C语言CRC-16 IBM格式校验函数

CRC-16校验产生2个字节长度的数据校验码，通过计算得到的校验码和获得的校验码比较，用于验证获得的数据的正确性。基本的CRC-16校验算法实现，参考： C语言标准CRC-16校验函数。

不同厂家通过对输入数据前处理和输出数据后处理的方式不同，又产生了不同的厂家校验函数，这里介绍IBM格式的CRC-16校验函数。IBM格式对输入数据，按照单个字节进行位反序。对于输出的校验码，进行整体位反序。

正向算法

正向算法是符合标准CRC-16的计算理论，从左向右计算，也即计算过程中移位时，向左移出。几种正向算法的实现如下：

CRC-16 IBM格式校验函数一(8位输入数据格式，64位装载计算)：

#include <stdio.h>
#include <stdlib.h>
uint16_t PY_CRC_16_IBM(uint8_t *di, uint32_t len)
{uint32_t crc_poly = 0x00018005;  //X^16+X^15+X^2+1 total 17 effective bits. Computed total data shall be compensated 16-bit '0' before CRC computing.uint8_t *datain;uint64_t cdata = 0; //Computed total datauint32_t data_t = 0; //Process data of CRC computinguint16_t index_t = 63;  ///bit shifting index for initial '1' searchinguint16_t index = 63;    //bit shifting index for CRC computinguint8_t rec = 0; //bit number needed to be compensated for next CRC computinguint32_t cn=(len+2)/6;uint32_t cr=(len+2)%6;uint32_t j;datain = malloc(len+2);for(j=0;j<len;j++){datain[j] = 0;for(uint8_t m=0; m<=7; m++){datain[j] |= ( ( di[j]>>(7-m) ) & 1 ) << m;}}datain[len] = 0; datain[len+1] = 0;//Compensate 16-bit '0' for input dataif(len<=6)   //Mount data for only one segment{for(j=0;j<=(len+1);j++){cdata = (cdata<<8);cdata = cdata|datain[j];}cn = 1;}else{if(cr==0){cr = 6;}else if(cr==1){cr = 7;}else if(cr==2){cr = 8;}else{cn++;}for(j=0;j<cr;j++){cdata = (cdata<<8);cdata = cdata|datain[j];}}do{cn--;while(index_t>0){if( (cdata>>index_t)&1 ){index = index_t;index_t = 0;data_t |= (cdata>>(index-16));{data_t = data_t ^ crc_poly;}while((index!=0x5555)&&(index!=0xaaaa)){for(uint8_t n=1;n<17;n++){if ((data_t>>(16-n))&1) {rec = n;break;}if (n==16) rec=17;}if((index-16)<rec){data_t = data_t<<(index-16);data_t |=  (uint32_t)((cdata<<(64-(index-16)))>>(64-(index-16)));index = 0x5555;}else{for(uint8_t i=1;i<=rec;i++){data_t = (data_t<<1)|((cdata>>(index-16-i))&1) ;}if(rec!= 17){data_t = data_t ^ crc_poly;index -= rec;}else{data_t = 0;index_t = index-16-1;index = 0xaaaa;}}}if(index==0x5555) break;}else{index_t--;if(index_t<16) break;}}if(cn>0) //next segment{cdata = data_t&0x00ffff;for(uint8_t k=0;k<6;k++){cdata = (cdata<<8);cdata = cdata|datain[j++];}data_t = 0;index_t = 63;  ///bit shifting index for initial '1' searchingindex = 63;    //bit shifting index for CRC computingrec = 0; //bit number needed to be compensated for next CRC computing}}while(cn>0);free(datain);uint16_t i_data_t = 0;for(uint8_t n=0; n<=15; n++){i_data_t |=  ( ( data_t>>(15-n) ) & 1 ) << n;}return i_data_t;
}

CRC-16 IBM格式校验函数二（8位输入数据格式）：

uint16_t PY_CRC_16_S_IBM(uint8_t *di, uint32_t len)
{uint16_t crc_poly = 0x8005;  //X^16+X^15+X^2+1 total 16 effective bits without X^16. Computed total data shall be compensated 16-bit '0' before CRC computing.uint32_t clen = len+2;uint8_t cdata[clen] ;for(uint32_t j=0;j<len;j++){cdata[j] = 0;for(uint8_t m=0; m<=7; m++){cdata[j] |= ( ( di[j]>>(7-m) ) & 1 ) << m;}}cdata[len]=0; cdata[len+1]=0;uint16_t data_t = (((uint16_t)cdata[0]) << 8) + cdata[1]; //CRC registerfor (uint32_t i = 2; i < clen; i++){for (uint8_t j = 0; j <= 7; j++){if(data_t&0x8000)data_t = ( (data_t<<1) | ( (cdata[i]>>(7-j))&0x01) ) ^ crc_poly;elsedata_t = ( (data_t<<1) | ( (cdata[i]>>(7-j))&0x01) ) ;}}uint16_t i_data_t = 0;for(uint8_t n=0; n<=15; n++){i_data_t |=  ( ( data_t>>(15-n) ) & 1 ) << n;}return i_data_t;
}

CRC-16 IBM格式校验函数三（16位输入数据格式）：

uint16_t PY_CRC_16_T16_IBM(uint16_t *di, uint32_t len)
{uint16_t crc_poly = 0x8005;  //X^16+X^15+X^2+1 total 16 effective bits without X^16. Computed total data shall be compensated 16-bit '0' before CRC computing.uint16_t data_t = 0; //CRC registeruint16_t cdata[len];for(uint32_t j=0;j<len;j++){cdata[j] = 0;for(uint8_t m=0; m<=7; m++){cdata[j] |= ( ( ( (di[j]>>8)>>(7-m) ) & 1 ) << m ) | ( ( ( ( (di[j]&0x00ff)>>(7-m) ) & 1 ) << m ) <<8 );}}for(uint32_t i = 0; i < len; i++){data_t ^= cdata[i]; //16-bit datafor (uint8_t j = 0; j < 16; j++){if (data_t & 0x8000)data_t = (data_t << 1) ^ crc_poly;elsedata_t <<= 1;}}uint16_t i_data_t = 0;for(uint8_t n=0; n<=15; n++){i_data_t |=  ( ( data_t>>(15-n) ) & 1 ) << n;}return i_data_t;
}

CRC-16 IBM格式校验函数四（8位输入数据格式）：

uint16_t PY_CRC_16_T8_IBM(uint8_t *di, uint32_t len)
{uint16_t crc_poly = 0x8005;  //X^16+X^15+X^2+1 total 16 effective bits without X^16. Computed total data shall be compensated 16-bit '0' before CRC computing.uint16_t data_t = 0; //CRC registeruint8_t cdata[len];for(uint32_t j=0;j<len;j++){cdata[j] = 0;for(uint8_t m=0; m<=7; m++){cdata[j] |= ( ( di[j]>>(7-m) ) & 1 ) << m;}}for(uint32_t i = 0; i < len; i++){data_t ^= cdata[i]<<8; //8-bit datafor (uint8_t j = 0; j < 8; j++){if (data_t & 0x8000)data_t = (data_t << 1) ^ crc_poly;elsedata_t <<= 1;}}uint16_t i_data_t = 0;for(uint8_t n=0; n<=15; n++){i_data_t |=  ( ( data_t>>(15-n) ) & 1 ) << n;}return i_data_t;
}

反向算法

反向算法是从由右向左计算，也即计算过程中移位时，向右移出。而计算过程中的输入数据高优先计算位和校验参数的对齐关系不变。因此把一个字节放在CRC计算寄存器的最低字节时，对于IBM格式，最右侧最低位实际上是高优先计算位，而校验参数要相应倒序，从而计算位置对照关系不变。

CRC-16 IBM格式校验函数五（反向算法，8位输入数据格式）：

uint16_t PY_CRC_16_T8_IBM_i(uint8_t *di, uint32_t len)
{uint16_t crc_poly = 0xA001; //Bit sequence inversion of 0x8005uint16_t data_t = 0; //CRC registerfor(uint32_t i = 0; i < len; i++){data_t ^= di[i]; //8-bit datafor (uint8_t j = 0; j < 8; j++){if (data_t & 0x0001)data_t = (data_t >> 1) ^ crc_poly;elsedata_t >>= 1;}}return data_t;
}

算法验证

5种算法结果相同：

通过在线CRC工具对照验证成功：

–End–