diff --git a/embassy-rp/Cargo.toml b/embassy-rp/Cargo.toml
index 303617ff..b4dab7f4 100644
--- a/embassy-rp/Cargo.toml
+++ b/embassy-rp/Cargo.toml
@@ -36,6 +36,8 @@ embassy-macros = { version = "0.1.0", path = "../embassy-macros", features = ["r
 atomic-polyfill = "0.1.5"
 defmt = { version = "0.3", optional = true }
 log = { version = "0.4.14", optional = true }
+nb = "1.0.0"
+cfg-if = "1.0.0"
 cortex-m-rt = ">=0.6.15,<0.8"
 cortex-m = "0.7.3"
 critical-section = "0.2.5"
diff --git a/embassy-rp/src/dma.rs b/embassy-rp/src/dma.rs
index 235e92d7..42c4fd13 100644
--- a/embassy-rp/src/dma.rs
+++ b/embassy-rp/src/dma.rs
@@ -1,5 +1,7 @@
 use core::sync::atomic::{compiler_fence, Ordering};
 
+use embassy_hal_common::impl_peripheral;
+
 use crate::pac::dma::vals;
 use crate::{pac, peripherals};
 
@@ -35,6 +37,10 @@ impl<T: Channel> Dma<T> {
     }
 }
 
+pub struct NoDma;
+
+impl_peripheral!(NoDma);
+
 mod sealed {
     use super::*;
 
diff --git a/embassy-rp/src/uart.rs b/embassy-rp/src/uart.rs
index b19f043f..db135d04 100644
--- a/embassy-rp/src/uart.rs
+++ b/embassy-rp/src/uart.rs
@@ -1,42 +1,199 @@
-use embassy_hal_common::{into_ref, PeripheralRef};
-use gpio::Pin;
+use core::marker::PhantomData;
 
-use crate::{gpio, pac, peripherals, Peripheral};
+use embassy_hal_common::{into_ref, PeripheralRef};
+
+use crate::gpio::sealed::Pin;
+use crate::gpio::AnyPin;
+use crate::{pac, peripherals, Peripheral};
+
+#[derive(Clone, Copy, PartialEq, Eq, Debug)]
+pub enum DataBits {
+    DataBits5,
+    DataBits6,
+    DataBits7,
+    DataBits8,
+}
+
+impl DataBits {
+    fn bits(&self) -> u8 {
+        match self {
+            Self::DataBits5 => 0b00,
+            Self::DataBits6 => 0b01,
+            Self::DataBits7 => 0b10,
+            Self::DataBits8 => 0b11,
+        }
+    }
+}
+
+#[derive(Clone, Copy, PartialEq, Eq, Debug)]
+pub enum Parity {
+    ParityNone,
+    ParityEven,
+    ParityOdd,
+}
+
+#[derive(Clone, Copy, PartialEq, Eq, Debug)]
+pub enum StopBits {
+    #[doc = "1 stop bit"]
+    STOP1,
+    #[doc = "2 stop bits"]
+    STOP2,
+}
 
 #[non_exhaustive]
+#[derive(Clone, Copy, PartialEq, Eq, Debug)]
 pub struct Config {
     pub baudrate: u32,
-    pub data_bits: u8,
-    pub stop_bits: u8,
+    pub data_bits: DataBits,
+    pub stop_bits: StopBits,
+    pub parity: Parity,
 }
 
 impl Default for Config {
     fn default() -> Self {
         Self {
             baudrate: 115200,
-            data_bits: 8,
-            stop_bits: 1,
+            data_bits: DataBits::DataBits8,
+            stop_bits: StopBits::STOP1,
+            parity: Parity::ParityNone,
         }
     }
 }
 
+/// Serial error
+#[derive(Debug, Eq, PartialEq, Copy, Clone)]
+#[cfg_attr(feature = "defmt", derive(defmt::Format))]
+#[non_exhaustive]
+pub enum Error {
+    /// Triggered when the FIFO (or shift-register) is overflowed.
+    Overrun,
+    /// Triggered when a break is received
+    Break,
+    /// Triggered when there is a parity mismatch between what's received and
+    /// our settings.
+    Parity,
+    /// Triggered when the received character didn't have a valid stop bit.
+    Framing,
+}
+
 pub struct Uart<'d, T: Instance> {
-    inner: PeripheralRef<'d, T>,
+    tx: UartTx<'d, T>,
+    rx: UartRx<'d, T>,
+}
+
+pub struct UartTx<'d, T: Instance> {
+    phantom: PhantomData<&'d mut T>,
+}
+
+pub struct UartRx<'d, T: Instance> {
+    phantom: PhantomData<&'d mut T>,
+}
+
+impl<'d, T: Instance> UartTx<'d, T> {
+    fn new() -> Self {
+        Self { phantom: PhantomData }
+    }
+
+    pub async fn write(&mut self, _buffer: &[u8]) -> Result<(), Error> {
+        todo!()
+    }
+
+    pub fn blocking_write(&mut self, buffer: &[u8]) -> Result<(), Error> {
+        let r = T::regs();
+        unsafe {
+            for &b in buffer {
+                while r.uartfr().read().txff() {}
+                r.uartdr().write(|w| w.set_data(b));
+            }
+        }
+        Ok(())
+    }
+
+    pub fn blocking_flush(&mut self) -> Result<(), Error> {
+        let r = T::regs();
+        unsafe { while r.uartfr().read().txff() {} }
+        Ok(())
+    }
+}
+
+impl<'d, T: Instance> UartRx<'d, T> {
+    fn new() -> Self {
+        Self { phantom: PhantomData }
+    }
+
+    pub async fn read(&mut self, _buffer: &mut [u8]) -> Result<(), Error> {
+        todo!();
+    }
+
+    pub fn blocking_read(&mut self, buffer: &mut [u8]) -> Result<(), Error> {
+        let r = T::regs();
+        unsafe {
+            for b in buffer {
+                *b = loop {
+                    let dr = r.uartdr().read();
+
+                    if dr.oe() {
+                        return Err(Error::Overrun);
+                    } else if dr.be() {
+                        return Err(Error::Break);
+                    } else if dr.pe() {
+                        return Err(Error::Parity);
+                    } else if dr.fe() {
+                        return Err(Error::Framing);
+                    } else if dr.fe() {
+                        break dr.data();
+                    }
+                };
+            }
+        }
+        Ok(())
+    }
 }
 
 impl<'d, T: Instance> Uart<'d, T> {
+    /// Create a new UARTE without hardware flow control
     pub fn new(
-        inner: impl Peripheral<P = T> + 'd,
+        uart: impl Peripheral<P = T> + 'd,
+        tx: impl Peripheral<P = impl TxPin<T>> + 'd,
+        rx: impl Peripheral<P = impl RxPin<T>> + 'd,
+        config: Config,
+    ) -> Self {
+        into_ref!(tx, rx);
+        Self::new_inner(uart, rx.map_into(), tx.map_into(), None, None, config)
+    }
+
+    /// Create a new UART with hardware flow control (RTS/CTS)
+    pub fn new_with_rtscts(
+        uart: impl Peripheral<P = T> + 'd,
         tx: impl Peripheral<P = impl TxPin<T>> + 'd,
         rx: impl Peripheral<P = impl RxPin<T>> + 'd,
         cts: impl Peripheral<P = impl CtsPin<T>> + 'd,
         rts: impl Peripheral<P = impl RtsPin<T>> + 'd,
         config: Config,
     ) -> Self {
-        into_ref!(inner, tx, rx, cts, rts);
+        into_ref!(tx, rx, cts, rts);
+        Self::new_inner(
+            uart,
+            rx.map_into(),
+            tx.map_into(),
+            Some(cts.map_into()),
+            Some(rts.map_into()),
+            config,
+        )
+    }
+
+    fn new_inner(
+        _uart: impl Peripheral<P = T> + 'd,
+        tx: PeripheralRef<'d, AnyPin>,
+        rx: PeripheralRef<'d, AnyPin>,
+        cts: Option<PeripheralRef<'d, AnyPin>>,
+        rts: Option<PeripheralRef<'d, AnyPin>>,
+        config: Config,
+    ) -> Self {
+        into_ref!(_uart);
 
         unsafe {
-            let p = inner.regs();
+            let r = T::regs();
 
             let clk_base = crate::clocks::clk_peri_freq();
 
@@ -53,49 +210,217 @@ impl<'d, T: Instance> Uart<'d, T> {
             }
 
             // Load PL011's baud divisor registers
-            p.uartibrd().write_value(pac::uart::regs::Uartibrd(baud_ibrd));
-            p.uartfbrd().write_value(pac::uart::regs::Uartfbrd(baud_fbrd));
+            r.uartibrd().write_value(pac::uart::regs::Uartibrd(baud_ibrd));
+            r.uartfbrd().write_value(pac::uart::regs::Uartfbrd(baud_fbrd));
 
-            p.uartlcr_h().write(|w| {
-                w.set_wlen(config.data_bits - 5);
-                w.set_stp2(config.stop_bits == 2);
-                w.set_pen(false);
-                w.set_eps(false);
+            let (pen, eps) = match config.parity {
+                Parity::ParityNone => (false, false),
+                Parity::ParityEven => (true, true),
+                Parity::ParityOdd => (true, false),
+            };
+
+            r.uartlcr_h().write(|w| {
+                w.set_wlen(config.data_bits.bits());
+                w.set_stp2(config.stop_bits == StopBits::STOP2);
+                w.set_pen(pen);
+                w.set_eps(eps);
                 w.set_fen(true);
             });
 
-            p.uartcr().write(|w| {
+            r.uartcr().write(|w| {
                 w.set_uarten(true);
                 w.set_rxe(true);
                 w.set_txe(true);
+                w.set_ctsen(cts.is_some());
+                w.set_rtsen(rts.is_some());
             });
 
             tx.io().ctrl().write(|w| w.set_funcsel(2));
             rx.io().ctrl().write(|w| w.set_funcsel(2));
-            cts.io().ctrl().write(|w| w.set_funcsel(2));
-            rts.io().ctrl().write(|w| w.set_funcsel(2));
+            if let Some(pin) = &cts {
+                pin.io().ctrl().write(|w| w.set_funcsel(2));
+            }
+            if let Some(pin) = &rts {
+                pin.io().ctrl().write(|w| w.set_funcsel(2));
+            }
+        }
+
+        Self {
+            tx: UartTx::new(),
+            rx: UartRx::new(),
         }
-        Self { inner }
     }
 
-    pub fn send(&mut self, data: &[u8]) {
-        unsafe {
-            let p = self.inner.regs();
+    pub async fn write(&mut self, buffer: &[u8]) -> Result<(), Error> {
+        self.tx.write(buffer).await
+    }
 
-            for &byte in data {
-                if !p.uartfr().read().txff() {
-                    p.uartdr().write(|w| w.set_data(byte));
+    pub fn blocking_write(&mut self, buffer: &[u8]) -> Result<(), Error> {
+        self.tx.blocking_write(buffer)
+    }
+
+    pub fn blocking_flush(&mut self) -> Result<(), Error> {
+        self.tx.blocking_flush()
+    }
+
+    pub async fn read(&mut self, buffer: &mut [u8]) -> Result<(), Error> {
+        self.rx.read(buffer).await
+    }
+
+    pub fn blocking_read(&mut self, buffer: &mut [u8]) -> Result<(), Error> {
+        self.rx.blocking_read(buffer)
+    }
+
+    /// Split the Uart into a transmitter and receiver, which is
+    /// particuarly useful when having two tasks correlating to
+    /// transmitting and receiving.
+    pub fn split(self) -> (UartTx<'d, T>, UartRx<'d, T>) {
+        (self.tx, self.rx)
+    }
+}
+
+mod eh02 {
+    use super::*;
+
+    impl<'d, T: Instance> embedded_hal_02::serial::Read<u8> for UartRx<'d, T> {
+        type Error = Error;
+        fn read(&mut self) -> Result<u8, nb::Error<Self::Error>> {
+            let r = T::regs();
+            unsafe {
+                let dr = r.uartdr().read();
+
+                if dr.oe() {
+                    Err(nb::Error::Other(Error::Overrun))
+                } else if dr.be() {
+                    Err(nb::Error::Other(Error::Break))
+                } else if dr.pe() {
+                    Err(nb::Error::Other(Error::Parity))
+                } else if dr.fe() {
+                    Err(nb::Error::Other(Error::Framing))
+                } else if dr.fe() {
+                    Ok(dr.data())
+                } else {
+                    Err(nb::Error::WouldBlock)
                 }
             }
         }
     }
+
+    impl<'d, T: Instance> embedded_hal_02::blocking::serial::Write<u8> for UartTx<'d, T> {
+        type Error = Error;
+        fn bwrite_all(&mut self, buffer: &[u8]) -> Result<(), Self::Error> {
+            self.blocking_write(buffer)
+        }
+        fn bflush(&mut self) -> Result<(), Self::Error> {
+            self.blocking_flush()
+        }
+    }
+
+    impl<'d, T: Instance> embedded_hal_02::serial::Read<u8> for Uart<'d, T> {
+        type Error = Error;
+        fn read(&mut self) -> Result<u8, nb::Error<Self::Error>> {
+            embedded_hal_02::serial::Read::read(&mut self.rx)
+        }
+    }
+
+    impl<'d, T: Instance> embedded_hal_02::blocking::serial::Write<u8> for Uart<'d, T> {
+        type Error = Error;
+        fn bwrite_all(&mut self, buffer: &[u8]) -> Result<(), Self::Error> {
+            self.blocking_write(buffer)
+        }
+        fn bflush(&mut self) -> Result<(), Self::Error> {
+            self.blocking_flush()
+        }
+    }
+}
+
+#[cfg(feature = "unstable-traits")]
+mod eh1 {
+    use super::*;
+
+    impl embedded_hal_1::serial::Error for Error {
+        fn kind(&self) -> embedded_hal_1::serial::ErrorKind {
+            match *self {
+                Self::Framing => embedded_hal_1::serial::ErrorKind::FrameFormat,
+                Self::Break => embedded_hal_1::serial::ErrorKind::Other,
+                Self::Overrun => embedded_hal_1::serial::ErrorKind::Overrun,
+                Self::Parity => embedded_hal_1::serial::ErrorKind::Parity,
+            }
+        }
+    }
+
+    impl<'d, T: Instance> embedded_hal_1::serial::ErrorType for Uart<'d, T> {
+        type Error = Error;
+    }
+
+    impl<'d, T: Instance> embedded_hal_1::serial::ErrorType for UartTx<'d, T> {
+        type Error = Error;
+    }
+
+    impl<'d, T: Instance> embedded_hal_1::serial::ErrorType for UartRx<'d, T> {
+        type Error = Error;
+    }
+}
+
+cfg_if::cfg_if! {
+    if #[cfg(all(feature = "unstable-traits", feature = "nightly", feature = "_todo_embedded_hal_serial"))] {
+        use core::future::Future;
+
+        impl<'d, T: Instance> embedded_hal_async::serial::Write for UartTx<'d, T>
+        {
+            type WriteFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a where Self: 'a;
+
+            fn write<'a>(&'a mut self, buf: &'a [u8]) -> Self::WriteFuture<'a> {
+                self.write(buf)
+            }
+
+            type FlushFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a where Self: 'a;
+
+            fn flush<'a>(&'a mut self) -> Self::FlushFuture<'a> {
+                async move { Ok(()) }
+            }
+        }
+
+        impl<'d, T: Instance> embedded_hal_async::serial::Read for UartRx<'d, T>
+        {
+            type ReadFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a where Self: 'a;
+
+            fn read<'a>(&'a mut self, buf: &'a mut [u8]) -> Self::ReadFuture<'a> {
+                self.read(buf)
+            }
+        }
+
+        impl<'d, T: Instance> embedded_hal_async::serial::Write for Uart<'d, T>
+        {
+            type WriteFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a where Self: 'a;
+
+            fn write<'a>(&'a mut self, buf: &'a [u8]) -> Self::WriteFuture<'a> {
+                self.write(buf)
+            }
+
+            type FlushFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a where Self: 'a;
+
+            fn flush<'a>(&'a mut self) -> Self::FlushFuture<'a> {
+                async move { Ok(()) }
+            }
+        }
+
+        impl<'d, T: Instance> embedded_hal_async::serial::Read for Uart<'d, T>
+        {
+            type ReadFuture<'a> = impl Future<Output = Result<(), Self::Error>> + 'a where Self: 'a;
+
+            fn read<'a>(&'a mut self, buf: &'a mut [u8]) -> Self::ReadFuture<'a> {
+                self.read(buf)
+            }
+        }
+    }
 }
 
 mod sealed {
     use super::*;
 
     pub trait Instance {
-        fn regs(&self) -> pac::uart::Uart;
+        fn regs() -> pac::uart::Uart;
     }
     pub trait TxPin<T: Instance> {}
     pub trait RxPin<T: Instance> {}
@@ -106,23 +431,23 @@ mod sealed {
 pub trait Instance: sealed::Instance {}
 
 macro_rules! impl_instance {
-    ($type:ident, $irq:ident) => {
-        impl sealed::Instance for peripherals::$type {
-            fn regs(&self) -> pac::uart::Uart {
-                pac::$type
+    ($inst:ident, $irq:ident) => {
+        impl sealed::Instance for peripherals::$inst {
+            fn regs() -> pac::uart::Uart {
+                pac::$inst
             }
         }
-        impl Instance for peripherals::$type {}
+        impl Instance for peripherals::$inst {}
     };
 }
 
 impl_instance!(UART0, UART0);
 impl_instance!(UART1, UART1);
 
-pub trait TxPin<T: Instance>: sealed::TxPin<T> + Pin {}
-pub trait RxPin<T: Instance>: sealed::RxPin<T> + Pin {}
-pub trait CtsPin<T: Instance>: sealed::CtsPin<T> + Pin {}
-pub trait RtsPin<T: Instance>: sealed::RtsPin<T> + Pin {}
+pub trait TxPin<T: Instance>: sealed::TxPin<T> + crate::gpio::Pin {}
+pub trait RxPin<T: Instance>: sealed::RxPin<T> + crate::gpio::Pin {}
+pub trait CtsPin<T: Instance>: sealed::CtsPin<T> + crate::gpio::Pin {}
+pub trait RtsPin<T: Instance>: sealed::RtsPin<T> + crate::gpio::Pin {}
 
 macro_rules! impl_pin {
     ($pin:ident, $instance:ident, $function:ident) => {