On 06-02-18, 11:48, Daniel Lezcano wrote:

On 06/02/2018 05:28, Viresh Kumar wrote:

quoted

Surely we can do one thing at a time if that's the way we choose to do it.

Easy to say :)

The current code is to introduce the feature without impacting the DT
bindings in order to keep focused on the thermal mitigation aspect.

There are still a lot of improvements to do after that. You are
basically asking me to implement the copy-on-write before the memory
management is complete.

Perhaps I wasn't clear. What I was trying to say is that we can do "one thing at
a time" if we choose to create a "combo device" (the way you proposed). I am not
trying to force you to solve all the problems in one go :)

Can you give an example? Or your understanding is incorrect or I missed
the point.

So I tried to write it down and realized I was assuming that different
cooling-maps can be provided for different cooling strategies
(cpufreq/cpuidle) and obviously that's not the case as its per device.
Not sure if it would be correct to explore the possibility of doing
that.

-- 
viresh