Testing for kernel performance regressions

From my experience developers are most responsive to comments on the latest changes (which is most to front of mind). I suggest prompt results against each rc release directed to the maintainer of each area would be more developer friendly.