Re: [PATCH 2/3] dt-bindings: timer: sifive,clint: add compatible for OpenC906

[Icenowy Zheng]

在 2022-12-05星期一的 10:36 +0000，Conor Dooley写道：
> On Fri, Dec 02, 2022 at 02:12:54PM +0800, Icenowy Zheng wrote:
> > 在 2022-12-01星期四的 19:18 +0000，Conor Dooley写道：
> > > On Wed, Nov 30, 2022 at 12:13:30PM -0600, Rob Herring wrote:
> > > > On Tue, Nov 22, 2022 at 03:41:27PM +0800, Icenowy Zheng wrote:
> > > > > 
> > > > > 
> > > > > 于 2022年11月22日 GMT+08:00 下午3:35:48, Krzysztof Kozlowski
> > > > > <krzysztof.kozlowski@xxxxxxxxxx> 写到:
> > > > > > On 22/11/2022 08:18, Icenowy Zheng wrote:
> > > > > > > 在 2022-11-21星期一的 11:06 +0100，Krzysztof Kozlowski写道：
> > > > > > > > On 21/11/2022 05:17, Icenowy Zheng wrote:
> > > > > > > > > T-Head OpenC906 is a open-source-licensed fixed-
> > > > > > > > > configuration of
> > > > > > > > > C906,
> > > > > > > > > which is now public and able to be integrated.
> > > > > > > > > 
> > > > > > > > > Add a compatible for the CLINT shipped as part of
> > > > > > > > > OpenC906, which
> > > > > > > > > should
> > > > > > > > > just be ordinary C9xx CLINT.
> > > > > > > > > 
> > > > > > > > > Signed-off-by: Icenowy Zheng <uwu@xxxxxxxxxx>
> > > > > > > > > ---
> > > > > > > > >  Documentation/devicetree/bindings/timer/sifive,clint
> > > > > > > > > .yam
> > > > > > > > > l | 1 +
> > > > > > > > >  1 file changed, 1 insertion(+)
> > > > > > > > > 
> > > > > > > > > diff --git
> > > > > > > > > a/Documentation/devicetree/bindings/timer/sifive,clin
> > > > > > > > > t.ya
> > > > > > > > > ml
> > > > > > > > > b/Documentation/devicetree/bindings/timer/sifive,clin
> > > > > > > > > t.ya
> > > > > > > > > ml
> > > > > > > > > index aada6957216c..86703e995e31 100644
> > > > > > > > > ---
> > > > > > > > > a/Documentation/devicetree/bindings/timer/sifive,clin
> > > > > > > > > t.ya
> > > > > > > > > ml
> > > > > > > > > +++
> > > > > > > > > b/Documentation/devicetree/bindings/timer/sifive,clin
> > > > > > > > > t.ya
> > > > > > > > > ml
> > > > > > > > > @@ -35,6 +35,7 @@ properties:
> > > > > > > > >            - const: sifive,clint0
> > > > > > > > >        - items:
> > > > > > > > >            - enum:
> > > > > > > > > +              - thead,openc906-clint
> > > > > > > > >                - allwinner,sun20i-d1-clint
> > > > > > > > 
> > > > > > > > Add entries sorted alphabetically. This should be
> > > > > > > > squashed
> > > > > > > > with
> > > > > > > > previous
> > > > > > > > patch.
> > > > > > > 
> > > > > > > I make it a seperated patch because I think it's a
> > > > > > > questionable
> > > > > > > approach.
> > > > > > > 
> > > > > > > If you think it's okay, I will just squash it and put it
> > > > > > > as
> > > > > > > the second
> > > > > > > patch in the next iteration, with adding openc906-plic as
> > > > > > > the
> > > > > > > first
> > > > > > > one.
> > > > > > 
> > > > > > What is a questionable approach? Why commit msg is not
> > > > > > saying
> > > > > > this?
> > > > > 
> > > > > Ah I mentioned it in the cover letter. The problem is just I
> > > > > doubt whether
> > > > > binding strings for single SoCs are necessary.
> > > > 
> > > > They are.
> > > > 
> > > > Unless all the quirks/bugs/features are somehow guaranteed to
> > > > be
> > > > exactly 
> > > > the same as other SoCs sharing the same compatible string, or
> > > > there
> > > > is 
> > > > another mechanism to identify the exact version (e.g. a version
> > > > register).
> > > 
> > > Icenowy,
> > > 
> > > Having thought about this a little - are we not *more* likely to
> > > see
> > > bug/quirk disparity between implementations of the OpenC906 stuff
> > > by
> > > the very nature of being an open-source IP?
> > 
> > It's an open-source edition of a specific version of the commercial
> > IP,
> > a fixed configuration.
> > 
> > In addition, maybe we can just retrieve the version infomation via
> > a T-
> > Head custom CPU configuration register, mcpuid. Despite the
> > implementation of this register is weird -- it contains 7 different
> > read-only values, with the most significant nibble behaving as an
> > index.
> 
> You lot all know the situation here a lot more than I do...
> I don't think "letting" people use the bare "thead,c900-foo" makes
> much
> sense as it gives us no chance to deal with quirks down the line.

Well, after rechecking the manual, I found it possible to handle quirks
-- T-Head has a custom "mcpuid" CSR (@ RISC-V CSR 0xFC0), which can be
used to retrieve some identification info of the core, including its
model ID, version, etc; and the T-Head PLIC/CLINT are part of their
C906 SoC design that there's another "mapbaddr" CSR that could be used
to retrieve the base address of them.

So I think it okay to just use "thead,c900-clint" here, and when
necessary, try to retrieve mcpuid for dealing with quirks.

> I don't think that using "thead,openc906-clint", "thead,c900-clint"
> makes all that much sense either, in case someone does something
> wacky
> with the open-source version of the core.
> 
> That leaves us with either:
> "vendor,soc-clint", "thead,openc906-clint", "thead,c900-clint"
> or:
> "vendor,soc-clint", "thead,c900-clint"
> right?
> 
> The first one seems like possibly the better option as you'd kinda
> expect that, in a perfect word, all of the open-source IP
> implementations would share quirks etc?
> 
> Thanks,
> Conor.
>