Re: [PATCH v3 01/13] dt-bindings: i3c: Add mipi-i3c-static-method to support SETAASA

On Tue, May 5, 2026 at 5:17 PM Guenter Roeck <[email protected]> wrote:
>
> On 5/5/26 14:32, Rob Herring wrote:
> > On Thu, Apr 23, 2026 at 11:23:11AM +0000, [email protected] wrote:
> >> Sashiko AI review found 3 potential issue(s):
> >
> > Does anyone read replies to the bot? If so, why was everyone hwmon list

everyone *but* the hwmon list...

> > dropped?
> >
>
> It appears that people either complain about getting AI feedback, or
> they complain that they don't. That is pretty much a lose-lose situation.
> For patches sent to linux-hwmon, Sashiko is configured to respond only
> to the mailing list and to the submitter in the hope to minimize the
> "I do not want to see this" feedback. That matches the configuration
> for several subsystems.
>
> As far as I can see it is not currently possible to limit AI review to
> patches actually associated with the given subsystem if sent to that
> subsystem's mailing list. I'll discuss with Roman to see if patches
> like this one can be filtered out when reviewing patches sent to
> hwmon mailing list.

No, I want the replies. I was just wondering why it is not reply-all
which is standard. I was also wondering why I was only sometimes
seeing the reviews. I have since figured that out and submitted a PR
to add DT. I set it to reply-all, so I guess there will be some
comment about that.

I don't think there's any solution if a patch hits 2 conflicting
configurations. Surely the folks that don't want to see it can figure
out how to delete or ignore emails... I have thousands of ignored
emails. It's not hard. :)

> >> - [Medium] The `mipi-i3c-static-method` schema incorrectly enforces `minimum: 0x1`, preventing devices from declaring no static address assignment support.
> >> - [Low] The `assigned-address` property description is now inaccurate and contradicts the new capabilities.
> >> - [Low] Missing macro definition for Bit 2 ('Other CCC') in `include/dt-bindings/i3c/i3c.h`.
> >> --
> >
> > This should go below the quoted commit msg.
> >
>
> Others (like me) want to see it at the very top to see a quick summary.
> Another lose-lose situation.

If there's anything kernel developers agree on, it is not top posting.
I don't see how sashiko is special here. Isn't below the commit
message close enough to the top? I don't mind so much about the list
of issues being at the top, but I really don't see why the commit
message is not quoted (I would guess that's just an artifact of
creating the reply out of the git tree it was applied to?). The commit
hash is also completely useless to anyone without access to the git
tree sashiko uses.

One thing I would love to see is reviews of the commit message itself.
I would love to never have to say "don't use 'This commit/patch'" or
"the commit msg should answer why, not what" again.

Rob