Lines Matching full:128

42  *  gf_wrap -- In our implementation, GF(2^128) is represented as a 16 byte
46  *                     X^128 + X^7 + X^2 + X + 1,
48  *  which leads to the reduction formula X^128 = X^7 + X^2 + X + 1. Indeed,
49  *  since 0 = (X^128 + X^7 + X^2 + 1) mod (X^128 + X^7 + X^2 + X + 1) and since
51  *  add X^128 to both sides to get
53  *       X^128 = (X^7 + X^2 + X + 1) mod (X^128 + X^7 + X^2 + X + 1)
58  *  This gets used in the following way. Doubling in GF(2^128) is just a left
60  *  case, the relation X^128 = X^7 + X^2 + X + 1 says that the high order bit
61  *  that overflows beyond 128 bits can be replaced by addition of
62  *  X^7 + X^2 + X + 1 <--> 0x87 to the low order 128 bits. Since addition
63  *  in GF(2^128) is represented by XOR, we therefore only have to XOR 0x87