This approach (hacking the LINT_KMODS out) seems kind of weird and fugly.
We've discussed just nuking lint altogether.
Is there some reason that this particular driver is especially tragic with lint?
I'm inclined to not make this change unless more information about why "fixing" the lint problems here is more trouble than its worth. (3rd party code maybe?)
Conversely, if lint as a tool has outgrown its usefulness, maybe its time to start the process of gutting it altogether?