Thanks for the answer. Maybe I should mention that I have implemented
CPUIDLE myself on my companys ARM-system.
To me it seems a little bit strange that there are two C0 states, "running"
and "not running" showed in powertop. The powertop screendumps I have found
on the web does not look like that.
I have added no special information to the sysfs entries "name" and "desc"
for my CPUIDLE states.

2008/12/11, Arjan van de Ven <[email protected]>:
>
> Rickard Andersson wrote:
>
>> Hi!
>>  I have compiled powertop 1.10 for my system and it displays my c states
>> as below:  Cn                Avg residency
>> C0 (cpu running)        ( 0.4%)
>> C0                6.8ms ( 0.4%)
>> C1              307.9ms (99.1%)
>>  I have two sleep states, C1 and C2, but powertop shows them as C0 and C1.
>> As I have understood it C0 should be "CPU running". When I do the change
>> below to function read_data_cpuidle(..) at line 368 in the file powertop.c
>> it all looks fine for me.
>>      /*sprintf(cnames[clevel], "C%i", clevel);*/
>>    sprintf(cnames[clevel], "C%i", clevel+1);
>>   Anybody has some comments?
>>
>>
>
> hi
>
> the patch unfortunately is not correct
> current kernels really have a C0 idle state in addition to a C0 running
> state...
> .... which means that whatever bug you are hitting is something else, yet
> undiagnosed...
>
_______________________________________________
Power mailing list
[email protected]
http://www.bughost.org/mailman/listinfo/power

Reply via email to