[dhalbert]

Board definition and testing by Claude. This note by @dhalbert.

Add adafruit_macropad_rp2040_zephyr board.

• Pin names same as the ports/raspberrypi board.
• Partition layout is from Add MacroPad RP2040 layout adaboot#11, same as Feather RP2040
• The upstream Zephyr board definition added fixed keys (gpio-keys) and encoder (gpio-qdec) nodes. These were disabled so keypad and rotaryio can claim their pins.
• Added CONFIG_DISPLAY=y to turn on the onboard display board.DISPLAY.
• The upstream Zephyr board definition puts the Zephyr console on uart1, which uses the STEMMA QT pins. The console is removed by default, as on the nRF52840 boards, and i2c0 is enabled as board.I2C() and board.STEMMA_I2C. Thanks @mikeysklar for the patch.
• A CONSOLE=1 or DEBUG=1 build puts the Zephyr console back on the STEMMA QT pins (TX on SDA, RX on SCL), without board.I2C() or board.STEMMA_I2C.

Other:

• I sorted board_aliases.cmake by board name for readability.
• The adaboot update also brings in Set up Raspberry Pi bootloader reboots once per chip, not per board adaboot#10, which applies RP2 boot-mode retention per SoC instead of per board.

Testing

On a MacroPad:

• CIRCUITPY works and is the right size
• on_next_reset(RunMode.UF2) resets into RPI-RP2.
• board.DISPLAY shows the REPL and displays a simple displayio test pattern.
• board.STEMMA_I2C() finds an MSA301 at 0x26, and is the same bus as board.I2C().
• With CONSOLE=1, the Zephyr boot banner and log output appear on an FTDI cable connected to the STEMMA QT pins.

These tests were done with #11649 (not yet merged) temporarily merged in:

• rotaryio gives one count per detent.
• keypad.Keys on all 12 keys and on the encoder switch works as expected.

[mikeysklar]

Tested on a MacroPad with #11649 merged in.

Keypad on all 12 keys, encoder switch and rotary all work.

I2C on STEMMA QT and USB serial both don't work as submitted.

The upstream dts puts the Zephyr console on uart1 at GPIO20/21.

Doing what the ItsyBitsy does fixes both: board.I2C() then finds a QT gamepad at 0x49, same as 10.3.1.

Could you use this instead?

--- a/ports/zephyr-cp/boards/adafruit/macropad_rp2040_zephyr/board.conf
+++ b/ports/zephyr-cp/boards/adafruit/macropad_rp2040_zephyr/board.conf
@@ -5,3 +5,7 @@ CONFIG_IOBROKER_PACKAGE_RP2040_QFN56=y
 
 # Enable the SH1106 OLED via the Zephyr display subsystem
 CONFIG_DISPLAY=y
+
+# No Zephyr UART console (see board.overlay), so SCL/SDA are free for I2C.
+CONFIG_UART_CONSOLE=n
+CONFIG_LOG_BACKEND_UART=n
--- a/ports/zephyr-cp/boards/adafruit/macropad_rp2040_zephyr/board.overlay
+++ b/ports/zephyr-cp/boards/adafruit/macropad_rp2040_zephyr/board.overlay
@@ -12,3 +12,19 @@
 		status = "disabled";
 	};
 };
+
+/* No Zephyr console, so SCL/SDA (GPIO21/20) are free for I2C0. */
+/ {
+	chosen {
+		/delete-property/ zephyr,console;
+		/delete-property/ zephyr,shell-uart;
+	};
+};
+
+&uart1 {
+	status = "disabled";
+};
+
+&i2c0 {
+	status = "okay";
+};

Also:

• board.STEMMA_I2C is missing.
• NeoPixels don't work on this board yet: neopixel_write on zephyr-cp is only implemented for nRF. RP2 support should be added using PIO, as a follow-up PR.

[dhalbert]

I2C should be fixed. I tested scanning an MSA301. If you compile with CONSOLE=1, the STEMMA port becomes a console port, as described in console.overlay.

NeoPixel support is To Be Done, as we probably want to use PIO, and we need to do PIO management.

Thanks for testing!

[mikeysklar]

Sounds good. Auto-merge enabled.

[mikeysklar]

LGTM