[GPU Lockup] Rare firmware robustness issues while recovering from faults

[Capta1nT0ad]

Hi,

I've just experienced a GPU lockup while moving a panel from the bottom to the top in Plasma (It appears I have also hovered over the 'Spacer' widget), after deleting a panel at the top. Before this, I was playing SuperTuxKart and SuperTux 2 while using the Falkon web browser. This was in a Wayland session. I cannot use Ctrl+Alt+number to change to a different TTY.

I was able to log into the system via SSH and run asahi-diagnose. Here is the output:

asahi-diagnose.txt

Here is how the screen looks:

I hope this helps.

[Capta1nT0ad]

When trying to reboot via SSH, this appears on the screen.

[asahilina]

Oh no! Thank you for this report. It looks like the firmware crashed... and I don't think I've ever seen this particular one before. Though I think I know where that fault address comes from, so I might be able to figure out what happened.

If you can figure out a reliable-ish way to reproduce it, that would be super helpful!

[Capta1nT0ad]

Oh no! Thank you for this report. It looks like the firmware crashed... and I don't think I've ever seen this particular one before. Though I think I know where that fault address comes from, so I might be able to figure out what happened.

If you can figure out a reliable-ish way to reproduce it, that would be super helpful!

I haven't found a way to reproduce it yet, even doing the exact thing I remember I was doing. However, the panel (and sometimes the whole screen) will briefly turn purple while trying to move the panel from the bottom to the top.

[asahilina]

The purple thing is a known issue, it affects other things too! But it shouldn't crash the firmware, so that worries me more...

[asahilina]

I managed to reproduce a (different!) crash by deliberately causing faults, so I'm going to investigate in that direction and see if I can figure out what's going wrong.