Can't get wayland to work with nvidia 570; can't roll back

Hello all,

Decided to give the 570 driver another try this morning. BIG MISTAKE!

Any ideas as to how to get Gnome on wayland to work with nvidia 570? I can disable wayland, but Gnome on X11 has problems …

When I try to re-install the 550 driver, I get

File ‘./x86_64/nvidia-compute-G06-32bit-550.144.03-lp156.30.1.x86_64.rpm’ not found on medium ‘Index of /opensuse/leap/15.6

I doubt that you really need the 32 bit package, I would uninstall and lock all nvidia-XXX-32bit packages and then try again.
Or you may install the 550.142 suite that still provides a 32bit package.

Thank @OrsoBruno ,

I doubt I actually need the 32-bit package … never really understood (and still don’t) why they get installed.

Right now I am trying to get noveau to work (doubtful). I am sooooo tired of nvidia garbage … might just but an ati and be done.

@oxwrongagain Works fine here, just packaging and delays that sometimes hit the third party repository…

Wed Mar 19 17:00:57 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.133.07             Driver Version: 570.133.07     CUDA Version: 12.8     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA T400                    On  |   00000000:01:00.0  On |                  N/A |
| 38%   32C    P8            N/A  /   31W |     229MiB /   2048MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

Wed Mar 19 17:01:14 2025       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 570.133.07             Driver Version: 570.133.07     CUDA Version: 12.8     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  Quadro RTX 4000                On  |   00000000:06:00.0 Off |                  N/A |
| 30%   36C    P8              9W /  125W |     170MiB /   8192MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

Hi @malcolmlewis ,

Not sure what you are saying … anyway … I installed 550 without 32-bit (had to do from a chroot environment as I accidentally removed nvidia kernel firmware … (killed my box good :slight_smile: ) … and am once more back to flaky, but usable. The best I can hope for these days. I don’t know if the problem is gnome, suse, nvidia, or some combination. I do know that 156 is the second worst suse release ever, at least for gnome users.

Anyway, would love to know if you have any ideas about getting gnome+wayland to work with 570 drivers. X11+gnome on suse is terrible, especially under high load.

Against all my better judgement, I am will to use ‘zypper rl again’ :slight_smile:

@oxwrongagain Wayland for GNOME, X11 is slowly on the way out :wink: So you must have an older card if nvidia kernel firmware is needed?

I use the run files here, AKA Hard Way

System 1; Systemd-boot, SELinux, only Nvidia T400 with open driver via cuda run file and NVIDIA run file, just a test box

System:
  Kernel: 6.13.6-1-default arch: x86_64 bits: 64 compiler: gcc v: 14.2.1
    clocksource: tsc avail: hpet,acpi_pm
    parameters: initrd=\opensuse-tumbleweed\6.13.6-1-default\initrd-0af0b907969584c7e0811b246aa1772d7c9b76ca
    root=UUID=0da94659-09fd-42b9-ba37-30c22453aa19 splash=silent
    mitigations=auto quiet intel_iommu=on fbdev=1 nvidia_drm.modeset=1
    security=selinux selinux=1 enforcing=1
    systemd.machine_id=07091d80dcc24a858f8a1c4a8b1cd1c9
  Desktop: GNOME v: 47.5 tk: GTK v: 3.24.48 wm: gnome-shell
    tools: gsd-screensaver-proxy avail: xscreensaver dm: GDM v: 47.0
    Distro: openSUSE Tumbleweed 20250317
Graphics:
  Device-1: NVIDIA TU117GLM [Quadro T400 Mobile] driver: nvidia v: 570.133.07
    alternate: nouveau,nvidia_drm non-free: 550/565.xx+ status: current (as of
    2025-01; 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: none off: DP-1 empty: DP-2,DP-3
    bus-ID: 01:00.0 chip-ID: 10de:1fb2 class-ID: 0300
  Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.6
    compositor: gnome-shell driver: gpu: nvidia,nvidia-nvswitch display-ID: 0
  Monitor-1: DP-1 model: Dell P2018H serial: <filter> built: 2018
    res: 1600x900 dpi: 94 gamma: 1.2 size: 434x236mm (17.09x9.29")
    diag: 494mm (19.4") ratio: 16:9 modes: max: 1600x900 min: 640x480
  API: EGL v: 1.5 hw: drv: nvidia platforms: device: 0 drv: nvidia gbm:
    drv: nvidia surfaceless: drv: nvidia wayland: drv: nvidia x11: drv: nvidia
  API: OpenGL v: 4.6.0 vendor: nvidia v: 570.133.07 glx-v: 1.4
    direct-render: yes renderer: NVIDIA T400/PCIe/SSE2 memory: 1.95 GiB
    display-ID: :0.0
  API: Vulkan v: 1.4.309 layers: 1 device: 0 type: discrete-gpu
    name: NVIDIA T400 driver: N/A device-ID: 10de:1fb2
    surfaces: xcb,xlib,wayland
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
    gpu: nvidia-settings,nvidia-smi x11: xprop,xrandr

System 2; Grub, Apparmor, Primary system (Development), Prime Render Offload with proprietary driver via cuda run file and NVIDIA run file.

System:
  Kernel: 6.13.6-1-default arch: x86_64 bits: 64 compiler: gcc v: 14.2.1
    clocksource: tsc avail: hpet,acpi_pm
    parameters: BOOT_IMAGE=/boot/vmlinuz-6.13.6-1-default
    root=UUID=1ab675ed-aba2-4593-9ebc-1c893015b607 splash=silent quiet
    security=apparmor intel_iommu=on iommu=pt fbdev=1
    nvidia_drm.modeset=1 mitigations=auto
  Desktop: GNOME v: 47.5 tk: GTK v: 3.24.48 wm: gnome-shell
    tools: gsd-screensaver-proxy avail: xscreensaver dm: GDM v: 47.0
    Distro: openSUSE Tumbleweed 20250317
Graphics:
  Device-1: Intel DG2 [Arc A380] vendor: ASRock driver: i915 v: kernel
    alternate: xe arch: Xe-HPG code: Alchemist process: TSMC n6 (7nm)
    built: 2022+ pcie: gen: 1 speed: 2.5 GT/s lanes: 1 ports:
    active: DP-1,DP-2,DP-3 empty: HDMI-A-1, HDMI-A-2, HDMI-A-3, HDMI-A-4
    bus-ID: 04:00.0 chip-ID: 8086:56a5 class-ID: 0300
  Device-2: NVIDIA TU104GL [Quadro RTX 4000] vendor: Hewlett-Packard
    driver: nvidia v: 570.133.07 alternate: nouveau,nvidia_drm
    non-free: 550/565.xx+ status: current (as of 2025-01; 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: none empty: DP-4, DP-5, DP-6, Unknown-1 bus-ID: 06:00.0
    chip-ID: 10de:1eb1 class-ID: 0300
  Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.6
    compositor: gnome-shell driver: X: loaded: modesetting,nvidia unloaded: vesa
    alternate: fbdev,intel,nouveau,nv dri: iris gpu: i915 d-rect: 5760x2160
    display-ID: 0
  Monitor-1: DP-1 pos: bottom-c model: Sceptre F24 built: 2023 res:
    mode: 1920x1080 hz: 100 scale: 100% (1) dpi: 93 gamma: 1.2
    size: 530x300mm (20.71x11.65") diag: 604mm (23.8") ratio: 16:9 modes:
    max: 1920x1080 min: 640x480
  Monitor-2: DP-2 pos: primary,top-left model: Sceptre F24 built: 2023 res:
    mode: 1920x1080 hz: 100 scale: 100% (1) dpi: 93 gamma: 1.2
    size: 530x300mm (20.71x11.65") diag: 604mm (23.8") ratio: 16:9 modes:
    max: 1920x1080 min: 640x480
  Monitor-3: DP-3 pos: top-right model: Sceptre F24 built: 2023 res:
    mode: 1920x1080 hz: 100 scale: 100% (1) dpi: 93 gamma: 1.2
    size: 530x300mm (20.71x11.65") diag: 604mm (23.8") ratio: 16:9 modes:
    max: 1920x1080 min: 640x480
  API: EGL v: 1.5 hw: drv: intel iris drv: nvidia platforms: device: 0
    drv: nvidia device: 1 drv: iris gbm: drv: iris surfaceless: drv: nvidia
    wayland: drv: iris x11: drv: iris
  API: OpenGL v: 4.6.0 compat-v: 4.6 vendor: intel mesa v: 25.0.1 glx-v: 1.4
    direct-render: yes renderer: Mesa Intel Arc A380 Graphics (DG2)
    device-ID: 8086:56a5 memory: 5.81 GiB unified: no display-ID: :0.0
  API: Vulkan v: 1.4.309 layers: 4 device: 0 type: discrete-gpu
    name: Quadro RTX 4000 driver: N/A device-ID: 10de:1eb1
    surfaces: xcb,xlib,wayland device: 1 type: cpu name: llvmpipe (LLVM
    19.1.7 256 bits) driver: N/A device-ID: 10005:0000
    surfaces: xcb,xlib,wayland
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo gpu: gputop,
    intel_gpu_top, lsgpu, nvidia-settings, nvidia-smi wl: wayland-info
    x11: xdpyinfo, xprop, xrandr

My install options for the proprietary driver is;

NVIDIA-Linux-x86_64-$(RUN_VERSION).run \
    --ui=none \
    --no-questions \
    --accept-license \
    --disable-nouveau \
    --no-install-libglvnd \
    --no-cc-version-check \
    --kernel-module-type=proprietary

For the open driver you don’t specify the kernel-module-type as default is open now.

My Nvidia module config used on both systems is;

/etc/modprobe.d/50-nvidia-tweaks.conf 

blacklist nouveau
options nouveau modeset=0
##Power Management
## Allow the GPU to go into its lowest power state when no applications are running
options nvidia NVreg_DynamicPowerManagement=0x02
## For suspending, make sure not using tmpfs!
options nvidia NVreg_PreserveVideoMemoryAllocations=1
options nvidia NVreg_TemporaryFilePath=/var/tmp
## Enable the PAT feature
options nvidia NVreg_UsePageAttributeTable=1

YEP! … NVIDIA GeForce GT 1030 … one of the first cuda gen cards. Cheap, low power consumption, and decent performance if you are not a gamer or CAD designer. In fact, as recent as year ago, a manufacturer in Japan released a brand new card on the same ancient platform. It is a chip set that refuses to die!

I must admit I am a little confused by your response, and perhaps more generally confused by the issues at hand.

When 570 first appeared in the 156 updates, I mindlessly updated. The only issue I had was that “sleep” or “screen blanking” did not work correctly. I downgraded to 550 and was once more dumb and happy, or at least as happy as a true nihilist can be :).

After trying 570 again this morning, I could not get gnome/gdm to work unless I disabled wayland (in /etc/gdm/custom.conf). I can run gnome with the 570 drivers, but only on X11, which sucks for MANY reasons, although “screen blanking” does work correctly. And this summarizes my entire experience with 156 nicely … a nasty game of whack-a-mole. Fix a problem here to make another there!

On a totally different note, since we last chatted here, I moved my entirely too large multi-media collection to ZFS. ZFS is the hottest thing since sliced bread! Really good, fast, and easy to use … much easier to set up and maintain than mdraid on volume manager on luks on ext4/xfs/NAME YOUR FS. I encourage you to experiment with ZFS if you have large (more 4 TB) of data to manage.

At this point, my biggest question is whether or not opensuse can deliver the stability I want in the next OS. It has certainly failed in 156.

@oxwrongagain I had one of those a few years back, the Quadro T400 replaced it :wink:

I’m running Wayland on both systems which is why I posted the inxi output.

So can you show the output from inxi -GSaz I also see in supported cards there is only one GT 1030 left… it will move to Legacy at some point…

Might be time to consider a GPU upgrade as Nvidia are concentrating on Turing+ and GSP.

Well since zfs is out of tree it has a habit of breaking :frowning_face: I’m happy with xfs on hardware RAID here (rsync over ssh for backup).

On a side note: GNOME 48 is about to be released in Tumbleweed…

dup in progress with much of Gnome to 48… :wink: (fingers crossed)

1 Like

Don’t put too much faith in gnome … some of these guys are really lost.

Back online on Gnome 48 wayland, everything good at first sight.
Nvidia still at 570.124.04 due to a missing file in the nvidia-non-free repo (see here )
But I’m using prime render offload, so the basic video runs on Intel…

1 Like

@OrsoBruno Likewise on my main system Intel/Nvidia, test system had to swap to the proprietary driver from the open one, output to monitor just disappeared, as in graphics still there but frozen… but all good now.

While I am often willing to take the time to try new things, to experiment, on most days I just want my system to work well so that I can work. I use Leap, not tumbleweed, because I value stability over having the latest and and too often not so greatest. I think the move to the 570 drivers was inappropriate for Leap. This is a beta driver and my experience with it only reinforces my desire not to be a beta tester unless I consciously decide to take the time …

While I genuinely love Gnome, the quality of their releases has suffered greatly in recent years. They are focused on enhancing the appearance of the system for all thirteen people who run gnome on their phones. If I were not willing and able to hack, Gnome would be useless to me. It is NOT the genuinely user friendly environment of years past. My gripes aside, I will give the Gnome team credit for pushing forward with wayland, rdp (is it really better than vnc??), etc. Will Gnome 48 be an improvement? I hope so, but I am fully prepared to be disappointed by bad design decisions yet again.

@malcolmlewis, for what it is worth:

root@orca# inxi -GSaz
System:
Kernel: 6.4.0-150600.23.42-default arch: x86_64 bits: 64 compiler: gcc
v: 7.5.0 parameters: BOOT_IMAGE=/vmlinuz-6.4.0-150600.23.42-default
root=UUID=f9ad0342-a904-43d2-a6d2-e60fe1696d68 splash=off preempt=full
quiet acpi_enforce_resources=lax noresume security=apparmor
mitigations=auto
Console: pty pts/3 wm: gnome-shell DM: GDM v: 45.0.1
Distro: openSUSE Leap 15.6
Graphics:
Device-1: NVIDIA GP108 [GeForce GT 1030] vendor: eVga.com. driver: nvidia
v: 550.144.03 alternate: nouveau,nvidia_drm non-free: 530.xx+
status: current (as of 2023-05) arch: Pascal code: GP10x
process: TSMC 16nm built: 2016-21 pcie: gen: 3 speed: 8 GT/s lanes: 4
ports: active: none off: HDMI-A-1 empty: DVI-D-1 bus-ID: 01:00.0
chip-ID: 10de:1d01 class-ID: 0300
Display: server: X.org v: 1.21.1.11 with: Xwayland v: 24.1.1
compositor: gnome-shell driver: X: loaded: nouveau
unloaded: fbdev,modesetting,vesa alternate: nv,nvidia
gpu: nvidia,nvidia-nvswitch display-ID: :0 screens: 1
Screen-1: 0 s-res: 2560x1440 s-size: <missing: xdpyinfo>
Monitor-1: HDMI-A-1 mapped: HDMI-1 note: disabled model: BenQ PD2500Q
serial: built: 2021 res: 2560x1440 hz: 60 dpi: 118 gamma: 1.2
size: 550x310mm (21.65x12.2") diag: 634mm (25") ratio: 16:9 modes:
max: 2560x1440 min: 640x480
API: OpenGL v: 4.6.0 NVIDIA 550.144.03 renderer: NVIDIA GeForce GT
1030/PCIe/SSE2 direct-render: Yes

If you have any ideas, I’ll give 570 another go at the weekend.

As you noted,

I also see in supported cards there is only one GT 1030 left… it will move to Legacy at some point…

Any ATI cards you might recommend? I am almost done with Gnome, but stubbornly hanging on because I know it well. I am definitely DONE with nvidia!

@oxwrongagain what about the current output from cat /proc/driver/nvidia/params | sort and as root user cat /sys/module/nvidia_drm/parameters/fbdev

I would get an Intel ARC GPU… I have a ARC A310 ELF 4GB and in this system a ARC A380 both are down on performance due to no rebar and PCIe 3.0 but still work fine for my needs. Does your current monitor have a Display Port connection, I have found they work better these days than HDMI.

root@orca# cat /proc/driver/nvidia/params | sort
DeviceFileGID: 485
DeviceFileMode: 432
DeviceFileUID: 0
DmaRemapPeerMmio: 1
DynamicPowerManagement: 3
DynamicPowerManagementVideoMemoryThreshold: 200
EnableDbgBreakpoint: 0
EnableGpuFirmware: 18
EnableGpuFirmwareLogs: 2
EnableMSI: 1
EnablePCIERelaxedOrderingMode: 0
EnablePCIeGen3: 0
EnableResizableBar: 0
EnableS0ixPowerManagement: 0
EnableStreamMemOPs: 0
EnableUserNUMAManagement: 1
ExcludedGpus: “”
GpuBlacklist: “”
IgnoreMMIOCheck: 0
ImexChannelCount: 2048
InitializeSystemMemoryAllocations: 1
KMallocHeapMaxSize: 0
MemoryPoolSize: 0
ModifyDeviceFiles: 1
NvLinkDisable: 0
OpenRmEnableUnsupportedGpus: 1
PreserveVideoMemoryAllocations: 1
RegisterPCIDriver: 1
RegistryDwords: “”
RegistryDwordsPerDevice: “”
ResmanDebugLevel: 4294967295
RmLogonRC: 1
RmMsg: “”
RmProfilingAdminOnly: 1
S0ixPowerManagementVideoMemoryThreshold: 256
TCEBypassMode: 0
TemporaryFilePath: “”
UsePageAttributeTable: 4294967295
VMallocHeapMaxSize: 0

root@orca# cat /sys/module/nvidia_drm/parameters/fbdev
Y

Hello @malcolmlewis ,

I hope the above output means more to you than it does to me :slight_smile: .

I am curious as to why you would recommend an intel card over an ati.

Basically, intel always work on linux :wink:

@oxwrongagain They have come a long way, built in hardware encoders/decoders working on the Xe driver, many improvements and all open source, many users have issues with AMD, way more here than Nvidia, my setup(s) have not stopped as I use the run file… Still some niggles here and there with Mesa, but overall my systems keep on running…

Thanks @malcolmlewis

I was under the impression that ati cards worked better with gnu/linux than nvidia; you state the opposite, unless I misunderstood.

Now that my frustration has subsided, I can admit that in 7 plus years of using this GT 1030, I have had very few issues and almost all have been at install time.

What is rebar?

Video playback is probably the most demanding thing I do. I am certainly not looking for a high dollar gaming rig.

Short for Resizable BAR
https://www.intel.com/content/www/us/en/support/articles/000090831/graphics.html

Thanks @hui