456 lines
16 KiB
Rust
456 lines
16 KiB
Rust
//! HAL interface to the UARTE peripheral
|
|
//!
|
|
//! See product specification:
|
|
//!
|
|
//! - nrf52832: Section 35
|
|
//! - nrf52840: Section 6.34
|
|
use core::cmp::min;
|
|
use core::mem;
|
|
use core::ops::Deref;
|
|
use core::pin::Pin;
|
|
use core::sync::atomic::{compiler_fence, Ordering};
|
|
use core::task::{Context, Poll};
|
|
use embassy::interrupt::InterruptExt;
|
|
use embassy::io::{AsyncBufRead, AsyncWrite, Result};
|
|
use embassy::util::WakerRegistration;
|
|
use embassy_extras::low_power_wait_until;
|
|
use embassy_extras::peripheral::{PeripheralMutex, PeripheralState};
|
|
use embassy_extras::ring_buffer::RingBuffer;
|
|
use embedded_hal::digital::v2::OutputPin;
|
|
|
|
use crate::fmt::*;
|
|
use crate::hal::ppi::ConfigurablePpi;
|
|
use crate::interrupt::{self, Interrupt};
|
|
use crate::pac;
|
|
|
|
// Re-export SVD variants to allow user to directly set values
|
|
pub use crate::hal::uarte::Pins;
|
|
pub use pac::uarte0::{baudrate::BAUDRATE_A as Baudrate, config::PARITY_A as Parity};
|
|
|
|
#[derive(Copy, Clone, Debug, PartialEq)]
|
|
enum RxState {
|
|
Idle,
|
|
Receiving,
|
|
}
|
|
|
|
#[derive(Copy, Clone, Debug, PartialEq)]
|
|
enum TxState {
|
|
Idle,
|
|
Transmitting(usize),
|
|
}
|
|
|
|
struct State<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi> {
|
|
uarte: U,
|
|
timer: T,
|
|
ppi_channel_1: P1,
|
|
ppi_channel_2: P2,
|
|
|
|
rx: RingBuffer<'a>,
|
|
rx_state: RxState,
|
|
rx_waker: WakerRegistration,
|
|
|
|
tx: RingBuffer<'a>,
|
|
tx_state: TxState,
|
|
tx_waker: WakerRegistration,
|
|
}
|
|
|
|
/// Interface to a UARTE instance
|
|
///
|
|
/// This is a very basic interface that comes with the following limitations:
|
|
/// - The UARTE instances share the same address space with instances of UART.
|
|
/// You need to make sure that conflicting instances
|
|
/// are disabled before using `Uarte`. See product specification:
|
|
/// - nrf52832: Section 15.2
|
|
/// - nrf52840: Section 6.1.2
|
|
pub struct BufferedUarte<
|
|
'a,
|
|
U: Instance,
|
|
T: TimerInstance,
|
|
P1: ConfigurablePpi,
|
|
P2: ConfigurablePpi,
|
|
> {
|
|
inner: PeripheralMutex<State<'a, U, T, P1, P2>>,
|
|
}
|
|
|
|
impl<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi>
|
|
BufferedUarte<'a, U, T, P1, P2>
|
|
{
|
|
pub fn new(
|
|
uarte: U,
|
|
timer: T,
|
|
mut ppi_channel_1: P1,
|
|
mut ppi_channel_2: P2,
|
|
irq: U::Interrupt,
|
|
rx_buffer: &'a mut [u8],
|
|
tx_buffer: &'a mut [u8],
|
|
mut pins: Pins,
|
|
parity: Parity,
|
|
baudrate: Baudrate,
|
|
) -> Self {
|
|
// Select pins
|
|
uarte.psel.rxd.write(|w| {
|
|
unsafe { w.bits(pins.rxd.psel_bits()) };
|
|
w.connect().connected()
|
|
});
|
|
pins.txd.set_high().unwrap();
|
|
uarte.psel.txd.write(|w| {
|
|
unsafe { w.bits(pins.txd.psel_bits()) };
|
|
w.connect().connected()
|
|
});
|
|
|
|
// Optional pins
|
|
uarte.psel.cts.write(|w| {
|
|
if let Some(ref pin) = pins.cts {
|
|
unsafe { w.bits(pin.psel_bits()) };
|
|
w.connect().connected()
|
|
} else {
|
|
w.connect().disconnected()
|
|
}
|
|
});
|
|
|
|
uarte.psel.rts.write(|w| {
|
|
if let Some(ref pin) = pins.rts {
|
|
unsafe { w.bits(pin.psel_bits()) };
|
|
w.connect().connected()
|
|
} else {
|
|
w.connect().disconnected()
|
|
}
|
|
});
|
|
|
|
// Enable UARTE instance
|
|
uarte.enable.write(|w| w.enable().enabled());
|
|
|
|
// Enable interrupts
|
|
uarte.intenset.write(|w| w.endrx().set().endtx().set());
|
|
|
|
// Configure
|
|
let hardware_flow_control = pins.rts.is_some() && pins.cts.is_some();
|
|
uarte
|
|
.config
|
|
.write(|w| w.hwfc().bit(hardware_flow_control).parity().variant(parity));
|
|
|
|
// Configure frequency
|
|
uarte.baudrate.write(|w| w.baudrate().variant(baudrate));
|
|
|
|
// Disable the irq, let the Registration enable it when everything is set up.
|
|
irq.disable();
|
|
irq.pend();
|
|
|
|
// BAUDRATE register values are `baudrate * 2^32 / 16000000`
|
|
// source: https://devzone.nordicsemi.com/f/nordic-q-a/391/uart-baudrate-register-values
|
|
//
|
|
// We want to stop RX if line is idle for 2 bytes worth of time
|
|
// That is 20 bits (each byte is 1 start bit + 8 data bits + 1 stop bit)
|
|
// This gives us the amount of 16M ticks for 20 bits.
|
|
let timeout = 0x8000_0000 / (baudrate as u32 / 40);
|
|
|
|
timer.tasks_stop.write(|w| unsafe { w.bits(1) });
|
|
timer.bitmode.write(|w| w.bitmode()._32bit());
|
|
timer.prescaler.write(|w| unsafe { w.prescaler().bits(0) });
|
|
timer.cc[0].write(|w| unsafe { w.bits(timeout) });
|
|
timer.mode.write(|w| w.mode().timer());
|
|
timer.shorts.write(|w| {
|
|
w.compare0_clear().set_bit();
|
|
w.compare0_stop().set_bit();
|
|
w
|
|
});
|
|
|
|
ppi_channel_1.set_event_endpoint(&uarte.events_rxdrdy);
|
|
ppi_channel_1.set_task_endpoint(&timer.tasks_clear);
|
|
ppi_channel_1.set_fork_task_endpoint(&timer.tasks_start);
|
|
ppi_channel_1.enable();
|
|
|
|
ppi_channel_2.set_event_endpoint(&timer.events_compare[0]);
|
|
ppi_channel_2.set_task_endpoint(&uarte.tasks_stoprx);
|
|
ppi_channel_2.enable();
|
|
|
|
BufferedUarte {
|
|
inner: PeripheralMutex::new(
|
|
State {
|
|
uarte,
|
|
timer,
|
|
ppi_channel_1,
|
|
ppi_channel_2,
|
|
|
|
rx: RingBuffer::new(rx_buffer),
|
|
rx_state: RxState::Idle,
|
|
rx_waker: WakerRegistration::new(),
|
|
|
|
tx: RingBuffer::new(tx_buffer),
|
|
tx_state: TxState::Idle,
|
|
tx_waker: WakerRegistration::new(),
|
|
},
|
|
irq,
|
|
),
|
|
}
|
|
}
|
|
|
|
pub fn set_baudrate(self: Pin<&mut Self>, baudrate: Baudrate) {
|
|
self.inner().with(|state, _irq| {
|
|
let timeout = 0x8000_0000 / (baudrate as u32 / 40);
|
|
state.timer.cc[0].write(|w| unsafe { w.bits(timeout) });
|
|
state.timer.tasks_clear.write(|w| unsafe { w.bits(1) });
|
|
|
|
state
|
|
.uarte
|
|
.baudrate
|
|
.write(|w| w.baudrate().variant(baudrate));
|
|
});
|
|
}
|
|
|
|
fn inner(self: Pin<&mut Self>) -> Pin<&mut PeripheralMutex<State<'a, U, T, P1, P2>>> {
|
|
unsafe { Pin::new_unchecked(&mut self.get_unchecked_mut().inner) }
|
|
}
|
|
|
|
pub fn free(self: Pin<&mut Self>) -> (U, T, P1, P2, U::Interrupt) {
|
|
let (mut state, irq) = self.inner().free();
|
|
state.stop();
|
|
(
|
|
state.uarte,
|
|
state.timer,
|
|
state.ppi_channel_1,
|
|
state.ppi_channel_2,
|
|
irq,
|
|
)
|
|
}
|
|
}
|
|
|
|
impl<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi> Drop
|
|
for BufferedUarte<'a, U, T, P1, P2>
|
|
{
|
|
fn drop(&mut self) {
|
|
let inner = unsafe { Pin::new_unchecked(&mut self.inner) };
|
|
if let Some((mut state, _irq)) = inner.try_free() {
|
|
state.stop();
|
|
}
|
|
}
|
|
}
|
|
|
|
impl<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi> AsyncBufRead
|
|
for BufferedUarte<'a, U, T, P1, P2>
|
|
{
|
|
fn poll_fill_buf(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Result<&[u8]>> {
|
|
self.inner().with(|state, _irq| {
|
|
// Conservative compiler fence to prevent optimizations that do not
|
|
// take in to account actions by DMA. The fence has been placed here,
|
|
// before any DMA action has started
|
|
compiler_fence(Ordering::SeqCst);
|
|
trace!("poll_read");
|
|
|
|
// We have data ready in buffer? Return it.
|
|
let buf = state.rx.pop_buf();
|
|
if !buf.is_empty() {
|
|
trace!(" got {:?} {:?}", buf.as_ptr() as u32, buf.len());
|
|
let buf: &[u8] = buf;
|
|
let buf: &[u8] = unsafe { mem::transmute(buf) };
|
|
return Poll::Ready(Ok(buf));
|
|
}
|
|
|
|
trace!(" empty");
|
|
state.rx_waker.register(cx.waker());
|
|
Poll::<Result<&[u8]>>::Pending
|
|
})
|
|
}
|
|
|
|
fn consume(self: Pin<&mut Self>, amt: usize) {
|
|
self.inner().with(|state, irq| {
|
|
trace!("consume {:?}", amt);
|
|
state.rx.pop(amt);
|
|
irq.pend();
|
|
})
|
|
}
|
|
}
|
|
|
|
impl<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi> AsyncWrite
|
|
for BufferedUarte<'a, U, T, P1, P2>
|
|
{
|
|
fn poll_write(self: Pin<&mut Self>, cx: &mut Context<'_>, buf: &[u8]) -> Poll<Result<usize>> {
|
|
self.inner().with(|state, irq| {
|
|
trace!("poll_write: {:?}", buf.len());
|
|
|
|
let tx_buf = state.tx.push_buf();
|
|
if tx_buf.is_empty() {
|
|
trace!("poll_write: pending");
|
|
state.tx_waker.register(cx.waker());
|
|
return Poll::Pending;
|
|
}
|
|
|
|
let n = min(tx_buf.len(), buf.len());
|
|
tx_buf[..n].copy_from_slice(&buf[..n]);
|
|
state.tx.push(n);
|
|
|
|
trace!("poll_write: queued {:?}", n);
|
|
|
|
// Conservative compiler fence to prevent optimizations that do not
|
|
// take in to account actions by DMA. The fence has been placed here,
|
|
// before any DMA action has started
|
|
compiler_fence(Ordering::SeqCst);
|
|
|
|
irq.pend();
|
|
|
|
Poll::Ready(Ok(n))
|
|
})
|
|
}
|
|
}
|
|
|
|
impl<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi>
|
|
State<'a, U, T, P1, P2>
|
|
{
|
|
fn stop(&mut self) {
|
|
self.timer.tasks_stop.write(|w| unsafe { w.bits(1) });
|
|
if let RxState::Receiving = self.rx_state {
|
|
self.uarte.tasks_stoprx.write(|w| unsafe { w.bits(1) });
|
|
}
|
|
if let TxState::Transmitting(_) = self.tx_state {
|
|
self.uarte.tasks_stoptx.write(|w| unsafe { w.bits(1) });
|
|
}
|
|
if let RxState::Receiving = self.rx_state {
|
|
low_power_wait_until(|| self.uarte.events_endrx.read().bits() == 1);
|
|
}
|
|
if let TxState::Transmitting(_) = self.tx_state {
|
|
low_power_wait_until(|| self.uarte.events_endtx.read().bits() == 1);
|
|
}
|
|
}
|
|
}
|
|
|
|
impl<'a, U: Instance, T: TimerInstance, P1: ConfigurablePpi, P2: ConfigurablePpi> PeripheralState
|
|
for State<'a, U, T, P1, P2>
|
|
{
|
|
type Interrupt = U::Interrupt;
|
|
fn on_interrupt(&mut self) {
|
|
trace!("irq: start");
|
|
loop {
|
|
match self.rx_state {
|
|
RxState::Idle => {
|
|
trace!(" irq_rx: in state idle");
|
|
|
|
let buf = self.rx.push_buf();
|
|
if !buf.is_empty() {
|
|
trace!(" irq_rx: starting {:?}", buf.len());
|
|
self.rx_state = RxState::Receiving;
|
|
|
|
// Set up the DMA read
|
|
self.uarte.rxd.ptr.write(|w|
|
|
// The PTR field is a full 32 bits wide and accepts the full range
|
|
// of values.
|
|
unsafe { w.ptr().bits(buf.as_ptr() as u32) });
|
|
self.uarte.rxd.maxcnt.write(|w|
|
|
// We're giving it the length of the buffer, so no danger of
|
|
// accessing invalid memory. We have verified that the length of the
|
|
// buffer fits in an `u8`, so the cast to `u8` is also fine.
|
|
//
|
|
// The MAXCNT field is at least 8 bits wide and accepts the full
|
|
// range of values.
|
|
unsafe { w.maxcnt().bits(buf.len() as _) });
|
|
trace!(" irq_rx: buf {:?} {:?}", buf.as_ptr() as u32, buf.len());
|
|
|
|
// Start UARTE Receive transaction
|
|
self.uarte.tasks_startrx.write(|w|
|
|
// `1` is a valid value to write to task registers.
|
|
unsafe { w.bits(1) });
|
|
}
|
|
break;
|
|
}
|
|
RxState::Receiving => {
|
|
trace!(" irq_rx: in state receiving");
|
|
if self.uarte.events_endrx.read().bits() != 0 {
|
|
self.timer.tasks_stop.write(|w| unsafe { w.bits(1) });
|
|
|
|
let n: usize = self.uarte.rxd.amount.read().amount().bits() as usize;
|
|
trace!(" irq_rx: endrx {:?}", n);
|
|
self.rx.push(n);
|
|
|
|
self.uarte.events_endrx.reset();
|
|
|
|
self.rx_waker.wake();
|
|
self.rx_state = RxState::Idle;
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
loop {
|
|
match self.tx_state {
|
|
TxState::Idle => {
|
|
trace!(" irq_tx: in state Idle");
|
|
let buf = self.tx.pop_buf();
|
|
if !buf.is_empty() {
|
|
trace!(" irq_tx: starting {:?}", buf.len());
|
|
self.tx_state = TxState::Transmitting(buf.len());
|
|
|
|
// Set up the DMA write
|
|
self.uarte.txd.ptr.write(|w|
|
|
// The PTR field is a full 32 bits wide and accepts the full range
|
|
// of values.
|
|
unsafe { w.ptr().bits(buf.as_ptr() as u32) });
|
|
self.uarte.txd.maxcnt.write(|w|
|
|
// We're giving it the length of the buffer, so no danger of
|
|
// accessing invalid memory. We have verified that the length of the
|
|
// buffer fits in an `u8`, so the cast to `u8` is also fine.
|
|
//
|
|
// The MAXCNT field is 8 bits wide and accepts the full range of
|
|
// values.
|
|
unsafe { w.maxcnt().bits(buf.len() as _) });
|
|
|
|
// Start UARTE Transmit transaction
|
|
self.uarte.tasks_starttx.write(|w|
|
|
// `1` is a valid value to write to task registers.
|
|
unsafe { w.bits(1) });
|
|
}
|
|
break;
|
|
}
|
|
TxState::Transmitting(n) => {
|
|
trace!(" irq_tx: in state Transmitting");
|
|
if self.uarte.events_endtx.read().bits() != 0 {
|
|
self.uarte.events_endtx.reset();
|
|
|
|
trace!(" irq_tx: endtx {:?}", n);
|
|
self.tx.pop(n);
|
|
self.tx_waker.wake();
|
|
self.tx_state = TxState::Idle;
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
trace!("irq: end");
|
|
}
|
|
}
|
|
|
|
mod sealed {
|
|
pub trait Instance {}
|
|
|
|
impl Instance for crate::pac::UARTE0 {}
|
|
#[cfg(any(feature = "52833", feature = "52840", feature = "9160"))]
|
|
impl Instance for crate::pac::UARTE1 {}
|
|
|
|
pub trait TimerInstance {}
|
|
impl TimerInstance for crate::pac::TIMER0 {}
|
|
impl TimerInstance for crate::pac::TIMER1 {}
|
|
impl TimerInstance for crate::pac::TIMER2 {}
|
|
}
|
|
|
|
pub trait Instance: Deref<Target = pac::uarte0::RegisterBlock> + sealed::Instance {
|
|
type Interrupt: Interrupt;
|
|
}
|
|
|
|
impl Instance for pac::UARTE0 {
|
|
type Interrupt = interrupt::UARTE0_UART0;
|
|
}
|
|
|
|
#[cfg(any(feature = "52833", feature = "52840", feature = "9160"))]
|
|
impl Instance for pac::UARTE1 {
|
|
type Interrupt = interrupt::UARTE1;
|
|
}
|
|
|
|
pub trait TimerInstance:
|
|
Deref<Target = pac::timer0::RegisterBlock> + sealed::TimerInstance
|
|
{
|
|
}
|
|
impl TimerInstance for crate::pac::TIMER0 {}
|
|
impl TimerInstance for crate::pac::TIMER1 {}
|
|
impl TimerInstance for crate::pac::TIMER2 {}
|