Re: [PATCH 2/2] arm64: dts: renesas: rzg2ul-smarc: Add /omit-if-no-ref/ to pinmux

[Lad, Prabhakar]

Hi Geert,

On Wed, Oct 26, 2022 at 8:29 AM Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
>
> Hi Prabhakar,
>
> On Wed, Oct 26, 2022 at 12:39 AM Lad, Prabhakar
> <prabhakar.csengg@xxxxxxxxx> wrote:
> > On Tue, Oct 25, 2022 at 9:13 AM Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
> > > On Thu, Sep 15, 2022 at 6:53 PM Prabhakar <prabhakar.csengg@xxxxxxxxx> wrote:
> > > > From: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx>
> > > >
> > > > In preparation to re-use the RZ/G2UL SMARC SoM and carrier DTS/I with the
> > > > RZ/Five add /omit-if-no-ref/ keyword to pinmux entries as the support for
> > > > RZ/Five SMARC EVK will be gradually added.
> > > >
> > > > Once we have full blown support for RZ/Five SMARC EVK we can get rid of
> > > > the /omit-if-no-ref/ keyword.
> > > >
> > > > Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx>
> > >
> > > Thanks for your patch!
> > >
> > > I finally had a deeper look at this...
> > >
> > > Why do you want to disable these nodes? While they are indeed not
> > > used yet on RZ/Five, they are valid hardware descriptions for the
> > > RZ/Five SMARC EVK, and their presence doesn't harm anything.
> > >
> > > I do see a valid use case for marking pin control subnodes with
> > > /omit-if-no-ref/: you can provide all possible configurations as a
> > > convenience for the user, so the user no longer has to look up the
> > > numeric parameters of the RZG2L_PORT_PINMUX() macros.
> > > But IMHO those would belong in the SoC-specific .dtsi, not in a
> > > board .dtsi.  See e.g. the massive use of /omit-if-no-ref/ in sunxi
> > > and rockchip .dtsi files.
> > >
> > > Am I missing something?
> > >
> > My intention was to keep the DTB as minimal as possible so that it
> > includes just the required pinmuxes which were enabled on the RZ/Five.
> > For example [0], [1] we do delete the pinctrl for the nodes which are
> > marked as disabled. Do you think we should drop it?
>
> You mean
>
>     /delete-property/ pinctrl-0;
>     /delete-property/ pinctrl-names;
>
> ?
> These do not delete pinctrl subnodes, but pinctrl properties in disabled
> device nodes pointing to pinctrl subnodes.  The actual pinctrl subnodes
> are still present.
>
Ahh right the pinctrl subnodes will still remain.

Cheers,
Prabhakar