On Mon, Sep 8, 2025 at 3:32 PM Alexei Starovoitov <alexei.starovoitov@xxxxxxxxx> wrote: > Posting a patch generated by AI without proof reading is not ok. > Wrong documentation is much worse than no documentation. > If you don't know what these fields do don't add random comments > to them. Read the code and document based on your understanding. > Copy pasting AI and throwing it at maintainers to review is inconsiderate. That's not what was done here. An AI has the advantage of reading mailing list posts and gathering wider kernel context. As the code is devoid of comments and I lack a fuller picture it was the best I could do. Were the code commented then I would have just migrated the comments that are missing. Here is the commit message that captures what was done: Recently diagnosing a regression [1] would have been easier if struct bpf_prog_info had some comments explaining its usage. As I found it hard to generate comments for some parts of the struct, what is here is a mix of mostly hand written, but some AI written, comments. [1] https://lore.kernel.org/lkml/CAP-5=fWJQcmUOP7MuCA2ihKnDAHUCOBLkQFEkQES-1ZZTrgf8Q@xxxxxxxxxxxxxx/ Anyway, you can decide to ignore this patch. Feel free to write a better one yourself. Take this patch as a bug report in the form of a patch trying to address the problem. I'm sorry that this has caused you such a great offense. Thanks, Ian
