FS#77685 - [amdgpu] Random major crashes on latest kernels

Attached to Project: Arch Linux
Opened by Giovanni Santini (ItachiSan) - Wednesday, 01 March 2023, 09:12 GMT
Last edited by Toolybird (Toolybird) - Wednesday, 19 April 2023, 22:15 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To No-one
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
The AMDGPU tends to crash on the latest 'linux' kernel.

My laptop has an integrated AMD GPU and a discrete NVIDIA one.
While using my laptop normally, the screen would start blanking and stop working as intended.
The video driver would panic, however all the rest of the computer would work e.g. I was able to talk in a videocall with no issue although the screen was unusable.

While having slightly worse performance, 'linux-lts' doesn't seem to have the issue.

Additional info:
* package version(s)
- linux: 6.2.1.arch1-1
- xf86-video-amdgpu: 23.0.0-1
* config and/or log files etc.: attached to this task
* link to upstream bug report, if any: none as of now, I can open some if needed.

Steps to reproduce:
1. Use the laptop
2. Wait
   20220301T0900_full.log (729 KiB)
This task depends upon

Closed by  Toolybird (Toolybird)
Wednesday, 19 April 2023, 22:15 GMT
Reason for closing:  Upstream
Additional comments about closing:  Plenty of activity in the upstream ticket, although no fixes or culprit found yet. Not much Arch can do here.
Comment by Giovanni Santini (ItachiSan) - Wednesday, 01 March 2023, 09:21 GMT
I can attach additional data if required. :)
Comment by Toolybird (Toolybird) - Wednesday, 01 March 2023, 20:59 GMT
It's clearly a kernel regression. General advice for these things here [1]. You are best advised to report it upstream. amdgpu issues can be reported here [2]. Please let us know what you find out.

[1] https://wiki.archlinux.org/title/Kernel#Debugging_regressions
[2] https://gitlab.freedesktop.org/drm/amd
Comment by Giovanni Santini (ItachiSan) - Thursday, 02 March 2023, 13:43 GMT
Thanks Toolybird!

So, in short, I should try to bisect the kernel and the AMDGPU package...?
Sounds like a lot of work ^^"
Comment by loqs (loqs) - Thursday, 02 March 2023, 21:23 GMT
You can find some prebuilt bisection kernels in  FS#77632 .
Comment by Giovanni Santini (ItachiSan) - Thursday, 09 March 2023, 12:25 GMT
So, the upstream issue is here:
https://gitlab.freedesktop.org/drm/amd/-/issues/2447

Currently working on a slightly modified version of 'linux-git' PKGBUILD as it requires v6.3 updates.

Now building the kernel and I will start bisecting from 'linux-rolling-stable'.
Comment by Giovanni Santini (ItachiSan) - Tuesday, 14 March 2023, 09:32 GMT
Currently not facing this same issue in 6.2.5.
Although I am facing massive lag sometimes, I do not have major crashes.
Comment by Johannes Deger (jaydi) - Friday, 17 March 2023, 12:19 GMT
I am facing the issue on 6.2.6-zen1-1-zen

Comment by Johannes Deger (jaydi) - Sunday, 19 March 2023, 14:09 GMT
I am facing the issue on 6.2.6-zen1-1-zen

Loading...