Coverage Report

Created: 2026-05-16 06:52

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/gmp/mpn/gcdext_lehmer.c
Line
Count
Source
1
/* mpn_gcdext -- Extended Greatest Common Divisor.
2
3
Copyright 1996, 1998, 2000-2005, 2008, 2009, 2012 Free Software Foundation,
4
Inc.
5
6
This file is part of the GNU MP Library.
7
8
The GNU MP Library is free software; you can redistribute it and/or modify
9
it under the terms of either:
10
11
  * the GNU Lesser General Public License as published by the Free
12
    Software Foundation; either version 3 of the License, or (at your
13
    option) any later version.
14
15
or
16
17
  * the GNU General Public License as published by the Free Software
18
    Foundation; either version 2 of the License, or (at your option) any
19
    later version.
20
21
or both in parallel, as here.
22
23
The GNU MP Library is distributed in the hope that it will be useful, but
24
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
25
or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
26
for more details.
27
28
You should have received copies of the GNU General Public License and the
29
GNU Lesser General Public License along with the GNU MP Library.  If not,
30
see https://www.gnu.org/licenses/.  */
31
32
#include "gmp-impl.h"
33
#include "longlong.h"
34
35
/* Here, d is the index of the cofactor to update. FIXME: Could use qn
36
   = 0 for the common case q = 1. */
37
void
38
mpn_gcdext_hook (void *p, mp_srcptr gp, mp_size_t gn,
39
     mp_srcptr qp, mp_size_t qn, int d)
40
18.6k
{
41
18.6k
  struct gcdext_ctx *ctx = (struct gcdext_ctx *) p;
42
18.6k
  mp_size_t un = ctx->un;
43
44
18.6k
  if (gp)
45
387
    {
46
387
      mp_srcptr up;
47
48
387
      ASSERT (gn > 0);
49
387
      ASSERT (gp[gn-1] > 0);
50
51
387
      MPN_COPY (ctx->gp, gp, gn);
52
387
      ctx->gn = gn;
53
54
387
      if (d < 0)
55
187
  {
56
187
    int c;
57
58
    /* Must return the smallest cofactor, +u1 or -u0 */
59
187
    MPN_CMP (c, ctx->u0, ctx->u1, un);
60
187
    ASSERT (c != 0 || (un == 1 && ctx->u0[0] == 1 && ctx->u1[0] == 1));
61
62
187
    d = c < 0;
63
187
  }
64
65
387
      up = d ? ctx->u0 : ctx->u1;
66
67
387
      MPN_NORMALIZE (up, un);
68
387
      MPN_COPY (ctx->up, up, un);
69
70
387
      *ctx->usize = d ? -un : un;
71
387
    }
72
18.2k
  else
73
18.2k
    {
74
18.2k
      mp_limb_t cy;
75
18.2k
      mp_ptr u0 = ctx->u0;
76
18.2k
      mp_ptr u1 = ctx->u1;
77
78
18.2k
      ASSERT (d >= 0);
79
80
18.2k
      if (d)
81
8.58k
  MP_PTR_SWAP (u0, u1);
82
83
18.2k
      qn -= (qp[qn-1] == 0);
84
85
      /* Update u0 += q  * u1 */
86
18.2k
      if (qn == 1)
87
11.7k
  {
88
11.7k
    mp_limb_t q = qp[0];
89
90
11.7k
    if (q == 1)
91
      /* A common case. */
92
9.24k
      cy = mpn_add_n (u0, u0, u1, un);
93
2.52k
    else
94
2.52k
      cy = mpn_addmul_1 (u0, u1, un, q);
95
11.7k
  }
96
6.52k
      else
97
6.52k
  {
98
6.52k
    mp_size_t u1n;
99
6.52k
    mp_ptr tp;
100
101
6.52k
    u1n = un;
102
6.52k
    MPN_NORMALIZE (u1, u1n);
103
104
6.52k
    if (u1n == 0)
105
0
      return;
106
107
    /* Should always have u1n == un here, and u1 >= u0. The
108
       reason is that we alternate adding u0 to u1 and u1 to u0
109
       (corresponding to subtractions a - b and b - a), and we
110
       can get a large quotient only just after a switch, which
111
       means that we'll add (a multiple of) the larger u to the
112
       smaller. */
113
114
6.52k
    tp = ctx->tp;
115
116
6.52k
    if (qn > u1n)
117
1.73k
      mpn_mul (tp, qp, qn, u1, u1n);
118
4.79k
    else
119
4.79k
      mpn_mul (tp, u1, u1n, qp, qn);
120
121
6.52k
    u1n += qn;
122
6.52k
    u1n -= tp[u1n-1] == 0;
123
124
6.52k
    if (u1n >= un)
125
6.52k
      {
126
6.52k
        cy = mpn_add (u0, tp, u1n, u0, un);
127
6.52k
        un = u1n;
128
6.52k
      }
129
0
    else
130
      /* Note: Unlikely case, maybe never happens? */
131
0
      cy = mpn_add (u0, u0, un, tp, u1n);
132
133
6.52k
  }
134
18.2k
      u0[un] = cy;
135
18.2k
      ctx->un = un + (cy > 0);
136
18.2k
    }
137
18.6k
}
138
139
/* Temporary storage: 3*(n+1) for u. If hgcd2 succeeds, we need n for
140
   the matrix-vector multiplication adjusting a, b. If hgcd fails, we
141
   need at most n for the quotient and n+1 for the u update (reusing
142
   the extra u). In all, 4n + 3. */
143
144
mp_size_t
145
mpn_gcdext_lehmer_n (mp_ptr gp, mp_ptr up, mp_size_t *usize,
146
         mp_ptr ap, mp_ptr bp, mp_size_t n,
147
         mp_ptr tp)
148
32.2k
{
149
32.2k
  mp_size_t ualloc = n + 1;
150
151
  /* Keeps track of the second row of the reduction matrix
152
   *
153
   *   M = (v0, v1 ; u0, u1)
154
   *
155
   * which correspond to the first column of the inverse
156
   *
157
   *   M^{-1} = (u1, -v1; -u0, v0)
158
   *
159
   * This implies that
160
   *
161
   *   a =  u1 A (mod B)
162
   *   b = -u0 A (mod B)
163
   *
164
   * where A, B denotes the input values.
165
   */
166
167
32.2k
  struct gcdext_ctx ctx;
168
32.2k
  mp_size_t un;
169
32.2k
  mp_ptr u0;
170
32.2k
  mp_ptr u1;
171
32.2k
  mp_ptr u2;
172
173
32.2k
  MPN_ZERO (tp, 3*ualloc);
174
32.2k
  u0 = tp; tp += ualloc;
175
32.2k
  u1 = tp; tp += ualloc;
176
32.2k
  u2 = tp; tp += ualloc;
177
178
32.2k
  u1[0] = 1; un = 1;
179
180
32.2k
  ctx.gp = gp;
181
32.2k
  ctx.up = up;
182
32.2k
  ctx.usize = usize;
183
184
  /* FIXME: Handle n == 2 differently, after the loop? */
185
589k
  while (n >= 2)
186
557k
    {
187
557k
      struct hgcd_matrix1 M;
188
557k
      mp_limb_t ah, al, bh, bl;
189
557k
      mp_limb_t mask;
190
191
557k
      mask = ap[n-1] | bp[n-1];
192
557k
      ASSERT (mask > 0);
193
194
557k
      if (mask & GMP_NUMB_HIGHBIT)
195
24.1k
  {
196
24.1k
    ah = ap[n-1]; al = ap[n-2];
197
24.1k
    bh = bp[n-1]; bl = bp[n-2];
198
24.1k
  }
199
533k
      else if (n == 2)
200
29.6k
  {
201
    /* We use the full inputs without truncation, so we can
202
       safely shift left. */
203
29.6k
    int shift;
204
205
29.6k
    count_leading_zeros (shift, mask);
206
29.6k
    ah = MPN_EXTRACT_NUMB (shift, ap[1], ap[0]);
207
29.6k
    al = ap[0] << shift;
208
29.6k
    bh = MPN_EXTRACT_NUMB (shift, bp[1], bp[0]);
209
29.6k
    bl = bp[0] << shift;
210
29.6k
  }
211
503k
      else
212
503k
  {
213
503k
    int shift;
214
215
503k
    count_leading_zeros (shift, mask);
216
503k
    ah = MPN_EXTRACT_NUMB (shift, ap[n-1], ap[n-2]);
217
503k
    al = MPN_EXTRACT_NUMB (shift, ap[n-2], ap[n-3]);
218
503k
    bh = MPN_EXTRACT_NUMB (shift, bp[n-1], bp[n-2]);
219
503k
    bl = MPN_EXTRACT_NUMB (shift, bp[n-2], bp[n-3]);
220
503k
  }
221
222
      /* Try an mpn_nhgcd2 step */
223
557k
      if (mpn_hgcd2 (ah, al, bh, bl, &M))
224
547k
  {
225
547k
    n = mpn_matrix22_mul1_inverse_vector (&M, tp, ap, bp, n);
226
547k
    MP_PTR_SWAP (ap, tp);
227
547k
    un = mpn_hgcd_mul_matrix1_vector(&M, u2, u0, u1, un);
228
547k
    MP_PTR_SWAP (u0, u2);
229
547k
  }
230
9.43k
      else
231
9.43k
  {
232
    /* mpn_hgcd2 has failed. Then either one of a or b is very
233
       small, or the difference is very small. Perform one
234
       subtraction followed by one division. */
235
9.43k
    ctx.u0 = u0;
236
9.43k
    ctx.u1 = u1;
237
9.43k
    ctx.tp = u2;
238
9.43k
    ctx.un = un;
239
240
    /* Temporary storage n for the quotient and ualloc for the
241
       new cofactor. */
242
9.43k
    n = mpn_gcd_subdiv_step (ap, bp, n, 0, mpn_gcdext_hook, &ctx, tp);
243
9.43k
    if (n == 0)
244
387
      return ctx.gn;
245
246
9.04k
    un = ctx.un;
247
9.04k
  }
248
557k
    }
249
31.8k
  ASSERT_ALWAYS (ap[0] > 0);
250
31.8k
  ASSERT_ALWAYS (bp[0] > 0);
251
252
31.8k
  if (ap[0] == bp[0])
253
960
    {
254
960
      int c;
255
256
      /* Which cofactor to return now? Candidates are +u1 and -u0,
257
   depending on which of a and b was most recently reduced,
258
   which we don't keep track of. So compare and get the smallest
259
   one. */
260
261
960
      gp[0] = ap[0];
262
263
960
      MPN_CMP (c, u0, u1, un);
264
960
      ASSERT (c != 0 || (un == 1 && u0[0] == 1 && u1[0] == 1));
265
960
      if (c < 0)
266
461
  {
267
461
    MPN_NORMALIZE (u0, un);
268
461
    MPN_COPY (up, u0, un);
269
461
    *usize = -un;
270
461
  }
271
499
      else
272
499
  {
273
499
    MPN_NORMALIZE_NOT_ZERO (u1, un);
274
499
    MPN_COPY (up, u1, un);
275
499
    *usize = un;
276
499
  }
277
960
      return 1;
278
960
    }
279
30.8k
  else
280
30.8k
    {
281
30.8k
      mp_limb_t uh, vh;
282
30.8k
      mp_limb_signed_t u;
283
30.8k
      mp_limb_signed_t v;
284
30.8k
      int negate;
285
286
30.8k
      gp[0] = mpn_gcdext_1 (&u, &v, ap[0], bp[0]);
287
288
      /* Set up = u u1 - v u0. Keep track of size, un grows by one or
289
   two limbs. */
290
291
30.8k
      if (u == 0)
292
77
  {
293
77
    ASSERT (v == 1);
294
77
    MPN_NORMALIZE (u0, un);
295
77
    MPN_COPY (up, u0, un);
296
77
    *usize = -un;
297
77
    return 1;
298
77
  }
299
30.8k
      else if (v == 0)
300
318
  {
301
318
    ASSERT (u == 1);
302
318
    MPN_NORMALIZE (u1, un);
303
318
    MPN_COPY (up, u1, un);
304
318
    *usize = un;
305
318
    return 1;
306
318
  }
307
30.5k
      else if (u > 0)
308
24.1k
  {
309
24.1k
    negate = 0;
310
24.1k
    ASSERT (v < 0);
311
24.1k
    v = -v;
312
24.1k
  }
313
6.32k
      else
314
6.32k
  {
315
6.32k
    negate = 1;
316
6.32k
    ASSERT (v > 0);
317
6.32k
    u = -u;
318
6.32k
  }
319
320
30.5k
      uh = mpn_mul_1 (up, u1, un, u);
321
30.5k
      vh = mpn_addmul_1 (up, u0, un, v);
322
323
30.5k
      if ( (uh | vh) > 0)
324
3.91k
  {
325
3.91k
    uh += vh;
326
3.91k
    up[un++] = uh;
327
3.91k
    if (uh < vh)
328
0
      up[un++] = 1;
329
3.91k
  }
330
331
30.5k
      MPN_NORMALIZE_NOT_ZERO (up, un);
332
333
30.5k
      *usize = negate ? -un : un;
334
30.5k
      return 1;
335
30.8k
    }
336
31.8k
}