Closed polloloco closed 1 year ago
When using this, I'm getting errors like
Jan 15 17:05:19 pve nvidia-vgpu-mgr[477123]: Failed to decode config: invalid type: string "202", expected u64 for key `vm` at line 11 column 1
Jan 15 17:05:19 pve nvidia-vgpu-mgr[477123]: Failed to apply profile override
My profile_override.toml, for reference:
[profile.nvidia-259]
num_displays = 1
display_width = 1920
display_height = 1200
max_pixels = 2304000
cuda_enabled = 1
#frl_enabled = 1
#framebuffer = 0x74000000
#framebuffer_reservation = 0xC000000
[vm.202]
#frl_enabled = 0
pci_device_id = 0x1E30
pci_id = 0x1E3012BA
Not sure why it's trying to parse 202
as a string instead of a number... unless table keys are always assumed to be strings?
@abraha2d I just fixed it in https://github.com/mbilker/vgpu_unlock-rs/pull/28
This pull request improves the user experience for proxmox users by allowing to specify per-VM overrides using a new
[vm.<id>]
section. The usual way to do per-VM overrides was by creating a section like this (in the example for the VM with id 100):Many users had problems with the long uuid string, and now they can do this instead (again, for a VM with the id 100):
Now, proxmox users don't have to bother with any uuids anymore :)