Trying to use an SPI epaper display gets stuck somewhere

[JohnTheCoolingFan]

Hello, I am making a project on an esp32 (esp-wroom-32, esp32 revision 1.0 from espflash) that uses a epaper display connected over SPI. It's a 2.9inch white/black/red display from weact, analogous to the similar model from waveshare.
I've generated a project and set up SPI according to the examples in esp-rs documentation, but the program gets stuck somewhere in the Epd2in9bc::new method call from epd-waveshare: https://docs.rs/epd-waveshare/latest/epd_waveshare/epd2in9bc/struct.Epd2in9bc.html#method.new

My suspicion is that there's some kind of deadlock happening but I don't know why or how.

My async_main.rs:

#![no_std]
#![no_main]

use core::cell::RefCell;

use embassy_executor::Spawner;
use embassy_time::{Duration, Timer};
use embedded_graphics::{
    prelude::{Point, Primitive},
    primitives::{Line, PrimitiveStyle},
    Drawable,
};
use embedded_hal_bus::spi::RefCellDevice;
use esp_backtrace as _;
use esp_hal::{
    gpio::{Input, Level, NoPin, Output, Pull},
    peripherals::SPI2,
    prelude::*,
    spi::{
        master::{Config, Spi},
        SpiMode,
    },
    Blocking,
};
use log::info;

extern crate alloc;

use epd_waveshare::{epd2in9bc::*, prelude::*};

#[main]
async fn main(spawner: Spawner) {
    let peripherals = esp_hal::init({
        let mut config = esp_hal::Config::default();
        config.cpu_clock = CpuClock::max();
        config
    });

    esp_alloc::heap_allocator!(72 * 1024);

    //let mut delay = embassy_time::Delay;
    let mut delay = esp_hal::delay::Delay::new();

    esp_println::logger::init_logger_from_env();

    let timer0 = esp_hal::timer::timg::TimerGroup::new(peripherals.TIMG1);
    esp_hal_embassy::init(timer0.timer0);

    info!("Embassy initialized!");

    /*
    let timer1 = esp_hal::timer::timg::TimerGroup::new(peripherals.TIMG0);
    let _init = esp_wifi::init(
        timer1.timer0,
        esp_hal::rng::Rng::new(peripherals.RNG),
        peripherals.RADIO_CLK,
    )
    .unwrap();
    */

    info!("Initializing pins");

    let cs = Output::new_typed(peripherals.GPIO5, Level::High);
    let busy_in = Input::new_typed(peripherals.GPIO4, Pull::None);
    let rst = Output::new_typed(peripherals.GPIO16, Level::High);
    let dc = Output::new_typed(peripherals.GPIO17, Level::Low);

    info!("Initializing spi bus");

    let spi_bus: Spi<'static, Blocking, SPI2> = Spi::new_typed_with_config(
        peripherals.SPI2,
        Config {
            frequency: 8.MHz(),
            mode: SpiMode::Mode0,
            ..Config::default()
        },
    )
    //.with_cs(peripherals.GPIO5)
    .with_sck(peripherals.GPIO18)
    .with_mosi(peripherals.GPIO23)
    .with_miso(NoPin)
    .with_cs(NoPin);
    let spi_bus = RefCell::new(spi_bus);

    info!("Initializing spi device");

    //let mut spi = SpiDevice::new(spi_bus, Output::new(peripherals.GPIO5, Level::High));
    let mut spi = RefCellDevice::new(&spi_bus, cs, delay).unwrap();

    info!("Initializing epd");

    let mut epd =
        Epd2in9bc::new(&mut spi, busy_in, dc, rst, &mut delay, None).expect("EPD creation error");

    info!("Drawing");

    let mut mono_display = Display2in9bc::default();
    mono_display.set_rotation(DisplayRotation::Rotate90);

    let _ = Line::new(Point::new(0, 120), Point::new(0, 200))
        .into_styled(PrimitiveStyle::with_stroke(Color::Black, 1))
        .draw(&mut mono_display);

    let mut chromatic_display = Display2in9bc::default();
    chromatic_display.set_rotation(DisplayRotation::Rotate90);

    let _ = Line::new(Point::new(15, 120), Point::new(15, 200))
        .into_styled(PrimitiveStyle::with_stroke(Color::Black, 1))
        .draw(&mut chromatic_display);

    info!("Updating color frame");

    epd.update_color_frame(
        &mut spi,
        &mut delay,
        mono_display.buffer(),
        chromatic_display.buffer(),
    )
    .unwrap();

    epd.display_frame(&mut spi, &mut delay).unwrap();

    info!("Pre sleep");

    epd.sleep(&mut spi, &mut delay).unwrap();

    info!("Post sleep");

    // TODO: Spawn some tasks
    let _ = spawner;

    loop {
        info!("Hello world!");
        Timer::after(Duration::from_secs(1)).await;
    }

    // for inspiration have a look at the examples at https://github.com/esp-rs/esp-hal/tree/v0.22.0/examples/src/bin
}

And these are the crate versions used:

[dependencies]
esp-backtrace = { version = "0.14.2", features = [
    "esp32",
    "exception-handler",
    "panic-handler",
    "println",
]}

esp-hal = { version = "0.22.0", features = [
    "esp32",
] }
esp-println = { version = "0.12.0", features = ["esp32", "log"] }
log = { version = "0.4.21" }
esp-alloc = { version = "0.5.0" }
embedded-io = "0.6.1"

embedded-io-async = "0.6.1"
embassy-net = { version = "0.4.0", features = [ "tcp", "udp", "dhcpv4", "medium-ethernet"] }

esp-wifi = { version = "0.11.0", default-features=false, features = [
    "esp32",
    "utils",
    "wifi",
    "esp-alloc",
    "log",
] }
heapless = { version = "0.8.0", default-features = false }
smoltcp = { version = "0.11.0", default-features = false, features = [
    "medium-ethernet",
    "proto-dhcpv4",
    "proto-igmp",
    "proto-ipv4",
    "socket-dhcpv4",
    "socket-icmp",
    "socket-raw",
    "socket-tcp",
    "socket-udp",
] }
embassy-executor = { version = "0.6.0",  features = [
    "task-arena-size-12288",
] }
embassy-time     = { version = "0.3.1",  features = ["generic-queue-8"] }
esp-hal-embassy  = { version = "0.5.0",  features = ["esp32"] }
static_cell      = { version = "2.1.0",  features = ["nightly"] }
critical-section = "1.2.0"
epd-waveshare = "0.6.0"
embedded-graphics = "0.8.1"
embassy-embedded-hal = "0.2.0"
embassy-sync = "0.6.1"
embedded-hal-bus = "0.2.0"

I am using embassy because I intend to also use wifi to make a small desktop calendar. What am I doing wrong?
Asked about this on Rust Community discord and on the epd-waveshare repo but haven't gotten useful responses besides one project example but I couldn't see what they did that I am doing wrong.

Please help me with this problem, I don't know what is going on on the inside and don't know how to debug to see the exact point the board gets stuck at.

[yanshay]

First, a real long shot based on an issue I currently have - how do you know it gets stuck? Because you don't see the info! output right after?
I currently have issues (that are new with 0.22) where what's getting stuck is the info! itself. You could try replacing all info! with esp_println::println! and remove the initialization of the logger to see if it's still happening to eliminate that option. If that's the case I'd be interested to know.

If it's not that, then inside the init there's a wait_until_idle() which waits (blocking, so not that great for an async app) so could be the device is not responding as expected and not necessary a 'deadlock'. Is this an example that is known to work? It's always best to start from an example that's known to work with such devices. When I have such issues I try to reduce the bus speed, even considerably, sometimes it get things working.

[JohnTheCoolingFan]

You could try replacing all info! with esp_println::println! and remove the initialization of the logger to see if it's still happening to eliminate that option. If that's the case I'd be interested to know.

Doing this doesn't stop the program from hanging in the same place.

Is this an example that is known to work?

A project by another person, I've used it to try and tweak my code in some places: https://github.com/GnomedDev/epaper-simplyplural-badge/blob/0440c798289a3b36dcfb42f4938d9fc3a9111da9/src/main.rs#L92
Not my starting point, but I don't see much differences in it:

• Uses async main (through src/main.rs instead of src/bin/async_main.rs though)
• Uses esp_hal Delay
• Epd setup is fundamentally the same as in my current code as far as I can see

[yanshay]

I would try to run that application and once it's working tweak it step by step to mine until it doesn't, to find the difference.

Sometimes it's due to small differences that at least I don't notice when reading the code.

[JohnTheCoolingFan]

The moment I've changed the display to epd2in9bc it suddenly stopped working! This might be an issue with epd_waveshare after all...

[JohnTheCoolingFan]

Yeah, this is an issue with the display driver library, I've opened an issue on their repository now that captures my latest findings in an clean form: caemor/epd-waveshare#226