On Thu, Sep 18, 2025 at 4:22 AM Dumitru Ceara <[email protected]> wrote:
>
> Hi Han,
>
> On 9/17/25 7:09 PM, Han Zhou wrote:
> >
> >
> > On Wed, Sep 17, 2025 at 8:17 AM Numan Siddique <[email protected]
> > <mailto:[email protected]>> wrote:
> >>
> >> On Wed, Sep 17, 2025 at 6:42 AM Dumitru Ceara <[email protected]
> > <mailto:[email protected]>> wrote:
> >> >
> >> > On 9/17/25 11:07 AM, Ales Musil wrote:
> >> > > On Tue, Sep 16, 2025 at 6:38 PM Mark Michelson via dev <
> >> > > [email protected] <mailto:[email protected]>> wrote:
> >> > >
> >> > >> Signed-off-by: Mark Michelson <[email protected]
> > <mailto:[email protected]>>
> >> > >> ---
> >> > >>  .../internals/contributing/submitting-patches.rst   | 13 +++++++
> > ++++++
> >> > >>  1 file changed, 13 insertions(+)
> >> > >>
> >> > >> diff --git a/Documentation/internals/contributing/submitting-
> > patches.rst
> >> > >> b/Documentation/internals/contributing/submitting-patches.rst
> >> > >> index c82e49124..1d6fa5e45 100644
> >> > >> --- a/Documentation/internals/contributing/submitting-patches.rst
> >> > >> +++ b/Documentation/internals/contributing/submitting-patches.rst
> >> > >> @@ -366,6 +366,19 @@ OVN please sign off your work.  The sign off
> >> > >> certifies the following:
> >> > >>
> >> > >>  See also http://developercertificate.org/ <http://
> > developercertificate.org/>.
> >> > >>
> >> > >> +AI-assisted Contributions
> >> > >> +-------------------------
> >> > >> +
> >> > >> +OVN is a Linux Foundation project, and the Linux Foundation's
> > policy for
> >> > >> +AI-generated patches can be found here:
> >> > >> +https://www.linuxfoundation.org/legal/generative-ai <https://
> > www.linuxfoundation.org/legal/generative-ai> .
> >> > >> +
> >> > >> +OVN allows the use of AI assistants in producing patches.
> > Contributions
> >> > >> +which have used an AI assistant should disclose the use of the
> > assistant
> >> > >> by
> >> > >> +using the "Assisted-by" tag. While AI-assisted patches are
> > allowed, the
> >> > >> author
> >> > >> +of the patch is ultimately responsible for ensuring that the AI-
> > generated
> >> > >> code
> >> > >> +has not violated any terms of the Developer's Certificate of
Origin.
> >> > >> +
> >
> > All looks good to me except that the standard of "Assisted-by" tag is
> > not very clear. For example, does use of auto-completion count as AI-
> > assisted? In the past auto-completion was very standard in IDE's but
> > nowadays it is powered by AI and is much more intelligent and can
> > generate small/big chunks of code. If so, developers using AI powered
> > IDE should always add the tag Assisted-by, and it seems not very useful
> > information. As for myself, I have been exploring Cursor recently and am
> > in a mixed mode of VIM and Cursor :)
> >
>
> I would guess it's up to the author of the patch to decide when the AI
> powered IDE did more than just "type what the author already wanted to
> type but just quicker".  Hence when to add the Asissted-by tag or not.
>
> > If the purpose of the tag is to remind reviewers to pay extra attention
> > to AI-generated code, it may be useful, but we should insist that no
> > matter if the tag is added or not, the author of the patch is ultimately
> > responsible for the quality of the code, and as a good practice, she/he
> > should carefully review/modify the code before submitting the patch.
> >
>
> I would argue that with any patch submission (AI assisted or not) the
> author should try to ensure the quality of their code is up to standard.
>  So I don't know if we really need to add it here.  But I'm not against
> doing that if you see value in it.
>
> > BTW, any suggestion for the value of the Assisted-by tag? For example,
> > should it be just the IDE (such as Cursor, Windsurf, ChatGPT etc.) or
> > the actual LLM model being used? Or just Assisted-by: AI?
> >
>
> Here are a couple of different policies from other projects.  I've been
> shown these examples internally:
>
>
https://github.com/avocado-framework/avocado/blob/master/docs/source/guides/contributor/chapters/ai_policy.rst#transparency-and-disclosure
>
>
https://github.com/patternfly/.github/blob/main/CONTRIBUTING.md#ai-assisted-development-guidelines
>
> Hope that helps.

Thanks Dumitru. Also discussed with Mark at the ovn meeting. Mark mentioned
that there are benefits such as "There's always a very slight chance it
could be a useful thing to know when doing analysis of bugs.". I think it
makes sense.

I think we can start with the current statement and see how it works. We
can refine the rules and the content of the tag in the future if necessary.

Acked-by: Han Zhou <[email protected]>

>
> Regards,
> Dumitru
>
> > Best regards,
> > Han
> >
> >> > >>  Feature Deprecation Guidelines
> >> > >>  ------------------------------
> >> > >>
> >>
> >> > >> --
> >> > >>
> >> > > Looks good to me, thanks. I have also added all OVN + few OVS
> > maintainers
> >> > > (as it might be useful for ovs too) to cc.
> >> > >
> >> > > Acked-by: Ales Musil <[email protected] <mailto:[email protected]>>
> >> > >
> >> >
> >> > Hi Mark, Ales,
> >> >
> >> > +Eelco
> >> >
> >> > Looks good to me too, thanks!
> >> >
> >> > Acked-by: Dumitru Ceara <[email protected] <mailto:[email protected]
>>
> >>
> >> Looks good to me too, thanks!
> >>
> >> Acked-by: Numan Siddique <[email protected] <mailto:[email protected]>>
> >>
> >> Numan
> >>
> >> >
> >> > Regards,
> >> > Dumitru
> >> >
> >> >
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to