BitBang enhancements (#286)

src/internal/NeoEspBitBangMethod.h

@@ -185,10 +185,10 @@ typedef NeoEspBitBangMethodBase<NeoEspBitBangSpeedSk6812> NeoEsp8266BitBangSk681
 typedef NeoEspBitBangMethodBase<NeoEspBitBangSpeed800Kbps> NeoEsp8266BitBang800KbpsMethod;
 typedef NeoEspBitBangMethodBase<NeoEspBitBangSpeed400Kbps> NeoEsp8266BitBang400KbpsMethod;
 
-typedef NeoEsp8266BitBangWs2812xMethod NeoEsp32BitBangWs2813Method;
+typedef NeoEsp8266BitBangWs2812xMethod NeoEsp8266BitBangWs2813Method;
-typedef NeoEsp8266BitBang800KbpsMethod NeoEsp32BitBangWs2812Method;
+typedef NeoEsp8266BitBang800KbpsMethod NeoEsp8266BitBangWs2812Method;
-typedef NeoEsp8266BitBangSk6812Method NeoEsp32BitBangLc8812Method;
+typedef NeoEsp8266BitBangSk6812Method NeoEsp8266BitBangLc8812Method;
-typedef NeoEsp8266BitBang400KbpsMethod NeoEsp32BitBangApa106Method;
+typedef NeoEsp8266BitBang400KbpsMethod NeoEsp8266BitBangApa106Method;
 #endif
 
 // ESP bitbang doesn't have defaults and should avoided except for testing

src/internal/NeoPixelEsp.c

@@ -36,7 +36,9 @@ License along with NeoPixel.  If not, see
 #define ICACHE_RAM_ATTR IRAM_ATTR
 #endif
 
-inline uint32_t _getCycleCount()
+static uint32_t _getCycleCount(void) __attribute__((always_inline));
+
+static inline uint32_t _getCycleCount(void)
 {
     uint32_t ccount;
     __asm__ __volatile__("rsr %0,ccount":"=a" (ccount));
@@ -53,29 +55,23 @@ inline uint32_t _getCycleCount()
 void ICACHE_RAM_ATTR bitbang_send_pixels_800(uint8_t* pixels, uint8_t* end, uint8_t pin)
 {
     const uint32_t pinRegister = _BV(pin);
-    uint8_t mask;
+    uint8_t mask = 0x80;
-    uint8_t subpix;
+    uint8_t subpix = *pixels++;
-    uint32_t cyclesStart;
+    uint32_t cyclesStart = 0; // trigger emediately
+	uint32_t cyclesNext = 0;
 
-    // trigger emediately
+    for (;;)
-    cyclesStart = _getCycleCount() - CYCLES_800;
-    do
-    {
-        subpix = *pixels++;
-        for (mask = 0x80; mask != 0; mask >>= 1)
     {
         // do the checks here while we are waiting on time to pass
-            uint32_t cyclesBit = ((subpix & mask)) ? CYCLES_800_T1H : CYCLES_800_T0H;
+		uint32_t cyclesBit = CYCLES_800_T0H;
-            uint32_t cyclesNext = cyclesStart;
+		if (subpix & mask)
+		{
+			cyclesBit = CYCLES_800_T1H;
+		}
 
         // after we have done as much work as needed for this next bit
         // now wait for the HIGH
-            do
+		while (((cyclesStart = _getCycleCount()) - cyclesNext) < CYCLES_800);
-            {
-                // cache and use this count so we don't incur another 
-                // instruction before we turn the bit high
-                cyclesStart = _getCycleCount();
-            } while ((cyclesStart - cyclesNext) < CYCLES_800);
 
         // set high
 #if defined(ARDUINO_ARCH_ESP32)
@@ -85,10 +81,7 @@ void ICACHE_RAM_ATTR bitbang_send_pixels_800(uint8_t* pixels, uint8_t* end, uint
 #endif
 		
         // wait for the LOW
-            do
+		while ((_getCycleCount() - cyclesStart) < cyclesBit);
-            {
-                cyclesNext = _getCycleCount();
-            } while ((cyclesNext - cyclesStart) < cyclesBit);
 		
         // set low
 #if defined(ARDUINO_ARCH_ESP32)
@@ -96,36 +89,48 @@ void ICACHE_RAM_ATTR bitbang_send_pixels_800(uint8_t* pixels, uint8_t* end, uint
 #else
         GPIO_REG_WRITE(GPIO_OUT_W1TC_ADDRESS, pinRegister);
 #endif
+		cyclesNext = cyclesStart;
+
+		// next bit
+		mask >>= 1;
+		if (mask == 0) 
+		{
+			// no more bits to send in this byte
+			// check for another byte
+			if (pixels >= end) 
+			{
+				// no more bytes to send so stop
+				break;
+			}
+			// reset mask to first bit and get the next byte
+			mask = 0x80;
+			subpix = *pixels++;
+		}
     } 
-    } while (pixels < end);
 }
 
 void ICACHE_RAM_ATTR bitbang_send_pixels_400(uint8_t* pixels, uint8_t* end, uint8_t pin)
 {
 	const uint32_t pinRegister = _BV(pin);
-    uint8_t mask;
+	uint8_t mask = 0x80;
-    uint8_t subpix;
+	uint8_t subpix = *pixels++;
-    uint32_t cyclesStart;
+	uint32_t cyclesStart = 0; // trigger emediately
+	uint32_t cyclesNext = 0;
 
-    // trigger emediately
+	for (;;)
-    cyclesStart = _getCycleCount() - CYCLES_400;
-    do
 	{
-        subpix = *pixels++;
+		// do the checks here while we are waiting on time to pass
-        for (mask = 0x80; mask; mask >>= 1)
+		uint32_t cyclesBit = CYCLES_400_T0H;
+		if (subpix & mask)
 		{
-            uint32_t cyclesBit = ((subpix & mask)) ? CYCLES_400_T1H : CYCLES_400_T0H;
+			cyclesBit = CYCLES_400_T1H;
-            uint32_t cyclesNext = cyclesStart;
+		}
 
 		// after we have done as much work as needed for this next bit
 		// now wait for the HIGH
-            do
+		while (((cyclesStart = _getCycleCount()) - cyclesNext) < CYCLES_400);
-            {
-                // cache and use this count so we don't incur another 
-                // instruction before we turn the bit high
-                cyclesStart = _getCycleCount();
-            } while ((cyclesStart - cyclesNext) < CYCLES_400);
 
+		// set high
 #if defined(ARDUINO_ARCH_ESP32)
 		GPIO.out_w1ts = pinRegister;
 #else
@@ -133,10 +138,7 @@ void ICACHE_RAM_ATTR bitbang_send_pixels_400(uint8_t* pixels, uint8_t* end, uint
 #endif
 
 		// wait for the LOW
-            do
+		while ((_getCycleCount() - cyclesStart) < cyclesBit);
-            {
-                cyclesNext = _getCycleCount();
-            } while ((cyclesNext - cyclesStart) < cyclesBit);
 
 		// set low
 #if defined(ARDUINO_ARCH_ESP32)
@@ -144,8 +146,24 @@ void ICACHE_RAM_ATTR bitbang_send_pixels_400(uint8_t* pixels, uint8_t* end, uint
 #else
 		GPIO_REG_WRITE(GPIO_OUT_W1TC_ADDRESS, pinRegister);
 #endif
+		cyclesNext = cyclesStart;
+
+		// next bit
+		mask >>= 1;
+		if (mask == 0)
+		{
+			// no more bits to send in this byte
+			// check for another byte
+			if (pixels >= end)
+			{
+				// no more bytes to send so stop
+				break;
+			}
+			// reset mask to first bit and get the next byte
+			mask = 0x80;
+			subpix = *pixels++;
+		}
 	}
-    } while (pixels < end);
 }
 
 #endif