implement a parallel interface for esp32 using the i2s peripheral (#2348)

* i2s_parallel for esp32 - partially working

* i2s_parallel for esp32 - works sync 16bit (still WIP)

* i2s_parallel - add 8-bit

* i2s_parallel - naive async implementation

* i2s_parallel - clippy

* add examples (mainly for testing now and we can reduce these before merge)

* set tx_wrx2 in 8 bit mode or it updates on half clocks!

* adjust clock config (still really a hack)

* better clock calculation/configuration

* no need to reset dma or interrupts in start, the DMA DmaTxBuf handles that

* seems to need a short delay during tx_reset

* use 20 clocks instead of 1us

* 1us delay between dma start and tx_start (like idf)

* changelog & documentation

* remove debugging functions

* no need for option in example

* make async example actually use async & remove Option

* implement Drop for I2sParallelTransfer

* fix signal offset comment for I2S0

* small cleanup in i2s_parallel examples

* added a note re I2S0 not supporting true 8bit

* update doc & fmt

* instace functions take &self

* fmt

* if run at 240MHz (CPU) delay needs to be bigger

* wait for fifo on start and fixes for 240MHz

* update for esp-hal-changes

* fmt

---------

Co-authored-by: Dániel Buga <bugadani@gmail.com>
This commit is contained in:
liebman 2024-10-30 11:55:47 -07:00 committed by GitHub
parent 9d4b8fdbc6
commit 448ed9a001
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
5 changed files with 895 additions and 0 deletions

View File

@ -21,6 +21,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
- `Spi::half_duplex_read` and `Spi::half_duplex_write` (#2373)
- `Cpu::COUNT` and `Cpu::current()` (#2411)
- `UartInterrupt` and related functions (#2406)
- I2S Parallel output driver for esp32. (#2348)
- Add an option to configure `WDT` action (#2330)
### Changed

714
esp-hal/src/i2s_parallel.rs Normal file
View File

@ -0,0 +1,714 @@
//! # Parallel Interface (via I2S)
//!
//! ## Overview
//! The I2S parallel interface allows for high-speed data transfer between the
//! ESP32 and external devices. It is commonly used to external devices such as
//! LED matrix, LCD display, and Printer. Only TX is implemented. Each
//! unit can have up to 8 or 16 data signals (depending on your target hardware)
//! plus 1 clock signal.
//!
//! ## Notes
//!
//! Data output is interleaved:
//! - 8bit: [A, B, C, D] is output as [C, D, A, B] (i.e., swapped as 16bit
//! words)
//! - 16bit: [A, B, C, D] is output as [B, A, D, C] (i.e., 16bit words are
//! swapped)
#![cfg_attr(esp32, doc = "")]
#![cfg_attr(
esp32,
doc = "I2S0 does not support true 8bit parallel output, so if you want to do 8bit"
)]
#![cfg_attr(
esp32,
doc = "you should use I2S1. If you have to use I2S0, it will only output the even"
)]
#![cfg_attr(esp32, doc = "bytes! so [A, B, C, D] will be output as [A, C]!!!!")]
#![cfg_attr(esp32, doc = "")]
//! ## Configuration
//!
//! The driver uses DMA (Direct Memory Access) for efficient data transfer and
//! supports various configurations, such as different data formats, standards
//! (e.g., Philips) and pin configurations. It relies on other peripheral
//! modules, such as
//! - `GPIO`
//! - `DMA`
//! - `system` (to configure and enable the I2S peripheral)
use core::{
marker::PhantomData,
mem::ManuallyDrop,
ops::{Deref, DerefMut},
};
use fugit::HertzU32;
use private::{calculate_clock, I2sClockDividers, TxPins};
use crate::{
dma::{
asynch::DmaTxFuture,
Channel,
ChannelTx,
DmaChannelConvert,
DmaEligible,
DmaError,
DmaTxBuffer,
PeripheralMarker,
Tx,
},
gpio::{
interconnect::{OutputConnection, PeripheralOutput},
OutputSignal,
},
peripheral::{Peripheral, PeripheralRef},
peripherals::{i2s0::RegisterBlock, I2S0, I2S1},
system::PeripheralClockControl,
Mode,
};
/// Represents a group of 16 output pins configured for 16-bit parallel data
/// transmission.
pub struct TxSixteenBits<'d> {
pins: [PeripheralRef<'d, OutputConnection>; 16],
}
impl<'d> TxSixteenBits<'d> {
#[allow(clippy::too_many_arguments)]
/// Creates a new `TxSixteenBits` instance with the provided output pins.
pub fn new(
pin_0: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_1: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_2: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_3: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_4: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_5: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_6: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_7: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_8: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_9: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_10: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_11: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_12: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_13: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_14: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_15: impl Peripheral<P = impl PeripheralOutput> + 'd,
) -> Self {
crate::into_mapped_ref!(
pin_0, pin_1, pin_2, pin_3, pin_4, pin_5, pin_6, pin_7, pin_8, pin_9, pin_10, pin_11,
pin_12, pin_13, pin_14, pin_15
);
Self {
pins: [
pin_0, pin_1, pin_2, pin_3, pin_4, pin_5, pin_6, pin_7, pin_8, pin_9, pin_10,
pin_11, pin_12, pin_13, pin_14, pin_15,
],
}
}
}
impl<'d> TxPins for TxSixteenBits<'d> {
fn bits(&self) -> u8 {
16
}
fn configure<I: Instance>(&mut self, instance: &PeripheralRef<'_, I>) {
use crate::private::Internal;
let bits: u8 = self.bits();
for (i, pin) in self.pins.iter_mut().enumerate() {
pin.set_to_push_pull_output(Internal);
pin.connect_peripheral_to_output(instance.data_out_signal(i, bits), Internal);
}
}
}
/// Represents a group of 8 output pins configured for 8-bit parallel data
/// transmission.
pub struct TxEightBits<'d> {
pins: [PeripheralRef<'d, OutputConnection>; 8],
}
impl<'d> TxEightBits<'d> {
#[allow(clippy::too_many_arguments)]
/// Creates a new `TxSEightBits` instance with the provided output pins.
pub fn new(
pin_0: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_1: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_2: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_3: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_4: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_5: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_6: impl Peripheral<P = impl PeripheralOutput> + 'd,
pin_7: impl Peripheral<P = impl PeripheralOutput> + 'd,
) -> Self {
crate::into_mapped_ref!(pin_0);
crate::into_mapped_ref!(pin_1);
crate::into_mapped_ref!(pin_2);
crate::into_mapped_ref!(pin_3);
crate::into_mapped_ref!(pin_4);
crate::into_mapped_ref!(pin_5);
crate::into_mapped_ref!(pin_6);
crate::into_mapped_ref!(pin_7);
Self {
pins: [pin_0, pin_1, pin_2, pin_3, pin_4, pin_5, pin_6, pin_7],
}
}
}
impl<'d> TxPins for TxEightBits<'d> {
fn bits(&self) -> u8 {
8
}
fn configure<I: Instance>(&mut self, instance: &PeripheralRef<'_, I>) {
use crate::private::Internal;
let bits: u8 = self.bits();
for (i, pin) in self.pins.iter_mut().enumerate() {
pin.set_to_push_pull_output(Internal);
pin.connect_peripheral_to_output(instance.data_out_signal(i, bits), Internal);
}
}
}
/// I2S Parallel Interface
pub struct I2sParallel<'d, I: Instance, DM: Mode> {
instance: PeripheralRef<'d, I>,
tx_channel: ChannelTx<'d, <I as DmaEligible>::Dma>,
mode: PhantomData<DM>,
}
impl<'d, I: Instance, DM: Mode> I2sParallel<'d, I, DM> {
/// Create a new I2S Parallel Interface
pub fn new<P, CH, CLK: PeripheralOutput>(
i2s: impl Peripheral<P = I> + 'd,
channel: Channel<'d, CH, DM>,
frequency: impl Into<fugit::HertzU32>,
mut pins: P,
clock_pin: impl Peripheral<P = CLK> + 'd,
) -> Self
where
CH: DmaChannelConvert<I::Dma>,
P: TxPins,
{
crate::into_ref!(i2s);
crate::into_mapped_ref!(clock_pin);
channel.runtime_ensure_compatible(&i2s);
let channel = channel.degrade();
PeripheralClockControl::reset(i2s.get_peripheral());
PeripheralClockControl::enable(i2s.get_peripheral());
// configure the I2S peripheral for parallel mode
i2s.setup(frequency, pins.bits());
// setup the clock pin
clock_pin.set_to_push_pull_output(crate::private::Internal);
clock_pin.connect_peripheral_to_output(i2s.ws_signal(), crate::private::Internal);
pins.configure(&i2s);
Self {
instance: i2s,
tx_channel: channel.tx,
mode: PhantomData,
}
}
/// Write data to the I2S peripheral
pub fn send<BUF: DmaTxBuffer>(
mut self,
mut data: BUF,
) -> Result<I2sParallelTransfer<'d, I, BUF, DM>, (DmaError, Self, BUF)> {
self.instance.tx_reset();
self.instance.tx_fifo_reset();
let result = unsafe {
self.tx_channel
.prepare_transfer(self.instance.get_dma_peripheral(), &mut data)
}
.and_then(|_| self.tx_channel.start_transfer());
if let Err(err) = result {
return Err((err, self, data));
}
crate::rom::ets_delay_us(1);
self.instance.tx_start();
Ok(I2sParallelTransfer {
i2s: ManuallyDrop::new(self),
buf_view: ManuallyDrop::new(data.into_view()),
})
}
}
/// Represents an ongoing (or potentially finished) transfer using the i2s
/// parallel interface
pub struct I2sParallelTransfer<'d, I, BUF, DM>
where
I: Instance,
BUF: DmaTxBuffer,
DM: Mode,
{
i2s: ManuallyDrop<I2sParallel<'d, I, DM>>,
buf_view: ManuallyDrop<BUF::View>,
}
impl<'d, I, BUF, DM> I2sParallelTransfer<'d, I, BUF, DM>
where
I: Instance,
BUF: DmaTxBuffer,
DM: Mode,
{
/// Returns true when [Self::wait] will not block.
pub fn is_done(&self) -> bool {
self.i2s.instance.is_tx_done()
}
/// Wait for the transfer to finish
pub fn wait(mut self) -> (I2sParallel<'d, I, DM>, BUF) {
self.i2s.instance.tx_wait_done();
let i2s = unsafe { ManuallyDrop::take(&mut self.i2s) };
let view = unsafe { ManuallyDrop::take(&mut self.buf_view) };
(i2s, BUF::from_view(view))
}
fn stop_peripherals(&mut self) {
self.i2s.instance.tx_stop();
self.i2s.tx_channel.stop_transfer();
}
}
impl<'d, I, BUF> I2sParallelTransfer<'d, I, BUF, crate::Async>
where
I: Instance,
BUF: DmaTxBuffer,
{
/// Wait for the transfer to finish
pub async fn wait_for_done(&mut self) -> Result<(), DmaError> {
DmaTxFuture::new(&mut self.i2s.tx_channel).await
}
}
impl<'d, I, BUF, DM> Deref for I2sParallelTransfer<'d, I, BUF, DM>
where
I: Instance,
BUF: DmaTxBuffer,
DM: Mode,
{
type Target = BUF::View;
fn deref(&self) -> &Self::Target {
&self.buf_view
}
}
impl<'d, I, BUF, DM> DerefMut for I2sParallelTransfer<'d, I, BUF, DM>
where
I: Instance,
BUF: DmaTxBuffer,
DM: Mode,
{
fn deref_mut(&mut self) -> &mut Self::Target {
&mut self.buf_view
}
}
impl<'d, I, BUF, DM> Drop for I2sParallelTransfer<'d, I, BUF, DM>
where
I: Instance,
BUF: DmaTxBuffer,
DM: Mode,
{
fn drop(&mut self) {
self.stop_peripherals();
// SAFETY: This is Drop, we know that self.i2s and self.buf_view
// won't be touched again.
let view = unsafe {
ManuallyDrop::drop(&mut self.i2s);
ManuallyDrop::take(&mut self.buf_view)
};
let _ = BUF::from_view(view);
}
}
mod private {
use fugit::HertzU32;
use crate::peripheral::PeripheralRef;
pub trait TxPins {
fn bits(&self) -> u8;
fn configure<I: super::Instance>(&mut self, instance: &PeripheralRef<'_, I>);
}
#[derive(Debug)]
pub struct I2sClockDividers {
pub mclk_divider: u32,
pub bclk_divider: u32,
pub denominator: u32,
pub numerator: u32,
}
#[cfg(any(esp32, esp32s2))]
const I2S_LL_MCLK_DIVIDER_BIT_WIDTH: usize = 6;
const I2S_LL_MCLK_DIVIDER_MAX: usize = (1 << I2S_LL_MCLK_DIVIDER_BIT_WIDTH) - 1;
pub fn calculate_clock(
sample_rate: impl Into<fugit::HertzU32>,
data_bits: u8,
) -> I2sClockDividers {
// this loosely corresponds to `i2s_std_calculate_clock` and
// `i2s_ll_tx_set_mclk` in esp-idf
//
// main difference is we are using fixed-point arithmetic here
// plus adjusted for parallel interface clocking
let sclk = crate::soc::constants::I2S_SCLK; // for now it's fixed 160MHz and 96MHz (just H2)
let rate_hz: HertzU32 = sample_rate.into();
let rate = rate_hz.raw();
let mclk = rate * 2;
let bclk_divider: u32 = if data_bits == 8 { 2 } else { 1 };
let mut mclk_divider = sclk / mclk;
let mut ma: u32;
let mut mb: u32;
let mut denominator: u32 = 0;
let mut numerator: u32 = 0;
let freq_diff = sclk.abs_diff(mclk * mclk_divider);
if freq_diff != 0 {
let decimal = freq_diff as u64 * 10000 / mclk as u64;
// Carry bit if the decimal is greater than 1.0 - 1.0 / (63.0 * 2) = 125.0 /
// 126.0
if decimal > 1250000 / 126 {
mclk_divider += 1;
} else {
let mut min: u32 = !0;
for a in 2..=I2S_LL_MCLK_DIVIDER_MAX {
let b = (a as u64) * (freq_diff as u64 * 10000u64 / mclk as u64) + 5000;
ma = ((freq_diff as u64 * 10000u64 * a as u64) / 10000) as u32;
mb = (mclk as u64 * (b / 10000)) as u32;
if ma == mb {
denominator = a as u32;
numerator = (b / 10000) as u32;
break;
}
if mb.abs_diff(ma) < min {
denominator = a as u32;
numerator = b as u32;
min = mb.abs_diff(ma);
}
}
}
}
I2sClockDividers {
mclk_divider,
bclk_divider,
denominator,
numerator,
}
}
}
#[allow(missing_docs)]
pub trait RegBlock: PeripheralMarker + DmaEligible {
fn register_block(&self) -> &'static RegisterBlock;
}
#[allow(missing_docs)]
pub trait Signals {
fn get_peripheral(&self) -> crate::system::Peripheral;
fn get_dma_peripheral(&self) -> crate::dma::DmaPeripheral;
fn ws_signal(&self) -> OutputSignal;
fn data_out_signal(&self, i: usize, bits: u8) -> OutputSignal;
}
#[allow(missing_docs)]
pub trait Instance: Signals + RegBlock {
fn rx_reset(&self) {
let r = self.register_block();
r.conf().modify(|_, w| w.rx_reset().set_bit());
r.conf().modify(|_, w| w.rx_reset().clear_bit());
}
fn rx_dma_reset(&self) {
let r = self.register_block();
r.lc_conf().modify(|_, w| w.in_rst().set_bit());
r.lc_conf().modify(|_, w| w.in_rst().clear_bit());
}
fn rx_fifo_reset(&self) {
let r = self.register_block();
r.conf().modify(|_, w| w.rx_fifo_reset().set_bit());
r.conf().modify(|_, w| w.rx_fifo_reset().clear_bit());
}
fn tx_reset(&self) {
let r = self.register_block();
r.conf().modify(|_, w| w.tx_reset().set_bit());
// without this delay starting a subsequent transfer will hang waiting
// for tx_idle to clear (the transfer does not start).
// While 20 clocks works for 80MHz cpu but 100 is needed for 240MHz!
xtensa_lx::timer::delay(100);
r.conf().modify(|_, w| w.tx_reset().clear_bit());
}
fn tx_dma_reset(&self) {
let r = self.register_block();
r.lc_conf().modify(|_, w| w.out_rst().set_bit());
r.lc_conf().modify(|_, w| w.out_rst().clear_bit());
}
fn tx_fifo_reset(&self) {
let r = self.register_block();
r.conf().modify(|_, w| w.tx_fifo_reset().set_bit());
r.conf().modify(|_, w| w.tx_fifo_reset().clear_bit());
}
fn tx_clear_interrupts(&self) {
let r = self.register_block();
r.int_clr().write(|w| {
w.out_done().clear_bit_by_one();
w.out_total_eof().clear_bit_by_one()
});
}
fn tx_start(&self) {
let r = self.register_block();
// wait for data to show up in the fifo
while r.int_raw().read().tx_rempty().bit_is_clear() {
// wait
}
// without this transfers are not reliable!
xtensa_lx::timer::delay(1);
r.conf().modify(|_, w| w.tx_start().set_bit());
while r.state().read().tx_idle().bit_is_set() {
// wait
}
}
fn tx_stop(&self) {
let r = self.register_block();
r.conf().modify(|_, w| w.tx_start().clear_bit());
}
fn is_tx_done(&self) -> bool {
self.register_block().state().read().tx_idle().bit_is_set()
}
fn tx_wait_done(&self) {
let r = self.register_block();
while r.state().read().tx_idle().bit_is_clear() {
// wait
}
r.conf().modify(|_, w| w.tx_start().clear_bit());
r.int_clr().write(|w| {
w.out_done().clear_bit_by_one();
w.out_total_eof().clear_bit_by_one()
});
}
fn set_clock(&self, clock_settings: I2sClockDividers) {
let r = self.register_block();
r.clkm_conf().modify(|r, w| unsafe {
w.bits(r.bits() | (crate::soc::constants::I2S_DEFAULT_CLK_SRC << 21))
// select PLL_160M
});
#[cfg(esp32)]
r.clkm_conf().modify(|_, w| w.clka_ena().clear_bit());
r.clkm_conf().modify(|_, w| unsafe {
w.clk_en().set_bit();
w.clkm_div_num().bits(clock_settings.mclk_divider as u8)
});
r.clkm_conf().modify(|_, w| unsafe {
w.clkm_div_a().bits(clock_settings.denominator as u8);
w.clkm_div_b().bits(clock_settings.numerator as u8)
});
r.sample_rate_conf().modify(|_, w| unsafe {
w.tx_bck_div_num().bits(clock_settings.bclk_divider as u8);
w.rx_bck_div_num().bits(clock_settings.bclk_divider as u8)
});
}
fn setup(&self, frequency: impl Into<fugit::HertzU32>, bits: u8) {
let frequency: HertzU32 = frequency.into();
self.set_clock(calculate_clock(frequency, bits));
// Initialize I2S dev
self.rx_reset();
self.tx_reset();
self.rx_fifo_reset();
self.tx_fifo_reset();
self.rx_dma_reset();
self.tx_dma_reset();
let r = self.register_block();
// clear all bits and enable lcd mode
r.conf2().write(|w| {
// 8 bit mode needs this or it updates on half clocks!
w.lcd_tx_wrx2_en().bit(bits == 8);
w.lcd_en().set_bit()
});
r.sample_rate_conf().modify(|_, w| unsafe {
w.rx_bits_mod().bits(bits);
w.tx_bits_mod().bits(bits)
});
r.fifo_conf().write(|w| unsafe {
w.rx_fifo_mod_force_en().set_bit();
w.tx_fifo_mod_force_en().set_bit();
w.rx_fifo_mod().bits(1);
w.tx_fifo_mod().bits(1);
w.rx_data_num().bits(32);
w.tx_data_num().bits(32);
w.dscr_en().set_bit()
});
r.conf1().write(|w| {
w.tx_stop_en().set_bit();
w.rx_pcm_bypass().set_bit();
w.tx_pcm_bypass().set_bit()
});
r.conf_chan().write(|w| unsafe {
w.rx_chan_mod().bits(1);
w.tx_chan_mod().bits(1)
});
r.conf().modify(|_, w| {
w.rx_mono().set_bit();
w.tx_mono().set_bit();
w.rx_right_first().set_bit();
w.tx_right_first().set_bit()
});
r.timing().reset();
r.pd_conf().modify(|_, w| {
w.fifo_force_pu().set_bit();
w.fifo_force_pd().clear_bit()
});
}
}
impl RegBlock for I2S0 {
fn register_block(&self) -> &'static RegisterBlock {
unsafe { &*I2S0::PTR.cast::<RegisterBlock>() }
}
}
impl Signals for I2S0 {
fn get_peripheral(&self) -> crate::system::Peripheral {
crate::system::Peripheral::I2s0
}
fn get_dma_peripheral(&self) -> crate::dma::DmaPeripheral {
crate::dma::DmaPeripheral::I2s0
}
fn ws_signal(&self) -> OutputSignal {
OutputSignal::I2S0O_WS
}
fn data_out_signal(&self, i: usize, bits: u8) -> OutputSignal {
// signals for 8bit and 16bit both start at an offset of 8 for I2S0
let offset = match bits {
8 => 8,
16 => 8,
_ => panic!("Invalid number of bits"),
};
match i + offset {
0 => OutputSignal::I2S0O_DATA_0,
1 => OutputSignal::I2S0O_DATA_1,
2 => OutputSignal::I2S0O_DATA_2,
3 => OutputSignal::I2S0O_DATA_3,
4 => OutputSignal::I2S0O_DATA_4,
5 => OutputSignal::I2S0O_DATA_5,
6 => OutputSignal::I2S0O_DATA_6,
7 => OutputSignal::I2S0O_DATA_7,
8 => OutputSignal::I2S0O_DATA_8,
9 => OutputSignal::I2S0O_DATA_9,
10 => OutputSignal::I2S0O_DATA_10,
11 => OutputSignal::I2S0O_DATA_11,
12 => OutputSignal::I2S0O_DATA_12,
13 => OutputSignal::I2S0O_DATA_13,
14 => OutputSignal::I2S0O_DATA_14,
15 => OutputSignal::I2S0O_DATA_15,
16 => OutputSignal::I2S0O_DATA_16,
17 => OutputSignal::I2S0O_DATA_17,
18 => OutputSignal::I2S0O_DATA_18,
19 => OutputSignal::I2S0O_DATA_19,
20 => OutputSignal::I2S0O_DATA_20,
21 => OutputSignal::I2S0O_DATA_21,
22 => OutputSignal::I2S0O_DATA_22,
23 => OutputSignal::I2S0O_DATA_23,
_ => panic!("Invalid I2S0 Dout pin"),
}
}
}
impl Instance for I2S0 {}
impl RegBlock for I2S1 {
fn register_block(&self) -> &'static RegisterBlock {
unsafe { &*I2S1::PTR.cast::<RegisterBlock>() }
}
}
impl Signals for I2S1 {
fn get_peripheral(&self) -> crate::system::Peripheral {
crate::system::Peripheral::I2s1
}
fn get_dma_peripheral(&self) -> crate::dma::DmaPeripheral {
crate::dma::DmaPeripheral::I2s1
}
fn ws_signal(&self) -> OutputSignal {
OutputSignal::I2S1O_WS
}
fn data_out_signal(&self, i: usize, bits: u8) -> OutputSignal {
// Because of... reasons... the 16-bit values for i2s1 appear on d8...d23
let offset = if bits == 16 { 8 } else { 0 };
match i + offset {
0 => OutputSignal::I2S1O_DATA_0,
1 => OutputSignal::I2S1O_DATA_1,
2 => OutputSignal::I2S1O_DATA_2,
3 => OutputSignal::I2S1O_DATA_3,
4 => OutputSignal::I2S1O_DATA_4,
5 => OutputSignal::I2S1O_DATA_5,
6 => OutputSignal::I2S1O_DATA_6,
7 => OutputSignal::I2S1O_DATA_7,
8 => OutputSignal::I2S1O_DATA_8,
9 => OutputSignal::I2S1O_DATA_9,
10 => OutputSignal::I2S1O_DATA_10,
11 => OutputSignal::I2S1O_DATA_11,
12 => OutputSignal::I2S1O_DATA_12,
13 => OutputSignal::I2S1O_DATA_13,
14 => OutputSignal::I2S1O_DATA_14,
15 => OutputSignal::I2S1O_DATA_15,
16 => OutputSignal::I2S1O_DATA_16,
17 => OutputSignal::I2S1O_DATA_17,
18 => OutputSignal::I2S1O_DATA_18,
19 => OutputSignal::I2S1O_DATA_19,
20 => OutputSignal::I2S1O_DATA_20,
21 => OutputSignal::I2S1O_DATA_21,
22 => OutputSignal::I2S1O_DATA_22,
23 => OutputSignal::I2S1O_DATA_23,
_ => panic!("Invalid I2S1 Dout pin"),
}
}
}
impl Instance for I2S1 {}

View File

@ -191,6 +191,8 @@ pub mod hmac;
pub mod i2c;
#[cfg(any(i2s0, i2s1))]
pub mod i2s;
#[cfg(all(esp32, any(i2s0, i2s1)))]
pub mod i2s_parallel;
#[cfg(any(dport, interrupt_core0, interrupt_core1))]
pub mod interrupt;
#[cfg(lcd_cam)]

View File

@ -0,0 +1,92 @@
//! This shows using Parallel IO to output 8 bit parallel data at 1MHz clock
//! rate with a delay of 10ms between each transfer.
//!
//! The following wiring is assumed:
//! - Data pins => GPIO16, GPIO4, GPIO17, GPIO18, GPIO5, GPIO19, GPIO12, and
//! GPIO14
//! - Clock output pin => GPIO25
//!
//! You can use a logic analyzer to see how the pins are used.
//% CHIPS: esp32
//% FEATURES: embassy embassy-generic-timers
#![no_std]
#![no_main]
use embassy_executor::Spawner;
use embassy_time::Timer;
use esp_backtrace as _;
use esp_hal::{
dma::{Dma, DmaPriority, DmaTxBuf},
dma_buffers,
gpio::Io,
i2s_parallel::{I2sParallel, TxEightBits},
prelude::*,
timer::timg::TimerGroup,
};
use log::info;
const BUFFER_SIZE: usize = 256;
#[main]
async fn main(_spawner: Spawner) {
esp_println::logger::init_logger(log::LevelFilter::Info);
info!("Starting!");
let peripherals = esp_hal::init(esp_hal::Config::default());
let dma = Dma::new(peripherals.DMA);
let io = Io::new(peripherals.GPIO, peripherals.IO_MUX);
let timg0 = TimerGroup::new(peripherals.TIMG0);
info!("init embassy");
esp_hal_embassy::init(timg0.timer0);
let dma_channel = dma.i2s1channel;
let i2s = peripherals.I2S1;
let clock = io.pins.gpio25;
let pins = TxEightBits::new(
io.pins.gpio16,
io.pins.gpio4,
io.pins.gpio17,
io.pins.gpio18,
io.pins.gpio5,
io.pins.gpio19,
io.pins.gpio12,
io.pins.gpio14,
);
let (_, _, tx_buffer, tx_descriptors) = dma_buffers!(0, BUFFER_SIZE);
let mut parallel = I2sParallel::new(
i2s,
dma_channel.configure_for_async(false, DmaPriority::Priority0),
1.MHz(),
pins,
clock,
);
for (i, data) in tx_buffer.chunks_mut(4).enumerate() {
let offset = i * 4;
// i2s parallel driver expects the buffer to be interleaved
data[0] = (offset + 2) as u8;
data[1] = (offset + 3) as u8;
data[2] = offset as u8;
data[3] = (offset + 1) as u8;
}
let mut tx_buf = DmaTxBuf::new(tx_descriptors, tx_buffer).expect("DmaTxBuf::new failed");
info!("Sending {} bytes!", BUFFER_SIZE);
loop {
let mut xfer = match parallel.send(tx_buf) {
Ok(xfer) => xfer,
Err(_) => {
panic!("Failed to send buffer");
}
};
xfer.wait_for_done().await.expect("Failed to send buffer");
(parallel, tx_buf) = xfer.wait();
Timer::after_millis(10).await;
}
}

View File

@ -0,0 +1,86 @@
//! This shows using Parallel IO to output 8 bit parallel data at 1MHz clock
//! rate with a delay of 10ms between each transfer.
//!
//! The following wiring is assumed:
//! - Data pins => GPIO16, GPIO4, GPIO17, GPIO18, GPIO5, GPIO19, GPIO12, and
//! GPIO14
//! - Clock output pin => GPIO25
//!
//! You can use a logic analyzer to see how the pins are used.
//% CHIPS: esp32
#![no_std]
#![no_main]
use esp_backtrace as _;
use esp_hal::{
delay::Delay,
dma::{Dma, DmaPriority, DmaTxBuf},
dma_buffers,
gpio::Io,
i2s_parallel::{I2sParallel, TxEightBits},
prelude::*,
};
use log::info;
const BUFFER_SIZE: usize = 256;
#[entry]
fn main() -> ! {
esp_println::logger::init_logger(log::LevelFilter::Info);
info!("Starting!");
let peripherals = esp_hal::init(esp_hal::Config::default());
let dma = Dma::new(peripherals.DMA);
let io = Io::new(peripherals.GPIO, peripherals.IO_MUX);
let delay = Delay::new();
let dma_channel = dma.i2s1channel;
let i2s = peripherals.I2S1;
let clock = io.pins.gpio25;
let pins = TxEightBits::new(
io.pins.gpio16,
io.pins.gpio4,
io.pins.gpio17,
io.pins.gpio18,
io.pins.gpio5,
io.pins.gpio19,
io.pins.gpio12,
io.pins.gpio14,
);
let (_, _, tx_buffer, tx_descriptors) = dma_buffers!(0, BUFFER_SIZE);
let mut parallel = I2sParallel::new(
i2s,
dma_channel.configure(false, DmaPriority::Priority0),
1.MHz(),
pins,
clock,
);
for (i, data) in tx_buffer.chunks_mut(4).enumerate() {
let offset = i * 4;
// i2s parallel driver expects the buffer to be interleaved
data[0] = (offset + 2) as u8;
data[1] = (offset + 3) as u8;
data[2] = offset as u8;
data[3] = (offset + 1) as u8;
}
let mut tx_buf: DmaTxBuf =
DmaTxBuf::new(tx_descriptors, tx_buffer).expect("DmaTxBuf::new failed");
info!("Sending {} bytes!", BUFFER_SIZE);
loop {
let xfer = match parallel.send(tx_buf) {
Ok(xfer) => xfer,
Err(_) => {
panic!("Failed to send buffer");
}
};
(parallel, tx_buf) = xfer.wait();
delay.delay_millis(10);
}
}