Hi,
I've send a revert for review: https://review.coreboot.org/c/coreboot/+/33244
It seems to me that the added reset isn't following the protocol (i.e. ignores another returned byte). I'm not familiar with these controllers and lack the time to look into it right now. So that's the best I can do atm.
In the long run, I think we might want to replicate Linux' i8042 driver behavior in _all_ payloads. Having a common implementation in the payloads and OS would considerably decrease the necessary testing and provide a cleaner cut what has to be done in coreboot and what in the payload.
Nico