Nvidia driver 520.56 and eGPU with RTX 3060

Reposting since I originally thought it was an upstream issue, but not so sure now.

The NVIDIA 520.56 driver installed from the RPMFusion repository OR DNF is completely broken for me. This may be an edge case scenario, but I am attempting to connect my Intel laptop to an eGPU with a RTX 3060 so I can both use multiple-monitors, and process GPU intensive projects. lspci and the laptop both recognize that it is connected to the GPU, however, the NVIDIA driver and module does not.

Fedora 37 will boot with the laptop BIOS logo frozen on one of the monitors & the other black, and will use the iGPU to render instead. The laptop straight up refuses to log in via Gnome xorg (even with nvidia drivers not installed?), and will instantly kick me back to the login screen if I do not launch via Wayland.

I have tried downloading a different Fedora ISO and reinstalling using that, with similar results.

Information that stood out to me:

  • The GPU “works” and passes rendering to the monitors on the live usb and a fresh install, but stops “working” the second I download the nvidia drivers and reboot.

  • With the driver not installed, “Graphics” in the “About” section of gnome-settings shows Software rendering / Mesa Intel® Graphics (ADL GT2)

  • With the driver installed, it only shows Mesa Intel® Graphics (ADL GT2)

  • lspci and the OS ARE CONNECTED to the GPU. The nvidia module and driver are not, for some reason.

Steps to reproduce:

  1. Install the latest nvidia drivers from the RPM Fusion repository on the Software program or CLI

  2. Reboot

Actual results:

Monitors black, nvidia GPU not recognized by driver and module.

Expected results:

GPU is fully functional, outputs to the monitors, and is utilized by the computer for graphics intensive processes.

Build date & hardware:

Build 520.56.06 on Fedora 37, kernel version 6.0.9-300.fc37.x86_64, CPU model Intel i5-1240P, GPU model RTX 3060

Additional information: Secure boot disabled.

'Important logs'
01:47:08 gdm-session-wor: GLib-GObject: g_object_unref: assertion 'G_IS_OBJECT (object)' failed
01:47:06 systemd: Failed to start app-gnome-user\x2ddirs\x2dupdate\x2dgtk-2896.scope - Application launched by gnome-session-binary.

01:47:00 pipewire: mod.x11-bell: XOpenDisplay() failed
01:46:21 systemd: Failed to start nvidia-powerd.service - nvidia-powerd service.
01:46:18 kernel: [drm:nv_drm_probe_devices [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000400] Failed to register device
01:46:18 kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000400] Failed to allocate NvKmsKapiDevice
01:46:15 kernel: 
01:46:15 virtqemud: Unable to open /dev/kvm: No such file or directory
01:46:15 kernel: cros-usbpd-charger cros-usbpd-charger.5.auto: Unexpected number of charge port count
01:46:15 bluetoothd: Failed to set mode: Failed (0x03)
01:46:13 thermald: Unsupported conditions are present
01:46:11 kernel: /usr/lib/systemd/system-generators/systemd-gpt-auto-generator failed with exit status 1.
01:46:11 kernel: Failed to dissect: Permission denied
01:46:09 kernel: usb usb10-port1: couldn't allocate usb_device
01:46:09 kernel: xhci_hcd 0000:09:00.0: HC died; cleaning up
01:46:00 kernel: tpm tpm0: [Firmware Bug]: TPM interrupt not working, polling instead
01:46:00 kernel: pci 0000:00:07.3: DPC: RP PIO log size 0 is invalid
'Sys logs'
02:09:12 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:09:12 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:09:12 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:09:11 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:09:11 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:09:10 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:09:10 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:09:09 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:09:09 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:08:36 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:08:36 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:08:35 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:08:35 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:08:34 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:08:34 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:08:32 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:08:32 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:19 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:19 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:18 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:18 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:17 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:17 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:16 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:16 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:11 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:11 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:10 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:10 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:09 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:09 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:07:08 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
02:07:08 kernel: NVRM: GPU 0000:04:00.0: RmInitAdapter failed! (0x26:0x56:1423)
02:02:04 kernel: usb 3-9: reset full-speed USB device number 3 using xhci_hcd
01:48:26 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:47:06 kernel: EXT4-fs (sda1): mounted filesystem with ordered data mode. Quota mode: none.
01:47:06 kernel: rfkill: input handler disabled
01:47:06 kernel: EXT4-fs (sdb2): mounted filesystem with ordered data mode. Quota mode: none.
01:47:04 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:59 kernel: rfkill: input handler enabled
01:46:53 kernel: usb 3-9: reset full-speed USB device number 3 using xhci_hcd
01:46:45 kernel: rfkill: input handler disabled
01:46:45 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:38 kernel: rfkill: input handler enabled
01:46:31 kernel: logitech-hidpp-device 0003:046D:4054.0009: HID++ 4.5 device connected.
01:46:30 kernel: usb 3-9: reset full-speed USB device number 3 using xhci_hcd
01:46:30 kernel: logitech-hidpp-device 0003:046D:4023.0008: HID++ 2.0 device connected.
01:46:30 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:27 kernel: usb 3-9: reset full-speed USB device number 3 using xhci_hcd
01:46:27 kernel: Bluetooth: RFCOMM ver 1.11
01:46:27 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:26 kernel: rfkill: input handler disabled
01:46:24 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:19 kernel: wlp166s0: Limiting TX power to 30 (30 - 0) dBm as advertised by 98:9d:5d:0c:6f:0e
01:46:19 kernel: IPv6: ADDRCONF(NETDEV_CHANGE): wlp166s0: link becomes ready
01:46:19 kernel: wlp166s0: associated
01:46:19 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:18 kernel: [drm:nv_drm_probe_devices [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000400] Failed to register device
01:46:18 kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000400] Failed to allocate NvKmsKapiDevice
01:46:18 kernel: NVRM: GPU 0000:04:00.0: rm_init_adapter failed, device minor number 0
01:46:17 kernel: intel_tcc_cooling: Programmable TCC Offset detected
01:46:16 kernel: RAPL PMU: hw unit of domain psys 2^-14 Joules
01:46:16 kernel: [drm] [nvidia-drm] [GPU ID 0x00000400] Loading driver
01:46:16 kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  520.56.06  Thu Oct  6 21:22:53 UTC 2022
01:46:16 kernel: nvidia-uvm: Loaded the UVM driver, major device number 507.
01:46:15 kernel: nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint.
01:46:15 kernel: thunderbolt 0-101: Razer Core X Chroma
01:46:15 kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module  520.56.06  Thu Oct  6 21:38:55 UTC 2022
01:46:15 kernel: nvidia 0000:04:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
01:46:15 kernel: 
01:46:15 kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 509
01:46:15 kernel: nvidia: module verification failed: signature and/or required key missing - tainting kernel
01:46:15 kernel: mei_pxp 0000:00:16.0-fbf6fcf1-96cf-4e2e-a6a6-1bab8cbe36b1: bound 0000:00:02.0 (ops i915_pxp_tee_component_ops [i915])
01:46:15 kernel: mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i915_hdcp_component_ops [i915])
01:46:15 kernel: intel_rapl_common: Found RAPL domain psys
01:46:15 kernel: Disabling lock debugging due to kernel taint
01:46:15 kernel: nvidia: module license 'NVIDIA' taints kernel.
01:46:15 kernel: cros-usbpd-charger: probe of cros-usbpd-charger.5.auto failed with error -71
01:46:15 kernel: cros-usbpd-charger cros-usbpd-charger.5.auto: Failing probe (err:0xffffffb9)
01:46:15 kernel: iwlwifi 0000:a6:00.0 wlp166s0: renamed from wlan0
01:46:15 kernel: Bluetooth: MGMT ver 1.22
01:46:15 kernel: iTCO_wdt iTCO_wdt: initialized. heartbeat=30 sec (nowayout=0)
01:46:15 kernel: thunderbolt 0-1: Razer Core X Chroma
1 Like

For completeness, can you also please note what debugging you’ve already done so far to conclude that this isn’t an upstream issue?

The RPM Fusion drivers are what Nvidia ship. It’s just their binary packaged as an RPM. If that doesn’t work, either it’s a configuartion type issue, or it’s indeed an issue with the proprietary Nvidia driver. It’s kinda hard to debug. If you’re open to reinstalling, perhaps you can try the nvidia binary blob from the website—if that works, it’ll tell us that it’s not an issue with the driver Nvidia ship, but if it doesn’t, it’ll pretty much confirm it’s there. (We don’t recommend folks use the nvidia blob from their website because it doesn’t play well with kernel updates etc—which is why RPM Fusion package it up as a nice rpm for us.)

Have you tried the bug-reporting bit listed here?

https://rpmfusion.org/Howto/NVIDIA#Bug_Report

A search for RmInitAdapter failed! gives me lots of links from the Nvidia developer forums:

I’ve read several issues some people have had with the 3060 graphic cards and nvidia with the 520.56 drivers. I’ve read where some have downgraded to the 470 driver and solved the issue. I don’t have that card so I don’t know for sure. As far as I know it’s not in rpmfusion stable yet but there is a new driver 520.60.11. As I don’t have that card I don’t know if it will fix the issues but nvidia is aware of it as a lot have had problems.Sorry I can’t be of more help.

My understanding is that the 3060 and newer cards are not supported by the 520.56 driver. (I may be wrong). The newer driver is supposed to support the newer cards once it has made its way to rpmfusion.

The message about software rendering is correct. The nouveau driver does not support hardware acceleration on any of the newer nvidia GPUs.

Downgrading to the 470 driver may help (I doubt it), but does not support wayland at all so is not an option for those who desire to use wayland for their desktop.

1 Like

I’ve spent several consecutive days trying to troubleshoot it. It’s more what I haven’t tried at this point.

Any updates or information that could help? I’m completely clueless at this point.

There is a new 525.60.11 driver in rpmfusion-nonfree-updates-testing

Also you were requested by Ankur to provide the informations needed in this link: Howto/NVIDIA - RPM Fusion

Please do so before anything else. (eventually using the bugzilla.rpmfusion.org bug tracker).
Also please consider reporting to upstream devtalk.nvidia.com

I would recommend installing the newer driver 525.60.11 noted above and reboot to see if it properly supports your system. If it does then a bug report is likely not necessary, but if it does not then the bug report would be a good thing.