This document describes the codec interface between controllers and codecs in Zircon. It is meant to serve as a reference for driver-authors, and to define the interface contract which codec drivers must implement and that controllers can use. The codec interface is a Banjo protocol exposed by codec drivers.
In this document:
<x0,x1,...,xn-1>, for example a vector with two elements 5 and 6 as
<<5,6>,<7,8>>represents a vector with 2 vectors in it.
|Codec||A real or virtual device that encodes/decodes a signal from|
: : digital/analog to/from analog/digital including all : : : combinations, e.g. digital to digital. Example codecs include : : : DAC-Amplifiers combos and ADC converters. : | Controller | The part of a system that manages the audio signals, for | : : example an SOC's audio subsystem or an independent sound card. : | DAI | Digital Audio Interface. Interface between controllers and | : : Codecs. For example an I2S or HDA link. :
The functionality provided by the codecs is divided into:
The controller is responsible for configuring and controlling the codecs. Codecs advertize capabilities and a controller determines how they are used as described below. Note that the codec drivers are expected to perform their own initialization and shutdown, just like any other driver. The controller can control the codec's state, such as through the reset function, but is not required to get codecs to an initialized state.
Codecs are composite devices that provide the codec protocol to controllers. It is expected that only one controller uses a codec's protocol, and one controller may use multiple codecs at once.
Many codec protocol operations are “fire-and-forget”, i.e. they do not expect a reply. Codec protocol operations with a reply are not considered completed until the reply of the function is received, and not considered completed successfully unless the reply contains a status
A codec can be reset by a controller at any time by issuing the
GetInfo function retrieves information from the codec including:
Before specifying the DAI format the controller must query the codec for its bridging capabilites. If the codec is bridgeable, then the controller must enable or disable bridging based on its knowledge of the system configuration. Note that this is a singular property of a codec, i.e. a codec either supports bridging or not, and it can be set in bridged mode or not. This protocol allows configuring as bridged only 2 channel stereo codecs, with the 2 outputs of the codec electrically bridged.
The DAI Format related protocol functions allow the codec to list its supported formats for the DAI. The supported formats may include multiple sample formats, rates, etc. Each codec advertises what it can support and the controller mandates what DAI Format is to be used for each codec.
To find out what formats are supported by a given codec, the controller uses the
GetDaiFormats function. The codec replies with a vector of
DaiSupportedFormats, where each
<2,4,6,8>. A stereo codec reports a vector with one element
<2>. Note that a codec that takes one channel and outputs its contents in all its outputs (e.g. 2 for a stereo amplifier) would report a vector with one element
<1>, if it supports either one or two input channels, it would report a vector with two elements
When not all combinations supported by the codec can be described with one
DaiSupportedFormats, the codec returns more than one
DaiSupportedFormats in the returned vector.
For example, if one
DaiSupportedFormats allows for 32 bits samples at 48KHz, and 16 bits samples at 96KHz, but not 32 bits samples at 96KHz, then the codec will reply with 2
<<16bits>,<96KHz>>. For simplicity, this example ignores parameters other than rate and bits per sample. In the case where the codec supports either 16 or 32 bits samples at either 48 or 96KHz, the codec would reply with 1
Additionally, it is assumed that bits per sample is always smaller or equal to bits per channel. Hence, a codec can report
<<16bits_per_channel,32bits_per_channel>,<16bits_per_sample,32bits_per_sample>> and this does not imply that it is reporting that 32 bits per sample on 16 bits samples is valid, it specifies only the 3 valid combinations:
Using the information provided by the codec in
GetDaiFormat, what is supported by the controller, and any other requirements, the controller specifies the format to use in the DAI with the
SetDaiFormat function. This functions takes a parameter that specifies:
<0,1>, i.e. both left and right channels are used. In bridged mode this will list only the one channel to be used by the codec, for example a codec’s stereo amplifier output bridged into one electrical mono output from the right channel of an I2S DAI would list only channel
<1>. If not bridged, a codec with multiple electrical outputs that is configured with one channel in
SetDaiFormatis expected to replicate the samples in this mono input on all its outputs.
SetDaiFormat is successful, the DAI format configuration is considered completed and samples can be sent across the DAI.
TODO(andresoportus): Add DAI format loss notification support once asynchronous notifications are added to Banjo.
Gain related support by any given codec is returned by the codec in response to a
GetGainFormat function in the
GainFormat structure. The controller can control gain, mute and AGC states in a codec using the
SetGainState function and the corresponding
GetGainState function allows retrieving the current state for the same.
The controller can query the plug detect state with the
GetPlugState function. The plug state includes hardwired and plugged states.
can_notify bool to
PlugState once asynchronous notifications are added to Banjo.