Coverage Report

Created: 2023-09-25 06:41

/src/openssl111/crypto/ec/ecp_nistp521.c
Line
Count
Source (jump to first uncovered line)
1
/*
2
 * Copyright 2011-2020 The OpenSSL Project Authors. All Rights Reserved.
3
 *
4
 * Licensed under the OpenSSL license (the "License").  You may not use
5
 * this file except in compliance with the License.  You can obtain a copy
6
 * in the file LICENSE in the source distribution or at
7
 * https://www.openssl.org/source/license.html
8
 */
9
10
/* Copyright 2011 Google Inc.
11
 *
12
 * Licensed under the Apache License, Version 2.0 (the "License");
13
 *
14
 * you may not use this file except in compliance with the License.
15
 * You may obtain a copy of the License at
16
 *
17
 *     http://www.apache.org/licenses/LICENSE-2.0
18
 *
19
 *  Unless required by applicable law or agreed to in writing, software
20
 *  distributed under the License is distributed on an "AS IS" BASIS,
21
 *  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
22
 *  See the License for the specific language governing permissions and
23
 *  limitations under the License.
24
 */
25
26
/*
27
 * A 64-bit implementation of the NIST P-521 elliptic curve point multiplication
28
 *
29
 * OpenSSL integration was taken from Emilia Kasper's work in ecp_nistp224.c.
30
 * Otherwise based on Emilia's P224 work, which was inspired by my curve25519
31
 * work which got its smarts from Daniel J. Bernstein's work on the same.
32
 */
33
34
#include <openssl/e_os2.h>
35
#ifdef OPENSSL_NO_EC_NISTP_64_GCC_128
36
NON_EMPTY_TRANSLATION_UNIT
37
#else
38
39
# include <string.h>
40
# include <openssl/err.h>
41
# include "ec_local.h"
42
43
# if defined(__SIZEOF_INT128__) && __SIZEOF_INT128__==16
44
  /* even with gcc, the typedef won't work for 32-bit platforms */
45
typedef __uint128_t uint128_t;  /* nonstandard; implemented by gcc on 64-bit
46
                                 * platforms */
47
# else
48
#  error "Your compiler doesn't appear to support 128-bit integer types"
49
# endif
50
51
typedef uint8_t u8;
52
typedef uint64_t u64;
53
54
/*
55
 * The underlying field. P521 operates over GF(2^521-1). We can serialise an
56
 * element of this field into 66 bytes where the most significant byte
57
 * contains only a single bit. We call this an felem_bytearray.
58
 */
59
60
typedef u8 felem_bytearray[66];
61
62
/*
63
 * These are the parameters of P521, taken from FIPS 186-3, section D.1.2.5.
64
 * These values are big-endian.
65
 */
66
static const felem_bytearray nistp521_curve_params[5] = {
67
    {0x01, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* p */
68
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
69
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
70
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
71
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
72
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
73
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
74
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
75
     0xff, 0xff},
76
    {0x01, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, /* a = -3 */
77
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
78
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
79
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
80
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
81
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
82
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
83
     0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
84
     0xff, 0xfc},
85
    {0x00, 0x51, 0x95, 0x3e, 0xb9, 0x61, 0x8e, 0x1c, /* b */
86
     0x9a, 0x1f, 0x92, 0x9a, 0x21, 0xa0, 0xb6, 0x85,
87
     0x40, 0xee, 0xa2, 0xda, 0x72, 0x5b, 0x99, 0xb3,
88
     0x15, 0xf3, 0xb8, 0xb4, 0x89, 0x91, 0x8e, 0xf1,
89
     0x09, 0xe1, 0x56, 0x19, 0x39, 0x51, 0xec, 0x7e,
90
     0x93, 0x7b, 0x16, 0x52, 0xc0, 0xbd, 0x3b, 0xb1,
91
     0xbf, 0x07, 0x35, 0x73, 0xdf, 0x88, 0x3d, 0x2c,
92
     0x34, 0xf1, 0xef, 0x45, 0x1f, 0xd4, 0x6b, 0x50,
93
     0x3f, 0x00},
94
    {0x00, 0xc6, 0x85, 0x8e, 0x06, 0xb7, 0x04, 0x04, /* x */
95
     0xe9, 0xcd, 0x9e, 0x3e, 0xcb, 0x66, 0x23, 0x95,
96
     0xb4, 0x42, 0x9c, 0x64, 0x81, 0x39, 0x05, 0x3f,
97
     0xb5, 0x21, 0xf8, 0x28, 0xaf, 0x60, 0x6b, 0x4d,
98
     0x3d, 0xba, 0xa1, 0x4b, 0x5e, 0x77, 0xef, 0xe7,
99
     0x59, 0x28, 0xfe, 0x1d, 0xc1, 0x27, 0xa2, 0xff,
100
     0xa8, 0xde, 0x33, 0x48, 0xb3, 0xc1, 0x85, 0x6a,
101
     0x42, 0x9b, 0xf9, 0x7e, 0x7e, 0x31, 0xc2, 0xe5,
102
     0xbd, 0x66},
103
    {0x01, 0x18, 0x39, 0x29, 0x6a, 0x78, 0x9a, 0x3b, /* y */
104
     0xc0, 0x04, 0x5c, 0x8a, 0x5f, 0xb4, 0x2c, 0x7d,
105
     0x1b, 0xd9, 0x98, 0xf5, 0x44, 0x49, 0x57, 0x9b,
106
     0x44, 0x68, 0x17, 0xaf, 0xbd, 0x17, 0x27, 0x3e,
107
     0x66, 0x2c, 0x97, 0xee, 0x72, 0x99, 0x5e, 0xf4,
108
     0x26, 0x40, 0xc5, 0x50, 0xb9, 0x01, 0x3f, 0xad,
109
     0x07, 0x61, 0x35, 0x3c, 0x70, 0x86, 0xa2, 0x72,
110
     0xc2, 0x40, 0x88, 0xbe, 0x94, 0x76, 0x9f, 0xd1,
111
     0x66, 0x50}
112
};
113
114
/*-
115
 * The representation of field elements.
116
 * ------------------------------------
117
 *
118
 * We represent field elements with nine values. These values are either 64 or
119
 * 128 bits and the field element represented is:
120
 *   v[0]*2^0 + v[1]*2^58 + v[2]*2^116 + ... + v[8]*2^464  (mod p)
121
 * Each of the nine values is called a 'limb'. Since the limbs are spaced only
122
 * 58 bits apart, but are greater than 58 bits in length, the most significant
123
 * bits of each limb overlap with the least significant bits of the next.
124
 *
125
 * A field element with 64-bit limbs is an 'felem'. One with 128-bit limbs is a
126
 * 'largefelem' */
127
128
14.8M
# define NLIMBS 9
129
130
typedef uint64_t limb;
131
typedef limb limb_aX __attribute((__aligned__(1)));
132
typedef limb felem[NLIMBS];
133
typedef uint128_t largefelem[NLIMBS];
134
135
static const limb bottom57bits = 0x1ffffffffffffff;
136
static const limb bottom58bits = 0x3ffffffffffffff;
137
138
/*
139
 * bin66_to_felem takes a little-endian byte array and converts it into felem
140
 * form. This assumes that the CPU is little-endian.
141
 */
142
static void bin66_to_felem(felem out, const u8 in[66])
143
999
{
144
999
    out[0] = (*((limb *) & in[0])) & bottom58bits;
145
999
    out[1] = (*((limb_aX *) & in[7]) >> 2) & bottom58bits;
146
999
    out[2] = (*((limb_aX *) & in[14]) >> 4) & bottom58bits;
147
999
    out[3] = (*((limb_aX *) & in[21]) >> 6) & bottom58bits;
148
999
    out[4] = (*((limb_aX *) & in[29])) & bottom58bits;
149
999
    out[5] = (*((limb_aX *) & in[36]) >> 2) & bottom58bits;
150
999
    out[6] = (*((limb_aX *) & in[43]) >> 4) & bottom58bits;
151
999
    out[7] = (*((limb_aX *) & in[50]) >> 6) & bottom58bits;
152
999
    out[8] = (*((limb_aX *) & in[58])) & bottom57bits;
153
999
}
154
155
/*
156
 * felem_to_bin66 takes an felem and serialises into a little endian, 66 byte
157
 * array. This assumes that the CPU is little-endian.
158
 */
159
static void felem_to_bin66(u8 out[66], const felem in)
160
1.49k
{
161
1.49k
    memset(out, 0, 66);
162
1.49k
    (*((limb *) & out[0])) = in[0];
163
1.49k
    (*((limb_aX *) & out[7])) |= in[1] << 2;
164
1.49k
    (*((limb_aX *) & out[14])) |= in[2] << 4;
165
1.49k
    (*((limb_aX *) & out[21])) |= in[3] << 6;
166
1.49k
    (*((limb_aX *) & out[29])) = in[4];
167
1.49k
    (*((limb_aX *) & out[36])) |= in[5] << 2;
168
1.49k
    (*((limb_aX *) & out[43])) |= in[6] << 4;
169
1.49k
    (*((limb_aX *) & out[50])) |= in[7] << 6;
170
1.49k
    (*((limb_aX *) & out[58])) = in[8];
171
1.49k
}
172
173
/* BN_to_felem converts an OpenSSL BIGNUM into an felem */
174
static int BN_to_felem(felem out, const BIGNUM *bn)
175
999
{
176
999
    felem_bytearray b_out;
177
999
    int num_bytes;
178
179
999
    if (BN_is_negative(bn)) {
180
0
        ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
181
0
        return 0;
182
0
    }
183
999
    num_bytes = BN_bn2lebinpad(bn, b_out, sizeof(b_out));
184
999
    if (num_bytes < 0) {
185
0
        ECerr(EC_F_BN_TO_FELEM, EC_R_BIGNUM_OUT_OF_RANGE);
186
0
        return 0;
187
0
    }
188
999
    bin66_to_felem(out, b_out);
189
999
    return 1;
190
999
}
191
192
/* felem_to_BN converts an felem into an OpenSSL BIGNUM */
193
static BIGNUM *felem_to_BN(BIGNUM *out, const felem in)
194
1.49k
{
195
1.49k
    felem_bytearray b_out;
196
1.49k
    felem_to_bin66(b_out, in);
197
1.49k
    return BN_lebin2bn(b_out, sizeof(b_out), out);
198
1.49k
}
199
200
/*-
201
 * Field operations
202
 * ----------------
203
 */
204
205
static void felem_one(felem out)
206
0
{
207
0
    out[0] = 1;
208
0
    out[1] = 0;
209
0
    out[2] = 0;
210
0
    out[3] = 0;
211
0
    out[4] = 0;
212
0
    out[5] = 0;
213
0
    out[6] = 0;
214
0
    out[7] = 0;
215
0
    out[8] = 0;
216
0
}
217
218
static void felem_assign(felem out, const felem in)
219
592k
{
220
592k
    out[0] = in[0];
221
592k
    out[1] = in[1];
222
592k
    out[2] = in[2];
223
592k
    out[3] = in[3];
224
592k
    out[4] = in[4];
225
592k
    out[5] = in[5];
226
592k
    out[6] = in[6];
227
592k
    out[7] = in[7];
228
592k
    out[8] = in[8];
229
592k
}
230
231
/* felem_sum64 sets out = out + in. */
232
static void felem_sum64(felem out, const felem in)
233
221k
{
234
221k
    out[0] += in[0];
235
221k
    out[1] += in[1];
236
221k
    out[2] += in[2];
237
221k
    out[3] += in[3];
238
221k
    out[4] += in[4];
239
221k
    out[5] += in[5];
240
221k
    out[6] += in[6];
241
221k
    out[7] += in[7];
242
221k
    out[8] += in[8];
243
221k
}
244
245
/* felem_scalar sets out = in * scalar */
246
static void felem_scalar(felem out, const felem in, limb scalar)
247
1.66M
{
248
1.66M
    out[0] = in[0] * scalar;
249
1.66M
    out[1] = in[1] * scalar;
250
1.66M
    out[2] = in[2] * scalar;
251
1.66M
    out[3] = in[3] * scalar;
252
1.66M
    out[4] = in[4] * scalar;
253
1.66M
    out[5] = in[5] * scalar;
254
1.66M
    out[6] = in[6] * scalar;
255
1.66M
    out[7] = in[7] * scalar;
256
1.66M
    out[8] = in[8] * scalar;
257
1.66M
}
258
259
/* felem_scalar64 sets out = out * scalar */
260
static void felem_scalar64(felem out, limb scalar)
261
297k
{
262
297k
    out[0] *= scalar;
263
297k
    out[1] *= scalar;
264
297k
    out[2] *= scalar;
265
297k
    out[3] *= scalar;
266
297k
    out[4] *= scalar;
267
297k
    out[5] *= scalar;
268
297k
    out[6] *= scalar;
269
297k
    out[7] *= scalar;
270
297k
    out[8] *= scalar;
271
297k
}
272
273
/* felem_scalar128 sets out = out * scalar */
274
static void felem_scalar128(largefelem out, limb scalar)
275
99.1k
{
276
99.1k
    out[0] *= scalar;
277
99.1k
    out[1] *= scalar;
278
99.1k
    out[2] *= scalar;
279
99.1k
    out[3] *= scalar;
280
99.1k
    out[4] *= scalar;
281
99.1k
    out[5] *= scalar;
282
99.1k
    out[6] *= scalar;
283
99.1k
    out[7] *= scalar;
284
99.1k
    out[8] *= scalar;
285
99.1k
}
286
287
/*-
288
 * felem_neg sets |out| to |-in|
289
 * On entry:
290
 *   in[i] < 2^59 + 2^14
291
 * On exit:
292
 *   out[i] < 2^62
293
 */
294
static void felem_neg(felem out, const felem in)
295
10.3k
{
296
    /* In order to prevent underflow, we subtract from 0 mod p. */
297
10.3k
    static const limb two62m3 = (((limb) 1) << 62) - (((limb) 1) << 5);
298
10.3k
    static const limb two62m2 = (((limb) 1) << 62) - (((limb) 1) << 4);
299
300
10.3k
    out[0] = two62m3 - in[0];
301
10.3k
    out[1] = two62m2 - in[1];
302
10.3k
    out[2] = two62m2 - in[2];
303
10.3k
    out[3] = two62m2 - in[3];
304
10.3k
    out[4] = two62m2 - in[4];
305
10.3k
    out[5] = two62m2 - in[5];
306
10.3k
    out[6] = two62m2 - in[6];
307
10.3k
    out[7] = two62m2 - in[7];
308
10.3k
    out[8] = two62m2 - in[8];
309
10.3k
}
310
311
/*-
312
 * felem_diff64 subtracts |in| from |out|
313
 * On entry:
314
 *   in[i] < 2^59 + 2^14
315
 * On exit:
316
 *   out[i] < out[i] + 2^62
317
 */
318
static void felem_diff64(felem out, const felem in)
319
169k
{
320
    /*
321
     * In order to prevent underflow, we add 0 mod p before subtracting.
322
     */
323
169k
    static const limb two62m3 = (((limb) 1) << 62) - (((limb) 1) << 5);
324
169k
    static const limb two62m2 = (((limb) 1) << 62) - (((limb) 1) << 4);
325
326
169k
    out[0] += two62m3 - in[0];
327
169k
    out[1] += two62m2 - in[1];
328
169k
    out[2] += two62m2 - in[2];
329
169k
    out[3] += two62m2 - in[3];
330
169k
    out[4] += two62m2 - in[4];
331
169k
    out[5] += two62m2 - in[5];
332
169k
    out[6] += two62m2 - in[6];
333
169k
    out[7] += two62m2 - in[7];
334
169k
    out[8] += two62m2 - in[8];
335
169k
}
336
337
/*-
338
 * felem_diff_128_64 subtracts |in| from |out|
339
 * On entry:
340
 *   in[i] < 2^62 + 2^17
341
 * On exit:
342
 *   out[i] < out[i] + 2^63
343
 */
344
static void felem_diff_128_64(largefelem out, const felem in)
345
278k
{
346
    /*
347
     * In order to prevent underflow, we add 64p mod p (which is equivalent
348
     * to 0 mod p) before subtracting. p is 2^521 - 1, i.e. in binary a 521
349
     * digit number with all bits set to 1. See "The representation of field
350
     * elements" comment above for a description of how limbs are used to
351
     * represent a number. 64p is represented with 8 limbs containing a number
352
     * with 58 bits set and one limb with a number with 57 bits set.
353
     */
354
278k
    static const limb two63m6 = (((limb) 1) << 63) - (((limb) 1) << 6);
355
278k
    static const limb two63m5 = (((limb) 1) << 63) - (((limb) 1) << 5);
356
357
278k
    out[0] += two63m6 - in[0];
358
278k
    out[1] += two63m5 - in[1];
359
278k
    out[2] += two63m5 - in[2];
360
278k
    out[3] += two63m5 - in[3];
361
278k
    out[4] += two63m5 - in[4];
362
278k
    out[5] += two63m5 - in[5];
363
278k
    out[6] += two63m5 - in[6];
364
278k
    out[7] += two63m5 - in[7];
365
278k
    out[8] += two63m5 - in[8];
366
278k
}
367
368
/*-
369
 * felem_diff_128_64 subtracts |in| from |out|
370
 * On entry:
371
 *   in[i] < 2^126
372
 * On exit:
373
 *   out[i] < out[i] + 2^127 - 2^69
374
 */
375
static void felem_diff128(largefelem out, const largefelem in)
376
99.1k
{
377
    /*
378
     * In order to prevent underflow, we add 0 mod p before subtracting.
379
     */
380
99.1k
    static const uint128_t two127m70 =
381
99.1k
        (((uint128_t) 1) << 127) - (((uint128_t) 1) << 70);
382
99.1k
    static const uint128_t two127m69 =
383
99.1k
        (((uint128_t) 1) << 127) - (((uint128_t) 1) << 69);
384
385
99.1k
    out[0] += (two127m70 - in[0]);
386
99.1k
    out[1] += (two127m69 - in[1]);
387
99.1k
    out[2] += (two127m69 - in[2]);
388
99.1k
    out[3] += (two127m69 - in[3]);
389
99.1k
    out[4] += (two127m69 - in[4]);
390
99.1k
    out[5] += (two127m69 - in[5]);
391
99.1k
    out[6] += (two127m69 - in[6]);
392
99.1k
    out[7] += (two127m69 - in[7]);
393
99.1k
    out[8] += (two127m69 - in[8]);
394
99.1k
}
395
396
/*-
397
 * felem_square sets |out| = |in|^2
398
 * On entry:
399
 *   in[i] < 2^62
400
 * On exit:
401
 *   out[i] < 17 * max(in[i]) * max(in[i])
402
 */
403
static void felem_square(largefelem out, const felem in)
404
582k
{
405
582k
    felem inx2, inx4;
406
582k
    felem_scalar(inx2, in, 2);
407
582k
    felem_scalar(inx4, in, 4);
408
409
    /*-
410
     * We have many cases were we want to do
411
     *   in[x] * in[y] +
412
     *   in[y] * in[x]
413
     * This is obviously just
414
     *   2 * in[x] * in[y]
415
     * However, rather than do the doubling on the 128 bit result, we
416
     * double one of the inputs to the multiplication by reading from
417
     * |inx2|
418
     */
419
420
582k
    out[0] = ((uint128_t) in[0]) * in[0];
421
582k
    out[1] = ((uint128_t) in[0]) * inx2[1];
422
582k
    out[2] = ((uint128_t) in[0]) * inx2[2] + ((uint128_t) in[1]) * in[1];
423
582k
    out[3] = ((uint128_t) in[0]) * inx2[3] + ((uint128_t) in[1]) * inx2[2];
424
582k
    out[4] = ((uint128_t) in[0]) * inx2[4] +
425
582k
             ((uint128_t) in[1]) * inx2[3] + ((uint128_t) in[2]) * in[2];
426
582k
    out[5] = ((uint128_t) in[0]) * inx2[5] +
427
582k
             ((uint128_t) in[1]) * inx2[4] + ((uint128_t) in[2]) * inx2[3];
428
582k
    out[6] = ((uint128_t) in[0]) * inx2[6] +
429
582k
             ((uint128_t) in[1]) * inx2[5] +
430
582k
             ((uint128_t) in[2]) * inx2[4] + ((uint128_t) in[3]) * in[3];
431
582k
    out[7] = ((uint128_t) in[0]) * inx2[7] +
432
582k
             ((uint128_t) in[1]) * inx2[6] +
433
582k
             ((uint128_t) in[2]) * inx2[5] + ((uint128_t) in[3]) * inx2[4];
434
582k
    out[8] = ((uint128_t) in[0]) * inx2[8] +
435
582k
             ((uint128_t) in[1]) * inx2[7] +
436
582k
             ((uint128_t) in[2]) * inx2[6] +
437
582k
             ((uint128_t) in[3]) * inx2[5] + ((uint128_t) in[4]) * in[4];
438
439
    /*
440
     * The remaining limbs fall above 2^521, with the first falling at 2^522.
441
     * They correspond to locations one bit up from the limbs produced above
442
     * so we would have to multiply by two to align them. Again, rather than
443
     * operate on the 128-bit result, we double one of the inputs to the
444
     * multiplication. If we want to double for both this reason, and the
445
     * reason above, then we end up multiplying by four.
446
     */
447
448
    /* 9 */
449
582k
    out[0] += ((uint128_t) in[1]) * inx4[8] +
450
582k
              ((uint128_t) in[2]) * inx4[7] +
451
582k
              ((uint128_t) in[3]) * inx4[6] + ((uint128_t) in[4]) * inx4[5];
452
453
    /* 10 */
454
582k
    out[1] += ((uint128_t) in[2]) * inx4[8] +
455
582k
              ((uint128_t) in[3]) * inx4[7] +
456
582k
              ((uint128_t) in[4]) * inx4[6] + ((uint128_t) in[5]) * inx2[5];
457
458
    /* 11 */
459
582k
    out[2] += ((uint128_t) in[3]) * inx4[8] +
460
582k
              ((uint128_t) in[4]) * inx4[7] + ((uint128_t) in[5]) * inx4[6];
461
462
    /* 12 */
463
582k
    out[3] += ((uint128_t) in[4]) * inx4[8] +
464
582k
              ((uint128_t) in[5]) * inx4[7] + ((uint128_t) in[6]) * inx2[6];
465
466
    /* 13 */
467
582k
    out[4] += ((uint128_t) in[5]) * inx4[8] + ((uint128_t) in[6]) * inx4[7];
468
469
    /* 14 */
470
582k
    out[5] += ((uint128_t) in[6]) * inx4[8] + ((uint128_t) in[7]) * inx2[7];
471
472
    /* 15 */
473
582k
    out[6] += ((uint128_t) in[7]) * inx4[8];
474
475
    /* 16 */
476
582k
    out[7] += ((uint128_t) in[8]) * inx2[8];
477
582k
}
478
479
/*-
480
 * felem_mul sets |out| = |in1| * |in2|
481
 * On entry:
482
 *   in1[i] < 2^64
483
 *   in2[i] < 2^63
484
 * On exit:
485
 *   out[i] < 17 * max(in1[i]) * max(in2[i])
486
 */
487
static void felem_mul(largefelem out, const felem in1, const felem in2)
488
478k
{
489
478k
    felem in2x2;
490
478k
    felem_scalar(in2x2, in2, 2);
491
492
478k
    out[0] = ((uint128_t) in1[0]) * in2[0];
493
494
478k
    out[1] = ((uint128_t) in1[0]) * in2[1] +
495
478k
             ((uint128_t) in1[1]) * in2[0];
496
497
478k
    out[2] = ((uint128_t) in1[0]) * in2[2] +
498
478k
             ((uint128_t) in1[1]) * in2[1] +
499
478k
             ((uint128_t) in1[2]) * in2[0];
500
501
478k
    out[3] = ((uint128_t) in1[0]) * in2[3] +
502
478k
             ((uint128_t) in1[1]) * in2[2] +
503
478k
             ((uint128_t) in1[2]) * in2[1] +
504
478k
             ((uint128_t) in1[3]) * in2[0];
505
506
478k
    out[4] = ((uint128_t) in1[0]) * in2[4] +
507
478k
             ((uint128_t) in1[1]) * in2[3] +
508
478k
             ((uint128_t) in1[2]) * in2[2] +
509
478k
             ((uint128_t) in1[3]) * in2[1] +
510
478k
             ((uint128_t) in1[4]) * in2[0];
511
512
478k
    out[5] = ((uint128_t) in1[0]) * in2[5] +
513
478k
             ((uint128_t) in1[1]) * in2[4] +
514
478k
             ((uint128_t) in1[2]) * in2[3] +
515
478k
             ((uint128_t) in1[3]) * in2[2] +
516
478k
             ((uint128_t) in1[4]) * in2[1] +
517
478k
             ((uint128_t) in1[5]) * in2[0];
518
519
478k
    out[6] = ((uint128_t) in1[0]) * in2[6] +
520
478k
             ((uint128_t) in1[1]) * in2[5] +
521
478k
             ((uint128_t) in1[2]) * in2[4] +
522
478k
             ((uint128_t) in1[3]) * in2[3] +
523
478k
             ((uint128_t) in1[4]) * in2[2] +
524
478k
             ((uint128_t) in1[5]) * in2[1] +
525
478k
             ((uint128_t) in1[6]) * in2[0];
526
527
478k
    out[7] = ((uint128_t) in1[0]) * in2[7] +
528
478k
             ((uint128_t) in1[1]) * in2[6] +
529
478k
             ((uint128_t) in1[2]) * in2[5] +
530
478k
             ((uint128_t) in1[3]) * in2[4] +
531
478k
             ((uint128_t) in1[4]) * in2[3] +
532
478k
             ((uint128_t) in1[5]) * in2[2] +
533
478k
             ((uint128_t) in1[6]) * in2[1] +
534
478k
             ((uint128_t) in1[7]) * in2[0];
535
536
478k
    out[8] = ((uint128_t) in1[0]) * in2[8] +
537
478k
             ((uint128_t) in1[1]) * in2[7] +
538
478k
             ((uint128_t) in1[2]) * in2[6] +
539
478k
             ((uint128_t) in1[3]) * in2[5] +
540
478k
             ((uint128_t) in1[4]) * in2[4] +
541
478k
             ((uint128_t) in1[5]) * in2[3] +
542
478k
             ((uint128_t) in1[6]) * in2[2] +
543
478k
             ((uint128_t) in1[7]) * in2[1] +
544
478k
             ((uint128_t) in1[8]) * in2[0];
545
546
    /* See comment in felem_square about the use of in2x2 here */
547
548
478k
    out[0] += ((uint128_t) in1[1]) * in2x2[8] +
549
478k
              ((uint128_t) in1[2]) * in2x2[7] +
550
478k
              ((uint128_t) in1[3]) * in2x2[6] +
551
478k
              ((uint128_t) in1[4]) * in2x2[5] +
552
478k
              ((uint128_t) in1[5]) * in2x2[4] +
553
478k
              ((uint128_t) in1[6]) * in2x2[3] +
554
478k
              ((uint128_t) in1[7]) * in2x2[2] +
555
478k
              ((uint128_t) in1[8]) * in2x2[1];
556
557
478k
    out[1] += ((uint128_t) in1[2]) * in2x2[8] +
558
478k
              ((uint128_t) in1[3]) * in2x2[7] +
559
478k
              ((uint128_t) in1[4]) * in2x2[6] +
560
478k
              ((uint128_t) in1[5]) * in2x2[5] +
561
478k
              ((uint128_t) in1[6]) * in2x2[4] +
562
478k
              ((uint128_t) in1[7]) * in2x2[3] +
563
478k
              ((uint128_t) in1[8]) * in2x2[2];
564
565
478k
    out[2] += ((uint128_t) in1[3]) * in2x2[8] +
566
478k
              ((uint128_t) in1[4]) * in2x2[7] +
567
478k
              ((uint128_t) in1[5]) * in2x2[6] +
568
478k
              ((uint128_t) in1[6]) * in2x2[5] +
569
478k
              ((uint128_t) in1[7]) * in2x2[4] +
570
478k
              ((uint128_t) in1[8]) * in2x2[3];
571
572
478k
    out[3] += ((uint128_t) in1[4]) * in2x2[8] +
573
478k
              ((uint128_t) in1[5]) * in2x2[7] +
574
478k
              ((uint128_t) in1[6]) * in2x2[6] +
575
478k
              ((uint128_t) in1[7]) * in2x2[5] +
576
478k
              ((uint128_t) in1[8]) * in2x2[4];
577
578
478k
    out[4] += ((uint128_t) in1[5]) * in2x2[8] +
579
478k
              ((uint128_t) in1[6]) * in2x2[7] +
580
478k
              ((uint128_t) in1[7]) * in2x2[6] +
581
478k
              ((uint128_t) in1[8]) * in2x2[5];
582
583
478k
    out[5] += ((uint128_t) in1[6]) * in2x2[8] +
584
478k
              ((uint128_t) in1[7]) * in2x2[7] +
585
478k
              ((uint128_t) in1[8]) * in2x2[6];
586
587
478k
    out[6] += ((uint128_t) in1[7]) * in2x2[8] +
588
478k
              ((uint128_t) in1[8]) * in2x2[7];
589
590
478k
    out[7] += ((uint128_t) in1[8]) * in2x2[8];
591
478k
}
592
593
static const limb bottom52bits = 0xfffffffffffff;
594
595
/*-
596
 * felem_reduce converts a largefelem to an felem.
597
 * On entry:
598
 *   in[i] < 2^128
599
 * On exit:
600
 *   out[i] < 2^59 + 2^14
601
 */
602
static void felem_reduce(felem out, const largefelem in)
603
962k
{
604
962k
    u64 overflow1, overflow2;
605
606
962k
    out[0] = ((limb) in[0]) & bottom58bits;
607
962k
    out[1] = ((limb) in[1]) & bottom58bits;
608
962k
    out[2] = ((limb) in[2]) & bottom58bits;
609
962k
    out[3] = ((limb) in[3]) & bottom58bits;
610
962k
    out[4] = ((limb) in[4]) & bottom58bits;
611
962k
    out[5] = ((limb) in[5]) & bottom58bits;
612
962k
    out[6] = ((limb) in[6]) & bottom58bits;
613
962k
    out[7] = ((limb) in[7]) & bottom58bits;
614
962k
    out[8] = ((limb) in[8]) & bottom58bits;
615
616
    /* out[i] < 2^58 */
617
618
962k
    out[1] += ((limb) in[0]) >> 58;
619
962k
    out[1] += (((limb) (in[0] >> 64)) & bottom52bits) << 6;
620
    /*-
621
     * out[1] < 2^58 + 2^6 + 2^58
622
     *        = 2^59 + 2^6
623
     */
624
962k
    out[2] += ((limb) (in[0] >> 64)) >> 52;
625
626
962k
    out[2] += ((limb) in[1]) >> 58;
627
962k
    out[2] += (((limb) (in[1] >> 64)) & bottom52bits) << 6;
628
962k
    out[3] += ((limb) (in[1] >> 64)) >> 52;
629
630
962k
    out[3] += ((limb) in[2]) >> 58;
631
962k
    out[3] += (((limb) (in[2] >> 64)) & bottom52bits) << 6;
632
962k
    out[4] += ((limb) (in[2] >> 64)) >> 52;
633
634
962k
    out[4] += ((limb) in[3]) >> 58;
635
962k
    out[4] += (((limb) (in[3] >> 64)) & bottom52bits) << 6;
636
962k
    out[5] += ((limb) (in[3] >> 64)) >> 52;
637
638
962k
    out[5] += ((limb) in[4]) >> 58;
639
962k
    out[5] += (((limb) (in[4] >> 64)) & bottom52bits) << 6;
640
962k
    out[6] += ((limb) (in[4] >> 64)) >> 52;
641
642
962k
    out[6] += ((limb) in[5]) >> 58;
643
962k
    out[6] += (((limb) (in[5] >> 64)) & bottom52bits) << 6;
644
962k
    out[7] += ((limb) (in[5] >> 64)) >> 52;
645
646
962k
    out[7] += ((limb) in[6]) >> 58;
647
962k
    out[7] += (((limb) (in[6] >> 64)) & bottom52bits) << 6;
648
962k
    out[8] += ((limb) (in[6] >> 64)) >> 52;
649
650
962k
    out[8] += ((limb) in[7]) >> 58;
651
962k
    out[8] += (((limb) (in[7] >> 64)) & bottom52bits) << 6;
652
    /*-
653
     * out[x > 1] < 2^58 + 2^6 + 2^58 + 2^12
654
     *            < 2^59 + 2^13
655
     */
656
962k
    overflow1 = ((limb) (in[7] >> 64)) >> 52;
657
658
962k
    overflow1 += ((limb) in[8]) >> 58;
659
962k
    overflow1 += (((limb) (in[8] >> 64)) & bottom52bits) << 6;
660
962k
    overflow2 = ((limb) (in[8] >> 64)) >> 52;
661
662
962k
    overflow1 <<= 1;            /* overflow1 < 2^13 + 2^7 + 2^59 */
663
962k
    overflow2 <<= 1;            /* overflow2 < 2^13 */
664
665
962k
    out[0] += overflow1;        /* out[0] < 2^60 */
666
962k
    out[1] += overflow2;        /* out[1] < 2^59 + 2^6 + 2^13 */
667
668
962k
    out[1] += out[0] >> 58;
669
962k
    out[0] &= bottom58bits;
670
    /*-
671
     * out[0] < 2^58
672
     * out[1] < 2^59 + 2^6 + 2^13 + 2^2
673
     *        < 2^59 + 2^14
674
     */
675
962k
}
676
677
static void felem_square_reduce(felem out, const felem in)
678
0
{
679
0
    largefelem tmp;
680
0
    felem_square(tmp, in);
681
0
    felem_reduce(out, tmp);
682
0
}
683
684
static void felem_mul_reduce(felem out, const felem in1, const felem in2)
685
0
{
686
0
    largefelem tmp;
687
0
    felem_mul(tmp, in1, in2);
688
0
    felem_reduce(out, tmp);
689
0
}
690
691
/*-
692
 * felem_inv calculates |out| = |in|^{-1}
693
 *
694
 * Based on Fermat's Little Theorem:
695
 *   a^p = a (mod p)
696
 *   a^{p-1} = 1 (mod p)
697
 *   a^{p-2} = a^{-1} (mod p)
698
 */
699
static void felem_inv(felem out, const felem in)
700
234
{
701
234
    felem ftmp, ftmp2, ftmp3, ftmp4;
702
234
    largefelem tmp;
703
234
    unsigned i;
704
705
234
    felem_square(tmp, in);
706
234
    felem_reduce(ftmp, tmp);    /* 2^1 */
707
234
    felem_mul(tmp, in, ftmp);
708
234
    felem_reduce(ftmp, tmp);    /* 2^2 - 2^0 */
709
234
    felem_assign(ftmp2, ftmp);
710
234
    felem_square(tmp, ftmp);
711
234
    felem_reduce(ftmp, tmp);    /* 2^3 - 2^1 */
712
234
    felem_mul(tmp, in, ftmp);
713
234
    felem_reduce(ftmp, tmp);    /* 2^3 - 2^0 */
714
234
    felem_square(tmp, ftmp);
715
234
    felem_reduce(ftmp, tmp);    /* 2^4 - 2^1 */
716
717
234
    felem_square(tmp, ftmp2);
718
234
    felem_reduce(ftmp3, tmp);   /* 2^3 - 2^1 */
719
234
    felem_square(tmp, ftmp3);
720
234
    felem_reduce(ftmp3, tmp);   /* 2^4 - 2^2 */
721
234
    felem_mul(tmp, ftmp3, ftmp2);
722
234
    felem_reduce(ftmp3, tmp);   /* 2^4 - 2^0 */
723
724
234
    felem_assign(ftmp2, ftmp3);
725
234
    felem_square(tmp, ftmp3);
726
234
    felem_reduce(ftmp3, tmp);   /* 2^5 - 2^1 */
727
234
    felem_square(tmp, ftmp3);
728
234
    felem_reduce(ftmp3, tmp);   /* 2^6 - 2^2 */
729
234
    felem_square(tmp, ftmp3);
730
234
    felem_reduce(ftmp3, tmp);   /* 2^7 - 2^3 */
731
234
    felem_square(tmp, ftmp3);
732
234
    felem_reduce(ftmp3, tmp);   /* 2^8 - 2^4 */
733
234
    felem_assign(ftmp4, ftmp3);
734
234
    felem_mul(tmp, ftmp3, ftmp);
735
234
    felem_reduce(ftmp4, tmp);   /* 2^8 - 2^1 */
736
234
    felem_square(tmp, ftmp4);
737
234
    felem_reduce(ftmp4, tmp);   /* 2^9 - 2^2 */
738
234
    felem_mul(tmp, ftmp3, ftmp2);
739
234
    felem_reduce(ftmp3, tmp);   /* 2^8 - 2^0 */
740
234
    felem_assign(ftmp2, ftmp3);
741
742
2.10k
    for (i = 0; i < 8; i++) {
743
1.87k
        felem_square(tmp, ftmp3);
744
1.87k
        felem_reduce(ftmp3, tmp); /* 2^16 - 2^8 */
745
1.87k
    }
746
234
    felem_mul(tmp, ftmp3, ftmp2);
747
234
    felem_reduce(ftmp3, tmp);   /* 2^16 - 2^0 */
748
234
    felem_assign(ftmp2, ftmp3);
749
750
3.97k
    for (i = 0; i < 16; i++) {
751
3.74k
        felem_square(tmp, ftmp3);
752
3.74k
        felem_reduce(ftmp3, tmp); /* 2^32 - 2^16 */
753
3.74k
    }
754
234
    felem_mul(tmp, ftmp3, ftmp2);
755
234
    felem_reduce(ftmp3, tmp);   /* 2^32 - 2^0 */
756
234
    felem_assign(ftmp2, ftmp3);
757
758
7.72k
    for (i = 0; i < 32; i++) {
759
7.48k
        felem_square(tmp, ftmp3);
760
7.48k
        felem_reduce(ftmp3, tmp); /* 2^64 - 2^32 */
761
7.48k
    }
762
234
    felem_mul(tmp, ftmp3, ftmp2);
763
234
    felem_reduce(ftmp3, tmp);   /* 2^64 - 2^0 */
764
234
    felem_assign(ftmp2, ftmp3);
765
766
15.2k
    for (i = 0; i < 64; i++) {
767
14.9k
        felem_square(tmp, ftmp3);
768
14.9k
        felem_reduce(ftmp3, tmp); /* 2^128 - 2^64 */
769
14.9k
    }
770
234
    felem_mul(tmp, ftmp3, ftmp2);
771
234
    felem_reduce(ftmp3, tmp);   /* 2^128 - 2^0 */
772
234
    felem_assign(ftmp2, ftmp3);
773
774
30.1k
    for (i = 0; i < 128; i++) {
775
29.9k
        felem_square(tmp, ftmp3);
776
29.9k
        felem_reduce(ftmp3, tmp); /* 2^256 - 2^128 */
777
29.9k
    }
778
234
    felem_mul(tmp, ftmp3, ftmp2);
779
234
    felem_reduce(ftmp3, tmp);   /* 2^256 - 2^0 */
780
234
    felem_assign(ftmp2, ftmp3);
781
782
60.1k
    for (i = 0; i < 256; i++) {
783
59.9k
        felem_square(tmp, ftmp3);
784
59.9k
        felem_reduce(ftmp3, tmp); /* 2^512 - 2^256 */
785
59.9k
    }
786
234
    felem_mul(tmp, ftmp3, ftmp2);
787
234
    felem_reduce(ftmp3, tmp);   /* 2^512 - 2^0 */
788
789
2.34k
    for (i = 0; i < 9; i++) {
790
2.10k
        felem_square(tmp, ftmp3);
791
2.10k
        felem_reduce(ftmp3, tmp); /* 2^521 - 2^9 */
792
2.10k
    }
793
234
    felem_mul(tmp, ftmp3, ftmp4);
794
234
    felem_reduce(ftmp3, tmp);   /* 2^512 - 2^2 */
795
234
    felem_mul(tmp, ftmp3, in);
796
234
    felem_reduce(out, tmp);     /* 2^512 - 3 */
797
234
}
798
799
/* This is 2^521-1, expressed as an felem */
800
static const felem kPrime = {
801
    0x03ffffffffffffff, 0x03ffffffffffffff, 0x03ffffffffffffff,
802
    0x03ffffffffffffff, 0x03ffffffffffffff, 0x03ffffffffffffff,
803
    0x03ffffffffffffff, 0x03ffffffffffffff, 0x01ffffffffffffff
804
};
805
806
/*-
807
 * felem_is_zero returns a limb with all bits set if |in| == 0 (mod p) and 0
808
 * otherwise.
809
 * On entry:
810
 *   in[i] < 2^59 + 2^14
811
 */
812
static limb felem_is_zero(const felem in)
813
115k
{
814
115k
    felem ftmp;
815
115k
    limb is_zero, is_p;
816
115k
    felem_assign(ftmp, in);
817
818
115k
    ftmp[0] += ftmp[8] >> 57;
819
115k
    ftmp[8] &= bottom57bits;
820
    /* ftmp[8] < 2^57 */
821
115k
    ftmp[1] += ftmp[0] >> 58;
822
115k
    ftmp[0] &= bottom58bits;
823
115k
    ftmp[2] += ftmp[1] >> 58;
824
115k
    ftmp[1] &= bottom58bits;
825
115k
    ftmp[3] += ftmp[2] >> 58;
826
115k
    ftmp[2] &= bottom58bits;
827
115k
    ftmp[4] += ftmp[3] >> 58;
828
115k
    ftmp[3] &= bottom58bits;
829
115k
    ftmp[5] += ftmp[4] >> 58;
830
115k
    ftmp[4] &= bottom58bits;
831
115k
    ftmp[6] += ftmp[5] >> 58;
832
115k
    ftmp[5] &= bottom58bits;
833
115k
    ftmp[7] += ftmp[6] >> 58;
834
115k
    ftmp[6] &= bottom58bits;
835
115k
    ftmp[8] += ftmp[7] >> 58;
836
115k
    ftmp[7] &= bottom58bits;
837
    /* ftmp[8] < 2^57 + 4 */
838
839
    /*
840
     * The ninth limb of 2*(2^521-1) is 0x03ffffffffffffff, which is greater
841
     * than our bound for ftmp[8]. Therefore we only have to check if the
842
     * zero is zero or 2^521-1.
843
     */
844
845
115k
    is_zero = 0;
846
115k
    is_zero |= ftmp[0];
847
115k
    is_zero |= ftmp[1];
848
115k
    is_zero |= ftmp[2];
849
115k
    is_zero |= ftmp[3];
850
115k
    is_zero |= ftmp[4];
851
115k
    is_zero |= ftmp[5];
852
115k
    is_zero |= ftmp[6];
853
115k
    is_zero |= ftmp[7];
854
115k
    is_zero |= ftmp[8];
855
856
115k
    is_zero--;
857
    /*
858
     * We know that ftmp[i] < 2^63, therefore the only way that the top bit
859
     * can be set is if is_zero was 0 before the decrement.
860
     */
861
115k
    is_zero = 0 - (is_zero >> 63);
862
863
115k
    is_p = ftmp[0] ^ kPrime[0];
864
115k
    is_p |= ftmp[1] ^ kPrime[1];
865
115k
    is_p |= ftmp[2] ^ kPrime[2];
866
115k
    is_p |= ftmp[3] ^ kPrime[3];
867
115k
    is_p |= ftmp[4] ^ kPrime[4];
868
115k
    is_p |= ftmp[5] ^ kPrime[5];
869
115k
    is_p |= ftmp[6] ^ kPrime[6];
870
115k
    is_p |= ftmp[7] ^ kPrime[7];
871
115k
    is_p |= ftmp[8] ^ kPrime[8];
872
873
115k
    is_p--;
874
115k
    is_p = 0 - (is_p >> 63);
875
876
115k
    is_zero |= is_p;
877
115k
    return is_zero;
878
115k
}
879
880
static int felem_is_zero_int(const void *in)
881
0
{
882
0
    return (int)(felem_is_zero(in) & ((limb) 1));
883
0
}
884
885
/*-
886
 * felem_contract converts |in| to its unique, minimal representation.
887
 * On entry:
888
 *   in[i] < 2^59 + 2^14
889
 */
890
static void felem_contract(felem out, const felem in)
891
1.17k
{
892
1.17k
    limb is_p, is_greater, sign;
893
1.17k
    static const limb two58 = ((limb) 1) << 58;
894
895
1.17k
    felem_assign(out, in);
896
897
1.17k
    out[0] += out[8] >> 57;
898
1.17k
    out[8] &= bottom57bits;
899
    /* out[8] < 2^57 */
900
1.17k
    out[1] += out[0] >> 58;
901
1.17k
    out[0] &= bottom58bits;
902
1.17k
    out[2] += out[1] >> 58;
903
1.17k
    out[1] &= bottom58bits;
904
1.17k
    out[3] += out[2] >> 58;
905
1.17k
    out[2] &= bottom58bits;
906
1.17k
    out[4] += out[3] >> 58;
907
1.17k
    out[3] &= bottom58bits;
908
1.17k
    out[5] += out[4] >> 58;
909
1.17k
    out[4] &= bottom58bits;
910
1.17k
    out[6] += out[5] >> 58;
911
1.17k
    out[5] &= bottom58bits;
912
1.17k
    out[7] += out[6] >> 58;
913
1.17k
    out[6] &= bottom58bits;
914
1.17k
    out[8] += out[7] >> 58;
915
1.17k
    out[7] &= bottom58bits;
916
    /* out[8] < 2^57 + 4 */
917
918
    /*
919
     * If the value is greater than 2^521-1 then we have to subtract 2^521-1
920
     * out. See the comments in felem_is_zero regarding why we don't test for
921
     * other multiples of the prime.
922
     */
923
924
    /*
925
     * First, if |out| is equal to 2^521-1, we subtract it out to get zero.
926
     */
927
928
1.17k
    is_p = out[0] ^ kPrime[0];
929
1.17k
    is_p |= out[1] ^ kPrime[1];
930
1.17k
    is_p |= out[2] ^ kPrime[2];
931
1.17k
    is_p |= out[3] ^ kPrime[3];
932
1.17k
    is_p |= out[4] ^ kPrime[4];
933
1.17k
    is_p |= out[5] ^ kPrime[5];
934
1.17k
    is_p |= out[6] ^ kPrime[6];
935
1.17k
    is_p |= out[7] ^ kPrime[7];
936
1.17k
    is_p |= out[8] ^ kPrime[8];
937
938
1.17k
    is_p--;
939
1.17k
    is_p &= is_p << 32;
940
1.17k
    is_p &= is_p << 16;
941
1.17k
    is_p &= is_p << 8;
942
1.17k
    is_p &= is_p << 4;
943
1.17k
    is_p &= is_p << 2;
944
1.17k
    is_p &= is_p << 1;
945
1.17k
    is_p = 0 - (is_p >> 63);
946
1.17k
    is_p = ~is_p;
947
948
    /* is_p is 0 iff |out| == 2^521-1 and all ones otherwise */
949
950
1.17k
    out[0] &= is_p;
951
1.17k
    out[1] &= is_p;
952
1.17k
    out[2] &= is_p;
953
1.17k
    out[3] &= is_p;
954
1.17k
    out[4] &= is_p;
955
1.17k
    out[5] &= is_p;
956
1.17k
    out[6] &= is_p;
957
1.17k
    out[7] &= is_p;
958
1.17k
    out[8] &= is_p;
959
960
    /*
961
     * In order to test that |out| >= 2^521-1 we need only test if out[8] >>
962
     * 57 is greater than zero as (2^521-1) + x >= 2^522
963
     */
964
1.17k
    is_greater = out[8] >> 57;
965
1.17k
    is_greater |= is_greater << 32;
966
1.17k
    is_greater |= is_greater << 16;
967
1.17k
    is_greater |= is_greater << 8;
968
1.17k
    is_greater |= is_greater << 4;
969
1.17k
    is_greater |= is_greater << 2;
970
1.17k
    is_greater |= is_greater << 1;
971
1.17k
    is_greater = 0 - (is_greater >> 63);
972
973
1.17k
    out[0] -= kPrime[0] & is_greater;
974
1.17k
    out[1] -= kPrime[1] & is_greater;
975
1.17k
    out[2] -= kPrime[2] & is_greater;
976
1.17k
    out[3] -= kPrime[3] & is_greater;
977
1.17k
    out[4] -= kPrime[4] & is_greater;
978
1.17k
    out[5] -= kPrime[5] & is_greater;
979
1.17k
    out[6] -= kPrime[6] & is_greater;
980
1.17k
    out[7] -= kPrime[7] & is_greater;
981
1.17k
    out[8] -= kPrime[8] & is_greater;
982
983
    /* Eliminate negative coefficients */
984
1.17k
    sign = -(out[0] >> 63);
985
1.17k
    out[0] += (two58 & sign);
986
1.17k
    out[1] -= (1 & sign);
987
1.17k
    sign = -(out[1] >> 63);
988
1.17k
    out[1] += (two58 & sign);
989
1.17k
    out[2] -= (1 & sign);
990
1.17k
    sign = -(out[2] >> 63);
991
1.17k
    out[2] += (two58 & sign);
992
1.17k
    out[3] -= (1 & sign);
993
1.17k
    sign = -(out[3] >> 63);
994
1.17k
    out[3] += (two58 & sign);
995
1.17k
    out[4] -= (1 & sign);
996
1.17k
    sign = -(out[4] >> 63);
997
1.17k
    out[4] += (two58 & sign);
998
1.17k
    out[5] -= (1 & sign);
999
1.17k
    sign = -(out[0] >> 63);
1000
1.17k
    out[5] += (two58 & sign);
1001
1.17k
    out[6] -= (1 & sign);
1002
1.17k
    sign = -(out[6] >> 63);
1003
1.17k
    out[6] += (two58 & sign);
1004
1.17k
    out[7] -= (1 & sign);
1005
1.17k
    sign = -(out[7] >> 63);
1006
1.17k
    out[7] += (two58 & sign);
1007
1.17k
    out[8] -= (1 & sign);
1008
1.17k
    sign = -(out[5] >> 63);
1009
1.17k
    out[5] += (two58 & sign);
1010
1.17k
    out[6] -= (1 & sign);
1011
1.17k
    sign = -(out[6] >> 63);
1012
1.17k
    out[6] += (two58 & sign);
1013
1.17k
    out[7] -= (1 & sign);
1014
1.17k
    sign = -(out[7] >> 63);
1015
1.17k
    out[7] += (two58 & sign);
1016
1.17k
    out[8] -= (1 & sign);
1017
1.17k
}
1018
1019
/*-
1020
 * Group operations
1021
 * ----------------
1022
 *
1023
 * Building on top of the field operations we have the operations on the
1024
 * elliptic curve group itself. Points on the curve are represented in Jacobian
1025
 * coordinates */
1026
1027
/*-
1028
 * point_double calculates 2*(x_in, y_in, z_in)
1029
 *
1030
 * The method is taken from:
1031
 *   http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#doubling-dbl-2001-b
1032
 *
1033
 * Outputs can equal corresponding inputs, i.e., x_out == x_in is allowed.
1034
 * while x_out == y_in is not (maybe this works, but it's not tested). */
1035
static void
1036
point_double(felem x_out, felem y_out, felem z_out,
1037
             const felem x_in, const felem y_in, const felem z_in)
1038
70.2k
{
1039
70.2k
    largefelem tmp, tmp2;
1040
70.2k
    felem delta, gamma, beta, alpha, ftmp, ftmp2;
1041
1042
70.2k
    felem_assign(ftmp, x_in);
1043
70.2k
    felem_assign(ftmp2, x_in);
1044
1045
    /* delta = z^2 */
1046
70.2k
    felem_square(tmp, z_in);
1047
70.2k
    felem_reduce(delta, tmp);   /* delta[i] < 2^59 + 2^14 */
1048
1049
    /* gamma = y^2 */
1050
70.2k
    felem_square(tmp, y_in);
1051
70.2k
    felem_reduce(gamma, tmp);   /* gamma[i] < 2^59 + 2^14 */
1052
1053
    /* beta = x*gamma */
1054
70.2k
    felem_mul(tmp, x_in, gamma);
1055
70.2k
    felem_reduce(beta, tmp);    /* beta[i] < 2^59 + 2^14 */
1056
1057
    /* alpha = 3*(x-delta)*(x+delta) */
1058
70.2k
    felem_diff64(ftmp, delta);
1059
    /* ftmp[i] < 2^61 */
1060
70.2k
    felem_sum64(ftmp2, delta);
1061
    /* ftmp2[i] < 2^60 + 2^15 */
1062
70.2k
    felem_scalar64(ftmp2, 3);
1063
    /* ftmp2[i] < 3*2^60 + 3*2^15 */
1064
70.2k
    felem_mul(tmp, ftmp, ftmp2);
1065
    /*-
1066
     * tmp[i] < 17(3*2^121 + 3*2^76)
1067
     *        = 61*2^121 + 61*2^76
1068
     *        < 64*2^121 + 64*2^76
1069
     *        = 2^127 + 2^82
1070
     *        < 2^128
1071
     */
1072
70.2k
    felem_reduce(alpha, tmp);
1073
1074
    /* x' = alpha^2 - 8*beta */
1075
70.2k
    felem_square(tmp, alpha);
1076
    /*
1077
     * tmp[i] < 17*2^120 < 2^125
1078
     */
1079
70.2k
    felem_assign(ftmp, beta);
1080
70.2k
    felem_scalar64(ftmp, 8);
1081
    /* ftmp[i] < 2^62 + 2^17 */
1082
70.2k
    felem_diff_128_64(tmp, ftmp);
1083
    /* tmp[i] < 2^125 + 2^63 + 2^62 + 2^17 */
1084
70.2k
    felem_reduce(x_out, tmp);
1085
1086
    /* z' = (y + z)^2 - gamma - delta */
1087
70.2k
    felem_sum64(delta, gamma);
1088
    /* delta[i] < 2^60 + 2^15 */
1089
70.2k
    felem_assign(ftmp, y_in);
1090
70.2k
    felem_sum64(ftmp, z_in);
1091
    /* ftmp[i] < 2^60 + 2^15 */
1092
70.2k
    felem_square(tmp, ftmp);
1093
    /*
1094
     * tmp[i] < 17(2^122) < 2^127
1095
     */
1096
70.2k
    felem_diff_128_64(tmp, delta);
1097
    /* tmp[i] < 2^127 + 2^63 */
1098
70.2k
    felem_reduce(z_out, tmp);
1099
1100
    /* y' = alpha*(4*beta - x') - 8*gamma^2 */
1101
70.2k
    felem_scalar64(beta, 4);
1102
    /* beta[i] < 2^61 + 2^16 */
1103
70.2k
    felem_diff64(beta, x_out);
1104
    /* beta[i] < 2^61 + 2^60 + 2^16 */
1105
70.2k
    felem_mul(tmp, alpha, beta);
1106
    /*-
1107
     * tmp[i] < 17*((2^59 + 2^14)(2^61 + 2^60 + 2^16))
1108
     *        = 17*(2^120 + 2^75 + 2^119 + 2^74 + 2^75 + 2^30)
1109
     *        = 17*(2^120 + 2^119 + 2^76 + 2^74 + 2^30)
1110
     *        < 2^128
1111
     */
1112
70.2k
    felem_square(tmp2, gamma);
1113
    /*-
1114
     * tmp2[i] < 17*(2^59 + 2^14)^2
1115
     *         = 17*(2^118 + 2^74 + 2^28)
1116
     */
1117
70.2k
    felem_scalar128(tmp2, 8);
1118
    /*-
1119
     * tmp2[i] < 8*17*(2^118 + 2^74 + 2^28)
1120
     *         = 2^125 + 2^121 + 2^81 + 2^77 + 2^35 + 2^31
1121
     *         < 2^126
1122
     */
1123
70.2k
    felem_diff128(tmp, tmp2);
1124
    /*-
1125
     * tmp[i] < 2^127 - 2^69 + 17(2^120 + 2^119 + 2^76 + 2^74 + 2^30)
1126
     *        = 2^127 + 2^124 + 2^122 + 2^120 + 2^118 + 2^80 + 2^78 + 2^76 +
1127
     *          2^74 + 2^69 + 2^34 + 2^30
1128
     *        < 2^128
1129
     */
1130
70.2k
    felem_reduce(y_out, tmp);
1131
70.2k
}
1132
1133
/* copy_conditional copies in to out iff mask is all ones. */
1134
static void copy_conditional(felem out, const felem in, limb mask)
1135
183k
{
1136
183k
    unsigned i;
1137
1.83M
    for (i = 0; i < NLIMBS; ++i) {
1138
1.65M
        const limb tmp = mask & (in[i] ^ out[i]);
1139
1.65M
        out[i] ^= tmp;
1140
1.65M
    }
1141
183k
}
1142
1143
/*-
1144
 * point_add calculates (x1, y1, z1) + (x2, y2, z2)
1145
 *
1146
 * The method is taken from
1147
 *   http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#addition-add-2007-bl,
1148
 * adapted for mixed addition (z2 = 1, or z2 = 0 for the point at infinity).
1149
 *
1150
 * This function includes a branch for checking whether the two input points
1151
 * are equal (while not equal to the point at infinity). See comment below
1152
 * on constant-time.
1153
 */
1154
static void point_add(felem x3, felem y3, felem z3,
1155
                      const felem x1, const felem y1, const felem z1,
1156
                      const int mixed, const felem x2, const felem y2,
1157
                      const felem z2)
1158
28.9k
{
1159
28.9k
    felem ftmp, ftmp2, ftmp3, ftmp4, ftmp5, ftmp6, x_out, y_out, z_out;
1160
28.9k
    largefelem tmp, tmp2;
1161
28.9k
    limb x_equal, y_equal, z1_is_zero, z2_is_zero;
1162
28.9k
    limb points_equal;
1163
1164
28.9k
    z1_is_zero = felem_is_zero(z1);
1165
28.9k
    z2_is_zero = felem_is_zero(z2);
1166
1167
    /* ftmp = z1z1 = z1**2 */
1168
28.9k
    felem_square(tmp, z1);
1169
28.9k
    felem_reduce(ftmp, tmp);
1170
1171
28.9k
    if (!mixed) {
1172
        /* ftmp2 = z2z2 = z2**2 */
1173
10.9k
        felem_square(tmp, z2);
1174
10.9k
        felem_reduce(ftmp2, tmp);
1175
1176
        /* u1 = ftmp3 = x1*z2z2 */
1177
10.9k
        felem_mul(tmp, x1, ftmp2);
1178
10.9k
        felem_reduce(ftmp3, tmp);
1179
1180
        /* ftmp5 = z1 + z2 */
1181
10.9k
        felem_assign(ftmp5, z1);
1182
10.9k
        felem_sum64(ftmp5, z2);
1183
        /* ftmp5[i] < 2^61 */
1184
1185
        /* ftmp5 = (z1 + z2)**2 - z1z1 - z2z2 = 2*z1z2 */
1186
10.9k
        felem_square(tmp, ftmp5);
1187
        /* tmp[i] < 17*2^122 */
1188
10.9k
        felem_diff_128_64(tmp, ftmp);
1189
        /* tmp[i] < 17*2^122 + 2^63 */
1190
10.9k
        felem_diff_128_64(tmp, ftmp2);
1191
        /* tmp[i] < 17*2^122 + 2^64 */
1192
10.9k
        felem_reduce(ftmp5, tmp);
1193
1194
        /* ftmp2 = z2 * z2z2 */
1195
10.9k
        felem_mul(tmp, ftmp2, z2);
1196
10.9k
        felem_reduce(ftmp2, tmp);
1197
1198
        /* s1 = ftmp6 = y1 * z2**3 */
1199
10.9k
        felem_mul(tmp, y1, ftmp2);
1200
10.9k
        felem_reduce(ftmp6, tmp);
1201
17.9k
    } else {
1202
        /*
1203
         * We'll assume z2 = 1 (special case z2 = 0 is handled later)
1204
         */
1205
1206
        /* u1 = ftmp3 = x1*z2z2 */
1207
17.9k
        felem_assign(ftmp3, x1);
1208
1209
        /* ftmp5 = 2*z1z2 */
1210
17.9k
        felem_scalar(ftmp5, z1, 2);
1211
1212
        /* s1 = ftmp6 = y1 * z2**3 */
1213
17.9k
        felem_assign(ftmp6, y1);
1214
17.9k
    }
1215
1216
    /* u2 = x2*z1z1 */
1217
28.9k
    felem_mul(tmp, x2, ftmp);
1218
    /* tmp[i] < 17*2^120 */
1219
1220
    /* h = ftmp4 = u2 - u1 */
1221
28.9k
    felem_diff_128_64(tmp, ftmp3);
1222
    /* tmp[i] < 17*2^120 + 2^63 */
1223
28.9k
    felem_reduce(ftmp4, tmp);
1224
1225
28.9k
    x_equal = felem_is_zero(ftmp4);
1226
1227
    /* z_out = ftmp5 * h */
1228
28.9k
    felem_mul(tmp, ftmp5, ftmp4);
1229
28.9k
    felem_reduce(z_out, tmp);
1230
1231
    /* ftmp = z1 * z1z1 */
1232
28.9k
    felem_mul(tmp, ftmp, z1);
1233
28.9k
    felem_reduce(ftmp, tmp);
1234
1235
    /* s2 = tmp = y2 * z1**3 */
1236
28.9k
    felem_mul(tmp, y2, ftmp);
1237
    /* tmp[i] < 17*2^120 */
1238
1239
    /* r = ftmp5 = (s2 - s1)*2 */
1240
28.9k
    felem_diff_128_64(tmp, ftmp6);
1241
    /* tmp[i] < 17*2^120 + 2^63 */
1242
28.9k
    felem_reduce(ftmp5, tmp);
1243
28.9k
    y_equal = felem_is_zero(ftmp5);
1244
28.9k
    felem_scalar64(ftmp5, 2);
1245
    /* ftmp5[i] < 2^61 */
1246
1247
    /*
1248
     * The formulae are incorrect if the points are equal, in affine coordinates
1249
     * (X_1, Y_1) == (X_2, Y_2), so we check for this and do doubling if this
1250
     * happens.
1251
     *
1252
     * We use bitwise operations to avoid potential side-channels introduced by
1253
     * the short-circuiting behaviour of boolean operators.
1254
     *
1255
     * The special case of either point being the point at infinity (z1 and/or
1256
     * z2 are zero), is handled separately later on in this function, so we
1257
     * avoid jumping to point_double here in those special cases.
1258
     *
1259
     * Notice the comment below on the implications of this branching for timing
1260
     * leaks and why it is considered practically irrelevant.
1261
     */
1262
28.9k
    points_equal = (x_equal & y_equal & (~z1_is_zero) & (~z2_is_zero));
1263
1264
28.9k
    if (points_equal) {
1265
        /*
1266
         * This is obviously not constant-time but it will almost-never happen
1267
         * for ECDH / ECDSA. The case where it can happen is during scalar-mult
1268
         * where the intermediate value gets very close to the group order.
1269
         * Since |ec_GFp_nistp_recode_scalar_bits| produces signed digits for
1270
         * the scalar, it's possible for the intermediate value to be a small
1271
         * negative multiple of the base point, and for the final signed digit
1272
         * to be the same value. We believe that this only occurs for the scalar
1273
         * 1fffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
1274
         * ffffffa51868783bf2f966b7fcc0148f709a5d03bb5c9b8899c47aebb6fb
1275
         * 71e913863f7, in that case the penultimate intermediate is -9G and
1276
         * the final digit is also -9G. Since this only happens for a single
1277
         * scalar, the timing leak is irrelevant. (Any attacker who wanted to
1278
         * check whether a secret scalar was that exact value, can already do
1279
         * so.)
1280
         */
1281
0
        point_double(x3, y3, z3, x1, y1, z1);
1282
0
        return;
1283
0
    }
1284
1285
    /* I = ftmp = (2h)**2 */
1286
28.9k
    felem_assign(ftmp, ftmp4);
1287
28.9k
    felem_scalar64(ftmp, 2);
1288
    /* ftmp[i] < 2^61 */
1289
28.9k
    felem_square(tmp, ftmp);
1290
    /* tmp[i] < 17*2^122 */
1291
28.9k
    felem_reduce(ftmp, tmp);
1292
1293
    /* J = ftmp2 = h * I */
1294
28.9k
    felem_mul(tmp, ftmp4, ftmp);
1295
28.9k
    felem_reduce(ftmp2, tmp);
1296
1297
    /* V = ftmp4 = U1 * I */
1298
28.9k
    felem_mul(tmp, ftmp3, ftmp);
1299
28.9k
    felem_reduce(ftmp4, tmp);
1300
1301
    /* x_out = r**2 - J - 2V */
1302
28.9k
    felem_square(tmp, ftmp5);
1303
    /* tmp[i] < 17*2^122 */
1304
28.9k
    felem_diff_128_64(tmp, ftmp2);
1305
    /* tmp[i] < 17*2^122 + 2^63 */
1306
28.9k
    felem_assign(ftmp3, ftmp4);
1307
28.9k
    felem_scalar64(ftmp4, 2);
1308
    /* ftmp4[i] < 2^61 */
1309
28.9k
    felem_diff_128_64(tmp, ftmp4);
1310
    /* tmp[i] < 17*2^122 + 2^64 */
1311
28.9k
    felem_reduce(x_out, tmp);
1312
1313
    /* y_out = r(V-x_out) - 2 * s1 * J */
1314
28.9k
    felem_diff64(ftmp3, x_out);
1315
    /*
1316
     * ftmp3[i] < 2^60 + 2^60 = 2^61
1317
     */
1318
28.9k
    felem_mul(tmp, ftmp5, ftmp3);
1319
    /* tmp[i] < 17*2^122 */
1320
28.9k
    felem_mul(tmp2, ftmp6, ftmp2);
1321
    /* tmp2[i] < 17*2^120 */
1322
28.9k
    felem_scalar128(tmp2, 2);
1323
    /* tmp2[i] < 17*2^121 */
1324
28.9k
    felem_diff128(tmp, tmp2);
1325
        /*-
1326
         * tmp[i] < 2^127 - 2^69 + 17*2^122
1327
         *        = 2^126 - 2^122 - 2^6 - 2^2 - 1
1328
         *        < 2^127
1329
         */
1330
28.9k
    felem_reduce(y_out, tmp);
1331
1332
28.9k
    copy_conditional(x_out, x2, z1_is_zero);
1333
28.9k
    copy_conditional(x_out, x1, z2_is_zero);
1334
28.9k
    copy_conditional(y_out, y2, z1_is_zero);
1335
28.9k
    copy_conditional(y_out, y1, z2_is_zero);
1336
28.9k
    copy_conditional(z_out, z2, z1_is_zero);
1337
28.9k
    copy_conditional(z_out, z1, z2_is_zero);
1338
28.9k
    felem_assign(x3, x_out);
1339
28.9k
    felem_assign(y3, y_out);
1340
28.9k
    felem_assign(z3, z_out);
1341
28.9k
}
1342
1343
/*-
1344
 * Base point pre computation
1345
 * --------------------------
1346
 *
1347
 * Two different sorts of precomputed tables are used in the following code.
1348
 * Each contain various points on the curve, where each point is three field
1349
 * elements (x, y, z).
1350
 *
1351
 * For the base point table, z is usually 1 (0 for the point at infinity).
1352
 * This table has 16 elements:
1353
 * index | bits    | point
1354
 * ------+---------+------------------------------
1355
 *     0 | 0 0 0 0 | 0G
1356
 *     1 | 0 0 0 1 | 1G
1357
 *     2 | 0 0 1 0 | 2^130G
1358
 *     3 | 0 0 1 1 | (2^130 + 1)G
1359
 *     4 | 0 1 0 0 | 2^260G
1360
 *     5 | 0 1 0 1 | (2^260 + 1)G
1361
 *     6 | 0 1 1 0 | (2^260 + 2^130)G
1362
 *     7 | 0 1 1 1 | (2^260 + 2^130 + 1)G
1363
 *     8 | 1 0 0 0 | 2^390G
1364
 *     9 | 1 0 0 1 | (2^390 + 1)G
1365
 *    10 | 1 0 1 0 | (2^390 + 2^130)G
1366
 *    11 | 1 0 1 1 | (2^390 + 2^130 + 1)G
1367
 *    12 | 1 1 0 0 | (2^390 + 2^260)G
1368
 *    13 | 1 1 0 1 | (2^390 + 2^260 + 1)G
1369
 *    14 | 1 1 1 0 | (2^390 + 2^260 + 2^130)G
1370
 *    15 | 1 1 1 1 | (2^390 + 2^260 + 2^130 + 1)G
1371
 *
1372
 * The reason for this is so that we can clock bits into four different
1373
 * locations when doing simple scalar multiplies against the base point.
1374
 *
1375
 * Tables for other points have table[i] = iG for i in 0 .. 16. */
1376
1377
/* gmul is the table of precomputed base points */
1378
static const felem gmul[16][3] = {
1379
{{0, 0, 0, 0, 0, 0, 0, 0, 0},
1380
 {0, 0, 0, 0, 0, 0, 0, 0, 0},
1381
 {0, 0, 0, 0, 0, 0, 0, 0, 0}},
1382
{{0x017e7e31c2e5bd66, 0x022cf0615a90a6fe, 0x00127a2ffa8de334,
1383
  0x01dfbf9d64a3f877, 0x006b4d3dbaa14b5e, 0x014fed487e0a2bd8,
1384
  0x015b4429c6481390, 0x03a73678fb2d988e, 0x00c6858e06b70404},
1385
 {0x00be94769fd16650, 0x031c21a89cb09022, 0x039013fad0761353,
1386
  0x02657bd099031542, 0x03273e662c97ee72, 0x01e6d11a05ebef45,
1387
  0x03d1bd998f544495, 0x03001172297ed0b1, 0x011839296a789a3b},
1388
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1389
{{0x0373faacbc875bae, 0x00f325023721c671, 0x00f666fd3dbde5ad,
1390
  0x01a6932363f88ea7, 0x01fc6d9e13f9c47b, 0x03bcbffc2bbf734e,
1391
  0x013ee3c3647f3a92, 0x029409fefe75d07d, 0x00ef9199963d85e5},
1392
 {0x011173743ad5b178, 0x02499c7c21bf7d46, 0x035beaeabb8b1a58,
1393
  0x00f989c4752ea0a3, 0x0101e1de48a9c1a3, 0x01a20076be28ba6c,
1394
  0x02f8052e5eb2de95, 0x01bfe8f82dea117c, 0x0160074d3c36ddb7},
1395
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1396
{{0x012f3fc373393b3b, 0x03d3d6172f1419fa, 0x02adc943c0b86873,
1397
  0x00d475584177952b, 0x012a4d1673750ee2, 0x00512517a0f13b0c,
1398
  0x02b184671a7b1734, 0x0315b84236f1a50a, 0x00a4afc472edbdb9},
1399
 {0x00152a7077f385c4, 0x03044007d8d1c2ee, 0x0065829d61d52b52,
1400
  0x00494ff6b6631d0d, 0x00a11d94d5f06bcf, 0x02d2f89474d9282e,
1401
  0x0241c5727c06eeb9, 0x0386928710fbdb9d, 0x01f883f727b0dfbe},
1402
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1403
{{0x019b0c3c9185544d, 0x006243a37c9d97db, 0x02ee3cbe030a2ad2,
1404
  0x00cfdd946bb51e0d, 0x0271c00932606b91, 0x03f817d1ec68c561,
1405
  0x03f37009806a369c, 0x03c1f30baf184fd5, 0x01091022d6d2f065},
1406
 {0x0292c583514c45ed, 0x0316fca51f9a286c, 0x00300af507c1489a,
1407
  0x0295f69008298cf1, 0x02c0ed8274943d7b, 0x016509b9b47a431e,
1408
  0x02bc9de9634868ce, 0x005b34929bffcb09, 0x000c1a0121681524},
1409
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1410
{{0x0286abc0292fb9f2, 0x02665eee9805b3f7, 0x01ed7455f17f26d6,
1411
  0x0346355b83175d13, 0x006284944cd0a097, 0x0191895bcdec5e51,
1412
  0x02e288370afda7d9, 0x03b22312bfefa67a, 0x01d104d3fc0613fe},
1413
 {0x0092421a12f7e47f, 0x0077a83fa373c501, 0x03bd25c5f696bd0d,
1414
  0x035c41e4d5459761, 0x01ca0d1742b24f53, 0x00aaab27863a509c,
1415
  0x018b6de47df73917, 0x025c0b771705cd01, 0x01fd51d566d760a7},
1416
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1417
{{0x01dd92ff6b0d1dbd, 0x039c5e2e8f8afa69, 0x0261ed13242c3b27,
1418
  0x0382c6e67026e6a0, 0x01d60b10be2089f9, 0x03c15f3dce86723f,
1419
  0x03c764a32d2a062d, 0x017307eac0fad056, 0x018207c0b96c5256},
1420
 {0x0196a16d60e13154, 0x03e6ce74c0267030, 0x00ddbf2b4e52a5aa,
1421
  0x012738241bbf31c8, 0x00ebe8dc04685a28, 0x024c2ad6d380d4a2,
1422
  0x035ee062a6e62d0e, 0x0029ed74af7d3a0f, 0x00eef32aec142ebd},
1423
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1424
{{0x00c31ec398993b39, 0x03a9f45bcda68253, 0x00ac733c24c70890,
1425
  0x00872b111401ff01, 0x01d178c23195eafb, 0x03bca2c816b87f74,
1426
  0x0261a9af46fbad7a, 0x0324b2a8dd3d28f9, 0x00918121d8f24e23},
1427
 {0x032bc8c1ca983cd7, 0x00d869dfb08fc8c6, 0x01693cb61fce1516,
1428
  0x012a5ea68f4e88a8, 0x010869cab88d7ae3, 0x009081ad277ceee1,
1429
  0x033a77166d064cdc, 0x03955235a1fb3a95, 0x01251a4a9b25b65e},
1430
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1431
{{0x00148a3a1b27f40b, 0x0123186df1b31fdc, 0x00026e7beaad34ce,
1432
  0x01db446ac1d3dbba, 0x0299c1a33437eaec, 0x024540610183cbb7,
1433
  0x0173bb0e9ce92e46, 0x02b937e43921214b, 0x01ab0436a9bf01b5},
1434
 {0x0383381640d46948, 0x008dacbf0e7f330f, 0x03602122bcc3f318,
1435
  0x01ee596b200620d6, 0x03bd0585fda430b3, 0x014aed77fd123a83,
1436
  0x005ace749e52f742, 0x0390fe041da2b842, 0x0189a8ceb3299242},
1437
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1438
{{0x012a19d6b3282473, 0x00c0915918b423ce, 0x023a954eb94405ae,
1439
  0x00529f692be26158, 0x0289fa1b6fa4b2aa, 0x0198ae4ceea346ef,
1440
  0x0047d8cdfbdedd49, 0x00cc8c8953f0f6b8, 0x001424abbff49203},
1441
 {0x0256732a1115a03a, 0x0351bc38665c6733, 0x03f7b950fb4a6447,
1442
  0x000afffa94c22155, 0x025763d0a4dab540, 0x000511e92d4fc283,
1443
  0x030a7e9eda0ee96c, 0x004c3cd93a28bf0a, 0x017edb3a8719217f},
1444
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1445
{{0x011de5675a88e673, 0x031d7d0f5e567fbe, 0x0016b2062c970ae5,
1446
  0x03f4a2be49d90aa7, 0x03cef0bd13822866, 0x03f0923dcf774a6c,
1447
  0x0284bebc4f322f72, 0x016ab2645302bb2c, 0x01793f95dace0e2a},
1448
 {0x010646e13527a28f, 0x01ca1babd59dc5e7, 0x01afedfd9a5595df,
1449
  0x01f15785212ea6b1, 0x0324e5d64f6ae3f4, 0x02d680f526d00645,
1450
  0x0127920fadf627a7, 0x03b383f75df4f684, 0x0089e0057e783b0a},
1451
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1452
{{0x00f334b9eb3c26c6, 0x0298fdaa98568dce, 0x01c2d24843a82292,
1453
  0x020bcb24fa1b0711, 0x02cbdb3d2b1875e6, 0x0014907598f89422,
1454
  0x03abe3aa43b26664, 0x02cbf47f720bc168, 0x0133b5e73014b79b},
1455
 {0x034aab5dab05779d, 0x00cdc5d71fee9abb, 0x0399f16bd4bd9d30,
1456
  0x03582fa592d82647, 0x02be1cdfb775b0e9, 0x0034f7cea32e94cb,
1457
  0x0335a7f08f56f286, 0x03b707e9565d1c8b, 0x0015c946ea5b614f},
1458
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1459
{{0x024676f6cff72255, 0x00d14625cac96378, 0x00532b6008bc3767,
1460
  0x01fc16721b985322, 0x023355ea1b091668, 0x029de7afdc0317c3,
1461
  0x02fc8a7ca2da037c, 0x02de1217d74a6f30, 0x013f7173175b73bf},
1462
 {0x0344913f441490b5, 0x0200f9e272b61eca, 0x0258a246b1dd55d2,
1463
  0x03753db9ea496f36, 0x025e02937a09c5ef, 0x030cbd3d14012692,
1464
  0x01793a67e70dc72a, 0x03ec1d37048a662e, 0x006550f700c32a8d},
1465
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1466
{{0x00d3f48a347eba27, 0x008e636649b61bd8, 0x00d3b93716778fb3,
1467
  0x004d1915757bd209, 0x019d5311a3da44e0, 0x016d1afcbbe6aade,
1468
  0x0241bf5f73265616, 0x0384672e5d50d39b, 0x005009fee522b684},
1469
 {0x029b4fab064435fe, 0x018868ee095bbb07, 0x01ea3d6936cc92b8,
1470
  0x000608b00f78a2f3, 0x02db911073d1c20f, 0x018205938470100a,
1471
  0x01f1e4964cbe6ff2, 0x021a19a29eed4663, 0x01414485f42afa81},
1472
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1473
{{0x01612b3a17f63e34, 0x03813992885428e6, 0x022b3c215b5a9608,
1474
  0x029b4057e19f2fcb, 0x0384059a587af7e6, 0x02d6400ace6fe610,
1475
  0x029354d896e8e331, 0x00c047ee6dfba65e, 0x0037720542e9d49d},
1476
 {0x02ce9eed7c5e9278, 0x0374ed703e79643b, 0x01316c54c4072006,
1477
  0x005aaa09054b2ee8, 0x002824000c840d57, 0x03d4eba24771ed86,
1478
  0x0189c50aabc3bdae, 0x0338c01541e15510, 0x00466d56e38eed42},
1479
 {1, 0, 0, 0, 0, 0, 0, 0, 0}},
1480
{{0x007efd8330ad8bd6, 0x02465ed48047710b, 0x0034c6606b215e0c,
1481
  0x016ae30c53cbf839, 0x01fa17bd37161216, 0x018ead4e61ce8ab9,
1482
  0x005482ed5f5dee46, 0x037543755bba1d7f, 0x005e5ac7e70a9d0f},
1483
 {0x0117e1bb2fdcb2a2, 0x03deea36249f40c4, 0x028d09b4a6246cb7,
1484
  0x03524b8855bcf756, 0x023d7d109d5ceb58, 0x0178e43e3223ef9c,
1485
  0x0154536a0c6e966a, 0x037964d1286ee9fe, 0x0199bcd90e125055},
1486
 {1, 0, 0, 0, 0, 0, 0, 0, 0}}
1487
};
1488
1489
/*
1490
 * select_point selects the |idx|th point from a precomputation table and
1491
 * copies it to out.
1492
 */
1493
 /* pre_comp below is of the size provided in |size| */
1494
static void select_point(const limb idx, unsigned int size,
1495
                         const felem pre_comp[][3], felem out[3])
1496
28.4k
{
1497
28.4k
    unsigned i, j;
1498
28.4k
    limb *outlimbs = &out[0][0];
1499
1500
28.4k
    memset(out, 0, sizeof(*out) * 3);
1501
1502
494k
    for (i = 0; i < size; i++) {
1503
465k
        const limb *inlimbs = &pre_comp[i][0][0];
1504
465k
        limb mask = i ^ idx;
1505
465k
        mask |= mask >> 4;
1506
465k
        mask |= mask >> 2;
1507
465k
        mask |= mask >> 1;
1508
465k
        mask &= 1;
1509
465k
        mask--;
1510
13.0M
        for (j = 0; j < NLIMBS * 3; j++)
1511
12.5M
            outlimbs[j] |= inlimbs[j] & mask;
1512
465k
    }
1513
28.4k
}
1514
1515
/* get_bit returns the |i|th bit in |in| */
1516
static char get_bit(const felem_bytearray in, int i)
1517
134k
{
1518
134k
    if (i < 0)
1519
99
        return 0;
1520
134k
    return (in[i >> 3] >> (i & 7)) & 1;
1521
134k
}
1522
1523
/*
1524
 * Interleaved point multiplication using precomputed point multiples: The
1525
 * small point multiples 0*P, 1*P, ..., 16*P are in pre_comp[], the scalars
1526
 * in scalars[]. If g_scalar is non-NULL, we also add this multiple of the
1527
 * generator, using certain (large) precomputed multiples in g_pre_comp.
1528
 * Output point (X, Y, Z) is stored in x_out, y_out, z_out
1529
 */
1530
static void batch_mul(felem x_out, felem y_out, felem z_out,
1531
                      const felem_bytearray scalars[],
1532
                      const unsigned num_points, const u8 *g_scalar,
1533
                      const int mixed, const felem pre_comp[][17][3],
1534
                      const felem g_pre_comp[16][3])
1535
237
{
1536
237
    int i, skip;
1537
237
    unsigned num, gen_mul = (g_scalar != NULL);
1538
237
    felem nq[3], tmp[4];
1539
237
    limb bits;
1540
237
    u8 sign, digit;
1541
1542
    /* set nq to the point at infinity */
1543
237
    memset(nq, 0, sizeof(nq));
1544
1545
    /*
1546
     * Loop over all scalars msb-to-lsb, interleaving additions of multiples
1547
     * of the generator (last quarter of rounds) and additions of other
1548
     * points multiples (every 5th round).
1549
     */
1550
237
    skip = 1;                   /* save two point operations in the first
1551
                                 * round */
1552
69.8k
    for (i = (num_points ? 520 : 130); i >= 0; --i) {
1553
        /* double */
1554
69.6k
        if (!skip)
1555
69.4k
            point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]);
1556
1557
        /* add multiples of the generator */
1558
69.6k
        if (gen_mul && (i <= 130)) {
1559
18.0k
            bits = get_bit(g_scalar, i + 390) << 3;
1560
18.0k
            if (i < 130) {
1561
17.9k
                bits |= get_bit(g_scalar, i + 260) << 2;
1562
17.9k
                bits |= get_bit(g_scalar, i + 130) << 1;
1563
17.9k
                bits |= get_bit(g_scalar, i);
1564
17.9k
            }
1565
            /* select the point to add, in constant time */
1566
18.0k
            select_point(bits, 16, g_pre_comp, tmp);
1567
18.0k
            if (!skip) {
1568
                /* The 1 argument below is for "mixed" */
1569
17.9k
                point_add(nq[0], nq[1], nq[2],
1570
17.9k
                          nq[0], nq[1], nq[2], 1, tmp[0], tmp[1], tmp[2]);
1571
17.9k
            } else {
1572
138
                memcpy(nq, tmp, 3 * sizeof(felem));
1573
138
                skip = 0;
1574
138
            }
1575
18.0k
        }
1576
1577
        /* do other additions every 5 doublings */
1578
69.6k
        if (num_points && (i % 5 == 0)) {
1579
            /* loop over all scalars */
1580
20.7k
            for (num = 0; num < num_points; ++num) {
1581
10.3k
                bits = get_bit(scalars[num], i + 4) << 5;
1582
10.3k
                bits |= get_bit(scalars[num], i + 3) << 4;
1583
10.3k
                bits |= get_bit(scalars[num], i + 2) << 3;
1584
10.3k
                bits |= get_bit(scalars[num], i + 1) << 2;
1585
10.3k
                bits |= get_bit(scalars[num], i) << 1;
1586
10.3k
                bits |= get_bit(scalars[num], i - 1);
1587
10.3k
                ec_GFp_nistp_recode_scalar_bits(&sign, &digit, bits);
1588
1589
                /*
1590
                 * select the point to add or subtract, in constant time
1591
                 */
1592
10.3k
                select_point(digit, 17, pre_comp[num], tmp);
1593
10.3k
                felem_neg(tmp[3], tmp[1]); /* (X, -Y, Z) is the negative
1594
                                            * point */
1595
10.3k
                copy_conditional(tmp[1], tmp[3], (-(limb) sign));
1596
1597
10.3k
                if (!skip) {
1598
10.2k
                    point_add(nq[0], nq[1], nq[2],
1599
10.2k
                              nq[0], nq[1], nq[2],
1600
10.2k
                              mixed, tmp[0], tmp[1], tmp[2]);
1601
10.2k
                } else {
1602
99
                    memcpy(nq, tmp, 3 * sizeof(felem));
1603
99
                    skip = 0;
1604
99
                }
1605
10.3k
            }
1606
10.3k
        }
1607
69.6k
    }
1608
237
    felem_assign(x_out, nq[0]);
1609
237
    felem_assign(y_out, nq[1]);
1610
237
    felem_assign(z_out, nq[2]);
1611
237
}
1612
1613
/* Precomputation for the group generator. */
1614
struct nistp521_pre_comp_st {
1615
    felem g_pre_comp[16][3];
1616
    CRYPTO_REF_COUNT references;
1617
    CRYPTO_RWLOCK *lock;
1618
};
1619
1620
const EC_METHOD *EC_GFp_nistp521_method(void)
1621
153
{
1622
153
    static const EC_METHOD ret = {
1623
153
        EC_FLAGS_DEFAULT_OCT,
1624
153
        NID_X9_62_prime_field,
1625
153
        ec_GFp_nistp521_group_init,
1626
153
        ec_GFp_simple_group_finish,
1627
153
        ec_GFp_simple_group_clear_finish,
1628
153
        ec_GFp_nist_group_copy,
1629
153
        ec_GFp_nistp521_group_set_curve,
1630
153
        ec_GFp_simple_group_get_curve,
1631
153
        ec_GFp_simple_group_get_degree,
1632
153
        ec_group_simple_order_bits,
1633
153
        ec_GFp_simple_group_check_discriminant,
1634
153
        ec_GFp_simple_point_init,
1635
153
        ec_GFp_simple_point_finish,
1636
153
        ec_GFp_simple_point_clear_finish,
1637
153
        ec_GFp_simple_point_copy,
1638
153
        ec_GFp_simple_point_set_to_infinity,
1639
153
        ec_GFp_simple_set_Jprojective_coordinates_GFp,
1640
153
        ec_GFp_simple_get_Jprojective_coordinates_GFp,
1641
153
        ec_GFp_simple_point_set_affine_coordinates,
1642
153
        ec_GFp_nistp521_point_get_affine_coordinates,
1643
153
        0 /* point_set_compressed_coordinates */ ,
1644
153
        0 /* point2oct */ ,
1645
153
        0 /* oct2point */ ,
1646
153
        ec_GFp_simple_add,
1647
153
        ec_GFp_simple_dbl,
1648
153
        ec_GFp_simple_invert,
1649
153
        ec_GFp_simple_is_at_infinity,
1650
153
        ec_GFp_simple_is_on_curve,
1651
153
        ec_GFp_simple_cmp,
1652
153
        ec_GFp_simple_make_affine,
1653
153
        ec_GFp_simple_points_make_affine,
1654
153
        ec_GFp_nistp521_points_mul,
1655
153
        ec_GFp_nistp521_precompute_mult,
1656
153
        ec_GFp_nistp521_have_precompute_mult,
1657
153
        ec_GFp_nist_field_mul,
1658
153
        ec_GFp_nist_field_sqr,
1659
153
        0 /* field_div */ ,
1660
153
        ec_GFp_simple_field_inv,
1661
153
        0 /* field_encode */ ,
1662
153
        0 /* field_decode */ ,
1663
153
        0,                      /* field_set_to_one */
1664
153
        ec_key_simple_priv2oct,
1665
153
        ec_key_simple_oct2priv,
1666
153
        0, /* set private */
1667
153
        ec_key_simple_generate_key,
1668
153
        ec_key_simple_check_key,
1669
153
        ec_key_simple_generate_public_key,
1670
153
        0, /* keycopy */
1671
153
        0, /* keyfinish */
1672
153
        ecdh_simple_compute_key,
1673
153
        0, /* field_inverse_mod_ord */
1674
153
        0, /* blind_coordinates */
1675
153
        0, /* ladder_pre */
1676
153
        0, /* ladder_step */
1677
153
        0  /* ladder_post */
1678
153
    };
1679
1680
153
    return &ret;
1681
153
}
1682
1683
/******************************************************************************/
1684
/*
1685
 * FUNCTIONS TO MANAGE PRECOMPUTATION
1686
 */
1687
1688
static NISTP521_PRE_COMP *nistp521_pre_comp_new(void)
1689
0
{
1690
0
    NISTP521_PRE_COMP *ret = OPENSSL_zalloc(sizeof(*ret));
1691
1692
0
    if (ret == NULL) {
1693
0
        ECerr(EC_F_NISTP521_PRE_COMP_NEW, ERR_R_MALLOC_FAILURE);
1694
0
        return ret;
1695
0
    }
1696
1697
0
    ret->references = 1;
1698
1699
0
    ret->lock = CRYPTO_THREAD_lock_new();
1700
0
    if (ret->lock == NULL) {
1701
0
        ECerr(EC_F_NISTP521_PRE_COMP_NEW, ERR_R_MALLOC_FAILURE);
1702
0
        OPENSSL_free(ret);
1703
0
        return NULL;
1704
0
    }
1705
0
    return ret;
1706
0
}
1707
1708
NISTP521_PRE_COMP *EC_nistp521_pre_comp_dup(NISTP521_PRE_COMP *p)
1709
0
{
1710
0
    int i;
1711
0
    if (p != NULL)
1712
0
        CRYPTO_UP_REF(&p->references, &i, p->lock);
1713
0
    return p;
1714
0
}
1715
1716
void EC_nistp521_pre_comp_free(NISTP521_PRE_COMP *p)
1717
0
{
1718
0
    int i;
1719
1720
0
    if (p == NULL)
1721
0
        return;
1722
1723
0
    CRYPTO_DOWN_REF(&p->references, &i, p->lock);
1724
0
    REF_PRINT_COUNT("EC_nistp521", x);
1725
0
    if (i > 0)
1726
0
        return;
1727
0
    REF_ASSERT_ISNT(i < 0);
1728
1729
0
    CRYPTO_THREAD_lock_free(p->lock);
1730
0
    OPENSSL_free(p);
1731
0
}
1732
1733
/******************************************************************************/
1734
/*
1735
 * OPENSSL EC_METHOD FUNCTIONS
1736
 */
1737
1738
int ec_GFp_nistp521_group_init(EC_GROUP *group)
1739
540
{
1740
540
    int ret;
1741
540
    ret = ec_GFp_simple_group_init(group);
1742
540
    group->a_is_minus3 = 1;
1743
540
    return ret;
1744
540
}
1745
1746
int ec_GFp_nistp521_group_set_curve(EC_GROUP *group, const BIGNUM *p,
1747
                                    const BIGNUM *a, const BIGNUM *b,
1748
                                    BN_CTX *ctx)
1749
153
{
1750
153
    int ret = 0;
1751
153
    BN_CTX *new_ctx = NULL;
1752
153
    BIGNUM *curve_p, *curve_a, *curve_b;
1753
1754
153
    if (ctx == NULL)
1755
0
        if ((ctx = new_ctx = BN_CTX_new()) == NULL)
1756
0
            return 0;
1757
153
    BN_CTX_start(ctx);
1758
153
    curve_p = BN_CTX_get(ctx);
1759
153
    curve_a = BN_CTX_get(ctx);
1760
153
    curve_b = BN_CTX_get(ctx);
1761
153
    if (curve_b == NULL)
1762
0
        goto err;
1763
153
    BN_bin2bn(nistp521_curve_params[0], sizeof(felem_bytearray), curve_p);
1764
153
    BN_bin2bn(nistp521_curve_params[1], sizeof(felem_bytearray), curve_a);
1765
153
    BN_bin2bn(nistp521_curve_params[2], sizeof(felem_bytearray), curve_b);
1766
153
    if ((BN_cmp(curve_p, p)) || (BN_cmp(curve_a, a)) || (BN_cmp(curve_b, b))) {
1767
0
        ECerr(EC_F_EC_GFP_NISTP521_GROUP_SET_CURVE,
1768
0
              EC_R_WRONG_CURVE_PARAMETERS);
1769
0
        goto err;
1770
0
    }
1771
153
    group->field_mod_func = BN_nist_mod_521;
1772
153
    ret = ec_GFp_simple_group_set_curve(group, p, a, b, ctx);
1773
153
 err:
1774
153
    BN_CTX_end(ctx);
1775
153
    BN_CTX_free(new_ctx);
1776
153
    return ret;
1777
153
}
1778
1779
/*
1780
 * Takes the Jacobian coordinates (X, Y, Z) of a point and returns (X', Y') =
1781
 * (X/Z^2, Y/Z^3)
1782
 */
1783
int ec_GFp_nistp521_point_get_affine_coordinates(const EC_GROUP *group,
1784
                                                 const EC_POINT *point,
1785
                                                 BIGNUM *x, BIGNUM *y,
1786
                                                 BN_CTX *ctx)
1787
234
{
1788
234
    felem z1, z2, x_in, y_in, x_out, y_out;
1789
234
    largefelem tmp;
1790
1791
234
    if (EC_POINT_is_at_infinity(group, point)) {
1792
0
        ECerr(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES,
1793
0
              EC_R_POINT_AT_INFINITY);
1794
0
        return 0;
1795
0
    }
1796
234
    if ((!BN_to_felem(x_in, point->X)) || (!BN_to_felem(y_in, point->Y)) ||
1797
234
        (!BN_to_felem(z1, point->Z)))
1798
0
        return 0;
1799
234
    felem_inv(z2, z1);
1800
234
    felem_square(tmp, z2);
1801
234
    felem_reduce(z1, tmp);
1802
234
    felem_mul(tmp, x_in, z1);
1803
234
    felem_reduce(x_in, tmp);
1804
234
    felem_contract(x_out, x_in);
1805
234
    if (x != NULL) {
1806
234
        if (!felem_to_BN(x, x_out)) {
1807
0
            ECerr(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES,
1808
0
                  ERR_R_BN_LIB);
1809
0
            return 0;
1810
0
        }
1811
234
    }
1812
234
    felem_mul(tmp, z1, z2);
1813
234
    felem_reduce(z1, tmp);
1814
234
    felem_mul(tmp, y_in, z1);
1815
234
    felem_reduce(y_in, tmp);
1816
234
    felem_contract(y_out, y_in);
1817
234
    if (y != NULL) {
1818
138
        if (!felem_to_BN(y, y_out)) {
1819
0
            ECerr(EC_F_EC_GFP_NISTP521_POINT_GET_AFFINE_COORDINATES,
1820
0
                  ERR_R_BN_LIB);
1821
0
            return 0;
1822
0
        }
1823
138
    }
1824
234
    return 1;
1825
234
}
1826
1827
/* points below is of size |num|, and tmp_felems is of size |num+1/ */
1828
static void make_points_affine(size_t num, felem points[][3],
1829
                               felem tmp_felems[])
1830
0
{
1831
    /*
1832
     * Runs in constant time, unless an input is the point at infinity (which
1833
     * normally shouldn't happen).
1834
     */
1835
0
    ec_GFp_nistp_points_make_affine_internal(num,
1836
0
                                             points,
1837
0
                                             sizeof(felem),
1838
0
                                             tmp_felems,
1839
0
                                             (void (*)(void *))felem_one,
1840
0
                                             felem_is_zero_int,
1841
0
                                             (void (*)(void *, const void *))
1842
0
                                             felem_assign,
1843
0
                                             (void (*)(void *, const void *))
1844
0
                                             felem_square_reduce, (void (*)
1845
0
                                                                   (void *,
1846
0
                                                                    const void
1847
0
                                                                    *,
1848
0
                                                                    const void
1849
0
                                                                    *))
1850
0
                                             felem_mul_reduce,
1851
0
                                             (void (*)(void *, const void *))
1852
0
                                             felem_inv,
1853
0
                                             (void (*)(void *, const void *))
1854
0
                                             felem_contract);
1855
0
}
1856
1857
/*
1858
 * Computes scalar*generator + \sum scalars[i]*points[i], ignoring NULL
1859
 * values Result is stored in r (r can equal one of the inputs).
1860
 */
1861
int ec_GFp_nistp521_points_mul(const EC_GROUP *group, EC_POINT *r,
1862
                               const BIGNUM *scalar, size_t num,
1863
                               const EC_POINT *points[],
1864
                               const BIGNUM *scalars[], BN_CTX *ctx)
1865
237
{
1866
237
    int ret = 0;
1867
237
    int j;
1868
237
    int mixed = 0;
1869
237
    BIGNUM *x, *y, *z, *tmp_scalar;
1870
237
    felem_bytearray g_secret;
1871
237
    felem_bytearray *secrets = NULL;
1872
237
    felem (*pre_comp)[17][3] = NULL;
1873
237
    felem *tmp_felems = NULL;
1874
237
    unsigned i;
1875
237
    int num_bytes;
1876
237
    int have_pre_comp = 0;
1877
237
    size_t num_points = num;
1878
237
    felem x_in, y_in, z_in, x_out, y_out, z_out;
1879
237
    NISTP521_PRE_COMP *pre = NULL;
1880
237
    felem(*g_pre_comp)[3] = NULL;
1881
237
    EC_POINT *generator = NULL;
1882
237
    const EC_POINT *p = NULL;
1883
237
    const BIGNUM *p_scalar = NULL;
1884
1885
237
    BN_CTX_start(ctx);
1886
237
    x = BN_CTX_get(ctx);
1887
237
    y = BN_CTX_get(ctx);
1888
237
    z = BN_CTX_get(ctx);
1889
237
    tmp_scalar = BN_CTX_get(ctx);
1890
237
    if (tmp_scalar == NULL)
1891
0
        goto err;
1892
1893
237
    if (scalar != NULL) {
1894
138
        pre = group->pre_comp.nistp521;
1895
138
        if (pre)
1896
            /* we have precomputation, try to use it */
1897
0
            g_pre_comp = &pre->g_pre_comp[0];
1898
138
        else
1899
            /* try to use the standard precomputation */
1900
138
            g_pre_comp = (felem(*)[3]) gmul;
1901
138
        generator = EC_POINT_new(group);
1902
138
        if (generator == NULL)
1903
0
            goto err;
1904
        /* get the generator from precomputation */
1905
138
        if (!felem_to_BN(x, g_pre_comp[1][0]) ||
1906
138
            !felem_to_BN(y, g_pre_comp[1][1]) ||
1907
138
            !felem_to_BN(z, g_pre_comp[1][2])) {
1908
0
            ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
1909
0
            goto err;
1910
0
        }
1911
138
        if (!EC_POINT_set_Jprojective_coordinates_GFp(group,
1912
138
                                                      generator, x, y, z,
1913
138
                                                      ctx))
1914
0
            goto err;
1915
138
        if (0 == EC_POINT_cmp(group, generator, group->generator, ctx))
1916
            /* precomputation matches generator */
1917
138
            have_pre_comp = 1;
1918
0
        else
1919
            /*
1920
             * we don't have valid precomputation: treat the generator as a
1921
             * random point
1922
             */
1923
0
            num_points++;
1924
138
    }
1925
1926
237
    if (num_points > 0) {
1927
99
        if (num_points >= 2) {
1928
            /*
1929
             * unless we precompute multiples for just one point, converting
1930
             * those into affine form is time well spent
1931
             */
1932
0
            mixed = 1;
1933
0
        }
1934
99
        secrets = OPENSSL_zalloc(sizeof(*secrets) * num_points);
1935
99
        pre_comp = OPENSSL_zalloc(sizeof(*pre_comp) * num_points);
1936
99
        if (mixed)
1937
0
            tmp_felems =
1938
0
                OPENSSL_malloc(sizeof(*tmp_felems) * (num_points * 17 + 1));
1939
99
        if ((secrets == NULL) || (pre_comp == NULL)
1940
99
            || (mixed && (tmp_felems == NULL))) {
1941
0
            ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_MALLOC_FAILURE);
1942
0
            goto err;
1943
0
        }
1944
1945
        /*
1946
         * we treat NULL scalars as 0, and NULL points as points at infinity,
1947
         * i.e., they contribute nothing to the linear combination
1948
         */
1949
198
        for (i = 0; i < num_points; ++i) {
1950
99
            if (i == num) {
1951
                /*
1952
                 * we didn't have a valid precomputation, so we pick the
1953
                 * generator
1954
                 */
1955
0
                p = EC_GROUP_get0_generator(group);
1956
0
                p_scalar = scalar;
1957
99
            } else {
1958
                /* the i^th point */
1959
99
                p = points[i];
1960
99
                p_scalar = scalars[i];
1961
99
            }
1962
99
            if ((p_scalar != NULL) && (p != NULL)) {
1963
                /* reduce scalar to 0 <= scalar < 2^521 */
1964
99
                if ((BN_num_bits(p_scalar) > 521)
1965
99
                    || (BN_is_negative(p_scalar))) {
1966
                    /*
1967
                     * this is an unusual input, and we don't guarantee
1968
                     * constant-timeness
1969
                     */
1970
0
                    if (!BN_nnmod(tmp_scalar, p_scalar, group->order, ctx)) {
1971
0
                        ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
1972
0
                        goto err;
1973
0
                    }
1974
0
                    num_bytes = BN_bn2lebinpad(tmp_scalar,
1975
0
                                               secrets[i], sizeof(secrets[i]));
1976
99
                } else {
1977
99
                    num_bytes = BN_bn2lebinpad(p_scalar,
1978
99
                                               secrets[i], sizeof(secrets[i]));
1979
99
                }
1980
99
                if (num_bytes < 0) {
1981
0
                    ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
1982
0
                    goto err;
1983
0
                }
1984
                /* precompute multiples */
1985
99
                if ((!BN_to_felem(x_out, p->X)) ||
1986
99
                    (!BN_to_felem(y_out, p->Y)) ||
1987
99
                    (!BN_to_felem(z_out, p->Z)))
1988
0
                    goto err;
1989
99
                memcpy(pre_comp[i][1][0], x_out, sizeof(felem));
1990
99
                memcpy(pre_comp[i][1][1], y_out, sizeof(felem));
1991
99
                memcpy(pre_comp[i][1][2], z_out, sizeof(felem));
1992
1.58k
                for (j = 2; j <= 16; ++j) {
1993
1.48k
                    if (j & 1) {
1994
693
                        point_add(pre_comp[i][j][0], pre_comp[i][j][1],
1995
693
                                  pre_comp[i][j][2], pre_comp[i][1][0],
1996
693
                                  pre_comp[i][1][1], pre_comp[i][1][2], 0,
1997
693
                                  pre_comp[i][j - 1][0],
1998
693
                                  pre_comp[i][j - 1][1],
1999
693
                                  pre_comp[i][j - 1][2]);
2000
792
                    } else {
2001
792
                        point_double(pre_comp[i][j][0], pre_comp[i][j][1],
2002
792
                                     pre_comp[i][j][2], pre_comp[i][j / 2][0],
2003
792
                                     pre_comp[i][j / 2][1],
2004
792
                                     pre_comp[i][j / 2][2]);
2005
792
                    }
2006
1.48k
                }
2007
99
            }
2008
99
        }
2009
99
        if (mixed)
2010
0
            make_points_affine(num_points * 17, pre_comp[0], tmp_felems);
2011
99
    }
2012
2013
    /* the scalar for the generator */
2014
237
    if ((scalar != NULL) && (have_pre_comp)) {
2015
138
        memset(g_secret, 0, sizeof(g_secret));
2016
        /* reduce scalar to 0 <= scalar < 2^521 */
2017
138
        if ((BN_num_bits(scalar) > 521) || (BN_is_negative(scalar))) {
2018
            /*
2019
             * this is an unusual input, and we don't guarantee
2020
             * constant-timeness
2021
             */
2022
0
            if (!BN_nnmod(tmp_scalar, scalar, group->order, ctx)) {
2023
0
                ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
2024
0
                goto err;
2025
0
            }
2026
0
            num_bytes = BN_bn2lebinpad(tmp_scalar, g_secret, sizeof(g_secret));
2027
138
        } else {
2028
138
            num_bytes = BN_bn2lebinpad(scalar, g_secret, sizeof(g_secret));
2029
138
        }
2030
        /* do the multiplication with generator precomputation */
2031
138
        batch_mul(x_out, y_out, z_out,
2032
138
                  (const felem_bytearray(*))secrets, num_points,
2033
138
                  g_secret,
2034
138
                  mixed, (const felem(*)[17][3])pre_comp,
2035
138
                  (const felem(*)[3])g_pre_comp);
2036
138
    } else {
2037
        /* do the multiplication without generator precomputation */
2038
99
        batch_mul(x_out, y_out, z_out,
2039
99
                  (const felem_bytearray(*))secrets, num_points,
2040
99
                  NULL, mixed, (const felem(*)[17][3])pre_comp, NULL);
2041
99
    }
2042
    /* reduce the output to its unique minimal representation */
2043
237
    felem_contract(x_in, x_out);
2044
237
    felem_contract(y_in, y_out);
2045
237
    felem_contract(z_in, z_out);
2046
237
    if ((!felem_to_BN(x, x_in)) || (!felem_to_BN(y, y_in)) ||
2047
237
        (!felem_to_BN(z, z_in))) {
2048
0
        ECerr(EC_F_EC_GFP_NISTP521_POINTS_MUL, ERR_R_BN_LIB);
2049
0
        goto err;
2050
0
    }
2051
237
    ret = EC_POINT_set_Jprojective_coordinates_GFp(group, r, x, y, z, ctx);
2052
2053
237
 err:
2054
237
    BN_CTX_end(ctx);
2055
237
    EC_POINT_free(generator);
2056
237
    OPENSSL_free(secrets);
2057
237
    OPENSSL_free(pre_comp);
2058
237
    OPENSSL_free(tmp_felems);
2059
237
    return ret;
2060
237
}
2061
2062
int ec_GFp_nistp521_precompute_mult(EC_GROUP *group, BN_CTX *ctx)
2063
0
{
2064
0
    int ret = 0;
2065
0
    NISTP521_PRE_COMP *pre = NULL;
2066
0
    int i, j;
2067
0
    BN_CTX *new_ctx = NULL;
2068
0
    BIGNUM *x, *y;
2069
0
    EC_POINT *generator = NULL;
2070
0
    felem tmp_felems[16];
2071
2072
    /* throw away old precomputation */
2073
0
    EC_pre_comp_free(group);
2074
0
    if (ctx == NULL)
2075
0
        if ((ctx = new_ctx = BN_CTX_new()) == NULL)
2076
0
            return 0;
2077
0
    BN_CTX_start(ctx);
2078
0
    x = BN_CTX_get(ctx);
2079
0
    y = BN_CTX_get(ctx);
2080
0
    if (y == NULL)
2081
0
        goto err;
2082
    /* get the generator */
2083
0
    if (group->generator == NULL)
2084
0
        goto err;
2085
0
    generator = EC_POINT_new(group);
2086
0
    if (generator == NULL)
2087
0
        goto err;
2088
0
    BN_bin2bn(nistp521_curve_params[3], sizeof(felem_bytearray), x);
2089
0
    BN_bin2bn(nistp521_curve_params[4], sizeof(felem_bytearray), y);
2090
0
    if (!EC_POINT_set_affine_coordinates(group, generator, x, y, ctx))
2091
0
        goto err;
2092
0
    if ((pre = nistp521_pre_comp_new()) == NULL)
2093
0
        goto err;
2094
    /*
2095
     * if the generator is the standard one, use built-in precomputation
2096
     */
2097
0
    if (0 == EC_POINT_cmp(group, generator, group->generator, ctx)) {
2098
0
        memcpy(pre->g_pre_comp, gmul, sizeof(pre->g_pre_comp));
2099
0
        goto done;
2100
0
    }
2101
0
    if ((!BN_to_felem(pre->g_pre_comp[1][0], group->generator->X)) ||
2102
0
        (!BN_to_felem(pre->g_pre_comp[1][1], group->generator->Y)) ||
2103
0
        (!BN_to_felem(pre->g_pre_comp[1][2], group->generator->Z)))
2104
0
        goto err;
2105
    /* compute 2^130*G, 2^260*G, 2^390*G */
2106
0
    for (i = 1; i <= 4; i <<= 1) {
2107
0
        point_double(pre->g_pre_comp[2 * i][0], pre->g_pre_comp[2 * i][1],
2108
0
                     pre->g_pre_comp[2 * i][2], pre->g_pre_comp[i][0],
2109
0
                     pre->g_pre_comp[i][1], pre->g_pre_comp[i][2]);
2110
0
        for (j = 0; j < 129; ++j) {
2111
0
            point_double(pre->g_pre_comp[2 * i][0],
2112
0
                         pre->g_pre_comp[2 * i][1],
2113
0
                         pre->g_pre_comp[2 * i][2],
2114
0
                         pre->g_pre_comp[2 * i][0],
2115
0
                         pre->g_pre_comp[2 * i][1],
2116
0
                         pre->g_pre_comp[2 * i][2]);
2117
0
        }
2118
0
    }
2119
    /* g_pre_comp[0] is the point at infinity */
2120
0
    memset(pre->g_pre_comp[0], 0, sizeof(pre->g_pre_comp[0]));
2121
    /* the remaining multiples */
2122
    /* 2^130*G + 2^260*G */
2123
0
    point_add(pre->g_pre_comp[6][0], pre->g_pre_comp[6][1],
2124
0
              pre->g_pre_comp[6][2], pre->g_pre_comp[4][0],
2125
0
              pre->g_pre_comp[4][1], pre->g_pre_comp[4][2],
2126
0
              0, pre->g_pre_comp[2][0], pre->g_pre_comp[2][1],
2127
0
              pre->g_pre_comp[2][2]);
2128
    /* 2^130*G + 2^390*G */
2129
0
    point_add(pre->g_pre_comp[10][0], pre->g_pre_comp[10][1],
2130
0
              pre->g_pre_comp[10][2], pre->g_pre_comp[8][0],
2131
0
              pre->g_pre_comp[8][1], pre->g_pre_comp[8][2],
2132
0
              0, pre->g_pre_comp[2][0], pre->g_pre_comp[2][1],
2133
0
              pre->g_pre_comp[2][2]);
2134
    /* 2^260*G + 2^390*G */
2135
0
    point_add(pre->g_pre_comp[12][0], pre->g_pre_comp[12][1],
2136
0
              pre->g_pre_comp[12][2], pre->g_pre_comp[8][0],
2137
0
              pre->g_pre_comp[8][1], pre->g_pre_comp[8][2],
2138
0
              0, pre->g_pre_comp[4][0], pre->g_pre_comp[4][1],
2139
0
              pre->g_pre_comp[4][2]);
2140
    /* 2^130*G + 2^260*G + 2^390*G */
2141
0
    point_add(pre->g_pre_comp[14][0], pre->g_pre_comp[14][1],
2142
0
              pre->g_pre_comp[14][2], pre->g_pre_comp[12][0],
2143
0
              pre->g_pre_comp[12][1], pre->g_pre_comp[12][2],
2144
0
              0, pre->g_pre_comp[2][0], pre->g_pre_comp[2][1],
2145
0
              pre->g_pre_comp[2][2]);
2146
0
    for (i = 1; i < 8; ++i) {
2147
        /* odd multiples: add G */
2148
0
        point_add(pre->g_pre_comp[2 * i + 1][0],
2149
0
                  pre->g_pre_comp[2 * i + 1][1],
2150
0
                  pre->g_pre_comp[2 * i + 1][2], pre->g_pre_comp[2 * i][0],
2151
0
                  pre->g_pre_comp[2 * i][1], pre->g_pre_comp[2 * i][2], 0,
2152
0
                  pre->g_pre_comp[1][0], pre->g_pre_comp[1][1],
2153
0
                  pre->g_pre_comp[1][2]);
2154
0
    }
2155
0
    make_points_affine(15, &(pre->g_pre_comp[1]), tmp_felems);
2156
2157
0
 done:
2158
0
    SETPRECOMP(group, nistp521, pre);
2159
0
    ret = 1;
2160
0
    pre = NULL;
2161
0
 err:
2162
0
    BN_CTX_end(ctx);
2163
0
    EC_POINT_free(generator);
2164
0
    BN_CTX_free(new_ctx);
2165
0
    EC_nistp521_pre_comp_free(pre);
2166
0
    return ret;
2167
0
}
2168
2169
int ec_GFp_nistp521_have_precompute_mult(const EC_GROUP *group)
2170
0
{
2171
0
    return HAVEPRECOMP(group, nistp521);
2172
0
}
2173
2174
#endif