Search results

in my case i use nomachine, and nomachine can utilize the GPU for hardware encoding, however it cannot stretch to 2 monitor afaik. i have an iGPU passthrough to a Linux vm, and two nvidia passthrough to two windows 11, all with nomachine and able to utilize the H264 hardware encoding with the gpu.

install the VM as you like, connect to the VM using VNC, or nomachine, or anydesk etc, make it full screen on your host.

set the display to none, and make sure you copy the i915_vbt. display and igpu passthrough does not play well together in latest kernel.

i have this same as yours /etc/modprobe.d/blacklist-nvidia.conf blacklist nouveau blacklist nvidia blacklist nvidiafb blacklist rivafb blacklist snd_hda_intel

nvidia card is easy to passthrough without additional setting, i am passing through it to a windows 11 VM since PVE8 and now PVE9. simply adding it to your VM should work just fine.

did you by any chance blacklisted i915 driver? another potential problem is to check your bios, some motherboard default setting will disable igpu when a PCIE GPU is installed.

for those using latest kernel in Linux VM and failed with iGPU passthrough, read here https://forum.proxmox.com/threads/proxmox-9-gpu-pass-through.169151/post-787947 you need to get the i915_vbt https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12106

i am not using AMD so not sure if what i did for intel can apply for AMD, or not. below was my step https://forum.proxmox.com/threads/alderlake-direct-passthrough-to-linux-vm-is-working-with-hdmi-output.138049 although when i wrote it was using 12th gen alderlake, but i confirm it is working...

could be AMD specific, I have 2 nvidias and 1 intel igpu, all passthrough with working HDMI and audio output.

I think you might be bump in to this https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12106 you need to manually solve it follow that discussion, and still latest 6.12 kernel still having this issues.

you can get the solution here https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12106 in summary, just get the i915_vbt from host (before passthrough activated) copy it the VM guest, and load it from there. ready through my post in gitlab then you should know how to do.

https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/12106#note_2553654 likely this is related.

for my experience, high IO delay only occur when intensive read write happen to HDD (but this can be solved if passthrough the SATA controller to VM).

This is just to share, if you are using intel iGPU passthrough to a Linux VM, you might encounter a problem that no HDMI output if you upgrade your Linux VM to kernel 6.10 onward. the solution will be here: https://bugzilla.kernel.org/show_bug.cgi?id=219193 i have Alderlake iGPU passthrough to...

set the Display to none in VM hardware tab. i have 3 passthrough working correctly. 2 nvidia and 1 intel igpu.

a checkbox to allow enable or disable the device, this will allow quick change without to add PCI or remove PCI in testing environment. appreciate if this can be considered.

ok, disabled secure boot in BIOS will still allow proxmox host to boot. no other configuration needed. i will just disable Secure Boot function.

I performed latest kernel update this morning for proxmox host, and went for host restart, now i ended up having this problem during BIOS boot up and the system shutdown immediately after this. i have secure boot enabled and what will happen if i reset secure boot? will it stop my system for...

it happen again today in the middle of the night where no one is using the server nor VM...... May 18 04:17:50 proxmox pvestatd[2081]: pbs: error fetching datastores - 500 Can't connect to 192.168.1.13:8007 (No route to host) May 18 04:18:00 proxmox pvestatd[2081]: pbs: error fetching...

the second reboot today which did not show any error: May 16 15:38:36 proxmox pvestatd[1871]: pbs: error fetching datastores - 500 Can't connect to 192.168.1.13:8007 (No route to host) May 16 15:38:43 proxmox kernel: [UFW BLOCK] IN=vmbr0 OUT= MAC=01:00:5e:00:00:01:c8:7f:54:c2:c1:30:08:00...