Rainer Dorsch composed on 2026-05-25 23:06 (UTC+0200):
> I run into an issue, which caused a hang of my Wayland session on
> Debian 13.
> journalctl revealed:
> radeon 0000:00:01.0: bo va 0x000010e594 conflict with (bo 0x000010e594
> 0x000010e5a3)
> (among other radeon reports).
> I have a Kabini APU (GCN 2nd gen / "Sea Islands"):
> rd@gigabyte:~$ lspci -nn | grep VGA 00:01.0 VGA compatible controller [0300]:
> Advanced Micro Devices, Inc. [AMD/ATI] Kabini [Radeon HD 8400 / R3 Series]
> [1002:9830] rd@gigabyte:~$
> I switched to the amdgpu kernel module:
> # echo "options amdgpu cik_support=1" | tee /etc/modprobe.d/amdgpu.conf
> # echo "options radeon cik_support=0" | tee -a /etc/modprobe.d/amdgpu.conf
> # update-initramfs -u -k all
> # reboot
> So far I did not observe any issues, it even resolved a framebuffer
> corruption
> issue after resume from suspend (had to switch to a virtual console and back
> to wayland to have a useful screen output again).
> Is amdgpu generally seen as the better kernel module for my hardware?
> If yes, would it make sense to make it the default in a stable Debian
> release?
> As a side note, I upgraded the system from previous Debian stable releases,
> so
> it might be that it is the default on new installations...
> If there is a better mailing list, please redirect the responses to that list.
On my GCN#2 GPUs my kernel cmdline parameters have included
radeon.cik_support=0 amdgpu.cik_support=1
since probably over a decade ago. Rumor had it that these were to have been made
default and thus unnecessary with kernel 6.19, but my experiences have mainly
suggested they haven't materialized as yet even in 7.0.
Running glmark2 with radeon vs. amdgpu kernel module in control it's been no
contest here — clearly the newer is the better.
--
Evolution as taught in public schools is, like religion,
based on faith, not based on science.
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata
