[PATCH v7 2/3] ASoC: qcom: dt-bindings: Add sc7180 machine bindings

[Cheng-yi Chiang]

On Thu, Sep 10, 2020 at 1:49 AM Rob Herring <robh at kernel.org> wrote:
>
> On Wed, Sep 9, 2020 at 3:24 AM Cheng-yi Chiang <cychiang at chromium.org> wrote:
> >
> > On Wed, Sep 9, 2020 at 4:34 AM Rob Herring <robh at kernel.org> wrote:
> > >
> > > On Mon, Sep 07, 2020 at 06:00:38PM +0800, Cheng-Yi Chiang wrote:
> > > > Add devicetree bindings documentation file for sc7180 sound card.
> > > >
> > > > Signed-off-by: Cheng-Yi Chiang <cychiang at chromium.org>
> > > > ---
> > > >  .../bindings/sound/qcom,sc7180.yaml           | 143 ++++++++++++++++++
> > > >  1 file changed, 143 insertions(+)
> > > >  create mode 100644 Documentation/devicetree/bindings/sound/qcom,sc7180.yaml
> > > >
> > > > diff --git a/Documentation/devicetree/bindings/sound/qcom,sc7180.yaml b/Documentation/devicetree/bindings/sound/qcom,sc7180.yaml
> > > > new file mode 100644
> > > > index 000000000000..ae809346ca80
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/sound/qcom,sc7180.yaml
> > > > @@ -0,0 +1,143 @@
> > > > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > > > +%YAML 1.2
> > > > +---
> > > > +$id: http://devicetree.org/schemas/sound/qcom,sc7180.yaml#
> > > > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > > > +
> > > > +title: Qualcomm Technologies Inc. SC7180 ASoC sound card driver
> > > > +
> > > > +maintainers:
> > > > +  - Rohit kumar <rohitkr at codeaurora.org>
> > > > +  - Cheng-Yi Chiang <cychiang at chromium.org>
> > > > +
> > > > +description:
> > > > +  This binding describes the SC7180 sound card which uses LPASS for audio.
> > > > +
> > > > +properties:
> > > > +  compatible:
> > > > +    const: qcom,sc7180-sndcard
> > > > +
> > > > +  audio-routing:
> > > > +    $ref: /schemas/types.yaml#/definitions/non-unique-string-array
> > > > +    description:
> > > > +      A list of the connections between audio components. Each entry is a
> > > > +      pair of strings, the first being the connection's sink, the second
> > > > +      being the connection's source.
> > > > +
> > > > +  model:
> > > > +    $ref: /schemas/types.yaml#/definitions/string
> > > > +    description: User specified audio sound card name
> > > > +
> > > > +  headset-jack:
> > > > +    $ref: /schemas/types.yaml#/definitions/phandle
> > > > +    description: phandle of the codec for headset detection
> > > > +
> > > > +  hdmi-jack:
> > > > +    $ref: /schemas/types.yaml#/definitions/phandle
> > > > +    description: phandle of the codec for hdmi jack detection
> > >
> > > You already have links to these devices. Why duplicate it here?
> > >
> > > What if you had 2 headsets? This doesn't scale.
> > >
> > Hi Rob, thanks for reviewing.
> > There was some discussion in
> > https://patchwork.kernel.org/patch/11737905/#23571643 about how to
> > specify the dailink that has a headset jack.
> > I would like to pass the information of headset jack and hdmi jack to
> > the machine driver so the machine driver can call
> > snd_soc_component_set_jack to set jack when init the corresponding link.
> > Headset jack and hdmi jack will be treated differently for button and
> > event type.
> > Because of this, we can not just set a property "jack" in the link.
>
> Don't design your binding around some driver architecture. Limitations
> of ASoC are not reasons for your binding.
>
> For DP and HDMI, we assume HPD is supported generally as that's a
> standard function for HDMI/DP controllers. There's a 'no-hpd' property
> for cases of broken HPD. That hardware description is part of the
> device HPD is connected to which is the HDMI/DP controller/bridge or
> the connector node in the case of a GPIO line. That doesn't belong in
> the virtual sound card.
>
> I would assume a codec to be similar. The codec node (the alc5682
> node) should have any jack related properties (or possibly implicitly
> support it by default).


Thanks for the explanation. Now I can follow your thoughts.
I see that the sound card driver should assume that there is jack/hpd
available on HDMI/codec unless it is specified in HDMI/codec node
(should not specified in sound card).

>
>
> > As for the 2 headsets case (I guess you mean hp jack and mic jack), on
> > this board we will not have this use case.
> > If someone really wants to build hp jack and mic jack on the board
> > based on this machine driver, we can add two more property hp-jack and
> > mic-jack to specify that,
> > as the machine driver will need to know the different jack types
> > anyway. What do you think ?
>
> I was thinking more of a case of having 2 of the same thing. Perhaps 2
> HDMI outputs. Or if you have DP instead of HDMI? Just going to add
> 'dp-jack'?


I agree that this may not scale. It will become an abuse of binding interface.
Sound card drivers should have more knowledge/control on this kind of
configuration.
I think this can be taken care of inside the sound card driver.
If a machine wants a different configuration, we can use compatible
strings to achieve that.
This will be more flexible than the limited binding interface.

Thanks for the discussion, I will remove headset-jack and hdmi-jack in v8.
>
>
> Rob