#include <avr/io.h>
#include <stdio.h>
#include <stdlib.h>
#include <stdbool.h>
#include <util/delay.h>
#include <avr/interrupt.h>
#include <avr/eeprom.h>

#define asm(x) __asm__ volatile(x)

uint8_t leds[9*3] = { 0 };

#if 0
const uint8_t fleds[9*3] = {
	0xbb, 0x00, 0x77, 
	0xbb, 0x77, 0x00, 
	0x77, 0xbb, 0x00, 
	0x00, 0xbb, 0x00, 
	0x00, 0xbb, 0x77, 
	0x00, 0x77, 0xbb, 
	0x00, 0x00, 0xbb, 
	0x77, 0x00, 0xbb, 
};
#else
const uint8_t fleds[9*3] = {
	0xbb, 0x00, 0x00,
	0xbb, 0x77, 0x00,
	0x77, 0xbb, 0x00,
	0x00, 0xbb, 0x00,
	0x00, 0xbb, 0x77,
	0x00, 0x77, 0xbb,
	0x00, 0x00, 0xbb,
	0x77, 0x00, 0xbb,
	0xbb, 0x00, 0x77,
};
#endif

int main(void)
{
	char j;
	uint8_t x, *p;
	uint8_t on, off;
	uint8_t end = (uint8_t)(uint16_t)(leds + sizeof(leds));
	uint8_t fade = 0;
	uint8_t counter = 0;
	uint8_t skip = 0;

	on = 1;
	off = 0;
	DDRB = on;

	for (;;) {
		counter++;
		if (++fade >= 0x80) {
			skip+= 3;
			if (skip > sizeof(fleds))
				skip-= sizeof(fleds);
			fade = 0;
		}
		for (x = 0;x != sizeof(leds);x++)
			leds[x] = (((fleds[(x + skip) % sizeof(fleds)] * (0x80 - fade)) >> 8) +
				   (fleds[(x + skip + 3) % sizeof(fleds)] * fade >> 8) +
				   (~counter & 7)) / 8;

		p = leds;
		_delay_us(50);

		for (;;) {
			x = *p++;
			for (j = 7;j >= 0;j--) {
				PORTB = on;
				asm("nop");
				asm("rjmp .+0");
				if (!(x & (uint8_t)(1 << j)))
					PORTB = off;
				asm("rjmp .+0");
				asm("rjmp .+0");
				asm("nop");
				PORTB = off;
				if (j != 0) {
					asm("rjmp .+0");
					asm("rjmp .+0");
				}
			}
			if ((uint8_t)(uint16_t)p == end)
				break;
		}
	}
}