[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJhJPsXOX16SA0bb8zWJ=wSxshjv38g038cR0b3u0CDDm1aUQw@mail.gmail.com>
Date: Wed, 18 Jan 2023 19:16:59 +0800
From: Kelvin Cheung <keguang.zhang@...il.com>
To: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
Cc: linux-clk@...r.kernel.org, devicetree@...r.kernel.org,
linux-mips@...r.kernel.org, linux-kernel@...r.kernel.org,
Michael Turquette <mturquette@...libre.com>,
Stephen Boyd <sboyd@...nel.org>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>
Subject: Re: [PATCH 1/2] dt-bindings: clock: Add binding for Loongson-1 clock driver
Hi Krzysztof,
Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org> 于2023年1月17日周二 18:47写道:
>
> On 17/01/2023 11:31, Kelvin Cheung wrote:
> >>> + "#clock-cells":
> >>> + const: 0
> >>> +
> >>> + compatible:
> >>> + enum:
> >>> + - loongson,ls1b-clk-pll
> >>> + - loongson,ls1b-clk-cpu
> >>> + - loongson,ls1b-clk-ahb
> >>> + - loongson,ls1c-clk-pll
> >>> + - loongson,ls1c-clk-cpu
> >>> + - loongson,ls1c-clk-ahb
> >>
> >> Are you registering single clocks? It looks like. No, make a proper
> >> clock controller.
> >
> > This binding contains two types of clock, pll-clk and div-clk.
> > Should I split the binding to two bindings files?
>
> No, you should register rather one clock controller. Why this have to be
> 3 separate clock controllers?
>
This sounds like a big change for the driver.
Could you please show me a good example of one clock controller?
Thanks very much!
> >>
> >>> +
> >>> + reg:
> >>> + maxItems: 1
> >>> +
> >>> + clocks:
> >>> + maxItems: 1
> >>> +
> >>> +required:
> >>> + - "#clock-cells"
> >>> + - compatible
> >>> + - clocks
> >>> + - reg
> >>> +
> >>> +additionalProperties: false
> >>> +
> >>> +examples:
> >>> + - |
> >>> + clocks {
> >>
> >> No, not really related to the binding.
> >
> > Should I remove the "clocks" section?
>
> Yes.
>
> >>
> >>> + #address-cells = <1>;
> >>> + #size-cells = <1>;
> >>> + ranges;
> >>> +
> >>> + xtal: xtal {
> >>
> >> Incorrect in this context. Missing unit address.
> >
> > XTAL doesn't have reg property.
>
> Yeah, but DTS is not correct now, is it? If you doubt, build your DTB
> with W=1.
>
No doubt.
I just want to know the right way to declare XTAL.
Could you please show me an example?
Thanks again!
> >>
> >>> + compatible = "fixed-clock";
> >>> + #clock-cells = <0>;
> >>> + clock-frequency = <33000000>;
> >>> + };
> >>> +
>
> Best regards,
> Krzysztof
>
--
Best regards,
Kelvin Cheung
Powered by blists - more mailing lists