1/*	$NetBSD$	*/
2
3#include <tommath.h>
4#ifdef BN_FAST_S_MP_SQR_C
5/* LibTomMath, multiple-precision integer library -- Tom St Denis
6 *
7 * LibTomMath is a library that provides multiple-precision
8 * integer arithmetic as well as number theoretic functionality.
9 *
10 * The library was designed directly after the MPI library by
11 * Michael Fromberger but has been written from scratch with
12 * additional optimizations in place.
13 *
14 * The library is free for all purposes without any express
15 * guarantee it works.
16 *
17 * Tom St Denis, tomstdenis@gmail.com, http://libtom.org
18 */
19
20/* the jist of squaring...
21 * you do like mult except the offset of the tmpx [one that
22 * starts closer to zero] can't equal the offset of tmpy.
23 * So basically you set up iy like before then you min it with
24 * (ty-tx) so that it never happens.  You double all those
25 * you add in the inner loop
26
27After that loop you do the squares and add them in.
28*/
29
30int fast_s_mp_sqr (mp_int * a, mp_int * b)
31{
32  int       olduse, res, pa, ix, iz;
33  mp_digit   W[MP_WARRAY], *tmpx;
34  mp_word   W1;
35
36  /* grow the destination as required */
37  pa = a->used + a->used;
38  if (b->alloc < pa) {
39    if ((res = mp_grow (b, pa)) != MP_OKAY) {
40      return res;
41    }
42  }
43
44  /* number of output digits to produce */
45  W1 = 0;
46  for (ix = 0; ix < pa; ix++) {
47      int      tx, ty, iy;
48      mp_word  _W;
49      mp_digit *tmpy;
50
51      /* clear counter */
52      _W = 0;
53
54      /* get offsets into the two bignums */
55      ty = MIN(a->used-1, ix);
56      tx = ix - ty;
57
58      /* setup temp aliases */
59      tmpx = a->dp + tx;
60      tmpy = a->dp + ty;
61
62      /* this is the number of times the loop will iterrate, essentially
63         while (tx++ < a->used && ty-- >= 0) { ... }
64       */
65      iy = MIN(a->used-tx, ty+1);
66
67      /* now for squaring tx can never equal ty
68       * we halve the distance since they approach at a rate of 2x
69       * and we have to round because odd cases need to be executed
70       */
71      iy = MIN(iy, (ty-tx+1)>>1);
72
73      /* execute loop */
74      for (iz = 0; iz < iy; iz++) {
75         _W += ((mp_word)*tmpx++)*((mp_word)*tmpy--);
76      }
77
78      /* double the inner product and add carry */
79      _W = _W + _W + W1;
80
81      /* even columns have the square term in them */
82      if ((ix&1) == 0) {
83         _W += ((mp_word)a->dp[ix>>1])*((mp_word)a->dp[ix>>1]);
84      }
85
86      /* store it */
87      W[ix] = (mp_digit)(_W & MP_MASK);
88
89      /* make next carry */
90      W1 = _W >> ((mp_word)DIGIT_BIT);
91  }
92
93  /* setup dest */
94  olduse  = b->used;
95  b->used = a->used+a->used;
96
97  {
98    mp_digit *tmpb;
99    tmpb = b->dp;
100    for (ix = 0; ix < pa; ix++) {
101      *tmpb++ = W[ix] & MP_MASK;
102    }
103
104    /* clear unused digits [that existed in the old copy of c] */
105    for (; ix < olduse; ix++) {
106      *tmpb++ = 0;
107    }
108  }
109  mp_clamp (b);
110  return MP_OKAY;
111}
112#endif
113
114/* Source: /cvs/libtom/libtommath/bn_fast_s_mp_sqr.c,v */
115/* Revision: 1.4 */
116/* Date: 2006/12/28 01:25:13 */
117