I did say this before, for me at least (and likely Darkslayer and vaporesso), the problem is caused by the bt-mice patch introduced with v46. If anyone is interested I can upload my build of v47 minus that specific patch somewhere for anybody who wants test whether that patch causes the same issue for them (I could do with a recommendation of where to upload it). I did intend to do further debugging on the issue, but I've been lazy and the thought of recompiling the kernel with bluetooth debugging patches and reflashing the kernel several more times - when I now have it running just fine - just to run tests didn't appeal much to me, but I might just set aside some time to do it tomorrow.