I can reproduce this as of linux-firmware-20241017. The “amdgpu.dcdebugmask=0x10” kernel param workaround works for now
1 Like
Occurs with my computer as well, AMD, Fedora 40, GNOME
Michael1 mentioned it occurs for them when watching video, and all the occurrences for me as well have been while watching youtube
just happened to me on Fedora 41 with gnome. Was on Youtube as well. Just applied the kernel param so hopefully that will be the end of it for now
Well, thought it was fixed after amdgpu firmware 20241017-2, but just had it happen twice tonight.
Oct 31 21:14:12 fedora kernel: amdgpu 0000:c1:00.0: [drm] ERROR dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
Oct 31 21:14:12 fedora kernel: amdgpu 0000:c1:00.0: [drm] ERROR dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
Oct 31 21:14:12 fedora kernel: amdgpu 0000:c1:00.0: [drm] ERROR dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
Oct 31 21:14:12 fedora kernel: sched: RT throttling activated
Oct 31 21:14:12 fedora kernel: amdgpu 0000:c1:00.0: [drm] ERROR dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
Oct 31 21:14:11 fedora kernel: amdgpu 0000:c1:00.0: [drm] ERROR dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
Oct 31 21:14:11 fedora kernel: amdgpu 0000:c1:00.0: [drm] ERROR dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
Forcing amdgpu reset as suggested upthread does make it go away without a restart.
This happens for me on arch / endevouros. So it’s not fixed in the bleeding edge either sadly.
Adding that this still occurs even with Fedora 41. No idea as to cause.
Just had it happen again with amdgpu.dcdebugmask=0x200 as param. Going to switch to 0x10 and see if that’s better I guess.
I haven’t seen anymore slowdowns and I’ve been using my fw daily for the last couple of weeks. Arch linux and all the latest fixes. I don’t know if I’m lucky or just have the right combination of fixes.
I’m on fully updated EndeavourOS but otherwise stock (I haven’t tried to fix it at all, I just reset the gfx card), and I got one this morning.
What various fixes have you implemented?
Nothing outrageous. Powertop auto tune. Latest bios. KDE plasma 6 on Wayland.
I put the laptop to sleep all the time, not hibernate, and it comes back everytime. I only noticed the slow down before plasma 6.2 and I haven’t seen it since.
Whenever the system becomes unresponsive aka slows down I open a terminal and run this to resolve. It closed pretty much everything, returns me to the login screen. I still restart though.
sudo cat /sys/kernel/debug/dri/1/amdgpu_gpu_recover
Just happened again to me on Fedora 41 with the “amdgpu.dcdebugmask=0x10” already being applied. Didn’t have any video or anything like that running in the foreground or background. Was just doing work on my machine sending emails.
Are you sure it was applied? /sys/module/amdgpu/parameters/dcdebugmask should contain 16.
If you need to recover without closing apps, you can hibernate (suspend to disk) and resume.
Apparently it’s not applied. dcdebugmask shows ‘0’. Can I just change it to ‘16’?? Haven’t ever applied a kernel param before so I clearly didn’t do it correctly 
No, the parameter isn’t writable and must be set when the module is loaded.
Happened to me as well, yesterday, for the first time. I was watching a video on the European Parliament website. Mouse cursor got incredibly slow and sluggish, I managed to shut down the system.
Framework 13" with AMD Ryzen 5 7640U.
Running Fedora Kinoite 41, with KDE Plasma 6.2.3.
I have the same issue but it only happens when I change Power Mode to something other than Performance. I upgraded from an Intel 11th gen board so maybe there’s some software discrepancy?
I’ve hit this exact same issue too. I’m running Fedora 40 with KDE, I’ve tried booting into all recent kernels (6.10.8, 6.10.12, 6.11.6) to no avail. As soon as I log in, the system grinds to a halt. I opened up the system monitor I’m seeing RAM usage go to max, at which point it freezes.
I’ve been meaning to update but I’ve installed AMD’s graphics drivers for Linux and haven’t had an issue since. More specifically I downloaded the .rpm file for AMD Radeon Graphics (not pro).
Let me know if that fixes things for you as well!
Thank you, but I forgot to mention that this is happening to me on an Intel Framework.