On 16-04-18, 09:44, Daniel Lezcano wrote: > Because we rely on the number to identify the cluster and flag it > 'processed'. The number itself is not important.
It is, because you are creating multiple groups (like cpufreq policies) right now based on cluster id. Which will be zero for all the CPUs. So in a big little system we will have two cpufreq cooling devices and one cpuidle cooling device. -- viresh
