port gpio from f4-hal

[burrbull]

No description provided.

[hannobraun]

Thank you, @burrbull, this looks good! Definitely better than the huge macro we have right now.

I reviewed this and couldn't spot any mistakes, but I also didn't have the time to check every detail (specifically I didn't double-check most of the register manipulation.

Can you confirm that this works? Happy to merge, if someone can confirm that this works as intended.

[burrbull]

Can you confirm that this works?

No, I can't. This works on F4 and F1 good enough, but I don't have any F7 device to test.

[hannobraun]

Okay, then let's wait until someone had a chance to test this. I probably won't get to that for a while.

[systec-ms]

I have tested the following cases on a STM32F767, everything worked fine.
For Ethernet, I used a modified stm32-eth crate.

• 2 Leds, GPIO{E3, E4}, set_low(), set_high() // Output PushPull

• 4 Buttons {B6, B5, C8, D13}, is_low(), is_high(), interrupt // Input PullUp

• 4 Leds {B1, D15, D14, D12}, set_low, set_high(), is_set_low() // Output PushPull

• UART {A9, A10}<Alternate>, into_alternate::<7>()

• Ethernet

  • MCO2 PC9.into_alternate::<0>()

  • gpioc.pc0.into_push_pull_output().set_high()

  • MDIO

    gpioa
        .pa2
        .into_alternate::<11>()
        .set_speed(stm32_eth::hal::gpio::Speed::VeryHigh)

  • MDC (clk)

  gpioc
      .pc1
      .into_alternate::<11>()
      .set_speed(stm32_eth::hal::gpio::Speed::VeryHigh);

  • ref_clk {A1}, crs {A7}, tx {B11, B12, B13}, rx {D0, D1} into_alternate::<11>().set_speed(VeryHigh);

[systec-ms]

In src/fmc_lcd/mod.rs the comments should still be adjusted:

stm32f7xx-hal/src/fmc_lcd/mod.rs

Line 153 in 7ad7efd

/// gpiod.pd14.into_alternate_af12(),

[hannobraun]

Thank you for testing, @systec-ms! And good catch on the doc examples that haven't been updated. I've opened #148, because that should really have been caught by the CI build.

I'm going to merge this pull request now. I think it would be unwise to hold it up because of the docs issue, as pretty much any change to GPIO going forward would introduce merge conflicts. Any pull requests that update the docs (and that solve #148) are of course appreciated!

Thanks again @burrbull!