Re: [PATCH] acpi/cpuidle: reinitialize power_usage values when adding/removingC-states

[Daniel Lezcano]

On 10/22/2012 07:13 PM, Julius Werner wrote:
>> Could we just say this is always true because state[i+1] consumes less
>> than state[i] ?
>>
>> And then just remove the 'set_power_state' function, and the field
>> 'driver->power_specified' ?
>>
>> That will cleanup the code and fix this problem, no ?
> 
> I totally agree with your analysis. Even if a driver were to set
> proper usage values (and the power_specified bit), none of the
> existing governors would care about those actual numbers (and since
> the vast majority of drivers uses fake values anyway, this is not
> likely to change in the future). This seems to be a classic example of
> unnecessary over-engineering.
> 
> I am mostly interested in getting that bug fixed right now, but
> removing unnecessary code is always a good thing. If you think it
> would have a good chance of getting merged, I would be happy to draft
> up a larger patch that refactors power_usage away completely.

I am in favor of removing the unnecessary code as it fixes a bug also
but I am not a maintainer, so I can't tell if it has a good chance to be
merged as a bug fix.

I think Rafael can tell us what approach he would prefer.

Thanks
  -- Daniel


-- 
 <http://www.linaro.org/> Linaro.org â Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

--
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/