atsamd5x: add SPI.TxN()

[sago35]

I created SPI.TxN() to repeat SPI.Tx() n times.

// pseudo code
SPI.TxN(tx, rx []byte, n int) {
    for i := 0; i < n; i++ {
        SPI.Tx(tx, rx)
    }
}

SPI.Tx() is not enough to run ili9341 on standard SPI for speed reasons.
In particular, SPI.TxN() is faster than SPI.Tx() in the following applications.

• Fill with a specific color in ili9341
  • li9341: change to use the standard SPI drivers#242

SPI.Tx() is slower because it waits for the SPI to finish communication for each function call.

testcode:

package main

import (
	"fmt"
	"machine"
	"time"
)

func main() {
	//led := machine.LED
	//led.Configure(machine.PinConfig{Mode: machine.PinOutput})

	spi := machine.SPI0
	spi.Configure(machine.SPIConfig{
		SCK:       machine.SPI0_SCK_PIN,
		SDO:       machine.SPI0_SDO_PIN,
		SDI:       machine.SPI0_SDI_PIN,
		Frequency: 48000000,
	})

	txbuf := make([]byte, 4)
	rxbuf := make([]byte, 4)

	for i := range txbuf {
		txbuf[i] = byte(i * 2)
		rxbuf[i] = byte(i * 2)
	}

	testNo := 0
	for {
		//led.Toggle()
		time.Sleep(time.Millisecond * 500)

		switch testNo % 4 {
		case 0:
			// txrx
			spi.TxN(txbuf, rxbuf, 16)

			for i := range txbuf {
				if txbuf[i] != rxbuf[i] {
					fmt.Printf("%d : got %02X want %02X\r\n", i, rxbuf[i], txbuf[i])
				} else {
					fmt.Printf(".")
				}
				rxbuf[i] = 0
			}
		case 1:
			// tx
			spi.TxN(txbuf, nil, 16)
		case 2:
			// rx
			spi.TxN(nil, rxbuf, 16)
		case 4:
			// skip
		}

		testNo++
	}
}

The speed is about the same compared to #1453.

txrxn:

txn:

rxn:

[deadprogram]

@sago35 is this PR something you are still working on, or should we close it?

[sago35]

@deadprogram
At this point, it would be better to adopt #3985. I will close this PR.

[deadprogram]

Thanks for the reply @sago35 now closing.