On Tue, Sep 23, 2014 at 06:08:04PM +0200, Vincent Guittot wrote:

This implementation of utilization_avg_contrib doesn't solve the scaling
in-variance problem, so i have to scale the utilization with original
capacity of the CPU in order to get the CPU usage and compare it with the
capacity. Once the scaling invariance will have been added in
utilization_avg_contrib, we will remove the scale of utilization_avg_contrib
by cpu_capacity_orig in get_cpu_usage. But the scaling invariance will come
in another patchset.

I would have expected this in the previous patch that introduced that
lot. Including a few words on how/why the cpu_capacity is a 'good'
approximation etc..

Finally, the sched_group->sched_group_capacity->capacity_orig has been removed
because it's more used during load balance.

That sentence is a contradiction, I expect there's a negative gone
missing someplace.