Hi all.

I think I need some help with debugging an issue I have.

I recently bought an "Apple Magic Trackpad 2" for reducing strain on my
hand (since I have some RSI[0] issues currently), which is supported in
Linux since version 4.20[1].

After connecting it to my laptop it worked instantly at first, but
"sometimes" un- and replugging the trackpad crashes my system.  Often it
works fine, but every now and then it triggers `general protection
faults` (often while opening big applications like Firefox), causes my
status bar to hang up, and usually ends in a freeze. Sometimes, just
before the freeze I get thrown back to the console and see a kernel
panic at the end of the log.

The laptop is a Thinkpad X200 running stock BIOS, Arch Linux and sway,
with microcode updates enabled. I could also reproduce the issue with
Gnome on Wayland. On Xfce I could not reproduce the issue, and also not
on Fedora (which I installed on the same laptop to the same SSD, for
testing).

I'm using the trackpad only via the USB cable, not via bluetooth.

Attached is a snippet from the dmesg output with debug logging enabled
(appended `debug` on the kernel cmdline) showing some of the protection
faults. I uploaded the full dmesg output here [2]. In this specific
dmesg output, most of the protection faults happened concurrently while
I was opening Firefox (several times to provoke the crash). A photo of
the kernel panic can be found here [3].

This kernel bug report[4] and related arch report[5] seem related to the
issue, maybe it's even the same issue? However, the kernel bug report
mentions that the issue only affects linux >=5.8.7. In my case, I could
reproduce the issue with the following linux versions (installed from
the Arch Linux archive):

        4.20.0
        4.20.1
        4.20.3
        4.20.7
        5.0.0
        5.0.13
        5.4.74
        5.2.0
        5.8.6
        5.8.7
        5.8.8
        5.9.2, 5.9.4
        5.10rc-2

The USB ID of the trackpad is 05ac:0265.

Because I'm very inexperienced with kernel debugging I would appreciate
any help/pointers to what would be wise to test next. If this is likely
to be an issue with the Linux kernel and not arch-related (unsure, since
I could not reproduce it on other distros), please tell me so and I will
post to the kernel bug tracker instead.

Thanks!

Merlin Büge


[0]: https://web.eecs.umich.edu/~cscott/rsi.html
[1]: 
https://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git/commit/?h=for-next&id=9d7b18668956c411a422d04c712994c5fdb23a4b
[2]: https://0x0.st/idkH.json
[3]: https://0x0.st/idkz.jpg
[4]: https://bugzilla.kernel.org/show_bug.cgi?id=209239
[5]: https://bugs.archlinux.org/task/67860




-- 
Merlin Büge

Attachment: amteth_arch_debuglog1_snippet
Description: Binary data

Reply via email to