Frequent Kernel Panics, segfaults, crashes on Alder Lake Laptop

Shapeshifter · 2024-07-16 12:06:36

It started a few months ago and has only gotten worse. This Lenovo ThinkPad X1 Yoga (Gen 7) runs for somewhere between 30 minutes and a few hours before it inevitably locks up. The lock-up happens at random, whether I'm doing something specific or nothing at all, and it comes in various flavors:

Full-on kernel panic: Caps Lock LED flashing, fans ramp up to full blast-off, nothing works anymore.
Lock-up without visible consequences, i.e. Caps Lock not flashing, but entirely locked up. Trying to log on via SSH immediately disconnects. Seems like it's still running but barely.
Keeps working somewhat, but can barely launch any new processes. Trying to do so just hangs. For example, can't log-in to the machine via SSH anymore. su sometimes still works. systemctl reboot usually just hangs. For example, at the end of below's paste, you can see that me trying to run thunderbird consistently segfaults, but in that instance, some other applications kept working.
Some application crashes randomly, but otherwise mostly OK.

If I can still use the machine as root, I can always see kernel problems in dmesg. Here's just a couple different errors (separated by "..."):

https://bpa.st/KOGQ.

In short:

General protection fault
Segfault
CPU stalls
Bad page map

What I already tried:

Memtest for over 24 hours. No issues. I highly doubts it's a memory problem, as the machine consistently crashes way earlier than 24 hours even under idle use.
Remove all kernel taints. I blacklisted all out-of-tree modules and removed i915.enable_dc=0 from the kernel (which I had added long ago for some reason). No taints, still crashes (line 323 in the paste). Some of the logs above are from before I did this but it evidently made no difference
Use Xorg instead of Wayland. I was wondering if it's a Wayland issue; apparently it's not.
Don't run X or Wayland, just log onto TTY and run some stress --vm 16 --vm-bytes 2300000000 and other busy things that don't use graphics. Everything is OK, no crashing after several hours of this.
Ran the Lenovo self-diagnostic software that ships onboard. Took many hours, all pass
Booted linux mint off a USB stick. Ran x11vnc and remoted in, while also running stress and glxgears. Nothing happens after several hours. I conclude it's definitely not a hardware problem and it is somehow tied to my arch installation.

So I have no idea what's wrong or what I can do. I think the RAM is fine. Maybe the CPU is fried, or some other board component, but I don't think so since without Wayland/X, it doesn't crash. I still want to believe that it is more likely a kernel bug, maybe in DRM/graphics. I'm not sure if it's relevant, but I feel like crashes are even more frequent if I use wayvnc or x11vnc, but that may be bias.

What can I do now?

Last edited by Shapeshifter (2024-07-17 15:29:39)

impossibleveins23 · 2024-08-05 12:26:18

* I would try a few kernels including the LTS to see if there is any effect.
* It would be helpful to see full journal from each of these use cases (if you had to reboot use -b -1 to get latest journal if it was written successfully to the disk).
* Also check this for how to extract more info regarding the panic

Arch Linux

#1 2024-07-16 12:06:36

Frequent Kernel Panics, segfaults, crashes on Alder Lake Laptop

#2 2024-08-05 12:26:18

Re: Frequent Kernel Panics, segfaults, crashes on Alder Lake Laptop

Board footer