The following C-language program can be used as a model to describe
the algorithm. It assumes that a char is one octet.
It also assumes that the entire file is available for one pass through
the function. This was done for simplicity in demonstrating
the algorithm, rather than as an implementation model.
static unsigned long crctab[] = {
0x00000000,
0x04c11db7, 0x09823b6e, 0x0d4326d9, 0x130476dc, 0x17c56b6b,
0x1a864db2, 0x1e475005, 0x2608edb8, 0x22c9f00f, 0x2f8ad6d6,
0x2b4bcb61, 0x350c9b64, 0x31cd86d3, 0x3c8ea00a, 0x384fbdbd,
0x4c11db70, 0x48d0c6c7, 0x4593e01e, 0x4152fda9, 0x5f15adac,
0x5bd4b01b, 0x569796c2, 0x52568b75, 0x6a1936c8, 0x6ed82b7f,
0x639b0da6, 0x675a1011, 0x791d4014, 0x7ddc5da3, 0x709f7b7a,
0x745e66cd, 0x9823b6e0, 0x9ce2ab57, 0x91a18d8e, 0x95609039,
0x8b27c03c, 0x8fe6dd8b, 0x82a5fb52, 0x8664e6e5, 0xbe2b5b58,
0xbaea46ef, 0xb7a96036, 0xb3687d81, 0xad2f2d84, 0xa9ee3033,
0xa4ad16ea, 0xa06c0b5d, 0xd4326d90, 0xd0f37027, 0xddb056fe,
0xd9714b49, 0xc7361b4c, 0xc3f706fb, 0xceb42022, 0xca753d95,
0xf23a8028, 0xf6fb9d9f, 0xfbb8bb46, 0xff79a6f1, 0xe13ef6f4,
0xe5ffeb43, 0xe8bccd9a, 0xec7dd02d, 0x34867077, 0x30476dc0,
0x3d044b19, 0x39c556ae, 0x278206ab, 0x23431b1c, 0x2e003dc5,
0x2ac12072, 0x128e9dcf, 0x164f8078, 0x1b0ca6a1, 0x1fcdbb16,
0x018aeb13, 0x054bf6a4, 0x0808d07d, 0x0cc9cdca, 0x7897ab07,
0x7c56b6b0, 0x71159069, 0x75d48dde, 0x6b93dddb, 0x6f52c06c,
0x6211e6b5, 0x66d0fb02, 0x5e9f46bf, 0x5a5e5b08, 0x571d7dd1,
0x53dc6066, 0x4d9b3063, 0x495a2dd4, 0x44190b0d, 0x40d816ba,
0xaca5c697, 0xa864db20, 0xa527fdf9, 0xa1e6e04e, 0xbfa1b04b,
0xbb60adfc, 0xb6238b25, 0xb2e29692, 0x8aad2b2f, 0x8e6c3698,
0x832f1041, 0x87ee0df6, 0x99a95df3, 0x9d684044, 0x902b669d,
0x94ea7b2a, 0xe0b41de7, 0xe4750050, 0xe9362689, 0xedf73b3e,
0xf3b06b3b, 0xf771768c, 0xfa325055, 0xfef34de2, 0xc6bcf05f,
0xc27dede8, 0xcf3ecb31, 0xcbffd686, 0xd5b88683, 0xd1799b34,
0xdc3abded, 0xd8fba05a, 0x690ce0ee, 0x6dcdfd59, 0x608edb80,
0x644fc637, 0x7a089632, 0x7ec98b85, 0x738aad5c, 0x774bb0eb,
0x4f040d56, 0x4bc510e1, 0x46863638, 0x42472b8f, 0x5c007b8a,
0x58c1663d, 0x558240e4, 0x51435d53, 0x251d3b9e, 0x21dc2629,
0x2c9f00f0, 0x285e1d47, 0x36194d42, 0x32d850f5, 0x3f9b762c,
0x3b5a6b9b, 0x0315d626, 0x07d4cb91, 0x0a97ed48, 0x0e56f0ff,
0x1011a0fa, 0x14d0bd4d, 0x19939b94, 0x1d528623, 0xf12f560e,
0xf5ee4bb9, 0xf8ad6d60, 0xfc6c70d7, 0xe22b20d2, 0xe6ea3d65,
0xeba91bbc, 0xef68060b, 0xd727bbb6, 0xd3e6a601, 0xdea580d8,
0xda649d6f, 0xc423cd6a, 0xc0e2d0dd, 0xcda1f604, 0xc960ebb3,
0xbd3e8d7e, 0xb9ff90c9, 0xb4bcb610, 0xb07daba7, 0xae3afba2,
0xaafbe615, 0xa7b8c0cc, 0xa379dd7b, 0x9b3660c6, 0x9ff77d71,
0x92b45ba8, 0x9675461f, 0x8832161a, 0x8cf30bad, 0x81b02d74,
0x857130c3, 0x5d8a9099, 0x594b8d2e, 0x5408abf7, 0x50c9b640,
0x4e8ee645, 0x4a4ffbf2, 0x470cdd2b, 0x43cdc09c, 0x7b827d21,
0x7f436096, 0x7200464f, 0x76c15bf8, 0x68860bfd, 0x6c47164a,
0x61043093, 0x65c52d24, 0x119b4be9, 0x155a565e, 0x18197087,
0x1cd86d30, 0x029f3d35, 0x065e2082, 0x0b1d065b, 0x0fdc1bec,
0x3793a651, 0x3352bbe6, 0x3e119d3f, 0x3ad08088, 0x2497d08d,
0x2056cd3a, 0x2d15ebe3, 0x29d4f654, 0xc5a92679, 0xc1683bce,
0xcc2b1d17, 0xc8ea00a0, 0xd6ad50a5, 0xd26c4d12, 0xdf2f6bcb,
0xdbee767c, 0xe3a1cbc1, 0xe760d676, 0xea23f0af, 0xeee2ed18,
0xf0a5bd1d, 0xf464a0aa, 0xf9278673, 0xfde69bc4, 0x89b8fd09,
0x8d79e0be, 0x803ac667, 0x84fbdbd0, 0x9abc8bd5, 0x9e7d9662,
0x933eb0bb, 0x97ffad0c, 0xafb010b1, 0xab710d06, 0xa6322bdf,
0xa2f33668, 0xbcb4666d, 0xb8757bda, 0xb5365d03, 0xb1f740b4
};
unsigned long memcrc(const unsigned char *b, size_t n)
{
/* Input arguments:
* const char* b == byte sequence to checksum
* size_t n == length of sequence
*/
register unsigned i, c, s = 0;
for (i = n; i > 0; --i) {
c = (unsigned)(*b++);
s = (s << 8) ^ crctab[(s >> 24) ^ c];
}
/* Extend with the length of the string. */
while (n != 0) {
c = n & 0377;
n >>= 8;
s = (s << 8) ^ crctab[(s >> 24) ^ c];
}
return ~s;
}
The historical practice of writing the number of "blocks" has been
changed to writing the number of octets, since the latter
is not only more useful, but also since historical implementations
have not been consistent in defining what a "block" meant.
Octets are used instead of bytes because bytes can differ in size
between systems.
The algorithm selected is based on that used by the ISO/IEC 8802-3:1996
standard (Ethernet) for the frame check sequence
field. The algorithm used does not match the technical definition
of a checksum; the term is used for historical reasons.
The length of the file is included in the CRC calculation because
this parallels inclusion of a length field by Ethernet in its
CRC, but also because it guards against inadvertent collisions between
files that begin with different series of zero octets. The
chance that two different files produce identical CRCs is much greater
when their lengths are not considered. Keeping the length
and the checksum of the file itself separate would yield a slightly
more robust algorithm, but historical usage has always been
that a single number (the checksum as printed) represents the signature
of the file. It was decided that historical usage was the
more important consideration.
Early proposals contained modifications to the Ethernet algorithm
that involved extracting table values whenever an intermediate
result became zero. This was demonstrated to be less robust than the
current method and mathematically difficult to describe or
justify.
The calculation used is identical to that given in pseudo-code in
the referenced Sarwate article. The pseudo-code rendition
is:
The pseudo-code is reproduced exactly as given; however, note that
in the case of cksum, A[i] represents a byte of
the file, the words X and Y are treated as a single 32-bit
value, and the tables f and f’ are a single
table containing 32-bit values.
