DXGI_ERROR_DEVICE_HUNG in Windows 11 VM with gpu passthrough

[jeroslav]

Hello,
I have successfully passed my RTX 2070 to a Windows VM, with the help of Proxmox documentation. Thanks for that. I am getting an error DXGI_ERROR_DEVICE_HUNG in various times of game play - sometimes in 5 minutes, other times in 20 minutes. I have looked through almost all the suggested solutions - removed drivers with DDU, modified registers, run Furmark to test stability of GPU. No signs of issues.
My specs are low-end but I don't think it should be an issue:
Ryzen 3 2200G
16GB RAM
RTX2070

Windows VM config:

args: -cpu host,-hypervisor,kvm=off, -smbios type=0,vendor="American Megatrends International",version=F52l,date="12/25/2023"
balloon: 0
bios: ovmf
boot: order=scsi0
cores: 4
cpu: host,hidden=1
cpuunits: 100
efidisk0: local-lvm:vm-100-disk-0,efitype=4m,pre-enrolled-keys=1,size=4M
hostpci0: 0000:01:00,pcie=1,x-vga=1
machine: pc-q35-8.1
memory: 10000
meta: creation-qemu=8.1.5,ctime=1714995054
name: windows-pc
net0: e1000=BC:24:11:93:E9:B7,bridge=vmbr0,firewall=1
numa: 0
ostype: win11
scsi0: local-lvm:vm-100-disk-1,iothread=1,size=250G
scsihw: virtio-scsi-single
smbios1: uuid=4ef58646-1dc2-4f5c-914c-7890ae205a44,manufacturer=R2lnYWJ5dGUgVGVjaG5vbG9neSBDby4sIEx0ZC4=,product=QUIzNTBNLURBU0g=,base64=1
sockets: 1
tpmstate0: local-lvm:vm-100-disk-2,size=4M,version=v2.0

Apart from the Windows VM I only have a small VM for Home assistant, etc.

balloon: 0
bios: ovmf
boot: order=scsi0;ide2;net0
cores: 1
cpu: x86-64-v3
efidisk0: local-lvm:vm-101-disk-0,efitype=4m,pre-enrolled-keys=1,size=4M
ide2: local:iso/debian-12.5.0-amd64-netinst.iso,media=cdrom,size=629M
machine: q35
memory: 2048
meta: creation-qemu=8.1.5,ctime=1715016751
name: smart-home
net0: virtio=BC:24:11:61:0D:87,bridge=vmbr0,firewall=1
numa: 0
ostype: l26
scsi0: local-lvm:vm-101-disk-1,iothread=1,size=120G
scsihw: virtio-scsi-single
smbios1: uuid=6645e4d9-b075-4531-ac0c-5b642f4c26cb
sockets: 1
vmgenid: 8876ecd7-6101-46b4-860b-384df98958e4

Any ideas where I should look, what settings should I change for the VM?

 

[gfngfn256]

My specs are low-end but I don't think it should be an issue

Not sure how you are convinced of this.

Any ideas where I should look, what settings should I change for the VM?

This problem appears pretty widespread amongst Windows users. How do you know it is HV/VM specific?

Have you run this same setup bare-metal Windows without incident?

If not, what I would try - to test - is run the same HW bare-metal Windows & see if this persists. So maybe write the VM's virtual disk to a spare HDD etc. boot up from it - (you will have to probably go through various Windows reconfig/setup/tweaks) - & test.

Side note: I would change in the VM, the memory from 10000 to 10240 to give it a perfect 10GB - some OSs don't like those incomplete RAM amounts.

 

[jeroslav]

Not sure how you are convinced of this.


This problem appears pretty widespread amongst Windows users. How do you know it is HV/VM specific?

Have you run this same setup bare-metal Windows without incident?

If not, what I would try - to test - is run the same HW bare-metal Windows & see if this persists. So maybe write the VM's virtual disk to a spare HDD etc. boot up from it - (you will have to probably go through various Windows reconfig/setup/tweaks) - & test.

Side note: I would change in the VM, the memory from 10000 to 10240 to give it a perfect 10GB - some OSs don't like those incomplete RAM amounts.

Thanks for the advices. I was not sure if I had some misconfiguration. It seems to run fine for some days, but then it occurs frequently until I reboot the node. I did not have much time recently, I just let my son play on the VM, but according to the graphs it seems like its pushing back and forth to swap and then it crashes.

Will need to upgrade the potato HW.

Thanks for the help nonetheless.