I already added "device hid" in NOTES, so please rebase your patch!

Please remove this \r character while at it :-)

Ditto.

Done

Done

I'd prefer we put this in sys/conf/NOTES as well.
And for mips/powerpc remove it in those notes files with nodevice if needed.
It's supported on all the 'mainstream' ports and there's nothing specific to the architecture in these files that I've noticed.
I think the duplication of the other options is worse than just having a nodevice or two if it turns out not to build there.

Is there some reason this is commented out? If it's not normally on in GENERIC, it should be here, if it is normally on and we want to support things with it off, it shouldn't be here, though it's a mixed bag if that's commented out with a comment about why or it simply not here. I prefer the former. But commented out w/o a reason will invite others to remove either the '#' or the whole thing when 'cleaning up' this.

This is unrelated and should be committed separately, here and elsewhere in this review (iichid.c)

HID-over-I2C depends on ACPI as it read configuration through _DSM method call (Microsoft specs requires that). That means it is architecture dependent at least for now.
If someone knows how to attach such a devices with FDT, it is easy to fix as all ACPI parts are isolated in probe() method.

This option only changes usbhid probe priority and adds/removes PnP IDs to different drivers. I will add comment about that.
IMO there is no difference if it is enabled or is disabled in LINT.

sysctl changes stays in separate commit in my repo.

Really, It is better to enable it in a case of "no difference". Just to not confuse others.