On Thu, Sep 12, 2019 at 2:26 AM Mark Brown broonie@kernel.org wrote:
On Wed, Sep 11, 2019 at 01:22:20PM -0700, Curtis Malainey wrote:
The source of the switch is commit af48f1d08a547 ("ASoC: rt5677: Support DSP function for VAD application") and does not explain the original intent of the switch. I believe the original intent of this commit is to keep the switch in sync with the VAD state. I do not believe we use the switch ourselves.
Well, I would assume that the control is used to allow users to enable and disable the VAD functionality at runtime. As with the routing if it's been exposed to users we should continue to let them control it.
I will work to add variable inputs, in the samus use case it doesn't make much sense to use the hotword without the pcm open since that audio needs to be captured. How would userspace received the detection without the pcm open?