Re: [PATCH v3 4/5] arm64: dts: arduino-imola: add support for sound

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] ASoC: redundant BE DAI links for identical q6afe CPU DAI
--

commit bd3dd7751e736b11cf0fd37d9d4991d90e6ee507
Author: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com>

arm64: dts: arduino-imola: add support for sound

Add support for sound on Arduino UNO Q board, which includes headset
playback and record, and lineout.

> diff --git a/arch/arm64/boot/dts/qcom/qrb2210-arduino-imola.dts b/arch/arm64/boot/dts/qcom/qrb2210-arduino-imola.dts
> index c472e13fceca4..e99d2a56dc287 100644
> --- a/arch/arm64/boot/dts/qcom/qrb2210-arduino-imola.dts
> +++ b/arch/arm64/boot/dts/qcom/qrb2210-arduino-imola.dts
[ ... ]
> +		ear-playback-dai-link {
> +			link-name = "Ear Out Playback";
> +
> +			codec {
> +				sound-dai = <&pmic4125_codec 0>, <&swr1 0>, <&rxmacro 0>;
> +			};
> +
> +			cpu {
> +				sound-dai = <&q6afedai RX_CODEC_DMA_RX_0>;
> +			};
[ ... ]
> +		hph-playback-dai-link {
> +			link-name = "Headphones Playback";
> +
> +			codec {
> +				sound-dai = <&pmic4125_codec 0>, <&swr1 0>, <&rxmacro 0>;
> +			};
> +
> +			cpu {
> +				sound-dai = <&q6afedai RX_CODEC_DMA_RX_0>;
> +			};
[ ... ]
> +		lo-playback-dai-link {
> +			link-name = "Line Out Playback";
> +
> +			codec {
> +				sound-dai = <&pmic4125_codec 0>, <&swr1 0>, <&rxmacro 0>;
> +			};
> +
> +			cpu {
> +				sound-dai = <&q6afedai RX_CODEC_DMA_RX_0>;
> +			};

[Severity: High]
Does defining multiple identical Backend (BE) DAI links cause conflicting
DSP driver operations?

This link, as well as ear-playback-dai-link and hph-playback-dai-link
defined above, map the exact same physical CPU DAI (RX_CODEC_DMA_RX_0) to the
exact same Codec DAIs.

When the ALSA DPCM core registers the card, it instantiates duplicate BE
runtimes. During audio playback, if the DPCM core triggers multiple of these
BEs concurrently or sequentially, could this interrupt active audio and
corrupt the shared hardware state? For instance, q6afe_dai_prepare() in
sound/soc/qcom/qdsp6/q6afe-dai.c checks is_port_started[dai->id]; if two
runtimes call prepare on the same DAI, the driver might abruptly stop and
restart the active DSP port.

Should logical branching to different endpoints (Ear, Headphones, Line Out)
be handled internally by the Codec driver's DAPM routing graph instead of
defining redundant BE bus links?

> +
> +			platform {
> +				sound-dai = <&q6routing>;
> +			};
> +		};

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/cover.1790114902.git.srinivas.kandagatla@oss.qualcomm.com?part=4