need inputs on enabling CoreOS on networking device

[Prabhu Balakannan]

Experts,
  I would like to get inputs on enabling CoreOS on our networking device.
We currently have networking device (Switch/Router) running proprietary OS. The current proprietary network OS we are running mainly interacts with hardware that contains switching and routing ASICs for programming perspective.
 We want to understand whether we can use CoreOS and include our KLM kernel module to control these devices in the hardware for programming and also we have a background SDK process that interacts with the KLM. Can the background SDK process run directly on CoreOS instead of container/docker?

-thanks
Prabhu.

[Brian Harrington]

Prabhu,

We have a number of users who over the users have made something like this work.  In these cases they're building the module from source.  If this is a driver for which you do not have any of the source code it will likely be a challenge.

http://www.emergingstack.com/2016/01/10/Nvidia-GPU-plus-CoreOS-plus-Docker-plus-TensorFlow.html
https://github.com/Avalanche-io/coreos-nvidia

http://tleyden.github.io/blog/2014/11/04/coreos-with-nvidia-cuda-gpu-drivers/

Traun (the gentleman who did the last post) has also done great work in making Couchbase "cloud native" atop Kubernetes as well.

In terms of having the background process talk to the kernel module, that shouldn't be an issue. The most important piece will be to identify which kernel capabilities are needed to do this (http://man7.org/linux/man-pages/man7/capabilities.7.html).  For the container which will load the module it will need CAP_SYS_MODULE.  For the container running the SDK process you would be better equipped to answer which capabilities are needed.

Hope this helps!

--Brian 'redbeard' Harrington

[Brian Harrington]

Pardon my confusing wording... I meant to say "We have had a number of users over the years who have made something like this work"

Long day and distractions while writing this!

[Prabhu Balakannan]

Thank you so much Brian for the links..
I had gone through few links some days back where it says, we need to download the source code and try compiling the module (on target device itself) with the relevant compiler and install and probe the device.
In our case, the ASIC will be the PCI endpoint, so the core pci driver should be able to enumerate the device. On top of that, we have few own KLM to program and read data
from asic. In that case, do we need to copy our source code to the container and compile it there and do insmod?

-thanks
Prabhu.