drivers: serial: stm32: add err_check

[GeorgeCGV]

Adds 'err_check' implementation for STM32 serial driver.

[codecov-io]

Codecov Report

Merging #13152 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master   #13152   +/-   ##
=======================================
  Coverage   51.96%   51.96%           
=======================================
  Files         309      309           
  Lines       45576    45576           
  Branches    10554    10554           
=======================================
  Hits        23683    23683           
  Misses      17083    17083           
  Partials     4810     4810

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5a56ee5...c4b8d63. Read the comment docs.

[GeorgeCGV]

Regarding this PR. It checks for erros consequentially and sets specific bits specified by the uart_rx_stop_reason. This contradicts will all other uart api implementations of the error check function.

Majority of implementations check for error sequentially but with one mistake...
They treat uart_rx_stop_reason values as corresponding bits.

e.g. gecko

	int err = 0;

	if (flags & USART_IF_RXOF) {
		err |= UART_ERROR_OVERRUN;
	}

	if (flags & USART_IF_PERR) {
		err |= UART_ERROR_PARITY;
	}

	if (flags & USART_IF_FERR) {
		err |= UART_ERROR_FRAMING;
	}

but because uart_rx_stop_reason values are simple integers the error won't be accumulative.
At the end we will have the last error reported and all flags cleared.

This looks like a mistake.. As either uart_rx_stop_reason values shall represent bits (e.g. UART_BREAK = 1 << 1, UART_ERROR_OVERRUN = 1 << 2) or err_check shall return as soon as it detects one of the flags and clears it, but this will affect the user as err_check would have to be called multiple times until all errors are cleared.

[GeorgeCGV]

Pull request #10820 brakes previously defined UART error bitmasks, which are now within uart_rx_stop_reason enum as simple integers.

[erwango]

Pull request #10820 brakes previously defined UART error bitmasks, which are now within uart_rx_stop_reason enum as simple integers.

Raised #13256 for this issue.
Let's see what will be the feedback before going further.

[erwango]

Minor comment