I’ve recently switched to kernel 6.12 because of drmModeAtomicCommit errors after the GNOME 48 update.
Yesterday I was playing a html game in Firefox when suddenly for the first time my system completely froze. I had no mouse cursor and I couldn’t even switch to a virtual terminal. The sound of the game played on though.
I didn’t know what to do so I ‘Magic SysRQed’ with REISUB and checked the journal. All I could see was:
kernel: amdgpu 0000:08:00.0: [drm] *ERROR* [CRTC:85:crtc-0] flip_done timed out
I’ve found a long running discussion in the AMD driver gitlab but no solution to fix the problem.
As the topic appears to be abandoned, I’ll close it to prevent any more people adding their own issues here as their hardware and software environment is not the same.
If you wish to have it re-opened, please PM @moderators and we’ll sort it, cheers.
For anyone reading this or being affected please consider reporting your issue in the linked AMD driver Gitlab rather than here in the Manjaro forums:
I can confirm I’m still affected by the random freezes every day even though kernel 6.12 as well as mesa and the linux firmware have been updated a couple of times.
According to a comment in the AMD driver gitlab, there’s hope, the bug/freezes might finally be fixed soon. Kernel 6.16rc6 with the added “VCN race fixes” is reported to run without any GPU freezes.
If this is indeed the fix for the problem, there are efforts being made to nominate the patches for inclusion into the stable kernels. I keep my fingers crossed.
I think we discussed this some place else but if I remember correctly, it was because you’re running an Intel-based chipset motherboard.
Based on the fact that the AMD Gitlab link I posted a second time in one of my last posts was still clicked over 50 times (just like the one in the first post), I think there’s a greater ‘dark figure’ of affected users simply not posting but reading the forum post.
And that’s just fine because I asked everyone to post any issues over in the AMD gitlab. I would like to assume this worked , because I am subscribed to a bunch of the AMD gitlab issues and especially #2950 was updated regularly with a constant stream of new users reporting the same issue.
EDIT (27/06/25):
When I read across multiple discussions in the AMD driver gitlab about the VNC race fixes, I really thought they finally found the ‘culprit’. It seems though, we shouldn’t get the hopes up too high because some people are still reporting freezes past kernel 6.16rc6.
I thought, I’ll add in two interesting links about the current development which I stumbled across in the AMD gitlab.
At this point you’re probably annoyed about reading of possible fixes while your screen still freezes daily (like mine), but there’s more progress/new hope, another ‘race condition’ might get fixed soon.
This time it’s in .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c (which a lot of the journals posted in the AMD gitlab point to as the source in the warning).
A new set of patches has been submitted on 30/07/2025, including one from ‘Aurabindo Pillai’ fixing a dmub access race condition. Since these have been submitted after the release of kernel 6.16, I’m not sure when/where the patches will finally appear, but they’re coming.
I still have my page flip errors regularly which means the issue hasn’t been fixed yet. Since I’m only affected once a day (lately it’s roughly just once every four days), I didn’t bother to disable any AMDGPU features like VRR, PSR or use any other GRUB commandline switches.
But I’ll take a closer look at these mitigations because atm I don’t know anything about it.
EDIT:
As far as I can see, the fixes I mentioned above haven’t been added to the stable kernel yet. I regularly check the changelog for kernel 6.12 on kernel.org but none of the fixes have been mentioned yet. No idea why it takes so long or why they refuse to add these VRR fixes to the stable kernel.
Just to be clear: I had them turned off and ran into this issue the first time, multiple time a day. Getting back to the default with all mitigations turned on, the system is stable again.
, because I am subscribed to a bunch of the AMD gitlab issues and especially 
EDIT (27/06/25):
It seems though, we shouldn’t get the hopes up too high because some people are 
(like mine), but there’s more progress/new hope,