DS4 touchpad is missing ID_INPUT_TOUCHPAD.

[Sur3]

Hi my DualShock 4 touchpad didn't work because ID_INPUT_TOUCHPAD was missing, I therefore wrote a rule to get the touchpad functional with libinput: SUBSYSTEM=="input", KERNELS=="0003:054C:09CC.", ATTR{device/name}=="Sony Interactive Entertainment Wireless Controller Touchpad", ENV{ID_INPUT_TOUCHPAD}="1" SUBSYSTEM=="input", KERNELS=="0005:054C:09CC.", ATTR{device/name}=="Wireless Controller Touchpad", ENV{ID_INPUT_TOUCHPAD}="1"

This is a 2016v2 version, don't know about the first version, but probably it will be: SUBSYSTEM=="input", KERNELS=="0003:054C:05C4.", ATTR{device/name}=="Sony Interactive Entertainment Wireless Controller Touchpad", ENV{ID_INPUT_TOUCHPAD}="1" #TODO: check name SUBSYSTEM=="input", KERNELS=="0005:054C:05C4.", ATTR{device/name}=="Wireless Controller Touchpad", ENV{ID_INPUT_TOUCHPAD}="1" #TODO: check name

See also: https://bugs.freedesktop.org/show_bug.cgi?id=105050#c10

Greetz Si

[whot]

fwiw, this shouldn't be fixed with a custom rule for this device, given the evdev codes this should always be tagged as touchpad anyway. a more generic version is needed (or a fix of whatever tags touchpads right now).

[blueness]

@whot agreed. @Sur3 I'm confused by your suggestion, wasn't this fixed upstream by the following: https://github.com/systemd/systemd/pull/8259

[Sur3]

@blueness, I'm not using systemd but OpenRC, therefore whatever that systemd bug fixed didn't affect me. Also that bug seems to be about not to assign "ID_INPUT_TOUCHPAD" and "ID_INPUT_MOUSE" at the same time, my problem is none of that were assigned at all, only the generic "ID_INPUT".

[blueness]

@Sur3 still @Whot is right, i can't add a custom rule for just this device.

[Sur3]

@blueness then we should research why the generic rule doesn't work.. @whot which part of the evdev codes exactly tells udev that it's a touchpad..? https://bugs.freedesktop.org/show_bug.cgi?id=105050#c2

[whot]

look at the systemd udev-builtin input_id.c please

[Sur3]

Ok I compared the input_id.c code from eudev and systemd, seems systemd also accepts ABS_MT_POSITION_X/Y for setting a dev as touchpad, whereas eudev checks only for ABS_X/Y and doesn't accept MT coordinates to set a device as touchpad..

[blueness]

systemd's test_pointers() in udev-builtin-input_id.c has quite a few changes which address joysticks, touchpads, touchscreens, etc. I'm going to have to sort through those and import. @Sur3 if I provide you with a patch, can you test it?

[blueness]

@Sur3 @whot can you testing the latest git head. commit dba4728 import's upstreams work on test_pointers(). Let me know if this properly addresses your issue.

[Sur3]

Ok thanks, I'll test that.

[Sur3]

Tried .9999.ebuild but all ID_INPUT* are missing, so even my mouse and keyboard were non-functional.

[blueness]

Okay, let me revert and test at my end before pushing again.

[blueness]

Actually there is another problem with -9999 from recent contributions which may be causing your failure with all ID_INPUT* missing. I know how to fix this, but I'm going to have to unravel some changes since 3.2.5.

[Sur3]

Yes or you could provide a patch, for the the new test_pointer so I can test that with 3.2.5 if that is easier for you..?

[blueness]

I've scheduled fixing this for Wednesday. There are some bad rules that were important uncritically by a contributor. I'll ping you then.

[blueness]

Okay I found the problem. Commit bb070c1 forgot to include the installation of 60-input-id.rules and 70-joystick.rules via the Makefile. Without this, commits dba4728 and 5df0137 leave the system without any useable input devices.

@Sur3 please test git/head again.

[Sur3]

Hi, mouse, keyboard and touchpad work fine now. Only problem so far is when connecting over bluetooth the gamepads udev rules seem to be only executed on first connect, if I remember correctly disconnecting and reconnecting the gamepad over bluetooth worked fine before, I'll test with 3.2.5 again if it works there..

[Sur3]

Nope same problem with 3.2.5, so it's probably not udevs fault but a bluetooth problem, because with usb everything works fine now. ^^

[blueness]

@Sur3 that's a different bug.