WS2812 Overhaul (#210)

* ARM - ws2812 bitbang (#7173)

* Initial ARM bitbang ws2812 driver

* Unify chibios platform to run rgblight_task

* Remove 'avr only' comments from ws2812 docs

* Remove 'avr only' comments from ws2812 docs

* Unify chibios platform to run rgblight_task - review comments

* Remove debug flags from keymap

* Add comments from review

* Add defines for STM32L0XX

* Attempt to get arm ws2812 working on multiple gcc versions

* Support RGBLIGHT_SLEEP when ChibiOS boards suspend (#7280)

Copypasta from the AVR suspend implementation with a Teensy-specific
hack removed

* Unify RGB and RGBW commands (#7297)

* Fix unicode in comments

Co-Authored-By: fauxpark <fauxpark@gmail.com>

* Remove separate RGBW implementation for a unified function

* Set White to 0 in RGBW LEDs

This is just to get this working, later, proper brightness can be handled elsewhere.

* Use us instead of nanoseconds(?) since it renders correctly on web

* Remove RGBW function from arm/ws2812.h

* Remove RGBW function from arm/ws2812.c

* Formatting changes

* Add doc info

* Remove force of debug on within rgblight - causes lockups waiting for hid_listen (#7330)

* Move Ergodox EZ RGB Light code to custom driver  (#7309)

* Move Ergodox EZ RGB code to custom driver

Also implements full addressing of Ergodox EZ's LED Strip, as written by seebs
Co-authored-by: Seebs <seebs@seebs.net>

* Make Clipping range accessible for custom drivers

* Remove RGBW_BB_TWI from driver and docs

* Revert changes to clipping range support

* Use just rgblight_set instead of full custom driver

* Convert to i2c_master commands

* Rename rgblight driver and clean up includes

* Use White channel on RGBW LEDs

* SPI DMA based RGB Underglow for STM32 (#7674)

* Initial stash of ws2812 spi driver

* Update comment, add sync backup plan

* Add testing notes to spi ws2812 driver

* Align RGBW error messages

Co-authored-by: Joel Challis <git@zvecr.com>
Co-authored-by: Jonathan Rascher <jon@bcat.name>
Co-authored-by: Florian Didron <fdidron@users.noreply.github.com>

1 files changed, 95 insertions, 1 deletions

diff --git a/drivers/arm/ws2812.c b/drivers/arm/ws2812.c
index 2094e50098..fa702fca98 100644
--- a/drivers/arm/ws2812.c
+++ b/drivers/arm/ws2812.c
@@ -1 +1,95 @@
-#error("NOT SUPPORTED")
\ No newline at end of file
+#include "quantum.h"
+#include "ws2812.h"
+#include "ch.h"
+#include "hal.h"
+
+/* Adapted from https://github.com/bigjosh/SimpleNeoPixelDemo/ */
+
+#ifndef NOP_FUDGE
+#    if defined(STM32F1XX) || defined(STM32F1xx) || defined(STM32F0XX) || defined(STM32F0xx) || defined(STM32F3XX) || defined(STM32F3xx) || defined(STM32L0XX) || defined(STM32L0xx)
+#        define NOP_FUDGE 0.4
+#    else
+#        error("NOP_FUDGE configuration required")
+#        define NOP_FUDGE 1  // this just pleases the compile so the above error is easier to spot
+#    endif
+#endif
+
+#define NUMBER_NOPS 6
+#define CYCLES_PER_SEC (STM32_SYSCLK / NUMBER_NOPS * NOP_FUDGE)
+#define NS_PER_SEC (1000000000L)  // Note that this has to be SIGNED since we want to be able to check for negative values of derivatives
+#define NS_PER_CYCLE (NS_PER_SEC / CYCLES_PER_SEC)
+#define NS_TO_CYCLES(n) ((n) / NS_PER_CYCLE)
+
+#define wait_ns(x)                                  \
+    do {                                            \
+        for (int i = 0; i < NS_TO_CYCLES(x); i++) { \
+            __asm__ volatile("nop\n\t"              \
+                             "nop\n\t"              \
+                             "nop\n\t"              \
+                             "nop\n\t"              \
+                             "nop\n\t"              \
+                             "nop\n\t");            \
+        }                                           \
+    } while (0)
+
+// These are the timing constraints taken mostly from the WS2812 datasheets
+// These are chosen to be conservative and avoid problems rather than for maximum throughput
+
+#define T1H 900           // Width of a 1 bit in ns
+#define T1L (1250 - T1H)  // Width of a 1 bit in ns
+
+#define T0H 350           // Width of a 0 bit in ns
+#define T0L (1250 - T0H)  // Width of a 0 bit in ns
+
+// The reset gap can be 6000 ns, but depending on the LED strip it may have to be increased
+// to values like 600000 ns. If it is too small, the pixels will show nothing most of the time.
+#define RES 10000  // Width of the low gap between bits to cause a frame to latch
+
+void sendByte(uint8_t byte) {
+    // WS2812 protocol wants most significant bits first
+    for (unsigned char bit = 0; bit < 8; bit++) {
+        bool is_one = byte & (1 << (7 - bit));
+        // using something like wait_ns(is_one ? T1L : T0L) here throws off timings
+        if (is_one) {
+            // 1
+            writePinHigh(RGB_DI_PIN);
+            wait_ns(T1H);
+            writePinLow(RGB_DI_PIN);
+            wait_ns(T1L);
+        } else {
+            // 0
+            writePinHigh(RGB_DI_PIN);
+            wait_ns(T0H);
+            writePinLow(RGB_DI_PIN);
+            wait_ns(T0L);
+        }
+    }
+}
+
+void ws2812_init(void) { setPinOutput(RGB_DI_PIN); }
+
+// Setleds for standard RGB
+void ws2812_setleds(LED_TYPE *ledarray, uint16_t leds) {
+    static bool s_init = false;
+    if (!s_init) {
+        ws2812_init();
+        s_init = true;
+    }
+
+    // this code is very time dependent, so we need to disable interrupts
+    chSysLock();
+
+    for (uint8_t i = 0; i < leds; i++) {
+        // WS2812 protocol dictates grb order
+        sendByte(ledarray[i].g);
+        sendByte(ledarray[i].r);
+        sendByte(ledarray[i].b);
+#ifdef RGBW
+        sendByte(ledarray[i].w);
+#endif
+    }
+
+    wait_ns(RES);
+
+    chSysUnlock();
+}