effe764..b01277b17946 100644
>> --- a/include/linux/energy_model.h
>> +++ b/include/linux/energy_model.h
>> @@ -175,33 +175,35 @@ void em_dev_unregister_perf_domain(struct device 
>> *dev);
>>   /**
>>    * em_pd_get_efficient_state() - Get an efficient performance state 
>> from the EM
>> - * @pd   : Performance domain for which we want an efficient frequency
>> - * @freq : Frequency to map with the EM
>> + * @table:        List of performance states, in ascending order
>> + * @nr_perf_states:    Number of performance states
>> + * @freq:        Frequency to map with the EM
>> + * @pd_flags:        Performance Domain flags
>>    *
>>    * It is called from the scheduler code quite frequently and as a 
>> consequence
>>    * doesn't implement any check.
>>    *
>> - * Return: An efficient performance state, high enough to meet @freq
>> + * Return: An efficient performance state id, high enough to meet @freq
>>    * requirement.
>>    */
>> -static inline
>> -struct em_perf_state *em_pd_get_efficient_state(struct em_perf_domain 
>> *pd,
>> -                        unsigned long freq)
>> +static inline int
>> +em_pd_get_efficient_state(struct em_perf_state *table, int 
>> nr_perf_states,
>> +              unsigned long freq, unsigned long pd_flags)
>>   {
>>       struct em_perf_state *ps;
>>       int i;
>> -    for (i = 0; i < pd->nr_perf_states; i++) {
>> -        ps = &pd->table[i];
>> +    for (i = 0; i < nr_perf_states; i++) {
>> +        ps = &table[i];
>>           if (ps->frequency >= freq) {
>> -            if (pd->flags & EM_PERF_DOMAIN_SKIP_INEFFICIENCIES &&
>> +            if (pd_flags & EM_PERF_DOMAIN_SKIP_INEFFICIENCIES &&
>>                   ps->flags & EM_PERF_STATE_INEFFICIENT)
>>                   continue;
>> -            break;
>> +            return i;
>>           }
>>       }
>> -    return ps;
>> +    return nr_perf_states - 1;
>>   }
>>   /**
>> @@ -226,7 +228,7 @@ static inline unsigned long em_cpu_energy(struct 
>> em_perf_domain *pd,
>>   {
>>       unsigned long freq, ref_freq, scale_cpu;
>>       struct em_perf_state *ps;
>> -    int cpu;
>> +    int cpu, i;
>>       if (!sum_util)
>>           return 0;
>> @@ -251,7 +253,9 @@ static inline unsigned long em_cpu_energy(struct 
>> em_perf_domain *pd,
>>        * Find the lowest performance state of the Energy Model above the
>>        * requested frequency.
>>        */
>> -    ps = em_pd_get_efficient_state(pd, freq);
>> +    i = em_pd_get_efficient_state(pd->table, pd->nr_perf_states, freq,
>> +                      pd->flags);
>> +    ps = &pd->table[i];
>>       /*
>>        * The capacity of a CPU in the domain at the performance state 
>> (ps)
> 
> Reviewed-by: Hongyan Xia <hongyan.xia@arm.com>
> 

Thank you Hongyan for the reviews!
I might address your NIT comment for the patch 2/24 when
I do the re-basing and sending the v8 (if there is a need).

Regards,
Lukasz