On Sat, Jan 09, 2021 at 10:16:16AM +0100, Marcus Glocker wrote: [...]
> So the problem doesn't seem to be in your new ujoy(4) code, but how the > dev/hid/hid.c:hid_is_collection() function tries to cope with the PS4 > controller. > > I'm not much in to HID, but when I sync up the hid_is_collection() > function with NetBSD, the PS4 controller attaches to ujoy(4) as > expected, and also can be accessed by usbhidctl(1), while my other > HID devices continue to work as before: Nice find. [...] > Once this has been fixed I would continue to do some more ujoy(4) > testing, and check about a possible import. We also should agree on > who/how to fix the ports part. While I can't speak for all 11,000 packages, I am yet to encounter a port that uses gamecontrollers/joysticks without going through sdl or sdl2. And for those 2 ports it's literally a 1-line diff each. I'm not sure if looking at all the ports with 'usbhid' in WANTLIB would be helpful. I have a list of 55 of those from sqlports, excluding anything with 'sdl' in its name.