Older families like F1 and F4 have a consistent naming for the CAN
peripherals: CAN when there is only one instance, CAN1/CAN2/CAN2 if
there are multiple instances.
Newer families like L4 and F7 use the name CAN1 even if there is only
one instance. The number of filter banks is incorrect for those.
Affected chips:
* STM32F722
* STM32F723
* STM32F730
* STM32F732
* STM32F733
* STM32L4P5
* STM32L4Q5
* STM32L4R5
* STM32L4R7
* STM32L4R9
* STM32L4S5
* STM32L4S7
* STM32L4S9
* STM32L431
* STM32L432
* STM32L433
* STM32L442
* STM32L443
* STM32L451
* STM32L452
* STM32L462
* STM32L471
* STM32L475
* STM32L476
* STM32L485
* STM32L486
* Generate a core index put into the PAC for the peripherals to use as
index into registers.
* Add EXTI v2 which uses CORE_INDEX to index exti registers
* Add IRQ-driven buffered USART implementation for STM32 v2 usart
* Implementation based on nRF UARTE, but simplified to not use DMA to
avoid complex interaction between DMA and USART.
* Implementation of AsyncBufRead and AsyncWrite traits
* Some unit tests to ring buffer
* Update polyfill version
* Update sub module to get usart IRQ fix
Portal is very niche, I've only ever used it in `nrf-softdevice` and in a very particular case:
sending event raw-pointers across tasks but "synchronously", because the destination task must process
it now, so it's not deallocated. For most usecases, Signal or channels is enough.
It's unclear to me whether it can be made ub-free. It has problems with reentrancy. It's also not leak-safe, which is quite annoying.
So, remove it for now. We can always add it back later.
Fixes#18
Syn considers doc comments to be attributes, so forwarding those fixes the issue.
It's probably a good idea to forward most attributes anyway.
