Re: [PATCH v7 0/8] [RFC] Rust device / driver abstractions

[Danilo Krummrich <dakr@redhat.com>]

On Wed, Mar 27, 2024 at 02:31:16PM +0100, Miguel Ojeda wrote:
> On Wed, Mar 27, 2024 at 12:25 PM Danilo Krummrich <dakr@redhat.com> wrote:
> >
> > In [2] I was asking about the preferred way to get some immediate discussions
> > going. When I proposed to send links to the corresponding branches to the
> > mailing list or send a patch series (for rust-device I ended up doing both) I
> > did not hear any contradiction.
> 
> Sending patch series directly is something you/Philipp proposed. And
> that is fine -- nobody can tell you to not send patches. But, from
> experience, we know sometimes it is best to get in the same page
> and/or room.
> 
> I specifically suggested reporting your progress in Zulip and/or the
> mailing lists for that reason, and because lifting code from other
> places like `rust` generally requires revisiting it first like Wedson
> mentioned.

Sorry, I wasn't aware that you prefer to have some extra process / stage of
revisiting / reviewing of existing patches / code that is picked up from
R4L/rust.

Maybe it would have been good to give me this pointer when I asked: "how (and
where) to get specific code discussed." [1]

Maybe I'm also misunderstanding what you mean by "revisiting".

Anyway, how can we proceed? Can we just continue with this series and improve
things by further review? Do you prefer to approach it differently?

> 
> In addition, we always suggest pinging directly the original authors
> too before submitting patches from them, to avoid this kind of thing.

As already mentioned, fully agree.

> 
> Moreover, I would recommend tagging more prominently the patches as
> staging/RFC/... (i.e. each of them). I know you have "RFC" in the
> cover letter one, but I am aware of at least one person that did not
> realize initially the patches were actually RFC. In addition, I would

Sure, I can try to make it more obvious by adding --subject-prefix="RFC: PATCH"
to git-format-patch.

> suggest that the cover letter explains where and how the patches where
> lifted from, so that people are aware of their state etc. I would also

I think that's actually explained in [2], which I referenced in the cover
letter. If you think there should be additional information, please let me know,
I'm happy to add it.

> recommend, for clarity and even if out of deference only, to mention
> whether the authors of the patches you are carrying were aware of this
> submission (i.e. if you explicitly heard from them).

Noted. But just to clarify (and I'm clearly not saying you implied something
else), not doing so is not meant to be understood as doing it without proper
deference. I was very careful in setting up proper authorship and co-authorship
(e.g. for fixes that I squashed into some commits).

> 
> Finally, it is not always possible "to get some immediate discussions
> going" (i.e. emphasis on "immediate"). It all depends on who / which
> subsystem / etc. you are dealing with.
> 
> Cheers,
> Miguel
> 

[1] https://rust-for-linux.zulipchat.com/#narrow/stream/415254-DRM/topic/Topic.20branches.20for.20staging.20Rust.20abstractions/near/426750399
[2] https://lore.kernel.org/dri-devel/Zfsj0_tb-0-tNrJy@cassiopeiae/T/#u