Which Linux distro are you using?
Debian
Which release version?
(if rolling release without a release version, skip this question)
(If rolling release, last date updated?)
02 May 2025 testing Trixie
Which kernel are you using?
6.12.22
Which BIOS version are you using?
JFP30.03.08
Which Framework Laptop 13 model are you using? AMD Ryzen™ 7 7840U Radeon 780M
My laptop randomly crashes the GPU with a kernel panic. The os is still working but the display is frozen.
9 out of 10 times it happens within 5 minutes of usage. The usage might just be using a terminal or typing.
When I don’t do anything it is fine but when I start using the mouse or keyboard it crashes.
Here the dmesg output:
[ 257.890619] amdgpu 0000:c1:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
[ 258.166195] amdgpu 0000:c1:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
[ 258.441756] amdgpu 0000:c1:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
[ 268.561732] amdgpu 0000:c1:00.0: [drm] *ERROR* [CRTC:79:crtc-0] flip_done timed out
Important to know it is always a fresh boot.
Suggestions are welcome because the laptop is now unusable.
You kinda wanna add this kernel param to disable panel self-refresh (psr):
sudo grubby --update-kernel=ALL --args="amdgpu.dcdebugmask=0x10"
This is a long-standing issue that kinda flip-flops b/w fixed and broken. A search for amdgpu on the forum will reveal numerous threads, (most) all recommending adding the kernel param noted above.
The latest regression is being tracked here.
1 Like
Thanks will test if this solves it for now.
I ran into the exact same issue (FW 13 AMD, Manjaro 6.14, UEFI 3.07): starting with some visual artifacts, then persistent screen freezes that would require a reboot when having an extra screen connected (via USB-C) and a GPU-demanding window opened on both (browser, electron app, etc.) the laptop and the external monitor.
Random freezes would happen multiple times a day, one monitor would generally continue working while the other was stuck, but no way to “unfreeze” the frozen monitor, and ultimately it would crash the entire laptop.
The freezes would not happen if I had a terminal app open on one of the monitors - although the visual artifcats would.
Initially thought this was related to a regression in Mesa 25 but even with the workarounds described there and the latest kernel, it would still happen (although the visual artifacts I had previously are now gone).
Then I found this thread, followed the regression guidance (essentially adding amdgpu.dcdebugmask=0x10 to boot options) and the freeze did not reproduce after a few days of uptime.
This looks promising, thanks for the info @ehsanj !