Re: [PATCH v1 5/7] ASoC: dt-bindings: wlf,wm8994: Update #sound-dai-cells property

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] ASoC: dt-bindings: wlf,wm8994: #sound-dai-cells const: 1 breaks existing DTs
--

commit c9447e4a210bccb381d9588b9e0794f2b80cb590
Author: Svyatoslav Ryhel <clamor95@gmail.com>

ASoC: dt-bindings: wlf,wm8994: Update #sound-dai-cells property

This commit updates the #sound-dai-cells property in the Wolfson
Microelectronics WM8994 audio codec binding from const: 0 to const: 1.

> diff --git a/Documentation/devicetree/bindings/sound/wlf,wm8994.yaml b/Documentation/devicetree/bindings/sound/wlf,wm8994.yaml
> index 0db04a90ac6b3..39bf1af7c3ab8 100644
> --- a/Documentation/devicetree/bindings/sound/wlf,wm8994.yaml
> +++ b/Documentation/devicetree/bindings/sound/wlf,wm8994.yaml
> @@ -69,7 +69,7 @@ properties:
>    SPKVDD2-supply: true
>  
>    '#sound-dai-cells':
> -    const: 0
> +    const: 1

[Severity: Medium]
Does replacing const: 0 with const: 1 break dtbs_check validation and
backwards compatibility for existing in-tree device trees?

All existing in-tree device trees using wlf,wm1811 or wlf,wm8994 specify
#sound-dai-cells = <0> and reference the codec with a 0-cell sound-dai
phandle:

arch/arm/boot/dts/samsung/exynos4212-tab3.dtsi:
    wm1811: audio-codec@1a {
        ...
        gpio-controller;
        #gpio-cells = <2>;
        #sound-dai-cells = <0>;

arch/arm/boot/dts/samsung/exynos4412-midas.dtsi:
    wm1811: audio-codec@1a {
        ...
        gpio-controller;
        #gpio-cells = <2>;
        #sound-dai-cells = <0>;

arch/arm/boot/dts/samsung/s5pv210-aries.dtsi:
    wm8994: audio-codec@1a {
        compatible = "wlf,wm8994";
        reg = <0x1a>;

        #sound-dai-cells = <0>;

The corresponding sound card bindings
(Documentation/devicetree/bindings/sound/samsung,midas-audio.yaml and
Documentation/devicetree/bindings/sound/samsung,aries-wm8994.yaml) and
machine drivers (sound/soc/samsung/midas_wm1811.c and
sound/soc/samsung/aries_wm8994.c) parse a bare codec phandle via
of_parse_phandle(codec, "sound-dai", 0) and hardcode the three codec DAIs.
Meanwhile, sound/soc/codecs/wm8994.c registers three DAIs without a custom
.of_xlate_dai_name callback, so snd_soc_get_dlc() in sound/soc/soc-core.c
supports both 0 cells and 1 cell.

Should #sound-dai-cells in
Documentation/devicetree/bindings/sound/wlf,wm8994.yaml use enum: [0, 1]
instead so that both existing 0-cell device trees and 1-cell consumers remain
valid?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261027120019.93611-1-clamor95@gmail.com?part=5