We definitely need to verify if/what impact these flags (namely RANDSTRUCT) has on each and every dkms package. I failed to verify that when going from 4.14.8 to 4.14.10 and learned my lesson by breaking nvidia.
The referenced nvidia patch can't hurt whether or not we actually employ RANDSTRUCT. Positional initialization is so 1990s ;-P All in all it's beginning to look as if we should write our own Void test procedures for kernel updates, or should we?