Keyboard Causes GUI Crash & Coredump

Support
,
1

This problem has been resolved; see post #6.

I use Solus/Budgie and Manjaro/Budgie almost exclusively. I also like to build custom keyboards.

I have a custom 66% Clueboard keyboard that I’ve been using for the past 5-6 years without any problems. Yesterday I decided to update the keymap in order to commonize its features with my other keyboards. As soon as I flashed the new keymap with QMK, my display went blank and I was unable to reboot the (Solus/Budgie Linux) machine with the Clueboard attached. Judging from the disk activity it booted, but the display was completely blank. The display did, however blink, as if the backlight was repeatedly switching on and off. I disconnected the Clueboard, rebooted and logged back into the machine. As soon as I connected the Clueboard, the display once again immediately went blank and began flashing.

I plugged the Clueboard into one of my laptops (Manjaro/Budgie Linux) and immediately the display went blank. It also would not reboot into the desktop with the Clueboard attached, but it did display a message that the Light Display Manager could not start. I also could not open a terminal.

Looking at the logs, the Solus machine didn’t capture anything useful. It doesn’t even capture the fact that the Clueboard was connected, only that I eventually pressed the the power button to power down the machine. But, in the Manjaro machine, journalctl records that the Clueboard was connected and then indicates that systemd immediately initiated a core dump.

Strangely, the Clueboard works perfectly when navigating the UEFI of these machines and it also works perfectly when attached to a Windows 10 machine, with all keyboard layers and functions working correctly. BTW, the Windows 10 machine dual boots the aforementioned Solus/Budgie distribution, so this is clearly not a hardware problem.

I installed the QMK keyboard firmware on the Windows 10 machine and reflashed the Clueboard and the problem with the Linux machines persists. Interestingly, I attached the Clueboard to a different laptop running the live Solus/Budgie ISO and later the same laptop running the live PCLinuxOS ISO and in both cases the Clueboard functioned flawlessly … presumably they aren’t using lightdm in the live environment … assuming, of course (big assumption) that lightdm sits at the root of the problem? When I plug the Clueboard into the laptop when running the Manjaro/Budgie live ISO, the GUI immediately disappears and some of the boot messages (just prior to Budgie starting) are displayed. When I unplug the Clueboard, the Budgie desktop reappears, but with a log-in prompt (I guessed that the password is manjaro and this turned out to be correct). I don’t know how old this ISO is, but it responds with “5.13.13-1-MANJARO” to a uname -r query. The Firefox version included on this ISO is 91.0.2.

I’ve posted the Manjaro/Budgie coredump on pastebin, in the event that someone can provide some insight as to what may be causing the problem. Coredump - Pastebin.com

I didn’t see any other posts here, relating to QMK keyboard firmware, so I was just wondering if this is a known issue, or if anyone has any insights, or suggestions. I’ll obviously be cross posting this issue to the QMK devs, but if possible, I’d kinda like to have some feedback regarding the coredump, before I do that. So, any help with that would be … helpful in pointing the QMK devs in the right direction.

Thanks in advance!

2

5.13 is EOL. Switch to a supported kernel:

sudo pacman-mirrors --fasttrack 5 && sudo pacman -Syyu 
sudo mhwd-kernel -i linux515

Then reboot into 5.15. Press Esc during boot to get into grub menu > advanced options and boot with 5.15. Double check your running kernel is not 5.13 then remove it:

uname -r #to check kernel version
sudo mhwd-kernel -r linux513
Also see

[HowTo] Provide System Information

[HowTo] Find error logs

3

You misunderstand. That is the kernel version of the live ISO, which is probably past its freshness date (I just don’t remember when I downloaded it). The kernels being used in the Solus/Budgie and Manjaro/Budgie installations are up to date.

Sorry if I was unclear.

4

:+1:

I went through the journal and only this part seems relevant to the problem as after this Xorg crashes:

Feb 25 09:22:40 copernicus kernel: usb 3-1: new full-speed USB device number 33 using xhci_hcd
Feb 25 09:22:41 copernicus kernel: usb 3-1: New USB device found, idVendor=c1ed, idProduct=2320, bcdDevice= 0.01
Feb 25 09:22:41 copernicus kernel: usb 3-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Feb 25 09:22:41 copernicus kernel: usb 3-1: Product: Clueboard 66%
Feb 25 09:22:41 copernicus kernel: usb 3-1: Manufacturer: Clueboard
Feb 25 09:22:41 copernicus mtp-probe[242470]: checking bus 3, device 33: "/sys/devices/pci0000:00/0000:00:14.0/usb3/3-1"
Feb 25 09:22:41 copernicus mtp-probe[242470]: bus: 3, device: 33 was not an MTP device
Feb 25 09:22:41 copernicus upowerd[710]: treating change event as add on /sys/devices/pci0000:00/0000:00:14.0/usb3/3-1
Feb 25 09:22:41 copernicus kernel: input: Clueboard Clueboard 66% as /devices/pci0000:00/0000:00:14.0/usb3/3-1/3-1:1.0/0003:C1ED:2320.0001/input/input20
Feb 25 09:22:41 copernicus kernel: hid-generic 0003:C1ED:2320.0001: input,hidraw0: USB HID v1.11 Keyboard [Clueboard Clueboard 66%] on usb-0000:00:14.0-1/inp>
Feb 25 09:22:41 copernicus kernel: input: Clueboard Clueboard 66% System Control as /devices/pci0000:00/0000:00:14.0/usb3/3-1/3-1:1.1/0003:C1ED:2320.0002/inp>
Feb 25 09:22:41 copernicus kernel: input: Clueboard Clueboard 66% Consumer Control as /devices/pci0000:00/0000:00:14.0/usb3/3-1/3-1:1.1/0003:C1ED:2320.0002/i>
Feb 25 09:22:41 copernicus kernel: hid-generic 0003:C1ED:2320.0002: input,hidraw1: USB HID v1.11 Device [Clueboard Clueboard 66%] on usb-0000:00:14.0-1/input1
Feb 25 09:22:41 copernicus kernel: usbcore: registered new interface driver usbhid
Feb 25 09:22:41 copernicus kernel: usbhid: USB HID core driver
Feb 25 09:22:41 copernicus mtp-probe[242498]: checking bus 3, device 33: "/sys/devices/pci0000:00/0000:00:14.0/usb3/3-1"
Feb 25 09:22:41 copernicus mtp-probe[242498]: bus: 3, device: 33 was not an MTP device
Feb 25 09:22:41 copernicus systemd-logind[432]: Watching system buttons on /dev/input/event20 (Clueboard Clueboard 66% System Control)
Feb 25 09:22:41 copernicus systemd-logind[432]: Watching system buttons on /dev/input/event21 (Clueboard Clueboard 66% Consumer Control)
Feb 25 09:22:41 copernicus systemd-logind[432]: Watching system buttons on /dev/input/event19 (Clueboard Clueboard 66%)
Feb 25 09:22:41 copernicus audit[492]: ANOM_ABEND auid=4294967295 uid=0 gid=0 ses=4294967295 subj==unconfined pid=492 comm="Xorg" exe="/usr/lib/Xorg" sig=6 r>
Feb 25 09:22:41 copernicus kernel: audit: type=1701 audit(1645802561.968:883): auid=4294967295 uid=0 gid=0 ses=4294967295 subj==unconfined pid=492 comm="Xorg>
Feb 25 09:22:41 copernicus systemd[1]: Created slice Slice /system/systemd-coredump.

Possible problem:

bus: 3, device: 33 was not an MTP device

Try following these guides regarding for the issue. If still not fixed report on qmk_firmware github:

Links

https://unix.stackexchange.com/questions/26842/mounting-usb-drive-that-is-not-recognized

External usb drives: "not an MTP device" - #2 by megavolt

Issues · qmk/qmk_firmware · GitHub

5

Thanks, I appreciate the suggestion. I noticed that line in the journal, but I disregarded it assuming that keyboards are not MTP devices. By MTP device I assumed (perhaps incorrectly) that this referred to something like a cell phone, or an Apple iPod.

Cheers!

6

This problem has been resolved. Xorg has an undocumented “feature” where it will segfault if a USB keyboard it connected which has a “%” character in its device name.

2 Likes
7

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.