Issue 2059 ringbuffers on gpdma

[tyler-gilbert]

No description provided.

[tyler-gilbert]

I still need to do some additional testing on the STM32U545 and STM32F429 before this is ready to merge.

[tyler-gilbert]

Everything is working on STM32U545. Need to check STM32F429 before merging.

[tyler-gilbert]

Tested and working on STM32F429 also (with dma). Not tested with BDMA but that has the same API as DMA. The only thing this adds to BDMA is the async stop and the write_immediate to pre-fill the buffer.

[tyler-gilbert]

Should be ready for review and to merge.

[tyler-gilbert]

@xoviat @Dirbaio

Please let me know if this needs anything else.

[Dirbaio]

This PR has quite a bit of unrelated things mixed:

• GPDMA ringbuffer support.
• New APIs stop and write_immediate.
• U5 RCC support for PLLP/Q
• SAI changes

Could you split them all in separate PRs, so we can discuss/review them independently?

Also, please rebase so you don't add merge commits to the history.

Additionally, now that this adds ringbuffers on GPDMA, we should enable the ringbuffered uart test for it. Could you do it?

• Remove this https:/embassy-rs/embassy/blob/main/tests/stm32/src/bin/usart_rx_ringbuffered.rs#L1
• Remove all mentions to not-gpdma in Cargo.toml https:/embassy-rs/embassy/blob/main/tests/stm32/Cargo.toml
• Run gen_test.py

[tyler-gilbert]

@Dirbaio,

Sounds good. I will work on that.

[tyler-gilbert]

STM32 U5 changes moved to: #2396
Add write_immediate() function: #2397
Add stop() function: #2399

I will create another MR with the GPDMA ring buffers once the above are merged. Then one more MR to address the tweaks to the SAI driver that add the write_immediate() and stop() functions.

[elagil]

@tyler-gilbert Interested in this as well, what is holding it back at the moment?
I can test on STM32H5 and H7 and maybe contribute something.