Adrian Figueroa
Adrian Figueroa
Regarding your above feature request, here are my first thoughts. ## Split `configure` and `start` The GPDMA linked-list table itself must not change its location in memory, after its pointer...
Hi @leftger! Yes, it's ready. I pinged @Dirbaio but he was swamped in work, so this is still pending a review. Also, I assume that reviewing this takes quite some...
Thank you all for having a look! I will rebase again and test on hardware - just to be sure that nothing broke.
> rename `ringbuffered.rs` into `ringbuffer.rs` Sounds reasonable!
## After rebasing - Pending changelog. - My local USB audio test with GPDMA -> SAI doesn't work anymore. I have to investigate.
Yes, I will add it. It is already a quite small example.
@leftger I added an H5 example. It is quite minimal and runs on 8 MHz SAI clock. I get no activity on either data line or left/right clock. Only the...
@leftger I found the problem, it was caused by refactoring some time ago... See the fix at https://github.com/embassy-rs/embassy/pull/3923/commits/62b4acb3ffc9be5d0d8dae5cfb94686b02daf02a The driver only made linked-lists for read-mode (peripheral to memory), even for...
Great, take your time! The fix is very minor. I will try to add an example for GPDMA ringbuffer UART as well, I think this will help others to test....
Tested UCPD on STM32H5 with the DMA changes, using this example: https://github.com/elagil/usbpd/tree/900326db415c189f216b4c8307b8dad24ac83d6b/examples/embassy-nucleo-h563zi Still works as before. See the opening post for a list of features and completed tests.