Skip to content

NVIDIA GTX 1070 (Pascal) not supported - driver fails to load #150

Description

@gd-harco

Note

As this is my first issue on a large open-source project, I used LLM assistance to draft this report and follow best practices. However, all technical information has been personally verified and is accurate. I apologize if this approach conflicts with repository guidelines—I didn't find a CONTRIBUTING.md file or similar to reference.

Description

After rebasing from bazzite-nvidia:stable to bazzite-dx-nvidia:stable, NVIDIA drivers fail to load. The system cannot communicate with the GPU, and nvidia-smi fails with driver communication error.

System Information

  • Previous image (working): ostree-image-signed:docker://ghcr.io/ublue-os/bazzite-nvidia:stable
    • Version: 43.20260101
  • Current image (broken): ostree-image-signed:docker://ghcr.io/ublue-os/bazzite-dx-nvidia:stable
    • Version: 43.20260120.1
  • GPU: NVIDIA GeForce GTX 1070 (Pascal architecture)
    • PCI ID: 10de:1b81
    • Bus address: 0000:1f:00.0

Steps to Reproduce

  1. Start with working bazzite-nvidia:stable installation
  2. Run: rpm-ostree rebase ostree-image-signed:docker://ghcr.io/ublue-os/bazzite-dx-nvidia:stable
  3. Reboot system
  4. Run: nvidia-smi

Expected Behavior

NVIDIA drivers should load successfully and nvidia-smi should display GPU information, as it did on the standard bazzite-nvidia image.

Actual Behavior

$ nvidia-smi
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

Kernel Logs

The following error repeats continuously in journalctl:

NVRM: The NVIDIA GPU 0000:1f:00.0 (PCI ID: 10de:1b81)
NVRM: nvidia.ko because it does not include the required GPU
NVRM: www.nvidia.com.
nvidia 0000:1f:00.0: probe with driver nvidia failed with error -1
NVRM: The NVIDIA probe routine failed for 1 device(s).
NVRM: None of the NVIDIA devices were initialized.
Full kernel log excerpt ``` Jan 24 12:19:38 Host-004 kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 508 Jan 24 12:19:38 Host-004 kernel: NVRM: The NVIDIA GPU 0000:1f:00.0 (PCI ID: 10de:1b81) NVRM: nvidia.ko because it does not include the required GPU NVRM: www.nvidia.com. Jan 24 12:19:38 Host-004 kernel: nvidia 0000:1f:00.0: probe with driver nvidia failed with error -1 Jan 24 12:19:38 Host-004 kernel: NVRM: The NVIDIA probe routine failed for 1 device(s). Jan 24 12:19:38 Host-004 kernel: NVRM: None of the NVIDIA devices were initialized. Jan 24 12:19:38 Host-004 kernel: nvidia-nvlink: Unregistered Nvlink Core, major device number 508 Jan 24 12:19:38 Host-004 (udev-worker)[892]: nvidia: Process '/usr/bin/nvidia-modprobe' failed with exit code 1. ``` (Pattern repeats multiple times)

Analysis

The error message "nvidia.ko because it does not include the required GPU" suggests that the NVIDIA driver version in bazzite-dx-nvidia does not include support for Pascal-generation GPUs (GTX 10xx series). This appears to be a regression from the standard bazzite-nvidia image, which works perfectly with the same hardware.

Workaround

Rolling back to the previous deployment restores GPU functionality:

rpm-ostree rollback
systemctl reboot

Additional Information

  • GTX 1070 is based on Pascal architecture (2016)
  • Standard bazzite-nvidia:stable works perfectly with this GPU
  • The issue appears specific to the bazzite-dx-nvidia variant

Thank you for your work on Bazzite DX! Happy to provide additional logs or testing if needed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions