MMm.. I wonder if we can just "code around it" in the kernel for the case where this chip is dead? Oh, I suppose not, since it also kills the HD controller for some reason.

Cheers