Yes, this isn't supported yet.

If anyone is interested in working on a "USART as SPI" driver and needs guidance on how to do this, feel free to ping me!

I'll gladly give this a shot in implementation, if you could aid a hand in the direction it needs to go in!

Currently, my confusion is just on how one would implement it, whilst also preventing usage of the USART for USART purposes.

whilst also preventing usage of the USART for USART purposes.

This part is actually very simple. The UsartSpi driver constructor consumes the USART peripheral, just like the current UART driver consumes it. Due to Rust's ownership rules, this means only one of the two can ever exist:

    let dp = atmega_hal::Peripherals::take().unwrap();
    let pins = atmega_hal::pins!(dp);
    let mut serial = Usart::new(
        dp.USART0, // <- USART0 is moved into `serial` here
        pins.pe0,
        pins.pe1.into_output(),
        Baudrate::<crate::CoreClock>::new(57600),
    );

    let mut spi = UsartSpi::new(dp.USART0, ...);  // <- compiler error because USART0 was moved previously and is no longer available

if you could aid a hand in the direction it needs to go in!

Generally, you'll need to add a usart_spi module in avr-hal-generic, similar in structure to the existing spi or usart modules. This defines the generic UsartSpi driver and a macro for the MCU-specific HALs. This macro is then instanciated in atmega-hal to implement USART as SPI for the available USART peripherals.

Roger on that. I do see that now, I was over-engineering that way to much...

I'll get to work on implementing the driver, however, would this preferred to be in the USART or SPI module?

I'd create a completely new module for USART as SPI. usart_spi I guess.

@CoolSlimbo Thanks for taking this one! I was missing this feature and pondered working on it, but my Rust skill is still lacking for it...

I'll be happy to give it a test!

@armandas If you could, that would be splendid.

I had completely forgot about this due to school, and my lack of multi USART devices, so testing it out and closing this up for me (us?) would be great.

@CoolSlimbo I just posted my results on the PR. There is a small compilation issue, so I think if you resolve it, the PR may be good ready for review.

Thank you.

As long as it runs the CI (which it should now), it'll be able to finally be merged!