[SOLVED] GPU Passthrough throws driver error 43

[JohnB17]

Hi, I am trying to passthrough an NVIDIA GPU, but it wouldn't work at all, so I followed the below link from Reddit. That at least made the monitor know there is a signal, but it is blank and it shows the error in the below image.
https://www.reddit.com/r/homelab/comments/b5xpua/the_ultimate_beginners_guide_to_gpu_passthrough/

 

[WesC]

I've been having errors and breakage with gpu passthrough on all my hosts I've upgraded to 7.2 with kernel 5.15.30. Downgrading the pve servers to kernel 5.13.x fixes the issue for me. I'm doing some testing and will file a bug when I have some concrete data to give PVE devs.

I'll also note that mainline kernels 5.15.33 to 5.15.36 report a lot of vfio and iommu related fixes and changes - I'll test those too in case they fix the issue in 5.15.30

 

[Hallow]

I've been having errors and breakage with gpu passthrough on all my hosts I've upgraded to 7.2 with kernel 5.15.30. Downgrading the pve servers to kernel 5.13.x fixes the issue for me. I'm doing some testing and will file a bug when I have some concrete data to give PVE devs.

I'll also note that mainline kernels 5.15.33 to 5.15.36 report a lot of vfio and iommu related fixes and changes - I'll test those too in case they fix the issue in 5.15.30

And I was wondering why my GPU passthrough broke after updating to 7.2..

 

[Hallow]

Apparantly aafter pinning my kernel to

proxmox-boot-tool kernel pin 5.13.19-6-pv

issue resolved it self.
Also I've stopped getting

May 07 18:34:41 zen-proxmox kernel: pcieport 0000:00:01.1: AER: Corrected error received: 0000:01:00.0
May 07 18:34:41 zen-proxmox kernel: nvme 0000:01:00.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)
May 07 18:34:41 zen-proxmox kernel: nvme 0000:01:00.0:   device [15b7:5006] error status/mask=00000001/0000e000
May 07 18:34:41 zen-proxmox kernel: nvme 0000:01:00.0:    [ 0] RxErr

errors

 

[JohnB17]

Nothing to do with the gpu and the node, it's the drivers detecting a vm, I got the wrong driver.

 

[trio198]

Nothing to do with the gpu and the node, it's the drivers detecting a vm, I got the wrong driver.

Hi @JohnB17

So which driver are you using to get it work?

Currently experiencing the same issue with proxmox 7.2, with rtx 3090 ti, grid driver 530 on host and regular quadro 516 driver on guest win10 vm.

 

[kromberg]

Getting this same error also with several GPU passthroughs:
pve-manager/7.2-7/d0dd0e85
5.13.19-6-pve, 5.15.39-3-pve, and 5.15.53-1-pve kernels.

 

[trio198]

Getting this same error also with several GPU passthroughs:
pve-manager/7.2-7/d0dd0e85
5.13.19-6-pve, 5.15.39-3-pve, and 5.15.53-1-pve kernels.

If only need GPU passthrough (i.e. not vGPU), this guide should be sufficient :

https://www.reddit.com/r/homelab/comments/b5xpua/the_ultimate_beginners_guide_to_gpu_passthrough/

But if you need vGPU, it's another story...:-D

vGPU still not support RTX 30 GPU. Even if you manage to install driver in both host and guest VM, error 43 will still occure.

 

[fahadysf]

Apparantly aafter pinning my kernel to

proxmox-boot-tool kernel pin 5.13.19-6-pv

issue resolved it self.
Also I've stopped getting

May 07 18:34:41 zen-proxmox kernel: pcieport 0000:00:01.1: AER: Corrected error received: 0000:01:00.0
May 07 18:34:41 zen-proxmox kernel: nvme 0000:01:00.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)
May 07 18:34:41 zen-proxmox kernel: nvme 0000:01:00.0:   device [15b7:5006] error status/mask=00000001/0000e000
May 07 18:34:41 zen-proxmox kernel: nvme 0000:01:00.0:    [ 0] RxErr

errors

Same thing, pinning to 5.13.19-6-pve works perfectly (for GPU Passthrough).

Full guide on my working setup in this post: https://forum.proxmox.com/threads/gpu-passthrough-code-43-unable-to-resolve.107996/post-506721