mbilker / vgpu_unlock-rs

Unlock vGPU functionality for consumer grade GPUs
MIT License
443 stars 52 forks source link

VM will not start and nvidia-vgpu-mgr spits errors #6

Closed SIGSTKFLT closed 2 years ago

SIGSTKFLT commented 2 years ago

Driver: NVIDIA-Linux-x86_64-510.47.03-vgpu-kvm with patch (tried with and without patch) GPU: GTX 1050 Ti Profile to override:

nvidia-286
    Available instances: 2
    Device API: vfio-pci
    Name: GRID P40-12C
    Description: num_heads=1, frl_config=60, framebuffer=12288M, max_resolution=4096x2160, max_instance=2

Error when booting vm:

mdev instance '00000000-0000-0000-0000-000000000100' already existed, using it.
kvm: -device vfio-pci,sysfsdev=/sys/bus/pci/devices/0000:06:00.0/00000000-0000-0000-0000-000000000100,id=hostpci0,bus=pci.0,addr=0x10: warning: vfio 00000000-0000-0000-0000-000000000100: Could not enable error recovery for the device
kvm: -device vfio-pci,sysfsdev=/sys/bus/pci/devices/0000:06:00.0/00000000-0000-0000-0000-000000000100,id=hostpci0,bus=pci.0,addr=0x10: vfio 00000000-0000-0000-0000-000000000100: failed to read device config space: Bad address
TASK ERROR: start failed: QEMU exited with code 1

systemctl status nvidia-vgpu-mgr

Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: Patching nvidia-286/pci_id: 456659844 -> 456135072
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: Patching nvidia-286/pci_device_id: 6968 -> 6960
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: Patching nvidia-286/framebuffer: 12012486656 -> 1968526677
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: Patching nvidia-286/frl_enabled: 1 -> 0
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: error: vmiop_log: (0x0): Guest BAR1 is of invalid length (g: 0x400000000, h: 0x10000000)
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: error: vmiop_log: (0x0): init_device_instance failed for inst 0 with error 1 (error setting vGPU configuration information from RM)
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: error: vmiop_log: (0x0): Initialization: init_device_instance failed error 1
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: error: vmiop_log: display_init failed for inst: 0
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: error: vmiop_env_log: (0x0): vmiope_process_configuration: plugin registration error
Apr 21 23:30:58 proxmox nvidia-vgpu-mgr[4090]: error: vmiop_env_log: (0x0): vmiope_process_configuration failed with 0x1f

/etc/vgpu_unlock/profile_override.toml

[profile.nvidia-286]
num_displays = 1
display_width = 1920
display_height = 1080
max_pixels = 2073600
cuda_enabled = 1
frl_enabled = 0
framebuffer = 1968526677
pci_id = 0x1B3011A0
pci_device_id = 0x1B30

Any help would be greatly appreciated.

SIGSTKFLT commented 2 years ago

Issue solved thanks to https://github.com/DualCoder/vgpu_unlock/issues/55. I had to override Q type profile 52.