RX 7800 XT issues after updating to kernel 6.7

I am having the same issues as listed here: https://bbs.archlinux.org/viewtopic.php?id=290707&p=3

I am kinda afraid of posting on there tbh… I’ll post here, i know ya’ll don’t bite!

To reproduce:

  1. reboot the system
  2. upon rebooting, I see plymouth prompting for a password, then the screen goes black when amdgpu drivers try to load, the screen never comes back on
  3. I hit the reset switch on the machine, rinse and repeat
  4. shutting down the machine, and cold booting works, the screen will go black momentarily, but comes back when amdgpu drivers have successfully loaded

i tried rebooting into the latest LTS kernel and it makes no difference, im not sure how rebooting from the reset switch does not work and the issue keeps persisting :\

How can I fetch dmesg logs when the monitor just switches off?

Edit: cold booting into LTS lets me reboot just fine, so as long as i keep rebooting into LTS

I’m also working on a bit similar problem on one of my machines (but much older equipment ).
As I haven’t found it yet, I’m using this tool to protect the filesystem between each reboot:
It’s by Mr. Kresimir …

By the time we find our problem, we might be able to at least save the disk from the cold reboots.

Just tough it would be a good tool for you in the meantime … :smile:

I’ll be following your thread to learn more. :wave:

Why not just downgrade the kernel and wait until the next kernel update for 6.7.

If you can boot with the USB installer, then it probably is not a hardware issue.

If and when you have booted with the USB installer, you could give some logs, e.g.

sudo journalctl -b -1 | eos-sendlog

and show the output here.

Edit: better if you can boot to TTY without the USB (press Ctrl-Alt-F3). Then you can do the command above and show the output here.

1 Like

I get the same behavior on reboot with my RX 7800 XT and kernel 6.7. I can also confirm that I do NOT have any issues with the 6.6 LTS kernel.

Here is the log from one of my attempted reboots while running 6.7: https://0x0.st/HURG.txt

there are some pretty clear errors being thrown by amdgpu around line 1055.


I guess it will be fixed in some kernel 6.7.x update. After all, this is a new stable kernel series, fixes will come relatively soon.

Anyway, reporting to the linux kernel devs may help get the fix sooner. And there may already be a similar report, haven’t checked.

Yeah, there is a full thread on archlinux forums bisecting the issue (referenced by OP in first post) and the bug is already being passed upstream… https://gitlab.freedesktop.org/drm/amd/-/issues/3062

… my current work around is to just avoid rebooting. If I need a reboot, it’s just some extra time to shut down and do a cold start.

1 Like

I’d just set the LTS kernel as default until the fix is in place for kernel 6.7.


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