Message-ID: <37181fc1-a485-427c-b3d2-2fcfb192e6e9@collabora.com>
Date: Thu, 29 Feb 2024 10:10:51 +0100
From: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
To: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>, broonie@...nel.org
Cc: wenst@...omium.org, lgirdwood@...il.com, robh@...nel.org,
 krzysztof.kozlowski+dt@...aro.org, conor+dt@...nel.org,
 matthias.bgg@...il.com, perex@...ex.cz, tiwai@...e.com,
 trevor.wu@...iatek.com, maso.huang@...iatek.com,
 xiazhengqiao@...qin.corp-partner.google.com, arnd@...db.de,
 kuninori.morimoto.gx@...esas.com, shraash@...gle.com, amergnat@...libre.com,
 nicolas.ferre@...rochip.com, u.kleine-koenig@...gutronix.de,
 dianders@...omium.org, frank.li@...o.com, allen-kh.cheng@...iatek.com,
 eugen.hristev@...labora.com, claudiu.beznea@...on.dev,
 jarkko.nikula@...mer.com, jiaxin.yu@...iatek.com, alpernebiyasak@...il.com,
 ckeepax@...nsource.cirrus.com, zhourui@...qin.corp-partner.google.com,
 nfraprado@...labora.com, alsa-devel@...a-project.org,
 shane.chien@...iatek.com, linux-sound@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 linux-arm-kernel@...ts.infradead.org, linux-mediatek@...ts.infradead.org,
 kernel@...labora.com
Subject: Re: [PATCH 18/22] ASoC: dt-bindings: mt8195: Document audio-routing
 and dai-link subnode

Il 29/02/24 09:25, Krzysztof Kozlowski ha scritto:
> On 27/02/2024 13:09, AngeloGioacchino Del Regno wrote:
>> Document the dai-link subnodes and the audio-routing property, allowing
>> to describe machine specific audio hardware and links in device tree.
>>
>> While at it, also deprecate the old properties which were previously
>> used with driver hardcoded configuration.
>>
>> Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
>> ---
> 
> 
>>   
>>     mediatek,adsp:
>>       $ref: /schemas/types.yaml#/definitions/phandle
>> @@ -45,12 +56,75 @@ properties:
>>         A list of the desired dai-links in the sound card. Each entry is a
>>         name defined in the machine driver.
>>   
>> +patternProperties:
>> +  ".*-dai-link$":
>> +    type: object
>> +    description:
>> +      Container for dai-link level properties and CODEC sub-nodes.
>> +
>> +    properties:
>> +      link-name:
>> +        description: Indicates dai-link name and PCM stream name
>> +        items:
> 
> That's not a list, but just enum.
> 

Ok! Will change for v2.

>> +          enum:
>> +            - DPTX_BE
>> +            - ETDM1_IN_BE
>> +            - ETDM2_IN_BE
>> +            - ETDM1_OUT_BE
>> +            - ETDM2_OUT_BE
>> +            - ETDM3_OUT_BE
>> +            - PCM1_BE
>> +
>> +      codec:
>> +        description: Holds subnode which indicates codec dai.
>> +        type: object
>> +        additionalProperties: false
>> +        properties:
>> +          sound-dai:
>> +            minItems: 1
>> +            maxItems: 2
>> +        required:
>> +          - sound-dai
>> +
>> +      dai-format:
>> +        description: audio format
>> +        items:
> 
> Ditto
> 
>> +          enum:
>> +            - i2s
>> +            - right_j
>> +            - left_j
>> +            - dsp_a
>> +            - dsp_b
>> +
>> +      mediatek,clk-provider:
>> +        $ref: /schemas/types.yaml#/definitions/string
>> +        description: Indicates dai-link clock master.
>> +        items:
> 
> Ditto
> 
>> +          enum:
>> +            - cpu
>> +            - codec
>> +
>> +    additionalProperties: false
> 
> This goes either to the top of the section (after type:object) for
> readability or after required: block below.
> 

I think I actually saw this in some other binding, that's why it's there... or
maybe I am confusing something else - whatever.

I'll move it, anyway - thanks

>> +
>> +    required:
>> +      - link-name
>> +
>>   additionalProperties: false
> 
>>   
>>   required:
>>     - compatible
>>     - mediatek,platform
>>   
>> +# Disallow legacy properties if dai-link-xxx nodes are specified
>> +if:
>> +  not:
> 
> I don't think this works. To test if node is present or node, you would
> need to use required.
> https://elixir.bootlin.com/linux/v6.4-rc7/source/Documentation/devicetree/bindings/net/qcom,ipa.yaml#L174
> 
> Are you sure this if:then: works as expected?
> 

Honestly, coming up with something that worked took me almost a full hour.
I was going a bit crazy over this one :-)

btw - Yes, it does work.

If you wanna give it a shot too, patch 21 and 22 are introducing the audio-routing
and (x)-dai-link nodes in (respectively) mt8195-cherry.dtsi and mt8186-corsola.dtsi
devicetrees - and those do validate just fine both with and without those patches.

I also tested this by adding the forbidden dptx-codec/hdmi-codec properties along
with the dai-link nodes in the example of this yaml file, and I've verified that
adding those makes the validation fail as expected.

Exactly, this is what I want and also what I've tested with dtbs/dt_binding_check:

fmt: (test nodes/properties) <- (result)

x-dai-link { ... }; mediatek,dptx-codec = <...>; <- FAIL (OK)
x-dai-link { ... }; mediatek,hdmi-codec = <...>; <- FAIL (OK)
x-dai-link { ... }; mediatek,dptx-codec = ..; mediatek,hdmi-codec .. <- FAIL (OK)

x-dai-link { ... }; (none of those two) <- PASS (OK)

(NO x-dai-link) mediatek,dptx-codec = <...> <- PASS (OK)
(NO x-dai-link) mediatek,hdmi-codec = <...> <- PASS (OK)
(NO x-dai-link) mediatek,dptx-codec = <...> mediatek,hdmi-codec = <...> <- PASS(OK)

..and of course, also tested the easy part
dai-link-x { ... }; <- FAIL (OK)    :-)

Cheers,
Angelo

> 
>> +    patternProperties:
>> +      ".*-dai-link$": false
>> +then:
>> +  properties:
>> +    mediatek,dptx-codec: false
>> +    mediatek,hdmi-codec: false
>> +
>>   examples:
>>     - |
>>   
> 
> 
> Best regards,
> Krzysztof
> 

Powered by blists - more mailing lists