We all know that introducing and supporting hardware in Linux mainline is challenging. At KernelCI, we want to understand how we can help move the ecosystem further upstream.
The challenges come from many angles:
- Invest in upstreaming and maintain the hardware support into Linux (and other projects) in the long-run
- Take IP protection in consideration, making sure that no critical information is shared in the open
- Time to market pressures for the product release
- Test mainline based products (continuously), making sure:
- The hardware remains functional
- No regressions appear
- Maintain test infrastructure, so testing can be timely and efficient
- Interact with the kernel community to improve the state of the art and fix issues
- And so much more
These challenges come from both technical and business aspects. On the one hand, there are still a lot of difficulties in interacting with the upstream community to implement drivers, review patches, and solve regressions. On the other hand, business stakeholders and decision-makers have a hard time understanding upstream practices, which contributes to insufficient investment to bring enough knowledge and resources for effective participation in upstream. KernelCI aims to help Hardware Vendors address both the technical and business obstacles.
With that in mind, KernelCI wants to start a discussion with all interested Hardware Vendors to share our experiences and pain points. We could then look ahead at how we can collaborate on improving the kernel integration processes and facilitate the maintenance of stable kernels in the long run. Stability and security needs are growing exponentially, with so much of our global infrastructure depending on the hardware and software we build.
If you are in Vienna, Austria on September 18th for Open Source Summit Europe and/or Linux Plumbers Conference, we invite you to join our in-person discussion to happen from 1:30pm to 3pm at LPC Room 1.34.
If you have any questions or comments, please contact us at kernelci-members@groups.io.