Re: [PATCH v1 1/2] dt-bindings: sound: add qcom,wsa885x-i2c

[Prasad Kumpatla]

On 6/11/2026 2:50 AM, Linus Walleij wrote:
> Hi Prasad,
>
> thanks for your patch!

Hi Linus Walleij,

Thanks for reviewing the patch and for the feedback.

> On Wed, Jun 10, 2026 at 5:57 PM Prasad Kumpatla
> <prasad.kumpatla@xxxxxxxxxxxxxxxx> wrote:
>
> > Document the Qualcomm WSA885X I2C smart amplifier binding.
> Skip I2C? We don't need to tell e.g. "PCI" in some device on PCI and
> there is no reason to mention I2C for this device, the fact that it sits
> on an I2C bus will be apparent later.

Ack, Will rename the file name.

> > Describe the required supplies, powerdown and interrupt GPIOs, the
> > optional battery configuration, and the optional init-table property
> > used to program the device during codec initialization.
> >
> > This matches the driver programming model and documents the DT data
> > needed to use the codec on platforms with Audio IF playback.
> >
> > Signed-off-by: Prasad Kumpatla <prasad.kumpatla@xxxxxxxxxxxxxxxx>
> Perhaps add
> Link: https://www.qualcomm.com/audio/applications/compute-and-mobile-audio/products/wsa8815
>
> (...)
Checking internal to get product Doc to publish.
> > ---
> >   .../bindings/sound/qcom,wsa885x-i2c.yaml      | 89 +++++++++++++++++++
> >   1 file changed, 89 insertions(+)
> >   create mode 100644 Documentation/devicetree/bindings/sound/qcom,wsa885x-i2c.yaml
> >
> > diff --git a/Documentation/devicetree/bindings/sound/qcom,wsa885x-i2c.yaml b/Documentation/devicetree/bindings/sound/qcom,wsa885x-i2c.yaml
> Drop the -i2c suffix on the files.
Ack, Will update.
> > new file mode 100644
> > index 000000000..1069f470d
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/sound/qcom,wsa885x-i2c.yaml
> > @@ -0,0 +1,89 @@
> > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/sound/qcom,wsa885x-i2c.yaml#
> Drop the -i2c suffix.
Ack, Will update.
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: Qualcomm WSA885x I2C smart speaker amplifier
> Drop I2C.
Ack, Will update.
> > +maintainers:
> > +  - Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxxxxxxxx>
> > +  - Prasad Kumpatla <prasad.kumpatla@xxxxxxxxxxxxxxxx>
> > +
> > +description: |
> > +  WSA885x is a Qualcomm Aqstic smart speaker amplifier with an I2C control
> > +  interface and a digital audio interface exposed through ASoC DAI callbacks.
> > +
> > +allOf:
> > +  - $ref: dai-common.yaml#
> > +
> > +properties:
> > +  compatible:
> > +    const: qcom,wsa885x-i2c
> Drop -i2c
Ack, Will update.
> > +  reg:
> > +    maxItems: 1
> > +
> > +  '#sound-dai-cells':
> > +    const: 0
> > +
> > +  powerdown-gpios:
> > +    description: GPIO controlling the SD_N powerdown pin.
> > +    maxItems: 1
> > +
> > +  interrupt-gpios:
> > +    description: GPIO used for the codec interrupt output.
> > +    maxItems: 1
> > +
> > +  vdd-1p8-supply: true
> > +
> > +  vdd-io-supply: true
> > +
> > +  qcom,battery-config:
> > +    $ref: /schemas/types.yaml#/definitions/uint32
> > +    description: Speaker battery configuration, 1 for 1S and 2 for 2S.
> What is a "1S" and a "2S"? Include description here.
Ack, Will Update more details for 1s and 2s in next version.
> > +    default: 1
> > +    enum: [1, 2]
> > +
> > +  qcom,wsa885x-init-table:
> > +    $ref: /schemas/types.yaml#/definitions/uint32-array
> > +    minItems: 2
> > +    maxItems: 256
> > +    description: |
> > +      Sequence of register/value pairs applied during codec hardware
> > +      initialization. Entries are encoded as alternating register address and
> > +      register value cells. The number of entries must be even (register/value
> > +      pairs); maxItems is 256 (128 pairs).
> Can this just be a table inside the driver, if it will be the same
> array for every user? If this is board-unique then it needs to describe
> what each value is actually doing well enough so engineers can use this
> documentation right here to configure their board without looking through
> register maps and what not.
>
> Something more abstract using SI units etc is probably needed here.

These values are part of a fixed hardware initialization sequence and 
are not board-specific.

I'll move them into the driver as a register table, making them easier 
to maintain and

avoiding opaque DT data.

Thanks,

Prasad

> Yours,
> Linus Walleij