CRC - OpenBSC

18 Jan 2012


      I am trying to generate CRC table mentioned in
openbsc/src/gprs/crc24.c file
(http://lingrok.org/source/xref/openbsc/openbsc/src/gprs/crc24.c)
I modified a CRC16 code to generate CRC24 (updated polynomial as per
LLC specification). But I can not get right table.
Below is source code. Can someone help me out?
//******************************************************************************
//
//    Copyright (c) 1999 Axon Instruments.
//    All rights reserved.
//
//******************************************************************************
// HEADER:  CRC16.cpp
// PURPOSE: Contains functions calculating a 16 bit CRC code.
// SOURCE:  http://www.brokersys.com/snippets/   file: CRC_16F.C
//
#define POLYNOMIAL 0xAD85DD
#define CRC_WIDTH 24
#define BITS_PER_BYTE 8
typedef unsigned int WORD;
typedef unsigned int UINT;
typedef char BYTE;
static WORD correctCRCtable[1<<BITS_PER_BYTE] =     // as per
openbsc/src/gprs/crc24.c
{
    0x00000000, 0x00d6a776, 0x00f64557, 0x0020e221, 0x00b78115,
0x00612663, 0x0041c442, 0x00976334,
    0x00340991, 0x00e2aee7, 0x00c24cc6, 0x0014ebb0, 0x00838884,
0x00552ff2, 0x0075cdd3, 0x00a36aa5,
    /* skip rest */
};
void InitCRCtab()
{
   int v, i;
   int b;
for ( b = 0; b <= (1<<BITS_PER_BYTE)-1; ++b )
   {
      int v = b << (CRC_WIDTH-BITS_PER_BYTE);
      int i;
      for (i = BITS_PER_BYTE; --i >= 0; )
         v = v&0x800000 ? (v<<1) ^ POLYNOMIAL : v<<1;
if((v & 0xFFFFFF) == correctCRCtable[b])
    	  correctCRCtable[b] = v;
      else {
    	  printf("Incorrect CRC table\n");
    	  printf("correctCRCtable[%d]=0x%06x calculated=0x%06x\n", b,
correctCRCtable[b], v&0xFFFFFF);
    	  exit(0);
      }
   }
}
int main()
{
    InitCRCtab();
    printf("Correct CRC table\n");
    return 0;
}
/*
Below is LLC specification:
The FCS field shall consist of a 24 bit cyclic redundancy check (CRC)
code. The CRC-24 is used to detect bit errors in the frame header and
information fields.
The FCS field contains the value of a CRC calculation that is
performed over the entire contents of the header and information
field, except for UI frames transmitted in unprotected mode, in which
case the FCS field contains the value of a CRC calculation that is
performed over the frame header and the first N202 octets (see
subclause 8.9.6) of the information field only (see subclause
6.3.5.5.2). The information over which the CRC is calculated is
referred to as the dividend in this subclause. Bit (1, 1) of the
dividend is the highest-order term in the calculation (see subclause
5.7.3). CRC calculation shall be done before ciphering at the
transmitting side, and after deciphering at the receiving side.
NOTE:	The definition below is different from that in 3GPP TS 24.022
[10] only with respect to the variable dividend length k of the LLC
frames. In 3GPP TS 24.022, the RLP frame has a fixed dividend length,
but the LLC frame has a variable dividend length.
The CRC shall be the ones complement of the sum (modulo 2) of:
-	the remainder of x^k (x^23 + x^22 + x^21 +... + x^2 + x + 1) divided
(modulo 2) by the generator polynomial, where k is the number of bits
of the dividend; and
-	the remainder of the division (modulo 2) by the generator polynomial
of the product of x^24 by the dividend.
The CRC-24 generator polynomial is:
G(x) = x^24 + x^23 + x^21 + x^20 + x^19 + x^17 + x^16 + x^15 + x^13 +
x^8 + x^7 + x^5 + x^4 + x^2 + 1
The result of the CRC calculation is placed within the FCS field as
described in subclause 5.7.3.
NOTE:	As a typical implementation at the transmitter, the initial
content of the register of the device computing the remainder of the
division is pre-set to all "1's" and is then modified by division by
the generator polynomial (as described above) of the dividend; the
ones complement of the resulting remainder is put into the FCS field.
As a typical implementation at the receiver, the initial content of
the register of the device computing the remainder of the division is
pre-set to all "1's". The final remainder, after multiplication by x24
and then division (modulo 2) by the generator polynomial of the
received frame, will be (in the absence of errors):
C(x) = x^22 + x^21 + x^19 + x^18 + x^16 + x^15 + x^11 + x^8 + x^5 + x^4
 */