Hi Biju,
On Wed, Nov 16, 2022 at 9:14 AM Biju Das <biju.das.jz@xxxxxxxxxxxxxx> wrote:
> > Subject: Re: [PATCH v2 1/3] dt-bindings: arm: renesas: Document Renesas
> > RZ/V2M System Configuration
> > On Mon, Nov 14, 2022 at 8:22 PM Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > wrote:
> > > > Subject: Re: [PATCH v2 1/3] dt-bindings: arm: renesas: Document
> > > > Renesas RZ/V2M System Configuration On Mon, Nov 14, 2022 at 6:05 PM
> > > > Biju Das <biju.das.jz@xxxxxxxxxxxxxx> wrote:
> > > > > > From: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx> On
> > > > > > 14/11/2022 17:44, Biju Das wrote:
> > > > > > >>>>> If the hardware manual said this is called "foo-whatever-
> > name"
> > > > > > >>>>> or "rz85736dfnx2", you would use it as well?
> > > > > > >>>>>
> > > > > > >>>>> Node names should be generic.
> > > > > > >>>
> > > > > > >>> I got the following feedback from LSI team.
> > > > > > >>>
> > > > > > >>> *The reason why we use "configuration" and not "controller"
> > > > > > >>> is the possibility that users might take it to mean
> > > > > > >>> "control" of system operation (product operation). This unit
> > > > > > >>> determines how the whole LSI
> > > > > > >> configure to operate LSI, so we use "configuration" for this
> > IP name.
> > > > > > >>>
> > > > > > >>> As per this it is not a controller. It is a Configuration
> > IP.
> > > > > > >>>
> > > > > > >>> What should be generic node name for system configuration
> > IP?
> > > > > > >>
> > > > > > >> Then it's closer to chipid and should be in hwinfo?
> > > > > > >
> > > > > > > You mean, sys: hwinfo@a3f03000 instead of sys: system-
> > > > > > configuration@a3f03000 ??
> > > > > >
> > > > > > Rather be put in hwinfo directory. Typical name is "chipid", but
> > > > > > I don't know if your device is a chipid.
> > > > >
> > > > > This IP is for sure not a chipid. We can detect SoC version. That
> > > > > is one of the functions provided by this IP.
> > > > >
> > > > > > > Also the same IP block is present on rz/v2ma as well?
> > > > > > >
> > > > > > > Shall I use the same file renesas,rzv2m-sys.yaml for both
> > > > > > > r9a09g011-sys and
> > > > > > r9a09g055-sys?
> > > >
> > > > Are they sufficiently similar? E.g. we have renesas,rst.yaml for the
> > > > various R-Car reset controllers, which have different register
> > layouts.
> > >
> > > I have been told both RZ/V2M and RZ/V2MA are exactly same, but later
> > does not have ISP module.
> >
> > OK...
> >
> > > > If the SoC version register is located at the same offset, it might
> > > > be worthwhile to add a family-specific compatible value, too, so the
> > > > soc_device driver doesn't have to contain a big list to match
> > against.
> > >
> > > The SoC version registers are located at the same offset.
> > > But there is no way to distinguish both the SoCs.
> >
> > So they really are the same SoC, with the ISP disabled (by fuses?).
> >
> > Then I guess the most sensible thing to do is to just use
> >
> > compatible = "renesas,r9a09g055", "renesas,r9a09g011"
> >
> > at the top level, and keep all other compatible values unchanged?
> > Cfr. what we did for R-Car Gen3e (r8a779mX).
>
> I started looking into PWM driver for V2M. For Linux we can use 7 channels
> Out of 16 as other 9 channels are reserved for ISP. Where as V2MA we can use all the 16 channels.
>
> So for clk, we need to use separate compatible, as we don't want to register clocks
> assigned for ISP to Linux.
>
> That is ok right??
Yes, that's fine.
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
