The MR looks mostly good for me, but I'm afraid that we need to do something about test failures. This happens very sporadically on win10 as well, but it seems consistent on win8 with this MR:
http://test.winehq.org/data/42811d775b1a299f1dddcbbae4a2b80cb1b77698/win22H2...
I don't know where those events come from, maybe reading and printing events in case wait succeeded would give us a hint.