CRC8 w C: Błąd w algorytmie transmisji SPI z wielomianem x8 + x2 + x + 1

    unsigned char crc8_temp = 0;
    unsigned char CRC8_TABLE[256];
    
    /*Obliczanie tablicy sumy kontrolnej CRC8*/
    unsigned char r; 
    int i, j; 
    for (i = 0; i < 256; i++) 
    { 
             r = (unsigned char) i; 
             for (j = 0; j < 8; j++) 
                 r = (unsigned char) ((r << 1) ^ ((r & 0x80) ? 0x07 : 0)); 
             CRC8_TABLE[i] = r; 
    }
    /*suma CRC8 z 3 bajtow*/
	crc8_temp = CRC8_TABLE[(crc8_temp ^ dat2) & 0xff];
	crc8_temp = CRC8_TABLE[(crc8_temp ^ dat1) & 0xff];
	crc8_temp = CRC8_TABLE[(crc8_temp ^ dat0) & 0xff];

unsigned char CRC8_TABLE[256] = {
	    0x00,0x07,0x0E,0x09,0x1C,0x1B,0x12,0x15,
	    0x38,0x3F,0x36,0x31,0x24,0x23,0x2A,0x2D,
	    0x70,0x77,0x7E,0x79,0x6C,0x6B,0x62,0x65,
	    0x48,0x4F,0x46,0x41,0x54,0x53,0x5A,0x5D,
	    0xE0,0xE7,0xEE,0xE9,0xFC,0xFB,0xF2,0xF5,
	    0xD8,0xDF,0xD6,0xD1,0xC4,0xC3,0xCA,0xCD,
	    0x90,0x97,0x9E,0x99,0x8C,0x8B,0x82,0x85,
	    0xA8,0xAF,0xA6,0xA1,0xB4,0xB3,0xBA,0xBD,
	    0xC7,0xC0,0xC9,0xCE,0xDB,0xDC,0xD5,0xD2,
	    0xFF,0xF8,0xF1,0xF6,0xE3,0xE4,0xED,0xEA,
	    0xB7,0xB0,0xB9,0xBE,0xAB,0xAC,0xA5,0xA2,
	    0x8F,0x88,0x81,0x86,0x93,0x94,0x9D,0x9A,
	    0x27,0x20,0x29,0x2E,0x3B,0x3C,0x35,0x32,
	    0x1F,0x18,0x11,0x16,0x03,0x04,0x0D,0x0A,
	    0x57,0x50,0x59,0x5E,0x4B,0x4C,0x45,0x42,
	    0x6F,0x68,0x61,0x66,0x73,0x74,0x7D,0x7A,
	    0x89,0x8E,0x87,0x80,0x95,0x92,0x9B,0x9C,
	    0xB1,0xB6,0xBF,0xB8,0xAD,0xAA,0xA3,0xA4,
	    0xF9,0xFE,0xF7,0xF0,0xE5,0xE2,0xEB,0xEC,
	    0xC1,0xC6,0xCF,0xC8,0xDD,0xDA,0xD3,0xD4,
	    0x69,0x6E,0x67,0x60,0x75,0x72,0x7B,0x7C,
	    0x51,0x56,0x5F,0x58,0x4D,0x4A,0x43,0x44,
	    0x19,0x1E,0x17,0x10,0x05,0x02,0x0B,0x0C,
	    0x21,0x26,0x2F,0x28,0x3D,0x3A,0x33,0x34,
	    0x4E,0x49,0x40,0x47,0x52,0x55,0x5C,0x5B,
	    0x76,0x71,0x78,0x7F,0x6A,0x6D,0x64,0x63,
	    0x3E,0x39,0x30,0x37,0x22,0x25,0x2C,0x2B,
	    0x06,0x01,0x08,0x0F,0x1A,0x1D,0x14,0x13,
	    0xAE,0xA9,0xA0,0xA7,0xB2,0xB5,0xBC,0xBB,
	    0x96,0x91,0x98,0x9F,0x8A,0x8D,0x84,0x83,
	    0xDE,0xD9,0xD0,0xD7,0xC2,0xC5,0xCC,0xCB,
	    0xE6,0xE1,0xE8,0xEF,0xFA,0xFD,0xF4,0xF3
	    };

r = (unsigned char) ((r << 1) ^ ((r & 0x80) ? 0x07 : 0)); 

((r << 1) ^ ((r & 0x80) ? 0x07 : 0));

void DAC_send_crc(unsigned char dat2, unsigned char dat1, unsigned char dat0) 
{
    unsigned int *wskaznik;
    unsigned int tablica[4];
    unsigned char crc8 = 0;
/*--------------------------------------------------*/
    /*Obliczanie sumy kontrolnej CRC 8*/
    crc8= CRC8_TABLE[crc8^dat2];
    crc8= CRC8_TABLE[crc8^dat1];
    crc8= CRC8_TABLE[crc8^dat0];
/*--------------------------------------------------*/
    /*Przypisanie adresow kolejnym 4 bajtom do wysylki w trybie DMA*/
    tablica[1]=dat2;
    tablica[2]=dat1;
    tablica[3]=dat0;
    tablica[4]=crc8;

    /*Ustawienie wskaznika na pierwszym bajcie tablicy oraz liczba bajtow do wysylki*/
    wskaznik = &tablica[1];
    unsigned int number_of_data=3; 

    while(!(BASE_SPI->SPI_SR & AT91C_SPI_ENDTX));        
    BASE_SPI->SPI_TPR =(unsigned int)wskaznik; 
    BASE_SPI->SPI_TCR = number_of_data; 

    clear_output(N_SYNC);	//start ramki 24 lub 32bitowej

    BASE_SPI->SPI_PTCR= AT91C_PDC_TXTEN;   //wlacz transmisje DMA
wroc:
    if (!(BASE_SPI->SPI_SR & AT91C_SPI_ENDTX)) goto wroc;	//zaczekaj na wysylke wszystkich bajtow
    set_output(N_SYNC);		//koniec ramki 24 lub 32bitowej
	        
    BASE_SPI->SPI_PTCR= AT91C_PDC_TXTDIS; 	//wylacz transmisje DMA
	        
    if ((BASE_PIOA->PIO_ODSR & 1<<N_PEC))
        {
        set_output(LED);		//N_PEC==1 to LED ON
        }
    else
        {
        clear_output(LED);
        }
}

u8 crc8_update(u8 crc, u8 byte)
{
	crc ^= byte;
	u8 i;
	for(i=0; i<8; i++)
		crc = (crc << 1) ^ ((crc & 0x80)?0x07:0);
	return crc;
}

u8 crc;
crc = 0;
crc = crc8_update(crc, pierwszy_bajt);
crc = crc8_update(crc, drugi_bajt);
crc = crc8_update(crc, trzeci_bajt);
// tutaj zmienna crc zawiera poprawne crc

u8 crc8_update(u8 crc, u8 byte)
{
	return CRC8_TABLE[crc ^ byte];
}

u8 crc8_update(u8 crc, u8 byte)
{
	crc ^= byte;
	crc ^= (crc >> 6) ^ (crc >> 7);
	crc ^= (crc << 1) ^ (crc << 2);
	return crc;
}

#include <stdint.h>

typedef union {
	uint32_t Qword;
	uint8_t Byte[ 4 ];
} u_Four_Bytes;

uint8_t DAC_CRC_Calc( const uint32_t Data )
{
	uint32_t CRC = Data;
	uint32_t Mask = 0x80000000;
	uint32_t Poly = 0x83800000; /* left-justified polymonial 0x107 ( x^8 + x^2 + x + 1) */
	
	while ( CRC >= 256 )
	{
		if ( CRC & Mask )
			CRC ^= Poly; /* if masked bit equal one, then XOR data with polymonial */
		Mask >>= 1; /* shift mask one position right */
		Poly >>= 1; /* shift polymonial one position right */
	}
	
	return ( uint8_t ) CRC;
}

int main( void )
{
    /* ...some code...*/
    
	u_Four_Bytes Data;
	uint8_t CRC;
	
	Data.Byte[ 3 ] = 0x01;
	Data.Byte[ 2 ] = 0x01;
	Data.Byte[ 1 ] = 0x04;
	Data.Byte[ 0 ] = 0; /* LSB always zeroed */
    
	CRC = DAC_CRC_Calc( Data.Qword );
	
	/* sending etc. */
	
	return 0;
}