Re: [PATCH v9 2/2] platform: Add initial synology microp driver

On April 26, 2026 11:04:55 AM GMT-03:00, Miguel Ojeda <[email protected]> wrote:
>On Sun, Apr 26, 2026 at 3:44 PM Markus Probst <[email protected]> wrote:
>>
>> On Sun, 2026-04-26 at 15:37 +0200, Miguel Ojeda wrote:
>> > On Fri, Apr 24, 2026 at 1:59 PM Markus Probst <[email protected]> wrote:
>> > >
>> > > Since this is an AI reply, I am not sure if I am supposed to reply to
>> > > it. Anyway, the answers:
>> >
>> > Up to you -- some people is just fixing the patch and providing a
>> > changelog in the next version. But it is great that you are taking a
>> > look at the report! I hope it is at least somewhat useful.
>> >
>> > We may eventually just send them to the mailing list like other lists
>> > in the kernel are doing (in which case replying may be more useful),
>> > or we may disable Sashiko completely if it is not worth our time. We
>> > are trying to see how people feel about it.
>> It would be an improvement if the bot had previous replies as context.
>> So If I reply that the found potential issue is a false positive, that
>> it won't appear in the next patch revision review again.
>
>Yeah, that could be nice -- not sure if there is such a feature or if
>Roman et al. (Cc'ing Roman) are thinking about a feature like that
>(especially for lists that send the report to the mailing list).
>
>To avoid using up a lot of context, perhaps it could be a single-line
>"command" in a reply, rather than sending all the replies to the LLM.
>
>For instance, if someone writes:
>
>    > Some lines written by Sashiko...
>    > ...
>    > ...
>
>    @sashiko: false positive
>
>then Sashiko could grep for that, pick the lines above that line (i.e.
>the quoted report from Sashiko) and give just those few lines and a
>note that that it was considered a false positives into the context.

But if the AI is unconvinced that it is a false positive?

Replying with a reason justifying why one thinks it's a false positive seems reasonable. 

I'm trying to avoid the, sometimes long, back and forth by using a local sashiko instance, using corporate keys (will try with local models soon) to reduce the traffic in the mailing list.

Plainly stating that's a false positive seems to harsh and potentially harmful, no? 

- Arnaldo

>
>> > From what I see here, there were at least 2 non-trivial true positive, right?
>> Yes.
>
>Thanks for confirming! I hope it was (is) useful.
>
>Cheers,
>Miguel
>

- Arnaldo