Re: [PATCH v9 03/23] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant

[AngeloGioacchino Del Regno]

Il 10/03/26 19:21, Rob Herring ha scritto:
> On Mon, Mar 9, 2026 at 5:04 AM Nicolas Frattaroli
> <nicolas.frattaroli@xxxxxxxxxxxxx> wrote:
> > On Saturday, 7 March 2026 19:01:17 Central European Standard Time Martin K. Petersen wrote:
> > > Nicolas,
> > >
> > > > > "ufs" is redundant as all the clocks are for UFS. Same comment on prior
> > > > > patch.
> > > >
> > > > Is this naming a big enough concern to block this series with two
> > > > explicit acks on this patch that fixes a wholly broken and useless
> > > > binding?
> > >
> > > It is if it comes from one of the DT maintainers.
> > >
> > > > I am trying to put out this dumpster fire of a downstream turd that
> > > > made its way into mainline as the review process has been completely
> > > > subverted, and is only getting worse with each passing month
> > >
> > > This has to stop. Please read Documentation/process/code-of-conduct.rst.
>
> I have little doubt that that is an accurate description of
> downstream. And if properties are getting added without bindings, then
> that is certainly a problem that should be complained about.
>
> > >
> >
> > I apologise for my tone, it's my frustration getting the better of me.
> >
> > I'll be handing off this series to someone else, so you won't have to
> > deal with me anymore.
> >
> > I do ask however that you don't apply patches from MediaTek blindly;
> > if there's code to read an OF property, and that OF property is not
> > in the binding, then the patch should be rejected, even if there's an
> > Ack from the MediaTek maintainer.
>
> There's functionality to find undocumented compatibles in kernel code
> (make dt_compatible_check), but not properties. Sounds like I need to
> add that.

Rob: yes please, that would help a lot in general, not just with this UFS driver.

Cheers,
Angelo

> Rob