vllm - 💡(How to fix) Fix [RFC]: Request to clarify attribution/authorship for the RISC-V CPU backend PR chain (#20292, #32405, #36538, #36578) [1 participants]

Motivation.

I am opening this RFC to request clarification of the attribution/authorship history for the RISC-V CPU backend work.

This is not a request to remove the merged functionality. The motivation is to ensure that the development history for this work is recorded accurately and transparently.

The concern is that the visible PR chain does not fully reflect the implementation lineage. In particular:

• #32405 explicitly stated that it continues and extends the work from #20292.
• Later PRs #36538 and #36578 continued the same line of RISC-V CPU backend work, but the visible attribution chain no longer acknowledged #32405.
• As a result, the current merged record may give future readers an incomplete understanding of who implemented the earlier follow-on work and how the final merged result evolved.

Regarding authorship, the core follow-on implementation after #20292 was originally developed by me during my internship at ISCAS. After I left, that work was further continued and submitted as #32405. Because of this history, I believe the current public record should be clarified so that earlier implementation work is not omitted from the attribution chain.

Proposed Change.

I would like maintainers to consider the following project-level change in how this PR chain is documented:

1. Add an explicit maintainer note, PR update, or linked issue comment that documents the implementation lineage for the RISC-V CPU backend PR chain: #20292 -> #32405 -> #36538 -> #36578.
2. Acknowledge that #36538 / #36578 carried forward earlier implementation work rather than presenting the merged result as entirely new work in the final PR.
3. Where possible, distinguish between:
   • earlier follow-on implementation work,
   • later fixes/rebasing/integration changes,
   • and the final merge-ready adjustments.
4. More generally, establish the expectation that when a later PR substantially carries forward a closed or superseded PR, the earlier implementation chain should be explicitly credited in the final visible record.

I am not requesting code removal or revert. I am requesting an attribution/authorship clarification so that the project history is more accurate for maintainers, contributors, and future readers.

Feedback Period.

No response

CC List.

@hansu2022

Any Other Things.

I recognize that this is not a traditional architecture RFC. I am using the RFC template because the current issue categories do not appear to provide a better project-level venue for discussing attribution/authorship clarification after the final PR has already been merged.

I am happy to help clarify which parts belonged to the original follow-on implementation and which parts were later fixes, rebasing work, or integration changes.

I am raising this as an RFC because the final PR in this chain has already been merged, so this is now a project-level question about attribution and contribution history rather than a review comment on a single open PR.

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.