View previous topic :: View next topic |
Author |
Message |
dede11 n00b

Joined: 11 Mar 2025 Posts: 2
|
Posted: Tue Mar 11, 2025 4:57 pm Post subject: integrated gpu not working amd ryzen 9 7950x3d |
|
|
Hello,
Trying to do gpu passthrough, but I am stuck at the same problem for 2 months.
I just need integrated GPU working, i will figure out the rest.
Integrated gpu is not fully loaded, in practice, that means when I tried to start for example hyprland, or any desktop environment, it crashes.
how i approached this problem.
I tried looking for similar problems online, I found some things like adding amdgpu.dc=0 to grub command line.
That caused hyprland to not crash the whole operating system, just program in terminal crashed and i got some output from hyprland itself that pointed me in some direction, complaining about amd gpu not being valid drm device. That was weird to me because the amd igpu showed up in /dev/dri as card0
So I tried getting some help on IRC
Some kind person there told me, that I should see following line in the dmesg, but i dont and thats why hyprland is complaining.
#amdgpu 0000:XX:XX.0: [drm] fb0: amdgpudrmfb frame buffer device
(copied from different thread)
I was told that this line is missing because nvidia card is somehow stealing the control over frame buffer device before integrated amd gpu gets chance to get it (forgive me for not using correct words)
But how can that be if the card is already blacklisted?
This kind person who had many years of experience with working with gentoo also sent me a custom crafted .config file cleaned of useless things based on my file that I sent.
That also difnt fix the problem, but at least I got the penguin logo working during boot
So, month later, I found a youtube tutorial:
https://www.youtube.com/watch?v=MaWkeq6aduc
the tutorial contained one interesting information,
If I try to force loading of vfio-pci driver into nvidia card by dracut initramfs, before main kernel loads, that could help.
I followed that youtube tutorial, and i used a script that put "vfio-pci" string into /sys/bus/pci/devices/0000:00:01:00.0/driver_override
That also didnt fix the problem, so im out of ideas. The only difference is, now the screen freezes and i cant even start hyprland. But I can still ssh into device and thats how I obtained the dmesg log.
Here is the dmesg im currently seeing -> https://logpaste.com/49dV7Vgz
here is the current .config file sent from that person ( i didnt edit it) -> https://logpaste.com/1v3VAZ6R
Another small detail,
I was told the file /proc/fb should contain some string related to amd.
Right now, when i do cat /proc/fb the file is empty
Before I forced loading of vfio-pci driver with dracut, the contents of /proc/fb was vg0
So, question is, can anyone help me to get the integrated GPU to work |
|
Back to top |
|
 |
NeddySeagoon Administrator


Joined: 05 Jul 2003 Posts: 55011 Location: 56N 3W
|
Posted: Tue Mar 11, 2025 7:03 pm Post subject: |
|
|
dede11,
Welcome to Gentoo.
Your kernel command line bothers me dmesg: | kernel-6-6-74-gentoo.img root=/dev/mapper/vg0-root ro rd.driver.pre=vfio-pci rd.luks.uuid=c32b911d-4824-4e81-b7de-c6195bb56303 rd.luks rd.lvm rd.lvm.vg=vg0 rootfstype=ext4 nvidia_drm.fbdev=0 root=/dev/mapper/vg0-root amd_iommu=on iommu=pt amdgpu.dc=0 |
From modinfo amdgpu
Code: | ...
dc:Display Core driver (1 = enable, 0 = disable, -1 = auto (default)) (int)
... |
That means amdgpu.dc=0 turns off the amdgpu driver.
Take the amdgpu.dc=0 away from the kernel command line. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
 |
dede11 n00b

Joined: 11 Mar 2025 Posts: 2
|
Posted: Tue Mar 11, 2025 8:35 pm Post subject: |
|
|
I just removed the amdgpu.dc=0, rebooted, glanced at the dmesg and remembered why I added it there in first place
now the dmesg is a bit different
[ Code: | 93.128447] amdgpu 0000:14:00.0: amdgpu: ring jpeg_dec uses VM inv eng 5 on hub 8
[ 93.129008] [drm] Initialized amdgpu 3.54.0 20150101 for 0000:14:00.0 on minor 0
[ 93.131520] amdgpu 0000:14:00.0: [drm] Cannot find any crtc or sizes
[ 93.131531] [drm] DSC precompute is not needed. |
thats why I added amdgpu.dc=0 in first place, i was trying to solve this
[drm] Cannot find any crtc or sizes
I have no idea what this means but some people online suggested adding amdgpu.dc=0 to command line to fix it
Take a look here
https://forum.manjaro.org/t/what-is-kernel-parameter-amdgpu-dc-0-doing-and-why-do-i-suddenly-need-it/32813/7
so maybe the kernel is broken?
edit: here is new dmesg https://logpaste.com/6pcLZJoJ
Code tags added by NeddySeagoon |
|
Back to top |
|
 |
NeddySeagoon Administrator


Joined: 05 Jul 2003 Posts: 55011 Location: 56N 3W
|
Posted: Tue Mar 11, 2025 9:05 pm Post subject: |
|
|
dede11,
Code: | [ 93.131520] amdgpu 0000:14:00.0: [drm] Cannot find any crtc or size |
The crtc is the Cathode Ray Tube Controller. It was named a long time ago. That's the bit of the GPU that drives the display.
Is unlikely to be a CRT today.
Sizes are usually provided bi the EDID data that is read from the display.
Pastebin all of dmesg please.
What display (make and model) is connected?
How is it connected?
== edit ==
Your 6.6 kernel predates your hardware but your IGPU has been around for a few years.
It won't do any harm to try the 6.12 kernel, which is the new stable.
== edit some more ==
Code: | CONFIG_VIDEO_NOMODESET=y | should be off. AMDGPU is a modesetting driver.
If you need to force that option for debug, there is a kernel command line option. From memory, nomodeset. It's only for debug as it prevents all the modestting drivers from working. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
 |
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|