Black screen on Nvidia after updating to 20260428

[Lioli7k]

Created a new general topic for this issue. Wanna see if AMD and Intel GPUs are affected by this.

[mrhine]

Hi,

I was having the same symptoms as the once described in your first message and, having a similar hardware, the same distro, kernel 7.0.3 and nvidia open driver G07, I was following the thread since the beginning.

I have also noticed the kernel message while activating the nvidia driver:

Unpatched return thunk in use. This should not happen!

mentioned in the boo#1263825

After having removed the nvidia open driver, the problem with SDDM stayed the same so I started to look elsewhere, finding the XSession log pointing to libcrypto.

It is possibile that, after having ruled out the driver activation problem, the libcrypto error in the Xsession ssh-agent activation script will manifest itself, as in my case.

[Lioli7k]

Alright. Let’s see if anyone else on X11 here have that libcrypto issue. I don’t remember anyone mentioning it so far.

[Lioli7k]

Got the same behaviour on 7.0.5 after updating to 20260509.

[FrankyU2]

Well, librcryto is working for me but the problem does become apparent around the time the partitions are being accessed in the boot process. I would guess systemd just happens to switch to framebuffer around this time by timing coincidence which is where the problem actually manifests. The switch to framebuffer is apparent because boot messages become bolder or brighter and the external monitor is activated. Or if I was on 7.0.3 this would be about the time the screen turns blue.

[conram]

The new kernel:7.0.5 doesn’t work with nvidia on my machine. I am using the .run driver.
I deleted the new kernel and revert back to the kernel:7.0.3
While installing the driver the last part is a message like “failed to build the initramfs”.
I tried building the initramfs manually but still no go. It boots then lost detection of the monitor.

[akontsevich]

7.0.5 works better for me: plymouth and sddm runs fine, however after login I constantly get KWin error notification every 30 secs: “The desktop effects settings were reset due to a graphics reset.” So definitely driver is not stable.

[Lioli7k]

I have found out that Arch Linux maintainers have made a patch during 7.0-rc1 phase. I’m gonna attempt porting it to OpenSUSE and test on my PC later.

[malcolmlewis]

@conram alas, all good here with the rpm;

(57/66) Installing: nvidia-open-driver-G07-signed-cuda-kmp-default-595.71.05_k7.0.5_1-2.2.x86_64

System:
  Kernel: 7.0.5-1-default arch: x86_64 bits: 64 compiler: gcc v: 15.2.1
    clocksource: tsc avail: hpet,acpi_pm
    parameters: initrd=\opensuse-tumbleweed\7.0.5-1-default\initrd-1dd5907d926a83557a0631bc6bf73d9af5e03ebb
    root=UUID=867ad03d-9a24-45c8-9e92-72a7f526fb5a splash=silent
    intel_iommu=on mitigations=auto quiet security=selinux fbdev=1
    nvidia_drm.modeset=1 rd.driver.blacklist=nouveau loglevel=2 psi=1
    rootflags=subvol=@/.snapshots/1/snapshot
  Desktop: GNOME v: 50.1 tk: GTK v: 3.24.52 wm: gnome-shell
    tools: gsd-screensaver-proxy dm: GDM v: 50.0 Distro: openSUSE Tumbleweed
    20260509
Graphics:
  Device-1: NVIDIA TU104GL [Quadro RTX 4000] vendor: Hewlett-Packard
    driver: nvidia v: 595.71.05 alternate: nouveau,nvidia_drm
    non-free: 550-580.xx+ status: current (as of 2025-11; EOL~2026-12-xx)
    arch: Turing code: TUxxx process: TSMC 12nm FF built: 2018-2022 pcie:
    gen: 1 speed: 2.5 GT/s lanes: 16 link-max: gen: 3 speed: 8 GT/s ports:
    active: DP-3 empty: DP-1,DP-2,Unknown-2 bus-ID: 65:00.0 chip-ID: 10de:1eb1
    class-ID: 0300
  Display: wayland server: Xwayland v: 24.1.11 compositor: gnome-shell
    driver: gpu: nv_platform,nvidia,nvidia-nvswitch display-ID: 0
  Monitor-1: DP-3 model: DT-156P-A1 serial: <filter> built: 2021 res:
    mode: 1920x1080 hz: 60 scale: 100% (1) dpi: 142 gamma: 1.2
    size: 340x190mm (13.54x7.64") diag: 395mm (15.5") ratio: 16:9 modes:
    max: 1920x1080 min: 640x480
  API: EGL v: 1.5 hw: drv: nvidia platforms: device: 0 drv: nvidia device: 2
    drv: swrast gbm: drv: nvidia surfaceless: drv: nvidia wayland: drv: nvidia
    x11: drv: nvidia inactive: device-1
  API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 595.71.05
    glx-v: 1.4 direct-render: yes renderer: Quadro RTX 4000/PCIe/SSE2
    memory: 7.81 GiB display-ID: :0.0
  API: Vulkan v: 1.4.341 layers: 3 device: 0 type: discrete-gpu
    name: Quadro RTX 4000 driver: nvidia v: 595.71.05 device-ID: 10de:1eb1
    surfaces: N/A device: 1 type: cpu name: llvmpipe (LLVM 22.1.4 256 bits)
    driver: mesa llvmpipe v: 26.0.6 (LLVM 22.1.4) device-ID: 10005:0000
    surfaces: N/A
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo gpu: gputop,
    intel_gpu_top, lsgpu, nvidia-smi wl: wayland-info x11: xprop

Did you try building the closed driver with the --kernel-module-type=proprietary option?

@Lioli7k that is for an old driver version?

[dart364]

Fine building the 580.159.03 .run driver here, can’t go any higher because of my old card. @conram does the nvidia-installer.log say anything interesting?

System:
  Kernel: 7.0.5-1-default arch: x86_64 bits: 64 compiler: gcc v: 15.2.1
    clocksource: tsc avail: hpet,acpi_pm
    parameters: BOOT_IMAGE=/boot/vmlinuz-7.0.5-1-default
    root=UUID=90e7b277-8c2a-4a99-8faa-8ba8efe76b20 plymouth.enable=0
    security=selinux selinux=0 noresume mitigations=auto
    systemd.show_status=yes
  Desktop: KDE Plasma v: 6.6.4 tk: Qt v: N/A info: frameworks v: 6.25.0
    wm: kwin_x11 tools: avail: gtklock, hypridle, hyprlock, lxlock, swayidle,
    swaylock, xscreensaver vt: 2 dm: 1: GDM v: 50.0 2: LightDM v: 1.32.0
    note: stopped 3: SDDM Distro: openSUSE Tumbleweed 20260509
Graphics:
  Device-1: NVIDIA GM107 [GeForce GTX 750 Ti] vendor: EVGA driver: nvidia
    v: 580.159.03 alternate: nouveau,nvidia_drm non-free: 550-580.xx+
    status: current (as of 2025-11; EOL~2026-12-xx) arch: Maxwell code: GMxxx
    process: TSMC 28nm built: 2014-2019 pcie: gen: 2 speed: 5 GT/s lanes: 16
    link-max: gen: 3 speed: 8 GT/s ports: active: none off: HDMI-A-1
    empty: DP-1,DVI-I-1 bus-ID: 01:00.0 chip-ID: 10de:1380 class-ID: 0300
  Device-2: Logitech HD Webcam C615 driver: snd-usb-audio,uvcvideo type: USB
    rev: 2.0 speed: 480 Mb/s lanes: 1 mode: 2.0 bus-ID: 1-1:2 chip-ID: 046d:082c
    class-ID: 0e02 serial: <filter>
  Display: x11 server: X.org with: Xwayland v: 24.1.11 compositor: kwin_x11
    driver: X: loaded: nvidia gpu: nvidia,nvidia-nvswitch display-ID: :0
    screens: 1
  Screen-1: 0 s-res: 1920x1080 s-dpi: 192 s-size: 254x143mm (10.00x5.63")
    s-diag: 291mm (11.48")
  Monitor-1: HDMI-A-1 mapped: HDMI-0 note: disabled
    model: Toshiba TOSHIBA-TV serial: <filter> built: 2012 res: mode: 1920x1080
    hz: 60 scale: 100% (1) dpi: 30 gamma: 1.2 size: 1600x900mm (62.99x35.43")
    diag: 1836mm (72.3") ratio: 16:9 modes: max: 1920x1080 min: 640x480
  API: EGL v: 1.5 hw: drv: nvidia platforms: device: 0 drv: nvidia device: 2
    drv: swrast gbm: drv: nvidia surfaceless: drv: nvidia x11: drv: nvidia
    inactive: wayland,device-1
  API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 580.159.03
    glx-v: 1.4 direct-render: yes renderer: NVIDIA GeForce GTX 750 Ti/PCIe/SSE2
    memory: 1.95 GiB
  API: Vulkan v: 1.4.341 layers: 4 device: 0 type: discrete-gpu name: NVIDIA
    GeForce GTX 750 Ti driver: nvidia v: 580.159.03 device-ID: 10de:1380
    surfaces: N/A device: 1 type: cpu name: llvmpipe (LLVM 22.1.4 256 bits)
    driver: mesa llvmpipe v: 26.0.6 (LLVM 22.1.4) device-ID: 10005:0000
    surfaces: N/A
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo de: kscreen-console,
    kscreen-doctor, xfce4-display-settings gpu: nvidia-settings,nvidia-smi
    wl: kanshi, swaymsg, wayland-info x11: xdriinfo, xdpyinfo, xprop, xrandr

[Lioli7k]

Yeah. Seems like it was applied to 595.58.03 and later disabled but not removed when 595.71.05 update dropped.

[conram]

@malcolmlewis

@malcolmlewis

Did you try building the closed driver with the --kernel-module-type=proprietary option?

Yep, so did this work for you?

[malcolmlewis]

@conram I have not, I’m trying to stick with the rpms at present… In the past I had an issue with building the open driver and switched to the closed which did build fine until a new nvidia driver dropped, but even that was only a few days…

[conram]

@malcolmlewis Ok, I will try the open driver tomorrow. I deleted the latest kernel.
Thanks

[malcolmlewis]

@conram There is a bug report… https://bugzilla.opensuse.org/show_bug.cgi?id=1264627

[pullasuti]

@Lioli7k

After few days of testing I believe I finally nailed it down. It is a GPU initialisation problem and it happens only on 7.0.x kernel.

I’ve tested pretty much everything I could think of. I even went as far as spinning up a CachyOS VM and compiling OpenSUSE kernel from source code just to build a Nvidia driver for OpenSUSE, without relying on OpenSUSE’s gcc compiler. No matter what I’ve tried outcome was the same… Until I tried to run sudo nvidia-smi --gpu-reset with no running graphical programs (nvidia-smi will refuse to reset GPU if anything is using it). Then all of the sudden I got a working system. Hopefully with this knowledge fix will be ready soon. This is for sure the wildest bug I ever dealt with.

Thanks for that. 7.0.5 update re-broke a working 4000-series installation (black screen with 20260428, but fixed with the first updates until now), and the gpu-reset indeed helps.

Unfortunately after 7.0.5 my second system green screening with a 5000 series card is now completely unresponsive, so can’t run the reset. Before I could ssh in, but now I guess just continue booting with LTS.

[LeonG]

Ah, as usual when I have any sort of problem with my system, it boils down to my NVidia card.

I’m on an NVidia 4070 Super. Did a distro upgrade, black screen after boot. There’s still input, but nothing displayed. GDM, Wayland, GNOME, yada yada.

Think I’m finally switching away from NVidia. Was going to wait and see, but every single time I have a problem with my system, it’s boiled down to NVidia being wonky.

[Zlinux1]

Perhaps replace sddm with gdm :
I was also experiencing crashes, black screens, KWin constantly resetting the visual effects, and the login screen crashing or failing to accept passwords.

My system is fully up to date with kernel 7.0.5 and the latest version of the Nvidia G06 driver.

Do you use SDDM to log in?

If so, try replacing it with GDM and then try logging in.

Why?

I was also having problems:

• crashes on the login screen

• KWin constantly resetting the visual effects

• KDE crashes.

Since I switched to GDM, I haven’t had any more problems.

I tested by switching back to SDDM: the problems started again.

So I switched back to GDM and haven’t had any problems since.

[opsusemaco]

That bug report seems to focus on systemd-boot…
I am very confused by this problem, is it sddm, systemd-boot, kernel, nvidia kernel modules, possibly combined with some particular hardware configuration? Too many variables.

I am also affected with the following config: Desktop PC with Tumbleweed Plasma (hence sddm), systemd-boot. My graphics card being now “legacy” (Maxwell), I am using G06 driver installed with zypper and the nvidia repositories, without tweaking. But maybe my problem is not the same. (?)

[Zlinux1]

What have you got to lose by giving it a go?

If it sorts out your problems, you’ll be happy.

If not, never mind – we’ll find a solution later.

I use Plasma (Wayland), the Nvidia G06 driver, and an Asus Nvidia RTX 5060 Ti graphics card.

Yes, I know, it’s one of the latest cards.

I can only say one thing: give it a go.