Re: [PATCH 04/18] cpufreq: Manage fallback policies in a list

On 01/27/2015 12:36 AM, Viresh Kumar wrote:

Policies manage a group of CPUs and tracking them on per-cpu basis isn't the best approach for sure.

The obvious loss is the amount of memory consumed for keeping a per-cpu copy of the same pointer. But the bigger problem is managing such a data structure as we need to update it for all policy->cpus.

To make it simple, lets manage fallback CPUs in a list rather than a per-cpu variable.

Can you explain why we need a fallback list in the first place? Now that we are not destroying and creating policy objects, I don't see any point in the fallback list.

-Saravana