modem: backend: uart: Add hw-flow-control for UART

[MarkusLassila]

Add Asynchronous UART implementation, which does not drop data when automatic hardware-flow-control is set in the device tree.

With automatic hardware flow control, the CTS pin will be automatically deactivated when there are no more asynchronous UART RX buffers available. After buffer space becomes available, and UART RX is restarted, the CTS pin will be activated.

[MarkusLassila]

Ping @SeppoTakalo

[bjarki-andreasen]

Could you try making the receive buffers smaller in the uart backend test suite

zephyr/tests/subsys/modem/backends/uart/src/main.c

Line 39 in 044a94f

static uint8_t backend_receive_buffer[4096];

and add a sleep here

zephyr/tests/subsys/modem/backends/uart/src/main.c

Line 137 in 044a94f

ret = modem_pipe_receive(pipe, receive_buffer, sizeof(receive_buffer));

to try to invoke the need for flowcontrol?

I have a suspicion bytes may be lost by disabling RX outright rather than using flow control (as disable may flush the rx buffer) :)

[SeppoTakalo]

I tried this a little bit.

Using nRF9151, I'm now able to reliably use 1 Mbit/s baud rate on the serial port without losing data on serial buffers.
So this kind of implementation significantly improves the upload performance.

[MarkusLassila]

@bjarki-andreasen, sorry about bit of a delay.

Currently the test does not actually test flow control and I don't think that I can change it as I do not believe that there are devices in the CI infrastructure that would have the suitable wiring. I tested this manually in MarkusLassila#1.

With hw-flow-control enabled, no data is dropped. The RX buffers are not flushed when UART is disabled.

[bjarki-andreasen]

Please cherry pick updated test from #89250 and add a test case which enables the new CONFIG_TEST_HW_FLOW_CONTROL option.

[MarkusLassila]

@bjarki-andreasen: For whatever the reason, it does not seem that the CI ran the modem.backends tests.

[bjarki-andreasen]

ping @SeppoTakalo @tomi-font

[tomi-font]

Reviewed all but modem_backend_uart_async_hwfc.c. Looks good overall, just some minor remarks.

[MarkusLassila]

@tomi-font, @SeppoTakalo: Seems that a minor change dismissed your reviews. Apologies.