1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
|
/*
* Copyright 2017-2022 The OpenSSL Project Authors. All Rights Reserved.
* Copyright 2014 Cryptography Research, Inc.
*
* Licensed under the Apache License 2.0 (the "License"). You may not use
* this file except in compliance with the License. You can obtain a copy
* in the file LICENSE in the source distribution or at
* https://www.openssl.org/source/license.html
*
* Originally written by Mike Hamburg
*/
#include "internal/e_os.h"
#include <openssl/macros.h>
#include "internal/numbers.h"
#ifdef UINT128_MAX
/* We have support for 128 bit ints, so do nothing here */
NON_EMPTY_TRANSLATION_UNIT
#else
# include "../field.h"
void ossl_gf_mul(gf_s * RESTRICT cs, const gf as, const gf bs)
{
const uint32_t *a = as->limb, *b = bs->limb;
uint32_t *c = cs->limb;
uint64_t accum0 = 0, accum1 = 0, accum2 = 0;
uint32_t mask = (1 << 28) - 1;
uint32_t aa[8], bb[8];
int i, j;
for (i = 0; i < 8; i++) {
aa[i] = a[i] + a[i + 8];
bb[i] = b[i] + b[i + 8];
}
for (j = 0; j < 8; j++) {
accum2 = 0;
for (i = 0; i < j + 1; i++) {
accum2 += widemul(a[j - i], b[i]);
accum1 += widemul(aa[j - i], bb[i]);
accum0 += widemul(a[8 + j - i], b[8 + i]);
}
accum1 -= accum2;
accum0 += accum2;
accum2 = 0;
for (i = j + 1; i < 8; i++) {
accum0 -= widemul(a[8 + j - i], b[i]);
accum2 += widemul(aa[8 + j - i], bb[i]);
accum1 += widemul(a[16 + j - i], b[8 + i]);
}
accum1 += accum2;
accum0 += accum2;
c[j] = ((uint32_t)(accum0)) & mask;
c[j + 8] = ((uint32_t)(accum1)) & mask;
accum0 >>= 28;
accum1 >>= 28;
}
accum0 += accum1;
accum0 += c[8];
accum1 += c[0];
c[8] = ((uint32_t)(accum0)) & mask;
c[0] = ((uint32_t)(accum1)) & mask;
accum0 >>= 28;
accum1 >>= 28;
c[9] += ((uint32_t)(accum0));
c[1] += ((uint32_t)(accum1));
}
void ossl_gf_mulw_unsigned(gf_s * RESTRICT cs, const gf as, uint32_t b)
{
const uint32_t *a = as->limb;
uint32_t *c = cs->limb;
uint64_t accum0 = 0, accum8 = 0;
uint32_t mask = (1 << 28) - 1;
int i;
assert(b <= mask);
for (i = 0; i < 8; i++) {
accum0 += widemul(b, a[i]);
accum8 += widemul(b, a[i + 8]);
c[i] = accum0 & mask;
accum0 >>= 28;
c[i + 8] = accum8 & mask;
accum8 >>= 28;
}
accum0 += accum8 + c[8];
c[8] = ((uint32_t)accum0) & mask;
c[9] += (uint32_t)(accum0 >> 28);
accum8 += c[0];
c[0] = ((uint32_t)accum8) & mask;
c[1] += (uint32_t)(accum8 >> 28);
}
void ossl_gf_sqr(gf_s * RESTRICT cs, const gf as)
{
ossl_gf_mul(cs, as, as); /* Performs better with a dedicated square */
}
#endif
|