After I run sudo dnf upgrade -y
which installed kernel-0:6.15.3-200.fc42.x86_64
my system is not longer booting unless during boot I choose the previous kernel 6.14.9-300.fc42.x86_64
.
The log from failed boot has the following entry:
cze 24 12:48:05 demon kernel: amdgpu 0000:0a:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
During the upgrade the following two packages were updated (in total 688 packages were altered) which might be of interest here:
Upgrade amd-gpu-firmware-0:20250613-1.fc42.noarch External User updates
Upgrade amd-ucode-firmware-0:20250613-1.fc42.noarch External User updates
(…)
Replaced amd-gpu-firmware-0:20250509-1.fc42.noarch External User @System
Replaced amd-ucode-firmware-0:20250509-1.fc42.noarch External User @System
I have AMD Ryzen 5 7600 CPU with integrated gpu.
Has anyone experienced something like this?
4 Likes
gnwiii
(George N. White III)
June 24, 2025, 11:32am
2
Which amd gpu do you have?
Yes, I had the same problem and downgrading those two packages fixed it.
#metoo . Did you solve it? By downgrading these two libraries perhaps (as suggested by @Masoud )?
For the record my GPU is an integrated AMD Radeon Raphael.
In the meantime I’ve set installonly_limit
in /etc/dnf/dnf.conf to 10, so the further updates won’t delete the working 6.14.* kernel
It’s not a kernel problem. It even breaks 6.14 if you upgrade those two.
Mine is an integrated one as well (Ryzen 9 9900X)
Yeah, probably Raphael then.
I beg to differ. Probably
I have these installed with 6.14. kernel and it works
Installed packages
amd-gpu-firmware.noarch 20250613-1.fc42 <unknown>
amd-ucode-firmware.noarch 20250613-1.fc42 <unknown>
Or am I missing something?
3 Likes
I might be making a mistake then. I couldn’t get past Luks prompt last night with 6.14 and these two upgraded.
I need to figure out which one it is to report it back.
FYI: dnf will never remove the kernel you are running on.
You can set the kernel to boot from by default using sudo grubby
.
Find the index of the kernel
sudo grubby --info=ALL
Set the default to that index
sudo grubby --set-default-index=X
Beware that a limit of 10 may mean that you fill up your /boot and cause issues becuase of running out of free space.
1 Like
Just checked, it’s Granite Ridge [Radeon Graphics] (according to cpu-x)
Awesome. I reverted the limit back to default and set default to the last working one.
1 Like
Great, thanks for checking. Will include this on my report back.
I’ve also update BIOS - AGESA to ComboAM5 1.2.0.3e Patch A (fresh and latest available for my mobo) but it doesn’t make a difference.
I have AMD Ryzen 5 7600 CPU with integrated gpu:
$ kinfo
Operating System: Fedora Linux 42
KDE Plasma Version: 6.4.0
KDE Frameworks Version: 6.15.0
Qt Version: 6.9.1
Kernel Version: 6.14.9-300.fc42.x86_64 (64-bit)
Graphics Platform: Wayland
Processors: 12 × AMD Ryzen 5 7600 6-Core Processor
Memory: 16 GiB of RAM (14.8 GiB usable)
Graphics Processor: AMD Radeon Graphics
#metoo after the update and I cannot even boot on 6.14.x.
Tried the solution from Arch forum, didn’t work either.
Any suggestions before I spend my day re-installing everything?
Did you try downgrading those libraries?