Plasma 6.7.x Wayland is failing to start on RX 580 (AMDGPU DRM device busy)

I tried upgrading my system from Plasma 6.6.5 to Plasma 6.7.x multiple times (first attempt with 6.7.0 before restoring, and just recently to 6.7.2 before restoring yet again). Each time I upgraded, Wayland stopped working.

Operating System: KDE neon User Edition
KDE Plasma Version: 6.6.5
KDE Frameworks Version: 6.26.0
Qt Version: 6.11.1
Kernel Version: 6.17.0-35-generic (64-bit)
Graphics Platform: Wayland
Processors: 12 × AMD Ryzen 5 5600G with Radeon Graphics
Memory: 32 GiB of RAM (31.1 GiB usable)
Graphics Processor: AMD Radeon RX 580 Series

Secondary GPU: NVIDIA GTX 750 Ti reserved for VFIO passthrough

Plasma 6.6.5 works normally with Wayland. 6.7.x doesn’t.

Problems:

  • Plasma Wayland session starts loading

  • Screen remains blank

  • After waiting, Plasma Wayland exits/shuts down (when running startplasma-wayland)

  • Plasma X11 session works (with 6.7.2; didn’t with 6.7.0, only showing the cursor)

I tested:

  • Removing VFIO kernel parameters (vfio-pci.ids=10de:1380,10de:0fbc) to see if there was some kind of conflict

  • Booting without the passthrough configuration

  • Creating a clean Plasma user profile (after logging into an X11 session)

The issue remained.

Relevant logs:

kwin_wayland:
Failed to open /dev/dri/card1 device (Device or resource busy)

Failed to open drm device /dev/dri/card1

No suitable DRM devices have been found

The device looks like it’s correctly detected:

/dev/dri/by-path/
pci-0000:01:00.0-card -> ../card1
pci-0000:01:00.0-render -> ../renderD128

DRIVER=amdgpu
PCI_ID=1002:67DF

X11 starts successfully and uses the same GPU (this wasn’t the case with 6.7.0).

Downgrading back to Plasma 6.6.5 restores Wayland functionality.

This seems to be a regression introduced somewhere between Plasma 6.6.5 and 6.7.x, possibly related to KWin DRM device selection or DRM device locking.

Anyway, I am done. I have a huge headache over trying to figure out why this upgrade doesn’t work for me.

Funny thing is, I also tested the same install I have on another PC, by restoring my desktop’s backup on a separate laptop using Intel graphics, and Plasma Wayland started normally there. This makes me suspect the issue is specific to the AMDGPU/AMD graphics stack rather than a general Plasma configuration issue.

Does anyone know the solution? Or is this something I need to wait out?

This might sound stupid but I had a similar error Failed to open /dev/dri/card1 device (Device or resource busy) some weeks ago after upgrading to plasma 6.7.x, and on Fedora back then the package akmod-nvidia-open was old in version (meta package which installs the actual Nvidia packages), while the actual driver component packages were newer, this lead to some bespoke scenario where the Plasma upgrade failed to boot properly and I had the exact same error.

The solution was to completely wipe out the Nvidia driver as documented in the RPM Fusion, and reinstall the coherent/compatible versions again.

I don’t know how the GPU driver packages are handled on KDE neon User Edition, but see if the packages have the same version (here for ex on Nvidia they’re all 595.80).

$ dnf list --installed '*nvidia*'

akmod-nvidia-open.x86_64                      3:595.80-1.fc44
kmod-nvidia-7.1.3-200.fc44.x86_64.x86_64      3:595.80-1.fc44 
kmod-nvidia-open-7.1.3-200.fc44.x86_64.x86_64 3:595.80-1.fc44 
nvidia-modprobe.x86_64                        3:595.80-1.fc44 
nvidia-persistenced.x86_64                    3:595.80-1.fc44 
nvidia-settings.x86_64                        3:595.80-1.fc44 
xorg-x11-drv-nvidia.x86_64                    3:595.80-1.fc44 
xorg-x11-drv-nvidia-cuda.x86_64               3:595.80-1.fc44 
xorg-x11-drv-nvidia-cuda-libs.i686            3:595.80-1.fc44 
xorg-x11-drv-nvidia-cuda-libs.x86_64          3:595.80-1.fc44 
xorg-x11-drv-nvidia-kmodsrc.x86_64            3:595.80-1.fc44 
xorg-x11-drv-nvidia-libs.i686                 3:595.80-1.fc44 
xorg-x11-drv-nvidia-libs.x86_64               3:595.80-1.fc44 
xorg-x11-drv-nvidia-power.x86_64              3:595.80-1.fc44 

Well, my GPU isn’t Nvidia. It’s AMD. I think there is an issue with the current version of Mesa being paired with this Plasma desktop upgrade. I’m not sure.

The latest KDE neon live ISO boots Plasma 6.7 normally on my rig with the RX580, but upgrading my installed system from Plasma 6.6.x to 6.7.x consistently results in a black screen.

So, I don’t know what’s going on. I’m just going to leave it.

Maybe after a week, I’ll make another backup and then try the upgrade again when 6.7.3 drops.

I don’t think this is a binding issue with VFIO regarding my GPU pass through setup with the secondary Nvidia card. But I’ve added these to /etc/initramfs-tools/modules file just in case:


vfio

vfio_pci

vfio_iommu_type1

vfio_virqfd

But if 6.7.3 fixes this problem, I know it was likely the KWin/DRM side of things that was giving me trouble.

Yup. Happens on an AMD laptop with an exact copy of my system drive with 6.7.3 also:

Jul 16 13:36:47 my-pc kernel: ACPI: bus type drm_connector registered
Jul 16 13:36:47 my-pc kernel: simple-framebuffer simple-framebuffer.0: [drm] Registered 1 planes with drm panic
Jul 16 13:36:47 my-pc kernel: [drm] Initialized simpledrm 1.0.0 for simple-framebuffer.0 on minor 0
Jul 16 13:36:47 my-pc kernel: simple-framebuffer simple-framebuffer.0: [drm] fb0: simpledrmdrmfb frame buffer device
Jul 16 13:36:47 my-pc systemd[1]: Starting modprobe@drm.service - Load Kernel Module drm...
Jul 16 13:36:48 my-pc systemd[1]: modprobe@drm.service: Deactivated successfully.
Jul 16 13:36:48 my-pc systemd[1]: Finished modprobe@drm.service - Load Kernel Module drm.
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] ATOM BIOS: 113-C91400-005
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] UVD is enabled in physical mode
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] VCE enabled in VM mode
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] Detected VRAM RAM=512M, BAR=512M
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] RAM width 64bits UNKNOWN
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] GART: num cpu pages 262144, num gpu pages 262144
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] PCIE GART of 1024M enabled (table at 0x000000F400600000).
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] Found UVD firmware Version: 1.43 Family ID: 15
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] UVD ENC is disabled
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: values for Engine clock
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 200000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 300000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 400000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 496560
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 600000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 685720
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 757900
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 800000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: Validation clocks:
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: engine_max_clock: 80000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: memory_max_clock: 93300
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: level : 8
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: values for Display clock
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 300000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 400000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 496560
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 626090
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 685720
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 757900
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 800000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 847060
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: Validation clocks:
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: engine_max_clock: 80000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: memory_max_clock: 93300
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: level : 8
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: values for Memory clock
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 667000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: 933000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: Validation clocks:
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: engine_max_clock: 80000
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: memory_max_clock: 93300
Jul 16 13:37:13 my-pc kernel: [drm] DM_PPLIB: level : 8
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] Display Core v3.2.369 initialized on DCE 11.0
Jul 16 13:37:13 my-pc kernel: amdgpu 0000:00:01.0: [drm] UVD initialized successfully.
Jul 16 13:37:14 my-pc kernel: amdgpu 0000:00:01.0: [drm] VCE initialized successfully.
Jul 16 13:37:14 my-pc kernel: amdgpu 0000:00:01.0: [drm] Using custom brightness curve
Jul 16 13:37:14 my-pc kernel: amdgpu 0000:00:01.0: [drm] Registered 2 planes with drm panic
Jul 16 13:37:14 my-pc kernel: [drm] Initialized amdgpu 3.64.0 for 0000:00:01.0 on minor 1
Jul 16 13:37:14 my-pc kernel: fbcon: amdgpudrmfb (fb0) is primary device
Jul 16 13:37:14 my-pc kernel: amdgpu 0000:00:01.0: [drm] fb0: amdgpudrmfb frame buffer device
Jul 16 13:38:00 my-pc kwin_wayland[1828]: No backend specified, automatically choosing drm
Jul 16 13:38:00 my-pc kwin_wayland[1828]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:01 my-pc kwin_wayland[1898]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:01 my-pc kwin_wayland[1898]: No backend specified, automatically choosing drm
Jul 16 13:38:01 my-pc kwin_wayland[1898]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:01 my-pc kwin_wayland[1907]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:01 my-pc kwin_wayland[1907]: No backend specified, automatically choosing drm
Jul 16 13:38:01 my-pc kwin_wayland[1907]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:01 my-pc kwin_wayland[1919]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:01 my-pc kwin_wayland[1919]: No backend specified, automatically choosing drm
Jul 16 13:38:01 my-pc kwin_wayland[1919]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:02 my-pc kwin_wayland[1923]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:02 my-pc kwin_wayland[1923]: No backend specified, automatically choosing drm
Jul 16 13:38:02 my-pc kwin_wayland[1923]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:02 my-pc kwin_wayland[1926]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:02 my-pc kwin_wayland[1926]: No backend specified, automatically choosing drm
Jul 16 13:38:02 my-pc kwin_wayland[1926]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:02 my-pc kwin_wayland[1934]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:02 my-pc kwin_wayland[1934]: No backend specified, automatically choosing drm
Jul 16 13:38:02 my-pc kwin_wayland[1934]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:03 my-pc kwin_wayland[1941]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:03 my-pc kwin_wayland[1941]: No backend specified, automatically choosing drm
Jul 16 13:38:03 my-pc kwin_wayland[1941]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:03 my-pc kwin_wayland[1943]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:03 my-pc kwin_wayland[1943]: No backend specified, automatically choosing drm
Jul 16 13:38:03 my-pc kwin_wayland[1943]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:03 my-pc kwin_wayland[1945]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:03 my-pc kwin_wayland[1945]: No backend specified, automatically choosing drm
Jul 16 13:38:03 my-pc kwin_wayland[1945]: Failed to open drm device /dev/dri/card1
Jul 16 13:38:03 my-pc kwin_wayland[1947]: Failed to open drm node /dev/dri/renderD128: Permission denied
Jul 16 13:38:03 my-pc kwin_wayland[1947]: No backend specified, automatically choosing drm
Jul 16 13:38:03 my-pc kwin_wayland[1947]: Failed to open drm device /dev/dri/card1

But 6.7.3 works on the Intel laptop.

Until, my current install boots up without an issue on the AMD laptop, I am not ever thinking of updating to 6.7.x. So, I guess I safely rule out that it’s a potential VFIO issue.

Wayland chokes on 6.7.3 for me too, and I am using Nvidia. For some reason reverting to nouveau results in my wifi driver failing to work. Since the main devel left due to what sounded like silly politics, NEON has been a bit of a disaster.

I really hate the idea of moving over to another distro. I’m only going to do it when the Ubuntu base gets really bad. Then maybe I’ll go to a less bleeding edge KDE distro. There are times why I wonder why I went with Neon.

But the real reason? Neon feels like a clean install of Windows 7 when you first install it. Maybe that’s why.

In any case, until this possible regression is fixed, I’m fine staying on 6.6.5. I’m not chasing after new features these days. Maybe that explains why I went back to the flip phone and MP3 player, huh? I’m tired of keeping up with tech; literally took away 15 to 20 years of my life away.