Implement HW SPI

[MrSurly]

Created to file a PR against.

I have this working. Need to test a bit more, and go over the code to make sure it's worthy.

Also, would like to get feedback on micropython/micropython #2986 before filing a PR

[dpgeorge]

Re the hspi driver: please call it something that mirrors the MCU. Do Espressif call it HSPI? Maybe hard_spi is a better name (see stmhal/spi.c at the bottom of that file for more hints).

[MrSurly]

Re the hspi driver: please call it something that mirrors the MCU. Do Espressif call it HSPI? Maybe hard_spi is a better name (see stmhal/spi.c at the bottom of that file for more hints).

I'm assuming this only applies to the .c and .h files, right? Presently, I'm calling them machine_hspi.c and machine_hspi.h. I personally don't care at all what they're called, but since there seemed to be some precedence for the "hspi" name, I kept that. Coincidentally (see below), one of the SPI controllers on the ESP32 is actually called HSPI.

My personal inclination would be to name them machine_hardware_spi.* or machine_hw_spi.*.

The ESP32 has two SPI "hosts" (as they're called) that can be accessed via the IDF: HSPI and VSPI. Interestingly, you can use any pins (or none!) that you like for your SPI host -- the ESP32 has an internal routing matrix (two, actually ...)

Is it safe to assume that a PR for this would be welcome in the near future?

[dpgeorge]

I'm assuming this only applies to the .c and .h files, right?

To everything, the file names and the C identifiers (function names etc) in those files.

The ESP32 has two SPI "hosts" (as they're called) that can be accessed via the IDF: HSPI and VSPI.

In that case the uPy bindings should not be called hspi because it would be confusing. The uPy bindings apply to both the HSPI and VSPI (I assume). Calling it machine_hw_spi sounds good.

Is it safe to assume that a PR for this would be welcome in the near future?

Yes!

[MrSurly]

Okay, the ESP-IDF implementation of HW SPI appears to have some issues: SPI: pin routing remains after disabling SPI, and spi_device_transmit doesn't return when changing host ID.

I'm hoping the ESP IDF team is responsive.

These two issues are probably blockers; HW SPI works fine as long as you configure it exactly once, which is probably OK for many applications.

[MrSurly]

PR #63