Tablet pressure stops working after a re-plug

[hyiltiz]

Goal

Tablet works after every re-plug after ...

• Ideally: some one-time only configuration has been done (maybe after writing some udev rules).
• Compromise: running some command that doesn't kill X11 before/after the re-plug (maybe manual unloading then reloading some kernel modules/drivers, flushing udev rules)
• Acceptable: running some command (including killing X11) before/after the re-plug (maybe manual unloading then reloading some kernel modules/drivers, flushing udev rules)
• Unacceptable: a reboot

Problem

I was able to set up my tablet (documented #296) without digimend drivers; even after uninstalling the digimend drivers, simply including the appropriate 50-tablet-xorg.conf file is sufficient to make it work, i.e. enable pressure sensitivity for the tablet after a fresh reboot Xorg.0.log in Debian Testing (tried kernels from 4.9 to 5.5.0). However, once I plug out the tablet USB, then there is nothing I can do to make it work upon plugging back in.

Things I have tried

Things I have tried before and after the re-plug in each combination:

1. Tried unloading / unload then reloading back in kernel drivers using modprobe -a/-r: hid, usbhid, hid-generic, hid-uclogic, usb-common, usbcore (Warning: most of them also leaves the keyboard and mouse unusable so be prepared);
2. Tried commenting out 50-tablet-xorg.conf file
3. Tried removing xserver-xorg-input-wacom package
4. Tried installing digimend drivers
5. Flush udev rules with: udevadm control --reload-rules && udevadm trigger

My hypothesis is that apparently the order that loads the drivers is important. hid-uclogic driver from digimend is not required for my tablet as simply including the 50-tablet-xorg.conf file is sufficient. However, even during a working state, usbhid seems to be driving the tablet rather than wacom regardless of what Xorg.0.log file included above indicates:

$ /sys/kernel/debug/usb/devices
...
T:  Bus=03 Lev=01 Prnt=01 Port=01 Cnt=01 Dev#=  5 Spd=12   MxCh= 0
D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=5543 ProdID=0061 Rev= 0.00
S:  Manufacturer=UC-Logic
S:  Product=A640
C:* #Ifs= 3 Cfg#= 1 Atr=a0 MxPwr=100mA
I:* If#= 0 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=01 Prot=02 Driver=usbhid
E:  Ad=81(I) Atr=03(Int.) MxPS=  64 Ivl=5ms
I:* If#= 1 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=01 Prot=02 Driver=usbhid
E:  Ad=82(I) Atr=03(Int.) MxPS=   8 Ivl=4ms
I:* If#= 2 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=00 Prot=00 Driver=usbhid
E:  Ad=83(I) Atr=03(Int.) MxPS=   8 Ivl=8ms
...

Here is a dmesg after a reboot (tablet working) then a re-plug (tablet stops working).

Would greatly appreciate if anyone could help.

[hyiltiz]

After some extensive research on kernel modules hid-uclogic hid-generic wacom, USB input interfaces usbhid, X input drivers libinput wacom, I found a solution. But first, it was hard to really identify the source of the problem. Once I got that right, fixing was somewhat straightforward.

Problem

I checked /sys/devices for reliable source of information. After a fresh reboot, the eventX file corresponding to a directory under sysfs which contained a device named UC-Logic A640 responded to physical stylus input events, when inspected with xxd /dev/input/eventX; however, after a re-plug (with or without killing Xorg), a different eventY file corresponding to a different directory for a device named UC-Logic A640 Mouse responded instead. Notice the change of the device name and sysfs directory for the same physical stylus.

Solution

I created a udev rules by repeatedly re-plugging/killing Xorg/rebooting while checking which /dev/input/eventX responded to the stylus inputs, and traverse the sysfs to locate its directory, then logged down udevadm info -a -p /sys/devices/pci0000:00/0000:00:14.0/usb1/1-2/1-2:1.0/0003:5543:0061.000B/input/input31/event10 (note the path name changes every time), finally comparing them until I could find the following matching rules. Note that # my stylus always showed up as one of event10, event12 or event14. You may need to adjust accordingly:

$ cat /etc/udev/rules.d/60-wacom.rules 
KERNEL!="event[0-9]*", GOTO="wacom_end"
KERNEL=="event[0-9]*", SUBSYSTEM=="input", ATTRS{name}=="UC-Logic A640 Keyboard", ATTRS{phys}=="usb-0000:00:14.0-2/input*", ATTRS{properties}=="0", SYMLINK+="input/wacom-keyboard"
KERNEL=="event1[024]" ,SUBSYSTEMS=="input", ATTRS{phys}=="usb-0000:00:14.0-2/input[01]", ATTRS{uniq}=="", ATTRS{properties}=="2", SYMLINK+="input/wacom-pen"
LABEL="wacom_end"

This creates the pad and the stylus devices in a stable location at /dev/input/wacom-keyboard and /dev/input/wacom-pen. Now we load the wacomX driver manually, setting the relevant device paths. This prevents X from getting confused by duplicated and buddy device names and types (there were several Mouse subdevices etc. as the result ofxinput list`):

$ cat /etc/X11/xorg.conf.d/50-tablet.conf
Section "InputDevice"
  Driver        "wacom"
  Identifier    "A640 stylus"
  Option        "Device"        "/dev/input/wacom-pen"        # USB ONLY
  Option        "Type"          "stylus"
  Option        "USB"           "on"                      # USB ONLY
EndSection

Section "InputDevice"
  Driver        "wacom"
  Identifier    "A640 pad"                                     # Tablet buttons
  Option        "Device"        "/dev/input/wacom-keyboard"        # USB ONLY
  Option        "Type"          "pad"
  Option        "USB"           "on"                      # USB ONLY
EndSection

Section "ServerLayout"
  Identifier    "X.org Configured"
  InputDevice   "A640 stylus"
  InputDevice   "A640 pad"
EndSection

With this, either flush udev with udevadm control --reload-rules && udevadm trigger or killall Xorg or simply reboot.

p.s. wdaemon may offer an alternative solution, but I'd rather keep one software out of the stack.
p.p.s hid-uclogic was never loaded somehow; it seems installing digimend via dkms was unnecessary.

References

On manually configuring X input driver for the tablet:
https://github.com/linuxwacom/xf86-input-wacom/wiki/Configuring-X

On manually detecting the input (sub)devices:
https://github.com/linuxwacom/input-wacom/wiki/Fixed-device-files-with-udev
http://reactivated.net/writing_udev_rules.html

If you feel advantageous, maybe try:
https://github.com/linuxwacom/xf86-input-wacom/wiki/wdaemon

[hyiltiz]

Can we learn something from this, and maybe abstract it into a Wiki tutorial page?

[spbnick]

@hyiltiz, I did a quick change to try and support your tablet with DIGImend drivers, please see if it works: https://github.com/spbnick/digimend-kernel-drivers/archive/add_support_for_parblo_a640.zip

PR #393 is opened for this.

[hyiltiz]

Thanks @spbnick ! I'll test it out and let you know. It turns out that my solution above wasn't robust at all; my udev rules wasn't robust against plugging it into a different USB port or the order of plugging in other USB devices (e.g. mouse). Glad to see it supported by digimend drivers.

[EDIT]
Just tested out. hid_uclogic is now loaded by usbhid so digimend is finally being used, but only pad device is recognized and doesn't respond to button presses either (detailed results in #393).