Hello all,

I’ve the task of implementing the following CRC computation and was wondering whether I could use boost::crc to save me a lot of hassle. 

 

------------------------

Default 32-bit Parity (CRC)

 

This 32-bit CRC is the same as that described in EKMS 304 ([6]) and EKMS 322 ([8]) and could be used as a CRC method in the keyAlgorithm Attribute (with OID { 1 2 826 0 1145 0 5 1 }). Let g(x) be the polynomial x³² + x²² + x² + x + 1. Let d(x) be the k bits of data to have parity applied to them. The first bit of data is the coefficient of the (k-1)st power of x etc. Let f(x) be the polynomial naturally represented by the initial fill 0xF8DD4258, i.e. x³¹ + x³⁰ + x²⁹ + x²⁸ + x²⁷ + x²³ + x²² + x²⁰ + x¹⁹ + x¹⁸ + x¹⁶ + x¹⁴ + x⁹ + x⁶ + x⁴ + x³. Then the output is

 

c(x) = x³²d(x) + x^kf(x) mod g(x).

 

The polynomial c(x) is then interpreted as a stream with the coefficient of x³¹ coming first and the coefficient of x⁰ coming last. The 32-bit parity is then c(x).

 

This calculation can be done using a simple shift register as shown below. The shaded stages represent ones in the initial fill. The logic steps k times to determine c(x), which is the final state of the register.

<diagram omitted>

------------------------

 

Any advice and or help would be appreciated.

Regards,

Paul