Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add support for >= 550.40.x drivers to the toolkit container #324

Closed
elezar opened this issue Jan 31, 2024 · 0 comments · Fixed by #330
Closed

Add support for >= 550.40.x drivers to the toolkit container #324

elezar opened this issue Jan 31, 2024 · 0 comments · Fixed by #330
Assignees

Comments

@elezar
Copy link
Member

elezar commented Jan 31, 2024

As part of generating management CDI specifications, the container-toolkit container creates control device nodes. This requires reading the device major number from /proc/devices. This fails in drivers from 550.40.x since the nvidia-frontend module was renamed nvidia.

See: https://www.nvidia.de/Download/driverResults.aspx/218124/en-us

Changed the name visible in /proc/devices of NVIDIA devices and theNVIDIA control device from "nvidia-frontend" to "nvidia" and "nvidiactl".Scripts which parse /proc/devices (such as udev rules) may need to beupdated. Note that the conventional /dev device paths like /dev/nvidia0and /dev/nvidiactl remain unchanged.

This can be triggered by starting the toolkit container with CDI enabled to trigger the code path but will also affect the nvidia-ctk system create-dev-char-symlinks command.

Note that nvidia-modprobe hardcodes the major number: https://github.com/NVIDIA/nvidia-modprobe/blob/d6bce304f30b6661c9ab6a993f49340eafca7a7e/modprobe-utils/nvidia-modprobe-utils.c#L85

We should probably do the same as this will ensure that we remain compatible across driver versions.

@elezar elezar changed the title Add support for 550.x drivers to the toolkit container Add support for >= 550.40.x drivers to the toolkit container Jan 31, 2024
@elezar elezar assigned elezar and tariq1890 and unassigned elezar Jan 31, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
2 participants