[SOLVED] Problem with GPU Passthrough

I don't know what helped, but I remove AMD card from vfio, updated system with apt update, rebooted the system and vendor-reset started to work, but id didn't help in my case, because the whole system still stack when the GPU glithches :( Anyway, I give it a try... Thank you leesteken for your help!
Never run apt update, always run apt dist-upgrade or use the GUI: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#system_software_updates
It should work fine on a clean Proxmox with vendor-reset if it is activated properly.
 
  • Like
Reactions: moxerproxer
Hello, leesteken. I made a clean install Proxmox 8 for proper vendor-reset for RX580 and can't find a repo where I can download right pve-headers, could you tell me where it is?
They should be available from the same repository as where you get your Proxmox kernel from. If you cannot find them, then use the new name: proxmox-headers. But the old name should also work: https://forum.proxmox.com/threads/install-pve-headers-uname-r-failure.138509/post-618229
 
  • Like
Reactions: moxerproxer
They should be available from the same repository as where you get your Proxmox kernel from. If you cannot find them, then use the new name: proxmox-headers. But the old name should also work: https://forum.proxmox.com/threads/install-pve-headers-uname-r-failure.138509/post-618229
That's what I get
Code:
apt install proxmox-headers-$(uname -r)
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
E: Unable to locate package proxmox-headers-6.2.16-3-pve
E: Couldn't find any package by glob 'proxmox-headers-6.2.16-3-pve'

UPD:
I've tried update and get this
Code:
apt update
Hit:1 http://ftp.ru.debian.org/debian bookworm InRelease
Get:2 http://ftp.ru.debian.org/debian bookworm-updates InRelease [52.1 kB]
Hit:3 http://security.debian.org bookworm-security InRelease
Err:4 https://enterprise.proxmox.com/debian/ceph-quincy bookworm InRelease
  401  Unauthorized [IP: 51.91.38.34 443]
Err:5 https://enterprise.proxmox.com/debian/pve bookworm InRelease
  401  Unauthorized [IP: 51.91.38.34 443]
Reading package lists... Done
E: Failed to fetch https://enterprise.proxmox.com/debian/ceph-quincy/dists/bookworm/InRelease  401  Unauthorized [IP: 51.91.38.34 443]
E: The repository 'https://enterprise.proxmox.com/debian/ceph-quincy bookworm InRelease' is not signed.
N: Updating from such a repository can't be done securely, and is therefore disabled by default.
N: See apt-secure(8) manpage for repository creation and user configuration details.
E: Failed to fetch https://enterprise.proxmox.com/debian/pve/dists/bookworm/InRelease  401  Unauthorized [IP: 51.91.38.34 443]
E: The repository 'https://enterprise.proxmox.com/debian/pve bookworm InRelease' is not signed.
N: Updating from such a repository can't be done securely, and is therefore disabled by default.
N: See apt-secure(8) manpage for repository creation and user configuration details.
 
Last edited:
That's what I get
Code:
apt install proxmox-headers-$(uname -r)
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
E: Unable to locate package proxmox-headers-6.2.16-3-pve
E: Couldn't find any package by glob 'proxmox-headers-6.2.16-3-pve'
Please stop using kernel version 6.2 as it does not get updates anymore. Maybe you could switch to the thread that I linked to as this is not specific to passthrough (and this thread is marked as solved).
 
  • Like
Reactions: moxerproxer
I've install everything (vendor-reset) without any troubles, but when I try to use this echo 'device_specific' > /sys/bus/pci/devices/<myGPUID>/reset_method I get Invalid argument What else I can do whith this?
 
I've install everything (vendor-reset) without any troubles, but when I try to use this echo 'device_specific' > /sys/bus/pci/devices/<myGPUID>/reset_method I get Invalid argument What else I can do whith this?
Is it installed correctly? Are you using the correct PCI ID? Is the module loaded? Try doing a modprobe vendor_reset first.
 
Is it installed correctly? Are you using the correct PCI ID? Is the module loaded? Try doing a modprobe vendor_reset first.
It was added after this command but after reboot it's showing "bus" again in /sys/bus/pci/devices/<myGPUID>/reset_method
 
Hello and a happy new year!

Has anyone managed to passthrough the iGPU of a 5600G?

My settings are:

cat /proc/cmdline
initrd=\EFI\proxmox\6.5.11-7-pve\initrd.img-6.5.11-7-pve root=ZFS=rpool/ROOT/pve-1 boot=zfs quiet amd_iommu=on iommu=pt initcall_blacklist=sysfb_init pcie_acs_override=downstream,multifunction textonly video=simplefb:eek:ff video=vesafb:eek:ff video=efifb:eek:ff video=vesa:eek:ff disable_vga=1 vfio_iommu_type1.allow_unsafe_interrupts=1 kvm.ignore_msrs=1

cat /etc/kernel/cmdline
root=ZFS=rpool/ROOT/pve-1 boot=zfs quiet amd_iommu=on iommu=pt initcall_blacklist=sysfb_init pcie_acs_override=downstream,multifunction textonly video=simplefb:eek:ff video=vesafb:eek:ff video=efifb:eek:ff video=vesa:eek:ff disable_vga=1 vfio_iommu_type1.allow_unsafe_interrupts=1 kvm.ignore_msrs=1

/etc/modprobe.d/vfio.conf
options vfio-pci ids=1002:1638,1002:1637,1002:1636 disable_vga=1
softdep radeon pre: vfio-pci
softdep amdgpu pre: vfio-pci
softdep snd_hda_intel pre: vfio-pci

I have also tried with just the first line of code

/etc/modprobe.d/blacklist.conf
blacklist amdgpu
blacklist radeon
blacklist snd_hda_intel
blacklist ccp

and with nothing blacklisted

lspci -nnks 06:00
06:00.0 VGA compatible controller [0300]: Advanced Micro Devices, Inc. [AMD/ATI] Cezanne [Radeon Vega Series / Radeon Vega Mobile Series] [1002:1638] (rev c9)
Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Cezanne [Radeon Vega Series / Radeon Vega Mobile Series] [1002:1636]
Kernel driver in use: vfio-pci
Kernel modules: amdgpu
06:00.1 Audio device [0403]: Advanced Micro Devices, Inc. [AMD/ATI] Renoir Radeon High Definition Audio Controller [1002:1637]
Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Renoir Radeon High Definition Audio Controller [1002:1637]
Kernel driver in use: vfio-pci
Kernel modules: snd_hda_intel

/etc/pve/qemu-server/106.conf
bios: ovmf
boot: order=ide1;ide0;ide2;net0
cores: 4
cpu: host,hidden=1
efidisk0: local-zfs:vm-106-disk-0,efitype=4m,pre-enrolled-keys=1,size=1M
hostpci0: 0000:06:00.0,pcie=1,romfile=vbios_1638.dat
ide0: local:iso/virtio-win-0.1.240.iso,media=cdrom,size=612812K
ide1: local-zfs:vm-106-disk-1,discard=on,size=128G,ssd=1
ide2: local:iso/en-us_windows_10_consumer_editions_version_22h2_updated_dec_2023_x64_dvd_3045f975.iso,media=cdrom,size=>machine: pc-q35-8.1
memory: 4096
meta: creation-qemu=8.1.2,ctime=1704478634
name: win10
net0: virtio=BC:24:11:0A:49:F5,bridge=vmbr0,firewall=1
numa: 0
ostype: win10
scsihw: virtio-scsi-single
smbios1: uuid=f71b4d72-3c4d-416b-bec2-8c3d26512246
sockets: 1
vga: none
vmgenid: 9f4ba1c7-b21c-4930-9637-04dfc370a850

The romfile was extracted from the manufacturer's bios file. I also tried this one but the VM doesn't boot

I have also created a fix_gpu_pass.sh like this:
fix_gpu_pass.sh
#!/bin/bash
echo 1 > /sys/bus/pci/devices/0000\:06\:00.0/remove
echo 1 > /sys/bus/pci/rescan

Any help would be greatly appreciated
 
I've been bashing my head against a wall, failing to get this to work, and I just had an epiphany.

I've been trying to pass through an old Quadro 2000 card. It's Fermi based. As I understand Fermi boards did not have UEFI support.

I'm rolling my eyes at myself right now. I'm going to stick the old 6GB Titan in there instead. It was Kepler. Hopefully that will work...
 
  • Like
Reactions: moxerproxer

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!