[SOLVED] /dev/dri: No such file or directory

[helio]

Hi,
Im running proxmox on a 6 x Intel(R) Core(TM) i5-9500T CPU with GPU passthrough.
I have one vm running and using GPU.
Stopping the VM and starting a lxc container with /dev/dri wont work
cd /dev/dri
/dev/dri: No such file or directory
when I look at Vga I see
00:02.0 VGA compatible controller [0300]: Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630] [8086:3e92]
DeviceName: Onboard IGD
Subsystem: Hewlett-Packard Company CoffeeLake-S GT2 [UHD Graphics 630] [103c:8595]
Kernel driver in use: vfio-pci
Kernel modules: i915

How can I make /dev/dri visible ,load?
Newbie here so sorry, Thanks
SOLUTION
Edit /etc/modprobe.d/vfio.conf and comment out whats in that file. Reboot and /dev/dri ARE BACK.

 

[alkaizer]

Hi,
Im running proxmox on a 6 x Intel(R) Core(TM) i5-9500T CPU with GPU passthrough.
I have one vm running and using GPU.
Stopping the VM and starting a lxc container with /dev/dri wont work
cd /dev/dri
/dev/dri: No such file or directory
when I look at Vga I see
00:02.0 VGA compatible controller [0300]: Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630] [8086:3e92]
DeviceName: Onboard IGD
Subsystem: Hewlett-Packard Company CoffeeLake-S GT2 [UHD Graphics 630] [103c:8595]
Kernel driver in use: vfio-pci
Kernel modules: i915

How can I make /dev/dri visible ,load?
Newbie here so sorry, Thanks

is the lxc privledged or unprivledged?

 

[helio]

is the lxc privledged or unprivledged?

Its privledged , I think this is related to another problem I m not able to "update-initramfs -u" I get this error:

root@proxmoxlandet:~# update-initramfs -u
update-initramfs: Generating /boot/initrd.img-6.8.12-1-pve
Running hook script 'zz-proxmox-boot'..
Re-executing '/etc/kernel/postinst.d/zz-proxmox-boot' in new private mount namespace..
No /etc/kernel/proxmox-boot-uuids found, skipping ESP sync.

 

[alkaizer]

Its privledged

Its privledged , I think this is related to another problem I m not able to "update-initramfs -u" I get this error:

root@proxmoxlandet:~# update-initramfs -u
update-initramfs: Generating /boot/initrd.img-6.8.12-1-pve
Running hook script 'zz-proxmox-boot'..
Re-executing '/etc/kernel/postinst.d/zz-proxmox-boot' in new private mount namespace..
No /etc/kernel/proxmox-boot-uuids found, skipping ESP sync.

go to /dev/dri/ in your proxmox host and run ls -l

look at the numbers after the group name should be something like 226, 0 and 226, 128

remember those numbers for card0 and render128 and then edit you lxc conf file
/etc/pve/lxc/###.conf

add
lxc.cgroup2.devices.allow: c 226:0 rwm
lxc.cgroup2.devices.allow: c 226:128 rwm
lxc.mount.entry: /dev/dri/card0 dev/dri/card0 none bind,optional,create=file
lxc.mount.entry: /dev/dri/renderD128 dev/dri/renderD128 none bind,optional,create=file

make sure you use the correct numbers for 226:0 and 226:128 based on what your output was for ls -l in /dev/dri

 

[helio]

go to /dev/dri/ in your proxmox host and run ls -l

look at the numbers after the group name should be something like 226, 0 and 226, 128

remember those numbers for card0 and render128 and then edit you lxc conf file
/etc/pve/lxc/###.conf

add
lxc.cgroup2.devices.allow: c 226:0 rwm
lxc.cgroup2.devices.allow: c 226:128 rwm
lxc.mount.entry: /dev/dri/card0 dev/dri/card0 none bind,optional,create=file
lxc.mount.entry: /dev/dri/renderD128 dev/dri/renderD128 none bind,optional,create=file

make sure you use the correct numbers for 226:0 and 226:128 based on what your output was for ls -l in /dev/dri

Thats the problem :
cd /dev/dri
/dev/dri: No such file or directory

 

[alkaizer]

Thats the problem :
cd /dev/dri
/dev/dri: No such file or directory

are you doing that in the lxc or on the proxmox host?

 

[helio]

are you doing that in the lxc or on the proxmox host?

On the proxmox host

 

[alkaizer]

On the proxmox host

did you try going to the VM that is shutdown and removing the gpu passthrough

 

[helio]

did you try going to the VM that is shutdown and removing the gpu passthrough

Yes I have.

 

[alkaizer]

Yes I have.

Yeah then Im not really sure what else without more info. If you didnt reboot the host after removing the igpu passthrough I would try that as a hail mary.

 

[helio]

Yeah then Im not really sure what else without more info. If you didnt reboot the host after removing the igpu passthrough I would try that as a hail mary.

Yes I have reboot the host after removing the igpu but no luck.

 

[alkaizer]

Yes I have reboot the host after removing the igpu but no luck.

ah damn well nevermind. Im all out of options

 

[helio]

ah damn well nevermind. Im all out of options

Thanks for trying to help .