netdev foundation

Hi!

We are pleased to announce that we have formed a foundation (under the auspices of Linux Foundation) to pay the bills for various systems which grew around netdev. The initial motivation was to move the NIPA testing outside of Meta, so that more people can help and contribute. But there should be sufficient budget to sponsor more projects.

We set up a GitHub repo to track proposed projects:

https://github.com/linux-netdev/foundation

The README page provides more information about the scope, and the process. We don't want to repeat all that information - please refer to the README and feel free to comment or ask any questions here.

And please feel free to suggest projects!

- netdev foundation TSC

To follow-up on Jakub's announcement:

As part of the process of establishing a lab for CI the TSC is looking for suppliers for HW hosting and (optionally) CI contracting services. So if anyone has contacts and so on to this end please feel more than welcome to contact Jakub and myself who are coordinating this activity within the TSC.

- Netdev Foundation TSC

I see a number of people opened the link (or maybe like most Internet traffic these days it's just AI scraping bots? :D) but no project proposals were added, yet.

I'd like to sincerely encourage proposing whatever you think would benefit the project, the community or simply make your life contributing easier. I spent some time last year trying to convince LF to create a "participatory budget", to make sure that the foundation acts on real community needs. That went nowhere. Perhaps "the community" doesn't have any coherent needs. IOW we don't usually have a way to pitch ideas to $foundations because there are no ideas to pitch, not the other way around. I'm hoping netdev foundation will disprove that.

Once again, please feel free to suggest things we can buy, sponsor, contract someone to build for us, etc. And perhaps less obviously - please comment / like the GH issues if you'd like to be involved in overseeing one of the projects. Unfortunately I don't think we can CC a public ML when we discuss the proposals with suppliers and such.

Hello Jakub,

I hope I didn't misunderstand what you said here. I keep thinking if there are other projects related to kernel development. IIUC, yes, there are some projects, something like xdp-project which heavily relies on the kernel but is equipped with its own standalone tests/benchmarks/documentations?

If you're referring to that kind of project, it would be easy to list a few potential directions for people on the mailing list to take:
1. RDMA impl/test.
2. Various upper layer stacks with the help of AF_XDP, which I am currently working on.
3. Useful debug tools/trace tools that are good to the real debug process (like writing bpftrace)

    // a few side notes here, let me put BPF timestamping as an

example. For now the selftests only show how it works overall, but not manifest how to write it is the best fit for the production load. I have a local one and hope more people can use it and polish it but I don't know where to put it except my personal github...
4. some other interesting kernel modules that are apparently not suitable for landing on the kernel.
...

I wonder if I'm digressing a bit...

Thanks,
Jason

Project in this context means something we can fund, for example adding a specific feature to a specific tool, like patchwork.