x11/nvidia-driver: Remove devclass from DRIVER_MODULE on recent main.

[jhb]

Summary

Initialize nvidia_devclass in nvidia_modevent instead.

[danfe]

Looks mostly fine. The patch header is not quite by our standards (only *.orig file timestamped, UTC-relative, seconds-precision) but that's not a blocker.

x11/nvidia-driver/files/extra-patch-src-nvidia_subr.c.in

2

You may want to regenerate the patch header with make makepatch (or by hand) so it stays consistent with other patches.

[jhb]

x11/nvidia-driver/files/extra-patch-src-nvidia_subr.c.in

2

make makepatch is not really practical for the nvidia port since it patches these files directly. This the output of diff -up. The porter's handbook simply says to use diff -u and doesn't say anything about trying to sanitizer or otherwise change patch headers?

[jhb]

x11/nvidia-driver/files/extra-patch-src-nvidia_subr.c.in

2

Hmm, I did manage to get something out of make makepatch I could use (by copying the header over from a patch it generated and massaging it to deal with %%NVSRC%%)

[danfe]

@jhb wrote:

make makepatch is not really practical for the nvidia port since it patches these files directly.

Yeah, I know, this is a bit of an annoyance, hence I did not insist. :-)

The porter's handbook simply says to use diff -u and doesn't say anything about trying to sanitizer or otherwise change patch headers?

Well, we try to follow the common standard, but it is not strictly enforced (albeit it probably should've mentioned in the PHB).

Hmm, I did manage to get something out of make makepatch I could use (by copying the header over from a patch it generated and massaging it to deal with %%NVSRC%%)

Right, that what I was thinking (I often do the same with non-trivial patches), thanks!