Drivers: spi: stm32: add support for DMA in asynchronous mode

[ajarmouni-st]

Adds support for DMA in asynchronous mode.

Tested on nucleo_f767zi& nucleo_f429zi successfully.
west build -p always -b nucleo_f429zi tests/drivers/spi/spi_loopback -T drivers.spi.stm32_spi_async_dma.loopback

drivers.spi.stm32_spi_async_dma.loopback test case passes on stm32h573i_dk, nucleo_f207zg, nucleo_f429zi, nucleo_f746zg
But fails on nucleo_wb55rg, nucleo_wl55jc, nucleo_g474re, nucleo_h743zi

[ajarmouni-st]

Cc @dgastonochoa

[777aker]

In the async callback if rx is NULL (ie I only want to use TX) no clean up is performed and therefore you can only send data once and it thinks the transaction is ongoing forever waiting for RX.

EDIT: I realized my fix creates problems with it being asynchronous

[777aker]

Ok, I fixed it. Basically, you enable interrupts and setup spi interrupts. When dma callback is called disable dma and enable spi interrupts. Then, when the transfer is complete and the spi interrupt is called disable spi interrupts enable dma and shift the dma buffers. My work is working on upstreaming this. Hopefully, I can post code soon.

Green line is the user thread doing work. Tan line is clock, red is mosi, yellow chip select.
 

[erwango]

@777aker @ajarmouni-st is no longer contributing to this project. Don't hesitate to take his commits in your PR for contribution.

Then, note that RTIO API is probably the way to go to support real async on this driver (see #67327)

[github-actions]

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.

[kaarelen]

@777aker will you finish this pull request? I'm ready to provide any possible help, but I'm kinda new into Zephyr, so I need some guidance.

I really want to make this feature happen!

[erwango]

@JarmouniA, @777aker Are you still working on this ?
Looking at this PR, I was actually wondering if it would make things any simpler to limit DMA support to SPI ASYNC API. Any opinion ?

[JarmouniA]

@JarmouniA, @777aker Are you still working on this ?

@erwango Not at the moment.

Looking at this PR, I was actually wondering if it would make things any simpler to limit DMA support to SPI ASYNC API. Any opinion ?

That would probably upset many users that only want DMA for higher transfer speeds, & don't care much about the asynchronous part.

[JarmouniA]

Also, I think the implementation should be reworked with this suggestion, it will make it more robust & probably much simpler.

When dma callback is called disable dma and enable spi interrupts. Then, when the transfer is complete and the spi interrupt is called disable spi interrupts enable dma and shift the dma buffers.

[github-actions]

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.

[github-actions]

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.