SDDM and AMD + NVIDIA GPU not working

llinux · September 10, 2026, 1:49pm

Hello

My laptop is an ASUS TUF GAMING A18 with integred AMD gpu and a NVIDIA card.

I have installed Opensuse Tumbleweed without problem.

That works with KDE Plasma Wayland and use the iGPU AMD.

Then i have installed the NVIDIA driver 595 from Nvidia and reboot.

After reboot, only black screen with mouse.

No possibilty to access a command line and CTRL + ALT + SUP does not work.

Then i have enabled dGPU as display mode in the bios and then i had received the login screen of SDDM.

To proof it’s a SDDM problem, i have installed GDM with Gnome environment.

Reboot and return in the bios to enable DYNAMIC as display mode.

Now i can have a login screen and Gnome, i can use apps with iGpu using AMD gpu AND can now click right on an app game to select the dGPU Nvidia.

But i can’t have the KDE environment : GDM give not the possibility.

Has anyone a solution to have SDDM working ?

nite_route · September 10, 2026, 4:26pm

yeah hybrid ASUS + fresh nvidia black-screening SDDM is a pain. GDM working on DYNAMIC while SDDM only works with dGPU forced usually means the greeter grabbed the wrong GPU.

on tumbleweed first thing I’d try is force SDDM onto X11 for a bit. drop a file in /etc/sddm.conf.d/ with:


[General]
DisplayServer=x11
```

reboot with DYNAMIC back on. if the greeter comes up you can still pick a Plasma Wayland session after login.

also make sure you're on the openSUSE nvidia packages (not a random .run) and `nvidia_drm.modeset=1` is on the cmdline. if X11 greeter still blacks out, check `/var/log/Xorg.0.log` right after for which PCI id it bound.

llinux · September 10, 2026, 5:40pm

@nite_route

Thanks for your reply.

And sorry for my bad english (2 years learning)

I had add “DisplayServer=x11” as you resquested.

But this don’t work.

I have make a reboot in dGPU mode display and i have see i have only the option “wayland” or “gnome” on the login screen of SDDM.

No X11 option

So i think it’s why it’s not installed.

Here my sddm.conf :

[General]
DisplayServer=x11
HaltCommand=/usr/bin/systemctl poweroff
InputMethod=
Namespaces=
Numlock=on
RebootCommand=/usr/bin/systemctl reboot

[Autologin]
Relogin=false
Session=
User=

[Theme]
Current=
CursorTheme=
DisableAvatarsThreshold=7
EnableAvatars=false
FacesDir=/usr/share/sddm/faces
Font=
ThemeDir=/usr/share/sddm/themes

[Users]
DefaultPath=/usr/local/sbin:/usr/local/bin:/usr/bin
HideShells=
HideUsers=
MaximumUid=60513
MinimumUid=1000
RememberLastSession=true
RememberLastUser=true
ReuseSession=true

[Wayland]
EnableHiDPI=false
SessionCommand=/usr/share/sddm/scripts/wayland-session
SessionDir=/usr/share/wayland-sessions
SessionLogFile=.local/share/sddm/wayland-session.log

[X11]
DisplayCommand=/usr/share/sddm/scripts/Xsetup
DisplayStopCommand=/usr/share/sddm/scripts/Xstop
EnableHiDPI=false
MinimumVT=1
ServerArguments=-nolisten tcp
ServerPath=/usr/bin/X
SessionCommand=/usr/share/sddm/scripts/Xsession
SessionDir=/usr/share/xsessions
SessionLogFile=.local/share/sddm/xorg-session.log
UserAuthFile=.Xauthority
XauthPath=/usr/bin/xauth
XephyrPath=/usr/bin/Xephyr

llinux · September 10, 2026, 5:57pm

@nite_route

If you will suggest to install X11, it’s not possible : Tumbleweed has no package name X11-video-nvidiaG07.

Only nvidia G05 and before.

Krovikan · September 10, 2026, 8:49pm

Gnome dropped the X11 session with release 50:

Plasma 6.8 will drop the X11 session too. Launch is in October 14th.

Regards

llinux · September 10, 2026, 9:23pm

@nite_route @Krovikan

I have made one other test : when start the boot, i added “systemd.unit=multi-user.target” to the boot command line.

I logged in as the “lorraine” user, then ran the command “sudo systemctl start sddm” and… I saw the login screen and logged into KDE.

Now, we know that:

SDDM doesn’t start properly when simply booting (black screen + mouse)
SDDM starts correctly when running the command “sudo systemctl start sddm”

But what’s the problem?

Krovikan · September 10, 2026, 11:36pm

I use the 610 series drivers (features branch, have one new driver from yesterday: first of 615 series) and months ago I used the 595 series.

These is my parameters in the kernel (yes, I use openSUSE TW):

I use the .run file. To install I need to clear in Grub (with e key) the parameter initcall_blacklist=sysfb_init, SDDM have a big resolution so you need to change to a console (Ctrl + Alt + F1 for example), and with root install the driver.

After reboot without clear any line in Grub (simply Enter in Grub) and it is all for rebuild the kernel when change the version every week.

You can use the repo Nvidia, but I don’t know about this.

Operating System: openSUSE Tumbleweed 20260904
KDE Plasma Version: 6.7.4
KDE Frameworks Version: 6.29.0
Qt Version: 6.11.2
Kernel Version: 7.2.2-1-default (64-bit)
Graphics Platform: Wayland
Processors: 14 × Intel® Core™ Ultra 5 245K
Memory: 32 GiB of RAM (30.9 GiB usable)
Graphics Processor: NVIDIA GeForce RTX 5070

You use Fish as default shell?

Regards

barryascott · September 11, 2026, 8:13am

Is this the recommended way to add nvidia drivers on tumbleweed?

For Fedora we advise to use nvidia drivers from fedora rmpfusion as the drivers from nvidia.com often cause issues.

595 are old drivers.

llinux · September 11, 2026, 9:48am

@Krovikan @barryascott

I realize there’s a lot I don’t know, since I’ve never faced so many problems—or ones of this level—before.

So, I apologize and thank you for your patience.

In response to @Krovikan, below is a screenshot of my Yast/bootloader.

As you can see, the resolution isn’t listed, even though my system is up to date (the last update was yesterday, September 10, 2026).

My shell is bash.

inxi -Ga
Graphics:
  Device-1: NVIDIA GB206M [GeForce RTX 5070 Max-Q / Mobile] vendor: ASUSTeK
    driver: nvidia v: 595.99.02 alternate: nouveau,nvidia_drm
    non-free: 550-6xx.xx+ status: current (as of 2026-07) arch: Lovelace
    code: AD1xx process: TSMC n4 (5nm) built: 2022+ pcie: gen: 1
    speed: 2.5 GT/s lanes: 8 link-max: gen: 5 speed: 32 GT/s lanes: 16 ports:
    active: none empty: DP-9,HDMI-A-1,eDP-2 bus-ID: 01:00.0 chip-ID: 10de:2d58
    class-ID: 0300
  Device-2: Advanced Micro Devices [AMD/ATI] HawkPoint1 vendor: ASUSTeK
    driver: amdgpu v: kernel arch: RDNA-3 code: Phoenix process: TSMC n4 (4nm)
    built: 2023+ pcie: gen: 4 speed: 16 GT/s lanes: 16 ports: active: eDP-1
    empty: DP-1, DP-2, DP-3, DP-4, DP-5, DP-6, DP-7, DP-8, Writeback-1
    bus-ID: 65:00.0 chip-ID: 1002:1900 class-ID: 0300 temp: 34.0 C
  Device-3: Shinetech ASUS FHD webcam driver: uvcvideo type: USB rev: 2.0
    speed: 480 Mb/s lanes: 1 mode: 2.0 bus-ID: 3-1:2 chip-ID: 3277:0096
    class-ID: fe01 serial: 200901010001
  Display: wayland server: X.org v: 1.21.1.24 with: Xwayland v: 24.1.13
    compositor: kwin_wayland driver: X: loaded: modesetting,nvidia
    unloaded: vesa alternate: fbdev,nouveau,nv dri: radeonsi gpu: amdgpu
    display-ID: 0
  Monitor-1: eDP-1 model: AU Optronics B180UAN01.2 built: 2024 res:
    mode: 1920x1200 hz: 144 scale: 100% (1) dpi: 126 gamma: 1.2
    size: 387x242mm (15.24x9.53") diag: 456mm (18") ratio: 16:10 modes:
    max: 1920x1200 min: 640x480
  API: EGL v: 1.5 hw: drv: nvidia drv: amd radeonsi platforms: device: 0
    drv: nvidia device: 2 drv: radeonsi device: 3 drv: swrast gbm: drv: nvidia
    surfaceless: drv: nvidia wayland: drv: radeonsi x11: drv: radeonsi
    inactive: device-1
  API: OpenGL v: 4.6.0 compat-v: 4.6 vendor: amd mesa v: 26.2.2 glx-v: 1.4
    direct-render: yes renderer: AMD Radeon 780M Graphics (radeonsi phoenix ACO
    DRM 3.64 7.2.3-1-default) device-ID: 1002:1900 memory: 500 MiB
    unified: yes display-ID: :1.0
  API: Vulkan v: 1.4.357 layers: 4 device: 0 type: integrated-gpu name: AMD
    Radeon 780M Graphics (RADV PHOENIX) driver: mesa radv v: 26.2.2
    device-ID: 1002:1900 surfaces: N/A device: 1 type: discrete-gpu
    name: NVIDIA GeForce RTX 5070 Laptop GPU driver: nvidia v: 595.99.02
    device-ID: 10de:2d58 surfaces: N/A device: 2 type: cpu name: llvmpipe
    (LLVM 23.1.0 256 bits) driver: mesa llvmpipe v: 26.2.2 (LLVM 23.1.0)
    device-ID: 10005:0000 surfaces: N/A
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
    de: kscreen-console,kscreen-doctor gpu: gputop, intel_gpu_top, lsgpu,
    nvidia-smi wl: wayland-info x11: xdpyinfo, xprop, xrandr

If it’s an NVIDIA driver issue and I need to update it, fine.

But how do I uninstall it (NVIDIA G07 or 595.99), remove the corresponding repository, and add the CUDA repository that has driver 610 or 615?

And I also need to know how to determine whether my graphics card supports 615 or only 610.

Krovikan · September 11, 2026, 10:14am

5070 Mobile or Desktop use the same drivers version.

Nvidia have problems with from kernel from 7.0 (someone says that only with openSUSE).

My tip is add initcall_blacklist=sysfb_init in the parameters after the rd parameter. You can add nvidia-drm.modeset=1 also.
Or better, try to add only nvidia-drm.modeset=1 and if don’t work try adding in addition initcall_blacklist=sysfb_init.

You need, after the tries proposed with 595, 610 driver (or 615 soon).

CUDA drivers are the G07. One zypper search:

Check you have the kernel of the driver. I see kernel 7.2.0 in CUDA driver, not 7.2.3 that is the current. ISOs go more slowly to change this.

Maybe adding the repo Nvidia in Repositories management of Yast2 (is in Comunitary repos).

The G07 non CUDA are 595.

Regards

llinux · September 11, 2026, 11:42am

@Krovikan

An advanced user from Tumbleweed says : “The “stable” branch for RTX5070 is still 595.99, the “New Feature” branches (610, 615) are for developers and early adopters.
I think your system is working and you may not wish to go down other rabbit holes with still beta drivers”

I will keep 595 for the moment.

Krovikan · September 11, 2026, 1:52pm

Is true. I mentioned here:

In stable branch is 595 yet and in features branch 610 and 615.

But you can try your 595 (as I mentioned in my previous post) adding first nvidia-drm.modeset=1 in kernel parameters and reboot, and if don’t work add (without clear nvidia-drm.modeset=1) initcall_blacklist=sysfb_init.

Regards

llinux · September 11, 2026, 5:14pm

I have found the problem

I found this line in journalctl -b :

Socket to launch DrKonqi for a systemd-coredump crash skipped, unmet condition check ConditionUser=!@system

And fing this here : https://bugs.kde.org/show_bug.cgi?id=502960

Have installed lightdm and this will not work : same problem

llinux · September 14, 2026, 3:25pm

@Krovikan @nite_route

Solution found here

101 · September 26, 2026, 11:44pm

After upgrading my system, I also had this issue of sddm-greeter-qt crashing immediately (black screen, only mouse). For me, using a 7 second (or 70 sec) systemd timer didn’t work, and downgrading nvidia/linux/sddm/other also didn’t work.

What worked for me: I solved it by updating BIOS to the latest non-beta version. I am also using an ASUS motherboard, AMD CPU and NVIDIA GPU. The previous BIOS version that I had installed was about 0.5-1 year old.

(I am writing this 12 days after the previous comment in case it helps anyone that find this page from searching online like I did.)

llinux · September 27, 2026, 2:32pm

@101

Oh, my laptop is ASUS TUF GAMING A18, buy this month.

Is yours the same ?

And, Since I’m a very novice Linux user, could you tell me how to update the BIOS, in case my solution stops working?

Thanks in advance

101 · September 27, 2026, 4:48pm

@llinux

I’m on a desktop, so I don’t have this laptop.

I can’t help you with how to update the BIOS on your laptop.

101 · September 29, 2026, 4:47pm

Important note. I wrongly assumed this fully fixed the issue for me, but updating BIOS changed SDDM’s behavior from never working properly to working after waiting a few seconds before starting it. I didn’t realize this immediately.

llinux · September 29, 2026, 4:56pm

@101

If it now works after a few seconds before launch, perhaps this is now due to the delay set in the timer (7 seconds?).