Re: [RESEND][PATCH] cpufreq: stats: Add 'load_table' sysfs file to show accumulated data of CPU

[Chanwoo Choi <cw00.choi@samsung.com>]

On 06/11/2013 02:06 PM, Viresh Kumar wrote:
> On 10 June 2013 17:43, Chanwoo Choi <cw00.choi@samsung.com> wrote:
>> On 06/07/2013 07:23 PM, Viresh Kumar wrote:
>>> On 5 June 2013 13:41, Chanwoo Choi <cw00.choi@samsung.com> wrote:
>> 1301500082290 800000 61 11 1 43
>>
>> ...
>>
>> When 1301500082290 ns:
>> cpu0's busy_cpu_threshold : 32 = 64 * (800000/1600000)
> s/64/61 :)

Sorry, my mistake.

>
>>> How are you getting loads different for all your cpus? I believe you
>>> are just recording these values for policy->cpu and all cpus share
>>> same policy on your platform.
>>>
>> I got the Per-CPU load by using cpufreq_notify_transition().
>> when cpufreq governor call dbs_check_cpu().
> I forgot to remove this line in my earlier reply. I understood this towards
> the end of patch.

OK.

>
>>>> +#ifdef CONFIG_CPU_FREQ_STAT_DETAILS
>>>> +       struct cpufreq_freqs freqs;
>>>> +#endif
>>> Why do you need this to be global?
>> I'll remove global variable and move 'freqs' in some structure.
> ??
>
> You can just make it a local variable in the only function it is used.

You are right. I'll fix it by using local variable.

>
> TIP: Always place a blank line before and after your reply to kernel
> mails, this makes it much more readable.

I will modify this patch according to your comment
and then resend it after merged below patch.
- [PATCH] cpufreq: stats: Remove CONFIG_CPU_FREQ_STAT_DETAILS

Thanks your comment and tip.

Best Regards,
Chanwoo Choi