| 1 |
#include <tommath.h> |
| 2 |
#ifdef BN_FAST_MP_MONTGOMERY_REDUCE_C |
| 3 |
/* LibTomMath, multiple-precision integer library -- Tom St Denis |
| 4 |
* |
| 5 |
* LibTomMath is a library that provides multiple-precision |
| 6 |
* integer arithmetic as well as number theoretic functionality. |
| 7 |
* |
| 8 |
* The library was designed directly after the MPI library by |
| 9 |
* Michael Fromberger but has been written from scratch with |
| 10 |
* additional optimizations in place. |
| 11 |
* |
| 12 |
* The library is free for all purposes without any express |
| 13 |
* guarantee it works. |
| 14 |
* |
| 15 |
* Tom St Denis, [email protected], http://libtom.org |
| 16 |
*/ |
| 17 |
|
| 18 |
/* computes xR**-1 == x (mod N) via Montgomery Reduction |
| 19 |
* |
| 20 |
* This is an optimized implementation of montgomery_reduce |
| 21 |
* which uses the comba method to quickly calculate the columns of the |
| 22 |
* reduction. |
| 23 |
* |
| 24 |
* Based on Algorithm 14.32 on pp.601 of HAC. |
| 25 |
*/ |
| 26 |
int fast_mp_montgomery_reduce (mp_int * x, mp_int * n, mp_digit rho) |
| 27 |
{ |
| 28 |
int ix, res, olduse; |
| 29 |
mp_word W[MP_WARRAY]; |
| 30 |
|
| 31 |
/* get old used count */ |
| 32 |
olduse = x->used; |
| 33 |
|
| 34 |
/* grow a as required */ |
| 35 |
if (x->alloc < n->used + 1) { |
| 36 |
if ((res = mp_grow (x, n->used + 1)) != MP_OKAY) { |
| 37 |
return res; |
| 38 |
} |
| 39 |
} |
| 40 |
|
| 41 |
/* first we have to get the digits of the input into |
| 42 |
* an array of double precision words W[...] |
| 43 |
*/ |
| 44 |
{ |
| 45 |
register mp_word *_W; |
| 46 |
register mp_digit *tmpx; |
| 47 |
|
| 48 |
/* alias for the W[] array */ |
| 49 |
_W = W; |
| 50 |
|
| 51 |
/* alias for the digits of x*/ |
| 52 |
tmpx = x->dp; |
| 53 |
|
| 54 |
/* copy the digits of a into W[0..a->used-1] */ |
| 55 |
for (ix = 0; ix < x->used; ix++) { |
| 56 |
*_W++ = *tmpx++; |
| 57 |
} |
| 58 |
|
| 59 |
/* zero the high words of W[a->used..m->used*2] */ |
| 60 |
for (; ix < n->used * 2 + 1; ix++) { |
| 61 |
*_W++ = 0; |
| 62 |
} |
| 63 |
} |
| 64 |
|
| 65 |
/* now we proceed to zero successive digits |
| 66 |
* from the least significant upwards |
| 67 |
*/ |
| 68 |
for (ix = 0; ix < n->used; ix++) { |
| 69 |
/* mu = ai * m' mod b |
| 70 |
* |
| 71 |
* We avoid a double precision multiplication (which isn't required) |
| 72 |
* by casting the value down to a mp_digit. Note this requires |
| 73 |
* that W[ix-1] have the carry cleared (see after the inner loop) |
| 74 |
*/ |
| 75 |
register mp_digit mu; |
| 76 |
mu = (mp_digit) (((W[ix] & MP_MASK) * rho) & MP_MASK); |
| 77 |
|
| 78 |
/* a = a + mu * m * b**i |
| 79 |
* |
| 80 |
* This is computed in place and on the fly. The multiplication |
| 81 |
* by b**i is handled by offseting which columns the results |
| 82 |
* are added to. |
| 83 |
* |
| 84 |
* Note the comba method normally doesn't handle carries in the |
| 85 |
* inner loop In this case we fix the carry from the previous |
| 86 |
* column since the Montgomery reduction requires digits of the |
| 87 |
* result (so far) [see above] to work. This is |
| 88 |
* handled by fixing up one carry after the inner loop. The |
| 89 |
* carry fixups are done in order so after these loops the |
| 90 |
* first m->used words of W[] have the carries fixed |
| 91 |
*/ |
| 92 |
{ |
| 93 |
register int iy; |
| 94 |
register mp_digit *tmpn; |
| 95 |
register mp_word *_W; |
| 96 |
|
| 97 |
/* alias for the digits of the modulus */ |
| 98 |
tmpn = n->dp; |
| 99 |
|
| 100 |
/* Alias for the columns set by an offset of ix */ |
| 101 |
_W = W + ix; |
| 102 |
|
| 103 |
/* inner loop */ |
| 104 |
for (iy = 0; iy < n->used; iy++) { |
| 105 |
*_W++ += ((mp_word)mu) * ((mp_word)*tmpn++); |
| 106 |
} |
| 107 |
} |
| 108 |
|
| 109 |
/* now fix carry for next digit, W[ix+1] */ |
| 110 |
W[ix + 1] += W[ix] >> ((mp_word) DIGIT_BIT); |
| 111 |
} |
| 112 |
|
| 113 |
/* now we have to propagate the carries and |
| 114 |
* shift the words downward [all those least |
| 115 |
* significant digits we zeroed]. |
| 116 |
*/ |
| 117 |
{ |
| 118 |
register mp_digit *tmpx; |
| 119 |
register mp_word *_W, *_W1; |
| 120 |
|
| 121 |
/* nox fix rest of carries */ |
| 122 |
|
| 123 |
/* alias for current word */ |
| 124 |
_W1 = W + ix; |
| 125 |
|
| 126 |
/* alias for next word, where the carry goes */ |
| 127 |
_W = W + ++ix; |
| 128 |
|
| 129 |
for (; ix <= n->used * 2 + 1; ix++) { |
| 130 |
*_W++ += *_W1++ >> ((mp_word) DIGIT_BIT); |
| 131 |
} |
| 132 |
|
| 133 |
/* copy out, A = A/b**n |
| 134 |
* |
| 135 |
* The result is A/b**n but instead of converting from an |
| 136 |
* array of mp_word to mp_digit than calling mp_rshd |
| 137 |
* we just copy them in the right order |
| 138 |
*/ |
| 139 |
|
| 140 |
/* alias for destination word */ |
| 141 |
tmpx = x->dp; |
| 142 |
|
| 143 |
/* alias for shifted double precision result */ |
| 144 |
_W = W + n->used; |
| 145 |
|
| 146 |
for (ix = 0; ix < n->used + 1; ix++) { |
| 147 |
*tmpx++ = (mp_digit)(*_W++ & ((mp_word) MP_MASK)); |
| 148 |
} |
| 149 |
|
| 150 |
/* zero oldused digits, if the input a was larger than |
| 151 |
* m->used+1 we'll have to clear the digits |
| 152 |
*/ |
| 153 |
for (; ix < olduse; ix++) { |
| 154 |
*tmpx++ = 0; |
| 155 |
} |
| 156 |
} |
| 157 |
|
| 158 |
/* set the max used and clamp */ |
| 159 |
x->used = n->used + 1; |
| 160 |
mp_clamp (x); |
| 161 |
|
| 162 |
/* if A >= m then A = A - m */ |
| 163 |
if (mp_cmp_mag (x, n) != MP_LT) { |
| 164 |
return s_mp_sub (x, n, x); |
| 165 |
} |
| 166 |
return MP_OKAY; |
| 167 |
} |
| 168 |
#endif |
| 169 |
|
| 170 |
/* $Source: /cvs/libtom/libtommath/bn_fast_mp_montgomery_reduce.c,v $ */ |
| 171 |
/* $Revision: 1.4 $ */ |
| 172 |
/* $Date: 2006/12/28 01:25:13 $ */ |