[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAPwEoQMyUMn0Fp9G68Axk3guYB_+DT9f3AkPJiBCXVin=Dck=A@mail.gmail.com>
Date: Mon, 5 Aug 2024 14:04:59 +0800
From: Stanley Chu <stanley.chuys@...il.com>
To: Krzysztof Kozlowski <krzk@...nel.org>
Cc: alexandre.belloni@...tlin.com, robh@...nel.org, krzk+dt@...nel.org,
linux-i3c@...ts.infradead.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, openbmc@...ts.ozlabs.org,
tomer.maimon@...oton.com, kwliu@...oton.com, yschu@...oton.com,
cpchiang1@...oton.com
Subject: Re: [PATCH v1 1/2] dt-bindings: i3c: Add NPCM845 i3c controller
Krzysztof Kozlowski <krzk@...nel.org> 於 2024年8月1日 週四 下午10:53寫道:
>
> On 01/08/2024 09:19, Stanley Chu wrote:
> > The npcm845 i3c devicetree binding follows the basic i3c bindings
> > and add the properties for allowing to adjust the SDA/SCL timing
> > to meet different requirements.
> >
> > Signed-off-by: Stanley Chu <yschu@...oton.com>
> > Signed-off-by: James Chiang <cpchiang1@...oton.com>
> > ---
> > .../bindings/i3c/nuvoton,i3c-master.yaml | 123 ++++++++++++++++++
>
> Use compatible as filename. Anyway word "master" was dropped.
>
> > 1 file changed, 123 insertions(+)
> > create mode 100644 Documentation/devicetree/bindings/i3c/nuvoton,i3c-master.yaml
> >
> > diff --git a/Documentation/devicetree/bindings/i3c/nuvoton,i3c-master.yaml b/Documentation/devicetree/bindings/i3c/nuvoton,i3c-master.yaml
> > new file mode 100644
> > index 000000000000..a40b37b16872
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/i3c/nuvoton,i3c-master.yaml
> > @@ -0,0 +1,123 @@
> > +# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/i3c/nuvoton,i3c-master.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: Nuvoton NPCM845 I3C master
>
> Use new terminology. Since 2021 there was a change... three years ago.
>
Hi Krzysztof,
Thanks for the review.
I will fix it in v2.
> > +
> > +maintainers:
> > + - Stanley Chu <yschu@...oton.com>
> > + - James Chiang <cpchiang1@...oton.com>
> > +
> > +allOf:
> > + - $ref: i3c.yaml#
> > +
> > +properties:
> > + compatible:
> > + const: nuvoton,npcm845-i3c
> > +
> > + reg:
> > + items:
> > + - description: I3C registers
> > + - description: GDMA registers
> > + - description: GDMA request control register
> > +
> > + reg-names:
> > + items:
> > + - const: i3c
> > + - const: dma
> > + - const: dma_ctl
> > +
> > + interrupts:
> > + maxItems: 1
> > +
> > + clocks:
> > + items:
> > + - description: system clock
> > + - description: bus clock
> > +
> > + clock-names:
> > + items:
> > + - const: pclk
> > + - const: fast_clk
> > +
> > + resets:
> > + maxItems: 1
> > +
> > + i3c-pp-scl-hi-period-ns:
> > + description: |
>
> Do not need '|' unless you need to preserve formatting.
>
> > + If need to configure SCL with required duty cycle, specify the clock high/low period directly.
> > + i3c-pp-scl-hi-perios-ns specifies the high period ns of the SCL clock cycle in push pull mode
> > + When i3c-pp-scl-hi-period-ns and i3c-pp-scl-lo-period-ns are specified, the i3c pp frequency is
> > + decided by these two properties.
>
> Wrap according to Linux Coding Style (and read coding style to figure
> the proper wrapping...).
>
I wiill revise the description and properties in v2.
> > +
> > + i3c-pp-scl-lo-period-ns:
> > + description: |
> > + The low period ns of the SCL clock cycle in push pull mode. i3c-pp-scl-lo-period-ns should not
> > + be less than i3c-pp-scl-hi-period-ns and the maximal value is i3c-pp-scl-hi-period-ns + 150.
>
> Everywhere: defaults, constraints.
>
> > +
> > + i3c-pp-sda-rd-skew:
> > + $ref: /schemas/types.yaml#/definitions/uint32
> > + description: |
> > + The number of MCLK clock periods to delay the SDA transition from the SCL clock edge at push
> > + pull operation when transfers i3c private read.
> > + maximum: 7
> > + default: 0
> > +
> > + i3c-pp-sda-wr-skew:
> > + $ref: /schemas/types.yaml#/definitions/uint32
> > + description: |
> > + The number of MCLK clock periods to delay the SDA transition from the SCL clock edge at push
> > + pull operation when transfers i3c private write.
> > + maximum: 7
> > + default: 0
> > +
> > + i3c-od-scl-hi-period-ns:
> > + description: |
> > + The i3c open drain frequency is 1MHz by default.
> > + If need to use different frequency, specify the clock high/low period directly.
> > + i3c-od-scl-hi-perios-ns specifies the high period ns of the SCL clock cycle in open drain mode.
> > + When i3c-od-scl-hi-period-ns and i3c-od-scl-lo-period-ns are specified, the i3c od frequency is
> > + decided by these two properties.
> > + i3c-od-scl-hi-period-ns should be equal to i3c-pp-scl-hi-period-ns or i3c-od-scl-lo-period-ns.
> > +
> > + i3c-od-scl-lo-period-ns:
> > + description: |
> > + The low period ns of the SCL clock cycle in open drain mode. i3c-od-scl-lo-period-ns should be
> > + multiple of i3c-pp-scl-hi-period-ns.
> > +
> > + enable-hj:
> > + type: boolean
> > + description: |
> > + Enable SLVSTART interrupt for receiving hot-join request.
>
> You described the desired Linux feature or behavior, not the actual
> hardware. The bindings are about the latter, so instead you need to
> rephrase the property and its description to match actual hardware
> capabilities/features/configuration etc.
>
> > +
> > + use-dma:
> > + type: boolean
> > + description: |
> > + Enable the i3c private transfers using DMA.
>
> Why wouldn't you enable it always? Where are dma properties for this?
>
I plan to implement a separate dma driver later, this property will be removed.
> > +
> > +required:
> > + - compatible
> > + - reg
> > + - interrupts
> > + - clock-names
> > + - clocks
> > +
> > +additionalProperties: true
>
> Nope, it cannot be true. Look how other bindings are doing this. Why
> implementing something entirely different?
>
> > +
> > +examples:
> > + - |
> > + i3c@...10000 {
> > + compatible = "nuvoton,npcm845-i3c";
> > + clocks = <&clk 4>, <&clk 26>;
> > + clock-names = "pclk", "fast_clk";
> > + interrupts = <0 224 4>;
> > + reg = <0xfff10000 0x1000>,
> > + <0xf0850000 0x1000>,
> > + <0xf0800300 0x4>;
>
> Order properties as in DTS coding style.
>
> > + reg-names = "i3c", "dma", "dma_ctl";
> > + resets = <&rstc 0x74 8>;
> > + #address-cells = <3>;
> > + #size-cells = <0>;
>
> Make your bindings complete - you skipped many properties. Why defining
> them if you are not using them and they do not have defaults?
>
Will fix it in v2.
> > + };
>
> Best regards,
> Krzysztof
>
Powered by blists - more mailing lists