Re: [PATCH 2/2] arm64: dts: renesas: r8a77961: ulcb-kf: Initial device tree

[Geert Uytterhoeven]

Hi Eugeniu,

On Thu, Oct 29, 2020 at 5:34 PM Eugeniu Rosca <erosca@xxxxxxxxxxxxxx> wrote:
> On Thu, Oct 29, 2020 at 03:09:10PM +0100, Geert Uytterhoeven wrote:
> > On Thu, Oct 29, 2020 at 2:38 PM Eugeniu Rosca <erosca@xxxxxxxxxxxxxx> wrote:
> > > Create a dedicated DTB for M3-ES3.0 + ULCB + Kingfisher combo.
> > > Inspire from the pre-existing ULCB-KF device trees:
> > >
> > > $ ls -1 arch/arm64/boot/dts/renesas/*ulcb-kf.dts
> > > arch/arm64/boot/dts/renesas/r8a77950-ulcb-kf.dts
> > > arch/arm64/boot/dts/renesas/r8a77951-ulcb-kf.dts
> > > arch/arm64/boot/dts/renesas/r8a77960-ulcb-kf.dts
> > > arch/arm64/boot/dts/renesas/r8a77965-ulcb-kf.dts
> > >
> > > Signed-off-by: Eugeniu Rosca <erosca@xxxxxxxxxxxxxx>
> >
> > Reviewed-by: Geert Uytterhoeven <geert+renesas@xxxxxxxxx>
> > i.e. will queue in renesas-devel for v5.11.
>
> Thank you for the prompt review!
>
> > > +       compatible = "shimafuji,kingfisher", "renesas,m3ulcb",
> > > +                    "renesas,r8a77961";
> >
> > Can you please send a patch to add this combo to
> > Documentation/devicetree/bindings/arm/renesas.yaml?
>
> I would happily do so if you resolve below concerns.
>
> Since the inception of the Kingfisher extension board description in
> v4.15-rc1 commit 5418a900412699 ("arm: shmobile: Document Kingfisher
> board DT bindings"), nobody attempted describing the SoC+ULCB+KF
> combinations in spite of four of such DT configurations being actively
> used and maintained, i.e. r8a779{50,51,60,65}-ulcb-kf.

I'm a bit confused. We do have:

      - description: Kingfisher (SBEV-RCAR-KF-M03)
        items:
          - const: shimafuji,kingfisher
          - enum:
              - renesas,h3ulcb
              - renesas,m3ulcb
              - renesas,m3nulcb
          - enum:
              - renesas,r8a7795
              - renesas,r8a7796
              - renesas,r8a77965

> So, if we start documenting the r8a77961-ulcb-kf combo as a board, this
> raises below questions:
>
>   => should the missing 4 SoC+ULCB+KF instances be documented as well?
>   => should a new compatible string be created for each such HW combo,
>      e.g. "renesas,<soc>-ulcb-kf"?
>
> I feel none of the above is really needed, based on the patterns
> established in Documentation/devicetree/bindings/arm/renesas.yaml,
> but I might be wrong. Thoughts/suggestions appreciated.

I don't think we want to add new compatible string to describe each
combo. Just add "renesas,r8a77961" to the last enum?

> IMHO one thing which is certainly worth clarifying and fixing is the
> KF revision currently documented in renesas.yaml, i.e. M03.
>
> Shimafuji released at least M04, M05 and M06 revisions of KF (nicely
> compared at https://elinux.org/R-Car/Boards/Kingfisher#Change_point).
>
> The question is, does the community intend to support M03 through M06
> (in which case all of them might need an entry in the documentation) or
> anything which is earlier than M06 has to be considered deprecated (in
> which case renesas.yaml would need a simple s/M03/M06/ update)?

I'm not that familiar with KingFisher and the various revisions.
Do these differences have an impact on the software side?
The diodes and filters probably don't.
The I2C repeaters are PCA9548ADB on M03, hence they use a Linux
driver. By what have they been replaced?
What's the nature of the MOST ↔ GPS Function select register change?

Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds