Re: [PATCH v3 6/7] thermal/drivers/cpu_cooling: Introduce the cpu idle cooling driver

[Martin Kepplinger]

On 25.10.19 16:45, Daniel Lezcano wrote:
> 
> Hi Martin,
> 
> On 25/10/2019 13:22, Martin Kepplinger wrote:
>> Hi Daniel,
>>
>> Quick note: You're missing a kref_init() in this version of the cooling driver.
>>
>> Also, as already mentioned, set_freezable() doesn't seem to work without
>> supporting suspend/resume.
> 
> Right, it is removed.
> 
>> And since cpuidle got somewhat completely reworked in 5.4, could you create
>> a new, rebased version of this? Let me know in case I missed an earlier update.
> 
> Yes, I'm on it. There is a bug in the cpuidle registering I'm
> investigating right now. If you want I can put a branch with the idle
> injection cooling device as soon as it is fixed.
> 
> Note most of the patches are now merged. The cpuidle cooling device was
> split into the idle injection mechanism in the drivers/powercap [1] and
> the cooling device initialization which is the part to be merged.
> 
> The idle injection resolution was increased [2].
> 
> The idle state injection selection is currently in the review process [3].
> 
> The remaining part is the thermal cooling device itself which is a
> separated cooling device from the existing cpu one.
> 
>   -- Daniel
> 
> [1] https://lwn.net/Articles/757393/
> [2] https://patchwork.kernel.org/patch/11074003/
> [3] https://lkml.org/lkml/2019/9/9/447
> 

Hi Daniel,

Putting up a branch may make sense and seems like it would make testing
easy. I'll definitely try to test early as soon as you have the pieces
in place one way or another :)

thanks for your work on this,
                                martin