I appreciate the offer, but I don't think that I'm cut out for the job. I do hope that the kernel-power team will find an extra pair of hands soon though, what you guys are pulling off with such a small amount of resources is, quite simply put, amazing.

By the way, there might be some patches in kernel-bfs that could be harvested for kernel-power. Patches worth mentioning are the updated UBIFS module (source), some of CK's responsiveness patches (e.g. mm-lru_cache_add_lru_tail-1.patch, mm-kswapd_inherit_prio-1.patch & mm-lots_watermark.diff, maybe mm-background_scan-2.patch), and the BFQ I/O scheduler.
I'm not sure whether all of the mentioned patches are too bleeding edge to be included in kernel-power though. At least the UBIFS patch is pretty well tested and the CK patches aren't too invasive.
The BFQ I/O scheduler should be approached with some caution however; we had some issues with functions that were depreciated in kernel 2.6.32 (from which the backport originates), but OTOH it can be built as a separate module, so it isn't forced upon anyone.