stm32f1xx-hal Rcc improvement

Mar 14 '22 12:03 AndreySmirnov81

Sorry. @burrbull, please explain why "impairment"?

Mar 14 '22 13:03 AndreySmirnov81

Because I don't see any "improvement".

If you delete RCC dependency, you need mark enable & reset with unsafe
What sense of &self requirement? (not sure about this)
steal(). Again.

Mar 14 '22 13:03 burrbull

@burrbull

What is the reason unsafe? I think that after switching to using bit-banding, the need for &rcc disappeared.
&self is needed so that only the owner of the corresponding peripheral can do enable and reset for it. This will help to avoid mistakes like this:

   pub fn initialize(
      ...
      tim2: pac::TIM2,
      ...
   ) -> Self {
      ...
      let rcc = unsafe { &(*pac::RCC::PTR) };
      pac::TIM1::enable(rcc);
      pac::TIM1::reset(rcc);
      ...
   }

A temporary way out of the situation. In the future, it is necessary to make a similar change in the stm32-usbd create: Usb Peripheral::enable() -> UsbPeripheral::enable(&self). Is steal() acceptable here https://github.com/stm32-rs/stm32f3xx-hal/blob/565b43ceab7b8255c122e82f4f1f1f2b3e85f11b/src/serial.rs#L1270 ? I believe that using steal() in certain situations is quite acceptable.

Mar 14 '22 13:03 AndreySmirnov81

Please review and merge.

Mar 18 '22 07:03 AndreySmirnov81

@burrbull Do you agree with the changes?

Sep 26 '22 13:09 AndreySmirnov81

@therealprof, @TheZoq2 Please review

Oct 07 '22 08:10 AndreySmirnov81