Call spi_write_blocking directly for RP2040

This is around a 6-13x speed up in graphicstest (at 62.5Mhz)

Adafruit_SPITFT.cpp

@@ -1326,7 +1326,13 @@ void Adafruit_SPITFT::writeColor(uint16_t color, uint32_t len) {
         hwspi._spi->write(lo);
       }
     } while (len);
-#else // !ESP8266
+#elif defined(ARDUINO_ARCH_RP2040)
+    spi_inst_t *pi_spi = hwspi._spi == &SPI ? spi0 : spi1;
+    color = __builtin_bswap16(color);
+
+    while (len--)
+      spi_write_blocking(pi_spi, (uint8_t *)&color, 2);
+#else // !ESP8266 && !ARDUINO_ARCH_RP2040
     while (len--) {
 #if defined(__AVR__)
       AVR_WRITESPI(hi);
@@ -2079,6 +2085,9 @@ void Adafruit_SPITFT::spiWrite(uint8_t b) {
     AVR_WRITESPI(b);
 #elif defined(ESP8266) || defined(ESP32)
     hwspi._spi->write(b);
+#elif defined(ARDUINO_ARCH_RP2040)
+    spi_inst_t *pi_spi = hwspi._spi == &SPI ? spi0 : spi1;
+    spi_write_blocking(pi_spi, &b, 1);
 #else
     hwspi._spi->transfer(b);
 #endif
@@ -2385,6 +2394,10 @@ void Adafruit_SPITFT::SPI_WRITE16(uint16_t w) {
     AVR_WRITESPI(w);
 #elif defined(ESP8266) || defined(ESP32)
     hwspi._spi->write16(w);
+#elif defined(ARDUINO_ARCH_RP2040)
+    spi_inst_t *pi_spi = hwspi._spi == &SPI ? spi0 : spi1;
+    w = __builtin_bswap16(w);
+    spi_write_blocking(pi_spi, (uint8_t *)&w, 2);
 #else
     // MSB, LSB because TFTs are generally big-endian
     hwspi._spi->transfer(w >> 8);
@@ -2437,6 +2450,10 @@ void Adafruit_SPITFT::SPI_WRITE32(uint32_t l) {
     AVR_WRITESPI(l);
 #elif defined(ESP8266) || defined(ESP32)
     hwspi._spi->write32(l);
+#elif defined(ARDUINO_ARCH_RP2040)
+    spi_inst_t *pi_spi = hwspi._spi == &SPI ? spi0 : spi1;
+    l = __builtin_bswap32(l);
+    spi_write_blocking(pi_spi, (uint8_t *)&l, 4);
 #else
     hwspi._spi->transfer(l >> 24);
     hwspi._spi->transfer(l >> 16);