examples/simple/simple.ino

@@ -38,13 +38,6 @@ supported boards. Notes have been moved to the bottom of the code.
   uint8_t clockPin   = 13;
   uint8_t latchPin   = 0;
   uint8_t oePin      = 1;
-#elif defined(ARDUINO_ADAFRUIT_FEATHER_ESP32C6) // Feather ESP32-C6
-  // not featherwing compatible, but can 'hand wire' if desired
-  uint8_t rgbPins[]  = {6, A3, A1, A0, A2, 0};
-  uint8_t addrPins[] = {8, 5, 15, 7};
-  uint8_t clockPin   = 14;
-  uint8_t latchPin   = RX;
-  uint8_t oePin      = TX;
 #elif defined(ARDUINO_ADAFRUIT_FEATHER_ESP32S2) // Feather ESP32-S2
   // M0/M4/RP2040 Matrix FeatherWing compatible:
   uint8_t rgbPins[]  = {6, 5, 9, 11, 10, 12};
@@ -66,11 +59,11 @@ supported boards. Notes have been moved to the bottom of the code.
   uint8_t latchPin   = 0;
   uint8_t oePin      = 1;
 #elif defined(_SAMD21_) // Feather M0 variants
-  uint8_t rgbPins[]  = {6, 5, 9, 11, 10, 12};
+  uint8_t rgbPins[]  = {6, 7, 10, 11, 12, 13};
-  uint8_t addrPins[] = {A5, A4, A3, A2};
+  uint8_t addrPins[] = {0, 1, 2, 3};
-  uint8_t clockPin   = 13;
+  uint8_t clockPin   = SDA;
-  uint8_t latchPin   = 0;
+  uint8_t latchPin   = 4;
-  uint8_t oePin      = 1;
+  uint8_t oePin      = 5;
 #elif defined(NRF52_SERIES) // Special nRF52840 FeatherWing pinout
   uint8_t rgbPins[]  = {6, A5, A1, A0, A4, 11};
   uint8_t addrPins[] = {10, 5, 13, 9};

library.properties

@@ -1,5 +1,5 @@
 name=Adafruit Protomatter
-version=1.7.0
+version=1.6.2
 author=Adafruit
 maintainer=Adafruit <info@adafruit.com>
 sentence=A library for Adafruit RGB LED matrices.

src/arch/esp32-c3.h

@@ -29,7 +29,7 @@
 #define _PM_portSetRegister(pin) (volatile uint32_t *)&GPIO.out_w1ts
 #define _PM_portClearRegister(pin) (volatile uint32_t *)&GPIO.out_w1tc
 
-#define _PM_portBitMask(pin) (1U << ((pin) & 31))
+#define _PM_portBitMask(pin) (1U << ((pin)&31))
 
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
 #define _PM_byteOffset(pin) ((pin & 31) / 8)

src/arch/esp32-c6.h

@@ -29,7 +29,7 @@
 #define _PM_portSetRegister(pin) (volatile uint32_t *)&GPIO.out_w1ts
 #define _PM_portClearRegister(pin) (volatile uint32_t *)&GPIO.out_w1tc
 
-#define _PM_portBitMask(pin) (1U << ((pin) & 31))
+#define _PM_portBitMask(pin) (1U << ((pin)&31))
 
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
 #define _PM_byteOffset(pin) ((pin & 31) / 8)

src/arch/esp32-s3.h

@@ -129,10 +129,12 @@ IRAM_ATTR static void blast_long(Protomatter_core *core, uint32_t *data) {}
 
 static void pinmux(int8_t pin, uint8_t signal) {
   esp_rom_gpio_connect_out_signal(pin, signal, false, false);
-  PIN_FUNC_SELECT(GPIO_PIN_MUX_REG[pin], PIN_FUNC_GPIO);
+  gpio_hal_iomux_func_sel(GPIO_PIN_MUX_REG[pin], PIN_FUNC_GPIO);
   gpio_set_drive_capability((gpio_num_t)pin, GPIO_DRIVE_STRENGTH);
 }
 
+#if defined(ARDUINO) // COMPILING FOR ARDUINO ------------------------------
+
 // LCD_CAM requires a complete replacement of the "blast" functions in order
 // to use the DMA-based peripheral.
 #define _PM_CUSTOM_BLAST // Disable blast_*() functions in core.c
@@ -160,23 +162,11 @@ IRAM_ATTR static void blast_byte(Protomatter_core *core, uint8_t *data) {
 
   // Timer was cleared to 0 before calling blast_byte(), so this
   // is the state of the timer immediately after DMA started:
-#if defined(ARDUINO)
   dmaSetupTime = (uint32_t)timerRead((hw_timer_t *)core->timer);
-#elif defined(CIRCUITPY)
-  uint64_t value;
-#if (ESP_IDF_VERSION_MAJOR == 5)
-  gptimer_handle_t timer = (gptimer_handle_t)core->timer;
-  gptimer_get_raw_count(timer, &value);
-#else
-  timer_index_t *timer = (timer_index_t *)core->timer;
-  timer_get_counter_value(timer->group, timer->idx, &value);
-#endif
-  dmaSetupTime = (uint32_t)value;
-#endif
   // See notes near top of this file for what's done with this info.
 }
 
-static void _PM_timerInit(Protomatter_core *core) {
+void _PM_timerInit(Protomatter_core *core) {
   // On S3, initialize the LCD_CAM peripheral and DMA.
 
   // LCD_CAM isn't enabled by default -- MUST begin with this:
@@ -252,9 +242,157 @@ static void _PM_timerInit(Protomatter_core *core) {
   desc.next = NULL;
 
   // Alloc DMA channel & connect it to LCD periph
-#if defined(CIRCUITPY)
+  gdma_channel_alloc_config_t dma_chan_config = {
-  if (dma_chan == NULL) {
+      .sibling_chan = NULL,
+      .direction = GDMA_CHANNEL_DIRECTION_TX,
+      .flags = {.reserve_sibling = 0}};
+  esp_err_t ret = gdma_new_channel(&dma_chan_config, &dma_chan);
+  (void)ret;
+  gdma_connect(dma_chan, GDMA_MAKE_TRIGGER(GDMA_TRIG_PERIPH_LCD, 0));
+  gdma_strategy_config_t strategy_config = {.owner_check = false,
+                                            .auto_update_desc = false};
+  gdma_apply_strategy(dma_chan, &strategy_config);
+  gdma_transfer_ability_t ability = {
+      .sram_trans_align = 0,
+      .psram_trans_align = 0,
+  };
+  gdma_set_transfer_ability(dma_chan, &ability);
+  gdma_start(dma_chan, (intptr_t)&desc);
+
+  // Enable TRANS_DONE interrupt. Note that we do NOT require nor install
+  // an interrupt service routine, but DO need to enable the TRANS_DONE
+  // flag to make the LCD DMA transfer work.
+  LCD_CAM.lc_dma_int_ena.val |= LCD_LL_EVENT_TRANS_DONE & 0x03;
+
+  _PM_esp32commonTimerInit(core); // In esp32-common.h
+}
+
+#elif defined(CIRCUITPY) // COMPILING FOR CIRCUITPYTHON --------------------
+
+// LCD_CAM requires a complete replacement of the "blast" functions in order
+// to use the DMA-based peripheral.
+#define _PM_CUSTOM_BLAST // Disable blast_*() functions in core.c
+IRAM_ATTR static void blast_byte(Protomatter_core *core, uint8_t *data) {
+  // Reset LCD DOUT parameters each time (required).
+  // IN PRINCIPLE, cyclelen should be chainBits-1 (resulting in chainBits
+  // cycles). But due to the required dummy phases at start of transfer,
+  // extend by 1; set to chainBits, issue chainBits+1 cycles.
+  LCD_CAM.lcd_user.lcd_dout_cyclelen = core->chainBits;
+  LCD_CAM.lcd_user.lcd_dout = 1;
+  LCD_CAM.lcd_user.lcd_update = 1;
+
+  // Reset LCD TX FIFO each time, else we see old data. When doing this,
+  // it's REQUIRED in the setup code to enable at least one dummy pulse,
+  // else the PCLK & data are randomly misaligned by 1-2 clocks!
+  LCD_CAM.lcd_misc.lcd_afifo_reset = 1;
+
+  // Partially re-init descriptor each time (required)
+  desc.dw0.size = desc.dw0.length = core->chainBits;
+  desc.buffer = data;
+  gdma_start(dma_chan, (intptr_t)&desc);
+  esp_rom_delay_us(1); // Necessary before starting xfer
+
+  LCD_CAM.lcd_user.lcd_start = 1; // Begin LCD DMA xfer
+
+  // Timer was cleared to 0 before calling blast_byte(), so this
+  // is the state of the timer immediately after DMA started:
+  uint64_t value;
+
+#if (ESP_IDF_VERSION_MAJOR == 5)
+  gptimer_handle_t timer = (gptimer_handle_t)core->timer;
+  gptimer_get_raw_count(timer, &value);
+#else
+  timer_index_t *timer = (timer_index_t *)core->timer;
+  timer_get_counter_value(timer->group, timer->idx, &value);
 #endif
+  dmaSetupTime = (uint32_t)value;
+  // See notes near top of this file for what's done with this info.
+}
+
+static void _PM_timerInit(Protomatter_core *core) {
+
+  // TO DO: adapt this function for any CircuitPython-specific changes.
+  // If none are required, this function can be deleted and the version
+  // above can be moved before the ARDUIO/CIRCUITPY checks. If minimal
+  // changes, consider a single _PM_timerInit() implementation with
+  // ARDUINO/CIRCUITPY checks inside. It's all good.
+
+  // On S3, initialize the LCD_CAM peripheral and DMA.
+
+  // LCD_CAM isn't enabled by default -- MUST begin with this:
+  periph_module_enable(PERIPH_LCD_CAM_MODULE);
+  periph_module_reset(PERIPH_LCD_CAM_MODULE);
+
+  // Reset LCD bus
+  LCD_CAM.lcd_user.lcd_reset = 1;
+  esp_rom_delay_us(100);
+
+  // Configure LCD clock
+  LCD_CAM.lcd_clock.clk_en = 1;             // Enable clock
+  LCD_CAM.lcd_clock.lcd_clk_sel = 3;        // PLL160M source
+  LCD_CAM.lcd_clock.lcd_clkm_div_a = 1;     // 1/1 fractional divide,
+  LCD_CAM.lcd_clock.lcd_clkm_div_b = 1;     // plus prescale below yields...
+#if LCD_CLK_PRESCALE == 8
+  LCD_CAM.lcd_clock.lcd_clkm_div_num = 7;   // 1:8 prescale (20 MHz CLK)
+#elif LCD_CLK_PRESCALE == 9
+  LCD_CAM.lcd_clock.lcd_clkm_div_num = 8; // 1:9 prescale (17.8 MHz CLK)
+#else
+  LCD_CAM.lcd_clock.lcd_clkm_div_num = 9; // 1:10 prescale (16 MHz CLK)
+#endif
+  LCD_CAM.lcd_clock.lcd_ck_out_edge = 0;    // PCLK low in first half of cycle
+  LCD_CAM.lcd_clock.lcd_ck_idle_edge = 0;   // PCLK low idle
+  LCD_CAM.lcd_clock.lcd_clk_equ_sysclk = 1; // PCLK = CLK (ignore CLKCNT_N)
+
+  // Configure frame format. Some of these could probably be skipped and
+  // use defaults, but being verbose for posterity...
+  LCD_CAM.lcd_ctrl.lcd_rgb_mode_en = 0;    // i8080 mode (not RGB)
+  LCD_CAM.lcd_rgb_yuv.lcd_conv_bypass = 0; // Disable RGB/YUV converter
+  LCD_CAM.lcd_misc.lcd_next_frame_en = 0;  // Do NOT auto-frame
+  LCD_CAM.lcd_data_dout_mode.val = 0;      // No data delays
+  LCD_CAM.lcd_user.lcd_always_out_en = 0;  // Only when requested
+  LCD_CAM.lcd_user.lcd_8bits_order = 0;    // Do not swap bytes
+  LCD_CAM.lcd_user.lcd_bit_order = 0;      // Do not reverse bit order
+  LCD_CAM.lcd_user.lcd_2byte_en = 0;       // 8-bit data mode
+  // MUST enable at least one dummy phase at start of output, else clock and
+  // data are randomly misaligned by 1-2 cycles following required TX FIFO
+  // reset in blast_byte(). One phase MOSTLY works but sparkles a tiny bit
+  // (as in still very occasionally misaligned by 1 cycle). Two seems ideal;
+  // no sparkle. Since HUB75 is just a shift register, the extra clock ticks
+  // are harmless and the zero-data shifts off end of the chain.
+  LCD_CAM.lcd_user.lcd_dummy = 1;          // Enable dummy phase(s) @ LCD start
+  LCD_CAM.lcd_user.lcd_dummy_cyclelen = 1; // 2 dummy phases
+  LCD_CAM.lcd_user.lcd_cmd = 0;            // No command at LCD start
+  LCD_CAM.lcd_user.lcd_cmd_2_cycle_en = 0;
+  LCD_CAM.lcd_user.lcd_update = 1;
+
+  // Configure signal pins. IN THEORY this could be expanded to support
+  // 2 parallel chains, but the rest of the LCD & DMA setup is not currently
+  // written for that, so it's limited to a single chain for now.
+  const uint8_t signal[] = {LCD_DATA_OUT0_IDX, LCD_DATA_OUT1_IDX,
+                            LCD_DATA_OUT2_IDX, LCD_DATA_OUT3_IDX,
+                            LCD_DATA_OUT4_IDX, LCD_DATA_OUT5_IDX};
+  for (int i = 0; i < 6; i++)
+    pinmux(core->rgbPins[i], signal[i]);
+  pinmux(core->clockPin, LCD_PCLK_IDX);
+  gpio_set_drive_capability(core->latch.pin, GPIO_DRIVE_STRENGTH);
+  gpio_set_drive_capability(core->oe.pin, GPIO_DRIVE_STRENGTH);
+  for (uint8_t i = 0; i < core->numAddressLines; i++) {
+    gpio_set_drive_capability(core->addr[i].pin, GPIO_DRIVE_STRENGTH);
+  }
+
+  // Disable LCD_CAM interrupts, clear any pending interrupt
+  LCD_CAM.lc_dma_int_ena.val &= ~LCD_LL_EVENT_TRANS_DONE;
+  LCD_CAM.lc_dma_int_clr.val = 0x03;
+
+  // Set up DMA TX descriptor
+  desc.dw0.owner = DMA_DESCRIPTOR_BUFFER_OWNER_DMA;
+  desc.dw0.suc_eof = 1;
+  desc.dw0.size = desc.dw0.length = core->chainBits;
+  desc.buffer = core->screenData;
+  desc.next = NULL;
+
+  // Alloc DMA channel & connect it to LCD periph
+  if (dma_chan == NULL) {
     gdma_channel_alloc_config_t dma_chan_config = {
         .sibling_chan = NULL,
         .direction = GDMA_CHANNEL_DIRECTION_TX,
@@ -269,9 +407,7 @@ static void _PM_timerInit(Protomatter_core *core) {
         .psram_trans_align = 0,
     };
     gdma_set_transfer_ability(dma_chan, &ability);
-#if defined(CIRCUITPY)
   }
-#endif
   gdma_start(dma_chan, (intptr_t)&desc);
 
   // Enable TRANS_DONE interrupt. Note that we do NOT require nor install
@@ -282,4 +418,6 @@ static void _PM_timerInit(Protomatter_core *core) {
   _PM_esp32commonTimerInit(core); // In esp32-common.h
 }
 
+#endif // END CIRCUITPYTHON ------------------------------------------------
+
 #endif // END ESP32S3

src/arch/esp32.h

@@ -32,7 +32,7 @@
 #define _PM_portClearRegister(pin)                                             \
   (volatile uint32_t *)((pin < 32) ? &GPIO.out_w1tc : &GPIO.out1_w1tc.val)
 
-#define _PM_portBitMask(pin) (1U << ((pin) & 31))
+#define _PM_portBitMask(pin) (1U << ((pin)&31))
 
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
 #define _PM_byteOffset(pin) ((pin & 31) / 8)

src/arch/nrf52.h

@@ -99,7 +99,7 @@ volatile uint32_t *_PM_portClearRegister(uint32_t pin) {
 #define _PM_pinInput(pin) nrf_gpio_cfg_input(pin)
 #define _PM_pinHigh(pin) nrf_gpio_pin_set(pin)
 #define _PM_pinLow(pin) nrf_gpio_pin_clear(pin)
-#define _PM_portBitMask(pin) (1u << ((pin) & 31))
+#define _PM_portBitMask(pin) (1u << ((pin)&31))
 
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
 #define _PM_byteOffset(pin) ((pin & 31) / 8)

src/arch/rp2040.h

@@ -26,8 +26,7 @@
 
 #pragma once
 
-#if defined(ARDUINO_ARCH_RP2040) || defined(PICO_BOARD) ||                     \
+#if defined(ARDUINO_ARCH_RP2040) || defined(PICO_BOARD) || defined(__RP2040__)
-    defined(__RP2040__) || defined(__RP2350__)
 
 #include "../../hardware_pwm/include/hardware/pwm.h"
 #include "hardware/irq.h"
@@ -106,13 +105,8 @@ void _PM_timerInit(Protomatter_core *core) {
 #elif defined(CIRCUITPY) // COMPILING FOR CIRCUITPYTHON --------------------
 
 #if !defined(F_CPU)     // Not sure if CircuitPython build defines this
-#ifdef __RP2040__
 #define F_CPU 125000000 // Standard RP2040 clock speed
 #endif
-#ifdef __RP2350__
-#define F_CPU 150000000 // Standard RP2350 clock speed
-#endif
-#endif
 
 // 'pin' here is GPXX #
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__

src/arch/samd-common.h

@@ -64,7 +64,7 @@ void _PM_IRQ_HANDLER(void) {
 #define _PM_pinInput(pin) gpio_set_pin_direction(pin, GPIO_DIRECTION_IN)
 #define _PM_pinHigh(pin) gpio_set_pin_level(pin, 1)
 #define _PM_pinLow(pin) gpio_set_pin_level(pin, 0)
-#define _PM_portBitMask(pin) (1u << ((pin) & 31))
+#define _PM_portBitMask(pin) (1u << ((pin)&31))
 
 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
 #define _PM_byteOffset(pin) ((pin & 31) / 8)

src/arch/stm32.h

@@ -28,7 +28,7 @@
 #include "timers.h"
 
 #undef _PM_portBitMask
-#define _PM_portBitMask(pin) (1u << ((pin) & 15))
+#define _PM_portBitMask(pin) (1u << ((pin)&15))
 #define _PM_byteOffset(pin) ((pin & 15) / 8)
 #define _PM_wordOffset(pin) ((pin & 15) / 16)
 
@@ -83,7 +83,7 @@ volatile uint16_t *_PM_portClearRegister(uint32_t pin) {
 // TODO: this is no longer true, should it change?
 void *_PM_protoPtr = NULL;
 
-static TIM_HandleTypeDef tim_handle;
+STATIC TIM_HandleTypeDef tim_handle;
 
 // Timer interrupt service routine
 void _PM_IRQ_HANDLER(void) {

src/core.c

@@ -95,9 +95,8 @@ ProtomatterStatus _PM_init(Protomatter_core *core, uint16_t bitWidth,
                            uint8_t addrCount, uint8_t *addrList,
                            uint8_t clockPin, uint8_t latchPin, uint8_t oePin,
                            bool doubleBuffer, int8_t tile, void *timer) {
-  if (!core) {
+  if (!core)
     return PROTOMATTER_ERR_ARG;
-  }
 
     // bitDepth is NOT constrained here, handle in calling function
     // (varies with implementation, e.g. GFX lib is max 6 bitplanes,
@@ -1200,24 +1199,18 @@ void _PM_convert_565_word(Protomatter_core *core, uint16_t *source,
             uint16_t upperRGB = upperSrc[srcIdx]; // Pixel in upper half
             uint16_t lowerRGB = lowerSrc[srcIdx]; // Pixel in lower half
             uint16_t result = 0;
-            if (upperRGB & redBit) {
+            if (upperRGB & redBit)
               result |= pinMask[0];
-            }
+            if (upperRGB & greenBit)
-            if (upperRGB & greenBit) {
               result |= pinMask[1];
-            }
+            if (upperRGB & blueBit)
-            if (upperRGB & blueBit) {
               result |= pinMask[2];
-            }
+            if (lowerRGB & redBit)
-            if (lowerRGB & redBit) {
               result |= pinMask[3];
-            }
+            if (lowerRGB & greenBit)
-            if (lowerRGB & greenBit) {
               result |= pinMask[4];
-            }
+            if (lowerRGB & blueBit)
-            if (lowerRGB & blueBit) {
               result |= pinMask[5];
-            }
               // Main difference here vs byte converter is each chain
               // ORs new bits into place (vs single-pass overwrite).
 // #if defined(_PM_portToggleRegister)
@@ -1330,24 +1323,18 @@ void _PM_convert_565_long(Protomatter_core *core, uint16_t *source,
             uint16_t upperRGB = upperSrc[srcIdx]; // Pixel in upper half
             uint16_t lowerRGB = lowerSrc[srcIdx]; // Pixel in lower half
             uint32_t result = 0;
-            if (upperRGB & redBit) {
+            if (upperRGB & redBit)
               result |= pinMask[0];
-            }
+            if (upperRGB & greenBit)
-            if (upperRGB & greenBit) {
               result |= pinMask[1];
-            }
+            if (upperRGB & blueBit)
-            if (upperRGB & blueBit) {
               result |= pinMask[2];
-            }
+            if (lowerRGB & redBit)
-            if (lowerRGB & redBit) {
               result |= pinMask[3];
-            }
+            if (lowerRGB & greenBit)
-            if (lowerRGB & greenBit) {
               result |= pinMask[4];
-            }
+            if (lowerRGB & blueBit)
-            if (lowerRGB & blueBit) {
               result |= pinMask[5];
-            }
               // Main difference here vs byte converter is each chain
               // ORs new bits into place (vs single-pass overwrite).
 // #if defined(_PM_portToggleRegister)