[PATCH v2 4/8] dt-bindings: sifive,ccache0: Support StarFive JH7110 SoC

[Rob Herring]

On Tue, Nov 22, 2022 at 11:35:28AM +0100, Emil Renner Berthing wrote:
> On Tue, 22 Nov 2022 at 11:16, Hal Feng <hal.feng at starfivetech.com> wrote:
> >
> > On Tue, 22 Nov 2022 10:01:30 +0000, Conor Dooley wrote:
> > > On Tue, Nov 22, 2022 at 05:55:57PM +0800, Hal Feng wrote:
> > > > On Tue, 22 Nov 2022 09:07:26 +0000, Conor Dooley wrote:
> > > > > On Tue, Nov 22, 2022 at 04:40:23PM +0800, Hal Feng wrote:
> > > > > > On Fri, 18 Nov 2022 19:39:52 +0800, Conor Dooley wrote:
> > > > > > > On Fri, Nov 18, 2022 at 11:37:50AM +0000, Conor Dooley wrote:
> > > > > > > > On Fri, Nov 18, 2022 at 09:17:10AM +0800, Hal Feng wrote:
> > > > > > > > > From: Emil Renner Berthing <kernel at esmil.dk>
> > > > > > > > >
> > > > > > > > > This cache controller is also used on the StarFive JH7110 SoC.
> > > > > > > >
> > > > > > > > "... and configured identically to that of the FU740"?
> > > > > > > > Anyways,
> > > > > > > > Reviewed-by: Conor Dooley <conor.dooley at microchip.com>
> > > > > > >
> > > > > > > Actually, after looking at the next patch - why can you not fall back to
> > > > > > > the fu740 one since you appear to have the same configuration as it?
> > > > > >
> > > > > > Right, I will drop this patch and use "sifive,fu740-c000-ccache" as
> > > > > > compatible in dts.
> > > > >
> > > > > Uh, that's not quite what I was suggesting. Rather than using that one
> > > > > in isolation, you can do the following in your dt:
> > > > > "starfive,jh7110-ccache", "sifive,fu740-c000-ccache"
> > > > >
> > > > > And then in the driver we need to make no changes - unless down the line
> > > > > we find some sort of issue that requires special handling etc. There's
> > > > > no harm in having a "starfive,jh7110-ccache" IMO.
> > > >
> > > > Just like what microchip did as blow?
> >
> > below
> >
> > > >
> > > > Documentation/devicetree/bindings/riscv/sifive,ccache0.yaml:
> > > > properties:
> > > >   compatible:
> > > >     oneOf:
> > > >       - items:
> > > >           - enum:
> > > >               - sifive,ccache0
> > > >               - sifive,fu540-c000-ccache
> > > >               - sifive,fu740-c000-ccache
> > > >               - starfive,jh7110-ccache
> > > >           - const: cache
> > > >       - items:
> > > >           - const: microchip,mpfs-ccache
> > > >           - const: sifive,fu540-c000-ccache
> > > >           - const: cache
> > >
> > > No, I don't think this is correct either. You'd do something like:
> > >
> > > >       - items:
> > > >           - const: starfive,jh7110-ccache
> > > >           - const: sifive,fu740-c000-ccache
> 
> For the record I don't think the line above should be there. The
> fu7400-c000 is a specific tapeout and pretending the JH7110 is that
> tapeout is not right. Especially when there is already the
> "sifive,ccache0" string for the generic IP.

All it really says is that this h/w will work with any client (OS) 
that understands 'sifive,fu740-c000-ccache'. Maybe 'sifive,ccache0' is 
sufficient too, IDK.

Rob